#rockbox 2012-07-23

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 2012-07-23

00:00:03		Join gxk [0] (~gxk@bzq-79-180-216-196.red.bezeqint.net)
00:04:52		Quit prof_wolfff (Ping timeout: 264 seconds)
00:05:35		Quit petur (Remote host closed the connection)
00:06:53	bertrik	freqmod, actually I think I'm missing everything from libopus/ogg
00:08:00		Quit rasher (Ping timeout: 248 seconds)
00:09:29		Quit kevku (Quit: KVIrc 4.2.0 Equilibrium http://www.kvirc.net/)
00:12:47		Quit ender` (Quit: War is a way of shattering to pieces, or pouring into the stratosphere, or sinking in the depths of the sea, materials which might otherwise be used to make the masses too comfortable, and hence, in the long run, too intelligent. -- George Orwell: 1984)
00:13:59	[Saint]	rocktriangle: bookmark on stop
00:15:18	[Saint]	Or, perhaps I'm misunderstanding. It might be best if you were to rephrase.
00:15:35		Join rasher [0] (~rasher@rockbox/developer/rasher)
00:17:43		Quit Rower85 (Quit: Hmmm...)
00:18:18		Join mystica555 [0] (~Mike@63-227-11-98.hlrn.qwest.net)
00:18:54	*	[Saint] hums.
00:19:57	[Saint]	I expected tonfind bookmarking settings in playback settings.
00:20:05	[Saint]	*to find
00:21:22	[Saint]	I guess the line between playback/general is a blurry one.
00:24:41		Quit Wardo (Quit: Blarglarg)
00:27:55		Quit XavierGr ()
00:31:27	rocktriangle	Alright then
00:31:29	rocktriangle	I'll try that
00:31:30	rocktriangle	Thank you
00:32:03	rocktriangle	By the way, I also got a warning, who could I go to to have it deciphered? At the top it said PANIC
00:32:05		Quit gxk (Ping timeout: 245 seconds)
00:32:59	bertrik	you could post it on the bug tracker
00:33:15	bertrik	there are several people here that can decipher it
00:33:54	gmaxwell	freqmod: if you really only have 9k of stack, then you're almost certantly going to need to go the NONTHREADSAFE_PSEUDOSTACK route, and stub in the parts required to get some heap to that global.
00:34:35	gevaerts	only? That's one of the larger stacks in rockbox!
00:35:50	gmaxwell	freqmod: Opus is recursive. And while I don't doubt that there are plenty of low hanging fruit to cut down on the stack usage, the PSEUDOSTACK switches almost all of it to heap.
00:35:53	bertrik	are we really that tight for memory that we can only afford 9k for stack?
00:36:16	gevaerts	We can increase it if really needed of course
00:36:32	freqmod	is it a problem to just create another thread with larger stack?
00:36:38	freqmod	with thread safty etc
00:36:46	Torne	the number of threads is hardcoded
00:36:48	gmaxwell	(downside being that it makes the codec totally non-threadsafe)
00:37:07	Torne	creating another one means there's another thread, with stack, in existence all the time
00:37:07	gevaerts	I don't think thread safety is an issue here
00:37:11	Torne	not just when that codec isbeing used
00:37:25	gmaxwell	(well, and its less well tested, because we've only really ever used it for the c55 DSPs which don't even work at the moment)
00:37:34	freqmod	i am going to a cabin without internet for the next week
00:38:01	gevaerts	gmaxwell: how much stack do you expect is needed normally?
00:38:18	freqmod	so i may try to fix something of it, but we are mostly trying to work to get the cabin freshly painted etc.
00:38:40	gevaerts	I mean, if it's 16K, we can increase it without too much discussion, if it's 128K, that's pretty much a dealbreaker
00:39:01	bertrik	freqmod, can you push the missing ogg stuff I mentioned?
00:39:14	freqmod	how much memory is available for codec heap?
00:39:16	gevaerts	Torne: are you sure about that?
00:40:08	gmaxwell	Probably more like 20-30k. Been a while since I checked. Certantly not 128k.
00:40:17	Torne	gevaerts: er, i thought so.
00:40:18	freqmod	bertrik: pushed
00:40:22	bertrik	thanks
00:40:27	Torne	gevaerts: do we allocate them with buflib now?
00:40:30	freqmod	and i have also made enough parsing so the codec starts on target
00:40:37	gmaxwell	(also I was mostly lookingat encode and decode; and you only care about decode at the moment)
00:40:38	freqmod	(and produces stack errors)
00:41:08	gevaerts	Torne: we have a fixed (maximum) number of threads, but IIRC for plugins and codecs those are more or less dynamic
00:41:13	amiconn	Torne. gevaerts: The maximum number of threads is hardcoded, the number of running threads is not
00:41:53	amiconn	There are a few spare slots, so codecs and plugins can start a background thread if necessary (e.g. mpa.codec does this on pp)
00:41:54	Torne	amiconn: sure, i didn't mean they were literally running all the time
00:42:02	Torne	but the stacks exist
00:42:05	*	freqmod read something in the codecs.h about spawing threads with the stack as a parameter
00:42:08	Torne	and their context
00:42:12	bertrik	I wonder if we can allocate the stacks with buflib now
00:42:13	amiconn	No they don't
00:42:17	gevaerts	Torne: the stack will be allocated in the codec buffer somewhere
00:42:23	amiconn	Only the context stores exist all the time
00:42:46	*	freqmod are going to sleep
00:43:15	amiconn	The main disadvantage with an extra thread is that the stack either cuts down codec iram, or is located in dram, which is a disadvantage as the main codec thread's stack is in iram already
00:43:43	gevaerts	freqmod: check CODEC_SIZE for the total (code+heap) size you get
00:44:44	gevaerts	I'd say that if this NONTHREADSAFE_PSEUDOSTACK works, it's the best solution
00:45:13	amiconn	While from this point it doesn't matter whether you create an extra thread with big iram stack or allocate them in iram another way (whatever that #define is), the latter is the easier way as you don't need an extra thread
00:46:35	gevaerts	Right, codec_stack is IRAM, so yes, increasing it is tricky
00:46:37	amiconn	A thread context isn't that large in rockbox, thanks to cooperative threading
00:47:02	amiconn	(meaning we don't have to save scratch regs)
00:49:24	gevaerts	Do we have numbers somewhere on how much IRAM we have left on various targets?
00:49:45	bertrik	freqmod, pretty cool to see it working in the sim, nice job!
00:50:29	gmaxwell	In any case if there are bugs w/ NONTHREADSAFE_PSEUDOSTACK, I'll fix them if you find them. We specifically put that in for memory limited platforms that didn't have allocators or multithreaded decode.
00:51:17		Join gxk [0] (~gxk@bzq-79-180-216-196.red.bezeqint.net)
00:55:06	saratoga	gevaerts: http://www.rockbox.org/wiki/CodecMemoryUsage
00:56:45	gmaxwell	HE AAC using 500k ?!@#! well, okay I'm no longer worried about opus.
