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 2010-02-02

00:00:37CIA-8New commit by pamaury (r24445): Various dircache improvements: reduce size of a path buffer to MAX_PATH and rewrote copy_path to improve stack usage. Also simplified get_entry.
00:01:28JdGordon_pamaury: if you're getting into the dircache code, there is a bug which would be good if you could look at
00:01:40moospamaury: I remenber discussion here about this big dircache reserve. I don't know if you know about this part?
00:02:17pamauryI don't know the details but it an area used by dircache when the number of entries increases no ?
00:03:34moosyeah sort of a temp buffer if there are at runtime modifications IIRC
00:03:47moosof the cache
00:04:01JdGordon_9506 is maybe the buffer one you are tihnking of?
00:04:27moosyou have it in the dircache debug menu
00:06:12 Quit tarbo (Ping timeout: 246 seconds)
00:06:35 Join matsl [0] (
00:07:12FlynDicefunman: (for the logs) I have managed to flash my shiny new clip+ using unmodified dualboot.S and it's working just fine. Will start trying buttons now!
00:08:19pamauryI've looked at the two patches. The 9506 seems small and relatively safe. The second one involes more changes in areas I don't master (buffering) so I won't touch it right now. I think there are others things to optimized. And there is at least one dircache bug I'm aware of which is embarassing
00:08:33JdGordon_FlynDice: yeah, thats as far as I got before getting a brick
00:08:42 Quit einhirn (Read error: Connection reset by peer)
00:08:55*pamaury points to all the green deltas for r24445
00:09:29 Nick Ypsy is now known as YPSY (
00:09:35FlynDiceJdGordon_: Well I got the 2gig for just that reason.... ;-)
00:20:50 Join tarbo [0] (~me@unaffiliated/tarbo)
00:21:52 Quit yosafbridge (Quit: Coyote finally caught me)
00:22:04*amiconn scratches head
00:22:10 Join yosafbridge [0] (
00:22:21amiconnThis coldfire doesn't behave as documented
00:22:41amiconn(or rather, the specs aren't precise enough)
00:23:19moospamaury: congrates for the deltas and the simplifications/improvements
00:23:46*moos proposes that amiconn update the coldfire docs :)
00:24:44pamaurymoos: let's hope I didn't break anything ! I have more green deltas in mind
00:25:08JdGordon_pamaury: dont go deleting whole features!
00:25:14moosgo go go !!! and if you break something, prepare you to have reports
00:25:23mooshaha :)
00:25:42*JdGordon_ thinks charcell is a good place to start thoguh if you do want to remove lots of code
00:26:13pamaurywhat is charcell ?
00:27:53 Quit bertrik (Quit: De groeten)
00:28:32 Join stooo [0] (
00:29:05 Join Sajber^ [0] (
00:29:07 Quit tarbo (*.net *.split)
00:29:08 Quit FlynDice (*.net *.split)
00:30:04 Join FlynDice [0] (
00:34:37 Quit stooo (Quit: Leaving.)
00:36:00 Quit pamaury (Quit: abort();)
00:37:06 Quit Stephen__ (Quit: Leaving)
00:39:23 Quit domonoky (Read error: Connection reset by peer)
00:40:32Strife89|DesktopDoes r24440 mean that we have DMA read and write on some PortalPlayer targets?;revision=24440
00:42:38 Quit ender` (Quit: Some people have one of those days. I've had one of those lives.)
00:49:51 Quit petur (Quit: Zzzz)
01:02:43 Join killan_ [0] (
01:02:43 Quit killan (Read error: Connection reset by peer)
01:05:21 Quit akur (Ping timeout: 246 seconds)
01:08:12 Join fdinel [0] (
01:09:02 Join tarbo [0] (~me@unaffiliated/tarbo)
01:09:59 Quit Casainho (Ping timeout: 264 seconds)
01:11:24 Quit pixelma (Ping timeout: 245 seconds)
01:11:55 Quit amiconn (Ping timeout: 260 seconds)
01:15:10 Quit tarbo (Ping timeout: 272 seconds)
01:15:15 Join dmb_ [0] (~Dmb@unaffiliated/dmb)
01:16:38 Quit matsl (Remote host closed the connection)
01:17:34 Join Horschti [0] (~Horscht2@xbmc/user/horscht)
01:18:44 Quit Horscht (Ping timeout: 256 seconds)
01:19:04 Nick fxb is now known as fxb__ (
01:22:25 Join akur [0] (
01:24:16 Join amiconn [0] (quassel@rockbox/developer/amiconn)
01:24:20 Quit dfkt (Quit: -= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.)
01:24:39 Join pixelma [0] (quassel@rockbox/staff/pixelma)
01:25:22 Join stripwax [0] (
01:25:24 Join hebz0rl [0] (
01:27:12 Quit robin0800 (Ping timeout: 272 seconds)
01:29:06 Join tarbo [0] (~me@unaffiliated/tarbo)
01:31:26 Join robin0800 [0] (
01:33:07 Quit akur (Quit: Leaving.)
01:34:11***Saving seen data "./dancer.seen"
01:36:23 Quit tarbo (Ping timeout: 264 seconds)
01:37:28 Quit dmb_ (Ping timeout: 240 seconds)
01:38:26 Quit efyx_ (Remote host closed the connection)
01:41:57CIA-8New commit by stripwax (r24446): Copied Tremor improvement (bitreverse) from Trunk
01:42:22stripwax^- copied to the mdctexp branch
01:47:47 Quit Farthen (Ping timeout: 264 seconds)
01:47:52 Quit pixelma (Ping timeout: 265 seconds)
01:48:06 Quit amiconn (Ping timeout: 272 seconds)
01:49:01 Join tarbo [0] (~me@unaffiliated/tarbo)
01:49:02 Join Farthen [0] (
01:51:47 Part toffe82
01:53:10 Join amiconn [0] (quassel@rockbox/developer/amiconn)
01:53:17 Join pixelma [0] (quassel@rockbox/staff/pixelma)
01:55:06CIA-8New commit by stripwax (r24447): Might as well copy over mdct2.c from Trunk as well
01:55:52 Quit tarbo (Ping timeout: 240 seconds)
01:57:58 Quit robin0800 (Remote host closed the connection)
02:00:01 Quit moos (Quit: ChatZilla 0.9.86 [Firefox 3.6/20100115144158])
02:08:01 Quit stripwax (Quit:
02:10:28 Join tarbo [0] (~me@unaffiliated/tarbo)
02:17:31 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
02:18:20 Quit tarbo (Ping timeout: 258 seconds)
02:20:36 Join Rob2222 [0] (
02:23:45 Quit Rob2223 (Ping timeout: 246 seconds)
02:31:14 Join tarbo [0] (~me@unaffiliated/tarbo)
02:37:39 Quit PaulJam (Ping timeout: 265 seconds)
02:38:16 Quit tarbo (Ping timeout: 240 seconds)
02:41:38 Quit JdGordon_ (Quit: Page closed)
02:44:35 Quit anewuser (Quit: =o)
02:48:09 Quit yosafbridge (Quit: Coyote finally caught me)
02:48:27 Join yosafbridge [0] (
02:52:08 Join tarbo [0] (~me@unaffiliated/tarbo)
02:55:03 Join JdGordon1 [0] (
03:00:09 Quit tarbo (Ping timeout: 246 seconds)
03:06:18 Join arbingordon [0] (
03:07:29CIA-8New commit by jdgordon (r24448): fix and promote fuze and e200v2 to stable
03:13:17 Join tarbo [0] (~me@unaffiliated/tarbo)
03:17:02mc2739JdGordon: tools/ also needs updating for Fuze and e200v2 status
03:18:10JdGordon1didnt that get updated just now also?
03:18:56mc2739no, only
03:19:47JdGordon1ffs... serves me right from checking in from my phone :/
03:20:03 Quit tarbo (Ping timeout: 258 seconds)
03:20:51CIA-8New commit by jdgordon (r24449): ditto
03:24:23 Quit JdGordon1 (Ping timeout: 240 seconds)
03:33:48 Join tarbo [0] (~me@unaffiliated/tarbo)
03:34:14***Saving seen data "./dancer.seen"
03:42:16 Quit tarbo (Ping timeout: 276 seconds)
03:46:10 Quit Sajber^ (Read error: Connection reset by peer)
03:47:53 Quit fyrestorm (Read error: Connection reset by peer)
03:49:12 Join Sparkeater [0] (
03:49:31Sparkeaterhi, can I be the 965th person to ask when the next release is?
03:50:56Sparkeaterthat was a meta-question.... my question is "when is the next release" ?
03:51:48JdGordontomorow or wedndedsay deending on your and our timezones
03:52:01Sparkeateralso I would like to report that superdom isn't working properly
03:53:43 Quit shaggy-h ()
03:54:19Sparkeaterwhen you assign how many farms/factories etc, you can have 3/3 for computer/player or 2/2 for each, but any other number seems to screw up completely and the player seems to get as many factories as farms no matter what.. like I tried to play it with 1 factory and 3 farms for each player, but it gives you 3 factories and 3 farms and gives the computer 1 factory and 3 farms
03:54:21 Join tarbo [0] (~me@unaffiliated/tarbo)
03:55:13Sparkeaterthat happens for pretty much any combination, except 4 of each for both players, then something screws up totally
03:55:54 Join shaggy-h [0] (
03:56:20Sparkeaterbut the only combinations it allows is like 1/1 factory/farm for both players, 2/2 or 3/3... it won't let you have 2/3 factories/farms or 1/3 or any other combo... it just goes random and assigns the human player too many of something
03:58:14 Quit MethoS- (Remote host closed the connection)
04:01:04 Quit tarbo (Ping timeout: 246 seconds)
04:14:36 Quit antil33t (Read error: Connection reset by peer)
04:14:43 Join antil33t [0] (
04:16:21 Join tarbo [0] (~me@unaffiliated/tarbo)
04:19:54JdGordonSparkeater: put all that in a bug report on the tracker
04:23:23 Quit tarbo (Ping timeout: 240 seconds)
04:31:38 Quit TheSeven (Killed (NickServ (GHOST command used by The_Seven)))
04:31:52 Join The_Seven [0] (~theseven@rockbox/developer/TheSeven)
04:32:02 Nick The_Seven is now known as TheSeven (~theseven@rockbox/developer/TheSeven)
04:32:13 Join BHSPitMonkey [0] (
04:32:13 Quit BHSPitMonkey (Changing host)
04:32:13 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
04:36:36SparkeaterI forgot my password so I can't put that on the tracker
04:37:18 Join tarbo [0] (~me@unaffiliated/tarbo)
04:44:49 Quit tarbo (Ping timeout: 272 seconds)
04:48:47krazykitso tell flyspray to reset the password. it's on there somewhere
04:52:48 Join Barahir_ [0] (
04:53:06SparkeaterI'm at a library computer and can't really screw around too much, if someone wants this computer I have to get off it
04:53:38 Quit Barahir (Read error: Operation timed out)
04:57:34 Join tarbo [0] (~me@unaffiliated/tarbo)
04:57:37 Join fyrestorm [0] (
04:58:34 Join DerPapst1 [0] (
04:58:35 Join S_a_i_n_t [0] (S_a_i_n_t@
04:59:48 Quit DerPapst (Ping timeout: 264 seconds)
05:02:47 Join yzflcyq [0] (
05:02:51 Quit fyrestorm (Quit: lamers envy me like they envy bill g -- main boot xp, just the way it should be!)
