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 2007-11-21

00:00:08 Join freqmod__nx [0] (
00:01:07 Join MethoS-- [0] (
00:01:07 Join thegeek_ [0] (
00:01:07 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
00:01:07NJoinArathis2 [0] (
00:01:07NJoinXerion [0] (
00:01:07NJoinSoap [0] (n=Soap@rockbox/staff/soap)
00:01:07NJoinlodesi_ [0] (
00:01:07NJoinWeiss_ [0] (
00:01:07NJoincrashd_ [0] (
00:01:07NJoincrwll [0] (
00:01:07NJoinatsea- [0] (i=atsea-@gateway/tor/x-b3d0c437d014db79)
00:01:07NJoinscorche|w [0] (n=42c007b2@rockbox/administrator/scorche)
00:01:07NJoinCasainho [0] (
00:01:07NJoinTMM [0] (
00:01:07NJoinjhulst [0] (n=jhulst@unaffiliated/jhulst)
00:01:07NJoinbillenium [0] (
00:01:07NJoinlee-qid [0] (
00:01:07NJoinmarkun [0] (n=markun@rockbox/developer/markun)
00:01:07NJoinFOAD [0] (
00:01:07NJoinprzemhb [0] (
00:01:07NJoinamiconn [0] (n=jens@rockbox/developer/amiconn)
00:01:07NJoinShaid [0] (
00:01:07NJoinhannesd [0] (
00:01:07 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
00:01:07NJoinDerPapst [0] (
00:01:07 Join Bagder [0] (n=daniel@rockbox/developer/bagder)
00:01:07NJoinGnu47 [0] (
00:01:07NJoinJon-Kha [0] (
00:01:07NJoinidnar [0] (n=mithrand@unaffiliated/idnar)
00:01:07NJoinnanok [0] (n=nanok@
00:01:07NJoinpixelma [0] (i=pixelma@rockbox/staff/pixelma)
00:01:07NJoindaurn [0] (n=daurnima@unaffiliated/daurnimator)
00:01:07NJoinmaddler [0] (
00:01:07NJoinroxfan[zzz] [0] (
00:01:07NJoinmale [0] (
00:01:07NJoinzicho [0] (
00:01:07NJoinjoshin [0] (n=josh@unaffiliated/joshin)
00:01:07NJoinparafin|away [0] (
00:01:07 Join GodEater [0] (n=bryan@rockbox/staff/GodEater)
00:01:07NJoinmoos [0] (
00:01:07 Join Rick [0] (i=rick@unaffiliated/rick)
00:01:07NJoinRondom [0] (
00:01:07NJoincourtc [0] (
00:01:07NJoindavina [0] (
00:01:07NJoinkrazykit [0] (
00:01:07NJoinJETC- [0] (
00:01:07NJoinsup [0] (
00:01:07NJoiniamben [0] (
00:01:07NJoinIsolinear [0] (
00:01:07NJoinalienbiker99 [0] (
00:01:07NJoinrasher [0] (n=rasher@rockbox/developer/rasher)
00:01:07NJoinXavierGr [0] (
00:01:07NJoinaustriancoder [0] (n=austrian@rockbox/developer/austriancoder)
00:01:07NJoinbagawk [0] (
00:01:07NJoinHellDragon [0] (i=jd@unaffiliated/helldragon)
00:01:07NJoinbnakiddmj [0] (i=0@
00:01:07NJoinBjoernErik [0] (
00:01:07NJoinlastebil [0] (
00:01:07NJoinsbeh [0] (n=sbeh@
00:01:07NJoinHadaka [0] (
00:01:07NJoinsneakums [0] (
00:01:07NJoinryanakca [0] (n=ryan@ubuntu/member/ryanakca)
00:01:07NJoinDataGhost [0] (
00:01:07NJoinSlasheri [0] (i=miipekk@rockbox/developer/Slasheri)
00:01:07NJoingromit` [0] (
00:01:07NJoinlids [0] (
00:01:07NJointierra [0] (
00:01:07NJoin[mbm] [0] (i=mbm@openwrt/developer/mbm)
00:01:07NJoingtkspert [0] (
00:01:07NJoinsslashes [0] (i=sslashes@
00:01:07NJoinmarcosource [0] (
00:01:07NJointuplanolla [0] (
00:01:07NJoinjmspeex [0] (
00:01:07 Join midkay [0] (n=midkay@rockbox/developer/midkay)
00:01:07NJoinGuerin [0] (
00:01:07NJoinFunkyELF [0] (
00:01:07NJoinivan` [0] (n=ivan`@unaffiliated/ivan/x-000001)
00:01:07NJoinSirFunk [0] (
00:01:07NJoinBrianHV [0] (
00:01:07NJoinjepler [0] (n=jepler@emc/developer/jepler)
00:01:07NJoinfxb [0] (
00:01:07NJoinTTThomas [0] (
00:01:07NJoinguyzmo [0] (
00:01:07NJoin[omni] [0] (
00:01:07NJoinfeisar [0] (
00:01:07NJoinlostlogic [0] (n=lostlogi@rockbox/developer/lostlogic)
00:01:07NJoinKohlrabi [0] (
00:01:07NJoinmaxkelley [0] (
00:01:07NJoinrvvs89 [0] (n=rvvs89@pdpc/supporter/active/rvvs89)
00:01:07NJoinDiDjCodt [0] (
00:01:07NJoinblithe [0] (
00:01:07NJoinshodanX [0] (
00:01:07NJoinJ [0] (
00:01:07NJoinpreglow [0] (n=thomj@rockbox/developer/preglow)
00:01:07NJoinAve [0] (
00:01:09***Server message 505: 'logbot :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
00:01:12***Server message 505: 'logbot :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
00:01:16***Server message 505: 'logbot :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
00:01:27 Join Rob2222 [0] (
00:01:27 Join snake_ [0] (
00:01:27NJoinannulus_ [0] (
00:01:27NJoinmaraz [0] (
00:01:27NJoinadvcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
00:01:27NJoinBigBambi [0] (n=Alex@rockbox/staff/BigBambi)
00:01:27NJoinLynx_ [0] (
00:01:27NJoinn1s [0] (
00:01:27NJoinNico_P [0] (n=nicolas@rockbox/developer/NicoP)
00:01:27NJoinqwm [0] (
00:01:27NJoinz35 [0] (
00:01:27NJoinjurrie [0] (
00:01:27NJoinDogBoy [0] (
00:01:27NJoinn17ikh|Lappy [0] (
00:01:27NJoinSeed [0] (
00:01:27NJoinSoap_ [0] (n=Soap@rockbox/staff/soap)
00:01:27NJoinjmworx [0] (
00:01:28NJoinATravelingGeek [0] (n=ATG@pdpc/supporter/student/ATravelingGeek)
00:01:28NJointedrock [0] (
00:01:28NJoinLangly [0] (
00:01:28***Server message 505: 'logbot :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
00:01:29BagderCasainho: the same name yes, but it's not used in the same builds
00:01:31CasainhoBagder: is that possible the same fucntion existe 2 times?
00:01:37Bagderit doesn't
00:01:39Casainhoah, okok :-)
00:02:17 Join dionoea_ [0] (
00:02:24 Join ender [0] (
00:03:16Nico_PjhMikeS: because in certain situations buffer_handle will only buffer one chunk
00:03:37jhMikeSsuch as? and what for?
00:03:42rasherI suspect r15724
00:03:46Casainhostatic void wheel_light_dimming_tick(void)
00:03:49Casainho //static bool flag_button_light = true;
00:03:51Casainho static unsigned char counter = 0;
00:03:53Casainho counter++;
00:03:54Casainho if (counter >= 25)
00:03:56Bagderplease stop that
00:03:56Casainho {
00:03:57Casainho counter = 0;
00:03:59Casainho }
00:04:02 Quit linuxstb (
00:04:02 Quit mirak (
00:04:02 Quit thegeek (
00:04:02 Quit ctaylorr (
00:04:02 Quit Rob222241 (
00:04:02 Quit roolku (
00:04:02 Quit weezerle (
00:04:02 Quit ender` (
00:04:02 Quit GodEater_ (
00:04:02 Quit newbyx86 (
00:04:02 Quit dionoea (
00:04:02 Quit freqmod_nx (
00:04:02 Quit DraX (
00:04:02Bagderuse a pastebin
00:04:02Casainho if (counter >= 19)
00:04:03Casainho {
00:04:03CtcpVersion from freenode-connect!freenode@freenode/bot/connect
00:04:04Casainho //Button light OFF.
00:04:04Mode"#rockbox +o Bagder " by ChanServ (ChanServ@services.)
00:04:05Casainho GPIOG_OUTPUT_VAL &=~ 0x80;
00:04:07Casainho }
00:04:11Casainho else
00:04:11Kick(#rockbox Casainho :Bagder) by Bagder!n=daniel@rockbox/developer/bagder
00:04:22rasherI suspect r15724 will fix my "file operations unavailable during database update" problem.
00:04:23 Join DraX [0] (
00:04:25***Server message 505: 'logbot :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
00:04:29rasherTime to test.
00:04:39rasher(on Sansa)
00:04:40Mode"#rockbox -o Bagder " by Bagder (n=daniel@rockbox/developer/bagder)
00:04:49Nico_PjhMikeS: such as when there are several messages in the buffering queue
00:05:05rasherWhat happened to making all committers ops?
00:05:06Nico_PI'm not sure these situations really happen
00:05:11 Join matsl [0] (
00:05:35 Quit matsl (Client Quit)
00:05:36 Join hunz [0] (
00:05:59jhMikeSif it needs to continue buffering, it should immediately go back to filling after a message really
00:06:05 Quit davina ("xchat on Ubuntu 7.04")
00:06:18 Join linuxstb [0] (n=chatzill@rockbox/developer/linuxstb)
00:06:18NJoinmirak [0] (
00:06:18NJointhegeek [0] (
00:06:18NJoinctaylorr [0] (
00:06:18NJoinRob222241 [0] (
00:06:18NJoinroolku [0] (
00:06:18NJoinweezerle [0] (
00:06:18NJoinender` [0] (
00:06:18NJoinGodEater_ [0] (n=bryan@rockbox/staff/GodEater)
00:06:18NJoinfreqmod_nx [0] (
00:06:18NJoindionoea [0] (
00:06:18NJoinnewbyx86 [0] (
00:06:20 Quit dionoea (Connection reset by peer)
00:06:35lostlogicjhMikeS: yeah, it should, like I said, just an optimization
00:06:36moosrasher: the more ops will be, the better...
00:07:25lostlogicjhMikeS: the (purely theoretical) problem would be if the handle that is playing is not the first handle on buffer then it won't be the first handle hit with buffering and it could be a long time before it gets hit again
00:07:59 Quit freqmod_nx (Connection reset by peer)
00:08:19scorche|wrasher: waiting on reply from t0mas
00:08:30scorche|wif anyone has a mail for him other than
00:09:25rasherEmail: firstname at (firstname: Tomas, lastname: Salfischberger ofcourse)
00:09:58scorche|wrasher: thanks
00:10:07rasherIt's on his wikipage :)
00:10:26jhMikeSlostlogic: seems possible in any case
00:11:41lostlogicjhMikeS: any easy way to make the queue check happen 1/2 as often as the buffer level / yield/sleep check? :)
00:13:18jhMikeSwhy? if that's an issue, then something more robust needs doing. I mean, keep a counter but I refuse to go that road. You could put any handle as first.
00:14:01jhMikeSjust because at least the guardbuf amout of data won't exist?
00:14:06 Join barrywardell [0] (n=barrywar@
00:15:07 Quit ender` (Connection timed out)
00:16:31 Quit Rob222241 (No route to host)
00:16:37 Quit thegeek (Connection timed out)
00:17:46 Join roolku_ [0] (
00:19:15lostlogicjhMikeS: ideally I'd like to have the buffer_handle() call take 2 parameters −− handle and size −− and ensure that at least size additional data is buffered on the requested handle before ti returns
00:19:16jhMikeSif the linked list is circular, the first handle can move around the circle without the problem occurring
00:19:20lostlogicregardless of the queue
00:21:39 Quit roolku (Connection timed out)
00:22:43 Quit n1s ()
00:22:44 Quit przemhb ("ChatZilla [Firefox]")
00:24:47preglowjhMikeS: seems almost all of the new voice problems are tool problems...
00:24:53preglowso at least you did a good job :>
00:25:22rasherWhy's voicebox still in the wiki anyway?
00:25:30jhMikeSpreglow: thanks, but you had more boring tool stuff to program
00:25:48preglowdamn straight
00:26:14preglowrasher: i'm more concerned with why the ml is full of reports on it not working
00:27:00jhMikeSVB = Vomit, Basic
00:27:19barrywardellBagder: did you see earlier that I fixed my build server problem?
00:27:43Bagderah no, great. thanks for poking me
00:28:18preglowjhMikeS: yeah, the fact that i can't try it doesn't make it any easier
00:28:19barrywardellI was just missing the zip program
00:28:26Bagderah, hehe
00:28:28preglowi think i'll just go finally bloody add rbutil support
00:28:58jhMikeSThat's not so bad since it's C++ at least
00:29:36zichoRockbox on Zune? Is it possible?
00:29:59Lloreanzicho: In theory, but nobody's working on it, so it'll probably never happen.
00:30:12zichoHaha, nobody wants zune.
00:30:13 Quit lee-qid (Read error: 110 (Connection timed out))
00:30:28Bagderwell, if you want rockbox don't buy a zune
00:30:57zichoI wont.
00:31:35barrywardellhmmm. maybe I should have bumped the version of e200rpatcher before building those binaries
00:32:43 Quit linuxstb ("ChatZilla 0.9.79 [Firefox]")
00:33:35 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
00:33:51linuxstbbarrywardell: We still could... I don't mind building another win32 binary.
00:34:07barrywardellyeah, I can build linux and mac binaries
00:34:27 Join Calcipher [0] (
00:34:43barrywardellwhat has changed in this version?
00:34:48linuxstbIf we can catch Bagder before he goes to sleep...
00:34:59Bagdersure, I'll stick around until you're done
00:35:03linuxstbNothing in e200rpatcher.c I think - just the addition of those extra checks in the ARM binary.
00:35:50barrywardellyes, ok. I'll bump the e200rpatcher version and build new binaries now...
00:36:09barrywardell0.2-svn OK for the new version?
00:36:13linuxstbOr just 0.2
00:36:24Bagder0.2 gets my vote
00:36:32barrywardell0.2 it is
00:36:45linuxstbWe're open source, so obviously can't reach 1.0 for a few years...
00:37:37barrywardelllinuxstb: committed. can you build the windows binary?
00:37:56 Quit MethoS-- (Read error: 110 (Connection timed out))
00:39:58linuxstbDone -
00:40:10 Quit ender (" Just because I have a short attention span doesn't mean I")
00:40:48 Part hunz
00:41:29linuxstbbarrywardell: Shall I tag the release?
00:41:37barrywardellyeah, good idea
00:42:17barrywardellhmm. can't login to my 64 bit linux machine right now to build that one
00:43:01barrywardellhere are the other two versions:">
00:44:44barrywardelland the bootloader binary:">
00:45:25Bagderthey're all in place now
00:45:57barrywardellthanks, I'll send the linux64 one in the morning
00:48:09 Nick roolku_ is now known as roolku (
00:49:45 Quit Nico_P (Remote closed the connection)
00:49:59 Join midgey [0] (
00:50:53 Quit mirak (Remote closed the connection)
00:51:06 Nick Weiss_ is now known as Weiss (
00:51:20preglowjmspeex: around?
00:51:26jmworxpreglow: yes
00:51:55preglowjmworx: i see that even speexenc followed by speexdec yields a slightly delayed output in comparison with the wav input before encoding, any way to avoid that?
00:52:28jmworxpreglow: your're seeing that all the time or just with the −−force-* options?
00:52:48preglowjmworx: well, just tried it now, not svn speexenc/dec, though
00:52:49preglowwb mode
00:53:06preglowno force, no nothing, as a matter of fact, just plain enc/dec
00:53:12jmworxpreglow: BTW, could you fill in this survey in case you haven't already?
00:53:29preglowjmworx: sure
00:54:17jmworxpreglow: if it's a short file, it's normal. For longer files (more than about 8 kB once encoded), it shouldn't happen though
00:55:01preglowjmworx: short file indeed, why does it happen?
00:55:57jmworxpreglow: basically, speexdec uses some computation based on the granulepos to discard samples from both the first and last page.
00:56:14preglowahh, yes, saw that
00:56:24jmworxUnfortunately, since there's only one granulepos per frame, if the first frame is also the last frame, we're screwed
00:57:12preglowthat's probably the reason, the clip is about 2kb compressed, heh
00:58:36 Quit moos ("Rockbox rules the DAP world")
00:58:44 Quit scorche|w ("CGI:IRC")
00:59:35preglowjmworx: so the decoder lookahead only compensates for decoder latency? you'll basically need to transmit encoder latency in some other way?
01:00:05jmworxpreglow: It's encoder latency that's transmitted indirectly through the granulepos.
01:00:10 Quit ompaul (Client Quit)
01:00:14preglowthat's what i expected, yes
01:00:17jmworx(because the decoder obviously knows how much latency it adds)
01:00:26preglowwell, ok
01:00:32preglowi guess i'll just ignore encoder latency
01:00:39preglowit's so short anyway
01:01:12jmworxwell, in the current implementation, latency is split half-half between encoder and decoder.
01:01:29preglowisn't the encoder latency slightly bigger?
01:01:32jmworx(it used to be that the encoder had more than the decoder)
01:01:35preglowahh, no
01:01:42preglowwindowsize is framesize+subframesize
01:01:56jmworxfor narrowband, latency is now split 5/5, but it used to be 10/0
01:02:27preglowwhich gives a latency of subframesize, same as for decoder
01:03:30preglowjmworx: btw, for the memset/cpy stuff, should i add a speex_memcpy? and what's the point with those wrappers anyway, isn't mem* always supposed to be pretty well optimized anyway?
01:03:57amiconnpreglow: Bah, didn't check all code paths and forgot to dim a variable. Trivial fix...
01:04:11 Quit rasher ("leaving")
01:04:13preglowamiconn: yeah, brian posted one the ml. you're on it?
01:05:02 Quit japc (Read error: 110 (Connection timed out))
01:05:04jmworxpreglow: I think memmove should be pretty much as fast as memcpy now, right?
01:05:23jmworx(i.e. it's just a simple test to figure out whether you can use the same algo as memcpy)
01:05:27preglowjmworx: memmove is basically memcpy + one other code path and an if, so yeah, more or less as fast
01:05:52preglowjmworx: but then again, another wrapper is four lines of code
01:06:24amiconnI've seen memmove() implementations of vastly varying quality
01:06:36preglowanother good reason to use memcpy when you can
01:06:51jmworxpreglow: wrapper doesn't make the code bigger because it's inline and it makes it easier for someone to entirely remove the dependenty on libc
01:06:53preglowjmworx: and almost all occurences of for () i've replaced in nb_celp and sb_celp can be done using memcpy
01:06:54amiconnA proper implementation should be as fast as memcpy(), with just the additional overlap/direction check
01:07:30amiconnSome cheap implementations resort to trivial byte copying on overlap
01:07:48preglowamiconn: i've seen one of those too...
01:08:08amiconnThe rockbox implementations are clear
01:08:56amiconnHowever, I still want to tame that monster named coldfire memcpy+memmove. It's fast, but a code monster...
01:09:05preglowyeah, it's pretty huge
01:09:08*amiconn wrote it :\
01:09:28amiconnAlmost 2KB for the whole thing iirc
01:10:21 Join rasher [0] (n=rasher@rockbox/developer/rasher)
01:10:43***Saving seen data "./dancer.seen"
01:10:56 Quit austriancoder ("Konversation terminated!")
01:12:36preglowbut it is fast :P
01:13:39 Join JdGordon [0] (
01:13:55jmworxpreglow: But how much does it slow down (or make other code slow down) because of I cache trashing?
01:14:53linuxstbIt's in IRAM - so not cached.
01:15:17amiconnYeah, it chips away ~4% of core iram :\
01:15:28preglowjmworx: using speex_resampler_skip_zeros() basically allows me to ignore resampler latency, yes?
01:15:32jmworxif you've got a high "i cache pressure", then it could be responsible for 4 kB of instruction transfer due to cache flushing every time you call it.
01:16:28jmworxpreglow: it doesn't make the latency go away, but it discards the leading zeros caused by latency
01:16:45 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
01:17:02preglowjmworx: yes, which does kind of make it not exist as far as the caller is concerned
01:17:12preglowas long as you're aware that the first block size is smaller, etc
01:19:51 Join japc [0] (
01:22:20preglowamiconn: did you upload a new voicebox?
01:23:22Soapdesowin (for the logs) - thank you for adding your battery bench to SansaRuntime.
01:23:46preglowamiconn: "make voice" is the only way to make voice files as of now, yes?
01:23:57*rasher makes some more noise in favour of putting voicebox in SVN
01:24:09rasherpreglow: Correct
01:24:28barrywardellBagder: if you're still around, I managed to build the linux64 version too. here's the updated tarball with t hem all:">
01:24:37preglowrasher: feel free to slap a gpl license on it and put it in svn
01:24:48preglowminus the binaries, of course
01:24:58 Join homielowe [0] (
01:25:01rasherConverting MakeVoices.vbs would be a nightmare, since it needs part of the build system (because of the features feature and target_id)
01:25:14rasherunless those informations are made available online or something
01:26:22amiconnWhy would we need to convert MakeVoices?
01:26:36preglowwe don't need that, no...
01:26:43preglowi'd rather make rbutil better
01:26:52amiconnThe current method works rather well in all supported build environments
01:27:21rasheramiconn: Well we don't need to at all. Some people would like it, is all
01:31:31*amiconn thinks that scripting is more flexible than an executable like rbutil
01:31:50amiconnAnd we need information from the buid system (features.txt etc)
01:32:08rasherpreglow: you want to add voice-making ability to rbutil?
01:32:17amiconnBtw, there's a dependency bug - features.txt doesn't get rebuilt when target features change
01:32:19Calcipherhey rasher, how does it look for commiting the db voice support?
01:32:51rasherCalcipher: not very good. The current solution is rather half-baked
01:33:54 Quit iamben (Read error: 104 (Connection reset by peer))
01:34:04preglowrasher: not really, no
01:34:04CalcipherI was never able to try it to see how well it performed, I never got past the strange errors
01:34:15preglowrasher: i'm just going to add rbspeexenc support
01:34:22rasherIt needs a fix either on the script side or in the code. I'm not sure which one is better or easier.
01:34:28preglowi'm _dying_ to dive back to more low-level stuff here
01:34:46rasherCalcipher: well, it can voice tags, but not stuff like "01 - title" −− only "title"
01:35:05rasherIe. any string that consists of more than a single tag won't be voiced
01:36:14 Join kkurbjun [0] (
01:36:40CalcipherI still would have liked to try it out, so how exactly is the voice data stored?, is it one large file for the entire db? that wouldn't work so I'm guessing its small clips per tag stored in that default location, thats a great idea btw, to centralize clips in a location
01:37:30Calcipherwell that too can be a problem for clean up, when you for example get rid of a folder, and then the voice data still remains in the default clip location
01:37:31rasherThat's what it does. In fact, the code already does look for voice clips for the database strings (just in / rather than inside a specific dir)
01:39:02preglowamiconn: didn't "make voice" need a c compiler before rbspeexenc?
01:39:20 Join iamben [0] (
01:39:39amiconnIn theory it should only require the preprocessor, if rbspeexnc is already compiled
01:39:55amiconnDue to another dependency bug, 'make voice' also compiles the bitmaps...
01:40:35rasherI've prodded Bagder about this, but judging by his reaction it won't be easy to fix
01:40:38amiconnrasher: Storing a ton of .talk clips sounds like an insane idea...
01:40:59amiconn...even more insane than file .talk clips, which I never use due to that very reason
01:41:17rasherWhy? It's not *that* many more than .talk clips
01:41:17amiconnEven a tiny file always takes a full cluster on FAT
01:41:45amiconnAnd a cluster can be as large as 32KB
01:41:57rasherSo you suggest creating a giant ball of concatenated clips?
01:42:13amiconnI would simply drop the idea of using clips for the database
01:42:20rasherIn favour of what?
01:42:22Calcipherwhat would be best for disk reads?
01:42:27amiconnThere's already spelling, and numbering
01:42:42rasherBoth vastly inferior to .talk clips
01:42:44amiconnBoth require no spinup and no additional work
01:42:52Calcipherso, you use spelling and numbering?
01:43:04amiconnOther than that, it would be an idea task for a tts engine
01:43:38LloreanWhy not create a .voice file for each of the database tcd files?
01:43:45amiconnCalcipher: I use file browser mode in 99.9% of all usage cases. For directories, I do use clips, for files, I just use numbers + file type announcing
01:44:13preglowamiconn: in theory, sure, but did it need a c compiler? there's a guy on the ml that doesn't seem to have a compiler installed that now complains voices won't build
01:44:23preglowalso, how do you install a c compiler in cygwin?
01:44:25amiconnFor me, the database is just a tool to spot typos in tags, and very rarely, duplicates
01:44:42rasheramiconn: so you don't use .talk clips and find them wasteful. Fine, but I don't see why it's so insane. If people want .talk clips, and they very clearly do, it should work for the database as well.
01:44:45CalcipherI see
01:45:08amiconnrasher: I do use them, but sparsely, i.e. for directories only
01:45:47amiconnpreglow: You just use the cygwin installer, which works like a typical, graphical linux package manager
01:46:05preglowwhat is the package named? i need something to point this guy to
01:46:25amiconnsomething gcc ...