00:57:22	saratoga	AAC HE uses a lot of memory
00:57:30	saratoga	also, libfaad is not very efficient
00:57:47	saratoga	although we've patched it up a lot
00:58:18	saratoga	aac in general is a pretty bloated format
00:58:37	saratoga	i guess design by committee and all that
01:00
01:00:57	gevaerts	It's mainly the IRAM bit that's really constrained here. We can't increase codec_stack much (if at all, I guess) without moving it to slower RAM (for every codec)
01:01:39		Join Scromple [0] (~Simon@119.225.209.134)
01:01:59		Quit gxk (Ping timeout: 255 seconds)
01:02:04	bertrik	we don't really sync codecs often with upstream, do we?
01:05:41	saratoga	bertrik: do you mind if I just disable the backtrace on AMS for now?
01:06:14	bertrik	no, I don't mind, seems to be giving more problems than solutions currently
01:11:03	saratoga	ok i put a FIXME comment in with the disable, hopefully the backtrace can be made to work better with AMS in the future
01:12:06	saratoga	on arm the panic screen calls "disable_interrupt(IRQ_FIQ_STATUS);" but never reenables interrupts, maybe we could enable them at the end just to be safe?
01:13:03	bertrik	I don't see why we need interrupts at that point anymore
01:13:47	bertrik	the only thing we do is wait for a button (and I suppose we can do that without interrupts)
01:14:38	saratoga	since the 30 second hold doesn't reboot the power, something is not getting enabled
01:14:48	saratoga	you think interupts aren't it?
01:15:01		Join gxk [0] (~gxk@bzq-79-180-216-196.red.bezeqint.net)
01:15:35	saratoga	oh crap
01:15:38	bertrik	luckily I haven't been in that situation myself yet, and I don't know exactly what happens
01:15:38	saratoga	its the ROM clock
01:15:48	bertrik	oh, I thought we re-enabled that already
01:15:50	saratoga	the ROM clock is off normally, and we turn it on during reboot
01:16:01	saratoga	so if you don't make it through reboot, you have no clock
01:16:08	saratoga	and hence no hold power to reboot
01:17:24	saratoga	I can #ifdef the ROM clock enable into the start of the abort handler (which is a bit ugly but safe), or just leave the clock on all the time
01:17:27	saratoga	what do you think
01:17:35	saratoga	maybe we should just leave it on to be safe ...
01:18:19	bertrik	I find it peculiar that we somehow seem to be able to disable the hardware power-down mechanism
01:18:56	bertrik	yes, I'd rather leave it on (and waste a miniscule amount of battery runtime) if it helps us from the current situation
01:18:58	saratoga	clearly the "reboot" doesn't fully reset the SOC's state
01:19:03	saratoga	ok i'll do tht
01:20:44	bertrik	which is a hardware design flaw IMHO
01:20:50		Quit gxk (Ping timeout: 245 seconds)
01:21:26	bertrik	IIRC, we allow the watchdog to reset us, can't get much cleaner than that I think
01:23:30	saratoga	yeah its bad design
01:24:00	CIA-5	Commit b61b14f in rockbox by Michael Giacomelli: Revert my previous commit disabling the internal ROM clock during normal
01:26:51	CIA-5	b61b14f build result: All green
01:30:50		Quit factor (Ping timeout: 245 seconds)
01:31:23		Quit MethoS- (Quit: Konversation terminated!)
01:34:42		Join factor [0] (~factor@r74-195-185-246.msk1cmtc01.mskgok.ok.dh.suddenlink.net)
01:36:17	***	Saving seen data "./dancer.seen"
01:40:33	saratoga	now i just have to wait and see if that fix worked
01:53:20		Join gxk [0] (~gxk@bzq-79-180-216-196.red.bezeqint.net)
01:57:49		Quit lebellium (Quit: ChatZilla 0.9.88.2 [Firefox 15.0/20120717110313])
01:58:24		Quit gxk (Ping timeout: 248 seconds)
02:00
02:04:23		Part rocktriangle
02:07:59		Quit anewuser (Read error: Connection reset by peer)
02:21:13		Join gxk [0] (~gxk@bzq-79-180-216-196.red.bezeqint.net)
02:39:35		Quit bertrik (Ping timeout: 245 seconds)
02:40:32		Quit gxk (Ping timeout: 248 seconds)
02:49:17		Quit TheLemonMan (Ping timeout: 260 seconds)
03:00
03:04:16		Join gxk [0] (~gxk@bzq-79-180-216-196.red.bezeqint.net)
03:17:56		Quit gxk (Ping timeout: 255 seconds)
03:36:21	***	Saving seen data "./dancer.seen"
03:44:06		Join gxk [0] (~gxk@bzq-79-180-216-196.red.bezeqint.net)
03:48:56		Quit gxk (Read error: Operation timed out)
04:00
04:03:31		Join TheSphinX_ [0] (~briehl@p579CCD90.dip.t-dialin.net)
04:06:44		Quit TheSphinX^ (Ping timeout: 248 seconds)
04:06:56		Join gxk [0] (~gxk@bzq-79-180-216-196.red.bezeqint.net)
04:12:58		Quit TheSeven (Disconnected by services)
04:13:05		Join [7] [0] (~quassel@rockbox/developer/TheSeven)
04:26:10		Quit [Saint] (Read error: Connection reset by peer)
04:26:15		Join [Saint_] [0] (~Saint]@unaffiliated/saint/x-8516940)
04:31:00		Join [Saint] [0] (~Saint]@101.98.158.103)
04:31:00		Quit [Saint] (Changing host)
04:31:00		Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
04:49:18		Quit amiconn (Disconnected by services)
04:49:18		Join amiconn_ [0] (amiconn@rockbox/developer/amiconn)
04:49:23		Nick amiconn_ is now known as amiconn (amiconn@rockbox/developer/amiconn)
04:50:15		Quit gxk (Ping timeout: 255 seconds)
04:50:21		Join pixelma_ [0] (pixelma@rockbox/staff/pixelma)
04:50:21		Quit pixelma (Disconnected by services)
04:50:23		Nick pixelma_ is now known as pixelma (pixelma@rockbox/staff/pixelma)
04:52:03		Join gxk [0] (~gxk@bzq-79-180-216-196.red.bezeqint.net)
05:00
05:05:24		Quit gxk (Ping timeout: 248 seconds)
05:18:08		Join gxk [0] (~gxk@bzq-79-180-216-196.red.bezeqint.net)
05:23:16		Quit gxk (Remote host closed the connection)
05:24:04		Join gxk [0] (~gxk@bzq-79-180-216-196.red.bezeqint.net)
05:32:36		Quit gxk (Ping timeout: 248 seconds)
05:36:22	***	Saving seen data "./dancer.seen"
05:39:23		Join gxk [0] (~gxk@bzq-79-180-216-196.red.bezeqint.net)
05:54:11		Quit gxk (Ping timeout: 252 seconds)
05:55:29		Join gxk [0] (~gxk@bzq-79-180-216-196.red.bezeqint.net)
06:00
06:09:41		Quit gxk (Ping timeout: 264 seconds)
06:16:23		Join gxk [0] (~gxk@bzq-79-180-216-196.red.bezeqint.net)
06:25:51		Quit mikroflops (Ping timeout: 245 seconds)
06:26:42		Join soap_ [0] (~soap@cpe-76-181-64-173.columbus.res.rr.com)
06:26:42		Quit soap_ (Changing host)
06:26:42		Join soap_ [0] (~soap@rockbox/staff/soap)
06:27:16		Quit soap_ (Client Quit)
06:28:09		Quit Totalled (Quit: PETTAN PETTAN, TSURUPETTAN!)
