Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2007-11-16

00:00:05Lloreankugel: Just file a proper bug report. I'm going to dinner.
00:00:07 Part Llorean
00:00:23bertrikjpt9: I think my .18 OF charges up to 4.0V
00:00:35 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
00:00:35BoD[]ok it seems to work :)
00:00:48jpt9is there any reason rockbox can't charge it the last bit?
00:02:08bertrikjpt9: no not really, but I think hardly anyone uses rockbox for charging anyway
00:02:34jpt9it seems kind of silly.
00:03:11jpt9it probably gives some people a bad impression... "They can't even get charging to work?"
00:03:46Bagderjpt9: feel free to fix
00:04:40bertrikthere isn't any kind of charging API I think
00:05:00bertrikor is there?
00:05:31 Join barrywardell [0] (n=barrywar@
00:05:39bertrikoh I see some stuff in powermgmt.h
00:07:06amiconnpowermanagement is a bit of a mess atm
00:07:22 Quit jgarvey ("Leaving")
00:07:25amiconnBut we can't charge with >100mA before getting usb to work
00:07:31amiconnThat's the main problem
00:10:17bertrikjpt9: controlling the charger looks really easy, the charging itself is done by a hardware circuit. The software only needs to configure the charge parameters like voltage and current
00:10:30PaulPositionHmm.. Can't take screendumps anymore. [H10] 'system->debug->screendumps' and then plug in the usb to trigger, right?
00:10:43bertrikAnd we need to be *really* sure what those parameters are
00:11:29jpt9so you don't have it bursting into flames and all.
00:12:09bertrikperhaps not flames, but at least too high voltage is damaging to the battery (and too high charging current probably too)
00:12:16 Join qweru [0] (
00:13:19bertrikI've never seen more than 4.0V just after charging with the OF, but the charge controller allows charging up to a voltage of 4.25V
00:13:20 Join Workaphobia [0] (
00:20:05preglowif you do a makefile target that has no deps, it'll always be executed, yes? <-
00:20:35bertrikjpt9: maybe as an intermediate step we could implement an USB device much simpler than USB mass storage, and use this to advertise our need for 500 mA
00:20:51preglowi've got "rbspeexdec : rbspeexdec.c libspeex" on a line here, yet it consistently fails to call the libspeex target
00:21:16linuxstbCan you show your makefile?
00:21:17bertrikperhaps something like a virtual com port which is quite easy as it just uses two bulk pipes with no complicated protocol
00:22:19preglowlinuxstb: it's tools/Makefile, look near the bottom
00:22:27preglowlinuxstb: it does try to build rbspeexenc, though
00:23:26 Quit bertrik ("bedtime")
00:23:34linuxstbI expect it's conflicting with the directory called libspeex you create
00:23:49preglowright, so i should phony it
00:23:50 Join lee-qid_ [0] (
00:23:57 Quit ender` (" MCSE = Moron Claiming System Expertise, Must Call Someone Else, or Minesweeper Champion, Solitaire E")
00:24:58linuxstbOr maybe better - the dependency should be libspeex/libspeex.a - that's what you're making...
00:25:00preglowlinuxstb: mno, doesn't seem like it, no
00:25:02 Quit Isolinear ()
00:25:21preglowso, i should call the libspeex target that too, i suppose?
00:25:28*preglow dislikes makefiles
00:25:35linuxstbYes, that's what I mean
00:26:06preglowlinuxstb: that got it working in a hurry
00:26:10PaulPositionBefore I fill in a bug report, anyone would confirm/infirm that screendump is broken? (my iriver h10 does detect usb in/out, it just doesn't generate a bmp)
00:26:30preglowbut now, back to the old problem, nb_celp.c seems to want consts that are in speex.h, but does not include speex.h, yet all compiles of libspeex but this one works...
00:26:52safetydanpreglow: include paths maybe?
00:27:27preglowsafetydan: well, it would have told me if it didn't find a file it tried to include, no?
00:27:31 Part agm3nt
00:28:01safetydanpreglow: true
00:28:10PaulPosition(Ahh.... interrestingly, when I have debug-screendump enabled, it does NOT detect usb out or, at least, the statusbar doesn't reflect it)
00:29:26linuxstbpreglow: nb_celp.h includes modes.h which includes <speex/speex.h> - so I think it could be include paths...
00:29:46linuxstbMaybe a conflict with an installed version of speex.h?
00:30:17preglowvery probably
00:30:47PaulJamPaulPosition: there is a report in the h10 forums on misticriver about broken screendump functionality.
00:31:11jhMikeSpreglow: hurry the fux up...I want to commit that SPC codec =;p
00:31:37PaulPositionPauljam - Indeed. But misticriver folks aren't really into submitting bugs to the tracker. −− Yet I don't want to be trigger-happy with it either.
00:33:03preglowi'm getting there...
00:33:35jhMikeS:D :PpPpPp ... need more liquor (/me does)
00:33:46preglowit sPEWS warnings
00:33:57preglowi definitely need more liquor
00:34:02preglowbut have none, so it's more makefile instead
00:34:17hcsjhMikeS, preglow, what's up w/ SPC?
00:34:35PaulJamPaulPosition: i'm not sure what you mean. i just mentioned it because you asked for confirmation of this bug.
00:35:49jhMikeShcs: it's my dual core version but voice swapping is blocking it
00:36:01PaulPositionPaulJam - Oh, okay.. Sorry, I wanted to know if other devices would show the same symptoms.
00:36:11preglowlinuxstb: so, how to solve that, really? swap all <> with "" ?
00:36:30BoD[]see you all!
00:36:38 Part BoD[]
00:36:50PaulJamPaulPosition: on h300 with r15627 screendump works.
00:37:13linuxstbpreglow: I'm not sure, but I would hope there is a better way than that...
00:37:58preglowlinuxstb: i can't possibly think of one, afaik, /usr/include is first in the include path, and if speex/speex.h exists there, that will be used first
00:38:05PaulPositionPaulJam - Thanks.. So I'll fill a report only for H10.
00:38:27 Quit midkay ("Leaving")
00:38:31preglowlinuxstb: so either drop speex/ or turn <> to "", both equally bad, afaik
00:39:03PaulPosition(after I read the wiki about bug reports, to make sure I don't behave like an idiot)
00:40:04preglowlinuxstb: there doesn't seem to be many
00:40:22preglowat least not in .h files
00:40:27 Quit lee-qid (Read error: 110 (Connection timed out))
00:42:34 Quit freqmod_nx (Remote closed the connection)
00:42:48 Join freqmod_nx [0] (
00:44:03preglowany better suggestions? :/
00:44:12 Quit jpt9 (Read error: 110 (Connection timed out))
00:44:14 Join linuxstb_ [0] (
00:44:35 Quit linuxstb (Nick collision from services.)
00:44:35 Nick linuxstb_ is now known as linuxstb (
00:45:20preglowlibspeex/SOURCES is turning into a right mess
00:45:24linuxstbAre you sure this is the problem? Reading the manpage, -I should add the directory to the head of the search path, and system directories come afterwards...
00:45:53ryanakcaWhy is it that every song gets included into the Database 4 times? Any way to correct this? iPod Nano with Tuesday's daily build.
00:46:14preglowlinuxstb: really, now
00:46:25kugeldid someone of the devs look at fs#8122 ?
00:47:08linuxstbpreglow: You could confirm by (temporarily) removing the system copy
00:47:47preglowok, i've got a libspeex
00:48:08 Part toffe82
00:48:49preglowlinuxstb: i might have ommited some include path here, yeah
00:49:06JdGordonBABECOEC ? what the heck is a COEC?
00:49:42krazykitthat's what happens when you call the wrong escort service, i think.
00:49:45preglowlinuxstb: it doesn't look like the speex makefile specifies -I. or anything...
00:49:50preglowyes, yes it does
00:49:51JdGordonprefetch abort at that address.. thats a stack oberflow or something yeah?
00:50:11linuxstbpreglow: Yes, I'm looking at that now... Does "make V=1" work in tools/ ?
00:51:04*n1s thought it was supposed to be C0EDBABE
00:51:14JdGordonthats why it thought it was funny
00:51:16preglowlinuxstb: make tools V=1 does
00:51:21PaulJamkugel: is this even a bug? i thought it is theis way, so you can use the format strings from the original tagnavi in your custom one. if you want different format strings in your custom config you can just use different names for them.
00:51:48 Join Isolinear [0] (
00:52:02JdGordoncrap.. doesnt crash the sim :(
00:52:11kugelthis user made his own format strings, and they were not recognized
00:52:12safetydanJdGordon: sounds like some protection cookie that got scrambled
00:52:37PaulPositionCould svn be used to go backward, say, two weeks... in search of when a bug may have been introduced.. how?
00:52:38kugelI guess because the %include of tagnavi.config is after the format strings
00:53:01preglowlinuxstb: figured it out
00:53:12kugelPaulPosition: you can check out an older SVN
00:53:15preglowlinuxstb: _I_ forgot to include -I.
00:53:34 Quit HellDragon (Client Quit)
00:54:19PaulJamPaulPosition: use for example 'svn up -r12345' to go to a specific revision
00:54:20PaulPositiondoh, I just found out how... missed it when I first read the wiki page.
00:54:24PaulPositionthanks :)
00:57:23 Quit scorche|w ("CGI:IRC (Ping timeout)")
00:57:36 Join HellDragon [0] (i=jd@unaffiliated/helldragon)
00:57:51JdGordonanyone have any ideas about FS #8163? it only seems to be the german lang.. ive tried hai and russion which are both larger so im not sure its an overflow
00:58:18 Join Thundercloud [0] (
00:58:43PaulJamkugel: hmm, then i don't understand this bugreport. i use custom formatstrings in my tagnavi_custom.config and there they work.
00:58:55 Join jmworx [0] (
00:59:38kugelSo answer this report and tell the way you made it
00:59:49 Quit lee-qid_ ("aufwiederbyebientotsayonara")
01:00:01linuxstbpreglow: So is it all good now?
01:01:19preglowlinuxstb: no, i still lack some stuff, plus, this thing vOMITS warnings
01:01:41preglowi don't know how much i'm going to care, really, it won't show up on the build tables much
01:02:07preglowand wtf, why isn't ranlib run?
01:02:10preglowlooks like it tries
01:02:36preglowprobably isn't defined for host based builds
01:03:49linuxstbShouldn't it be the same as a sim build?
01:04:29linuxstbNo, it shouldn't (to answer my own question...)
01:05:05preglowand for some reason, big parts og nb_celp.c can't find speec_alloc and speex_free... and os_support.h is included just fine
01:05:15preglowforget me
01:05:39jmworxpreglow: what do you mean "can't find"?
01:05:46PaulJamkugel: i just did it the way the wiki decribes it. i don't use any tricks to make it work.
01:06:14kugelthe wiki says that the %include is before everything else (just comments before)
01:06:22kugelso the wiki is kinda wrong
01:06:30preglowjmworx: nevermind this, this is all my fault
01:06:37preglowjmworx: i'm trying to coax our svn libspeex to build natively
01:06:49jmworxwhat do you mean?
01:07:10preglowjmworx: i use speex_resampler in our own mini-encoder, and most people don't have that new a libspeex yet
01:07:23preglowjmworx: so i'm using the libspeex we already have in svn to build the encoder app
01:07:25PaulJamkugel: maybe it is just a leftover from before the formatstings were introduced. you shoud ask Slasheri, he is the Database expert.
01:08:04kugelI talked to the one who has this problem
01:08:08***Saving seen data "./dancer.seen"
01:08:39kugelhe made his format strings in the tagnavi_custom, and they aren't recognized
01:08:58kugelthat's all I can tell. I don't have this problem (but just since I don't have my own format strings)
01:09:00 Quit DerPapst ("So Long And Thanks For All The Fish!")
01:09:30 Quit n1s ()
01:09:38preglowlinuxstb: so, how to make that makefile run ranlib in a good way?
01:10:00 Join jpt9 [0] (
01:10:01PaulJamkugel: ask him to upload his file to the tracker, maybe this helps understanding the problem.
01:10:13preglowjmworx: and yes, right now i very much curse myself for having removed all that encoder stuff in weird ways :)
01:10:33linuxstbpreglow: Maybe just add AR=ar to the 'make -C' line ?
01:10:35preglowthere we go
01:10:57preglownow let's see if this bastard works
01:10:57 Quit jpt9 (Client Quit)
01:11:19jmworxpreglow: You're aware that speex 1.2beta3 is moving the resampler to libspeexdsp?
01:11:24preglowlinuxstb: it doesbn't call AR, it calls AR+ANLIB
01:11:36preglowjmworx: sure, but that doesn't matter much for us since we don't use your build system anyway
01:12:00preglowjmworx: all i need from libspeexdsp is resample.c, so i just include that in libspeex for this occasion
01:12:05preglowit won't be visible system-wide anyway
01:12:12jmworxthen why does it matter that "most people don't have that new a libspeex yet"?
01:12:32linuxstbpreglow: Are you sure - the Makefile echos 'AR+RANLIB' but then only does $(AR)
01:12:38jmworxpreglow: are you defining OUTSIDE_SPEEX?
01:12:52preglowjmworx: no, should it be necessary? it's built as part of speex
01:14:10preglowlinuxstb: that worked just fine
01:14:10 Join Llorean [0] (
01:14:21 Quit Zagor ("Client exiting")
01:14:23preglowlinuxstb: so all that's left of the annoyances now are the warnings, which might be 64 bit related
01:15:57jmworxpreglow: Well, OUTSIDE_SPEEX makes it use different symbols so that if someone eventually links with libspeexdsp as well, there's no symbol duplication. I guess in your case, you might not care about that...
01:15:58preglowlinuxstb: the resulting binary even works
01:16:22preglowjmworx: nah, we control our environment so well we don't have to worry about that
01:17:08preglowwell, cool, that was the last piece
01:17:17preglowwe should be ready to deploy some testing now
01:18:48PaulJamJdGordon: about FS #8163, it looks like in this screen the title is made to scroll artificially (even though it doesn't make sense in this context) and after it has scrolled for a while there are two square characters at the end of the title string (those that are displayed when the glyph is not available). so maybe the problem has to do with the title being made longer to scroll.
01:18:53 Join radinp [0] (
01:19:43 Join kugel_ [0] (
01:21:10JdGordonPaulJam: but it happens when exiting the screen...
01:24:28 Part Workaphobia
01:25:15 Quit kugel (Nick collision from services.)
01:25:18 Nick kugel_ is now known as kugel (
01:25:40PaulPositionShould I delete everything in my build folder before I 'make' a new (older) revision?
01:26:16 Quit kugel ("ChatZilla [Firefox]")
01:26:36PaulPositionor whatever.. :p
01:27:35mud_i would do a 'make clean' or yes delete everything if you're getting weird behavior...
01:27:58 Part pixelma
01:28:30PaulPositionmud_ - Hmm, gonna read up on 'make clean'.. It's just that RMing eveything by hand is, uh..., long. Thanks.
01:29:38linuxstbThat's all "make clean" does...
01:30:02PaulJamJdGordon: i have no sansa to test, i just thougt it might be related because these strange characters at the end of the padded title don't look like they belong there.
01:34:07PaulJamJdGordon: oh, my h300 just froze too when exiting this setting, but for me it happened with english language while i had no problem with german.
01:34:08 Join TheCollector [0] (
01:35:00 Quit weezerle ("...und tschĂĽss!")
01:36:24 Quit grndslm (Read error: 104 (Connection reset by peer))
01:37:21TheCollectorI need some help fixing my BL and firmware on my sansa e250. I've followed all the steps, but I can't seem to make any permanent progress.
01:38:24TheCollectorafter flashing the origional bl, it doesn't seem to be written
01:39:19LloreanThere shouldn't be anything like "flashing" in the e200 install.
01:40:06TheCollectorsorry, I'm just using that as a generic term for writing the bootloader
01:40:22TheCollectorthis is the sansa one, though
01:40:50mud_TheCollector: you're attempting to uninstall rockbox? or update the original firmware or something?
01:41:31TheCollectormud_: no, I accidentally nuked the first partition (the hidden one) while trying to get the microsd card to work
01:41:33linuxstbTheCollector: How are you "flashing" the bootloader? Just transferring it with e200tool doesn't write it.
01:41:45preglowjhMikeS: around?
01:42:08TheCollectorlinuxstb: what else needs to be done then?
01:42:11preglowjhMikeS: should we integrate this in your patch as well, then?
01:42:23linuxstbTheCollector: What the "unbrick" page says...
01:42:35jhMikeSpreglow: add eveything in that can be
01:43:03preglowjhMikeS: aight's, updated patch coming soon
01:43:03linuxstbTheCollector: Sorry, I've just read that page, and it's not clear...
01:43:10TheCollectorok, great
01:43:28TheCollectorI was hoping ther ewas just something I missed
01:43:49jhMikeSpreglow: i have a few tweaks of my own here
01:44:08 Quit radinp (Read error: 110 (Connection timed out))
01:45:08preglowjhMikeS: this SHOULD constitute a full kit, just testing should be needed
01:45:16preglowand stuff like agreeing on encoder options
01:45:21preglow-q 5 might be unnecessarily high
01:46:15TheCollectorlinuxstb: after I use e200tool to recover the bl, do I need to run it again using the 'write' command?
01:47:20jhMikeSpreglow: the quality sounds really high here
01:47:52preglowjhMikeS: hold on, found some bugs
01:48:39jhMikeSI'm going to adjust the stack size down and aim for say 90 ,95%? coldfire uses the most stack.
01:49:50preglowwhy not, stack size should be constant for a given speex mode, like i said
01:50:00linuxstbTheCollector: See this page for some instructions - (start at step 6)
01:51:08TheCollectorah, guess I was missing copying the bl to that 16mb partition
01:51:47linuxstbYes. I'm surprised the wiki doesn't say that.
01:51:49jhMikeS1824 bytes will be < 95% on coldfire
01:52:15 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
01:57:11 Join powr-toc [0] (
01:57:19preglowjhMikeS: apply and please try a voice build
01:57:36preglowjhMikeS: just use configure like before, then choose advance, then voice, then the rest should be self-explanatory
01:58:05jhMikeSok, gotta clean up. got the speex encoder or is that included?
01:58:25preglowthat patch does what the old one does, but adds speex encoder
01:58:31preglowi forgot the encoder itself
01:58:38preglowdo i need to svn add for that?
01:58:51preglowprobably, if using svn diff, i expect
01:59:12preglowhmm, i'll just smack a license onto it first
01:59:53linuxstbMinor point (it can be fixed later), but why does configure need to check for the presence of rbspeexenc? Couldn't it just be built when you do "make voice"? (similar to how the other tools/ things are built for a normal build)
02:00:19preglowjhMikeS: updated
02:00:35preglowlinuxstb: does configure check for that?
02:00:45linuxstbIf I'm reading your diff correctly...
02:00:49preglowright, i added it to the tools list
02:01:06linuxstbThe very first hunk in the diff...
02:01:10preglowwell, i guess it could... i'm all for not ever touching the build system again
02:01:39preglowlinuxstb: well, you'll be pleased to hear that's what happens
02:01:41jhMikeSpreglow: really love the voice that you made for h100...quite audible through everything
02:01:46preglowlinuxstb: that segment is run on "make voice"
02:01:49powr-tocI get a cannot load rockbox.ipod error
02:02:12linuxstbpowr-toc: That means you either forgot to extract to your ipod, or did it wrongly.
02:02:22*preglow tries to remember which voice that was
02:02:53jhMikeSpreglow: is it ready yet?
02:03:02preglowjhMikeS: the patch? yes
02:03:07 Quit barrywardell ()
02:03:29safetydanFor some reason having Speex for voice seems is the most interesting thing to happen to Rockbox lately. At least for me. :)
02:03:31jhMikeShow much disk space does generating the pcm take?
02:03:38*safetydan is probably just weird
02:03:39linuxstbpreglow: Then I think you can simply remove that check from configure.
02:03:42preglowjhMikeS: dunno, not much
02:03:51preglowjhMikeS: btw, i removed a couple of more files in speex here, might give a kb or so less space taken
02:04:06jhMikeSgot a link to something to use? I have no such tools
02:04:21preglowlinuxstb: well, it looks better than just puking when it's not there for some reason, doesn't it?
02:04:24jhMikeSand yeah you mentioned it earlier but I'm too lazy to go logging
02:04:26JdGordonPaulJam: nuts, ok
02:04:33preglowjhMikeS: windows? linux?
02:04:35powr-toclinuxstb: I've extracted it so I have a .rockbox in my ipod root dir
02:04:39 Quit Nico_P (Remote closed the connection)
02:04:50preglowjhMikeS: espeak, festival, flite
02:04:54jhMikeSI use VMWare so I guess that's linux :)
02:04:58linuxstbpreglow: No - it _is_ there, it's just not built until you type "make voice" (which is after you run configure...)
02:05:21preglowlinuxstb: afaik, i think it's built with all the other tools as it is now...
02:05:25ryanakcaWhy is it that every song gets included into the Database 4 times? Any way to correct this? iPod Nano with Tuesday's daily build.
02:05:33preglowlinuxstb: if you use voice or no
02:05:37linuxstbpreglow: Yes, but not until _after_ you run configure...
02:05:43 Join midkay [0] (n=midkay@rockbox/developer/midkay)
02:05:58preglowlinuxstb: oh, sure
02:06:15jhMikeSpreglow: do I have to remove the last building patch?
02:06:39preglowjhMikeS: yes
02:06:42preglowjhMikeS: this one includes all of it
02:07:00jhMikeSand are these synths and apt-get thing or download somewhere?
02:07:00preglowjhMikeS: they'll continue to do so if i update them, so do keep them hanging around for reversal
02:07:04linuxstbpowr-toc: And do you have a file called "rockbox.ipod" inside .rockbox?
02:07:05preglowjhMikeS: apt
02:07:10 Join mud__ [0] (
02:07:22 Nick mud__ is now known as mud-rb (
02:07:30jhMikeShope I have space for's tight on my image
02:08:04preglowthe gen itself should be tight enough
02:08:22preglowlinuxstb: so building rbspeexenc no matter if you want voice or not is a-ok?
02:08:33powr-toclinuxstb: yes
02:08:35linuxstbryanakca: Have you tried deleting the database files (*.tcd inside .rockbox) and rebuilding?
02:08:37safetydanryanakca: check that you don't have duplicates of your songs in hidden directories
02:09:57jhMikeSpreglow: which do you recommend of the voice synths
02:10:25linuxstbpreglow: No, I'm not saying that. I'm saying that if you download a clean svn tree, run tools/configure, and select voice, you'll get an error saying rbspeexenc doesn't exist. (IIUC)
02:10:43ryanakcasafetydan: ok, will do.
02:10:51ryanakcalinuxstb: nope, I'll try that too :)
02:10:56preglowlinuxstb: true
02:11:02preglowyes, yes it is
02:11:10preglowlinuxstb: i think i was assuming it'd be a binary or something when i wrote that
02:11:15powr-toclinuxstb: seems to work now... I think it might have been a problem with writes not having flushed to disk... I'm using an automount.
02:11:26preglowjhMikeS: i don't really know, i like espeak for its pronounciation, but it sounds metallic
02:11:32safetydanryanakca: look in your trash folders, so Recycle Bin on Windows or .trash for Macs
02:11:36preglowjhMikeS: it's also the smallest
02:11:52jhMikeSmight the H100 one...I like that darth vader thing
02:12:10preglowflite and festival try to sound natural
02:12:38linuxstbpowr-toc: Yes, "automount" isn't "autounmount"...
02:12:50preglowespeak sounds like some half drunk guy you found in a pub and had talk into your ring modulator
02:13:36powr-toclinuxstb: yeah... I forget sometimes :-\
02:13:43powr-tocthanks for your help anyway :)
02:13:46 Quit powr-toc (Remote closed the connection)
02:14:17ryanakcasafetydan: eh? why would I look in my trash folder? (Linux btw)
02:14:45ryanakcathey wouldn't be on my iPod, would they?
02:15:19*ryanakca thought they were in $HOME/.Trash ... hrm.
02:15:59mud-rbryanakca: it depends on the setup, but mostly they are on the same partition that you're deleting stuff from i believe, otherwise it would have to do copies all the time when you moved to trash...
02:16:13ryanakcamud-rb: ah
02:16:32*ryanakca checks
02:17:44safetydanryanakca: ah well the majority of ipod users tend not to be linux users. Bad assumption on my part.
02:18:04preglowjhMikeS: did the rb binary get any smaller, btw?
02:18:33ryanakcasafetydan: hehe :D
02:19:10jhMikeShaven't built it yet...was having problems with apt-get wanting a new bloody kernel
02:20:19jhMikeSI do a quicky e200 make bin
02:21:20jhMikeSjust got an error about inttypes.h in by
02:21:33TheCollectorlinuxstb: have you ever installed the new fw and bl and had it fail even though it says upgrade successful?
02:21:42linuxstbI don't own a Sansa
02:22:04 Nick fxb is now known as fxb__ (
02:22:22jhMikeSapt-get totally borked my build environment
02:22:36jhMikeSthat or it killed the path env variable
02:22:51TheCollectorit always goes back to manufacturing mode after it says upgrade finished
02:23:30preglowjhMikeS: \o/
02:24:18TheCollectorI wish - it just goes into recovery mode again
02:24:18preglowi usually start screaming and swearing
02:24:23preglowdoesn't solve anything, but that's what i usually do
02:25:01 Quit mud_ (Connection timed out)
02:25:41jhMikeSugh...better just finish the upgrade
02:27:26jhMikeSok, this things gonna pipe me through a whole list of upgrades...could be awhile
02:29:37ryanakcasafetydan: `du -a /media/IPOD/ | grep -i mp3 | gev -v kpod | sed -e 's@.*/media@/media@g' | sort` only lists two codec files in .rockbox/ and a couple image files for the themes... so I don't think it's any files hidden away.
02:30:22ryanakcaAs for the database, I've tried having only one song on the iPod, and recreating the database, but it still shows the song 4 times in the database.
02:31:22safetydanryanakca: do you mean it appears four times in the same list?
02:31:41safetydanalso, do you know if it is actually the same song? You might have some songs with duplicate tags.
02:31:58 Join tarehart [0] (
02:32:48 Part linuxstb
02:33:10ryanakcayes. say I have ABC and DFG on the iPod, the list goes 'ABC, ABC, ABC, ABC, DFG, DFG, DFG, DFG'. Because I removed all music from my iPod with amaroK and transfered one song to it. And, before, when I had a pile of songs on there, each and every song was listed 4 times in the database browser/list.
02:33:48tarehartI'm having trouble with a sansa e250R installation, can anybody help?
02:34:29advcomp2019tarehart, what are you having trouble with
02:34:35jhMikeSthey must be joking...they give me sources.list entries for etch then it can't stat it :\
02:34:53tarehartI'm using the automated installer and it can't find the player
02:35:25*preglow isn't too fond of deban
02:35:28advcomp2019you need to do the e200r steps
02:35:32 Join austriancoder [0] (n=austrian@rockbox/developer/austriancoder)
02:35:50tarehartI need to go manual?
02:35:57scorchetarehart: the manual says where you need to go
02:36:05scorcheit links to a page on the wiki
02:36:26ryanakcatarehart: have you mounted your sansa?
02:36:42tarehartbootloader patching, gotcha
02:37:01tarehartI'll take this as far as I can and come back the next time I make a blunder
02:38:14 Quit safetydan (Read error: 104 (Connection reset by peer))
02:38:29jhMikeSok, if you get that do the apt-get update thing _after_ adding the sources :\
02:40:10Mode"#rockbox +o scorche " by ChanServ (ChanServ@services.)
02:40:16Mode"#rockbox -b *!* " by scorche (i=Blah@rockbox/administrator/scorche)
02:40:22Mode"#rockbox -o scorche " by ChanServ (ChanServ@services.)
02:41:43 Quit Thundercloud (Remote closed the connection)
02:42:37 Join WalkGood [0] (i=WalkGood@unaffiliated/walkgood)
02:43:17 Join Gnu47 [0] (
02:44:26 Join Segadude [0] (
02:44:39 Join Febs [0] (
02:44:44Segadudecan rockbox play MIDIs
02:45:29mud-rbSegadude: yes
02:46:29mud-rbSegadude: same way as it plays everything else i believe
02:47:21thegeeka cool dude wrote a midi decoder/codec
02:47:38SegadudeHow do i get it?
02:47:59thegeekI think it comes with rockbox by default?
