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

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

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

#rockbox log for 2009-08-14

00:01:02 Quit petur (Remote closed the connection)
00:03:58 Quit domonoky (Read error: 104 (Connection reset by peer))
00:09:13 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
00:09:38 Quit linuxstb (Nick collision from services.)
00:09:41 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
00:09:55CIA-6New commit by zagor (r22304): Don't claim negative upload time estimates.
00:10:18LambdaCalculus37ej0rge: Have you gotten to try out FS #10431 yet?
00:11:39rasherZagor: oh dear
00:11:47rasherThat's a bit of a tragic diff
00:12:19Zagorrasher: well, lastultime is used in the scheduling. it's just not interesting in the log.
00:13:04ej0rgeLambdaCalculus37: No. Probably saturday.
00:13:43LambdaCalculus37ej0rge: Okay, let me know how it works for you. Some of the plugins may crap out when building because of missing graphics.
00:14:49 Join Strife89 [0] (
00:15:28 Join webguest63 [0] (n=c73d19fe@gateway/web/cgi-irc/
00:15:38 Quit webguest63 (Client Quit)
00:17:19 Quit truthtaco_ (Read error: 110 (Connection timed out))
00:17:32 Join truthtaco_ [0] (
00:22:32 Quit bmbl ("Bye!")
00:28:35 Quit Zagor ("Leaving")
00:30:16 Join aaron424 [0] (
00:30:20 Quit CaptainKwel ("Page closed")
00:43:16 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
00:44:12 Join Strife1989 [0] (
00:44:48 Quit ehntoo (Read error: 104 (Connection reset by peer))
00:46:32 Quit Strife89 (Nick collision from services.)
00:46:39 Nick Strife1989 is now known as Strife89 (
00:49:07*amiconn just checked fat.c more closely regarding the time increment on non-rtc targets
00:49:27 Join ehntoo [0] (
00:50:17amiconnThis increment is no longer 01h 15min like it was at some point, but just 10 minutes, which means the minute marker idea needs adjustment
00:52:22amiconnEither we switch to 01h increments, which means files which are frequently written to will increment much faster, or we use :11:11, :22:22 etc (12 minutes increase on average, i.e. not using :00:00)
00:54:45 Quit n1s ("Lämnar")
01:00:19JdGordon|why such large arbitrary movements? why not just 1 minute?
01:00:28JdGordon|the point is to just make sure the time changes... isnt it?
01:01:44 Quit GeekShadow (Read error: 104 (Connection reset by peer))
01:06:06amiconnIf we'd use one minute, we couldn't set a kinda obvious marker anymore
01:09:41JdGordon|yeah, I still dont see the point of a marker thats not really obvious... the date is going to be wrong no matter what it is so you're just misleading the user...
01:10:02*amiconn already explained most aspects
01:11:12amiconnWe need to increment to indicate change. We must never adjust the timestamp backwards. Both points together mean that we cannot use the date as a marker for pre-existing files when changing them in rockbox
01:12:00amiconnThat leaves the possibility to use the time (minutes + seconds) as a marker that the timestamp might not be correct
01:13:08amiconnThis is independent of the choice for the initial date of new files. In this case I just *prefer* it to be roughly correct (i.e. using the build date)
01:14:18***Saving seen data "./dancer.seen"
01:17:25 Quit LambdaCalculus37 ("Fwump")
01:17:56 Quit thegeek (Read error: 104 (Connection reset by peer))
01:18:01 Join thegeek_ [0] (
01:19:06 Quit thegeek_ (Read error: 104 (Connection reset by peer))
01:19:09 Join thegeek [0] (
01:21:06 Join AsaelReiter [0] (n=5f56414f@gateway/web/cgi-irc/
01:22:00 Quit AsaelReiter (Client Quit)
01:22:14 Quit JdGordon| ("Miranda IM! Smaller, Faster, Easier.")
01:22:18 Join AsaelReiter [0] (n=5f56414f@gateway/web/cgi-irc/
01:25:06 Quit AsaelReiter (Client Quit)
01:25:20 Join CaptainKwel [0] (
01:29:12 Quit Thundercloud (Remote closed the connection)
01:29:35 Join Strife1989 [0] (
01:29:40 Quit ender` (" I do not fear death. I had been dead for billions and billions of years before I was born and had not suffered the slightes")
01:30:07 Quit Strife89 (Nick collision from services.)
01:30:38 Nick Strife1989 is now known as Strife89 (
01:31:20 Part toffe82
01:36:20 Join JdGordon_ [0] (i=63cb1c31@gateway/web/freenode/x-njfdblrkmvscopod)
01:42:59 Quit thegeek (Read error: 104 (Connection reset by peer))
01:44:06 Join robin0800 [0] (
01:55:11 Quit robin0800 ("ChatZilla 0.9.85 [Firefox 3.5.2/20090810133528]")
01:55:34 Join robin0800 [0] (
02:01:33 Quit Rondom (Nick collision from services.)
02:01:43 Join Rondom [0] (
02:03:52 Quit JdGordon_ (Ping timeout: 180 seconds)
02:05:24 Quit robin0800 ("ChatZilla 0.9.85 [Firefox 3.5.2/20090810133528]")
02:05:55 Join robin0800 [0] (
02:11:20 Quit mt (Read error: 104 (Connection reset by peer))
02:12:31 Quit aaron424 (Remote closed the connection)
02:13:15 Quit liar_ (Read error: 60 (Operation timed out))
02:13:25 Join aaron424 [0] (
02:15:33 Quit jgarvey ("Leaving")
02:16:15 Quit efyx (Remote closed the connection)
02:30:37 Quit robin0800 ("ChatZilla 0.9.85 [Firefox 3.5.2/20090810133528]")
02:31:06 Join robin0800 [0] (
02:31:41 Join liar_ [0] (
02:32:18 Quit robin0800 (Client Quit)
02:35:04 Quit aaron424 ("ChatZilla 0.9.85 [Firefox 3.5.2/20090810133513]")
02:42:57 Join safetydan [0] (n=deverton@rockbox/developer/safetydan)
03:11:01 Join funman [0] (n=fun@rockbox/developer/funman)
03:11:50funmanthe AMS guy didn't call today but we exchanged via SMS and he should call me tomorrow at 18h (UTC+2)
03:14:22***Saving seen data "./dancer.seen"
03:17:35 Quit martian67 ("out")
03:17:56 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
03:42:08 Join TheCoolGman [0] (n=446e5e7d@gateway/web/cgi-irc/
03:42:55TheCoolGmanHi I'm have an issue when compiling with cygwin and svn. I've done it on a clean build too:
03:44:09funmanTheCoolGman: svn cleanup; svn diff|patch -p0; make clean; make
03:45:03 Join polobricolo [0] (n=polobric@
03:46:04funmanwhich model are you building for?
03:47:25TheCoolGmaniPod 5g
03:49:09funmanthat is video?
03:49:59TheCoolGman Yes
03:52:24 Quit polobricolo (Read error: 54 (Connection reset by peer))
03:53:42funmanipod video builds fine here, try to clean your ccache cache ?
03:54:57TheCoolGmanYour are using Tortise SVN right? Or dose that not make a diffrence? Also I'm not sure how to do that.
03:56:16 Quit funman ("free(random());")
03:58:44TheCoolGmanNow I'm alone
04:05:34 Quit TheSeven (Nick collision from services.)
04:05:48 Join The_Seven [0] (
04:05:53 Nick The_Seven is now known as TheSeven (
04:44:02 Nick truthtaco_ is now known as TruthTaco (
04:56:54 Quit Strife89 (Read error: 104 (Connection reset by peer))
05:02:10 Join tarbo_ [0] (n=me@unaffiliated/tarbo)
05:11:37 Quit tarbo (Connection timed out)
05:14:26***Saving seen data "./dancer.seen"
05:34:23 Quit Horschti ("Verlassend")
05:42:04CIA-6New commit by mt (r22305): Fix a (stupid) mistake in libatrac/atrac3.c
06:18:39JdGordonUnhelpful: mercy ping?
06:42:40 Join qurvel [0] (
06:49:32 Quit jernejovc (Connection reset by peer)
06:49:41 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
06:51:07 Join jernejovc [0] (
06:55:31 Part qurvel
07:00:41 Join qurvel [0] (
07:00:42 Part qurvel
07:06:17 Quit ehntoo (Read error: 110 (Connection timed out))
07:06:44 Quit TheCoolGman ("CGI:IRC (EOF)")
07:09:20 Join qurvel [0] (
07:14:27***Saving seen data "./dancer.seen"
07:21:25 Quit BlakeJohnson86 (Read error: 110 (Connection timed out))
07:21:56 Join BlakeJohnson86 [0] (
07:24:59 Part qurvel
07:41:02 Join ender` [0] (
07:54:08 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
07:59:51 Quit moos (Read error: 145 (Connection timed out))
08:04:27 Quit safetydan ("Leaving.")