06:30:57		Join mikroflops [0] (~yogurt@h-34-21.a238.priv.bahnhof.se)
06:32:53		Quit gxk (Ping timeout: 255 seconds)
06:34:20		Join gxk [0] (~gxk@bzq-79-180-216-196.red.bezeqint.net)
06:52:51		Quit gxk (Ping timeout: 252 seconds)
06:56:25		Join gxk [0] (~gxk@bzq-79-180-216-196.red.bezeqint.net)
07:00
07:00:20		Quit pixelma (Disconnected by services)
07:00:20		Join pixelma_ [0] (pixelma@rockbox/staff/pixelma)
07:00:22		Nick pixelma_ is now known as pixelma (pixelma@rockbox/staff/pixelma)
07:00:26		Quit amiconn (Disconnected by services)
07:00:26		Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
07:00:31		Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
07:05:29		Join Totalled [0] (~Totalled@199.58.84.5)
07:06:16		Quit gxk (Ping timeout: 245 seconds)
07:06:22		Join [Saint__] [0] (~Saint]@unaffiliated/saint/x-8516940)
07:09:29		Quit [Saint_] (Ping timeout: 260 seconds)
07:15:13		Quit [Saint__] (Ping timeout: 248 seconds)
07:16:47		Join [Saint_] [0] (~Saint]@unaffiliated/saint/x-8516940)
07:23:21		Quit guymann (Quit: brb)
07:26:32		Quit [Saint_] (Remote host closed the connection)
07:28:39		Join gxk [0] (~gxk@bzq-79-180-216-196.red.bezeqint.net)
07:30:26		Join Totalled_ [0] (~Totalled@199.58.84.5)
07:31:04		Quit Totalled (Ping timeout: 260 seconds)
07:31:04		Nick Totalled_ is now known as Totalled (~Totalled@199.58.84.5)
07:35:33		Join Rower85 [0] (husvagn@v-413-alfarv-90.bitnet.nu)
07:36:24	***	Saving seen data "./dancer.seen"
07:41:30		Join guymann [0] (~c@108-237-202-52.lightspeed.wlfrct.sbcglobal.net)
07:43:16		Quit gxk (Ping timeout: 248 seconds)
07:51:06		Join gxk [0] (~gxk@bzq-79-180-216-196.red.bezeqint.net)
07:56:36		Quit gxk (Ping timeout: 240 seconds)
07:59:29		Join gxk [0] (~gxk@bzq-79-180-216-196.red.bezeqint.net)
08:00
08:03:36		Quit gxk (Ping timeout: 240 seconds)
08:05:30		Join gxk [0] (~gxk@bzq-79-180-232-237.red.bezeqint.net)
08:41:34		Join Zagor [0] (~bjst@sestofw01.enea.se)
08:41:34		Quit Zagor (Changing host)
08:41:34		Join Zagor [242] (~bjst@rockbox/developer/Zagor)
08:46:59		Quit pixelma (Remote host closed the connection)
08:46:59		Quit amiconn (Remote host closed the connection)
08:48:24		Join pixelma [0] (pixelma@rockbox/staff/pixelma)
08:48:24		Join amiconn [0] (amiconn@rockbox/developer/amiconn)
08:50:18		Join ender` [0] (krneki@foo.eternallybored.org)
08:59:53		Join einhirn [0] (~Miranda@bsod.rz.tu-clausthal.de)
09:00
09:00:57		Join kevku [0] (x@heaaqi4aafadxgaamaafaadyaby.dyn.reverse.name)
09:03:57		Join mgottschlag [0] (~quassel@HSI-KBW-091-089-250-086.hsi2.kabel-badenwuerttemberg.de)
09:04:01		Quit mgottschlag (Changing host)
09:04:01		Join mgottschlag [0] (~quassel@reactos/tester/phoenix64)
09:33:33	[Saint]	Zagor: Something interesting happened with the wiki yesterday.
09:33:43		Join stoffel [0] (~quassel@pD9E41A46.dip.t-dialin.net)
09:33:56	[Saint]	Someone/thing managed to get write access without being present in a usergroup.
09:35:26	Zagor	aha
09:35:48	Zagor	much damage?
09:36:06	[Saint]	Just a spam page.
09:36:17	Zagor	ok
09:36:26	***	Saving seen data "./dancer.seen"
09:36:47	[Saint]	Which makes me think it was automated, if a human did it they probably would've realized quite quickly they could've done a lot more.
09:38:07		Quit dionoea (Ping timeout: 246 seconds)
09:39:09		Join dionoea [0] (~dionoea@oyp.chewa.net)
09:39:23	[Saint]	But, ...it seems to have removed itself. Or, someone else did so.
09:39:43	[Saint]	it was at http://www.rockbox.org/wiki/ProofReading18
09:39:51	Zagor	the front page says JdGordon removed it
09:39:57	[Saint]	Ah.
09:40:28	[Saint]	I wanted to leave it there once I realized there was no user present by that name in the wikiusersgroup.
09:40:45	[Saint]	In case it somehow helped address how it was achieved.
09:41:08	Zagor	I don't think it would
09:41:13	[Saint]	I was keen to remove it myself until I noticed there was no permissions granted in order to be able to revoke them.
09:43:52	kugel	Torne: the stack of a coded can be in the coded, without occupying space in the core
09:44:14	kugel	codec*
09:44:38		Quit Totalled (Quit: PETTAN PETTAN, TSURUPETTAN!)
09:45:00		Join Totalled [0] (~Totalled@c-98-245-9-211.hsd1.co.comcast.net)
09:45:30	[Saint]	though, there's "http://www.rockbox.org/wiki/PeterPan", that seems to be created similarly.
09:45:54	[Saint]	It was added yesterday, and the wikiusersgroup has had no additions since the 12th of July
09:45:59	[Saint]	Zagor: ^
09:46:19	[Saint]	It seems there's a way in someone//thing knows about :-S
09:46:25	kugel	Torne: ah, i see you have been schooled already?
09:46:41	Zagor	[Saint]: indeed
09:46:55	kugel	s/?/:-)/
09:48:30	[Saint]	Zagor: Although it would be a HUGE task (which, I'm prepared to volunteer time towards if need be), do you think it would be wise to make a snapshot of the wiki and start over, send out a mail to the users/developers list saying "if you had access and still want it, you'll need to request it again"?
09:48:53	[Saint]	It'd be a huge job, but the wiki is pretty sick, and it'd give a chance to do a general cleanup too.
09:49:11	Zagor	to reduce the size of WikiUsersGroup, you mean?
09:49:26	[Saint]	And refresh the wiki in general.
09:49:35	[Saint]	Some pages are outright f**cked
09:49:50	Zagor	well that won't really be fixed by kicking off all users, will it?
09:50:10	JdGordon	[Saint]: dont get your panties in a twist.... like I said, the spam was done bcause users can edit their own page without being in the wikiusersgroup
09:50:14	JdGordon	PeterPan is me
09:50:33	JdGordon	nothing to worry about, no wiki hacking
09:50:41	JdGordon	just annoying permissions
09:50:50	[Saint]	Oh....right. I see, you should remove that page then :)
09:50:56	Zagor	ah, the spam was done in the user's own page?