02:48:04PaulPositionDepending on the target, it does cough on files with more than 3 or 4 tracks in my experience..
02:48:26Segadudeyeah the one i;m trying to play has 33
02:48:43PaulPositionIt does come with rockbox by default, but you need to find a patchset (an instrument library..) The wiki page on PluginMidiPlayer has info and links to patchset..
02:49:02Segadudeok thanks
02:49:34PaulJamJdGordon: did you get my PM?
02:49:39 Part WalkGood
02:51:10preglowi don't make clean properly either
02:54:05tarehartok, I've got a new problem with installation
02:54:14jhMikeSthis is tottaly fucked up now...must be a way out of this circular crap
02:54:29preglowusing windows? :P
02:54:33tarehartthis pertains to sansa e250R bootloader patching
02:55:01jhMikeSusing apt-get...the dependencies are circular
02:55:10JdGordonPaulJam: soz, was afk.. i just saw it
02:55:14tarehartwhen I connect in manufacturer mode, windows vista isn't letting me install the driver I want
02:55:31 Join hoobop [0] (n=user@unaffiliated/hoobop)
02:55:41TheCollectorjhMikeS: I quit using debian because of that
02:55:46preglowjhMikeS: i've had some crazy shit like you've never seen before with apt-get
02:55:55TheCollectorjust traded that trouble for others when I switched to slack
02:56:16preglowit truly is capable of just turning itself into useless fluff
02:56:51jhMikeSah well, I'm not ready to redo everything right I'll have to figure my way through it I suppose
02:56:55 Quit mud-rb (Remote closed the connection)
02:57:31advcomp2019tarehart, those drivers are for xp not vista at least
02:57:54tarehartthat would explain it
02:58:12tarehartdo I have any recourse, I wonder
02:59:09jhMikeSpreglow: well, if you fixed the crazy shit, any pointers?
02:59:23preglowjhMikeS: a couple so severe i just had to reinstall
02:59:43advcomp2019tarehart, if you do not have a xp machine you might have to burn a linux livecd and do the linux steps
02:59:43preglowand no, i usually just break down whenever it happens
02:59:58tarehartthat would be exciting
03:00:05preglowjhMikeS: if i used some virtualization stuff, i'd just back up a known working image...
03:00:17tarehartwell, thanks for the help so far
03:01:02jhMikeSbeh...well I guess I could get a new image...copy the files over and restart
03:04:07preglowwhat about using windows?
03:04:22preglowsounded like you had a dev env running there anyway
03:04:47 Quit austriancoder (Remote closed the connection) it I think :)
03:07:33jhMikeSugh...maybe noty
03:08:09***Saving seen data "./dancer.seen"
03:09:58 Quit J3TC- (Read error: 110 (Connection timed out))
03:10:02preglowfew things annoy me more than having to fight with my operating system
03:15:42 Nick ord is now known as qwm (
03:17:10 Quit Segadude ("Quitting!")
03:20:29ryanakcato reboot on nano after building database, does that mean menu + select or play/pause followed by select?
03:20:40jhMikeSwhich frontend is usually installed with the VMWare image?
03:21:22scorcheas in WM?
03:21:50jhMikeSIt shows Dialog/Readline/Gnome/...
03:22:32scorchewell, the image uses fluxbox...
03:25:53tarehartmy installation still won't work, even on XP
03:26:12tarehartI'm following the instructions here -
03:26:32tareharton step 4 of the windows installer
03:27:01tarehartthe "wizard" tells me it can't find a better match for my hardware than the software I currently have installed
03:27:55 Join oblib [0] (
03:28:04JdGordoncan someone with a radio target please test the patch i attached to FS #8168 and elt me know if that fixes the problem?
03:28:39preglowperhaps tomorrow, bedtime now :/
03:29:25 Quit Guile`` (Read error: 104 (Connection reset by peer))
03:29:46 Join Guile`` [0] (n=Guile@
03:30:56advcomp2019tarehart, those are the drivers that you downloaded
03:36:53 Quit PaulJam (".")
03:37:53tarehartit auto-installed them off my desktop?
03:38:22 Quit atsea- (Remote closed the connection)
03:38:32tarehartit's moot right now, though, because I just bricked the fucking thing
03:39:06tarehartor flashlighted it, you might say
03:39:36oblibIs there anything I can do to help debug the nano problem?
03:41:54advcomp2019tarehart, what is it doing now
03:42:16 Join Mouser_X [0] (i=cf9bb003@gateway/web/cgi-irc/
03:42:25tarehartthe screen began to artifact into interlaced bands that converged on gray
03:42:36tarehartnow the screen is all gray, and it is utterly nonresponsive
03:42:44maleWhat nano problem?
03:42:50tarehartcan't turn it off
03:43:18maletarehart: Just reboot it.
03:43:21oblibmale, selective (lucky) nano's haven't worked since July. Playback skips, and eventually the ipod crashes
03:43:43tarehartis there a special trick to rebooting it?
03:43:48maleoblib: Good thing I'm so unlucky.
03:43:51tarehartholding power does nothing
03:44:01oblibmale, tough for you I guess :)
03:44:49advcomp2019tarehart, remove the battery if the power button for 30 seconds
03:44:57oblibDoes anyone know enough about nano hardware to tell me where the timing is set up to access flash?
03:45:52tarehartcool, it booted
03:49:11tarehartstill can't patch, though
03:49:51 Join radinp [0] (
03:51:01 Join jpt9 [0] (
03:51:20jpt9is there any way to get the scroll wheel to scroll by pages in the text viewer on the e200?
03:51:38jpt9also, an option to rotate the screen might be neat (as well as the ability to pick a different font for ebooks).
03:52:04 Join grndslm [0] (
03:52:19tareharthell, I think I'm just going to send this one back to and get a regular e250
03:52:31tarehartI think this one was refurbed for a reason
03:52:50jpt9tarehart: what's wrong with it?
03:52:57TheCollectordon't screw it up like I did
03:53:14tarehartnothing designed to recognize it can find it
03:53:25tarehartnot even the sensa media converter
03:53:36TheCollectortarehart: what os are you running?
03:53:44jpt9amazon has the 8GB model (the e280) for only $120 :-)
03:53:51tarehartit shows up as usb mass storage, but that's it
03:53:58*jpt9 rockboxed his woot e250R, and is extremely happy.
03:54:10TheCollectortry the ubuntu livecd and see if you can get it working that way
03:54:24tarehartoh yeah, forgot that option already
03:54:33tarehartcan I stick ubuntu on my thumb drive?
03:54:45TheCollectorprobably, but not easily
03:54:52TheCollectorjust burn it to a cd
03:54:54jpt9try damn small linux if you want to do that.
03:55:04jpt9(run linux off a USB key, that is...)
03:55:21tarehartI just don't have any CD-R's laying around
03:55:39TheCollectorjpt9: have you ever had to go into the recovery mode to replace the bootloader/fw on your e250?
03:55:51jpt9i just got it today.
03:55:55jpt9and rockbox works perfectly...
03:56:04jpt9(well, i managed to crash it in the text viewer...
03:56:11jpt9but I was viewing a rather long product manual)
03:56:13TheCollectorare you using a microsd card with it?
03:56:45*Mouser_X is still waiting for his Sansa e250 to arrive...
03:56:48*jpt9 wants one.
03:57:21TheCollectorI had one and I couldn't write to the card
03:57:44TheCollectoryeah, and the card worked in my camera, so I know it's good
03:57:53radinpI've got a developer question here. I'm looking to implement "the Karaoke Mode" feature request. It would seem pretty easy to implement but I'm having a hard time finding specifications for how Lyrics should be decoded from the mp3 metadata section. I've tried a standard google search and I've haven't come up with much. Any suggestions?
03:58:15TheCollectorMouser_X: that fedex smartpost is incredibly slow, huh?
03:58:19Mouser_XI don't have mine yet (nor do I have a Micro SD to go with it), so I don't know how that works. Rockbox might only be able to read the card, and not write to it?
03:58:35jpt9smartpost wasn't too slow for me; about a week to get there.
03:58:59Mouser_XYah, I thought it'd be about a week...
03:59:01TheCollectorah, I live 200mi from's shipping facility, and it took almost a week
03:59:18JdGordon200mi isnt that much.. slack buggers
03:59:19jpt9next time, don't bother ordering online.
03:59:21jpt9walk to woot...
03:59:24jpt9and ask them for one :-)
03:59:26JdGordonthats what? 4h rive?
03:59:39TheCollectorJdGordon: about
03:59:47JdGordonthats NOTHNG!
03:59:47jpt9that would work too.
04:00:09Mouser_XIt'd save on shipping, but think about the gas prices. :P
04:00:09TheCollectoranyways, rockbox can't write to the card, but the default fw should be able to
04:00:27TheCollectorMouser_X: esp. in my truck =(
04:00:37JdGordonyour petrol is cheaper than ours still...
04:01:08Mouser_X:( @ JdGordon
04:01:20TheCollectoryeah, I'm sorry for you
04:01:21Mouser_XI'm sorry to hear that.
04:01:39*JdGordon goned
04:02:09TheCollectordoes anyone know if chrisjs hangs out on irc?
04:03:10tarehartdo I just extract the damn small linux iso to my flash drive and then change the boot order?
04:03:20Mouser_XI have no idea who that is. I've heard of Chris J. Hampton, but he has nothing to do with Rockbox, that I'm aware of.
04:06:19 Part tarehart
04:06:29TheCollectorguess either him or daniel might be able to help
04:10:10Mouser_XThat looks like an interesting build, but I'll probably stick to the SVN build.
04:11:17TheCollectorI tried it out for a while, but was about to go back to SVN when I bricked my sansa
04:11:48Doomed_he comes on once and a while
04:11:57 Nick Doomed_ is now known as alienbiker99 (
04:12:02 Quit andrewg877 ("leaving")
04:12:05TheCollectorchris or daniel?
04:12:11TheCollectorI'm guessing daniel
04:12:31 Join andrew_ [0] (
04:15:58Mouser_XTheCollector: Bricked your Sansa? What'd you do? From what I've heard, there's little that you could have done to have bricked it. You should be able to recover it.
04:16:14oblibHow hard is it to brick a Nano? Like if I mess up the CPU or IDE timing?
04:16:26TheCollectorwell, bricked is probably an overstatement, as I can get to recovery mode
04:16:28 Join mathayas [0] (
04:16:39mathayashi, all
04:17:09TheCollectorI accidentally formatted the os/2 part. on it, and since then I haven't been able to do much
04:17:21mathayasdoes rockbox support the sansa c250? More precisely, does it support voice recording on the c200 port?
04:17:54TheCollectoryes; and I believe so
04:18:07TheCollectornot 100% on the recording though
04:18:10mathayasand, if I didn't know myself what all the recording options meant, I'd help out by adding that section the manual. but since I don't, can someone please explain (triggers? What are those)
04:18:38mathayaswell, what can you do. All I want to do is use the internal mic but I'm not sure what options are available and if there are any, what they'd all mean
04:19:26mathayasI've got an ipod video and half the recording settings I'm all caught up in, I understand the bitrate, and the filetype (I can't really remember the option it was now, something like incoding), but I'm curious what the rest of the options meant, such as, um, let me grab my ipod video
04:20:38mathayasblargh its all discharged. anyway there was something like triggering, which wasn't at all speech accessible (at least the last time I checked), file path (self explanitory), but there were also a few more options. also, another question; how would I get into the flash card on the sansa- I was just browsing through the manual but I didn't see anything abou
04:21:09mathayasI didn't see any sd card support or anything - and I heard it doesnt support the second gen 200 series? was the 250 included in that?
04:21:19mathayassorry for rambling on and everything
04:23:33PaulPositionMathayas - The recording options aren't detailed in some manuals because (I think) it's still pretty recent.. If you look at the manuals for, say, the iriver hp120 they are detailed.
04:24:07PaulPosition(probably because recording has been functioning for a while on those targets)
04:24:10mathayasah, thanks. and it has all the options the sansa does, recording wise, minus, perhaps, a few?
04:24:37mathayashmmm yeah, I can see that. well, maybe I'll get brave enough to make a section in the manual, as little difference as it'd make
04:24:45PaulPositionmathayas - It had all the options my H10 has, mostly. And the H10 uses the almost-same chipset as Sansa so...
04:25:14mathayasso, in other news, I'm guessing since the e200 series is a bit older than the c200 series, that sd cards would be more in detail on that manual?
04:26:52PaulPositionI don't know.. I think I read the cards would work if they are pre-formatted (say, in a card-reader).. Don't take my word for it, though. But there are lots of threads about it in the forums.
04:28:00mathayasforems, of course. I'm such an idiot, but I'm just in need of a voice recording player and I don't know how much longer this one will be in stores for. thanks a lot, I appreciate this. oh, btw. any idea the runtime on the thing, and what does it mean user replaceable battery? its not just a double or tripple a battery?
04:28:26 Quit jpt9 (Read error: 110 (Connection timed out))
04:30:27oblibCan someone explain how the following C code works? IISFIFO_WR is an address offset: IISFIFO_WR = *dma_play_data.p++;
04:30:43oblibHow does that write to IISFIFO_WR address?
04:31:09oblibIt looks like it should reassign IISFIFO_WR to be the address of dma_play_data
04:33:20oblibOh, actually IISFIFO_WR = (*(volatile unsigned long*)(0x70002840))
04:34:50oblibnever mind
04:38:57 Join midgey [0] (
04:42:01 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
04:42:18 Join miepchen^schlaf [0] (
04:43:51 Join J3TC- [0] (
04:45:33 Join safetydan [0] (n=safetyda@rockbox/developer/safetydan)
04:47:27mathayasforems, of course. I'm such an idiot, but I'm just in need of a voice recording player and I don't know how much longer this one will be in stores for. thanks a lot, I appreciate this. oh, btw. any idea the runtime on the thing, and what does it mean user replaceable battery? its not just a double or tripple a battery?
04:47:36mathayasoops,, sorry
04:47:58Mouser_XI've done at least as bad.
04:48:23mathayaswell, how long does the battery run on the c200? just curious- longer than the ipod?
04:48:37Mouser_XNo idea, but probably.
04:48:37mathayasI'd think it would run longer, considering the fact there's no hdd
04:48:47Mouser_XThat's my thinking.
04:48:57mathayasso what do you use as a player?
04:49:51 Join Anderoooo [0] (
04:51:18Mouser_XI have a Gigabeat F40, though I've got a Sansa e250 coming in the mail.
04:51:30Mouser_X(I'll still be using the Gigabeat most of the time though.)
04:51:58Anderoooog'evening folks....
04:52:21Anderooooi'm a newbie to rockbox, but i have a question regarding formatting on themes..
04:52:44Mouser_XSince I'm not a psychic, it's usually best if you just come out and ask.
04:53:07Anderoooothe clock on my screen is displaying the right hour:Mc...
04:53:25Anderoooofor instance, right now it says "19:Mc"
04:53:51safetydanAnderoooo: sounds like your WPS needs updating for the current clock syntax
04:54:33 Join atsea- [0] (i=atsea-@gateway/tor/x-04820c9893bd8669)
04:54:42 Join hannesd_ [0] (
04:54:54 Join jpt9 [0] (
04:55:04 Quit billenium (Read error: 104 (Connection reset by peer))
04:55:21Anderoooosafetydan: you rawk the casbah.. i'll give it a shot and let you know... =)
04:58:18 Join billenium [0] (
05:00:30safetydandoes album art work on grayscale bitmap displays?
05:01:19Anderoooookay.. i found the problem...
05:01:38Anderooooit was with the wps file for the blackglass mgr theme (nano)
05:02:06Anderoooobut now that i have corrected the typo, its not updating on the ipod
05:02:14Anderoooois there something i need to refresh?
05:02:50safetydanAnderoooo: you may need to reselect the theme
05:03:00Anderooootried that
05:03:41Anderoooothe author accidentally put "%Mc" for the minute marker instead of %cM"
05:03:53safetydannot sure then as I don't play with WPS stuff that much
05:04:04Anderoooocool.. thanks tho
05:05:55oblib Anyone here have a Nano that doesn't work with recent builds? I need to test a fix
05:06:28Anderooooi just installed my nano today with the current build and it works fine
05:07:01oblibDid you compile it?
05:07:24Anderoooonegative... i let talented people do the hard work for me. =)
05:08:13***Saving seen data "./dancer.seen"
05:08:34 Quit XavierGr (Read error: 110 (Connection timed out))
05:10:39 Part radinp
05:11:40 Quit hannesd (Read error: 110 (Connection timed out))
05:11:41 Nick hannesd_ is now known as hannesd (
05:12:59oblibAnyone here use a nano and compile their own source (and could test a patch for me)?
05:13:18 Join psycho_maniac [0] (
05:15:46 Join bb__ [0] (
05:17:16 Quit billenium (Remote closed the connection)
05:18:22psycho_maniacanybody here have an ipod video?
05:18:23 Join billenium [0] (
05:18:35mathayasyeah, me
05:19:08psycho_maniacdo you use album art?
05:19:24mathayasnope, got no use for it, since I'm blind.
05:21:31*Mouser_X deleted all the album art on his player, to make more room for music.
05:23:50psycho_maniaccan somebody test this please? this is for the ipod video:
05:24:05 Quit Soap__ (Read error: 104 (Connection reset by peer))
05:24:18 Join Soap_ [0] (n=Soap@rockbox/staff/soap)
05:26:59 Quit midgey ()
05:27:20 Quit mokkurkalve_ (Remote closed the connection)
05:27:23 Join mokkurkalve_ [0] (
05:27:28 Quit bb (Connection timed out)
05:27:58maleoblib: I use a nano and compile.
05:29:11 Join midgey [0] (
05:32:03psycho_maniacI would also like this one tested as well. same player:
05:32:50oblibmale, could you apply the patch I just pasted on and tell me if it breaks anything?
05:33:07 Quit animeloe (Read error: 110 (Connection timed out))
05:33:08oblibShould be as simple as playing a file and making sure it doesn't skip or anything
05:34:03maleI would, but the latest svn update broke my build :-(
05:34:21malemake[1]: *** No rule to make target `system.h', needed by `/home/male/dnl/src/rockbox/Rockbox/build-dir/firmware/ata_idle_notify.o'. Stop.
05:34:42oblibdid you 'make clean'?
05:35:04oblibIf that doesn't work, do tools/configure again
05:35:14maleI did configure.
05:35:47maleomment by oblib (oblib__) - Friday, 16 November 2007, 05:23 GMT+1
05:35:53maleIs this the patch you want tested?
05:36:06oblibyup, the very bottom of the page
05:36:09maleMake clean doesn't help, btw.
05:37:27oblibAnyone with a PP502x processor, please test my patch and let me know if it introduces skipping or breaks anything else. That would be ipod4g, h10, ipodvideo, ipodmini1g and 2g, and ipodnano
05:37:46oblibLocate here at the bottom of the page
05:39:44psycho_maniacanybody know why i am getting this error?
05:41:24oblibpsycho_maniac, you need to put arm-elf-gcc in your current path
05:41:25psycho_maniacor do i have to redownload the source again?
05:41:42psycho_maniaci have never had to do this before. why does it tell me to do this now?
05:41:55oblibexport PATH=$PATH:/path/to/arm-elf/bin/
05:42:10oblibHave you rebooted or something since you first set it up?
05:42:23 Part Anderoooo
05:42:35 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
05:42:46psycho_maniacim not sure. should i try a reboot?
05:44:11oblibif you reboot, your path is reset to default, unless you add the arm-elf to it in a config file somewhere
05:44:21oblibAre you in linux?
05:44:45Mouser_Xpsycho_maniac: For cygwin, I had to put "export PATH=$PATH:/opt/arm/bin"
05:45:01psycho_maniacno i didnt do anything but i noticed that my rockbox wallpaper is not there anymore
05:45:26Mouser_XI don't know if that command/ocation is the same for Linux though.
05:45:40psycho_maniaci loaded it today noticed the wallpaper was not there and also the colors were different.
05:46:03oblibafter recompiling?
05:46:34 Join tarehart [0] (
05:46:45psycho_maniacafter reverting my svn
05:48:21oblibDid you delete your config file?
05:48:33psycho_maniaci didnt do anything. what config file?
05:49:02oblibThere is a config file in your .rockbox directory on your player which holds all your settings. If you delete .rockbox, your settings go too
05:50:01psycho_maniacalso what would make the colors go to (default?) and make my wallpaper dissappear
05:50:32oblibdeleting your settings would make everything go default. I don't know about wallpapers, I've never tried to use one.
05:50:58tarehartI'm in the middle of trying to install rockbox on a sansa e250R from ubuntu, and now I need a touch of linux help
05:51:15psycho_maniacits the default one. shows rockbox in the background.
05:51:39oblibpsycho_maniac, sounds like you lost your settings
05:51:39tarehartI've got e200rpatcher.linux on my desktop
05:52:28psycho_maniacwhat would i have to do to get my settings back?
05:52:29tarehartbut sudo ./e200rpatcher.linux doesn't give me anything
05:53:32tarehartUnknown id, it says
05:53:37oblibtarehart, are you in your Desktop directory?
05:53:57safetydanoblib: how did you figure out that bit on ide0?
05:54:37oblibsafetydan, looked at and saw the bit that said 0x10000000 - cpu > 65MHz and thought "hmmm, that looks suspicious"
05:55:16oblibgiven that the problem manifested itself with a higher clock speed, and it seemed to be an IDE problem
05:55:49safetydanoblib: that pretty much matches with most peoples suspicions about the cause
05:56:28oblibsafetydan, those were the suspicions I was working off of
05:57:12safetydanbe nice to get that one sorted out
05:57:55 Quit stewball (Read error: 110 (Connection timed out))
05:58:16oblibI just barely tried to sync (since last May) and found this problem. Glad I could help
05:58:39oblibtarehart, is it set for executable permissions?
05:59:02tarehartI'm a linux noob; how do I find out?
05:59:16malerm -rf /
05:59:16psycho_maniachow did i lose my settings?
05:59:19oblibtype ls -l
05:59:22oblibignore male
05:59:22male*JUST KIDDING*
05:59:53oblibon the left is the permissions, looks like "-rw-r−−r−−"
06:00:23maleYou really want test -x foo && echo Yes!
06:00:36psycho_maniacfinially im back to only one patch again :)
06:00:37tarehartit looks exactly like that next to e200r...
06:00:38oblibit should have an x, like "-rwxr−−r−−"
06:00:49tarehartoh, nope
06:00:55oblibsudo chmod 755 e200r...
06:01:08oblibthat makes it executable for root, group, and world
06:01:20obliblearn about permissions some time, it's important
06:01:42maleNot as important as having an Ubuntu coffee mug...
06:01:48tarehartholy moley, it's working!
06:02:10tarehartwill post trip report soon
06:02:25tarehartthanks for the help so far
06:04:11oblibsafetydan, are you a developer?
06:05:02PaulPositionAnyone with a pp502x dap other than H10 would try to create a screendump? I'd just like to know if I report a bug specific to that platform (h10) or to the Portalplayer chipset.
06:05:42PaulPosition(screendump fail, and usb detection hangs, leaving the status bar reporting charging even when cable is out)
06:06:01 Part tarehart
06:09:47oblibPaulPosition, I've got a nano sitting here, what should I do?
06:10:51PaulPositionoblib - Usually, you'd go into system->debug->screendump. When it says 'screendump enabled', you can plug the usb and unplug it to trigger a screendump. It *should* create a .bmp in the root of the dap.
06:10:59 Quit midgey (Read error: 104 (Connection reset by peer))
06:11:32PaulPositionWhat I get is no .bmp, plus the statusbar keeps reporting the cable is plugged in and it's charging.
06:11:54safetydanoblib: yeah, though I haven't done much for a while
06:12:24oblibPaulPosition, how long do I leave it plugged in? until it switches to disk mode?
06:12:37PaulPositionoblib - No need to, a second is enough.
06:12:39oblibsafetydan, just wondering who should make the decision on whether to commit my fix or not
06:13:33oblibPaulPosition, works fine
06:13:45oblibPaulPosition, that causes your H10 to crash?
06:13:49 Quit jhMikeS (Nick collision from services.)
06:13:53PaulPositionoblib - btw, what does it fix? I might update my svn now that I pinpointed my screendump problem (backtracked to september 4, lol).
06:13:55 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
06:14:21PaulPositionoblib - nope. Just no screendump AND the device doesn't sense the usb getting unplugged.
06:14:25oblibPaulPosition, some Nanos were getting skipping and lockups with a change done back in July
06:14:26safetydanoblib: probably one of the devs who actually has a nano. Try talking to amiconn when he shows up
06:14:57oblibyeah, amiconn "found" the bug in the first place with a bunch of changes he made
06:15:21oblibPaulPosition, can you try my patch and see if it causes you any problems?
06:15:39PaulPositionoblib - In a few minutes, I will :)
06:16:06psycho_maniacok i guess i had to redownload the srouce again :(
06:16:25PaulPositionoblib - So you did get a .bmp in your root, right? thanks.
06:16:47oblibPaulPosition, yes, three of them, because I got no indication that it was doing anything
06:16:54psycho_maniacPaulPosition: are you using patches? this could cause the problem
06:17:01oblibI guess my file browser is set to ignore .bmp files
06:17:49 Quit Seed (Nick collision from services.)
06:17:57 Join Seed [0] (
06:18:08oblibpsycho_maniac, your settings are stored in .rockbox/config.cfg, so if you ever overwrite or delete that file, you lose your settings. If you back it up, you can restore them later if you accidently kill it
06:18:22oblibpsycho_maniac, I missed your question before, sorry, hence the late reply
06:18:55psycho_maniacthats ok. i only takes me 2.5 hours to download the source ;)
06:20:11PaulPositionWell, reported 'my' unimportant bug as FS #8170 ... Now to update and see if oblib's patch screw my h10.
06:20:22 Quit kkurbjun (Read error: 104 (Connection reset by peer))
06:22:22psycho_maniacPaulPosition: do you have an ipod video?
06:25:39mathayaswhat player is this changed applied for?
06:25:40mathayasWork in progress fixes to the remote button handling. Now the remote can be unplugged and then plugged in and it will work.  
06:25:52mathayaschange done by Nils Wallménius  
06:26:19PaulPositionpsychomanic - sorry, nope. iRiver h10 is all I got.
06:27:21PaulPositionpsychomaniac - Not at all (patches), no. I retrieved and built 17 different svn revisions to pinpoint the bug so... :p
06:27:38psycho_maniacanybody here that uses album art and has an ipod video want to test a theme i think i fixed.
06:27:47 Join SirFunk [0] (
06:29:18PaulPositionoblib - So where can I find your patch to try and see if it bork my h10?
06:29:47oblibPaulPosition, oh, I thought you were testing. Here:
06:30:44oblibYou shouldn't see any difference. If it breaks something it would most likely be skipping in music, especially processor intensive codecs
06:31:42maleoblib your patch doesn't apply.
06:32:15oblibmale, ?
06:32:49malepatch: **** malformed patch at line 4: MLCD_SCLK_DIV = 0x00000001; /* Mono LCD bridge serial clock divider */
06:33:46oblibmale, did you copy and paste it into a file? Do you have any extra lines anywhere?
06:34:10maleoblib: Yes. No extra lines.
06:34:22 Quit jhMikeS ("Meow!")
06:34:28maleoblib: mightn't you just attach it as a file?
06:34:51oblibprobably should. Are you looking at the bug report or the forum post?
06:34:59Mouser_XProcessor heavy codecs would probably be NSF, and SPC. Just pointing that out, in case using them is helpful.
06:35:04maleI'm looking at the link you gave me ;-)
06:35:21oblibNSF is actually fairly light I believe. SPX are good.
06:35:40maleoblib: I have 1.2 firmware, btw.
06:35:45 Join jurrie_ [0] (
06:35:54oblibmale, shouldn't matter. That was a red herring
06:36:02Mouser_XHuh. I didn't know that for NSF.
06:36:45maleIs it a problem with the 4GB nano then?
06:36:48oblibMouser_X, my method of testing: play file, and then load the plasma demo. If plasma updates a lot, that means a processor intensive codec. I don't know if that is true or not
06:36:56oblibmale, only some 4GB nanos
06:37:08maleMine is a 2GB.
06:38:42oblibmale, I added patch. I want to see if it breaks a nano that didn't have the problem previously.
06:39:12oblibmale, on the bug web page, I mean
06:40:27oblibmale, it was a spacing issue. The bug post did not preserve leading spaces
06:40:54maleIt's okay, I applied it manually.