08:07:19 Quit CaptainKwel (Remote closed the connection)
08:15:00 Join Rob2223 [0] (
08:26:59 Join Lss [0] (
08:33:01 Quit Rob2222 (Read error: 110 (Connection timed out))
08:39:53 Join petur [50] (n=petur@rockbox/developer/petur)
08:52:50 Quit Lss (Read error: 104 (Connection reset by peer))
08:53:07 Quit Ridayah (Read error: 110 (Connection timed out))
09:13:28 Join flydutch [0] (
09:14:29***Saving seen data "./dancer.seen"
09:16:46 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
09:27:34 Join pamaury [0] (
09:30:02 Join einhirn [0] (
09:32:09 Join n1s [0] (n=n1s@rockbox/developer/n1s)
09:33:05 Join Thundercloud [0] (
09:57:37 Join CQ [0] (
09:58:52CQhello, I have an ipod 4g clickwheel where the battery doesn't charge to more than 55% or so, and discharges veery quickly... the battery is set to the correct size... does this mean battery dead, or can there be another cause? Same happens with orig OS and RB3.3
10:04:15 Quit kachna (Read error: 113 (No route to host))
10:04:26 Quit Thundercloud (Remote closed the connection)
10:04:41linuxstbCQ: That does sound like a dying battery to me...
10:05:54CQI was afraid of that...
10:06:14CQplenty of ipod disassembly instructions out there that I'm not too much looking forward to using
10:12:32CIA-6New commit by mt (r22306): Replace av_log() with DEBUGF and drop some now-useless files.
10:12:50 Join mt [0] (n=MTee@rockbox/developer/mt)
10:24:42CIA-6New commit by mt (r22307): Move main() outside atrac3.c and create atrac3.h
10:31:38 Join daurnimator [0] (i=daurnima@freenode/staff/daurnimator)
10:34:58n1sJdGordon: looked at fs#9886 I don't understand why we want to buffer_alloc a statically sized buffer or will the buffer get a dynamic size later?
10:35:00linuxstbmt: Nice progress with atrac. So the fixed-point conversion is done and working fine?
10:35:13 Join efyx [0] (
10:35:28mtlinuxstb: Yep.
10:35:57*linuxstb needs to search for his single atrac-in-rm file...
10:36:49n1sJdGordon: also s/wps_buffer_alloc/skin_buffer_alloc/ in two comments ;)
10:37:06mtlinuxstb: Or - :)
10:37:49linuxstbmt: I'm just looking the list of codecs on this page - - does Rockbox support both raac (LC-AAC) and racp (HE-AAC) ?
10:38:58mtlinuxstb: It should but I hadn't really found racp samples for testing.
10:39:32linuxstbOK, so the code is there, but it's untested?
10:40:45mtYes, I guess it's a quesion of libfaad with he-aac though since parsing and packet extraction shouldn't be different.
10:41:13linuxstbThe aac-in-mp4 codec plays he-aac, so it should be OK.
10:41:18markunisn't HE-AAC disabled on some targets?
10:42:21*linuxstb doesn't know.
10:43:12*linuxstb wonders why that wikipedia article excludes mp3-in-rm
10:44:14linuxstbSo there seems to be 10 codecs used in realaudio, and when atrac3 is working, Rockbox will support 5. Not bad ;)
10:44:53markunreally good in fact :)
10:45:24gevaerts50% is barely a passing grade in schools :)
10:45:26linuxstbAnd the first two should be possible as well - "lpcj" and "28.8" have ffmpeg decoders IIRC.
10:45:56mtsipro too has an ffmpeg decoder.
10:46:11markunmt: then what are you waiting for :P
10:46:19mtmp3 : only one sample is known to exist, that's why it's excluded.
10:46:24linuxstbGreat! You have 3 days left until the firm pencils-down date...
10:47:01mtmarkun: :D .
10:47:01gevaertsDo we support ape-in-rm?
10:47:07linuxstbSo really it's just realaudio lossless that's missing. And I would doubt many people use that...
10:47:19markungevaerts: we could beat RM to their own game! :)
10:47:27gevaertsmarkun: exactly!
10:47:34gevaerts"no samples are known to exist"
10:47:44markunwe can make some samples
10:48:20mtlinuxstb: I guess also ralf isn't much used yes. Only one sample on mplayer.
10:49:04mtgevaerts: ape-in-rm ? :)
10:49:10gevaertsdoesn't matter. If there's a codec around that's called Ralf, we obviously have to support it
10:49:14*linuxstb thought he recalled a radio station streaming things in mp3-in-rm, but it was actually mp3-in-asf
10:49:40linuxstbA friend of Sid's?
10:49:41gevaertsmt: why not? :)
10:49:42mtgevaerts: You do ralf then. :P
10:50:49*gevaerts shuts up before he starts promising things :)
10:50:59mtBecause ape can't be in rm. :)
10:51:17gevaertswhy not?
10:51:32linuxstbWell, it would have to be called "rape"..
10:51:46mtYikes ! That's one good reason ..
10:53:20 Quit CQ (Read error: 110 (Connection timed out))
10:55:37 Join BdN3504 [0] (n=4e3434f9@gateway/web/cgi-irc/
11:14:32***Saving seen data "./dancer.seen"
11:15:40mtShould I add rockbox's gpl headers to atrac's fixed point files ?
11:16:08linuxstbWho wrote them?
11:19:39mtMe and saratoga.
11:22:24linuxstbmt: Then you'll need to agree a license with saratoga, but it might be nice to use the LGPL, to make it easier for someone to port the code back to ffmpeg at some point.
11:24:47 Quit BHSPitMonkey (Remote closed the connection)
11:26:57 Quit BdN3504 ("CGI:IRC (EOF)")
11:44:50 Quit TheSeven ("ChatZilla 0.9.85 [Firefox 3.5.2/20090729225027]")
11:48:10 Join TheCoolGmam [0] (n=446e5e7d@gateway/web/cgi-irc/
11:49:05TheCoolGmamWhat do you guys think about FS #5153 ? It seem's it could use some tiny improvements and that's it.
11:50:12 Join DarkDefender [0] (
11:50:46linuxstbTheCoolGmam: Hi. Are you the person on the forums having line-ending problems with cygwin/tortoisesvn?
11:51:02 Join AlexP_ [0] (n=alex@rockbox/staff/AlexP)
11:51:03 Quit AlexP (Read error: 104 (Connection reset by peer))
11:51:12TheCoolGmamYeah check it again to see what I posted
11:51:37linuxstbYes, I've seen that.
11:51:51TheCoolGmamwhat should I do?
11:51:57 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
11:52:00TheCoolGmamDelete the folder?
11:52:05linuxstbAccording to that wiki page, you shouldn't need to reinstall cygwin, just re-run setup.exe Have you rebooted since doing that?
11:52:48linuxstbAnd as gevaerts said, deleting and then recreating the build directory is also a good idea.
11:53:03TheCoolGmam*TheCoolGman realizes he made an Epic Fail
11:53:19TheCoolGmamI should restart heh
11:54:02 Join antil33t [0] (n=Mudkips@
11:54:05TheCoolGmamI'll restart and tell you what I get
11:54:20linuxstbIf that does fix it, it would be useful to post to that forum thread saying so - so people searching for the same problem find a solution.
11:54:25 Quit TheCoolGmam ("CGI:IRC (EOF)")
11:55:30 Join SUSaiyan` [0] (
11:55:37 Join Rob2222 [0] (
11:55:41 Quit SUSaiyan (Read error: 104 (Connection reset by peer))
11:56:03linuxstbmt: In the README.rockbox for libatrac, you refer to svn r18079 - what date is that from? It would be useful to add that info, so people have an idea of how that it is.
11:56:03 Quit Rob2223 (Read error: 104 (Connection reset by peer))
12:05:57 Join CIA-61 [0] (n=CIA@
12:07:19 Quit CIA-6 (Read error: 145 (Connection timed out))
12:22:42 Join _Kopfgeldjaeger [0] (
12:23:14 Join courtc_ [0] (n=court@unaffiliated/courtc)
12:28:43 Join Lynx_ [0] (
12:28:48 Quit Kopfgeldjaeger (Read error: 104 (Connection reset by peer))
12:28:57 Nick _Kopfgeldjaeger is now known as Kopfgeldjaeger (
12:34:24 Quit courtc (Read error: 111 (Connection refused))
12:36:23 Join _Shaid [0] (
12:36:23 Quit Shaid (Read error: 104 (Connection reset by peer))
12:36:30 Nick _Shaid is now known as Shaid (
12:37:59 Join TheCoolGman [0] (n=446e5e7d@gateway/web/cgi-irc/
12:38:27 Quit TheCoolGman (Client Quit)
12:38:55 Join TheCoolGman [0] (n=446e5e7d@gateway/web/cgi-irc/
12:39:31TheCoolGmanDamn! I restarted and typed "echo > a.txt" and it still came out 1 kb
12:43:04TorneEr, I just read that thread, and it's.. wrong
12:43:28TorneTelling cygwin to translate line endings is a really bad idea
12:44:27TheCoolGmanI just wish you guys would still give out the current build
12:45:05Tornewe do
12:45:25TheCoolGmanI'm talking about the source
12:45:55 Join Ridayah [0] (
12:46:03TheCoolGmanit was always update into a 7z until you guys stopped updating it
12:46:12 Join moos [0] (i=mustapha@rockbox/staff/moos)
12:46:31Torneif yo ucan manage to compile it under cygwin, why is it so hard to use svn?
