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

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

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

#rockbox log for 2009-06-25

00:00:15Bagderyes, 1:50 for upload sounds too much
00:00:33Bagderanyway, I need to make the server reject such completions
00:02:16rasherAlso something to investigate why it didn't kill it
00:02:50MikachuBagder: here's my output if it somehow helps
00:03:46Bagderclip fuzesim h10_5gb h300 iaudiox5 ipodmini1g mrobe500 sansae200sim sansae200v2 were the dupes
00:04:46rasherHm, I think the system wastes time
00:05:02rasherMy client started 6 builds, but only finished one
00:05:08rasherthe rest got killed halfway through
00:05:14rashergranted, it's not a fast client
00:05:40Bagderright, the order of how to hand out the build may need further tweaking
00:05:54rasherBut it seems like it could have at least finished another one if it didn't keep getting killed
00:06:16rasherIs it even necessary to give the same build to multiple hosts before the end?
00:06:25Bagderit doesn't
00:06:41rasherAh. Then my client just really really sucks :)
00:07:00Bagderit sorts the builds first on handedoutcount and then on "weight"
00:07:26Bagderthen it moves through the list of clients in weight order and picks a build from that list
00:07:34Bagderuntil all clients have 3 builds
00:08:06Bagderin this case we had 10 clients doing 59 builds
00:08:35Bagderso first it'd had out the top 30 builds, the ones considered slowest
00:09:16Bagderand for each completed build being reported, it hands out another from the build list
00:10:41CtcpPing from gevaerts!n=fg@rockbox/developer/gevaerts
00:11:31pixelmafor further reference: the "Sleep Timer" went into the "Time & Date" "menu" inside "System" - the manual still mentions the old place only (directly inside "System")
00:12:46pixelmahmm... where did I look?
00:13:40CIA-71New commit by mcuelenaere (r21501): Fix some forgotten OR's in some button-target.h files
00:13:43 Nick fxb is now known as fxb__ (
00:14:03pixelmaah, took a manual of a non-RTC target. Sorry...
00:14:50 Quit mirak ("Ex-Chat")
00:16:22mcuelenaerecan anyone tell me why BUTTON_O|BUTTON_BACK is in BUTTON_MAIN for iRiver H10 targets but those buttons aren't defined?
00:21:56CIA-71New commit by bagder (r21502): fix the counting of build clients, make sure that duplicate completions are ...
00:22:03 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
00:23:48CIA-71New commit by mcuelenaere (r21503): Also fix the iRiver H10
00:29:09 Quit evilnick ("Page closed")
00:30:35 Quit ender` (" I spilled Spot Remover on my dog... Now he's gone.")
00:32:45 Join n00b81 [0] (n=taylor@unaffiliated/n00b81)
00:34:10 Quit bmbl ("Bye!")
00:36:22 Quit domonoky (Read error: 104 (Connection reset by peer))
00:38:46 Join perrikwp [0] (
00:45:05 Part n00b81 ("Leaving")
00:47:01***Saving seen data "./dancer.seen"
00:48:05 Quit notlistening ("Leaving")
00:50:47 Quit foaly ("leaving")
00:50:52 Quit bertrik ("De groeten")
00:53:37 Quit mcuelenaere ()
00:56:32 Join n00b81 [0] (n=Taylor@unaffiliated/n00b81)
00:57:21 Part n00b81 ("Leaving")
01:10:23 Quit tessarakt2 ("Client exiting")
01:20:57CIA-71New commit by rasher (r21504): Statements end with semicolon...
01:23:24 Quit JdGordon| ("Miranda IM! Smaller, Faster, Easier.")
01:25:42 Quit Thundercloud (Remote closed the connection)
01:28:48 Quit faemir (Connection reset by peer)
01:37:02 Quit barrywardell ()
01:38:49 Quit gevaerts (Nick collision from services.)
01:39:01 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
01:39:34 Join Hillshum [0] (i=4ba5ebce@gateway/web/freenode/x-7047cda176f7f414)
01:41:25CIA-71New commit by rasher (r21505): Modify the number of bogomips if not all cores are used (specified at commandline)
01:52:17 Join evilnick [0] (i=620ec27e@gateway/web/freenode/x-d4f70ee1ed2c8b5d)
01:54:17 Part toffe82
01:56:23 Quit Hillshum (Ping timeout: 180 seconds)
02:15:40 Quit dmb (Read error: 113 (No route to host))
02:15:52 Join senpai3330 [0] (n=4b8ff2ac@gateway/web/cgi-irc/
02:17:40 Quit PaulJam_ (".")
02:17:40 Quit senpai3330 (Client Quit)
02:17:44 Join perrikwp1 [0] (
02:19:37 Quit kugel (Read error: 104 (Connection reset by peer))
02:35:37 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/freenode/x-e539bdc7f2ac465f)
02:35:57 Quit perrikwp (Read error: 110 (Connection timed out))
02:38:03 Join Strath [0] (
02:47:02***Saving seen data "./dancer.seen"
02:52:36 Quit evilnick ("Page closed")
02:56:42 Quit Zarggg (Read error: 60 (Operation timed out))
02:57:00 Join Zarggg [0] (
02:57:54 Quit FlynDice (Remote closed the connection)
03:02:28 Quit jgarvey ("Leaving")
03:04:13 Quit funman ("free(random());")
03:15:02 Quit BlakeJohnson86 (Remote closed the connection)
03:18:25 Quit robin0800 ("Leaving")
03:23:00 Join BlakeJohnson86 [0] (
03:27:46 Quit BlakeJohnson86 (Remote closed the connection)
03:29:05 Join BlakeJohnson86 [0] (
03:31:38 Quit Sir_Brizz ("leaving")
03:36:48 Join robin0800 [0] (
03:38:29 Join perrikwp [0] (
03:39:03 Quit robin0800 (Client Quit)
03:39:22 Join robin0800 [0] (
03:39:31 Join perrikwp2 [0] (
03:40:26 Quit robin0800 (Client Quit)
03:40:44 Join robin0800 [0] (
03:42:35 Quit robin0800 (Client Quit)
03:42:54 Join robin0800 [0] (
03:44:43 Quit perrikwp (Read error: 60 (Operation timed out))
03:46:26 Quit robin0800 (Remote closed the connection)
03:48:26 Quit amiconn (Nick collision from services.)
03:48:30 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
03:48:39 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
03:49:16 Quit pixelma (Nick collision from services.)
03:49:19 Join pixelma_ [50] (n=pixelma@rockbox/staff/pixelma)
03:49:36 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
03:55:12 Quit perrikwp1 (Read error: 110 (Connection timed out))
03:55:57 Quit jfc (Remote closed the connection)
03:57:03 Join jfc [0] (
04:11:31 Quit cool_walking_ (
04:11:31 Quit saratoga (
04:18:08 Join perrikwp [0] (
04:21:07 Quit Sajber^ (Read error: 54 (Connection reset by peer))
04:22:47 Quit BlakeJohnson86 ("Leaving.")
04:23:49 Quit perrikwp2 (Read error: 60 (Operation timed out))
04:25:18 Join BlakeJohnson86 [0] (
04:25:37 Quit kachna|lappy (Read error: 113 (No route to host))
04:38:10 Join perrikwp1 [0] (
04:43:22 Quit fyrestorm (Read error: 104 (Connection reset by peer))
04:43:22 Quit BlakeJohnson86 (Read error: 104 (Connection reset by peer))
04:44:20 Join BlakeJohnson86 [0] (
04:47:04***Saving seen data "./dancer.seen"
04:49:41 Join anomoly [0] (
04:56:16 Join linuxguy3 [0] (
04:56:38 Quit perrikwp (Read error: 110 (Connection timed out))
04:59:38 Quit Galois ("Leaving")
05:07:04 Quit jkl (Read error: 60 (Operation timed out))
05:11:18 Join perrikwp [0] (
05:15:53 Join FlynDice [0] (n=FlynDice@
05:21:41 Join jkl [0] (
05:22:17 Join dmb [0] (n=dmb@unaffiliated/dmb)
05:22:50 Join Galois [0] (i=djao@
05:29:34 Quit perrikwp1 (Read error: 110 (Connection timed out))
05:32:18 Join Blue_Dude [0] (
05:33:01Blue_DudeFixed a playback bug: FS #10157 : inaccurate seeking while paused
05:33:34Blue_DudePlease also see FS #10377 - PCM buffer remainder is played under too many conditions
05:33:42 Quit Blue_Dude (Client Quit)
05:36:04 Quit Strath ("Leaving")
05:39:26 Join perrikwp1 [0] (
05:50:15 Quit Horscht ("Verlassend")
05:57:29 Quit perrikwp (Read error: 110 (Connection timed out))
06:03:35 Join jkl__ [0] (
06:03:58 Quit jkl (Read error: 113 (No route to host))
06:07:05 Join martian67_ [0] (n=martian6@2001:470:b:356:221:91ff:fe8c:d8a7)
06:08:07 Quit martian67_ (SendQ exceeded)
06:08:09 Quit martian67 (Read error: 113 (No route to host))
06:09:05 Join martian67_ [0] (n=martian6@2001:470:b:356:221:91ff:fe8c:d8a7)
06:13:27 Join Xerion_ [0] (
06:15:54 Quit jkl__ ("Leaving")
06:16:05 Join jkl__ [0] (
06:16:08 Nick jkl__ is now known as jkl (
06:19:40 Join perrikwp [0] (
06:31:02 Quit Xerion (Read error: 110 (Connection timed out))
06:31:02 Nick Xerion_ is now known as Xerion (
06:38:14 Quit perrikwp1 (Read error: 110 (Connection timed out))
06:39:31 Quit timc (Remote closed the connection)
06:44:39 Join timc [0] (n=aoeu@
06:47:05***Saving seen data "./dancer.seen"
06:53:42 Quit kkurbjun ("Leaving.")
