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).

#rockbox log for 2005-11-09

00:00:06amiconn...but writing has to be done with 16bit width
00:00:34amiconnHow should we handle such special conditions?
00:04:29LinusNdo we have to read them as char?
00:06:19preglowi wonder if i should just remove support for sv4-6 musepack streams
00:06:41amiconnLinusN: Yes, the datasheet says so
00:07:28LinusNso silly...
00:07:36amiconnIn fact, you have to write to TCSR and TCNT with a 16 bit access to the same address (0x05FFFFB8), the high byte determines which register is written
00:07:55amiconnHigh byte 0x5A -> TCNT, 0xA5 -> TCSR
00:08:22amiconnRead address for TCSR is 0x05FFFFB8, for TCNT it is 0x05FFFFB9
00:09:14LinusNlinuxstb: the lowlevel functions in lcd-16bit.c should be moved to lcd-ipod.c or something...
00:09:53LinusNlots of linking errors for the h300 as it is now
00:10:03amiconnLinusN: You're right, but then lcd-recorder.c needs some cleanup too. Today it contains the gmini lowlevel code
00:10:13BagderLinusN: but nothing really serious, is there?
00:10:32 Quit ender` (Read error: 104 (Connection reset by peer))
00:10:49LinusNBagder: no, but i'd have to #ifdef a lot in the lcd-16bit.c to make it build
00:11:09LinusNit's very ipod-centric as it is now
00:11:11linuxstbLinusN: Feel free to change lcd-16bit.c - I haven't made any changes locally.
00:11:39linuxstbThere's also lots of dead code just copied from the h100 driver - as an example of how those functions should be implemented.
00:12:55XShocKjust a quick equstion. gcc 4.0.2 is not supported, right? rockbox builds and strarts, but no plugins work..
00:13:11linuxstbamiconn: What a stange post...
00:13:30LinusNXShocK: you're on your own there...
00:14:06BagderXShocK: start fixing! ;-)
00:14:33 Quit linuxstb_ ("CGI:IRC (Ping timeout)")
00:15:15XShocKok. will actually try now. :)
00:16:06amiconnLinusN: Any idea how to handle the SH1 watchdog? Separate #defines fpr reading and writing??
00:17:58LinusNamiconn: i guess so
00:21:46 Quit Moos ("Glory to Rockbox")
00:21:50XavierGrLinusN: did you read my comment about those mp3s that showed wrong lenght earlier?
00:22:04preglowhmm, i just tried a bootloader build for the nano, and it failed
00:22:47preglowriiiight, probably the tools dep error again
00:23:39LinusNXavierGr: no
00:24:02XavierGrwell it seems that the same bug is reproducable in foobar
00:24:11XavierGrwhile in foobar is way worse
00:24:24XavierGrtags cannot be shown in these files
00:25:08XavierGrif the wrong lenght is smaller than the true it will skip tracks midways.
00:25:13amiconnHmm, something is indeed wrong with the recorder v1 charging. Now that I charged my batteries in an external charger, runtime is already ~13h and still going
00:25:19preglowthe box is really going to have a run for its money with new ipod builds as well, heh
00:25:29amiconnWith internal charging, I get 6...7 hours of runtime
00:25:41XavierGrep0ch said to upload the files and check them
00:25:47XavierGrany idea where to upload some?
00:26:34XavierGr(ep0ch said to upload the files, and he will check them) just to not misunderstand
00:26:37BagderXavierGr: I'll offer a link in /msg
00:28:26*preglow got the ipod money
00:28:57preglowi turned out to use the same bank as zagor, so went swift enough
00:28:57 Join linuxstb_ [0] (
00:29:07linuxstb_Any shops still open in Oslo?
00:29:18preglowseriously doubt it
00:32:18 Join DJDD [0] (
00:32:21preglowi'll see about buying one tomorrow
00:33:49LinusNamiconn: from what i can see, the long-delta check is too generous in the algorithm
00:36:29*LinusN runs the first rockbox code on the h300
00:37:02LinusNjust single stepping the init
00:41:02amiconnIt seems the automated builds got a lot slower than they used to be
00:41:15amiconnAlmost an hour now
00:41:24amiconn...and still running
00:41:37Bagderits due to other processes running on the host
00:42:15amiconnYes, you mentioned that. So the number of other processes increased a lot recently?
00:42:39 Quit XavierGr (Read error: 104 (Connection reset by peer))
00:42:57Bagderactually, I don't think the build time has increased to an hour at all times
00:43:09preglowit doesn't
00:43:22preglowtwelve 'o clock is usually prime cron time
00:44:24Bagderthe last 4 builds averaged on 1875 seconds build time
00:44:54Bagdera little over 31 minutes
00:45:09preglowi wonder what it would be without ccache
00:46:08Bagderwell I'm not gonna try ;-)
00:46:58amiconnLinusN: The comment in crt0.S is now wrong.
00:47:52markunIs it possible to get admin rights for the patch tracker?
00:48:55LinusNamiconn: which one?
00:49:16amiconn. /* Platform: iRiver H120/H140 */
00:49:23Bagdermarkun: indeed, hang on
00:49:35markunBagder: ok, thanks
00:50:02markunBagder: sourceforge login is "marcoen"
00:50:08 Quit matsl (Remote closed the connection)
00:51:41preglowwhat's up with the main.c commit? :P
00:52:22preglowfigured that was testing code
00:53:17Bagdermarkun: you should now have all the super powers there
00:53:29amiconnUsing the watchdog for reset on SH1 works
00:54:22*amiconn has a convenient test plugin "lawbreaker", that just enabled memory guard, then tries to read from location zero
00:55:27markunBagder: Too bad the summary can't be changed..
00:55:50***Saving seen data "./dancer.seen"
00:55:54amiconnTiMiD: There's a bug in the new list browser...
00:56:15amiconnIf you delete the lst file in the list, the list cursor vanishes.
00:59:21 Join XavierGr [0] (
01:00:43Bagderyahoo web crawler crawling down the viewcvs...
01:02:04BagderI've edited the robots.txt now
01:03:53XavierGrep0ch, Linus, or others that are interested: this link has 3 tracks which have the length bug with rockbox ->
01:04:35XavierGrI have others too but mostly classical so the size of them is prohibited for upload with my connection.
01:06:53XavierGris it easy for someone to explain me what's the purpose of yield() function?
01:07:02preglowit passes control on to the next thread
01:07:16XavierGrwith no parameters?
01:07:19preglowrockbox uses cooperative threading
01:07:34preglowif there are other threads running (which there always are), you pass the control on to them
01:07:37XavierGrif we have 3 threads how can I switch to a specific one?
01:07:39preglowyou can't
01:07:47preglowrockbox uses a simple round robin scheduler
01:08:07XavierGrso every time I yiled another thread takes control?
01:08:15preglowyes, that's the entire point of it
01:08:22preglowyou're not supposed to care what thread is run
01:08:32preglowif you do, you're doing something wrong
01:08:33 Join ashridah [0] (
01:09:03XavierGrso what will happen if I call yield twice in a row?
01:09:38XavierGrwill the scond time be executed?
01:09:50TiMiDamiconn: I'm investigating
01:10:14BagderXavierGr: yes, when it has run all the other threads one lap
01:10:14amiconnAlso, there's still the player default icon problem
01:10:25TiMiDit's solved (I think)
01:10:42TiMiDI will commit if i find a way to solve this one too
01:11:00amiconnOn target, the icon isn't always blank, sometimes it shows a garbage icon
01:11:24XShocKomg. i can't replicate the error i got with gcc 4.0.2... it now works as it should....
01:11:41TiMiDthe problem is just that when deleting, tree.c doesn't put the cursor to the right place
01:15:00amiconnBagder: Does the server do one automated build for every commit, even when several commits queued up during the previous build "round"?
01:15:37XShocKanybody got the same result with gcc 4.0.2 ? "rockbox loads, but can't load plugins."
01:16:03XShocKsince i built it again, put in the player, and it works perfectly...
01:16:18Bagderamiconn: after a build is made, it waits a minute and checks for changes
01:16:28Bagderany number
01:16:33Bagderand if any, it rebuilds
01:16:41Bagderif not, it waits a minute and rechecks
01:19:05amiconnSounds like it should be :)
01:19:33TiMiDwhat was the previous behaviour when deleting a file ? selecting the first item of the dir, the next item, the previous one ?
01:20:13amiconnIiuc the cursor was placed on the next file, unless the last one was deleted
01:21:23TiMiDI will look at the old code to see how it was handled
01:21:24linuxstb_That sounds natural anyway - i.e. it keeps the cursor in the same position and shifts all subsquent files up one place.
01:21:54TiMiDsince there is no indication given to the filetree that a file has been deleted
01:22:35LinusNthat should be enough, shouldn't it?
01:23:21TiMiDI wonder what trick do they made to make it work :(
01:25:10TiMiDI see a way to fix it, maybe it was the one previously used
01:25:36linuxstb_All you need to do is remember the cursor position in the list, reload the directory, and then check if it's still within the bounds. If not, then set it to the last file.
01:27:02TiMiDexacly what I thought ;)
01:27:44TiMiDit works :)
01:29:44*preglow kicks musepack
01:30:34XShocKok. IHP120 works perfectly with gcc 4 ( tested everything except audio files, since only have mp3s..)
01:31:00linuxstb_preglow: You seem very determined to hammer musepack into shape.
01:31:39preglowlinuxstb_: yes, after i realized i knew it well enough to do it
01:31:45preglowbut it is remarkably resistant
01:34:06linuxstb_XShocK: Can you test a change to libfaad with gcc4? It's part of the code that causes an internal compiler error with 3.4.x
01:34:43linuxstb_You just need to change "#if 0" to "#if 1" in line 74 of apps/codecs/libfaad/fixed.h and then compile
01:35:20preglowremarkable, the stack is _JUST_ not big enough for me to stuff a couple of arrays into it
01:35:40preglowany reason why the codec stacks are exactly 1152*4*2 ? :)
01:35:40linuxstb_XShocK: Thanks. I'm curious if gcc4 is better than gcc3 in that respect.
01:39:22XShocKit compiled on gcc 4
01:39:30XShocKand on gc 3.4.3 it failed
01:39:59linuxstb_Maybe that means we should all upgrade then...
01:39:59LinusNtime to sleep, nite all
01:40:00preglowperhaps it's time to move on to gcc4 for m68k
01:40:03preglowLinusN: gnite
01:40:07 Part LinusN
01:43:22linuxstb_XShock: Could you upload a compiled with gcc4 to somewhere? I can test all the codecs.
01:44:29linuxstb_Thanks. I don't expect any problems, but you never know.
01:51:21 Join Midgey34 [0] (
01:53:47linuxstb_Just downloading it now. Give me 2 minutes to test.
01:53:57 Quit linuxstb_ ("CGI:IRC")
01:55:39preglowseems there's some dep problems with the map files or codec plugins
01:56:29linuxstbWow - AAC is realtime :)
01:56:46linuxstbAlmost... It played for 20 seconds and then a tiny pause.
01:57:07linuxstbThen another 20 seconds, and another tiny pause.
01:57:30 Join _user_ [0] (
01:57:31preglowperhaps i should build myself a gcc4 toolset myself
01:57:33linuxstbIt's very very close now.
01:57:41linuxstbYes - I think we should all upgrade.
01:58:10linuxstbWould your multiplication cause this improvement, or do you think it's also gcc itself?
01:58:56preglowi tend to believe the last
01:59:03preglowMUL_C isn't used that often outside of sbr
01:59:17linuxstbIn which case, most of the codecs should see an increase then.
01:59:43preglowit's worth a try
01:59:53preglowperhaps you'd like to do some quick boost ratio comparisons? :)
02:00:35linuxstbhehe. Just give me 200 hours...
02:00:52 Quit novimon (Read error: 110 (Connection timed out))
02:00:57linuxstbSorry, you meant boost ratio, not runtime?
02:02:13linuxstbNo noticeable difference with ALAC.
02:03:31preglowwhen i want to replace an array in a struct with a pointer to some outside array without modifying any source, i just replace a int hehe[400]; by int *hehe; and assign the pointer somewhere else. what if i want to do the same for 'int hehe[30][30]'? there's nothing i can do like that then, no?
02:04:02linuxstbI've thought about the same problem - without thinking of a solution.