05:04:30yzflcyqI found it diffcult to make rockbox all Chinese UI.
05:04:49 Quit tarbo (Ping timeout: 248 seconds)
05:06:33JdGordonare you using a font with the chinese characters?
05:07:12yzflcyqThat's not the point.I mean that something can't be translated.
05:08:12JdGordonwhats the problem?
05:11:04 Quit Horschti (Quit: Verlassend)
05:12:30yzflcyqFor example,'Touchscreen Settings' cannot be translated into Chinese.
05:14:04krazykityou mean where you'd expect the translation for "touchscreen settings", you see the english?
05:14:48 Nick Blast_Tyrant is now known as BlastTyrant (
05:16:23krazykiti don't know much about how the translations work, but it sounds like there are translations of strings missing for the chinese lang
05:17:57yzflcyqWell I have translated all into Chinese and commited a diff file.
05:18:28 Join tarbo [0] (~me@unaffiliated/tarbo)
05:24:27 Quit yzflcyq (Quit: CGI:IRC)
05:26:12 Quit tarbo (Ping timeout: 258 seconds)
05:32:51Sparkeateris the new rockbox going to include much new stuff?
05:34:17***Saving seen data "./dancer.seen"
05:35:59S_a_i_n_tsparkeater: you could try it and see for yourself ""
05:36:17Sparkeateris it finished?
05:36:26S_a_i_n_tnot quite...
05:36:35S_a_i_n_tthat's the release candidate build.
05:36:49S_a_i_n_tfollow the link.
05:36:50SparkeaterI'll wiat until tomorrow for the final
05:37:07Sparkeaterthat'll be more exciting
05:37:54S_a_i_n_tfair enough, I always use SVN builds myself...if you're using 3.4 at the moment then you'll see a fair bit of difference.
05:38:22SparkeaterI'm using a build from 3-4 weeks ago
05:38:52 Join tarbo [0] (~me@unaffiliated/tarbo)
05:39:16S_a_i_n_tthen you won't see too much difference.
05:39:43SparkeaterI hope someone has fixed the game superdom
05:39:59Sparkeateror I hope this version of superdom is build-specific-broken
05:40:17S_a_i_n_twhat's wrong with it?
05:40:23*S_a_i_n_t never plays it.
05:41:27SparkeaterI play it all the time, it kickas ass. except you can give like 2/2 factories/farms to the human and computer, or 3/3, or 1/1, but any other combination screws up and it gives the human player too many factories
05:42:34Sparkeaterlike I want to play with 1 factory and 3-4 farms for each player, but it just screws up and gives some random amount of factories to the human player (more than the computer)
05:43:34 Part froggyman
05:45:32SparkeaterI'll be back a bit later
05:45:38 Quit Sparkeater (Quit: CGI:IRC)
05:49:23 Join thegeek [0] (
05:49:36 Quit tarbo (Ping timeout: 264 seconds)
05:54:59 Quit BlastTyrant (Quit: Skynet has taken control)
05:58:51 Join tarbo [0] (~me@unaffiliated/tarbo)
06:05:43 Quit tarbo (Ping timeout: 240 seconds)
06:09:41 Join ender` [0] (
06:16:32 Join Rob2223 [0] (
06:19:45 Quit Rob2222 (Ping timeout: 245 seconds)
06:20:17 Join tarbo [0] (~me@unaffiliated/tarbo)
06:21:46 Join BlastTyrant [0] (
06:24:48 Quit fdinel (Quit: Miranda IM! Smaller, Faster, Easier.
06:26:35 Quit arbingordon (Quit: `)
06:27:25 Quit tarbo (Ping timeout: 272 seconds)
06:31:04 Quit scorche (Read error: Connection reset by peer)
06:31:34 Join scorche [0] (~scorche@rockbox/administrator/scorche)
06:36:55 Quit hebz0rl (Remote host closed the connection)
06:40:22 Join tarbo [0] (~me@unaffiliated/tarbo)
06:47:25 Quit tarbo (Ping timeout: 256 seconds)
07:00:24 Join tarbo [0] (~me@unaffiliated/tarbo)
07:08:23 Quit tarbo (Ping timeout: 240 seconds)
07:23:29 Join tarbo [0] (~me@unaffiliated/tarbo)
07:26:50 Quit liar (Ping timeout: 245 seconds)
07:31:00 Quit tarbo (Ping timeout: 264 seconds)
07:34:18***Saving seen data "./dancer.seen"
07:46:00 Join tarbo [0] (~me@unaffiliated/tarbo)
07:54:11 Quit tarbo (Ping timeout: 272 seconds)
07:57:03*GodEater does the release day dance
08:09:07 Join tarbo [0] (~me@unaffiliated/tarbo)
08:10:47 Quit BHSPitMonkey (Remote host closed the connection)
08:14:51 Join arbingordon [0] (
08:15:55 Join Zagor [0] (
08:15:56 Quit Zagor (Changing host)
08:15:56 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
08:16:23 Quit tarbo (Ping timeout: 240 seconds)
08:21:58 Join dmb_ [0] (~Dmb@unaffiliated/dmb)
08:27:46 Join PaulJam [0] (
08:31:23 Join tarbo [0] (~me@unaffiliated/tarbo)
08:32:22 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
08:40:01 Quit tarbo (Ping timeout: 246 seconds)
08:45:40 Join flydutch [0] (
08:45:40 Join petur [0] (~petur@
08:45:40 Quit petur (Changing host)
08:45:40 Join petur [0] (~petur@rockbox/developer/petur)
08:47:45 Quit AndyI (Ping timeout: 248 seconds)
08:53:52 Join AndyI [0] (~pasha_int@
08:54:06 Join tarbo [0] (~me@unaffiliated/tarbo)
09:01:52 Quit tarbo (Ping timeout: 260 seconds)
09:16:42 Join tarbo [0] (~me@unaffiliated/tarbo)
09:23:53 Quit tarbo (Ping timeout: 240 seconds)
09:31:49 Join davidb_csh [0] (davidb@
09:32:38 Join maruk [0] (
09:34:20***Saving seen data "./dancer.seen"
09:36:11 Join Rob2222 [0] (
09:38:34 Join tarbo [0] (~me@unaffiliated/tarbo)
09:38:55 Quit Rob2223 (Ping timeout: 245 seconds)
09:46:00 Quit tarbo (Ping timeout: 264 seconds)
09:55:19 Quit BlastTyrant (Ping timeout: 258 seconds)
10:00:17 Join S_a_i_n_t_ [0] (S_a_i_n_t@
10:01:01 Join tarbo [0] (~me@unaffiliated/tarbo)
10:02:48 Quit S_a_i_n_t (Ping timeout: 264 seconds)
10:02:52 Nick S_a_i_n_t_ is now known as S_a_i_n_t (S_a_i_n_t@
10:03:09AlexPThere are a few language updates on flyspray - if anyone fancies committing them that'd be great, if not I'll do it tonight when I get home.
10:05:35 Quit perfectdrug_ (Ping timeout: 245 seconds)
10:08:38 Quit tarbo (Ping timeout: 252 seconds)
10:17:56pixelmasomeone *replaced* one language string with another - not deprecating the first and adding the second :\
10:21:19pixelmaand the most confusing there is that he also changed it (to the english phrase) in *some* lang files. As German was one of those, I didn't see the change and only realised just now looking over rasher's translation stats (source and dest being the same)
10:22:26pixelmaLANG_COMPRESSOR if anyone is interested
10:23:26 Join tarbo [0] (~me@unaffiliated/tarbo)
10:26:28pixelmaDanish, German, French, Italian, Polish, Swedish, Tagalog and Walloon are affected (see r22832)
10:29:51rasherYeah, I wanted to complain about that but forgot
10:30:14 Join perfectdrug [0] (
10:31:08 Quit tarbo (Ping timeout: 276 seconds)
10:31:22GodEaterthere's a language called walloon?
10:32:21pixelmaI think it's a French dialect (much like we have a Brasilian Portugese)
10:33:20pixelmadialect might be not exactly correct
10:39:26topikit's the french spoken in the south of belgium
10:39:53gevaertsYes and no
10:40:00topikmostly regular french but differences like numbers
10:40:10gevaertsno, that's now Walloon
10:40:12gevaertsThat's French
10:40:20 Join LinusN [0] (~linus@rockbox/developer/LinusN)
10:40:40gevaertsAlso, the numbers are exactly the same as in many bits of France :)
10:45:18 Join tarbo [0] (~me@unaffiliated/tarbo)
10:45:42topikyou're the regional expert :). slighly north it seems wrongly considered french-ish
10:46:32*gevaerts isn't actually. He lives in the bit where they don't speak French
10:47:37topikthe good bit
10:47:49topikmore expert than most though
10:48:20gevaertswell anyway, my point is that Walloon is not fr_BE
10:52:02 Quit tarbo (Ping timeout: 240 seconds)
10:53:15pixelmahow to deal with this commit - revert it, so that the translators are aware? German could be fixed to use "Kompressor" in dest and voice, if anyone has time now
10:54:21gevaertsI'd remove the wrong translations
10:55:17pixelmayou'd have to remove the wrong "source" in the translation files too, so that the string is found by genlang
10:59:45 Quit perfectdrug (Ping timeout: 272 seconds)
11:02:31GodEaterare these issues going to hold up the release today, or are they only in trunk, and not in the release branch?
11:04:42gevaertsthey're in both
11:05:04GodEatercue dramatic music
11:07:44 Join tarbo [0] (~me@unaffiliated/tarbo)
11:09:26topikis that the music for your release day dance?
11:09:47 Quit rphillips (Ping timeout: 256 seconds)
11:10:08GodEatergevaerts: what's your grammar kicking regex set to now?
11:15:44 Quit tarbo (Ping timeout: 252 seconds)
11:16:55 Join rphillips [0] (
11:28:31 Join tarbo [0] (~me@unaffiliated/tarbo)
11:34:21***Saving seen data "./dancer.seen"
11:37:09 Quit elcan (Read error: Connection reset by peer)
11:38:22 Quit tarbo (Ping timeout: 240 seconds)
11:43:12 Join elcan [0] (
11:47:15 Quit S_a_i_n_t (Ping timeout: 272 seconds)
11:47:40 Join S_a_i_n_t [0] (S_a_i_n_t@
11:49:39 Join tarbo [0] (~me@unaffiliated/tarbo)
11:56:49 Quit tarbo (Ping timeout: 256 seconds)
11:57:57 Join Sajber^ [0] (
12:00:36 Quit PaulJam (Ping timeout: 260 seconds)
12:07:22 Quit gibbon_ (Quit: leaving)
12:09:09 Quit TheSeven (Quit: ChatZilla 0.9.86 [Firefox 3.5.7/20091221164558])
12:09:46 Join tarbo [0] (~me@unaffiliated/tarbo)
12:16:19 Quit tarbo (Ping timeout: 246 seconds)
12:18:30 Join moos [0] (
12:18:44 Quit moos (Changing host)
12:18:44 Join moos [0] (moos@rockbox/staff/moos)
12:21:41 Join moos_ [0] (
12:24:03 Quit moos (Ping timeout: 258 seconds)
12:24:08 Nick moos_ is now known as moos (
12:24:45 Quit moos (Changing host)
12:24:45 Join moos [0] (moos@rockbox/staff/moos)
12:29:11rasherpixelma: do you have time to look at/commit FS #10950 and FS #10951?