01:46:35CalcipherI don't see why the database is ignored by so many users, It feels like a much easier and organized way to get to the music you want, for example you can play all of one artists tracks even if they're in separate directories, and this requires no creation of a playlist, or just listen to an entire album
01:46:48amiconnI think you can go without a crosscompiler if you just want to build voice, if you select a sim
01:47:03*amiconn disagrees with Calcipher
01:47:16lostlogicsansa has pretty nice battery life.
01:47:29rasherDatabase is an additional tool - it can do some things that the filetree can't, but not everyone wants those things.
01:47:34*preglow isn't cut out for mailing lists
01:47:51*scorche gets along famously with filetrees
01:47:53rasherpreglow: the package just named 'gcc'
01:48:31amiconnpreglow: But if a crosscompiler is needed, provides cygwin packages for those
01:48:55amiconnCompiling those yourself can take a few hours...
01:49:09preglowahhhh, cygwin...
01:49:10amiconnAnd for some reason, the multilib patch for arm doesn't work on cygwin...
01:49:11rasheron a fast computer.
01:49:40amiconnrasher: For me, it's around 1.5 hours per architecture iirc
01:49:47Calcipheryeah it took me about an hour and change when I compiled
01:49:53amiconnOn a 4 years old laptop
01:50:42 Part pixelma
01:50:56preglowamiconn: btw, going to add narrowband mode in the encoder soon
01:51:02preglowgot it going here, just need to test a bit more
01:51:45*amiconn wonders what's up with the multilib patch
01:51:53preglowas a matter of fact, i think i'll just commit it now
01:52:02preglowlinuxstb: could you provide me with a windows binary?
01:52:09amiconnIt compiles without errors, but 'make instal' just decides there's nothing to do for most folders
01:52:43amiconnpreglow: If you commit, providing a windows binary is a simple thing
01:52:50 Quit hcs ("Leaving.")
01:52:55amiconnThe next 'make voice' on my side would rebuild it
01:52:57 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
01:53:20preglowtestet with nb mode now, and it works ok, but the clicking is back :/
01:53:29preglowi'll need to figure that out before i do anything else
01:53:31amiconnCrosscompiled win32 binaries tend to be larger for some reason (but maybe not when they're stripped)
01:54:35rasherJdGordon: I don't think that "ipod+car" task you just closed was accessory related
01:56:28 Quit mokkurkalve (Read error: 110 (Connection timed out))
01:57:07 Join EternalRains [0] (
02:00:44 Part EternalRains ("Konversation terminated!")
02:03:19preglowjust compensating for encoder lookahead isn't enough for nb mode
02:04:32preglowamiconn: this silence thing really doesn't seem to be working out on swcodec right now, i wonder why
02:05:01JdGordonrasher: sure it is.....
02:05:39rasherJdGordon: then why did he mention usb?
02:05:51JdGordoncoz thats how you connect it to the car?
02:05:57JdGordonunless... ohh..... umm...
02:06:03 Quit scorche (Read error: 104 (Connection reset by peer))
02:06:23rasherNot that it isn't a support request, but I think it's simply one of those car stereos that have a usb plug
02:06:32 Join scorche [0] (n=scorche@rockbox/administrator/scorche)
02:06:34JdGordonyeah could be...
02:06:36JdGordonoh well...
02:06:42*JdGordon gone
02:07:07preglowjmworx: basically, if i just make sure to add 'lookahead' samples of zeros at the end of a clip i encode, everything in that clip should be encoded correctly, right?
02:07:19 Quit zicho (Remote closed the connection)
02:10:52preglowjhMikeS: around?
02:11:23markunpreglow: jmworx or jmspeex?
02:11:49markun(or is it the same person?)
02:12:02markunah, never mind
02:12:12 Join FOAD_ [0] (
02:12:13preglowmarkun: same person, different context :)
02:12:20*markun goes back to sleep :)
02:12:21 Join johnf1911 [0] (
02:13:12jhMikeSpreglow: yah
02:13:21johnf1911I have an iRiver H120 with boatloader v5, when I try to install a recent version of rockbox, I invariably get a invalid instruction error
02:13:36johnf1911should I update the bootloader? what bootloader is required for what version of firmware, and how can I tell?
02:13:56jhMikeSmight be a good idea to update
02:14:13johnf1911what is the current version? I haven't been able to figure out how it works
02:14:35 Join UncleRem1s [0] (
02:15:05jhMikeSI have v6 on mine and it's just fine.
02:15:50preglowjhMikeS: you sure the silence clip stuff works?
02:16:15 Quit hcs (Remote closed the connection)
02:16:32jhMikeSif it gives it to the thread to play, then it should.
02:17:06preglowhrm, i wonder why this clicking occurs :/
02:17:11 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
02:17:49 Quit billenium ("Ex-Chat")
02:18:03jhMikeSdid debugf show that the silence if clause was being reached?
02:18:39 Join billenium [0] (
02:18:44johnf1911bah, it's clearly going to be a lot easier to patch the firmware on windows
02:18:45preglowjhMikeS: it did, just as it should
02:19:02 Quit UncleRemus (Read error: 110 (Connection timed out))
02:19:06preglowi've gotten rid of the clicking by properly taking lookahead into account in rbspeexenc, but it doesn't work in nb mode
02:19:13 Quit billenium (Client Quit)
02:19:27jhMikeSever try comparing the output to concatenating an encoded clip with a silence clip and decoding that outside the fw?
02:20:11preglowi actually just did, and even that result surprised me somewhat, the filter went on ringing far longer than one frame
02:20:54jhMikeSno way to force mute the filter softly?
02:21:00 Join billenium [0] (
02:21:13preglowwell, it should just die out after a frame or so
02:21:50jhMikeSit does that in the non-rockboxed libspeex code?
02:22:10preglowdidn't try, but i really haven't changed that part
02:22:25preglowbut anyway, if i encode an extra zero frame, then even the nb mode doesn't click
02:22:50preglowbut after taking into account the lookahead, that zero frame really shouldn't be necessary because of the silence clip
02:23:34 Nick fxb is now known as fxb__ (
02:24:36 Join iamben_ [0] (
02:24:52 Join midkay_ [0] (
02:25:12 Quit iamben (Read error: 104 (Connection reset by peer))
02:25:17jhMikeScan you pick and choose original vs. asm filters to narrow it down to something. I mean it not supposed to ring that long, so...hmmm
02:25:21preglowapparently, there is no end to the voicebox complaints
02:25:41jhMikeSwhat's the problem now?
02:25:45preglowjhMikeS: i'm pretty sure that has nothing to do with it, i've comparaed both qmf and iir to c results, and they're bit identical
02:25:51preglowjhMikeS: some shortcut creation stuff
02:26:34preglowno idea
02:27:03 Quit DerPapst ("So Long And Thanks For All The Fish!")
02:27:53 Quit FOAD (Read error: 110 (Connection timed out))
02:27:53 Nick FOAD_ is now known as FOAD (
02:29:52LloreanWhy does voice creation require GCC again?
02:30:03preglowright now, for compiling the encoder
02:30:09LloreanOr more specifically, why can't it be made to run without GCC assuming the presence of a (language).lang file?
02:30:28LloreanI thought the .voice file required a dev environment for something else too?
02:30:37preglowfor parsing features
02:30:37LloreanDetermining device capabilities or something?
02:30:45preglowi guess
02:30:52rasherThat's done with cpp, I believe
02:31:08LloreanCould it be bypassed with the presence of a .lang?
02:31:28LloreanFor the people who just want to create .voices, and nothing else, and don't want cygwin?
02:31:32rasherthat would require to completely rework how voice files are built
02:31:56LloreanAh, okay.
02:32:39LloreanWhere do the voice strings come from, then?
02:33:26rasherWell, they come from the lang file, but in combination with genlang and a target-id (and features)
02:33:40rasheryou *can* run directly and get a voicefile
02:33:52LloreanWith langv2 aren't .lang files also target-specific?
02:34:02LloreanOr, wait, that's .lng rather?
02:34:13rasherBut then .lng doesn't contain the voice strings
02:34:24 Quit Arathis2 ("Bye, bye")
02:35:14LloreanNevermind me then.
02:35:37preglowi wonder if i should just do a new rbspeexenc release for the fixed wb clips and put in a note that nb mode is glitchy
02:35:42rasherYou can do something like -V -l=english -t=h120:recording:otherfeatureshere -i=7 -e=rbspeexenc -E="-c10 -q 4" -s=flite -S=""
02:35:46rasherand get a working voicefile
02:35:56rasherUsing only genlang, rbspeexenc and english.lang
02:36:45amiconnPlus wavtrim, voicefont, and for windows, sapi_voice.vbs
02:36:45rasherPiecing together the target-string isn't trivial, but could be made available online
02:36:50rasherah yes
02:36:54rasherand of course
02:37:34johnf1911upgrading the bootloader to V6 won't introduce problems booting my current rockbox image, will it?
02:37:39LloreanOr batch files that already include the target string?
02:38:04rasherLlorean: sure, it's only a matter of saving features.txt from the daily builds
02:38:05amiconnThe features list might change from time to time
02:38:19amiconnHappened at least 2 times in the last weeks
02:38:19rasheror current builds, of course
02:38:31 Quit midkay (Read error: 110 (Connection timed out))
02:38:38LloreanJust wondering, because Cygwin seems to be mildly horrible for blind users, from the sounds of it
02:39:12rasheramiconn: the build server could piece together a decent .bat file that would allow users to build voice files fairly easily given the needed tools
02:39:28amiconnI think it would be better to provide a more complete set of voice files (e.g. more languages than just english)
02:39:47rasherNot really mutually exclusive, or even related
02:39:52Lloreanamiconn: But it's likely there will always be voices people want to use that we don't have, or differing rate desires, etc.
02:40:08amiconnYes, sure.
02:40:13LloreanThough, wasn't there talk of time compression for playback recently?
02:40:47preglowyes there was
02:41:46preglowamiconn: eh? it's possible to install gcc and not cc1.exe in cygwin?
02:41:52 Quit weezerle ("...und tschüss!")
02:42:00amiconnNot normally
02:42:09preglowwell, this guy has apparently managed it
02:42:29amiconnPeople manage to do all sorts of weird things
02:44:02 Quit midkay_ (Read error: 110 (Connection timed out))
02:46:38amiconnpreglow: Shortcut problem fixed as well, btw. Mrf :\
02:47:19preglowupdated wiki?
02:47:23 Join webguest73 [0] (i=cfe966c8@gateway/web/cgi-irc/
02:47:38amiconnUnfortunately vbscript does only find undeclared variables if you're actually executing that code path
02:48:05 Quit webguest73 (Client Quit)
02:48:56preglowmany languages share that trait
02:49:49 Quit Rondom (Read error: 110 (Connection timed out))
02:49:58rasherHm, need the target_id as well as apps/feature. I wonder if the daily-build script could pick this out and put it in a .bat file
02:50:57preglowamiconn: but you updated the wiki, yes? i don't know what time it is in the server place, so can't check the time stamp :/
02:51:30amiconnThe server is in Sweden hence the same time zone as you (and me)
02:51:37amiconnBut yes, it's updated
02:52:00preglowamiconn: the server time is most definitely not the same as my time
02:52:08rasherBut parts of the website is utc
02:52:35amiconnThe script that puts the time stamp onto the front page didn't notice the change yet
02:52:36preglowand which parts i do not know
02:53:11johnf1911I can now run current builds
02:53:14johnf1911thanks a lot for the help!
02:53:19preglowamiconn: also, some people say that selecting speex as an encoder still produces mp3 files, any idea about that?
02:53:39amiconnThen the clips on my swcodec targets wouldn't work...
02:53:40preglowi only know what they say
02:53:46preglowWhen I run the new voice box that was posted, I've noticed that I have two radio buttons for speex but they seem to make mp3 format files that don't play on my mp3 player. am I doing anything wrong?
02:54:26rasherAt least this tells me that I was right in creating with the assumption that different encoders would be needed eventually
02:55:46preglowsure, i just needed to get around to it :) i've been talking about speex voice files for over a year now
02:56:12*amiconn retries
02:56:34 Quit Thundercloud (Remote closed the connection)
02:58:13Soapamiconn, I know you are neck deep in other tasks, but would you like Santa to drop one of these: into your Christmas stocking?
02:58:48preglowamiconn: but ok, do you have time to compile a new rbspeexenc shortly?
02:59:37preglowjust one more wb test here, and i'll commit it
03:02:46preglowforget that, i still get clicks for some rare cases...
03:03:22 Join midkay [0] (n=midkay@rockbox/developer/midkay)
03:03:30 Join daurnimator [0] (n=daurn@
03:03:40 Quit daurn ("Cyas")
03:08:02amiconnpreglow: Must be the second problem then; both speex and mp3 clip creation works as it should here
03:09:21 Quit roolku ()
03:10:32preglowi assume so
03:10:40*preglow wishes he could have speex not click :/
03:10:47***Saving seen data "./dancer.seen"
03:14:19 Quit jhMikeS (Nick collision from services.)
03:14:25 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
03:15:05preglowjmworx: why do you sometimes copy an excitation array before passing it to iir_mem16? it's capable of taking a separate in and out array parameter, so copying first shouldn't be necessary
03:15:28jhMikeSpreglow: then bail out and don't allow nb for now :p
03:15:43preglowjhMikeS: i've made wb mode click again
03:15:59preglowway more seldom than before, but i can make it do so
03:15:59jhMikeSdon't say that! :\
03:17:01preglowthis is nothing an extra encoded zero frame can't fix, but it'd like to hear if jmworx has any clever ideas first
03:17:06preglowbut seems like that'll be tomorrow, bedtime now
03:17:46preglowi still wonder why silence clips don't work, though
03:18:05jhMikeSshould really try a manual concat for that
03:18:18preglowi guess so, but that'll also be tomorrow
03:18:20*preglow bails out
03:23:13jmworxpreglow: what do you mean about the iir_mem16 copy?
03:24:54preglowjhMikeS: ok, i just had to check that out before going to bed, it works fine with a manual concat
03:25:35preglowjmworx: well, nb_celp.c:1242, for example, why not just pass st->exc as the first parameter?
03:25:41preglowthe copy seems kind of redundant
03:26:49jmworxpreglow: exc is a 32-bit var isn't it?
03:26:50preglowthe same is done a couple of other places
03:27:12preglowjmworx: well, weird, iir_mem16 takes a spx_word16_t* as param
03:27:34jmworxoh, ok it no longer is, which is probably why...
03:27:43 Quit homielowe (Read error: 110 (Connection timed out))
03:27:54preglowboth exc and out are spx_word16_t
03:28:55jmworxYeah, I changed a lot of code around beta1 and beta2 and might have left some redundant/useless stuff
03:29:23preglownice to have me trawling through it, then :)
03:30:35preglowjmworx: it looks to me like the filter can be left ringing for more than a frame if only null modes are used after an ordinary frame, is that so?
03:32:15jmworxcan be left??
03:33:31preglowwell, will be ringing
03:34:21jmworxpreglow: in theory, it'll ring to infinity because it's an IIR filter. In practice, I'm pretty sure you can cut after one frame, no?
03:34:37preglowjmworx: well, sure, but you seem to be limiting the coefs for each frame with bw_lpc
03:35:30jmworxbw_lpc has nothing to do with ringing and it's an encode-only thing to approximate psychoacoustics stuff
03:35:41 Join psycho_maniac [0] (
03:35:51jhMikeSpreglow: hmmm...the bit cursor is being reset so that's the only diff I can imagine
03:36:45preglowjmworx: encode only it's not, it's used for null mode decoding
03:36:56preglowjmworx: but yeah, i did misunderstand what it does
03:38:00jmworxpreglow: Oh, on decoding it has a completely different purpose. Basically, it moves all the poles in-wards so that ringing is shorter
03:38:05preglowjhMikeS: weird, then, "twenty" clip plus silence on target clicks, the same concatted by hand here and decoded with what is essentially the voice thread code gives a nice fade
03:38:39jhMikeSyou sure that use of speex_bits_remaining() < 8 is the right thing?
03:39:09jmworxspeex_bits_remaining() < 8 looks wrong
03:39:18jmworxA valid frame can be only 5 bits
03:39:46preglowvery true
03:40:16jhMikeSspeex_bit_remaining() < 0 should end the clip, correct?
03:40:18 Join btwizt [0] (
03:40:38jhMikeSlooking in speex.c, that is what's done
03:40:39preglowbut we might also have 0-7 bits left which are just left over
03:40:57preglowand don't represent a frame at all
03:41:26 Join iamben [0] (
03:41:31jhMikeSI see what's up with that. I need to check for end of stream
03:41:39jmworxjhMikeS: how the hell can speex_bit_remaining() < 0 ??
03:41:56jmworxjhMikeS: you're aware that speex_decode_int(0 actually tells you when you're done, are you?
03:42:01preglowjhMikeS: basically, only one bit of that last byte might be valid data
03:42:03jmworxi.e. handles all that logic
03:42:43preglowjmworx: well, it does, but that way we might get an extra null frame we didn't actually encode
03:42:47preglowbut i guess that's not much of a problem...
03:42:54preglowand as it is right now, i don't see how we can avoid it
03:42:56jhMikeSheh. I went for speex.c on the API reference in part.
03:43:01jmworxpreglow: No, you won't get anything you didn't encode
03:44:39*preglow goes to find out how speex_bits_write() works
03:45:03jmworxpreglow: When you call speex_bits_write, it appends a "terminator code" that speex_decode knows about
03:45:11preglowthen yes
03:45:16preglowwe don't need that < 8 stuff
03:45:19jmworxSo if you want to decode everything, you just call speex_decode() in a loop until it returns -1.
03:45:20preglownot at all
03:45:52jmworxWhen you get -1, it means "couldn't decode anything else" (and *not* it's the last one).
03:48:47preglowoh well, it didn't help for this clip anyway
03:49:06preglowbut we'll see if it makes a difference on targets
03:50:32jhMikeStesting speex_bits_decode() < 0 is a safe test? I want to get the next data from the callback on both end and error to avoid if (ret = -1 || ret = -2) cruft.
03:50:58jmworxjhMikeS: testing speex_bits_decode() < 0 is a stupid test
03:51:05jmworxoh sorry
03:51:22jmworxI thought you had written speex_bits_remaining :-)
03:51:36jhMikeSno, no :-) getting rid of that.
03:51:43jmworxyes, it's fine, although there's two things that can be < 0
03:51:58jmworx-1 means nothing left, whereas -2 means there was some corruption
03:52:06jhMikeSI'm not concern with which reason in this particular case
03:52:21jhMikeSsince the same action is needed
03:53:50preglowjmworx: so let me see if i've gotten it straight, there should be no need to encode anything more than 'lookahead' samples extra to ensure the encoder has output all you'll need to reconstruct your audio clip?
03:54:39preglowgood, then any remaining problem has to do with us not decoding enough null frames
03:54:46 Quit iamben_ (Read error: 110 (Connection timed out))
03:54:57jmworxactually, there's also the decoder lookahead
03:55:14preglowyeah, but that's solved by decoding null frames, yes?
03:55:31jmworxi.e. if you encode encode_lookahead more samples, you've really encoded everything but the decoder may not give you everything back because of its own lookahead
03:55:57jmworx...which can be solved by decoding a null frame or encoding a few more samples
03:56:02preglowyup, sure, just want to make sure i've got the encoder side right, since that's more important
03:56:41jmworxpreglow: I'd actually recommend encoding an extra frame at regular rate just to make everything easy.
03:56:54jmworxIt's not like a few bytes are really going to hurt you
03:57:22preglowjmworx: and you can rest assured i'll do so if the silence clip mechanism we already have continues to not work :)
03:57:41preglowi've already tried it out encoding an extra zero frame, and it solved all clicking i could find
04:00:03jhMikeSpreglow: want to test a patch or should I just commit since it's the correct way anyhow? (need to check it myself first)
04:00:08 Part johnf1911
04:00:47preglowcommit away
04:00:55preglowi seriously need to go to bed now
04:01:03preglowwill check it out tomorrow, gnight
04:04:25jmworxpreglow: still working on a memmove/memset patch?
04:06:52XavierGrwhy make install takes ages these days?
04:08:22 Join sarixe [0] (
04:08:24 Join iamben_ [0] (
04:09:14rasherXavierGr: probably primarily because of make zip
04:09:28XavierGrany commiter online, I have a lang update
04:10:01XavierGrrasher: I remember it was way faster in the old days
04:10:06btwiztHey Guys, doesna nyone here have any experience with the Sansa c200 series and if so may i aska few quiestions?
04:10:07 Quit barrywardell ()
04:10:39rasherXavierGr: could be something else I guess, but I think make zip is what takes most time
04:11:00XavierGrrasher: so it is slow for you too?
04:11:10advcomp2019btwizt, ask and we can try to answer
04:11:54rasherXavierGr: ffairly slow, yes
04:12:05btwiztThanks ill start with my..crappy.. situation
04:12:30btwiztI got a Sansa c250 V2 2gb for my birthday yesterday, and it worked perfectly
04:12:59btwizttoday i tried the sansa firmware updater, which updated it, but it removed all text from the sansa interface
04:13:26rasherThat's not really Rockbox related. Rockbox doesn't even support the c200 v2
04:13:30btwiztim jsut wondering if using Rockbox will let me use my C250 again
04:13:37btwiztit doesnt?
04:13:43btwiztit says it does on tehw ebsite
04:13:56btwiztor is that only the v1
04:13:59rasher"SanDisk: Sansa c200, e200 and e200R series (not the v2 models)"
04:14:17btwiztwell my bad...guess im waiting for it to be supported or for sansa to fix it
04:14:22btwiztsorry tow aste your time
04:14:36 Join FEL [0] (i=18c8b7b4@gateway/web/cgi-irc/
04:14:43FELoké guy i need help
04:14:48rasherYou could try the sansa forums
04:14:56FELi bought my e260 and i got the box in my hand
04:15:03FELand before open i wanna be sure its not and e260R
04:15:09FELsooo how do i know its not an R
04:15:25FELon the box it only say E260 ..
04:16:03btwiztIve already tried teh sansa forums, no one is helping
04:16:13FELbut it also say playforsure
04:16:16btwiztand ive emailed sansa support directly, no help from them
04:16:25FELsoo that make me thinkin its an R
04:16:37btwiztall sansa players are PLayforsure
04:16:41advcomp2019FEL, if it does not have the R on it, it is a plain one
04:16:46FELo really ?
04:16:59btwiztonly the R comes witht eh Rhapsody stuff
04:17:24rasherAs far as i know, the only way to know for sure is to turn it on and see if it has a rhapsody menu entry.
04:17:35FELyeah but
04:17:38FELits brand new
04:17:44rasherRhapsody models have been sold with non-R cases and vice-versa
04:17:46FELand before open it i wanna be sure
04:17:52btwiztso i guess i ahve no hope of ever getting rockbox on my c250 v2
04:17:54FELomg shit
04:17:56advcomp2019btwizt, most of the sansa players does both MSC mode and MTP mode
04:18:07FELare you serious ?
04:18:10rasherFEL: why is this so important?
04:18:15FELfor rockbox ?
04:18:25FELi heard that u cant installl rockbox on the R version
04:18:38rasherFEL: Rockbox does support the R models. Although the install is more complicated
04:18:39Soapthe e200R not only works now with rockbox, we have the technology to turn an R into a non-R
04:18:56FELis it complicated ?
04:19:02FELyes lol?
04:19:06 Join J3TC- [0] (
04:19:12jhMikeSstupid thing is that Rhapsody mode = MSC and Plays for Sure = MTP ... messed me up for a bit
04:19:13FELomg shit that crap ;D
04:19:30FELthat idiot why they messed boxx lol
04:20:01rasherFEL: please stop swearing, and unless your box says Rhapsody, odds are that it's not an R. But you can't know for absolute sure
04:20:25FELsoo i must open the brand new box i guess
04:20:41FELso u guy wanna know if im lucky or not ? ;p
04:21:26 Quit iamben (Read error: 110 (Connection timed out))
04:21:38FELwhat look R menu ?
04:22:03advcomp2019there should be Rhapsody Channels in it
04:22:11marazFEL: Please pay attention to your grammar and spelling. It's really difficult to understand what you're saying.
04:22:48advcomp2019i think that is what you are asking
04:22:57psycho_maniacyou can also have more then one thought in a sentence. the enter button is not a space bar
04:23:00FELsorry dude we are not all speaking english.
04:23:13FELin this world .
04:23:30marazHey, I'm Finnish and I'm still capable of doing so.
04:23:43SoapFEL, regardless of your native tongue, you are expected to cut out the "u", "lol", and other non-english words.
04:24:09FELah sorry.
04:24:22marazBut before this turns into more rivalry, let me state the IRC guidelines as stated in the wiki (check out the topic): Use clear, grammatical, correctly-spelled English. "Words" such as "thx", "u", and "kewl" are not acceptable. This includes l337speak and odd usages of capital letters as well.
04:25:11FELthe box smell new xD
04:25:20*maraz sighs
04:25:38maleWhat's IRC without leetspeak?
04:25:42FELsansa e200 quick start guide.
04:25:44LloreanFEL: This is an on-topic channel. If it's not important and related to Rockbox, please just think it and don't type it.
04:26:04maleLlorean: Awesome advice. Put that in the topic.
04:26:27Lloreanmale: It more or less is.
04:26:36scorchemale: it is...and if that reference isnt enough, the link says it...there isnt a need to have a looooooooooooooooooooooong topic
04:27:02maleIs looooooooooooooooooooooong a proper English word?