02:04:23preglowi can fit the entire mpc_decoder struct in iram now, but i need to outsource a set of arrays like that :/
02:06:33amiconnPointers to multi-dimensional arrays are possible, but don't ask me how to insert these into a struct
02:06:54linuxstbNot sure if this helps:
02:06:57amiconnIirc TiMiD's fire plugin used these before I did my optimising work
02:07:56preglowamiconn: yes, they are possible, but you need to edit the source using the struct after substituting the original multidim array for a pointer
02:08:45preglowlibmad uses them all the time
02:14:14 Quit XShocK ("Leaving")
02:15:11*preglow has tamed musepack
02:15:40preglowthat is, if i can find a good solution to the above problem
02:15:50preglowit hardly boosts at all now
02:18:08preglow10% boost ratio on a standard file
02:18:11 Join XShocK [0] (
02:18:22TiMiDgood night here
02:18:25preglowgod knows how fast this could have been without the retarded math
02:18:32preglowit would never have boosted, that's for sure
02:22:04linuxstbXShocK: Everything seems to be working fine with your gcc4 build.
02:23:35preglowthink i'll try myself that build as well
02:26:54preglowrealtime aac
02:26:56preglowso it is possible
02:27:15 Quit tvelocity ("Leaving")
02:30:36linuxstbXShock: I'm assuming this build includes the "#if 1" change?
02:36:00linuxstbnight all.
02:48:46 Join BirdFish [0] (n=bradbox8@
02:53:39XavierGramiconn: unfortunately the fix you made didn't do much on the ticking issue...
02:55:12XavierGrand I found something new and weird...
02:55:51***Saving seen data "./dancer.seen"
02:56:58XavierGrif I press with my finger the remote jack (while pluged-in, vertical to the player, not parallel to the ports) these ticks are almsot gone....
02:58:28 Join Midgey31 [0] (
03:09:27 Quit Midgey31 ("Download Gaim:")
03:09:51 Join Midgey31 [0] (
03:12:49 Quit _user_ ("CGI:IRC (EOF)")
03:14:35 Quit Midgey34 (Read error: 110 (Connection timed out))
03:15:20 Quit DJDD (Read error: 110 (Connection timed out))
04:11:26XavierGrstupid windows SIM, I can't manage to run a single plug-in of my own, while it runs perfectly on the target....
04:22:36 Quit solexx (Read error: 104 (Connection reset by peer))
04:28:00 Join solexx [0] (
04:29:23Midgey31XavierGr: which plugin is this?
04:41:46XavierGrwell it is a plugin that I am working to it right now.
04:41:54 Quit hardeep ("BitchX: now Y2K compatible!")
04:42:25Midgey31oh mysterious... I wait until you reveal more
04:42:33XavierGrthe problem is that as soon as I call a function from the plug-in starting point, the sim crashes, while the target continues.
04:43:18Midgey31hmm, well I don't think can be much help, no h100 to test with and very limited knowledge of C (a fact I hope to soon correct)
04:43:45XavierGrdo you have an mp3 player at all?
04:44:07XavierGrhmm nice
04:44:07Midgey31my bother has a iAudio X5
04:44:17Midgey31and the other has a creative zen muvo 2
04:44:18XavierGrnice too
04:44:34XavierGrno UMS
04:44:42XavierGror it?
04:44:56XavierGrI think muvo 2 is UMS, no?
04:44:58Midgey31umm, I'm not too familiar with it
04:45:03Midgey31but he seems to like it
04:45:16XavierGr4GB right?
04:45:59Midgey31its actually the MuVo 2 Fm if that makes any difference
04:46:23Midgey31the fact that it can only play MP3 and WMA is why I tend to stay away from it
04:46:28XavierGryes it is UMS as far as I can remember
04:46:39XavierGrit was the first DAP i considered bying
04:47:04Midgey31I did the same because it was so much cheaper than an iPod
04:47:05XavierGrthen I saw a ZEN extra 60GB but I realised that without UMS capabilities I couldn't live
04:47:15Midgey31I really like the Zen
04:47:21Midgey31although it is a big large
04:47:30XavierGrthen I saw an iHP-140 I fell in love with it and still I am...
04:48:01Midgey31they were discontinued at the time I got my h320 and I don't have found memories with ebay
04:48:35XavierGrwell I am considering to buy an h340 as an alternative if my unit fails...
04:49:01XavierGrbut they are not easy to find and I don't think I can afford another 400euros
04:49:42XavierGrbut even in thought that one day it will be the most capable Rockbox mp3 player...
04:50:04Midgey31for the time being
04:50:11Midgey31who knows what may come along
04:50:18Midgey31but it is a powerful little bugger
04:50:22XavierGrIn fact I will not buy another mp3 player if it isn't supported by Rockbox...
04:50:55Midgey31I haven't gotten to experience rockbox, but I'm pretty sure I'll feel the same once I do
04:51:50Midgey31by the way, isn't it _really_ late/early over in Greece right now?
04:52:22XavierGrvery late, I should sleep in 1 hour and 30 mintues I should be awake...
04:52:38Midgey31I don't think I could ever do that
04:52:58Midgey31I have to get up in 7 hours for class and I'm not looking forward to that...
04:53:00XavierGrnow that you mention it, I don't know if I can too.
04:53:29Midgey31screw sleep, load up on caffeine and you'll be good to go
04:55:05XavierGrmaybe I will just lye down on my bed for the rest of my time.
04:55:07XShocKant there any custom firmwares for RCA Lyra?
04:55:16XShocKJukebox 20gb
04:55:32XShocKit has DSP DM320 from TI>
04:55:54***Saving seen data "./dancer.seen"
04:56:46XShocKor DM230, not sure.. anyway i doubt there are any firmwares around.. i couldn't even find datasheet for that DSP... only analog, C55x..
05:36:56 Join Lost-ash [0] (
05:36:56 Quit ashridah ("Leaving")
05:38:44 Quit Midgey31 ("Download Gaim:")
05:39:58 Nick Lost-ash is now known as ashridah (
05:45:11 Quit XShocK (Remote closed the connection)
06:16:30XavierGrI just ran flat my H140s battery.
06:17:21XavierGrI forced it to work until it reached a 2.73V level. Is that normal? I thought that this is li-ions internal limit...
06:18:11 Join _user_ [0] (
06:19:30 Join DJDD [0] (
06:33:49 Quit _user_ ("CGI:IRC (EOF)")
06:44:55 Quit RotAtoR ()
06:52:31 Join XShocK [0] (
06:55:56***Saving seen data "./dancer.seen"
07:11:17 Quit DJDD (Read error: 110 (Connection timed out))
07:36:37 Join B4gder [0] (
07:50:36 Join DJDD [0] (
08:19:34 Join bbad [0] (n=bbad@
08:29:06 Join einhirn [0] (
08:42:49 Join ender` [0] (i=ychat@
08:56:00***Saving seen data "./dancer.seen"
09:08:02 Join Paul_The_Nerd [0] (
09:08:37 Join LinusN [0] (
09:24:20Zagorneat, ebay phishing attempt to rockbox-dev
09:39:00Paul_The_NerdThe sad thing is that enough people fall for them that it's worth their time to keep doing it.
09:42:29 Quit DJDD (Read error: 110 (Connection timed out))
09:43:37ashridahPaul_The_Nerd: the annoying thing is that even if they send out 1 million emails and only get 10 hits, it's economical for them to do so.
09:44:03Paul_The_NerdIt's probably economical even if they get 1 hit.
09:46:17 Part LinusN
09:51:01 Quit Lynx_ (" reboot")
09:53:58 Join Lynx_ [0] (
09:57:22 Quit korpse ("brb changing encodings")
09:59:40 Join korpse [0] (
10:03:41 Quit korpse (Client Quit)
10:12:32linuxstbWould anyone be against requiring gcc4 to build the Coldfire version of Rockbox? Switching to it would appear to solve the ICE problems we've been experiencing with libfaad and it also appears to give a noticeable speed improvement (AAC is very very close to realtime when compiled with gcc4).
10:18:11B4gderwhy would we need to require it?
10:18:18B4gdercan't we just recommend it?
10:18:44 Join Jungti1234 [0] (n=jungti12@
10:19:31merbananB4gder: like "gcc4 is recommended coz then Rockbox works instead of not" :)
10:19:34 Quit Kohlriba ("Leaving")
10:19:56B4gderlike that
10:20:08B4gderI bet most people don't have aac songs anyway
10:24:32linuxstbBecause gcc 3.4.x fails with ICEs in parts of libfaad. e.g. one of preglow's optimisations caused an ICE.
10:24:34Paul_The_NerdThey're more likely to once the iPod port is done though
10:24:48Paul_The_NerdHave AACs, I mean, B4gder
10:24:58linuxstbSupporting 3.4.x and making full use of gcc 4 would mean #ifdefs
10:25:03ashridahyeah, anyone buying music off itunes potentially has AAC's don't they?
10:25:31ashridahdoes gcc 4 build all currently supported platforms without any detrimental effects?
10:25:35linuxstbBut I think I would propose at least a transitional period where gcc4 is recommended but gcc3 would still work.
10:25:41B4gderashridah: no
10:25:54*ashridah imagines there's code size issues
10:27:16 Quit mbr ("leaving")
10:27:17B4gderthere are more issues than just the size on sh1
10:28:05linuxstbAt the moment it is just one small function in libfaad that would need an #ifdef around it, but it wouldn't surprise me if we hit more ICE problems when working with libfaad - it seems very fragile in that respect.
10:29:36 Join mbr [0] (
10:31:02 Join Kohlrabi [0] (
10:34:13preglowwe wouldn't need to use gcc4 for sh1 anyway
10:34:19preglowwe pretty much can't
10:41:28linuxstbAnyone know a reason not to use gcc4 for ARM?
10:43:58 Join ep0ch [0] (n=ep0ch@
10:44:44preglowi'd like some code size and speed tests with gcc4, but the speed part is of course not straight forward as it is
10:45:04preglowbut i guess the fact that it at least compiles the code we need it to counts for something
10:45:19ep0chXavierGr: I've got your mp3s
10:46:02ep0chXavierGr: What tags are missing? In foobar2000 0.9b10 i can see tags
10:46:52ep0chXavierGr: the bitrate is being reported a bit wrong though :s
10:56:02***Saving seen data "./dancer.seen"
11:01:46 Quit Jungti1234 ("Http://www.ZeroIRC.NET Zero IRC Ver 2.8")
11:04:27ep0chI have a suspicion the GCC4 build is boosting mp3 more than before.
11:09:33ep0chyeah my 112 kbps mp3 boosts with 6% with that GCC4 build. Bleeding edge build doesnt boost at all.
11:10:41ep0chnow for Vorbis!
11:12:18linuxstbSo it seems it's not going to be straightfoward choice then...
11:13:30preglowi also got the impression libmad was slower, which was why i wanted speed tests :/
11:15:31preglowyeah, i can confirm that, i've got a 112kbs file that now suddenly boosts
11:15:46ep0chVorbis@137: GCC3 28-29% boost, GCC4: 34% boost
11:15:48preglowi assume this isn't the result of lear's synth.c hack to make gcc4 work, though
11:16:32preglowm68k doesn't exactly seem to be an important target for gcc these days
11:17:33ep0chflac aint boosting though :)
11:20:17preglowwhat about codec sizes?
11:24:38 Join Jungti1234 [0] (n=zeroirc@
11:27:48 Quit linuxstb ("Leaving")
11:28:00 Part Jungti1234
11:29:57XavierGrep0ch are you here?
11:30:09XavierGrI use foobar 0.8.3
11:30:25XavierGrwhat length do you have for the first song?
11:35:24 Join Jungti1234 [0] (n=jungti12@
11:36:26ashridahhm. if i use gcc 4.0.2, do i need to update binutils?
11:37:45XavierGrep0ch: Yes foobar 0.9b10 seems to solve the tag problem
11:37:54 Quit ashridah ("Leaving")
11:38:05 Join ashridah [0] (
11:38:07XavierGrep0ch though the track length still remains.