12:30:29 Join tarbo [0] (~me@unaffiliated/tarbo)
12:31:10 Join moos_ [0] (
12:32:40 Quit moos (Ping timeout: 245 seconds)
12:32:51 Nick moos_ is now known as moos (
12:33:00 Quit moos (Changing host)
12:33:00 Join moos [0] (moos@rockbox/staff/moos)
12:35:00 Join dfkt [0] (dfkt@unaffiliated/dfkt)
12:37:19 Quit tarbo (Ping timeout: 246 seconds)
12:44:33 Join MethoS- [0] (~clemens@
12:50:39 Join tarbo [0] (~me@unaffiliated/tarbo)
12:57:53 Quit tarbo (Ping timeout: 248 seconds)
12:58:57 Quit shaggy-h ()
13:11:19 Join tarbo [0] (~me@unaffiliated/tarbo)
13:18:34 Quit tarbo (Ping timeout: 265 seconds)
13:20:58 Join wind [0] (
13:23:54 Join S_a_i_n_t_ [0] (S_a_i_n_t@
13:26:18 Quit S_a_i_n_t (Ping timeout: 265 seconds)
13:26:28 Nick S_a_i_n_t_ is now known as S_a_i_n_t (S_a_i_n_t@
13:27:02 Quit S_a_i_n_t (Client Quit)
13:31:46 Join tarbo [0] (~me@unaffiliated/tarbo)
13:34:25***Saving seen data "./dancer.seen"
13:34:32 Join PaulJam [0] (
13:38:13 Join einhirn [0] (
13:38:44 Quit tarbo (Ping timeout: 252 seconds)
13:44:12 Quit Sajber^ (Read error: Connection reset by peer)
13:46:29 Quit einhirn (Read error: Connection reset by peer)
13:51:47 Join tarbo [0] (~me@unaffiliated/tarbo)
13:58:13peturLinusN: I gather my h120 is still in the state it was at arrival?
13:58:17 Quit tarbo (Ping timeout: 240 seconds)
13:59:08LinusNexcept for the bdm connector, yes (hides in shame)
14:00:25 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
14:00:30peturI was just thinking about it, since Bagder and Zagor are coming to fosdem...
14:00:48peturI guess it is too late now to ask you (short notice)
14:05:21 Quit dfkt (Read error: Connection reset by peer)
14:05:55dionoeaoh nice, I hadn't noticed that there was a Rockbox talk at 2 PM on Sunday
14:06:27petur2PM? I thought it was later
14:06:43CIA-8New commit by mc2739 (r24450): Complete Czech language update ...
14:06:52dionoeaaccording to the pdf schedule from the fosdem website it's at 2
14:06:58peturah no, 2PM it is
14:07:24CIA-8New commit by mc2739 (r24451): Complete Czech language update ...
14:07:53 Join dfkt [0] (dfkt@unaffiliated/dfkt)
14:08:06B4gder indeed
14:08:53dionoeaB4gder: what kind of talk will that be? just an introduction to rockbox or something technical?
14:09:58B4gderabout rockbox the project, but also how we work to reverse engineer products to get rockbox on them
14:10:20B4gderto give an idea what we're doing
14:11:20 Quit killan_ (Read error: Connection reset by peer)
14:11:22B4gderbut we're meeting a few rockboxers there so there will be off-the-schedule talks, no doubt ;-)
14:12:19 Join tarbo [0] (~me@unaffiliated/tarbo)
14:13:20 Join TheSeven [0] (~theseven@rockbox/developer/TheSeven)
14:14:44 Part maruk
14:19:02pixelmarasher: not until today in the evening
14:19:25pixelmaoh, mc2739 was quicker :)
14:20:07 Quit tarbo (Ping timeout: 260 seconds)
14:20:14mc2739any idea why I got yellow with lowmem sansa ams?
14:21:56 Join darkham [0] (
14:22:02 Quit darkham (Client Quit)
14:23:42 Join teru [0] (
14:27:57 Join BlastTyrant [0] (
14:28:05 Join killan [0] (
14:32:18 Join tarbo [0] (~me@unaffiliated/tarbo)
14:32:47 Join froggymana [0] (
14:33:25pixelmamc2739: seems to a missing quote in czech.lang (line number see build log)
14:35:23mc2739pixelma: I do not see a missing quote in that area. And why only those three targets?
14:42:16 Quit tarbo (Ping timeout: 246 seconds)
14:47:52gevaertsZagor: does b0hoon have dev access on flyspray?
14:48:38Zagorhe does now
14:49:13 Quit TheSeven (Read error: Connection reset by peer)
14:49:28 Join The_Seven [0] (~theseven@rockbox/developer/TheSeven)
14:49:38 Nick The_Seven is now known as TheSeven (~theseven@rockbox/developer/TheSeven)
14:50:44 Quit TheSeven (Read error: Connection reset by peer)
14:51:03 Join TheSeven [0] (~theseven@rockbox/developer/TheSeven)
14:53:16 Join tarbo [0] (~me@unaffiliated/tarbo)
14:56:13 Quit TheSeven (Read error: Connection reset by peer)
14:56:31 Join The_Seven [0] (~theseven@rockbox/developer/TheSeven)
14:56:41 Nick The_Seven is now known as TheSeven (~theseven@rockbox/developer/TheSeven)
14:56:50 Quit moos (Remote host closed the connection)
14:59:41pixelmamc2739: LANG_EQUALIZER_GAIN has crossfade as feature in the source and swcodec in the dest and voice part (crossfade etc. was disabled on the lowmen Sansas because it takes too much RAM. Just my guess though
15:00:11pixelmaline 998 is the end of that phrase
15:00:33 Quit tarbo (Ping timeout: 265 seconds)
15:03:23pixelmathough I'm a bit confused abour advanced EQ being all swcodec then again (following phrases) and I can't remember seeing those changes when working on deutsch.lang
15:03:58 Join shaggy-h [0] (
15:04:42mc2739pixelma: thanks, I'll try changing that and see if it corrects the prblem
15:08:49pixelmathat phrase has swcodec as feature everywhere in deutsch.lang, looks like they did a bit too much in czech
15:09:43CIA-8New commit by mc2739 (r24452): fix yellow - correct feature designation
15:14:29 Join tarbo [0] (~me@unaffiliated/tarbo)
15:14:40 Join moos [0] (
15:14:52 Quit moos (Changing host)
15:14:52 Join moos [0] (moos@rockbox/staff/moos)
15:16:01 Quit moos (Client Quit)
15:18:46mc2739pixelma: ok, I guess I went the wrong way then.
15:21:21 Quit tarbo (Ping timeout: 240 seconds)
15:24:16LinusNhehe, check step 12 on this list:
15:24:51 Join moos [0] (moos@rockbox/staff/moos)
15:25:33 Join knine [0] (
15:27:44pixelmaZagor: I've heard you have the power to change the CIA bot settings
15:28:05ZagorI have the authority. I don't know if I have the knowledge :-)
15:29:01 Quit kaniini (Ping timeout: 240 seconds)
15:29:26pixelmait would be nice if the branch could be announced too. I saw it in other channels so I think it is possible somehow and it might be interesting now that we have trunk, www, mdct and release branch
15:32:39CIA-8New commit by mc2739 (r24453): More feature designation corrections
15:34:08Zagorlet's see how this looks
15:34:28 Join jgarvey [0] (
15:34:30***Saving seen data "./dancer.seen"
15:35:00 Join Zarggg_ [0] (
15:35:04 Quit Zarggg (Read error: Connection reset by peer)
15:35:31 Join tarbo [0] (~me@unaffiliated/tarbo)
15:36:34CIA-8New commit by mc2739 to (r24454): fix yellow - correct feature designations
15:37:23 Nick fxb__ is now known as fxb (
15:38:05 Quit TheSeven (Ping timeout: 252 seconds)
15:38:05 Quit froggymana (Quit: CGI:IRC (EOF))
15:39:16Zagorit looked strange, that's how...
15:39:48 Part LinusN
15:40:23Zagornew test
15:42:28 Quit tarbo (Ping timeout: 246 seconds)
15:56:31 Join liar_ [0] (~53af7f68@gateway/web/freenode/x-mocsxwdxlbohdvfn)
15:56:41 Join tarbo [0] (~me@unaffiliated/tarbo)
16:03:29 Quit liar_ (Ping timeout: 248 seconds)
16:03:43 Quit tarbo (Ping timeout: 260 seconds)
16:14:58 Quit wind (Quit: CGI:IRC (EOF))
16:17:46 Join tarbo [0] (~me@unaffiliated/tarbo)
16:22:10 Quit bmbl (Quit: Bye!)
16:24:17 Quit tarbo (Ping timeout: 252 seconds)
16:37:09 Join toffe82 [0] (~chatzilla@
16:38:10 Quit teru (Quit: Quit)
16:38:41 Join tarbo [0] (~me@unaffiliated/tarbo)
16:41:47 Quit FlynDice (Remote host closed the connection)
16:42:13rasherRemind me, what's the format for commit messages from FS?
16:43:23gevaertsrasher:;revision=24450 has an example
16:44:07CIA-8New commit by rasher to / (r24455): Portuguese translation update. ...
16:45:53 Quit tarbo (Ping timeout: 258 seconds)
16:45:57 Join FlynDice [0] (
16:45:58CIA-8New commit by rasher to / (r24456): Portuguese translation update. ...
16:46:43rasherThat should be /branches/v3_5?
16:46:59Zagorthat was intended to be, yes...
16:49:08Zagorsomeone with a bit of spare time is welcome to read up on and suggest modifications to
16:50:09dionoeawhy do you need the "New commit by" and "to" bits?
16:50:23dionoeacan CIA send messages for other stuff than commits ?
16:50:42Zagordionoea: yes, you can send pretty much any message to it
16:52:13*Unhelpful still wants to see it whine at the committer for reds :)
16:52:56Zagoryes, that's planned. and that too is something anyone is encouraged to look into.
16:53:11Zagorwe have the red detection. we just need a simple trigger client.
16:55:11CIA-8New commit by unhelpful to / (r24457): Correct ARMv5/6 signed divider, which was not saving return address to stack before using lr for calculation.
16:57:24Unhelpfulhm: /home/chshrcat/build/rockbox/apps/lang/portugues.lang:8391:1: warning: empty dest before line in non-deprecated phrase!
16:59:51 Join tarbo [0] (~me@unaffiliated/tarbo)
17:01:58rasheroh my
17:02:42 Quit soap (Quit: soap)
17:02:55Unhelpfulit certainly sounds important :)
17:04:45rasheroh.. he sure missed one
17:05:16 Nick fxb is now known as fxb__ (
17:06:35 Quit tarbo (Ping timeout: 258 seconds)
17:08:15rasherHm, should I remove them to fix the yellow or just wait for him to reply with a translation on FS?
