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

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

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

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

#rockbox log for 2008-03-18

00:00:29 Join stripwax [0] (
00:01:29 Join svenVI [0] (
00:04:32 Quit ender` (" I was in the grocery store. I saw a sign that said "pet supplies." So I did. Then I went outside and saw a sign that said "")
00:05:10 Quit tvelocity_ (Remote closed the connection)
00:09:44ali_asamiconn, just occured to me if you could multiply the number by 10000000000000001b, that ought to work, but I'm nolonger sure my maths is right.
00:10:11ali_asAnd the required integer is invalid in one instruction anyway.
00:10:30amiconnAnd it would be much much slower, especially on arm
00:10:52ali_asOh, it would be agonisingly slow, but it would have solved the problem as stated.
00:13:23 Quit floatsaka (Read error: 110 (Connection timed out))
00:13:53ali_asMy high half word would be out by one if the input was negative.
00:13:58 Quit mcuelenaere ()
00:14:48 Quit PaulJam (".")
00:17:24 Quit stripwax_ (Read error: 110 (Connection timed out))
00:19:31 Quit OlivierBorowski (Remote closed the connection)
00:20:56 Join stripwax_ [0] (
00:27:06 Quit Arathis2 ("Bye, bye")
00:28:20 Quit bughunter2 ("Leaving.")
00:31:32 Join FOAD_ [0] (
00:39:25 Part bertrik
00:40:07 Quit stripwax (Read error: 110 (Connection timed out))
00:43:41 Join hd [0] (i=jd@unaffiliated/helldragon)
00:46:15Bagdernow at 4 mentors: me, petur, scorche and LinusN
00:47:15Bagder <= apply here
00:47:22*Bagder heads away to sleep
00:47:48 Quit FOAD (Read error: 110 (Connection timed out))
00:47:49 Nick FOAD_ is now known as FOAD (
00:50:03amiconnBagder: The build system didn't pick up a commit...
00:53:19*preglow goes looking for a project list
00:54:27 Quit gevaerts ("Zirconium Zzz..")
00:54:40 Join cool_walking_ [0] (
00:55:46 Join fasmaie [0] (
00:57:19***Saving seen data "./dancer.seen"
01:00:34*JdGordon|uni thought he already did the mentor signup
01:01:24*preglow doesn't think he'll mentor this year, thanks to uncertain circumstances this summer
01:05:36*linuxstb sees ffmpeg are setting a high standard with their compulsory qualification tasks -
01:05:58 Quit jhMikeS (Nick collision from services.)
01:06:04 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
01:09:29preglowlinuxstb: well, ffmpeg is notorious
01:10:20preglowall ffmpeg code seems to be thrown around for ages before it's commited, really does make our trying to merge wma fixed point stuff back look bleak
01:10:41preglowsince it's bloody boring work, and they're extremely hard to please
01:12:16JdGordon|unianyone still awake enough to comment on 8753? plugging the headphones back in after manully pausing playback shouldnt restart it?
01:12:20scorche|shholy wow
01:13:29preglowJdGordon|uni: premise sounds good
01:13:33preglowthen again, i hate that entire feature
01:13:42preglowwell, hate is a strong word...
01:17:13linuxstbJdGordon|uni: I think you could argue for either behaviour... But like preglow, I avoid that feature, so don't care...
01:20:08JdGordon|uniyeah, I dont use the feature either, but the fix shuold be dead simple which is why im looking at it
01:20:34amiconnThe build system needs a good kicking
01:20:42linuxstbJdGordon|uni: Maybe people who do use it should comment on the desired behaviour before you "fix" it?
01:21:01JdGordon|unii was hoping for that with my message before
01:21:54scorche|shanyone have any ideas on what should be done about qualification for our GSoC projects?...i do like the idea of an entry interview of sorts either here in IRC or on something like skype
01:22:36JdGordon|unithere was a bit of a discussion about that yesterday.. Bagder came up woth some questions for them
01:23:01linuxstbAren't students encouraged to get to know the project in advance? i.e. if they hang around here for a while, we probably won't need interviews of any kind.
01:23:06scorche|shsorry...i dont usually read the logs unless i get highlighted or some other such
01:23:26JdGordon|unias long as they can already code in c and compile the source I think thats enough...
01:23:52linuxstbPlus not planning on a 3 month holiday...
01:23:53scorche|shlinuxstb: they are, but i wouldnt mind a bit more of an official interview or at least introduction/minimal questions about them and their proposed project
01:23:54JdGordon|uniit will really only be a problem when we ned to narrow down the projets we want to the 3 or 4 we might get
01:24:26scorche|shespecially as im sure there will be plenty of folk who wont come in here before they submit their projects
01:24:34scorche|shalthough i would like to make that a requirement
01:24:39*svenVI happens to be interested in a soc project.... though doubts his idea is feasible.
01:24:41linuxstbscorche|sh: Then those are probably the kind of people we don't want...
01:24:52linuxstbsvenVI: What's your idea?
01:25:21svenVII want to hack around with an iPod Classic so that custom firmware can be run on it.
01:25:38 Join sdoyon [0] (
01:25:43scorche|shlinuxstb: i can accept that some people may have some apprehensions about coming in here, but if we make it a requirement in bold bright letters on the ideas page or some such, they dont really have an excuse to not come in here
01:25:54linuxstbIt would be great if google would pay you to crack Apple's encryption... ;)
01:26:07scorche|shsvenVI: unless you already have some success, i wouldnt suggest that as a GSoC project
01:26:14svenVII was thinking they'd reject it on those grounds. :P
01:26:32 Quit amiconn (Nick collision from services.)
01:26:38 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
01:26:41scorche|shsvenVI: reason being would be that it wouldnt be that fair to accept a project that we are not sure can be completed
01:26:48linuxstbWell, the project would just be described as "port to the ipod Classic".
01:26:51amiconnNo Swede around to kick the build system, and still no method for committers to kick it remotely :(
01:26:54*amiconn sincerely hopes that he didn't break anything, as he won't be able to fix it during the next ~18 hours
01:27:20svenVIYeah, I agree, I don't know what sort of work it would entail. I haven't begun researching the hardware differences yet.
01:27:57svenVI(GSoC or not, I was intending on doing that this summer anyways.)
01:28:09kkurbjunscorche, I liked the idea of having an entry test to make sure they can compile the code and present at least a demo plugin to verify they at least have the basic setup that we discussed at the SOC conference.
01:28:14JdGordon|uniamiconn: shouldnt it restart after 20min?
01:28:16scorche|shhonestly, unless you have lots of prior experience and already have at least some minor success, i dont see it being accepted as a GSoC project
01:28:36scorche|shbut dont let that stop you from working on it evne outside of GSoC =P
01:28:53scorche|shkkurbjun: i did too, but i think that some sort of interview in here would do nicely
01:28:59amiconnJdGordon: Then it should have restarted quite a while ago from my 1st commit (around 45 minutes ago)
01:29:21kkurbjunI think that would be something that would be included with it
01:29:24ali_assvenVI, do you have much/any experience with encryption or protected chipsets?
01:29:44svenVIali_as: I have none, I was going to start with this project.
01:30:02scorche|shbut plugin hello-world-esque that could be done in less than an hour with some wiki searching on how to make a plugin would be fine
01:30:13kkurbjunat my job it's a pretty extensive interview process with 2 technical sections and a basic personality interview - I don't see why we wouldn't want to do something similar
01:30:43scorche|shi dont see why we cant do a basic interview at the least
01:30:56scorche|sheither way, we should figure this out quickly :)
01:31:11scorche|shon that note, /me -> home
01:31:38ali_asThe iphone is being hacked on a regular basis, I wonder if any of those methods would apply to the ipod classic.
01:31:49 Quit JdGordon|uni (" ajax IRC Client")
01:32:16 Join Davide-NYC [0] (
01:32:53Davide-NYCScalable Fonts in GSo page?
01:33:13linuxstbmake: *** No targets specified and no makefile found. Stop.
01:33:19preglowali_as: more likely they apply to the ipod touch
01:33:26Davide-NYCWhat do you guys think of scalable fonts as a project?
01:33:44*Davide-NYC ducks behind his desk.
01:33:57preglowlinuxstb: will you respond if i do "ls" too? :)
01:34:12linuxstbSure, but I'm in an empty directory
01:34:17preglowDavide-NYC: would be cool if optimized enough
01:34:21preglowrm -rf /
01:34:42krazykitsorry, you're not root ;)
01:34:43Davide-NYCmaybe it should be added to the GSoC wiki page? (I'm certainly not going to do it)
01:35:07preglowi don't really think we need it, though
01:35:14preglowi doubt any other mp3 players use scalable fonts
01:35:48linuxstbMaybe an SoC qualification code could be to submit a patch that optimises something - either for binsize or speed.
01:36:16preglowdo we really need qualification tasks?
01:36:17ali_asThat might be hard to judge.
01:36:26kkurbjunlinuxstb, that could be a hard requirement for someone completely unfamiliar with the project
01:36:55kkurbjunpreglow, the idea was just to make sure that they have at least a basic setup before we accept them, I think markuns student ran into that problem
01:37:33preglowif you can't get a basic setup in one day, you're either incompetent or not trying
01:37:36preglowi don't see it as a problem
01:37:37linuxstbkkurbjun: Not really - it just shows they understand C, and Rockbox coding requires someone who understands those concepts.
01:37:52linuxstb(coding with limited resources)
01:38:01*amiconn thinks that at least some possible optimisations are easier to find for an "outsider"
01:38:43kkurbjunpreglow, some students are simply going through the list of projects and applying to multiple ones, it would at least weed out people who are serious enought to setup a basic environment
01:38:52ali_asIf you shave one cycle by implimenting a C function in assembler, have you really succeeded in optimising it though.
01:39:13 Quit Nico_P (Remote closed the connection)
01:39:24preglowkkurbjun: true enough
01:39:50preglowbut really, i think qualification tasks should be specific, "optimize something" doesn't cut it
01:40:02kkurbjunlinuxstb: I think it would defiantly be worthwhile if we have enough applicants
01:40:15kkurbjunenough to make the requirements more difficult that is
01:40:42kkurbjunwe could even do something like take file x.c and optimize function y in a way that you see fit
01:40:43linuxstbpreglow: My idea was just trying to find something one-step above "compile rockbox"
01:40:56amiconnali_as: Beating gcc isn't hard, and often delivers significant performance gains, way more than just a few cycles
01:40:59kkurbjunand then evaluate their method
01:41:13preglowif you can't beat gcc on our targets, you'are also incompetent :)
01:41:58linuxstbI wasn't thinking of asm optimisation though - just finding code which could be more efficient (in C)
01:42:09amiconnThat seems to be the case for all our target architectures in varying degree. Gcc might be harder to beat when it comes to x86 - but I have no experience with that
01:42:13ali_asamiconn, that wasn't my point. x10 speed increase over gcc is often possible, my point was that 'optimise' is rather a subjective goal.
01:43:05kkurbjunwhen can students start applying to projects?
01:43:10ali_asIf the point is to make a modification that doesn't crash the system, and compile, then it is satisfactory.
01:43:20kkurbjunand when do interviews/acceptances take place?
01:43:58amiconnlinuxstb: Yeah, C code can often also be optimised (although often not as much, because gcc already tries to optimise a bit)
01:44:19Davide-NYCFWIW, I can compile RockBox and even modify the code *somewhat* and I am a moron. Vet these people.
01:44:30ali_asAlgorithmic optimisation should always be the first step anyway.
01:44:43*amiconn 's first contributions to rockbox were in fact ASM optimisations
01:45:20ali_asThat does not surprise me from your mov/orr solution :)
01:45:41amiconnOptimisations of what already was assemblerised, that is. And I learned SH ASM while doing it
01:45:54 Join RoC_MasterMind [0] (
01:45:56amiconnali_as: Back then we had no ARM targets, only SH1
01:46:04 Quit RoC_MasterMind (Client Quit)
01:46:09ali_asSH, is that a hitachi chip?
01:46:44SSnakeand it should be called assembly, but nobody cares these days
01:47:29ali_asNot met an SH chip yet.
01:47:34kkurbjunffmpeg's qualification tasks are pretty extensive:
01:47:53preglowi hate windows :/
01:47:53linuxstbmake: *** No targets specified and no makefile found. Stop.
01:48:48kkurbjunI like this one: take an existing patch and make it suitable for inclusion in SVN
01:48:56linuxstbkkurbjun: I was just about to suggest that...
01:49:01kkurbjunthey picked a specific patch that was nearly ready to be included
01:49:05ali_asamiconn, were you an acorn nut btw?
01:49:27 Part pixelma
01:49:28linuxstbkkurbjun: But it would need us to go through and select patches that were (in theory) suitable for inclusion, and comment on them.
01:49:38amiconnali_as: hmm?
01:49:45kkurbjunyep, how close are some?
01:49:53kkurbjunI don't keep up on the tracker
01:49:56ali_asOwner of an archimedes for example.
01:50:00linuxstbkkurbjun: Nor do I...
01:50:05kkurbjunthere's that NES plugin, has anyone looked at that
01:50:16linuxstbDid that actually make it as far as a patch?
01:50:25kkurbjunI think so..
01:50:37linuxstbThere are probably plugins that still need adapting to some newer targets...
01:51:13amiconnali_as: Ah, that's what you mean. Nope, never even saw one
01:51:15 Join RoC_MasterMind [0] (
01:51:19kkurbjunyep, it's in here :
01:51:48kkurbjunit died in october though so it may be a way off from even compiling
01:52:02*amiconn started with a ZX Spectrum, later had an Amiga
01:52:25ali_asThe Amiga was always the arch rival to the Archimedes.
01:52:27amiconnI did code some Z80 asm on the Spectrum, but didn't code much at all on the Amiga
01:52:40kkurbjunthe doom scrollwheel patch could be one actually
01:52:51kkurbjunthat should be pretty trivial
01:52:59kkurbjunand I would be happy to answer questions
01:53:10ali_asI dabbled with the 6502 on the BBC B, but that was a horrible thing to program in assembler.
01:53:35ali_asARM was the first assembler I learned properly, really friendly.
01:53:41 Nick fxb is now known as fxb__ (
01:53:50linuxstbkkurbjun: I quite like that idea - and it serves a useful purpose as well..
01:54:28ali_asYou could also mix assembler and BASIC, which is genius.
01:54:56kkurbjunlinuxstb, it could also be interesting to see if a potential student is willig to contact us to ask for feedback
01:55:07kkurbjunit would be nice to have someone who is open to communicate
01:55:42SSnakeyou make me remember endless nights lost peeking and poking hex values without an assembler on my C64... ARGHH
01:56:04 Quit RoC_MasterMind (Client Quit)
01:56:57*amiconn ->sleep
01:57:06ali_asGnite amiconn.
01:58:45 Join lastebil [0] (
01:59:20SSnakegood night everyone
01:59:27 Part SSnake
02:06:33 Join Llorean [0] (
02:09:45ashesdo any of you have a lot of expirence encoding video for iriver h300's?
02:10:00ashesi'm doing fairly well with trial and error
02:10:45ashesbut there are some glitches... sometimes the video freezes for a couple seconds, or it goes haywire for a bit
02:10:53ashesthe audio never skips and is very good
02:11:07ashesi'm using mpeg1 and mp2/192k
02:11:30asheseven if i tone down the video bitrate to 150, there are still glitches
02:12:05ashesi have a lot of video i want to encode, and i want to perfect it before i encode everything i have
02:12:31ashesi'm encoding from authentic/original dvd source
02:12:34 Quit davina (Remote closed the connection)
02:13:05 Quit BigBambi (Read error: 113 (No route to host))
02:13:14Lloreanashes: Have you tried simply using the ffmpeg parameters from the WinFF presets on the wiki?
02:13:42ashesive been using mplayer, with help from the wiki
02:14:45ashesthe wiki example for mencoder works well, but not perfectly
02:15:36LloreanWhat framerate are you encoding at?
02:15:50scorche|shi was thinking of something a good bit simpler for the qualification task...just mainly set up a compiling environment, make a very simple (hello world-esque) plugin and compile...involves a few things and we can probably judge a good bit of how capable they are beyond that from an interview
02:17:15 Join JdGordon|uni [0] (i=82c20d67@rockbox/developer/JdGordon)
02:17:30ashesfps is set to 25, but its actually getting between 9 and 12 fps duing encoding
02:17:39JdGordon|uniLlorean: do you use the headphone pause feature?
02:19:34LloreanJdGordon|uni: Nope
02:19:37 Join Febs [0] (
02:19:52Lloreanashes: That's the encode speed, completely unrelated to the playback speed. Try setting to 20 fps.
02:26:18LloreanJdGordon|uni: IMO, if you don't want it to unpause on insert, stop it instead. :-P
02:27:00*JdGordon|uni just wants to decide how to close the patch... its a simple fix if it is indeed worth fixing
02:27:59LloreanIn all seriousness, I see how it could go either way. This is one case, I wouldn't mind a forum poll (or ML poll) just to see what the users want.
02:33:11 Join d34df00d [0] (n=d34df00d@
02:42:40JdGordon|uniive put it on the dev ml, i guess the forums would be good also
02:42:42*JdGordon|uni stas away from the user ml :p
02:42:57LloreanProbably the wisest plan.
02:43:00LloreanThey're very touchy there
02:43:16LloreanEven when I try to be very polite, *and* answer the question, if I even hint that their might be guidelines they chew me out.
02:44:10JdGordon|uniUI or general?