12:46:56TheCoolGmanIt's not Im issues try to learn to use SVN iwth Cygwin
12:47:06TorneIf you stop using tortoise the problem will go away
12:47:22TheCoolGmanAnd the Daily build source IS STILL r21932
12:47:27TorneUsing cygwin's CR/LF line ending mode is not recommended by the cygwin folks because it breaks loads of other stuff
12:47:41TheCoolGmanyeah I notice that when patching
12:47:42Tornerun cygwin setup and pick to install subversion
12:47:48Tornethen just type svn co svn:// rockbox
12:47:51Tornein a cygwin window.
12:47:54Torneyou will get the latest source.
12:48:18TheCoolGmanShould I use that instead of Trotise SVN
12:50:17Tornethat should do it with the right line endings, yes
12:50:30Tornei can't test it atm though, this machine can't connect to svn://
12:50:52TheCoolGmanSo I have the setup open and all I have are these opitons Unix/Binary or DOS/text
12:56:21krazykitTheCoolGman, have patience in waiting for an answer. only so many people actually use cygwin
12:57:12Torneok, look. you already tried to rerun setup and change the line ending mode, right?
12:57:16Torneopen a cygwin window and type mount
12:57:19TheCoolGmanReally? I didn't know
12:57:44Tornedo all the mount points say binmode at the end
12:58:00Tornealso, I am at work so pardon me if I don't answer you instantaneously :)
12:58:25TheCoolGmanHold so chose Unix/Binary? in setup
12:58:51TorneDon't choose anything in setup
12:58:54TorneExit setup
12:59:06TorneRun a cygwin window and type mount, and see if the lines all say binmode or not
12:59:45 Join intrados_ [0] (
12:59:53Torneall of them?
13:00:06TorneThen you've already changed it to DOS/text mode.
13:00:25 Quit intrados_ (SendQ exceeded)
13:01:17TheCoolGmanso then why was I getting that error with alarm_menu.c?
13:02:16 Join intrados_ [0] (
13:03:41TorneI have no idea.
13:04:05 Quit intrados (Read error: 60 (Operation timed out))
13:04:20TheCoolGmanI will try a clean build again
13:05:28 Quit Rob2222 ()
13:05:40linuxstbTorne: If using CR/LF mode isn't recommended, then I guess we should remove that wiki page saying how to use tortoisesvn, or replace it with a page saying not to use it...
13:05:50Tornelinuxstb: Cygwin doesn't handle this consistently
13:05:55linuxstbBut presumably petur is using it fine, otherwise he wouldn't have added that wiki page...
13:06:01TorneCR/LF mode only works for apps which actually open files in text mode
13:06:06Tornemost unix programs don't bother :)
13:06:07linuxstbTorne: I'm not surprised... I can't imagine how it can.
13:06:31TorneI don't know of a way to make tortoise use a specific line ending
13:06:33linuxstbI never knew there was such a thing in Unix...
13:06:42Tornelinuxstb: there is, it's in the ANSI C API :)
13:06:53Torneit's just a no-op on UNIX
13:07:04Tornewell, normally.
13:07:20linuxstbDo you mean the O_BINARY flag? Or something else?
13:07:46linuxstbBut anyway, we're straying off-topic...
13:07:49Tornewell, that and "rb" as fopen() :)
13:07:51Tornebut yes
13:08:06TorneI have heard nothing but tales of woe about cygwin installs that use crlf
13:08:31Torneif i recall correctly, if you set cygwin to unix mode and just run cygwin's command line svn, it will check out with unix line endings
13:08:42Torneand everything will be happy and great until you come to use a win32 editor to edit files :)
13:08:58linuxstbYes, which is exactly what TheCoolGman would have got with the source archive.
13:09:01Tornebut i can't test this right now because i can't access svn:// urls from behind this wirewall
13:09:14Torneand i can't find an http svn repo anywhere to test which has eol-style:native ;)
13:09:25TheCoolGmanI'm suprised you are able to use IRC
13:09:45TorneI can HTTP CONNECT to anywhere I like, but svn doesn't know how to do that.
13:09:55Torneand i cba to install a tunneling program
13:10:35TheCoolGmanI'm almost done downloading the latest build
13:10:36Tornelinuxstb: anyway your advice doesn't seem to work; bash redirection doesn't open files in text mode
13:10:48Torneso echo > foo.txt doesn't tell you what line ending mode you have. :)
13:11:09linuxstbTorne: It was just a guess... If that doesn't work, then indeed cygwin's CR+LF handling is broken.
13:11:41Tornelinuxstb: they're doing their best to comply with the relevant standards.. :) They can't do CR/LR translation for all file IO or it would be impossible to work with actual binary files.
13:14:33***Saving seen data "./dancer.seen"
13:18:57peturI'm no longer using tortoisesvn since I moved to linux
13:19:40TheCoolGmanI'm using windows But I could just SVN from Cygwin now
13:20:07gevaertsTheCoolGman: are you making a new build directory every time?
13:20:29TorneTheCoolGman: If you want to try that, then set the line ending mode back to unix
13:20:41Torneand start completely from scratch using the cygwin version of subversion
13:20:56TheCoolGmanI sorry to ask this because I'm a noob at this but what do you mean exactly make new folder if so then yes
13:22:17 Join GeekShado_ [0] (
13:23:00 Join Rob2222 [0] (
13:27:45TheCoolGmanIt's compiling
13:28:23TheCoolGmanAnd I still get the same error
13:29:17TheCoolGmanIs it just me or the build because at this point I have no idea
13:31:44TheCoolGmanAlright may be when I get backon you guys can tell me whats goin on I have to sleep
13:31:52TheCoolGmanthanks for the help
13:32:26 Quit TheCoolGman ("CGI:IRC (EOF)")
13:33:07 Join ehntoo [0] (
13:38:29 Quit GeekShadow (Read error: 110 (Connection timed out))
13:45:24 Quit gevaerts (Nick collision from services.)
13:45:33 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
13:48:30 Join AndyI [0] (i=AndyI@
13:51:38 Join robin0800 [0] (
13:57:43 Quit Rondom (
13:58:40NJoinRondom [0] (
14:01:06 Quit AndyIL (Read error: 110 (Connection timed out))
14:05:32 Quit ehntoo (Read error: 110 (Connection timed out))
14:13:29 Quit pamaury ("Quitte")
14:14:07 Join pamaury [0] (
14:29:32*n1s wonders why the same rev of the source 7zip is recreated every day?
14:32:38 Quit AndyI ()
14:33:26 Join AndyI [0] (i=AndyI@
14:36:45 Quit petur (Read error: 104 (Connection reset by peer))
14:36:58 Join petur [50] (n=petur@rockbox/developer/petur)
14:37:07martian67n1s, bit rot!
14:37:45 Join LambdaCalculus37 [0] (i=44a0430d@rockbox/staff/LambdaCalculus37)
15:01:56 Quit robin0800 ("ChatZilla 0.9.85 [Firefox 3.5.2/20090810133528]")
15:02:29 Join robin0800 [0] (
15:14:34***Saving seen data "./dancer.seen"
15:16:00 Quit robin0800 ("ChatZilla 0.9.85 [Firefox 3.5.2/20090810133528]")
15:16:20 Join robin0800 [0] (
15:17:30UnhelpfulJdGordon: hm?