17:14:28 Join piotrekm [0] (~pm@unaffiliated/piotrekm)
17:14:48mc2739rasher: maybe use the English version for the empty dest
17:15:12rasherWell if he doesn't turn up in time for release, yes
17:16:37 Join Sajber^ [0] (
17:18:22piotrekmwhen compiling nano2g firmware with eabi enabled i get this error , any ideas what's wrong?
17:20:52mc2739piotrekm: maybe Unhelpful or kugel can help with that, I think they are the developers with the most experience in compiling Rockbox with eabi
17:21:01 Join tarbo [0] (~me@unaffiliated/tarbo)
17:28:09 Quit tarbo (Ping timeout: 246 seconds)
17:30:06Unhelpfulpiotrekm: i'm looking at it now
17:33:31Unhelpfulpiotrekm: at a guess i'd say it has to do with the unused attribute on dma_callback
17:33:55Unhelpfulor... maybe not. hm.
17:34:32***Saving seen data "./dancer.seen"
17:40:58 Join bmbl [0] (
17:40:58 Quit bmbl (Changing host)
17:40:58 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
17:42:36 Join tarbo [0] (~me@unaffiliated/tarbo)
17:45:04Unhelpfulpiotrekm: dma_callback in pcm-s5l8700.c has __attribute__((unused)). there is no reference to it in C (only in inline asm) and as it is static it is dropped from the object file by gcc, apparently.... although the assembler notes the reference and marks it as an external symbol. this happens even if i simply remove the unused attribute
17:45:28piotrekmseems logical
17:45:48piotrekmbut everyting goes alright in the standard arm compilation
17:45:51Unhelpfulchanging the attribute to "used" causes the function to be retained, and then compilation completes
17:45:59piotrekmoh, thanks a lot
17:46:44Unhelpfulthat's also a different compiler version, perhaps the newer one is more aggressive about unreachable static functions
17:48:52piotrekmUnhelpful: hm... now it complains about ape -
17:49:05Unhelpfuli don't really want to change it myself without knowing why TheSeven wanted it unused
17:49:43 Quit tarbo (Ping timeout: 265 seconds)
17:52:53Unhelpfulpiotrekm: the ape decoder uses a customized divider that uses certain hacks to speed things up a bit (for ape). one of these is a big table of precalculated reciprocals. the table is sized to use all iram not used by other code or buffers, but since some of what is in iram is C code, the size of that code may have changed
17:53:43 Join Sajber^1 [0] (
17:54:19piotrekmi see
17:54:46Unhelpfultry changing recip_max at apps/codecs/lib/udiv32_arm.S:102 from 13184 to 12672
17:55:15 Quit Sajber^ (Ping timeout: 260 seconds)
17:55:58piotrekmUnhelpful: thanks again, worked;D
17:56:15Unhelpfuli suppose i need to either come up with a solution to calculate the correct size at runtime, or i need to #ifdef a completely separate set of constant sizes for eabi
17:57:07Unhelpfulerm, not runtime, compile/link-time
17:59:17 Join fyrestorm [0] (
18:00:43 Quit krazykit (Ping timeout: 246 seconds)
18:00:51piotrekmUnhelpful: on last question i hope: do i need the bootloader also compiled as eabi?
18:04:02 Join Tomis2 [0] (~Tomis@
18:04:44 Quit Tomis (Ping timeout: 245 seconds)
18:04:45 Nick Tomis2 is now known as Tomis (~Tomis@
18:04:47 Join tarbo [0] (~me@unaffiliated/tarbo)
18:06:31Unhelpfulpiotrekm: shouldn't matter. i'd be afraid to use an eabi bootloader on a target where i can't be sure of recovery if it fails ;)
18:07:31 Join krazykit [0] (~kkit@
18:08:01Unhelpfulthe vast majority of things have worked unchanged with eabi... but one or two things have needed correction (or still do, mpegplayer doesn't work on eabi on my beast)
18:11:26piotrekmhm... hung up at the splash screen
18:11:28 Quit tarbo (Ping timeout: 240 seconds)
18:13:09 Join _zic [0] (
18:13:38 Join liar [0] (
18:17:55piotrekmis there any way to turn on some on-screen debugging to be somehow able to know at which point the firmware hung up?
18:19:54 Join Casainho [0] (~chatzilla@
18:25:31 Join tarbo [0] (~me@unaffiliated/tarbo)
18:27:06 Join GeekShado_ [0] (
18:27:16 Quit petur (Quit: *plop*)
18:29:19 Quit GeekShadow (Ping timeout: 245 seconds)
18:32:26 Quit tarbo (Ping timeout: 252 seconds)
18:32:53 Join Kitar|st [0] (
18:33:45 Join domonoky [0] (
18:33:46 Quit domonoky (Changing host)
18:33:46 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
18:36:16 Quit Kitr88 (Ping timeout: 252 seconds)
18:37:34 Quit Kitar|st (Ping timeout: 265 seconds)
18:37:40 Join Kitar|st [0] (
18:40:22 Join bertrik [0] (
18:42:09piotrekmhow can I disable building of a certain app?
18:42:27Unhelpfulpiotrekm: that's why i wouldn't try building the bootloader with eabi yet. ;) when kugel and i started this he narrowed down the crashes on e200 by inserting code to turn the wheel light or the backlight on and off at various points during startup, to see where things fell apart. you'll probably need to try a similar method, although if the splash screen works you may be able to write lines to the screen
18:42:45Unhelpfulpiotrekm: a certain plugin? edit apps/plugins/SOURCES or apps/plugins/SUBDIRS
18:46:27Unhelpfulbut a plugin isn't why it's hanging, unless you've contrived to run one on startup
18:48:05 Join kugel [0] (kugel@rockbox/developer/kugel)
18:48:08 Join pamaury [0] (
18:48:45kugelUnhelpful: __attribute__((used)) should be fine. I remember that TheSeven was just as curious about how to get it compile
18:49:55kugelis release party today?
18:50:23piotrekmhmm... i built the firmware in debug and logf mode (by wild guess), and the screen no says *PANIC* \n mutex_unlock->wrong thread(<some cut-off text> 'n <three unprintable chars> != main)
18:50:31piotrekm*now says
18:50:57bertrikthat might be a bug
18:51:40piotrekmi only wonder whether this is the description of the problem which occured woth the no-debug eabi firmware, or is it something new
18:53:07kugelas we do :)
18:55:23 Join stoffel [0] (
18:57:28 Quit shaggy-h (Ping timeout: 240 seconds)
19:05:31 Join tarbo [0] (~me@unaffiliated/tarbo)
19:06:12 Quit tarbo (Client Quit)
19:07:08 Join shaggy-h [0] (
19:10:27 Quit Zambezi (Read error: Connection timed out)
19:10:43 Quit grndslm (Read error: Operation timed out)
19:11:24 Quit z35 (Ping timeout: 246 seconds)
19:11:35 Join grndslm [0] (
19:14:52 Join rhodan [0] (
19:19:10 Quit Casainho (Ping timeout: 252 seconds)
19:20:43 Join evilnick_B [0] (~0c140464@rockbox/staff/evilnick)
19:22:46 Join Horscht [0] (~Horscht2@xbmc/user/horscht)
19:23:13FlynDiceTired of beating my head & will now put ignorance on display for all..... Does this line in the dualboot makefile "define" SANSA_CLIPPLUS for the #ifdef's in dualboot.s? −−> arm-elf-gcc -DSANSA_CLIPPLUS -c -o dualboot_clipplus.o dualboot.S
19:25:38pixelmarasher: what to do with the lang string that wouldn't show for some translators? I'd suggest reverting the part in the wrongly touched languages but then we have a few languages not 100% up-to-date for the release. Of course I could fix for German while at it, if you are around you could fix Danish, I sure hope some Swedish giy could help today too, French could be possible and maybe Italian too (as the maintainer seems to be quite quick)
19:26:57kugelFlynDice: yes. you could double check with #ifdef SANSA_CLIPPLUS\n#error\n#endif
19:29:16moospixelma: I tried to set a dev env on this laptop, and then update the french translation. IIRC it doesn't need much...
19:29:33moos*I will try
19:30:05FlynDicekugel: Thanks, I just couldn't figure out where it was being defined in the code and thought perhaps this was it. A makefile is like a big black box that just "works" to me right now...
19:31:00 Quit dfkt (Read error: Connection reset by peer)
19:31:33pixelmamoos: there is an error in some languages where someone put the English strings into the phrase so that it wouldn't even show up as "source differs" in genlang
19:31:51 Quit kugel (Ping timeout: 260 seconds)
19:31:59 Join meandyou [0] (~chatzilla@
19:32:10meandyouHi there!
19:32:29moosI will track the svn history, but frenc is already almost up to date
19:33:03meandyouI am using Rockbox for years now... and i like the new version with the colors and all that!!! BUT it seems to crash a lot more than the previous ones...
19:33:12meandyoui am using an ipod 60G
19:33:24meandyouis it me or is it known as a bug?
19:33:36pixelmamoos: yes, but if we revert that part of the commit it won't be (again). LANG_COMPRESSOR is what I'm talking about btw.
19:33:55domonokymeandyou: it shouldnt crash. what does it do when it crashes ?
19:34:12meandyoubecause i did a reinstall after a format but the problem is still the same: WHen i listen to an Mp3 and i clikc on the central button, i freezes
19:34:20meandyoui have to do a full reboot,
19:34:23meandyoui mean by that
19:34:33***Saving seen data "./dancer.seen"
19:34:43moospixelma: Hah, you spoke about this string... I would be around this night and then update at last minute if needed
19:34:45meandyouCentral button + top button for 3 seconds....
19:34:59 Join Strife89 [0] (~michael@
19:35:01pixelmamoos: great
19:35:25moosno problem
19:36:24 Join dfkt [0] (dfkt@unaffiliated/dfkt)
19:36:25domonokymeandyou: you could try the ipod video 30gb build. there are reports of 60gb ipod videos which are in reality 30gb ipods with bigger hd. And using the 60gb build on a 30gb ipod video will produce crashes (the other way round is safe)
19:36:28moosmeandyou: which revison/version do you use?
19:36:43meandyoui think it is the latest one...
19:36:46 Join rhodan_ [0] (
19:36:57meandyoubut with the previsou one it was doing the same...
19:36:59 Join paulk [0] (
19:37:09meandyoui mean since i am using v3 of rockbox
19:37:32 Join soap [0] (
19:37:55 Quit rhodan (Ping timeout: 260 seconds)
19:37:57pixelmaleaves Polish, Tagalog, Walloon and maybe Italian... reverting is a bit weird too though because then they'll get the old translation (since the original string changed and wasn't deprecated and added as new (then they would get the right thing, even if it is English) :\
19:38:30moospixelma: ouch :(
19:38:51paulkhello! I'd like to know how I can get "button events" while I scroll without stoping ("rb->button_get" just give one "scroll event" until I stop scrolling and restart)
19:39:10JdGordonI've built release zips"> if someone wants to copy them to the servers? ( B4gder, Zagor)
19:39:21 Join hebz0rl [0] (
19:40:24pixelmamaybe fix for German with a big remark in the commit message?
