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 2009-06-13

00:03:25 Join isforinsects [0] (n=isforins@unaffiliated/isforinsects)
00:07:29 Join moos [0] (n=mustapha@rockbox/staff/moos)
00:08:34 Join r0b- [0] (
00:11:00 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
00:11:47 Join notlistening [0] (
00:13:18notlisteningbluebrother: bit of a side question, can the rbutility be used to generate voiced manus for a unsupported build anyway?
00:13:22 Quit bluebrother (Read error: 113 (No route to host))
00:13:56notlistening:D timing
00:14:49 Quit froggyman (" HydraIRC -> <- In tests, 0x09 out of 0x0A l33t h4x0rz prefer it :)")
00:16:30Blue_DudeI'm putting the custom build define in the config.h file. We'll see if that does the trick.
00:22:23CIA-37New commit by amiconn (r21267): Fix the charcell viewport defaults. The statusbar is offscreen on charcell. This fixes e.g. the yesno screen when deleting files.
00:25:59 Join Chesteta [0] (
00:29:53 Quit petur ("Zzzzz")
00:31:37 Quit moos (Read error: 104 (Connection reset by peer))
00:32:00 Join moos [0] (n=mustapha@rockbox/staff/moos)
00:37:17Blue_DudeNew version of FS #10324 is up. This one limits the WAV output function to simulator builds.
00:37:27 Quit bertrik (Read error: 60 (Operation timed out))
00:39:31CIA-37New commit by amiconn (r21268): (Release branch) Fix the charcell viewport defaults. The statusbar is offscreen on charcell. This fixes e.g. the yesno screen when deleting files.
00:40:34***Saving seen data "./dancer.seen"
00:43:28 Quit Nico_P (Remote closed the connection)
00:48:24 Quit midgey|w (" ajax IRC Client")
00:56:07Chestetahey is bertrik here? I was going to test the fm patches you mentioned... would that need a bootloader rebuild or just rb?
00:56:17amiconnJdGordon: Speaking about the Delete? screen - it seems there something redundant.
00:57:11amiconnFirst, the yesno screen asks whether to delete the item. If the answer is yes, the yesno screen prints the "Deleting.." message, still together with the path
00:57:55JdGordon|... ok, so whats the problem?
00:58:01amiconnBut afterwards the delete_handler in onplay.c shows a splash "Deleting...." with no further data. That seems redundant and unnecessary
00:58:02JdGordon|it might take a while to delete no?
00:58:22JdGordon|probably a VERY old bug... you're welcome to fix it
00:58:38amiconnI think this might be a leftover from pre-yesno screen times
00:58:40 Quit ender` (" You don't have to burn books to destroy culture. Just get people to stop reading them. -- Ray Bradbury")
00:58:55JdGordon|which is before me.. so yeah...
00:59:29amiconn...and could hence go away. The 'yes' message (Deleting... / pathname) shows more information than this splash
00:59:56 Part isforinsects
01:00:41amiconnHmm, but if you delete a folder, the remove_dir() function will show further splashes
01:04:09 Quit flydutch ("/* empty */")
01:04:17 Quit BHSPitLappy (Remote closed the connection)
01:06:09 Quit matsl ("Leaving")
01:15:26 Join stripwax [0] (
01:18:40 Quit JdGordon| (" ajax IRC Client")
01:21:43 Quit amiconn (Nick collision from services.)
01:21:45 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
01:21:53 Join pixelma_ [50] (n=pixelma@rockbox/staff/pixelma)
01:21:53 Quit pixelma (Nick collision from services.)
01:22:05 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
01:22:12 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
01:24:39 Quit Chesteta ()
01:28:06 Quit Lss ()
01:28:21 Join Lss [0] (
01:30:55 Part toffe82
01:33:34 Quit stripwax (Read error: 110 (Connection timed out))
01:39:00 Join JdGordon| [0] (i=460144d0@rockbox/developer/JdGordon)
02:02:01 Quit Lss (Read error: 104 (Connection reset by peer))
02:04:17 Quit rodan ()
02:10:18 Quit linuxguy3 (Read error: 104 (Connection reset by peer))
02:12:37 Join linuxguy3 [0] (
02:18:35 Quit barrywardell ()
02:19:25 Quit krazykit (Read error: 113 (No route to host))
02:20:46 Quit JdGordon| (" ajax IRC Client")
02:40:35***Saving seen data "./dancer.seen"
02:43:03 Quit moos ("Rockbox rules the DAP world")
02:54:01 Join delt [0] (i=1000@
02:54:05deltHello all
02:54:19deltquestion about "preliminary support for Apple's Ipod Accessory Protocol"
02:54:46deltipod nano, rockbox-current. will it stop charging the battery when it's full?
02:55:52deltor do i still have to put on the "lock" switch before plugging it into the charger/dock if i don't want to kill my battery in a few months?
02:59:31notlisteningthink most of the knowledgable ones are sleeping ;)
03:00:03deltprobably, it must be around 3 in the morning over there... the netherlands i think?
03:02:01 Quit Blue_Dude ("ChatZilla 0.9.84 [Firefox 3.0.10/2009042316]")
03:06:17 Join krazykit [0] (
03:06:43 Quit notlistening (Remote closed the connection)
03:08:00 Join Dauron [0] (
03:12:10 Quit _Auron_ (Read error: 60 (Operation timed out))
03:12:10 Nick Dauron is now known as _Auron_ (
03:25:25deltanyway bbl
03:25:26 Part delt
03:30:57 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
04:01:55 Quit JdGordon (Read error: 104 (Connection reset by peer))
04:04:00 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
04:23:05 Quit r0b- (Read error: 60 (Operation timed out))
04:23:20 Join r0b- [0] (n=nnscript@
04:38:53 Quit efyx (Remote closed the connection)
04:40:38***Saving seen data "./dancer.seen"
04:53:32 Quit perrikwp (" ajax IRC Client")
05:06:22 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
05:15:45 Join Horscht86 [0] (
05:16:16 Quit robin0800 (Read error: 104 (Connection reset by peer))
05:18:37 Join robin0800 [0] (
05:32:00 Quit Horschti (Read error: 110 (Connection timed out))
05:53:46r0b-os amypme awale!
06:05:30 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
06:08:00 Quit robin0800 (Read error: 104 (Connection reset by peer))
06:40:40***Saving seen data "./dancer.seen"
06:52:45 Join robin0800 [0] (
06:59:35 Quit antil33t ()
07:01:44r0b-robin0800 are you knowledgable
07:07:22robin0800r0b-: just ask
07:17:37r0b-is there any way to get scrolling in the dict plugin on the e250
07:20:33 Quit HBK ()
07:27:17 Join intrados [0] (
07:32:43robin0800r0b-: don't know just tried on my c240 and it crashes
07:34:45r0b-i wish they would finalize that :(
07:38:12r0b-robin0800 may i pm you?
07:41:41 Join beta_ [0] (
07:42:11 Join Horschti [0] (n=Horscht2@xbmc/user/horscht)
07:48:23 Quit kadoban (Read error: 104 (Connection reset by peer))
07:49:21 Join kadoban [0] (
07:50:03 Quit robin0800 (Remote closed the connection)
07:51:38 Join antil33t [0] (n=Mudkips@
07:58:51Lloreanr0b-: "they"? Remember, Rockbox is all volunteer effort, if you want something done work on it.
07:58:53 Quit Horscht86 (Read error: 110 (Connection timed out))
07:59:00LloreanVery little work has been done on dict in a long, long time
08:00:46r0b-i am working on a windows pc
08:01:59LloreanDevelopment can be done on Windows as well. If you don't want to work on it, that's fine. My point, though, was that if the people who *want* the feature won't do it, there's not really much point expecting someone else to.