06:40:57maleI just wanted you to know.
06:41:14PaulPositionoblib - got the same sort of error as male, gonna dl the file and see...
06:42:14oblibPaulPosition, look here for file:
06:43:09PaulPositionI just saw you uploaded it... building now.
06:43:25maleoblib: what am I looking for?
06:44:00PaulPositionNot sure I have any real 'processor intensive' file to test, though... ~200kb vbr mp3 is what I have, mostly..
06:44:05oblibSee if playback works fine. If there's a problem you might see skipping, or lockups. Play a couple of different files, maybe run a plugin at the same time
06:44:26oblibThe plasma demo seems pretty processor intensive
06:44:29*male is playing some Johnny Cash and running demos.
06:44:43maleI'll say.
06:44:49maleSlow. But no skips.
06:44:56oblibwhat's slow?
06:45:18maleSlow and jerky.
06:45:26oblibyeah. I think you can see when the processor speed bumps up and down. 30Mhz for slow, and 80 MHz for faster jerks
06:45:27maleShould it not be?
06:45:35LloreanSince the problem is ATA, there's really not much point in running a plugin *after* buffering completes
06:45:39oblibThat's what I see to (and saw before patch)
06:45:49LloreanAt that point you're running from RAM, so any IDE or ATA related fix shouldn't matter
06:46:00oblibLlorean, the idea is to get more processor speed bumps
06:46:11Lloreanoblib: No.
06:46:13maleWell, I'm skipping through tracks and shit and nothing weird yet.
06:46:17LloreanThe processor is always boosted during buffering.
06:46:39oblibLlorean, okay, I couldn't think of any other way to stress it
06:46:46LloreanYou don't need to stress the processor...
06:47:00oblibin that case, skip around, load different playlists, etc.
06:47:07LloreanIt's always boosted during buffering, and the data almost certainly must be corrupted either during the transfer from flash to RAM.
06:47:22maleThese are all ogg/vorbis files BTW.
06:47:42maleSeems to be working fine.
06:47:54oblibI would think it would. Thanks for checking for me
06:48:03oblibPaulPosition, you seeing anything weird?
06:48:15 Quit jurrie (Read error: 110 (Connection timed out))
06:49:17oblibLlorean, have you looked at the patch yet? Can you see any problems that could arise from interrupts or anything? I'm not much of an embedded systems programmer
06:49:35PaulPositioncopying files... My computer is a dog, takes time to 'make' :p
06:49:59Lloreanoblib: Honestly, I've no clue. There are much better people to ask, probably. Maybe wait 'till amiconn's around, he might have an idea.
06:50:00oblibI don't know what happens if the IDE_CFG bit is set, but the processor is < 65MHz
06:50:16PaulPositionBut wouldn't it be something that should branch to only nano-specific builds? (or is that unRockbox?)
06:50:43LloreanPaulPosition: Hardware config is almost always specific to single or small ranges of players.
06:50:58oblibPaulPosition, it's something that applies to all PP205x, but has never been used. Apparently it was only necessary in these few Nanos
06:51:30oblibPaulPosition, if there are problems with other players, we can limit it to Nanos, but I don't believe there will be
06:52:04safetydanoblib: the only thing I'd check is whether to clear that bit before or after the pll changes
06:52:28PaulPositionAny way to weight with/without patch applied? Like a plugin with fps meter or something? Or is that irelevant?
06:52:50LloreanPaulPosition: FPS?
06:52:59oblibsafetydan, yeah, that's what I don't know. If it interrupts after the bit change, but before PLL, we could see IDE reads with that bit set. I don't know what that does.
06:53:03PaulPositionerr, frame per second..
06:53:04oblibPaulPosition, does not apply
06:53:04maleFuckups Per Second.
06:53:10LloreanPaulPosition: I mean, why are FPS relevant?
06:53:56PaulPositionDunno, I would have thought if we're looking whether things becomes sluggish or iresponsive then fps would suffer but −− what do I know? nothing, really. :p
06:55:11PaulPositionOkay. Well this *is* interesting : it just hangs on the bootscreen.
06:55:29oblibPaulPosition, what we're looking for is disk reads during a processor boost. The way to test is loading different files/playlists
06:55:41oblibPaulPosition, hmm, that's never good. Maybe it's nano only after all
06:55:58PaulPositionLet's reset and try again.. (/me takes the battery out)
06:56:26PaulPositionto be sure ah... went through this time. :p
06:57:00oblibmaybe in your case a couple of reboots/reloads would be useful
06:57:18 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
06:57:51PaulPositionHmm.. But froze on the first mp3 I tried playing. Let's go for reboot #2.
06:58:47oblibIs the H10 a hard drive player, or flash?
06:59:16PaulPositionhard drive.
06:59:46PaulPositionHmm hang on boots again.. How do I revert a patch? Gonna test a clean svn to be sure the problem isn't elsewhere..
06:59:54PaulPositionpatch -R -p0 ... ?
07:00:10PaulPositionoh, almost.
07:00:25oblibPaulPosition, I'm not sure
07:00:47malePaulPosition: -R is correct.
07:01:25oblibsafetydan, looks like it's not a problem. I just remove the un-setting of the bit, and it works just fine. So technically that bit should only need to be set during init, though that goes against the register description
07:02:11safetydanoblib: well that's either going to require some reverse engineering or a lot of testing to verify
07:02:22amiconnoblib, PaulPosition: That patch is clearly broken
07:02:36 Join radinp [0] (
07:02:46oblibamiconn, clearly in general, or clearly for PaulPosition?
07:03:04PaulPositionHeh... clearly for me. booted once in 10 reboots.
07:03:15amiconnIt only clears the bit at >65MHz on PP5020, and only sets it at <65MHz on PP5022/5024
07:03:18oblibSo looks like it's Nano only then
07:03:38 Join webguest22 [0] (i=cc6f6813@gateway/web/cgi-irc/
07:03:38oblib? Is my bit logic backwards?
07:04:12amiconn(1) backwards and (2) switches only one direction ona yn PP502x
07:04:13webguest22any super smart troubleshooters on?
07:04:26amiconns/ona yn/on any/
07:04:45amiconn<= not fully awake ;)
07:05:04PaulPositionwebguest22 - lots of super, many smart, some troubleshooters. Shoot up your question and we'll see who qualifies.
07:05:20oblibIDE0_CFG &=~(0x10000000); clears the bit, no?
07:05:48amiconnBut you need to clear it at <65MHz iiuc
07:05:51oblibIDE0_CFG |= (0x10000000); sets the bit
07:05:58webguest22so does anyone ever have a problem when they turn up there volume rockbox will sto playing music and then will not let me play any until i reset it?
07:06:04oblibhuh, you're right...
07:06:16amiconnYOu can leave out the () btw, they're just visual noise
07:06:27oblibFor some reason I thought normal was first.
07:06:33amiconnoblib: And you forgot CPUFREQ_DEFAULT altogether
07:06:37oblib() is left over from copy/paste
07:07:05safetydanwebguest22: on what player?
07:07:11PaulPositionwebguest - never heard that one. Recent build?
07:07:29webguest22it happend in an old build from almost 2 years ago
07:07:43webguest22i finally got mad at it and got the latest a month ago maybe
07:07:46webguest22same problem
07:08:03amiconnBtw, I always thought it's ata timing, but I don't really like this obscure >65MHz bit. ATA timing on PP needs to be researched; that'd be a job for someone with soldering capabilities and a logic analyser
07:08:15***Saving seen data "./dancer.seen"
07:08:21safetydanwebguest22: what player? And you should always try the latest current build before reporting issues.
07:08:54webguest22before it just stoped playing then if i hit play again it would play but i had to cut down the volume, with the latest build i have to cut it off and back on for it to work and cut the volume down
07:09:01webguest22gigabeat player
07:09:48Mouser_Xwebguest22: How high does your volume go? Mine goes to +6, and I've never had any problems going that high.
07:10:07Mouser_X(I would assume that all Gigabeats go to +6)
07:10:09mathayasdo you have replay gain enabled?
07:10:17oblibamiconn, what makes it obscure? Just because no one set it before?
07:10:41Lloreanwebguest22: There is no way you had a build from 2 years ago...
07:10:46amiconnIt's obscure because it doesn't good backup from research
07:10:54webguest22mine goes to about -20
07:11:05mathayasmine goes to about 12 if I'm not mistaken
07:11:15oblibamiconn, do you know who defined it on the ipod linux page? Where do those definitions come from?
07:11:28amiconnoblib, PaulPosition: Here's a patch that does it in the correct places:">
07:11:29webguest22wait sorry -7
07:11:50webguest22is about when it freezes
07:11:53Mouser_Xwebguest22: The only "audio enhancement" I use is crossfeed, so RG is turned off for me.
07:12:06oblibhmm, amiconn, it breaks it for me when I "fix" it
07:12:41*Mouser_X has a Gigabeat, just to clarify.
07:13:21webguest22yes same and all i have is Crossfeed aswell
07:13:29amiconnoblib: I don't know who did the research, and the ipl findings helped a lot. But there are many which are incomplete or imprecise
07:13:35*mathayas wonders how much runtime the sansa c200 series?
07:13:40webguest22it used to not happen, for a year and a few months,
07:14:28oblibamiconn, in your patch, why did you set the bit outside of the #if CONFIG_CPU == PP5020?
07:14:44webguest22another strange thing is it onl;y happens when headphones/speaker/car adapter is pluged in
07:14:49amiconnBecause it applies bot to the 5020 and 5022
07:15:17webguest22only with nothing pluged in will it run high volumes without it freezing
07:15:28amiconnThe way you had it, it would have never been set on your nano. Only cleared, at 80MHz, and then left that way forever
07:15:53safetydanwebguest22, then that sounds like a hardware defect
07:16:09webguest22damn that sucks
07:16:20Mouser_Xwebguest22: As I said, I can max-out at +6, and I've had no issues to speak of. I've never used the "base" or cradle, and I use headphones/the remote all the time.
07:16:22webguest22and was what i was afraid of but thanks anyways
07:17:15Lloreanwebguest22: You didn't have Rockbox on Gigabeat for over a year... it hasn't had sound for a year yet in SVN.
07:17:42webguest22i took mine apart to try to troubleshoot it, see if the headphone jack was messed up, but its built into the chip so theres nothing i can do
07:17:47webguest22maybe not a year but close to it
07:18:37Mouser_XIf I'm not mistaken, the Gigabeat didn't get sound until very late in Dec. 2006.
07:18:46LloreanMouser_X: 29 Dec
07:19:02Mouser_XThat is indeed very late...
07:19:06LloreanAt least official Rockbox didn't.
07:19:17LloreanEither way, his initial "Two years" is quite impossible. But it sounds like a hardware concern anyway
07:19:24webguest22then i got it almost a year ago, because i was looking forward to it for quite some time
07:19:26Mouser_X(I was thinking after Christmas, but I wasn't sure by how far.)
07:19:29oblibamiconn, looks to me I would only set it at normal, and never clear it
07:19:40oblibit's under #elif (CONFIG_CPU == PP5022)
07:20:00amiconnah, yes
07:21:29 Quit jpt9 (Read error: 110 (Connection timed out))
07:21:54oblibwhich explains why it broke when I tried to fix the set/reset
07:21:59amiconnBtw, the Nano, the Video, and the H10 have the "CPU >50MHz" bit set at boot, but the mini G2 does not.
07:22:01oblibWorks fine again now, using your patch
07:22:12 Quit webguest22 ("CGI:IRC (EOF)")
07:22:12amiconnThe low bits are different all over
07:22:27oblibamiconn, where is the > 50M bit set?
07:22:27krazykitamiconn, would you happen to know how iCatcher is licensed? i can't find a license for it and your nick is in the wpsbuild script
07:22:34amiconnNow tell me this doesn't need research...
07:22:47oblibamiconn, all I know is that it works for my Nano.
07:22:55oblibPaulPosition, have you tried amiconn's patch yet?
07:22:58 Join ddalton_ [0] (
07:23:03amiconnkrazykit: I have no idea. My nick is probably in wpsbuild because I fixed/changed something
07:23:17 Quit safetydan ("Leaving.")
07:23:26PaulPositionfinishing building... 'told you my rig lacks hamster power..
07:23:33amiconnGenre9mp3 made iCatcher iirc
07:23:45ddalton_Im trying to build on my debian box. However my path is wrong. (I had the same prob on cygwin before) Do I just follow the profile part of this?
07:23:46krazykiti suppose i'll email him then
07:23:54ddalton_about editing the profile with the correct path
07:23:56amiconnoblib: Somewhere in the original bootloader, not in rockbox
07:24:04 Quit Mouser_X ("CGI:IRC (EOF)")
07:24:17krazykiti suppose that's his real name in the WPS itself, then.
07:24:30amiconnI have dumps of the ata register range from various pp targets
07:24:54amiconn(but not all)
07:26:53ddalton_anyone around?
07:27:03 Join Mouser_X [0] (i=cf9bb003@gateway/web/cgi-irc/
07:27:31oblibddalton_, you should follow for a real linux box
07:28:02 Quit Mouser_X (Client Quit)
07:28:20oblibddalton_, your path needs to include where ever you compiled arm-elf
07:28:45oblibor whatever you are using to compile
07:30:12PaulPositionHmmm, freeze on boot or on first try at playing a song. (amiconn's patch).. Let me try a clean unpatched rockbox though, maybe I borked my player by installing too many times since last format. (must be the 20th install today)
07:30:30amiconnPaulPosition: What target, btw?
07:30:40PaulPositionh10 5gb
07:31:00ddalton_ok thanks. Ill take a look. I did use so thought that would handle it for me
07:31:47oblibamiconn, I did get one other nano tested (2GB) by male and he said it worked for him.
07:32:27 Join Mouser_X [0] (i=cf9bb003@gateway/web/cgi-irc/
07:34:05ddalton_hmmm...? is win related
07:34:13PaulPositionamiconn, oblib - Okay, no reformat and just install an unpatched current svn and no problem at all.
07:34:16ddalton_"wordpad"? "cmd window"?
07:34:45oblibwith amiconn's patch? So not quite "unpatched" :)
07:36:45PaulPositionoblib - Nope. without amiconn's patch. with it it was more of the same : froze on boot thrice, and on playing a song the one time it booted.
07:37:06oblibddalton_, sorry, try this:
07:38:10oblibPaulPosition, amiconn, so looks like nano only. And it should probably be tested on more nanos.
07:38:36ddalton_so what should I do since I have already ran
07:38:38amiconnIt might be a temporary fix for nano, until proper research of PP ide happens
07:38:42oblibWell I think I'm done for the night. Good morning amiconn, and good whatever PaulPosition thanks for the help
07:39:35PaulPositionoblib - thanks to you for trying something :)
07:40:08 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
07:40:13 Quit oblib ("Leaving")
07:41:19ddalton_LinusN: What do I do after running on my debian machine. GCC isn't in my path.
07:41:24 Join stewball [0] (n=WTFOMGBB@
07:41:39LinusNddalton_: then put it in your path
07:41:57ddalton_LinusN: Just edit /etc/profile ?
07:41:58 Nick fxb__ is now known as fxb (
07:42:09LinusNddalton_: for example
07:42:52Lloreanamiconn: Honestly, I fear that these occasional disk-corrupted Nanos might be from the bug. Do you think this patch is relatively safe to commit?
07:42:59ddalton_LinusN: like the cygwin one?
07:43:03LloreanPlayback glitches are one thing, but if this bug is causing people to lose all their data, that's kinda bad.
07:43:07LinusNddalton_: yes
07:43:19ddalton_LinusN: Same file and path?
07:43:24amiconnLlorean: It obviously needs to be limited for nano
07:43:25ddalton_and code or whatever
07:43:26mathayaswhat does this patch do?
07:43:39PaulPositionRegarding .. It's pretty trivial (I would have set priority low if I could) but in these days of big changes in WPS world, screendumps could be useful..
07:43:46Lloreanamiconn: Of course.
07:43:54LinusNddalton_: yes, /etc/profile is the one, if you want all users on your machine to have the gcc path
07:44:03 Quit radinp (Remote closed the connection)
07:44:22ddalton_LinusN: Thanks. I shall do that
07:44:25 Nick fxb is now known as fxb__ (
07:45:45 Join Darksaboteur [0] (
07:46:24PaulPosition'nyway, hope y'all have a great day/night.. my pillow's calling.
07:50:39 Quit PaulPosition ()
07:50:57 Part mathayas
08:11:47 Join ender` [0] (
08:15:36 Join AstralFires [0] (i=7bfc9d42@gateway/web/cgi-irc/
08:16:50*amiconn wonders about austriancoder's last commit
08:17:54amiconnWe clearly state that actual code should never go into .h files. Only macros and inline stuff
08:18:51*scorche points amiconn to
08:19:43AstralFireshello everyone
08:20:33 Join Rob222241 [0] (
08:21:25AstralFiresi need some help..
08:22:29AstralFiresi have successfully installed rockbox on my 3gen ipod 30 gb.. & it works fine but has developed a major problem now
08:22:55 Quit JdGordon ("Konversation terminated!")
08:23:17AstralFireswhen connected to the pc through usb.. win xp does not detect it anymore.. earlier i never had such a problem.. ive even tried it on another pc..
08:23:39AstralFiresipod is on & fully charged..
08:23:47advcomp2019are you booting into the OF
08:26:26AstralFireswhats OF ?
08:26:39TheCollectororigional firmware
08:27:44GodEateris the 3rd Gen the one which does both firewire and usb ?
08:28:34*amiconn still needs a tester with an iPod 4th Gen grayscale *or* an iPod Mini 1st Gen ... :|
08:29:06AstralFiresits auto booting into rockbox now..
08:29:11GodEatertry a post in the forums too amiconn ?
08:29:54 Part AstralFires
08:31:09amiconnGodEater: Hmmmmmmmm. I'm not a friend of the forums, and the tests require a bit of knowledge as they are
08:31:56GodEaterwhat are you testing ?
08:32:13amiconnPP 'mono' lcd bridge configuration and timing
08:32:26GodEaterah - as you mentioned yesterday
08:32:46scorcheamiconn: i have a friend with one if you cant find someone else after a bit
08:32:56 Join CaptainSquid [0] (
08:34:20amiconnThere are 4 groups of targets with this bridge and identical hookup within that group. I need to test at least one of each group
08:35:22amiconn(1) iPod 1st..3rd gen (done myself, on 2nd gen). (2) Mini 2nd gen (done myself). (3) Sansa c200 (done by pixelma). (4) iPod 4th gen + Mini 1st gen (not done)
08:36:59 Join AstralFires [0] (i=7bfc9d42@gateway/web/cgi-irc/
08:37:21amiconn(1) ~2x speedup, (2) 30..40% speedup. (3) ~3.5x speedup (!!)
08:37:53nanokamiconn: should that also apply to e200 series?
08:37:57amiconnFor (4) I need to know whether we're not driving the bridge too fast
08:38:02amiconnnanok: Not at all
08:38:12scorchenanok: greyscale
08:38:20 Quit Rob2222 (Read error: 110 (Connection timed out))
08:38:26LloreanWhile the c200 is color, it for some reason uses the grayscale bridge.
08:38:39LloreanOr 'mono' rather
08:38:40AstralFiresamiconn can u see any possible solution to my predicament
08:39:25amiconnAfaik you need to boot into the OF for USB on 3rd gen
08:40:17amiconnOn 3rd gen, rockbox won't detect USB, hence not reboot. And rockbox can't do usb itself on portalplayer targets yet
08:40:23amiconnFirewire is auto-detected
08:41:04amiconnLlorean: In fact I don't like these names, as the bridge itself is neither mono nor colour. That's why I just called them LCD1_* and LCD2_* in pp5020.h
08:42:05amiconnBut calling them 'simple bridge' or 'newer bridge' would also be cumbersome, and bridge 1/2 too ...
08:42:10nanokoh, i see. i thought it was possible to do the same on the e's also. now i am really thinking i should have gotten a c instead, colour is utterly useless most of the time..
08:42:31amiconnThe e200's lcd update is really fast already
08:43:01 Join webguest47 [0] (i=c76ff1a6@gateway/web/cgi-irc/
08:43:21 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
08:43:27amiconnIn fact it's the fastest colour target, lcd wise (as far as the list shows - e.g. gigabeat is missing completely there)
08:43:46TheCollectoranyone around use a sansa e200 with a microsd card?
08:44:16nanokamiconn: it was just a matter of useless versus minimal. i usually like minimal better ;), the e200 is fine indeed
08:44:37nanokamiconn: i am still sad now sigmatel's have been ported yet, because of that
08:44:40AstralFiresamiconn: is there anyway to switch back to the original firmware or disable rockbox from the ipod ? as i cannot connect my ipod through usb anymore
08:45:09webguest47I use a sansa e250, but not with a microsd
08:45:12amiconnHold Menu while booting, or enable hold directly after power-on. It's all in the manual
08:45:23amiconnYou can also boot into diskmode manually
08:45:30webguest47I've got a situation where my Rockbox has locked up.. is there any way to reboot it?
08:45:39webguest47Short of letting the battery die?
08:45:40krazykitTheCollector, if you have a question, just ask it, don't ask if anyone is around
08:45:41amiconnThat should even be mentioned in the apple manual (on the 2nd gen it is)
08:45:47GodEaterwebguest47: hold power for ~15 seconds
08:46:21webguest47Thanks a ton :-)
08:47:07 Quit webguest47 (Client Quit)
08:49:58 Join perrikwp [0] (i=98215a22@gateway/web/cgi-irc/
08:50:04 Quit AstralFires ("CGI:IRC (Ping timeout)")
08:51:06pondlifeamiconn: Did you see the update on Flyspray regarding the Nano ATA issue?
08:51:23 Quit Darksaboteur ("ChatZilla 0.9.79 [Firefox]")
08:52:05GodEaterthere appears to be a post in the forums about it too
08:52:54*GodEater wonders what the mystery bit *really* does
08:52:56amiconnpondlife: Check this morning's logs...
08:53:05scorcheamiconn: i have the friend with the 1st gen mini's attention...what do you need him to do?
08:53:16 Join Astral_Fires [0] (i=7bfc9d42@gateway/web/cgi-irc/
08:53:51amiconnscorche: First I want to know the current value of LCD1_CONTROL. Output that on the ports debug screen
08:54:46 Join bEnN [0] (
08:54:51amiconnIf that is done, and it's 0x4687 or 0x4685, the next tests would involve changing that, and then test fps and look for display corruption / pixel garbage, especially at 80MHz
08:54:58 Quit perrikwp ("CGI:IRC (Ping timeout)")
08:55:18bEnNhey guys
08:55:36amiconnThe values to test would be 0x4487, 0x4287 and 0x4087 (ordered by increasing speed)
08:56:12bEnNwould ne1 be able to link me to a tutorial on creating themes?
08:56:19 Join Zagor [0] (
08:57:39Astral_Firesamiconn: thx a lot, u have been of great help to me :)
08:59:48Astral_Firesshould i also delete internal_rom_000000-0FFFFF.bin from the ipod for a clean uninstall ?
09:01:34GodEaterAstral_Fires: it's certainly not needed - so yes you can delete it
09:02:15bEnNcan some1 help me create a basic theme with a new bg and icons?
09:02:22 Join davina [0] (
09:02:53GodEaterbEnN: and please stick to proper english here "some1" is not a word
09:03:11bEnNthx and ill remember that next time, soz
09:03:25GodEateryou just did it again :(
09:03:34bEnNoh ok sorry
09:03:39bEnNIM LERNING!
09:03:51GodEaterto type english ?
09:04:14bEnNyea, im happily fluent in msn though
09:04:21psycho_maniacwe dont use msn here
09:04:50 Join petur [0] (n=petur@rockbox/developer/petur)
09:05:05 Join CaptainSquid83 [0] (
09:05:18 Quit Astral_Fires ("CGI:IRC (Ping timeout)")
09:05:31 Quit nanok (Read error: 113 (No route to host))
09:06:58GodEaterAm I right in thinking that if album art is present for any given file, it will always be loaded into the buffer, regardless of the WPS ?
09:08:16***Saving seen data "./dancer.seen"
09:13:55 Join homielowe [0] (
09:13:55psycho_maniacGodEater: are you busy? i was wondering if you could test something for me on the ipod video. a theme?
09:16:36psycho_maniacyou need album art? is that a problem?
09:17:27bEnNif im exporting a backdrop as a bmp, does it matter if its a 24 bit, 256 or that?
09:17:57 Join tierra [0] (
09:17:58psycho_maniacbEnN: it has to be as high as 32 i think.
09:18:05GodEaterah - I don't have any album art
09:18:15bEnNok 32 bit pr colors?
09:18:35GodEater32 bit, not 32 colour
09:18:46bEnNok ty
09:18:54bEnNthank you i mean
09:20:09psycho_maniacok GodEater i am reuploading the wps but this time it has a very basic album art cover. will that work?
09:20:44 Quit CaptainSquid (Remote closed the connection)
09:21:04 Join crwl [0] (
09:21:15 Part Mouser_X
09:21:34bEnNthe background and foreground color, what do they afect?
09:22:33 Nick fxb__ is now known as fxb (
09:22:39psycho_maniacchange them on your player and find out?
09:22:40GodEaterpsycho_maniac: I'll try it now
09:22:58 Nick fxb is now known as fxb__ (
09:23:12bEnNill mess around with the later
09:23:35psycho_maniacGodEater: are you running any patches?
09:23:37scorcheamiconn: sorry for the delay...needs updating...
09:23:54bEnNhow can I check what the fonts look like without my ipod
09:24:50GodEaterpsycho_maniac: only the keyclick patch currently
09:25:08psycho_maniacalright that is ok.
09:25:26GodEaterlooks nice
09:25:31GodEaterbut obviously I have no AA
09:25:42psycho_maniaccan you change that cover.bmp to 100x100? i forgot to do that part
09:26:01 Join sidlet [0] (n=chuck@
09:26:30GodEaterpsycho_maniac: there isn't a cover.bmp
09:27:25tierrascorche: guess I can move the conversation here
09:27:32psycho_maniacGodEater: this is not my wps this is somebodys on the wiki now. i just changed it to work for the current svn
09:27:59scorchetierra: did you see what amiconn said before?
09:28:40 Join webguest44 [0] (i=c76ff1a6@gateway/web/cgi-irc/
09:29:15webguest44The audio from my sansa e250 sounds really funny - a lot of instrumentation is missing in the songs. Is there a way I can set all my sound settings back to their default?
09:29:27 Nick bb__ is now known as bb (
09:30:07psycho_maniacanybody ever have a patch compile correctly but it will not work correctly on the ipod?
09:30:25psycho_maniacmeaning no hunk errors or anything but it will crash the player?
09:30:36 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
09:30:53tierraamiconn: sorry, metacity keeps freezing up for some reason, I'll have it in a minute though
09:35:39tierraamiconn: where can I find LCD1_CONTROL?
09:35:46markunamiconn: did you see
09:35:55scorchetierra: settings, debug, view I/O ports
09:35:55bEnNif i copy some1 elses wps and put in my own bmp it should work right
09:36:13tierrathought so, but it's not listed
09:36:46tierraI see plenty of GPIO*, GPO32, and DEV_EN#
09:36:48tierranothing else
09:37:10markunbEnN: why don't you stop the "some1"?
09:37:18*scorche wonders if amiconn meant to add it to the screen literally
09:37:27bEnNsorry, habit
09:38:41psycho_maniacbEnN: that would work.
09:38:44 Quit webguest44 ("CGI:IRC (Ping timeout)")
09:39:06tierrarockbox version reported is: r15632-071115 btw
09:40:01tierrascorche: I should update more often, looks like quite a bit has changed... heh
09:40:11*scorche nods
09:40:42bEnNdont know if this is the place to report it, but i mught have found a problem with the new nano release
09:41:00LinusNbEnN: shoot
09:41:46bEnNi was running loader2, with ipod linux and rockabox and when I ran rockbox the load bar would fill then nothing would happen
09:42:03bEnNit worked when I ran r14152
09:42:25scorchebEnN: does it work with the latest rockbox bootloader?