19:40:48 Quit soap (Changing host)
19:40:48 Join soap [0] (~4cb54aea@rockbox/staff/soap)
19:41:36 Join TheSeven [0] (~theseven@rockbox/developer/TheSeven)
19:41:49soapsince when is access to #rockbox-community blocked in the web-client?
19:42:17pixelmaor do what should have been done before (deprecating, but I don't know how to name the new string then)
19:45:49 Join guest [0] (
19:46:04JdGordonsoap: you cant get in?
19:46:20guesthow much time is left before 3.5 is released?
19:46:27 Join webguest25 [0] (
19:47:32 Quit gevaerts (Killed (NickServ (GHOST command used by gevaerts_!
19:47:41 Join gevaerts [0] (~fg@rockbox/developer/gevaerts)
19:47:45 Quit webguest25 (Client Quit)
19:48:35soapcan't get in, jd
19:48:49JdGordonI just did :) did you spell it correctly?
19:48:55pixelma14 minutes?
19:49:06JdGordon12 now
19:49:06paulkWhat's new in rb 2.5 ?
19:49:19JdGordonin 3.5 however
19:49:19guestthanks! me waiting for it n am sure lot more people as well
19:49:28paulkoh yes, 3.5 :)
19:49:38pixelmahuh? I didn't know there was an exact time (hour) set
19:49:40 Quit moos (Quit: reboot)
19:50:17JdGordonguest: you can download it from">
19:50:47guestthanks gordon! appreciate it
19:51:18B4gderwe still need voices, manuals, tarballs...
19:51:26B4gdertarball rather
19:52:11pixelmawill there be anything done to make Archos voices fit?
19:52:33B4gdernot by me
19:53:16B4gderJdGordon: can you make me a single huge zip with them all in?
19:53:16 Quit GeekShado_ (Ping timeout: 252 seconds)
19:53:47gevaertsWere all remaining translation updates committed?
19:53:54 Join efyx_ [0] (
19:54:06pixelmadoesn't festival allow speeding up voices or lowering quality?
19:54:53 Quit piotrekm (Quit: Leaving.)
19:56:22 Nick rhodan_ is now known as rhodan (
19:57:06pixelmagevaerts: there's only one very new (turkish) that could be worth it (and maybe the "compressor" fix if I get some opinions)
19:58:18pixelmathat I found in a quick tracker search for translation and language
19:58:28gevaertsoh, FS #10956 needs to go in I suspect
19:59:00pixelmaalready has
19:59:05 Join moos [0] (moos@rockbox/staff/moos)
19:59:24pixelmaoh, hmm, maybe not
19:59:45gevaertsit's ten minutes old
19:59:45guesthey, can someone pls look at
20:00:42moosis the release really planed 20h precise?
20:01:06pixelmagevaerts: but it reverts mc2739's later fixes (see LANG_EQUALIZER_GAIN)
20:01:24*gevaerts has a look at it
20:02:41guest@gevaerts you responded to me?
20:03:19mc2739gevaerts: FS #10956 does need to go in. What I committed earlier did not convert properly according to the author.
20:03:54gevaertsmc2739: the question is how to fix it with the least amount of work
20:04:36 Quit PaulJam (Ping timeout: 246 seconds)
20:05:24 Quit soap (Quit: CGI:IRC (Ping timeout))
20:06:55 Join Adnyxo [0] (
20:08:28mc2739gevaerts: I can work on the new patch
20:08:37gevaertsmc2739: I think I can handle it
20:11:21pixelmahmm... if I could fix all those languages at once. * looking around for B4gder, Zagor, rasher, b0hoon, some Italian guy, someone who speaks Walloon and some more *
20:12:17Zagorpixelma: ?
20:12:21mc2739gevaerts: Also, I think rasher was going to do another update on the Portuguese translation
20:12:53pixelmaZagor: I'd need a translation of "Compressor" for svenska.lang
20:13:32 Join phanboy4 [0] (
20:13:37gevaertsmc2739: I don't mind more updates, but I care most about this Czech update now. That one is just bad
20:14:09CIA-8New commit by gevaerts (r24458): (hopefully) fix encoding (FS #10956)
20:14:31Zagorpixelma: the swedish word seems to be "Kompressor"
20:14:54pixelmanice and easy as it is the same in German :)
20:15:54CIA-8New commit by gevaerts (r24459): (hopefully) fix encoding (FS #10956)
20:16:33pixelmaZagor: and not necessary as it has been translated already. Duh, sorry
20:17:22pixelmawell, it was not given as genlang wouldn't have found it
20:17:37pixelmaI should have checked before though
20:17:58Zagordon't worry about it. it took me 20 seconds to look it up.
20:19:54pixelmaok. It made me look up the other languages too and it looks like French and Italian was also fixed. Leaves Danish, Polish, Tagalog and Walloon
20:20:40 Join Zambezi [0] (~Zulu@
20:21:49 Join mitk [0] (
20:22:10gevaertsI think we should commit the Turkish update. We can always add more bits later
20:23:06AlexPpixelma: I say go ahead and get it right
20:23:25mitkpixelma: I don't know currently wassup but how can I help you with polish?
20:23:39gevaertsAlexP: you proposed waiting. What do you think?
20:23:46 Quit bmbl (Quit: Bye!)
20:23:58AlexPgevaerts: For what?
20:24:09gevaertsAlexP: Turkish. Wait longer, or commit now?
20:24:33AlexPgevaerts: Oh right. I say commit - then as much as possible is done for the release
20:24:51AlexPIf he happens to show up when we are still arsing around, then we can add to it
20:25:21AlexPmitk: Good timing :)
20:25:52mitkAlexP: Thanks but still don't know wassup
20:26:03pixelmamitk: there was a problem with the translation and I'd like to fix it today before the release. I'd need a translation for "Compressor" (sound compressor)
20:26:30pixelmaor a Polish committer
20:26:44mitkpixelma: Simply "Kompresor" :)
20:27:06gevaertsone s?
20:27:20CIA-8New commit by gevaerts (r24460): Update Turkish translation ...
20:27:27mitkgevaerts: Yes, for sure
20:27:33pixelmawithout any special characters?
20:27:59mitkpixelma: It's strange also for me, but yes
20:28:14CIA-8New commit by gevaerts (r24461): Update Turkish translation ...
20:28:39 Join kugel [0] (~kugel@rockbox/developer/kugel)
20:28:45guestI am getting "arm-elf-gcc" n "arm-elf-ld" not found while building. where as i hv both in path
20:28:50guestcan someone help?
20:29:15AlexPguest: We are preparing to release, so you might have to hang on a minute or two :)
20:29:18pixelmamitk: nice, thanks. I think I saw your real name in the tracker so I could mention it (and add it to the language file)
20:29:41guestokay, sure
20:29:53gevaertsrasher: how often is the translation page updated?
20:30:00rashergevaerts: once an hour
20:30:08rasherI can do a manual update if you like
20:30:56mitkpixelma: I'm in language file already.
20:31:05pixelmarasher: shall I fix "Compressor" in Danish too while at it or would you have time later in the evening and want to do it yourself?
20:31:17pixelmamitk: alright :)
20:32:00gevaertsrasher: don't bother. I can wait that long :)
20:32:25 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
20:33:24GodEaterguest: so typing "which arm-elf-gcc" returns the path where they are installed?
20:33:53 Quit amiconn (Killed (NickServ (GHOST command used by amiconn_)))
20:33:53 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
20:34:15 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
20:34:40rasherpixelma: I'll look at it now.
20:35:04guest@GodEater: no.. the package i downloaded for windows had gcc n ld separately.. not prefixed with "arm-elf"
20:35:25gevaertsguest: please use real words. "n" doesn't mean much
20:35:28 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
20:35:28guest@GodEater do i need some specific gcc
20:35:29 Quit pixelma (Killed (NickServ (GHOST command used by pixelma_)))
20:35:32moospixelma: I canot get yet a dev env on this computer (connection problems). According to, there are just those 2 base skin strings missings
20:35:41moosfor french
20:35:46 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
20:35:59GodEaterguest: you downloaded a package for windows? Which guide were you following for getting a build environment set up?
20:36:09AlexPmoos: I think "Compressor" is missing too
20:36:15pixelmamoos: yeah, French looks like it was fixed
20:36:22AlexPah right
20:36:25AlexPmy mistake
20:36:40moosALexP: I did fix that long time ago
20:36:44AlexPmoos: cool :)
20:37:09AlexPmoos: I guess easiest is to just give the two missing translations here and someone can commit them
20:37:20AlexPMaybe pixelma along with her other changes
20:37:22pixelmait wouldn't show up on the missing strings part though if it wasn't because it was done incorrectly
20:37:29guest@godeater: i referred : i use windows and this guide does not have information about gcc package. so i googled for arm-elf-gss windows binaries and downloaded one such package
20:37:44CIA-8New commit by rasher (r24462): Translate LANG_COMPRESSOR
20:37:47moosAlexP: I didn't think something good yet :)
20:38:01GodEaterguest: we only support one true windows environment, and that's cygwin, I take it that's not what you're using?
20:38:08AlexPmoos: There is still time :)
20:38:19CIA-8New commit by rasher (r24463): Fix LANG_COMPRESSOR
20:38:31GodEaterguest: The rockbox build environment is *very* picky about specific versions of gcc for each architecture, so downloading some random one from the web is likely not going to work.
20:38:37moosbut I guess, instead to doing something in hurry, best to leave in english for the realease (unless some french people could find inspiration)
20:38:46Unhelpfulpiotrekm: you really need to find out how far it gets before hanging. you know it hangs somewhere before the splash is cleared, so that's a start...
20:39:04guest@GodEater : I am using cygwin, but that did not have gss and ld. that was shown by errors when i ran "/tools/configure". it said arm-elf-gcc not found
20:39:05AlexPmoos: Whatever would look/feel best to you
20:39:09pixelmaalright. The only thing more I could do now if I'm lucky is Tagalog
20:39:31gevaertsguest: did you follow ?
20:39:50guest@gevaerts : no. i will look at this now.
20:39:53GodEaterguest: do as gevaerts suggests and read that, we provide specific cygwin packages for gcc and ld
20:40:06gevaertsit's linked to from SimpleGuideToCompiling :)
20:40:10moosAlexP: I don't like to reuse english for a translation (I made lot of effort to avoid it in french)
20:40:29GodEaterwe should retire that SimpleGuideToCompiling - it's anything but
20:40:47guestgevaerts and GodEater : I guess i overlooked it. I thought cygwin configuration was simple :P
20:40:57bluebrotherGodEater: call it "NotSoSimpleGuideToCompiling"? ;√-)
20:41:05GodEaterguest: it is really :)
20:41:10GodEaterbluebrother: hehe :)
20:41:30moosAlexP: I like the "habillage" one
20:41:44mooswe could reuse it
20:41:45AlexPIt sounds not too bad to me as a non-native speaker
20:41:47 Quit flydutch (Quit: /* empty */)
20:42:08moosAlexP: want ot do the changes? if so I can help you :)
20:42:20moosI canot have dev env here
20:42:41AlexPmoos: pixelma has some others to go in, so I guess it makes sense to tell her - pixelma, is that OK with you?