15:19:42 Quit robin0800 (Client Quit)
15:19:58 Join robin0800 [0] (
15:20:20 Join fdinel [0] (
15:23:40 Quit robin0800 (Client Quit)
15:24:05 Join robin0800 [0] (
15:29:44 Quit robin0800 ("ChatZilla 0.9.85 [Firefox 3.5.2/20090810133528]")
15:30:02 Join robin0800 [0] (
15:35:21 Join Lynx [0] (
15:35:48 Nick Lynx is now known as Guest5888 (
15:36:51 Join Omlet [0] (
15:39:27 Join evilnick [0] (i=0c140464@gateway/web/freenode/x-uhpbvzcjcmxwedpp)
15:39:48 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
15:40:34 Quit robin0800 ("ChatZilla 0.9.85 [Firefox 3.5.2/20090810133528]")
15:40:58 Join robin0800 [0] (
15:41:14 Quit martian67 ("out")
15:41:47 Quit robin0800 (Client Quit)
15:42:10 Join robin0800 [0] (
15:43:03 Join avacore [0] (
15:43:03 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
15:45:15 Quit martian67 (SendQ exceeded)
15:46:17 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
15:47:31 Quit Omlet ()
15:51:10 Quit Lynx_ (Success)
15:51:10 Nick Guest5888 is now known as Lynx_ (
15:54:20 Quit avacore^ (Read error: 110 (Connection timed out))
15:56:33 Join Omlet [0] (
15:59:51 Part LinusN
16:08:28 Quit Sajber^ (Read error: 104 (Connection reset by peer))
16:13:05 Nick Omlet is now known as Omlet[AFK] (
16:14:05 Nick Omlet[AFK] is now known as Omlet (
16:15:32 Nick Omlet is now known as Omlet_Away (
16:15:41 Quit Omlet_Away ()
16:15:50 Join Omlet [0] (
16:16:08 Quit Omlet (Client Quit)
16:19:07 Join Omlet [0] (
16:27:59 Join jgarvey [0] (
16:30:54pamaurygevaerts: are you there ?
16:31:16 Join captainkwel [0] (i=2669ecc2@gateway/web/freenode/x-nqntwynepvydjspn)
16:31:52gevaertsin a way
16:32:28pamauryI have some horrible news about the usb_serial bug
16:35:04pamauryMy tests suggest that (at least on my sansa e200) there is a direct link between the 96 limit of the usb_drv_arc and the needs_exclusive_storage flag
16:35:45pamauryI have rewritten a usb driver and it fail to send a >=97 packet BUT if I tweak usb.c and usb_core.c to grant is exclusive storage THEN it succeeds !
16:36:07gevaertsah, that explains things
16:36:12pamaurywhich mean that with exclusive storage I have a working a usb_serial driver
16:36:24pamauryhow is this link possible ?
16:37:12gevaertsbecause needs_exclusive_storage also triggers cpu boost
16:37:32pamaurywhat is cpu boost ? it scales cpu frequency ?
16:38:05gevaertsyes. Normally we run at 30MHz, but large packets don't work then. Boosting makes it run at 80MHz, which we do for storage for precisely this reason
16:38:13*gevaerts should have resalised this before
16:38:58 Join funman [0] (n=fun@rockbox/developer/funman)
16:39:46pamauryIt took me some time to realize it, only when I completely destroyed the usb_storage driver I realized the difference
16:40:37gevaertswell, we now know what's going on. I suspect that just limiting the amount transmitted is the right solution as long as we only do outgoing traffic
16:45:43pamauryOk, I'll come back later, hope this information will help. Perhaps a commit would be useful to change the comment is usb_serial. Also determine on which target it's needed.
16:46:38gevaertsIf you feel bored, you could play with the stream disable bit in REG_USBMODE to see if it makes a difference
16:47:00gevaertsI probably won't do much on this today. I'm travelling tomorrow, and I need to pack
16:47:44gevaertspamaury: if you can find MCF5251RM.pdf, it seems to describe the same USB controller in a bit more detail for some registers
16:51:30 Join toffe82 [0] (n=chatzill@
16:54:20 Join aaron424 [0] (
16:55:17 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
16:57:11 Join yawny [0] (i=user36@
16:59:34 Quit yawnage (Read error: 113 (No route to host))
17:01:00 Join vodi [0] (
17:06:32JdGordonn1s: the reason its bufalloced is because I still like the idea of (when this is finished) allowing the user to decide (or the theme based on previous boots) how much is needed.. really there is no need to waste xKB on a 2mb target when only 1/3 of it is being used...
17:06:59 Join BryanJacobs [0] (n=bryanjac@
17:07:10JdGordonalso there really is no difference between bufallocing a buffer, or just making a static one..
17:07:36gevaertsbufalloc hides it from the delta table :)
17:08:09LambdaCalculus37My Clip is behaving a little strange. I just updated my build on it, and now it shuts off immediately after I turn it on and it displays the Rockbox logo for about half a second before shutting off.
17:08:16n1sJdGordon: ok, makes sense, to me buffer_alloc implies we might want different sizes so i think for a static size a static buffer is clearer
17:08:18LambdaCalculus37Any attempts to turn it on aren't working.
17:08:50JdGordongevaerts: and that :)
17:09:06JdGordonwell.. I still like the idea of making that buffer customisable (in the future)...
17:09:43JdGordonand I'm pretty certain more will agree once radio+statusbar gets skiniified
17:09:50 Join kugel_ [0] (
17:10:01n1sideally it should auto-resize to the min size needed :)
17:10:04 Quit petur ("beer time")
17:10:06 Nick kugel_ is now known as kugel (
17:10:44kugelthat's what I'm thinking too
17:11:19JdGordonyes and no....
17:11:35LambdaCalculus37I wonder if something went kaput with it...
17:11:54JdGordonLambdaCalculus37: dead battt?
17:12:22 Quit aaron424 ("ChatZilla 0.9.85 [Firefox 3.5.2/20090810133513]")
17:13:19CIA-61New commit by mt (r22308): Factor out bytestream reading functions needed in rm parsers/codecs.
17:13:23LambdaCalculus37JdGordon: I thought so as well, but it still does this even after a charge.
17:14:07 Join Blue_Dude [0] (
17:14:35***Saving seen data "./dancer.seen"
17:16:30kugelLambdaCalculus37: update the bootloader?
17:16:46UnhelpfulJdGordon: was i being pinged for a reason? :)
17:17:27JdGordonyeah, I thought my bug had something to do with not loadnig mbps right (or something like that) and ws hoping you could help... turns out im just an idiot
17:18:41Unhelpfulexcellent, carry on then. :)
17:19:14 Join FOAD_ [0] (
17:19:20kugelUnhelpful: I was seeing some functions yesterday which could be also in the common lcd file (set/getfont, get_width/_height) I haven't looked at each lcd-* file to check it though
17:19:24LambdaCalculus37kugel: The OF is working fine, so what I think I'm going to do is just uninstall the bootloader with a clean OF update, then reinstall it again.
17:19:40kugelLambdaCalculus37: no need for a clean of update
17:20:33 Join daurn| [0] (
17:22:11 Quit daurnimator (Read error: 60 (Operation timed out))
17:22:53 Join ehntoo [0] (
17:23:12Unhelpfulkugel: thanks for the heads-up, i was pretty sure there'd be more to put in it. i want to pull in the charcell scroll_fn as well... and some of the drawing functions that mostly just wrap lower-level functions and are not themselves aware of the native bitmap format
17:25:50domonokyBagder: can you take a look at this patch for genlang.cgi ? I tested it locally (modified so it works for me) but i am unsure if its fine for the server..
17:26:07CIA-61New commit by mt (r22309): Use the bytestream reading functions form librm and drop bytestream.h and libavutil/
17:29:12JdGordonkugel: what was the issue with the recorder not being able to show a top item?
17:29:43kugelit can't flip the buttons properly if the screen is flipped
17:29:57kugelamiconn knows more about that
17:30:16 Quit FOAD (Read error: 110 (Connection timed out))
17:30:16 Nick FOAD_ is now known as FOAD (
17:30:40funmanamiconn knows everything
17:31:10JdGordonwell... too bad.. the code you want to add to get around that issue is not worth it... if the recorder is the only target with that isssue, and only when its in upsidedown mode (1 person *maybe* on the planet) then big bloody deal
17:31:10kugeland there was an issue with at least one of the h100 remotes also
17:32:56pixelmaif I remember it can flip the buttons but one of the then needed combos wouldn't be possible in case the original way of operating the quickscreen there wouldn't work (electrical restrictins), not sure though
17:33:15kugelJdGordon: I completely agree, but amiconn and Llorean objected for that reason
17:33:46pixelmaeh... if the original way of operating it would be restored
17:34:34kugelisn't flipping buttons handled in software as for any other target?
17:35:46pixelmawell, flipping buttons is not the problem itself. Impossible combos due to hardware restrictions is
17:36:19kugelthere are no combos involved in the quickscreen IIUC
17:37:01kugeljust plain up/down/left/right (and play to exit) −− for the recorder
17:37:01pixelmathe original way of operating the quickscreen was quite different...
17:37:13pixelmaand quicker
17:38:50kugelbut this isn't anymore, apparently
17:39:31 Quit ender` (" I do not fear death. I had been dead for billions and billions of years before I was born and had not suffered the slightes")
17:39:53 Join motorbit [0] (
17:40:50motorbitim a bit lost with rockbox on a d2+ . i got it all wokring, but there are some quistions i culd not find answers for within the forum.