08:02:05r0b-ill look and see what i can do
08:03:09r0b-Llorean will Visual Studio 2008 work?
08:03:32LloreanNo, there's instructions for getting the proper development environment set up in the wiki.
08:16:36 Join Rob2223 [0] (
08:18:26 Join r0b_- [0] (
08:18:45 Quit r0b- (Nick collision from services.)
08:18:55 Nick r0b_- is now known as r0b- (
08:19:32r0b-looking at the code it shouldnt be hard
08:19:52r0b-i just need to figure out how the keypad controls are assigned
08:34:26 Quit Rob2222 (Read error: 110 (Connection timed out))
08:40:44***Saving seen data "./dancer.seen"
08:50:48 Join linuxguy4 [0] (
08:51:04 Quit linuxguy3 (Read error: 113 (No route to host))
08:55:50 Join bertrik [0] (n=bertrik@
09:09:06 Nick amiconn is now known as amiconn_ (n=jens@rockbox/developer/amiconn)
09:09:35 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
09:10:56 Quit pabs (Read error: 60 (Operation timed out))
09:19:18 Nick crwll is now known as crwl (
09:31:45 Join bluebrother [0] (
09:32:03*bluebrother has a single exe beastpatcher now :)
09:32:37linuxstbbluebrother: What was it before? Ah, you mean static linking?
09:33:13bluebrothernot exactly static linking −− building everything together.
09:34:35linuxstbSo you've got rid of the lib completely?
09:35:34bluebrotherI just added all sources from beastpatcher and MTP_DLL to one project and build it. After some small adjustments of course. And adding the libraries for linking that MTP_DLL uses.
09:36:12bluebrothernot sure if it's worth the work to make MTP_DLL a static library and link that instead. I'd need to figure how that works first ;)
09:36:16linuxstbIs all the licensing for beastpatcher clear now? I forget what was concluded about the Creative example code...
09:36:34linuxstbbluebrother: No, I wouldn't bother with that.
09:37:35bluebrotheras far as I know the creative code is considered free for use. See the README in the MTP_DLL folder.
09:41:26bluebrotherbtw, any thoughts on moving bin2c out of the sansapatcher folder? As it is now used by different tools having it in a more common location sounds reasonable to me. Maybe in tools? Or a separate folder below rbutil?
09:48:22linuxstbbluebrother: Yes, perhaps rbutil/tools/
09:56:44 Join tvelocity [0] (
09:59:01 Join petur [0] (
10:01:21 Quit intrados (Remote closed the connection)
10:12:45CIA-37New commit by bluebrother (r21269): Some more MTP_DLL cleanup. ...
10:18:01linuxstbbluebrother: Have you thought about the old sendfirm? Does that still build for Windows?
10:18:49bluebrotheryes, it does.
10:19:27 Quit petur ("gone swimming...")
10:19:32bluebrotherbeastpatcher can now send arbitrary files too ...
10:20:00linuxstbWhy? Isn't that sendfirm's job?
10:20:40amiconnWhat happened to the "recover original nk.bin" idea?
10:21:24 Join rodan [0] (
10:21:33bluebrotherwell, IMO that's similar to ipodpatcher: it allows writing an arbitrary binary file too, so I don't see a reason why beastpatcher shouldn't be able to do it too
10:24:56linuxstbamiconn: I think simply that no-one has taken it forward. I think it's the way to go, but somehow doing it on-target, rather than from a host computer.
10:25:08linuxstbbluebrother: OK.
10:25:39bluebrotherthere's also still the issue with some beasts rejecting the single-boot bootloader :/
10:26:49amiconnMaybe it is just a size issue?
10:26:50linuxstbDo we have a flyspray task for that? It would be nice to try and narrow down the cause - e.g. OF version, size of bootloader, ...
10:27:12bluebrotherpossible. Haven't investigated.
10:27:31amiconnOF version shouldn't be involved, but there might be different versions of the loader (in flash)
10:27:53linuxstbI seem to remember jhMikeS attempting to build single-boot bootloaders with padding, but can't recall the outcome...
10:28:46linuxstbamiconn: Yes, that's what I meant. Upgrading to 1.3 of the OF with Toshiba's tool seemed to cause issues for people, and I'm assuming that updated the loader.
10:29:21bluebrotherlinuxstb: FS #9778
10:30:58 Join DanHibiki [0] (n=Miranda@
10:35:51 Part DanHibiki
10:40:48***Saving seen data "./dancer.seen"
10:58:51CIA-37New commit by bluebrother (r21270): Build beastpatcher by incorporating the MTP_DLL code instead of linking the DLL.
10:59:48 Join pabs [0] (
11:13:01 Join BdN3504 [0] (n=55b20235@gateway/web/cgi-irc/
11:18:04BdN3504Anyone here? i just made something like an in-depth tutorial on WPS creation. Do you think i shall add it to the wiki somewhere?
11:20:53 Quit tessarakt (Read error: 110 (Connection timed out))
11:21:24 Join tessarakt [0] (
11:36:45 Join notlistening [0] (
11:38:28notlisteningbluebrother, with rbutil when a wav file is generated by a speech engine is it immediately processed into mp3/spex before the next sound clip is generated
11:41:10 Join matsl [0] (
11:41:45bluebrothernotlistening: is that a question or a statement?
11:42:54notlisteningquestion sorry
11:46:33bluebrotheras far as I can see from the code (at least for voicefile creation) all strings are voiced, then encoded.
11:46:50bluebrotherdomonoky is more familiar with voice code (he wrote it ;-)
11:48:03notlisteningright will wait for him then thanks
11:48:41notlisteningi had another question you might know, is there a way to generate the voiced menus for an unsupported build?
11:49:12bluebrotherunsupported as in modified or not-yet-released target?
11:50:19amiconnnotlistening: If you are fiddling with an unreleased target, you need a build environment anyway. Just use that to build voice files instead of rbutil then
11:50:58*amiconn always builds all voice files he needs in cygwin
11:51:20 Join ender` [0] (
11:52:41notlisteningamiconn, that sounds like what I am after but was not sure how to do that?
11:53:14amiconnJust create a build directory, and the use 'configure' to configure a voice build (it's under (A)dvanced)
11:53:20amiconnThen 'make voice'
11:53:37bluebrotherrbutil needs the output of genlang (via the website). I could imagine that the website will happily handle that for an unreleased target
11:54:18notlisteningBrillant. teach me not to look into the options ;)
11:54:53notlisteningbut the issue is can you specify an unrleased target in rbutil somehow?
11:55:27pixelmabluebrother: could it also provide this for modified builds (such as RTC modded H100 etc.)?
11:57:09bluebrotherpixelma: in case of the RTC mod it should work −− it uses features.txt
11:57:59pixelmaI can't follow your reasoning it seems
11:58:18bluebrotherthinking about it it should even work without further changes: building voicefiles uses the features stated in rockbox-info.txt
11:59:02pixelmaah, and RBUtil uses the one found on the player?
11:59:40bluebrotherwell, that genlang is a script on the server that gets asked about the strings. It needs language, target, features and revision. Thus it could work if you pass it a target that is in svn but not supported by rbutil
12:03:16bluebrotheryes, it uses the features defined in the file on the player.
12:06:30notlisteningSo in theroy it just works?
12:06:47bluebrotherwell, as far as I understand it ;-)
12:07:39notlisteningOk, well as yet i have not got my player to work with the voice fukes but i reserve any further commentuntil i investigate more
12:08:13pixelmaif that theory is correct, then it should also work with unsupported targets as soon as you have a build on your player that correctly includes the features.txt
12:08:59bluebrotherpixelma: no, because the target is taken from the settings instead of the player. AFAIK.