09:42:32bEnNim testing that now
09:42:42bEnNit might have just been my setup but im not sure'
09:42:51bEnNI had rockbox.ipod on my root and in the older
09:43:02bEnNmy loader.cfg pointed to the root one though
09:43:31scorchetierra: well, i need to go to bed...if you are going to stay up, amiconn should be around...if you need to change the source, i can make you some builds tomorrow
09:43:40tierrascorche: wow, the entire DB stuff is all new, and something I'd expect, and yeah, very awesome
09:43:44scorcheamiconn: tierra has a mini 1st gen for testing
09:44:00tierraalright, night
09:44:01scorcheyour previous build didnt even ahve the database? o_O
09:44:10tierrayeah, must have been pretty old
09:44:39scorcheit was likely just didnt know about it
09:44:50bEnNoh has scorche left?
09:45:16scorchei *think* tagcache was in before the ipods
09:45:41bEnNo scorche, im about to test it with the rockbox bootloader
09:46:01scorchebEnN: someone else can help you...i am moving to bed :)
09:46:07bEnNoh ok
09:48:29 Quit ddalton_ (Read error: 110 (Connection timed out))
09:48:59bEnNhow do I update rockbox, I copied the files but it still has the old ver
09:49:55psycho_maniacextract ALL of the to the root of your player
09:50:27scorchedelete the rockbox file you have in the root of the device
09:50:32scorche...but i am not here..
09:50:54 Join mrkiko [0] (n=mrkiko@
09:52:14psycho_maniacGodEater: that wps did work? did you play a song longer then 10 seconds?
09:53:41pixelmapsycho_maniac: because I saw this wrong information from you here in the logs and in the forums and by the way - you gave 2 different answers. The backdrop doesn't need to be in a specific bit depth... Rockbox is able to scale it down to what the display can handle (dithered if necessary)...
09:54:38psycho_maniacok. thanks pixelma. i got the wrong info.
09:54:39pixelmaI know there are places in the wiki that still suggest this
09:55:41bEnNis there an auto installer for rockbox nano?
09:56:13psycho_maniaci think i was refering to the album art .bmp
09:56:26pixelmabEnN: there is an installer which is explained in the manual
09:57:12bEnNok, just so people know, the new nano release doesnt work with loader 2 i dont think'
09:57:23bEnNits opk with rockbox boot loader though
09:58:11pixelmapsycho_maniac: the same applies there. Just out of interest (and curiousity how it looks) I experimented with album art on my (greyscale) M5 - it works with the prepared colour (usually 24-bit) cover.bmps
09:58:22psycho_maniacok now i keep getting a data abort when i load a theme with album art.
09:58:30 Quit bEnN ("ChatZilla [Firefox]")
09:58:48GodEaterpsycho_maniac: yes - it worked fine, other than me having no AA as I said
10:00:20psycho_maniacif you have an album art theme with cover.bmp over the specified demsions of the wps it will crash. EVEN with bmp-resize
10:00:59psycho_maniacAND i just realized how much i love 7738. i just ran the current build and it scrolls so slow lol
10:02:54 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
10:07:13psycho_maniacMay I have Wiki write permissions?
10:07:45LinusNof course
10:07:50LinusNwhat's your wikiname?
10:08:07LinusNspeaking of wiki, i really like this MarcGuay guy
10:08:24psycho_maniachaha yes did you read his wiki profile?
10:08:55 Part pondlife ("Gone")
10:09:03LinusNJerryLange now has edit powers
10:09:25psycho_maniacThank you.
10:10:47pixelmaspeaking of the wiki... someone knows why saratoga deleted the "USB otg" part on the NoDo?
10:11:18GodEaterhe decided he's going to implement it ? ;)
10:11:23 Part sidlet ("Ex-Chat")
10:11:49pixelmaGodEater: on all targets? ;)
10:12:00GodEaterthat would be challenging certainly!
10:12:19psycho_maniacis there a way to change a file on the wiki?
10:12:22GodEatermaybe he's making a little solder kit available for targets without the right hardware ?
10:13:17GodEaterpsycho_maniac: no - the accepted way to do that is to "hide" the original file, and upload another one
10:13:39pixelmathat would need a new page UsbOtgMod or something ;)
10:14:55LinusNnobody asks about otg on the archos targets anyway, so we can safely delete it :-)
10:15:14LinusNit was worse a few years ago
10:15:29GodEaterLinusN: you're right, that MarcGuay chap has done a lot of hard work on the wiki it seems :)
10:15:41LinusNi'm impressed and thankful
10:15:51GodEatercan we award him a gold star ?
10:15:57LinusNwe should :-)
10:16:17TheCollectorthe "unbrick" page for the sansa e200 needs some updating - if I make an account, will I be given write privledges?
10:16:48psycho_maniacso i would name it something different?
10:16:59amiconnmarkun: (1) I don't read the forums anymore, so I won't know any thread unless someone points me to it. But (2) read this morning's logs
10:17:15GodEaterpsycho_maniac: yes, that's the idea
10:17:44markunamiconn: (1) that's why I point you to posts from time to time (2) ok
10:17:45GodEaterTheCollector: yes, you'll get write privileges if you make an account
10:18:02GodEaterTheCollector: the only reason we make people ask is to stop spambots signing up and ruining the wiki
10:18:15GodEaterso unless you're spambot - you'll get permission
10:18:19GodEateryou're not a spambot are you ?
10:18:27*GodEater peers suspiciously at TheCollector
10:20:11hcsquick, the Voight-Kampff test!
10:20:18TheCollectorwell, I guess that all depends
10:20:36TheCollectorif spambots can install rockbox, then maybe
10:20:42 Join [potato] [0] (
10:20:43 Quit micols_ (Read error: 104 (Connection reset by peer))
10:21:18TheCollectorok, I just got my account confirmed
10:22:51LinusNTheCollector: done
10:27:35psycho_maniacGodEater: can you test another theme?
10:27:57GodEaterpsycho_maniac: not right now - bit tied up sorry
10:29:29psycho_maniacanybody else want to test a theme on the ipod video?
10:30:19 Quit [potato] ()
10:32:13LinusNpsycho_maniac: what's special about it?
10:33:14psycho_maniaci fixed the album art but for some reason the text does not update and i want to know if its just my player.
10:34:06amiconntierra ?
10:34:55 Join PaulJam [0] (
10:38:18hcsok, I presume that it isn't my fault that the recent build is broked, despite is being triggered by my commit
10:38:56hcsconsidering that I only touched rockblox.c
10:39:56amiconnhcs: ?
10:40:42hcsamiconn: I just did my first commit in a while and I'm noticing the broken M-Robe build
10:40:58hcsjust a bad build server?
10:41:39 Join TMM [0] (
10:44:24amiconnDo you read the dev ml? Build server owners need to update arm-elf-gcc with the multilibs patch
10:44:36amiconnUnfortunately many of them did not do so yet
10:45:45hcsyeah, I had read that, but didn't recall all the details
10:46:02 Join Clam [0] (
10:52:27 Join inteliwasp [0] (
10:54:04 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
10:55:06inteliwaspi have a 4g ipod that is giving me errors when i try to boot it, it keeps giving me the sad ipod icon or some folder icon
10:59:01homieloweintelliwasp: Is this related with a rockbox install ?
11:00:34inteliwaspyes, it is a daily build from about a month ago
11:00:41 Join weezerle [0] (
11:01:12inteliwaspthe last time i tried it, it got to rb but it said that the it could not boot
11:01:25homieloweSad icon could mean a bad HD possibly, if its the folder icon , you might be able to recover it with disk mode.
11:01:57psycho_maniacyay blackglass and cobalt themes for the ipod video now work without patches. !!
11:02:01inteliwaspand i just got a new battery
11:02:32inteliwaspi wonder if one of the wires got loose...
11:08:17***Saving seen data "./dancer.seen"
11:09:39mrkikoHi all! I' seeing big rockbox changes!
11:10:26 Part Clam
11:13:00 Quit rasher ("need more rams")
11:13:33mrkikobitwise.c:177: internal compiler error: in cgraph_expand_function, at cgraphunit
11:14:31mrkikogcc says to submit a full bug-report... but :) I' m not alive again
11:19:40 Quit inteliwasp ("Leaving")
11:25:38psycho_maniacwhat does "quietsave" mean?
11:27:55GodEaterI'm trying to rework FS7738 to remove the settings entries for it. Where is the best place to store the values for wheel_accel_start and wheel_accel once they're no longer required in the menu system ?
11:28:55GodEatersilly me - ignore that question
11:31:27 Join mf0102 [0] (n=michi@
11:38:27GodEaterFS7738 has now had it's settings menu stuff removed. Is it good enough for a commit now ?
11:38:36GodEaterplease test it and let me know
11:40:18*GodEater crosses his fingers and hopes
11:44:06 Join Onj [0] (
11:44:53amiconnIt's still button_clickwheel only.#
11:45:53amiconnAnd I wonder why the 200 does quadratic acceleration, while the ipods use power of 4
11:46:05OnjHello. Could I ask a question about the latest SVN? I seem to have a problem building a voice for my target at the moment.
11:46:11GodEaterno clue - you'd have to ask jhMikes / buschell
11:46:25GodEaterOnj: what problem would that be ?
11:47:52 Quit psycho_maniac (" HydraIRC -> <- Nine out of ten l33t h4x0rz prefer it")
11:47:52OnjWEll two days ago I built voices for Iriver H120 which worked fine, but Ipod Video 30 and 80 did not. Then, it was suggested to me that I delete the entire rockbox directory and get SVN to redownload the entire thing, which allowed me to build for Ipod30 and 80, but now my own target doesn't work. How silly. However, I can use one of the voices I built two days ago and not yesterday, with today's build, and that works. My system menu is extremely sc
11:47:52Onj like, copy, cut, move, a/b repeat, and so on.
11:49:32linuxstbGodEater: I've only had a quick look, but maybe remove "DEFAULT_" from the names of the #defines, as they can't be changed any more. Also, I don't think they need to be passed as parameters to the button_apply_acceleration() function any more - that code can just use the #defines. It follows from this that the "factor" variable in that function can also become a #define
11:50:07GodEaterlinuxstb: hmm - ok
11:52:22GodEaterlinuxstb: do you mean move whatever #define I turn factor into out into the same files where the ACCEL_START #defines are currently ?
11:52:28 Join barrywardell [0] (n=barrywar@
11:53:02OnjAlso, when I did build the broken voices for Ipod targets, the system menu was missing completely. I've never seen anything like it.
11:55:44linuxstbGodEater: IIUC, factor is based on DEFAULT_WHEEL_ACCEL_START, so you can #define WHEEL_ACCEL_FACTOR in button.c, based on the WHEEL_ACCEL_START which is set in the individual target config files.
12:08:16 Quit scorche (Read error: 110 (Connection timed out))
12:10:27 Join moos [0] (
12:12:03GodEaterlinuxstb: ok - all changed
12:12:58 Join lee-qid [0] (
12:16:30 Join desowin [0] (
12:18:18*GodEater pokes other ipod owners to get them to try it
12:19:50OnjI wonder if the menu disappeared just for the voice, or on-screen as well?
12:22:28GodEaterOnj, have you tried a daily build and it's matching voice file from the rockbox site ?
12:23:10OnjSure. but I never liked festival anyway, and I used to build them under windows. so I got Cygwin and I can build them just fine, no compile errors, none of that, but for some reason, something is wrong.
12:23:30GodEaterso the one from our site works ok ?
12:23:32 Join Thundercloud [0] (
12:23:33GodEaterand doesn't miss menus out ?
12:24:30OnjI haven't tried this morning, but yesterday everything seemed fine. It's now my H120 that doesn't work, but my friend's Ipod does, with the same parameters when I build a voice. Very strange.
12:24:57OnjMaybe i will have to delete the rockbox dir yet again and run the SVN co thing.
12:24:58GodEaterwell if you try with today's, then we can rule out whether it's your build environment, or something we've broken
12:25:12 Quit Soap ()
12:27:37 Nick fxb__ is now known as fxb (
12:28:25 Join rasher [0] (n=rasher@rockbox/developer/rasher)
12:28:31pixelmaOnj: It's now my H120 that doesn't work, but my friend's Ipod does, with the same parameters when I build a voice. <- What do you mean with the "same parameters"? How do you build them (in different build directories etc.)?
12:28:57OnjNo same build directories, my own personal lame choices though.
12:30:18Onjok the festival voice from the site works. I literally just got through building my own voice, so now I test that.
12:30:52 Quit barrywardell ()
12:31:20pixelmastill unsure if I understand you correctly... you are aware that voice files are target specific? (just making sure)
12:31:37 Join Bagder_ [0] (
12:32:19OnjWow. That really does not work. The first item in the system menu is blank. then move. show indacies. Interesting choices, but definitely incorrect. I always run svn update before I build, is there something else I should do as well? I feel quite stupid for this.
12:32:49 Join barrywardell [0] (n=barrywar@
12:33:12OnjYeah. I know all that. I run the ../tools/configure, choose my target, etc etc. Build the voice, run it, and it fails. I was the unofficial rockbox voice builder for windows until August fifth this year, when it switched to target-specific voices.
12:35:08linuxstbAre you sure you're updating the build on your player correctly? Maybe that part isn't working, so you're using a new voice file with an old build.
12:35:24 Quit Nico_P (Remote closed the connection)
12:35:34OnjAbsolutely. I just downloaded it. The build does say today's date in the system info, so I know that's correct also.
12:36:57 Join james_malone [0] (n=James@
12:38:48pixelmaso are you building the voice files in the same directory and just "switch" by rerunning configure again?
12:39:27Onjyeah, i cut and paste them to another folder after build. I've had this working before. I built for h10 20 gb and those voices work, the Ipod voices work with this current svn, but H120 no go.
12:39:47OnjTwo days ago it was reversed.
12:41:04 Quit james_malone (Client Quit)
12:43:26pixelmaand do you also run "make clean" before configuring again? I once had a problem when I chose another voice in the tts options and a make clean soled it... if it's not that than I'm out of ideas unfortunately
12:46:17OnjOh, no. I didn't know to do that. I will try that. Thanks very much. Building again now.
12:46:36linuxstbOnj: I don't know about voice, but when building Rockbox itself, it should be done in separate build directories for different targets. IIUC, you can have a common "pool" folder somewhere that caches voice clips.
12:47:19OnjAah. Well. er, right then. I wasn't aware of this
12:47:55 Quit Bagder (Read error: 110 (Connection timed out))
12:47:58pixelmathat could probably make it easier :)
12:50:06 Quit mrkiko (Read error: 104 (Connection reset by peer))
12:50:20OnjWhat do you know! It worked. So different build folders for different targets. I'm ea windows user, thus stupid when it comes to Linux. I figured that the configure would clean up before and after itself, but I guess not.
12:50:31*Onj looks stupid
12:51:22GodEaternot really
12:52:18 Join mrkiko [0] (
12:54:49 Join Casainho [0] (
12:55:45Casainhohello people! :-)
12:56:28linuxstbOnj: All configure does is create the "Makefile" and "autoconf.h" files in the build directory - nothing else.
12:56:58Casainhobarrywardell: hello!
12:57:19Casainhobarrywardell: can you please tell me how is sansa USB developing?
12:57:37barrywardellall I know is what Zagor has posted to the tracker
12:57:43OnjI don't have to run a configure every time though now. I can make one directory for each target, configure it once and just run make voice when I need it. That saves a lot of time. Thanks.
12:58:46ZagorCasainho: the last two weeks nothing much has happened. I've been stuck on a hw/driver problem I don't know how to solve. some new ideas came up yesterday that I'll try as soon as I find the time.
12:59:20CasainhoZagor: sorry, I can't help - I am asking because I am just curious...
13:00:07CasainhoZagor: USB on Sansa is OTG? - what will It be able to do? - justo to connect to computer?
13:00:26Zagormy code is device mode only, i.e. connect to computer
13:00:46Zagorotg can potentially connect two sansas together and swap files
13:01:00Zagorbut that is quite a bit more complex than what I'm doing
13:01:34Zagorotg is conceptually host and device capability on the same port, with auto-negotiating
13:02:03CasainhoOTG does depends on hardware?
13:02:28Casainhoso, Sansa have USB OTG(?) :-) :-)
13:02:57 Join Foofty [0] (i=d4dbbf87@gateway/web/cgi-irc/
13:02:58 Join webguest62 [0] (i=3b6412ee@gateway/web/cgi-irc/
13:03:01Zagoryes it does
13:03:06 Join MethoS- [0] (
13:03:19Casainhookok - good luck with USB :-) ;-)
13:03:56CasainhoZagor, where can I find in TWiki info about PortalPlayer? - I wouldl ike to know waht DAC It have...
13:03:58webguest62i want to start making a new flavor of rockbox
13:04:08webguest62were do i start
13:04:19Zagorwebguest62: you mean a new port?
13:04:43Casainhofrom drivers in source, there are a few drivers for DAC but I don't know what is for Sansa... - Is that information on TWiki?
13:04:43webguest62based of the original
13:04:58barrywardellZagor: is it at the stage where could add a charging mode for the sansa (ie. request 500mA, no extra endpoints)?
13:05:11Zagorbarrywardell: yes we could
13:05:28Fooftywebguest: you want to modify rockbox or port it to new hardware?
13:06:21webguest62just make a version that works a little better for a Ipod Nano 1st gen
13:07:04barrywardellZagor: nice. can your code easily deal with selecting between msc and charging mode?
13:08:02Zagorbarrywardell: sure, we could add a switch to inhibit sending endpoint descriptors
13:08:19***Saving seen data "./dancer.seen"
13:08:26 Quit mrkiko (Remote closed the connection)
13:08:55markunIs there any use for a 9+18x18 unicode font?
13:10:42linuxstbwebguest62: What kind of changes do you have in mind?
13:11:40ryanakcaWhy is it that every song gets included into the Database 4 times? Any way to correct this? iPod Nano with Tuesday's daily build.
13:12:13ryanakcaI've tried removing the database files and recreating the database. I've also tried looking for any duplicates of the songs on the device, none
13:13:36linuxstbryanakca: It could be a side-effect of the problems some Nanos are experiencing. Do you have any other problems (e.g. freezing during playback)? Also, have you checked the filesystem for errors (I can't remember if I asked that last time you were here...)
13:13:43 Join Siku [0] (
13:13:50ryanakcalike, say I have ABC and DFG on the iPod, the list goes 'ABC, ABC, ABC, ABC, DFG, DFG, DFG, DFG'. When I removed all music from my iPod with amaroK and transfered one song to it. And, before, when I had a pile of songs on there, each and every song was listed 4 times in the database browser/list.
13:14:38ryanakcalinuxstb: no... fsck.vfat ?
13:15:06linuxstbI guess so. Or you could simply try a reformat, just to be sure.
13:15:07ryanakca(btw, dunno if this changes anything, but it's a first generation Nano)
13:15:25linuxstbWe assumed so - that's the only Nano Rockbox works on.
13:15:29ryanakcalinuxstb: ok... and then run upstairs to a windows comp and do a restore of the firmware?
13:15:32ryanakcaah :)
13:15:49linuxstbNo, no need to restore the firmware. Just format /dev/sdX2 (the firmware is on /dev/sdX1)
13:16:03webguest62linuxstb: i just want to make a interface that my mom can understand
13:16:28linuxstbwebguest62: How would you change it? Remove options?
13:16:28 Quit Foofty ("CGI:IRC (EOF)")
13:16:46*ryanakca will try fsck.vfat, and if that doesn't work, I'll reformat it tonight when I get back from school
13:17:37linuxstbI'm assuming you're not trusting Amarok when it tells you it's deleted all files? i.e. you're looking inside the iPod_Control directory on your Nano, and making sure no files are still there?
13:17:54webguest62first of all i'd like to change the main menu list into a list of icons
13:18:25webguest62nice pretty pictures
13:19:19GodEaterthat's a pretty big change =/
13:19:59ryanakcalinuxstb: yes... I ran... *checks* 'md5sum -b `du -a /media/IPOD/ | grep -i kpod | sed -e 's@.*/media@/media@g' | sort` >> music-md5sums' and looked threw the list for matching sums, and I also ran 'fdupes -r /media/IPOD'.
13:20:17ryanakcaand as for that, yes
13:20:21Zagorwebguest62: you do know you can add icons to the top menu?
13:20:38ryanakcawhen there aren't any files, there aren't any, and when there are some, there aren't any duplicates.
13:21:23webguest62yes but i want a easyer to see interface
13:21:26 Join ddalton_ [0] (
13:23:11Zagorwebguest62: easier than a list of icons?
13:23:30ryanakcajust make the font bigger
13:23:45linuxstbwebguest62: Out of curiousity, has your mom tried Rockbox?
13:23:52webguest62its not the look i want
13:23:59ryanakcabut she might not mind it?
13:24:18*ryanakca quite likes it actually :D
13:24:20webguest62she doesn't like it
13:24:21Zagorwebguest62: you probably want to write down your ideas first, and make mockup screens. it will help you work faster than starting pulling at the code.
13:24:49 Quit desowin ("use linux")
13:24:52ryanakcalinuxstb: anyways, thanks, I'll format it tonight and start from scratch. Thanks ;)
13:26:09webguest62i found the new build to be a little unfinished
13:26:18webguest62it looks good but
13:26:24linuxstbryanakca: Is there a reason you use Amarok? You know you can just copy the files to your Nano's drive (with cp) and Rockbox will play them?
13:26:45webguest62some of the fuctionality is a bit glitchy
13:26:55 Join n1s [0] (
13:27:25Zagorwebguest62: we most certainly welcome additional help
13:27:41Zagordon't assume nobody else wants your modifications
13:27:42linuxstbAnything in particular? If you can't fix them, bug reports are also very welcome.
13:29:06webguest62i'm just wondering if theres a fix for the USB connativity
13:29:18 Join spiorf [0] (
13:30:06LinusNwebguest62: no, because it requires a lot more than a "fix"
13:30:09 Join Nico_P [0] (
13:30:35webguest62so the problem is known?
13:31:00 Join jpt9 [0] (
13:31:35LinusNwebguest62: ehum, wait a sec
13:31:59LinusNwhat exactly is the problem?
13:35:04 Join desowin [0] (
13:37:11 Quit jhMikeS (Nick collision from services.)
13:37:17 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
13:41:19webguest62when i link up my ipod nano to my computer it starts up and freazes
13:41:34preglowdoesn't happen with mine
13:42:07webguest62only if i have it turned off when i link it up
13:42:16preglownot then either
13:42:32webguest62the old firmware didn't do it
13:42:58webguest62its only the past two updates
13:42:59 Join MethoS-- [0] (
13:43:13linuxstbI don't get it on my ipod either - but I know others have reported it (both pixelma and amiconn I think). IIRC, they've said it happens with some builds, but not others, seemingly at random.
13:43:20 Quit jhMikeS (Nick collision from services.)
13:43:25preglowwe need proper usb :/
13:43:26 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
13:43:36*LinusN looks at Zagor
13:43:42 Quit barrywardell ()
13:44:00webguest62USE 2.0??
13:44:16linuxstb2.0 what?
13:44:21webguest62sometime i've only got 1.1
13:44:24preglowwebguest62: no, our own usb driver
13:44:37 Quit ddalton_ ("I was using BOFHNet IRC version 1.2 by fmillion - get your copy today from !")
13:46:22linuxstbwebguest62: Which Rockbox build are you currently using? Maybe preglow could test the same build on his ipod, to see if he also gets the usb freeze?
13:50:08jhMikeSpreglow: I'm was trying to build with the last patch applied and building the encoder fails because it can't find inttypes.h (I don't know about other files yet). SVN builds fine.
13:51:51amiconnlinuxstb: I get this freeze rather often on all PP502x targets, but the probability depends on the actual target
13:52:11amiconnOn mini g2 it happens *very* often, on video not that often
13:52:26amiconnAlso not often on H10
13:55:23 Quit jpt9 (Read error: 110 (Connection timed out))
13:56:25 Join DrMoos [0] (
13:57:29preglowjhMikeS: is that so
13:57:58preglowjhMikeS: what, don't you have inttypes.h, then?
13:58:01 Quit moos (Read error: 104 (Connection reset by peer))
13:58:18pixelmalinuxstb: it freezes when I try connecting from a running Rockbox (not when it was turned off) - often on my c200, first experienced on amiconn's Mini. With certain builds, they can even be made with the same computer
13:59:00preglowpixelma: freezes _in_ rockbox, you mean?
13:59:13pixelmathen it just "hangs" in Rockbox's USB screen and doesn't reboot to the OF
13:59:17preglowyeah, i've had that plenty of times
14:01:02webguest62not here
14:01:56 Quit MethoS-- (Connection timed out)
14:02:19 Nick DrMoos is now known as moos (
14:02:39 Quit lee-qid ("aufwiederbyebientotsayonara")
14:02:53 Quit stewball (Read error: 104 (Connection reset by peer))
14:03:31jhMikeSpreglow: I have two places but not on some global include path apparently. It's a fresh debian image (trying to install espeak messes things up, will try others).
14:04:10preglowjhMikeS: well, i don't know what to do about that, you should have inttypes
14:04:16 Quit Casainho (Remote closed the connection)
14:04:36 Quit MethoS- (Read error: 110 (Connection timed out))
14:06:16jhMikeSI do, I the #includes. Just a path thing I think. has anyone else tried this on debian? I invite them to.
14:06:54jhMikeSNot the #includes, in the "/include" directory on the image.
14:07:22preglowwell, i developed it on ubuntu, which basically is debian in this regard
14:07:37 Nick fxb is now known as fxb__ (
14:08:02LinusNjhMikeS: which patch?
14:08:28 Quit webguest62 ("CGI:IRC")
14:08:52linuxstbjhMikeS: Try "apt-get install libc6-dev" - that should give you /usr/include/inttypes.h
14:09:28linuxstbjhMikeS: Or am I misunderstanding, and the file's there?
14:09:43preglowi haven't looked at .talk file generation at all
14:09:52preglowis there anything else i should have a look at?
14:10:06preglowi don't use voice, remember
14:10:10jhMikeSI have it...if I compile the binary directly with gcc from the command line, it finds it
14:10:40jhMikeSwhere it doesn't find is when compiling libspeex
14:10:51OnjOK. All targets I've built with today using separate build directories work. So I'd like to say thanks to everyone who's helped me out today, it's much appreciated.
14:12:40*jhMikeS added an #include <inttypes.h> to the #includes in rbspeexenc.c to check and ran gcc on that
14:15:14preglowi wish people would stop writing shell scripts!"
14:15:15linuxstbspeex/speex_config_types.h is including "inttypes.h"
14:16:31LinusNso the include path is wrong then?
14:17:15LinusNi believe it could be <inttypes.h> without causing any problems
14:17:34jmspeexlinuxstb: where did you see that? You must be using a modified version of Speex.
14:17:52preglowhrm, how should i handle the entire swcodec/hwcodec in i can't detect what kind of target the user has, so i have to ask about something. should i ask whether to use rbspeexenc or lame, or should i ask if the user has a hwcodec target or not?
14:17:57jhMikeSI think it's just an include path problem yeah.
14:18:46 Quit Onj ("I can get up and I can get out, so I did.")
14:19:05LinusNpreglow: i think a hwcodec/swcodec question could do, or rather ask if the user has an Archos Player/Recorder/Ondio
14:19:25LinusNthe user isn't likely to know what "hwcodec" is
14:20:05preglowyeah, exactly
14:20:21preglowplayer/recorder/ondio is the full range of archos stuff we support?
14:20:43LinusNbut the player is sometimes called "studio"
14:21:04 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
14:21:05preglowall add that too, then
14:21:07LinusNand the recorder comes in 3 flavours, "Recorder", "V2 Recorder" and "Fm Recorder"
14:21:39LinusNbut i guess "studio/player/recorder/ondio" should suffice
14:22:00preglowyeah, me too
14:22:23preglowbtw, what do you think would be a good way of handling this change wrt the users?
14:22:31preglowi guess some kind of advance notice would be helpful
14:22:39preglowespecially in case of problems, and such
14:22:58LinusNperhaps, but i don't think many will notice anyway until after the fact
14:23:23preglowif everything goes smoothly, then no
14:23:36preglowbut i guess i should do an ml post telling that voice file format has changed
14:24:00LinusNoh yes, we should of course inform them about the change
14:24:10LinusNi suddenly had an interesting idea
14:24:40preglowis it still there? :>
14:26:08markunwe have a new unicode font!
14:26:20preglowroyal plural? :V
14:26:44LinusNthe old voice file could contain a clip that says "pleas upgrade the voice file" that is played when the voice file is too old
14:27:14LinusNand even the speex version of rockbox should be ably to play back that single clip somehow
14:27:21 Join linuxstb__ [0] (
14:27:22markunLinusN: is it possible yet to correct svn commit messages?