04:27:07maleI'm so confused.
04:28:25marazmale: No, it's just a reference to a certain feline mammal.
04:28:43FELshould i install rockbox right now ?
04:28:59psycho_maniacFEL: only if you want
04:29:01scorchefeisar: that is up to you...
04:29:02marazYou probably should verify that the player functions as it is.
04:29:14FELyes it work ..
04:29:40psycho_maniacthen install it by following the instructions in the manual
04:30:44FELdont get detected -_-
04:31:01scorcheok...i wont
04:31:16 Quit btwizt ("Wake me up when september ends")
04:31:25Soapwhat is your native tongue FEL? French?
04:32:28marazEnglish isn't that far from French, but that's a topic not quite suited for this channel.
04:32:35 Quit w0rd54 (Client Quit)
04:33:10FELOk let me Explain do i need to do something before install rockbox with the autoinstall ... like put it in a certain mod
04:33:29Soapits all in the manual.
04:34:03scorcheincluding instructions if you have an R
04:35:27 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
04:35:40 Quit advcomp2019 (Nick collision from services.)
04:35:51 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
04:37:36 Join w0rd54 [0] (
04:39:54 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
04:40:16 Join miepchen^schlaf [0] (
04:40:41 Join tdoggette [0] (
04:40:55FELin the rockbox auto installation at : select your device in the filesystem , what do i put ?
04:45:08tdoggetteI'm trying to install rockbox on a clean (just reset, latest Apple firmware) 4th gen grayscale iPod using Ubuntu 7.10.
04:45:22maraztdoggette: follow the instructions in the manual.
04:45:25tdoggetteI downloaded and installed using the QT utility
04:45:40tdoggetteand it went smoothly, but rockbox doesn't start
04:45:54marazwhat does it say?
04:46:16tdoggetteselect+menu just reboots into the apple firmware
04:46:31 Quit FEL ("CGI:IRC (EOF)")
04:46:31Soapsounds like you installed the firmware, but not the bootloader.
04:46:43tdoggetteit said I already had the latest
04:46:59marazthe apple firmware doesn't really do anything
04:47:12maraz... in a rockbox boot procedure.
04:47:51tdoggetteI installed to bootloader, it works like a charm
04:51:02 Quit jhulst (Read error: 113 (No route to host))
04:51:55 Quit XavierGr (Read error: 110 (Connection timed out))
04:52:54 Join kryptoz [0] (n=kryptoz@
05:02:32 Part kryptoz
05:03:16 Quit TTThomas ("Lost terminal")
05:09:39tdoggetteOkay, something weird just happened. I was copying files to the iPod, and my computer warned of "unsafe drive removal." I can it to recognize the iPod again by rebooting it, but now the file system is read-only. What's going on?
05:09:56tdoggette*I can it -> I got it
05:10:20 Join homielowe [0] (
05:10:48***Saving seen data "./dancer.seen"
05:11:08tdoggetteThe utility says my configuration is invalid now.
05:14:02tdoggetteShould I just reformat the sucker and start over, or is there a solution to the "read-only" thing?
05:16:12 Quit animeloe ("Leaving")
05:16:12psycho_maniacmaybe try a reinstall first ? what player?
05:17:11tdoggette4th gen iPod
05:17:20tdoggetteI can't reinstall, though−− read only.
05:19:08psycho_maniachave you tried to get it into disk mode and then pluging it into the computer?
05:19:45tdoggettedisk mode is select+play, right?
05:20:03psycho_maniacyes right when the apple logo appears
05:20:56 Quit sarixe ("(EE) Failed to load "quit" module")
05:21:09tdoggetteThat works.
05:21:23tdoggetteI guess I'll do that when I cant to load things on
05:27:53 Join zack [0] (i=59a8103d@gateway/web/cgi-irc/
05:28:13psycho_maniachello zack
05:28:33zacki can't play wma on my ipod with rockbox?
05:28:48krazykityou can.
05:28:50zackmy nephew manage to make it work today
05:28:55scorcheyou can as long as they arent infested with DRM
05:29:00krazykitor rather, you should be able to play unprotected wma
05:29:24zackmy nephew gave me an unprotected wma
05:29:26krazykitdrag the files on the ipod, play as normal
05:29:43zackstill my ipod detect the file, but freeze
05:30:02krazykitare you using the latest build?
05:30:07psycho_maniacwhat model?
05:30:07zacki'm under rockbox wich i live
05:30:15zackipod 30gb
05:30:51 Join Mouser_X [0] (
05:31:11 Quit krazykit ("brb.")
05:31:17psycho_maniacdo you have the latest default build for your player? there updated more then once a day.
05:31:36zackversion 15047
05:31:53zackhow do you update once you already install the new version
05:32:12zackcan you update live
05:32:16psycho_maniacjust extract the zip file to the root of your player.
05:32:17zackvia interent
05:32:26psycho_maniaci dont not know what "update live" means
05:32:51zacklike plug your ipod rockbox and then it does it for you
05:33:15zackok, because i installed it yesterday
05:33:28psycho_maniacwhere did 1504 come from?
05:33:44zackand my wma file doesn't work, even though it works on my ubuntu laptop
05:34:12zackis it the latest version
05:34:26psycho_maniacthe newest is 15731 your version is a bit old
05:34:50zackok maybe that explain, but i downloaded it from the rockbox website
05:35:18zackwhere can i find it
05:35:18psycho_maniacjust update it now and it will have the most current build for your player.
05:35:57Mouser_XGet it from the "Current Build" page.
05:36:15 Join krazykit [0] (
05:36:15 Quit zack ("CGI:IRC (EOF)")
05:36:22psycho_maniacgo to look on the left and see "current build"
05:36:26 Join zack [0] (i=59a8103d@gateway/web/cgi-irc/
05:36:33psycho_maniacgo to look on the left and see "current build"
05:36:49zackok no problem i'll see you in a minute to check with you guys!
05:41:14 Join btwizt [0] (
05:41:26btwiztHey guys its me again teh didiot witht eh c200 v2
05:44:10btwiztok guess now is not teh time for my brain storm
05:45:23Mouser_XJust because no one says anything, does not mean no one is here.
05:45:42 Nick Mouser_X is now known as No_one (
05:45:55 Nick No_one is now known as Mouser_X (
05:46:01zackit works !!!
05:46:08zackthank you man
05:46:13zackall sorted
05:46:22btwiztWell im thinking that since the only way i can see for my C200v2 to get fixed is for Sansa to send me the firmware, what if i give that firmware tehy send me to teh Rockbox team to help the development for this aprticular version
05:46:29psycho_maniacnow on the main page it will show you what has been updated in the current build
05:46:46 Quit scorche (Nick collision from services.)
05:47:07 Join mokkurkalve [0] (
05:47:12psycho_maniacjust read the "what" part and you should be able to understand it.
05:47:14 Join scorche [0] (n=scorche@rockbox/administrator/scorche)
05:47:31btwiztsorry im typing fast
05:47:43btwiztodd mixture of being really pissed off and genius
05:49:38btwiztjsut thought maybe having the firmware would help them figure out if Rockbox can run on a c200 V2
05:49:54Mouser_XQuestion: I saw that there's a Soundwave MP3 player for about $100. I'm honestly (*Very* slightly) considering buying one, and donating it to Rockbox...
05:50:11Mouser_XIt'd be awesome to say "I Rockboxed Soundwave!"
05:50:22Mouser_X(Yes, I'm refering to the Transformer.)
05:50:36 Join Workaphobia [0] (
05:50:46lostlogicMouser_X: any idea what's in it?
05:50:55btwiztI rocked Soundwaves Box
05:51:00Mouser_XThat's the problem. No idea at all, in anyway whatsoever.
05:51:16Mouser_XIt doesn't appear to have a display though.
05:51:25 Quit btwizt ("Wake me up when september ends")
05:51:27Mouser_X(That'd make navigation difficult.)
05:51:31lostlogicwe have voice :-P
05:51:47Mouser_Xlol. Awesome.
05:54:27 Join briantumor [0] (
05:54:48briantumori have a sansa c250
05:54:58briantumori don't know wether or not it's a v2
05:55:30psycho_maniacshouldnt it say in the manual of the player?
05:55:52briantumordoesn't say which version
05:55:55krazykitbriantumor, did the box advertise audible support?
05:56:08briantumoraudible support?
05:56:10Mouser_XMaybe he has no manual? While it's very likely that my Gigabeat came with a manual, I don't actually remember ever seeing it...
05:56:19advcomp2019what version of firmware do you have
05:57:09mokkurkalveit say's on the wiki that the c200 also has a "v2" on the backside if it's a... v2
05:58:01briantumoryeah, i didn't see a v2 anywhere
05:58:33briantumorin the version.sdk file.. it says... Product: c250 \FW: 01.00.03A \Region: America \Build Date: 2006.10.04
05:58:54advcomp2019that is a v1 not a v2 then
05:58:59mokkurkalveThe warning against the v2 are now also on the e200 wiki page. I was wondering of getting a e280 as my H300 are too big for exercice situation...
05:59:13briantumorok, so it's compatible then? ^_^
05:59:35mokkurkalveGuess that means I have too look for the right version (v1) then...?
05:59:47briantumori got mine for free ^_^
05:59:55advcomp2019yea, it should be if it has 1.00.03a
05:59:57briantumorfrom dell when i ordered my laptop
06:00:43psycho_maniacgood deal briantumor
06:01:19mokkurkalveSo do you know if the e200 series have changed chip...? Should I hasten to get an old one?
06:03:37krazykiti don't know that the e200v2 has even been seen in the wild
06:03:58JdGordonI tinhk the v2 not is more for the c200v2
06:04:51mokkurkalveI do't know anything −− I just saw that the same varning as on c200 wiki page was now up on e200 wiki page also...
06:04:57briantumorso does the voice recorder save to wav like the original firmware?
06:05:20krazykityou can change the recording format
06:05:27briantumorcool, to what?
06:05:36krazykitwav, wavpack, and mp3 i think.
06:05:43briantumornice ^_^
06:05:54briantumorcan i also play doom on it? ^_^
06:05:54mokkurkalveand aiff
06:06:01advcomp2019i think the e200v2 series are not out yet
06:06:03psycho_maniacbriantumor: i suggest you read the rockbox manual for your player.
06:06:10briantumorok sorry :\
06:08:56jhMikeShmmm...fmradio exit on e200 doesn't work now. (supposed to be power (menu) release but nothing happens)
06:09:05briantumorvideo support??
06:09:13mokkurkalvewell, if I want a rockboxed e280 I should probably purchace it now then...
06:09:34psycho_maniacif you read the manual it will tell you.
06:09:48briantumori know.. i'm reading :S
06:09:51*jhMikeS and you can't turn it off! :P
06:11:07jhMikeSah, misread, can exit can't shut the radio off
06:11:23 Quit zack ("CGI:IRC (EOF)")
06:11:43LloreanjhMikeS: Power should stop the radio.
06:11:50jhMikeSit's isn't working
06:12:06LloreanWas the radio stoppable before?
06:12:44LloreanThere's a glitch
06:12:49jhMikeSis has the previous key as BUTTON_UP
06:12:50LloreanI didn't remove the PRE_ dependency
06:13:08LloreanI'll go ahead 'n fix that then, shall I?
06:13:24jhMikeSabsolutely...was driving me nuts :)
06:13:29LloreanI don't think it's a very good button combo. ;)
06:14:20 Quit J3TC- (".UPP.")
06:16:08 Join iamben [0] (
06:16:24LloreanUnless I typoed, should be all better.
06:16:35LloreanThat being said, does "Power" stop the FM radio if you leave the radio screen and go to the menu?
06:18:49jhMikeSthat file needs $Id$ in there too
06:19:04 Join J3TC- [0] (
06:19:15jhMikeSACTION_FM_POWER should actually stop the radio
06:20:24 Quit Workaphobia (Read error: 110 (Connection timed out))
06:21:50LloreanAnd, ACTION_FM_POWER? while I'm asking questions without verbs.
06:21:59LloreanOh, wait.
06:22:06LloreanI should read everything you say before questioning.
06:22:12LloreanStill, I'm not sure what you're asking
06:22:31jhMikeSall good now
06:23:09LloreanI should've fixed that with my initial keymap change, but being the complete doofus I am, I just didn't think "New screens have been added since I originally wrote this patch"
06:23:15jhMikeS$Id$ gets replace with such as $Id: pcmbuf.c 15668 2007-11-18 17:12:19Z jethead71 $
06:23:43LloreanSo I apparently deliberately created more inconsistency in the keymap with my original change. =/
06:23:58 Join radinp [0] (
06:24:20jhMikeShehe ... I'll check recording
06:24:24 Join TradeJack [0] (
06:24:57LloreanOh, that's an I not an L
06:25:06LloreanIn the big comment at the top of the file, gotcha
06:25:21LloreanI'm a little off tonight, I think
06:26:05*jhMikeS too ... can't get a single constant correct the first time :p
06:26:44LloreanThe only change in the recording screen is making the menu come up in a single press of BUTTON_DOWN rather than holding it
06:27:20LloreanEverything else is actually consistent with how the new radio and WPS style mappings are anyway, pretty much
06:27:46LloreanOr at least as reasonably as one can expect while maintaining usability and vastly disparate functions
06:28:45 Quit iamben_ (Read error: 110 (Connection timed out))
06:28:47LloreanI'm kinda surprised it had a long-press function without a short-press function, honestly
06:28:58jhMikeSI think scan/preset using the record button is a bit odd though that never changed. Would make sense to just jump to record.
06:29:11LloreanjhMikeS: I agree, actually
06:29:22LloreanBut I think the radio controls are a bit odd in general
06:30:06jhMikeSnot sure what would be scan/preset then...long down or long up? only choices really
06:30:45LloreanACTION_FM_PRESET sets a preset?
06:30:51jhMikeSthe long play button is often used for that
06:31:20LloreanDoes the recording screen even have a shortcut-to-recording function yet?
06:31:22jhMikeSyeah, but I'm talking about ACTION_FM_MODE
06:31:53LloreanWell yeah, I was just trying to figure out if some other key was swappable
06:31:56jhMikeSactually that's the presets list
06:32:17LloreanSeems odd on Select then
06:32:40LloreanDo you know if two buttons can have the same action, btw?
06:32:45jhMikeSnot unique though
06:33:04jhMikeSyes is is
06:33:08LloreanI'd really, _really_ like to have both _LEFT and _POWER be ACTION_STD_CANCEL
06:33:13LloreanSince _POWER is completely unused in _STD
06:33:39LloreanMan, if I were un-blind, I'd see _STD_OK defined twice
06:33:44jhMikeScan't cancel radio with left :)
06:33:51LloreanRadio isn't STD though
06:33:55LloreanBut the Text Editor is.
06:34:17LloreanAnd it uses STD_CANCEL to exit, and I'd kinda like to leave Left being able to cancel out of things, but power too.
06:34:48jhMikeSthe button is the key to the table, not the action
06:34:55LloreanAlright, should I put FM_MODE on long-play while I'm in there then?
06:35:07LloreanRemove it from Rec so it's freed up for later use and people can get used to it now?
06:35:10jhMikeSyeah, much more like others
06:35:14briantumorwhat's voice ui?
06:35:14LloreanWill do, then
06:35:17briantumorvoice command?
06:35:25Lloreanbriantumor: Reads the menus to you
06:35:40LloreanIt speaks the menus
06:35:44LloreanFor blind users
06:37:25jhMikeSdaily voices aren't being built yet in speex? hmmm.
06:37:39LloreanjhMikeS: They aren't? It should've happened automatically
06:37:59jhMikeSI dont't see the links except for hwcodec
06:39:11LloreanThen it's probably just broken entirely
06:39:19LloreanMaybe the script is still trying to pass it an encoder, and mucking things up?
06:40:14jhMikeSNo idea. Still alot of scrambling to straighten out the tools I guess.
06:41:15 Quit radinp (Read error: 113 (No route to host))
06:41:36 Nick billenium is now known as BilleniumZzZ (
06:43:05 Join iamben_ [0] (
06:44:31briantumorhrm... sansa c200 firmware is ok for c250 as well?
06:44:57briantumorkk :)
06:45:06LloreanAlright, I think I'm on my last keymap change for the moment.
06:45:11LloreanLeft something out on the gigabeat actually
06:48:48 Quit joshin (Read error: 113 (No route to host))
06:49:43 Quit iamben (Read error: 104 (Connection reset by peer))
06:51:02briantumoron the features list it says mp4.. does that include video as well?
06:51:35LloreanWhat features list, specifically?
06:51:50jhMikeSisn't that AAC?
06:51:57LloreanMP4 is a container
06:52:19LloreanIt can contain an awful lot of stuff, so none of our features lists should specifically say it's a supported format of anything.
06:55:51LloreanI suspect he also doesn't realize that when I ban him, it won't just be his username.
06:56:00briantumorso.. if i put rockbox on an ipod video.. it would lose it's ability to play mp4 videos?
06:56:17lostlogicyou could boot to the original firmware to play them.
06:56:23lostlogicrockbox does not play them though.
06:56:28Mouser_XRockbox couldn't play the MP4 videos, but the OF would still work.
06:56:30briantumoroh.. you can dual boot?
06:57:06Mouser_XIn most cases, yes. You have a c200, right? If yes, then yes, it dual-boots.
06:57:17briantumori mean for an ipod video
06:57:21briantumormy friend wants to try it
06:57:27lostlogicdual boots by default on the ipods as well
06:57:32Mouser_XYes, the iPods all dual-boot as well.
06:58:02lostlogic(until you fsck up the OF part of the partition as I have, I wonder if I'll ever restore that...)
06:58:23Mouser_Xlostlogic: What player? Sansa e250?
06:58:37 Quit scorche (Nick collision from services.)
06:58:38lostlogicMouser_X: ipod video
06:58:41lostlogicwhere I don't need the OF
06:59:06 Join scorche [0] (n=scorche@rockbox/administrator/scorche)
06:59:06*Mouser_X hasn't seen the OF of his Gigabeat in nearly a year.
06:59:19Mouser_XIt's been about 10 months.
07:00:05*jhMikeS has never seen gigabeat OF
07:00:25Mouser_XDumb question: With very little knowledge of C, how difficult would it be to port something to Rockbox, if you have the source code? I assume it'd be difficult, if even doable at all.
07:00:37jhMikeSwhy the assumptions?
07:00:45 Quit Calcipher ("I-n-v-i-s-i-o-n 2.0 Build 3515 with A Pack Fix By")
07:00:52psycho_maniacon the gigabeats can i plug it into usb and then turn it on? i think this is a problem on the ipods. but can you do this on the F40s?
07:01:01jhMikeSI suppose you'll get the C knowledge working on it.
07:01:23Mouser_XActually, I should say "With very little knowledge of programming langauges in general, let alone C,"
07:01:50lostlogicMouser_X: it's be a long process, but not inherently difficult
07:01:54jhMikeSWell, I mean if you succeed you'll definitely have the knowledge :)
07:01:58Mouser_Xpsycho_maniac: Yes, you can.
07:02:00lostlogicMouser_X: you'd hafta first read pretty much all of the OF and rockbox' code
07:02:08Mouser_XIt'll go into bootload USB mode.
07:02:45*Llorean is an idiot. AGAIN
07:03:22Mouser_Xlostlogic: Why would have to read the OF code?
07:03:27Mouser_X(And, how?)
07:03:50lostlogicwait, I misunderstood, I read bad
07:03:57Mouser_XI see.
07:04:36lostlogicit depends on what it is then... some things make use of lots of libraries that wouldn't be available on rockbox, others use only lower level functions for which rockbox has an analogous call to use instead
07:05:48Mouser_XIt's a Winamp plugin that plays Wonderswan music rips (it's an emulated format, thus it'd be atomic).
07:06:06Mouser_XReally, nothing special at all. More curiousity than anything.
07:08:07lostlogicMouser_X: for codecs it depends largely on whether they are an integer or floating point implementation to start with
07:08:20 Join radinp [0] (
07:08:53lostlogicif they start out integer based then it's usually quite easy to port a codec to rockbox by replacing file reading and buffer writing calls with those on the codec api, removing the use of malloc (as a rule) and yielding after frames
07:09:21 Quit qweru ("moo")
07:09:49*Llorean wants someone to ban him from making commits for the rest of tonight after the red is cleared up.
07:09:50 Nick BilleniumZzZ is now known as BilleniumzZz (
07:10:50***Saving seen data "./dancer.seen"
07:11:32*lostlogic offers Llorean a beer
07:12:40*krazykit thinks Llorean needs to hit xkcd's magic drunk level for coding
07:13:00Lloreanlostlogic: This is all because I decided to be nice and leave "Select" as an alternate pause button in Mpegplayer on gigabeat.
07:13:08LloreanI should've just said "Screw you guys, it's the same as the WPS" and been done with it
07:13:47LloreanBut I said "Hey, I should go ahead and leave it, it's not like there's something else for that to do" and didn't think of the other little tweaks. It's a pathetically simple patch, which is of course why I pathetically didn't notice the other bits
07:13:48scorchelostlogic: we tried at devcon...remember? :(
07:14:03LloreanI've found a beer I don't detest finally.
07:14:13 Quit J3TC- (Read error: 113 (No route to host))
07:14:20lostlogicdo tell!
07:14:25LloreanGuinness Extra Stout
07:14:32LloreanAs it doesn't have that whole _beer_ taste going for it
07:14:41lostlogicooh, good choice at least.
07:14:49scorcheLlorean: try Fat Tire
07:14:52lostlogicyeah, it doesn't taste hoppy
07:14:55Lloreanscorche: I've tried Fat Tire.
07:14:58 Join J3TC- [0] (
07:15:12LloreanIt's got a minimal detestation factor, but it's still in the "Assuming I had a choice, I'd avoid it" category does the job well for me
07:15:24LloreanThey have it on tap at a movie theatre here
07:15:27LloreanI think it's on tap.
07:15:30scorchewell, we know what to get you next devcon :)
07:15:58lostlogicugh, copying 15 gigs of replacement rips to an ipod takes a while
07:20:02psycho_maniaclostlogic: try doing over 30 :(
07:20:17radinpSo, what does everyone think of the recent encryption thread on the dev mailing list?
07:20:25lostlogicpsycho_maniac: I'll try to avoid it −− this is the first time I've reloaded my video since I've had it
07:20:43LloreanOkay, I've resolved my red.
07:20:48LloreanNo more commits while I'm doing other things
07:21:01Lloreanradinp: I think it's silly
07:21:05lostlogicnaw, red builds are fUN
07:21:17LloreanIf you want data on your DAP encrypted, encrypt it while putting it on the DAP
07:21:19lostlogicit's like christmas
07:21:38psycho_maniacred and green
07:22:29radinpLlorean: really? It seems to have potential. i could imagine it being used as a portable certificate container.
07:22:59Lloreanradinp: What exact use do you have in mind for it?
07:25:32radinpLlorean: Well, in my case I use OpenVPN quite extensively. It would be useful to leave my certificates and keys on rockbox. This eliminates the needs to store keys on the host computers.
07:25:40 Part TradeJack
07:26:28radinpLlorean: But I do agree that performing encryption on rockbox seems unnecessary.
07:26:55psycho_maniacwhat if the cia steals your dap for data on it :_
07:29:33radinpOk, well in that case we'll need to encrypt the keys themselves. However, any DAP should be able to handle a small amount of encryption.
07:33:10 Nick BilleniumzZz is now known as BilleniumZzZ (
07:36:06 Quit midgey ()
07:36:40 Join midgey [0] (
07:40:23lostlogicugh, album art is kinda neat and I hate myself for saying it. I really hate myself.
07:41:56Guerini hate you too
07:42:23lostlogicnow if only all of the existing WPSs using it that support stock builds didn't suck.
07:42:35LloreanI was thinking that I might do something else with the album art space.
07:42:56lostlogicon buffer? on screen?
07:42:59LloreanWell, I mean the file is cover.bmp, but nothing says it has to be album art
07:43:05psycho_maniacim tryiing to edit a wps and i do not know what this means %pb|height|leftpos|rightpos|toppos| what does the |height| do? i try editing it and it does nothing.
07:43:21LloreanFor example, a lot of my music is videogame remixes, I could show the logo of the gamesystem it came from, or a sprite from in-game.
07:43:29LloreanMaybe some sort of genre logo for "real" music
07:43:34lostlogicpsycho_maniac: how many pixels high the progress bar is in theory
07:44:16lostlogicLlorean: oh sure −− and dont' forget that you can specify the album art image in other ways too (based on the 'artist' tag for instance)
07:44:18Mouser_XIt works for me. I've used it to place the progress bar at a specific location on the WPS before.
07:44:39Lloreanlostlogic: Yeah, I thought of that. Portraits of the various artists.
07:45:04LloreanI just want something more interesting than an album cover is to me
07:45:32lostlogicI'm going to have album art and at least one album art theme on my players so that I can 'show off' rockbxo to friends ;)
07:53:34psycho_maniacthis is my problem with the progress bar.
07:54:00lostlogicpsycho_maniac: you want to adjust the toppos (top position)
07:54:13Lloreanchange toppos to about 3 or 4 less than it is
07:55:42psycho_maniacall mine shows is %pb|12|0|320|
07:55:51psycho_maniacso i would go %pb|12|0|320|4| ?