11:58:10 Join ender1 [0] (i=ychat@
12:12:02B4gderashridah: no, you can use the same binutils
12:13:03 Quit ender` (Read error: 110 (Connection timed out))
12:17:05ashridahB4gder: tah. i've already compiled it up and copied it onto my iriver anyway, it bootedokay, but haven't tested plugins yet
12:25:50 Quit XavierGr ("Trillian (")
12:31:16 Join DJDD [0] (
12:32:18ep0chpreglow: aac and flac codec size got a tiny bit smaller with gcc4
12:32:31ep0chpreglow: mpa, mpc, vorbis and wavpack got bigger
12:33:01ep0chi can paste the actual sizes if you want
12:33:38ep0chyeah aac lost a whole 23k
12:35:16 Join XavierGr [0] (
12:37:15markundamn, this bejeweled game is keeping me from working.. level 10 now.
12:38:03 Join DJDD_ [0] (
12:39:05Paul_The_NerdI always run out of moves around 4 or 5
12:39:25markunI want to quit, but I don't want to loose the score.
12:39:52Paul_The_NerdYou can save the game, you know
12:40:01markunYes, I know :)
12:40:29Paul_The_NerdOh, so that's "Just an excuse"
12:41:24markunlevel 11..
12:49:41 Join korpse [0] (
12:51:03Jungti12342005-11-09 01:35
12:54:07 Quit DJDD (Read error: 110 (Connection timed out))
12:55:59preglowseems they only have 2 gig nanos left around here
12:56:03***Saving seen data "./dancer.seen"
12:56:08preglowoh well
12:57:02markunWill you order a 4gig one?
12:57:15preglownah, i want it today :>
12:58:18XavierGrimpatient like me preglow...
12:58:27preglow2 gigs should do for my use anyway
12:58:43XavierGrwhen an idea is tuck in my head I can't get it out!
12:59:18markunI wonder how good the hardware of the iPod 4g is compared to H120.
13:00:57markunI might buy one for my girlfriend when is runs rockbox
13:01:49XavierGrI am annyoed by this iPod outbreak
13:02:17XavierGrnow I can't boast that my player is one of the unique ones that runs Rockbox
13:02:24markunWait till linus finished the H300 loader, there will be another outbreak
13:02:46XavierGrwell h300 is the same as h100 so... ;|
13:02:58XavierGron the other hand
13:03:08ep0chi don't mind the H300 outbreak, might mean some more people working on codecs
13:03:09XavierGrit will be a fair clash between H100 and an iPod
13:03:33korpsewhy would there be an outbreak?
13:03:37korpse(for the uninformed)
13:03:42Paul_The_NerdThe iPod still doesn't have a few of the features that the H100 has though? Right?
13:03:48markunbejeweled is finally over at 1263 points..
13:03:49XavierGrep0ch: wanted to tell me something?
13:03:50Paul_The_NerdFM, Optical, most notably.
13:04:04XavierGrPaul_The_Nerd: A lot not some
13:04:05markunPaul_The_Nerd: yes, that's true
13:04:12preglowcan't say i care what rockbox runs on
13:04:21XavierGrthat's why I said a fair clash
13:04:40XavierGriPod users were backing their reaons on the database and ease of use
13:05:00XavierGrbut now that it will run rockbox it will be chopped
13:05:03Paul_The_NerdYeah, but neither of those will be consistent with Rockbox.
13:05:09XavierGrno reckording no radio, no to a lot
13:05:11ep0chXavierGr: not really looked at those mp3s much. did you compress them yourself?
13:05:16Paul_The_NerdThough one could conceivably support the iPod database on Rockbox
13:05:37Paul_The_NerdHeck, you could even support iTunes encrypted files if you *really* wanted to, though it'd probably be a much better idea to decrypt them before copying over
13:05:46XavierGrep0ch: yes, I just don't remember with which encoder...
13:05:50XavierGrbut the timings are wrong no?
13:06:06ep0chi'll try and find out about the encoder
13:06:12ep0chyes times are very wrong
13:06:24 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
13:06:25Jungti1234markun: Is some because I play bejeweled game in msn messenger?
13:06:35XavierGrthough winamp and iriver fw show it correctly...
13:06:41ep0chand something wrong with the bitrate too
13:07:07ep0chprobably down to dodgy length
13:07:32XavierGrwell I have made a wird outcome with this
13:07:48XavierGron my player the durations are swapt between tracks
13:08:00XavierGre.g the next track holds the duration of the previous one!
13:09:12XavierGrfucked up situation
13:09:27XavierGrI still don't get why winamp can show it correctly
13:09:39B4gderbecause winamp prolly scan the whole file
13:10:40XavierGrthough Linus insisted it was a rockbox bug.
13:10:52B4gderI didn't exclude that possibility
13:11:04 Nick paugh is now known as AliasCoffee (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
13:11:05B4gdercomparing to winamp however is always a weak argument
13:11:21XavierGrand fw iriver
13:11:22B4gdersince winamp spends loads of memory and loads of cpu that Rockbox just doesn't
13:11:26XavierGrbut that weak too
13:11:36ep0chXavierGr: compressed with lame 3.95 apparantly
13:11:43XavierGrI will try other players too.
13:11:44B4gderI think the comparison to the original much better in this case
13:12:22XavierGrhmm which program did I use for this.... it must have been this idiotic dbpowerAMP I can never trust it again.
13:12:53XavierGrand the problem is that I don't know how many mp3s are like that on my collection...
13:13:02 Join amiconn_ [0] (
13:13:05XavierGrreripping is not an option.
13:13:48Paul_The_NerddbPowerAMP has been pretty reliable for me, but I will admit I don't use MP3
13:14:21XavierGryes I use it for lossless but I can't really trust it on mp3
13:14:30XavierGrI will use only EAC from now on...
13:15:45Paul_The_NerdDon't they both use LAME for their encoding though?
13:16:23XavierGrwell I had to download lame for both of the programs
13:16:29XavierGrat least latest version
13:16:40XavierGrbut EAC is the sure thing anyway.
13:16:52XavierGrit will copy wav on disc and then encode
13:16:53ep0chXavierGr: was you're original source from CD then?
13:17:03XavierGrand check for errors and f any reports them
13:17:52Paul_The_NerdThe EAC error checking is just to verify the sector reads, it'd prevent skips in the MP3 and such primarily.
13:18:05Paul_The_NerdIt's not likely to fix encoding issues like that if I understand...
13:18:27Paul_The_NerdI do the opposite really... dbPoweramp for OGG lossy, and EAC when I want to make an archival copy of a CD.
13:18:31XavierGryes but still this is very usefull
13:19:26XavierGrI have many mp3s that have clips and bips (due to software) and I never got informed about that, only when listening, now it is too dificult for me to track down all these tracks
13:23:42ep0chXavierGr: i just used Mp3 repair tool 1.5, install it, and select "Remove everything after last frame"
13:23:48ep0chseems ok in foobar now
13:24:09XavierGrdid it removed audio data or tags?
13:24:13ep0chno thats a lie
13:24:14 Join linuxstb [0] (
13:24:28ep0chi told it to remove 1 frame at the begining
13:25:02ep0chwell the tags still seem to be there
13:25:23XavierGrlinuxstb: I can say that your nick is contraddicting from now on! :)
13:25:43XavierGrLINUXstb to make apple ipod ports, nooooo
13:26:06preglowwhat, it's not like he's porting mac os x or something :)
13:26:12XavierGrep0ch: so did it worked?
13:26:13B4gderwhat is the stb fore anyway?
13:26:33linuxstbThe only features missing on the iPod are digital I/O and radio. It has recording (with varying samplerates up to 96KHz) but line in/out are only available via the proprietory dock connector.
13:27:19linuxstbB4gder: It comes from a project I started and abandoned years ago - to develop DVB software for a linux set-top box.
13:27:19XavierGrfanboy :D
13:27:27linuxstbXavierGr: Advocate. I want help :)
13:28:18preglowtime for me to see if there's an ipod available around here
13:28:18 Quit Kohlrabi ("Leaving")
13:28:58Paul_The_NerdRockbox support would be the decision influencing my purchase of a Nano if I do.
13:29:48linuxstbMy main motivation for the port is that ipods are _very_ easily available. Unlike the iriver devices now. There are millions all over the planet.
13:29:53 Quit amiconn (Read error: 110 (Connection timed out))
13:29:54 Nick amiconn_ is now known as amiconn (
13:29:54preglowand perhaps you'll see the first photos afterwards :)
13:29:57preglowlinuxstb: exactly
13:30:03Paul_The_NerdYeah, that is a good reason
13:30:09ep0chhas the nano got that cool clicky wheel?
13:30:14XavierGrcome to think if this port is succeesfull and finished it will be the only target that remains in the market
13:30:16ep0chfor scrolling etc
13:30:19XavierGrin fact brand new..
13:30:30preglowep0ch: i should most certainly hope so
13:30:52ep0chdidnt think it was big enough to act as a scroll wheel
13:30:59linuxstbThe color ipod that I have is already obsolete - it's now replaced by the video ipod. The Nano is the only model which can still be considered current.
13:31:13korpse(with regard to the Nano)
13:31:28DBUGEnqueued KICK Jungti1234
13:31:28*Jungti1234 ̽ʴϴ.(wait)
13:32:13Paul_The_NerdApparently, korpse, Apple is already replacing the ones with actually defective screens
13:32:40korpseoh okay
13:32:52preglowwell, yes, with screens that are said to be not too good either
13:32:59preglowsome people claim all nanos scratch too easily
13:32:59SlasherHmm, i really don't understand why i can't get any response from the eeprom chip.. I think i will go and open the iriver again and hook up a logic analyzer
13:33:20Paul_The_NerdThe defect is that the LCD actually cracks beneath and gets these horrid burn-looking thigns on 'em
13:33:34preglowall ipods i've seen are scratched like hell anyway
13:33:41preglowsince people refuse to use protective cases with them
13:33:45Paul_The_NerdBut yeah, I've heard they do scratch easily. But supposedly they're made of the same stuff as normal iPods
13:33:52korpsei'd like to get a portable audio player, but all of them seem to have rather annoying drawbacks
13:33:57markunSlasher: What
13:33:57Paul_The_NerdThey just have a sleeker surface so it's more apparent.
13:34:06linuxstbkorpse: Then buy three or four :)
13:34:23korpselinuxstb: can i bill you? :)
13:34:24markunSlasher: What's in the eeprom? The code that starts the firmware from flash?
13:34:37Slashermarkun: trying to get eeprom memory working but it doesn't answer anything to commands..
13:34:54Slashermarkun: it's optional setting memory, 1 KiB in size
13:35:21korpsethe only iRiver HDD player still in production is the H10, right?
13:35:27Slasherwe could save the config structure in there instead of disk for example
13:35:41korpsehow does the H10 stack up against the iPod family (of audio players)
13:35:57Paul_The_NerdThe H10 makes kittens cry.
13:36:09markunSlasher: would be nice, then we can save the config every time and don't have to worry about the drive spinning up
13:36:27Paul_The_Nerd(Honestly, I don't really know. But all I've heard are various complaints about battery life, and other things I haven't paid attention to)
13:36:35korpsePaul_The_Nerd: hmm
13:36:49Slashermarkun: yes
13:36:56korpsewhat's the development time for the iPod version of Rockbox looking like?
13:36:57Paul_The_NerdI have an h120, so when I read negative things about the H10 I kind of tend to skim over them.
13:37:00XavierGrSlasheri: I have a suggestion that came to my mind yesterday....
13:37:07korpseactually, let me just consult the wiki
13:37:33XavierGrWhile the dircache works magnificently the same doesn't applies for the playlist viewer.
13:37:39Paul_The_NerdThere's no real timeline with any of this, korpse, but judging from linuxstb's obvious manic obsession, it's probably gonna go pretty well.
13:37:42markunI believe that H10 stays in standby mode when you turn it off (just like the ipod) and slowly drains the battery
13:37:49XavierGrevery time I tried to see the playlist viewer the HD would kick in.
13:38:01SlasherXavierGr: Hmm, true. The playlist is a whole separate thing
13:38:22SlasherIt reads part of the playlist file from disk to its own cache
13:38:29korpsePaul_The_Nerd: in terms of what you can actually buy (off the shelf) at the moment, do you think there is anything that beats/competes with the iPod?
13:38:36XavierGrI see no good reason to load names from the disk for the playlist
13:38:53B4gderkorpse: iAudio
13:39:02Paul_The_NerdiAudio X5 is supposed to be pretty impressive.
13:39:23korpsemmm, time to whip out google again :P
13:39:30XavierGrSlasher: Do you think that i caching of playlist happens it will have a major drawback?