09:51:01	JdGordon	yes
09:52:06	Zagor	there's another far simpler fix for the bulging WikiUsersGroup page
09:52:09	[Saint]	I'm pretty sure it never used to be that way, though. He wasn't in the wikiusersgroup, so shouldn't have had a users page to begin with.
09:52:57	Zagor	with the latest foswiki upgrade, there's a new feature which inverts the access scheme. so instead of adding people who are allowed to write, new registrants are added to a block list from which we remove those who are approved
09:53:09	[Saint]	As always, there's a chance I'm misremebering, but as I recall the user's page should be generated on their first login.
09:53:37	Zagor	[Saint]: that was a bug. all registrants should get a user page immediately.
09:53:58	Zagor	one can argue wether he should be allowed to edit it though, since that allows spam.
09:54:04	[Saint]	Ah, hmmm. It was a "nice" bug, then :)
09:54:59	[Saint]	Well, that's good. I guess. I can't say I didn't want an excuse to do a mass wiki cleanup though :)
09:56:11	JdGordon	I say leave it as is unless it becomes a bigger problem
09:56:19	JdGordon	one in 5 years is not an issue worth worrying baout
09:57:56	[Saint]	Well, yeah. As I said above, though, I kinda wanted the excuse to do a big purge of the wiki. Make sure its all up-to-date and functional, remove outdated content, and users that haven't ever made a singular edit.
09:58:29	[Saint]	Its no biggie, it would've been a daunting task, but the wiki is definitely the area of the project that gets the least attention.
10:00
10:00:09	[Saint]	Primarily, the user registration needs to be functional (well, "need" is a strong word, but it sure would be nice). As it is, its pure luck whether or not adding a user to the usersgroup actually grants them write access.
10:01:13	[Saint]	And if you select "edit" on the wikiusersgroup (in Chrome, for sure) the "quick add" dialogue disappears and 17 miles of whitespace gets added to the bottom of the page.
10:03:19	[Saint]	JdGordon: you're not by chance also BabisManouselis, are you? :)
10:07:16		Quit dongs (Ping timeout: 248 seconds)
10:08:05		Quit Scromple (Read error: Connection reset by peer)
10:17:04		Join bertrik [0] (~bertrik@ip117-49-211-87.adsl2.static.versatel.nl)
10:17:04		Quit bertrik (Changing host)
10:17:04		Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
10:25:26		Quit stoffel (Remote host closed the connection)
10:34:15		Join wodz [0] (~wodz@iwl138.internetdsl.tpnet.pl)
10:41:37		Join dongs [0] (~dongs@l212168.ppp.asahi-net.or.jp)
10:48:22	bertrik	how do you run the sim in gdb in a practical way?
10:50:55	wodz	you need to compile it with −−sdl-threads
10:51:16	wodz	passed to configure. Otherwise it bombs out on thread switch.
10:51:55	wodz	There is some command you can pass to gdb to workaround though (can't remember the command)
10:52:28	bertrik	it looks like my executable is slightly borked too (my fault)
11:00
11:03:24	gevaerts	Zagor, Bagder: do you know why the html manual links seem to be dead?
11:04:02		Join TheLemonMan [0] (~LemonBoy@adsl-ull-13-208.50-151.net24.it)
11:04:03		Quit TheLemonMan (Changing host)
11:04:03		Join TheLemonMan [0] (~LemonBoy@unaffiliated/thelemonman)
11:36:30	***	Saving seen data "./dancer.seen"
11:40:43	wodz	gevaerts: so basically there is no improvement on e200 with v8 of bflt patch, right?
11:45:03		Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org)
11:49:16	gevaerts	wodz: that's what it looked like, yes
11:53:32	wodz	so it must have to do with .ncdata and .ncbss as nano2g and coldfire choke only on doom
11:55:04	wodz	gevaerts: Do you know the status of e200 emulator (can't remember the link but its on FS)? Does it run current version of rb?
11:55:33	gevaerts	No idea
12:00
12:04:26	wodz	its a pitty I don't have working PP target :/
12:04:32		Join stoffel [0] (~quassel@pD9E41A46.dip.t-dialin.net)
12:04:39		Join lebellium [0] (~chatzilla@e179140140.adsl.alicedsl.de)
12:05:04	gevaerts	I could send you one
12:06:30	wodz	Maybe I can ask bohoon. That way we will avoid international shipping
12:15:29	[Saint]	I have a couple of iPod Colors surplus to requirement...
12:15:58	gevaerts	Yes, but you're not ideally located shipping-wise :)
12:17:26	gevaerts	I seem to have 8 PP devices, I'm sure I can miss one of them
12:17:47	wodz	by the way, currently plugins use main stack AFAIK. This will be problem if we consider multiple plugins at the same time
12:18:25	wodz	well, I can always ask Torne to send me back my mini1g
12:19:54	Torne	Oh, yeah.
12:19:55	Torne	sorry about that.
12:20:14	Torne	i can do that whenever, really
12:20:38	gevaerts	That's an issue to keep in mind, yes. It being a problem will depend on how we do multiple plugins. I'm e.g. not convinced we want more than one "regular" plugin, I think we're more interested in e.g. battery bench not having to stop, or wps visualisation plugins (both of which will use either separate threads or callbacks), or moving out bits of the core (which would work more as a library I guess)
12:22:10	wodz	currently I completely ignored library support
12:22:21	gevaerts	Oh, of course
12:22:21	wodz	I mean the got and friends
12:22:50	Torne	wodz: well, we don't actually ahve todo it that way
12:22:58	Torne	in fact we probably won't
12:23:04	gevaerts	When I say "more as a library" I also don't necessarily mean "the way a shared library tends to work"
12:23:05	Torne	since i doubt we want link by name
12:23:51	gevaerts	In fact, I'm not sure if "library" is the proper word to use there :)
12:24:34	Torne	i would imagine we would do that kind of thing using a single symbol that pointed to a struct like plugin_api
12:26:56	Torne	anyway we can work that stuff out later :p
12:27:29	wodz	sure. First thing is to have working all plugins on all platforms
12:36:05		Quit Topy44 (Ping timeout: 264 seconds)
12:39:57		Join Topy44 [0] (~Topy44@f048203231.adsl.alicedsl.de)
12:56:25		Quit stoffel (Read error: Operation timed out)
13:00
13:22:11		Join dfkt [0] (dfkt@unaffiliated/dfkt)
13:36:32	***	Saving seen data "./dancer.seen"
13:37:21		Join [Saint_] [0] (~Saint]@unaffiliated/saint/x-8516940)
13:38:11		Join einhirn [0] (~Miranda@bsod.rz.tu-clausthal.de)
13:45:17		Join [Saint__] [0] (~Saint]@unaffiliated/saint/x-8516940)
13:45:39	gevaerts	Looks like the html manuals stopped building (on the server) in early June
13:49:03		Quit [Saint_] (Ping timeout: 260 seconds)
13:50:58		Join MethoS- [0] (~clemens@134.102.106.250)
13:54:24		Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
13:54:33	gevaerts	OK, the html manual still builds for me. This seems to be server-side
13:54:43	gevaerts	Zagor, Bagder: a job for you?
13:55:04	kugelp	how about one thread per plugin?
13:55:04	Zagor	yep, I'll try to get a peek at it today
13:55:41	gevaerts	kugelp: doesn
13:56:19	gevaerts	kugelp: doesn't the drawing API assume it's only called from the main thread?