17:40:52pixelmakugel: yes, but someone wanted to restore that
17:40:59motorbitdoes anyone here knows about rb on d2?
17:41:43JdGordonpixelma: it hasnt happened in 3 years... its not going to happen
17:41:57*JdGordon is happy to suggest disabling the top item completly on the rec
17:42:18kugelno need if it's really only about combos
17:42:36 Quit funman ("free(random());")
17:42:59amiconnThe current *qucik*screen doesn't really deserve its name
17:43:46JdGordonneither the quick screen :)
17:44:15*pixelma wouldn't be upset if the quickscreen was abandoned completely :P
17:44:42JdGordonits useful for testing :p
17:44:49JdGordonbut yeah, I wouldnt be upset either
17:44:54kugelI use it a lot, so I would
17:45:03JdGordonwe can live with that though :D
17:45:20kugelamiconn: what's the issue with the recorder?
17:45:44JdGordonf1+up cant be held together?
17:45:49domonokymotorbit: just ask your question, if someone knows he will answer.
17:46:31amiconnF2 (or F3) + Up can't be detected simultaneously as they're connected to the same ADC channel
17:46:39pixelmamotorbit: I guess the best info you could find is in the forums, not many people with D2s are around I think (let alone D2+)
17:46:40motorbithow to make the touchscreen of teh d2 work in such a way it wont get up menue all the time when i press onto the right side of the screen?
17:46:44motorbiti calibrated it
17:46:44amiconnThis has nothing to do with display flipping
17:46:52motorbitits still unusable
17:47:41kugelI don't think that's a blocker for a top item though
17:49:17amiconnIt is, if you consider the original, really quick way to operate the quickscreen
17:49:24pixelmamotorbit: did you try setting the "absolute" (direct) touch screen mode? Just a guess, I don't know much about those targets
17:49:36motorbitits unsupported for the d2 atm
17:50:00kugelamiconn: but I don't consider this way
17:50:02motorbitbut k, looks as if this alpha still needs some work there, thats ok
17:50:07motorbitanother quesiton is:
17:50:23 Quit cg_ (Read error: 110 (Connection timed out))
17:50:35motorbitthe d2 original firmware is having a problem sorting id tags by id tag tracknumber if an album has more than 30 tracks.
17:50:44motorbitso i removed them along with the id tag filename
17:50:52kugeland even then, just meaning the up item is unusable isn't worth useless extra code for all other targets IMO
17:51:18motorbitthat works verry well with the original firmewar, however rockbox will only show "no filename" when using the database
17:51:37motorbitis there a way to make the database show the filename instead of a "missing tag" error?
17:53:44 Join kugel_ [0] (
17:54:00 Quit kugel (Nick collision from services.)
17:54:04 Nick kugel_ is now known as kugel (
17:54:27kugelamiconn: right now, extra logic would be needed to disable a top item for the recorder only
17:54:58amiconnWhy does that need extra logic? Just one or two ifdefs
17:55:22kugelifdef RECORDER in apps/ ?
17:55:40 Join funman [0] (n=fun@rockbox/developer/funman)
17:55:59motorbitmy english sucks: did none understod me or does noone knows the answer ? :D
17:56:02kugelanyway, that's simply uninteresting for the current quickscreen
17:56:02*amiconn thinks that other targets might have similar peculiarities when it comes to the quick quickscreen mode
17:56:02 Join Lss [0] (
17:56:27amiconnIt's also a matter of screen size
17:56:31kugelpossibly, the clip also has some impossible combos
17:56:51funmanm200v4 also I think
17:57:42funmanmotorbit: i think you have to tag your files to use the database, else there is no use for it
17:58:05funman"database" really is a "tag databse"
17:58:08motorbitso i can either tag them for original firmware or for rb :/
17:58:10pixelmaor use the filebrowser
17:58:13funmanfor filename browsing you can use the file browser
17:58:22motorbitbut it sucks
17:58:27motorbitbecause with the sd
17:58:28funmanmotorbit: the original firmware doesn't read id3?
17:58:36motorbiti wuld have to browse a long way each time
17:58:41motorbitit does
17:58:52motorbitbut it will show the filename if no tag is given
17:59:06motorbitinstead of a "no tag" error
17:59:46motorbitthe files do have tags... for album, genre, etc
17:59:52motorbitjust not for tracknumber and trackname
18:02:27 Join JdGordon_ [0] (i=ad8097f1@gateway/web/freenode/x-jcdksgpchzmpbfrw)
18:03:38 Join kachna [0] (
18:03:53 Quit robin0800 ("ChatZilla 0.9.85 [Firefox 3.5.2/20090810133528]")
18:04:20 Join robin0800 [0] (
18:04:59kugelamiconn: btw, i just committed r22228, it wasn't my code entirely :)
18:05:49pixelmakugel: but then you took responsibility for this code ;)
18:07:04 Quit robin0800 (Client Quit)
18:07:22 Join robin0800 [0] (
18:07:46kugelbut nice find indeed
18:08:26 Quit robin0800 (Client Quit)
18:08:44 Join robin0800 [0] (
18:09:37linuxstbmotorbit: We simply don't want the database to guess tags where they don't exist - it's the user's job to tag their files. What works for you will break things for others.
18:11:21 Quit robin0800 (Client Quit)
18:11:26kugelamiconn: I think I managed do a good job w.r.t to the screen size
18:11:42 Join robin0800 [0] (
18:12:21kugelit's fully functional on a clip even with 12-nimbus
18:12:46Unhelpfulamiconn: do you know if there is a greater-than-instruction-size alignment for functions on coldfire? the per-function deltas for r22289 for M3 total zero, but it somehow comes to +190...
18:13:39pixelmaUnhelpful: it's the pyros-roolku server producing this delta, look at the later delta jumps
18:14:25motorbiti can understand this one
18:14:27motorbitbut than
18:14:47motorbitthere is no guessing when using the aviable informations if the prefeared information is not given ;)
18:15:08kugelpixelma: captiankewl alsoo on the clip
18:15:34linuxstbmotorbit: But it sounds like your problem is a bug in the Cowon firmware...
18:15:36 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
18:15:47pixelmamotorbit: seeing FA34 won't help people on Ipods
18:15:52kugelthis binsize fluctuation is really getting bad
18:15:55motorbiti always had my files soted like that
18:16:17motorbitand there is this bug :)
18:16:38motorbiti just cant see where a error code does more sense than the filename
18:16:47Unhelpful190 is pretty big for somebody's toolchain being messed. :/
18:16:50motorbitsome fallback is not bad all the time
18:17:11pixelmakugel: what? ... ah you mean the small deltas - if I saw correctly they are on differnt clients though whereas the +190 bytes for the M3 is only that one
18:17:34kugelthere are +-20,22,32 and 94 also
18:17:51motorbitwho cares about ipods ? ^^
18:18:16linuxstbmotorbit: Now that's a logical argument that will win the day...
18:18:42*pixelma read that as an ironic statement
18:18:51motorbitwith one? that ipod one? or that fallback one?
18:18:51 Join spacetraveler [0] (
18:19:07funmandamn, still no call
18:19:30 Quit spacetraveler (Client Quit)
18:19:50moosfunman: send him an sms reminder?
18:20:20kugelJdGordon_: I laughed a bit when you said "also, using a default which is target dependant is a bad idea from a support POV"
18:21:29JdGordon_I was going to add a tongue-in-cheek remark about the stupidity f making it customisable at all :)
18:22:34kugelI mean I agree of course, do you have an idea for a good default?
18:22:57JdGordon_does it need one?
18:23:18kugelall settings have a default
18:23:31kugeland yes, I think it needs one to make clear there's a top item now
18:24:12JdGordon_i was going to say brightness but thats target dependant
18:24:31kugelbrightness is totally useful though
18:24:36JdGordon_start screen?
18:24:38kugelI have this in my qs for ages
18:25:13*kugel wouldn't like to choose an item which you probably change once very three month, if it all
18:25:25kugelwe already have one of this as default
18:25:31funmanmoos: good idea, i just did it
18:25:35JdGordon_like shuffle, repeat mode, and show files?
18:25:41kugelmaybe "warn on deleting dynamic playlist"? :p
18:26:15 Join domonoky1 [0] (
18:26:43 Quit Lss (Read error: 131 (Connection reset by peer))
18:27:13 Quit ehntoo (Read error: 110 (Connection timed out))
18:27:41kugelluckily settings_list.c is an #ifdef hell, that makes finding a all-target setting at least adventurous
18:27:43 Quit pamaury ("Quitte")
18:27:56 Join pamaury [0] (
18:29:07JdGordon_well... I suppose there isnt any real reason to not have a target dependant option there
18:29:08kugelfade on stop, or party mode?