13:39:33SlasherI don't exactly know how the playlist works, but for somebody who has worked with it, it might be possible to save the playlist entirely on memory on iriver
13:39:54SlasherXavierGr: i am not sure
13:39:58crwliAudio X5 is really quite cool
13:40:10SlasherIt eats more memory but not sure how much
13:40:17crwlit's mostly like h300, i think, but not so freakin' huge
13:40:22korpseis there an iAudio Rockbox port?
13:40:29B4gderkorpse: there's a start of one
13:40:40B4gderbut no more than so
13:40:48korpseah okay
13:41:08B4gderan opportunity to show off!
13:41:19B4gderdo the port
13:41:33preglowbut i'm out
13:41:50korpsehaha, the most low-level coding i've ever done was programming some crappy chip at school
13:43:40 Quit AliasCoffee ("Leaving")
13:46:58 Quit Paul_The_Nerd (Read error: 104 (Connection reset by peer))
13:48:04korpsewhat is the X5L?
13:48:29crwlwasn't that just a thicker model with larger battery?
13:48:51korpsenot a clue
13:49:00korpsethat could easily be right though
13:49:09korpseis the M5 an older model?
13:50:05korpsecrwl: looks as though that is the case
13:53:32XavierGranyone that knows rockbox thread internals?
13:54:05XavierGrI am trying to set up a tsr plugin that will benchmark the battery.
13:54:22ep0chm5 is in fact a newer model
13:54:29ep0chwith less features
13:54:33XavierGrI am on something but it is buggy and I can tell its response, cause my lack of knowledge on the matter
13:54:49korpseep0ch: ah, thank you
13:55:03XavierGr^^can't tell its response
13:55:10B4gderXavierGr: what's your question about the threads then?
13:55:18ep0chand yeah X5L just has a bigger battery if i recall
13:55:59XavierGrok I call a thread after I have set the stack with rb->create_thread and get the id
13:56:33XavierGrwill that run the thread code? or do I have to call the function with the treah explicitly?
13:58:04B4gdercreate_thread() adds a new thread to the chain
13:58:05korpsedoes the iAudio's AC adapter do 110-240v? or should i look for a replacement adapter locally?
13:58:40B4gderkorpse: I don't think you should expect very detailed iAudio knowledge in here
13:59:00XavierGrBagder: and one of the arguments is the threads name (a function) will it run the code too, or do I have to call the function?
13:59:35korpseB4gder: just thought maybe someone owned one
13:59:41B4gderXavierGr: no, it adds that thread to the list of threads and then it will be run in the round-robin fashion all threads are run
14:00:09korpseB4gder: i know this isn't #audioplayerreviews or whatever, you guys just seem more knowledgeable than anybody else i could ask :)
14:00:27B4gderkorpse: I'm just saying, I might be proved wrong
14:00:50XavierGrhmm got it and what rb->plugin_tsr exaclty does?
14:01:38B4gderregisters a callback
14:02:01B4gderthat will be called when another plugin wants the memory
14:02:16B4gdersince only one plugin can be run at a single moment
14:03:18XavierGrso when a user has started a tsr plugin and then chooses another one, rb->plugin_tsr will run?
14:03:34B4gderthe function you set with plugin_tsr() will be run, yes
14:04:12XavierGrand what if the user selects the tsr plugin that was previously started?
14:04:21B4gderit makes no difference
14:04:41XavierGr(e.g. selects alpine_cdc and the selects it again)
14:04:52B4gderif a plugin is started when a tsr is running, that callback is run before the new plugin is loaded
14:05:16XavierGrso it will just call the tsr function and end the thread?
14:06:02B4gderit will call the function
14:06:03XavierGror whatever we program in plugintsr( function)
14:06:19XavierGrnow I understand
14:06:19B4gderthe ending of thread or whatever is made in the function itself
14:06:34XavierGryes that's what I meant
14:06:34B4gderall this is clearly visible in the alpine_cdc plugin
14:07:04XavierGrwell... not so clear for my inexperianced eyes :(
14:07:12XavierGrbut thanks anyway!
14:07:19B4gderalthough I believe it bugs...
14:07:24XavierGrnow I have a more clear view of it
14:07:38B4gderno, I take that back
14:07:48XavierGrwhat that plugin is for anyway?
14:08:01B4gderits for a cd changer emulation
14:08:47XavierGremulate what exactly, why someone would want this?
14:10:08linuxstbI'm guessing that it is to interface an Archos device running Rockbox with an in-car CD player. I'm sure hardware modifications are needed.
14:10:39XavierGrthat's what I thought but that sounds too specific
14:10:43B4gder"This is a feasibility study for Archos emulating an Alpine M-Bus CD changer."
14:10:51B4gderthis is all mentioned in the source
14:11:21linuxstbYou're right, it is very specific. But it doesn't do any harm to have it in the standard Rockbox.
14:11:24B4gder"You need to make an adapter with an 8-pin DIN plug for the radio..."
14:11:49B4gderand it serves as a good example on how to write a tsr
14:12:05B4gderas it is the only one afaik
14:12:08XavierGrit is the only tsr plugin right?
14:12:54XavierGrrockbox == endless possibilities! :D
14:13:39 Join einhirn_ [0] (i=Miranda@
14:14:21 Quit DJDD_ ("Trillian (")
14:17:19 Quit einhirn (Read error: 113 (No route to host))
14:19:15 Join DJDD [0] (
14:32:16preglownow let's see what we can see
14:32:32ep0chyou got it already?
14:32:49preglowthere's a mac shop ten minutes from here
14:33:00preglowentering mac shops is like entering a new bloody dimension
14:33:01ep0chyou've opened it up before even pressing play havent you :D
14:33:17preglowwhere everything is sleek and everyone are constantly fascinated at what new things the prophet has conjured up
14:33:33Zagorit's not a shop, it's a church
14:33:42preglowthe guy selling it to me insisted i had a look at this new program mac had just released
14:33:52preglowwork must be a blast to him
14:35:06preglowbloody horrid packaging
14:35:53preglowi think i like this thing
14:36:17preglowit has the scroll wheel
14:36:19preglowworks just fine
14:36:57ep0chwell, hurry up and get RB working on it, i want one now :)
14:37:59ep0chthis will mean you have less time to optimise the codecs for ihp now :(
14:38:29ep0chtake it back take it back!
14:38:53XavierGrpreglow: how much?
14:40:39linuxstbpreglow: I'm going to eat now - I expect photos when I get back :)
14:40:55preglowof it, or it running your bootloader? :P
14:42:34preglowjust need to find out why it makes no sound first
14:54:03 Quit ashridah ("Leaving")
14:56:04***Saving seen data "./dancer.seen"
14:56:20markunpreglow: is it broken or did you get sound now?
14:56:26preglowgot sound
14:56:49preglowsound level is a bit weak for my tastes, but nothing major
14:57:06markunThe quality is ok?
14:57:13preglowyeah, i think it's perfectly decent
14:57:21preglowbut it badly needs some charge, i see
14:57:33linuxstbpreglow: I think I read that Apple artificially limit the volume in their EU models.
14:57:43preglowlinuxstb: then hip hooray
14:57:54preglowfor that shall be fixed
14:58:19preglowlinuxstb: i'll just give it some charge, then boot to linux and see what's up
15:02:58linuxstbpreglow: I can't see any differences in the IPL source between the normal 4G ATA driver and the Nano ATA driver - but you will be the first person to run it....
15:09:25preglowlinuxstb: we'll see
15:09:37XavierGrwow the tsr plugin procedure works indeed, I just made a void thread (starts and exits)... now what should I write inside the thread....?
15:09:52preglowif the ipl guys haven't broken an ipod yet, so i wont be the first
15:09:53XavierGrbut better take some sleep first.
15:09:54preglowi hope
15:11:55korpsehow does one upgrade the firmware on an iAudio?
15:12:20 Quit DJDD (Read error: 110 (Connection timed out))
15:12:29preglowlinuxstb: ipod has no partition type on the boot part?
15:13:01linuxstbSPAM WARNING: Han Soloo has registered on the Wiki.
15:13:09preglowyeah, i noticed
15:13:39linuxstbpreglow: That's right, it's type 0 (zero).
15:13:40preglowboot part of eighty megs, seems a bit hefty
15:13:57linuxstbIf you change it to any other type, your iPod won't work any more.
15:14:04preglowcan i change that if i feel adventurous?
15:14:06linuxstbAs I discovered...
15:14:49linuxstbpreglow: I'm sure you can. I would suggest doing a "dd if=/dev/sda | gzip -9 > mynano.bin.gz" to backup the whole drive.
15:15:07preglow>This is your Apple iPod. You probably do not want to boot from it!
15:15:08linuxstbYou should then be able to restore it just by dumping that back.
15:15:13preglownice message in the mbr
15:17:39linuxstbpreglow: Are you about to test the bootloader?
15:17:48preglowlinuxstb: still backing up
15:17:59preglowbut yes, that's next
15:21:08korpseJungti1234: thanks, i must've missed it when i was looking
15:22:43linuxstbHas our Wiki been taken down by a nice webmaster or by a nasty spammer?
15:23:12Jungti1234korpse: english-
15:23:12B4gdertaken down?
15:23:19linuxstbIt appears unavailable....
15:23:35linuxstbSorry - it's back now...
15:23:58B4gderI'm trying to figure out why our deny of hansolo's IP address doesn't work properly
15:24:32linuxstbSo he/she is repeatedly using the same IP address?
15:24:53B4gderright now at least he's using the same IP we deny in the general site
15:25:56preglowmy, gzip is slow
15:27:07 Join _FireFly_ [0] (
15:28:30preglowi can't connect to ipl cvs
15:28:54linuxstbDo you want anything in particular?
15:29:02preglowmake_fw, nothing much else
15:29:06 Join uncledrax [0] (
15:29:39Jungti1234How many people is Korean who take part in H100 project?
15:30:35B4gderJungti1234: exactly zero
15:31:15B4gdermaybe we should offer those ipl guys to host their cvs? B-]
15:32:32B4gderokay, I believe I've blocked HanSolo slightly better now
15:32:52B4gderI guess future will tell how well I've managed
15:35:49preglowlinuxstb: i get a bad gen error on the last make_fw line
15:35:53preglowin the wiki
15:36:06ZagorB4gder: or how many other IPs he's got :-)
15:36:18B4gderyes, funny!
15:38:21 Join Kohlrabi [0] (
15:39:43uncledraxyay, ipod rockbox!!! damn, i only have a mini 2nd gen. guess ill go out and get a nano :)
15:39:59preglowlinuxstb: i replaced -g 3 with -g nano
15:40:05preglowlinuxstb: would that be a fatal mistake? :/
15:43:05linuxstbpreglow: That should be fine. My mistake - they recently changed that make_fw.c option
15:43:30preglowok, then i'm good to go now
15:43:38preglowand btw, how do i make the ipod stop displaying Do not disconnect?
15:43:54preglowdoes it even matter?
15:45:59preglowpumounting the partition did not help
15:46:46korpsewho is hansolo?
15:47:15preglowsome spammar
15:47:25preglowi probably need some modprobe command here, no?
15:48:36korpseah alright
15:49:21Zagorpreglow: sounds like some windows/apple special hoo-daa. just yank the cable if you have unmounted.
15:49:22preglowinsmod, i mean
15:49:45preglowsome ipodlinux pages does insmod -r on the firewire driver
15:50:00preglowsure, why not
15:50:20_FireFly_preglow: rmmod should also do it to remove a module :)
15:50:26Zagorlinux does not disable device drivers while the device is connected
15:51:40Zagoralso insmod -r is quite difficult if it's compiled-in... :-)
15:51:45Jungti1234Is no there a person who is developing here unicode rockbox?
15:51:53_FireFly_afaik you can't deactivate a single device if you have several devices connected which uses the same driver
15:51:57_FireFly_under linux
15:52:02preglowok, nothing much happened at all
15:52:02_FireFly_Zagor: :)
15:52:15preglowi see the apple logo for some seconds, then the apple firmware boots
15:52:18_FireFly_preglow: simple unplug the device after unmounting
15:52:18preglowi must have done something wrong
15:52:30CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
15:52:30*Jungti1234 ߿ ƿ̽ϴ.(wait)
15:53:35ZagorJungti1234: Frank (phaedrus961) committed an update to the unicode patch yesterday
15:54:01Jungti1234Did you complete?