12:09:41 Quit ender` (Read error: 54 (Connection reset by peer))
12:09:43notlisteningthats what i thought that rbutil was using the predefined selection
12:10:43pixelmahmm, forgot that there are some target specific voice strings
12:11:43notlisteningwas there a reason for this behaviour and not taking it from the player?
12:12:25bluebrotherthe device selection? Well, it's already selected by the user, hence already available. No need to read it again.
12:12:58pixelmatarget etc. are also mentioned in the rockbox-info.txt
12:13:48bluebrotherwell, it's definitely arguable what's the best way to retrieve those values.
12:15:08notlisteningokay we are after slightly different things, I think the way you have it working for supported build only is just fine but maybe an advanced option fir those in development and using custom build would be nice to specify rockbox-info.txt
12:19:54bluebrotherI'm not sure if re really want this. Rockbox Utility is the end-user installation tool. That unfortunately leaves out use cases like an almost-releasted target.
12:20:37bluebrothernot sure what's the best way to handle this.
12:21:32 Join ender` [0] (
12:21:44bluebrothergtg now. bbl.
12:21:50 Quit bluebrother ("Leaving")
12:27:55 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
12:28:28 Part pondlife
12:30:11 Quit matsl (Read error: 60 (Operation timed out))
12:40:43 Join flydutch [0] (
12:40:49***Saving seen data "./dancer.seen"
12:41:41 Join gregzx [0] (
12:46:09notlisteningamiconn, should the build generate the voice file for you and add into the zip at the end ?
12:47:03pixelmayou "make voice" seperately and will get a <yourselectedlanguage>.voice in the build folder
13:00:31 Nick Horschti is now known as Horscht (n=Horscht2@xbmc/user/horscht)
13:11:20notlisteningthanks pixelma
13:14:28 Join B4gder [241] (n=daniel@rockbox/developer/bagder)
13:19:40 Quit BdN3504 ("CGI:IRC (EOF)")
13:27:39 Quit linuxstb (Read error: 113 (No route to host))
13:44:03 Nick fxb__ is now known as fxb (
13:44:43 Quit flydutch ("/* empty */")
13:45:06 Join petur [0] (
14:07:56 Join efyx [0] (
14:22:22 Nick fxb is now known as fxb__ (
14:23:14 Quit fyrestorm (Read error: 54 (Connection reset by peer))
14:27:47 Join fyrestorm [0] (
14:40:10 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
14:40:27 Part pondlife
14:40:50***Saving seen data "./dancer.seen"
14:45:34 Join HBK [0] (
14:48:14 Quit petur ("real life....")
14:57:54 Join fdinel [0] (
15:06:22 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.10/2009042316]")
15:06:33 Join robin0800 [0] (
15:07:06Bagderopenal ? as in ?
15:07:25gevaertssounds like that to me
15:07:27 Quit B4gder ("It is time to say moo")
15:08:28*Bagder decides to just avoid that thread
15:14:56 Join tomers [0] (
15:37:04 Join nibbler_ [0] (n=Nibbler@
15:42:24gevaertstomers: I'll see if I can find out what's happening with HID one of these days. I do have an analyzer here (MQP Packet-Master USB480), so I should be able to get something done
15:47:12tomersgevarts: Thanks. I believe I just missing something in these HID reports...
15:49:24 Join stripwax [0] (
15:52:23KohlrabiHi, is there any way I could help in developing rockbox for Sansa Clip v2 target?
15:58:30bertrikyes, probably
15:58:54bertrikmake sure to read the ams sansa forum thread and the wiki entries on ams sansas and the clip wiki in particular
15:59:28bertrikI think funman also looked a bit at the clipv2, not sure if he already tested out the dual-bootloader on it
15:59:43Kohlrabi'kay, thanks
16:01:05bertrikthe thread is at and the wiki page i was referring to is
16:01:38KohlrabiThe missing recovery mode sounds dangerous :I
16:02:46bertrikthere's a patch for clipv2 here
16:03:20 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
16:03:34 Join funman [0] (n=fun@rockbox/developer/funman)
16:03:35CIA-37New commit by bluebrother (r21271): Refactor sendfirm functionality in beastpatcher and do some code police.
16:04:26funmanKohlrabi: hi, clipv2 dualboot works fine here, on the patch bertrik mentioned there is also code for lcd and buttons, and some work on SD driver
16:04:48funmanI was just looking at AS3531 doc on the internet, but nowhere I can find which SD controller is used :(
16:04:49bertrikKohlrabi, IIUC some players got bricked in the initial dual-bootloader work, but otherwise the odds of bricking turn out to be quite small
16:05:59bertrikfunman, my first guess would be that it is the same as in the v1
16:06:00Kohlrabiso, first step would of course be to install the DB-loader?
16:07:05funmanbertrik: no it's not, some peripherals are the same (dma for example), but definitely not SD
16:07:32funmanKohlrabi: you must patch a clipv2 firmware each time you want to run code because we don't have access to storage atm
16:07:39bertrikyeah, but first get the latest sources from svn, compile the cross-compiler, patch rockbox with the clipv2 patch, compile the bootloader and create a firmware update file with it using mkamsboot
16:07:52funmanmkamsboot in svn supports clipv2, but i didn't merge/sync the clipv2 patch yet
16:08:04Kohlrabisounds like the sane solution would be to set up a dev-environment on my debian box then
16:08:25bertrikyes, it's all a lot faster and easier on linux IMO
16:12:37bertrikfunman, by the way I tested the timers in 32-bit mode on ams and they work as expected (committed already)
16:14:27 Join funman_ [0] (
16:14:36funman_I found mention of as353x and linux on a deleted page of ams chinese website, perhaps we can look if ams submitted patches to lkml (i doubt it since the as3525 patches were not worked after they had been rejected)
16:14:47funman_bertrik: i noticed, thanks :)
16:15:53CIA-37New commit by bluebrother (r21272): Make bin2c compile with VS2005 and move it to a separate folder as its getting used by different tools now.
16:16:06Kohlrabibertrik, funman: do I need to follow the manual guide here?
16:16:15KohlrabiDo I need to patch my gcc?
16:16:17 Quit funman ("Reconnecting")
16:16:37bluebrotherKohlrabi: you should be able to use tools/ which does everything for you
16:17:53bertrikKohlrabi, I recommend the automatic build with
16:18:18 Join moos [0] (n=mustapha@rockbox/staff/moos)
16:18:44bertrikit patches only the target specific gcc, your host gcc stays the same
16:20:58 Join wincent [0] (
16:21:56 Nick fxb__ is now known as fxb (
16:22:15Kohlrabiok, thanks
16:23:04funman_my linux has a MACH_AS353X entry in arch/arm/tools/mach-types
16:23:49Kohlrabione day with the clip, and I'm already annoyed by lack of gapless support :(
16:23:54funman_only in this file though, I don't know how these entries get added
16:30:15 Join meyermerenstein [0] (
16:31:47 Nick fxb is now known as fxb__ (
16:31:49 Nick fxb__ is now known as fxb (
16:34:32 Join toffe82 [0] (
16:35:02 Quit patmulchrone (Read error: 110 (Connection timed out))
16:40:46 Quit funman_ ("leaving")
16:40:53***Saving seen data "./dancer.seen"
16:42:47 Quit meyermerenstein ("leaving")
16:44:36 Quit nibbler_ (Read error: 110 (Connection timed out))
16:45:10Kohlrabibertrik: I tried t patch with the clibv2 patch, and it tells me "reversed patch detected"
16:46:08bertrikhm, then the patch is partially out of date probably
16:46:24KohlrabiStill works?