06:56:49 Quit jkl ("Leaving")
07:17:27 Quit perrikwp (Read error: 110 (Connection timed out))
07:23:08 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
07:23:42 Join framo4 [0] (
07:27:56 Quit slam_ (Read error: 110 (Connection timed out))
07:39:15 Join AndrewRB [0] (
07:40:14AndrewRBhi, anybody awake?
07:42:12AndrewRBi'll take that as a "no" then. oh well. goodnight.
07:42:28 Quit AndrewRB (Client Quit)
07:46:23 Join slam_ [0] (
07:53:24 Quit Unhelpful (Read error: 101 (Network is unreachable))
08:01:52 Join Unhelpful [0] (n=Militant@rockbox/developer/Unhelpful)
08:13:28 Nick martian67_ is now known as martian67 (n=martian6@2001:470:b:356:221:91ff:fe8c:d8a7)
08:25:20 Join n1s [0] (n=n1s@rockbox/developer/n1s)
08:34:53 Join bertrik [0] (
08:37:05 Join petur [50] (n=petur@rockbox/developer/petur)
08:37:51bertrikmarkun, I got RTC reading on the meizu m3 working last night
08:38:04bertrikthe data seems to be bitswapped :/
08:39:10 Quit safetydan ("Leaving.")
08:39:21amiconnHmm, and other i2c transfers are not bitswapped?
08:39:29bertrikI'm not sure what endian mode we are running with this processor
08:39:51bertrikamiconn, this is the first i2c communication I have working, so I don't know
08:40:01amiconnEndianess wouldn't matter. It defines byte order, not bit order
08:40:19 Join flydutch [0] (
08:41:08bertrikhm, yes makes sense
08:42:17amiconnMaybe there's a setting in the controller that defines bit order. However, selectable bit order doesn't make sense for i2c imo
08:42:49amiconnIs that i2c controller part of a more complex serial i/o controller?
08:43:02 Join Rob2223 [0] (
08:43:29bertrikit's built into the s5l8700
08:43:34 Join kachna|lappy [0] (
08:43:46bertrikwhich is the main SoC of the meizu m3
08:46:28 Join einhirn [0] (
08:47:09***Saving seen data "./dancer.seen"
08:53:23amiconnI know
08:53:51amiconnWhat I mean are the modules within the SoC
08:54:27 Part anomoly
08:56:09bertrikno this is an external RTC, the S35390
08:58:20bertrikI think I'll clean up the i2c code and try to read from the codec later tonight and see if that is bit swapped too
09:00:35 Quit Rob2222 (Read error: 110 (Connection timed out))
09:07:58 Nick fxb__ is now known as fxb (
09:09:55 Nick fxb is now known as fxb__ (
09:09:58 Nick fxb__ is now known as fxb (
09:23:03 Quit bertrik ("De groeten")
09:28:26 Join Thundercloud [0] (
09:31:59 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
09:44:32 Join moxfyre [0] (
09:46:11 Nick fxb is now known as fxb__ (
09:47:27 Quit moxfyre (Client Quit)
09:47:46 Join moxfyre [0] (
09:48:31 Join ender` [0] (
09:49:18 Nick daurn is now known as daurnimator (n=daurnima@unaffiliated/daurnimator)
09:50:57 Quit Thundercloud (Remote closed the connection)
09:51:50 Join nibbler_ [0] (
09:52:09 Quit BHSPitMonkey ("Ex-Chat")
10:02:32 Join barrywardell [0] (n=barrywar@
10:02:36 Quit barrywardell (Remote closed the connection)
10:08:24 Quit kachna|lappy (Read error: 113 (No route to host))
10:12:34 Quit moxfyre ("Ex-Chat")
10:12:51 Join moxfyre [0] (
10:16:16 Quit framo4 ("muuh")
10:29:40 Nick J-23 is now known as audiofap (
10:29:50 Nick audiofap is now known as J-23 (
10:38:09 Join krazykit` [0] (
10:38:23 Join planetbeing [0] (
10:40:12 Quit planetbeing (Client Quit)
10:40:49 Quit krazykit (Remote closed the connection)
10:47:14***Saving seen data "./dancer.seen"
11:00:31 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
11:01:37mtlinuxstb, saratoga : The bug about id3v1buf[0] and -buf[1] is fixed. RMContext was big (~1148) so copying it to id3->toc overwrote id3v2buf then id3v1buf[0] and 1. Modified the code so that RMContext is now 124 bytes and I'm using id3v2buf now instead of toc to copy RMContext.
11:03:16mtI'm now also using the same bitstream.c/h and bswap.h files from libwma with slight modifications so that they would compile with cook, to be sure the target problems aren't from those.
11:03:23 Nick zitune[afk] is now known as zitune (n=zitune@bearstech/zitune)
11:03:34mtStill getting undefined instruction though.
11:09:03 Quit bmbl (Read error: 110 (Connection timed out))
11:25:58 Join kachna|lappy [0] (
11:31:50 Join robin0800 [0] (
11:48:22 Join fml [0] (n=4fd3c824@gateway/web/cgi-irc/
11:49:16 Join Lss [0] (
11:49:18fmlHm... two very promising patches: FS #10157 and FS #10377 Sadly, I can't tell whether they make sense
11:53:19 Join merbanan [0] (n=banan@
11:54:52 Join Sajber^ [0] (
11:55:41 Quit flydutch ("/* empty */")
11:58:59 Part lazysod
12:10:13linuxstbmt: Do you still need all that information in RMContext? Getting it down to 100 bytes would be useful, as then you can use the otherwise unused toc field.
12:20:22 Quit fml ("CGI:IRC 0.5.9 (2006/06/06)")
12:36:43 Quit nibbler_ (Read error: 110 (Connection timed out))
12:41:20 Quit FlynDice (Remote closed the connection)
12:43:14 Join PaulJam [0] (
12:47:18***Saving seen data "./dancer.seen"
12:50:49 Join notlistening [0] (
12:57:28 Quit BlakeJohnson86 (Remote closed the connection)
12:57:58 Join BlakeJohnson86 [0] (
13:00:43 Quit BlakeJohnson86 (Remote closed the connection)
13:02:09 Join BlakeJohnson86 [0] (
13:13:09 Nick zitune is now known as zitune[afk] (n=zitune@bearstech/zitune)
13:13:19 Quit notlistening (Remote closed the connection)
13:17:04 Join lee321987 [0] (
13:17:12 Quit lee321987 (Client Quit)
13:27:41 Nick krazykit` is now known as krazykit (
13:31:46 Nick zitune[afk] is now known as zitune (n=zitune@bearstech/zitune)
13:41:55 Join AndrewRB [0] (
13:45:44 Join flydutch [0] (
14:05:57AndrewRBwould anybody mind taking a quick look at this: and telling me which Rockbox target is the most similar? =)
14:09:13gevaertsmr500 also has a TMS320
14:09:52gevaertsI do think your 512MB RAM is probably 512Mb though
14:10:03AndrewRBi'm not so sure
14:10:23gevaerts512MB seems like an awful lot for a DAP
14:10:38AndrewRBunless I don't understand which capital refers to which term
14:10:50AndrewRBbytes = B right?
14:11:17AndrewRBits not a dap its a PMP, focused around playing video
14:11:40AndrewRB(i know the two are blurred, but still i feel dirty calling it a DAP)
14:12:12*gevaerts points out that rockbox is a DAP firmware, so anything rockbox gets ported to is by definition a DAP ;)
14:12:55gevaertsAnyway, even then 64MB sounds much more likely.
14:12:55*AndrewRB wonders who defines these terms in the first place
14:12:57Erant512MB? No way.
14:13:20ErantThere's hardly a SoC out there that can address 512MBs of RAM anyway
14:13:46AndrewRBok, clearly i'm wrong then. lemme try and see where i got that info from to start with
14:15:54 Join robin0800_ [0] (
14:16:37AndrewRBi misread the datasheet
14:17:42 Quit robin0800_ (Client Quit)
14:18:01 Join robin0800_ [0] (
14:19:12AndrewRBso the ZVM and MR500 both share the same SoC? huh
14:20:08gevaertsI'm not sure if TMS320 is a single soc or a family
14:20:25AndrewRBtms320 is a family - the DM320 bit is the specific chip
14:20:56AndrewRBwhich it seems both share. the O2 which i'm hoping to get somewhere with has a tms320dm644
14:22:03AndrewRBso anyway, i'm currently learning arm assembly, and reading through this book; ... anything else you guys could suggest I read/learn?
14:22:54 Quit robin0800_ (Client Quit)
14:23:13 Join robin0800_ [0] (
14:23:21gevaertsyou're familiar with c?