20:43:00 Quit Tomis (Ping timeout: 264 seconds)
20:43:15pixelmaI could do that yes
20:43:27moosnice, that's cool
20:43:28 Quit dmb_ (Ping timeout: 248 seconds)
20:43:42pixelmaI'm not touching French with the other update anyways
20:44:01moosyup that what I thought, still best to separate things
20:44:49*moos will use the rasher's interface for the 1st time :)
20:45:17guest@GodEater : :) downloading newer packages from rockbox website. will come here again, after a while
20:46:06bluebrotherdomonoky: how's the status for a new rbutil release?
20:46:09 Join JdGordon_ [0] (~836b0065@gateway/web/freenode/x-spghemtfwndxgkkw)
20:46:53 Join Bullet` [0] (
20:46:55domonokybluebrother: dont know. didnt noticed any bugs when i last tested it :-)
20:47:41 Quit paulk (Quit: Ciao)
20:47:55amiconnUnhelpful: For the improved filters I need to "borrow" 380 entries from the recip table on pp5002
20:48:05*kugel proposes to backport lang changes during the branch phase at the end in a big svn diff apps/lang/*.lang
20:48:17amiconnOf course I'll runn speed tests before commit
20:48:27domonokybluebrother: but i think we should release a new rbutil as fast as we can. do you know any open issues ?
20:49:03bluebrotherbtw, I've noticed that mkamsboot doesn't check the md5sum of the generated firmware file. Is this intentional?
20:49:38domonokybluebrother: that sounds like a bug.
20:49:51bluebrotherwell, unless I missed that of course ;-)
20:50:10rasherkugel: I'm not entirely convinced that's a good idea
20:50:12moospixelma: here we go
20:51:00bertrikWe check the md5 of the original firmware to make sure that it is compatible with the dual bootloader, but I don't see why we should check the generated firmware file
20:51:15domonokybluebrother: ah, i miss remembered. it was left out on purpose. so new bootloaders dont neccessary need new mkamsboot.
20:51:53domonokyand it isnt so dangerious then hxx0 players, so we left the last check out.
20:52:23bluebrotherit is not? I thought those players can really be bricked by putting a wrong file on it
20:52:26 Join casainho [0] (
20:52:52domonokybluebrother: yes, but the dangerious code is not the bootloader, but the dualboot code.
20:53:33domonokyas long as the dualboot code is fine, and patches at the correct place (thats why we check the input OF), it should be recoverable.
20:53:36 Quit BlastTyrant (Ping timeout: 246 seconds)
20:53:47bluebrotherah, ok
20:54:27*bluebrother goes updating the german translation of rbutil and things that "OF" shouldn't get used in rbutil at all as abbreviation
20:54:31 Quit mitk (Quit: CGI:IRC)
20:56:20AlexPbluebrother: No, it should be spelt out I think
20:56:45bluebrotherI won't change that before the release though. It's not a big issue anyway.
20:56:57AlexPno, just something for next time
20:57:17bluebrotherstrings seem to be somewhat inconsistent to me anyway, so a review of _all_ strings might be a nice idea
20:58:58amiconnUnhelpful: Since udiv32_arm.S is now tuned for libdemac, shouldn't it be moved (back) there?
21:01:16bluebrotheranyone want to update another rbutil translation?
21:02:14 Quit Zambezi (Read error: Connection timed out)
21:02:51gevaertspixelma: both Walloon and Tagalog miss 37 other strings, so I wouldn't feel too bad about one more
21:03:23CIA-8New commit by bluebrother (r24464): Update german rbutil translation.
21:03:58pixelmagevaerts: yes, but this string won't be found by genlang so I'd like to fix it before it will be forgotten
21:04:28gevaertspixelma: I know. I think removing it is the best fallback option
21:04:34AlexPpixelma: Can you just make it so that they appear as missing?
21:05:35rasherYeah, simply removing the phrase works
21:06:22rasherFor portuguese, cowond2 and mrobe500 targets should be removed if he doesn't show up with a fix
21:06:40rashereh, from LANG_CONFIRM_WITH_BUTTON
21:06:45pixelmaI'll try and fix if possible in the next hours. If not I'll remove it
21:07:23bluebrotherdomonoky: I'm not aware of any serious issues. My plan is to disable the D2 as discussed yesterday, then tag and release −− unless someone wants to fix some translations the next couple of hours
21:08:26domonokybluebrother: sounds good.
21:09:01 Join piotrekm [0] (~pm@unaffiliated/piotrekm)
21:09:12 Quit stoffel (Remote host closed the connection)
21:09:44rasherpixelma: Simply removing those targets from both source and dest in that phrase should make it come up as untranslated, yet still retain the translations we do have
21:11:05pixelmado I understand correctly that I should keep the id?
21:11:33Unhelpfulamiconn: was it ever tuned for / used by anything else? i'd favor moving it there, it would also make it easier to do a rebuild-with-exact-size handling of the reciprocal table
21:11:33bluebrotherdomonoky: what do you think about a "show disabled players" option in the configuration? Might be useful for svn binaries.
21:12:27 Join Zambezi [0] (Zulu@
21:12:39domonokybluebrother: might be usefull for svn binarys, but do we want such a option in a release build ?
21:13:04bluebrotherdomonoky: no, but that could easily be handled by some -config release switch
21:13:14Unhelpfulthere's no real "tuning for ape" of the armv5/6 divider, other than having another copy of the divider in iram if iram is used
21:13:20 Join Blue_Dude [0] (
21:13:35CIA-8New commit by pixelma (r24465): Update French translation by Mustapha Senhaji
21:13:36Blue_DudeHi guys...
21:14:23Blue_DudeFound a teeny bug/feature: The WPS background stays loaded on the bookmark page when stopping playback.
21:14:47pixelmaheh... the build table looks quite yellowish
21:16:28JdGordon_Blue_Dude: how recent is your build?
21:16:40kugelBlue_Dude: welcome back anyway :)
21:16:46Blue_DudeA few days ago. I'm building the current one right now.
21:16:48pixelmaone warning is from Unhelpful's commit and the others from the czech and turkish updates
21:16:56Blue_DudeThanks! Been in hibernation.
21:17:01gevaertsyes, I'm obviously innocent!
21:17:28 Quit Zambezi (Changing host)
21:17:28 Join Zambezi [0] (Zulu@unaffiliated/zambezi)
21:17:45JdGordon_Blue_Dude: that should have been fixed in the last day or two
21:17:48JdGordon_or made worse :)
21:18:07kugelBlue_Dude: is that in 3.5 also?
21:18:12Blue_DudeOK, no harm done.
21:18:33Blue_DudeI haven't loaded the 3.5 build yet. Only the current svn branch.
21:18:34gevaertsok, I should obviously have left the Turkish translation to someone who knows about lang files
21:19:02CIA-8New commit by gevaerts (r24466): Fix missing quote (again?)
21:19:26CIA-8New commit by bluebrother (r24467): Disable D2 in rbutil as discussed in IRC. ...
21:19:30CIA-8New commit by bluebrother (r24468): Add italian to the list of language names.
21:19:33CIA-8New commit by gevaerts (r24469): Fix missing quote (again?)
21:19:40Unhelpfulpixelma: it is? i don't see that any new warnings came in w/ r24457...
21:19:56guesthi, i was able to run ../tools/configure successfully, but 'make' command fails with this output : "make: makefile: line 3: Error −− Expecting macro or rule defn, found neither" can someone guide
21:20:12gevaertsguest: which version of make?
21:20:18pixelmaUnhelpful: oh, sorry - off by one error
21:20:47gevaertspixelma, rasher: can you look at Turkish? I suspect that the file doesn't handle missing translations correctly, but I don't really know enough about them to fix
21:21:36guest@gevaerts : which make outputs : /cygdrive/c/PROGRA~1/MKSTOO~1/mksnt/make
21:22:21gevaertsguest: I mean, what does make −−version say?
21:22:46pixelmagevaerts: maybe in a bit (quite busy currently and still have to eat dinner...)
21:23:35CIA-8New commit by pixelma (r24470): Update French translation by Mustapha Senhaji
21:25:05guest@gevaerts : it throws 30-40 lines of text. you need that here?
21:25:19moosmerci pixelma
21:25:29 Quit toffe82 (Read error: Connection reset by peer)
21:26:09gevaertsguest: is it 3.80? If so, you need to upgrade. Rockbox needs 3.81 to build
21:26:49kugelBlue_Dude: any progress in pcmbuf rework?
21:27:08Blue_DudeI was in hibernation for a reason. :)
21:27:19Blue_DudeBut not that one.
21:27:31guest@gevaerts : ok, where can i get 3.81 version? would google results help?
21:27:38Blue_DudeIn a word, no. I am stuck.
21:29:17gevaertsguest: no idea, I don't use cygwin
21:29:27 Quit moos (Quit: ChatZilla 0.9.86 [Firefox 3.6/20100115144158])
21:29:49guest@gevaerts: okay, i am trying to pull it using cygwin installer
21:30:19 Quit Kitar|st ()
21:34:35***Saving seen data "./dancer.seen"
21:36:42JdGordon_are we done with lang updates in the 3.5 branch?
21:37:46 Nick knine is now known as kaniini (
21:38:42 Quit Farthen (Killed (NickServ (GHOST command used by Farthen_!
21:39:08 Join Farthen_ [0] (
21:39:09 Nick Farthen_ is now known as Farthen (
21:39:14 Join Kitar|st [0] (
21:39:16 Quit krazykit (Quit: rebooting)
21:39:46Blue_DudeJdGordon_: Bug still there with r24470...
21:39:50AlexPJdGordon_: I just tried the latest build, and the backdrop is completely corrupted - mainly black screen but with crazy colours on bits
21:40:00AlexPJdGordon_: 3.5 branch is fine
21:40:17JdGordon_eek crap?
21:40:26JdGordon_Blue_Dude: put it on flyspray please
21:41:00AlexPJdGordon_: I'll try with some other backdrops, but the one I tried is fine with the release branch
21:41:25JdGordon_yeah, no surprise there. backdrops handling changed dramatically after the branch
21:41:43 Join krazykit [0] (~kkit@
21:42:53AlexPHuh, I just tried to set a backdrop through the context menu on a bmp and my beast just shut down
21:43:36AlexPSame again, this time with the cabbiev2 backdrop
21:44:03 Join toffe82 [0] (~chatzilla@
21:44:13JdGordon_24365 works so its sometime after that :/ although thats a M build so it could be later even
21:44:15AlexPHmmm, the cabbiev2 backdrop is fine if I load it via the theme (i.e. load the entire theme)
21:45:24AlexPFor some reason it just seems to be one particular backdrop, but why does it work with v3.5 and not current? Especially when the built in backdrops work with current
21:46:02JdGordon_the 3.5 branch and trunk dont share any background logic
21:46:22Blue_DudeJdGordon_: FS #10957.
21:46:39JdGordon_can you try r24434?
21:46:42AlexPSo maybe there is something about this .bmp that 3.5 didn't mind but trunk doesn't? Seems a bit odd though
21:46:48AlexPJdGordon_: yep, can do
21:46:58pixelmaJdGordon_: no, we are not done and warnings have to be fixed too
21:47:14Blue_DudeGot to go. I'll check back later and troubleshoot if I can.