16:47:03bertrikno idea, I haven't worked on clipv2 myself. A lot of changes for ams sansas have recently been committed
16:48:23Kohlrabifunman's gone
16:48:29KohlrabiI'll check back later :)
16:49:08 Join nibbler_ [0] (n=Nibbler@
16:54:04 Join kugel [0] (n=kugel@rockbox/developer/kugel)
16:54:49kugelbertrik: I think you have a typo in the 32bit timer commit
16:55:56kugelin timer.c it says "TIMER1_CONTROL = (TIMER2_CONTROL & (1<<4)) |", shouldn't it be TIMER1_CONTROL?
16:56:36bertrikyes, ah stupid
16:57:18kugelsorry for telling so late, I already noticed it when it was on the tracke :(
16:57:20bertrikI'll fix that right away
16:59:20bertrikno problem, it's my bug :|
17:01:28CIA-37New commit by bertrik (r21273): Fix bug in earlier commit of FS #10317: don't mix TIMER1 and TIMER2 settings
17:09:30 Join DarkDefender [0] (
17:17:36 Quit moos (Read error: 110 (Connection timed out))
17:21:14 Join nibbler__ [0] (n=Nibbler@
17:25:23 Quit nibbler_ (Read error: 110 (Connection timed out))
17:25:42 Join einhirn [0] (
17:26:35 Join TBOL3 [0] (i=180ac800@gateway/web/ajax/
17:29:27tomersgevaerts: What can you tell about using Wireshark for USB capture? Does it work only on Linux?
17:30:15gevaertsit's linux-only as far as I know, yes. There are also some windows things thoug
17:30:36gevaertsThe problem with them is that they do nearly no decoding, so you have to do that yourself
17:32:23tomersI am googling for more info on how to set it up. Is it required to re-compile the kernel (I am using vanilla Ubuntu)
17:33:14 Part nibbler__ ("Ex-Chat")
17:34:03 Quit einhirn (Read error: 104 (Connection reset by peer))
17:34:35kkurbjuntomers you shouldn't have to recompile your kernel
17:34:45kkurbjunubuntu has it as a loadable module
17:35:02gevaertsah, good :)
17:35:04*gevaerts wasn't sure
17:35:36kkurbjunlike gevaerts said though, it doesn't do much decoding.. I guess if you want to see the full packet size with extra debug messages you would have to recompile
17:36:05kkurbjunI think you have to manually set some things outside of make config if you want the full packets
17:36:19r0b-does rockbox automaticly stop charging when the battery is full?
17:36:26gevaertsif you're used to the software that comes with decent hardware tracers, you are going to be disappointed
17:36:31kkurbjunI was able to see everything I needed for debugging mass storage though
17:38:23kkurbjuntomers: also as a heads up I don't think that the latest ubuntu kernel release has usbmon as a module, I boot in the older kernel 2.6.27-11
17:38:36kkurbjunthat is with ubuntu 9.04
17:39:36tomerskkburbjun: Ubuntu 9.04 is 2.6.28-11-generic
17:39:37kkurbjungevaerts: yeah, it doesn't do much.. It only seems to do any decoding if you are connnected at high speed too, it doesn't decode anything at fullspeed
17:40:49kkurbjunlet me check my laptop, I clearly havn't updated the kernel on this one in a while (since 8.10 apparently).. the laptop had usbmon, but on an older version..
17:41:53kkurbjunyeah, it's 2.6.26-11 on the laptop as well, latest is 2.6.28-11, but that didn't have usbmon as a compiled module
17:43:33gevaertsat least you don't have to patch libpcap and wireshark anymore...
17:44:02tomersSo how can you get usbmod on 2.6.28-11?
17:45:10kkurbjunI would gess that you have to recompile the kernel, or at least the module, I've never compiled a module seperate from doing a complete kernel build though... your computer doesn't have the older kernel installed with 2.6.28-11?
17:45:43kkurbjunI installed 9.04 fresh on the laptop (it was when if first came out) and it had 2.6.26-11 on it also
17:46:27kugelyou can build modules only with make modules, but I don't know if it requires the kernel to be build already nor whether you can select single modules
17:46:33tomersI didn't install Ubuntu yet on that computer. I am about to install Ubuntu 9.04 vanilla, so I just wonder if there is anything special I have to do in order to have wireshark capturing USB for me
17:47:00kkurbjunoh, gotcha, I think that it installs 2.6.26-11 as well, at least it did for me
17:48:59Mikachuas long as it's the same kernel source you can probably copy /proc/config.gz to .config, add the module, build, and copy the .ko to /lib/modules/... and modprobe it
17:49:39Mikachushould be the same compiler version too
17:50:01 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
17:50:21TBOL3Maybe you guys can help me, for some reason, Rockbox has started acting strangely as of the past two days.
17:50:46kkurbjunMikachu: is /proc/config.gz standard? I don't seem to have it on my system
17:50:58gevaertstry /boot/config*
17:51:02Mikachukkurbjun: well, it's a config option :)
17:51:09TBOL3First, bookmarks, for some reason, I can't make them. When I go to create bookmark, the player freezes, I can hold down the power button, as if I was going to make a hard reset, and ocasionally, it will say bookmark created
17:51:37kkurbjunyeah, boot definately has backups of the config files
17:51:39TBOL3but then no bookmark was made. Next, the player won't auto turn off, but I think I might have done that.
17:53:01TBOL3I am using r21139-090531
17:53:18gevaertsTBOL3: which DAP is this?
17:53:19TBOL3Hmm, that's interesting, I updated the firmware yesterday.
17:53:25TBOL3Oh, sorry, sansa e260.
17:54:18TBOL3Hmm, yes, it says it should poweroff after 1 min. of being idle, but it doesn't.
17:54:42kkurbjunI'm looking through the config file on the 2.6.28-11 kernel and it looks like USB mon might just be built in without having to modprobe it
17:55:06kkurbjunI was just trying a modprobe on my laptop saw that the module wasn't there and went to the older kernel
17:55:20gevaertsTBOL3: one thing that might be worth trying is checking the filesystem
17:55:58kkurbjunso tomers you probably don't have to do anything on 2.6.28-11 other than mount debugfs
17:56:01 Join patmulchrone [0] (
17:56:08TBOL3Do you mean check to see if it's Fat32?
17:56:27gevaertsno, check if it's corrupted
17:56:47TBOL3Okay, sorry for being a bit of a n00b, but how would I do that in kde.
17:56:55tomerskkurbjun: Thanks
17:57:56tomerskkurbjun: How does debugfs relates to USB monitoring?
17:58:26kkurbjunsorry, usbfs
17:58:29gevaertsTBOL3: fsck.vfat
17:58:39kkurbjunI was using debugfs for some additional information
18:01:12gevaertsif it's built-in, just running wireshark and starting a trace on the right interface should be enough
18:01:48gevaertstomers: I've made a trace of a HID connection and added it to your FS entry
18:02:32kkurbjungevaerts: I think you have to mount usbfs first otherwise wireshark doesn't see the usb interfaces
18:02:46gevaertshm, maybe
18:02:58kkurbjuntomers: I just double checked wireshark on 2.6.28-11 after mounting usbfs and it works fine
18:03:44kkurbjunso all I did was sudo mount -t usbfs /dev/bus/usb /proc/bus/usb and then launched wireshark, picked the bus and I can monitor the packets
18:03:47tomerskkurbjun: Great. So I can install Ubuntu now... Will 9.10 also support USB monitoring in Wireshare?
18:04:37kkurbjuntomers: I would guess yes, but really I have no idea, that's up to the ubuntu developers
18:04:39Nico_Pkugel: here?
18:05:40Nico_PI just read a post you made on anythingbutipod about the ubuntu 9.04 bug with usb
18:06:02Nico_Pyou said gphoto package have been updated, but I'm not seeing that update. do you know how I can get it?