14:27:34LinusNdon't think so
14:27:40 Join Calcipher [0] (
14:27:43markunwell, doesn't matter so much
14:27:59preglowno? i'm pretty sure i've done it once
14:28:02preglowbut perhaps that was with cvs
14:28:23jhMikeScvs , yes. you need the svn hook installed for that.
14:28:53 Join jpt9 [0] (
14:29:05 Quit jpt9 (Client Quit)
14:29:28 Join jpt9 [0] (
14:29:34 Quit linuxstb (Nick collision from services.)
14:29:39 Quit linuxstb_ (Nick collision from services.)
14:29:43 Nick linuxstb__ is now known as linuxstb (
14:29:44jpt9how often is rockbox supposed to crash?
14:29:58jpt9(on a refurb Sansa e250R)
14:30:29jpt9i've had it freeze up completely a few times in the last day or so.
14:30:39jpt9(I just got the Sansa from woot)
14:30:40jhMikeSoften, it's a feature
14:30:57jpt9unfortunately, I haven't been writing down the error messages; I'll start doing that.
14:31:10jpt9jhMikeS: so there isn't anything wrong with the Sansa?
14:31:31krazykithe was kidding...
14:31:44jpt9i knew he was kidding...
14:31:46jhMikeSI just got one and I suppose I'll see myself
14:32:24jpt9if I use the "Complete Installation" button in Rockbox Utility, will it leave all my settings/themes/configuration/etc intact?
14:32:45linuxstbpreglow: Why not simply make the script ask for the file format (i.e. MP3 or Speex), and add a note saying that MP3 is for Archos devices, Speex is for all others?
14:32:47LinusNjpt9: yes
14:33:03jpt9i assume it gets the latest version?
14:33:43markunyay, my new font is even bigger than unifont (1.8MB)
14:33:55jpt9is there any way to permanently disable the "refresh database" crap from the OF?
14:33:59preglowLinusN: because the format should be irrelevant and the info about archos should be there anyway
14:34:00jpt9it's *really* annoying.
14:34:08preglowlinuxstb: that was to you...
14:34:12GodEaterjpt9: not until we have our own usb stack
14:34:48linuxstbpreglow: IMO it's very relevant - asking which device the user has seems to be hiding what's going on behind the scenes, which I don't like.
14:35:22preglowlinuxstb: i don't why it matters, for archos you have one choice: mp3, for swcodec you have once choice: speex, and you will never, ever need to know what format you are using
14:35:33preglowbut sure, i don't care, i can hang on two format names
14:35:43jhMikeSobviously this encoder building should run on debian. where might I look for the best place to deal with include details?
14:35:55*amiconn wonders what is :>
14:36:56*amiconn uses voiceBox for generating .talk clips
14:37:01*jhMikeS 's problem is building the speex encoder itself in tools
14:37:10jpt9oh yeah...
14:37:24linuxstbjhMikeS: Did you change the "inttypes.h to <inttypes.h> ?
14:37:43jpt9sometimes Rockbox won't let me shut off my Sansa if the USB cable is plugged it. (it's attached to one of those USB power adapters).
14:37:53jhMikeSlinuxstb: in the speex files? no, will try that.
14:38:00preglowamiconn: what/where is that?
14:38:12preglowamiconn: i don't use this shit myself so i have no idea what i have to fix....
14:38:15linuxstbjhMikeS: Yes, see the logs from 23 minutes ago...
14:38:41n1sjpt9: the sansa can not be shut down while connected to external power (afaik)
14:38:50jpt9does anyone know how to get rid of the Windows "Your USB device you attached is almost full. We think it might be new from the manufacturer, so we'll offer you the option to delete all the data on it. While stealing the focus from where you're currently typing. Have a nice day." thing?
14:41:53krazykitjpt9, use a better operating system?
14:41:59jhMikeSbah, still not working
14:42:05jpt9i will...
14:42:11jpt9once I get a decent disk defragmenter...
14:42:16linuxstbjhMikeS: Is this patch on flyspray? I can try on Debian.
14:42:33preglowlinuxstb: the patch in question is mine, you had a look at it yesterday
14:42:34jpt9Diskeeper Home (I'm at RPI; it came with the laptop) can't shove all the files to one end of the disk.
14:43:00jhMikeSlinuxstb: not one with the encoder, can make a patch though
14:44:49jpt9also, is there any way you could have the wheel on the Sansa move the keyboard cursor left/right (or better yet, if you're not using Rec for anything, have it swap the X/Y key controls)?
14:44:55jhMikeSyep, definitely compiles libspeex in a normal build context. I verified to make sure it's not just me.
14:45:08krazykitjpt9, not without compiling your own build
14:45:16linuxstbpreglow: So I should just apply that patch and type "make rbspeexenc" in tools?
14:45:40preglowlinuxstb: just do make tools
14:45:45 Join thegeek_ [0] (
14:45:49linuxstbOuch, doing that attempts to do "mkdir /libspeex"...
14:45:53amiconnpreglow: It's not in SVN, but in the wiki:
14:46:04amiconnIt's a vbscript + html application
14:46:24preglowlinuxstb: feel free to fix, i absolutely hate hacking this shit
14:46:33amiconnI don't think fixing it will be a problem. We just need to provide an encoder executable in the package
14:46:55preglowi'll see what i can do later
14:47:10LinusNpreglow: tools/configure looks for tools/speexenc when configuring, but it isn't built until after the configuration is done
14:47:11preglowi'll probably boot windows once during the day
14:47:17preglowLinusN: i know, fixed locally
14:47:58*amiconn boots windows approximately once a week, or even less often
14:48:10amiconnI just let it run...
14:48:13LinusNpreglow: rbspeexenc.c:261: error: redefinition of 'get_long_le'
14:48:48preglowi must be accidentally including some rockbox stuff, then, i assume the system doesn't have that
14:49:37LinusNwow, looks like the source is included twice
14:49:54linuxstbjhMikeS: It compiles fine for me - I did "mkdir build-voice ; cd build-voice; (ran configure for voice); make tools" and it worked.
14:50:06LinusNmust be because i tried to apply both your and jhmikes patch
14:50:20LinusNnow it compiles fine
14:50:31preglowlinuxstb: why the hell did it try to mkdir /libspeex ?
14:50:47linuxstbpreglow: Because TOOLSDIR isn't defined if you type "make" manually inside the tools directory.
14:50:58preglowlinuxstb: but you're never suppĂosed to do that, are you?
14:51:17jhMikeSlinuxstb: the image i'm running is straight off the wiki link...0 modifications
14:51:47linuxstbIt works for the other tools... And the user might want to build rbspeexenc (e.g. for talk clips) without doing a configure
14:51:52LinusNpreglow: would your patch break anything if we commit it now?
14:52:00preglowLinusN: yes, voice building
14:52:01linuxstbjhMikeS: I'm using a "real" Debian install.
14:52:07LinusNpreglow: why?
14:52:15preglowLinusN: because i removed the encoder option from configure
14:52:17*jhMikeS is using VMWare
14:52:29preglowLinusN: so you/bagder have to fix that on the web end of things
14:52:47LinusNpreglow: you removed it?
14:52:47linuxstbpreglow: I'm happy to try and fix it though - leave it with me...
14:52:54preglowlinuxstb: aye
14:53:00preglowLinusN: yes i did
14:53:05preglowLinusN: there is no real choice anyway
14:53:26preglowif $swcodec is set, i set encoder to lame, if not, rbspeexenc
14:53:27LinusNbut we must still support mp3
14:53:37preglowthe other way around...
14:54:10*preglow glues another "read _before_ pressing enter" note to his display
14:55:09LinusNpreglow: can you commit everything but the configure cgange?
14:55:13LinusNchange even
14:55:26preglowLinusN: i still haven't fixed .talk clip generation either
14:55:28preglownot even tested it
14:55:44linuxstbpreglow: I think it may be cleaner if we create a "rbspeexenc" directory in tools/, and put rbspeexenc.c in there, along with a new Makefile. Typing "make" in that directory (or "make -C ../tools/rbspeexenc") should then be capable of building the encoder.
14:55:45preglowLinusN: if i don't commit the configure change, no one will be able to make voice files
14:55:48LinusNit woudl still feel better if the speex encoder was commited
14:55:54preglowLinusN: go ahead...
14:55:57preglowlinuxstb: go ahead...
14:56:02preglowyou guys need new nicks!!!
14:56:35 Nick linuxstb is now known as DaveC (
14:56:48*DaveC sees this is registered though...
14:57:01 Nick DaveC is now known as linuxstb (
14:57:04LinusNyes, i regged it to stop you from changing your nick
14:57:05preglowi'll just make my nick completer choose randomly, will probably give better results
14:57:36 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
14:57:38preglowLinusN: ok, so what would you rather happen now?
14:58:10LinusNi'd want the speex encoder committed, but still build voices the old way
14:58:35preglowokiedoke, that we can do as soon as linuxstb fixes the stuff he wanted fixed
14:59:17linuxstbAre nicks case-sensitive?
14:59:20LinusNi'm a little concerned when the patches become so big
14:59:37LinusNi like incremental change
14:59:39*linuxstb wonders why xchat's nick-completion ignores case
15:00:24preglowLinusN: sure, no point in not doing it incrementally when possible either
15:01:19LinusNthat way we can prepare for speex voice building in a nicer way
15:03:30 Quit thegeek (Read error: 110 (Connection timed out))
15:04:11LinusNjhMikeS: could you perhaps break out the type conversions in your speex patch and commit them separately, like right now?
15:05:02LinusNthat way it is easier to review the speex specific changes in the patch
15:05:43jhMikeSLinusN: I suppose so
15:06:01LinusNwould be great
15:06:23GodEaterxchat's nick completion can be set to complete nearest match of whoever spoke last
15:06:38GodEaterand useful :)
15:06:57markunLinusN: btw, when are we switching over to git? :)
15:08:14*GodEater waves his "commit 7738" placard
15:08:20***Saving seen data "./dancer.seen"
15:09:14preglowwhat is this for?
15:10:40preglowok, currently both and just hard code both encoder and tts engine in variables, can i continue doing this, pretty please? :)
15:10:54preglowit seems they're not made to be user friendly anyway
15:12:55 Quit Febs (Read error: 110 (Connection timed out))
15:13:11LinusNGodEater: look my latest comment in 7738
15:13:19preglow$0 in a shell script means argv[0], yea?
15:13:30preglowlinuxstb: where will rbspeexenc end up? in tools/ or tools/rbspeexenc?
15:14:04GodEaterLinusN: it varies by target
15:14:17LinusNGodEater: yes, but not in runtime
15:14:28preglowthat would be exciting though, wouldn't it
15:14:33LinusNit should be handled by the preprocessor
15:15:15GodEatererm - ok - I'll look into it again
15:16:26LinusNi'm having a look now
15:16:43GodEaterthe default case can go completely too I guess
15:18:26GodEaterah - no it can't - that's the one the e200 uses
15:19:07LinusNthat's WHEEL_ACCELERATION == 1, right?
15:19:44*GodEater still doesn't understand most of the maths in the patch - that was all jhMikeS and buschell's hard work
15:22:09preglowanother thing we should agree on: -q parameter to encode with
15:22:35LinusNpick one :-)
15:23:48preglow-q 4?
15:23:53preglowamiconn: didn't you have some suggestion for -q ?
15:24:44preglowi guess i'll just tune that later
15:25:05preglowbut default for -c (encoder complexity, more quality at same bitrate, but slower encode) can be set to max, yes?
15:25:16preglowit's not excessively slow
15:25:33GodEaterI could have sworn one of the cases came back 2 as well - but apparently not
15:26:39jhMikeSLinusN: separate commit done
15:26:53LinusNjhMikeS: lovely
15:27:32preglowjhMikeS: i'm commiting some smaller stuff contained in my patch to svn now, just so you don't get any trouble
15:29:03 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
15:31:57 Join Frazz [0] (
15:32:15 Quit Frazz (Client Quit)
15:32:28linuxstbpreglow: In tools/rbspeexenc/
15:33:29preglowjust spam me the makefile when you're done
15:33:38preglowhow do i remove something i have svn added but not commited, again?
15:34:25LinusNsvn revert <file>?
15:34:39LinusNjust a guess
15:34:40amiconnpreglow: -q 4 or even -q 3
15:34:41jhMikeSthat reverts a file
15:34:45jhMikeSsvn rm
15:35:00GodEaterLinusN: it looks like the check for WHEEL_ACCELERATION==2 is not required
15:35:01jhMikeSsave it if you want it though
15:35:07GodEaterI can't find a config which uses it
15:35:09pondlifeLinusN: Any chance your menu.c fix has resolved ?
15:35:10 Join farhank [0] (n=furk@
15:35:13 Join Arathis [0] (
15:35:34LinusNGodEater: let's keep it anyway, one day we might find a target that benefits from it
15:35:53GodEaterI could have sworn the Nano was set up to use 2
15:35:57LinusNpondlife: nope :-(
15:35:57GodEaterI must have imagined it
15:36:33LinusNpondlife: my fix only appiled to the padding for the scrolling demo text
15:36:38pondlifeAh, ok
15:37:01farhankDoes anyone know when the wma patch for fastforwarding and rewinding is going to be included so I can generate a checked out source and update?
15:37:18LinusNfarhank: which patch #?
15:37:30farhanki'm not sure, someone was talking about it in here yesterday.
15:38:15farhankI just know that they have it, and I wondered when it was going to be included, as i'm leaving for london tomorrow and will be stuck with some not foastforwarding wmas if I don't convert them.
15:38:24LinusNfarhank: i found it, but it looks like it isn't finished
15:39:09pondlifeGodEater: Sorry if I'm being obtuse and pestery, but could you just recheck FS #7307. Verify that keyclick_3.patch definitely causes the beepfromhell, and that it doesn't happen with keyclick_4.patch. I find it hard to believe, sadly.
15:39:33pondlifeApparently Sansa still suffers with _4...
15:39:50GodEaterum - ok
15:40:25LinusNjhMikeS: time to resync your speex patch :-)
15:41:14farhankseeking is broken in wma, but is that part of why when you stop playing a file and go back to it later it doesn't bookmark your position?
15:41:21pondlifeGodEater: I want to be sure before I start hacking away madly... it seems most likely that you were lucky yesterday.
15:41:57jhMikeSLinusN: deed
15:42:19 Quit jumijoze ("Some say the world will end in fire; some say in segfaults.")
15:42:19LinusNfarhank: i believe so
15:42:45*LinusN is excited about speex voice
15:42:50preglowjhMikeS: i'll resync my part as soon as linuxstb sends me the makefile stuff
15:43:23farhankspeex voice? it isn't going to generate mp3 files?
15:43:29linuxstbpreglow: I'm almost there - the problem now is an undefined reference to le2h32() in speex_header.c
15:43:38preglowlinuxstb: btw, latest builld seemed to add about 25kb to arm builds, will be less for coldfire
15:43:42preglowi don't think that's too bad
15:43:48GodEaterpondlife: building with v3 at the moment
15:43:50preglowlinuxstb: why didn't i get that?
15:44:07 Quit farhank ()
15:44:08pondlifeOK. I'm fairly sure you can get that to fail still...
15:44:08LinusNfarhank: we are working towards using speex instead for the targets that can support it
15:44:13preglowLinusN: btw, latest builld seemed to add about 25kb to arm builds, will be less for coldfire
15:44:27preglowand yes, i enjoy telling you linuxstb the same things
15:44:46jhMikeSpreglow: still have a few of those minor changes left...made more than I realized
15:44:47linuxstbpreglow: I don't know... Maybe I'm not passing the correct -D options or something... Let me upload what I've got.
15:44:51LinusN25k is nothing
15:45:06preglowi can shave off more too
15:45:23linuxstbpreglow: My Makefile is here -
15:45:27preglowi'm just a bit hesitant to go #ifdef crazy
15:45:48linuxstbpreglow: Create a new directory (rbspeexenc) in tools/ move rbspeexenc.c into it, and put that Makefile there as well. Then "cd" into into, and type "make"
15:45:51pondlife25kcode gain, but freeing up 512k swap buffers?
15:45:56 Join Febs [0] (n=chatzill@
15:46:16jhMikeSyeah but the 512K buffer was only allocated when voice was required
15:46:17preglowpondlife: more than that
15:46:23preglowiram swap buffer too
15:46:31preglowthe greatest gain is less complexity, if you ask me
15:46:46*pondlife agrees
15:46:47jhMikeSyeah, the buffer handling is way simpler
15:46:49preglowlinuxstb: ok, so this thing replaces the libspeex/Makefile for this scenario?
15:47:18linuxstbYes. It's a completely self-contained Makefile, but it uses SOURCES in the libspeex directory.
15:47:45preglowlinuxstb: i get the same
15:47:47linuxstbI think the problem is that we have no endianness check...
15:48:05linuxstbI think it worked by chance for you - it was compiling with the same endianness as the target...
15:48:18preglowi never use the header module
15:48:37preglowisn't letoh32 a rockbox function?
15:48:51preglowi know where the bug is
15:49:29preglowok, how do i detect if i'm building as part of rockbox?
15:49:39preglowany nice define?
15:50:02linuxstbI think there's a -DROCKBOX define. If not, just add one to your libspeex Makefile
15:50:28preglowifdef ROCKBOX fixed it
15:50:29linuxstbIn fact, that's something I invented for libdemac - I use it there.
15:50:40preglowso i should go #ifdef ROCKBOX?
15:50:54linuxstbYes - but you need to add -DROCKBOX to the libspeex Makefile
15:51:01GodEaterpondlife: I cannot get the earsplitting shriek with v3
15:51:05GodEaterI've tried really hard
15:51:15pondlifeSo keyclick is clicking?
15:51:20GodEaterfor the most part
15:51:27GodEaterI get the occasional *short* beep
15:51:30pondlifeThe occasional beerp
15:51:36GodEaterbut not the never ending one
15:51:48preglowlinuxstb: kiedoke, will commit a fix so you can continue working, the makefile still isn't 100% good
15:51:49GodEaterstill want me to try with v4 ?
15:51:49pondlifeAre you using SVN as a base?
15:51:55GodEaterpondlife: of course
15:52:16pondlifeGodEater: Maybe jhMikeS has resolved it? I'll look at his pcmbuf work shortly...
15:52:27linuxstbpreglow: Which Makefile?
15:52:31pondlifeNo need to mess with v4 unless you can get a brain-fryer out of v3.
15:52:33preglowlinuxstb: yours
15:52:40GodEaterpondlife: have you seen the changes LinusN and I have made to 7738 yet ?
15:52:44linuxstbpreglow: OK...
15:53:09preglowlinuxstb: in speex_header.c, just replace the #if 1 with #ifdef ROCKBOX
15:53:15preglowlinuxstb: that should have you on your way for now
15:53:22pondlifeBut I'm unlikely to commit it as I don't have an iPod
15:53:43GodEaternot sure who is going to in that case
15:53:48linuxstbArgh, I did "svn update" and now have conflicts...
15:53:52pondlifeI just think that it should be committed as the previous time I played with an iPod in Rockbox, the wheel didn't feel great.
15:53:55GodEaterI've ripped all the menu entries out now
15:54:09GodEaterand just set the defaults that had been picked before as constants
15:54:18pondlifeCool. So you just need to pester a dev who has an iPod :)
15:54:19preglowlinuxstb: at least this makes the libspeex/Makefile a lot nicer...
15:54:33GodEaterpondlife: I already have - they're mostly ignoring me ;)
15:54:43pondlifehaha, that old problem.
15:54:59GodEateramiconn isn't keen on it because it's clickwheel ipods only
15:55:01*LinusN installs the scceleration patch on his G5
15:55:32pondlifeWell, it can be extended to the other iPods once it's in SVN...
15:56:02linuxstbpreglow: I think I've messed up my tree, and your patch doesn't apply any more to SVN - can you update it?
15:56:11preglowlinuxstb: kiedoke
15:56:59preglowlinuxstb: i won't include any rbspeexenc.c stuff, then
15:57:01*pondlife is happy that jhMikeS worked on the crossfade code and never removed it ;)
15:58:36LinusNGodEater: so have we tested 7738 on all targets that it affects?
15:58:49 Quit CaptainSquid83 ("Miranda IM!")
15:59:10GodEaterI'm not sure. Llorean tested on Nano, linuxstb tested on a color, and I've tested on a 5.5G
16:00:00*jhMikeS should probably remove preglow's previous voice building patch before this becomes a mess
16:00:08preglowoh yes
16:00:09preglowi did warn you :>
16:00:33GodEaterthe e200 bit has been test by a lot of users afaik
16:00:50linuxstbpreglow: Hmm, strange error - it builds rbspeexenc but then gives an error?
16:00:52GodEaterthe only unconfirmed targets at the 2g mini, and the 4G greyscale
16:01:04jhMikeSpreglow: libspeex changes too? some are required here.
16:01:35preglowlinuxstb: it does build libspeex, yes, it complained about the resampler here
16:01:44preglowlinuxstb: do you defined ROCKBOX_VOICE_ENCODER also when processing SOURCES?
16:01:54*jhMikeS can just svn revert /tools
16:02:08linuxstbpreglow: Where are you typing "make" - you should do it inside tools/rbspeexenc/ for now
16:02:12preglowlinuxstb: did
16:02:41linuxstbI don't build libspeex... it just builds all the .o files and links them with rbspeexenc,o to create rbspeexenc (which works for me)
16:02:48linuxstbOr is that what you meant?
16:03:10preglow gcc -o rbspeexenc $(OBJS) -lm
16:03:28preglowdoesn't that line need rbspeexen.c too?
16:03:57preglowcc -DHAVE_CONFIG_H -DROCKBOX_VOICE_ENCODER -I ../../apps/codecs/libspeex -iquote ../../apps/codecs/libspeex -fomit-frame-pointer -Wno-unused-parameter -c -o all.o
16:04:00preglowcc: no input files
16:04:02preglowi get that here
16:04:26linuxstbYes, but before that, it's finished and created rbspeexenc ?
16:04:40 Quit spiorf (Remote closed the connection)
16:04:40 Quit Zagor ("Client exiting")
16:05:08LinusNthe file browser is still hopelessly sluggish on my g5 :-(
16:05:21preglowlinuxstb: it did indeed
16:05:28amiconnLinusN: eh?
16:05:57LinusNamiconn: yes, the cursor lags a lot
16:05:58amiconnI never actually had the impression that the browser is slow on G5, even before my lcd speedup
16:06:14GodEaterit's only speedy when the cpu is boosted
16:06:25amiconnWith the speedup, it's completely decent, even at 24MHz
16:06:36GodEaterstill choppy here
16:06:38linuxstbpreglow: This is odd - I'm assuming it's trying to create all.o because it's making "all"...
16:06:55linuxstbpreglow: If you type "make rbspeexenc" then it works perfectly.
16:06:56LinusNamiconn: the menus are like lightning, but not the file browser
16:08:23LinusNi'm surprised that the file browser is so much slower
16:08:32LinusNi mean, what's the difference?
16:08:46LinusNit's horrible
16:08:47linuxstbpreglow: Hmm, adding ".PHONY : all" fixes it, but I wonder why that's needed...
16:09:13GodEaterLinusN: is this the same degree of slow you mentioned a month of so back ?
16:09:18 Join spiorf [0] (n=spiorf@
16:09:27GodEaterI believe I got it too
16:09:34GodEaterbut I can't remember when it got fixed
16:09:36GodEateror why
16:09:54preglowlinuxstb: don't ask me, i would have been frothing at the mouth already if dealing with this
16:10:21jhMikeSpreglow: now I'm getting undefined refs to codec_realloc, speex_uwb_mode, speex_default_user_handler, speex_inband_handler :\
16:10:36linuxstbpreglow: OK, Makefile updated on my website - same URL
16:10:53GodEaterin any case, my file browser is now no longer slow
16:10:56preglowi'll check it out
16:11:13linuxstbpreglow: But we still need to integrate this into tools/Makefile somehow...
16:11:57preglowlinuxstb: and once again i am more than content with leaving it in your able hands :>
16:12:40preglowlinuxstb: i'll add some optimizations and stuff
16:12:42 Join Bagder [0] (n=daniel@rockbox/developer/bagder)
16:12:55LinusNthe default wps is buggy on my g5
16:13:13LinusNit clears a part of the screen for no reason
16:13:48linuxstbpreglow: Would you prefer rbspeexenc to be created in tools/ ? uclpack something similar - the source (and makefile) is in tools/ucl but the executable is created as tools/uclpack
16:15:32preglowlinuxstb: well, it sounds more logical, but i don't really care
16:16:10preglowjhMikeS: ahh
16:16:13pixelmaLinusN: is that when you had a wps with album art before? I think something like this was reported here and Nico_P committed a fix for it a few days ago, maybe it's still not completely fixed?
16:16:18linuxstbOK, let's rename the rbspeexenc directory to rbspeex, and then I'll build rbspeexenc in tools/
16:16:32preglowjhMikeS: add back speex_callbacks.c in SOURCES and see if that makes a few disappear
16:16:33LinusNpixelma: i have never had that, and i just reset my settings
16:16:49pixelmaweird, never seen that
16:17:41preglowjhMikeS: i don't get why you're getting codec_ things, though
16:18:32jhMikeSdo I need those for the core?
16:19:10preglowno, not at all, you fixed those last by commenting out a couple of allocs
16:19:18preglowperhaps you've deleted that code again?
16:19:20preglowit was in bits.c
16:19:26jhMikeSthat was malloc, free
16:19:38preglowright, so it's realloc now
16:19:44Nico_PLinusN: do you have album art files?
16:19:58jhMikeSI'll just use #ifndef ROCKBOX_VOICE_CODEC. BTW, are those tables to stay in IRAM for non-core?
16:19:59linuxstbpreglow: OK, final version of Makefile now in same place, and a diff for tools/Makefile is here -
16:20:16linuxstbpreglow: You now need to rename the tools/rbspeexenc directory to tools/rbspeex
16:20:22pixelmaLinusN: I take it back, I have that too now on my c200. It looks as if it's my old wps but with the default font and without the bitmaps
16:20:24LinusNNico_P: weird, a reboot solved the issue
16:20:58LinusNpixelma: yes, that sounds like what i had too
16:21:07preglowlinuxstb: excellent
16:21:29pixelmayes, looks like it's not reset correctly - reboot helped here too
16:21:41amiconnLinusN: File browser isn't slow at all here on the G5.5 ...
16:21:53LinusNamiconn: really odd
16:22:11LinusNamiconn: i wonder what on earth can cause that
16:22:54 Join jpt9_ [0] (
16:23:12preglowlinuxstb: 404 on tools.diff
16:23:28linuxstbI'll upload it then... ;)
16:23:36linuxstbThere now.
16:23:47pixelmaLinusN: wasn't there something about the gradient bar settings last time you discussed that here?
16:23:59Nico_PLinusN: do you have filetype colouring enabled?
16:24:04jhMikeSpreglow: one more time: ought I to retain those tables in IRAM for non-core?
16:24:13LinusNpixelma: yes, that was my first suspicion, but it is the same with a solid bar
16:24:25LinusNNico_P: is that enabled by default?
16:24:38Nico_PLinusN: no, I don't think it is
16:24:43*GodEater has both file coloring and the gradient bar in use on his 5.5
16:24:48linuxstbThe file browser seems slow on my 5g as well... I'm not sure how old the build is though...
16:25:01*pixelma wonders why bookmarking sometimes works on c200 and sometimes not... :\
16:25:16preglowbookmarking needs work...
16:25:19*pondlife wants bookmarking to be per-playlist, not per-folder
16:25:39preglowjhMikeS: most of the tables can be left out at minimal cost, should i check that out?
16:25:50GodEateropinion seems divided on where the bookmark file should end up too
16:25:57pixelmawow, and now it hung on the "bookmark saved" screen
16:26:12pondlifeI'd use it much more if I could save a playlist and have the resume point live in the .m3u8 file...
16:26:26pondlifeSome kind of bookmarks as #EXTM3U maybe
16:26:39jhMikeSpreglow: I think so...I say squeeze it to the minimum. I don't think speed is an issue here at all.
16:26:42GodEaterproprietary playlist format ?
16:27:01preglowjhMikeS: okiedoke, then i'll do that after commiting the encoder
16:27:05preglowlinuxstb: thanks, btw
16:27:34pondlifeGodEater: Is there a standard way I could achieve this?
16:27:35 Quit hcs ("Leaving.")