15:54:37Zagorsorry, not committed. submitted, to the patch tracker.
15:55:02Jungti1234Is no there a person in seat?
15:55:15Zagorin seat? you mean here in the channel?
15:55:49Jungti1234A person who can talk now.
15:55:49Zagori guess not, since nobody answered.
15:56:08Jungti1234Thank answer.
15:56:23 Quit ep0ch (Read error: 104 (Connection reset by peer))
15:56:34ZagorJungti1234: you need to sign up for evening classes in english :-)
15:57:20linuxstbpreglow: Yes, just unplug the ipod when you have finished. The "do not disconnect" message will then disappear.
15:57:22preglowno, i can't get this bootloader to work at all
15:57:26Jungti1234English does not really.
15:57:28preglowall i see is the apple logo
15:57:32preglowthen the original firmware
15:57:41Jungti1234So difficult.
15:58:08linuxstbpreglow: Try extending the usleep command in bootloader/ipod.c - change it to 15000000 (15 seconds). Maybe the bootloader is running, but not displaying anything on the LCD.
15:58:30linuxstbIt's currently set to 5000000 (5 seconds).
15:59:51preglowlinuxstb: does it reboot when i unplug usb?
16:01:16linuxstbIt should do, yes.
16:02:45preglowlinuxstb: yup, what you said is correct
16:02:49preglowlinuxstb: doens't display shit
16:04:05 Join Febs [0] (
16:08:58Zagoropen it! open it!
16:09:07preglowyou can't open these without breaking them in two
16:09:24Zagorperhaps there are enough photos on the net already
16:09:40preglowi think so, but that still doesn't help me
16:09:49preglowi think the nicer thing to do is just check out the ipl sources, heh
16:13:44FebsCan someone with more technical ability than me take a quick look at this revised FAQ and let me know if I accurately captured what Linus has reported over the last few days?
16:13:46Jungti1234Do not you become Hangul input support in unicode rock box?
16:18:17Jungti1234 - Can you reference this picture?
16:20:32Jungti1234Can no one answer?
16:22:52Jungti1234good night
16:22:59Jungti1234see you next time
16:23:44 Part Jungti1234
16:23:53 Join ryanpg [0] (
16:25:51ryanpghi all... rockbox looks awesome, as an owner of an iaudio M5 I'm hoping for a succesful port, any hackers here know if the M5 is on the list for support? I'm unsure how much it differs from the X5 or M3 which are both mentioned on
16:26:39B4gderwell, there's no work on any of them atm
16:27:34ryanpgoh... that didn't come across from reading the webpage
16:27:55ryanpgbtw, all the M3 image links broke here ->
16:28:07B4gderfeel free to fix
16:28:17B4gderwe had a wiki breakdown a while ago
16:28:23B4gderwe still suffer from it
16:29:47ryanpghmm... purchasing an M3, scanning pics... seems like a bit much for me to tackle right now
16:30:14ryanpgoh well, good luck all
16:30:18 Part ryanpg ("Leaving")
16:34:19linuxstbpreglow: Sorry, I'm busy working at the moment. But you're saying the bootloader is running, but the LCD doesn't change?
16:34:34linuxstbDoes the bootloader at least turn the backlight on?
16:56:07***Saving seen data "./dancer.seen"
16:56:52preglowseems to
16:57:11 Join LinusN [0] (
16:58:49preglowbut i think the apple one did that as well
16:59:21linuxstbBut if it turns it on before the 15 second delay, then that's a good sign.
16:59:28preglowoh yes
16:59:29preglowlong before
16:59:36linuxstbThen the bootloader is doing it.
16:59:47preglowwhat happens is, i see the apple logo
16:59:50preglowthen backlight turns on
16:59:53preglowthen delay
17:00:11preglowtakes about a sec before backlight turn son
17:00:51linuxstbThe flash firmware displays the apple logo (without backlight), then reads the contents of the boot partition into RAM, and then jumps to the entry point to transfer control to the program loaded from the boot partition.
17:01:04linuxstbi.e. the rockbox bootloader.
17:05:23preglowi'll have a look and see if i see something obvious
17:06:51 Quit uncledrax (Read error: 104 (Connection reset by peer))
17:08:36linuxstbI've just compared the IPL code and the Rockbox code, and can't see anything obvious. It _should_ work.
17:09:15preglow /* iPodLinux doesn't appear have any LCD init code for the Nano */
17:09:23linuxstbIt's true :)
17:09:25preglowperhaps it's just well hidden?
17:09:39linuxstbHave you got a copy of the IPL cvs yet?
17:09:44preglowanywho, shouldn't it be init by the apple firmware loader anyway
17:09:56Rickdoesn't the nano like, not have an LCD?
17:10:03Rickor am I think of that other one
17:10:04linuxstbpreglow: Yes, you would think so.
17:10:06preglowRick: shuffle
17:10:09LinusNthat's the shuffle
17:10:21Rickoh right, nano is the tiny scratchable one
17:10:25linuxstbThat's completely different, and iPodLinux have ignored it.
17:10:40Rickhavn't been following players much lately
17:10:53linuxstbI'm sure the shuffle would benefit from a voice UI though....
17:11:10LinusNyou tend to ignore the rest of the market when you have a rockbox player
17:11:28RickI havn't updated rockbox in like 5 months so I wouldn't say that either :P
17:11:38*Rick hides
17:11:48B4gderbad Rick, baaaad baaaad Rick
17:11:56preglowshame! shame!
17:12:25_FireFly_but "never touch a running system" ;)
17:12:27Rickmy player will probably be gathering dust for another 2 months until next semester :P
17:12:37linuxstbpreglow: The IPL kernel framebuffer driver:
17:13:05preglowstill can't login to cvs
17:13:08Rickeww, mixed code formatting, evile
17:13:18*Rick grins
17:13:42Ricklooks neat though
17:13:53linuxstbpreglow: The Nano is identified by ipod_hw_ver==0xc in that code.
17:14:01preglowyes, figured
17:14:10FebsYou know, all of the Mistic River H300 fanboys are despondent enough about the fact that the new iPod plays video ...
17:14:23Febsif you get Rockbox running on an iPod before the H300 it could send them over the edge.
17:14:25preglowand better than h3x0 ever will
17:14:40preglowit already _is_ running better on ipod than h3x0 :P
17:14:46preglowthat will last up until linus finishes the bootloader
17:15:02linuxstbpreglow: Looking in init_lcd(), there is no code executed for the nano.... 0x6 is my ipod - the color/photo
17:16:04preglowso i see
17:16:14preglowperhaps i should try the ipodlinux bootloader just for kicks
17:16:27preglowthis would be greatly facilitated by being able to use cvs
17:16:32linuxstbThe ipodlinux bootloader doesn't display anything.
17:16:41Rickthe nano's lcd is bigger than the standard and mini lcds?
17:16:44linuxstbBut you could try booting the kernel - that will display console messages.
17:17:06preglowlinuxstb: your bootloader is able to do that currently, yes?
17:17:19linuxstbThe Rockbox bootloader will load the kernel for you - find a linux.bin, put it on your FAT32 partition and enable it in my bootloader
17:17:38linuxstbpreglow: Yes - you need to change a #if 0 to enable it.
17:18:25linuxstbGet a kernel here:
17:18:37linuxstbGet the very newest one, obviously.
17:22:25 Join mashalla [0] (
17:23:00preglowipodlinux runs
17:23:04preglowup until it panics
17:23:17linuxstbUsing my bootloader?
17:23:29linuxstbCool - that confirms the ATA driver is working.
17:24:02preglowcan one get decent protective cases for these things?
17:24:09preglowit doesn't surprise me if it scratches quickly
17:25:09linuxstbpreglow: I've found the stupid mistake :)
17:25:30linuxstbCONFIG_LCD is wrong in config-ipodnano.h
17:25:53preglowi just need to remove ipodlinux now
17:25:59preglowdoesn't seem to me that disk mode works
17:26:41 Quit einhirn_ ("Miranda IM! Smaller, Faster, Easier.")
17:28:04linuxstbpreglow: It should be fixed in CVS now.
17:28:13linuxstbHave you got it in disk mode yet?
17:28:25linuxstbDoes the reboot work?
17:28:39preglowyes, sure
17:28:45preglowbut the menu and play thing doesn't
17:28:53linuxstbMaybe it's different for the Nano...
17:29:09preglowlet's just hope the mechanism exists
17:29:11preglowif not, i'm fucked
17:29:45linuxstb - scroll down to "nano and Button combinations"
17:30:10preglowkind of hard to make a booting linux look like a manufacturer error ;)
17:30:27linuxstbIt should be select+play - not menu and play
17:30:35linuxstbWhat did I write on the wiki?
17:30:47preglowmenu and aply
17:30:57linuxstbMy mistake again then...
17:31:18preglowhad me worried for a sec there, heh
17:31:28linuxstbIf you update the firmware directory from CVS, rebuild the bootloader and reinstall, it should work....
17:32:34preglowlinuxstb: ok, it works, but the picture is very garbled
17:32:54preglowi think i can see what's wrong, though
17:33:11linuxstbIt the text OK, or is that garbled as well?
17:33:26preglowand the graphics stops 4/5 at the side of the screen
17:33:33preglowand it seems like the drawing overflows onto the next line
17:33:45preglowso i guess you've just got erronous line lengths somewhere
17:36:37_FireFly_can i assume, if the target has no BITMAP-lcd that there is no remote ??
17:37:43B4gderif you want a remote with lcd, check HAVE_REMOTE_LCD
17:39:57_FireFly_hmm it seams that the dimension of the remote-lcd is on all target(which has an remote width lcd) the same
17:40:19B4gderright now, yes
17:41:09_FireFly_then can i use my created default wps for all remote-lcd
17:42:10 Quit B4gder ("time to say moo")
17:43:03_FireFly_but i can assume that if HAVE_BITMAP_LCD is set that then the remote has also an lcd
17:47:47 Join Mongey| [0] (
17:48:35Mongey|anyone here
17:49:22Mongey|i better leave if thers not
17:50:13Mongey|LinusN; are you on
17:50:28preglowif you'v got someting to ask, just ask it
17:50:33preglowyou've, yes
17:51:07Mongey|no just here to concradulat linus on his h300 work
17:51:51Mongey|well i do have 1 question
17:52:01linuxstbpreglow: Any ideas about your LCD?
17:52:29*Mongey| is burning a suse iso now
17:52:48preglowlinuxstb: looking through code now
17:53:15linuxstbThere are obviously two potential problems - the writing into lcd_framebuffer and the lcd_update() function. You could try writing a pattern into lcd_framebuffer to see if that gets displayed properly.
17:58:04 Part mashalla
17:59:39 Join mashalla [0] (
18:02:34Mongey|well thats one cd wasted
18:02:58_FireFly_Mongey|: ??
18:05:03linuxstbpreglow: What is weird?
18:05:24preglowi'm trying to see a pattern in how it draws wrong
18:06:09Mongey|i was trying to burn a live cd thing of linux; Suse and i did it wrong cause i dont know what to do
18:06:17preglowfirst, the entire display seems to be shifted
18:06:36preglowi try to write at position zero, and the pixel appears in the middle of the screen
18:07:03_FireFly_funny ;)
18:07:14preglowand the position overflows, when i try to write a straight vertical line, there are several lines spaced by several pixels
18:07:44preglowabout 1/4 of the right part of the screen is missing
18:07:46linuxstbHow are you drawing? Don't use the build-in lcd_???? functions.
18:07:59preglowlinuxstb: i write directly to the framebuffer
18:08:40preglowfor (i = 0; i < LCD_HEIGHT; ++i) { lcd_framebuffer[i][0] = 0xff; lcd_framebuffer[i][1] = 0xff; }
18:08:43preglowthen update
18:08:48Mongey|Anyone know what ive to burn on for the live cd
18:08:53 Nick ender1 is now known as ender` (i=ychat@
18:09:00preglowi shouldn't use update?
18:09:31linuxstbYes - use lcd_update(). But most of the other functions are not implemented yet. Only the ones used by the font rendering
18:09:45preglowfigured that
18:09:59preglowi see what i do is correct, but the stride and offset seems to be completely off the bat
18:09:59Mongey|good bye
18:10:32_FireFly_wrong slash ;)
18:10:36 Part Mongey|
18:11:14 Quit Febs ("CGI:IRC (EOF)")
18:11:47 Join Febs [0] (
18:11:59linuxstbDoes lcd_clear_display() followed by lcd_update() display correctly? You should get a screen full of blue.