18:06:27kugelwell, the update manager installed an update that day
18:06:39kugelbut it apparently didn't contain the usb fixes :(
18:06:59Nico_Pok :(
18:07:17Nico_Pdo you know what the easiest fix is?
18:07:47kugelI compiled my own libgphoto
18:07:47TBOL3Okay, I used the command fsck.vfat /media/Sansa\ e260/
18:08:22TBOL3And the results, dosfsck 3.0.1, 23 Nov 2008, FAT32, LFN Read 512 bytes at 0:Is a directory
18:08:43kugelI downloaded the source.deb, made sure the fix is included (it wasn't when I downloaded it), then I applied the ubuntu diff and used build-pkg (or dpkg-build or something)
18:08:44TBOL3Oh, and ubuntu updated HAL yesterday, it might have made a difference, but I've been on kde since.
18:09:32Nico_Pkugel: alright, I'll do that then. thanks
18:12:51 Join BryanJacobs [0] (
18:12:54tomersgevaerts: Can you please upload a recording of the PowerPoint keys? BTW Ellisys Visual USB is far berret TMO...
18:12:55kugelNico_P: hm, I think I checked their svn out instead of downloading the source.deb
18:13:16Nico_PI found a patch that applies on the source.deb
18:16:35kugelI guess the libgphoto deb for 9.10 also has the fix
18:20:00r0b-where might i find additional rockbox plugins?
18:21:00krazykitall the ones currently suitable for inclusion are included
18:21:16krazykitthere are others in various stages of development on flyspray, though
18:21:46gevaertstomers: I'll do that later today. I know this isn't perfect, but ellisys was also about 8 times as expensive :)
18:22:07*gevaerts has to go now
18:27:46 Join ademille [0] (
18:28:06 Join flydutch [0] (
18:28:58 Part ademille
18:31:20 Quit TBOL3 (" ajax IRC Client")
18:34:12 Quit ender` (" error compiling committee.c: too many arguments to function")
18:40:54***Saving seen data "./dancer.seen"
18:41:11 Join BlakeJohnson861 [0] (
18:43:42 Join TBOL3 [0] (i=180ac800@gateway/web/ajax/
18:43:45 Join ender` [0] (
18:47:16BryanJacobsoops, sorry
18:53:49bluebrotherr0b-: there are no additional plugins. There are a few additional plugins as patches in the tracker, but you need to compile those yourself
18:55:01TBOL3Okay, I think i have the trouble sorted out, it was really weird.
18:55:13TBOL3If any of you care, I'll put what I think happened here.
18:55:32TBOL3First, do any of you know if the rockbox utility uses the gtk+ library?
18:56:21krazykitit uses qt
18:56:32TBOL3Hmm, that's interesting.
18:56:33krazykitqt4, more precisely
18:57:01TBOL3Anyway, because of the gphoto bug, I downloaded kubuntu-desktop, and have been using that to connect to my e260.
18:57:12 Quit BlakeJohnson86 (Read error: 110 (Connection timed out))
18:57:34TBOL3But any time I tried to connect using an app that used gtk+ (in particular, gpodder), the device would mysteriously dissapear.
18:58:09TBOL3I noticed that, even though I updated to the latest current build yesterday, rockbox still claimed it was a month old build.
18:58:23krazykitthen you didn't update
18:58:28*bluebrother wonders if the Rockbox Utility binary would be "rbutilqt" if it used Gtk+
18:58:56TBOL3And because I updated using the rockbox utility, in kubuntu, somehow, it wasn't connecting
18:59:05TBOL3Oops, wow, I didn't realize that.
18:59:18bluebrotheror why it has a "About Qt" menu entry :P
18:59:46TBOL3But, either way, I booted to windows, and connected it to the rockbox utility there, and updated, and now, for some reason, it works fine.
19:00:24TBOL3Hmmm, I wonder what was happening then with qt, or maybe I inadvertantly had something mixed in there that made some type of unstable connection.
19:01:00kugelUnhelpful: what do you think of using exit() in pf?
19:01:30bluebrotherQt has nothing to do with the connection to the player. It "just" does the GUI stuff (and some system API abstraction)
19:02:44martian67bluebrother, it does a bit more than that
19:02:53martian67it has some multimedia functionality
19:04:23kugelit has much more than plain gui
19:04:33kugelbut that's not used much in rbutil
19:05:02TBOL3bluebrother: I know it's just a desktop, but because people could conenct using kubuntu, but not with ubuntu, I assumed that ubuntu, somehow using gphoto, was preventing the device from being mounted as msc
19:05:35TBOL3Oh well, I'm happy it worked, and I guess I'll go report this as a bug to launchpad, thank you
19:08:05 Quit TBOL3 (" ajax IRC Client")
19:14:15 Join intrados [0] (
19:14:23martian67yea QT is much more of a cross platform c++ framework
19:14:29martian67than a X11 widget set
19:17:10 Join nibbler_ [0] (n=Nibbler@
19:24:02kugelGodEater: heh, I thought whether I should ask him for the full name in the thread or in the task :)
19:24:18kugelI decided for task :)
19:31:55CIA-37New commit by rmenes (r21274): Fix a couple of spelling errors pointed out by Melba Sitjar.
19:34:23 Quit stripwax ("")
19:35:58 Join saratogahome [0] (n=41becb3b@gateway/web/cgi-irc/
19:43:41 Quit notlistening (Remote closed the connection)
19:46:17CIA-37New commit by rmenes (r21275): Commit FS #10328 by Melba Sitjar: add two missing strings to Tagalog.
19:47:02 Join Horschti [0] (n=Horscht2@xbmc/user/horscht)
19:47:39CIA-37New commit by rmenes (r21276): A combination of FS #10328 and additional fixes by Melba Sitjar to ...
19:48:12 Join gartral [0] (n=Gartral@
19:48:52 Quit saratogahome ("CGI:IRC (EOF)")
19:50:06gartralmpegplayer will not play... it displays first frame of video and hangs the device (E250 v1) tryed with r21270 and 21273, files were encoded with FFMpeg and work without errors on my computer
19:50:55CIA-37New commit by mcuelenaere (r21277): Add build* and output to svn:ignore of the root folder (FS #10314)
19:57:05 Quit Nico_P (Remote closed the connection)
20:02:15 Quit Horscht (Read error: 110 (Connection timed out))
20:03:22kugelwhy not sim* also? :(
20:05:25 Quit robin0800 (Remote closed the connection)
20:05:48gartralkugel: didnt think of it, ok i'll build a sim
20:05:59 Quit flydutch (
20:05:59 Quit linuxguy4 (
20:05:59 Quit Kopfgeldjaeger (
20:05:59 Quit CaptainKwel (
20:05:59 Quit daurnimator (
20:05:59 Quit courtc (
20:05:59 Quit FOAD (
20:05:59 Quit dionoea (
20:05:59 Quit Unhelpful (
20:05:59 Quit preglow (
20:05:59 Quit SUSaiyan (
20:06:10kugelgartral: I was refering to the most recent commit
20:06:35NJoinflydutch [0] (
20:06:35NJoinlinuxguy4 [0] (
20:06:35NJoinKopfgeldjaeger [0] (
20:06:35NJoinCaptainKwel [0] (
20:06:35NJoindaurnimator [0] (n=daurnima@unaffiliated/daurnimator)
20:06:35NJoincourtc [0] (n=court@unaffiliated/courtc)
20:06:35NJoinFOAD [0] (
20:06:35 Join dionoea [0] (n=dionoea@videolan/developer/dionoea)
20:06:35 Join Unhelpful [0] (n=Militant@rockbox/developer/Unhelpful)
20:06:35NJoinpreglow [0] (i=thomj@rockbox/developer/preglow)
20:06:35NJoinSUSaiyan [0] (
20:09:26gartralohh.. well.. testing the files in a sim is a good idea anyway
20:24:09CIA-37New commit by kugel (r21278): Fix typo in comment.