07:56:31lostlogicpsycho_maniac: -4 probably
07:56:42LloreanI think toppos is an absolute y coordinate, but I'm not sure
07:56:53lostlogicoh, then 200ish
07:57:00lostlogicand then adjust from there
07:57:01 Join OlivierBorowski [0] (
07:57:17psycho_maniacalright will try that. thank you
07:57:18lostlogicpsycho_maniac: by default the progress bar is placed at a _line_ as if it were text
07:57:38lostlogicpsycho_maniac: so now you are going to position it by pixels (if LLorean is right) as though it were a bitmap
07:57:52Mouser_XSometimes, you'll need to place it a line above, and then modify the coordinates to get it low enough.
07:58:15Mouser_X(Or, a line below, or whatever.)
08:00:40psycho_maniaci understand now
08:01:07*Llorean kinda wishes the progressbar just took x1, y1, x2, y2
08:01:18psycho_maniacthat would be a lot easier
08:01:41LloreanThe problem is that they were just tacked on the end of it, as it evolved, so as not to break previous generation WPSes
08:02:10LloreanThen again, when viewports happens, %m will probably be removed, which means if WPSes are broken anyway, it'll be a chance to fix it
08:02:27*scorche sees "when"
08:02:28 Quit OlivierBorowski (Read error: 104 (Connection reset by peer))
08:02:51Lloreanscorche: My assumption has always been that it _will_ happen, given an infinite timeline
08:03:20*scorche tries not to make assumptions
08:04:14LloreanSometimes you have to assume one way or another.
08:04:47LloreanFor example, if you say "if" then you're making the assumption that it *might* not happen, but it's still an assumption based on the fact that it's been said that it will happen.
08:04:53*Llorean shrugs
08:05:04*Llorean disappears to do other things
08:06:06scorchei dont consider accepting that there is a possibility of a binary situation that can go either way, going either way an assumption
08:06:15scorche(if you could understand that sentence...)
08:11:23 Quit nanok (Read error: 110 (Connection timed out))
08:11:51 Quit thegeek_ (Read error: 110 (Connection timed out))
08:14:40 Quit GodEater (Read error: 110 (Connection timed out))
08:15:48 Quit BigBambi ("Please insert girder")
08:26:33 Join GodEater [0] (n=bryan@rockbox/staff/GodEater)
08:30:27 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
08:31:28 Quit psycho_maniac (Read error: 110 (Connection timed out))
08:39:34 Quit n17ikh|Lappy ()
08:39:52briantumorgot my sansa c250 pimped out with rockbox ^_^
08:39:58briantumorthanks :)
08:43:12 Join FOAD_ [0] (
08:44:23 Part Llorean
08:45:01 Join Rob222241 [0] (
08:47:35 Join CaptainSquid [0] (
08:48:01 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
08:50:37 Quit hcs (Remote closed the connection)
08:50:47 Join psycho_maniac [0] (
08:51:36 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
08:52:45 Join stripwax [0] (
08:57:18 Join ender` [0] (
08:58:40 Quit psycho_maniac (" HydraIRC -> <- Wibbly Wobbly IRC")
08:59:02 Quit stripwax ("Miranda IM! Smaller, Faster, Easier.")
08:59:56 Quit BHSPitMonkey ("Leaving")
09:00:11 Join BHSPitMonkey [0] (n=stephen@
09:00:13 Quit FOAD (Read error: 110 (Connection timed out))
09:00:13 Nick FOAD_ is now known as FOAD (
09:02:52 Quit Rob2222 (Read error: 110 (Connection timed out))
09:03:29 Quit Guerin (Read error: 110 (Connection timed out))
09:05:13 Join petur [0] (n=petur@rockbox/developer/petur)
09:05:50 Quit hcs ("Leaving.")
09:06:12pondlifeJdGordon: ping
09:06:21LinusNthe CrossCompiler wiki page is a complete mess
09:06:44 Join Shaid` [0] (
09:06:52LinusNwhy do we write so much about all the compiler versions that we *don't* recommend?
09:07:16LinusNdamn, we should just say what versions we support and that's that
09:07:23 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
09:07:32peturso that next time somebody wonders why we're not using it, we still know why ;)
09:08:02LinusNwe could still add some footnotes about the other versions, but not clutter the entire instruction with it
09:08:06GodEaterput that on a different page somewhere
09:10:07LinusNfor example, we say in the table on top that you should use binutils-2.17, and then later on say that you can't use it to build iriver targets
09:10:09 Join thegeek [0] (
09:10:54***Saving seen data "./dancer.seen"
09:11:31JdGordonpondlife: hey
09:14:26JdGordonlinuxstb: you round?
09:18:41 Quit shodanX (Remote closed the connection)
09:18:45 Join shodanX [0] (
09:25:08 Quit Shaid (Read error: 110 (Connection timed out))
09:25:08 Nick Shaid` is now known as Shaid (
09:26:26pondlifeJdGordon: Back now, sorry
09:26:33pondlifeAbout the squeal...
09:27:00pondlifeI'd guess it's either caused by overlapping beeps, or very short beeps
09:27:00*GodEater prods at Bagder again
09:27:24pondlifeDoes the problem go away if you make the beep length much larger (100, say , rather than 2)
09:27:40JdGordonill try
09:27:45JdGordona tad busy atm though
09:27:47pondlifeNo use as a keyclick, but good for diagnosis
09:27:51pondlifeOK, whenever
09:27:54pondlifeReport back on FS
09:28:55LinusNi guess i can remove the calmrisc stuff from CrossCompiler?
09:29:08 Quit hcs ("Leaving.")
09:29:14JdGordonpondlife: which file was the beep call in again?
09:29:25*JdGordon will have a play now while he waits for this bloody file to downlaod
09:29:41JdGordonaction.c ...
09:30:04pondlifeThe second parameter to pcmbuf_beep() is the length
09:30:19pondlifeI don't have the code with me right now though
09:30:26JdGordonfound it :)
09:30:32 Join Zagor [0] (
09:30:39JdGordonmy sansa takes bloody ages to boot into OF now for some reason :(
09:30:43BagderGodEater: yessir!
09:32:53JdGordonhmm.. rolo seems to be stuffed again ? :(
09:36:35JdGordonpondlife: setting it to 100 makes it better.. but its still there
09:37:13pondlifeDoes it still happen if you (a) disable repeat and (b) are careful to let a beep finish before you click again?
09:37:23pondlifei.e. avoid any overlap
09:38:27JdGordonyes and yes, but very very infrequently
09:39:41pondlifeHmm...I'm out of ideas :/
09:40:09pondlifeI guess you'd need to try and debug in pcmbuf...
09:40:25pondlifeSomewhat hard without a target that shows the problem.
09:40:53JdGordonI have no problems with you commiting it and disable it by default for sansa if you want...
09:41:03pondlifeStupid idea, but what happens if you comment out the pcmbuf_beep call (i.e. replace it with a semi-colon)?
09:41:16JdGordonthat worked fine last time i tried it
09:41:20pondlifeAh, ok
09:41:25briantumorwhat's the best video resolution for sansa c250?
09:41:28pondlifeNo click = no squeal
09:41:43pondlifeWell, maybe I should exclude Sansa from the patch :/
09:42:04pondlife^ to JdGordon
09:42:14JdGordonis it only me thats noticed it?
09:42:15 Join OlivierBorowski [0] (
09:42:22JdGordonit might just be my sansa?
09:42:39pondlifeGodEater had it on his iPod Video, but it "went away"...
09:42:50pondlifeI suspect all PP targets suffer.
09:43:08pondlifeBut with a sample size of 4 devices, I can't be very sure
09:43:27GodEaterpondlife: oi - I tried and tried to get it to happen again
09:43:35pondlifeGigabeat: fine, H340: fine, Sansa: borked, iPod Video: fine
09:44:06pondlifeGodEater: I know, but there's no obvious reason why it wouldn't happen
09:44:28pondlifeProbably depends on dircache ;)
09:45:13pondlifeGodEater: So v5 is definitely not squealing for you?
09:45:22GodEaterI don't think I've tried v5
09:45:28linuxstbJdGordon: I'm around now.
09:45:34JdGordonpondlife: is it supposed to have a different beep for repeats?
09:46:10JdGordonlinuxstb: hey, i was thinking about viewports and the screen api again and I dont tinh it will cause a problem at all... just add the set_viewoprt call to the api and everything should be fine
09:46:11 Join davina [0] (
09:46:34JdGordonas long as any screens using multiple vp's set it to the default vp on exit
09:46:36pondlifeJdGordon: Nope, just the one click
09:46:51pondlifeGodEater: It's the same as v3
09:46:57JdGordonok, im pretty sure the repeat one sounds diferent.. but i couldbe wrong
09:47:11briantumorthis is awesome!
09:47:15briantumori love you guys
09:47:25JdGordoncommit it... if more people complain then we can try fixing it... it doesnt happen fery often at 100 and sounds fine
09:47:26pondlifeJdGordon: There's only one call to pcmbuf_beep and all parameters are constants.
09:47:41pondlifeIt needs to be a length of 2
09:47:42pondlifeNot 100
09:47:53pondlife100 sounds too beepy
09:47:58pondlife2 is more of a click
09:48:10pondlifeBut you're right, it ought to work.
09:48:18 Quit briantumor ("Leaving")
09:48:27pondlifeIf it shows up a problem in the underlying layers, then committing it might help get those fixed ;)
09:48:32JdGordonI also couldnt get it to squeel whe music was playing
09:48:45pondlifeNope, pcmbuf_beep takes different paths if playing or not.
09:48:58 Quit radinp (Read error: 113 (No route to host))
09:49:03pondlifeWhen not playing it starts and stops DMA just for the beep
09:49:14pondlifeOr something like that...
09:49:20pondlifejhMikeS will know.
09:49:24JdGordonmaybe thats the problem then?
09:49:57pondlife"pcmbuf_beep is borked on Sansa, and maybe on all PortalPlayer" would be the FlySpray headline
09:51:27pondlifeProblem is, it'll just get reported as "keyclick is broken".
09:52:13pondlifeI think I need to get a pcmbuf guru involved before commit.
09:52:20 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
09:54:28 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
10:04:40 Quit TMM ("Ex-Chat")
10:05:35 Join Casainho [0] (
10:05:50CasainhoGood morning to all :-)
10:05:53 Join waran [0] (
10:06:46Casainhocan anyone tell me If I can have a Task running at a higher frequency than the Tick_task, 100HZ?
10:09:14 Quit midgey ()
10:09:57Bagderwhy would you?
10:10:06 Quit OlivierBorowski (Remote closed the connection)
10:10:10Bagderthen you should have a proper thread I guess
10:10:20 Join OlivierBorowski [0] (
10:10:52 Quit tdoggette ("Leaving")
10:13:17Casainhobecause I tried to make a PWM signal... and the 100Hz are very low frequency...
10:14:09Casainhois there any PWM signal in RockBox? - I saw some references to PWM on source but I didn't understand... - If there are PWM in RockBox, how is that implemented?
10:14:38 Join Rob2222 [0] (
10:15:43CasainhoI used the add_tick_task(); ...
10:17:27linuxstbThe tick_task is 100Hz. If you want a higher frequency, you can use the user timer - see metronome for an example (rb->timer_register)
10:17:39pixelmais someone here with a c200 and a microSDHC (Zagor)?
10:18:13Zagorit's only a 1GB card though
10:18:34Casainholinuxstb: Where can I read about metronome?
10:18:42pixelmaZagor: same as me, that's why I was asking for a SDHC card :)
10:19:36pixelmajust wondering if someone could try that, because there's a report in the forums that a 6GB card wouldn't get detected on a c240
10:23:08 Join JdGordon_ [0] (
10:23:51 Quit Rob222241 (Read error: 113 (No route to host))
10:24:46 Quit JdGordon (Read error: 104 (Connection reset by peer))
10:25:25linuxstbCasainho: apps/plugins/metronome.c
10:25:56linuxstbCasainho: You could also search the source for other places register_timer() is used.
10:26:04linuxstbI mean timer_register()...
10:27:40amiconnlinuxstb, Casainho: A better reference would probably be the software backlight fading pwm
10:28:29Casainhoamiconn: is there PWM for some device?
10:28:55CasainhoI can't find info about working of timer_register()
10:30:22linuxstbCasainho: Have you looked at the current places it's used in Rockbox?
10:30:56CasainhoI am looking... I would like to find the function, with some help comments...
10:31:41CasainhoIt just says: /* Register a user timer, called every <cycles> TIMER_FREQ cycles */
10:35:04Casainhookok - I will look trough source and try to understand the function of timer_register()
10:38:44Casainhothank you :-)
10:42:01linuxstbamiconn: Is gcc still refusing to work on cygwin with the multilibs patch?
10:42:17amiconnI tried like 3 or 4 times
10:42:28amiconnIt compiles and "installs" with no error
10:42:59amiconnBut 'make install' decides there's nothing to do for many sub-dirs, even if I install into a clean location
10:44:01linuxstbDoes gcc itself get installed? i.e. is it just the multilibs it is skipping?
10:45:12*linuxstb wonders if anyone else has tried it
10:45:42BagderI think LinusN did...
10:46:29amiconnlinuxstb: No. The installed gcc doesn't work
10:46:48amiconnBut wiothout the multilibs patch, gcc can be built and installed fine on cygwin
10:46:49linuxstbBagder: Successfully, or the same result as amiconn?
10:46:56BagderI don't know
10:47:06 Join FOAD_ [0] (
10:47:39 Join lee-qid [0] (
10:49:22 Join BeChris [0] (i=509c2c01@gateway/web/cgi-irc/
10:49:54 Quit BeChris (Client Quit)
10:50:27 Join Aware [0] (
10:51:25linuxstbamiconn: Are the multilibs built? Typing "find . -name libgcc.a" in the gcc build directory should give lots of matches (I get 24)
10:51:46amiconnHmm, don't remember
10:52:12 Quit atsea- (Remote closed the connection)
10:54:28 Join spiorf [0] (
10:54:56Awarea quick question: ipod video now has hardware controlled backlight brightness. where can i adjust this?
10:56:12linuxstbProbably somewhere like Display -> LCD Settings
10:56:27waranI found a brightness-setting for my Nano(1st). But I wonder if changing this would drain more battery than it would save (because there is no hardware implementation; if I remember correctly).
10:57:23linuxstbWe (well, amiconn) found that the Nano has hardware-controlled brightness, and that's what he implemented for both the Nano and Video.
10:57:35waranahh, yay :D
10:57:44Awarei can't find it in lcd settings
10:58:12linuxstbAware: Are you running a recent (and official) version of Rockbox?
10:58:25Awareyes i am
10:58:41linuxstbWhich version? (it should tell you in the System menu)
10:59:17 Quit FOAD (Read error: 110 (Connection timed out))
10:59:18 Nick FOAD_ is now known as FOAD (
11:01:10linuxstbThat's not recent - the current version is r15736 and backlight brightness was added with r15599...
11:01:22pixelmaI wouldn't call that recent (current build is at 15736), backlight brightness got implemented in 15599... bah
11:01:35AwareARGH, sorry
11:01:42linuxstb"recent" in Rockbox terms means within the last few hours...
11:01:49Awarei see :)
11:02:15Awarei remember updating especially for this functionality... i'm going crazy
11:03:12waranI use yesterday's version =)
11:08:03Awareupdated and working nicely
11:09:18waranAware, if you are going to update your RockBox 3 times per day now just make sure to maintain at least one working backup. If the "current" version isn't running smoothly.
11:10:42pixelmawaran: the last 30 daily builds are archived and available for download (not saying that a backup is a bad idea)
11:10:55***Saving seen data "./dancer.seen"
11:11:05waranmhm, then
11:11:11Awareusually i only update when something on the majorchanges page interests me but your advice certainly is valid
11:12:58waranMaybe Apple should make RockBox the default OS for the iPod. Would offer about 100 times more functionality :)
11:13:22Awareturn it into a device worth using
11:14:09 Join zicho [0] (
11:14:14Awarei wonder how many ipods they have sold to people who plan to solely run rockbox
11:15:48waranIm one. It was a long time ago when I thought about which MP3-Player it will be. Then I did some research and found out: iPod is nice (the wheel) - and RockBox would make it useable. So I decided to buy an Nano with RockBox in my mind :)
11:17:01zichoWhich generation nano?
11:17:07waranThe "big points" I was looking for is 1) no iTunes-shit and 2) Vorbis playback.
11:17:14 Join barrywardell [0] (n=barrywar@
11:17:33waranzicho, 1st gen. RB does only work with the first one since Apple uses annother mainboard for the following generations.
11:17:45zichoAh, does it work well?
11:17:45 Join atsea- [0] (i=atsea-@gateway/tor/x-3b79af7ab6d898e1)
11:18:26waranzicho, are you kidding? It outperforms the originial firmware 100 times :)
11:18:27Casainhoanyone can tell me what means "rb->timer_register(...." - I am not good at programing... :-)
11:18:44Awarei was going to get an x5 but i couldn't get over its ugliness so i went the 5.5g. sold my soul...
11:19:05*petur kindly points to #rockbox-community
11:19:25waranpetur, is *-community the smalltalk channel?
11:19:40waranok, cu :)
11:20:56 Join irony [0] (i=c3488401@gateway/web/cgi-irc/
11:21:07ironyomg lots of ppl
11:21:25 Quit barrywardell (Client Quit)
11:21:51markunirony: "lol"
11:22:26pixelmaI don't see a single "ppl" around here ;)
11:22:50markunirony: anyway, welcome to the wonderful world of rockbox
11:22:55Bagderthat's short for "portal player lamer" ;-)
11:23:10ironymarkun: i was here a long time a go :)
11:23:30ironymarkun: i came up witht the name :)
11:23:40 Join barrywardell [0] (n=barrywar@
11:23:41markunwow, nice
11:23:42ironythats pretty much the only thing ive contributed with heh
11:23:54ironywell thats pretty much the only thing i did
11:23:57ironydamn lag
11:24:11markunare you still using your archos?
11:24:17ironyi had the jb6000
11:24:25ironyat the time, fall 2001
11:24:33ironynow i have an ipod video actually
11:25:19ironyi have rockbox on it of course
11:26:16markunof course :)
11:26:22ironywhat nich has daniel now again
11:26:31ironywhat was it
11:26:51ironyoh sorry
11:26:52*amiconn points to the IrcNicks wiki page
11:27:03markunirony: Zagor's his brother
11:27:08barrywardellBagder: here's the linux64 version of e200rpatcher:">
11:28:05BagderI am Daniel
11:28:57 Quit Gnu47 (Read error: 110 (Connection timed out))
11:30:57linuxstbNo, I am Daniel
11:31:34 Quit Aware ()
11:35:05 Quit jmspeex ("さようなら")
11:38:00Nico_Pthere's another USB question on the ML...
11:38:11Nico_Pat least this time things will be clear
11:38:59 Join moos [0] (
11:43:37 Join J-23 [0] (
11:45:55 Join bloody_devil666 [0] (
11:46:17 Join roolku [0] (
11:47:13roolkuNico_P: morning. :) did you see FS #8201 and do you have an idea what is causing it?
11:48:00Nico_Proolku: I've seen it but haven't had time to investigate... maybe this evening
11:48:25 Join Bagder_ [0] (
11:49:15ironyBagder: hi
11:49:41ironyi found you while looking for info on Vood boxes
11:49:46roolkuNico_P: cool - I got lost in playback.c :(
11:49:49ironydo you know a lot aboput them
11:49:52Bagder_Vood boxes?
11:50:08Nico_Proolku: it has improved :)
11:50:10Bagder_that's not me
11:50:37ironyBagder: tilgin adsl/router/voip linux based device
11:50:54Bagder_Mood you mean?
11:51:05 Quit stewball (Read error: 110 (Connection timed out))
11:51:10ironyqt3/embedded on Linux frame buffer on a sh4-linux for Tilgin.
11:51:21Bagder_yeah, I've done that
11:51:22ironynow i realize it wasnt for these devices
11:51:31ironyits for their set top boxes or?
11:51:32roolkuNico_P: yes I know, but I don't understand all this special casing with previous and current track and how they are the same sometimes
11:51:52Bagder_irony: yes, I poked mainly on their Mood400 thing
11:52:00irony understand
11:52:04 Nick fxb__ is now known as fxb (
11:52:17roolkuNico_P: (with regards to elapsed time)
11:52:25ironyI just got this vood adsl/router/voip box and it seems to be linux based, so i wanted to know what i can do with it ;)
11:52:35ironyits from telgin
11:53:05Nico_Proolku: ah yes... you need to know what the static structs are for and when they get copied from/to the main buffer (the bufgetid3 calls)
11:53:52Bagder_irony: I wouldn't even count on being able to login to it...
11:54:29Bagder_but it's a bit too off-topic for this channel, let's stop
11:54:36 Quit bloody_devil666 ()
11:54:42 Nick fxb is now known as fxb__ (
11:54:47 Part J-23
11:57:57*Nico_P is off
11:57:59 Quit Nico_P (Remote closed the connection)
11:59:52 Quit barrywardell ()
12:00:39 Quit waran (Read error: 110 (Connection timed out))
12:02:50 Quit Bagder (Nick collision from services.)
12:02:55 Nick Bagder_ is now known as Bagder (
12:03:50 Quit pondlife ("Read error: 110 (Connection slimed out)")
12:04:27ironyBagder: you're right, just wanted to ask you if you had experience with it. I can log in to it through the web interface and change loads of stuff.... well, thanks anyway :)
12:06:28 Quit irony ("CGI:IRC")
12:07:24 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
12:07:28 Part pondlife ("Gone")
12:08:37 Join ionvas [0] (
12:10:34 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
12:10:36 Part pondlife ("Gone")
12:11:14 Join Thundercloud [0] (
12:21:18 Quit Soap ()
12:24:27 Join J-23 [0] (
12:30:22 Part LinusN
12:31:43 Part J-23
12:34:14 Join PaulJam [0] (
12:38:59 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
12:39:09 Quit J3TC- (Read error: 113 (No route to host))
12:46:19 Join Webguest-Fully13 [0] (i=7cb74c05@gateway/web/cgi-irc/
12:47:22Webguest-Fully13Hi nerds.
12:47:56Webguest-Fully13No ones home..
12:48:31 Join Gnu47 [0] (
12:49:40PaulJamWebguest-Fully13: if you have a specific (rockbox related) question, just ask.
12:49:47 Join evert [0] (n=evert@
12:50:00everthello everybody :)
12:50:11Webguest-Fully13No time for hello.
12:50:14Webguest-Fully13Get to work on the Gigabeat S.
12:51:04 Quit OlivierBorowski (Remote closed the connection)
12:51:04evertI'm in search for a new mp3 player, i'm only having linux, so after a bit of research i saw rockbox. anybody here with a sandisk sansa e2x0 player ?
12:51:22Webguest-Fully13Whats your price range?
12:51:34evert150€ approx
12:51:38Webguest-Fully13Any prefered storage capacitys?
12:52:02evertas long as it is flash, i'd prefer 4gb (or more)
12:52:27Webguest-Fully13Can't think of anything really..
12:52:34Webguest-Fully13An Ipod video would be in your price range..
12:52:36evertin the local store i saw a sandisk sansa e270 with 6Gb and a extra case included
12:52:40Webguest-Fully13Although its not flash memory.
12:52:51Webguest-Fully13Rockbox is already released for the Video..
12:53:02evertWebguest-Fully13: hmm, but i'm not standing really positive towards ipods
12:53:03 Join OlivierBorowski [0] (
12:53:29Webguest-Fully13Arrr neither am I..
12:53:34Webguest-Fully13Zune possibly?
12:53:41evertthat sandisk is priced 140€
12:53:44Webguest-Fully13Not sure if it's linux compatible though.. Probably not.
12:53:55everti don't think zune is linux compatible ;)
12:54:17PaulJamalso, there is no rockbox for the zune
12:54:23Webguest-Fully13Gigabeat X-F?
12:54:41evertprevious i had a creative, but in my experience those players are not that good, bad firmware, no rockbox, not very good linux support
12:54:51Webguest-Fully13Old i know, but with 20-40gb storage capacitys, for around 20 pounds, you can't go wrong :P
12:55:29Webguest-Fully13Creatives are horrible...
12:55:36evertthat sandisk player is looking good, but how good is the player's case ? Can i drop it from a meter height without that i can trash it ?
12:55:52everti've never seen a sandisk player before, so i really don't know how good they are
12:55:55Webguest-Fully13Anything with Flash memory will withstand alot.
12:56:09 Join Febs [0] (
12:56:16Webguest-Fully13I've dropped my Gigabeat S several times..
12:56:44evertso that 6gb sandisk player for 140€ is a good deal (it will be cheaper online, but for mp3 i prefere the local store)
12:57:13Webguest-Fully131 pound is roughly $3 AUS yeah?
12:58:42PaulJamevert: if you get the sansa, make sure it is not a v2 (with audible support). i think i read that they differ from the normal e2x0 and rockbox doesn't run (yet) on those.
12:58:43Webguest-Fully13Right, converted it..
12:59:06Webguest-Fully13Doesn't seem like you're getting much for your cash..
12:59:26evertPaulJam: how can i see that ? The description says:
12:59:28evertSandisk SANSA E270 6GB ***INTERNET PRIJS***+ GRATIS CASE***
12:59:45evertWebguest-Fully13: what player would you suggest then for my price range ?
12:59:45 Quit Webguest-Fully13 ("CGI:IRC (EOF)")
13:00:06 Join Arathis [0] (
13:01:23 Join stewball`ghost [0] (n=WTFOMGBB@
13:02:57 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
13:05:08PaulJamevert: i think on the the package audible support is mentioned, so if you buy at a local store you can look at it. but i'm not really sure (mybe the v2 isn't even released). i'm not interested in the sansa myself and just thought i should mention it, so you are aware of the potential problem and can do some research yourself.
13:05:42evertok, so i've to make sure i've got a sansa player without that audible support ?