18:29:37JdGordon_party mode maybe.. but the point of that one i so its hard to disable
18:29:43kugelor "interpret numbers while sorting" :D
18:30:31linuxstbBryanJacobs: Hi.
18:31:14pixelmaspeaking of target dependent default... JdGordon_ when are you going to change all other cabbiev2 versions to display volume in dB in place of the icon when changing volume?
18:31:51BryanJacobslinuxstb: Hi!
18:31:55JdGordon_do I haaaaaave tooo? :'(
18:32:03*BryanJacobs is still working
18:32:04*JdGordon_ runs to his room
18:33:14 Join DarkSpectrum [0] (
18:34:12pixelmait's a bit weird now that only one target has it and you committed this one (it's something I'd wanted to see on other targets too but changing it isn't the simplest, there are even versions that don't use viewports at all yet)
18:34:44 Quit DarkSpectrum (Read error: 104 (Connection reset by peer))
18:34:48 Join DarkSpectrum [0] (
18:34:54JdGordon_I dont think all targets really need to be identical
18:35:01kugelpixelma: I'm going to put up a new PLA patch soon, wanna test a bit over the weekend?
18:35:12 Quit robin0800 ("ChatZilla 0.9.85 [Firefox 3.5.2/20090810133528]")
18:36:40linuxstbBryanJacobs: How are things progressing? Are you able to post something to flyspray soon?
18:36:53pixelmaJdGordon_: in this case it would be very possible to be consistent and it's a valuable information in my eyes. I coul understand difference that are due to different screen size (or depth) but not this one
18:37:13linuxstbBryanJacobs: Your patch on flyspray buffers things 1:1 - have you fixed that in your current implementation?
18:38:03 Quit JdGordon_ (Ping timeout: 180 seconds)
18:38:04BryanJacobslinuxstb: well, my current implementation is based on the chunky stuff
18:38:18BryanJacobslinuxstb: and I'm planning on delivering on time :-)
18:39:04BryanJacobsthe "conservative" chunky implementation that has no fragmentation in cases where a rebuffer wouldn't be needed is what I'm focusing on
18:39:31linuxstb"On time" was the 10th August ;)
18:40:39 Join robin0800 [0] (
18:41:10linuxstbI'm not sure I understand your last sentence... What do you mean by "where a rebuffer wouldn't be needed" ?
18:42:20funmanremember students often get smarter than mentors :-)
18:42:31BryanJacobslinuxstb: hey, pencils down hasn't come yet
18:42:44 Quit domonoky (Read error: 110 (Connection timed out))
18:42:46BryanJacobsI mean where the data don't bump into another handle
18:43:04BryanJacobswhich could theoretically happen
18:43:18linuxstbWhy is that different? I would expect for most files, the data would bump into itself (i.e. be larger than the audio buffer).
18:44:58 Join JdGordon| [0] (n=Miranda@nat/microsoft/x-ndqlnhlyazdjjzki)
18:46:04BryanJacobslinuxstb: no, not where it bumps into itself, where it bumps into something else
18:46:23linuxstbThat's what I'm asking - why are those two cases different?
18:46:35BryanJacobsbecause if it bumps into itself it won't cause fragmentation
18:46:45BryanJacobschunky handles can interleave with eachother
18:47:49 Quit funman ("free(random());")
18:47:54 Quit robin0800 (Remote closed the connection)
18:48:01linuxstbWell, it will cause fragmentation as the early chunks in the files are used up.
18:48:12 Join robin0800 [0] (
18:48:41BryanJacobswhen we add a normal handle to the buffer the chunky ones are finished
18:48:48BryanJacobsthey become normal handles
18:49:32linuxstbYes, but as they're used up, couldn't the memory be fragmented?
18:49:33 Nick yawny is now known as roflcon (i=user36@
18:49:49BryanJacobsso the only time there will be fragmentation when a non-chunky handle is in the buffer (which is when we care) is *before* those early bits get used
18:50:30BryanJacobsand the memory pressure it causes will be at most as bad as that caused by a single file being as large as the multiple ones that were chunky
18:51:29 Quit robin0800 (Remote closed the connection)
18:51:48 Join robin0800 [0] (
18:52:10linuxstbBut anyway, going back to "bumping into something else", doesn't that just mean the buffer is full, and you stop buffering until more space is free?
18:53:29BryanJacobswhen the "something else" is not a chunky handle, yes
18:53:34kugelwhat does this fragmentation cause?
18:53:49BryanJacobskugel: space in the buffer that is not usable until something is freed
18:54:06kugelwhat if we try to play that back?
18:54:09motorbitk than, thanks
18:54:10 Quit motorbit ("ChatZilla 0.9.85 [Firefox 3.5.2/20090729225027]")
18:54:15BryanJacobskugel: that won't happen...
18:54:25BryanJacobseach handle "knows" how much stuff it has in it
18:54:39BryanJacobsthe fragmentation is when a handle is "full" but there's a gap before the next one
18:54:58kugelthere's still the problem on the clip where buffering just seems to stop at some point (will not rebuffer). virtually every file is bigger than the whole audio buffer
18:55:19BryanJacobskugel: maybe this will fix that :-P
18:55:31 Quit moos ("Rockbox rules the DAP world")
18:55:56 Quit robin0800 (Remote closed the connection)
18:56:14 Join robin0800 [0] (
19:01:19 Quit jgarvey ("Leaving")
19:03:02Unhelpfulsince the data is accessed by handle, might it be possible to compact by moving the data down?
19:04:06 Quit SUSaiyan` (No route to host)
19:04:44*JdGordon| requests everyone test 9886
19:05:02JdGordon|as usual... if it works you shouldnt notice anything different
19:05:44Unhelpfulis the latest patch "complete"?
19:06:14JdGordon|for the time being yes
19:06:40Unhelpfuli meant more "do i need to apply other patches in a set first?"
19:06:50JdGordon|oh, no, just the last one
19:07:52kugelI don't see any good reason for using buffer_alloc()
19:08:18kugelthe size is still static, that's just going to hide the binsize hit of JdGordon's upcoming experiments
19:08:52JdGordon|:) no, I'll put it back to try and see the delta... but I would imaguine it will eventully go to that anyway
19:09:50Unhelpfulum... i don't have apps/skin_engine/skin_parser.c ?
19:09:57linuxstbJdGordon|: That would seem more sensible - if the size becomes dynamic, move it to buffer_alloc(), but not before...
19:10:31JdGordon|ah fooeey... missed an svn add? :'(
19:10:52 Quit pamaury ("Quitte")
19:10:53 Join pamaury_ [0] (
19:10:57 Quit pamaury_ (Client Quit)
19:11:11 Join pamaury [0] (
19:12:19JdGordon|hmm... its already under versoin control apparently
19:12:35JdGordon|its not in the diff?
19:12:40Unhelpfulidunno, your diff contains edits to it, there are deletions as well as additions, and it's named wps_parser.c in my repo... i just did a fresh pull.
19:13:02JdGordon|its a copy/rename from wps_parser
19:13:02 Join funman [0] (n=fun@rockbox/developer/funman)
19:13:19funmanjust talked with the AMS guy on the phone, we pushed back the real talk to sunday
19:13:25kugelwhy didn't you just do the rename in a seperate commit?
19:13:49kugelfunman: he's testing your integrity!
19:13:56JdGordon|obviously to cause problems for everyone :/
19:13:59JdGordon|funman: aww
19:14:38***Saving seen data "./dancer.seen"
19:14:41 Quit efyx (Remote closed the connection)
19:14:58 Join efyx [0] (
19:16:03 Join ehntoo [0] (
19:17:09 Quit pamaury ("Quitte")
19:18:04Unhelpfulsvn diff should be generating a removal diff for the old file and a new-file diff for the new one if it's supposed to apply correctly at all :P
19:18:08funmani really should write a paper with what i want to say because i felt a bit lost
19:18:25Unhelpfulotherwise anybody trying to apply that to a clean archive has to do the rename themselves first. :P
19:19:22linuxstbfunman: Wouldn't it be easier to just do it by email?
19:19:34Unhelpfulactually, it looks like it *has* a removal diff for wps_parser.c... i wonder how it can also have deletions in skin_parser.c?
19:20:08JdGordon|I did svn cp to keep the history
19:20:26Unhelpfulshouldn't that be svn mv?
19:20:38JdGordon|i mean yes..t hats what i did
19:20:51JdGordon|which is just a svn cp followed by svn mv apparently anyway
19:21:44JdGordon|you want me to do a build for you? might be easiest
19:22:16linuxstbOr maybe just upload skin_parser.c?
19:22:34Unhelpfulif you can make a patch that's against "regular" svn that ought to do it...