14:23:25 Part LinusN
14:24:07gevaertsin that case, just the datasheets I guess
14:24:20 Quit robin0800_ (Client Quit)
14:24:37 Join robin0800_ [0] (
14:24:55AndrewRBsure. do you think I should add the hardware info from my forum post to a new rockbox wiki page? or would I be best leaving it till I'v made some sort of progress
14:26:27gevaertsdo it now. That way it doesn't get lost if you'd happen to lose interest :)
14:27:44AndrewRBhaha. if i lose interest, i can't see anybody else taking it on as a rockbox target - it isn't a very popular player. but sure, will do. gotta get used to TWiki first. any suggestions as for the article title? "Cowon O2 Hardware Info" perhaps?
14:27:54 Quit robin0800_ (Client Quit)
14:28:12 Join robin0800_ [0] (
14:28:31gevaertssounds good
14:28:53 Nick fxb__ is now known as fxb (
14:29:23 Quit robin0800_ (Client Quit)
14:29:36AndrewRBgreat. by the way, judging by the gallery pics, devcon looks like it was fun =)
14:29:42 Join robin0800_ [0] (
14:29:58gevaertsit was :)
14:30:32 Join MarcGuay [0] (
14:30:50AndrewRBwhat was the final tower height? (minus the non-dap on the top =P)
14:32:52gevaertsI don't know. Last year I looked closely at the pictures, DeviceChart and other sources to work it out, but I haven't done that yet for this year's Tower
14:34:41 Quit AndrewRB ("XChat beats Konqueror any day...")
14:35:41linuxstbgevaerts: Or you could project a horizontal line to the person standing nearby, and then ask him to measure himself...
14:36:02gevaertshow accurate would that be? :)
14:36:09*linuxstb _thinks_ it's a him..
14:36:29linuxstbgevaerts: Depends how accurate your line projection is...
14:36:47linuxstbThere are a few straight lines for references...
14:36:48gevaertslinuxstb: which picture are you looking at?
14:36:50MarcGuaylinuxstb: Did you write sansapatcher?
14:37:11linuxstbMarcGuay: That depends on whether you have a follow-up question ;)
14:37:15gevaertsI think that's kugel
14:38:26MarcGuaylinuxstb: More of a bug report. :) Is that better? After it writes the bootloader it doesn't seem to send a "close write" command - the sansa firmware continues to display "writing".
14:38:56 Quit DataGhost ()
14:39:00MarcGuayProbably not a big deal as most people should be using RButil...
14:39:11linuxstbDoes it not happen in rbutil then?
14:39:23MarcGuayDidn't check.
14:39:38linuxstbAnd which OS are you using?
14:39:40gevaertshm, could this explain the issues on OSX?
14:39:53MarcGuayWindows XP
14:40:05linuxstbgevaerts: Which issues?
14:40:35gevaertsnot sure. Wasn't there something where you had to unplug because otherwise the OS doesn't see some changes?
14:40:45*gevaerts doesn't remember this very precisely
14:40:56MarcGuayWhere doesn't RButil store it's logs?
14:41:18 Join LambdaCalculus37 [0] (i=44a0430d@gateway/web/freenode/x-0c3b8e504ffe7955)
14:41:23linuxstbIt doesn't store them behind the sofa.
14:41:32MarcGuayIt installed the bootloader fine.
14:41:32scorche|shor in the refrigerator
14:41:45MarcGuayIt claims to have logged the install, though.
14:41:45LambdaCalculus37Or in the car trunk.
14:42:11scorche|shor underneath the bed
14:42:38robin0800_they are not here
14:44:43MarcGuayZagor, Bagder: "Devcon 2009 is in progress"
14:45:52linuxstbIt was true at the time the news was posted.
14:46:37 Nick robin0800_ is now known as robin0800__ (
14:47:21***Saving seen data "./dancer.seen"
14:58:29 Quit PaulJam (Nick collision from services.)
14:58:36 Join PaulJam_ [0] (
14:59:56 Quit robin0800__ ("Leaving")
15:00:22 Join robin0800_ [0] (
15:02:25 Quit robin0800_ (Client Quit)
15:02:44 Join robin0800_ [0] (
15:04:17domonokyMarcGuay: rbutil stores what it has installed in .rockbox/rbutil.log on your player.
15:04:27 Quit robin0800_ (Client Quit)
15:04:45 Join robin0800_ [0] (
15:05:23 Quit robin0800_ (Client Quit)
15:05:41 Join robin0800_ [0] (
15:06:28 Quit robin0800_ (Client Quit)
15:06:46 Join robin0800_ [0] (
15:12:24 Join evilnick [0] (i=0c140464@gateway/web/freenode/x-c00888d5a6f15d79)
15:13:38 Quit robin0800_ ("Leaving")
15:13:56 Join robin0800_ [0] (
15:15:08 Quit robin0800_ (Client Quit)
15:15:26 Join robin0800_ [0] (
15:16:02 Join cmwslw [0] (
15:16:23 Quit Sajber^ (Read error: 104 (Connection reset by peer))
15:16:32 Part cmwslw ("Ex-Chat")
15:17:30 Quit robin0800_ (Client Quit)
15:17:48 Join robin0800_ [0] (
15:18:33MarcGuaydomonoky: Is it possible to see the CL output of sansapatcher?
15:19:21 Quit robin0800_ (Client Quit)
15:19:32 Quit n1s (Read error: 110 (Connection timed out))
15:19:39 Join robin0800_ [0] (
15:20:17domonokyMarcGuay: i am not sure, maybe if you start rbutil via commandline.
15:21:37 Quit robin0800_ (Client Quit)
15:21:56 Join robin0800_ [0] (
15:22:59 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
15:26:12CIA-71New commit by mcuelenaere (r21506): Lua: add the package library
15:32:16 Quit slam_ (Read error: 110 (Connection timed out))
15:33:25 Quit einhirn (Read error: 104 (Connection reset by peer))
15:34:12 Join Blue_Dude [0] (
15:35:30MarcGuaydomonoky: Doesn't seem to be any different...
15:35:52MarcGuayIs rbutil.log the "installation log" it speaks of?
15:35:58 Quit robin0800_ ("Leaving")
15:36:46 Join robin0800_ [0] (
15:40:38 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
15:42:08CIA-71New commit by bagder (r21507): expanded help text for the options
15:42:47 Quit robin0800_ ("Leaving")
15:43:09 Join robin0800_ [0] (
15:44:16Bagderone client connects badly and gets disconnected by the server
15:44:23Bagderonly to reconnect badly again immediately
15:45:29gevaertsI'm seeing disconnects
15:45:49Bagderit's rasher's cygwin client
15:46:05Bagderuh and I see why
15:46:12Bagderit claims 0 bogomips
15:46:37gevaertsthat's probably realistic though :)
15:49:40rasherYeah, /proc/cpuinfo doesn't have bogomips lines
15:50:26gevaertsBagder: maybe initialise the bogomips value to 1 instead of 0
15:50:44Bagderexactly, 1 if there's no /proc/cpuinfo
15:51:19CIA-71New commit by bagder (r21508): no /proc/cpuinfo will now claim 1 bogomips, the server rejects clients that ...
15:52:08 Nick zitune is now known as zitune[afk] (n=zitune@bearstech/zitune)
15:52:55Bagderrasher: will you update your cygwin client to this?
15:54:08 Join jgarvey [0] (
15:55:56 Quit kachna|lappy (Success)
15:56:08rasherWhen I get the chance (not for several hours..)
15:56:25Bagderit might be a good test
15:56:39rasherIt also means it detects no cores, but you can override that
15:57:01 Join n1s [0] (n=n1s@rockbox/developer/n1s)
15:58:08rasherBut maybe that should be based on another line
15:58:20rasher.. less likely to be absent
15:58:38rasherHow about uname in the HELLO string?
15:59:14gevaertswhich bits are you missing?
15:59:18BagderI check the bogomips one only because its the last in the line and I split all args on space
15:59:20 Join notlistening [0] (
15:59:41BagderI guess I can check for "" instead
15:59:58notlisteningthanks for looking at that domonoky i will see where the problem lies
16:00:16rasherMost lines were there, but not bogomips. Check for Cpuflags maybe?
16:00:34 Join Hillshum [0] (i=cd7ae8fa@gateway/web/freenode/x-7131833a18ea38dc)
16:01:21rasherPretty certain that was there
16:02:03rasherBagder: wait a minute - you should be able to update it!