13:05:52evertthanks for mentioning it :)
13:07:02 Quit japc (Read error: 110 (Connection timed out))
13:09:36everti'm going to get a look in the local store, if it isn't the v2 (with audible support) i'm going to buy it :)
13:09:42everti hope it won't be a 'bad buy' :)
13:10:59***Saving seen data "./dancer.seen"
13:12:13 Quit lee-qid ("aufwiederbyebientotsayonara")
13:13:43 Join BeChris [0] (i=509c2c01@gateway/web/cgi-irc/
13:16:17linuxstbevert: The Sansa e200 series is one of the most popular Rockbox targets, so I don't think you'll be disappointed. (I don't own one though, so can't comment - I prefer hard-disk players).
13:16:56 Quit BeChris (Client Quit)
13:17:09linuxstbLinusN: Did you try building arm-elf-gcc under Cygwin with the latest multilibs patch?
13:17:14 Join BeChris [0] (i=40bf328a@gateway/web/cgi-irc/
13:18:12 Nick ionvas is now known as ionvas__ (
13:18:43ArathisJdGordon_: soo, any chance of committing your ".no_index" patch soon? :)
13:19:18 Quit spiorf (Remote closed the connection)
13:20:45 Join webguest70 [0] (i=509c2c01@gateway/web/cgi-irc/
13:20:46 Quit webguest70 (Client Quit)
13:21:32 Join kugel [0] (i=kugel@unaffiliated/kugel)
13:22:40 Quit BeChris ("CGI:IRC (Ping timeout)")
13:31:35 Quit moos ("Rockbox rules the DAP world")
13:31:36 Join J3TC- [0] (
13:39:34 Join AceNik_ [0] (n=AceNik@
13:42:03AceNik_guys i started using the equalizer on my h10[20 GB] recently its works fantastic without crossfade, but the deal is , battery levels are goofing around, a moment i have battery indicated at 47 the next moment after a song its 19 or something, can someone tell me whats up, i hav only 2.30 - 3 Hrs of playback
13:43:27 Quit AceNik_ (Client Quit)
13:43:31 Quit daurnimator (Read error: 110 (Connection timed out))
13:44:21preglowperhaps the eq sucks it down
13:44:34preglowit's not too easy on the pp cpus
13:46:00pixelmahmm... looking at the just committed greek.lang diff - there are phrases (especially in the beginning) that don't use the "feature:" in the dest and voice part (sometimes containing two lines with "*: "). Isn't that needed there too?
13:48:05LinusNlinuxstb: yes i have, and i am working on updating the cygwin packages as we speak
13:48:25linuxstbAh, so you didn't have the problems amiconn has reported?
13:48:35linuxstb(make install doing nothing)
13:49:10amiconnIt does something, just not in every sub-dir
13:52:06LinusNwell, it created all the libs, and i can build the mrobe binary
13:52:33 Quit J3TC- (".UPP.")
13:54:26 Quit Nico_P (Remote closed the connection)
13:57:08 Join J3TC- [0] (
14:08:44 Quit Casainho ("ChatZilla 0.9.79 [Firefox]")
14:08:54 Join OlivierBorowski_ [0] (
14:12:50preglowjhMikeS: around?
14:15:28LinusNok, the cygwin packages are updated with newer versions of the compiler, including the multilib arm compiler
14:15:50LinusNit might take a while before the download mirrors have caught up though
14:16:02amiconnHow did you manage to do that??
14:16:56JdGordon_Arathis: I tihnk we just need a better filename to check for and it can go in...
14:17:21 Nick JdGordon_ is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
14:17:25LinusNamiconn: the normal way
14:17:50LinusNi downloaded gcc and patched it with the multilibs patch
14:19:52LinusNamiconn: how can i tell if it fails?
14:20:31 Nick fxb__ is now known as fxb (
14:20:33LinusNthe gcc dir is full of libs and i can build m:robe
14:21:09 Nick fxb is now known as fxb__ (
14:21:47JdGordonany objections to FS #5960?
14:23:58roolkuJdGordon: does it check inside the excluded folders? I would like to exclude root and include only my music folder
14:24:28JdGordonno, it excludes the whole tree under the excluded folder
14:24:54 Quit scorche (Read error: 104 (Connection reset by peer))
14:26:11Tokihi! possibly a stupid question, but... how to disable recycle bin on a rockboxed player?
14:26:21roolkumy preference would be an include mechanism, rather then an exclude one (or a combination of both)
14:26:27 Quit OlivierBorowski (Read error: 110 (Connection timed out))
14:27:03roolkualso a less intrusive file (shorter, lower case and ideally hidden)
14:27:05 Join scorche [0] (n=scorche@rockbox/administrator/scorche)
14:27:18JdGordonI've changed the filename to database.ignore
14:27:48JdGordoni agree a white/black list would be better, but I think this is about as good as we would want to add
14:28:27JdGordonits purpose is really to just ignore recordings, podcasts and badly tagged folders.. so a proper list isnt really needed
14:29:30ArathisJdGordon: perhaps you could use a format similar to .htaccess where +/-indexes says if the index of a folder is shown or not. that way we could exclude and include dirs to get a more complex structure and the ability to just include /Music for example.
14:29:56JdGordonthats the white/black lists I was saying.. but I dont tihnk its worth adding
14:30:00peturToki: offtopic.... but as a hint: in windows, right-click the recycle-bin and select properties...
14:30:31amiconnLinusN: I can build nothing when installing a multilib-patched arm-elf-gcc ...
14:30:31JdGordonI'm also not sure how well it would cope with building from multiple "top levels"
14:31:02Tokipetur: oki tks :)
14:31:12LinusNamiconn: then i suggest you install the new packages
14:32:43preglowjmworx: yeah, i am working on a patch, should i eliminate all the unneeded copies i see too?
14:32:44amiconnAre they built with the same prefix I am using?
14:32:52roolkuJdGordon: I suppose it is a compromise, but I think it is a unclean trick to define a block just to be able to declare variables. :)
14:32:55amiconnAnd btw, I don't absolutely need multilibs...
14:33:40JdGordonroolku: yeah, I agree...
14:33:48roolkuJdGordon: I don't think it is neccessary to clutter up .rockbox with an database.ignore file - seems fairly unlikely someone will store audio files in it
14:34:35JdGordonllorean suggested (in march when this first came up) that the file should go in so people dont have to try and remember the filename
14:35:21roolkufair enough
14:35:46 Join mf0102 [0] (n=michi@
14:36:14LinusNamiconn: the prefix is /opt/arm, /opt/m68k and /opt/sh
14:37:48JdGordonbuggerit.. in it goes
14:43:55markunJdGordon: why don't we do a opt-in instead of opt-out system?
14:44:14LinusNmarkun: because then it won't work out of the box
14:44:41*pixelma wonders whether clarifying her 1-hour-ago-question would help getting it answered or if it's because no-one who knows is available at the moment..
14:44:43markunLinusN: why not? We could have the root as the default
14:45:26 Quit crwll ("tuunaus")
14:45:29LinusNmarkun: well, that might of course work
14:45:35*JdGordon wonders what pixelma asked an hour ago :p
14:45:46 Join Casainho [0] (
14:45:50CasainhoHello :-)
14:46:00markunCasainho: hi again
14:46:22JdGordonmarkun: because im not sure how the db building works and I'm not certain that you can call build_tagcache() multiple times with different paths
14:46:47Casainhoah... I would like to know how settings are saved.. from when we turn off player to torn on...
14:47:33pixelmaJdGordon: I wondered a bit about parts of the greek.lang patch that's been committed earlier today...
14:47:37Casainhoare settings saved in flash memory (Sansa example) ?
14:47:44JdGordonpixelma: yeah, you might be right... those extra *: none lines look wrong
14:47:46LinusNJdGordon: "len" is a funny name for a file handle ;-)
14:47:48markunJdGordon: I just think it's more anoying that I now have to clutter my tree with files when I just want to index my albums dir
14:48:22JdGordonyou dont have to
14:48:41markunJdGordon: just for my folders which also contain audio files?
14:48:48JdGordonpixelma: it probably works because the 2nd *: is overwriting the none part
14:49:01JdGordonmarkun: just for folders with audio you dont want indexed
14:49:14JdGordonand only the top folder if there is a whole tree you want skipped
14:49:38pixelmaJdGordon: but it shouldn't on targets that don't have the specific feature
14:50:24JdGordonif that feature is missing then the *: none is correct
14:50:30JdGordonmaybe i misunderstood you?
14:51:06pixelmamaybe, I see that I'm not explaining very well...
14:51:31JdGordonyou mean when the source and dest features dont match up?
14:51:39*JdGordon just noticed that
14:52:47JdGordonbah, stupid red delta
14:53:02pixelmayes, for example the complete LANG_REMOVE_MMC phrase in greek.lang
14:54:43JdGordonit depends how the .lng generation is done.. but I'd say revert it and have it fixed propeerly...
14:54:51JdGordonthose doubnle *: lines are definatly wrong
14:55:10*amiconn wonders what a fodler is ;)
14:56:10*JdGordon is happy to add some smiles to the crowd
14:56:30JdGordonthink how boring the front page would be if it wasnt for my imaginative spelling :)
14:57:41*preglow hides his fodlers
14:59:09 Join joshin [0] (n=josh@unaffiliated/joshin)
14:59:48*roolku is surprised how skipping fodlers fixes memory corruption when using .talk clips :p
15:00:07roolku(FS #5690)
15:01:19JdGordonwhat what with the who who?
15:01:43JdGordonoh fuck!
15:02:04*JdGordon goes to bed.. my brain is obviously alseep anyway
15:02:56 Quit JdGordon ("Konversation terminated!")
15:05:53preglowalseep, sounds like an indian name
15:09:27 Join weezerle [0] (
15:09:53 Join crwl [0] (
15:10:07 Nick BilleniumZzZ is now known as Billenium (
15:11:01***Saving seen data "./dancer.seen"
15:16:12 Nick parafin|away is now known as parafin (
15:24:44markunjhMikeS, linuxstb: any of you interested in his sample?
15:28:34 Join n1s [0] (
15:29:33linuxstbmarkun: mpegplayer is a different plugin to how i remember it now, especially with the resume feature...
15:29:40 Quit scorche (Nick collision from services.)
15:30:09 Join scorche [0] (n=scorche@rockbox/administrator/scorche)
15:30:53roolku...which jhMikeS is completely changing now anyway
15:31:40preglowjhMikeS: spcs sound gREAT on nano now
15:31:57*preglow bestows jhMikeS with "knight of rockbox" title
15:32:03preglowit comes with a funny hat
15:32:24 Quit atsea- (Read error: 104 (Connection reset by peer))
15:35:58preglowsweet, sweet spc love
15:36:54*linuxstb can't remember seeing someone love a codec so much...
15:37:13 Quit CaptainSquid (Remote closed the connection)
15:39:12preglowhmm, no seek support in spc yet
15:39:20preglowi wonder how well the spc emulator is performing
15:40:24markunpreglow: are you looking into seeking for wma?
15:40:26 Join atsea- [0] (i=atsea-@gateway/tor/x-bb57130b539960a8)
15:43:02preglowmarkun: not in the least
15:43:21preglowmy next project for wma is making it faster and syncing it with ffmpeg
15:44:19linuxstbAre there any codecs left we don't support? I can think of Realaudio (Cook) and ATRAC, neither of which seem very useful...
15:44:35linuxstbs/codecs/useful codecs/
15:45:33 Join AceNik_ [0] (n=AceNik@
15:46:10AceNik_guys how does the new update by Jonathan Gordon work regarding the database.ignore
15:46:16preglowi'll concentrate on improving what we have, that's for sure
15:46:46LinusNAceNik_: i'm sure he will update the manual shortly :-)
15:47:14AceNik_guys i started using the equalizer on my h10[20 GB] recently its works fantastic without crossfade, but the deal is , battery levels are goofing around, a moment i have battery indicated at 47 the next moment after a song its 19 or something, can someone tell me whats up, i hav only 2.30 - 3 Hrs of playback
15:47:28preglowthis thing pulls every spc off perfectly
15:47:29DogBoylinuxstb, realaudio not being very useful doesn't keep a lot of content providers from using it
15:47:30markunpreglow: this might have to be reverted as it's not mathematically correct:
15:47:40linuxstbAceNik_: If you create a file called 'database.ignore" in folder, then the database ignores all files in that folder (and all sub-folders).
15:48:08LinusNAceNik_: in short, for every folder (including subfolders) you want to exclude from the database search, add a file called "ROCKBOX_DATABASE_IGNORE_FOLDER"
15:48:30preglowmarkun: well, ok, i haven't checked it out too closely
15:48:33AceNik_linuxstb: but it would be just a file, containing nothing inside right? 0kb ? jus named "database.ignore" ?
15:48:38preglowmarkun: how incorrect is it?
15:48:40linuxstbLinusN: The name changed between flyspray and SVN...
15:49:19linuxstbAceNik_: Yes, Rockbox just checks if the file exists, so it can contain anything (or nothing)
15:49:56markunpreglow: I didn't hear any differences, but didn't compare the resulting wavs either
15:49:58AceNik_linuxstb, LinusN : thanks, can you help me with the EQ & battery prob on my h10 ?
15:50:01Arathiswith JdGordons latest commit I can't export my db anymore and my player won't shut down. I need to reset it and than it give's me the "building database [x/9]" screen
15:50:10preglowmarkun: wavs should always be compared when doing codec changes, imho
15:50:21markunyes, I should have
15:50:38preglowAceNik_: the problem is that you get less battery with eq enabled?=
15:50:55markunpreglow: still would be nice to have 1 fast imdct we could reuse
15:50:57linuxstbArathis: That doesn't sound good...
15:51:12Arathisbut at least it works properly itself. the dirs to be excluded from the database are excluded
15:51:27 Join MethoS- [0] (
15:51:32linuxstbDogBoy: Do "a lot" of content providers still use it? I thought almost everyone used wma (or mp3) nowadays?
15:51:41Arathislinuxstb: it's not good, but after the reset it works
15:52:00AceNik_preglow: yes i agree, that is the case, but battery levels are jumpy, & also sometimes they arent fully empty but they show empty & shutdown, but afetr 10 mins the player can be turned on again, & the player shows 30-40% battery remaining
15:52:17preglowlinuxstb: aac in particular needs shaping up
15:52:35n1slinuxstb: we still don't have xm and mod support!
15:52:36preglowlinuxstb: i'm in the process of syncing faad, but speex stuff keeps me busy
15:52:47linuxstbpreglow: Sure. I'm not thinking about existing codecs that aren't perfect, just things we're missing....
15:53:00preglown1s: and that would be the last codec i want that i'd use myself
15:53:04linuxstbMaybe add some missing codecs to SoC2008...
15:53:06preglowsomeone code it already :>
15:53:45Arathislinuxstb: when build the db in the settings the busy icon disappears and when trying to export the db it says failure. when going straight to the db from the main menu and building it there it stops when (possibly) all files are counted
15:53:49 Quit scorche (Nick collision from services.)
15:54:05linuxstbArathis: I know nothing about the DB (and don't even use it...)
15:54:05n1spreglow: i think i will look into midi as a proper codec (well no so much dec) soonish so any nasty architecturat difficulties should be removed :)
15:54:18 Join scorche [0] (n=scorche@rockbox/administrator/scorche)
15:54:30 Nick fxb__ is now known as fxb (
15:54:31AceNik_guys do you think for making EQ & battery more functional on h10 we could re bench the h10's & use those values for further builds for determinign battery
15:55:12 Join MethoS-- [0] (
15:55:16linuxstbpreglow: Are you planning a speex encoder for Rockbox?
15:55:20preglowAceNik_: the fact that eq sucks down battery can't be helped
15:55:22preglowlinuxstb: some day, yes
15:55:29preglowlinuxstb: but it's not really a priority
15:55:51preglowwill require an ocean of optimized code
15:55:54pixelman1s: since you worked at langV2... can you comment on the greek.lang changes in some phrases (see tracker entry)?
15:56:01AceNik_preglow; so you tink the eq is best disabled on the h10 then ? ok
15:56:18n1spixelma: sure, I'll have a look, gotta eat something first tho
15:56:20preglowAceNik_: i always think it is best disabled, but then again i don't even want to use it
15:56:46linuxstbpreglow: Do you use _anything_ you've coded for Rockbox? ;)
15:56:51pixelman1s: sure, thanks
15:57:00AceNik_ok thanks, but trust me it sure does boost quality of music to another level
15:57:21preglowlinuxstb: well, wakeup alarm :>
15:57:35preglowand i do enjoy my mp3 opts from time to time
15:59:01Zagorand here's my old friend the oops again...
15:59:12preglowZagor: give him my regards
15:59:19 Quit AceNik_ ("bye guys have fun , enjoy !!!!!")
15:59:23 Quit MethoS-- (Read error: 104 (Connection reset by peer))
15:59:28 Join japc [0] (n=japc@
16:01:44 Join MethoS-- [0] (
16:04:09 Join scorche|w [0] (n=42c007b2@rockbox/administrator/scorche)
16:04:51 Quit kkurbjun ("Leaving.")
16:11:34DogBoylinuxstb, anybody with any sense doesn't use it but I've ran into content I wanted to play on my portable that I needed better cook support
16:11:42*preglow vanishes
16:14:08 Quit MethoS- (No route to host)
16:14:56DogBoyarchives at for one
16:15:48DogBoyI would bet that's a common situtation too −−that a content provider would switch at some point to mp3 but not archives of their older shows
16:17:26jhMikeSpreglow: (for when you unvanish) Any change to the speex clicking situation?
16:24:43linuxstbDogBoy: What do you mean by "better" cook support? Does any portable support cook at all?
16:25:09 Quit Zagor ("Client exiting")
16:38:57jhMikeSpreglow: so I should be expecting a funny hat in the mail? :)
16:39:14 Quit evert ("Lost terminal")
16:39:25markunDogBoy: the same reason why I'm happy we support WMA now
16:42:53 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
16:46:57 Quit jhulst (Read error: 104 (Connection reset by peer))
16:49:11 Quit MethoS-- (Connection timed out)
16:54:04 Join ramon8 [0] (
16:56:49 Part LinusN
16:58:43 Join MethoS-- [0] (
17:00:53 Quit Casainho ("ChatZilla 0.9.79 [Firefox]")
17:03:49 Join XavierGr [0] (
17:04:53 Quit J3TC- (Read error: 110 (Connection timed out))
17:07:26 Join evert [0] (n=evert@
17:07:58evertand i'm back, i've bought a 8Gb sansa for 130€, not very cheap, but i can directly mount it under linux so i'm already quite happy with it :)
17:08:12evertwhat advantages would rockbox give me on it ? :)
17:08:25evertand what are the risks of putting rockbox on it ?
17:08:31lostlogicsee the why rockbox page on our site
17:08:47evertok, sry i've to rtfm first, i was a bit to enthousiastic ;)
17:08:50lostlogicthere is technically some risk of bricks, but the rockbox brick rate is vanishingly small
17:09:16evertlostlogic: when it would brick, can i return it to the shop and get a new of is it 'lost' then ?
17:09:25linuxstbAnyone else think "voice-driven interface" is misleading on the WhyRockbox page?
17:10:03lostlogicevert: in the unlikely event that you brick it beyond being able to restore through either manufacturer or recovery mode, I don't know, but that would be unethical so I'll not comment.
17:10:06jhMikeSlinuxstb: indeed
17:10:26linuxstbjhMikeS: Any suggestions?
17:10:48 Join The-Compiler [0] (
17:10:50evertlostlogic: ok, i hope it won't brick then, would be sad for a mp3 player what i can use 'out-of-the-box' under linux ;)
17:10:52jhMikeSvoiceable interface?
17:11:02***Saving seen data "./dancer.seen"
17:12:08GodEaterdo we actually know of anyone who has bricked their sansa to the state of being completely unrecoverable ?
17:12:25lostlogic"Speaking menus" "Speaking interface" "Spoken menus" "Spoken interface"
17:12:30linuxstbSome people seem to be incapable of recovering it - does that count?
17:12:38n1soptional talking interface?
17:12:40GodEateror is it only the case the cluesless newbies (or enthusiastic beginners) just can't be bothered to follow the instructions ?
17:12:55lostlogiclinuxstb: haha, I thought for a whiel I'd turned myself into one of those :-P
17:13:03everti'm not a linux newb or such, so following a guide can't be that difficult ( i hope :p)
17:13:19GodEaterevert: then your risk of "bricking" is extremely low
17:13:23lostlogicevert: if you follow it letter by letter it works fine, if you try to be smart (I did) you'll end up with a semi-brick ;)
17:13:27*linuxstb votes for "Optional spoken interface"
17:14:05evertok, thanks for the answers, i'm going to read the wiki pages to get a bit knowledge of rockbox and such :)
17:14:14evertit looks very cool and promising :)
17:14:30lostlogicevert: once you're onboard (you will be ;)) we welcome bug reports and suggestions.
17:14:34*GodEater will wait for the inevitable "ewww, it's SO ugly" comment then
17:14:45n1sI almost bricked my h300 today as i dropped it while riding amy bike and it was nearly run over by a truck ;)
17:14:51lostlogiclinuxstb: I like it −− might want to ask the blind users community what they would be most likely to search for too?
17:14:53evertGodEater: what do you mean ?
17:15:11GodEaterevert: it's what most people think of the default theme :)
17:15:14lostlogicevert: you'll see :)
17:15:39evertok, i'll see, i'm not giving a 'sh*t' about the looks, fucntionality is priority for me :)
17:16:49evert# USB handler ALERT! - The Sansa recognises when the USB has been plugged in, but does nothing more.
17:16:54evertwhat does it mean ?
17:17:12GodEaterit means Rockbox has no native USB stack at the moment
17:17:29GodEaterso in order to transfer content to it you have to use the original firmware (OF)
17:17:49evertok :)
17:18:08GodEatermind you, I have no clue where you read that - so I'm only guessing that's what it means
17:18:33everti read it on the 'rockbox sansa e200 port' page
17:18:36 Join H10_007quick [0] (
17:18:59lostlogicyeah, that's what it means
17:19:12lostlogicI can't wait until we get PP USB support, specially now that I have _two_ PP players.
17:19:24evertPP ?
17:19:26GodEaterI assume it'll work on the Gigabeat too ?
17:19:31AvePP is soon extinct
17:19:41GodEatersadly ?
17:19:51Avegreat little platform, no?
17:19:51*GodEater won't miss one more undocumented platform
17:19:55Avesad ti see it go..
17:20:04AveI mean really, what else IS there
17:20:13GodEaterall the other ones we support ?
17:20:15Aveeverything we get from now on is even MORE closed and drm-bull
17:20:48GodEatertell that to the telechips porters...
17:20:49Aveit looks like that
17:20:57n1sAve: pp is only open now because of some major RE work
17:21:04AveI hope the world opens up, but apple and microsoft make the rules
17:21:16GodEaterlike hell they do
17:21:17Avebut tha fect is that soon you can no longer get PP devices at all
17:21:25H10_007quickI'm trying to setup the compilers on my new Ubuntu install and every time I run the script it gives me an error message of "Patch required to run this script please install and re-run script" What patch do I need?
17:21:48n1sH10_007quick: the tool named "patch"
17:21:52GodEaternot _a_ patch - it means you need the binary called "patch"
17:22:10GodEaterH10_007quick: sudo apt-get install build-essential
17:23:09H10_007quickthankyou very much
17:23:57 Quit The-Compiler (Read error: 104 (Connection reset by peer))
17:24:09 Join The-Compiler [0] (
17:24:48 Join Siku [0] (
17:25:06Avenot too long ago, your player options were basically 1) el cheapo silicon mp3-only players 2) PP based ones 3) iriver-expensiveness
17:25:17 Quit H10_007quick ("ChatZilla 0.9.79 [Firefox]")
17:25:21Ave#1 also plays back wma, yippee
17:25:32GodEatermy iriver was cheaper than the equivalent ipod of the time =/
17:25:50Avesome brands even have vorbis support these days out of the box which is amazing
17:25:59GodEateras did my iriver at the time....
17:26:32Avebut rb gives me two features that no one else does at this pricepoint and availability: simple USM-file system and vorbis support
17:26:38GodEaterit had a better remote control than the ipod too
17:26:43AveI also like gapless playback
17:27:02GodEaterboth of which the Rio Karma also did
17:27:08Aveums even
17:27:14Avehas rio gone belly up?
17:27:18Avepast tense
17:27:33GodEaterdon't think so
17:28:22Avegonna have to check the markets, once you start filtering by say, vorbis playback and no need for special upload programs, the selection shrinks to near zero
17:28:31GodEateralthough I'm not sure they still make DAPs
17:31:48markunlinuxstb: I'm still puzzled by this:
17:32:34markunhis bootloader gives him "error: -1" but those errors are now replaced by actual text messages in the latest bootloader which he claims to have
17:32:38GodEatermaybe his ISP is caching the old bootloader ?
17:33:29GodEaterdoesn't he also need to remember to flick the battery off/on if he attempts to upgrade the bootloader ?
17:33:55markunGodEater: good point
17:34:02markunGodEater: can you tell him that?
17:34:09GodEaterhehe - you scared now ? :)
17:34:36markunno, just drank too much beer and afraid to say stupid things :)
17:35:00markun(only drank 1 beer actually, but with very little food all day)
17:35:43GodEaterdone :)
17:35:48 Quit MethoS-- (Remote closed the connection)
17:36:05*linuxstb waits for petur to arrive
17:36:15GodEatertwo pings
17:36:19GodEatercan't be long now :)
17:36:26*petur is too busy to even respond :(
17:36:35linuxstbNo you're not...