20:30:43bluebrothermartian67: of course I simplified a bit. For the multimedia stuff it uses phonon of the KDE project.
20:32:07bluebrotherso it's not strictly Qt itself. Same with webkit. The network module (which we use) and the xml module (which we don't use) however are plain Qt afaik.
20:32:16bluebrotherso that's non-GUI stuff
20:40:58***Saving seen data "./dancer.seen"
20:48:59 Quit tvelocity (Remote closed the connection)
20:50:55 Join OPP2 [0] (
20:51:10OPP2hi i have a 5G ipod video what kind of battery life can I exepect with rockbox 3.2
20:56:54 Quit Tuplanolla (Read error: 110 (Connection timed out))
21:00:13 Quit gartral ("Lost terminal")
21:01:26 Join delt [0] (i=1000@
21:01:28deltHello all
21:02:07delt"preliminary support for Apple's Ipod Accessory Protocol" −− ipod nano, rockbox-current. will it stop charging the battery when it's full?
21:02:24deltor do i still have to put on the "lock" switch before plugging it into the charger/dock if i don't want to kill my battery in a few months?
21:02:58Mikachui don't think any of those three are related?
21:03:20Mikachuaccessory protocol, charging, lock switch
21:03:35deltlock switch == boot into stock firmware
21:03:54deltwhich stops charging the battery once it's full
21:04:35deltif you keep charging a full battery, eventually it loses its zero point and holds less and less charge
21:05:38deltso if i keep my ipod on the little dock/charger thingy using rockbox 3.0, the battery life will be less and less
21:07:42deltactually its already starting to die :/
21:08:42deltipod nano, bought a few months ago from someone who used stock firmware, when i got it it was fine
21:10:47gevaertsdelt: I at least never heard this (but I'm not an ipod specialist, so that doesn't mean much).
21:11:41deltwhat about players other than ipod? (they have batteries too?)
21:12:08gevaertscharging is very hardware dependent, so DAPs tend to have their own charging code
21:12:39gevaertscan you submit a bug report with all information you have?
21:13:19 Quit nibbler_ (Read error: 110 (Connection timed out))
21:13:24deltisn't much information... i don't have the equipment to test voltage and current between the player and the dock
21:13:49 Join FlynDice [0] (
21:13:51gevaertsmaybe, but without a bug report nobody is going to look at it at all
21:14:01Mikachuif you boot back to the OF and charge, you still get short battery time then?
21:14:58delthard to say... i'm not talking about how long -one- charge lasts... i mean the battery's performance in general
21:15:11Mikachuyou said you noticed it has gotten worse
21:15:49deltso if i charge it using either software up to full, won't make a difference immediately, BUT i strongly suspect rockbox keeps charging the battery after it's full, which damages it over time
21:16:11Mikachuso if it isn't battery time, what have you actually noticed?
21:16:49deltthe battery's performance has deteriorated.
21:17:29kugelgoing to wps from plugins requires a tad bit of hacking in the core
21:17:35deltif i charge it up to 100%, and then disconnect the ipod, leave it turned off, a few hours later i turn it back on and it says 80% battery
21:18:03deltthe next day it might even be down to 50% or 60% (not being used at all)
21:18:41kugelGodEater: frederik full is his realname?
21:18:54kugelI suspect he's kidding us since we asked for the full name
21:19:09gevaertsdelt: I assume it stays that low? Some batteries need a bit of time to get to their proper voltage
21:19:44deltgevaerts: yeah i noticed when i turn it back on, after a few minutes it goes up about 4 or 5%
21:20:00DhraakellianI've found that my Sansas don't seem to charge as quickly under Rockbox
21:20:13deltmaybe the battery is just old.. but if i change it, i want to make sure the same thing won't happen to the new one
21:20:58deltcurrently the only way to do that is to boot the ipod with the stock firmware when charging
21:21:27delt(ie. lock switch when booting, or press i think select + play at the apple logo)
21:21:48Dhraakellianare the boot-into-OF keys different on each player?
21:21:49gevaertsI do know that ipods tend not to charge with a decent current currently, but I never heard of it not stopping properly
21:21:53Dhraakelliantarget, rather
21:22:13gevaertsDhraakellian: well, *keys* are different on each target...
21:22:31deltif i connect it to a USB port there's no problem because it automatically goes into OF
21:22:37delt"disk mode"
21:22:52DhraakellianI guess I had just assumed that |<< was fairly standard
21:23:14Dhraakellianthen again, I only have experience with sansas
21:23:14 Join robin0800 [0] (
21:23:39deltbut it's kind of annoying when i have it on my alarm clock playing and have to remember to remove it after like 30 minutes
21:24:06delt(alarm clock/radio/ipod charger/dock)
21:24:07Dhraakelliandelt: incentive to get up within the 30 minutes?
21:24:47deltwell, if i listen to say 2 or 3 albums in a row from the charger i know i'm damaging the battery
21:25:05gevaertsanyway, if you suspect that something is wrong, please submit a bug report
21:25:07deltwell, strongly suspect rather, as i said before
21:25:19 Join aaron424 [0] (n=410d02d8@gateway/web/cgi-irc/
21:25:44deltcan't be sure because rockbox doesn't indicate if it's charging or not
21:26:04gevaertsyou've said that a few times now...
21:26:49deltso how do i submit a bug report?
21:28:14delti guess it's "bugs" in the "tracker" section?
21:28:36robin0800delt If you charge it in rockbox i'm sure with the cabbie v2 will show charging on its battery icon
21:28:57Mikachuit shows charging, but it continues showing that as long as the cable is in
21:29:00Mikachuregardless of battery level
21:29:39deltno, i use that theme (the default one i believe) and iirc it just shows the battery being more full, like it thinks the dock's power is the battery
21:29:49delt2sec i'll check again
21:30:00robin0800Mikachu: No it dosen't it has both charging and connecting icons they are different
21:30:55gevaertsyou can see the battery state in the debug menu
21:31:33Dhraakelliangevaerts: ah, so you can see what the actual battery level is rather than the 100% that the statusbar shows whenever it's plugged in?
21:32:00Dhraakelliannice. I shall have to remember that!
21:32:35Dhraakellianalso for right after I unplug and the statusbar battery indicator is still slowly falling back to the correct value
21:32:43deltnope... there's no way to tell if it's connected and/or charging just by looking at the screen
21:34:23deltoh, was also wondering.. this alarm clock charger can auto-turn-on the ipod and start it playing at the alarm time.... does this work with current build?
21:35:16Mikachudelt: rockbox can start the ipod without any extra hardware on a set time (and optionally date)
21:35:19robin0800delt: Thats with Cabbie corect if you look in the backgrounds folder for the theme you should see them they are on the battery strip
21:37:51deltin the bmp file i see the battery sign with a zigzag and another one with an arrow. i've never seen those on the ipod screen.
21:39:37robin0800delt: Well either the wps has an error or the rockbox code has an error it works ok on a sansa so perhaps its the wps
21:41:04deltthe "wps" file is a kind of script that says what goes where on the screen, right?
21:41:28Mikachudelt, robin0800: most likely we don't know how to get info about how the charging is going on ipods, so it can't be displayed
21:41:33Mikachubut i don't know details of charging stuff
21:42:28delt..... and we can't base it on current battery level, since when plugged in it thinks the battery is full
21:43:05deltor is there another value (haven't checked the debug info yet)
21:43:26gevaertsthe debug info is the *only* place to get reliable information...