02:46:12LloreanLet's go General
02:46:29LloreanI'd almost even classify it Playback, but General is probably better for getting attention on this one
02:50:09 Quit fasmaie (Read error: 113 (No route to host))
02:51:40 Quit JdGordon|uni (" ajax IRC Client")
02:54:39 Quit sdoyon ("ircII EPIC4-2.6 -- Are we there yet?")
02:54:41 Join [CBR]Unspoken|w [0] (n=cbr@
02:55:28 Join pandrew [0] (n=andrew@
02:57:20***Saving seen data "./dancer.seen"
03:01:23 Quit XavierGr ()
03:01:42 Quit Davide-NYC ("ChatZilla 0.9.81 [Firefox]")
03:02:25 Join MobileShaid [0] (
03:02:42 Quit perrikwp (" ajax IRC Client")
03:04:26 Join perrikwp [0] (i=98215aef@gateway/web/ajax/
03:05:47 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
03:08:31Beta2KJdGordon, I use the headphone pause, why?
03:10:51 Quit CyBergRind|w (Connection timed out)
03:19:44 Quit Shaid (Read error: 110 (Connection timed out))
03:19:45 Nick MobileShaid is now known as Shaid (
03:20:28 Join perrikp [0] (i=98215aef@gateway/web/ajax/
03:20:39 Quit perrikwp (" ajax IRC Client")
03:20:57 Nick perrikp is now known as perrikwp (i=98215aef@gateway/web/ajax/
03:24:38 Quit perrikwp (Client Quit)
03:31:30 Join zwj [0] (n=chatzill@
03:38:13zwjcould i port rockbox as an Application to the OpenMoko platform
03:38:48krazykitzwj, certainly you could. the simulator uses SDL, which OpenMoko presumably has
03:40:26zwjmake rockbox can run in the sdl is ok?
03:41:10scorchethe sim already uses just need to adapt it to OpenMoko
03:43:25 Quit ali_as ("zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzZZZZZZZZZZZZZZZZZZZZZZZz")
03:47:37 Join sarixe [0] (
03:49:26 Quit sarixe (Client Quit)
03:52:37 Quit Seed (Nick collision from services.)
03:52:43 Join Seedy [0] (
03:53:23 Join sarixe [0] (
03:58:15 Quit hd (Remote closed the connection)
04:05:48 Join perrikwp [0] (i=98215aef@gateway/web/ajax/
04:20:33 Nick Seedy is now known as Seed (
04:20:54 Quit sarixe ("(EE) Failed to load "quit" module")
04:32:30 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
04:32:58 Join miepchen^schlaf [0] (
04:41:54 Quit merbzt (Read error: 110 (Connection timed out))
04:42:08 Join hd [0] (i=jd@unaffiliated/helldragon)
04:45:32 Quit Thundercloud (Remote closed the connection)
04:46:02 Join DC1 [0] (
04:46:53 Quit kugel ("ChatZilla 0.9.81 [Firefox]")
04:47:05 Quit DC1 (Client Quit)
04:54:13 Join mchua [0] (n=mchua@conference/pycon/x-d470f6eb7a369d4e)
04:57:24***Saving seen data "./dancer.seen"
04:59:26 Quit Rick (Read error: 104 (Connection reset by peer))
04:59:30 Join Rick [0] (
05:17:47 Join toffe82 [0] (
05:31:28 Join argumentD [0] (
05:32:43 Quit Horscht ("IRC is just multiplayer notepad")
05:33:13 Join lymeca [0] (i=lymeca@unaffiliated/lymeca)
05:33:25lymecaWhat video formats can Rockbox play on 5th gen iPods?
05:35:17cool_walking_See the PluginMpegplayer wiki page.
05:54:33 Quit zwj (Read error: 110 (Connection timed out))
05:59:31 Quit CaptainKewl (Read error: 110 (Connection timed out))
06:04:34 Join ol_schoola [0] (
06:06:55 Join FOAD_ [0] (
06:19:44 Join zwj [0] (n=chatzill@
06:21:34 Quit wisp_sh (Read error: 110 (Connection timed out))
06:24:01 Quit FOAD (Read error: 110 (Connection timed out))
06:24:01 Nick FOAD_ is now known as FOAD (
06:44:50pikhqCongrats on getting Google SoC slots.
06:45:34 Quit JdGordon ("Konversation terminated!")
06:47:40 Quit ol_schoola ()
06:54:16 Part toffe82
06:57:25***Saving seen data "./dancer.seen"
07:02:41 Join JdGordon [0] (
07:03:55 Join aliask [0] (n=chatzill@rockbox/developer/aliask)
07:11:26 Quit argumentD (Read error: 113 (No route to host))
07:14:33 Quit pandrew (Remote closed the connection)
07:18:36 Join wisp_sh [0] (
07:25:25 Quit JdGordon ("Konversation terminated!")
07:25:58 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
07:27:26 Part kkurbjun
07:27:30 Join kkurbjun [0] (
07:27:37 Part kkurbjun
07:30:40amiconnLinusN: The build system didn't pick up 2 commits at least, and they also don't appear on the frontpage...
07:30:55LinusNoh, badness
07:32:39 Join BrianInMaine [0] (n=38006718@gateway/web/cgi-irc/
07:32:55 Quit otih (Remote closed the connection)
07:37:12 Quit zwj (Connection timed out)
07:40:49LinusNfor some reason, the svn update was stuck since yesterday...
07:41:06amiconnThat already happened several times iirc
07:41:23*amiconn is kinda annoyed by those repeated build system failures
07:41:30LinusNme too
07:43:02amiconnYou never know whether the next commit works, and if none of you 3 is around, it may be that the committer doesn't have time to fix eventual problems when the build system is working again
07:44:18amiconnAnd you also don't see what delta your commit caused if several commits queued up meanwhile
07:44:36LinusNthat too
07:46:24amiconnWill you kick off a build?
07:46:33LinusNworking on it
07:49:40 Join zwj [0] (n=chatzill@
07:51:24 Quit BrianInMaine ("CGI:IRC")
07:54:12LinusNamiconn: booooh! red! :-)
07:55:00amiconnYeah, forgot to commit one file, it seems...
07:58:37 Quit Shaid (Read error: 110 (Connection timed out))
08:04:08 Join Rob2222 [0] (
08:22:03 Quit Rob2223 (Read error: 110 (Connection timed out))
08:22:06 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
08:23:29 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
08:24:03 Join scev0la [0] (n=c11efa52@gateway/web/cgi-irc/
08:25:19 Quit BigBambi (Remote closed the connection)
08:37:08 Join merbzt [0] (
08:40:38 Join Bger [0] (n=Bager@
08:49:49 Join ch4os_ [0] (n=ch4os@unaffiliated/ch4os/x-059673)
08:50:40KeypadWow, 32 GB Ipod touch... When did that come out ?
08:51:32 Join ender` [0] (
08:55:53 Join bughunter2 [0] (
08:57:09 Quit cool_walking_ (Remote closed the connection)
08:57:27***Saving seen data "./dancer.seen"
09:03:22GodEaterhas anyone come up with a plan for things a potential gsoc student must be able to do to be accepted yet ?
09:06:23 Quit zwj (Connection timed out)
09:11:01 Join JdGordon [0] (
09:11:04LloreanMaybe we should just ask for a resume, or whatever you Euro folks call such a thing. :-P
09:11:05 Quit stripwax_ (Read error: 104 (Connection reset by peer))
09:11:16LloreanI mean, their qualifications will somewhat depend on what task they undertake
09:11:39LloreanNew port requires somewhat different skills than Rockbox as an app which is somewhat different from usability study and redesign
09:13:33 Join Bagderr [0] (n=daniel@gateway/web/cgi-irc/
09:14:31 Join petur [50] (n=petur@rockbox/developer/petur)
09:14:43 Quit JdGordon (Client Quit)
09:15:03GodEaterLlorean: I was just looking at the ideas pages for a few of the other accepted organisations
09:15:15GodEaterand the xmbc one includes a list of things you should be able to do
09:15:26GodEaterand mostly, apart from their language of choice being C++
09:15:29GodEaterit looks a good fit
09:15:50LloreanI'm clearly tired, because when I first read that sentence, I thought you were saying they were all things *I* should be able to do. :-P
09:15:58LloreanA basic list would be good, though, yes.
09:16:10 Join JdGordon [0] (
09:16:11 Quit JdGordon (Client Quit)
09:16:13GodEaterhaha - yes, not your personally
09:16:39LloreanYeah, that looks like a very good list.
09:16:55LloreanI'm not sure I agree with that last one.
09:17:28GodEaternot entirely no
09:17:41GodEaterif you can have one which still gives a good level of commitment it wouldn't bother me so much
09:17:57LloreanRent still needs to be paid, etc, etc. And as long as they'll have the 35 hours / week, I don't mind if they're spending another 20 or more delivering pizzas, filing papers, or whatever other job they have as a student.
09:18:12LloreanThe key is "have X hours a week" and "be able to do the work"
09:18:12GodEateryeah, no issues with that here either
09:19:28LinusNstill, it doesn't hurt if the applicant really want the feature to be implemented properly, rather than just complete the project
09:19:45LinusNif you see the difference
09:20:03LloreanOne possibility is that we could just, rather than requiring "a patch", require that applicants include with their application their own proof of qualification relating to Rockbox, and let them decide the specifics (with suggestions such as "patch" or "documentation showing what you've researched already into this project, showing you're aware of the tasks and difficulty" or whatnot)
09:20:17GodEateryou mean write good quality code, rather than "aim to finish inside the deadline" ?
09:20:20LloreanLinusN: I agree, the applicant should be invested in the project.
09:20:31LinusNGodEater: something like that
09:21:19LloreanI especially like the "Project Tasks and Deliverables: Break your project into tasks or pieces. Describe in detail what you plan to accomplish and what the results will be." part of their application
09:21:48GodEaterthis quick check on their page also revealed that XBMC now works on linux too - which is interesting in itself :D
09:23:44 Join JdGordon [0] (
09:26:39 Join simonrvn_ [0] (i=simon@unaffiliated/simonrvn)
09:28:47 Quit simonrvn (Nick collision from services.)
09:28:48 Join simonrvn1 [0] (
09:28:59 Nick simonrvn1 is now known as simonrvn (
09:29:01 Quit mchua ("Holy shiny distraction, Batman!")
09:29:02 Quit simonrvn_ (Client Quit)
09:30:00 Join roolku [0] (
09:30:00 Quit JdGordon ("Konversation terminated!")
09:34:29 Join mrkiko [0] (
09:36:41 Join zwj [0] (n=chatzill@
09:39:02 Quit mrkiko (Remote closed the connection)
09:41:12 Join JdGordon [0] (
09:41:18 Join kies [0] (
09:49:10 Nick fxb__ is now known as fxb (
09:51:00 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
10:03:48 Quit JdGordon ("Konversation terminated!")
10:04:36LloreanBagder: Could we get a "These builds are generated from actively developed source, and at times may be buggy or unusable. We appreciate your feedback on any issues you may encounter." to the current build page? (Or something similar)
10:05:24Bagderra good idea, yes
10:06:14LloreanI think it'd go a long way to helping manage peoples expectations and reactions to problems.
10:06:26LloreanMaybe a few less "Rockbox sucks, what kind of product is this crap" statemens. ;)
10:06:54pondlifeOne important point for GSoC students is to emphasise how important communication is, at least with their mentor, if not with us all.
10:07:27pondlifeDon't know if that's been mentioned yet, but perhaps we can rework that into a pre-requisite.
10:07:28GodEaterLlorean: you know that guy you're arguing with in the iPod install thread is a journalist right ?
10:08:26peturpondlife: amen
10:08:49 Join JdGordon [0] (
10:08:58LloreanGodEater: I think I stated my opinion clearly, and do welcome his responses to my question. There's really no effective way for us to keep the current status of every port on the front page.
10:09:32linuxstbLlorean: IMO we should think about removing the 1G build from the supported list - if we had known about that problem before adding it, we wouldn't have done...
10:09:41GodEaterI realise that of course, I'm just pointing out that his experience as a tester is coming to Rockbox and the very first build he downloaded didn't work
10:09:48GodEaterhe has no idea that this is a *just* a bug
10:10:04Lloreanpondlife: What I'd personally like to see is "every day that you work on Rockbox, you should email your mentor with your status at end of work (multiple emails if you stop and start again are welcome, but not necessary if you just want to send one) and you're expected to work at least 5 days a week (like a full time job)
10:10:12GodEaterassuming it is just a bug of course
10:10:39Lloreanlinuxstb: I'd be okay with removing it from the supported list if it never worked.
10:10:55GodEaterthe main thing though, is that I wouldn't like to see him walking away from this discussion with a really negative view of our project, and then writing about it in his magazine.
10:10:58GodEaterthat would suck a lot
10:11:18pondlifeDid amiconn confirm that his 1G works? Making a score of 2-1 to the not-workings....
10:11:27Bagderrhe did, yes
10:11:28GodEaterhe did say his 1G works fine apparently
10:12:02linuxstbLlorean: That's the point - it _did_ never work, on certain 1Gs (IIUC).
10:12:18pondlifeSo, maybe the front page should say "some 1Gs"?
10:14:03 Part LinusN
10:14:08 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
10:15:05 Part LinusN
10:15:08LloreanGodEater: I've tried to explain a little more, does that sound better?
10:15:36 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
10:16:30linuxstblostlogic: ping
10:19:38LloreanGodEater: There's something like 36 unique install methods and/or hardware distinctions. It'd really be quite hard to keep some sort of constantly updated usability chart on all of them unless it was on the wiki, I think. And anywhere, including a wiki, is likely to have outdated warnings on it if we try
10:20:01LloreanI think a general clarity that "this is in development, while it should work, you shouldn't *expect* it to work" is a better solution, if we can make that clear in advance
10:20:24*linuxstb notices discussion about the headphone plug/unplug commit in #rockbox-community... now has a little top "disclaimer"
10:22:21 Quit JdGordon ("Konversation terminated!")
10:23:10 Join JdGordon [0] (
10:24:10 Join bughunter3 [0] (
10:24:11 Quit bughunter2 (Read error: 104 (Connection reset by peer))
10:26:51 Join MethoS- [0] (
10:28:05 Quit JdGordon (Remote closed the connection)
10:28:23 Join JdGordon [0] (
10:28:37 Quit JdGordon (Remote closed the connection)
10:28:47*Llorean just thinks attempting to have current status of various targets on the front page will lead to more confusion, not less.
10:28:55 Join JdGordon [0] (
10:29:19*Bagderr agrees
10:29:33LloreanI've probably explained it poorly though
10:29:44*Llorean added a disclaimer that it's all personal opinion, and not project policy
10:29:56dionoeaHello. You need to fix the link on
10:29:58 Quit JdGordon (Remote closed the connection)
10:30:04peturwell he has a point, if some models of 1gen do not work properly, it should be mentioned
10:30:09dionoea(Looks like it's missing a http://)
10:30:13LloreanI think if you start saying "This target is not usable right now" people will expect that you're testing, and thus all other targets *are* usable
10:30:14 Join JdGordon [0] (
10:31:53 Quit aliask ("ChatZilla 0.9.81 [Firefox]")
10:32:09peturLlorean: I feel this issue is not 'just' a bug, but a version of 1gen that is not working properly...
10:32:35linuxstbIMO, this kind of issue doesn't arise very often, and in this case there is no point leaving the target on the front page - we have enough users to test it, and it obviously isn't what we would like to call "supported".
10:32:47linuxstbSo I would say just remove 1G...
10:33:28 Quit XavierGr (Nick collision from services.)
10:33:39 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
10:33:40LloreanI agree that in this case there's no reason to leave it supported.
10:35:23LloreanBut I don't think every time a target stops working we should be toggling its presence on the supported page, or trying to keep it updated with current usability.
10:35:30 Quit JdGordon (Remote closed the connection)
10:35:46 Join JdGordon [0] (
10:35:53LloreanBuilds break for a day or a few regularly enough, and half the time they're probably fixed before someone who can change the front page even knows they're broken.
10:35:58linuxstbLlorean: I think we just need to treat things on a case-by-case basis...
10:36:46 Join UncleRemus [0] (
10:38:31LloreanWell, I think "Supported" should mean "It's supposed to work, file bug reports", and if it's not in the "Supported" list, that means "We don't want you filing bug reports, or asking for help outside of development-related channels"
10:39:18peturthat is not entirely true imho
10:39:26LloreanWhich bit?
10:39:44 Join Rob2223 [0] (
10:40:01peturwe could ask for help for ports in progress that aren't fully supported yet
10:40:35LloreanYes. In here and in the new-port thread.
10:40:56LloreanBut, for example, people shouldn't be asking how to install in the iFP on the user mailing list.
10:41:20Bagderrimo, "supported" means it has a supposed level of functionality that makes it useful
10:41:34peturwe need unsupported/inprogress/supported ;)
10:41:45Bagderraren't all in progress?
10:42:30LloreanBagderr: I've always seen it more as "we're willing to give support." The M:robe 100 has the level of functionality to make it useful, but it's not in the "supported" list yet, for example
10:42:50Bagderrwell, it'll be in the supported list once there's a manual for it
10:42:59Bagderrso yes, it is related to the support part
10:43:19Bagderrbut I'm thinking of ports that may exist but lack major funtions such as fm tuner, recording or similar
10:44:59 Quit zwj (Read error: 110 (Connection timed out))
10:44:59LloreanBut, for those targets, we provide support still. So they're "supported" in that sense.