21:47:20JdGordon_ok, well and refuse to work so arg
21:47:25JdGordon_someone else might have to build them
21:47:38 Quit Blue_Dude (Quit: ChatZilla 0.9.86 [Firefox 3.5.7/20091221164558])
21:47:41JdGordon_24435 looks like the only change which might have caused problems
21:48:27 Join dmb_ [0] (~Dmb@unaffiliated/dmb)
21:49:37CIA-8New commit by pamaury (r24471): FS #8967: Fix autoscore computation overflow when the playtime is huge.
21:51:04kugelwe have no ReleaseNotes35 guys :)
21:51:29JdGordon_we havnt released yet so its fine
21:52:07*kugel set up the last two release notes, maybe another person volunteers this time
21:53:03 Join Tomis [0] (~Tomis@
21:53:31JdGordon_but you're practiced!
21:54:38 Join BlastTyrant [0] (
21:56:08pamauryTorne: you closed "FS #8982 - Database update progress stuck to -1% " saying that debug menu is for developpers only but you didn't close "FS #6761 - Database progress reads above 100%". Didn't you see it or is there a good reason for that ?
21:56:27Tornei haven't seen it
21:56:36Tornesomeone drew my attention to the other
21:57:04 Quit dmb_ (Ping timeout: 240 seconds)
21:57:28pamauryDo you think we can safely close this one also ? (there is no simple fix afaik and it's not necessary to have an accurate reading in the debug menu)
21:58:26Torneyeah, i don't thinkit's interesting
21:58:46Tornewe keep debating taking the debug menu away entirely from users, so that would fix it ;)
21:58:47 Quit Strife89 (Quit: Going home.])
22:01:05JdGordon_only out of release builds
22:04:03kugeland we probably would've done it already if JdGordon_ bothered to finish his patch!
22:04:50JdGordon_yeah well, nuking my tree is a good way for me to lose interest in a patch :p
22:11:08pamauryThe patch is easy (to remove debug_menu), no ?
22:11:59pamaurySlasheri: are you there ?
22:13:48JdGordon_yes, but the patch I was working on was to #define a RELEASE which would then #ifdef out the debug menu and do a few other things, and build all the needed files also
22:14:40 Part domonoky
22:15:18kugelit's on fs, isn'ti t?
22:15:55Slasheripamaury: nice fix :) i was just reading that
22:16:55pamaurySlasheri: I have a discovered a "bug" in the dircache/tagcache interaction and I would like your opinion. Do you have some time ?
22:17:20Slasheriwhat kind of a bug?
22:17:37pamauryIt leads to dircache disabling itself
22:17:55Slasheriok, interesting
22:18:02 Quit _zic (Quit: Ex-Chat)
22:18:25Slasheriplease tell more :)
22:18:42 Join dmb_ [0] (~Dmb@unaffiliated/dmb)
22:19:39 Join Tomis2 [0] (~Tomis@
22:20:12CIA-8New commit by bluebrother (r24472): Fix stupid typo in my last commit. How did I miss that?
22:20:47 Join froggyman [0] (
22:21:03pamauryHere is my theory, I can reproduce it deterministically and gevaerts also reproduce it. Say you have dircache enabled and you start to play a playlist. The core opens the .playlist_control file and dircache binds to say, fd=1. Now, as you are evil, you go into settings menu and ask tagcache to rebuild the cache. But of course, tagcache is even more evil than you and steals dircache buffer ! Later one, tagcache finishes its buil
22:21:03pamauryd and then you go back to play another playlist. The core now write it to .playlist_control. The core updates the filesize accordingly and notify it to dircache. But dircache buffer was stealed so fd bindings are not available anymore. So dircache disables itself
22:22:47kugelwhat dircache buffer?
22:22:58pamaurydircache entries buffer
22:23:09pamaurykugel: dircache is cache afterall
22:23:17Slasherihmm, indeed. But shouldn't dircache be already disabled when the buffer was stolen?
22:23:32 Quit Tomis (Ping timeout: 256 seconds)
22:23:33 Nick Tomis2 is now known as Tomis (~Tomis@
22:23:36kugelsure but it cannot just release its cache to get it overwritten, can it?
22:24:16pamaurySlasheri: it is. But for some reason, either it is renabled after or the dircache update function don't check if dircache is enabled or not
22:24:44pamaurykugel: yes it can, thanks to dircache_steal_buffer ;)
22:25:00Slasheripamaury: yes, thet cache will be rebuilt after the tagcache commit is finished
22:25:08 Quit guest (Quit: CGI:IRC (Ping timeout))
22:25:10pamaurySlasheri: yes it will be
22:25:20 Quit GodEater (Ping timeout: 248 seconds)
22:25:38Slasheripamaury: do you mean that the problem was that the cache gets re-enabled when it shouldn't?
22:26:00kugelpamaury: doesn't it give only the reserve buffer (so that dircache can add entries without rebooting)?
22:26:15pamauryNo, the problem is that when it gets reenabled, the updates (filetime, filesize) are notified to dircache but the fd bindings do not reflect the actual bindings
22:26:22Slasherikugel: tagcache requests the whole cache buffer
22:26:27 Join casainho_ [0] (
22:26:48kugelwhat if I run tagcache without dircache?
22:27:00 Quit casainho (Read error: Operation timed out)
22:27:01pamauryit won't use any buffer for that task
22:27:10 Nick casainho_ is now known as casainho (
22:27:20pamauryBut if dircache is available, it will use the cache to commit
22:27:28Slasheripamaury: ah, indeed!
22:28:15Slasheripamaury: iirc, i hit into a similar problem a long time ago and added fdbindings buffer to hold some entries until the cache rebuild is complete so that the cache can be updated
22:29:07 Join GodEater [0] (
22:29:08 Quit GodEater (Changing host)
22:29:08 Join GodEater [0] (~bibble@rockbox/staff/GodEater)
22:29:39Slasheripamaury: if possible, maybe the fd_bindings part of code could be fixed to handle that special case also
22:30:22pamauryPerhaps, perhaps not. fd_bindings stores entry points. When dircache is rebuilding, some pointers don't make sense
22:30:38pixelmaI can't seem to find out how to fix the warnings in the Portuguese and Turkish translations. rasher: do you have an idea?
22:31:31 Join stripwax [0] (
22:31:45pamaury(completely unrelated) did someone with a c250 looked at FS #8659 - Sansa c250 - bug in the metronome application (won't play!) ? There is a patch and it's so simple that it should take a minute to check if it works or not
22:32:07 Quit rhodan (Remote host closed the connection)
22:32:11kugelwasn't it just a keymap problem?
22:32:13Slasheripamaury: hmm.. but similar idea somehow, to hold changed filesize and fd pointers
22:32:26pixelmapamaury: the fix is in the simplify PLA patch
22:33:00pamauryI don't know anything about it, it's just to have people look at it :)
22:33:04Slasheripamaury: the fd_bindings cache was created because some files may change during the cache is rebuilt
22:33:42Slasheriso those changes will be committed to the fresh cache immediately when the rebuild is done, so the cache would match to the fs
22:35:35pamauryI think I understand, but it doesn't handle already bound fds, only the one bound during rebuild
22:36:32pixelmapamaury: the fix is only a band-aid fix. Maybe it would be ok for the release but then the c200 is not the only one with keymap problems in metronome
22:36:56Slasheripamaury: yes, that is a bug
22:37:36Slasheripamaury: maybe the fdbind_queue (fdbind_cache) should be repopulated when disabling the dircache temporarily
22:37:53pamaurypixelma: ok
22:38:31Slasheripamaury: then fd bindings would be rebound automatically when the rebuild is done
22:38:44pamaurySlasheri: yes
22:39:05pamaurySlasheri: also, how does fdbind_cache handle filetime/filesize updates and fd unbinding ?
22:40:07Slasheripamaury: it seems it doesn't (when looking at the code). I remember i was going to add handling for that but hit some problem and didn't do it yet
22:40:50pamauryHum, I'm stupid, dircache doesn't unbind, but it doesn't handle filetime/size update during rebuild
22:40:53Slasheribut i don't remember what kind of a problem or is there any issues to do that now
22:42:14Slasheripamaury: i think it should be possible to just add some extra fields to the fdbind_queue to handle filetime/size
22:42:58pamaurySlasheri: yes, it seems simple
22:43:50Slasheripamaury: and on dircache_bind() it should panicf() if fdbind_idx >= MAX_PENDING_BINDINGS (instead of silently ignoring)
22:44:23pamaurySlasheri: disagree
22:44:34*pamaury finishes his apple before explaining why
22:45:03Slasherihmm, really? :)
22:45:24Slasherii would consider that like a "stack overflow"
22:46:05Slasherisecond option would be to abort dircache build..
22:46:15Slasheribut i think panic is more robust and simpler
22:46:17pamaurySlasheri: with MAX_PENDING_BINDINGS=MAX_OPEN_FILES and little intelligence, no overflow can arise
22:46:35pamaurybecause bindings twice to same fd should overwrte and not queue
22:46:44Slasheriwell, true.. but just to be sure, that's what panics are for
22:47:08pamauryYes but here we can avoid it ;)
22:47:20 Join JustJohnny68 [0] (
22:47:44Slasheripamaury: hmm.. not true! because fdbind_cache increases on every open()/bind action
22:47:57Slasheriso it can overflow
22:47:59pamaurySlasheri: yes but it shouldn't
22:48:26Slasheribut you never know that :)
22:48:29*pamaury is only saying that it's not cleverly coded ^^
22:48:40Slasherimaybe there is some hungry background process that likes opening lot of files
22:49:35pamauryWhat I'm saying it say if someone binds pathA to FD and then binds pathB to FD, no new entry should be created, the old one should be reused
22:49:38Slasheripamaury: close() won't reduce fdbind_cache size
22:50:05 Join Omlet [0] (
22:50:19pamauryThe current code can overflow but it could be coded in way so that it doesn't
22:50:37*pamaury feels he is not clear enough
22:50:56Slasheripamaury: but how would you hold path names in memory without fear or overflow?
22:51:20pamaurySlasheri: how does the current code do ?
22:51:26Slasheridircache needs those path strings when rebounding those entries
22:51:57Slasherifd id numbers will change but paths shouldn't
22:52:34pamauryI'm puzzled, I think you didn't understand me. Let's try again
22:52:43Slasheri| for (i = 0; i < fdbind_idx; i++)
22:52:43Slasheri| dircache_bind(fdbind_cache[i].fd, fdbind_cache[i].path);
22:52:49Slasheriplease find that piece of code :)
22:52:57pamauryI know this piece of code :)
22:53:31Slasheriso there it tells dircache to bind that path to this fd
22:53:50casainho(sorry worng channel)
22:53:57Slasheriok, in current implementation that could be solved indeed
22:54:01pamauryYes but if during dircache rebuild, two files are bound to the same fd, the first one can be forgotten, and the entry reused !
22:54:12Slasheribut when you add the filesize/changetime, you couldn't
22:54:39Slasheripamaury: yes, now i understand what you meant :)
22:54:51pamauryah !