17:39:48amiconnSince when does eq boost the quality of the music? :>
17:40:37 Join daurnimator [0] (n=daurnima@unaffiliated/daurnimator)
17:41:06 Join daurnimas [0] (n=daurn@
17:42:19 Nick daurnimas is now known as zz_daurnimas (n=daurn@
17:44:50markunamiconn: didn't you know? :)
17:45:46markunit can even boost it to another level
17:46:45GodEaterindeed, I can listening to a mere tune, set my eq boost to "phat" and then I'm listening to a clearly superior tune :)
17:47:22*GodEater wonders if linuxstb's aoler.lang file ever included that term
17:48:30pixelmamarkun: another level doesn't tell much - could be quite low too ;)
17:48:45 Quit _pill ("changing servers")
17:49:16markunpixelma: but boosting something to a lower level also sounds strange :)
17:50:20supGodEater: you have bad hardware
17:50:30markunchanging speakers and headphones can make music sound quite different to me, so the EQ could probably too to some extent
17:50:33GodEateris that bad as in good ?
17:51:00markunbadest hardware!
17:54:21 Join nanok [0] (n=nanok@
17:54:43preglowjhMikeS: working on it now
17:56:09preglowjust tried with the new code now, some pieces still click
17:58:01 Join przemhb [0] (
17:58:38jhMikeSpreglow: Don't know what else I can do on that end of things. It does play the silence clip (some delay before music volume is brought back up).
17:59:33preglowjhMikeS: and there's no reset upon seeing -1 returned from decode_int, no? it just goes straight to a new clip if there's more?
18:00:20jhMikeSTry adding one in before setting the bits after getting the next clip and see if it changes things.
18:01:39jhMikeSI guess copy line 233 to line 327
18:02:32 Join _pill [0] (
18:03:10preglowwhy would i want to add a reset?
18:03:29 Quit petur ("stkov")
18:03:34 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
18:03:44jhMikeSyou were just asking about one? I wondered if you though that might be a reason.
18:04:08 Part hcs
18:04:23jhMikeSthat's where it will be after a negative value from decode_int
18:04:40preglowi was more thinking about removing one
18:04:51preglowjust wanted to make sure there's no reset between a clip and the silence clip
18:05:12jhMikeSthere won't be. only on explicit startup
18:06:52preglowno, the clip "file" is deinitely an init, followed by two clips
18:07:05preglowthen "two" comes along with an init, and another clip
18:07:13preglowi'm guessing each is followed by the silence clip
18:07:32 Nick ionvas__ is now known as ionvas (
18:07:58jhMikeSwhere are the clicks in that?
18:08:47jhMikeSI have another thought though. If the filter will ring, skipping the lookahead after decoding may remove a soft transition.
18:08:48preglowdepends on the clip
18:09:09 Quit idnar (Nick collision from services.)
18:09:10preglowjhMikeS: very good point
18:09:11 Join idnar_ [0] (n=mithrand@unaffiliated/idnar)
18:09:14preglowi'll try to remove that
18:09:27preglowset the lookahead only on init, that's the correct thing to do anyway
18:09:45 Quit linuxstb ("Leaving")
18:09:53jhMikeSmake sure to set it to 0 though
18:09:54preglowjhMikeS: weird, one doing "file" + "twenty" + numeral, sometimes there are three inits, sometimes only two
18:10:26jhMikeSso it's sending Q_VOICE_PLAY instead of just queueing multiple clips to play in sequence?
18:10:38preglowjhMikeS: that was it
18:10:48preglowproblem solved
18:11:04preglowi'll try with nb now
18:11:55preglowjhMikeS: i just removed the GET_LOOKAHEAD in the get_more() stage, shouldn't that suffice? that way it'll only be set after init
18:12:56 Join spiorf [0] (n=spiorf@
18:13:07preglowjhMikeS: you really do write very nice threading code, this stuff looks bulletproof
18:14:14preglowjhMikeS: that fixed nb mode too!
18:14:42 Join Frazz [0] (
18:14:59 Join jpt9 [0] (n=jpt9@
18:15:03jhMikeSsweet, bug fixed by deleting stuff...the best way
18:15:26 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
18:15:40preglowbut that should suffice, yeah? just deleting that one decoder_ctl line?
18:16:06jhMikeSmake sure lookahead is explicity set 0 and that ought to work
18:16:32preglow td.lookahead -= MIN(VOICE_FRAME_SIZE, td.lookahead);
18:16:37preglowdoesn't that line handle that?
18:16:51jhMikeSline 329, td.lookahead = 0
18:16:58preglowthat should have happened by the time any clip gets queued via get_more anyway
18:17:12jhMikeSnot nescessarily
18:17:55jhMikeSwell, true, it's paranoid but the cost is cheap to guarantee it
18:18:36jhMikeSmaybe not, it may not finish inserting frames into the DSP before it has to look at another clip
18:18:37*preglow is the bit counting type of person :>
18:18:54preglowbut sure, i'm putting it in anyway
18:19:20jhMikeSif the lookahead is ever > 1 frame, it won't have finished counting down
18:19:46preglowsure, might as well write that part in the paranoid spirit too
18:19:59preglowlookahead won't be > 1 frame unless something very big happens to speex
18:20:19jhMikeSjmspeex already warned it not to be assumed
18:20:27preglowsure, you should assume as little as possible
18:20:41preglowi should know, i had to write code to handle that exact case for mpa.c ...
18:21:45preglowamiconn: around?
18:22:05 Join linuxstb [0] (n=chatzill@rockbox/developer/linuxstb)
18:22:23*preglow needs a mingw person
18:24:00jhMikeSpreglow: this little state-machine paradigm is easy to follow? (I have to ask since of course I can follow it having written it :p)
18:24:15preglowjhMikeS: i almost always find state machines easy to follow
18:24:53jhMikeSit makes complex behavior simple to implement actually
18:26:16 Quit jpt9 ("Leaving")
18:26:45preglowlinuxstb: around?
18:27:34preglowlinuxstb: able to make a new rbspeexenc windows binary for me?
18:28:15preglowstrip it too
18:28:23preglowlinuxstb: just make sure to svn up
18:30:00 Join kclaf [0] (n=kclaf@
18:30:55preglowlinuxstb: stellar, thanks
18:31:28preglownew exe is smaller than the old, nice
18:31:37preglowdid you try to run it?
18:32:44preglowno biggie, it probably works
18:32:51linuxstbUnstripped was about 125KB
18:32:57preglowsounds right
18:32:59 Quit mf0102 ("Verlassend")
18:33:03preglowthe old was around that size
18:33:44jhMikeSpreglow: would not skipping lookahead only on the silence clip have worked too?
18:33:48 Join BigBambi [0] (n=alex@rockbox/staff/BigBambi)
18:34:21preglowjhMikeS: lookahead really should be skipped only after initing the encoder anyway
18:34:28preglowjhMikeS: and afaik, that's what i just made it do
18:34:58 Join Traveler5 [0] (
18:35:07Traveler5hey hey
18:35:09preglowjhMikeS: and how do i know which clip is the silence clip?
18:35:11jhMikeSyeah, just wondering if gaps could show up if the lookahead gets long or there's a shift?
18:35:52jhMikeSpreglow: I suppose I'd have a way for the voice thread to find that out and compare pointers.
18:36:29preglowjhMikeS: just do it like this till people complain?
18:36:40preglowafaik, this should work out fine
18:37:36Traveler5looked at the hello world example. if I make a game for rockbox I prettymuch have my variables initialized then do a while(exit == 0) loop?
18:38:01 Quit Febs (Read error: 110 (Connection timed out))
18:38:10jhMikeSTraveler5: whatever suits what you're implementing
18:39:08Traveler5well the hello world plugin terminates after running once, right?
18:39:14jhMikeSthe plugin stays active until the entrypoint returns somehow
18:39:30Traveler5so it will loop automatically?
18:39:32jhMikeSnever looked at it
18:41:15n1sTraveler5: no it doesn't, many games do what you want though so looking at them might be of more interest to you
18:42:08jhMikeSplugin is running until plugin_start returns. what conditions you decide to have it return on is up to your design.
18:42:26Traveler5alright then I'll stick with my original plan of a looped 'main method'
18:43:03 Join Rondom [0] (
18:43:04Traveler5well my layout would be setup, main loop, return probably
18:44:34preglowand will almost certainly work just great
18:44:46preglowcode it however you like, as long as you don't return from plugin_main, you'll be fine
18:44:58jhMikeSyou have to to stop it :)
18:45:16 Quit BigBambi (Remote closed the connection)
18:45:47*jhMikeS wonders if it's time to discuss TSR plugins now :)
18:46:01Traveler5well the program will be too goodto end :P
18:47:43jhMikeSTerminate and Stay Resident
18:48:04jhMikeSbattery_bench is one of those
18:51:59preglowjhMikeS: seems i have a new fun problem with dithering
18:52:50*jhMikeS gets too many associations upon hearing "dithering"
18:52:55preglowjhMikeS: a guy uploaded one of thoC[C[C[Cse mastered-to-pieces metal tracks, and with crossfeed enabled, it actually seems to exhaust the 4 bit headroom of the mp3 output format...
18:53:09pregloweh, that's supposed to be "those", there, god knows what happened
18:53:26preglowi have no idea how that is possible, and i shall have to investigate
18:53:41jhMikeSI thought the 4-bit headroom was slightly too small
18:53:48preglowi thought it should be plenty
18:54:24preglowbut of course, that is assuming none of it is in use after decoding
18:54:27preglowand that's not true at all
18:54:32jhMikeSLimiting decoder output to 24-bits would be just fine imho
18:54:37preglowmp3 can overshoot that to an amazing degree
18:56:35jhMikeSThe problem is just an ARM problem I take it?
18:57:31preglownot in the least
18:57:40preglowdithering doesn't use emac
18:57:46jhMikeSbut it does clamp
18:57:55jhMikeSor is it wrapping noise?
18:58:07preglowif the sample it gets is maxed out, dithering will overflow it
18:58:14preglowand with emac, it can be maxed out
18:58:23preglowbecause we have saturation for that enabled
18:58:46jhMikeSI did want to do an emac version of that
18:58:48preglowbut yes, another couple of bits of headroom would not hurt in the least
18:59:11preglowcodecs like wma definitely need to be fixed
18:59:15preglowi think it has one bit of headroom
18:59:16jhMikeSthen check if the sample is already as close to saturation as the amount you want to add to it first
18:59:31preglowand this will all break on arm, of course
18:59:37preglowi should try the track on arm, actually
18:59:37 Join midgey [0] (
19:00:06jhMikeSI think > 24 bits should be forbidden and fail with an irritating beep noise
19:00:06 Quit midgey (Read error: 104 (Connection reset by peer))
19:00:17 Join midgey [0] (
19:01:06jhMikeSor should be do 64-bit samples just in case? :p
19:01:06preglowi just tested
19:01:12preglowand crossfeed alone doesn't overflow
19:01:15preglowbut with dithering, it does
19:01:19preglowso there has to be something else wrong here
19:02:02jhMikeSdithering of course can increase amplitude later if amplitude fell short previously (can't express this right)
19:04:29 Join lee-qid [0] (
19:07:32 Nick ionvas is now known as ionvas__ (
19:08:49preglowno, you did fail quite severely there :)
19:09:28 Join J3TC- [0] (
19:11:04***Saving seen data "./dancer.seen"
19:11:49jhMikeSthe error delta will sum back later and if it matches the sign of the input, amplitude will go up
19:13:16jhMikeSdither random is a nice purple noise generator.../me thinks UV noise is better :p
19:13:40 Quit UncleRem1s ("leaving")
19:14:04 Nick idnar_ is now known as idnar (n=mithrand@unaffiliated/idnar)
19:14:37preglowi have no idea what kind of noise we use right now
19:14:52preglowi think it's rand - prevrand or something, which looks like tri-pdf, but isn't
19:15:03 Quit atsea- (Read error: 104 (Connection reset by peer))
19:15:03preglowand also looks like highpassed noise
19:16:35 Quit midgey ()
19:17:55jhMikeSdifferentiated white noise = purple noise
19:19:10 Join atsea- [0] (i=atsea-@gateway/tor/x-390a8b4c66aecd88)
19:19:37 Quit Traveler5 ("Java user signed off")
19:20:08jhMikeSso from nyquist down it should drop -6 db / octave
19:21:27 Join radinp [0] (
19:22:06 Quit kclaf (Remote closed the connection)
19:24:41 Join bertrik [0] (
19:24:58 Join woodensoul [0] (n=woodenso@
19:25:32woodensoulHey guys, can I get permission to edit the Wiki so I can upload WPSs?
19:25:47preglowjhMikeS: sounds like what we use
19:26:04radinpIWhat are WPS's?
19:26:34preglowjhMikeS: do you know what the rules are for determining at amplitude to add dithering noise?
19:26:39woodensoulradinp: New to Rockbox? While Playing Screen
19:26:50preglowjhMikeS: rect-pdf noise should be added at 0.5 lsb, tri-pdf at 1 lsb, but i don't know why
19:27:10bertrikit's even in wikipedia ??? :P
19:27:15linuxstbwoodensoul: What's your wiki name?
19:27:46 Join kclaf [0] (i=kclaf@
19:28:08radinpwoodensoul: yeah I'm a new a n00b, but I have started developing a regular expression library for rb.
19:28:10linuxstbwoodensoul: Done.
19:28:26woodensoulthanks linuxstb
19:28:38woodensoulexpression library?
19:30:30 Join Lear [0] (i=chatzill@rockbox/developer/lear)
19:30:57 Join Bagder_ [0] (
19:31:14radinpwoodensoul: A regular expression library would allow string matching using a predefined set of operators on expressions. E.g. (*) or (+).
19:31:35 Quit SirFunk (Remote closed the connection)
19:32:09radinpThis would pave way to the addition of a search function in the text_editor.
19:32:56linuxstbYou make it sound as if you could have a search function without regexps...
19:34:21*preglow wonders how people can stand using text_editor
19:34:22radinpSure, you could use the standard string library and use strcmp
19:34:23woodensoulsounds good radinp
19:35:28radinpI'm sure this regular expression library could be useful for other plugins but text_editor is the only one I can think of at the moment.
19:37:23 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
19:39:14rasherpreglow: It's not *that* bad once you get to know it. I wrote a (simple) WPS in it. Faster than endless usb connect/disconnect loops
19:39:14jhMikeSpreglow: rect-pdf? where's that?
19:41:04*jhMikeS is being dumb (thinking pdf doc) and has no little knowledge of actual terms for things.
19:41:48preglowjhMikeS: pdf = probability distribution function
19:42:10preglowjhMikeS: really just decides how likely the different sample values are to occur, rectangular pdf means all samples occur with the same probability
19:42:14 Quit freqmod__nx (Remote closed the connection)
19:43:29*jhMikeS would normally say "uniform distribution"
19:43:34 Join freqmod_nx [0] (
19:44:04jhMikeSI can say why...the area of a triangle is 1/2 the area of a rectangle
19:44:25preglowsounds logical, yes...
19:44:53preglowbut can you then explain at what amplitude highpass triangular dither should be applied? :P afaik, that's what we use
19:47:20 Join briantumor [0] (
19:47:30briantumordo they have tron for rockbox??
19:47:36briantumorthat would be nice
19:47:44briantumori'd never put down my sansa :\
19:48:28scorche|wall plugins that are deemed "done" by us are included with rockbox
19:48:54 Quit Bagder (Read error: 110 (Connection timed out))
19:48:58briantumorthere aren't any other repositories?
19:49:12krazykitno, except what's on the tracker.
19:49:21preglowthere are
19:49:23jhMikeSpreglow: I'm ashamed of the lack of specificity of my previous remark required to make that statement true :[p
19:49:23preglowbut not many
19:49:38krazykitthere are several GPL tron clones that you could probably port.
19:49:52 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
19:50:09 Join miepchen^schlaf [0] (
19:53:28*jhMikeS studies this dithery gadget
20:06:17radinpSo what is the advantage of using Git over Subversion for development? It seems like they both perform the same functionality.
20:06:34 Quit japc (Read error: 110 (Connection timed out))
20:06:48 Quit J3TC- (Read error: 110 (Connection timed out))
20:07:23 Join Llorean [0] (
20:07:52scorche|wradinp: that might be something to talk about in #git (or whereever they have their channel)
20:08:05maleI wouldn't do it.
20:08:16maleIt's a pretty stupid question.
20:08:35maleAnd they won't take kindly to it...
20:09:33maleThey short answer is the svn "performs the functionality" of CVS, and git is nothing like CVS.
20:09:53males/They/The; s/the/that/
20:10:06 Quit kclaf (Remote closed the connection)
20:10:20scorche|wmale: my point was that it is offtopic here..
20:10:34maleIs it? rockbox uses svn.
20:10:45maleI assume that's why he brought it up.
20:11:36radinpSorry about that question. I think it's an appropriate question. As it's integral to the development of rockbox.
20:12:00krazykitradinp, i'd search the IRC logs. i believe it's been discussed in this channel before.
20:12:02scorche|wradinp: ah, but you didnt phrase it as if it applied to rockbox ;)
20:12:36scorche|wmale: the question is offtopic...asking about how it would impact rockbox is on-topic
20:12:38jhMikeSpreglow: where does this distribution come in? shaping more strongly near the x axis? and less at the extreme amplitudes?
20:12:40rasheramiconn: would you mind testing a patch to I've just gone for the low-hanging fruit, so I'm not sure it makes any difference in speed
20:13:13scorche|wradinp: this may be of use:
20:13:15malescorche|w: Are discussions of what is and what isn't on-topic, on-topic/
20:14:07radinpscorche|w: Thanks, I knew I saw that page previously.
20:14:15maleUsing git to track an svn repo is better than having to use svn directly, but it's nothing like using the git model alone.
20:14:54scorche|wmale: is there something you are trying to get at with that question?
20:15:51markunscorche|w: that we should switch to git of course :)
20:16:20scorche|wmarkun: the question...not the statement ;)
20:16:21malescorche|w: You spam the channel more with this on-topic, off-topic bull**** than would just letting small asides slide.
20:16:51scorche|wmale: it prevents further spam
20:16:53markunouch ;)
20:17:18scorche|wit may not do it immediately, but it does
20:17:23tierraThis discussion was only going to devolve into an argument about which VCS is better anyway, which never gets anywhere.
20:18:01radinpGit looks very interesting, but it looks like it will be a pain to use in conjunction with subversion. Looks like Git makes a lot of sense if you're starting a project from scratch.
20:18:03markuntierra: why not? That's how we switched from cvs to subversion, which I don't think anyone regrets
20:18:40scorche|wthe current position is mainly: "yes we see that git offers many benefits, but is it really worth the effort to change?"
20:20:05Nico_Pyeah I'm not sure git would bring a benefit as main VCS for us
20:20:31scorche|wwel, it would..
20:20:44Nico_Pyeah, but lots of complications too
20:20:59maleIt isn't much effort to change to git at all.
20:21:01Nico_Pit is very nice for local dev though
20:21:04preglowjhMikeS: mno
20:21:14maleThe hard part is convincing everybody that the svn repo is shutdown ;-)
20:21:28Nico_Pmale: changing the repo isn't hard, but teaching everyone to use it is
20:21:31preglowjhMikeS: dithering function output doesn't have anything to do with signal amplitude
20:21:37n1smale: we have a buch of scripts that need adjusting etc...
20:21:52malen1s: True.
20:21:58radinpSo, with this distributed approach where would a backup copy of the rockbox source sit?
20:22:13scorche|wa backup copy?
20:22:16markunmale: well, we also have some scripts which need to be changed
20:22:19maleEveryone has a copy in git.
20:22:26Nico_Pmale: I use it and love it... but I think it would be overkill for us as a main repo
20:22:29*amiconn would hate switching to git
20:22:40markunamiconn: why?
20:22:52Nico_Pmarkun: would have the main repo
20:22:53scorche|wwell, first off, there isnt a windows client
20:22:55amiconnBecause of slowness
20:22:58markunNico_P: so?
20:23:00maleThat's bullshit.
20:23:01Nico_Pscorche|w: it's coming
20:23:11maleThey go to a lot of trouble to support windows.
20:23:12scorche|wNico_P: yes
20:23:13preglowgit is hardly slow
20:23:21male(god knows why)
20:23:23Nico_Pmarkun: sorry, that was for radinp (same nick color)
20:23:31amiconnAnd I really don't get why git should be superior to svn
20:23:34scorche|wmale: it sint bullshit...there is not currently a proper windows client
20:23:38amiconnsvn does all we need...
20:24:04malescorche|w: Proper meaning.... Completely monolithic and standalone?
20:24:11Nico_Pscorche|w: and there is no tortoisegit ;)
20:24:20preglowamiconn: it's easier to do several branches and stuff with git, but we don't do that much anyway
20:24:22scorche|wmale: see Nico_P for what i mean
20:24:28Nico_Pmale: you have to admit windows support is pretty poor ATM
20:24:50Nico_Ppreglow: maybe we would if it were easy ;)
20:25:05markunamiconn: it's easier for me to work on multiple areas of rockbox with git than with svn
20:25:06amiconnI don't need a native windows client, cygwin would be sufficient, but the cygwin git port is reported to be sloo-ow
20:25:06Nico_Pamiconn: git is faaast one linux
20:25:14preglowNico_P: why? i thought we decided against using several branches
20:25:25Nico_Ppreglow: did we?
20:25:31jhMikeSpreglow: just wonder atm what is distributed in a triangular fashion...the values for error?
20:25:36 Join MethoS- [0] (
20:25:37preglowNico_P: sure, that's why all the mrobe development is happening in the main repo right now
20:26:03preglowjhMikeS: the values for dither, the random output
20:26:12Nico_Ppreglow: IIRC we decided to use branches to maintain releases, and mrobe dev is in trunk because merging is a PITA in svn
20:26:12*amiconn didn't and doesn't use those tortoise* thingies
20:26:16scorche|wpreglow: partly because it is such a pain to merge
20:26:23*scorche|w is too slow
20:26:31 Join thgz [0] (
20:26:34preglowscorche|w: no, that wasn't part of the reason at all
20:26:49preglowjhMikeS: a rect-pdf dither will spit out values from -1 to +1 with equal probability, where tri-pdf will tend to spit out values around 0
20:27:01Nico_Ppreglow: it was the reason why the S devs merged their branch back
20:27:02jhMikeSrandom alone surely isn't
20:27:10tierraWas I not correct? This discussion simply devolved into an argument...
20:27:14markunpreglow: well, it *was* a pain to merge the gigabeat branch back
20:27:14radinpI'm starting to see the benefits of Git. It seems very similar to bittorrent approach.
20:27:25tierrano-one here is going to come out ahead
20:27:27amiconnNico_P: Yeah, but not everybody uses linux. Even though I have a linux box, and I know linux is faster at building rockbox, I'm still doing most of my development in cygwin
20:27:30tierrayou should all get back on topic
20:27:38preglowNico_P: but not the reason we decided to put all future ports in the main branch, that was because we feel new port development is better integrated that way
20:27:47markuntierra: which topic do you want to talk about?
20:27:59preglowNico_P: with new port devs having to take into consideration all the other ports from an early stage, simply because they live in the same tree and do the same daily builds
20:28:00Nico_Pamiconn: I'm not saying it's the main argument, but it feels real good and when the windows version comes to that level you might like it too ;)
20:28:00rasheramiconn: did you see my question about
20:28:04linuxstbmarkun: Looking back, would it have been a problem to have developed the F port in the main Rockbox CVS from the start?
20:28:31tierrawell, I don't want to talk about anything in here, but there are probably others here that are being hesitant about interupting your pointless discussion with a request for help with Rockbox
20:28:33 Quit evert ("leaving")
20:28:42markunlinuxstb: maybe not, but people were experimenting a lot in their own trees and didn't have commit rights
20:28:47Nico_Ppreglow: I agree ports are better suited in the trunk because they don't hurt anyone, but branches are meant for unstable dev
20:28:58bertrikWould it be acceptable to post 10 sec music clip on the bug tracker, to demonstrate a bug?
20:29:09preglowbertrik: sure
20:29:11amiconnNico_P: Well, so far I did work with svn (rockbox), but not yet with git. And since I see no benefit in git over svn, I won't start using it until forced (which I would hate)
20:29:19Lloreantierra: This is a rockbox development channel. Since we're discussing how choice of VCS affects the future of Rockbox Development, how is that off topic?
20:29:21tierrabesides, if there was to be any decision on the matter, it would half to be discussed on either your mailing lists or forums, right?
20:29:25markuntierra: it will quiet down in an hour :)
20:29:33tierraand you would all just be repeating yourselves
20:29:35preglowi don't care what source code management tools we do use at all
20:29:39Nico_Ppreglow, amiconn: I would've gone crazy doing MoB without git
20:29:40preglowsvn is ok, git will be ok, whatever
20:29:48Lloreantierra: No, actually the people who make the decision are most likely to be here. And absolutely least likely to ever see anything in the forums, which are more for support.
20:30:02Nico_Pit could have been something else but it was git
20:30:12thgzHi again
20:30:14radinpSorry, it's all my fault
20:30:17bertrikhi thgz
20:30:20radinpI started the Git topic.