10:45:43 Quit ch4os_ ("Lost terminal")
10:45:48peturlets give it a 'supported' number (1-5) or color?
10:45:48Bagderrright, so they are "supported" once we have a manual and core functionality and stability that makes them useful
10:46:04 Join gevaerts [0] (
10:46:12LloreanBagderr: One more web-type request. Could, in the table on the front page listing players, we get a second link to the TargetStatus page that says something like "View the status of supported and upcoming players" (or even replace that with this)
10:46:19LloreanThe TargetStatus page would be great for showstopper bugs, I think
10:47:10Bagderrwe had it there before, zagor removed it...
10:47:27LloreanI think it's rather invaluable, personally, but ah well.
10:47:35LloreanBagderr: I think we're discussing in circles. You're saying "we'll give support once they're supported (by your definition)" and I'm saying "if the point we call them 'supported' is when we're willing to give support, isn't that what the word ends up meaning?"
10:47:57 Quit JdGordon (Remote closed the connection)
10:48:09peturso set criteria for 'supported'
10:48:11BagderrI'm not arguing, I'm only trying to put into words when a port is to be considered "supported"
10:48:14 Join JdGordon [0] (
10:48:42 Join Buschel [0] (
10:48:55LloreanWell, I agree that there's a technical-status point when a target becomes "supported"
10:49:09LloreanBut I think it also means the point where we open up to people posting bug reports on it.
10:50:06pondlifeShould we lock the headphone-unpause vote on ? FWIW I misread the question, so one of the yes votes should be a no :)
10:50:11 Join MethoS-- [0] (
10:50:44linuxstbpondlife: I also misread, and I voted no instead of yes, so it's OK ;)
10:50:57pondlifeIt's fixed now anyway
10:53:16 Join OlivierBorowski [0] (
10:54:18markunWhen I pause music and then select a new file in the browser, the new file is loaded, but playback is still paused. Is this intended?
10:54:52linuxstbYou mean start a new playlist?
10:55:17pondlifemarkun: Yes, I think that's correct
10:55:28markunlinuxstb: yes
10:55:30pondlifeIf you manually paused, then it should stay paused until you resume
10:55:40markunI expected it to start playing my new song
10:55:56gevaertsNow that you mention it, that's been bothering me as well. I usually listen to half-hour radio recordings, followed by some garbage (start of the next program,...), so when I get to the end I pause and then go to look for what I want next
10:55:57pondlifeWhy not stop then
10:57:14markunpondlife: the fact that it can be solved in another way doesn't mean this is good behaviour.
10:57:24pondlifePersonally, I'd be happy to scrap pause completely, if we could get STOP to resume a bit more intelligently (i.e. don't clear PCM buffers)
10:57:31***Saving seen data "./dancer.seen"
10:57:43pondlifeFunctionally, STOP and PLAY could work the same as pause/unpause.
10:57:44 Quit Rob2222 (Read error: 110 (Connection timed out))
10:57:48markunI wouldn't mind that as well
10:57:54LloreanThere's a distinction
10:58:17LloreanIf you "Insert" while stopped, you get a new playlist. If you "Insert" while paused it doesn't clear.
10:58:35pondlifeI also would prefer not to have that distinction
10:58:40pondlifei.e. don't clear
10:58:45roolkupause should stay on the wps and stop return to the file/database tree
10:58:46pondlifeBut that's a preference
10:58:47LloreanThen how do you start a new, single song playlist?
10:58:58pondlifeSelect the song
10:59:07LloreanThat starts a dynamic playlist based on the whole folder
10:59:09LloreanNot just a single song.
10:59:14roolkuotherwise there is no indication what is going to resume
10:59:24LloreanAs well, how do you start a playlist with a tree of folders?
10:59:26pondlifeI rarely start a single song... as I said, that's a preference
10:59:29LloreanFor example, all of "Weird Al"
10:59:47LloreanSo, you're forcing people to not be able to do that with filetree?
11:00:00pondlifeIt's just my preference
11:00:09LloreanI know, but you're suggesting your preference be implemented.
11:00:10markunanyway.. I still think playback should be unpaused when you select a new file
11:00:13pondlifeNo I'm not#
11:00:26pondlifeI am not suggesting anything changes
11:00:48 Join tvelocity [0] (
11:01:05gevaertsHow about adding a new settings menu titled 'Pause behaviour' ? We have resume on headphone plug, resume on new song, ...
11:01:06markunindeed, just stating what would make you happy :)
11:01:09*gevaerts ducks under the table
11:01:10pondlifeI'm saying purely that I would prefer the simplicity. I don't use pause much as it breaks voice
11:01:13Lloreanmarkun: I agree, "select" on a new file should be treated as if you were starting playback fresh.
11:01:19LloreanThe previous state should be irrelevant.
11:01:41pondlifeWhy not have context options for both Insert... and New With... or whatever?
11:02:05Lloreanpondlife: It's a solution, yes.
11:02:08pondlifeI tend to use one never-ending dynamic playlist in general use
11:02:23JdGordonsomeone needs to come up with a good playlist UI to replace the current mess....
11:02:25LloreanMy point was merely "you can't just scrap the distinction between pause and stop, without making up for it, or you've cut a significant amount of related functionality"
11:02:26pondlifei.e. I keep it playing and insert stuff that I think I might like to hear soon.
11:02:43pondlifeI agree, but I was just stating a personal preference
11:02:56pondlifeNo plans to change it here, although it could do with simplifying
11:03:16markunand I thought that Llorean was Mr Semantics...
11:03:20LloreanWell, it was your statement thFunctionally, STOP and PLAY could work the same as pause/unpause." that doesn't read at all like "preference" since it's discussing actual functionality rather than personal use cases.
11:03:37pondlifeWhatever, sorry for the lack of clarity early in the day
11:04:00*Llorean figures if you're gonna call me Mr Semantics, I ought to live up to it.
11:04:16pondlifeI particularly dislike how easy it can be to clear a playlist by accident.
11:04:18LloreanAnyway, yeah, I agree it *could* be changed, and even think if done well it could make playlisting less confusing for users.
11:04:27pondlifeAnd the dynamic playlist warning has holes in it, I think
11:06:43 Quit MethoS- (Read error: 113 (No route to host))
11:06:53LloreanI think that you should be able to have two playlists "open" at once. The currently playing one, and one you're editing.
11:07:17pondlifeMultiple playlists open at once, ideally.
11:07:21JdGordonnot nice from a coding POV
11:07:32LloreanIf you want to start a new playlist, you can just start with a blank "editing" one, then make it active once it's ready. You can do this from "stop" without losing your previous dynamic one, by swapping them, and you can still have the blank list functinality.
11:07:42pondlifeBetter than special-casing a "dynamic playlist"...
11:08:04LloreanI'd say only two. The one that belongs to playback and a second one for editing if so desired.
11:08:31pondlifeWhy limit it? Why not have the playlist catalog integrated better?
11:08:43LloreanWhy have more than two in RAM at any given time anyway?
11:08:57LloreanYou can only edit one at once, and playback can only play one at once.
11:09:13LloreanIf you're moving tracks from one to the other, it can close one and open the other.
11:09:25pondlifeWell, technically only the playback one needs to be in RAM, no?
11:09:33pondlifeBut this is a UI thing
11:09:48 Join Rob2222 [0] (
11:09:48LloreanWell, you could make it appear like several are "available" sure. ;)
11:10:07pondlifeExactly - I'd like the UI to be identical no matter where the playlist is stored
11:10:16pondlifeAnd also not too dependent on playback state
11:10:20LloreanBut the "edit" one needs to be in RAM so you can "Insert" into it (some way of setting whether insert goes to the playback list or the active editable list)
11:10:22 Join merbz1 [0] (
11:10:47 Quit amiconn (Nick collision from services.)
11:10:53 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
11:11:23pondlifeThis is an aside though - my main concern is purely ease-of-use (as *I* see it).
11:11:52pondlifeLet's focus on getting a single playlist working perfectly first.
11:12:17 Join PaulJam [0] (
11:13:25pondlifeHmm, could "auto-change directory" become a playlist addition feature, rather than a playback one?
11:13:41pondlifeA bit like "recursive"
11:14:00JdGordonit should be there
11:14:35pondlifeauto-change dir doesn't make much sense when you're playing an m3u, and on some database queries.
11:14:57LloreanI thought it wouldn't work for either of those anyway
11:15:10pondlifeI attempts to work for the database, badly.
11:15:32pondlifeAnd, to be fair, it doesn't work properly with directoriesl at the moment.
11:15:44*pondlife spells badly
11:20:17*JdGordon wants to setup playlist so it will go on forever without removing the old tracks untill it runs out of memory for the new ones coming up
11:20:46JdGordonproblem is if the list is shuffled at all it will need fiddling when we want to add new files or mem will get fragmented
11:22:13 Join MethoS- [0] (
11:24:11pondlifeJdGordon: The playlist buffer is a dedicated area of memory though, right? Size set in system settings...
11:24:20pondlifeSo is fragmentation a problem?
11:24:29 Quit merbzt (Read error: 110 (Connection timed out))
11:24:34JdGordoncourse it is...
11:24:56JdGordonif we keep removing tracks from one end and adding to the other its fine
11:25:21JdGordonbut if we shuffle the list and leave the filenames in place (so we have pointers into the buffer) when removing/adding its gets fidly
11:25:51 Quit Rob2223 (Read error: 110 (Connection timed out))
11:26:00pondlifeShuffle would rewrite the whole list, maintaining only the current playback position, right?
11:26:18JdGordonbut thats slow
11:26:35JdGordoncompared to just shuffling indicies
11:27:02JdGordonalso, I want to add in a random mode which is shuffle but not shuffle the playlist
11:28:01pondlifeYou mean random jumping?
11:28:28pondlifeWhat advantage does that have over shuffle?
11:28:36JdGordonthe playlist doesnt get fudged
11:28:53pondlifeI like the next track info to be correct... and the playlist viewer
11:29:11JdGordonthe next track would be... playlist viewer wouldnt though
11:29:20JdGordonPV would show the playlist in the list order
11:29:23pondlifeSeems like a second shuffle method is over-complication to me...
11:29:27Bagderrand random mode doing "prev track" is also messy
11:29:42JdGordonthis is btw NOT to replace shuffle...
11:29:42BagderrI don't see any point with that either
11:30:01pondlifeI'm hoping to move auto-change dir out of playlist so that playback doesn't have to do these things.
11:30:03Bagderrbut not kiss...
11:30:13pondlifeout to playlist, I meant there.
11:30:24JdGordonplans to do it soon?
11:30:32pondlifeNo, as always...!!
11:30:36JdGordon:) ok cool
11:30:38pondlifeYou'll likely beat me to it..
11:30:56pondlifeBut any simplification is good.
11:31:40JdGordonalso, I want to split playlist.c up into current playlist handling and .m3u handling
11:31:47pondlifeGood idea
11:31:55pondlifeplaylist_() and playlistfile_()
11:32:22JdGordonnot sure how we will get away with being able to modify the playlist and arbitrary m3u's at the same time though
11:34:46pondlifeNo, we only have one buffer
11:34:49 Join fasmaie [0] (
11:35:32JdGordondump the current one to disk, then load the m3u, modify then write it back and reload the other one?
11:35:54pondlifeSounds unpleasant
11:36:04pondlifeJust have one at a time for now
11:36:34*JdGordon happy to do that :)
11:37:10pondlifeBut get the playlist API cleared out as much as possible, move auto-change dir, then look at the UI stuff... enough to keep you busy? :)
11:37:53JdGordon*grumbles something about not being paid enough* :p
11:39:09PaulJamif you work on the auto-change dir feature it would be nice if you could adress the issue in FS #7077
11:40:46JdGordonhmm... might be able to fix that now thanks to the event stuff Slasheri added :)
11:41:13JdGordonassuming i didnt misunderstand the report
11:41:16JdGordonwhich is highly possible
11:41:26pondlifeHopefully that, and Follow Playlist, will both be fixed by the simpler solution
11:41:32pondlifei.e. a loong playlist
11:42:27pondlifeDoes this mean that the recursively insert directories and auto-change dir options become combined?
11:42:54pondlifeNone / Recursive / Next / Ask or something
11:43:09 Quit MethoS-- (Read error: 110 (Connection timed out))
11:44:56JdGordongrr... how do you dir change on the e200 from the wps?
11:45:04JdGordonI tohught it was power+right?
11:45:13pondlifeShort right then long right?
11:46:17JdGordonbah, woops.. didnt have a dirplay list going ?:p
11:47:06JdGordonPaulJam: is there a reliable way to trigger that?
11:49:18PaulJamJdGordon: see the 4th comment for a way to reproduce. it always happens when the random mechanism selects a dir that contains no music iself, but one of the subdirectorys, so itisnt reliably reproduceable.
11:50:27JdGordonit is if my folder file only contains folders like that...
11:51:46PaulJamif it helps understanding, i could pastebin the changes i did in my source to fix this (i didn't post a patch because of the realname policy).
11:52:08 Nick fxb is now known as fxb__ (
11:53:19 Join ctaylorr [0] (
12:01:14 Quit MethoS- (Read error: 113 (No route to host))
12:01:35 Join Mathiasdm [0] (
12:04:29 Join Arathis [0] (
12:09:14 Join JdGordon|uni [0] (i=d31c5f5a@rockbox/developer/JdGordon)
12:12:48 Quit JdGordon|uni (Client Quit)
12:16:44 Join sbeh [0] (
12:16:53sbehis it possible to disabled recording?
12:17:23linuxstbYes, if you edit the source code and compile your own Rockbox build
12:17:52sbeham always compiling by my own
12:18:18sbeh(this could be an setting)
12:18:23sbeh(where are the rss-feeds?)
12:18:26linuxstbRemove HAVE_RECORDING from firmware/export/config-?????.h
12:19:31sbehlinuxstb: thx
12:20:13pondlifesbeh: Why would you want to do that?
12:21:47*Bagderr walks over to the "why?" camp and looks at sbeh
12:21:50sbehbecause i often accidentally press the record-button of my e200
12:22:40*petur always wondered about the need to couple the recording screen to a short-press button
12:22:58LinusNi think a long press is better
12:23:00pondlifeIt should be long press on all targets
12:23:24*pondlife finds it hard to use USB charging on H300 still ... ;)
12:23:35*petur seconds pondlife
12:23:50PaulJambtw, i notices in my h300, when i hold the REC button during startup to clean the settings, it enters the recording screen after boot. is this intended? i don't think this has always been the case.
12:24:03pondlifeNo, that's a similar bug
12:24:48sbehis there already a bugrequest for long-press-on-record?
12:24:50sbeh(where are the rss-feeds?)
12:25:03sbehbugrequest oO
12:25:09Bagderrthere are rss feeds all over the internet!
12:25:35sbehi think, they are only in the web
12:25:41 Quit Mathiasdm ("Yuuw!")
12:25:56Bagderrif web you mean http then yes, feeds are there
12:26:06Bagderrdid you perhaps mean any specific rss feeds?
12:26:09sbehif i say web, i mean web :]
12:26:28sbehBagderr: no, i dont meant an specific one, that is because i am asking here
12:26:29*Bagderr looks at sbeh
12:26:42*Bagderr stops talking to sbeh and goes back to work
12:26:53pondlifeI can't see a bug report, yet...
12:27:01sbehok, will create one
12:28:14*pondlife would rather not have a button for recording, even if it's marked with a "record" symbol.
12:28:15 Join thegeek_ [0] (
12:29:52sbehis central europe gmt+1?
12:30:25sbehi always forget such nonrelevant things
12:31:24Bagderrif it is irrelevant, then why ask?
12:31:30 Quit thegeek (Read error: 110 (Connection timed out))
12:31:46*gevaerts assumes that flyspray asks that on registration
12:32:30sbehbecause flyspray asked me
12:33:39 Join mrkiko [0] (
12:37:11 Quit ctaylorr (Read error: 110 (Connection timed out))
12:38:46 Quit Arathis ("Bye, bye")
12:40:16sbehsometimes greylisting sucks
12:40:41sbeh*waiting on registration email happy about disabled recording-button*
12:41:25Bagderrjust note that it'll end up with the other 500 feature requests
12:42:53 Quit fasmaie (Read error: 113 (No route to host))
12:43:29sbehi dont need that feature, because my record-button is disabled now, but the other way it looks cleaner to me :P
12:51:19 Part sbeh
12:52:57 Part Bagderr
12:56:05 Quit jhMikeS (Nick collision from services.)
12:56:11 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
12:57:34***Saving seen data "./dancer.seen"
12:57:38 Quit PaulJam (".")
12:57:52 Quit bblack (Remote closed the connection)
12:57:52 Quit perrikwp (Remote closed the connection)
12:57:52 Quit GodEater (Remote closed the connection)
12:59:05 Join GodEater [0] (i=c2cbc962@rockbox/staff/GodEater)
13:01:54GodEaterhave we had (m)any applications from potential gsoc students yet ?
13:02:41GodEateralso, is there much point in people who aren't on the commiters list volunteering to be mentors ?
13:03:27dionoeaApplications for SoC aren't open yet. (starts on the 20th i think)
13:03:38 Join webguest25 [0] (n=1485000d@gateway/web/cgi-irc/
13:03:44dionoeaErr, make that 24th.
13:03:46GodEaterdionoea: sorry, by "application" I meant "interest in applying"
13:04:12dionoeaAh ok :) Sorry.