16:02:12Bagderhehe, indeed
16:03:02 Join tom_ [0] (
16:04:07 Quit notlistening (Remote closed the connection)
16:04:07 Quit tom_ (Read error: 104 (Connection reset by peer))
16:06:33Bagdernice test anyway, one client connects/disconnects every second during the entire round ;-)
16:06:55Bagder"ALERT: this build was already completed"
16:07:04Bagdernow detecting the situation I found yday
16:08:31 Quit robin0800_ ("Leaving")
16:10:38 Join robin0800_ [0] (
16:10:52Hillshumrasher: now I see what I didn't yesterday in english.lang
16:12:25Hillshumthe clip fits under *
16:13:13 Quit robin0800_ (Client Quit)
16:14:04 Join robin0800_ [0] (
16:14:28 Join daurn [0] (n=daurnima@unaffiliated/daurnimator)
16:16:21 Quit robin0800_ (Client Quit)
16:16:48 Join robin0800_ [0] (
16:17:41 Quit daurnimator (Read error: 60 (Operation timed out))
16:23:44 Quit robin0800_ ("Leaving")
16:24:02 Join robin0800_ [0] (
16:24:21 Quit n1s (Read error: 110 (Connection timed out))
16:24:25 Join fyrestorm [0] (
16:34:01 Join n1s [0] (n=n1s@rockbox/developer/n1s)
16:34:27 Join Blue_Dude_ [0] (
16:36:30 Join toffe82 [0] (n=chatzill@
16:47:22***Saving seen data "./dancer.seen"
16:52:19 Quit Blue_Dude (Read error: 110 (Connection timed out))
16:54:18rasherBagder: I also get "error at line 302.", I assume where it does scalar() on something that doesn't have a value yet
16:59:33rasherBagder: Should I update the cygwin client?
16:59:52Bagderplease do
17:00:28rasherI don't think it'll help
17:00:40rasheropen CPUINFO, "</proc/cpuinfo" will succeed
17:01:33rasherMaybe just my $bogomips = 1;
17:01:57Bagderyeah, that should work fine enough
17:01:57rasherA pity-bogomips :)
17:05:07*rasher tries to fix the cores thing as well
17:08:04 Quit n1s (Read error: 110 (Connection timed out))
17:08:30CIA-71New commit by rasher (r21509): Set the default number of bogomips to 1, to make sure the server doesn't reject the client. ...
17:09:57Bagderworks fine
17:10:41*rasher notices that uname output is already in the hello command... wonder how I missed that
17:11:22CIA-71New commit by bagder (r21510): check for duplicate client names, and use client+user names internally to ...
17:11:36rasherI think it'd be a good idea to keep a few cygwin clients around, since stuff sometimes breaks on cygwin only
17:12:06Bagderit'll be interesting to see how such clients perform
17:12:17rasherI suspect very poorly
17:15:25linuxstbCould they be too slow to get anything finished, meaning we'll never know the results?
17:15:44*linuxstb hasn't been following the build system changes...
17:16:01Bagderyes, that's a risk
17:17:32CIA-71New commit by bagder (r21511): change the uploaded file names to use the user name embedded as well, as that's ...
17:17:59CIA-71New commit by bagder (r21512): use dash between client and user name so that we can match the upload file ...
17:19:22 Join GodEater- [0] (
17:20:03 Join Riku [0] (
17:20:52 Quit Lss (
17:20:52 Quit Neovanglist (
17:20:52 Quit GodEater (
17:20:52 Quit J-23 (
17:21:49NJoinNeovanglist [0] (i=Neovangl@
17:25:10 Quit cg_ (Remote closed the connection)
17:25:41 Nick GodEater- is now known as GodEater (
17:30:16rasherBagder: Time for an automatic script upgrade?
17:38:05 Quit robin0800_ ("Leaving")
17:39:00 Nick fxb is now known as fxb__ (
17:43:01 Join robin0800_ [0] (
17:44:25 Quit robin0800_ (Client Quit)
17:45:48 Join Cory [0] (
17:46:16 Nick Cory is now known as Guest53002 (
17:46:18 Quit Guest53002 (Client Quit)
17:46:37 Join Cory|XChat2 [0] (
17:51:55 Join cg_ [0] (
17:52:26 Join AndyI [0] (i=AndyI@
17:55:07Blue_Dude_rasher: should we use langtool to add items? Or is adding items to english.lang sufficient?
17:55:33rasherBlue_Dude_: No and yes, in that order
17:55:48Hillshumwe have to get them to all languages with langtool though
17:55:50Blue_Dude_Ah thanks.
17:56:23rasherHillshum: no
17:56:27rasherThat'd be genlang
17:56:29 Quit robin0800 (Read error: 110 (Connection timed out))
17:56:40Hillshumoh yeah
17:59:13 Join robin0800 [0] (
18:04:50 Join JdGordon| [0] (i=63cb3d60@gateway/web/freenode/x-f5abdc2ee71f7a00)
18:07:25 Quit HellDragon (Client Quit)
18:07:44 Quit PaulJam_ (Nick collision from services.)
18:07:49 Join PaulJam [0] (
18:08:48 Join simjoko [0] (
18:11:57 Quit petur ("work->home")
18:12:09simjokoanyone interested in some benchmarks on ogg/mp3/flac on gcc-4.0.3/4.4.0/trunk? had too much time to kill :P
18:13:12Hillshumsimjoko: if you've done them already, sure
18:13:22simjokolet me paste them somewhere
18:15:34rasher0.00MHz needed for realtime. That doesn't sound right
18:15:55simjokofor some reason ogg did not play on gcc 4.4
18:15:59simjokobut on 4.0.3 and trunk oO
18:16:13Hillshumcan we get that on Monkey's Audio?
18:16:37simjokocan do that
18:17:05Hillshumsimjoko: I meant the 0.00MHz decode
18:17:18Hillshumbut if you want to benchmark it, go ahead
18:17:18rasherSo trunk is slower than 4.0.3 it seems
18:17:39simjokoah, didn't get you at first :)
18:17:49 Join saratoga [0] (i=9803c6dd@rockbox/developer/saratoga)
18:17:55 Join faemir [0] (n=faemir@
18:18:17saratogaso trunk is roughly the same as 4.04 in everything tested
18:18:36saratogaAAC would be interesting too since its less well optimized and has less assembly then ogg/flac/mp3
18:18:54simjokodoes anyone know what went wrong when I encoded the file?
18:19:00simjokomplayer plays it fine
18:19:04saratogaalso i suppose bin sizes would be good to know
18:19:39 Quit timc (Read error: 110 (Connection timed out))
18:20:04 Part daurn
18:21:15markunja/en 誰よりもママを愛す
18:21:27markunoops, sorry
18:21:55 Join timc [0] (n=aoeu@
18:22:09obosaratoga: I've been trying to get the Views bootloader to run on the ARM emulator (FS #6524), but at the moment I can't even get the e200 firmware to run, it crashes when trying to write to the COP... any clues?
18:22:45saratogaobo: that emulator doesn't support the COP
18:23:02saratogai think all it could run was the bootloader, since that didn't use COP
18:23:48saratogaits not really all that close to complete, it was mostly just useful for looking at the very first hardware init stuff in the e200 bootloader if I recall correctly
18:25:00oboFor me it crashes just with the e200 bootloader
18:25:28saratogai was able to run it far enough to see the sandisk splash screen years ago
18:25:41 Quit merbanan (Remote closed the connection)
18:26:17Hillshumobo: OF or RB bootloader
18:26:31oboHillshum: OF
18:26:41 Join Lss [0] (
18:26:41 Join J-23 [0] (
18:27:43 Join moxfyre2 [0] (
18:27:58obosaratoga: well even hardware init would be good for me at the moment - I'm a little lost in a sea of function pointers in the disassembly, unable to see where functions are being called from, or what particular registers might be set to.
18:28:00 Join notlistening [0] (
18:28:25toffe82LambdaCalculus37: I wish I could, I will check if I have enough miles to pay the plane ticket ;)
18:29:45saratogathe binary in that zip works for me
18:29:58saratoga(found by randomly searching my vs build folder)
18:30:25obothanks :)
18:37:14 Join HellDragon [0] (
18:41:14 Quit JdGordon| (Ping timeout: 180 seconds)
18:41:41obosaratoga: yup, that one displays the boot splash for the e200, has a register overflow for the view, but neither give any log output - but thanks for digging it up.
18:44:20 Quit Hillshum ("Page closed")
18:45:07 Join bertrik [0] (
18:46:48rasherbertrik: hey, sorry about not paying attention before commenting on that patch
18:46:53 Join JdGordon| [0] (n=Miranda@nat/microsoft/x-d62c0c1a93b972c0)
18:47:23***Saving seen data "./dancer.seen"
18:47:40bertrikno problem, you were right about it anyway
18:50:19 Join akur [0] (n=akur@
18:52:20akurDoes anyone knows how is going that issue with ipod charching with Rockbox?
18:52:43 Join FlynDice [0] (n=FlynDice@
18:54:48Torneakur: we discussed this yesterday actually. there's a patch, and it's probably no worse than how we handle charging on any othe rplatform, but it's not been committed yet.
18:55:03Tornea decision wasn't noticably actually reached.
18:55:30akurso I still have to reboot into the OF for charging, right?
18:56:02Torneit does charge very slowly, as long as you are not using the disk or playing anything
18:56:08Torneslightly faster if the lcd is off
18:56:24Tornebut even in the best case it takes something over 12 hours to fully charge
18:56:30Torne(without the patch from FS #8802)
18:57:12 Quit Lss (Read error: 60 (Operation timed out))
18:57:31Torneadd yourself to the watch list for FS #8802 if you want to see when it gets fixed :)
18:58:07FlynDiceCan anyone tell me what AFE stands for. I've been seeing it in the as3525 datasheet and cannot deduce what it is so far.