13:56:41	kugelp	no
13:58:13	kugelp	the scrolling thread draws too, even concurrently
13:58:40	gevaerts	Anyway, I'm not sure what problem this would solve exactly. For fullscreen plugins, only one can run usefully at a time anyway, so others would have to be notified in some way about not doing anything, while for hypothetical other (new) plugin types things like button handling haven't been worked out at all yet
13:59:10	gevaerts	Unless proven otherwise I'm still of the opinion that some sort of callback system will work better
14:00
14:04:24	wodz	Figuring out why quite a few plugins misbehave on PP is more important for now.
14:04:39	kugelp	just saw you talked about problematic stacks
14:04:57	wodz	I just pointed out that this will be something to talk about
14:05:45	wodz	It is not for now. For 'many' plugins at once we will need buflib support to allocate space also
14:09:37		Quit bertrik (.net .split)
14:09:37		Quit mgottschlag (.net .split)
14:09:38		Quit [7] (.net .split)
14:09:38		Quit funman (.net .split)
14:09:39		Quit mc2739 (.net .split)
14:09:39		Quit soap (.net .split)
14:09:40		Quit CIA-5 (.net .split)
14:09:40		Quit gmaxwell (.net .split)
14:09:56		Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
14:10:12		Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
14:10:12		Join mgottschlag [0] (~quassel@reactos/tester/phoenix64)
14:10:12		Join gmaxwell [0] (~greg@wikimedia/KatWalsh/x-0001)
14:10:12		Join funman [0] (~fun@rockbox/developer/funman)
14:10:12		Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
14:10:12		Join soap [0] (~soap@rockbox/staff/soap)
14:10:12		Join CIA-5 [0] (~CIA@cia.atheme.org)
14:18:56		Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
14:21:35		Quit bertrik (.net .split)
14:21:35		Quit mgottschlag (.net .split)
14:21:35		Quit funman (.net .split)
14:21:36		Quit mc2739 (.net .split)
14:21:36		Quit soap (.net .split)
14:21:36		Quit CIA-5 (.net .split)
14:21:36		Quit gmaxwell (.net .split)
14:21:45		Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
14:22:31		Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
14:22:31		Join mgottschlag [0] (~quassel@reactos/tester/phoenix64)
14:22:31		Join gmaxwell [0] (~greg@wikimedia/KatWalsh/x-0001)
14:22:31		Join funman [0] (~fun@rockbox/developer/funman)
14:22:31		Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
14:22:31		Join soap [0] (~soap@rockbox/staff/soap)
14:22:31		Join CIA-5 [0] (~CIA@cia.atheme.org)
14:24:26		Quit [Saint__] (Quit: Quit)
14:24:36		Quit mystica555 (Ping timeout: 245 seconds)
14:24:44		Quit linuxstb (Quit: This computer has gone to sleep)
14:24:51		Join [Saint_] [0] (~Saint]@unaffiliated/saint/x-8516940)
14:40:48		Quit [Saint_] (Quit: Quit)
14:41:11		Join [Saint_] [0] (~Saint]@unaffiliated/saint/x-8516940)
14:52:10		Join T44 [0] (~Topy44@f049046158.adsl.alicedsl.de)
14:55:33		Quit Topy44 (Ping timeout: 260 seconds)
14:57:41		Quit mikroflops (Ping timeout: 264 seconds)
15:00
15:01:36		Quit MethoS- (Ping timeout: 252 seconds)
15:04:38		Join petur [0] (~petur@rockbox/developer/petur)
15:13:55	*	TheSeven directs slowdow here (backlog: http://pastie.org/4306096 )
15:14:32	TheSeven	backlight thread is still alive, i.e. usb state machine deadlock?
15:14:37	TheSeven	gevaerts? anyone?
15:15:55		Join slowdow [0] (~slowdow@172.Red-81-37-67.dynamicIP.rima-tde.net)
15:16:30	TheSeven	slowdow: I already posted the backlog here, so you don't need to explain again
15:17:29	slowdow	ok, Thanks TheSeven
15:17:41	slowdow	so. help! :)
15:17:55	wodz	well usb is badly borked recently
15:18:43	slowdow	i should point out at that i'm happily using rockbox every day on 2 devices, and that i love it.
15:18:51	TheSeven	wodz: so you suspect a core bug, not a target driver one?
15:18:55	slowdow	this Ipod is perfectly usaable.
15:19:03	slowdow	i just use the fallback image to copy data
15:19:12	slowdow	and then boot to the new builds for use. no biggie
15:19:17		Quit dfkt (Quit: -= SysReset 2.55=- Sic gorgiamus allos subjectatos nunc.)
15:19:27	slowdow	but it is a bit annoying, and for your common or garden user, a bit of a faff?
15:19:55	slowdow	the fallback image mounts first time every time.
15:20:18	TheSeven	too bad these builds are like half a year apart
15:20:25	TheSeven	so we can't just pinpoint a change easily
15:20:34	wodz	TheSeven: There are reports of misbehaving on amsv2
15:20:47	TheSeven	well, amsv2 is the same OTG, right?
15:20:54	wodz	other soft usb targets are rare unfortunately
15:21:00	slowdow	ive got a small pile of builds i've archived, the standard builds were working up to a point a few months ago
15:21:27	TheSeven	does the classic currently use funman's shared driver?
15:21:41	slowdow	errm.
15:21:54	slowdow	I've no idea. tell me how to find out?
15:22:40		Join LinusN [0] (~linus@giant.haxx.se)
15:22:52	*	TheSeven tries to remember whether the current fallback build was before or after funman's OTG rework
15:23:04	wodz	TheSeven: nano2g is broken not by funman's work. It was some general usb core rework
15:23:06		Quit mgottschlag (Ping timeout: 240 seconds)
15:23:20		Join mikroflops [0] (~yogurt@h-34-21.a238.priv.bahnhof.se)
15:23:43	slowdow	Fallback is R31516-120101
15:24:04	TheSeven	wodz: IIRC there were some issues with funman's stuff as well, no idea if those were fixed in the meantime
15:24:41		Join mystica555 [0] (~Mike@71-211-209-19.hlrn.qwest.net)
15:25:02	TheSeven	hm, so this is a "vanilla" build without any special patches (I think I resorted to 3.9 for the nano2g because of USB trouble)
15:28:08	TheSeven	I've been watching closely recently... so SW USB is all the way borked?
15:28:45	wodz	TheSeven: funman tried transplanted old driver over HEAD and this didn't work as well. Bisect points to some general usb core rework
15:28:47	TheSeven	I haven't been*
15:29:13		Join Neelark [0] (5af6f49d@gateway/web/freenode/ip.90.246.244.157)
15:29:44	TheSeven	If I would have the time, I'd probably go ahead and rewrite the USB core and OTG driver from scratch
15:30:24		Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org)
15:30:35	slowdow	Mounting official rockbox-ipod6g 0823817-120718 in Ubuntu
15:30:54	slowdow	via vm
15:31:06	TheSeven	I have a fairly well-working monolithic driver implementation for that OTG (for another project, on another SoC)
15:31:28	TheSeven	slowdow: if it fails, please pastebin dmesg output
15:32:08	slowdow	or maybe not.