18:12:31preglowentire screen is init to blue before i start drawing at least
18:13:47_FireFly_btw. about slash and back-slash ;)
18:16:07linuxstbThis could be related to the old problem I used to have where the screen would be corrupted for no apparent reason. Changing unrelated code in the bootloader would fix it.
18:16:36linuxstbI suddenly stopped experiencing this, but I never did find out the real reason.
18:17:00linuxstbI think I talked about it here a few weeks ago.
18:17:01preglowyou don't remember the nature of the corruption?
18:17:37linuxstbRows of pixels seemed to be duplicated - i.e. displaying in their proper position, and then again about 4 rows further down.
18:18:23preglowthis might perhaps look like that
18:18:34preglowi would take a photo, but i'm in linux, and my camera needs a windows program :/
18:18:43linuxstbgphoto2 ?
18:19:00linuxstbIs it debian?
18:19:19preglowbut my camera uses some proprietary transfer stuff
18:19:45preglowwell, there you go
18:19:48preglowit supports my camera :P
18:19:50linuxstbJust try making the bootloader do different things - comment out some of the lcd_puts and sprintf lines.
18:22:54preglowgphoto2 works like a charm
18:24:49_FireFly_nice ;)
18:24:54linuxstbThat looks very different to my corruption though.
18:25:23korpsenice one, preglow
18:25:29preglownice and nice...
18:25:39preglowwould be nicer with a couple of intelligeble logos
18:25:40linuxstbWeird that only about 3/4 of the width contains anything.
18:26:12 Quit ender` (Read error: 104 (Connection reset by peer))
18:26:13_FireFly_maybe the framebuffer doesn't cover the whole screen-width
18:26:18preglowit does
18:26:58_FireFly_it doesn't seams that :)
18:26:59korpsethere is some shape
18:27:02linuxstbCan you compare the update() functions in fb.c from IPL and in Rockbox. Hopefully the mistake is there.
18:27:14preglowtry some other stuff first
18:27:24preglowipl update does the same thing for photo and nano
18:27:33preglowso i don't see why it shouldn't work for me if it works for you
18:28:02linuxstbThat's assuming that lcd_type is correctly set to 1 (it should be). You could try manually setting lcd_type=1 somewhere in lcd-16bit.c
18:28:07preglowi tried that
18:28:09preglowdidn't work
18:28:26preglowin the code where you set lcd_type, i just edited all the lines to say lcd_type = 1
18:28:27_FireFly_is the screen-width of the photo bigger or the same as the nano ??
18:28:47linuxstbpreglow: But I think that code is only compiled for the color/photo
18:29:03_FireFly_or the 1/4 smaller ??
18:29:06preglowin that case lcd_type should be set to 1 anyway
18:29:15linuxstb_should be_
18:29:47linuxstb_FireFly_: From memory, color/photo is 220x176, the Nano is 176x132
18:32:06preglowi drew a line from x = 0 to x = max
18:32:08preglowat h = 0
18:32:14preglowjust a small bit of it appeared
18:32:28 Join justsomeperson [0] (
18:34:00_FireFly_why are these function doesn't "exported" through playback.h although these function are used in wps ??
18:34:01_FireFly_void audio_next_dir(void);
18:34:01_FireFly_void audio_prev_dir(void);
18:34:30_FireFly_has it somethinng to do with thw code-size ?
18:36:02linuxstbpreglow: I think I've found the problem again.
18:36:14preglowspill it
18:36:33linuxstbLook in fb.c - the ipod_update_photo() function.
18:37:00linuxstbThe code to calculate the drawing region is different for the Nano - but I didn't copy that code.
18:37:16preglowsmall wonder, then
18:37:18preglowi'll deal with it
18:37:42preglowof course
18:37:56preglowthis has to be it
18:38:00preglowexplains the error just nicely
18:38:18linuxstbLet me just check the rest of that function.
18:39:22linuxstbYes - there's another check later in that function for ipod_hw_ver==0x6 which I've removed.
18:39:49linuxstbYou should remove the line "rect2=rect4;" for Nano builds.
18:40:19linuxstbBut that's it - two places.
18:43:19_FireFly_gonna to eat something
18:43:31 Join ender` [0] (i=ychat@
18:48:04linuxstbpreglow: Any luck?
18:48:06preglowi swear, this usb plug feels like it's going to die on me tomorrow
18:52:51linuxstbA nice sight.
18:53:00preglowi'll commit the changes
18:56:10***Saving seen data "./dancer.seen"
18:57:59linuxstbpreglow: That should be LCD_IPODCOLOR....
18:58:20preglowi forgot to even verify that, haha
18:58:25 Join Mindship-02 [0] (
18:58:27preglowi'll fix
18:59:04Mindship-02Is there a list of features and requests? And can I send in one?
18:59:24linuxstb - check the menu links on the left side of the screen.
19:00:08Mindship-02linuxstb: sorry, I feel so stupid, I just didn't see it!
19:00:08Mindship-02Can Rockbox record, and can it record from the radio?
19:00:18linuxstbYes and Yes.
19:00:33linuxstbBut on the iriver, the recording is still experiemental.
19:00:50 Join hardeep [0] (
19:01:15Mindship-02ah, okay
19:01:41Mindship-02the whole Rockbox idea is quite perfect
19:01:54hardeepTiMiD: around?
19:01:56_FireFly_what about my question about the two functions any comments ??
19:02:14Mindship-02is it developped purely by educated programmers or too by hobby-ing people?
19:02:33linuxstbMost people are both.
19:03:13linuxstbBut we all do it for fun in our spare time.
19:03:40TiMiDhardeep: yes ?
19:03:50_FireFly_i'm currently studying "Informatik" computer-science :)
19:04:30Mindship-02Is anyone working on request 955078? I want to have (text file containing my name and adress) to be showed on boot...
19:04:57Mindship-02Since I have this device two more persons bought it. I want everyone to know which one is mine...
19:05:01hardeep_FireFly_: Yeah, that's not correct. Both those functions should be in audio.h just like audio_next_track()
19:05:22hardeepTiMiD: Are you working on porting the playlist viewer to the new guilist?
19:05:44_FireFly_yepp he does
19:05:47hardeepTiMiD: I was going to look into it but didn't want to duplicate the work if you're already working on it =)
19:05:58linuxstbMindship-02: I don't think so. On the iriver, changing the bootloader is very dangerous - so it's not likely to have any "luxury" features like that one.
19:06:00TiMiDI'm doing it yes
19:06:05hardeepokay, cool
19:06:09TiMiDare you the one who wrote the code ?
19:06:14TiMiDthe initial code ?
19:06:52TiMiDJust that I have some problems to understant exactly how it works ...
19:07:15Mindship-02linuxstb: is it dangerous to compile the bootloader? somewhere the Rockbox splash must be coded in!
19:07:41LinusNthe splash isn't in the boot loader
19:07:46TiMiDwhy to store 2 lists in the same array for example (If I understand correctly)
19:07:47hardeepTiMiD: anything I can help with?
19:08:03hardeep2 lists?
19:08:16Mindship-02LinusN: so what about the words of linuxstb?
19:08:25linuxstbMindship-02: Ignore my answer. The Rockbox splash is displayed by Rockbox itself, not the bootloader. So yes, in theory that screen could be easily changed to display anything.
19:08:35linuxstbMindship-02: I was wrong.
19:08:51TiMiDI mean the first_index
19:09:05Mindship-02ah, now it went from "not likely" to "easily"!
19:09:06TiMiDmut maybe I didn't understood it at all
19:09:21hardeepdo you mean the first index versus first display index?
19:10:04TiMiDjust the first index
19:10:08linuxstbMindship-02: But it still needs a developer to do it, and it might not be considered an acceptable patch for general Rockbox use.
19:10:13TiMiDI don't understand at all :/
19:10:54Mindship-02linuxstb: how are you with your time, feel like guiding (teaching) me with this little project? :-P
19:10:55hardeepthe first index in the viewer is used to identify the first entry in the playlist that's currently loaded in the viewer
19:11:10_FireFly_hardeep: i have moved the two function definition to audio.h i
19:11:16hardeepthe viewer doesn't load the entire playlist to save memory
19:11:40TiMiDand the rest of the playlist is stored where ?
19:11:42hardeepwhenever we go beyond the first or end index, we load more information
19:11:43_FireFly_so i have a problem less to worry about it by my try to make a wps-widget ;)
19:11:53linuxstbMindship-02: Feel free to ask questions here. Most of the time somebody will answer.
19:12:06hardeepwe get all the info from the playlist code (playlist.c)
19:12:25LinusNMindship-02: begin with setting up a development environment
19:12:53Mindship-02LinusN: any example DevEnv-s?
19:13:03hardeephowever, the actual track names are stored on disk and we only read on demand (the playlist code saves the seek point in it's table)
19:13:42TiMiDso each time you move inside your list, you read from disk oO
19:14:00TiMiDor is the entire playlist loaded into memory ?
19:14:03hardeepeach time you go beyond what's loaded
19:14:15TiMiDI understand
19:14:23hardeepwe limit the amount we load to a max of 200 entries or when the buffer is full
19:14:47TiMiDI wouldn't have understood that by myself :)
19:14:51*amiconn appears, reading up logs...
19:14:57Mindship-02LinusN: Thanks for the link, I'll study that first! :-)
19:15:02hardeepon the archos devices we don't have enough memory to cache all the playlist track names in memory
19:15:18hardeepon the iriver it may be possible
19:15:21TiMiDI wrote some code, it seems to work, excepted for some things like remove :/
19:15:52hardeepwhat happens with remove?
19:16:10TiMiDstrange things :)
19:16:14hardeepto the viewer, remove is just a call to the playlist code to remove the file... after which we just need to update the display
19:16:25TiMiDto remove the file ?
19:16:33hardeepfrom the playlist
19:16:36TiMiDI thought it was just to remove the entry in the playlist
19:16:54hardeepsorry, yeah, that's what it does
19:17:08amiconnhardeep: I'd rather not cache all names in memory even on iriver
19:18:06TiMiDif it's your code you could perhaps do it with my list :)
19:18:06 Quit Febs ("CGI:IRC (EOF)")
19:18:18 Join Febbs [0] (
19:18:47TiMiDsince I have some problems to understand your code :/
19:19:00hardeepamiconn: i like it the way it is too... i just saw some people complaining about the disk access on the viewer
19:19:28amiconnThink of a 20000-entry playlist with 100+ chars per entry
19:19:38amiconn> 2 MB
19:19:41hardeepTiMiD: heh, i guess my documentation isn't good enough =)
19:19:46TiMiDalso there is a load_playlist_entries and load_playlist_entries_r
19:20:07hardeepTiMiD: yeah, one loads from the top down and the other from the bottom up
19:20:15hardeepdepending on which way we're moving
19:20:18 Join dpassen1 [0] (n=dpassen1@resnet-233-61.resnet.UMBC.EDU)
19:20:22TiMiDyou could have done it in one fn no ?
19:20:36TiMiDjust loading from pos-200 or pos
19:21:07hardeepexcept if we are unable to load 200 tracks
19:21:16hardeepsay the buffer can only hold 100
19:21:22hardeepit depends on the size of the track names
19:21:52_FireFly_maybe this kind of offset could be calculated
19:21:56TiMiDmore complicated than I expected
19:22:26hardeep_FireFly_: not without actually reading the track names =)
19:22:29 Quit Mindship-02 ("shutdown")
19:23:26TiMiDI will try to do it now since I have more informations
19:23:27_FireFly_but how do you get the right parameter for the two function??
19:24:21hardeepwhich two functions?
19:24:33_FireFly_load_playlist_entries and load_playlist_entries_r ;)
19:25:18hardeepif you're scrolling down the list then we use load_playlist_entries() and pass the first track currently displayed
19:25:36hardeepif you're scrolling up then we use load_playlist_entries_r() and use the last track currently being displayed
19:25:46_FireFly_maybe we could join these two function to one with two parameters the secend could tell the function if the value of the first parameter gives the start index or the end index
19:27:05hardeepit's possible to join the two together, it just seemed easier to keep them separate
19:27:24amiconnMaybe a parameter that indicates the search direction?