18:58:21Torneif any? :)
18:58:29akurok thanks :)
18:58:52FlynDice00: SCLK and LRCK from I2SOUTIF (used if AFE sends data)
18:58:56TorneFlynDice: one I know is "analogue front end"
18:59:07FlynDiceThat makes sense thanks
18:59:18Tornesounds plausible, yah
18:59:23Tornethough not very specific :0
19:00:09bertrikFlynDice, part of the AS3525 is basically the same as the AS3514 codec/powermanagement/RTC used in the v1 sansas
19:00:35bertrikit handles the analogue stuff (and some digital like RTC too)
19:01:25FlynDicebertrik: yes, that's what I'm reading about but I couldn't find a reference to figure out what it stood for ;)
19:01:57 Part akur
19:03:06FlynDiceBack to the day job...
19:03:16 Quit FlynDice (Remote closed the connection)
19:07:46 Join slam_ [0] (
19:19:28 Quit PaulJam (Nick collision from services.)
19:19:35 Join PaulJam_ [0] (
19:20:57saratogawhats wrong with the c200 manual:
19:21:57agaffneylooks fine to me
19:22:07agaffneyeach of those links shows additional content
19:25:09 Join atrus [0] (
19:27:18saratogaits not fine
19:27:27saratogaunless you're seeing a lot more chapters then I
19:28:04gevaertsindeed it's not
19:29:57 Join jkl [0] (
19:30:01 Nick fxb__ is now known as fxb (
19:31:55Mikachuwhen i build it locally, it complains about the multirow package missing
19:32:00Mikachumaybe the same holds for the build server
19:32:20gevaertsno. It's a real problem
19:32:35 Quit HellDragon (Client Quit)
19:32:37gevaertsSome targets don't have \ActionTreeStop, and there's no nopt
19:33:30bertrikmarkun, I managed to read from the RTC over i2c on the meizu m3 last night
19:34:29LambdaCalculus37bertrik: Nice!
19:36:21 Join HellDragon [0] (
19:36:30bertrikthat should probably work for the other meizus too
19:36:45CIA-71New commit by gevaerts (r21513): c200 doesn't have \ActionTreeStop anymore since r21373, so don't try to use it.
19:37:01gevaertssaratoga: should be fixed tomorrow
19:37:04saratogagood to see the Meizu ports progressing again
19:37:42 Quit Zambezi (Remote closed the connection)
19:38:59bertrikyou can also thank gevaerts for lending me his m3 and markun for helping with the backlight
19:39:06gevaertssaratoga: indeed. Giving that m3 to bertrik seems to have been a good move :)
19:40:12bertrikthe other stuff like flash and USB will be much harder than this I think
19:40:26 Join n1s [0] (n=n1s@rockbox/developer/n1s)
19:40:45 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
19:40:54gevaertsUSB shouldn't be too hard I expect. It's the same controller as in the tccs
19:41:16gevaertsand while that driver isn't finished yet, it does mostly work
19:41:45bertrikbut we have only a bootloader running, I'm not sure if the kernel is even doing anything yet
19:42:42 Quit moxfyre2 ("Leaving")
19:42:54bertrikgevaerts, by the way, how does the startup by USB-plugin and shutdown by USB-plugout work anyway?
19:43:02 Join petur [0] (
19:43:23bertrikAFAIK there is a way to bring it in a state that requires a full battery drain, how do I get into this state (and how to avoid it)?
19:44:07gevaertslast time we got it into that state by interrupting OF-flashing halfway
19:44:39bertrikah ok, I'm usually very careful not to touch or move anything during flash
19:44:58gevaertsme too, but we had the required buttons wrong :)
19:45:17bluebrotherdomonoky: IMO FS #10380 is not a bug −− if the Kubuntu installation is messes up and returns the wrong bin path it's nothing we can do anything about. That need to work as otherwise it would be impossible to have multiple Qt installations on the same machine
19:46:56gevaertsbertrik: there's a GPIO pin that you have to set to make it keep running. Otherwise it just turns off after releasing the power ( button or unplugging USB
19:47:05 Join nibbler__ [0] (
19:47:53gevaertspossibly the PWRON pin (P1.3)
19:48:11bertrikok thanks
19:48:59 Quit Cory|XChat2 (Read error: 104 (Connection reset by peer))
19:49:03 Quit Cory` ()
19:49:21gevaertsThat's actually why we couldn't get anything to work at devcon2008. The code was fine, but we didn't do this and we unplugged...
19:50:38markunand didn't hold the play button (which would have worked as well)
19:54:01bertrikhm, so this delayed development for a year ? :)
19:54:18gevaertsnot a year. Just three or four months
19:54:19markunno :)
19:54:53 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
20:08:40 Quit flydutch ("/* empty */")
20:13:48 Join Hillshum [0] (i=cd7ae838@gateway/web/freenode/x-ed4a71c84e83f454)
20:17:23 Join taylor_ [0] (n=taylor@unaffiliated/n00b81)
20:18:54 Nick taylor_ is now known as n00b81 (n=taylor@unaffiliated/n00b81)
20:19:38CIA-71New commit by mcuelenaere (r21514): Update Lua manual page
20:20:28 Join Grahack [0] (
20:20:46CIA-71New commit by bagder (r21515): make the master run external scripts at certain events: ...
20:21:53 Part n00b81 ("Leaving")
20:22:44Bagderthe new system should be able to update the table before the entire round is complete
20:27:41 Quit Grahack ("Leaving.")
20:31:12 Join Thundercloud [0] (
20:33:33JdGordon|can it kill the run so a fix commit can get it going quickly if we see an obvious break?
20:34:17Mikachuthen you wouldn't see which commit broke the tree if two people commit close to eachother
20:34:52Hillshumless acountabilty! sweet!
20:35:11gevaertsI guess it can, but someone would have to push the button (I wouldn't do this automatically), so who?
20:35:27Bagderyeah, a KILLALL command should be easy to do
20:36:23Bagderbut I'm not sure we should use such
20:39:07Mikachuit would probably be more fun to push a test commit via git to the build system without having the commit clutter up the main branch history
20:39:07Bagderyeah, I've pondering on allowing "branch builds" somehow
20:39:07 Quit MarcGuay (Read error: 110 (Connection timed out))
20:39:22gevaertsyou need pretty solid svn cleanliness detection then, and some way to handle an unclean tree
20:39:43Bagderyeah, there are all sorts of problems involved
20:40:06Bagderbut for example trying out an FS patch with it would be rather neat
20:40:19markunBagder: FreeBSD's uname also doesn't support the -o option
20:40:50Bagderroom for improvement there then
20:41:33*LambdaCalculus37 should see if Mac OS X's uname supports the -o option as well
20:41:41Bagder"BUILD 21515"
20:42:46CIA-71New commit by bagder (r21516): append is .=, not = ;-) and some extra info on bad HELLOs
20:43:05 Join mirak [0] (
20:43:26LambdaCalculus37Bagder: No -o option on OS X's uname, either.
20:43:56Bagder99 builds not complete, 10 clients. 30 builds in progress
20:47:09BagderI'm ready to accept more build clients btw, if anyone sits waiting
20:47:09JdGordon|you can add mine... or where is instructions to add them?
20:47:09gevaertsthat sounds like a full run
20:47:09Bagdergevaerts: it is, I want to try building a table out of the results
20:47:26***Saving seen data "./dancer.seen"
20:47:41 Join No_Mo_Ro [0] (i=IIIeJI_B@
20:47:47No_Mo_Rohello all =)
20:47:54 Join Cory` [0] (
20:48:34JdGordon|hmm, whats the best way to have this script run on startup?
20:49:03gevaertsJdGordon|: an @reboot cronjob?
20:50:53Horschtinit script
20:51:17JdGordon|dont want it running as root though
20:51:22Mikachuman su
20:51:46Horschtman linux
20:51:49Mikachuor make the script +s
20:52:32 Quit Cory` (Client Quit)
20:52:47 Part No_Mo_Ro
20:53:02 Join Cory` [0] (
20:53:15 Quit Hillshum ("Page closed")
20:53:32*JdGordon| slaps Bagder for forcing curl :)
20:54:09JdGordon|ok, new client added... and building straight away
20:54:17Bagder"1 builds not complete, 12 clients. 11 builds in progress"
20:54:22*gevaerts thinks that depending on curl is the right decision
20:55:13*Mikachu has a feeling he will be compiling in vain from this point
20:55:47bertrikhm, my bogomips rating varies with frequency ... does actually make sense
20:55:48*gevaerts wants to see statistics :)
20:57:02Mikachu\o/ dead last
20:57:05Bagderit's... a bit surprising
20:57:22gevaertswow, cygwin managed an entire build!
20:57:30gevaertsHow long did it take?
20:57:34Mikachui am, or something else?
20:58:51BagderBuild Time: 514
20:58:56BagderBuild Type: recorder
20:59:19Bagderthat's the single cygwin build on rasher's
20:59:59gevaertsdon't give it an arm build :)
21:00:15JdGordon|or a leg!
21:00:37gevaertsHow long did the entire run take? I think we have most of the fast hosts in that list
21:00:58Bagderit took a depressing 784 seconds
21:01:43gevaertsmaybe we need another go with the new faster builders added
21:02:06CIA-71New commit by bagder (r21517): bump revision since we changed the uploaded file names
21:02:31rasherDid you do an update?
21:02:46bertrikShould we restart the client?