16:27:36preglowjhMikeS: but really, i don't get those new errors
16:29:23GodEaterpondlife: don't think so
16:29:42GodEaterwhich is why we invented the bookmark file in the first place isn't it ?
16:30:19preglowok, encoder commit coming up
16:30:56pondlifeYep, but I'd like to be able to have my "all random tracks" playlist for use when I'm feeling indecisive, then listen to a particular album and later return to my original playlist,.
16:31:09pondlifeSo a per-playlist bookmark would do that neatly, I think
16:31:25pondlifeThe per-folder method is a special case of per-playlist, no?
16:31:42GodEaterI guess so
16:31:54GodEaterbut the bookmark naming relies on the folder of the currently playing track
16:32:05pondlifeWhich sucks for database users.
16:32:10GodEaterregardless of whether you're actually playing a folder or a playlist you've built yourself
16:32:25GodEaterI think it sucks for file browser users too
16:32:31pondlifeYep :)
16:32:35jhMikeSpreglow: building a core decoder you don't?
16:32:37GodEaterI don't like that the .bmark gets saved into the parent folder
16:32:41GodEaterbut apparently some people do
16:32:47GodEaterso I didn't bother submitting a patch to change it
16:33:07pondlifeI like the idea that I could copy my bookmarks, but they should be associated with the playlist, not the folder.
16:33:16GodEaterTo be honest - I think we'd be better off with a bookmark catalog
16:33:19GodEatermuch as we have for playlists
16:33:22 Join karashata [0] (
16:33:24GodEaterperhaps even use the same folder
16:33:28pondlifeOr build it into the playlist catalog
16:33:42preglowjhMikeS: hmm?
16:33:54pondlifeThe standard resume tracking should just be a bookmark too.
16:33:58GodEaterI don't like the idea of the resume info in that actual playlist though
16:34:22GodEaterit stops the playlist being portable =/
16:34:30pondlifeCould be a .bmark file alongside the playlist. The dynamic playlist would need a dynamic bookmark
16:34:31jhMikeSpreglow: for building the decoder for the core, you get no errors? don't see how that's possible myself. Anyway...I don't need the user callback stuff for sure.
16:34:53GodEateryeah that would work nicely
16:34:57preglowjhMikeS: i haven't tried yet, i'll need a new patch from you
16:34:59pondlifeI agree, I wouldn't want to lose my bookmark(s) because I did some m3u maintenance in another program
16:35:06preglowjhMikeS: btw, do remove my old patch around now
16:35:13preglowall the encoder stuff is going in
16:35:14jhMikeSwhich parts
16:35:21pondlifeBut .bmark alongside .m3u makes some sense.
16:35:22jhMikeSI reverted /tools
16:35:46jhMikeSI guess the default user handler does the bit advancing...hmmm
16:35:48preglowtons of stuff will go down in libspeex/ now as well
16:35:55PaulJamdoesn't rockbox do it tha way when you play a m3u8? if i play a playlist and create a bookmark, then the bookmark is created in the same dir as the playlist and has the same name, so to resume you just habe to load the bmark file instead of the m3u8.
16:36:04jhMikeSjust hold on a sec
16:36:12preglowtoo late, but you don't have to update
16:36:37LinusNamiconn: so you are saying that the file browser is just as responsive as the menus?
16:36:40GodEaterPaulJam: I didn't think it did - the routine to name the .bmark file seemed to look at the location of the currently playing file
16:36:45GodEaterI'll check again
16:36:58jhMikeSit's not much...I'll just revert that tree
16:37:16PaulJamGodeater: this happens afaik when you play a dynamic playlist
16:37:39GodEateryou appear to be 100% correct
16:37:44pondlifeSo if I save my playlist first, it should work
16:37:45GodEaterI obviously didn't understand the code properly
16:37:55GodEaterit's a little unwieldy mind you
16:38:02pondlifeBut when I resume it goes back to being a dynamic one....:/
16:38:14*pondlife will play
16:38:25preglowback in ten, brb
16:38:40GodEateryeah - that's not ideal
16:38:47GodEateralso - the naming could be better
16:39:15GodEaterwhen you're viewing "supported" files - the .bmark file appears to be called <playlist name>.m3u8
16:39:22 Quit jpt9 (Read error: 110 (Connection timed out))
16:41:46pondlifeThe current playlist filename should be preserved if I shut down, it doesn't seem to...
16:41:56pondlifei.e. if I resave it should default to the same name
16:42:26*GodEater has just tried bookmarking a playlist
16:42:45GodEaterand loading it back after playing a different playlist gets the whole playlist back
16:42:58GodEaternot just the folder the bookmark was in
16:44:53 Join jgarvey [0] (
16:48:01 Join XavierGr [0] (
16:50:35 Join scorche|w [0] (n=42c007b2@rockbox/administrator/scorche)
16:50:56 Quit jpt9_ (Read error: 110 (Connection timed out))
16:51:52 Quit Febs ("ChatZilla 0.9.79 [Firefox]")
16:51:55preglowwell, that was smooth
16:51:57*jhMikeS will just hold off trying to sync any patches for a little bit
16:52:08preglowjhMikeS: nothing more from me now
16:52:29 Part tkooda ("thanks")
16:54:34 Quit Bagder ("It is time to say moo")
16:54:34amiconnLinusN: I really can't compare, just no noticeable lag
16:54:54LinusNin my case it is so much slower in the browser
16:55:09amiconnI am using solid colour bar, no background image, no filetype colouring, and no custom icons
16:55:22LinusNamiconn: same here, i use the default settings
16:55:26amiconnJust rockbox blue background and rockbox orange selector
16:55:55LinusNi have tried different settings, no difference in lag
16:55:57jhMikeSpreglow: ok, now I'm just getting the stuff about the various heap calls.
16:56:34LinusNi use the default settings, as per a settings reset
16:56:57jhMikeSpreglow: should I do the same for the realloc as the others?
16:59:09*amiconn has to compare again whether he can feel a difference between menus and browser
17:01:03LinusNamiconn: it is easier to see if you enter a directory with many items, and use a small font
17:01:09preglowjhMikeS: jwell, why not
17:01:43preglowlinuxstb: oh yeah, forgot to credit you for that, i assume you'll survive :>
17:02:25linuxstbpreglow: No problem ;) BTW, I think line 75 in rbspeexenc be fread, not read?
17:02:35jhMikeSin speex_bits_read_from and speex_bits_read_whole_bytes, it reallocs bits->chars...not sure how to handle this.
17:02:36LinusNamiconn: i get the same lag if i browse fonts
17:03:48preglowlinuxstb: i LOVE how c just lets bugs like that through
17:04:04linuxstbpreglow: You may want to add -Wall to the CFLAGS...
17:04:11linuxstb(that's how I found it...)
17:04:30jhMikeShmmm...doesn't seem to care if I just remove the group functions alltogether.
17:05:29preglowlinuxstb: are you sure the makefile deps are ok? if i change rbspeexenc.c, it doesn't rebuild when called from "make tools"
17:05:39LinusNamiconn: hmm, i changed to nimbus-12, and now the menus are almost as laggy as the browser :-)
17:05:46*jhMikeS should just start removing functions and if compiles it's good if not, keep it
17:06:47LinusNamiconn: well, the browser is still a lot worse than the menus
17:07:18 Join tarehart [0] (
17:07:23 Part LinusN
17:07:34 Part tarehart
17:08:23***Saving seen data "./dancer.seen"
17:08:23linuxstbpreglow: The problem is in tools/Makefile - we'll need to make rbspeexenc .PHONY
17:08:57 Quit TMM ("Ex-Chat")
17:09:02 Quit petur ("connection reset by (loads of) beer")
17:09:14linuxstbpreglow: I think the same issue will happen with uclpack...
17:09:29jhMikeSwhat does .PHONY do? Where's the makefile manual? Use "man"?
17:10:43preglowlinuxstb: ok, so any time i change rbspeexenc, people will need to actually make clean first
17:10:47preglowthat's very nasty
17:10:56preglowjhMikeS: gnu hates man, you need to use "info make"
17:11:03preglowjhMikeS: but i'd just go for the web manual
17:11:08linuxstbThe only version of the info is better - ask google... -
17:11:35linuxstbpreglow: No, adding the .PHONY means that "make -C rbspeex" will always be run.
17:12:04linuxstbThe problem is that currently, "make" sees rbspeexenc already exists, and that there are no dependencies, so does nothing.
17:12:18linuxstb(in tools/Makefile)
17:12:36lostlogicjhMikeS: clip_sample_16 seems to do the wrong thing with the sign bit... what am I missing?
17:12:54preglowwell, end result is that changes i make to rbspeexenc.c are ignored, yes?
17:13:07preglowit certainly seems like it
17:13:11jhMikeSlostlogic: same code as in DSP and SPC codec
17:13:27jhMikeSit should be fine unless I munged it somehow
17:13:40preglowlinuxstb: even a hard "make tools" ignores anything i do to rbspeexenc.c
17:14:18jhMikeSlostlogic: which part looks wrong? I don't see a problem.
17:14:50lostlogicjhMikeS: I must not be understanding it −− why is the sign bit xor'd with the _low_ bit of the final sample?
17:15:27jhMikeS0xffffffff ^ 0x7fff = 0x8000, 0x00000000 ^ 0x7fff = 0x7fff
17:15:50linuxstbpreglow: Fix committed.
17:15:53 Join OlivierBorowski [0] (
17:15:54jhMikeSwoops 0xffff8000
17:16:26lostlogicoh, there's my stupid −− >> is signed.
17:16:32jhMikeSyeah :)
17:18:01 Join PaulPosition [0] (
17:22:44 Join Febs [0] (n=chatzill@
17:23:56 Quit n1s ()
17:24:10pondlifelinuxstb: Your red?
17:24:11preglowlinuxstb: most excellent
17:24:54preglownow, that's a fancy error
17:24:59pixelmalooks like problems on bygg again?
17:25:12*preglow helps bygg to its feet
17:25:38*preglow takes the booze bottle away from bygg
17:26:05 Join salty-horse [0] (n=ori@pdpc/supporter/active/salty-horse)
17:26:43salty-horsewhen playing a song in the file browser, and the dir is turned into a playlist, is it possible to have it automatically sorted by track number?
17:26:44*linuxstb blames the red on bygg as well...
17:27:36linuxstbIt should be sorted by filename - so if the filenames contain the track number at the start, it will sort correctly. Otherwise you will need to use the database.
17:28:15 Quit pondlife (Read error: 104 (Connection reset by peer))
17:29:33 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
17:29:58 Quit idnar (Nick collision from services.)
17:30:00 Join idnar_ [0] (i=mithrand@unaffiliated/idnar)
17:35:40salty-horsereasonable enough, linuxstb. thanks
17:42:01PaulPositionlinuxstb - Don't waste any time on this, but : a (long) time ago, you commited some changes to the usb detection on PP devices (0/1 to USB_EXTRACTED/USB_INSERTED).. Since then the screendump doesn't work, at least on the h10 devices (see FS #8170).. Is it possible that debug's screendump didn't get the notice?
17:42:38 Nick idnar_ is now known as idnar (i=mithrand@unaffiliated/idnar)
17:43:53linuxstbPaulPosition: It sounds like you've proved my commit broke it... ;)
17:43:56 Quit Xerion (" ")
17:44:06*linuxstb tests on his ipod
17:44:20*preglow got a t-shirt
17:45:08linuxstbPaulPosition: I've just tried a screendump on my ipod (Color), and it works fine...
17:45:46PaulPositionlinuxstb - It might be something specific to h10s, I wouldn't know. But I started hunting the bug (through downloading 20+ different svn versions) when a few folks on misticriver told me about that.
17:47:30PaulPositionlinuxstb - Been reading source file after source file, but I'll be damned if I understand anything. As I said, it's pretty trivial anyway, so if there's no 'eureka' moment, just forget about it and keep working on making rockbox work. :)
17:47:51linuxstbWhat happens if you hold RIGHT when inserting the USB cable?
17:48:01PaulPositiongonna try that.
17:49:50PaulPositionlinuxstb - Well, well.. :o So the 'right/select' button would serve like the 'menu' button for that.. It does work.
17:50:45amiconnHmm, odd
17:50:56PaulPosition(although it kind of make 'screendump' useless :p )
17:51:11amiconnThe browser *is* a bit laggy compared to the menu on G5.5 (and btw I'm using Nimbus-19)
17:51:25amiconnBut it seems there is a stray boost somewhere
17:51:39amiconnI have a boost count of 1 right after boot
17:52:13linuxstbPaulPosition: I think this is because the H10 reverses the USB logic - it only reboots into disk mode if you're holding RIGHT (IIUC)
17:52:54PaulPositionlinuxstb - It would make sense. I guess.
17:53:26linuxstbPaulPosition: I'm still looking at it though...
17:55:31linuxstbAh yes, that is the reason. The H10 defines the "charge only" button as BUTTON_NONE, which means the usb detect function is returning USB_POWERED, not USB_INSERTED when nothing is pressed.
17:55:43linuxstbAnd the screendump is only done on USB_INSERTED...
17:56:04amiconnLooks like the lagginess simply depends on string length
17:56:22PaulPositionlinuxstb - But as I said, it's a trivial thing. WPS designers still have simulator for creating screendumps and most debug info don't react on pressing right, so there's no big loss.
17:56:22amiconnSo if there are lots of long filenames in the list, it will slow down
17:56:31 Join localhero [0] (n=dinzdale@
17:57:27amiconnI think that means mono bitmap drawing is in need of optimisation
17:57:28linuxstbPaulPosition: Yes, I can't think of a nice fix... Let's keep waiting for Zagor's usb code...
17:57:38 Part localhero
17:57:46linuxstbPaulPosition: I'll comment on your task though.
17:58:12PaulPositionlinuxstb - Thanks for taking some time looking at it. And commenting, if only so no one starts chasing windmills.
17:58:17amiconnOf course it could pay off to check for nasty things like multiple redraws in the list code as well
17:59:27PaulPositionamiconn - Is that specific to the G5.5 or is it something you'd like to test on (completely different) targets?
17:59:56*PaulPosition 's only forte is having time to test.
18:01:03amiconnIt probably affects all 16 bit colour targets (if it's the mono bitmap drawing), or all targets (if it's the list code)
18:01:23amiconnBut larger screens will suffer more
18:01:55amiconn(unless that larger screen is backed by a faster cpu, i.e. gigabeat)
18:01:55PaulPositionheh.. I have the smallest screen.
18:02:03amiconnI doubt that...
18:03:04amiconnSmallest screen (both in terms of pixels and in terms of framebuffer bytes) would be the archoses
18:03:08pixelmaeven not the smallest colour screen
18:04:26PaulPositionTrue enough, I didn't think of the mono targets. But is there a colour one with less than 128x128 ?
18:04:39pixelma132x80 (c200) is smaller than 128x128
18:04:39amiconnc200: 132x80
18:04:56PaulPositionSheesh, c200. Forgot about this new kid.
18:05:39 Join webguest24 [0] (i=82cf39eb@gateway/web/cgi-irc/
18:06:55 Quit webguest24 (Client Quit)
18:07:51amiconntierra: Around now?
18:08:28amiconnMeh, who did this strange "optimisation" in lcd-16bit.c: lcd_mono_bitmap_part() that actually slows things down?
18:08:36*amiconn checks svn
18:10:14amiconnLooks like gregw...
18:10:49amiconnPutting a switch() into an inner loop is clearly not a way to speed things up...
18:14:57linuxstbI'm also curious about his ARM optimisation to lcd_bitmap_transparent_part...
18:15:54pondlifeHmm, how should one "artificially heat" a Nano? ;-)
18:18:09PaulJammicrowave for example...
18:18:15krazykitprobably a hairdryer
18:18:21 Part salty-horse ("Leaving")
18:18:45krazykitactually, put some oil in the skillet... add onions and garlic, when those are near done, put in the nano
18:19:17pondlifeoblib's very specific with that "CPU > 65MHz bit" comment.
18:19:36pondlifeWhy 65MHz, I wonder?
18:21:48 Join japc [0] (n=japc@
18:25:18amiconnThe bit is named that way by the ipl people
18:26:09amiconnBut it seems that it's not recommended on PP in general; PaulPosition tried it on H10 and got freezes
18:26:10Nico_Pdoes this mean FS #7510 is fixed?
18:26:27amiconnThe PP ata controller definitely needs research
18:26:29 Quit Calcipher ("—I-n-v-i-s-i-o-n— 2.0 Build 3515 with A Pack Fix By")
18:26:54amiconnI might look into that later, for now I'm quite busy with lcd & backlight
18:27:25*amiconn *still* needs a tester with some basic build knowledge and an ipod 4th gen grayscale or a mini 1st gen
18:27:51*amiconn also thinks a gfx benchmark plugin would be a good idea
18:28:41preglowjhMikeS: so, what next?
18:28:42pondlifeSomeone here must have a 4th gen grayscale or a mini 1st gen
18:28:49Nico_Pamiconn: what would it do compared to testfps?
18:29:19amiconnTest things like lcd_drawline(), lcd_fillrect(), lcd_mono_bitmap_part(), lcd_bitmap_part() ...
18:29:38pondlifeMandelbrot with an FPS readout?
18:29:56amiconnMandelbrot isn't that gfx heavy
18:30:07preglowit's more cpu heavy
18:30:11linuxstbamiconn: I agree, a gfx benchmark plugin would be very helpful....
18:30:54amiconnBut lcd-16bit.c: lcd_mono_bitmap_part() is clearly in need of optimisation, and in fact t shouldn't be difficult
18:31:21amiconnpondlife: test_fps tests hardware. A gfx benchark is a pure software benchmark
18:31:40amiconnIt could even run in the sim (althoguh that probably doesn't make much sense)
18:32:14amiconnpondlife: For mono bitmaps, logo.rock with the sleep() removed would be a better benchmark
18:32:44 Part linuxstb ("Leaving")
18:32:52*amiconn wonders whether greg white really benched his optimisations on gigabeat
18:33:13*amiconn suspects that it slows down things on all 16 bit targets, including gigabeat
18:34:10 Join linuxstb [0] (n=chatzill@rockbox/developer/linuxstb)
18:34:33 Join MethoS- [0] (
18:34:49pixelmaPaulJam: as I understand it, a microwave won't help you much as long as you don't apply some water as well.. ;)
18:35:44PaulJampixelma: well, the metal parts should heat up, no?
18:36:00preglowor fat :P
18:36:29krazykitand/or spark like mad.
18:36:37Nico_PPaulJam: it is a *very* bad idea to put metal in a microwave
18:37:01PaulJami didn't intend to do it
18:37:43*scorche|w recommends the hair dryer method of heating up a nona
18:37:55scorche|wthat also works for nanos >_>
18:37:57pixelmathat nona again... :P
18:38:18scorche|widop nona
18:38:28Nico_Pdoes the patch posted on FS #7510 fix the problem?
18:42:25linuxstbIt seems to...
18:43:06Nico_Pthat's excellent news! :)
18:43:32linuxstbI guess it needs testing on other PP targets before committing though.
18:43:36PaulPosition...on nano's that is.
18:44:12linuxstbNo, all targets I think... (well, all that use the ATA controller)
18:44:18PaulPositionlinuxstb - that patch + h10 == freezes on boot 8 out of 10 times, freezes on (trying to) playing song the rest of the time.
18:45:01pixelmaamiconn said that about 20 minutes ago...
18:45:04PaulPositionUnless he changed it a lot since I tried it ~10 hours ago.
18:45:15*linuxstb goes to read the logs...
18:47:47preglowwhat, a patch on that page??
18:48:48preglownow, lookie here
18:48:52preglowi wonder where this info was found
18:49:34preglowbtw, what's the status on ipod hw eq?
18:51:50rasherLooks like the nano problem was pinched. Well done, whoever that was (I just quickly skipped through -sf mails)
18:52:35 Quit Guile`` ("KVIrc 3.2.0 'Realia'")
18:52:40 Join JoeBorn [0] (
18:52:46JoeBornhi all
18:52:54JoeBornthought you might find this interesting:
18:53:14JoeBornTI is releasing a free DSP compiler for the C54x for open source work
18:54:25JoeBornwith the DSP bridge, it should be a lot easier to port rockbox to many of the TI devices
18:54:29jhMikeSamiconn: those are optimizations?
18:54:30rasherNot bad. Pity they had to be picky about who uses it
18:54:50JoeBornit's a good step, IMHO
18:55:04rasherThe distinction between opensource and commercial is a bit weird
18:55:17linuxstbrasher: That's what struck me as well...
18:55:27linuxstbJoeBorn: Do you have a link to the actual license of that compiler?
18:55:59JoeBornit's an attachment to the top link
18:56:17amiconnjhMikeS: The commit message says so...
18:56:24JoeBornwe negotiated it for months.
18:56:37JoeBornyou should have seen the first version of the license.
18:56:41*jhMikeS would have thought the diff engine reversed panels
18:56:56 Quit moos ("Rockbox rules the DAP world")
18:56:58rasherI can't find that attachment
18:57:18jhMikeSyeah, I'm looking at that and saying wtf
18:57:25linuxstbI can't see the attachments either...
18:57:49linuxstbJoeBorn: Are attachments only visible to logged in users?
18:58:12JoeBornshouldn't be, let me see.
18:58:43amiconnjhMikeS: Well, the idea to implement often needed draw modes directly instead of calling the pixelfuncs might be helpful - but the switch must happen outside the inner loop then
18:58:47JoeBorntry again, should be fixed
18:59:11preglowJoeBorn: now this is an unexpected turn of events
18:59:54PaulPositionpreglow - hw eq? Is someone working on getting rockbox to use hardware eq on some targets?
18:59:54rasherNow it's there..
19:00:07preglowPaulPosition: i'm trying to have it removed and replaced with tone controls on ipod video
19:00:15 Join toffe82 [0] (
19:00:15linuxstbI wonder why they only made the Linux version available though - lots of open source devs use Windows, and I'm sure vice-versa...
19:01:09PaulPositionpreglow - Oh. I *so* don't know about the ipod platform. I just wish I could have some functional eq on my h10 without ui slowing to a crawl. :p
19:01:28amiconnJoeBorn: What host archs/OSes is this compiler available for?
19:01:28JoeBornlinuxstb: we asked for linux, since that's what most of our folks are using
19:01:34 Join bertrik [0] (
19:01:44JoeBornamiconn: linux for now, but windows might be coming
19:01:47JoeBornwe're working on it.
19:01:51amiconnWe'd need at least linux-x86 and linux-x86_64
19:02:03JoeBorncomment on the blog that you'd like 64
19:02:06amiconnWindows would be helpful too (cygwin)
19:02:40preglowlinux-x86_64 isn't strictly needed
19:02:47preglowi'd go for windows first, fo shizzle
19:02:55amiconnWell, several build servers are running that
19:03:05preglowyeah, but they can also run linux-x86
19:03:07preglowno trouble at all
19:03:22amiconnYes - but only if the 32 bit compatibility libs are installed
19:03:40*amiconn prefers to keep his system 64 bit clean
19:04:00preglowsure, me too, but enabling it is very simple, so it really is no biggie, if you ask me
19:04:04amiconnAnyway, not all build servers would need it - those would need to be excluded from TI builds
19:07:30 Quit nicktastic (Read error: 113 (No route to host))
19:08:27***Saving seen data "./dancer.seen"
19:09:16 Join webguest80 [0] (i=8071369a@gateway/web/cgi-irc/
19:13:35jhMikeSpreglow: what's left to do in the voice switchover besides daily builds using speex? *getting impatient*
19:13:58rasherWhat needs to be changed?
19:14:03rasherWon't it automatically use speex?
19:14:35JoeBornamiconn: well comment on the blog, TI will be listening
19:14:43jhMikeSTo commit the changed playback of course but can't do that until the other build support is in place.
19:14:43 Quit webguest80 ("CGI:IRC (EOF)")
19:15:11rasherBut what build support? Has anything changed to build a voicefile?
19:15:25jhMikeSnew format = new encoder
19:15:40rasherYeah but isn't that built automatically?
19:15:59jhMikeSI don't think that was put in place yet in a config.
19:16:17rasherso you have to manually build rbspeexenc?
19:16:54jhMikeSyeah, atm.
19:17:22rashercouldn't it be built in the same way as voicefont and wavtrim?
19:18:31jhMikeSThere was some trouble. one thing was on Debian VMWare it couldn't find certain #includes, I guess a recent commit fixed another.
19:19:09rasherBut shouldn't "make voice" at least attempt to build rbspeexenc in the same way the other tools are built?
19:20:24 Join linuxstb_ [0] (
19:20:49 Quit billenium ("Ex-Chat")
19:21:01 Quit jhulst ("Konversation terminated!")
19:22:19rasherAdding \$(SILENT)\$(MAKE) -C \$(TOOLSDIR)/rbspeex CC=\$(HOSTCC) all to the Makefile when building voice for swcodec
19:22:21jhMikeSMaybe. I didn't work on that part but when will the build servers start generating the speex voice for dailies. The core commit and the format change have to happen together.
19:22:25rasherIf I got that right
19:22:58preglowjhMikeS: little, i was asking you the same
19:22:59rasherThe format change should happen because the commit makes it happen, surely?
19:23:05preglowjhMikeS: i am going to have a test, that's for sure
19:23:08preglowand that's going to happen now
19:23:22preglowand i still need to fix
19:23:24jhMikeSpreglow: I'm ready to push the button at a moment's notice
19:23:44preglowwell, we need to get a hold of a swede too
19:23:53preglowdaily voice builds
19:23:54rasherpreglow: create support in instead
19:23:56preglowthey need to change the script
19:23:58 Quit linuxstb (Nick collision from services.)
19:24:02rasherWhat needs to be changed?
19:24:02 Nick linuxstb_ is now known as linuxstb (
19:24:04preglowrasher: i have created support in
19:24:11rasherpreglow: no I mean for generating talk clips
19:24:19preglowrasher: just the basic shit, use lame for hwcodec, rbspeexenc for swcodec
19:24:23preglowthat's that, basically
19:24:29rasherpreglow: I thought that happened automatically in configure?
19:24:36rasherI don't see why it wouldn't
19:24:43preglownot for, surely?
19:24:48preglowi don't really know, i never use this stuff
19:25:03rasherI think our problem is that we're talking about two things at once.
19:25:16rasherFirst, what needs to be changed in the voice build scripts?
19:25:22 Quit Thundercloud (Remote closed the connection)
19:25:29rasherDaily voice, that is.
19:26:05preglowrasher: codec selection
19:26:07preglowi removed one prompt
19:26:14preglowcodec choice
19:26:19rasherthere was no codec choice
19:26:21preglowwhich basically was just "lame" before
19:26:31pregloware you certain about that? i'm quite sure configure asked for codec
19:26:38rasherIt doesn't.
19:26:50rasher# option=`input`
19:26:51rasher option="l"
19:27:02preglowright, then we're good to go
19:27:03rasherIt used to, though
19:27:42rasherAssuming rbspeexenc will build automatically and succesfully, it should work
19:27:54rasherAs far as I can tell anyway
19:28:09preglowi still have to fix and test talk clips
19:28:19preglowand create a build for amiconn's voicebox
19:28:40 Join japc_ [0] (n=japc@
19:28:54rasherYou should just code talkclips support in and deprecate
19:29:30 Join linuxstb_ [0] (
19:29:33preglowi'm not doing that now :)
19:29:43preglowthis shell script work is killing my interest slowly and certainly
19:29:51preglowi was in this for the codec work :P
19:29:52rasherBut.. perl!
19:30:05preglowit's not enough anymore :/
19:30:17preglowwhat do most people use for voice files anyway?
19:30:18 Nick fxb__ is now known as fxb (
19:30:39rasherVoice files are created by is dead?
19:31:01rasherI don't think the vbscript solution has caught up, and is dead as a dodo
19:31:04preglowwhy is it still in svn? :)
19:31:27rasherBecause no one has deleted it yet
19:32:09rasherI guess I'll have a go at stuffing talkclip generation into
19:32:51rasherWill probably wait until after your commit
19:34:03preglowok, so talk clips will be dead until you do? :>
19:34:35rasherI don't think that'd be terribly unreasonable
19:34:46rasherI doubt many people use anyway
19:34:47 Quit linuxstb_ (Read error: 104 (Connection reset by peer))
19:36:22amiconnpreglow: voicebox probably needs minimal adjustment, and an rbspeexenc.exe
19:36:32preglowamiconn: i'll get you that today
19:36:36rasherpreglow: and you removed the encoder argument from encodewav?