13:05:20*gevaerts wishes he was a student
13:05:28*GodEater does too
13:05:49GodEateralthough only in this case. I wouldn't actually want to go back to studying =/
13:06:25gevaertsThere is that, unfortunately
13:06:59*gevaerts has to study anyway. Only now it's called a "Specification" instead of a "Textbook"
13:09:05 Quit webguest25 ("CGI:IRC (Ping timeout)")
13:09:09GodEaterI tend to find that feels less like studying to me
13:09:54 Quit feisar ("leaving")
13:10:26 Join Feisar [0] (
13:17:06 Join Thundercloud [0] (
13:19:38 Quit Buschel ()
13:26:04 Quit mrkiko ("leaving")
13:27:03 Quit GodEater (" ajax IRC Client")
13:27:34 Join GodEater [0] (i=c2cbc962@rockbox/staff/GodEater)
13:33:23 Join CaptainKewl [0] (
13:33:55 Join CaptainKwel [0] (
13:34:11*Nico_P wishes he had time for the GSoC this year
13:35:00 Join Stuffuh [0] (
13:35:57 Join Buschel [0] (
13:36:49 Join _thomas__ [0] (
13:37:03 Nick _thomas__ is now known as Vector (
13:37:39StuffuhIs there a way to search the irc log?
13:38:02 Quit Vector (Client Quit)
13:38:21GodEaterStuffuh: <your query here>
13:38:56 Join tilmann [0] (
13:39:52StuffuhDid you guys get a lot of questions when woot had that e250 sale?
13:40:32GodEaternot really
13:41:39 Join louse [0] (
13:44:51StuffuhGenerally speaking, which player has the best possible version of Rockbox available for it, and why?
13:45:19GodEaterwe get asked that question a lot
13:45:21Bagder"best" is a point of view
13:45:27GodEaterand there's a thread in the forums covering the answers
13:46:23StuffuhOk. thanks.
13:46:44GodEaterthe answer varies on who gives it, and what features are important to them.
13:48:14StuffuhIs there a page that has a wiki style comparison chart on it?
13:48:32Bagdersee BuyersGuide and DeviceChart
13:48:59pondlifeHmm, I get no pictures on ...
13:49:50Bagderthat's the new web site structure's fault...
13:52:11 Quit CaptainKewl (Read error: 110 (Connection timed out))
13:53:21 Join FOAD_ [0] (
13:57:22 Join seablue [0] (
14:01:22GodEaterBagder: do you have an opinion on mentors not being commiters? I'd offer my services as one as an interested and semi-competent person, but won't bother if you don't think it's worth it?
14:02:00BagderI'm fine with non-committers being mentors, as long as we're confident they're good
14:02:04 Join mf0102 [0] (n=michi@
14:02:49*GodEater notes this makes no mention of Badger's opinion on how "good" he is.
14:03:15*Bagder is silent
14:03:41GodEaterI shall sign up then, and wait and see whether I am accepted or not. I shall infer your opinion from this :)
14:03:55Bagderwe're currently on 7 listed mentors
14:05:26GodEater'tis done
14:05:54 Join LambdaCalculus37 [0] (n=44a0430f@gateway/web/cgi-irc/
14:05:54pondlife7 mentors? Any students?
14:06:24linuxstbApplications can't be made until the 24th (I think...)
14:06:33pondlifeAh, ok
14:06:39*pondlife should read the logs...
14:06:43LambdaCalculus37pondlife: Are there any sort of qualifications for students?
14:07:13Bagderwe'll just ask them a few questions
14:07:23GodEaterLambdaCalculus37: their IRC nicks mustn't start with L, or end in 7 :)
14:07:24pondlifeThere are. Google has their terms, plus Bagder has some questions.
14:07:30LambdaCalculus37GodEater: :P
14:08:09*GodEater thinks any sort of application requirements / questions should be put on the wiki page
14:09:02Bagderyes, I'll take them to the dev list too for a beating
14:09:32 Part d34df00d ("Konversation terminated!")
14:09:54linuxstbI think we need to be clear from the start about what criteria we'll be using to fail people. That seemed a problem last year - a reluctance to fail people...
14:10:17Bagdertrue, we should write up a "how to work on gsoc for Rockbox" page
14:10:23 Quit FOAD (Read error: 110 (Connection timed out))
14:10:23 Nick FOAD_ is now known as FOAD (
14:10:45 Join zwj [0] (n=chatzill@
14:11:11 Quit zwj (Client Quit)
14:11:20Bagderwith good inspiration from this =>
14:11:34linuxstbI was about to ask if you could remember that URL...
14:11:48markunlinuxstb: at google they told us to be really strikt, because almost none of the people who pass the first period even though they didn't do enough work make it to the end.
14:12:17GodEaterthat seems sensible
14:12:34GodEaterI think that any projects we accept should have very definite milestones to go with them
14:12:47GodEater(as definite as we can realistically make them anyway)
14:13:17markunBagder: do you think it could be that TEA5767 chips are labeled "TEA5760UK". It's sounds unlikely to me, but it's what kgb2008 who is REing the Meizu firmwares concludes.
14:13:30Bagderwhile we're eager to adjust from the somewhat laxed last year, we must not make this year too strict...
14:13:48GodEaterno, that too would be foolish
14:13:55Bagdermarkun: sounds very unlikely to me, but that's of course possible
14:13:56GodEaterwe're not out to discourage applications :)
14:14:14Bagderexactly, we must first lure them in
14:14:19Bagderthen tighten the screws!
14:14:54GodEater"yes, you're very welcome to come and work on Rockbox. Now please sign here, and initial here (in blood)."
14:15:06GodEater"thank you, we now own your soul."
14:15:10*LambdaCalculus37 pricks his finger and signs
14:15:16*gevaerts is suddenly happy he's not a student any more
14:15:18Bagder"use this container to send in a blood sample to the following address"
14:15:19 Quit Febs (Read error: 110 (Connection timed out))
14:15:37LambdaCalculus37What is your bidding, my Masters?
14:15:54 Join nplus [0] (
14:16:04pondlifeLambdaCalculus37: You must rewrite playback.c.
14:16:04Beta2KDont' forget to do a drug test
14:16:22BagderBeta2K: so what drugs are compulsory? :-)
14:16:24pondlifeIndeed, can't be having sobriety.
14:16:27LambdaCalculus37I... OBEY...
14:16:44Beta2KOnly the natural ones
14:16:52*linuxstb wants to know if mentors can get their students to make the coffee
14:17:07LambdaCalculus37linuxstb: I can make a mean pot of coffee.
14:17:07pondlifeBagder: Rocks, but not cracks
14:17:13Beta2KAnd they must be old enough to buy the mentors beer
14:17:26*petur takes note
14:17:32GodEaterthat sets an odd age range depending on where they are
14:17:49GodEaterwe'll end up with post-grads in the US
14:18:12pondlifeI can't imagine a place where students don't buy beer ;)
14:18:28GodEaterI can imagine a place where they don't buy it legally ;)
14:18:32lostlogicblah, data abort on my ipod 5g overnight
14:18:47LambdaCalculus37lostlogic: How'd that happen?
14:18:54lostlogicbeats me
14:19:05GodEatertrying to calculate Pi to the nth digit again eh ?
14:19:59gevaertsYou can only calculate that for even n
14:20:29lousehow does a binaural beat plugin sound? (project idea)
14:20:31pondlifeRelated to the forum report of an iPod deciding to spin the HD when stopped?
14:21:03lostlogicpondlife: don't think so −− it was playing
14:21:08pondlifeAh, ok
14:21:41lostlogicI should fsck it and reproduce before I bitch more, because I have been using the new USB stack ;)
14:22:01pondlifeEven so, it shouldn't data abort
14:22:47*gevaerts hopes the USB stack won't be blamed
14:23:29GodEaterlouse: sounds a little simplistic to me.
14:23:33LambdaCalculus37lostlogic: Is your iPod video an older 5G or a 5.5G?
14:23:54lostlogicgevaerts: probably not, although last evening when I tried to connect via rockbox USB, it hosed my usb-storage driver and required a system reboot when I tried to mount it.
14:23:58lostlogicLambdaCalculus37: 5.0G
14:24:33LambdaCalculus37lostlogic: Mine's a 5.5G. But the USB stack isn't giving me any nasty issues.
14:24:53GodEaterLambdaCalculus37: are you still using Full Speed? Or High Speed ?
14:25:00louseGodEater: yeah it might be hard to stretch that over 3 months :)
14:25:25LambdaCalculus37GodEater: Just Full Speed.
14:25:29lostlogicHmm, I'm using full speed, so didn't 'spect any issues, since I'd used ot successfully before.
14:26:03pondlifegevaerts: Any point in you mentoring a High Speed USB project? Or are you just missing a bit of PP magic?
14:26:43*pondlife mentions H300, USBOTG and GSoC in the same sentence, just for the fun of it.
14:27:12*petur slaps pondlife for triggering him
14:27:21pondlifeI didn't say beer!
14:27:28*petur slaps again
14:27:39*pondlife slaps again, too
14:27:42pondlifePoor again!
14:27:49gevaertspondlife: it's mostly missing some PP magic.
14:29:23GodEaterthere's probably not enough work in getting the stack working on the Gigabeat S either I assume ?
14:30:00gevaertsI hope not (same thing, some enable magic)
14:30:50gevaertsAnd anyway I'll be mostly (maybe entirely) offline for a month from about 2008-07-12
14:31:30*gevaerts doesn't mean by that that USB projects are impossible, just that he is probably not useful as a mentor
14:32:01peturthere can be two mentors...
14:32:10GodEaterno - but if that's the case, they're probably not useful as GSoC projects (at least the PP and S projects)
14:32:17GodEatersince the break through could happen in the first week, or not at all.
14:34:04lostlogicthe correct config for 'safe' USB Stack is to jsut define USE_ROCKBOX_USB right?
14:34:06*LambdaCalculus37 reads through the SummerOfCode2008 wiki page
14:34:36gevaertsIndeed. And even if a student spends three months looking for it, and succeeds in the end, he will have a successful project that produced 10 lines of code. I don't think anyone is looking for that
14:34:38LambdaCalculus37lostlogic: USE_ROCKBOX_USB is Full Speed.
14:34:59lostlogicright, high speed is an additional define and not 'safe'...
14:35:42GodEatergevaerts: well it's not that those lines wouldn't be useful, it's just that it would be hard to justify spending a load of google's money on something which may not produce a result at all, or all the work being done and dusted inside a few hours.
14:35:43gevaertslostlogic: It's 'safer', not 'safe'.
14:36:12gevaertsGodEater: indeed. And if I were a stdudent I would be a bit disappointed about it as well.
14:37:01 Quit Rob2222 ()
14:37:44*gevaerts notices he likes saying "indeed"
14:37:52GodEaterwhich is not to say that there might not be other USB based things worth doing
14:38:13GodEaterbut I think the AAP is already mentioned on the wiki page
14:38:27lostlogic:( connecting to rockbox USB and trying to do anything definitely seems to cause ... significant unhappiness ... on my linux box ATM.
14:39:33GodEaterlostlogic: which module is picking it up?
14:39:33GodEateruhci, or ehci ?
14:40:12lostlogicehci is for high speed, right?
14:40:17gevaertsGodEater: Ind^H^H^H Yes. There's still MTP and maybe Audio (I've started on that, but if someone wants it as a gsoc project I can find plenty of other things)
14:40:20lostlogicI don't have uhci built.
14:40:33GodEaterI could have sworn uhci is what grabs it on mine
14:40:49lostlogicthat depends on your USB controller iirc
14:40:55gevaertsGodEater: ohci and uhci are/were competing standards for USB 1.1
14:41:21GodEaterah ok
14:41:33gevaertslostlogic: did it work before ? What revision are you using ?
14:42:33lostlogicusing latest revision and yes, it has worked in the past, but I couldn't tell you exactly what rev did work :(
14:42:42GodEaterIs there real interest from people in Rockbox supporting MTP ?
14:43:58gevaertsThe main advantage of MTP over MSC is that you don't export the entire block device, so the FAT driver can stay active, which means the player can remain functional while copying files.
14:44:19Beta2KHow about Firewire for the iPods for GSoC? <ducks>
14:44:33GodEaterdoesn't it kind of fly in the face of the policy we have for allowing people to copy music to wherever they like on the device ?
14:44:40GodEateror am I being slightly naive here?
14:45:29gevaertsGodEater: I don't really know how MTP works, but I think you can still do that.
14:45:45GodEateryeah, I guess I'm going on experience of how other vendors implement it
14:46:05gevaertsBeta2K: why not ? The controller spec is available (but of course there might be "magic-enable-sequence" problems)
14:46:09LambdaCalculus37Beta2K: Actually, FireWire support would be useful as well, but depends on how easy it is to implement.
14:46:11GodEaterthinking about it logically, I suppose there's no reason why you can't actually keep that as part of the functionality if that's what you really want
14:46:34*LambdaCalculus37 had a notion about video playback
14:47:06Beta2KJust seemed to be a sore point from my pervious readings
14:47:20LambdaCalculus37Since we have Vorbis, FLAC, and Speex, what about another Ogg format: Theora?
14:47:20LinusNi think any project involving reverse engineering would be bad
14:47:21gevaertsGodEater: I'm not sure how MTP handles file names, but if it keeps them, MTP would be something analogous to NFS or SMB
14:47:22Beta2KToo many people asking and no one with the tiem to do it
14:47:41*gevaerts has an excuse : no firewire-player
14:48:13GodEatergevaerts: I think MTP *can* keep the filenames - it's just current implementations I've seen choose not to.
14:48:23GodEaterI could be entirely wrong though :D
14:48:25 Join Febs [0] (n=chatzill@
14:48:59linuxstbGodEater: I could be wrong, but my understanding is that the device has the freedom to store the files in any way it wishes.
14:49:14linuxstb(it may not even be given the original filenames - just the metadata...)
14:49:29GodEaterlinuxstb: that's my understanding too. My original question is null and void because I hadn't thought about it before I asked.
14:49:40*GodEater gives himself a slap on the wrist
14:49:46gevaertsGodEater: does it present a directory-tree like structure to the OS that's not absolutely based on album/title ? If so, it keeps them...
14:50:03*gevaerts has never used mtp
14:50:11GodEatergevaerts: the only recent implementation I've seen is that on the gigabeat S
14:50:28GodEaterwhich stores the files in one folder, with seemingly random (or possible sequence numbered) names
14:50:37linuxstbDon't some MTP devices also have a "data" folder - where users can copy datafiles to?
14:50:41GodEaterbut I assume it also updates a database on the player somewhere
14:50:54LambdaCalculus37linuxstb: I think the Creative devices do.
14:51:16linuxstbRockbox could simply create an "artist/album/nn-track.ext" file structure though.
14:51:25gevaertsI can imagine that if you do MTP mainly for DRM-reasons, and you use a database exclusively, you don't bother with keeping filenames
14:51:26GodEaterso perhaps an MTP implementation *would* be a good project for gsoc
14:51:34GodEaterif anyone were interested
14:51:53linuxstbGodEater: I think it would, especially as it seems to be on its way to becoming a genuine USB device class.
14:51:54 Join perrikwp [0] (i=98213b96@gateway/web/ajax/
14:52:12 Join perrikp [0] (i=98213b96@gateway/web/ajax/
14:52:22GodEaterfingers crossed for a student keen on doing it then
14:52:49gevaertsI'd say yes, if we can get USB to work properly.
14:52:55 Quit perrikwp (Client Quit)
14:53:08linuxstbI think it's quite a big project as well - the student won't just need to get the USB parts of it working, but also integrate it nicely with the Rockbox database.
14:53:40GodEaterlinuxstb: surely the database integration would be a Part 2 of the project ?
14:53:43LambdaCalculus37Also get media players to see the device and sync with it.
14:53:44*gevaerts still hopes someone will come along "today" (for daily-changing values of "today") with the correct magic
14:53:46GodEaterit doesn't *need* to integrate from the get go
14:53:56 Quit CaptainKwel (Read error: 110 (Connection timed out))
14:54:47GodEaterBagder: shame about curl :(
14:56:44 Join Rob2222 [0] (
14:57:38***Saving seen data "./dancer.seen"
14:58:00*GodEater decides it's time to go to the gym.
15:03:55 Quit perrikp (" ajax IRC Client")
15:04:13 Join perrikwp [0] (i=98213b96@gateway/web/ajax/
15:04:55 Join shotofadds [0] (n=1485000d@gateway/web/cgi-irc/
15:06:08 Quit XavierGr (Nick collision from services.)
15:06:19 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
15:07:55 Quit shotofadds (Client Quit)
15:10:06 Quit thegeek_ (Read error: 110 (Connection timed out))
15:14:27 Join SSnake [0] (
15:14:54SSnakehi all
15:16:52 Join desowin [0] (n=desowin@atheme/developer/desowin)
15:17:33peturall is not here atm
15:18:49SSnake...little news: xrick runs ok also on grayscale targets
15:19:14*petur likes the quote in gevaert's mail :)
15:19:35SSnakeanyone interested in trying it :P ?
15:20:16*gevaerts would like that quote not to be true, but knows from experience that it is (at least for mere mortals like himself)
15:21:08pondlifeIt's definitely true
15:21:13 Nick SSnake is now known as TheItalianJob (
15:21:18pondlifeI refer you to playback.c (again)
15:21:49pondlifeTheItalianJob: Did you get any further on the legal issues?
15:22:38pondlifeIt would be nice to get a patch up on Flyspray, once you've worked out how to externalise any copyrighted material.