21:02:50Bagderit should now happen magically when I set 8 as a requirement in the server
21:03:01Mikachunothing happened yet
21:03:15Mikachunow it did
21:03:24Bagderyay, nice magic
21:03:32rasherEven worked on cygwin
21:07:16*JdGordon| notes his box isnt in the stats above... so OF COURSE its going to be dog slow :D
21:12:29amiconnBagder: Hmm, the build client doesn't do svn update?
21:12:40Bagderit doesn't? hehe
21:12:43 Quit Cory` ("Ex-Chat")
21:12:46Bagderthat's.. silly
21:13:06amiconnAt least I don't see any output which would indicate that it does
21:13:14MikachuHELLO 7 arm,sdl mikachu:choo fartmonstret i686 32 GNU/Linux 4169
21:13:14MikachuUpdate to 21517
21:13:16Mikachuisn't that it?
21:13:39amiconnI started my client and got builds right away
21:13:43 Quit simjoko ("KVIrc Insomnia 4.0.0, revision: , sources date: 20090520, built on: 2009/06/06 12:36:26 UTC")
21:13:49MikachuChecked out revision 21516.
21:13:51amiconnThe tree is unmodified, but not up to date
21:13:54Mikachusvn info says 21517 though
21:14:17Bagderyes, it does svn up
21:14:27Mikachubut maybe not always?
21:14:37Bagderunconditionally actually
21:14:44Bagderline 255 in
21:14:46JdGordon|I assume it has no problems with clients dropping off mid build yeah?
21:14:51BagderJdGordon: correct
21:14:55amiconnHmm, it is up to date
21:15:04amiconnI would have expected some output from the script
21:15:06BagderJdGordon: clients may come and go
21:15:21JdGordon|duplicate client name?
21:15:33BagderJdGordon: will be rejected if from the same user
21:15:37*JdGordon| slaps self
21:15:37Mikachumy gigabeats build is up to 5 minutes now
21:15:50rasheramiconn: it should be in the log at least
21:15:56rasherWhich will be gone by now
21:15:57Mikachuwoo, it finished
21:16:37MikachuBuild Status: Failed
21:16:42Mikachubut it still uploaded the zip file
21:16:49JdGordon|Bagder: it doesnt like client reconnecting quickly?
21:17:01BagderJdGordon: so its the same client?
21:17:05Bagderit could be a bug
21:17:30MikachuERROR: /usr/local/ipod-chain/lib/gcc/arm-elf/4.0.3/libgcc.a(_udivsi3.o) uses FPA instructions, whereas /tmp/a/rockbox/build-31732/apps/plugins/midi/midi.elf does not
21:17:35JdGordon|arg... does runclient spawn new proc's or something?
21:17:36*Mikachu has no idea what that means
21:17:54*amiconn added two clients to the mix, one on his linux box (amd64 bit, capable of building everything) and one in a linux vm (32 bit, all except mipsel)
21:17:55BagderJdGordon: yes, it forks to build in a separate process
21:18:13MikachuBagder: why does it upload failed builds? :)
21:18:25rasherMikachu: To get the log
21:18:32Mikachuit even prints "no 'rockbox.gigabeat'"
21:18:32Bagderyes, but the zip...
21:18:56rasherMaybe failed builds could be retried on another host?
21:19:00Bagderthe client shouldn't do that
21:19:19Bagderrasher: we've discussed that, but I think not
21:19:45JdGordon|so what do I need to kill to kill the client completly?
21:20:08MikachuRemoving /tmp/a/rockbox/build-5236
21:20:11Mikachubut that dir still exists
21:20:20Mikachuas does another one
21:20:21Bagderyeah, that's a bug in the client
21:20:33Bagderyou'll gain bonus points if you find and fix it ;-)
21:20:43Mikachui had a vague memory you said you fixed it, but obviously not then
21:20:58Bagderzagor thought he fixed it, but it proved to be wrong
21:21:16Bagderthen he took off to Greece to escape from it!
21:21:27Mikachumaybe if you redirect the rm output to a log file
21:23:15 Join kachna|lappy [0] (
21:23:52rasherI saw "cannot remove directory for /home/Jonas/rockbox/build-3456/apps/lang: Directory not empty at line 521" (perhaps output is different in cygwin)
21:24:11Bagderseems related
21:24:21amiconn"0-length pipe msg from 4!" ??
21:25:41saratogaany reason the directory file limit is so small on flash targets like the e200v1?
21:25:57saratogai think we could safely make it much larger and save users the occasional error message with no ill effects
21:26:09amiconnAfaik it's 1000 on all swcodec targets
21:26:12saratogamanual says it defaults to 400
21:26:22amiconnThe manual is outdated on this
21:26:32amiconn400 is a sane default for lowmem targets
21:26:41*amiconn sets it even lower on most targets
21:27:15JdGordon|I got that 0-length pipe message also... but I was tryign to kill the client so it could just be pebkac here...
21:28:31Bagderit's a sign of a problem too
21:28:59JdGordon|also... not to nitpick or anything, but shouldnt the args for the .pl be using −− instead of - ?
21:29:07 Join moxfyre2 [0] (
21:29:15Bagderwe just took the easy route of using perl -s
21:29:37Bagderit wants the arguments like that
21:29:43JdGordon|ok, fair enough
21:29:46*amiconn got this 0-length pipe message 3 times already
21:29:58Bagderyes, it's common
21:29:58JdGordon|start another run!
21:30:19BagderI get it often after a kill
21:30:31Bagderor perhaps only then
21:31:03amiconnAh yes. Killed build... removing.... 0-length...
21:31:45Bagderactually, I think we can just ditch that output
21:32:02*amiconn wonders how the build client copes with time deviation
21:32:18Bagderthe client doesn't care about time
21:33:13 Quit jkl (Read error: 113 (No route to host))
21:33:23rasherBagder: Why not retry failed builds? I see that it might make "red lines" slower, but failed builds often fail quickly anyway, in my experience
21:33:28 Quit tvelocity[a] (Remote closed the connection)
21:33:44saratogacan I change that 400 to 1000 in the manual or will that screw up LOWMEM?
21:33:57Bagderwell, most failed builds are actually true build errors independent of the build client
21:34:16Bagderand not all red builds are fast
21:34:21Bagderbasically those reasons
21:34:46amiconnAre there client stats available somewhere?
21:34:59Bagderno, things aren't yet that stable
21:35:03 Join Blue_Dude [0] (
21:35:31 Join jkl [0] (
21:35:46saratogahaving failed builds reported quickly would be nice, nothing more annoying then waiting around the build screen for 10 minutes waiting to see if you broke something
21:35:53saratogaretrying them would be annoying
21:36:08BagderI'm planning to get the build table updated while the build round is in progress
21:36:58amiconnDoes the client reconnect properly when the server is restarted?
21:37:34saratogawho should i talk to about the manual?
21:41:06Mikachudidn't gevaerts fix the c200 thing after you mentioned it?
21:41:22gevaertsthat was an easy one :)
21:41:24LambdaCalculus37saratoga: Me, AlexP or pixelma.
21:41:55CIA-71New commit by bagder (r21518): upped the client req, changed the check for dupes since the previous way ...
21:42:14 Join simjoko [0] (
21:42:16saratogaLambdaCalculus37: is there someway to make a number conditional on how much memory a target has?
21:42:51LambdaCalculus37saratoga: Not really sure. There may be a way. What is it you want to do?
21:43:28saratogachange the directory limits to be 1000 on bigmem targets
21:43:33saratogaright now it says 400 for everything
21:44:11JdGordon|isnt the usual way to add a line in features.txt?
21:44:25JdGordon|or is that lang only?
21:44:33LambdaCalculus37JdGordon|: I think that's lang only.
21:44:56gevaertsit works for the manual as well I think, but you tend to end up with lots of features
21:44:59rasherPretty sure the manual can use features
21:45:01LloreanWhy exactly do we need that limit? (Or rather, why with that limit in place will it refuse to show the whole folder, rather than spinning up to read the rest of it if you scroll too far?)
21:45:28saratogait'll error out if you try to open a folder above the limit
21:45:56LloreanYes, but why?
21:46:16LloreanIs it impossible to load part of the list, then if we need some filename not in the list, load it?
21:46:16saratogapresumably the memory to store the directory contents from the FAT is statically allocated
21:46:46Bagderyes, we have no code for reading "the next part"
21:46:48simjokoI don't seem to be able to create an aac file rockbox will play. I tried faac and the nero linux encoder. Do I miss sth.?
21:46:53JdGordon|no, I tihnk the issue is in the list handling part..
21:46:56amiconnLlorean: You can't display the dir partially because of sorting
21:47:14Bagderwell, you _could_, it just would be confusing
21:48:00saratogaencountering the error could be made to increase the limit I suppose
21:48:11Lloreanamiconn: Clearly sorting doesn't take *that* long or browsing would be really slow. if we had a limit of 400 names, and we had 500 entries, couldn't we just sort and show the first 400, then if someone wanted the list to show #401, re-do the sort and go ~200 in each direction from where the user triggered the re-sort?
21:48:43amiconnYou don't know which are the first 400 until you sort all of them
21:48:53LloreanYes, but you don't need to keep *all* of them while sorting.