20:30:21preglownow that the major annoyances of cvs is gone, i'll be just fine with svn
20:30:28preglowbut if someone decides we are to use git, i'm a-ok with that
20:30:31*bertrik agrees with preglow
20:30:31Nico_Ptierra: we tend to not use the ML enough
20:30:32tierraso anyone of the devs that this might matter to looses out if they just didn't happen to be here while it's being discussed? that's pretty dumb
20:30:34amiconnWell, what would be the benefit? We do want a central repo, hence no multiple official repos. And I really don't see why I would need an own repo
20:30:55preglowamiconn: a local repo would be fun, though
20:31:01Nico_Pamiconn: I didn't either until I had one
20:31:03scorche|wtierra: we keep logs, and many read them....this partly why i enforce many rules here
20:31:06amiconnI have a working copy, and that's sufficient. In *very* rare cases, I use a second working copy
20:31:07thgzWhenever I start SIM with english.voice included, I don't get any speech output in the UI
20:31:19linuxstbNico_P: Would Rockbox using git officially have made any difference to you? From what it sounds, SVN centrally and git locally seems to work well for people...
20:31:19thgzI'm buildin on Sansa E200 series
20:31:20Lloreantierra: First, the decision can't be made instantly. Second, there are logs. Third, if a dev isn't interested enough in development to at least keep his toe in what's going on, isn't that his own fault?
20:31:23maleamiconn: Because it makes it easier on you to develop features with other programmers without polluting the main repo.
20:31:39Nico_Plinuxstb: honestly, not much diff, no... I'm fine the way it is
20:31:51maleamiconn: It also makes it easier for you to track the development of others.
20:31:54amiconnpreglow: Why? Just more overhead to handle...
20:31:55tierrayou don't find the logs harder to filter through than a mailing list discussion?
20:32:03radinpEspecially developers like myself who are inexperienced with the rb source.
20:32:05Nico_Pamiconn: my local repo is probably smaller than *one* of your working copies and I have 4 branches
20:32:14 Quit woodensoul ()
20:32:44linuxstbNico_P: So how do you work with local branches? Surely you need local checkouts in order to compile?
20:32:49*amiconn doesn't care much about diskspace
20:33:04Lloreantierra: I didn't say they weren't easier to filter through. But with our number of devs, were git realistically likely to happen, first word would spread, then people would ask for reference to the time in the logs to see what the arguments for and against were, then a decision would be made.
20:33:10amiconnI rather use up some extra diskspace than handle extra overhead
20:33:11Nico_Pamiconn: me neither, but it's still a strong point IMO
20:33:11malelinuxstb: The checkouts are ephemeral.
20:33:23Nico_Plinuxstb: I have one local checkout
20:33:44linuxstbSo how does that work?
20:33:53amiconnWhy would I need more than the trunk? I really don't understand...
20:34:10Nico_Pamiconn: working on several things...
20:34:10maleAnd that's the key problem.
20:34:26Lloreantierra: Your assumption that any major development happens on the mailing list in relations to this project is mostly false. But a message saying "We're contemplating a change to git, and are currently in favour" would go out surely long before a change would happen.
20:34:27maleIf everybody did a little reading about git I think there would be a lot more enthusiasm here.
20:34:30tierraLlorean: they've still missed their chance to get their word in edge-wise with the people involved with the initial discussion
20:34:35bertrikI don't see any problem at all
20:34:35Lloreantierra: Which is no reason against discussing it realtime.
20:34:39Nico_Plinuxstb: you just checkout the branch you want (equivalent to svn switch)
20:34:43tierrathat's not really a "discussion"
20:34:48jhMikeSpreglow: I am missing part of the picture here though I do understand that the tri-pdf will of course spit out values with the highest p at the center (peak) of the window (well it's a triangular dist. afterall ;)
20:35:13Lloreantierra: Delineating the advantages and disadvantages of it in relation to the needs of Rockbox as a project with multiple inputs is not a discussion?
20:35:24amiconnNico_P: And why does working on several things need several branches?
20:35:30n1sMy opinion is that git is a solution in search of a problem...
20:35:32preglowjhMikeS: then what part of the picture is missing?
20:35:41amiconnI can do that fine within a single working copy..
20:35:45maleamiconn: Because it makes it easier on you.
20:35:50Nico_Pamiconn: it doesn't always, but it can become very nice if they are big changes
20:35:52tierraI can only see that process resulting in biased decisions being made based on who happens to be in the channel during the discussion
20:35:56maleamiconn: And git has no penalty for branching.
20:35:59Lloreantierra: Were you specifically aware that git would be disadvantageous because in relation to new ports it more readily allows those developing them to work outside of the automated build system, and ignore their impact on existing ports?
20:36:02Nico_Pamiconn: I'm curious as to how you manage your work
20:36:15Lloreantierra: Then frankly, you're an idiot for assuming the whole decision will happen here.
20:36:20amiconnmale: I'm not afraid of a penalty in the tool, but the penalty of a branch in itself
20:36:25Lloreantierra: If an email goes to the list, SOMEONE has to write it.
20:36:31*amiconn prefers KISS
20:36:37maleamiconn: Git is pure KISS.
20:36:47Lloreantierra: If the email is based on a single person's desire for GIT, is it more or less likely to be biased compared to a discussion where some people are against GIT, and some for?
20:37:03maleamiconn: It's just that the minimum level of simplicity is more complex than you're used to.
20:37:15maleamiconn: (for distributed SCM)
20:37:21Nico_PLlorean: I have to admit I kinda agree with tierra... we underuse the ML
20:37:25amiconnNico_P: I just work on source files, and commit those changes which belong to the part of the work I want to commit...
20:37:36tierrawell, sure, so they even get their chance to explain their point of view completely, then you see everyone's reply to that
20:37:54maleamiconn: I'm not aware of anyone who has switch from svn to git and decided to go back.
20:37:59tierraalso, generally speaking, posts to mailing lists are much more thought out responses
20:38:00Nico_Pamiconn: what do you do when you have large changes you can't commit progressively? and want to work on another small change?
20:38:00LloreanNico_P: My point isn't about the use of the ML. As I said, this discussion can be linked to in a post to the ML and you know as well as I do, the ML gets used before major changes to "how things work"
20:38:01scorche|wNico_P: i think so too...especially when you compare with other projects
20:38:24tierrayou aren't going to get these quick interjections with stupid comments like "git sucks!"
20:38:24amiconnNico_P: Then I just leave the big change uncommitted, and work on & commit the small one
20:38:34maleamiconn: That's bad and confusing!
20:38:35amiconnBut that's so rare it's not even worth discussing
20:38:35Lloreantierra: Obviously, you've never seen our user list.
20:38:45maleamiconn: With git you can commit every change without penalty.
20:38:47tierrayou don't have a development list?
20:38:52LloreanWe have a development list
20:38:59preglowrasher: nice
20:39:00amiconnPerhaps it is due to how I do stuff that requires big changes
20:39:03LloreanAre you making access is restricted to it?
20:39:04Nico_Pamiconn: what do you do if the small change is in one of the file you've already changed? it happens to me
20:39:05thgzThe problem may be in my Building environment, I try to use a prebuilt binary
20:39:11Lloreanmaking the assumption that, rather
20:39:19rasherpreglow: What?
20:39:35amiconnNico_P: It *never* happened to me that I had non-trivial changes from 2 features within one source file...
20:39:59 Join Arathis2 [0] (
20:40:01krazykitthgz, you're trying to use the sim with a prebuilt .voice?
20:40:02preglowrasher: buildzip patch
20:40:11jhMikeSpreglow: some background knowledge is missing to put all together
20:40:12amiconnThe only exceptions are the config-*<target>.h files, sometimes
20:40:12bertrikthgz: I also use a prebuilt gnuarm tool chain with no ill effect so far (I've tried something like 100 revisions)
20:40:44Nico_Pamiconn: well, you're lucky :) and how do you compile to test? I'm sure you manage all those things I'm asking you about, but git makes them so natural it's hard to imagine working without it
20:40:45amiconnThere it might happen that I have 2 features enabled for a target, and in the commit only one feature must be enabled
20:40:59amiconnNico_P: configure && make && make zip
20:40:59Lloreantierra: Many of the people who'd comment on the -devel list are just as likely to make irrelevant comments. Either because they're in favour of it because it makes unsupported builds easier, or they're against it for philosophical reasons, or whatever. Being a mailing list doesn't automatically make them more well thought out reasons than being IRC makes them less thought out.
20:41:14Nico_Pamiconn: of course... but when you have both changes in
20:41:19amiconn(the configure step just once every few weeks)
20:41:27amiconnThen I have both changes in....
20:41:49rasherpreglow: ah yes. I suspect the current patch won't change speed much though. It's the "find" runs that cost most, presumably.
20:42:02Lloreantierra: As well, do you know how many of the core, active, commit access having developers aren't involved in this discussion?
20:42:06amiconnAs already said, maybe it's due to the way I do stuff that requires big changes
20:42:11Nico_Pamiconn: I did manage perfectly before I started using git, but it does help me a lot
20:42:23jhMikeSaw fsck! :)
20:42:32Nico_Pit became a must for me when I had to do offline work
20:43:10amiconnI'm usually thinking it through as far as possible without changing a single line of code (except a few lines of test code here & there), and only if I am sure what to do, start writing the actual code
20:43:21 Nick snake_ is now known as Snake (
20:44:10 Join midgey [0] (
20:45:35thgznot working
20:45:35preglowrasher: and that's used for what?
20:46:07 Part thgz
20:46:40rasherpreglow: well, find is executed a bunch of times to copy rocks, codepages and codecs into the dir
20:46:51linuxstbNico_P: How do you manage your build directories?
20:47:14rasherpreglow: each time executing cp for each file. Causing the hilariously huge slowdown on cygwin
20:47:23Nico_Plinuxstb: I only have one (well, one for the sim, one for the target and a third for various things)
20:47:25linuxstbNico_P: I have separate checkouts for different things I'm working on, with their own sets of build directories.
20:47:33Nico_Plinuxstb: but I reuse the same ones each time
20:47:44linuxstbSo you need to do a full make each time?
20:47:54Nico_Plinuxstb: basically it's like I had only one svn co
20:48:14Nico_Plinuxstb: yes, but most of the time, not much files change and I do make -j 3 bin so it's quite fast
20:48:24linuxstbI mean when you switch checkouts...
20:48:31linuxstbOr rather, switch branches.
20:48:33*preglow checks out "make bin"
20:48:33Nico_Pyeah that's what I meant to
20:49:04 Quit ompaul (Read error: 113 (No route to host))
20:49:11preglowmake bin only makes rockbox.* ?
20:49:17Nico_Ppreglow: yes
20:49:22 Quit krazykit ("leaving")
20:49:43Nico_Plinuxstb: also I don't switch branches every 2 minutes
20:49:55linuxstbpreglow: "make help" ;)
20:50:05jhMikeSsuper simple
20:50:17sup: ]
20:50:20supthank you
20:50:39jhMikeSthe highpass is just serendipity
20:50:55 Join krazykit [0] (
20:51:10preglowdoesn't sansas do backlight fading?=
20:51:29bertrikpreglow: no, not that I know of, would like to have it
20:51:30jhMikeSnot enough levels that it would look remotely nice
20:51:42briantumori can't find a gpl 2d tron clone :\
20:52:03 Quit Arathis (Nick collision from services.)
20:52:05*amiconn has ~30 build dirs
20:52:08 Quit midgey ()
20:52:15amiconnThe voice build dirs not included
20:52:18 Join TMM [0] (
20:52:18bertrikif you do it fast, maybe you won't notice the limited fade levels
20:52:24preglowjhMikeS: levels? most backlights only have two levels but work fine with pwm
20:52:32 Join kclaf [0] (i=kclaf@
20:52:37jhMikeSit's done with i2s on the AS3514
20:52:44 Nick Arathis2 is now known as Arathis (
20:53:19preglowok, so you can't make it switch quickly enough?
20:53:20Nico_Pwhat I really loved with git was incremental history of my local MoB dev
20:53:22bertrikI know there is a patch for backlight fade, but IIRC it violates some rules regarding i2c use in a tick task
20:53:34briantumoroh yeah... you can adjust the brightness?
20:53:53preglowjhMikeS: does e200 retailos do fading?
20:54:13bertrikyes, there is a 16-step controllable current source, which should result in a linear backlight brightness range
20:54:16jhMikeSbertrick: yeah, using any blocking functions (like mutx_wait) from an ISR is a strict nono
20:54:38jhMikeSpreglow: no
20:55:02preglowok, just wondering since i see that wheel light dimming patch/code uses the user timer
20:55:11amiconn16 steps would be quite ugly for fading
20:55:14jhMikeSthat's just an on/off through GPIO
20:55:20preglowbut i guess that's a-ok if it's not in use by pwm fading
20:55:54amiconniPod Video has 32 brightness levels in hardware. The OF does use it for fading, and the steps are visible
20:56:33bertrikvisible ok, but annoyingly so?
20:56:44amiconnHence rockbox uses software pwm fading *in addition* to using hardware brightness for brightness
20:57:17 Join woodensoul [0] (n=woodenso@
20:57:18preglowamiconn: btw, did you figure out if pwm is used for 5g/nano backlight intensity?
20:57:35woodensoulHow can I delete files I've uploaded to the Wiki?
20:57:42amiconnThis is possible because there's a gpio pin for switching the light, independently of the hardware brightness circuit
20:57:59bertrikoh, I was wrong about 16 levels, it's 32 level (including off)
20:58:18amiconnpreglow: I would think so. I googled a bit and found that pulse controlled backlight pwm chips are quite common
20:58:30preglowamiconn: has to use a really high freq, then
20:58:42amiconnThe hardware pwm is most probably in the kHz range
20:59:00preglowbtw, any new bootloaders for 5g/nano?
20:59:37preglowrockbox blinks and glitches very nicely when booting now, thanks to backlight intensity changes
21:01:11Nico_Pwoodensoul: you can't
21:01:11 Join MethoS-- [0] (
21:01:26Nico_Pwoodensoul: you can just hide them, only admins can delete files
21:01:27amiconnIt goes min brightness (apple loader) -> max brightness (rockbox bootloader) -> min brightness (default in main rb) -> configured brightness
21:01:53preglowapple loader doesn't seem to use the backlight at all for me
21:01:58amiconnImo the rockbox bootloader should also use min brightness
21:02:03amiconnOn Video it does
21:02:33linuxstbIt pretty much has to, as it takes about 10 minutes to boot...
21:03:10preglowamiconn: well, we should match it to what apple uses
21:03:22pregloweverything just to avoid the glitching, it looks really bad
21:03:29woodensoulwell is there an admin around so I can request they be deleted?
21:03:35linuxstbIs there anything else we want to change the ipod bootloaders?
21:03:39bertrikmaybe do not nothing at all to the brightness
21:04:11bertrik*delete not
21:04:14preglowamiconn: there is no backlight at all here when the apple logo is displayed
21:04:30amiconnJust recompiling the bootloader should fix that behaviour, but I didn't try that yet
21:05:15amiconnpreglow: You're right. However, I would prefer a little bit of light in our bootloader
21:05:21jhMikeSshouldn't the brightness be at the default when booting then just be switch to the user setting after the logo? other stuff does that.
21:05:35amiconnjhMikeS: It does
21:06:01 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
21:06:15amiconnBut the default is now lowest, and the latest bootloader official bootloader does not yet have brightness control, so it sets full brightness
21:06:24 Join stripwax [0] (
21:06:26Lloreanpreglow: The backlight in the apple bootloader being off, to the Rockbox one being on is actually helpful I think
21:06:30amiconn(which is blindingly bright in dark environments)
21:06:31 Quit stripwax (Client Quit)
21:06:53LloreanRight now, once the Apple logo is lit, you know you're in Rockbox code, and turning hold on at this point is "too late" and will clear settings instead of going to the Apple FW
21:06:58amiconnLlorean: Yeah, hence I'd just recompile the Video + Nano bootloaders
21:07:12preglowLlorean: too bad it's so damned ugly, then
21:07:19amiconnThat should give the default, hence no glitching when switching over to main rockbox
21:07:25 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
21:07:25preglowthe light glitching around looks really, really unprofessional
21:07:49LloreanI'm not saying something doesn't need to be done. I'm just saying, I'd prefer if whatever it is, it's still clear when apple hands off to our code. ;)
21:09:14jhMikeSpreglow: the sansa has a nice white display for a bit...then black, then rb logo. :)
21:09:56jhMikeSnice = not so nice here :)
21:10:34LloreanSpeaking of logos, why not have the Rockbox bootloader display a logo?
21:10:46rasherWe need a pretty logo first..
21:10:52*rasher ducks and runs
21:10:53bertrikbarrywardell made a patch for that I think
21:11:02amiconn"Yes is no and no is yes. Delete all files? (n):"
21:11:05***Saving seen data "./dancer.seen"
21:11:07LloreanSpecifically, the mono-version of the Rockbox logo, in the same location that the color version now displays, so it's still small, but looks nice?
21:11:49amiconnLlorean: Mainly we need a *small* logo, as on some targets, the bootloader must not exceed a size limit which is quite low
21:11:52LloreanAnd a version number, so when asking people their bootloader version, we can get a useful answer.
21:12:02bertrik FS #7977, I haven't tested it myself
21:12:08Lloreanamiconn: Would a mono fit on those targets?
21:12:12amiconnMy idea was using a compressed bitmap
21:12:55amiconnLlorean: Would probably look strange...
21:13:54 Quit The-Compiler ("Verlassend")
21:14:38jhMikeSforget bitmaps, use YUV420 then ucl compress that. It starts out 3/4 the size then uncompressed. :)
21:14:53 Join hannesd_ [0] (
21:16:30bertrikwith colored bars on the background while checking the application's CRC
21:17:51amiconnsvn bootloader crashes...
21:17:57preglowjhMikeS: oh, wait, you can dct it, quantize the coefs, then run-length encode that too!!
21:19:09preglowreally, what encoding is the filename in fopen() supposed to be encoded like? the current locale?
21:19:45 Quit MethoS- (Read error: 101 (Network is unreachable))
21:20:18amiconnWhy does that bootloader just crash?
21:20:20jhMikeSpreglow: think you're on to something there! :)
21:22:42n1spixelma: I checked out using different targets in a translation vs english, turns out english decides what goes in (as I thought) possibly could be a problem for voice file generation though...
21:22:52*jhMikeS next step: how does a series of differentiation of rect-pdf give you tri-pdf? probably a ridiculously simple answer too...eck studying. :P
21:23:48pixelmajhMikeS: "nice white display"? I only see the sandisk logo + "lil monsta" thing and it's also flickering (backlight turning on and off a few times) =)
21:24:19linuxstbamiconn: Have you tried other targets, or just the Video?
21:24:26amiconnJust Video
21:25:24preglowjhMikeS: our dither uses random - prevrandom, that looks more like highpassed rect-pdf to me, but is actually called hp tri
21:25:27pixelman1s: thanks for checking. That's what I had imagined too...
21:25:40jhMikeSpixelma: just adds exitement to the boot process :)
21:26:12pixelmausually I'm not looking anyways ;)
21:26:57bertrikis booting the OF supposed to have this weird white blooming effect?
21:27:13krazykitbertrik, it seems to happen with some people
21:27:56bertrikI see it everytime I boot the OF and it looks a bit unhealthy
21:28:40jhMikeSbasically LCD refresh has been stopped for a little while
21:28:50amiconnHmm, r15598 bootloader (last pre-brightness) works..
21:31:00 Quit hannesd (Read error: 110 (Connection timed out))
21:31:02 Nick hannesd_ is now known as hannesd (
21:31:07 Join FOAD_ [0] (
21:31:19amiconnEek, and r15600 (first proper past-brightness) crashes...
21:31:48 Join jhulst_ [0] (n=jhulst@unaffiliated/jhulst)
21:33:10 Join stripwax [0] (
21:33:43 Join MethoS- [0] (
21:34:19 Quit radinp (Read error: 110 (Connection timed out))
21:36:59linuxstbamiconn: If it helps, SVN bootloader works fine on my ipod Color. On my Video, I get a hard crash (disk powers down with a click) and the apple logo turns into a white square...
21:37:17amiconnYes, with some pixel gibberish in it
21:37:34linuxstbClever effect...
21:39:16 Quit amiconn (Nick collision from services.)
21:39:24 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
21:41:02BilleniummIRC on linux ftw?
21:41:07 Quit Billenium (Remote closed the connection)
21:41:27roolkuNico_P: any ideas regarding FS #8201? *nag* *nag* :)
21:41:48preglowlinuxstb: oh, btw, what status does the hw eq removal patch have?
21:41:59Nico_Proolku: sorry, no
21:42:14linuxstbpreglow: Last I saw was that there was a bug that needed squashing...
21:42:57linuxstbamiconn: Are you enabling interrupts in the backlight function?
21:43:23amiconnI am disabling them, and restore the old value again
21:43:34 Quit jhulst_ ("Konversation terminated!")
21:43:41amiconnSo if they already are disabled, they stay disabled
21:45:21 Quit jhulst (Connection timed out)
21:45:46 Join Casainho [0] (
21:45:52Casainhohello guys :-)
21:46:25CasainhoI am looking for help about buttonlight wheel on Sansa...
21:46:48CasainhoI need to know wich part of the code tunr of the light off after time out time...
21:47:42 Quit FOAD (Read error: 110 (Connection timed out))
21:47:43 Nick FOAD_ is now known as FOAD (
21:48:14 Join billenium [0] (
21:48:58jhMikeSwell isn't that the sweetest little thing.../me made a triangular distrution generator with a graph
21:49:18 Join barrywardell [0] (n=barrywar@
21:50:37 Nick ionvas__ is now known as ionvas (
21:51:02 Quit MethoS-- (Read error: 110 (Connection timed out))
21:51:51bertrikCasainho: check out backlight.c
21:51:53 Quit barrywardell (Client Quit)
21:52:17 Quit woodensoul ()
21:52:43Casainhobertrik: this one? : static void buttonlight_update_state(void)
21:52:45bertrik_buttonlight_on/off is the hardware function that actually turns it on or off
21:53:06bertrikfor sansa it's implemented in backlight-c200_e200.c
21:53:35bertrikCasainho: yes, as far as I know
21:54:49Nico_Proolku: have you seen a relation between the place of the tracks in the buffer and the fact that their info isn't saved?
21:54:55Casainhobertrik: I made this code: comment 3:
21:54:57Nico_PI mean frist/last
21:55:24 Quit weezerle (Read error: 104 (Connection reset by peer))
21:55:42 Join J3TC- [0] (
21:55:58 Join Fraser [0] (
21:56:33 Join barrywardell [0] (
21:56:46Casainhoand in static void buttonlight_update_state(void), I put "wheel_light_quantity = 0;" instead of the original "_buttonlight_off();"...
21:57:02CasainhoIt works partiality...
21:57:06preglowjhMikeS: ordinary tri generator that graphs output occurences?
21:57:14bertrikHow fast do you update it?
21:57:31Casainhoat 100 micro secounds...
21:57:55bertrikI think that's even faster than the fiq handler for audio ....
21:58:21 Quit scorche (Nick collision from services.)
21:58:23CasainhoĨt's working nice, but It don't turn off at time off, but turns off If I select OFF in menu settings...
21:58:47Casainhoehehe - I don't know nothing about fiq handler for audio... :-)
21:58:50 Join scorche [0] (n=scorche@rockbox/administrator/scorche)
21:58:59roolkuNico_P: it's the last one in the loop of call's to the unbuffer call backs
21:59:49jhMikeSpreglow: yeah
22:00:01Casainhobertrik: I want to know from where comes the time off, because time off is not offing the wheel light... :-(
22:00:02bertrikas far as I know, the audio DMA interrupt runs at less than 1 kHz and it's asm-optimised for performance
22:00:08preglowjhMikeS: lemme see
22:00:10Nico_Proolku: ok
22:00:18Nico_PI'll try to look when I finish supper
22:00:22roolkuNico_P: so assuming 5 files get buffered, the loop is executed 4 times when the rest of file 5 is buffered and file 5 will not have its elapsed time set
22:00:23preglowjhMikeS: also, could you graph the output of what we use? rand - prevrand
22:00:32bertrikso you're probably wasting some battery life just for the wheel fade
22:00:33jhMikeSthat's what it's graphing
22:00:46roolkuNico_P: ... and file 4 will not have its elapsed time set
22:00:53preglowjhMikeS: i'd love to see the graph
22:00:55CasainhoI can lower the frequency :-)
22:01:06jhMikeSI made it do fancy things...hold on
22:02:38bertrikthere are better ways than PWM to do duty-cycle modulation that may help to lower the frequency and still keep it from flickering
22:03:19amiconnlinuxstb: I think I know what's going on, abot to test...
22:03:28jhMikeSjust a stupid excel thing - the script:
22:03:35 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
22:03:49 Nick Bagder_ is now known as Bagder (
22:03:51bertrikI disabled the wheel light by the way, it's too damn bright
22:03:52preglowjhMikeS: excel, oh, that microsoft windows thing
22:03:59*preglow strokes his linux
22:04:13amiconnbertrik: The software pwm for backlight fading on several ipods and iriver h1x0 uruns at 200Hz
22:04:18amiconn(i.e. 400 ints/sec)
22:04:23amiconnCompletely flicker free
22:05:04jhMikeSpreglow: graph coming
22:05:06roolkuNico_P: i have DEBUGFed prev_track_elapsed = curtrack_id3.elapsed;
22:05:13bertrikok, and how many steps can it make?
22:05:29amiconnlinuxstb: The problem was that _backlight_hw_enable uses sleep()
22:05:42linuxstbI thought sleep was faked?
22:05:51amiconnIt seems like it's not
22:05:54roolkuNico_P: and they both seem to have the right time (well a time that is not 0), but not the recored that is pointed to by track_ridx
22:05:55Nico_Proolku: what's probably happening is that curtrack_id3 for the last song isn't written back to the main buffer
22:06:05linuxstbBut doesn't the ATA driver need sleep()?