15:22:39TheItalianJobabsolutely not. I though to put all dumped data in an external file...
15:23:26LambdaCalculus37Good idea.
15:23:36TheItalianJobwell actually i'm not so sure graphics data is really copyrighted..... i mean: those are a bunch of datas taken with dumped with an emu
15:23:55pondlifeGraphics are normally copyrighted.
15:24:08TheItalianJobmaybe graphics data it's not exacly graphics data cointened in the original game
15:24:19pondlifeGraphics, sound, level layouts/maps...
15:24:58gevaertsTheItalianJob: whether they are copied from emulator screenshots or from the datafiles doesn't make any difference.
15:25:49TheItalianJobmmm yes, you're right. It's just a reproducition of copyrighted material, so copyrighted too
15:25:55LambdaCalculus37TheItalianJob: Why not do what Doom and Packbox do now? Put the plugin into source, and have people get the graphics data on their own.
15:26:26TheItalianJobbecause i need some extra time to do that.
15:26:44TheItalianJobok i'll try to do that in the next few days
15:26:56 Quit perrikwp (" ajax IRC Client")
15:27:47TheItalianJoboh let me say many thanks to the writer of the grayscale library: it needed no particular effort to port to grayscale
15:28:02pondlifeThat'll be amiconn, mainly, I think..
15:28:17TheItalianJobjust looked to plasma.c and all went good
15:29:06gevaertslostlogic: (sorry, got distracted) can you try a build before r16694 ? I think it's unlikely that that caused the problems, but it's the only change in usb storage in 5 days...
15:30:46 Join perrikwp [0] (i=98213cab@gateway/web/ajax/
15:31:17TheItalianJobabout the dsp function i tried to use: it seems dsp_process went ok if i write 1024 samples, but whenever i try to write more samples it seems to write over the source buffer pointer i give
15:32:23 Nick TheItalianJob is now known as TheItalianPianis (
15:32:35 Nick TheItalianPianis is now known as ItalianPianist (
15:33:20ItalianPianist(...looking for some nice nick to register on the wiki :P suggestions appreciated )
15:33:32TuplanollaItalianPianist would be awesome
15:33:46 Nick ItalianPianist is now known as Italian_Pianist (
15:34:04gevaertsThe wiki rules say you have to use your real name there...
15:34:18Italian_Pianistoh ok. So let's change it
15:34:27 Nick Italian_Pianist is now known as PierluigiVicinan (
15:35:16desowinyou have to use real name in wiki, in irc you can use nick ;)
15:35:47PierluigiVicinanoh so i remembered good. I meant ircname on wiki registration...
15:36:01 Nick PierluigiVicinan is now known as Italian_Pianist (
15:36:40desowinand nick changes aren't much appreciated, I think
15:39:07louse/usr/local/arm-elf/lib/gcc/arm-elf/4.0.3/../../../../arm-elf/bin/ld: cannot find -lspeex-rockbox −−any clues?
15:41:33linuxstblouse: Is this a clean build from the current source code, or have you updated from an older copy of the source?
15:42:35louselinuxstb: let me try a clean build
15:42:44lostlogicgevaerts: will when I get to work.
15:43:15linuxstblouse: Make sure you re-run ../tools/configure
15:44:59lostlogicgevaerts: wait, no, I've ahd this problem with before and after that build −− I'm... a super slow reporter, so it'd hafta be further back. Will look more when I'm at work, plus at work on my windows laptop it doesn't force me to reboot when it breaks :)
15:45:37 Join ali_as [0] (
15:46:25 Quit desowin ()
15:48:04*LambdaCalculus37 tests FS #8680 on his iPod video
15:48:16gevaertslostlogic: ok. The culprit could also be r16656.
15:48:17 Join toffe82 [0] (
15:48:37 Quit corevette (Remote closed the connection)
15:49:30louselinuxstb: so far so good
15:50:00 Join desowin [0] (n=desowin@atheme/developer/desowin)
15:51:04louselinuxstb: bingo, thanks a lot!
15:51:42*gevaerts sees toffe82 and gets reminded of the gigabeat connectors...
15:52:54toffe82the price is in the wiki , there
15:53:21gevaertstoffe82: I've seen it. You should get a payment in a few minutes
15:54:15*gevaerts is extremely good at forgetting about mails that sit in his inbox if he doesn't handle them immediately
15:54:35markuntoffe82: I'll transfer the money later
15:56:17 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
16:04:49Nico_Ppondlife: I discovered FS #6958 because you just closed it, but I like the idea and I could adapt it
16:05:19Nico_Pmaybe a separate debug screen would be better though
16:05:21pondlifeI thought it was probably more use when playback/buffering was more complicated
16:05:50pondlifeAlso, it should scroll - we're no longer so limited on the number of files.
16:06:04Nico_Pyeah I was thinking of using a list widget
16:06:33pondlifePerhaps if you press RIGHT on the buffering debug screen?
16:06:52pondlifeBut that would be logf only; we don't need red deltas
16:07:40pondlifeOn reflection, I think it's not worth it.
16:07:46Nico_Pmaybe some functionalities of the debug menu (or the whole menu) should be in an ifdef of its own
16:08:24Nico_Pmaybe not... I dare hope buffering is working as expected ;)
16:09:07pondlifeDid you see the report with a long FLAC and "repeat one" giving odd results?
16:09:18pondlifeSound buffering-y
16:09:23pondlifeSounds.. even
16:09:28Nico_Pah yes
16:10:06Nico_Pamiconn: do you recall having any long FLAC files on the ipod?
16:13:22 Quit mf0102 ("Verlassend")
16:18:08 Part LinusN
16:23:08 Nick fxb__ is now known as fxb (
16:25:50 Quit scev0la ("CGI:IRC")
16:30:18 Quit nplus ("Leaving")
16:32:54 Join nplus [0] (
16:34:47 Quit Febs (Read error: 110 (Connection timed out))
16:36:34 Quit toffe82 (Read error: 104 (Connection reset by peer))
16:38:19 Join toffe82 [0] (
16:39:38 Nick fxb is now known as fxb__ (
16:46:43 Part pondlife
16:48:31 Join silvio_ [0] (
16:52:53 Quit wisp_sh (Read error: 110 (Connection timed out))
16:53:22 Join ol_schoola [0] (
16:57:40***Saving seen data "./dancer.seen"
17:07:06 Part Italian_Pianist
17:13:57 Quit perrikwp (" ajax IRC Client")
17:15:47 Join Hammer89_ [0] (
17:17:04Hammer89_does anyone know if I'll need to update my e250's bootloader? (haven't installed a new bootloader since sometime last year)
17:17:45 Join perrikwp [0] (i=98213cab@gateway/web/ajax/
17:18:57 Join daurn [0] (i=daurn@unaffiliated/daurnimator)
17:20:56*amiconn disagrees regarding the G1 supportedness
17:21:57amiconnIf rockbox doesn't work on some G1's, that's a (severe) bug. It works without problems on mine. Last year we had a similar problem with the Nanos, and then we also didn't remove them from the supported targets list
17:22:19 Join Stuffuh^ [0] (
17:23:30 Join mcuelenaere [0] (
17:25:10*amiconn would really like to see this bug fixed of course
17:26:13amiconnUnfortunately I don't have an idea what might cause it. It would be possible to provide 2 special test builds though. One that directly enters the HW info screen, to check the hardware revision
17:26:41linuxstbDoesn't the bootloader display the hw revision?
17:26:42amiconn...and a second one that directly enters the ports debug screen, and has wheel power saving disabled
17:27:28 Join PaulJam [0] (
17:27:54 Quit daurnimator (Connection timed out)
17:28:14amiconnlinuxstb: no
17:28:40amiconnBtw, does only the wheel not work, or the buttons as well?
17:29:13amiconnIf the buttons also don't work, it wouldn't be possible to display anything in the bl because you couldn't switch to verbose mode...
17:29:14linuxstbLine 272 of bootloader/ipod.c...
17:29:50amiconnAH, yes
17:30:09linuxstbAssuming it's not cut off the screen...
17:30:32amiconnI tried it on target. It goes away too quickly to be readable
17:31:13linuxstbFrom the forum thread - "the scroll wheel doesn't work. There's a response from the four surrounding buttons and the centre button (the backlight switches on when any of these are pressed), but the scroll wheel isn't being read at all."
17:31:33amiconnHalf a second or so. If you know where it will appear, you might manage to read it
17:31:36linuxstbThe user could delete the rockbox.ipod file though.
17:31:48linuxstbOr hold PLAY - to not load IPL
17:32:00 Join EspeonEefi [0] (i=espeonee@STRATTON-SIX-FOURTY-NINE.MIT.EDU)
17:32:18amiconnAh, yes, that works
17:32:26 Join Zagor [242] (
17:32:35linuxstbI'll post a message asking them to hold PLAY and report their hw revisions.
17:33:20amiconnIirc wheel powermanagement disabling was tested already?
17:33:29 Join Capn_Fish [0] (
17:34:17Capn_FishIs the FM tuner on my iRiver H10 5GB supported?
17:34:32ZagorI just ran into an annoying bug: I recorded about 20 minutes on my c200 and when I tried to stop the recording it did nothing. since "stop recording" is MENU|REPEAT on c200, I held it down. but a bit too long, causing the player to shut down. and now I have no file with the recording. fsck.vfat refuses to touch the partition. anyone got any ideas how to rescue my recording?
17:35:41linuxstbamiconn: The flyspray task had a patch which claimed to disable it, and that patch didn't solve the problem -
17:36:25linuxstbZagor: I thought you were the FAT expert? ;)
17:37:05peturchkdsk /f
17:37:34Zagorpetur: is chkdsk more foregiving than fsck.vfat? fsck says "Invalid disk format in boot sector"
17:38:08Zagorcan't hurt trying, I guess
17:38:37peturI don't know fsck.vfat very well, chkdsk has always helped me
17:38:52Zagoroh crap. "refresing database". I wonder how many of my clusters it stole :-(
17:38:53linuxstbamiconn: What's your 1G's hardware revision?
17:38:59peturif windows wants to mount it...
17:39:19peturZagor: ouch
17:39:22 Quit Stuffuh (Read error: 110 (Connection timed out))
17:39:26 Quit mcuelenaere ()
17:39:35 Join Stuffuh [0] (
17:40:03linuxstbZagor: I would have said to "dd" your c200's disk to your PC first...
17:40:11 Join mcuelenaere [0] (
17:40:15 Quit Stuffuh^ (Read error: 110 (Connection timed out))
17:40:35linuxstbZagor: You could maybe even just load that dump into an audio editor - assuming you recorded as WAV
17:41:21 Join Stuffuh^ [0] (
17:41:21 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
17:41:32 Quit ol_schoola ()
17:41:46LambdaCalculus37Capn_Fish: Yes, it's supported.
17:42:44Zagorlinuxstb: mp3
17:42:56Capn_FishLambdaCalculus37: Wonderful! Thanks much.
17:43:15 Quit Capn_Fish ("Leaving")
17:44:13linuxstbZagor: Maybe Rockbox never even flushed the buffer to disk then...
17:44:50Zagorhow much ram does the c240 have? it was ~20 minutes @128 kbit
17:45:02linuxstbI think it's 32MB
17:45:17Zagorcould be that then
17:46:28amiconnlinuxstb: That patch should have worked, although for a quick test it wouldn't be necessary to change so much
17:46:39amiconnI would just have killed 2 lines
17:46:47 Join wisp_sh [0] (
17:47:05amiconnMy 1st Gen has hw rev 0x00010001
17:47:25Zagorhuh "dd: reading `/dev/sda1': Input/output error" after 208 MB
17:47:40linuxstbCorrupted partition table?
17:48:29ZagorI don't think so. sd 3:0:0:0: [sda] 2006528 512-byte hardware sectors (1027 MB)
17:48:50 Join mcuelenaere [0] (
17:49:34ZagorI must say having any function on MENU|REPEAT is a really bad idea
17:50:03Zagorthe difference between activating the function and shutting off the player is something like 0.5 seconds
17:50:22amiconnlinuxstb: Btw, mine was a 5GB model originally. I just swapped the HDD for a 10GB one, because the original hdd was broken
17:50:52linuxstbZagor: The ipod is similar - long press on PLAY is "stop", very long press is shutdown. But it doesn't seem to be an issue for anyone there.
17:51:08amiconnThe problem is probably UI responsiveness
17:51:34amiconnThat seems to have gotten worse during the last time
17:51:39Zagorlinuxstb: how long is "very long"?
17:51:51amiconn3 seconds or so
17:51:57GodEaterI personally don't like that stop / power off are on the same button on the ipod
17:52:03GodEaterI very rarely use stop as a result
17:52:22Zagorthe c200 shuts off after about a second
17:52:27gevaertsZagor: the sansa c200 OF handles bad FAT bootsectors very badly for USB. I have no idea why, only that I didn't find a way to get out of it using 'standard' tools
17:52:27GodEateralthough I accept there are few other places to put it :)
17:52:33amiconnGodEater: It's the same on several other targets... and I never had the problem to shutdown by accident due to that
17:52:48linuxstbZagor: Seems to be a bug - POWEROFF_COUNT (in button-target.h) is 40 for the ipod, but only 10 for the c200
17:52:54amiconnI had that once in the jpeg viewer though, where Play scrolls down...
17:52:54 Quit petur ("real life")
17:53:09GodEateramiconn: this is likely because you avoid bitmapped themes, and therefore the ui is much more responsive for you
17:53:22Zagorlinuxstb: sounds ripe for change, yes
17:53:46*Zagor runs away
17:53:59amiconnGodEater: The UI responsiveness has gotten worse even with bitmapless themes. Like, why does rockbox spinup the disk *before* actually stopping audio when pressing stop??
17:54:17GodEatersearch me ;)
17:54:32GodEaterbut when using a bitmapped theme, it's hard to tell when stop has been registered
17:54:48amiconnAnd why is the statusbar delayed so much in lists, especially (but not only!) with voice enabled
17:54:51GodEaterand so it's easy to over hold the button and end up turning it off
17:55:46amiconnIt's always hard to tell, but since rockbox queues button events, you don't *have* to wait until you see/hear the reaction
17:55:51GodEaterthe only alternative I can think of really is to have an "off" in the menu somewhere
17:55:53GodEateralthough I realise this is not a popular suggestion
17:56:36linuxstbI assume you mean instead of very-long-play?
17:56:56amiconnThat's why I prefer buttons over touch interfaces any day - you can queue up input events pretty reliably
17:57:11GodEaterit hasn't bothered me enough for me to implement it in my own build yet though
17:58:14amiconnIt's just extremely convenient e.g. on archos Recorder to quickly press Off->Off->Off for stop+shutdown, and then put it away knowing that it *will* shutdown for sure
17:58:33*GodEater notices it is now officially "escape from work time"
17:59:13*LambdaCalculus37 still has four hours to go in his day =/
18:00:13 Quit Stuffuh (Read error: 110 (Connection timed out))
18:01:24 Quit EspeonEefi ("さよなら")
18:03:12 Join EspeonEefi [0] (i=espeonee@STRATTON-SIX-FOURTY-NINE.MIT.EDU)
18:05:54lostlogicgevaerts: 16656 seems to be the culprit for me
18:06:40 Quit Hammer89_ ()
18:07:02gevaertslostlogic: ok. I'll try to find a way to handle this better. Not sure how, as the previous handling was far from perfect as well
18:07:20lostlogicit's boost always versus boost only for transfers?
18:07:31 Join Hammer89_ [0] (
18:08:13 Join theli_ua [0] (n=theli@
18:08:20 Quit Hammer89_ (Client Quit)
18:08:49lostlogicI'm really irritated at linux' usb-storage for having _zero_ ability to recover from a _slightly_ missed timing...
18:09:04lostlogichaving to reboot linux on failure but not windows is pretty heinous.
18:09:11*amiconn thinks that a bootloader build should build the actual bootloader file, requiring no extra steps to make it usable (except on targets where we have to inject it into an OF file, i.e. the irivers)
18:09:35amiconnRight now it doesn't do that at least for the iaudios
18:10:07amiconnThe 'scamble' call must be different from main builds on those
18:10:20amiconn...and the actual output filename too
18:11:00gevaertslostlogic: always or "large" transfers. In your case apparently it also needs it for small transfers, which is only doable as "always" as far as I can see.
18:11:38gevaertslostlogic: And I haven't had to reboot linux at all since I started on this
18:11:47 Part theli_ua
18:14:03lostlogicgevaerts: hmm, what kernel? what modules? my usb-storage module gets hard locked every time...
18:14:09lostlogiceven rmmod -f won't free it up
18:14:45gevaertslostlogic: I'm on debian 2.6.24-1-686.
18:15:08lostlogicI'm on vanilla, so that should be quite similar
18:15:37 Quit XavierGr (Nick collision from services.)
18:15:48 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
18:15:57lostlogicdo you do any usb magic that I might not know to release the driver when the device acts up?
18:16:45gevaertsthe most I've had to do was unplug the device. What ehci chipset do you have ?
18:17:04 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
18:18:23lostlogicit's an ALI M1563 south bridge
18:18:44*amiconn had no problem with the usb driver locking up either on linux
18:19:17amiconnGiven my previous experience regarding linux and usb, that's a good thing
18:20:25gevaertslostlogic: I'm now on intel ICH6, and I have had pretty good experiences with NEC as well. I have seen VIA fail a few times on a specific project.
18:21:07lostlogichmm, interesting. I think I have a USB controller PCI card that I could use instead which might behave better.