15:32:53	slowdow	because the device mounts on the host system first, then when i select to mount it to the VM it unmounts and remounts, i think it's allready crashed
15:32:59	slowdow	i get a win error message
15:33:14	TheSeven	ok
15:34:07	slowdow	One of the USB devices connected to this Pc has malfunctioned and windows does not recognise it
15:34:31	slowdow	hang on, let me find another laptop, i'll boot it to pendrive and plug in the iPod.
15:35:44		Join einhirn [0] (~Miranda@bsod.rz.tu-clausthal.de)
15:36:36	***	Saving seen data "./dancer.seen"
15:41:32		Join WalkGood [0] (~4@unaffiliated/walkgood)
15:41:47		Join [Saint__] [0] (~Saint]@unaffiliated/saint/x-8516940)
15:42:57		Quit [Saint_] (Ping timeout: 248 seconds)
15:43:11		Nick [Saint__] is now known as [Saint_] (~Saint]@unaffiliated/saint/x-8516940)
15:43:32		Nick [Saint_] is now known as [Sinner] (~Saint]@unaffiliated/saint/x-8516940)
15:44:32		Nick [Sinner] is now known as [Sinner_] (~Saint]@unaffiliated/saint/x-8516940)
15:44:38		Nick [Sinner_] is now known as [Sinner__] (~Saint]@unaffiliated/saint/x-8516940)
15:44:47		Nick [Sinner__] is now known as [Sinner] (~Saint]@unaffiliated/saint/x-8516940)
15:51:15	slowdow	Do you want a dmsg output of the working/fallback, and then the fail/new. or just the new.
15:53:18		Join amayer [0] (~amayer@mail.weberadvertising.com)
15:58:47	TheSeven	just the failing one
16:00
16:00:42	slowdow	humm. ok. so. first time i plugged in the ipod, on 0823817-120718 it failed. Dmsg saved.
16:00:56	slowdow	rebooted ipod to fallback - mounted. worked
16:01:06	slowdow	Rebooted ipod to 0823817-120718 - Mounted.. WORKED!
16:05:29		Nick [Sinner] is now known as [Saint_] (~Saint]@unaffiliated/saint/x-8516940)
16:05:59	TheSeven	odd... probably just luck and some e.g. timing related bug
16:08:14	slowdow	OT - good linux IRC client?
16:09:01	TheLemonMan	slowdow: weechat / irssi
16:10:38	[Saint_]	Xchat/Quassel
16:11:24	slowdow	thnks
16:11:30		Quit [Saint_] (Quit: almost home)
16:12:07	slowdow	54meg of irc client. thats a chunky client.
16:13:23	TheSeven	which one is that?
16:13:51	TheSeven	quassel, if you count all the kde/qt libs it requires?
16:15:41	slowdow	yes
16:15:53	*	gevaerts points to #rockbox-community
16:15:58	slowdow	dumped it for xchat , its on its waym slowwwwly
16:16:13	*	TheSeven tells gevaerts to fix usb instead of pointing at random channels :P
16:18:16	slowdow	so. pastebin. http://pastebin.com/kf5FtDRh
16:18:29	slowdow	first 2 lines = the Ubuntu error message on connection.
16:19:09	gevaerts	TheSeven: what's wrong with USB?
16:19:15	TheSeven	slowdow: that was two tries?
16:19:44	slowdow	no, that was one.
16:20:15	slowdow	ive got the one for 2 tries, and the third one where it mounted 0823817-120718
16:20:31	TheSeven	it seems to have connecteed 2 times, exactly 90 seconds apart
16:20:35	slowdow	its still mounted btw, im loath tp unplug it just because its working :)
16:21:14	wodz	TheSeven: once as high speed and when it failed as full speed
16:21:24	slowdow	thats deffinately one try. it took me that long to copy and paste the error message
16:21:25	wodz	typical fallback mechanism
16:22:04	slowdow	I have to go away for a while, I'll chack back later if there's anything else you want.
16:22:17	TheSeven	wodz: the full speed attempts failed, and then another highspeed attempt was made an succeeded a little bit before failing as well in the end
16:23:00	slowdow	not that it makes much difference, but i spent 3 years doing QA for playstation/dreamcast/pc games, if there's anything you want testing, tell me the test conditions and i'll do what i can, Thanks again for the help.
16:23:14		Quit Neelark (Quit: Page closed)
16:23:59	TheSeven	oh, we have a QA volunteer? :P
16:24:05	wodz	cool
16:24:08	wodz	:-)
16:24:21	slowdow	if it'd help
16:24:36	slowdow	ive only got access to Ipod Classic 6g + Iriver H340
16:24:49	slowdow	but, if it'll help. no probs.
16:25:22	slowdow	BBL
16:29:47		Part LinusN
16:30:00		Join lebellium_ [0] (~chatzilla@e179067146.adsl.alicedsl.de)
16:30:17	slowdow	- it ejected/unmounted without crashing too, i'll reboot the linux box and try remounting it on a fresh boot when i get back.
16:31:48		Quit lebellium (Ping timeout: 260 seconds)
16:31:55		Nick lebellium_ is now known as lebellium (~chatzilla@e179067146.adsl.alicedsl.de)
16:32:12		Quit petur (Quit: plop)
16:40:28		Join prof_wolfff [0] (~prof_wolf@82.159.1.234.dyn.user.ono.com)
16:47:35	amiconn	gevaerts: The drawing api does not assume being called from a single thread only. The functions directly accessing the lcd do though
16:48:44		Join y4n [0] (~y4n@unaffiliated/y4ndexx)
16:48:57	amiconn	lcd_update and friends
16:49:58		Join MethoS- [0] (~clemens@134.102.106.250)
16:50:17	amiconn	This is e.g. why one has to be especially careful when using the greylib (it has a couple of comments regarding this)
16:50:49	*	gevaerts nods
16:58:03		Quit Zagor (Quit: Clint excited)
17:00
17:01:14		Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org)
17:05:50		Quit wodz (Quit: Leaving)
17:09:05		Quit efyx (Remote host closed the connection)
17:20:07		Join Neelark [0] (5af6f49d@gateway/web/freenode/ip.90.246.244.157)
17:20:12		Quit soap (Read error: Connection reset by peer)
17:30:58	kugelp	amiconn: that's not actually the case, though
17:34:34		Quit Neelark (Ping timeout: 245 seconds)
17:34:59		Quit saratoga (Ping timeout: 245 seconds)
17:36:37	***	Saving seen data "./dancer.seen"
17:37:23		Join soap [0] (~soap@cpe-76-181-64-173.columbus.res.rr.com)
17:37:23		Quit soap (Changing host)
17:37:23		Join soap [0] (~soap@rockbox/staff/soap)
17:43:35		Join dfkt [0] (dfkt@unaffiliated/dfkt)
17:46:56		Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
17:52:37		Quit dfkt (Ping timeout: 255 seconds)
17:53:40	kugel	amiconn: the scroll engine calls these, and many LCD drivers have mutexes for this reason
17:55:46		Join efyx [0] (~efyx@lap34-1-82-225-185-146.fbx.proxad.net)
17:57:12		Join evilnick [0] (~evilnick@rockbox/staff/evilnick)
18:00
18:00:08		Join dfkt [0] (dfkt@unaffiliated/dfkt)
18:03:47		Join XavierGr [0] (~xavier@rockbox/staff/XavierGr)
18:06:18		Quit TheLemonMan (Ping timeout: 260 seconds)
18:24:21		Quit soap (Quit: soap)
18:30:06		Quit ehntoo (Ping timeout: 246 seconds)
18:31:06		Join ehntoo [0] (ehntoo@gateway/shell/mtu-lug/x-ekohdnhznjrygbqx)
18:32:13		Join Mir_ [0] (~Mir@pool-71-109-217-252.lsanca.dsl-w.verizon.net)
18:32:54		Quit Mir (Ping timeout: 246 seconds)
18:33:22		Join Wardo [0] (~Mirandaha@176-120-190-109.dsl.ovh.fr)
18:38:13	slowdow	So, on ubuntu - after that first 0823817-120718 Usb mount fail, then booting to fallback = success, now 3X working mounts on 0823817-120718. (power on ipod, wait for disk access to stop - plug in, connects instantly)
18:46:33		Join wodz [0] (~wodz@89-76-32-53.dynamic.chello.pl)
18:48:30		Join mgottschlag [0] (~quassel@HSI-KBW-091-089-250-086.hsi2.kabel-badenwuerttemberg.de)
18:48:30		Quit mgottschlag (Changing host)
18:48:30		Join mgottschlag [0] (~quassel@reactos/tester/phoenix64)
18:50:52	amiconn	kugel: While a mutex might be a good idea in a high color driver, it definitely isn't in an 1 bit or 2 bit driver
18:51:24	amiconn	The reason is that the greylib (has to) write to the lcd in isr context
18:53:09		Join Strife89 [0] (~Strife89@207.144.201.128)
18:53:40	kugel	just saying that "lcd_update is called by a single thread" isn't true
18:57:46	slowdow	on the 2nd pc, (the same pc as mounted in Ubuntu), the ipod is now mounting in win7. it mounts as "rockbox media player" - however, when plugging it into my main laptop it still fails, and mounts as Apple iPod Classic USB Device" with 0bytes.