22:55:26Slasheripamaury: and that would work until we have to hold those file attributes related to the path in memory
22:56:17AlexPJdGordon_: OK, the problem with the backdrop was having a UI viewport defined in the theme .cfg. Comment that out and it works
22:56:42 Part JustJohnny68
22:58:13pamaurySlasheri: the handling of filesize/filetime would be a little bit more tricky because we would need to know if that quantity had been updated during the rebuild but that's not a real problem. I think everything can be fixed relatively easily
22:58:38pamauryExcept if you think of another problem
22:58:54Slasheripamaury: in my code, i have a diff for fdbind_queue in dircache.h:
22:58:55Slasheri|+ int fd; /* -1 if closed. */
22:58:55Slasheri|+ /* To be used if fd has been closed. */
22:58:55DBUGSent KICK Slasheri to server
22:58:55Slasheri|+ long filesize;
22:58:55Slasheri|+ long startcluster;
22:58:55Kick(#rockbox Slasheri :No flooding!) by logbot!
22:58:58 Quit Omlet (Quit: ( :: NoNameScript 4.22 :: ))
22:59:06 Join Slasheri [0] (miipekk@rockbox/developer/Slasheri)
22:59:13Slasherithat sucks
22:59:19pamaurypastebin it
22:59:29Slasherii have no idea why i haven't finished it
22:59:34Slasheribut you got the idea
22:59:40pamaury logbot a expulsť Slasheri de #rockbox (No flooding!)
22:59:51pamauryYou were kicked beause of flooding
22:59:58Slasheriyes, i now
23:00:03AlexPJdGordon_: However I seem to be seeing that the WPS background doesn't completely replace the backdrop (with r24468)
23:00:09Slasherii think those rules are too strict
23:00:39AlexPJdGordon_: But I'll have another play tomorrow
23:02:09Slasheripamaury: but if you are willing to do, please use that idea and add the modify time in the struct too :)
23:02:22pixelmaSlasheri: you can't read many lines of code in an IRC client, and pastebin (alike) page exist and are easy to use
23:02:26 Join robin0800 [0] (
23:02:44 Quit Farthen (Killed (NickServ (GHOST command used by Farthen_!
23:03:07AlexPAre we waiting for anything for release? Langs?
23:03:09 Join Farthen_ [0] (
23:03:11 Nick Farthen_ is now known as Farthen (
23:03:15Slasheripixelma: that's true, but i think it was more clear to paste a few lines very related code here that to use hard to use links for that
23:03:21pamaurySlasheri: I'm seeing another problem. If someone opens a file, modify the size, and close it, we need to keep an entry if the queue to be sure that the final dircache entry will be valid !
23:03:58pamauryOtherwise, that entry could end up with a bad filesize, ... I think that why your code mention /* To be used if fd has been closed. */
23:04:24pixelmaAlexP: fix the warnings (and some lang guru who is able to) and I had hoped to get some answers but could go the remove route
23:04:27Slasheripamaury: at least i am little reluctant to click on all links, and when using a text based terminal or without a working copy/paste, those links can be a pain
23:05:29Slasheripamaury: ah, yes!
23:05:44pamaurySo you're right, even with a clever code, we could overflow
23:05:47Slasheripamaury: that was the problem and reason i didn't finish the code yet! (at least i think so)
23:05:51JdGordon_AlexP: thats with the rev I asked?
23:06:23JdGordon_hmm, having a ui vierwport defined shouldnt cause any backdrop issues
23:06:26AlexPJdGordon_: With 24431 - 24434 won't boot on the beast
23:06:35AlexPJdGordon_: And with trunk incidently
23:06:41JdGordon_ok, I'll investigate
23:06:55Slasheripamaury: because if we have many R/W fds open to the same file.. we would have to find out the latest modification
23:07:04Slasherihmm.. am i thinking this too hard :)
23:07:13AlexPJdGordon_: Before you spend lots of time, I'll package and upload the theme for you
23:07:28AlexPJdGordon_: But that'll have to be tomorrow (my time) now
23:08:09Slasheripamaury: but that was the reason why fd_cache keeps growing
23:08:32pamauryit grows each time a file is opened
23:08:59Slasheriyes, but wont the entry be kept in the cache?
23:09:27pamaurySlasheri: I don't understand
23:09:39 Quit Zagor (Quit: Clint excited)
23:09:48Slasheripamaury: so what was the problem with that?
23:11:11pamauryThe problem is that when a file a bind to a fd F and then another file gets bound to fd F, we need to keep the first one in the queue because if it's filesize was modified, we will have to make the change in the final dircache entry (when the rebuild is finished). That's the main reason why we can't work with a constant space
23:11:18pamaury*is bound
23:11:43*bertrik started a sansa clip battery benchmark, with mp3 instead of ogg this time
23:11:56pamaurySlasheri: agree ?
23:13:00Slasheripamaury: yes, exactly. that was what i was trying to say earlier :)
23:13:23Slasheriand currently the code should work that way
23:13:33pamauryOk ! We managed to misunderstand each other :)
23:13:37Slasherihehe :)
23:13:58pamauryOk, so the current code lacks two things:
23:14:14pamaury1) It should put in the queue the bindings present before the rebuild
23:14:24pamaury2) It doesn't save filetime/filesize changes during the rebuild
23:14:33Slasheriyes, very true
23:14:49pamaury1) is easy to fix. 2) is what you started
23:16:40Slasheriindeed. I thinks the thing 2) might need special case handling if there are more than one fd pointing simultaneously to the same file
23:16:52Slasheri(would that be ever likely to happen?)
23:17:18*pamaury is doing the same remark
23:17:40pamauryIt would be embarassing that the same file be opened twice at the same time in rockbox
23:18:15Slasheriindeed.. maybe we could assume that would be too unlikely to happen?
23:19:17pamauryYes I think.
23:19:30kugelSlasheri: there's this 30k buffer with dircache that's statically allocated. could that be buffer_alloc'd instead?
23:19:31Slasheri(simultaneous read-only fds wouldn't be a problem)
23:19:43pamaurySlasheri: read-only won't change filesize/filetime !
23:20:04pamaurykugel: I think there is a patch is FS for this
23:20:26Slasherikugel: which buffer?
23:20:37kugelI also wondered if it's possible to disable the entire thread if dircache is deactivated
23:20:40pamauryCan someone reproduce this: FS #9654 - A song that is played the 1st time cannot be rated ? I don't manage.
23:21:05pamaurykugel: also in FS, same patch irrc (I think it was made by JdGordon)
23:21:13Slasherikugel: i think so :)
23:21:26Slasheribut now sleep everyone!
23:21:33pamaurykugel: FS #9506 - buffer_alloc the dircache stack
23:22:06JdGordon_yep, thats my patch
23:22:21kugelSlasheri: I tried that, very strange this happened though ;)
23:22:21Slasheripamaury: i added those as todo in in my dircache code not to forget.. in case nobody else remembers :)
23:22:31JdGordon_doing that makes it damn near impossible to ever have reallocable buffer_alloc()s so we can toggle features without restarting though
23:22:49pamaurySlasheri: ok, so I let you fix this as you already have some code
23:23:26Slasheripamaury: i try to remember :) i think there is some time from my last commit :)
23:24:09pamauryJdGordon: does you bug report (FS #9654 - A song that is played the 1st time cannot be rated) still holds ?
23:25:57pamauryoops, sorry you're aren't the one who posted, you just modified it !
23:27:11JdGordon_I do belive it does
23:27:20pamauryI can't reproduce it.
23:27:27pamauryCan you try to do so ?
23:27:33 Quit ender` (Quit: Trying to establish voice contact ... please yell into keyboard.)
23:27:38JdGordon_one min, I'll have a go on my mini
23:28:00*JdGordon_ is initializing the db
23:28:50JdGordon_yep, no "set rating" wps context menu
23:29:12pamauryHum, why can't I reproduce it !
23:29:27JdGordon_dircache is off iirc
23:29:40JdGordon_no, dircache is on
23:30:12*pamaury is reinitializing the db...
23:30:13JdGordon_db lload to ram -> off, auto update -.> no, FUCK gather runtime data -> no :p
23:30:16 Quit HBK ()
23:30:41JdGordon_with that enabled it works
23:30:43pamauryyou have a problem with gather runtime data ?
23:30:43JdGordon_shock horror :p
23:31:04JdGordon_shouldnt that be enabled by default?
23:31:59pamauryAh you're right, I don't have it also :(
23:32:14 Quit bertrik (Quit: De groeten)
23:32:29pamaury(I never use ratings)
23:33:15pamauryBut the next/prev tricks doesn't work for me
23:34:09 Join HBK [0] (
23:34:37***Saving seen data "./dancer.seen"
23:34:41pamauryHum, even when I reboot I can't rate it
23:35:20pamauryPerhaps that's normal with "gather runtime data" to off
23:36:40pamauryJdGordon: are you sure you can reproduce it when gather runtime data is on ? When it's off, it's normal that's it doesn't appear, right ?
23:39:47 Quit toffe82 (Read error: Connection reset by peer)
23:39:53JdGordon_no, I cant repro with gather runtime on
23:40:00JdGordon_which makes sense so its not actually a bug
23:40:05JdGordon_is PEBKAC
23:40:14pamauryPEBKAC ?
23:40:22 Quit dmb_ (Ping timeout: 272 seconds)
23:41:06JdGordon_user stupidity
23:41:41pamauryWhy this acronym/word ?
23:41:50 Quit GeekShadow (Quit: The cake is a lie !)
23:42:48 Quit jgarvey (Quit: Leaving)
23:43:49JdGordon_Problem Exists Between Keyboard And chair
23:43:55JdGordon_(doesnt really work with a DAP though)
23:45:11 Quit piotrekm (Quit: Leaving.)
23:47:55 Quit evilnick_B (Quit: Page closed)
23:49:26pamauryJdGordon_: do I close it ?
23:50:23CIA-8New commit by amiconn (r24473): APE codec: Speed up decoding of -c2000 and higher on ARMv4 and coldfire by fusing vector math for the filters. Speedup is roughly 3.5% for -c2000, 8% ...
23:50:32pamauryBecause the bug report says that it displays '<no info>' with gather runtime data one
23:53:11pamaurygevaerts: could you try to reproduce a bug ? It take sone minute. FS #9654
23:53:12 Join dmb_ [0] (~Dmb@unaffiliated/dmb)
23:53:47gevaertspamaury: I was just about to go to sleep :\
23:54:04*gevaerts also has no idea how to rate a song
23:54:13pamauryHold select in WPS screen
23:54:20*pamaury learned it a few minutes agp
23:54:29pamauryIn the menu there is a Set Rating item
23:55:04gevaertsWhat does "Copy a new song into your library" mean?
23:55:22pamauryPick a song that you never played
23:55:26pamauryOn that device
23:55:31pamauryAlternativaly, reset the DB
23:56:02gevaertsare you sure that that's what he means?
23:56:37pamauryNo but that's how I understand it
23:56:48amiconnNano G2 has code in IRAM?
23:56:48pamauryI tried both, copy a new song on the device and reinit DB
23:57:24gevaertsI'd assume it means "copy a new song and don't bother updating the database"
23:57:45 Quit blithe (Remote host closed the connection)
23:58:50pamauryI mean I tried both. Copy a new song, try it. FAIL. Reinit DB, try it, FAIL.

Previous day | Next day