19:23:03Unhelpfuli can't apply this one with git apply or with patch
19:23:45JdGordon|is it only that file?
19:24:52Unhelpfuli'm not sure... which files did you rename? perhaps it will apply if i duplicate those manually first...
19:26:03JdGordon|I dont remember... anything that was wps_<something> to skin_<something>
19:26:08JdGordon| is skin_parser.c
19:26:15JdGordon|which target, ill build for you
19:27:06JdGordon|except, i dont have a httpd at home so uploading could be difficultt
19:28:10funmanlinuxstb: answers to my emails happen to be very late and empty
19:28:30 Join Marvin_ [0] (n=5abfca64@gateway/web/cgi-irc/
19:28:46Marvin_i'm the guy with the nano 1g that forze and seemed to die
19:28:55Marvin_did not reboot etc
19:29:02Unhelpfuli think i've got it, just duplicating wps_parser.c as skin_parser.c made the patch apply... it's rather odd that the duplication of the entire file isn't part of what svn put in the patch. it should be treating it as the insertion of all lines in the file, rather than as the difference between the changed skin_parser.c and the original wps_parser.c
19:29:15Unhelpfulalso you have whitespace-only diffs :)
19:29:19Marvin_well after first 3 days it did nothing, tried to plug it in to charge, nothing
19:29:42Marvin_i was on a vacation for a week and NOW tried to plug it in, and it is ALIVE, started to charge
19:30:36Marvin_just for information, some of you guyes here adviced that it should run out of juice in a night or so and then shut down and get out of a dead loop, it took much longer
19:31:12CIA-61New commit by mt (r22310): Revert the asm fixmul stuff for now.
19:31:26 Quit funman ("leaving")
19:31:31amiconnUnhelpful: Doing an 'svn mv' and editing the same file in a single commit isn't handled very well by svn. svn warns you if you mv a file that has local edits, but it cannot warn you in the other case
19:31:36 Join webguest25 [0] (n=d5e9675c@gateway/web/cgi-irc/
19:32:18Unhelpfulah... doing that makes svn diff rather useless. :/
19:32:39 Join gtkspert_ [0] (
19:37:01CIA-61New commit by mt (r22311): Support for playback of atrac3 audio in rm, in sim.
19:43:22CIA-61New commit by mt (r22312): Fix yellow.
19:43:30UnhelpfulJdGordon: loaded icatcher on my beast, a bunch of the icons were solid pink, then a data abort. :D
19:44:16JdGordon|for some reason the icatcher backdrop has those pink areas to show where the images should go so they arnt beign drawn wrong...
19:44:23JdGordon|that said, that shouldnt happen :p
19:44:27JdGordon|where is the abort?
19:44:49Unhelpfuli can't get into that right now. :)
19:44:53 Quit gtkspert (Read error: 101 (Network is unreachable))
19:45:20 Quit linuxstb (Read error: 110 (Connection timed out))
19:47:19JdGordon|hmm... some alignement issues in icatcher on my video... doubt thats my patch though
19:49:44Unhelpfulas in pixmap alignment or data alignment? (the latter can trigger a data abort, after all)
19:50:28JdGordon|the volume/power/time lines
19:50:52 Join pamaury [0] (
19:54:15kugelsvn also clearly fails horribly when you have editted a file while the file is moved upstream
19:58:28Unhelpfulkugel: git is not too nice about that either
20:05:26mtatrac3 works on target too now (e200) !! .. but it's incredibly slooow.
20:06:13Unhelpfulif you delete a file, and then it's modified upstream, your rebase will be, um, bad.
20:07:17 Quit webguest25 ("CGI:IRC (EOF)")
20:07:20 Join cg_ [0] (
20:10:56 Quit Blue_Dude ("ChatZilla 0.9.85 [Firefox 3.5.2/20090729225027]")
20:12:36kugelUnhelpful: I can imagine, the way how rebase works is too simply for that ;)
20:13:29Unhelpfulalso, touch thou not anything under .git, so sayeth the torvalds. seriously, that's the only time i've irretrievably ruined my git repo. :)
20:13:36kugelbut svn silently keeps the non-moved file (appears in svn diff etc), while the moved file has no modifications
20:16:36amiconnUnhelpful: Coldfire sections are 32 bit aligned. Instruction size is variable, but is a multiple of 16 bits. No further alignment is applied to individual functions
20:17:20Unhelpfulamiconn: it turns out to be a dodgy build client... the next few builds had small or 0 deltas (on the same client), then it moves clients and gets a -190
20:17:45amiconnYeah, got that, but did that check anyway
20:26:18 Quit kugel (Remote closed the connection)
20:41:59 Quit pamaury ("exit(*(int *)0 / 0);")
20:43:09 Quit BryanJacobs ("Java user signed off")
21:00:36mtheh 32.30% realtime
21:04:52 Join webguest52 [0] (n=44621cc8@gateway/web/cgi-irc/
21:05:38 Quit webguest52 (Client Quit)
21:07:27JdGordon|mt: just wait, that will be 150% pretty soon... just wait for some asm magic :)
21:09:25 Join DarkSpectrum- [0] (
21:09:25 Quit DarkSpectrum (Read error: 104 (Connection reset by peer))
21:09:28 Nick DarkSpectrum- is now known as DarkSpectrum (
21:11:57mtJdGordon|: Yeah hopefully :) .. there are also a couple of memcpy's that are applied on each audio frame, I guess getting rid of those should have a significant effect.
21:14:42***Saving seen data "./dancer.seen"
21:14:52 Quit flydutch ("/* empty */")
21:20:35 Join aaron424 [0] (
21:28:47*Unhelpful wonders if it might *be* a good idea make the buffer less file-data-specific, and more like a malloc-by-handle... it would reduce the allocation overhead for things that are using bufalloc for non-file buffers, and "special" buffer types could be handled by having a type/flags int in place of all of the other header data, which could be followed by a struct containing whatever header data applies to that type
21:29:22 Quit LambdaCalculus37 ()
21:33:42 Quit DarkDefender (Remote closed the connection)
21:36:45 Quit amiconn (Nick collision from services.)
21:36:46 Join amiconn_ [0] (i=quassel@rockbox/developer/amiconn)
21:37:06 Nick amiconn_ is now known as amiconn (i=quassel@rockbox/developer/amiconn)
21:39:01 Quit pixelma (Nick collision from services.)
21:39:01 Join pixelma_ [0] (i=quassel@rockbox/staff/pixelma)
21:39:19 Nick pixelma_ is now known as pixelma (i=quassel@rockbox/staff/pixelma)
21:41:32 Join jernejovc_ [0] (
21:42:18 Join funman [0] (n=fun@rockbox/developer/funman)
21:42:42 Quit DarkSpectrum (Read error: 104 (Connection reset by peer))
21:43:04 Quit jernejovc (
21:43:04 Quit z35 (
21:43:04 Quit jhulst (
21:43:04 Quit n17ikh (
21:43:04 Quit togetic (
21:43:04 Quit GodEater (
21:43:04 Quit AfterDea1h (
21:43:04 Quit lostlogic (
21:43:04 Quit Hadaka (
21:43:04 Quit Tuplanolla (
21:43:12NJoinn17ikh [0] (
21:43:59 Join Jester [0] (n=Jester@unaffiliated/jester)
21:44:16 Join GodEater [0] (n=bibble@
21:44:33NJoinTuplanolla [0] (
21:44:43NJoinjhulst [0] (
21:45:06 Join AfterDeath [0] (n=icxcnika@freenode/weird-exception/network-troll/afterdeath)
21:45:45NJoinlostlogic [50] (n=lostlogi@rockbox/developer/lostlogic)
21:49:00JdGordon|Unhelpful: which uses do you have in mind?
21:49:06JesterHi. Sorry if this has been answered somewhere but I couldn't find it in the FAQ or google. I'm wondering if installing rockbox on an ipod will affect the ability of external devices to control the ipod. For example a car audio unit might require a certain firmware on an ipod to be able to control it, so it's not far-fetched to think rockbox might destroy that feature
21:49:18Unhelpfulok, so, my argument i guess here is that a dynamic allocator is a *good* thing when it's not used thoughlessly - so that themes, for example, always take up the space they need.
21:49:32JdGordon|Jester: bing is better :)
21:49:46gevaertsJdGordon|: please stay on-topic!
21:49:50Unhelpful(and no more)
21:49:51JdGordon|Jester: yes, rockbox hasnt had much luck with external controllers... but its getting better
21:50:11JesterJdGordon|: that's too bad
21:50:21funmanJester: you can choose to run the OF anytime instead of rockbox (dual boot), so you could use your external devices with the OF
21:50:22Unhelpfulhowever, you'd still be able to boot the OF for that purpose... although of course the OF can only see files loaded with itunes, and can only play certain formats.