19:27:34_FireFly_that was my secend parameter ;)
19:27:35TiMiDI think I will rewrite some code here since now there isn't drawings, some fn looks a little useless
19:37:43preglowoh well
19:37:51preglowlinuxstb: what's next on your list+
19:39:26linuxstbA proper interrupt-driven button driver and also implementing current_tick via a timer interrupt.
19:39:32linuxstbi.e. interrupts :)
19:39:50linuxstbBut this is something I'm not very familiar with.
19:40:48linuxstbThe other big jobs before Rockbox itself is anywhere near bootable is thread context switching and writing the startup code for crt0.S
19:41:24linuxstbOnce those are working, Rockbox should be close to working.
19:43:00preglowthread context switching should be pretty easy
19:43:02linuxstbThere is some basic button detection code in the ipodlinux bootloader, but when I tried it a while ago, it wasn't very reliable at detecting a keypress on startup.
19:43:22linuxstbThe button driver from the kernel is the place to take the code from.
19:43:42linuxstbHow is your knowledge of ARM assembler?
19:44:17linuxstbThere should hopefully be a #warning in all the places we need to write code.
19:45:19preglowmy knowledge of arm assembler is ok, but far from complete
19:45:26preglowhoping to improve, though
19:48:54linuxstbMy knowledge is still almost zero.
19:51:22linuxstbDo you have any plans?
19:53:20preglowwhatever needs to be done
19:53:39preglowthat is, i plan to test ipl, if that's possible
19:55:37linuxstbGood luck with ipl. Back later.
19:55:41 Quit linuxstb ("Client Exiting")
20:01:03preglowneed to partition
20:01:09preglowwell, i'm not jumping through hoops for it yet
20:02:01 Quit hardeep (" Like's GUI? Then try HydraIRC -> <-")
20:03:02 Quit justsomeperson ("CGI:IRC")
20:08:19preglowmy, is this thing thin
20:08:32 Join arkascha [0] (
20:10:13 Part LinusN
20:28:29amiconn_FireFly_: You can't assume whether the remote has an lcd or not based on what the main lcd is
20:29:24amiconnThe archos remote has no lcd, and is usable for both the player (charcell main lcd) and recorder v1 (bitmap lcd)
20:31:34_FireFly_k but the code for the remote standard wps have i ifdefed with checking NB_SCREENS = 2 so in this case i can assume that the lcd in the remote is a bitmap one ;)
20:32:01_FireFly_because my search on the <target>-config.h showed me that when a remote-lcd exist then it is a bitmap one
20:36:11amiconnI would use #ifdef HAVE_REMOTE_LCD
20:36:27amiconnBut yes, all (current) remote lcds are bitmapped
20:37:11amiconnIn fact there is only one remote lcd type so far, so there is no CONFIG_REMOTE_LCD
20:40:48Febbspreglow: as predicted, the fanboy hysteria begins: (post 1651)
20:40:57 Nick Febbs is now known as Febs (
20:41:28 Join hardeep [0] (i=hardeeps@norge.freeshell.ORG)
20:42:50amiconnpreglow: Congrats for your successful ipod nano bootloader run
20:44:48HClipod nano bootloader
20:44:50HClwow i missed a lot
20:49:59TiMiD_FireFly_: just use NB_SCREENS since it could be set to 1 even with remote plugged for debug purpose for example
20:51:04_FireFly_yepp i use it :)
20:53:49_FireFly_i have it mostly done the widget :)
20:55:24preglowamiconn: all credits go out to linuxstb for that one
20:56:12amiconnTiMiD: I found a way how to make the binary smaller on all platforms where NB_SCREENS == 1
20:56:14***Saving seen data "./dancer.seen"
20:56:31TiMiD#ifdef instead of for ?
20:57:44amiconnUsing #ifdefs everywhere will probably be rather ugly, I have a better suggestion instead :)
20:58:00preglowFebs: what post would that be? i can't find any numbering
20:58:29amiconnA macro like FOR_ALL_SCREENS(i) , i.e. taking the loop variable as an argument.
20:58:41amiconnFor NB_SCREENS > 1, it would expand to:
20:58:53amiconnfor (i = 0; i < NB_SCREENS; i++)
20:59:12amiconnbut when NB_SCREENS == 1, it would simply expand to:
20:59:16amiconni = 0;
20:59:35amiconnSimple and elegant, imho
21:01:18amiconnI tried this replacement (manually) in one single place - it saved 48 bytes on archos.
21:01:42amiconnNow, how many occurences of for (i = 0; i < NB_SCREENS; i++) are there in your code? :)
21:04:36_FireFly_i think many ;)
21:06:49Febspreglow: it's the last post in that thread. There is no actual hysteria yet, but give it time ...
21:07:37FebsWhoops. Now second-to-last, by djray.
21:09:24 Quit Febs ("CGI:IRC")
21:09:39 Join Febs [0] (
21:12:51*Febs must learn not to hit the "back" button while using the web-based IRC client.
21:19:40 Quit Mxm`Pas`Bien (Read error: 110 (Connection timed out))
21:19:41TiMiDpretty good :)
21:19:52TiMiDthe macro idea
21:32:32 Join Acksaw [0] (
21:37:58 Join RotAtoR [0] (
21:38:53amiconnTiMiD: grep tells me this for() occurs 30 times. I want those 1440 bytes back :)
21:39:36amiconnTiMiD: Do you want me have a go at it?
21:40:52amiconnBetter call it FOR_NB_SCREENS(i)
21:41:21hardeepi'm surprised the compiler doesn't catch this
21:43:08 Quit bbad ("KVIrc 3.2.0 'Realia'")
21:44:28XavierGrhardeep: I was the one that made the suggestion. I said that having in mind that if the dircache is <400kb then a playlist can't be that big. Also you metnion that 200 entries are stored on ram, in my test yesterady I remember disk activity every time I entered the playlist viewer...
21:44:51 Join manuPsi [0] (n=poupet@
21:45:21XavierGralso yesterday I intentionally made my player to play as long as it could reaching a 2.73V level, isn't that exhasting for the battery?
21:45:36XavierGrwhat is the limit for li-ion?
21:45:45hardeepXavierGr: Yeah, that happens because we need to read the playlist to get the tracknames
21:45:56 Part manuPsi
21:46:37XavierGrso unless we get to the viewer nothing is stored on ram, right?
21:46:55hardeepand it's cleared as soon as we exit the viewer
21:47:11hardeep(we share the buffer with the plugins)
21:47:58XavierGrthat could be a problem if I tried to change that for my self.
21:49:22hardeepyeah =) the purpose of the 200-track cache in the viewer is just to avoid constantly accessing the disk while in the viewer. there's no global place where track names are stored in memory except when you're playing from disk
21:49:55hardeepi guess we could try and use the in_ram buffer for caching as much as we can
21:50:42XavierGrthat would be very good IMHO, I think that rockbox should ty to prevent any sort of disk activity more than once
21:51:04XavierGrbut if this trend continues I can understand that we will run out of buffer for theaudio
21:51:31XavierGrat least it is safe and doable for targets that have large ram.
21:52:18 Join Moos [0] (
21:55:12Acksawim off
21:55:15Acksawsee ya guys later
21:55:51 Quit Acksaw ()
22:00:31 Quit Kohlrabi (Nick collision from services.)
22:00:34 Join Kohlrabi [0] (
22:01:32*ehntoo is back
22:02:55amiconnhardeep: Hmm, some questions and thoughts:
22:04:00amiconn(1) If the playlist viewer uses the plugin ram to buffer track names, would it be possible to just extend the maximum line count in the buffer? The plugin buffer is 768KB instead of 32KB
22:04:05amiconn...on iriver
22:05:24amiconn(2) Does the playlist viewer pfn_tsr_callback() before using the buffer? Otherwise rockbox might crash badly if a tsr plugin is active and the playlist viewer is called
22:06:01amiconnpfn_tsr_exit() I mean
22:07:28 Join Maxime [0] (
22:08:27 Join linuxstb [0] (
22:11:03hardeepamiconn: 1. yeah, this is definitely doable. I'm going to look into this
22:11:44hardeep2. didn't know about this. will fix this
22:17:18hardeepfor 2, i guess we'll need to add a new api to the plugin code to force a tsr to exit
22:22:48amiconnThat's exactly what this callback is for
22:23:25amiconnA tsr plugin calls plugin_tsr() to register this callback before returning
22:23:39amiconnThe plugin loader calls it before loading a new plugin
22:24:31amiconnThis mechanism was implemented by [IDC]Dragon. I'm sure he would have added this handling to the playlist viewer code as well - if he would have known that the playlist viewer reuses the plugin buffer
22:24:58amiconnI didn't know this as well until today...
22:30:27hardeepyeah, the playlist viewer is the only non-plugin that makes use of the plugin buffer so it's easy to miss
22:30:57XavierGrwhy not change that?
22:31:10XavierGrwhy not use the same memory as the dircache?
22:31:22hardeepa couple of reasons
22:31:29hardeep1. there is no dircache on the archos devices
22:31:46hardeep2. this would overwrite the dircache
22:31:55hardeepactually, maybe 2 isn't an issue
22:32:03XavierGr1 is an issue though
22:32:17amiconn3. the dircache can be disabled
22:32:40XavierGrhow much memory is the plugin buffer for archos ?
22:33:06 Join Febbs [0] (
22:33:06 Quit Febs ("CGI:IRC (EOF)")
22:33:10 Join Philip_0729 [0] (
22:33:15hardeepamiconn: are there any tsr plugins currently?
22:33:21preglowi wonder if we should deal with this ipod clickwheel thing
22:33:29*Febbs STILL needs to learn not to hit the back button when using the web IRC client.
22:33:31amiconnYes, exactly one, and only for archos so far
22:33:37hardeepwhich one?
22:33:41 Nick Febbs is now known as Febs (
22:33:43amiconn[IDC]Dragon's alpine_cdc
22:34:15hardeepah, okay
22:34:17amiconnPerhaps we'll have battery_log soon
22:34:21 Quit _FireFly_ ("Leaving")
22:34:25XavierGrI am just prepearing another one for battery benchmarks, though I don't know if Ican manage to finish it
22:34:43XavierGramiconn: are you working on it?
22:34:52amiconnBut thinking about it this would mean that using the playlist viewer would stop battery logging
22:35:03XavierGryes indeed.
22:35:24XavierGrbut anyway if a true benchmark is done the playlist viewer shouldn't be loaded.
22:35:44XavierGr(by the user)
22:35:45amiconnXavierGr: Not yet, but it should be quite easy, perhaps except the part that monitors disk activity and takes the chance to flush the log when the disk is spinning
22:35:59XavierGryes that my concern
22:36:10XavierGrtoday I manged to set the thread correctly
22:36:12hardeepactually, instead of killing the tsr, we just need to make sure the plugin buffer size doesn't get reset to 0
22:36:12amiconnI also did some tsr experiments in the past
22:36:29XavierGrbut right now the thread is empty and I need to think something for it
22:36:36hardeepthen the playlist viewer will only be using the buffer not being used by the tsr
22:36:42amiconn...logging thread round-trip to get an idea how fast rockbox thread scheduling is, and other things
22:37:14amiconnhardeep: Ah yes, that sounds like a good option.
22:37:29amiconntsr plugins should usually be small
22:37:38 Join _FireFly_ [0] (
22:37:47XavierGrso are you working this right now, or soon, because if you do then I will have to drop my tries on that?
22:38:07amiconnRight now I'm going to try my FOR_NB_SCREENS macro idea
22:38:14amiconnShouldn't take long...
22:38:24XavierGrI am no match for you, though I still do it to learn more on programming and rockbox internals
22:38:30*amiconn starts reading gcc docs on macros
22:41:48linuxstbamiconn: I like your macro idea, but don't you think it obfuscates the code a little?
22:41:58 Quit arkascha ("Konversation terminated!")
22:43:06amiconnWhat else would you suggest?
22:44:41 Join DangerousDan [0] (
22:45:01Philip_0729Hi everyone :), just read recent CVS and noticed some activety with H300
22:46:39preglowyes, linus was on the prowl yesterday
22:47:06 Quit _FireFly_ ("Leaving")
22:49:18Philip_0729do we have a more accurate idea of when it will be ready for us less-technically minded ??
22:49:49preglowwhen he finishes the bootloader, things should happen pretty quickly
22:50:07Philip_0729but its close now?