18:21:29amiconnWhere could I find that info in linux?
18:21:41gevaertsamiconn: lspci
18:22:38amiconnWhat I thought, it's built into the chipset: ATI IXP SB400 USB host controller
18:25:05joshinHas the new USB stack been made the default yet?
18:26:48joshinOk, thanks. I'll keep making my own builds with it enabled.
18:28:29gevaertsjoshin: what player are you using ?
18:28:47*gevaerts is trying to get more info on what works and what doesn't
18:28:59joshine200 series.
18:29:12 Join quitte [0] (
18:29:48lostlogicgevaerts: for reference, the problem that I mentioned is specific to my ipod 5g, I've not seen it on my sansa\
18:30:04quittehi. I just bought a creative zen V. of course it'S not supported. so the first thing I dd was open it. next thing was crashing it ;)
18:30:27quitteis there a wiki page to write down the chip specs. and a place to search for similar devices?
18:31:45lostlogic is a good place to start
18:31:55lostlogicI don't see the Zen V listed there at all yet though.
18:32:10quitteright. that's why I went straight here
18:32:21 Quit LambdaCalculus37 ("CGI:IRC (Ping timeout)")
18:32:44 Join LambdaCalculus37 [0] (n=44a0430f@gateway/web/cgi-irc/
18:32:44lostlogicso just add it to the pre-development targets list and start a wiki page about the hardware you find.
18:32:46quittei'm waiting for the soldering iron to heat up to remove the shielding - then I can tell what this actually is
18:33:03quitteok. and just search the wiki for the hardware I find
18:34:29quittein case there are german hackers here that are interested: saturn hast them for 49€, the 4G version
18:34:29 Quit Buschel (Read error: 104 (Connection reset by peer))
18:34:39quittejust missed one :)
18:35:39lostlogichehe, my mom has one and I'm sure she wouldn't mind a port, because she already uses rockbox on her iRiver H120
18:36:48LambdaCalculus37lostlogic: Rockbox: Fun for the Whole Family! :)
18:36:53 Join Buschel [0] (
18:37:00quittecreative ca0209-HBG. nothing else inside except sdram and flash
18:37:35lostlogicheh, so now we wonder if they actually made their own SoC or whose they rebranded
18:40:21amiconnI need an idea how to nicely integrate a bitmap into 'scramble', without duplicating too much code
18:40:31amiconnFormat is the same as our mono bitmap format
18:41:17amiconnThat bitmap needs to be put into the iaudio bootloader header
18:41:32amiconnIt's the gfx that appears when flashing an M3
18:42:33amiconnRight now, you get an empty screen with just a progress bar when flashing a bootloader on M3 - a newbie might be scared...
18:43:33 Quit gevaerts ("work->home")
18:44:19*amiconn wants to get M3 bootloader v1 out the door
18:44:41 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
18:44:54 Join dan_a [0] (n=dan-mirc@
18:45:00 Join GodEater_ [0] (i=57507940@rockbox/staff/GodEater)
18:45:01amiconnADC is reliable now; I booted like 30 times and rolo'ed like 50 times without a single button failure
18:45:32amiconnBefore adc_close(), failure rate was about 50%
18:46:08quittemight be a sandisk e200v2
18:46:50 Join bblack [0] (i=81ae709a@gateway/web/ajax/
18:47:54 Join Buschel_ [0] (
18:48:00*Buschel_ kicks his WLAN adapter again and again
18:48:15 Quit mcuelenaere (Read error: 110 (Connection timed out))
18:48:28 Join csc` [0] (n=csc@archlinux/user/csc)
18:53:11 Join Stuffuh [0] (
18:56:34 Quit nplus (Read error: 110 (Connection timed out))
18:56:38preglowBuschel_: what you get for using wlan :)
18:56:54Buschel_mobility :)
18:57:42***Saving seen data "./dancer.seen"
18:59:40ali_asOn ARM is there a way of accuratly timing short sequences of instructions? I'm reading the datasheets and only finding low resolution CCLK timers.
19:02:58Buschel_preglow: and my wife hates cables lying all around the living room
19:03:00 Quit XavierGr (Read error: 110 (Connection timed out))
19:03:23Buschel_preglow: could you check the dsp routines?
19:03:58 Quit Buschel (Read error: 110 (Connection timed out))
19:05:03 Join mcuelenaere [0] (
19:05:22 Nick Buschel_ is now known as Buschel (
19:07:00 Quit perrikwp (" ajax IRC Client")
19:09:11 Join gevaerts [0] (
19:09:16 Join jgarvey [0] (
19:11:21 Quit Stuffuh^ (Read error: 110 (Connection timed out))
19:16:39 Join perrikwp [0] (i=98213cab@gateway/web/ajax/
19:16:49 Quit perrikwp (Client Quit)
19:17:02 Join perrikwp [0] (i=98213cab@gateway/web/ajax/
19:19:44 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
19:22:58preglowBuschel: i can check them, yes, but not actually try them, but i guess you have that covered
19:25:49Buschelpreglow: yes, they all seem to work fine from my tests.
19:25:50 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
19:26:58 Quit perrikwp (" ajax IRC Client")
19:27:28Buschelpreglow: btw, when taking a look into dsp.c I see a yield() which is called each 128 samples. Is this necessary? Does it harm anything or can this be moved outside the while()-loop?
19:29:22preglowBuschel: only one way to find out
19:29:46preglowBuschel: dsp should yield more often than each block, though, dsp can easily be required to handle blocks of several thousand samples
19:30:25 Join moos [0] (
19:32:13jhMikeSBuschel: I've had been thinking about that since I've been working on threading stuff. I thought perhaps a timed yield would be more least every tick or timeouts for threads will be too jittery.
19:32:28Buschelpreglow: even removing it did not show any negative effect −− at least not for playback.
19:33:30 Join mcuelenaere [0] (
19:33:51Buschelwhen assuming 44100Hz sampling rate we call yield() each 3ms, even more often while decoding into the buffer
19:36:25jhMikeSyeah, that's too much and only an average rate, not a burst rate. yielding when the tick changes would keep everything in line and not SR dependent.
19:36:25 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
19:36:58 Join perrikwp [0] (i=98213cab@gateway/web/ajax/
19:37:21preglowa timed yield would of course be the right thing to do
19:37:53 Quit svenVI ("Part messages are funny.")
19:38:09 Quit Buschel (Read error: 104 (Connection reset by peer))
19:38:33pixelmaZagor: reading the logs a bit, I have a few changes to the c200 keymap (vkeyboard, recording and radio screen, and a resume) here. I'm still not very satisfied with it though I prefer using it in my own builds, so there must be some advantage for me... and it has a short press of power to stop the recording. Wanted to put it into the tracker - at least for discussion - but somehow I keep forgetting :\
19:40:33 Quit desowin (Read error: 110 (Connection timed out))
19:43:30 Join mcuelenaere [0] (
19:43:34amiconnjhMikeS, preglow: mandelbrot yields once per tick. So if heavy zooming in mandelbrot doesn't make playback skip, once per tick should be enough
19:43:35LambdaCalculus37pixelma: I wouldn't mind trying out the changes to the keymap if you post them in the tracker.
19:44:57LambdaCalculus37What changes did you make to the keymap?
19:44:57 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
19:45:36jhMikeSamiconn: I read in the logs you had a theory about the str instruction alignment issue?
19:46:23 Join Buschel [0] (
19:47:03pixelmaLambdaCalculus37: if you remind me... ;) guess IRC it's not the place to explain them, would become too long an explanation
19:47:04amiconnjhMikeS: Yes. I guess there's a hardware bug that involves the cache controller (why doesn't that surprise me?)
19:47:07*Buschel cooled his WLAN adapter
19:47:23 Join RexDart [0] (n=cd9cbcfe@gateway/web/cgi-irc/
19:47:28 Join thegeek [0] (
19:47:36amiconnIf the core is put to sleep right at the start of a cache line, the cache controller might not be done fetching that line
19:48:18 Join newnick [0] (n=cd9cbcfe@gateway/web/cgi-irc/
19:48:18jhMikeSamiconn: but what about switch_thread being IRAM not making a difference? (iirc you said it didn't matter)
19:48:19amiconn...and it doesn't seem to like being put to sleep with unfinished work...
19:48:20 Quit RexDart (Client Quit)
19:48:32newnickAlrighty ladies
19:48:54amiconnThat's a question I'm not able to answer yet
19:49:02 Nick newnick is now known as RexDart (n=cd9cbcfe@gateway/web/cgi-irc/
19:49:22amiconnMaybe I did a mistake, or the cache controller and iram controller are somehow related
19:49:35RexDartConcerning rockbox official builds....user patches/hacks are sometimes implemented into the mainstream builds, right?
19:50:02amiconnIf that's the case, the bug should also occur with the cache disabled if switch_thread() is located in iram
19:50:05RexDartLike the one where you can hold the menu button on an ipod to prevent it from going to simple mode / data transfer mode when plugging in a USB cable
19:51:07RexDartIf so...I was reading about that particular hack as its own patch, and it was said that you could reverse the behavior of the menu button in that respect...should I grab that patch and apply it, or is there a simpler way to make that switch?
19:51:15jhMikeSamiconn: well, I do know no trouble occurs with a disabled cache though it could be from a decoding stage on whether or not to cache data.
19:51:34jhMikeS..or code
19:51:45amiconnDid you also test with caches disabled *and* switch_thread() in iram?
19:51:50 Join mcuelenaere [0] (
19:52:23jhMikeSnever both
19:52:27quittedamn. i deleted the demo videos on the zen V before looking what codec and resolution they are using. can someone tell me?
19:53:06pixelmagoogle? ;)
19:53:13jhMikeSI had done enough testing that I had to just put it away for awhile
19:53:19domonokyquitte: only rockbox talk here, ask somewhere else..
19:53:31quitteonly found a bunc of encoders for windows that supposedly work
19:53:48amiconnI might test that, but it will have to wait a while
19:54:02 Join gevaerts_ [0] (
19:54:07 Quit gevaerts_ (Client Quit)
19:54:11amiconnI'm still busy with the M3 port, and then there's that weird 1st Gen scrollwheel problem...
19:54:12 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
19:54:57pixelmaRexDart: guess you have to apply the patch and compile your own build. Depending on xour device you might be able to find an unsupported build for your player that already has it. But keep in mind, those are called unsupported for a reason...
19:56:02 Quit perrikwp (" ajax IRC Client")
19:56:14 Join austrian_guest [0] (n=5b7137a8@gateway/web/cgi-irc/
19:57:14 Quit GodEater_ (" ajax IRC Client")
19:57:54jhMikeSamiconn: incidentally, switch_thread was first moved out of IRAM in r15134 and that's when the problem started
19:58:25gevaertsRexDart: I'm not sure if you can actually reverse-apply an existing patch for that specific behaviour (the code has changed a lot in that area)
19:58:39*jhMikeS just check the history on that
19:59:06amiconnlinuxstb: Seems we have an answer why the scrollwheel fails....
19:59:32amiconnThat ipod reports a hw revision of 0xffffffff - meaning that the button code won't enable the wheel
19:59:58amiconnThis is because the code "thinks" it's a 2nd (or 3rd) gen, and those need no wheel enable
19:59:58 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
20:00:11 Join GodEater_ [0] (i=57507940@rockbox/staff/GodEater)
20:00:14amiconnA ROM dump of that ipod would be really interesting
20:00:44amiconnPerhaps it's a very early model, or the flash rom contents is somewhat damaged?
20:01:15amiconnCould it be possible that this ipod was never flash-updated?
20:01:52GodEater_anyone else here subscribe to Linux Journal (the american version) ?
20:02:12GodEater_there appears to be a 3 page article on Rockbox in this month's issue :)
20:03:22pixelmagevaerts: I thought RexDart talks about a patch he already found in the tracker (I think there is one that switches the behaviour of USB connect or charging while holding select or not on Ipods)
20:03:35gevaertslostlogic: I'm rereading the logs now to make sure I understand the usb vs boost problems. Do you get these problems at full speed (i.e. usb 1.1 speeds), or am I misreading things ?
20:04:50 Join desowin [0] (n=desowin@atheme/developer/desowin)
20:04:54*gevaerts waits for RexDart to clarify this, and won't say anything more on the subject until he knows what we're talking about
20:05:59lostlogicgevaerts: I am not using high speed.
20:06:17gevaertslostlogic: ok. just making sure.
20:06:31 Join Absinthe [0] (
20:06:48RexDartpixelma, I was wondering about rockbox official builds incorporating patches because it looks like this one has been incorporated
20:06:50amiconnHmm. There might also be a bug in the bootloader... I wonder what version bidmead installed
20:07:06RexDartwhich could mean a quick hex edit or something to switch this behavior around
20:07:46 Join mcuelenaere [0] (
20:07:48 Join Kreativethinkr [0] (n=485c7c61@gateway/web/cgi-irc/
20:07:48GodEater_hmm - do we claim somewhere that our music playback is better quality than the OF on the ipod ?
20:08:01 Join Mathiasdm [0] (
20:08:03RexDartbecause currently, ipod roxbox behaves as the description of that patch states
20:08:07pixelmaRexDart: could you clarify what exactly you are talking about? I'm not sure anymore that I understood correctly.
20:08:10preglowBuschel: are you absolutely certain that doing two samples at once will never do anything bad in the case of odd sample numbers?
20:08:18Kreativethinkrhow to get my pic flow it is not work ing right ?
20:08:43preglowBuschel: in any way, there should be a comment explaining that the code intentionally ignores odd cases
20:08:46GodEater_Kreativethinkr: define "not working right"
20:08:58 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
20:09:06Kreativethinkrsay no album art but i have alot
20:09:15amiconnThe bootloader bug theory can't be correct, as it also doesn't work in main rockbox
20:09:17RexDartHere, I finally found the patch I'm talking about:
20:09:17Kreativethinkrno album art found
20:09:32preglowBuschel: and please use either only "fd" or "ia" as a stm/ldm postfix, not both, i think "ia" is used most in rockbox code
20:09:52GodEater_Kreativethinkr: and your album art is arranged as per the AlbumArt guide in the wiki ?
20:09:58jhMikeSpreglow: how can you not have odd sample cases with pitch shifting?
20:10:47gevaertsRexDart: that patch won't apply any more. Too much changed since then
20:10:47 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
20:10:54preglowjhMikeS: it's not that, it's just that the code does everything in pairs, and ignores the odd case, it shouldn't matter, though, since bufer sizes as are always even
20:10:54pixelmaRexDart: you mean it reboots into disk mode when you hold select and plug the cable? Or do you experience that it's not rebooting even if you don't hold a button (which most probably has a different reason)?
20:11:02preglowjhMikeS: or...
20:11:18domonokyKreativethinkr: are you sure you have the album art in the correct way ? (take a look at the wiki)
20:11:35GodEater_Kreativethinkr: unless your AlbumArt is available on your player in the way described on the page - Picture Flow will not find it
20:11:56KreativethinkrThank i will lokk
20:12:39RexDartpixelma, if I hold Menu and insert the cable, it simply charges. If I insert the cable without holding Menu, it goes into disk mode
20:12:49preglowBuschel: also, adding and subtracting before shifting when doing the (a +/- b)/2 operations might make the data overflow
20:13:03jhMikeSpreglow: resampling can produce odd numbers of mono samples or stereo pairs on the output
20:13:08preglowthat's why it's done like a/2 + b/2 in the c code
20:13:10amiconnpreglow: Iirc we do use both forms. 'fd' for the stack, and 'ia' otherwise
20:13:38pixelmaRexDart: that sounds like it does what it is supposed to do (and it's been this way for a looong time)
20:13:42preglowamiconn: i just think the inc/dec before/after forms are easier to understand
20:14:10*amiconn thinks 'fd' is actually easier for the stack
20:14:31preglowwell, it's all down to habit, i guess
20:14:34preglowi don't care much anyway
20:14:39preglowat least as long as it's consistent
20:14:46preglowand it seems to be
20:15:17 Join OlivierBorowski_ [0] (
20:15:18jhMikeSresampling can produce 0 output samples but then the loop just ends there so the asm can assume at least one is available
20:16:09gevaertsRexDart: If you do your own builds, that's pretty easy to reverse (it should just be a 1-character change in the source).
20:16:16Buschelpreglow/jhMikeS: that's why I love reviews
20:16:19preglowBuschel: you should be able to do a/2 + b/2 without additional instructions, so that's preferred
20:16:22 Join mchua [0] (n=mchua@conference/pycon/x-e71c2facd6c413da)
20:16:51 Part mchua ("Holy shiny distraction, Batman!")
20:17:43*jhMikeS already went through the ringer doing the CF stuff
20:18:50 Join mcuelenaere [0] (
20:20:47 Quit GodEater_ (" ajax IRC Client")
20:21:42Buschelpreglow: yep, will do so (tomorrow). same should be possible for a/2 - b/2.
20:21:48preglowBuschel: yup
20:21:55preglowBuschel: apart from that, things look golden this far
20:23:28jhMikeSwho wanted karaoke mode in the first place anyway?
20:25:30 Quit Mathiasdm ("Yuuw!")
20:25:31 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
20:25:37 Join herrwaldo [0] (
20:25:42 Quit joshin (Read error: 110 (Connection timed out))
20:28:29gevaertsIs there an easy way to disable _all_ interrupts except USB, preferably for a longer period ? Would either the sansa sd driver or the ipod ata driver then still work ?