18:58:31	slowdow	is there a way i can force my other laptop to mount the ipod as "rockbox media device" .. maybe that'll work and its a problem between the Apple driver and the later versions of rockbox?
19:00
19:01:34	slowdow	strange this is, i plugged it into another laptop, that's never had itunes or anything fruit company related installed, while it was running rockbox, and it still installed the driver as "Apple iPod Classic USB Device"
19:03:05	yuriks	slowdow: http://support.microsoft.com/kb/315539
19:03:20	yuriks	slowdow: do that and remove the device, then let it detect again when you plug it in
19:13:53	slowdow	the device isnt shown, i've run set devmgr_show_nonpresent_devices=1, and sleected show hidden
19:14:23	slowdow	its only shown when its connected. is it ok to remove the driver while its connected and therefore visible?
19:22:58	slowdow	ok, removed the apple driver, re-plugged, it's now called "Rockbox media player" (bad capitalisiation included) but it still mounts as a 0byte drive.
19:23:29	slowdow	and crashes the Ipod.
19:25:39		Nick Mir_ is now known as Mir (~Mir@pool-71-109-217-252.lsanca.dsl-w.verizon.net)
19:25:55	slowdow	Rockbox fallback R31516-120101 still mounts first time. as "Rockbox media player"
19:27:14	slowdow	so... now, i've got one machine that mounts 0823817-120718 in win7 and Linux. and one that won't mount in win7. (typically, the fail machine is my main machine)
19:35:28		Quit bluebrother (Ping timeout: 248 seconds)
19:36:37		Quit fs-bluebot (Ping timeout: 255 seconds)
19:36:39	***	Saving seen data "./dancer.seen"
19:37:58		Join fs-bluebot [0] (~fs-bluebo@g225255108.adsl.alicedsl.de)
19:38:07		Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
20:00
20:14:42		Join soap_ [0] (~soap@cpe-76-181-64-173.columbus.res.rr.com)
20:14:43		Quit soap_ (Changing host)
20:14:43		Join soap_ [0] (~soap@rockbox/staff/soap)
20:18:29		Join linuxstb [0] (~linuxstb@host86-181-136-95.range86-181.btcentralplus.com)
20:23:34		Quit linuxstb (Quit: This computer has gone to sleep)
20:24:51		Join stoffel [0] (~quassel@pD9E4371C.dip.t-dialin.net)
20:29:23		Quit WalkGood (Quit: ♪ ♫ ♪ ♫ ♪ ♫ ♪)
20:35:16		Join Neelark [0] (5af6f49d@gateway/web/freenode/ip.90.246.244.157)
20:45:34	bertrik	is FLAC also typically encoded in an ogg encoder?
20:45:39	bertrik	how is meta-data done with FLAC?
20:52:30		Quit stoffel (Ping timeout: 245 seconds)
20:54:00	wodz	at least flacs I have start with fLaC magic string which I assume mens it is not packed in ogg container
20:54:51	wodz	http://flac.sourceforge.net/faq.html#general__native_vs_ogg
20:56:27	wodz	bertrik:^
21:00
21:11:16	yuriks	bertrik: they use ogg tags
21:12:10		Join n1s [0] (~n1s@nl118-168-30.student.uu.se)
21:12:10		Quit n1s (Changing host)
21:12:10		Join n1s [0] (~n1s@rockbox/developer/n1s)
21:16:01		Quit T44 (Read error: Connection reset by peer)
21:16:14		Join Topy [0] (~Topy44@f049046158.adsl.alicedsl.de)
21:18:52		Join stoffel [0] (~quassel@pD9E4371C.dip.t-dialin.net)
21:36:40	***	Saving seen data "./dancer.seen"
21:43:18		Nick soap_ is now known as soap (~soap@rockbox/staff/soap)
21:43:32	amiconn	"ogg tags" are called vorbis comments afaik
21:45:59	n1s	yes, at least in ogg/vorbis files
21:46:41	gmaxwell	n1s: They're called vorbis-comment elsewhere too.
21:47:51	n1s	gmaxwell: ok
21:48:15	n1s	bertrik: ogg/flac is rather uncommon and not as widely supported as native flac
21:50:45		Quit MethoS- (Quit: Konversation terminated!)
21:54:38		Join anewuser [0] (~anewuser@190.207.196.98)
21:54:39		Quit anewuser (Changing host)
21:54:39		Join anewuser [0] (~anewuser@unaffiliated/anewuser)
21:58:16		Quit perrikwp (Read error: Connection reset by peer)
21:59:32		Join perrikwp [0] (~quassel@cpe-024-163-024-033.triad.res.rr.com)
22:00
22:08:42		Quit Neelark (Quit: Page closed)
22:08:57		Quit Wardo (Read error: Connection reset by peer)
22:16:54		Join T44 [0] (~Topy44@f049046158.adsl.alicedsl.de)
22:16:54		Quit Topy (Read error: Connection reset by peer)
22:21:13		Join Scromple [0] (~Simon@119.225.209.134)
22:29:25		Quit evilnick (Quit: Leaving)
22:30:56		Quit T44 (Read error: Connection reset by peer)
22:31:04		Join linuxstb [0] (~linuxstb@host86-181-136-95.range86-181.btcentralplus.com)
22:31:25		Quit y4n (Quit: PANTS OFF!)