21:49:00JdGordon|Llorean: they are sorted in a statically sized buffer
21:49:02amiconnyou do
21:49:10saratogayou can't sort without having every element
21:49:29JdGordon|it could be done *very* slowly without any sort buffer...
21:49:35gevaertsLlorean: you don't, but it would get slow
21:49:35LloreanIf I can hold only 15 cards in my hand, I can still make sure I end up with 4 aces, 4 2s, 4 3s, and 3 4s after going through the whole deck.
21:50:11rasherSlow is better than not doing it at all. But might not be worth it, in terms of complexity
21:50:17gevaertsnot entirely the same. What if you don't know how many aces there are?
21:50:41Lloreangevaerts: I always keep the lowest numbered cards I find, so if I have a full hand, and draw a 3, I discard my current 5.
21:51:04 Quit bmbl ("Bye!")
21:51:08gevaertstrue. yes, you can do it, but it will be horribly slow
21:51:09 Quit Blue_Dude_ (Connection timed out)
21:51:19JdGordon|you could do it in the list filename callback, it would be crazy slow thouhgh and a bitch to make 100% correct
21:51:23JdGordon|but it could be done
21:51:23Llorean"slow" is better than "these 1200 files on your disk are unusable now"
21:51:24gevaertsEverytime you scroll down you have to spin up again
21:51:40amiconnLlorean: That means comparing the current card with *every* card in your hand
21:51:49Lloreanamiconn: Not if your hand is already sorted
21:51:59saratogai doubt the FAT is sorted in some way we care about
21:52:20Mikachuwhy not just show the entries unsorted?
21:52:30amiconnThe directories on fat are just in order of file/dir creation
21:52:34rasherMikachu: that could work
21:52:43rasherWith a splash saying why
21:52:50amiconn...with new ones intermixed if there were gaps from previous deletions
21:53:14LloreanOn a related note - is dircache limited in the same way?
21:53:35amiconnAlso you cannot just seek N entries forward because each vfat entry takes a varying number of actual directory entries
21:54:05*amiconn doesn't see why we would need to change anything
21:54:45LloreanBecause the current splash doesn't help a user much at all, and if they don't have a manual on hand they're basically stuck.
21:54:54LloreanToo many people seem to like the "put all my files in one folder" solution.
21:54:56amiconnThe 400 (resp. 1000) entry limit is just the default - you can increase or decrease it, depending on how much memory you want to set aside for large dirs
21:55:00rasherWhy not? Surely showing the files in some way (very slowly sorted, or quickly unsorted) is better than not doing it
21:55:15amiconnRockbox will display a splash if there are more entries than the current limit
21:55:34 Join tessarakt [0] (
21:55:36rasheramiconn: you're way overestimating users
21:55:39Lloreanamiconn: Yes, but that doesn't tell the user the limit can be changed, or where to find it.
21:55:49LloreanLimits are often enough hard coded in these players.
21:55:56LloreanLike the 9999 song limit on H100 was, for example
21:55:57amiconnA simple case of rtfm...
21:56:47LloreanThe manual isn't always available (or right, for that matter). For some things it can be expected as necessary, but that doesn't mean you shouldn't try to solve its necessity where possible.
21:57:06LloreanIf someone doesn't know that there's a setting for changing the limit, how are they going to know to look for it in the manual?
21:58:43 Quit LambdaCalculus37 ()
22:00:41 Quit nibbler__ (Read error: 113 (No route to host))
22:07:12 Quit Bagder ("*plopp*")
22:07:50rasheramiconn seems to be living in a world where everyone reads the manual front to back before installing Rockbox
22:08:06 Join Bagder [241] (n=daniel@rockbox/developer/bagder)
22:08:29 Quit simjoko (Read error: 110 (Connection timed out))
22:08:39saratogagreping i don't see that error message printed
22:08:47saratogais it in the lang file maybe?
22:08:56Lloreanrasher: While I've always been in the "users need to RTFM first" camp, there is a real problem when users may not realize it's even a question to ask.
22:09:42rasherLlorean: We can be in that camp all we like, but at the end of the day, we'll just be pissing off users for no really good reason
22:10:13LloreanWell if it's something like "How do I change themes?" or similar - an actual question then I think it's okay.
22:10:14saratogalets just improve the error message . . .
22:10:16 Quit Blue_Dude (Read error: 110 (Connection timed out))
22:10:51Lloreansaratoga: Ideally I'd like error messages like that not to be splashes, but a prompt screen: "File buffer is set too low. Press select to go to the setting, any other key to cancel"
22:14:26saratogawell if i could find the error screen i'd see what can be done
22:14:47 Join _lifeless [0] (n=lifeless@
22:16:47bertrikgevaerts, your meizu m3 is an 'old' model, that means it uses an uda1380 codec, right?
22:17:34gevaertsbertrik: I'd suspect that, but I'm not sure
22:17:49 Join Hillshum [0] (i=cd7ae838@gateway/web/freenode/x-f93fd1d094d13c02)
22:18:29saratogahuh in the code the max for lowmem is 200
22:18:34markunwe should have checked when it was open
22:18:35saratogaso the number in the manual is wrong for all targets
22:18:55bertrikdidn't scorche take a picture of it?
22:19:02bertrikoh that was the sansa e200v2
22:19:04saratogaanyone mind if I change it to 1000 since thats more correct for most targets?
22:19:51 Quit bluebrother ("leaving")
22:22:06bertrikhm, reading the exernal rtc on the meizu m3 seems to work fine, but I can't get anything from the codec
22:22:16bertrikthe rtc returns bitswapped values :/
22:22:37saratogawell i'm going to up the limit for flash targets with lots of memory since theres no harm in increasing the default
22:22:52bertrikit could be my i2c driver, reading the codec works a bit different (with a repeated start). Repeated start is not mentioned in the data sheet
22:24:36 Join Blue_Dude [0] (
22:25:59amiconnsaratoga: There's potential harm in increasing static allocations on all targets. Not necessarily for music playback on flash targets, but for several plugins
22:26:28saratogaamiconn: can you elaborate?
22:27:22amiconnThere are plugins which need to load large files completely into memory, e.g. the jpeg viewer
22:27:46saratogahow much memory does one file entry take?
22:28:17amiconnfile size + decoded size(s) (as yuv bitmaps)
22:29:01saratogasorry i meant for the directory size, looks like 40 bytes
22:29:17 Quit Hillshum ("Page closed")
22:29:39saratogai think increasing it by 1000 on flash targets with lots of memory should be pretty safe, thats only a 40KB change in bin size
22:29:52amiconnIt's almost certainly more than just 40 bytes
22:29:56saratogaand of course users are free to change it if they need more
22:31:14saratogaplaylist->buffer_size = AVERAGE_FILENAME_LENGTH * global_settings.max_files_in_dir;
22:31:52saratogaso really that setting isn't declaring how many files are available, but actually how many 40 byte slots there are (which i guess allows for files that span slots?)
22:31:54 Join shotofadds [0] (n=rob@rockbox/developer/shotofadds)
22:32:22amiconnThere's also struct entry
22:33:29saratogaand another for the database
22:33:39 Quit parafin ("So long and thanks for all the fish")
22:33:45 Quit linuxstb (Read error: 110 (Connection timed out))
22:33:45 Join parafin [0] (
22:34:35rasherSo you could have a dir of 300 files that didn't fit in what's supposedly a 400-file buffer?
22:34:36amiconnstruct entry is 12 bytes on a 32 bit target (i.e. all) - it's not packed
22:34:43saratogayes i think so
22:34:57amiconnYes, if all filenames in there are very long
22:35:02saratogathe error message is "Dir Buffer is Full!" which is a poor message IMO
22:36:34saratogaperhaps we could change it to "Directory is too large, see General Settings"
22:36:51amiconnWell, if you have a better idea that is still short enough...
22:36:58saratogahow much space do i have?
22:37:18amiconnIt's a splash. Available space depends on display size and font
22:37:28amiconnWorst case is 22 chars of course
22:37:50rasherI'm pretty sure there are plenty splashes >22 chars
22:38:42amiconnYes, and those are pretty bad on the Player. Translations are often even worse than English in that respect
22:39:39LloreanMaybe "File Limit set low"
22:39:45LloreanShort, but suggests it's a setting.
22:40:00LloreanOr perhaps "too low" rather than just "low"
22:40:00amiconnThere's this idea of having longer splash texts auto-scroll on the Player (only works for timed splashes, obviously)
22:40:11saratogahow about Err:File Limit set low
22:40:17saratoga22 chars and explains the problem
22:40:39rasherFejl: Filgrænse sat lavt
22:40:41rasher24 already
22:40:42amiconnsplash() word-wraps
22:40:44LloreanNot ideal by any stretch, but a whole lot better.
22:41:04saratogadrop the space like i did and you're only 1 over
22:41:17*rasher claims it's impossible to do all splashes in 22 characters and make sense
22:41:18saratogai saw some messages that long in the source anyway
22:41:46saratogawell i have to go, but if someone wants to fix it, the define in lang is LANG_SHOWDIR_BUFFER_FULL
22:41:58 Quit slam_ (Read error: 60 (Operation timed out))
22:41:59shotofaddsUnhelpful (if you're here): r21345 results in garbled JPEG AA on my target (Cowon D2, ARMv5). Seems fine in the sim, and the JPEG viewer on target is fine too. Looks like one of the optimisations has broken something...