20:29:10Beta2KI doubt that either would work after
20:29:18 Quit OlivierBorowski (Read error: 110 (Connection timed out))
20:29:25jhMikeSgevaerts: such as?
20:30:00amiconnInterrupts should never be disabled for extended times. This would break several things (which ones depends on the target): buttons, timer tick + user timer, ...
20:30:05herrwaldogot a little question ... i'm using rockbox on a sansa c250 (i had a 240 before wich didn't had this issue but i broke it) when the screen goes in standy mode and i pres a button sometimes the backlight goes on but the lcd stays blank .. checked the original firmware and it doesn't have this behaviour ... a bug ? or am i doing something wrong
20:30:27amiconnAnd sleep() relies on the tick, hence pretty much any driver that uses sleep() for delays would stop working
20:30:36 Quit Rob2222 ()
20:30:58gevaertsI want to test if usb high speed gets more reliable if there's nothing else running. I don't really care if I have to hard reset afterwards, but I need disk access to do the test without too many changes in the code
20:31:43amiconnI doubt that it will work
20:32:27 Join argumentD [0] (
20:32:57herrwaldothe music keeps playing, the controls still work i can turn it off 'softly' and when i restart it it is fixed
20:33:04 Join mcuelenaere [0] (
20:33:05pixelmaherrwaldo: maybe that's somehow related to this bug If so you could try the attached patch or build added today
20:33:53RexDart<pixelma> RexDart: that sounds like it does what it is supposed to do (and it's been this way for a looong time)
20:34:18RexDartHow long? I'm not being picky, but I only remember it being in since a couple of months
20:34:30RexDartAnd yeah, it's supposed to...but I want to reverse it
20:34:49gevaertsRexDart: can you do your own builds ?
20:34:50 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
20:34:55*domonoky just applyed as a mentor for rockbox gsoc, although it depends a bit on the students project if i am of use..:-)
20:34:58RexDartI haven't tried yet
20:35:00pixelmaherrwaldo: because it was described the same (backlight comes on but screen is black)
20:35:14 Quit DataGhost ("NTOSKRNL.EXE caused a buffer overflow in System Idle Proce")
20:35:23RexDartBut I've got some minor programming/syntax experience and access to whatever tools are necessary
20:35:29pixelmaRexDart: what would you want to reverse it to? (What should it do)?
20:35:41 Join joshin [0] (n=joshin@VDSL-130-13-252-76.PHNX.QWEST.NET)
20:35:42herrwaldopixelma: indeed, sounds quite alike ... and is quite irritating
20:36:10RexDartcable + menu button = disk mode, cable - menu button = charge
20:36:27herrwaldopixelma: i've been using the orig firmware for a bit now and it seems like it doesn't have it so hardware seems ok
20:37:10 Join Rob2222 [0] (
20:38:03gevaertsRexDart: you need to change firmware/usb.c line 190 (if((button_status() & ~USBPOWER_BTN_IGNORE) == USBPOWER_BUTTON)). Change the == to != and it should do what you want. You still need to compile then of course
20:38:41RexDartawesome, thanks
20:38:47RexDartthat'll do it
20:39:22*domonoky thinks it would be good to have a seperate forum section for discussion with potential gsoc students.. :-)
20:40:05 Quit spiorf (Remote closed the connection)
20:40:46domonokynot only for potential students, but also while gsoc is running, to track progress... they could post status reports there... this would be more concentrated than somewhere in the ml..
20:42:16pixelmaRexDart: ah, at least I know now why I had a hard time to understand you. I thought you were talking about reversing a patch and not reversing the behaviour and as far as I know the official builds never had it the way you wanted it (so nothing to reverse code wise...)
20:42:44 Join mcuelenaere [0] (
20:42:56linuxstbamiconn: Interesting news about the 1G problem. I do know that the very early 1Gs had a slightly different firmware format on disk, and ipodpatcher can't handle it. So they most likely have been updated if ipodpatcher worked.
20:43:30amiconnHmm. What about the flash rom format?
20:43:36*amiconn wants to see a dump
20:43:49linuxstbI've no idea about htat.
20:45:18herrwaldopixelma: copied the build from the url u gave me, went in and out of standy like 20 times already ... seems it was exactly what i needed, tnx
20:46:43gevaertsherrwaldo: please add that information to the tracker if at all possible (together with the symptoms you saw), so the people working on that know about it.
20:46:57 Quit BigBambi (Remote closed the connection)
20:48:19austrian_guestHi - does anyone here knows how recording actually works?
20:48:55austrian_guestIn case anyone remembers me, I'm still writing an instrument tuner plugin
20:49:45herrwaldoaustrian_guest: Man that would be awsome !
20:49:52RexDartpixelma, it was hard to describe..I wasn't even sure if that particular patch was involved or not...
20:50:11austrian_guestBut when I try to record I get an error message I don't understand (DMA_REC_ERROR_DMA) which is -1
20:50:47gevaertsI wonder if I should apply as a gsoc mentor. I'll be unavailable from 2008-07-11 to 2008-08-10, and I'm still a new developer. On the other hand, I might be useful if someone wants to do a USB related project.
20:50:49jhMikeSaustrian_guest: I'm not sure why you'd get DMA errors
20:51:06austrian_guestAnd I don't have a f**king clue why I get that or how to avoid it... when i ignore it my X5 crashes with "illinstr..."
20:51:26preglowgevaerts: apply and see if anything pops up
20:51:47austrian_guestWell, actually I get -1 callback status, and in some other plugin this error is defined as DMA_RED_ERROR_DMA
20:51:49preglowgevaerts: absence can be handled by other people here, as long as you're available outside of that time
20:52:49jhMikeSaustrian_guest: IRAM init issue?
20:53:01austrian_guestI don't use iRAM at all
20:53:40scorche|shit would be nice to have enough mentors so that we can essentiall have 2 per student (even though only one can be the official one)
20:53:46 Join Stuffuh^ [0] (
20:54:04jhMikeSpreglow: do you have that flanger example online?
20:54:15austrian_guestI already got it
20:54:23gevaertsOK. I'll register
20:54:33austrian_guestbut it does not compile..
20:54:52scorche|shjust because you register, doesnt means you have to if you register, you get a feww shirt! ;)
20:54:54jhMikeSthere's an updated verion that should I believe preglow put up
20:55:13 Join spiorf [0] (n=spiorf@
20:55:57 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
20:56:07 Quit linuxstb (Nick collision from services.)
20:56:09 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
20:56:58*jhMikeS has gtg...bbl
20:57:04 Quit PaulJam (".")
20:57:46***Saving seen data "./dancer.seen"
20:58:04 Quit LambdaCalculus37 ("CGI:IRC 0.5.9 (2006/06/06)")
21:00:44 Quit spiorf (Remote closed the connection)
21:00:45 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
21:01:41 Join ViperShark [0] (n=vipersha@
21:02:00*gevaerts just applied as the tenth mentor
21:02:17 Join shotofadds [0] (i=51016aa7@gateway/web/ajax/
21:02:35 Quit toffe82 (Read error: 104 (Connection reset by peer))
21:04:07*shotofadds has interrupts working on TCC780x (D2) :-)
21:04:11 Join toffe82 [0] (
21:04:53markunshotofadds: congratulations!
21:05:47shotofaddsthis is just me messing around in the bootloader, so it'll need a bit of working into the main build. Still good though!
21:06:19 Join mcuelenaere [0] (
21:06:28austrian_guestcongratulations ;)
21:06:48 Quit seablue ("life, death, life, death")
21:08:39shotofaddshopefully not much more to do before buttons work in the menu :-)
21:08:45shotofaddsnow I just need a working NAND drive :/
21:08:46 Join merbzt [0] (
21:08:50 Quit Stuffuh^ ("( :: NoNameScript 4.2 :: )")
21:09:06 Join spiorf [0] (n=spiorf@
21:11:03 Quit Stuffuh (Read error: 110 (Connection timed out))
21:11:48 Join Buschel_ [0] (
21:13:55 Join piga [0] (n=leonardo@
21:13:59 Join low_light [0] (i=c730180b@gateway/web/ajax/
21:15:55austrian_guestOK, big progress! Recording works (well, at least a little bit), the pich of the input is calculated and displayed, but after ~10 secs rockbox crashes and shows me a dircache stack overflow error!
21:16:08austrian_guestWhy would this happen? I don't access the disc at all!
21:16:33linuxstbThat means your code is overwriting the stack used by the dircache thread.
21:16:48low_lightamiconn: did the datasheets provide any insight into the c200 lcd issue I mentioned yesterday?
21:17:09amiconnNo, unfortunately. It almost looks like a timing issue
21:17:23amiconnDid you try with just one NOP?
21:17:40low_lightMy patch is in FS #8552. For me it fixes the issue with mpegplayer and jewels
21:17:59low_lightyes, one nop
21:18:05 Join DrMoos [0] (
21:18:11pixelmalow_light: herrwaldo had a similar problem earlier, tried your patched build and said it fixed the problems
21:18:34kslaterhey, just loaded the latest rockbox for Sansa e200 and I notice that windows now sees it as USB mass storage
21:18:49 Quit moos (Read error: 104 (Connection reset by peer))
21:19:02 Nick DrMoos is now known as moos (
21:19:05gevaertskslater: in the official build ?
21:19:14 Quit merbz1 (Read error: 110 (Connection timed out))
21:19:20low_lightpixelma: yes, I see he commented on FS
21:19:21kslaterlatest one for sansa e200, yes
21:19:32 Quit piga ("Leaving")
21:19:33kslaternot, the latest stable build
21:19:49kslaterhowever, doesn't seem that I can see into the drive yet
21:20:09gevaertskslater: doesn't it reboot immediately ?
21:20:36kslaterreboot immediately if I open the drive folder?
21:21:53gevaertskslater: the USB mass storage support is still experimental, and should still be disabled by default. If you have it on an official build, that means something went wrong with the 'disabled' part
21:22:06*gevaerts investigates
21:22:28kslaterseems to be there and on
21:22:29 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
21:23:12low_lightpixelma: if your c200 doesn't have these problems, perhaps you can test the patch to be sure that it doesn't mess up anything else
21:23:42 Quit Horscht ("electromagnetic radiation from satellite debris")
21:24:06gevaertslow_light: I'll test as well, after I've found out why UMS seems to be enabled on kslater's e200
21:24:14pixelmalow_light: will do - and my c250 doesn't show these problems
21:24:48kslaterthat would be ultra-sweet if usb mass storage support worked
21:25:20low_lightok. shouldn't do anthing bad, it's a NOP afterall
21:25:40gevaertskslater: it does partly, but there are still bugs. See for details
21:27:00kslatergevaerts: this e260 I have with me is funky anyway. It doesn't read the battery state correctly. Doesn't disconnect properly with the orig fw, etc.
21:27:08Buschel_preglow: couldn't wait for tomorrow :o) new patch is up.
21:27:34kslaterI wanted to swap the display from this one to my other e250 with the good mb and bad display, but that looks like it would be really tricky
21:29:18gevaertskslater: there might be some tricky interaction there. The official build still behaves as expected on my c250, so unless I hear about someone else seeing this I'll assume it's related to dodgy hardware
21:29:33 Quit Kreativethinkr ("CGI:IRC")
21:29:46kslatergood plan. I'll load the same stuff onto my other device and see how it responds.
21:29:51kslaterthanks for the input
21:30:32 Join mcuelenaere [0] (
21:30:35linuxstbLlorean: - an attempt at a summary table for the current status of "supported" targets.
21:32:10Lloreanlinuxstb: Might change "No known issues" to "No issues preventing playback" or something similar. I'm not sure how to word it, but I'm afraid it might quickly fill up with regular bugs, like the misc playback bugs currently affecting all targets, etc.
21:32:54linuxstbI was thinking about that - maybe "No major issues". We should probably define what the different statuses are.
21:34:03LloreanOr maybe just have a "Currently Usable" column, with a check or an X, then "Description of issue" for the second column
21:34:25Nico_Plinuxstb: excellent idea (the table). how detailed do you think the notes should be (e.g. is RoLo not working on the F worth mentioning?)
21:34:54linuxstbNico_P: The intention is just for "show-stopper" bugs - like the Nano flash problems we had, or the current 1st gen issues.
21:35:10LloreanNico_P: I think it's for bugs like the PP5002 issue recently, the Nano issue before that, and the current 1st Generation iPod Wheel issue.
21:35:48linuxstbBut there's the opportunity for the target name to be a link to a target-specific status page if one exists. But the problem of course is the more we try to document things, the more things there are to keep up to date...
21:36:07*linuxstb goes to eat dinner - feel free to edit it
21:36:14 Quit Buschel_ ()
21:36:21*gevaerts edits linuxstb's dinner
21:36:54shotofaddsI wonder if those !s before 'Sandisk' are intentional :p
21:38:13Nico_Pshotofadds: I guess they were to prevent the word from appearing as a broken link
21:39:52amiconnSandisk won't become a link. If it is written SanDisk, it would, and then the ! is useful
21:40:48Nico_Pit's written "SanDisk", but for some reason the ! still appears
21:40:55shotofaddsbut shouldn't the ! be invisible?
21:41:11Nico_Pit should. maybe it's because of the | right next to it
21:42:09shotofaddshmm. do we have a twiki scratchpad page somewhere?
21:42:10 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
21:42:15LloreanIsn't there also a verbatim tag?
21:42:34shotofaddsverbatim makes text comes out in monospace, though
21:42:37RexDartgeez kslater
21:42:47RexDartlooks like the proper solution for you is to stop buying sansa players :P
21:43:12Lloreanshotofadds: Well, there's some tag that prevents linking, but doesn't change it to monospace, other than the !
21:43:19Nico_Padding a space between the '|' and the '!' makes the '!' disappear
21:43:21*Llorean has used it before.
21:43:28Nico_PLlorean: it's <nop>
21:43:37LloreanThere you go.
21:43:44LloreanFor if you don't want to add that space.
21:44:05pixelmalow_light: something specific I should watch out for?
21:44:11*Nico_P just added the space
21:44:30gevaertsIs the manual supposed to have the virtual keyboard keybindings in section 4.1.3 ? If so, the c200 manual doesn't have them
21:44:49 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
21:45:07Nico_PLlorean: what do you think of making each target name a link to the port page?
21:45:44LloreanNico_P: Probably a good idea.
21:46:33LloreanThough I'd put a warning at the top of the port pages to the tune of "This page is often out of date, as port pages usually aren't maintained frequently once a port becomes supported"
21:48:21Nico_PLlorean: I'll add that warning underneath the table... can't say I feel like editing individual port pages right now
21:49:04low_lightpixelma: probably nothing ;)... mpegplayer and jewels were the two plugins that triggered the bug on my player
21:49:39*low_light watched elephant's dream to test mpegplayer
21:49:43 Join mcuelenaere [0] (
21:49:45low_lightI don't get it
21:51:33gevaertslow_light: seems to work fine on my c250
21:52:17amiconndomonoky: Could you add M3 bootloader installation to rbutil?
21:52:47domonokysure.. is it the same as m5/x5 ?
21:53:16amiconnThe file follows a slightly different naming scheme: cowon_m3.bin
21:53:23amiconnOtherwise it's the same
21:53:40domonokynice, so i can reuse the the install routine :-)
21:53:52amiconnI guess we need some detection info?
21:54:09amiconnIs iaudio detection based on usb ids?
21:54:19domonokyyes, USB Ids would be good..
21:54:45pixelmagevaerts: re. the virtual keyboard table - yes, it's supposed to be there and I'm quite sure it once was (I added it), not sure when it vanished again...
21:54:47domonokyusb ids is the first we check, if that doesnt find a device, we use other methods..
21:54:55amiconnI'll just commit logo embedding in scramble.c for iaudio, then check the id
21:55:31gevaertspixelma: it made it difficult to put my usb-enabled build back after testing the official build without booting the OF.
21:55:45low_lightgevaerts: good, thanks.
21:55:48 Quit Absinthe ("Leaving")
21:55:49*gevaerts doesn't edit text that often on the c250
21:57:40domonokywhats the screen size of the m3 ?
21:57:57pixelmaeh, the statusbar in the mpegplayer plugin is inverted for me (white on blackor maybe it's the now default lightgrey on black) while I use black on white. I guess that it's not related to your patch low_light - maybe some viewport in list bug (haven't used mpeplayer since)
21:58:13pixelmac250, of course
21:58:51pixelmadomonoky: 0x0x0 (or 128x96x2 with attached remote) :)
21:59:23amiconnThe remote LCD is handled as the main LCD in rockbox
21:59:27domonokythere are no themes for 0x0x0 :-)
21:59:41*gevaerts volunteers to make those
21:59:45amiconnThere are still a few quirks to iron out, but it's already usable
22:00:02 Quit Nevtus (Remote closed the connection)
22:00:19domonokynow i only need a usbid for the m3 (only had to edit the rbutil.ini for m3 support) :-)
22:00:23amiconnquirks are e.g. that the main hold switch controls the 'backlight on hold' status
22:01:00amiconnAlso the vkeyboard doesn't work (probably just a keymap issue)
22:01:06 Join Nevtus [0] (n=Nevtus@unaffiliated/nevtus)
22:01:26*amiconn goes digging in the wiki for the usbids tools
22:01:44*gevaerts recommends lsusb in linux for that
22:01:50pixelmalow_light: other than that, I can't find any problems. So probably your patch works fine.