19:36:43amiconnIirc it already allows to choose swcodec or hwcodec, because of format support
19:36:45preglowrasher: encodewav??
19:36:52rasherpreglow: inside
19:37:00amiconnThe hwcodec setting only generates clips for .mp2 and .mp3 files
19:37:24preglowrasher: i don't believe i've done any such thing, no
19:37:33rashercould've sworn..
19:37:50amiconnvoicebox is mostly a vbscript, hence easy to fix compared to shell scripts ;)
19:38:04rasherpreglow: ah right, nevermind
19:38:57 Quit linuxstb (Nick collision from services.)
19:38:57rasherI'm just talking rubbish
19:39:34amiconnThe user will need to be aware of the change though
19:40:26amiconnBy default, voicebox only creates missing clips, but will not recreate existing ones. There's a checkbox to override this; the user will have to do this once to replace the old .talk clips in mp3 format
19:40:46amiconnThe script cannot know whether the clip is in old or new format
19:41:36amiconnThat reminds me - the core voice code must be robust enough to not crash when it's fed with mp3 data
19:42:30jhMikeSit is
19:43:46 Join merbanan [0] (n=banan@
19:44:25pixelmahmm... isn't rbutil also able to create .talk clips? Should probably adapted then too...
19:44:58 Join linuxstb [0] (n=chatzill@rockbox/developer/linuxstb)
19:45:09rasherGosh, all these options
19:45:31rasherpreglow: waste no time in deleting and friends ( and
19:46:49preglowrasher: commiting a new dude now?
19:47:22 Quit MethoS- (Read error: 110 (Connection timed out))
19:52:48rasherpreglow: Huh?
19:53:08 Join MethoS- [0] (
19:53:31 Join Xerion [0] (
19:55:58 Join webguest98 [0] (i=51dbe728@gateway/web/cgi-irc/
19:56:54amiconnbygg seems to be cursed...
19:57:34rasherEh.. are perl dirhandles global?
19:58:29linuxstbIsn't everything global in Perl, unless prefixed by "my" ?
19:59:11rasherDirhandles are different beasts
19:59:31rasherAh, there's a core module which seems somewhat saner
20:06:10webguest98hi. when in file browser the result of pressing SELECT button and FF button (on iaudio m5) are the same. where can I look for file to change it? i mean eg. pressing SELECT will insert whole folder but pressing FF will insert only one selected file.
20:09:54 Join Frazz [0] (
20:09:57preglowrasher: you told me to delete the bash scripts, i assumed you were working on a replacement for
20:10:13rasherpreglow: yeah, I'll implement it inside
20:10:22rasher(as planned from the start - I just never got round to it)
20:12:22 Join nanok [0] (n=nanok@
20:13:29preglowdo most people just download their voice files?
20:14:14rasherI'd assume so
20:14:22preglowamiconn: you around?
20:16:13 Quit japc (Read error: 110 (Connection timed out))
20:16:16 Quit japc_ (Read error: 110 (Connection timed out))
20:17:26 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
20:21:24preglowamiconn: i'm going to try an rbspeexenc build with mingw now
20:21:53 Quit Xerion (" ")
20:22:03PaulPositionwebguest98 - Have you tried a long press of select? Context menu would allow choosing the sort of insert you want (I think)..
20:22:04linuxstbpreglow: s/gcc/$(CC)/ in the Makefile...
20:22:05amiconnImho it would be best to build it in cygwin, but as pure win32 executable
20:22:13amiconnWe already do this for several tools
20:22:39amiconnFor sapi voice building, this is a necessity
20:22:55preglowamiconn: well, that's what mingw does...
20:23:02amiconn...because the encoder executable is started from the vbscript, not directly from cygwin
20:23:12preglowmingw is a pure windows env
20:23:17 Join Thundercloud [0] (
20:23:19preglowi never use cygwin, i hate the stuff
20:23:23amiconnYes, but it would be nice to have it built automatically on demand
20:23:32*linuxstb curses Windows for requiring a .exe extension and looks at the Makefiles again...
20:23:37preglowwell, i'm not installing cygwin
20:24:20 Join Zagor [0] (
20:25:16 Join Clam [0] (
20:25:35amiconnHmm. Once it builds in mingw, it's probably just a question of CFLAGS to make it build in cygwin
20:25:43 Join Xerion [0] (
20:25:49amiconncygwin gcc -mno-cygwin essetially *is* mingw
20:26:00*linuxstb wonders how uclpack works in cygwin - the target is "uclpack", even for Windows builds (IIUC)...
20:26:20ClamI'm having a bit of trouble with the WPS. When I use the Time tags, and a graphic on the same "line" the graphic will display for a millisecond then disapear. If I use any other text on the same line, the graphic stays...
20:26:35preglowwhat is -iquote?
20:26:38preglowgcc 3.4.5 does not have it
20:27:00linuxstbpreglow: It's the search path for #include "filename.h"
20:27:41preglow-I also includes that, no?
20:28:25amiconnlinuxstb: cygwin is clever enough to know about the .exe extension - in most cases
20:29:01linuxstbpreglow: I didn't think so - I thought that's the whole point if <> and "" when #incliding
20:29:32linuxstbamiconn: Hmm... In tools/configure, the target is "uclpack", so would "make" look for "uclpack.exe" ?
20:29:43linuxstbI mean tools/Makefile
20:30:24preglowthis didn't work too well...
20:30:36webguest98PaulPosition - yes I can do that but it's slow. the problem is that by pressing SELECT and FF exactly the same actions are taken. it'll be a small improve to differentiate these actions.
20:31:02amiconnlinuxstb: Not only uclpack - that applies to all tools
20:31:15preglowlinuxstb: just doing CC=i586-mingw32msvc-gcc does not work well and gives strange errors...
20:31:36linuxstbpreglow: Did you see the fix I mentioned (replacing gcc with CC) ?
20:31:39 Join Fraser [0] (
20:31:48preglowapplied it too
20:31:51PaulPositionwebguest - Then I *think* you'd have to set up a dev environment, edit source code and compile your own build.. On my device, FF/RW are used for page-up/down so I can't really relate. :p
20:31:56linuxstbpreglow: That worked fine for me...
20:32:04preglowi586-mingw32msvc-gcc -DHAVE_CONFIG_H -DROCKBOX_VOICE_ENCODER -I ../../apps/codecs/libspeex -O3 -fomit-frame-pointer -Wno-unused-parameter -Wall -c -o bits.o
20:32:07preglowi586-mingw32msvc-gcc: no input files
20:32:08 Nick fxb is now known as fxb__ (
20:32:09preglowthat happens here
20:32:45ClamI uploaded some examples: as opposed to:
20:33:44linuxstbpreglow: Are you doing that in the tools/rbspeex directory?
20:33:49preglowlinuxstb: yessir
20:33:56webguest98PaulPosition: yes, I want do compile but I simply don't know where I can find the file responlisbe for these actions.
20:34:06linuxstbpreglow: I've just done it, and get a .exe (well, a rbspeexenc file without the .exe extension...)
20:34:16linuxstbAnd you did 'make clean" ?
20:34:20preglowwhat failed here?
20:34:21LloreanNico_P: You here?
20:35:26LloreanNico_P: Anyway, I can't stay at the computer long. Just wanted to ask if the Album Art code then takes into account someone changing from a WPS with one size AA to a different size AA (up or down) safely.
20:35:29 Join BigBambi [0] (n=alex@rockbox/staff/BigBambi)
20:35:39linuxstbpreglow: "CC=i586-mingw32msvc-gcc make" is all I typed... My i586-mingw32msvc-gcc is v4.2.1
20:35:47Nico_PLlorean: I'm here, yeah
20:36:10Nico_PLlorean: depends on what you call safely
20:36:52 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
20:37:08krazykitwebguest98, apps/keymaps/keymap-yourmodel.c
20:37:28preglowlinuxstb: i didn't know such a mingw gcc existed...
20:37:44Nico_PLlorean: off the top of my head, everything should go smoothly... I think I tested too
20:37:49linuxstbpreglow: It's just the Debian one - "apt-get install mingw32"
20:37:58LloreanNico_P: "Doesn't cause crashes or glitchy-looking data"
20:37:58preglowmine is the ubuntu one
20:38:09linuxstbWhat version is yours?
20:38:12preglowlinuxstb: anyway, all is well, relay that exe to amiconn, doing it properly can come later
20:38:15preglowlinuxstb: 3.4.5
20:38:18Nico_PLlorean: I pretty sure it doesn't
20:38:38LloreanJust thought of it, and was curious
20:38:58ClamWell, it seems to only happen when using the "%pc" tag. Has this got something to do with the constant updating of the current time in the track, but the graphics won't get refreshed?
20:39:32 Join Arathis2 [0] (
20:39:35linuxstbamiconn: (my .exe compiled with mingw32)
20:39:40PaulPositionwebguest98 - So that would be keymap-x5.c which is used for both the x5 and m5.. Btw, I think that 'extra' select has been put in to facilitate one-handed operation of the dap.
20:40:10LloreanSoap_: If you're around at any point, there's a "fix" for the Nano issue, but one person's reported problems with it not saving settings on shutdown, or playing videos in mpegplayer, wondering if you could see what happens for you. Tracker entry 7510
20:40:19 Quit Frazz (Read error: 110 (Connection timed out))
20:40:40*Llorean is going to try it on his "working" nano, just as soon as he runs off and does some errands.
20:43:15preglowjhMikeS: have you tested .talk clips?
20:43:45jhMikeSno...not sure how to pull it off with the trouble I'm having
20:44:25jhMikeSwith the includes, or is that fixed?
20:44:34preglowwhat, you still can't compile?
20:44:44jhMikeSI'll try again
20:44:57preglowi haven't tried lately, i just assumed you had it fixed when you posted a patch
20:45:10preglowoh, those includes
20:45:44linuxstbjhMikeS: You're talking about building rbspeexenc ?
20:46:17linuxstbCan you update to latest SVN and try running "make" inside tools/rbspeex/ ?
20:46:51webguest98krazykit - this file controls buttons, but BUTTON_RIGHT and BUTTON_SELECT are connected both to ACTION_STD_OK. If it might work I should add eq. ACTION_insert and bind connect with BUTTON_RIGHT. and I have to change some *.c file.
20:47:09jhMikeSunrecognized option -iquote
20:47:35webguest98eq = e.g.
20:47:59webguest98without 'bind'
20:48:04krazykitwebguest98, sorry, i'm not really familiar with that code.
20:48:49linuxstbjhMikeS: So which native gcc does that image come with?
20:48:57webguest98ok. but thanks you
20:49:50*linuxstb sighs...
20:50:31preglowjhMikeS: go windows? :>
20:50:43preglowouch, old image?
20:50:53linuxstbjhMikeS: OK, just try removing the -iquote option completely in the Makefile
20:50:55jhMikeSlast one on the Wiki
20:51:29linuxstbjhMikeS: I don't think that was necessary anyway - it builds fine for me without it.
20:51:31rasherGah, I'm getting files... That's not what I wanted
20:51:37 Quit Arathis (Nick collision from services.)
20:51:42 Nick Arathis2 is now known as Arathis (
20:52:57webguest98PaulPosition: did you understand what my problem is? any advice?
20:53:10jhMikeScc: ../../apps/codecs/libspeex: linker input file unused because linking not done
20:53:14jhMikeSbunches of those
20:53:42 Join ToHellWithGA [0] (
20:54:15linuxstbjhMikeS: Hmm... Can you do "make clean" and then "make &> log" and upload the log file somewherE?
20:55:17preglowi got one of those with mingw too
20:55:37ToHellWithGAhey yall. i have the most recent 30GB black ipod video. rockbox runs great but when i plug in to usb without putting the device in the "hold" position it neither charges nor syncs. am i doing something wrong?
20:55:57rasherpreglow: would it be okay if I committed .talk file generation before your commit?
20:56:03 Quit bertrik ("bye")
20:56:04preglowrasher: 100%
20:56:09preglowrasher: i in fact welcome it
20:56:16PaulPositionwebguest98 - I "kind of" understand, but I'm not familiar with rockbox source.. I guess you'd have to browse the files #included for the call you're looking for.. dunno, really. I'm no programmer, I just have a basic understanding.
20:56:42rasherI think I have it working, but I bet there'll be bugs. Those can be sorted out in time I guess
20:56:54preglowi'm concocting a small mail to the users list now
20:57:14PaulPositionwebguest98 - Maybe start with browsing the various 'menu' files looking for what happens when you insert from context menu and see if you can call it right away from a button? I'm lost, really. :p
20:59:34*linuxstb just compiled rbspeexenc on Mac OS X - no warnings, worked perfectly...
21:01:29linuxstbjhMikeS: Did you remove the $(SPEEXSRC) after -iquote when you removed -iquote?
21:02:18rasherpreglow: there we go
21:03:01preglowrasher: excellent, delete the bash ones while you're at it?
21:06:15jhMikeSuh, nope
21:06:42jhMikeSbut I you saying I should, or you thought I did?
21:06:54 Join hannesd_ [0] (
21:07:25rasherpreglow: done
21:07:29linuxstbI'm saying you should - it was a parameter to the iquote option... Or just do svn update - I've committed the change to SVN.
21:07:38jhMikeSah, np
21:08:28 Join Falen [0] (
21:08:30***Saving seen data "./dancer.seen"
21:08:59preglowjhMikeS: ok, some testing on me now, and i'm good to go
21:09:08jhMikeSok, now I've got one
21:09:34linuxstbjhMikeS: rbspeexenc compiled for you?
21:10:00jhMikeSyes...a bit verbosely :)
21:10:46linuxstbThere are too many source files there... But I guess we could shut it up.
21:10:46preglowdiff still pukes on memswap
21:10:58preglowclaims it detects a reversed patch
21:11:19jhMikeSthat's weird
21:12:03preglowit's just a plain svn diff?
21:12:40jhMikeSyeah, should be. I don't think I changed the file then did 'svn rm' or would've complained about local changes.
21:12:50preglowif i tell it to apply it nonetheless, it fails on the entire hunk
21:13:22jhMikeSblah, just pull those last hunk out then
21:14:00amiconnping (Bagder_, Zagor)
21:14:35jhMikeSfrankly it doesn't matter, the files won't be compiled
21:14:42amiconnZagor: seems to have problems
21:14:55*amiconn thinks it's cursed
21:14:57preglowjhMikeS: yeah, know
21:15:03preglowjhMikeS: but ok, we're going for this now soon?
21:15:09preglowi'll fire a mail if so
21:15:30*linuxstb rehearses his speex-voice-codec dance
21:15:40jhMikeSsynchronize watches and begin at T- what?
21:16:04Zagoramiconn: looking
21:16:12linuxstbDon't we need Bagder to perform some build-server magic?
21:16:17*amiconn would like to do the lcd speedup dance, but cannot do so without a G4/mini G1 tester :(
21:16:29*jhMikeS is poised over the launch key
21:16:29amiconntierra: around now?
21:16:41tierrayeah, but I'm at work
21:16:46rasherlinuxstb: depends whether rbspeexenc will compile automatically
21:16:57 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
21:16:58rasherI'm still a bit unclear on whether or not that happens
21:17:16linuxstbrasher: It should do. At least, that's what I think preglow setup.
21:17:26rasherThen there should be no changes needed
21:17:40jhMikeSwas that committed? I thought the configure addition wasn't
21:17:59linuxstbAh probably not yet...
21:18:05preglowlinuxstb: i was wrong, it turns out
21:18:15tierraamiconn: while I have a lot of experience with programming (mostly C++), I am just a Rockbox user, I've never touched the source, just so you're aware... if the build system isn't easy to use, I'm less likely to help with testing past the daily builds on the site
21:18:39amiconntierra: Ohkay, I'll wait if nobody else jumps in before
21:19:05amiconnI can also provide test builds
21:19:45tierrathat would probably be helpful, but if it is just a simple autoconf and make, then that won't be a problem
21:19:50preglowlinuxstb: i added rbspeexenc to all the swcodec toolsets in configure, that should do it, non?
21:19:57 Quit Nico_P (Remote closed the connection)
21:20:03preglowtierra: you'll need new compilers
21:20:15linuxstbpreglow: I don't think that's the right place - it's only needed for voice files...
21:20:31preglowlinuxstb: configure allows you to make voice files
21:20:34PaulPositionHmm.. Is it normal that I'd get only +20db or 0db Gain in recording screen of H10 ? (also, automatic gain isn't implemented for that target?)
21:20:50linuxstbpreglow: Yes, but we don't want to build rbspeexenc when doing a normal build
21:20:54 Join bistouri [0] (i=58a10615@gateway/web/cgi-irc/
21:21:03linuxstb(only when doing a "make voice")
21:21:07jhMikeSPaulPosition: That's all the WM8731 provides
21:21:16ToHellWithGAdo yall have any thoughts on why i can't connect or charge via USB when running rockbox but can when my default apple firmware boots?
21:21:30amiconnpreglow: rbspeexenc compiles as-is on cygwin, but is probably not an independent win32 executable (i.e. requires cygwin1.dll)
21:21:37rasherToHellWithGA: Rockbox doesn't support USB on the ipod
21:21:40PaulPositionjhmikes - Ah, well good to know. I'm a bit surprised but, heh! :p thanks.
21:21:42preglowlinuxstb: then where to put it :/
21:21:45amiconnWill test, and adjust if I can figure it out
21:21:49 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
21:21:57tierraamiconn: I'm on Gentoo at home btw
21:21:57 Quit bistouri (Client Quit)
21:21:59ToHellWithGArasher: horseshit
21:22:03jhMikeSPaulPosition: The gains are the same as on not really unusual
21:22:04ToHellWithGAit's worked fine before
21:22:17rasherToHellWithGA: Then you must have magic hardware.
21:22:19 Quit hannesd (Read error: 110 (Connection timed out))
21:22:20 Nick hannesd_ is now known as hannesd (
21:22:39tierrathough I have a Windows install with Cygwin on it as well
21:22:42amiconnpreglow: It needs cygwin1.dll atm
21:22:48ToHellWithGAit charged by usb and allowed the device to be mounted if i didn't hold a button when it was plugged in
21:22:53ToHellWithGAnow it doesn't
21:23:06preglowamiconn: no need for that, it should work just dandy without
21:23:12Zagoramiconn: same problem as before, tolls/ucl/src/libucl.a was old/wrong version. I only cleaned out old .o files, and forgot the .a file.
21:23:17rasherpreglow: line 1975 (but only when building voice)?
21:23:23PaulPositionjhmikes - Oh I believe you. Not saying it's surprising, just that *I* am surprised. ;)
21:23:41amiconnpreglow: Well, atm it doesn't. Needs special parameter for gcc to avoid this
21:23:47*preglow goes for his first speex voice test spin
21:23:58rasherToHellWithGA: so you mean that it doesn't reboot when you plug usb?
21:24:02ToHellWithGAso what you're saying is the image of a usb cable that was on the screen of my ipod while rockbox was connected via usb was some figment of my imagination
21:24:05linuxstbpreglow: I'll look at it...
21:24:18ToHellWithGAit doesn't mount when i plug usb and i'm in the rockbox firmware
21:24:32ToHellWithGAit mounts when i plug usb and i'm in the apple firmware
21:24:40ToHellWithGAit used to not care :/
21:24:47amiconnpreglow: In fact all tools built by tools/Makefile are like that - but that means the .exe will work for .voice building within cygwin
21:25:17 Quit jhulst (Read error: 110 (Connection timed out))
21:25:19amiconn...because wavtrim and voicefont are built the same way, i.e. require cygwin1.dll - and they work too
21:25:48preglowwell, that wasn't exactly a success
21:25:49 Quit desowin ("use linux")
21:26:18preglowit says absolutely nothing
21:26:32jhMikeSpreglow: what's wrong? do the files have version 4.00?
21:26:34rasherTime to fire up a logf build..
21:26:43 Join Jeton [0] (n=chatzill@
21:26:47preglowjhMikeS: i have no idea?
21:26:54preglowif that is something i need to fix: no
21:27:05jhMikeSchange the first 32-bit word to 00 00 01 90
21:27:11jhMikeSin the voice file
21:27:39preglowyes, but i just generated it, do i need to change voicefont.c or something to make it do so all the time?
21:27:48preglowyeah, lookie here
21:27:52preglowit's 3.0
21:28:02jhMikeSthe patch should have the update
21:28:04 Quit MethoS- (Read error: 104 (Connection reset by peer))'d that not get included?
21:29:20amiconnlinuxstb: Do you have an idea what would be the best way to introduce another CFLAGS flag for some tools on cygwin?
21:30:01jhMikeSheh. when I reverted the tools directory I didn't change that back to 400 :p
21:30:29preglowstill no dice
21:30:51preglowthe first 32 bit word is 00 00 01 2c
21:30:57 Join bertrik [0] (
21:31:00 Quit Clam ("...")
21:31:01jhMikeSmake it 190 then
21:31:11jhMikeSyou mean after voicefont?
21:31:17preglowyeah, but why didn't my update work
21:31:29preglow value = SWAP4(400); /
21:31:32preglowshouldn't that work?
21:32:02jhMikeSsays if you alter talk, then sync it with voicefont.c
21:32:49ryanakcalinuxstb: ok, so, format /dev/sdf2... and then reinstall rockbox? Will that mess up the apple firmware?
21:32:50preglowi thought make voice overwrote
21:32:51preglowbut no
21:32:52preglowit doesn't
21:33:36ToHellWithGArasher: could it be an issue of using usb rather than usb2 or vice-versa?
21:33:40preglowhint: roloing rombox.iriver doesn't work
21:34:08preglowjhMikeS: \o/
21:34:13rasherToHellWithGA: Quite unlikely. I'm probably not much help in your issue, as I know next to nothing about the ipods.
21:34:21preglowthis bloody rules
21:34:23preglowsounds really nice
21:34:31rasherNow to make rbspeexenc build automatically on make voice...
21:34:55rasherCan't you simply add it to $toolset in configure?
21:35:19 Quit amiconn (Nick collision from services.)
21:35:27 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
21:36:17preglowlucky onj
21:36:24preglowhe'll have to recreate those files
21:36:34amiconnlinuxstb: Adding -mno-cygwin to both compilation and linking in rbspeex/Makefile works. (1) Is there a reason why linking doesn't use CFLAGS? (2) What woul dbe the best way to add this for cygwin only?
21:37:03preglowbut ok, now i need to test .talk file generation
21:37:11 Part webguest98
21:37:16amiconnpreglow: 'make voice' creates a new file with a numeric modifier
21:37:33preglowamiconn: yeah, i noticed :)
21:37:37rasherpreglow: if [ "$swcodec" = "yes" ]; then toolset="${toolset} rbspeexenc"; fi
21:37:44amiconnIt has always been that way...
21:37:47linuxstbrasher: I'm looking at the Makefile for voice to build rbspeexenc - I want to avoid building it for normal builds.
21:38:03rasherlinuxstb: that line should do it, I believe
21:38:16rashereh, inside some voice-related block
21:38:17linuxstbrasher: Won't that always build it?
21:38:21amiconnlinuxstb: We don't do this for the other tools, e.g. voicefont, wavtrim
21:38:39rasherlinuxstb: depends where you place it :)
21:38:39preglowamiconn: well, sure, but i have never before made a voice file
21:39:03linuxstbamiconn: Any reason why not? I would want them to be a dependency of voice:, and nothing else.
21:39:09amiconnThe tools are only built once unless they're updated, so I wouldn't care much
21:39:32preglowrasher: so, the way to make .talk files is by direct invocation of ?
21:39:39rasherlinuxstb: just do it in line 305
21:39:39amiconn'make zip' being a shell-in-perl script is way more annoying....
21:39:54LloreanSpeaking of voices, do we have a plan to provide official voices in different synthesizers, or do we want to pick one and go with it?
21:39:59rasherpreglow: that was the way with as well
21:40:21rasherlinuxstb: just after toolset="${toolset} voicefont wavtrim"
21:40:26amiconnLlorean: Different languages would be even more important than different synthesizers imho
21:40:34preglowrasher: " -C -e=rbspeexenc -s=espeak /media/myplayer" or something?
21:40:48Lloreanamiconn: Different languages would be very nice.
21:40:49rasherpreglow: I believe you need to pass -E and -S as well (even if empty)
21:40:55rasherGuess it could be more flexible about that
21:41:15rasherpreglow: But otherwise, yes
21:41:35preglowseems to work
21:41:39amiconnhrrrm :/
21:42:02Lloreanamiconn: But our current voice files sound like crap, even in English, and I was wondering if we wanted to try to pick a good synth, or just let people pick from a selection of bags of shit, or what.
21:42:19rasherLlorean: subjective..
21:42:24preglowouch, core dumo
21:42:27 Quit jgarvey (Read error: 104 (Connection reset by peer))
21:42:49preglowit dumped core all over my laĂp here
21:42:50Lloreanrasher: I don't think anyone _really_ thinks Espeak, Festival or Flite sound "good"
21:43:02LloreanIt's generally a selection of least bad.
21:43:02rasherLlorean: do we have any other options?
21:43:31LloreanI don't know.
21:43:54LloreanWe have espeak+mbrola obviously. But there are also windows voices, if we wanted to do a daily voice build and could get some sort of volunteer.
21:45:24preglowrasher: what's it supposed to create, really?
21:45:28preglowi didn't get this
21:45:36preglowit made a ton of before coredumping
21:45:38Zagorumm, can I use UNCACHED_ADDR for iram addresses?
21:45:51rasherpreglow: the .talk files are expected. The core dump is not.
21:45:52Zagor(I know I don't have to, but will it work?)
21:45:56preglowbut all of them just say "talk mp3 clip"
21:46:06amiconnLlorean: Afaik the only free (beer) voices are the (english) microsoft voices, plus some really old L&H voices
21:46:06jhMikeSZagor: that won't work
21:46:06rasherThat's what they speak?
21:46:22ZagorjhMikeS: ok, thanks
21:46:37rasherpreglow: hahaha, well, I'll have a look at that. Not sure what to do about the core dump - I doubt it's my fault, to be honest
21:47:00amiconnThe latter sound like crap, but there are voices for english, french, german and spanish iirc
21:47:36Lloreanamiconn: We don't necessarily need free (beer) voices. We can also look for non-free voices that absolutely clearly allow redistribution, and see about getting donations to purchase one for whoever the voice builder ends up being.
21:48:02amiconnIs the OSTYPE env var available anywhere?
21:48:24*pondlife likes the old L&H voices best...
21:48:50linuxstbrasher: This is what I had in mind for tools/configure - - it only builds the voice-related tools when doing "make voice"
21:49:03pondlifeYes, clearer than any MS voice .
21:49:21preglowrasher: actually, it looked like _sh_ did the core dumping
21:49:42amiconnThe L&H voices are so incredibly slow they are unbearable, imho
21:49:45rasherlinuxstb: I don't see why you added voicetoolset?
21:49:58amiconnAnd they're sapi4, so building a voice file takes ages
21:50:04 Join bEnN [0] (
21:50:14rasherlinuxstb: toolset only includes wavtrim and voicefont when building voices.. Why not simply add rbspeexenc if building voices AND swcodec is set?
21:50:19linuxstbrasher: Because voice tools are independent of the normal build tools.
21:50:26pondlifeamiconn: I have an L&H Michael which is SAPI5
21:50:35 Join ttuttle [0] (n=tom@pdpc/supporter/student/gentoo.contributor.ttuttle)
21:50:41amiconnThere seem to be sapi5 versions as optional components for the microsoft reader, but I couldn't get those to work
21:50:45ttuttleHey, how can I encode video for the Sansa e200 running rockbox?
21:50:50pondlifeYep, that's the one
21:50:51linuxstbrasher: The principle elsewhere in tools/configure is to only build the tools required for that build - my patch extends that principle to voice.
21:50:57Lloreanttuttle: See the PluginMpegplayer wiki page.
21:51:00pondlifeMuch better than the SAPI4 version
21:51:13ttuttleLlorean: I did. It tells me what format but not how.
21:51:35Lloreanttuttle: Odd, because it not only mentions how but provides links to at least 2 or 3 programs you can use...
21:51:47preglowjhMikeS: what happens if we encounter old .talk clips now?
21:52:10amiconnlinuxstb: Could you check whether $OSTYPE exists on macos?
21:52:12ttuttleLlorean: Oh. I'm sorry, I didn't see the part at the bottom.