22:50:57preglowi have no idea
22:51:01preglowdon't hold your breath
22:51:04preglowlinus has a lot on his plate
22:51:20Philip_0729needs a bigger plate.....
22:51:33preglowi think he feels he's got a plate big enough
22:51:56preglowbut if you've got some way of extending his day beyond 24 hours, he might want to talk to you about that
22:52:13Philip_0729well there is this one thing.....
22:52:13preglowi would too, as a matter of fact
22:52:22Philip_0729but nah
22:52:29Philip_0729i'll just wait
22:52:33preglowi can't code on speed :>
22:52:40linuxstbamiconn: I have no better suggestion, just something to think about. Maybe the meaning of the macro will be obvious when it is seen in context.
22:53:22Philip_0729i just can't code.. :(
22:53:26amiconnOkay I just did it. The real gain is a little less than expected, 896 bytes on archos
22:53:48amiconnIt doesn't look very obfuscated to me just an arbitrary example:
22:53:58amiconn FOR_NB_SCREENS(i)
22:53:58amiconn {
22:53:58amiconn gui_textarea_clear(&screens[i]);
22:53:58DBUGEnqueued KICK amiconn
22:53:58amiconn }
22:55:03 Quit dpassen1 (Read error: 110 (Connection timed out))
22:56:07preglowsounds decent
22:56:20preglowcan't think of a better solution than that
22:56:21***Saving seen data "./dancer.seen"
22:56:21linuxstbThe only alternative would be to include NB_SCREENS as a parameter - e.g. FOR_LOOP(i,NB_SCREENS) but I think I prefer your version.
22:56:54amiconnHmm, in fact that's a good idea too
22:57:23amiconnIt would allow to automatically strip single-cycle for loops
22:57:42linuxstbOr FOR_LOOP(i,0,NB_SCREENS)
22:57:43amiconn*arbitrary single-cycle for loops
22:58:54amiconnThe question is where such a universal macro should reside
22:59:11 Join _FireFly_ [0] (
23:01:33 Join Midgey34 [0] (
23:01:35Bagdergosh, a new codec
23:02:03Bagderpatch tracker
23:02:14linuxstbMmm. Seeking hell
23:02:18linuxstbBut a nice addition.
23:02:40linuxstbI hope they chose a decoder with a compatible license.
23:02:49Bagder"from the ffmpeg project"
23:03:00linuxstbLGPL IIRC
23:04:07preglowso, who'll deal with that patch?
23:04:25linuxstbI'm happy to.
23:04:26CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
23:04:26*Bagder steps backwards and whistles
23:04:39preglowlinuxstb: btw, have you got any reports on doing dealing with ipods through cygwin dd?
23:04:50XavierGrwhich patch is it I can't find that number?
23:05:00preglowXavierGr: the patches page isn't updated yet
23:05:07linuxstbpreglow: No. All I know is what amiconn told me the other day.
23:05:18XavierGrwhich codec?
23:05:21preglowXavierGr: shorten
23:05:24preglowXavierGr: lossless thingie
23:05:31XavierGrnever heard of it
23:05:35preglowwell, i have
23:05:41preglowthe more the merrier, if you ask me
23:05:48XavierGrof course
23:05:54 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
23:06:07*XavierGr wonders when someone will start midi coding again! :)
23:06:46amiconnpreglow: I know cygwin dd is working well, I used that several times to make images of the Ondio flash disk for analysis
23:06:59preglowamiconn: how do i determine device names?
23:07:15ender`isn't shorten kinda like zip - inefficient, it's missing many features modern formats have (tagging), but it's sticking around because everybody else (well, at least among traders) uses it?
23:07:30preglowender`: probably
23:07:52linuxstbYes, a perfect description of shorten.
23:08:20linuxstbI download quite a lot of shorten files, but I always immediately convert them to flac and delete the originals.
23:08:57preglowlinuxstb: any thoughts on how we're going to deal with the ipod buttons?
23:09:29preglowjust pretend an increment in wheel position means left or right button press?
23:09:54amiconnpreglow: Open "computer management", go to the "device manager" In the lower main pane, all disk and cd-rom devices are listed
23:10:08amiconnEach disk device has a number associated with it.
23:10:38amiconnDevice 0 equals /dev/sda in cygwin, device 1 equals /dev/sdb etc.
23:11:02preglowahh, right
23:11:05_FireFly_argh my wps widget crashes rb on target but not in the sim
23:11:09amiconnI'm not 100% sure about the windows names, german windows here
23:11:44preglowi wonder if they've got an equivalent of ssh-agent on windows
23:11:58preglowthe thing i love the most about developing on another box, is all the passwords
23:12:01amiconnI used that with sf cvs
23:12:39preglowcygwin component?
23:12:40_FireFly_preglow: you can use the pagent from putty
23:12:41amiconncygwin does have the ssh agent
23:12:49preglow_FireFly_: i tried that once, and it flaming didn't work
23:13:02preglowit used some kind of wise-ass key format
23:13:04Bagdercygwin has openssh and its tools
23:13:17_FireFly_you must convert the ssh-private key to putty-key
23:13:21preglowor perhaps i should just start developing with linux
23:13:23_FireFly_that it works
23:13:29XavierGr_FireFly_: It is usually the other way around for me
23:13:37XavierGrSim crashes targets go on...
23:14:20amiconnXavierGr: I'm not sure whether tsr plugins will work in the sim. Nobody tried that so far
23:14:21XavierGrI mean when I test my stuff...
23:14:49XavierGramiconn and emoty thread that just exits works like a charm
23:15:00_FireFly_i think I0C is a segmentation fault am i right ??
23:15:06linuxstbpreglow: The keypad is going to be interesting. If we keep the scrollwheel functionality, then we only have up/down (or left/right - but not at the same time), and the MENU/NEXT/PREV/PLAY keys
23:15:13XavierGrbut in my previous attempt a while loop will run for some time and then eventually crash
23:15:22XavierGrbut maybe I have something wrong there.
23:15:25amiconnA thread function should never ever exit the normal way
23:15:39amiconnThis will confuse the rockbox scheduler
23:15:51XavierGryes this was a test to see what made the sim crash
23:15:51preglowlinuxstb: i think we need to define another set of keys
23:15:58preglowlinuxstb: just called BUTTON_INC, BUTTON_DEC or something
23:16:21preglowlinuxstb: which every key handler can handle as fits best to the context
23:16:58preglowand we can alias those onto existing keys for platforms that don't have them, so we don't waste any space on those
23:17:03linuxstbSounds good.
23:17:20XavierGramiconn: actually it is a thread with a while that in the first run calls gTread.exiting and the exits the thread.
23:17:21preglowbut i need some input from gurus on this, i really have _no_ idea aboyt how these parts of rockbox work
23:17:27preglowlinuxstb: i think i should also be able to handle threading
23:18:28amiconnWhy invent new button definitions? Do the ipods have ordinary up/down buttons in addition to th ewheel?
23:18:46amiconnOtherwise the wheel could just use BUTTON_UP/BUTTON_DOWN
23:19:12amiconnIt would be possible to use different names, but this will mean more #ifdefs in the code
23:19:12preglowamiconn: no
23:19:31preglowamiconn: yes, but the wheel might mean left/right in one screen and up/down in another
23:19:38amiconnYes, and?
23:19:38preglowamiconn: it depends on context what they should mean
23:19:53amiconnBUTTON_LEFT/BUTTON_RIGHT would be undefined
23:20:10preglowi keep forgetting this is done per platform
23:20:20amiconn...and most major places use button aliases
23:20:23preglowbut it pretty much has to
23:20:36amiconn#define TREE_UP BUTTON_UP
23:20:48amiconn...depending on the keypad
23:21:39 Part Philip_0729
23:22:17 Quit mashalla ()
23:26:39XavierGractually amiconn I think you were right
23:27:01linuxstbI haven't tried it yet, but the Shorten patch looks good.
23:28:26XavierGrSim does not accept a thread that will run together with other threads
23:28:49XavierGrI just test it on target and it handled quite well an the new thread.
23:30:22amiconnlinuxstb: Hmm, it seems a general FOR_LOOP macro isn't possible :(
23:31:09amiconnIt would require preprocessor conditionals within the macro body. Afaik this doesn't work
23:33:05preglowi think the preprocessor does several passes
23:33:11preglowbut that might just be for macro expansion
23:33:41Bagderyou could also simply do #ifdef define macro #else define macro #endif
23:33:51Bagderinstead of the other way around
23:34:06amiconnNot for linuxstb's idea
23:34:23Bagderdidn't pay attention to the details
23:34:23amiconnThe macro body would have to depend on the macro arguments
23:34:39amiconnThe idea was to have FOR_LOOP(a, b, c)
23:35:00preglowlinuxstb: what do you think about just renaming it to libffmpeg?
23:35:16amiconnwhich should expand to for(a = b; a < c; a++) if c - b > 1
23:35:28preglowthat wont work, no
23:35:32amiconnand simply to a = b; if c - b == 1
23:35:48linuxstbpreglow: It would be nice to rename it, but it makes a mess in CVS.
23:35:57preglowBagder: subversion!
23:36:08*Bagder ducks
23:36:14amiconnThe simple idea works
23:36:38amiconnBagder: The original idea is to reduce code size for platforms with NB_SCREENS == 1
23:37:26linuxstbThere is one side-effect of not running a for loop - i will be equal to 0 at the end of the loop, not equal to NB_SCREENS.
23:37:29amiconnI regained 896 bytes on archos recorder
23:37:53preglowlinuxstb: i think i'll just start filling in the lcd driver blanks
23:38:09amiconnI think I should commit it
23:38:50linuxstbpreglow: OK. Linus seems to be wanting to use lcd-16bit.c for the H3x0. So he suggested moving the ipod specific parts into lcd-ipod.c
23:39:02preglowso do i
23:39:23linuxstbSo you could do that - it will only be 2 or 3 functions in lcd-ipod.c
23:42:57 Quit ehntoo (Read error: 110 (Connection timed out))
23:43:24linuxstbWe should probably do the same with lcd-h100.c at some point. i.e. move all the generic code to lcd-2bit.c and only keep the low-level stuff in lcd-h100.c We could then quite easily add support for the 4G b/w ipods.
23:44:10amiconnYes, and lcd-recorder.c could be split in lcd-1bit.c and lcd-gmini100.c
23:44:35amiconnparts of the recorder and h100 lcd code are in lcd.S though
23:45:06linuxstbIs that "low" or "high" level code? Or both?
23:45:28amiconnVery low level - the data and command transfer routines
23:45:53amiconnlcd.S is also used for the player
23:46:22linuxstbThat shouldn't be a problem. Just means those LCDs require 3 files in SOURCES
23:47:36linuxstbThis is all assuming that the framebuffers should be organised the same for all LCDs with the same bitdepth
23:48:19preglowi can't think why not
23:48:26amiconnThat's not sure
23:49:12preglowBagder, Zagor: thanks for letting me have an ipod, btw, appreciated! :>
23:49:13amiconnThe 1 and 2 bit lcds we encountered so far use the same organisation, but I know that the gmini 2xx lcd controller (4bit grey) is different
23:49:39linuxstbBut that's 4-bit - so it must be different.
23:49:42Bagderpreglow: that thank is really to the donors, not us
23:49:55preglowBagder: i know, but the planning is up to you
23:50:03*preglow thanks the donors
23:50:12Bagderwe know you do lots of goodness
23:50:21amiconnlinuxstb: Yes, but that's not what I mean. Our 1bit and 2bit lcd use pixel "line blocks", with the bytes within the block oriented vertically
23:50:28preglowlet's hope it extends to ipod as well
23:51:13amiconnThe gmini 2xx controller uses horizontal orientation.
23:52:35amiconnOtoh, the gmini 1xx uses the same orientation as the recorder, and also the h1x0 remote lcd
23:54:04XavierGrin some years we will hold methanol bottles instead of water...
23:54:10linuxstbamiconn: Yes, but so far it's not a problem. I'll try and find out what the ipod greyscale LCD is like.
23:55:25 Join unique311 [0] (
23:57:27XavierGrwtf according to the spec a bottle of water fulled with methanol will provide 375 days of continuous usage!
23:58:59linuxstbMmm. I'm not 100% sure, but it looks like the ipod 2-bit LCD organises data horizontally.

Previous day | Next day