21:50:41gevaertsJester: there's only one way to find out if your particular device works with rockbox
21:51:01funmansome devices work with rockbox, so nothing says yours shouldn't
21:51:32Jesterfunman: really. That's pretty cool, unfortunately my main use for rockbox is to be able to play flac files in my car in combination with an audio unit that can control the ipod :/
21:51:46funmanjust try
21:51:53UnhelpfulJdGordon: another case i had where we'd get an improvement from bufalloc-as-malloc is cover art. right now we load album art for each file. we *could* rather easily see if the file in question has been loaded, and re-use it, but right now we don't have a way to free that single buffer usefully - unless we could compact the buffer. :)
21:52:14Unhelpfulthis would save AA_size * (num_tracks - 1) when loading a whole album
21:53:09JdGordon|for that case, I rekon adding some smarts to buffering.c would be a better solution
21:53:21 Join Hadaka [0] (
21:53:30gevaertsJdGordon|: yes, but IIUC that's what Unhelpful is getting at
21:53:41*funman owes a beer to anyone who helps fixing the playback stopping on sansa clip
21:56:21JdGordon|Unhelpful: but yeah, thats the same issue as buffering the same track more than once if its in the playlist more than once
21:56:30gevaertsUnhelpful: loading a file to the audio buffer *is* dynamic allocation, so I don't think there should be a general veto against replacing that with a different sort of dynamic allocation, at least not a veto based on the new solution using dynamic allocation
21:57:20NJointogetic [0] (n=togetic@unaffiliated/ibuffy)
21:58:35JdGordon|Unhelpful: buffering prsumably stores all the active handles, so you could add a id field or something to the AA type handle and then do some magic to reuse the handle isntead of doing a actual file buffer
21:58:47 Join Sajber^ [0] (
21:58:57JdGordon|I dont think turning alloc into a more mallocy system is a idea which will go down well
22:00:25 Quit funman ("free(random());")
22:01:12 Quit Sajber^ (Read error: 104 (Connection reset by peer))
22:01:25 Join SUSaiyan [0] (
22:03:31 Join Sajber^ [0] (
22:03:33UnhelpfulJdGordon: except that we probably have multiples of an AA much more often than multiples of a track.
22:04:06JdGordon|yeah ture
22:04:17 Quit Sajber^ (Read error: 54 (Connection reset by peer))
22:04:47JdGordon|which is why I'd be all for adding the code especially for AA
22:04:48 Join Xqtftqx1 [0] (
22:04:52UnhelpfulJdGordon: i'm really pretty much only suggesting that we add a few things - the ability to reclaim fragmented free space, and possibly the ability to free some no-longer-needed audio data on demand
22:05:33 Join Sajber^ [0] (
22:06:12 Join spacetraveler [0] (
22:06:45Unhelpfulthis would be a win for dynamic WPS allocation also, because if it can use bufalloc then you don't need to stop playback... as i understand it you need to stop playback for buffer_alloc?
22:07:24JdGordon|how could you reclaim free space without breaking pointers?
22:08:31UnhelpfulJdGordon: by having a firm requirement that pointers be re-fetched any time that a yield might have happened, etc. or possibly with a "may move" mutex, which you'd need something like anyway if the handle is used by a COP process.
22:08:33 Quit Sajber^ (Read error: 104 (Connection reset by peer))
22:09:02JdGordon|yeah, that adds way too much overhead for alot of what uses buf_alloc
22:09:06 Quit spacetraveler (Client Quit)
22:09:14JdGordon|especially when I tihnk at east 1 thread is put there
22:09:53 Join Omlet05 [0] (
22:09:58 Join pamaury [0] (
22:10:34 Quit fuzzie (Read error: 104 (Connection reset by peer))
22:11:12JdGordon|we have talked about the idea where everything gets dumped and needs to be alloc'ed again.. but that has just as many nasty issues
22:12:02Unhelpfulmay-move is the only COP-safe method anyway. the way i see that working would be that a mutex is held while working on the buffer data, with the understand that it must be released regularly. any time the mutex is released, the buffer thread is permitted to compact, and anything with a handle must re-fetch
22:14:57Unhelpfulproblems with that method include that that there could only be one thing using that method, because of the mutex... unless we have more complicated thread synchronization primitives we could use in a similar fashion.
22:15:25JdGordon|I cant see this being useable with anything that uses buffer_alloc() today, but yeah, it sems ok for the apps/ level buffering (buffering.c stuff)
22:17:36Unhelpfuli'm not saying we *need* a malloc badly... i'm just saying that perhaps dynamic allocation for some items would save more space to buffer audio, and/or remove some artificial limits as well, possibly to the degree of being worth the overhead.
22:19:39UnhelpfulWPS is one place where we see trouble at both ends of the spectrum - themers who have actually hit the artifical limits we have, and also large amounts of wasted space for many not-so-big-WPS
22:19:52 Quit Omlet (Read error: 110 (Connection timed out))
22:22:40 Join Zagor [242] (
22:22:47JdGordon|yeah, but there is another problem... the way allocation is done (in my path) for things like images and stuff... a regular pointer is tored, to allow it to be more flexible everything would need to be converted to offsets or handles, neither is very nice
22:23:56 Join TheSeven [0] (
22:24:50Unhelpfulthat's true. if we used something buflib-like, where there's a handle table, it would greatly reduce the cost of re-fetching... however, buflib is not very ideal if you might, say, make a lot of allocations, then free *most* of them while holding one for a long time
22:25:36*JdGordon| slides the conversation sideways a bit
22:25:52Unhelpfulhandles in our linked-list buffer go away when their associated allocations do. handles in buflib are free as handles immediately on free, but only free as space if no used handles of higher index exist.
22:26:23JdGordon|this is why I want to make the skin buffer user customisable.. apparently some people like the idea of changing themes 15 times a day, while others maybe only once ever, so only using the absolute minimum ram is good for them but bad for the first
22:29:32JdGordon|Unhelpful: we already move handles around dont we?
22:31:19 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
22:34:48UnhelpfulJdGordon: as near as i can tell, we don't move buffers very often, and we don't do it automatically to defragment free space
22:35:41 Quit evilnick ("Page closed")
22:36:13domonoky1Zagor: can you take a look at FS #10525 (genlang fixes) ? :-)
22:37:25JdGordon|Unhelpful: I dont know how often they move, and was under the impression the buffering wsnt supposed to get into a fragmented state?
22:37:37Zagordomonoky1: a patch! how about simply looking for a '.' in $rev and then treat it like a release? liming revisions is bound to backfire one day :)
22:37:41 Nick roflcon is now known as alexja (i=user36@
22:38:03 Quit aaron424 (Remote closed the connection)
22:38:09 Join jgarvey [0] (
22:38:15 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
22:38:43domonoky1Zagor: good idea (me goes searching for perl tutorials) :-)
22:39:03Zagorthat's the spirit
22:43:21 Join thegeek [0] (
22:43:59 Join thegeek_ [0] (
22:44:00 Quit thegeek (Client Quit)
22:45:13JdGordon|domonoky1: if you're playing around with stuff for release... there is a patch of mine in the tracker to get the release number into a RB_RELEASE define in the code if you want to take over that one also?
22:48:27 Quit GeekShado_ (Read error: 110 (Connection timed out))
22:51:33domonoky1 Zagor: patch updated :-)
22:53:19domonoky1JdGordon|: not really interessted :-)
22:54:00JdGordon|well fine then.. :'(
23:03:38 Join petur [0] (
23:08:50 Nick courtc_ is now known as courtc (n=court@unaffiliated/courtc)
23:10:06 Quit captainkwel ("Page closed")
23:11:12CIA-61New commit by roolku (r22313): mrobe100: power off HDD (courtesy of lowlight) ~2h extra playtime
23:12:41 Quit bubsy (Read error: 60 (Operation timed out))
23:14:31tmztPalm Pre Starts To Get More Virtual With Its Keyboard
23:14:31DBUGSent KICK tmzt to server
23:14:31tmztBy Evan Selleck on Friday, Aug 14th 2009 No Comments
23:14:31tmztWorth Reading?
23:14:31Kick(#rockbox tmzt :No flooding!) by logbot!n=bjst@rockbox/bot/logbot
23:14:47***Saving seen data "./dancer.seen"
23:14:49 Join tmzt [0] (n=tmzt@
23:14:59tmztI appologize for that
23:20:23 Join wincent [0] (
23:24:06 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
23:34:36 Quit pamaury ("exit(*(int *)0 / 0);")
23:35:35 Quit ehntoo (Read error: 110 (Connection timed out))
23:39:56 Join ehntoo [0] (
23:43:54 Join Thundercloud [0] (
23:56:32 Quit bmbl ("Bye!")

Previous day | Next day