22:31:37		Join T44 [0] (~Topy44@f049046158.adsl.alicedsl.de)
22:32:23		Join MethoS- [0] (~clemens@134.102.106.250)
22:36:54		Quit MethoS- (Client Quit)
22:44:47	funman	hi gmaxwell
22:45:13	gmaxwell	funman: greetings.
22:45:42	funman	gmaxwell: bertrik said he wanted to look at opus for rockbox a few days ago
22:46:29	funman	bertrik: i think flac-in-flac-container uses vorbis comments (afaik metadata is not part of ogg container)
22:46:51	gmaxwell	He and freqmod have, and there is something now which is "running" in the sim.
22:47:21	gmaxwell	There were some challenges wrt stack usage and I made a suggestion about the opus build config for that, not sure where it stands this second.
22:47:54	funman	ah you progress to fast for me to follow :)
22:48:48		Quit Xerion (Quit: )
22:51:19		Quit stoffel (Remote host closed the connection)
22:51:30		Join Xerion [0] (~xerion@5419F5F4.cm-5-2d.dynamic.ziggo.nl)
22:53:54		Join stoffel [0] (~quassel@pD9E4371C.dip.t-dialin.net)
22:54:24		Quit kevku (Ping timeout: 248 seconds)
22:55:26	funman	is there a patch on gerrit ?
22:56:52	gmaxwell	freqmod/rockbox-opus">https://github.com/freqmod/rockbox-opus
23:00
23:04:23		Quit [Saint] (Ping timeout: 260 seconds)
23:06:01		Part amayer
23:07:01		Join perrikwp_ [0] (~quassel@cpe-024-163-024-033.triad.res.rr.com)
23:07:03		Quit perrikwp (Read error: Operation timed out)
23:07:10	bertrik	funman, I fixed some stuff regarding metadata parsing and elapsed time (so the progress bar shows correct progress), but haven't been able to make it run on target yet
23:07:38	bertrik	I think I can't just push that to the repo of freqmod
23:07:56	bertrik	freqmod is away this week
23:09:29		Quit kugel (Remote host closed the connection)
23:10:20	bertrik	but the latest stuff from his repo runs in the simulator
23:11:54	gmaxwell	I see he's still using VAR_ARRAYs.. so it's probably running into the stack usage issues that were mentioned on the actual target.
23:13:26	funman	bertrik: maybe try running the sim in valgrind to see errors non-fatal in the sim but fatal on target
23:14:01	gmaxwell	I mean I don't think there is any doubt. if you only have 9k stack then I'm pretty sure that isn't enough for a VAR_ARRAY configured build.
23:14:14	bertrik	gmaxwell, so the new plan should be to use the NONTHREADSAFE_PSEUDOSTACK thing instead of VAR_ARRAYS, right?
23:14:35		Quit dfkt (Quit: -= SysReset 2.55=- Sic gorgiamus allos subjectatos nunc.)
23:14:37	gmaxwell	Right, and it'll need a little twiddling because you'll need to allocate some more heap for that.
23:15:23	bertrik	unfortunately, the whole memory mapping thing on rockbox is a bit of a mystery to me
23:15:39	bertrik	we have all kinds of weird hacks to steal memory here and there
23:15:53	gevaerts	We do?
23:16:18	gmaxwell	well, it allocates heap for the decoder state already. Stubbing out the allocation functions in NONTHREADSAFE_PSEUDOSTACK and setting the global to point to some more heap that comes out of that codec_malloc call should be enough.
23:17:28	bertrik	gevaerts, what does codec_malloc do?
23:19:00	gevaerts	bertrik: at first sight, it malloc()s memory from the leftover codec buffer
23:19:13	bertrik	running the sim with gdb always gives me weird issues
23:19:19		Join TheLemonMan [0] (~LemonBoy@adsl-ull-13-208.50-151.net24.it)
23:19:19		Quit TheLemonMan (Changing host)
23:19:19		Join TheLemonMan [0] (~LemonBoy@unaffiliated/thelemonman)
23:20:12	funman	bertrik: i think you need sdl-threads
23:20:23	gevaerts	sdl threads are a bit easier, yes
23:20:29		Join kugel [0] (~kugel@141.45.176.104)
23:20:29		Quit kugel (Changing host)
23:20:29		Join kugel [0] (~kugel@rockbox/developer/kugel)
23:24:16		Quit rasher (Read error: Connection reset by peer)
23:24:52	*	bertrik is waiting for doom to compile again ...
23:25:52	funman	bertrik: make bin && make codecs ?
23:26:17	bertrik	false
23:26:44		Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
23:26:49	funman	(to skip building rocks)
23:27:52		Quit stoffel (Remote host closed the connection)
23:28:23		Join rasher [0] (~rasher@rockbox/developer/rasher)
23:30:09	wodz	gevaerts: dict returning 'Failed to allocate buffer' is super weird. Are you sure you tested v8 build not the previous one once again?
23:31:16		Quit T44 (Read error: Connection reset by peer)
23:31:36		Join T44 [0] (~Topy44@f049046158.adsl.alicedsl.de)
23:33:43	bertrik	hm, gdb isn't complaining, let's try on target
23:34:31	funman	gdb will not complain unless rockbox crashes
23:35:01	funman	and even then it won't complain, but just let you navigate the code before the crash
23:35:03		Quit kugel (Read error: Operation timed out)
23:35:05	bertrik	oh, I should try valgrind
23:35:53	bertrik	when compiling for a target, I still have a malloc popping up somewhere
23:35:58	n1s	valgrind seems to not be very good at figuring out our homegrown malloc stuff
23:36:22	n1s	or well, not our but tlsf
23:36:41	***	Saving seen data "./dancer.seen"
23:36:55	n1s	bertrik: some codecs use a more or less real malloc now
23:38:21		Quit n1s (Quit: Ex-Chat)
23:38:52	funman	valgrind will not track our malloc at all for memory leaks but it should catch overreads/overwrites
23:40:20	bertrik	malloc in opus should be replaced by a different mechanism with the proper #defines now as far as I understand
23:41:03		Join WilliamC [0] (~William@nc-184-3-21-227.dhcp.embarqhsd.net)
23:41:12	WilliamC	My mp3 player isn't charging over 50%
23:42:06	WilliamC	Any idea as to why?
23:44:16	bertrik	which player?
23:44:24	WilliamC	Sansa Clip Zip
23:44:44	bertrik	then I have no idea why
23:48:32	WilliamC	It was fully charged, something was bugged in the theme
23:48:51	WilliamC	or not
23:49:10	WilliamC	Would buggy Rockbox firmware keep it from charging?
23:49:12		Quit mc2739 (Quit: reboot)
23:51:08	funman	you can take a look at charging debug screen
23:51:26	WilliamC	How?
23:51:34	WilliamC	I cannot access that when it's charging
23:51:36		Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
23:52:19		Quit [Saint] (Ping timeout: 248 seconds)
23:53:34		Quit wodz (Quit: Leaving)
23:55:58	funman	hm yes you can, just press any button while you plug the cable and it'll go in charging only mode (no USB data connection)
23:56:08		Join kugel [0] (~kugel@141.45.176.104)
23:56:09		Quit kugel (Changing host)
23:56:09		Join kugel [0] (~kugel@rockbox/developer/kugel)
23:56:46	WilliamC	I'll do that when I'm done using it
23:58:20	gevaerts	wodz: I checked the version number a few times when I started noticing the similar issues

Previous day | Next day