22:02:07domonokyamiconn: bottom of this:
22:02:08 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
22:02:47low_lightpixelma: here status bar is correct for mpegplayer, but it's inverted in the sudoku menu
22:03:30pixelmasudoku menu is fine for me
22:05:01preglowBuschel: excellent
22:05:09amiconndomonoky: And now the ID is on that page :)
22:06:02preglowBuschel: did you have a look at the sv8 musepack source code, btw?
22:06:19amiconnVery simple numbering scheme...
22:06:25pixelmalow_light: chopper menu is inverted too, and my personal sokoban has lost its colour...
22:06:30domonokyamiconn: and now rbutil supports it.. :-)
22:07:13pixelmagevaerts: if you still have it, could you test that too? I wonder if it's a fault on my side (playing with the graphics) or low_light's patch
22:07:30amiconnNow we just need that bootloader on the downloader server(s)
22:07:50Buschelpreglow: no, not yet. I downloaded, but never unzipped it
22:07:52pixelmagevaerts, low_light: I now think it's my fault
22:07:55low_lightpixelma: is the status bar color hardcoded? seems if i switch themes mpegplayer always has a black statusbar bg while sudoku always has a white statusbar bg
22:08:34 Join DefineByte [0] (
22:08:35Buschelpreglow: did you?
22:08:57amiconnI think the statusbar uses what the default viewport has set
22:09:06pixelmano idea, it probably shouldn't ... but that would explain why it seems correctly for me in most cases
22:09:10amiconnThe lists still don't behave
22:09:12*Nico_P looks at the BuyersGuide and notices the H10 are dirt cheap
22:10:05 Join mcuelenaere [0] (
22:10:28low_lightchopper and star seem to always have a black statusbar bg
22:11:50pixelmalow_light, amiconn: the "white" on black is actually lightgrey on black (I can see a slight difference from an angle compared to my plain white background) - as far as I know the default colour settings are now lightgrey on black (cabbiev2)
22:11:54low_lightanyway, thanks for the feedback...gtg
22:12:00 Quit low_light (" ajax IRC Client")
22:12:38pixelmathat is in the plugin statusbars that use usual menus
22:13:10Bagderis it time to make the mrobe100 builds more official?
22:13:13pixelmahmm... or only in some (e.g. in the wormlet menu)
22:15:10Bagder9 mentors in the list
22:15:17Bagder(gsoc I should add)
22:16:18preglowBuschel: barely, but long ago
22:16:28preglowBuschel: i think i remember thinking it wouldn't be too hard to port, thiugh
22:16:43 Quit austrian_guest ("CGI:IRC (Ping timeout)")
22:16:48Buschelpreglow: just browsing the repository
22:16:54preglowBuschel: anyway, personally won't start merging before it's out of beta
22:17:34amiconnBagder: Hi, could you put the M3 bootloader on the download server(s)?
22:17:38Buschelpreglow: sysnthesis filter stayed the same :o) so, all the optimizations I did are still valid
22:18:21Bagderamiconn: do you have a binary around, or should I build one?
22:18:55amiconnThe different file naming scheme is correct
22:19:32amiconnThere are also v3 bootloaders for X5 and M5, but those aren't that important, and I'll soon make a v4
22:19:53Bagderok, m3 bootloader is now uploaded
22:19:53amiconn...which will have proper hold switch handling (M3 has it right from the start)
22:20:08Buschelpreglow: I wonder why EQ_TAP, FIR_BANDS, ... are still defined in the code. I've added these some years ago for a winamp-plugin −− only used for "fast EQ", nothing to do with the decoder or encoder
22:20:13preglowBuschel: i think only the huffman part was touched, and some bitstream changes
22:20:17amiconnI can't test X5V though, do you know who could?
22:20:29preglowBuschel: yeah, i've noticed
22:20:30gevaertsI just added CPU_INT_EN = USB_MASK; CPU_INT_CLR = ~USB_MASK; to the USB init sequence, and then copied 1GB from the sansa with only one reset. It needed a hard reset afterwards of course.
22:20:37Bagderamiconn: no, I don't
22:20:45BagderI can't recall any dev with one of those
22:20:57amiconnMaybe Linus added X5V support blindly?
22:21:17amiconnI think only the id string and filename are different
22:21:17Bagderyes we did and someone verified it
22:21:31Bagdercould've been on iaudiophile or similar
22:21:33*amiconn will do a binary compare on the v2 bootloaders
22:21:33preglowgevaerts: high speed?
22:21:40gevaertshigh speed
22:21:47Bagderit's only the strings and file names iirc
22:21:49moosamiconn: aren't suposed to be strctly the same? (just fm radio)
22:21:56preglowsucks if we have to disable interrupts for transfers :/
22:22:15Bagdermoos: the bootloaders don't support radio ;-)
22:22:18amiconnpreglow, gevaerts: I doubt that this is needed
22:22:37moosBagder: that's why I thought that's the same :)
22:22:37 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
22:22:49amiconnThe fact that there's one reset tells that interrupts aren't the real problem
22:22:55Bagdermoos: they differ only because the hw requires them to
22:22:59amiconnThere should be no reset at all
22:23:10 Quit toffe82 (Read error: 104 (Connection reset by peer))
22:23:16gevaertsI doubt it as well, but it _is_ some kind of progress. I'll try with a hub next
22:23:55roolkubagder: could you put a new m:robe bootloader on the server please?
22:24:16Bagdersure, you have one around?
22:24:19 Join toffe82 [0] (
22:24:27roolkuBagder: just a sec
22:26:23roolkubagder: the manual is done as well - if you want to make the port official
22:27:07DefineBytedoes anyone hand out cigars when a new target is supported?
22:27:55moosbeers are more common here ;)
22:28:03 Join Absinthe [0] (
22:28:03BagderI do want that
22:28:12Bagderis there even a small pic for it?
22:28:30roolkuBagder: I have committed one in the smallpics folder
22:28:44pixelmaroolku: nice work filling out those button tables. I guess the mrobe100 manual is more complete in this regard than some others. I remember how surprised I was to find so many blank ones (especially in the Gigabeat and H10 manual) when updating the c200 manual
22:29:08roolkuBagder: I didn't dare messing with the perl scripts as I can't test them locally
22:30:26roolkuroolku: thanks, I used the gigabeat manual as a comparison so I spotted a few things here and there
22:30:31 Quit silvio_ ("Verlassend")
22:30:31 Join mcuelenaere [0] (
22:30:37roolkudoh, that was for pixelma
22:31:05pixelmaroolku: talking to yourself? ;)
22:31:37roolkupixelma: I think there might be still a mistake here and there, but hopefully someone will complain
22:32:55roolkupixelma: the viewer keymap seems a little out of date for all targets...but I couldn't face sorting that out at the time
22:33:18pixelmawell, the manual has some mistakes (in general)
22:34:08pixelmaroolku: the complete text viewer chapter is a bit out-of-date... :\
22:35:39 Join jac0b [0] (
22:35:53 Join Febs [0] (n=chatzill@
22:36:08Bagdermrobe100 is now present on the build download page
22:36:14roolkunice :)
22:36:19 Part jac0b ("Ex-Chat")
22:36:26Bagderthe picture should appear too soonish
22:37:10pixelmaI hope to be able to draw an mrobe100 and m3 svg the next days...
22:37:11 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
22:37:50roolkupixelma: that would be nice. my effort at labelling looks a little clumsy
22:38:27Buschelgood night all
22:38:36roolkuBagder: will you do the link to the manual as well? I guess it is not being built automatically yet
22:38:38 Quit Buschel ()
22:38:49amiconnBagder: The id string is indeed the only difference between the X5 and X5V bootloaders
22:38:51Bagderyes, I'll fix that as well
22:41:32moosBagder, roolku: I guess a change in the rockbox frontpage is needed too
22:43:00 Quit tilmann ("leaving")
22:43:45 Join linuxstb_ [0] (
22:47:16Nico_PBagder: does the www server use files directly from SVN now?
22:48:35 Quit EspeonEefi ("さよなら")
22:50:12Bagderit does, but it doesn't update automatically but only on demand
22:50:33amiconnBagder: Brand new v4 bootloaders for x5, x5v and m5 :)
22:50:56Bagderok, I'll update
22:51:19 Join robin0800 [0] (
22:54:49 Join devurandom [0] (
22:55:54 Quit jgarvey ("Leaving")
22:56:23devurandomIs this rbutil bug know? "utils.cpp:59: error: ‘getenv’ was not declared in this scope"
22:57:40linuxstb_Where are you compiling it? (Windows, Mac OS X, Linux, ...)
22:57:49***Saving seen data "./dancer.seen"
22:58:01Zagorpixelma: please post your button patch. I'm also rather annoyed at the c200 radio screen, so I'd like to see how you've changed it
22:58:32*linuxstb_ wonders why the WPS needs both %x and %xl/%xd
22:59:02roolkuBagder: did you see the last post in the mrobe 100 forum thread? - is it possible that he has an OF mi4 file with a different (unknown to rockbox) checksum? Maybe I should ask for it for analysis
22:59:33Bagderyes please do and check it with mi4code
22:59:49Bagderor I can check it
23:03:01gevaertsamiconn: this disable-interrupts thing doesn't solve the hub issue at all. I also tried again without hub, and I got 2 resets. Testing with interrupts enabled gave 22. On ipod it locked up compeletely after a few megabytes. I guess that one doesn't like having no interrupts
23:04:27amiconnAs I suspected...
23:04:53 Join EternalRains [0] (
23:05:03gevaertsI think it does provide more "proof" that it's not a voltage issue anyway
23:05:28amiconnWhat makes you think so?
23:06:04gevaertsMaybe I should tone that down. It's not _only_ a voltage issue. Timing or clocking seems to be involved as well.
23:06:10 Join einhirn [0] (
23:06:31 Join kugel [0] (i=kugel@unaffiliated/kugel)
23:06:48Bagderroolku: manual present!
23:07:06 Quit desowin ()
23:07:09roolkunice :)
23:08:03pixelmalinuxstb: %x is (should be) for showing images statically (I mostly use that for "background" pictures in 1-bit display WPSs), %xl/%xd when using images in conditionals.
23:08:13amiconngevaerts: Hmm. Maybe there's a register that selects the clock source for the usb controller
23:08:45Bagderamiconn: that's not unlikely, I've worked with an unrelated chip that had such an option
23:09:27pixelmaZagor: will try to do it tomorrow (will have more time), now I'm quite tired and would probably write a lot of nonsense or do typos... feel free to remind me
23:09:32linuxstb_pixelma: Yes, but can't you do the same thing with %xl/%xd ?
23:09:38 Join Horschti [0] (
23:10:24 Quit Horscht (Nick collision from services.)
23:10:33pixelmanot sure what's behind (how the code is handled) but this way I only need one line and not two
23:10:40EternalRainsis %%x in any way optimized for static display?
23:10:48pixelmalinuxstb: as a WPS writer
23:11:10gevaertsamiconn: So it comes down again to someone studying OF code I guess
23:12:06linuxstb_pixelma: Yes, I can see that. But my problem is that the %x tag isn't stored as tokens in the loaded WPS, so I don't know where it was in the WPS (i.e. which viewport it belonged to).
23:12:15Bagderamiconn: iaudio bootloaders updated on the download server
23:12:26amiconnThanks :)
23:12:50linuxstb_But I guess I can just store the viewport separately for each %x image...
23:13:05Bagderamiconn: the "history" table on the IaudioBoot misses the v4
23:13:28Bagderoops ;-)
23:13:40*Bagder shuts up
23:13:56linuxstb_Bagder: Have you seen the new table on the TargetStatus page?
23:14:25Bagderno... but I see it now
23:15:44amiconnBah, I can't attach the test build to a forum post :(
23:15:50linuxstb_The "in development" text for the link no longer seems appropriate on the front page - I was thinking maybe it could be a "current status" link. And maybe also have a link from the current builds page?
23:15:59devurandom[22:57] <linuxstb_> Where are you compiling it? (Windows, Mac OS X, Linux, ...) << Linux, GCC 4.3
23:16:01amiconnForum only allows 128KB, and rockbox.ipod is ~205KB zipped
23:16:21devurandomamiconn: Tried 7zip already?
23:17:24 Quit RexDart ("CGI:IRC (EOF)")
23:18:23 Join MethoS-- [0] (
23:19:26devurandomlinuxstb_: Reason is probably that GCC 4.3 cut down the headers to a minimum. Every non-necessary header is now not anymore included by default.
23:20:06 Quit DefineByte ("Bye all")
23:20:42linuxstb_devurandom: getenv should be in stdlib.h - I wouldn't have thought that was omitted...
23:21:19 Quit Absinthe ("Leaving")
23:21:35amiconnSo, where to post that test build?
23:21:36devurandomlinuxstb_: Are you the maintainer?
23:21:52devurandomI guess you have to #include <cstdlib>...
23:22:00linuxstb_devurandom: No, I'm not.
23:22:20 Quit ViperShark (Remote closed the connection)
23:22:21linuxstb_domonoky is one of the maintainers...
23:22:43linuxstb_amiconn: I can host it if you don't want to put it on your webspace.
23:23:12*amiconn almost forgot about his webspace (not the dyndns one)
23:23:16Bagderdevurandom: so you're saying they took getenv() out of that header?
23:23:28linuxstb_amiconn: Or the flyspray task - I think bigger attachments are allowed there.
23:23:50Bagdersounds like a g++ bug to me
23:23:52amiconnDoes flyspray keep a history?
23:24:10devurandomBagder: rbutil/rbutilqt/utils.cpp doesn't include cstdlib...
23:24:14amiconnI thinks so, and that test build isn't worth to be kept. It's just for harvesting a rom dump
23:24:42linuxstb_amiconn: I _think_ you can remove attachments, but external webspace is probably easiest.
23:24:55domonokydevurandom: whats not working ?
23:25:06devurandom[22:57] <linuxstb_> Where are you compiling it? (Windows, Mac OS X, Linux, ...)
23:25:14devurandomOups, sry... copy&paste
23:25:22Bagderthen why did it ever work?
23:25:25devurandom"Is this rbutil bug know? "utils.cpp:59: error: ‘getenv’ was not declared in this scope""
23:25:26linuxstb_domonoky: rbutil gives an error about getenv when compiling on Linux with gcc 4.3
23:25:37 Quit Febs (Read error: 104 (Connection reset by peer))
23:25:56devurandomBagder: Because before several c++ headers included unnecessary headers.
23:26:18domonokyand is this easily fixable by including the right header ? :-)
23:26:27devurandomI think so, yes.
23:26:28Bagderlucky for me I avoid C++ ;-)
23:27:13domonokydevurandom: if you find out what to change, could you put please a patch into flyspray ? (i will then take a look, and commit it)
23:27:19 Quit moos ("good night")
23:27:29devurandomdomonoky: kk.
23:27:39*domonoky compiles rbutil on windows.. :-)
23:28:35 Join alienbiker99 [0] (
23:28:51 Quit domonoky (Read error: 104 (Connection reset by peer))
23:30:00 Join JdGordon|uni [0] (i=82c20d6a@rockbox/developer/JdGordon)
23:30:42 Join MethoS- [0] (
23:30:45devurandomOugh... Cant I file a bug there without registering?
23:32:20 Quit tvelocity (Remote closed the connection)
23:32:21 Quit pikhq ("Rebooting stuff.")
23:38:00 Quit ompaul (Read error: 113 (No route to host))
23:38:35 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
23:39:53 Join corevette [0] (
23:43:05 Join Klevi [0] (
23:43:16 Quit quitte ("Lost terminal")
23:43:44KleviQuestion: Rockboy is left out of Rockbox now by default (for the 4th gen Ipod)
23:44:28linuxstb_I mean Yes....
23:44:36linuxstb_Seems it's never been included.
23:45:28KleviI saw an IRC log of you saying it worked nicely this morning. (google turned it up)
23:45:45Kleviand, I looked for it this morning, i mean =)
23:46:32KleviWhy was it left out? Rockboy is awesome.
23:47:44 Join bluebrother [0] (n=dom@rockbox/staff/bluebrother)
23:48:25linuxstb_It's not left out as such, it's just never been ported to the 4G greyscale's LCD (and 1G-3G)
23:48:45KleviWhich isnt easy to accomplish?
23:49:44linuxstb_It's not that hard, but isn't trivial. But the main problem is that no-one capable of doing it is interested in Rockboy.
23:50:00 Quit EternalRains ("ChatZilla 0.9.81 [Firefox]")
23:50:02amiconnOr doesn't have the device
23:50:17 Join EspeonEefi [0] (i=espeonee@STRATTON-SIX-FOURTY-NINE.MIT.EDU)
23:50:24amiconnI could do it, but rockboy certainly doesn't have priority for me
23:50:44 Quit MethoS-- (Read error: 113 (No route to host))
23:50:49 Join MethoS-- [0] (
23:50:51amiconnActually we might want two variants of the gfx scaling code, one for the 160x128 iPods, and one for the minis
23:51:19amiconnThe 160x128 ipods could offer the same flexibility as the H1x0 version of rockboy
23:52:56KleviWrong tab/box?
23:53:13Lloreanamiconn: I assume you're referring to the forum post. In which case I'll again assume and say he was probably responding to a quite earlier post.
23:53:49KleviOh ^.^ Paitence is a lesson learned slowly, clearly.
23:54:09*gevaerts waits for someone to remove soap's clearly off-topic post ;)
23:55:58Klevi... that cant be a good smell im smelling..
23:56:03 Nick Klevi is now known as Klevi_BRB (
23:57:42*scorche|sh looks around for said post

Previous day | Next day