21:43:27deltanyway... thanks for your assistance, will be back later with (hopefully) more information
21:43:32kkurbjunis there a high quality reference for the images that the plugins use? I am trying to scale the graphics for the various plugins to look appropriate on a 640x480 screen size - at the moment I am just doing cubic interpolation, but if there's a better reference then the SVN graphics used that would be ideal
21:43:33 Part delt
21:43:49 Quit aaron424 ("CGI:IRC (Ping timeout)")
21:44:26kkurbjunI was also looking at the scale2x algorithm but at least on the card decks I think that cubic looks better with some sharpening applied
21:44:39robin0800Mikachu: Even if While charging the voltage is unknown when chargeing is complete it switches fom charging to connected
21:44:52Mikachu... on the sansa
21:45:43robin0800Mikachu: I think only the C200's
21:48:20robin0800Mikachu: For unknown voltage only I think all portal players switch as described
21:51:47 Join petur [50] (n=petur@rockbox/developer/petur)
21:54:10 Join merbanan [0] (
21:55:46 Quit OPP2 (Read error: 110 (Connection timed out))
22:00:09kugelcontext menus should behave consistently with normal lists with regards to going to wps and stopping playback, correct?
22:02:03kugelyou can't go to wps from a simplelist or context menu
22:04:30 Join tvelocity [0] (
22:06:22LloreanIdeally, yes.
22:07:05kugelI've tackled that locally
22:07:42 Quit intrados (Remote closed the connection)
22:07:59 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
22:12:03 Join intrados [0] (
22:20:07 Join webguest55 [0] (n=50791b52@gateway/web/cgi-irc/
22:20:07 Quit webguest55 (Client Quit)
22:20:12 Join Z-mann [0] (n=50791b52@gateway/web/cgi-irc/
22:21:51Z-mannhallo allezusammen. hello everyone.
22:25:57 Join Tuplanolla [0] (n=jani@unaffiliated/tuplanolla)
22:40:07 Part Llorean
22:40:22 Join Llorean [0] (
22:40:59***Saving seen data "./dancer.seen"
22:42:06 Quit wincent (Read error: 110 (Connection timed out))
22:45:07 Join yessss [0] (
22:52:41CIA-37New commit by kkurbjun (r21279): Add appropriate sized cards for 640x480 screens based on svg-cards-2.0. SOURCES is setup for 640x480 on a number of plugins as well, but graphics are ...
22:55:37 Quit yessss ("leaving")
22:59:07CIA-37New commit by kkurbjun (r21280): M:Robe 500: Fix up 256 color blitting when in portrait mode, fix a mistake in the plugin linker script, reduce the amount of memory used by ...
22:59:24 Part rodan
23:03:23kugelkkurbjun: "when in portrait mode"?
23:03:47kugeldoes it support 90° somehow?
23:03:53kugel90° flipping
23:04:17 Join froggyman [0] (
23:04:17kkurbjunkugel: the lcd on the m:rboe is naturally 480x640 .. not through hardware, normally the screen is rotated to landscape mode in software
23:04:56kkurbjunI have been making sure both work since portrait mode is much faster, but it is a pain to use when it's on it's cradle
23:05:10kugelso portrait is default, but mpegplayer rotates (as usual on portrait mode targets)?
23:05:40kugelI'd love to see 90° flipping for other targets as well
23:06:16kkurbjunlandscape is default right now in the builds, mpegplayer is always in landscape mode, the YUV blitting doesn't have any code for portrait mode
23:06:34kkurbjunit's not a setting that you can do in the menu, you have to change the config file and build your own
23:07:09kkurbjunit would be nice for it to be a setting, but that would take quite a bit more work with all the hard coded widths and heights
23:07:47kkurbjunand I'm not sure what the performance penalty would be if those were variables instead of hard coded
23:07:54kugelbut the screen api has lcdwidth and lcdheight members, so it's just a matter of looking through all places and convert
23:09:29kkurbjunyeah, plugins would be really tough with all the bitmaps
23:09:46kkurbjunsince they are built at compile time with the appropriate sized images
23:10:19kugelyea, not that easy then
23:10:41gevaertsYou could build all plugins twice...
23:11:27kkurbjunI think it would be better to have one large image that is compiled in the plugins and then use the scaler to downsize for the screen
23:11:35kkurbjunbut that's again, alot of work :-D
23:11:56kugelscaling isn't even necessary, rotating is
23:12:13gevaertsI don't think that will work well for some of these small pixel-painted bitmaps like e.g. the sokoban things
23:12:39kkurbjunyeah, true...
23:12:55kkurbjunrotating is done in the lcd blitting routines
23:13:18kkurbjunso you don't need to do rotation in the plugin
23:16:04kugelanother question, why aren't we using __atribute__((unused)) instead of doing (void)param all over the place?
23:16:22gevaertsgood point
23:16:59kkurbjunwhat's the difference?
23:17:20kugelI've looked for the impact. I was having a empty function which just does (void)param, it should be inlined but it wasn't inlined
23:17:36kugelthe disassembly showed that gcc made a function call doing nothing but that read
23:17:59gevaerts(void) actually reads the variable?
23:18:11kugelkkurbjun: (void)param is actually an instruction, while __attribute__((unused)) just silences the warning
23:18:22gevaertsi.e. it loads a memory address somewhere?
23:18:35Bagder(void)param isn't strictly an instruction
23:18:54gevaertsIn that case I certainly vote for __attribute__((unused))
23:19:13Bagderwe should probably switch to UNUSED(var) or similar
23:19:24Bagderso that we can define it to use whatever we please
23:19:33kugel(I did it for the sim) it did push eax, mov eax, [var], pop eax (something like that)
23:19:45gevaertsMaybe that can be this year's huge-devcon-commit-that-touches-nearly-everything :)
23:20:33kugelI've seen gtk programs which have lots of "int func(G_UNUSED void* param)"-like functions
23:21:42gevaerts309 affected files
23:22:26kugel"int debug_menu(__attribute__((unused)) void* ignored)" doesn't give any warning
23:23:55gevaertskugel: that's a bit hard to do everywhere though. You often don't know there that it will be unused, only in the function body
23:24:16kugelit doesn't matter
23:24:24kugelyou can use it just fine, it just kills the warning
23:24:43gevaertsIt also kills it in places where it should stay then
23:25:25kugelI don't think so
23:25:40Bagderah, you need it in the prototype?
23:25:45kugelany function which has a (void)param wants to kill the warning
23:25:53kugelBagder: yes
23:26:03gevaertskugel: often there's an #ifdef around that
23:26:06Bagderkugel: but (void) is quite easily conditional
23:26:24Bagderyou can do that in the proto as well, but it'll look a bit diifferent
23:26:27kugelyes, but where's the problem?
23:26:52CIA-37New commit by kkurbjun (r21281): Fix red
23:27:02gevaertssuppose you have a function taking two parameters, one of which is only used on HWCODEC, so it's (void)param on #ifdef SWCODEC
23:27:15 Join mt [0] (n=MTee@rockbox/developer/mt)
23:27:32kugelthat's a usual case, I know
23:27:43gevaertsNow someone rewrites the HWCODEC bit to no longer need that parameter. In your proposal, nobody may notice that the parameter can now be dropped
23:28:12kugelexcept for the guy that rewrites it?
23:28:16Bagderwith the __attribute__, you'd #ifdef the actual prototype
23:28:25gevaertsif the code is convoluted enough, nobody will notice
23:28:35kugelI don't think so
23:28:46kugelBagder: you do that in the function definition
23:28:55Bagderyse, that's what I meant
23:29:01kugellike I showed above, not sure about the declaration
23:29:04Bagderbut that's messy imho
23:29:26Bagderand thinking about it. Is there any cases where we do (void) like this unconditionally?