22:42:09amiconnThere's nothing that stops us from having different texts for different targets, but it increases the work for translators if not used carefully
22:42:33rasheramiconn: That still leaves the problem of actually coming up with something that is less than 22 characters
22:43:08amiconnOr that scrolling splash idea...
22:43:21rasherThat sounds much better
22:46:04rasherWould be interesting if someone with a lot of time on his hands figured out length-limits for language IDs (on player)
22:46:46rasherI'm sure there are a lot of translations that break these limits
22:47:29***Saving seen data "./dancer.seen"
22:47:49 Join flydutch [0] (
22:48:15 Join goffa_ [0] (n=goffa@
22:48:21 Quit Blue_Dude ("ChatZilla 0.9.85 [Firefox 3.0.11/2009060215]")
22:48:28amiconnMe too. The German translation does that for sure
22:48:54 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
22:51:43gevaertsBagder: curl: (26) read function returned funny value
22:52:14Bagdernow that is an error message!
22:52:51*rasher blames curl developer
22:56:45Bagder20090625-20:52:06 Cancel: build fuze client bertrik-laptop-user seconds 138
22:56:49Bagderthat hurts
22:57:38rasherClient disappeared?
22:57:49Bagder20090625-20:54:19 Cancel: build sansae200v2 client fartmonstret-mikachu seconds 271
22:57:59Bagderno, it's a killed build after that many seconds
22:58:08Mikachuall my builds were probably broken anyway
23:00:01rasherBagder: what happens if a client connects with an old version?
23:00:11Bagderit gets told to update
23:00:35Bagderand isn't allowed to party
23:00:47 Quit n1s ("Lmnar")
23:00:52 Quit goffa__ (Read error: 110 (Connection timed out))
23:02:15*rasher is now running a version 5 client
23:02:43Bagdernow you don't ;-)
23:03:13Bagder"Client disconnect (asked to update), removing client maia-rasher"
23:03:20amiconnrasher: Your bogomips fix offsets the bogomips for all clients
23:03:38Bagderah yes it only adds
23:03:42rasheramiconn: doesn't matter
23:04:04rasherThe order will be the same
23:05:02 Nick fxb is now known as fxb__ (
23:06:33gevaertsBagder: `make -k -j$c $log`; runs the thing in a subshell, right?
23:07:16gevaertsI've been doing some tests here, and it looks to me as if a make running in a shell that gets killed with -9 keeps going
23:07:44 Join Sajber^ [0] (
23:07:55gevaertswhich would explain the directories not being empty, as they keep getting filled while rmtree is busy
23:08:27Bagder is the first speed chart
23:08:40Bagderof the latest 53 builds round
23:09:41rasherI wonder if scoring builds by "buildtime per bogomips" or something like that would rank builds usefully
23:10:35Bagderpossibly. Me and zagor discussed using a single machine and do ccache-less builds to rank them
23:11:48rasherProbably more reliable, but less magic
23:12:56Mikachu"but"? :)
23:13:20*rasher likes magic
23:13:34rasherMy way would mean new builds got auto-ranked!
23:14:46JdGordon|Bagder: is this output expected?
23:15:06BagderI see that too
23:15:28rasherLooks familiar
23:15:52Bagderthe "Starting client" part shows that it restarted
23:16:58*gevaerts tries to think of a reliable way to kill make
23:17:14Bagder is a log of the 53-builds build round
23:17:14rasherwhile true; do killall make; done ?
23:17:28Mikachuwhile killall make; do true; done
23:17:47*gevaerts tries to think of a reliable way to kill exactly one carefully selected make process
23:19:14lilltigerps -A
23:19:17lilltigerfind it
23:19:24lilltigerthen kill it
23:19:27BagderI have an idea on how to hand out builds in a way that might be more favorable to the slowest machines
23:19:28Mikachukill <tab>
23:19:30*gevaerts tries to think of a reliable way to kill exactly one carefully selected make process from within the build script
23:19:34lilltigerlike kill pid
23:19:58lilltigernever seen the build script :p
23:21:17 Join fyre^OS [0] (
23:21:30 Quit martian67 ("out")
23:21:43Bagderthe last build is fun
23:21:50Bagder13 builds were killed
23:21:54 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
23:22:25gevaertsBagder: I think you just need to use the negative of the pid
23:22:32 Quit petur ("Zzzzz")
23:22:42gevaertsThat should kill the process group
23:23:48*gevaerts reads up a bit more on this
23:23:57Bagderit's easy to test of course
23:24:12BagderI mean that it at least doesn't make things worse
23:24:27gevaertsit worked in my synthetic perl-less test
23:24:54 Part evilnick
23:25:28 Join evilnick [0] (i=0c140464@gateway/web/freenode/x-17c618ecfed61fb8)
23:26:41gevaertsI suspect that you're supposed to create a process group first, and that the fact that it works without that on linux may be a coincidence
23:27:05 Quit jgarvey ("Leaving")
23:28:05gevaertsif that's the case, we'd need a setpgrp; just after the fork
23:29:00JdGordon|Bagder: is something funny happening around line 135 of that last paste?
23:29:22JdGordon|jakorasia-cg gets 2 builds immediatly on connect.... why not just one?
23:29:41JdGordon|ditto at line 180 also
23:29:44notlisteningwho maintains the configure script specifically the voice element of it?
23:29:45 Quit mt (Read error: 104 (Connection reset by peer))
23:30:08BagderJdGordon: all clients get 3 builds actually
23:30:10 Quit Rob2223 ()
23:30:50rashernotlistening: I don't think there's a maintainer as such, but I wrote most of it
23:30:53 Join Rob2222 [0] (
23:31:01Bagderit needs more than one build to be able to start on the next while uploading the previous
23:31:04gevaertsBagder: I think should do it (untested)
23:31:22rasherBagder: couldn't it just ask for a new one before uploading?
23:31:59Bagderwe have no way of asking for builds actually ;-)
23:32:00notlisteningrasher, thanks would it be okay to try and intergrate the SAPI engiine i am working on under linuc and can i come groweling to you for help?
23:32:49rasherSure, give it your best. I'll try to help. The most interesting bits will probably be in I imagine
23:33:39notlisteningrasher, thanks will do my bed ;)
23:33:47 Join mt [0] (n=MTee@rockbox/developer/mt)
23:34:08gevaertsBagder: I've now started a client with that one, so if you fire off a new round we'll see if it helps
23:34:14Bagdercurl | patch -p0
23:34:30mtlinuxstb: What's the problem with using id3v2buf instead of toc ?
23:34:31notlistening* best - with my typing skills it might not be good enough
23:34:42*amiconn added another client
23:34:54amiconnCygwin 1.7...
23:38:19 Quit fyrestorm (Connection timed out)
23:38:37 Quit flydutch ("/* empty */")
23:40:39 Join safetydan [0] (n=deverton@rockbox/developer/safetydan)
23:40:57pixelmais it worthwhile to add my vm on the laptop (same as during DevCon) back to the server farm?
23:41:46 Quit moxfyre2 ("Leaving")
23:42:42bertrik\o/ markun helped me with getting the meizu m3 to talk to the codec so we can now access the codec!
23:43:15evilnickWhat's the protocol for deleting a forum reply that adds nothing to a post: ?
23:43:25evilnickShould I PM the user and then delete?
23:43:52 Quit fyre^OS (Read error: 110 (Connection timed out))
23:44:14JdGordon|you linked the thread not the post...
23:44:33linuxstbmt: It's not really a problem, it's just that you will then be leaving toc unused.
23:44:38evilnickSorry, the very last reply is only echoing the second to last post
23:44:51gevaertsBagder: according to, the signal needs to be negative, not the pid...
23:45:42*gevaerts wrongly assumed that it would behave the same as kill(2) on linux
23:46:16*rasher has some unfavourable opinions about Perl to share
23:47:40mtlinuxstb : And if I use toc, I will be leaving id3v2buf unused ;)
23:48:07mtBut anyway, I'm trying to focus on getting the code to work reliably on taget first
23:48:30mtI think this could be done as part of the
23:48:41mt* as part of the optimizations
23:49:18rashergevaerts: that's reversed
23:49:42gevaertsok, so curl|patch -p0 -R :)
23:49:43Bagderyeah, it confused me for a second too
23:50:04*gevaerts did that on purpose
23:50:05rasherMine updated with this change
23:51:11mtBy the way, can I use splashes in a metadata parser ?
23:55:27pixelmaso - no interest in another test machine which is a bit more on the slow side (upload speeds also, except for small files)?
23:55:40rasherpixelma: The more the merrier
23:55:47gevaertspixelma: it's bound to be faster than our cygwin friends :)
23:55:59*rasher isn't running his cygwin client
23:56:06pixelmaI could add one of those too...
23:56:24mtlinuxstb: how would you go about debugging this codec on target ? (I think the bug is in the parser btw)
23:56:50linuxstbmt: Isn't the id3v2buf used for the other metadata?
23:57:22mtlinuxstb : No. v1buf is the one used for metadata.
23:58:08linuxstbMaybe using logf over usb would help. Is your device the e200?
23:58:19*pixelma needs to update the build script though , most probably
23:58:56mtlinuxstb: Yes.

Previous day | Next day