22:06:10amiconnI replaced it with udelays for bootloaders, and that fixes the bootloader
22:06:37Casainhobertrik: with 100us, I can have 1%, 100 steps -> 100 * 100us = 10ms">
22:07:16roolkuNico_P: yes, I thinks so too
22:07:33preglowjhMikeS: that looks an awful lot like a triangle, yes
22:07:40Casainhobertrik: what are others ways for doing duty-cycle and lower frequency - to save battery?
22:07:47bertrikIf 200 Hz works for the backlight, then I think it should also work for the wheel light, unless the backlight has some hardware filter that the wheel has not
22:08:02preglowjhMikeS: so there are two ways to make a triangle pdf, apparently
22:08:10pixelmaroolku, Nico_P: I noticed a weird effect if a playlist ends and the player is turned off through idle power off, then when I start again and try to resume it plays the last seconds of the last track again, that happens only once. All next tries show the "nothing to resume" splash correctly. It only seems to happen if the playlist is longer so it has to rebuffer at some point. Could it be related, or not?
22:08:13preglowjhMikeS: and this one has less noise around dc than the other one
22:08:36CasainhoI did chose an hight freq, 4 times the 25hz for human eye don't perceive
22:08:42 Quit billenium (Remote closed the connection)
22:08:45preglowjhMikeS: does this reveal to you at which amplitude it should be mixed with the audio, though?
22:09:40bertrikCasainho: you can use a kind of accumulator that you add the desired light level to. The overflow bit is then the bit to enable/disable the wheel light
22:10:01 Quit ender` (Read error: 104 (Connection reset by peer))
22:10:19bertrikfor example, if you add 128 to the accumulator, it will overflow once every two cycles, making a 50% duty cycle
22:10:30bertrik(overflow to the 8-th bit I mean)
22:10:34amiconnCasainho: Setting the timer to 100us permanently is a waste
22:10:37jhMikeSwell, tri will produce the same max amplitude as rect but twice the width (to get the same area)
22:10:37 Quit Frazz (Read error: 110 (Connection timed out))
22:10:48amiconnYOu could just do what the backlight software pwm does
22:11:24jhMikeSwait, actually tri will produce twice the amplitude and twice the dist width
22:11:30Casainhowell, I don't know how backlight pwm work, I didin't look for that...
22:11:46 Quit Fraser ("Leaving")
22:12:13amiconnIt programs the timer for each phase, so even though the backlight pwm uses half the period (5ms repeat cycle and 1% steps), it just fires 400 interrupts/sec, not 10,000
22:12:13bertrikthe way of duty-cycle modulation I'm proposing always results in the highest possible switching frequency
22:12:19Casainhobertrik: thanks for that infor about accumulator
22:12:35amiconnThat said, a permanent software pwm will most likely be rejected for svn inclusion
22:12:58jhMikeSI guess it needs appropriate scaling...haven't looked at the full pipeline of stuff yet
22:13:07bertrikamiconn: ah ok, you use a hardware timer to trigger the interrupts
22:13:22amiconnbertrik: Yes, sure. There is no other way
22:13:30CasainhoI understand - I just wanted to learn and have wheel light with less bright...
22:13:36 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
22:13:48 Quit advcomp2019 (Nick collision from services.)
22:13:54 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
22:14:20linuxstbamiconn: Seems there is a special case for the bootloader in sleep(), so it seems that's broken?
22:14:46Casainhobertrik: don't you want to implement your way for this wheel light PWM?
22:14:51amiconnlinuxstb: I'm not sure. The ata driver does use sleep() - and it doesn't break...
22:14:57roolkupixelma: I wouldn't say related, but I think I have seen that behaviour as well actually
22:15:26 Join stripwax_ [0] (
22:15:26amiconnbertrik: I can't see what other method than pwm would deliver higher frequency at a given average interrupt rate
22:15:56linuxstbamiconn: But doesn't it only use sleep() if something goes wrong? IIRC the ipod bootloaders worked for a long time with a broken sleep(), and we only noticed when testing on the 5.5g?
22:16:07roolkupixelma: the missing elapsed time happens for every rebuffer, not just the playlist end
22:16:31 Quit scorche (Nick collision from services.)
22:16:45bertrikCasainho: no, the way I explained works differently than the way the backlight is PWM'd
22:17:00 Join scorche [0] (n=scorche@rockbox/administrator/scorche)
22:17:19amiconnlinuxstb: Even ata_power_on() uses sleep()...
22:17:40Casainhobertrik: how is backlight PWM'd? - at waht frequency?
22:17:43bertrikI was under the impression that the interrupt was increasing a counter, with something like 255 interrupts for one PWM cycle
22:17:59jhMikeSthe output from the history really looks more gaussian...from d it's triangular
22:17:59amiconnbertrik: Not the backlight pwm - see backlight.c
22:18:05bertrikbut the backlight uses a different method that can get 1 cycle for two interrupts
22:18:14amiconnIt reprograms the timer for the lowest possible interrupt rate
22:18:45amiconnBut even the we do not want that permanently enabled
22:18:55preglowjhMikeS: what history?
22:19:18Nico_Ppixelma: I don't think it's related, but when you say "last seconds", do you really mean it or is it shorter? I've seen something similar but the played part was very short
22:19:45jhMikeSpreglow: error history
22:19:56 Join ender` [0] (
22:20:10preglowjhMikeS: oh, right, that might easily be gaussian, yes
22:20:11jhMikeSI'm just looking at each stage in itself atm
22:20:16Casainhoamiconn: not permanently enabled, why?
22:20:22amiconnCPU load
22:20:31preglowjhMikeS: like lpc estimation errors are usually exponentially distributed
22:20:34 Quit Arathis ("Bye, bye")
22:21:10Casainhoso, making PWM, for backlight or others lights, will not be desired? - because of CPU load?
22:21:22amiconnYes, not permanently
22:21:29amiconnFor fade in/fade out it's okay
22:21:54amiconnAnd btw you cannot easily use it for multiple ligts
22:22:00Casainhoah, I understand, because fade in/out is just sometimes and not always genertaing PWM...
22:22:10preglowamiconn: e200 doesn't have backgliht fading, so it's ok
22:22:39Casainhowhy can't for multiple lights??
22:22:57bertrikI think I would already be happy with backlight fades using hardware brightness control
22:23:45amiconnlinuxstb: ehhhh... I think I found the real problem now...
22:23:54Casainhois there hardware for brightness control on Sansa e200??
22:24:01amiconnsleep() *is* faked - but it calls switch_thread()
22:24:11linuxstbThat's what I was just looking at...
22:24:19amiconnAnd the backlight init is called before kernel_init()....
22:25:26linuxstbIs there any reason to call switch_thread in the bootloader? Do any PP bootloaders create threads?
22:25:38amiconnE.g. the ata thread
22:26:02pixelmaNico_P: haven't measured, it's not a very long time.
22:26:08amiconnI'll move backlight_init() past kernel_init() and test
22:26:10bertrikwow, backlight level can go up to 31, but the current maximum 12 is already quite bright
22:26:40Nico_Ppixelma: in my case it was less than a second I think
22:27:02amiconnlinuxstb: Working nicely...
22:27:04Nico_Ppixelma: I filed for it
22:27:51Casainhookok - bye bye :-) have good night :-)
22:27:56 Quit Casainho ("ChatZilla [Firefox]")
22:29:28pixelmahave to try again, I could reproduce it on both c200 and M5 but not under all circumstances. My impression was that it has to do with playlist longer than buffer or not - not a 100% sure if that's the thing in common.
22:29:32 Quit MethoS- ("Konversation terminated!")
22:29:57pixelmaNico_P: ok, maybe I'll comment on it if I got some more details
22:30:21amiconnpreglow: You could 'svn up' now and build a Nano bootloader. On Video, the commit fixes the backlight glitching around
22:30:26pixelmaBigBambi: wise pixels? ;D
22:31:03amiconnJust off (apple loader) -> minimum brightness (rockbox bootloader, rockbox start) -> configured brightness
22:31:58jhMikeSpreglow: you want the amplitude at which the triangular noise should be mixed (the rand output)?
22:33:46 Quit stripwax (Read error: 110 (Connection timed out))
22:34:09BigBambipixelma: Gah
22:34:27LearNico_P: Looks similar to what happens when you skip (back?); first you get the fragment from the file you skipped from, before the new track starts playing. PCM buffers not flushed properly?
22:34:42Nico_PLear: yeah, maybe
22:34:45BigBambipixelma: I'm stil having problems with switching between QWERTY and AZERTY keyboards each day
22:35:39jhMikeSpreglow: looks like it should be twice what it is
22:35:50preglowjhMikeS: and what makes you say that?
22:36:20Nico_PLear: the pcm buffer is shown as empty in the debug screen though
22:36:43 Join BlackFrog [0] (
22:37:26jhMikeSpreglow: not entirely sure yet
22:37:50jhMikeSright now it's only at the fractional level, right?
22:38:27BlackFrogThe web server at is sending the mime type for .dmg files to "application/x-apple-diskimage" which someone just made up (its not registered with the iana), it shoudl be "application/octet-stream"
22:38:36LearNico_P: Well, I'm not familiar with the low-level audio stuff... :) Could be in the driver level too.
22:38:41 Quit J3TC- (".UPP.")
22:38:47Nico_PLear: me neither ;)
22:39:03 Quit BlackFrog (Client Quit)
22:39:19bertrikNico_P: I saw some large negative numbers in the buffering thread screen yesterday
22:39:21*jhMikeS back after idea inducing smoke
22:39:33Nico_Pbertrik: oooh, not good
22:39:44jhMikeSnot the fancy kind of smoke though :p
22:40:00preglowi thought you said you wanted ideas!
22:41:31 Quit Lear ("ChatZilla 0.9.79 [Firefox 3.0b1/2007110904]")
22:41:33briantumorno tron for rockbox :(
22:41:44briantumorand i can't find a 2d tron that doesn't use x :\
22:41:57preglowwell, you'd have to port it, yes
22:42:12briantumorwell.. i mean.. i just know ktron
22:42:19briantumorthat uses kde.. how would i port that?
22:42:21stripwax_briantumor it would not be difficult to write one from scratch
22:42:25*roolku gasps - well that was unexpected :(
22:42:33 Join BigBambi_ [0] (n=Alex@rockbox/staff/BigBambi)
22:42:41pixelmaNico_P: were you referring to FS #8160 (with Lear)?
22:42:55 Join J3TC- [0] (
22:43:16briantumorso there's a rockbox dev package?
22:43:17Nico_Ppixelma: no, FS #8102
22:43:40 Join Zagor [0] (
22:44:43*amiconn points roolku to apps/recorder/recording.c, line 593ff
22:44:45pixelmaah, ok. But I saw the behaviour described in 8160 too, yesterday - and my impression was the same as hcs's
22:45:00bertrikNico_P: I can reproduce it now
22:45:18 Quit linuxstb (Read error: 104 (Connection reset by peer))
22:45:59bertrikAs far as I can see it happens when skipping backwards with cross-fade enabled, I'll investigate further tomorrow
22:47:00 Quit Rondom (Read error: 110 (Connection timed out))
22:47:34bertrikOh it also happens without cross-fade
22:47:45Nico_Pskipping backwards can be bad
22:47:52Nico_P(FS #8092)
22:47:53*Bagder is not that negative about a crypto-plugin
22:47:59*preglow neither
22:48:04preglowi think it's a pretty good idea
22:48:13Bagderfine for keeping PINs etc safe
22:48:13Nico_Proolku: nice commit :) I'm touched that you used my version too :p
22:48:30*Nico_P agrees... I don't get why the idea is bashed
22:48:45 Quit Siku ()
22:49:08LloreanI'm just curious what the intent for it is, with multiple strong encryptions for various file types, etc.
22:49:14preglowbecause some people obviously feel the fact that these devices are daps means we should only play music, but i think that's kind of silly
22:49:42BagderI could use it to store PINs for cards/entrances etc that I have in my life
22:49:44Nico_Proolku: it looks to me like the callback is simply skipped for some tracks
22:51:03 Join linuxstb [0] (n=chatzill@rockbox/developer/linuxstb)
22:51:49roolkuNico_P: no, the callback is executed, but the check for elapsed==0 returns early and never stores the data
22:52:07 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
22:52:24Nico_Proolku: I set a breakpoint on tagtree_unbuffer_event and it looked like some tracks were never touched by it
22:52:34 Quit linuxstb (Remote closed the connection)
22:52:47 Join billenium [0] (
22:52:50Nico_Proolku: which does seem logical to me, looking at audio_clear_track_entries
22:53:14 Join linuxstb [0] (n=chatzill@rockbox/developer/linuxstb)
22:53:16amiconnBagder: I think there are 2 problems with a password safe on a dap: (1) the cumbersome password input. The "master" password should be stronger than any password protected with it. (2) the passwords might be protected from revealing, but not from losing access (when the dap gets broken/stolen etc)
22:53:21Nico_Pbut in some cases the elapsed value will still be 0 even now that I improved that
22:53:41preglowamiconn: none of those are showstoppers, if you ask me
22:53:43amiconnThat said, it's just a plugin, and there are other plugins which are quite cumbersome (e.g. the text editor)
22:53:43LloreanBagder: I have a code wallet on my phone that I like, but it's the intent for encrypting various files and stuff that I don't really see.
22:53:46Bagderamiconn: sure, but they're no showstoppers
22:53:51preglowof course i'd never keep them only on the dap
22:53:55preglowas i would never only keep them on a pc
22:54:08preglowand i can think of plenty of nice input methods
22:54:17roolkuNico_P: maybe that is an additional problem
22:54:18preglowyou don't have to use a virtual keyboard
22:54:24preglowgestures would be fine
22:54:27Bagderbesides, nobody is forcing anyone to use it
22:54:27LloreanSurely even up, up, down, down, left, right, left, right as a password to getting to your various door codes and such is more secure than a sticky note in your wallet, for someone with very poor memory. :)
22:54:42bertrikuse the wheel as a virtual safe dial perhaps, and select to "open" the safe
22:54:46amiconnpreglow: You need a long sequence of gestures then...
22:54:55Nico_Proolku: what were you gasping about btw... the red delta?
22:55:01preglowamiconn: you think someone is going to sit with the dap and try to brute force your combination?
22:55:08preglowamiconn: their thumbs would fall off
22:55:23Bagderthey'd get the file of to a PC and _then_ brute-force
22:55:30amiconnBut the encrypted data is a file, which can be copied, and brute-forced on a pc
22:55:30roolkuNico_P: in my experiment track 6 would call the callback and be rejected at line 683 when track 7 which was only partially buffered would spin up the disk to fully buffer
22:55:35preglownot with an explosive device in the ata connector :D
22:55:52preglowthat'd be default
22:56:05roolkuNico_P: yes. the red delta :( must be the overhead from the plugin interface
22:56:11preglowbut yes, good point
22:56:14preglowit's way better than nothing
22:56:16 Quit BigBambi (Read error: 113 (No route to host))
22:56:24amiconnroolku: The plugin api is just +8 bytes (2 function pointers)
22:57:01Bagderand most people that would steal a DAP wouldn't be able to crack an encrypted file, even if it was done with only 5 gestures...
22:57:09amiconnBut check that other place which should be an ideal candidate for dir_exists()
22:57:11 Quit bertrik ("bye")
22:57:14roolkuamiconn: nod
22:57:19preglowbut i most certainly don't get the bashing of the idea, the intent and use of such a plugin is clear
22:57:29Nico_Proolku: seen the check_dir func in recording.c amiconn pointed you to?
22:57:49roolkuNico_P: sorry, must have missed that...
22:57:56 Nick parafin is now known as parafin|away (
22:58:05Nico_Pdon't apologize ;)
22:58:28amiconnpreglow: I won't bash it. It has its drawbacks, but why not...
22:58:50preglowsure it has its drawbacks, but its way better than nothing
22:58:57preglowit'd be my only way of carrying around encrypted information, really
22:58:57amiconnA plugin is just a plugin. No harm done if one doesn't want to use it
22:59:20preglowi'm sure someone will code a retina scanner plugin soon anyway :)
22:59:34preglowor perhaps voice identification
22:59:35preglowthat'd rock
22:59:37amiconnMight even be possible on the m:robe...
22:59:47LloreanVoice ID, plus USB-OTG for a thumbprint verifier?
23:00:23amiconnAnd preglow's explosive device...
23:02:03 Nick BigBambi_ is now known as BigBambi (n=Alex@rockbox/staff/BigBambi)
23:02:06linuxstbDidn't someone try and write a some kind of crypto plugin in the past?
23:03:04amiconnI think even a strong algorithm wouldn't be too slow on our targets if it's used for reasonable amounts of data
23:03:10briantumorso where should i start if i want to port games for rockbox?
23:03:22Nico_Proolku: the red delta is probably caused by the fact of having the functions plublic (and a plugin doesn't affect binsize)... but IMHO it's still a very positive change
23:03:37amiconnPGP is usable on Amiga
23:04:14*linuxstb searches the IRC logs and finds bger talking about it about 2 years ago...
23:07:31 Quit ramon8 (Read error: 110 (Connection timed out))
23:07:59PaulJambriantumor: can you program in C?
23:09:28briantumorwell.. i've worked with c++ mostly.. but i know c
23:10:52Nico_Proolku: I think I have a proper fix
23:11:02PaulJami would suggest to start by setting up a dev environment for rockbox, getting the sourcecode and looking at how the other games and plugins work.
23:11:07***Saving seen data "./dancer.seen"
23:13:19linuxstbAnything Unix-like - Linux, Mac OS X, Cygwin...
23:13:28briantumorok :)
23:14:12*barrywardell points out that we even have encryption already in the pp-mi4 bootloaders - it's used to boot the H10 OF
23:15:15Lloreanbarrywardell: And the OF takes forever to decrypt on the Sansa at least.
23:15:22LloreanRunning a decrypted OF is drastically faster.
23:16:33barrywardellis the cpu boosted? that makes a big difference
23:16:44LloreanNo clue
23:17:40LloreanAnd the firmware is a fairly big file, too. So that doesn't help
23:17:42linuxstbNo, I don't think it is. Which also makes me wonder about that IDE0_CONFIG change - we're setting the so-called "> 65MHz bit" in the bootloader now. Was the Nano bootloader tested with that change?
23:17:48LloreanDon't know how intensive the algorithm is.
23:18:24Lloreanlinuxstb: Look for oblib working with someone else in the logs.
23:18:29barrywardellgiven the sansa bootloader has to decrypt the OF mi4, it can't be that bad
23:18:41briantumorthanks :)
23:20:03Lloreanlinuxstb: I seem to recall some discussion of the issue, but was doing other things and not involved.
23:20:38 Nick dionoea_ is now known as dionoea (
23:22:00roolkuNico_P: for FS #8201? cool
23:22:17Nico_Proolku: want to test the patch?
23:23:03Lloreanlinuxstb: I think amiconn is the one who said it should be set when initing device on the nano. I can't remember why though
23:23:04Nico_Proolku: you'll probably know better than me if it fixes the problem
23:24:17roolkuNico_P: will take a bit, I am currently doing a test compile for the check_dir thing amiconn pointed out
23:24:33roolkuNico_P: but, yes please. :)
23:24:42linuxstbLlorean: I'm guessing oblib didn't test the bootloader, as it was broken with amiconn's backlight work (which has just been fixed). But I guess we'll find out when someone tests the latest Nano bootloader...
23:25:12amiconnLlorean: The OF bootloader being slow doesn't tell much about the algorithm... remember archos boot time with archos loader + OF?
23:25:45amiconnThe descramble *can* be made fast, as our rolo code shows...
23:25:57Lloreanamiconn: I was judging "Rockbox loading encrypted OF vs decrypted OF"
23:26:18LloreanAt least, I think those were the two cases I'm judging on. :)
23:26:25linuxstbI don't think anyone has tried to optimise our mi4 decrypt function either.
23:26:34amiconnah, ok
23:26:43Lloreanamiconn: So, about linuxstb's question. >65mhz is set in the bootloader, is this a bad thing?
23:26:48amiconnBut the OF is quite large, I guess
23:27:04LloreanYeah, large enough that it's not a representative file size for what someone's likely to want to encrypt with a dap.
23:27:10amiconnLlorean: I don't know. I'm not even sure whether this bit is named correctly
23:27:18Nico_Proolku: do you think FS #8196 is related?
23:28:29amiconnA test should tell. That's the nice thing about the ipod bootloaders - there's no big risk in trying somethin gout
23:28:59 Join JdGordon [0] (
23:29:30roolkuNico_P: it shouldn't really, as this calls tagcache_update_numeric directly and doesn't rely on the callback - but who knows
23:29:53Nico_Proolku: really? where from?
23:30:06barrywardellLlorean: decrypting the H10 bootloader takes just a couple of seconds here.
23:30:42Lloreanbarrywardell: This is the full OF though, not just a bootloader.
23:30:49roolkuNico_P: onplay.c:962
23:30:58barrywardelli meant to say full OF
23:31:53Nico_Proolku: ah yes
23:32:02LloreanThe OF in the sansa could be said to be "a few second" but it's long enough to feel annoying to me compared to a decrypted in advance OF
23:32:21barrywardellI think it's the cpu frequency that makes a big difference. As far as I remember, the H10 starts at a high frequency
23:32:55 Quit stripwax_ ("Miranda IM! Smaller, Faster, Easier.")
23:33:56briantumorwhy is there a vmware image and not one for qemu?
23:34:11Bagderbriantumor: we're waiting for you to provide one
23:34:30 Nick ionvas is now known as ionvas__ (
23:37:09preglowjmworx: i've got a patch for mem* usage and elimination of some redundant copies now
23:38:21 Join thegeek_ [0] (
23:38:54jhMikeSpreglow: what is the long term error distribution supposed to be from this function
23:39:38preglowjhMikeS: no idea
23:40:05 Join petur [0] (
23:40:06jhMikeSif I don't differentiate the random, then it's triangular
23:41:17 Quit atsea- (Read error: 104 (Connection reset by peer))
23:43:42 Join johnf1911 [0] (
23:44:04 Join matsl [0] (
23:44:58 Quit zicho (Remote closed the connection)
23:45:07 Quit matsl (Client Quit)
23:45:14johnf1911question" last night I upgraded my H120 to a recent release of rockbox, from one dating from 2k5/2k6; I listened to it during the day, things are generally good; however: I used to get 16 hours of run time on a full charge, when I used it for an hour today on my trip home, it reported an estimated time remaining of 6h00; have there been major batter regressions either in recent release, or compared to ancient release?
23:45:40preglowjhMikeS: rockbox should last longer than ever
23:45:45johnf1911it could be that time estimate is inaccurate / different than it used to be
23:45:46preglowjohnf1911: rockbox should last longer than ever
23:45:47johnf1911irssi what :)
23:45:52preglowjohnf1911: the estimate might be off
23:46:02 Quit Zagor ("Hello mr Oops")
23:46:05*preglow kicks nick completer
23:46:06n1sjohnf1911: that estimate is a notorious lier ;)
23:46:36johnf1911that is true
23:46:43johnf1911I will see how long it really lasts
23:48:04 Join jmspeex [0] (
23:48:11preglowjohnf1911: you should get more than 16 hours now
23:48:18preglowjohnf1911: even i do, and my battery is old and smells bad
23:48:22johnf1911ok, cool
23:48:28johnf1911I'll see what my actual experience is
23:48:32johnf1911and come back if it sucks :)
23:48:42johnf1911thanks again
23:49:03preglowjmspeex: i've got a patch for mem* usage and elimination of some redundant copies now
23:49:11jhMikeSpreglow: :)
23:49:28preglowirssi needs an ai based nick completer
23:49:40jhMikeSif that's supposed to have a triangular ditribution, it doesn't. it's rather gaussian in SVN
23:50:14jmspeexpreglow: can you email it to me?
23:50:26 Join Zagor [0] (
23:50:30 Join atsea-63 [0] (i=atsea-@gateway/tor/x-0472cfeb7316f37b)
23:50:37preglowjmspeex: can i stuff it on my web site instead? my mail server is shit slow right now
23:50:59linuxstbZagor: How are things going?
23:51:10linuxstb(apart from the oopses...)
23:51:51 Quit Jon-Kha (Remote closed the connection)
23:52:47preglowjmspeex: tested with speexenc/dec, seems to work fine
23:53:00Zagorlinuxstb: not much progress. I should probably commit the basics needed for charging before venturing further into the 64-byte test.
23:54:38jmspeexpreglow: most important, does it make a difference in performance?
23:54:47scorche|wZagor: not to interrupt your work, but when you get the time, can you remove the bright yellow color as an option for nicks in the IRC log script?
23:55:09jmspeexpreglow: or code size
23:55:35preglowjmspeex: haven't checked performance, but unless the target mem* sucks badly, it should be faster
23:55:37 Quit thegeek (Read error: 110 (Connection timed out))
23:55:38preglowcan check code now
23:56:04jmspeexpreglow: It could also be the same speed if the copies are negligible...
23:57:23Zagorscorche|w: the colours are created using a simple algorithm using first/mid/last nick character for rgb values.
23:57:33preglowjmspeex: binary size is smaller on amd64, by as much as 0.5kb for the encoder, 150b or so for the decoder
23:58:08roolkuNico_P: hm, my sim refuses to initialise the database now - what have I broken...
23:58:16preglowjmspeex: some of those copies are framelength size, those'll surely be faster
23:58:22Zagorscorche|w: so I can't fix a specific colour, I/we have to fix the algorithm

Previous day | Next day