23:29:50kugelit's less messy than (void)param (which often alone requires an #ifdef), also we can make it shorter with #define UNUSED __attribute__((unused))
23:29:54gevaertsprobably, in some "generic" functions like the one implemented by various drivers for the same class of hardware
23:29:58kkurbjunthe m66591 does a void unconditionally
23:30:08kkurbjunthe driver that is
23:30:13Bagdergevaerts: ah yes, that's right. We probably have a few such
23:30:15kugelBagder: there are many places. like many event callbacks
23:30:34gevaertsyes, those too
23:30:57gevaertsAnyway, those can be changed without problems I think
23:31:09mtI know a weekly status update of a gsoc project is required, but there hasn't been almost any news this week, so no updates from me.
23:31:43kugela mail saying no progress would be fine too
23:33:05kugelgevaerts: re your example. if someone rewrites it without seeing SWCODEC doesn't need this either, he'll not remove the parameter no matter of __attribute__((unused))
23:33:10gevaertsfrom what I can see, there seem to be 23 #ifdefed blocks of (void) things, and 605 unconditionals
23:33:36 Join rodan [0] (
23:35:11gevaertskugel: let's start by changing all easy cases, and think a bit more about the others. If there really are only about 20 or 30 left, it's not that big a deal
23:35:30kugelthat's fine with me
23:38:00kkurbjunI have a question on licenses - I used the svg-cards-2.0 for the 640x480 screen size card deck and it is licensed under the LGPL-2.1 is there anything I need to do with respect to documenting those card's use in the source?
23:38:44kkurbjunsince it's just a binary image I can't put the license in a header
23:38:55kugelBagder: our AMSes rely on that (void)x is an instruction somewhere in the memory init
23:39:19 Quit DarkDefender ("Leaving")
23:39:23kugelkkurbjun: I think a README or LICENSE would do it
23:39:54kugel"(void)tmp; /* we just need to read from this location */"
23:40:03kkurbjunwhere do you think that would be appropriate to place, the files are in plugin/bitmaps/native
23:40:13Bagderkugel: UGLY
23:40:32kugelI have no idea why this is
23:40:44Bagderand how does 'tmp' read from "this location" ?
23:40:59kugelunsigned int tmp = *(volatile unsigned int*)(0x30000000+0x2300*MEM);
23:41:51Bagderright, then 'tmp' holds an integer. but what address does '(void)tmp' then read from?
23:41:58 Join srnt [0] (
23:42:22kugeltmp holds that address, (void)tmp loads the data at that address into a register
23:42:45gevaertsI think that a bit of inline asm would be cleaner in this sort of case
23:42:57Bagderno, the line you showed read the data off that address into tmp
23:43:06Bagderbut sure that itself can be an address
23:43:14kugelas I said, the x86 disassembly of such a thing looked like "push eax;mov eax, [tmp];pop eax"
23:43:39BagderI find such code next to disgusting
23:43:59srntHi. How does rockbox choose which image display during a song playback?
23:44:19kugelsure, but the comment above says "/* this part is required, if you know why please explain */" so we probably have to live with it
23:44:39kkurbjunkugel: did you try that with different versions of GCC, is it guaranteed that (void) will do some type of read?
23:44:48Bagderthis part being that particular (void) ?
23:44:48kugelno idea
23:44:56 Join animatorgeek [0] (n=4b53ec3e@gateway/web/cgi-irc/
23:44:56gevaertsI wouldn't count too much on it...
23:45:00Bagderme neither
23:45:14animatorgeekOkay, I think I've found a bug in the configure/makefile
23:45:19srntHow is Sansa Fuzz supported?
23:45:21kugelit will probably break with -O too
23:45:24gevaertsJust do it in asm
23:45:35animatorgeekI just got this error when building: make: *** No rule to make target `/rockbox/build-sim/apps/bitmaps/native/rockbox /build-simlogo.160x53x2.o', needed by `/rockbox/build-sim/rockboxui.exe'. Stop.
23:46:10kugelkkurbjun: no I didn't. my gcc is 4.3.3 here
23:46:17animatorgeekNotice that the file it's trying to build is called build-simlogo.160x53x2.o
23:46:45animatorgeekAs far as I know it should be building rockboxlogo.160x53x2.o
23:47:15gevaertsanimatorgeek: I seem to remember that having your rockbox directory in the root broke things, but this may have been fixed...
23:47:35animatorgeekLooks like something is going wonky, substituting the build directory's name at the beginning of that file, rather than "rockbox"
23:47:40krazykitsrnt, for your album art question, refer here:
23:47:50Bagderanimatorgeek: if it was that serious our build table would not be green, as it happens to be atm
23:48:08Bagderbut sure, it might be that it's a root dir
23:48:12animatorgeekI don't have it in my root dir,
23:48:27gevaertsdid you change the paths you reported here?
23:48:27animatorgeekI have it in c:\cygwin\rockbox\biuild-sim
23:48:39gevaertsthat's the cygwin root
23:49:06animatorgeekWell, in cygwin I have it in rockbox/build-sim
23:49:21*bluebrother thinks when working on cygwin one should use the paths shown by cygwin, not explorer
23:49:23animatorgeekIs that what you mean? Having the Rockbox source dir as a subdir of the cygwin root?
23:50:04animatorgeekSo I should move it somewhere else? Like to my home dir perhaps?
23:50:24gevaertsthat's one possible place, yes.
23:50:45animatorgeekIs there a recommended place, or just somewhere not-on-root?
23:50:48gevaertsI'm not entirely sure that this is the problem, but I'm pretty sure that it was a problem at some point
23:50:55gevaertsthat's up to you
23:51:32animatorgeekI'm compiling from 3.2, not the current codebase, but I had a similar problem when I tried compiling the current code a few days ago.
23:52:06animatorgeekThe weird thing is that when I switched to 3.2 the problem fixed itself. Now it's back again when I try to build the sim
23:52:34rodanDoes rockbox run on sansa vuze?
23:52:35srntkrazykit, thank you
23:52:52krazykitrodan, do you see it on the front page?
23:52:53bluebrotherthere is no sansa vuze.
23:52:55srntrodan, sansa is the best, I just bough fuzz
23:53:04srntmaybe fuzz?
23:53:11gevaertsrockbox runs on neither the sansa vuze nor on the sansa fuzz. There is work in progress on the sansa fuze however
23:53:12srntthat is
23:53:16bluebrotherwhat's this fuzz about the sansa?
23:53:24bluebrotherh100 rulez!
23:53:39animatorgeekI'm with you, Bluebrother :)
23:53:50animatorgeek(well, h120, that is)
23:53:57rodanon the mainpage it mentions nothing about the Clip, but i heard rockbox runs on the sansa clip
23:54:05bluebrotherh100 = h100 series :)
23:54:15bluebrotherRockbox does not run on the clip.
23:54:34bluebrotherit's being worked on, but until it's in a state it can get called supported it does not run.
23:54:51animatorgeekWell, thanks for the help re directories.
23:55:02srntrodan, this page says it runs on Fuze and Clip
23:55:26 Join saratogahome [0] (n=41becb3b@gateway/web/cgi-irc/
23:56:45krazykitsrnt, it's not in a state where it should be run by non-developers
23:57:49*bluebrother points out that there is a difference between running and running. And between running and supported.
23:57:52rodanthank you guys
23:58:04rodanjust curious.
23:58:49saratogahomeif you're curious you could read the wiki
23:58:58saratogahometheres even a status link with everything you asked on the front page

Previous day | Next day