21:52:16linuxstbamiconn: To answer your earlier questions - 1) No reason for not including CFLAGS - it works fine for me if you add it to the linking step; 2) Can't you just do the same check for cygwin that you did for ipodpatcher, and add -mno-cygwin to CFLAGS?
21:52:18ttuttleLlorean: That's actually a wonderful wiki page.
21:52:39Lloreanttuttle: I personally recommend WinFF, as long as you use version .29
21:52:44rasherlinuxstb: I suppose
21:52:51preglowjhMikeS: you need to check out all the other stuff as well, .talk clips and dir spelling works funny here
21:52:52ttuttleLlorean: Linux here.
21:52:56preglowjhMikeS: filename spelling doesn't work at all
21:53:03*amiconn checks
21:53:03Lloreanttuttle: WinFF is actually a Windows or Linux front end for ffmpeg.
21:53:04 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
21:53:12 Quit Falen ("ChatZilla [Firefox]")
21:53:18ttuttleLlorean: Oh. That's a strange name for a program that works on Linux.
21:53:22linuxstbamiconn: Yes, it exists and for me is "darwin8.0"
21:53:26LloreanIt was originally a Windows program, I guess.
21:53:27preglowjhMikeS: ah, no, it works if i disable .talk clips
21:53:29ttuttleLlorean: Ah.
21:53:40jhMikeSpreglow: it simply plays what's asked of it. it does spell the keyboard and the info screens ok.
21:53:48amiconnMaybe checking $OSTYPE would be better (in general) than spawning uname?
21:53:49LloreanEither way, you do encounter problems with too-new ffmpegs on Linux. I keep planning to update the presets for .31 of WinFF but haven't done it yet.
21:53:57ttuttleBy the way, RockBox is awesome.
21:54:07amiconnOn linux it says 'linux-gnu', on cygwin just 'cygwin'
21:54:09ttuttleHave you received the influx of Woot customers who just got their players?
21:54:13preglowmy first hang :P
21:54:23ttuttlepreglow: It's fun. Hold down power forever and it will reboot.
21:54:28ttuttlepreglow: (Or at least power off.)
21:54:42amiconnMight be bash-only...
21:54:52preglowttuttle: nah, needed the paper-clip
21:54:52ttuttleLlorean: Hey, can ffmpeg run on two CPU's? Or at least split video and audio across CPU's? I'm dual core.
21:54:55ttuttlepreglow: Ah.
21:55:01preglowand it's my first voice hang, i've been hanging steadily for years now :P
21:55:01amiconnAnyone who uses dash (ubuntu users)?
21:55:05preglowamiconn: yes...
21:55:20*BigBambi sticks up the ubuntu hand (does that mean I use dash :))
21:55:20jhMikeSpreglow: hang doing what?
21:55:26preglowjhMikeS: browsing menus
21:55:34amiconnpreglow: 'echo $OSTYPE' ?
21:55:40Lloreanttuttle: Varies from codec to codec, unfortunately. As opposed to splitting the video/audio decoding, it'll attempt to encode one or the other using two threads, but depends on the codec supporting it
21:55:42 Join Shoot [0] (
21:55:47preglowthomj@pocketsandwich:~$ 'echo $OSTYPE'
21:55:47preglowbash: echo $OSTYPE: command not found
21:55:48ttuttleLlorean: okay
21:55:49ShootJust wondering
21:55:51LloreanAfaik, mpeg2 support and mp3 support in ffmpeg aren't multithreaded
21:55:54ttuttleLlorean: what about the stock encoding format they suggest?
21:55:56ttuttleLlorean: ok
21:55:58preglowamiconn: wait a sec...
21:56:05ShootCan anyone help me with my bricked Sansa?
21:56:07preglowamiconn: dash says the same
21:56:07jhMikeSpreglow: really shouldn't be possible from the thread's pov.
21:56:10BigBambiamiconn: does ubuntu use dash by default?
21:56:13ShootI'll give you all the details
21:56:16ShootCan anyone help me with my bricked Sansa?
21:56:26ttuttleShoot: Don't spam please.
21:56:30Shoot1. Recovery Mode and Manufacturing Mode are not functioning
21:56:31Lloreanttuttle: If you look at the presets.xml for WinFF, and find the right entry in it, it's almost exactly what command line parameters to use with ffmpeg from the terminal
21:56:34preglowamiconn: without the quotes it just gives me blank space
21:56:40BigBambiIf so: alex@MaximusLappy:~$ echo $OSTYPE
21:56:48preglowjhMikeS: perhaps it's not voice's fault
21:56:50Lloreanttuttle: The big difference in newer ffmpegs is that they renamed some of the codecs between versions
21:56:55jhMikeSpreglow: which player?
21:56:57rasherpreglow: odd.. talk clips seem to work here
21:56:59amiconnIirc someone here said it does (rasher?). I am running debian on my linux box
21:57:09Shoot2. Is there any way to fix my Sansa?
21:57:13ttuttleLlorean: Ah.
21:57:16ShootI'm downloading Ubunto now
21:57:21BigBambiamiconn: Silly question - how do I check?
21:57:24LloreanShoot: How do you know Manufacturing mode isn't functioning?
21:57:25Nico_PBigBambi: echo $0 should tell you what shell you have
21:57:34jhMikeSrasher: I wonder if talk clips properly override directory spelling
21:57:36preglowjhMikeS: h120
21:57:38ShootLlorean: I tried Hold/Select/USB
21:57:39BigBambiNico_P: That'll be bash then
21:57:42Lloreanttuttle: mp3 became lamemp3 or mp3lame or something like that, breaking front ends.
21:57:47ttuttleLlorean: libmp3lame ;-)
21:57:51LloreanThat's it.
21:57:51ShootLlorean: It didn't work =\
21:57:54ttuttleLlorean: Yeah, I tried mp3 first and it didn't work.
21:57:56scorche|wshoot: and what happened?
21:57:58LloreanShoot: "Didn't work" is vague. What happened?
21:58:00jhMikeSpreglow: couldn't manage that all day myself...really odd
21:58:06rasherjhMikeS: I don't think that's my area
21:58:16ShootLlorean: I placed an OF.bin file in 16MB-format
21:58:20ttuttleLlorean: ffmpeg is surprisingly easy to use. mencoder is far more complicated.
21:58:21Nico_PBigBambi: but on my system, /bin/sh links to dash
21:58:25ShootLlorean: And now it won't turn on
21:58:27jhMikeSrasher: preglow said it acted odd having them both enabled at the same time
21:58:27Lloreanttuttle: In fact, a find/replace on mp3 to libmp3lame should actually fix the existing presets.xml to work with newer ffmpegs, assuming unlike ubuntu yours *has* MP3 support
21:58:31*BigBambi checks
21:58:38ShootLlorean: I think I actually damaged a .Rom file
21:58:40LloreanShoot: Manufacturing mode doesn't have a 16MB-format partition.
21:58:53ttuttleLlorean: Yeah, mine has MP3 support. And it's Debian. Which means Ubuntu purposefully broke it.
21:58:53LloreanShoot: Again, what does "Doesn't work" mean when you try manufacturing mode?
21:59:06rasherpreglow: I suspect massive weirdness. Both directories and filename talk clips contain whatever they should, here
21:59:08ShootLlorean: Manufacturing mode is hold + select + USB?
21:59:13Lloreanttuttle: I'm Ubuntu, but I complie ffmpeg from svn anyway. :)
21:59:28rasherpreglow: I've not tested it on target, but playing the .talk files with mplayer yields expected results
21:59:42rasherjhMikeS: well, then I blame rockbox speech code, which I have never touched
21:59:55ShootLlorean: I tried that =\
22:00:03LloreanShoot: Please, JUST answer my question: WHAT HAPPENS.
22:00:03scorche|wand what happened?
22:00:05ShootLlorean: Now the wheel light won't even turn on
22:00:16ShootLlorean: Nothing, if I turn on manu, wheel light doesn't work
22:00:20ShootLlorean: Display as well
22:00:29LloreanDisplay's not supposed to come on.
22:00:34ShootLlorean: Yeah
22:00:43 Part hoobop ("leaving")
22:01:00ShootLlorean: You need any other details?
22:01:16amiconnHmm, OSTYPE seems to cause trouble sometimes. Looks like I should go for the old uname method
22:01:25LloreanIf your wheel light doesn't come on in that situation, you're basically out of luck.
22:01:29ShootLlorean: I've tried the search function on forums; I found a neat thread
22:01:33ShootLlorean: FUCK
22:01:37LloreanPlease, watch your language.
22:01:42ShootLlorean: Sorry
22:01:54jhMikeSare these "e200r" things sold as plain-vanilla really e200r or some other odd combo?
22:02:02LloreanAnd in the future, don't do stupid things like putting files other than those our instruction say in recovery partitions. It's never a good idea to experiment if you don't have a way to undo the damage.
22:02:09preglowrasher: you're probably doing the right thing here
22:02:13preglowrasher: the files vary in size
22:02:17preglowbut all that is said is "talk mp3 clip"
22:02:25preglowperhaps there's some "voice file type" bug going on here?
22:02:25ShootLlorean: Sansa's warranty is 1 year?
22:02:57rasherpreglow: Hrm, I've no idea, but it sounds like it isn't my bug, and I really know very very little about the actual voice code
22:03:07*jhMikeS hasn't tried a rockbox install on his woot special yet
22:03:11ttuttleLlorean: Does the MPEG2 video player ever crash?
22:03:18ttuttlejhMikeS: you should. is it rhapsody?
22:03:23scorche|wjhMikeS: mine was a genuine R
22:03:26 Quit krazykit ("Connection reset by beer")
22:03:35LloreanjhMikeS: They seem to either be real e200Rs, or vanilla e200s with an e200R physical shell.
22:03:39jhMikeSit say "R" on the back and has Rhapsody fw
22:03:45scorche|was was krazykit's
22:03:51Lloreanttuttle: In Rockbox? If it does, it's a bug...
22:03:52ttuttleLlorean: Mine doesn't say 250R, and isn't 250R.
22:03:57ttuttleLlorean: Yeah. I found a file.
22:03:59scorche|wwell, if it has the rhapsody firmware, it is an R
22:04:00amiconnlinuxstb: Thanks, it works :)
22:04:02preglowjhMikeS: well, test with .talk files
22:04:03ttuttleLlorean: Unfortunately I can't send it to you to test.
22:04:18ttuttleLlorean: (Copyright 'n' stuff...)
22:04:22*amiconn forgot that he added such a check not too long ago :\
22:04:29BigBambiNico_P: /bin/sh is indeed a link to dash, but gnome-terminal (which is what applications -> accessories -> terminal points to) is bash
22:04:32Lloreanttuttle: Well, most files that will crash it will do so because the file has stream errors, I hear.
22:04:37ttuttleLlorean: hmm
22:04:39pregloware the different options supposed to be cascaded in some way, so that if i have "voice file .talk clips" and "voice filenames" enabled, both are said?
22:04:41LloreanIt's more that our player is less error-resistant than others.
22:04:48jhMikeSpreglow: you locked with .talk files? I don't want to destroy my debian install at this time. I'll just say things into a mic and be happy.
22:05:04ttuttleLlorean: Plays fine on my PC.
22:05:18Lloreanttuttle: That may just mean that your PC based player is better about ignoring errors.
22:05:23 Join donutman25 [0] (n=chatzill@
22:05:23ShootHmm, does anyone know how long Sansa e200's warranty lasts?
22:05:30ttuttleLlorean: It's mplayer. It usually prints an error.
22:05:33scorche|wShoot: ask them...not us
22:05:35*Nico_P is off
22:05:37ttuttleShoot: Check the Woot blog, it usually says. Or email Woot.
22:05:40Lloreanttuttle: Did you encode the file yourself?
22:05:44ttuttleLlorean: Yeah, using ffmpeg.
22:05:46ShootWoot blog?
22:05:48ShootI'll google it
22:05:51Lloreanttuttle: ps or ts?
22:05:52preglowjhMikeS: i don't the talk files had anything to do with that
22:05:53ttuttleShoot:, click on "blog".
22:05:56ttuttleLlorean: Huh?
22:06:00*BigBambi points at the topic
22:06:01preglowjhMikeS: but .talĂk files don't seem o work here anyway
22:06:16Lloreanttuttle: program stream or... transfer I think stream?
22:06:24ShootAlso : Does anyone here know if e200R > e200?
22:06:25ttuttleLlorean: I don't know what either of those mean.
22:06:35ttuttleShoot: e200r < e200, unless you want Rhapsody.
22:06:50scorche|wit doesnt really matter, as you change change them...
22:07:00scorche|wcan change
22:07:13Lloreanttuttle: Basically, did you use the exact same command line parameters I used in the presets.xml on the wiki, or did you add or change any?
22:07:20ttuttleLlorean: I think the same. Lemme try again.
22:07:52Lloreanttuttle: If they are the same, file a proper bug report, and try to generate a short clip from content you feel comfortable sharing that reproduces the crash.
22:08:02preglowamiconn: how will "make voice" work on cygwin?
22:08:06jhMikeSpreglow: maybe the .talk code was trying to be crafty to control voice but not voice just instantly plays whatever is fed to it.
22:08:16amiconnpreglow: What do you mean?
22:08:23preglowamiconn: will it work fine as it is?
22:08:37ShootWhat is Rhapsody?
22:08:42amiconnThe svn version (w/o rbspeexenc) is working fine
22:08:57amiconnI'm using it for all my voice files (using at&t voices)
22:09:13scorche|wShoot: ask google...
22:09:37preglowamiconn: well, i'm wondering if it'll work fine with it
22:09:47ttuttleLlorean: Argh. Suddenly it's not accepting -acodec libmp3lame
22:10:01Lloreanamiconn: Assuming the changes were wrapped in the proper #if or #ifdef can you think of any problems the "Nano IDE0 fix" might cause, assuming it seems to work for testers over the next couple days?
22:10:01*Llorean would rather get it in asap if possible.
22:10:23amiconnIf it works on all 'reachab
22:10:42amiconnle' nanos, both affected ones and unaffected ones, I think it should go in
22:10:50ShootErm o___o how does the Sansa e200r connect with Rhapsody? USB?
22:11:16Lloreanamiconn: It certainly works on my unaffected Nano fine.
22:11:25amiconn(#ifdef IPOD_NANO, and with a comment saying this is somewhat hackish)
22:11:38scorche|wShoot: please stay on topic...that has nothing to do with rockbox
22:11:47ttuttleShoot: Also o___o is not a word
22:12:11amiconnpreglow: If added properly it should just work
22:12:34ttuttleLlorean: Okay, ffmpeg is saying "Unknown decoder 'libmp3lame'" if I use -acodec libmp3lame, and "Unknown encoder 'mp3'" if I use -acodec mp3. I'm confused. I got it to work before!
22:12:40amiconnThe only difference is that on cygwin, the encoder (and also wavtrim) is spawned from the vbscript instead of the perl script
22:12:47Lloreanttuttle: Emoticons can be considered modern virtual punctuation useful in replacing the expressiveness in human body language, facial expression, and intonation for the purpose of online communication, assuming they aren't abused. :-P
22:12:50amiconnThat saves the cygwin process creation overhead
22:13:18Lloreanttuttle: Honestly, no clue. :) Sorry.
22:13:21ryanakcalinuxstb: hmmm... formatting fixed it... for now :)
22:13:24ttuttleLlorean: I know. I consider any text that matches \w+ in a regular expression to not be an emoticon ;-)
22:14:11ttuttleLlorean: If I put -acodec libmp3lame at the end it works. Who knew?
22:14:52 Join Snake_ [0] (
22:14:56ttuttleOoh, I'm getting almost 300 FPS encoding.
22:15:32Snake_Hey guys, quick question: Are there any USB Host applications for Rockbox that you know of?
22:15:58BigBambiSnake_: That would surely require us having a software USB stack?
22:16:51Snake_BigBambi: no idea :P im trying to get my PSP to connect to it, and the guys in the PSP forums told me to find some host for the ipod in order to connect to it
22:17:04BigBambiSnake_: I ws implying we don't
22:17:41Snake_BigBambi: would it be difficult to recomplie usbhostfs for it??
22:17:48BigBambiFor rockbox?
22:17:58BigBambiYou do know rockbox is not based on linux right?
22:18:17BigBambiTo compile anything for rockbox requires it to be ported to it
22:18:19ttuttleBigBambi: There seems to be a bit of a rumor. Someone else thought that.
22:18:22Snake_Oh thats right
22:18:40preglowrasher: btw, i get some warnings on now
22:18:42BigBambittuttle: Many people think that. Despite the rumours, they are wrong
22:18:46ttuttleBigBambi: I know.
22:18:56ttuttleBigBambi: It would have displayed a penguin when it booted if it were Linux :-b
22:19:01BigBambittuttle: And I know you know :)
22:19:03 Nick Bagder_ is now known as Bagder (
22:19:09BigBambiIndeed so
22:19:10ttuttleBigBambi: And I know you know I know... :stack overflow:
22:19:13Snake_Well I knew it wasnt linux, but I forgot ;-)
22:19:32*ttuttle tries another MPEG2 file.
22:19:39Snake_ive used both and rockbox is way better
22:19:44Snake_more usable, more purpose
22:19:54ttuttleStill hanging.
22:19:56*BigBambi doesn't own an iPod
22:20:42 Part Jeton
22:20:48Lloreanttuttle: With an SVN build from about 30 minutes ago, I can watch a video I encoded with ffmpeg r10481 just fine. =/
22:21:03rasherpreglow: about $V not being set?
22:21:06ttuttleOkay. Lemme build my own ffmpeg.
22:21:16ttuttleLlorean: Do you know the svn repos offhand?
22:22:08ttuttleLlorean: Found it.
22:22:11Lloreanttuttle: No, I found them somewhere on their site and promptly forgot
22:22:11LloreanApparently I'm 100 revisions old on ffmpeg. Heh.
22:23:35preglowrasher: aye
22:24:19rasherpreglow: I noticed that - should be harmless (but I'll fix)
22:25:22 Part pondlife ("Gone")
22:25:29ttuttleLlorean: Wow, this takes a while to build.
22:25:36 Quit Siku ()
22:25:41ttuttleLlorean: How hard is it to build Rockbox?
22:25:51ttuttleLlorean: Or should I just download it again?
22:26:07Lloreanttuttle: It's kinda off-topic here, though
22:26:27*BigBambi declares rockbox easy to build
22:26:28LloreanIf you download from the "current build" page you'll have the most recent SVN revision of Rockbox.
22:26:35LloreanBut it's darn easy to build yourself too.
22:26:47amiconnrasher: Hmm, looks like the .talk clip creation in doesn't check file extensions at all
22:26:53ttuttleLlorean: Okay.
22:27:00amiconnIt would even create .talk clips for .talk clips....
22:27:07rasheramiconn: no
22:27:11amiconnVoiceBox only voices supported filetypes
22:27:18ttuttleLlorean: I think I didn't encode as mp3.
22:27:23rasheramiconn: it checks for =~ /\.talk$/
22:27:29preglowrasher: found a bug
22:27:37ttuttleLlorean: It simply won't accept either acodec option. One is invalid for encoding, one is invalid for decoding (why is it decoding?)
22:27:38rasheramiconn: I noticed only supported filetypes are voiced, but why on earth is that?
22:28:07ttuttleLlorean: Will -acodec mp2 work?
22:28:11amiconnWhy would you want an unsupported file with a .talk clip?
22:28:31Lloreanttuttle: It should.
22:28:32rasheramiconn: because you might be showing all filetypes in the browser?
22:28:35ttuttleLlorean: Okay.
22:28:42LloreanDon't forget the sample rate _must_ be 44100 unless things have changed.
22:28:48amiconnThen you can still have it numbered or spelt
22:28:56rasher.. but not voiced
22:29:08Lloreanamiconn: I think the option for voicing all files should still be offered, if someone has the space and desire.
22:29:33rasheramiconn: then you might as well say the same for all files. Why offer .talk clips at all when you can just have it numbered or spelt?
22:29:36ttuttleLlorean: Hmm.
22:29:40ttuttleLlorean: Yeah, it is.
22:29:59amiconnrasher: In fact I don't use file .talk clips, only dir clips
22:30:17ttuttleLlorean: Hmm. I wish MP3 would work.
22:30:25amiconnFile .talk clips are wasting too much space imho
22:30:27rasheramiconn: That's you, though. Not everyone.
22:31:47ttuttleLlorean: Could it be because my input file has MP3 audio as well?
22:33:12preglowjhMikeS: so, have you got some .talk stuff going?
22:33:34ShootGuys, does removing the battery off my Sansa e200 help it recover from the bricking?
22:33:45ShootSomeone on the Sandisk help line told me to
22:34:01preglowall i can make it say is "talk mp3 clip"
22:34:19rasherpreglow: and you're sure you're not selecting an actual talk clip?
22:34:45Lloreanttuttle: Never been a problem for me.
22:34:45Lloreanttuttle: What player are you encoding for?
22:35:09ttuttleLlorean: Sansa e250.
22:35:13preglowrasher: what do you mean by that? i'm hovering over a talk file when it does that, yeah,but why does it do that?
22:35:33preglowrasher: it doesn't explain to me what any of the other file types are, and i don't even have that feature enabled, not that it makes a difference
22:35:53Lloreanttuttle: -acodec libmp3lame -ab 128 -ar 44100 -vcodec mpeg2video -s 224x176 -b 320kb -strict -1 works for me, with -s adjusted if you're doing 16:9
22:36:17LloreanFor encoding, haven't tried it on the player.
22:36:18rasherpreglow: well, talk files don't have talk clips generated for them - hovering over the corresponding "regular" file should voice the contents of that .talk file
22:36:38rasherpreglow: only supported (mp3, ogg, flac, etc) files get voiced though
22:36:54ttuttleLlorean: What version of ffmpeg?
22:36:56preglowit still shouldn't say "talk mp3 file" unless i have "voice file types" enabled, though
22:37:04preglowrasher: this reminds me, there probably is some language to fix now
22:37:04ttuttleLlorean: I'll rollback mine.
22:37:35preglowany good alternatives to ".talk mp3 clip" ?
22:38:04preglow"talk voice clip" doesn't make much sense
22:38:07preglowwhy not just "voice clip" ?
22:38:08Lloreanttuttle: That's with SVN current ffmpeg.
22:38:11ttuttleLlorean: Hmm.
22:38:16rasherpreglow: Why not indeed.
22:38:18 Join mog [0] (
22:38:19ttuttleLlorean: I ran that, and I got "Unknown decoder 'libmp3lame'".
22:38:35ttuttleLlorean: And "WARNING: The bitrate parameter is set too low. It takes bits/s as argument, not kbits/s"
22:38:48mogI have a question about the Gigabeat player
22:38:55ttuttlemog: Ask it.
22:38:58BigBambimog: Just ask
22:39:02mogI see that rockbox is supported on the F series, but what about the S series?
22:39:03preglowrasher: as usual i'm not too clear on the context in voice files
22:39:06BigBambiNot yet
22:39:07preglowso i'll just delay changing it
22:39:11jhMikeSpreglow: trying to do _something_ here
22:39:17BigBambimog: Join in and make it happen
22:39:26LloreanMine accepts kbits/s
22:39:28jhMikeSI don't even know how to use them so have to go read up anyway
22:39:29mogBigBambi: I'm thinking about it
22:39:38toffe82mog: super
22:39:41Lloreanttuttle: Note that I did say 320kb for the bitrate, not just 320
22:39:49ttuttleLlorean: I copied your line verbatim.
22:39:49BigBambimog: You are very welcome
22:39:52preglowjhMikeS: ditto here
22:40:03mogBigBambi: but I got a lot going on, college stuff and the like, I'll be back around if I can, thanks for the support
22:40:08Lloreanttuttle: Odd, it works for me just fine. =/ You're best off asking ffmpeg guys at this point
22:40:13 Part mog ("Time makes no sense")
22:40:54amiconnHrmm. I thought there already was an option in voiceBox, but there isn't
22:41:01ttuttleLlorean: I'm asking there.
22:42:02amiconnrasher: I'd just use '.talk clip'
22:42:25preglowamiconn: the . isn''t pronounced
22:42:40amiconnyeah, sure
22:42:56preglowso it should be "dot talk clip" or soemthing
22:43:05amiconnThere's also a standard language string
22:43:39amiconnImho the dot is redundant for voice - thats why there are both <dest> and <voice> in .lang files
22:45:08 Quit merbanan (Remote closed the connection)
22:45:10 Quit jhulst (Connection timed out)
22:46:24ttuttleLlorean: I'm trying mencoder but it's *not even real-time*!
22:46:43ttuttleLlorean: It's saying 21 minutes to encode an 11 minute clip :-(
22:46:57 Join DerPapst [0] (
22:50:08 Quit karashata ("I will *SO* make you regret that... Later... *is a lazy dragon*")
22:51:25ShootLlorean: Ok. I removed the battery and put it back in.
22:51:32ShootLlorean: Now the mousewheel is working.
22:54:33preglowdo we need anymore?
22:55:26Bagderpreglow: iirc, it can still generate a db on host (for mp3 and ogg)
22:55:39ShootCan anyone explain to me how to run e200tool?
22:55:42LloreanDon't we have an executable for that now, though?
22:55:56BagderLlorean: yes, if that still builds...
22:56:13BagderI'm just saying, I have no opinion
22:56:20BagderI don't use the database
22:56:54rasherhow do you build that tool anyway? (perhaps it should be part of the build table or something?)
22:57:36Bagder"make database" in tools/
22:57:54Bagderand no, it doesn't build...
22:58:50LloreanI like that tool better since it uses Rockbox's own database code, so as long as it's kept building it should always build a valid database, right?
22:58:54ShootLlorean: What should I do from now? The mousewheel is working now.
22:59:11BagderLlorean: yes, and I agree that it is better than the perl version in almost all ways
22:59:25linuxstbLlorean: I'm not even sure it ever worked completely... But it shouldn't need much massaging.
22:59:44 Join mog [0] (
22:59:47LloreanShoot: Follow the instructions on the SansaE200Unbrick wiki page
22:59:49rasherLlorean: plus it uses the same metadata parsing, so it supports all files, and gives the same results
23:00:12linuxstbAnd hopefully the same bugs...
23:00:16preglowBagder: so it has been updated for tagcache?
23:00:27rasherpreglow: yup
23:00:33ShootLlorean: I've tried; I'm at the step where I need to use e200 tool. I've downloaded Ubuntu and the Linux binary. Now what?
23:00:45Lloreanrasher, linuxstb: Honestly, I wouldn't mind seeing that code either incorporated into RBUtil or give it the ability to launch it if the executable is found.
23:01:05LloreanShoot: Follow the instructions on the page...
23:01:21rasherhaving it inside rbutil (using Rockbox code) would be neat
23:01:29LloreanI'm honestly equally capable of reading a wiki page as you..
23:01:41preglowjhMikeS: now?
23:02:09linuxstbLlorean: Are you sure about that Unbrick wiki page? It doesn't tell users to put their e200 into recovery mode (using the bootloader transferred via e200tool), and then having to copy a bootloader and mi4 file to the recovery partition...
23:03:05jhMikeSpreglow: I have to get a way together to generate this stuff
23:03:16jhMikeSwhat exactly was going on?
23:03:23preglowjhMikeS: i can upĂload yu a ton of files, no worries
23:03:36preglowjhMikeS: btw, are talk files supposed to just be raw speex? no header or nothing?
23:04:30jhMikeSheck if I know. the thread takes raw speex and plays whatever the heck is sent to it.
23:04:42Lloreanlinuxstb: Hm, I thought that was the proper recovery page. Shoot, well now I have no clue what page I was thinking of. =/
23:04:47 Quit bEnN (Read error: 104 (Connection reset by peer))
23:04:58linuxstbLlorean: I couldn't find it either...
23:05:14preglowjhMikeS: that's the way the script works, at least
23:05:20LloreanI swear there was a page that told the proper process for recovering a completely bricked Sansa, and I thought that was it
23:05:36linuxstbLlorean: But I think it would make sense for that Unbrick page to be improved - that's where we always send people. (I don't own a Sansa though...)
23:05:40preglowjhMikeS: anyway, then you can just rbspeexenc a couple of wavs and rename them to
23:06:06jhMikeStried that. I might see the calls mp3_info in talk_file
23:06:16Lloreanlinuxstb: I've never actually managed to brick my Sansa, but I bet there's a lot of stuff I could copy and paste from the old e200R install page or something
23:06:20preglowwell ouch, then
23:06:31preglowmake it stop that