#rockbox 2012-01-09

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 2012-01-09

00:03:34		Quit Topy44 (Ping timeout: 240 seconds)
00:05:53		Quit Kiwi_CAM (Ping timeout: 248 seconds)
00:07:53		Join anewuser [0] (~anewuser@186.93.206.86)
00:07:53		Quit anewuser (Changing host)
00:07:53		Join anewuser [0] (~anewuser@unaffiliated/anewuser)
00:08:10		Quit crwl (Ping timeout: 255 seconds)
00:12:00		Join crwl [0] (~crwlll@dsl-jklbrasgw1-ffb9c300-103.dhcp.inet.fi)
00:17:57		Join Kiwi_CAM [0] (~Kiwi_Cam@189.60.247.184)
00:24:51	kugel	saratoga: codesourcery fails too
00:27:38		Join robin0800 [0] (~robin0800@cpc2-brig8-0-0-cust848.3-3.cable.virginmedia.com)
00:31:04		Quit dfkt (Quit: -= SysReset 2.55=- Sic gorgiamus allos subjectatos nunc.)
00:35:37		Quit liar (Ping timeout: 252 seconds)
00:35:47	*	kugel has a suspicion
00:41:05		Join perrikwp_ [0] (~quassel@cpe-024-163-024-033.triad.res.rr.com)
00:41:32		Join patheticbliss [0] (~calvin@75-134-134-13.dhcp.roch.mn.charter.com)
00:42:54	patheticbliss	Hey guys quick question: I take it the Fuze v2 USB is stable now? I don't see anything about it on the wiki anymore.
00:43:28		Quit bertrik (Ping timeout: 240 seconds)
00:43:41		Quit perrikwp (Ping timeout: 240 seconds)
00:43:50		Quit robin0800 (Quit: Leaving)
00:46:28	jhMikeS	patheticbliss: Not "stable" but "enabled". Whether it actually stays in a release build is another matter.
00:53:52		Quit saratoga (Ping timeout: 258 seconds)
00:54:43		Quit t0rc (Quit: WeeChat 0.3.6)
00:55:02	kugel	grr
00:55:11	kugel	I cant work out why the code crashes
00:55:21	kugel	I'm actually considering some cpu errata
00:55:35	JdGordon	thats a cop out!
00:56:27	kugel	it's plain asm, and the same code works perfectly fine on other armv6
00:57:23		Join saratoga [0] (9803ec71@gateway/web/freenode/ip.152.3.236.113)
00:57:59	kugel	saratoga: i tried with the same gcc options as for android now also
00:58:09	kugel	still not working...
00:58:20	kugel	I'm thinking it could possibly a cpu bug?
00:58:52	kugel	otoh there's lots of other code doing lots of smlal as well
00:58:58		Quit lebellium (Quit: ChatZilla 0.9.88 [Firefox 10.0/20120104111456])
00:59:15	saratoga	kugel: all that code does is load, store and mul, so i don't see how a cpu bug could happen
00:59:29	saratoga	plus the gcc generated code is very, very similar
01:00
01:00:12	saratoga	the fact that its seems to vary when it crashes makes me think memory corruption
01:00:32	***	Saving seen data "./dancer.seen"
01:01:13	kugel	I checked the parameters, they look alright
01:01:43	kugel	and it shouldnt crash if it's loading corrupted audio data, just sound wrong, no?
01:02:04	saratoga	kugel: correct
01:02:13	saratoga	you can feed it random numbers and it won't crash
01:02:30	saratoga	its literally just a filterbank, no assumptions are made about the data
01:02:45	kugel	so I don't see how memory corruption can make it crash if the parameters are still correct
01:03:08	saratoga	how did you check the parameters?
01:03:27	kugel	printf
01:03:45		Quit patheticbliss (Quit: leaving)
01:04:14	saratoga	you can do that on target?
01:04:30	kugel	I can printf and redirect to a file
01:05:05	saratoga	can you print the pointer values before each ldr/str operation?
01:05:14	saratoga	i'm assuming it has to be a load/store that crashes
01:05:35	kugel	I could, but inserting printf calls into asm is tricky
01:08:04		Quit ender` (Quit: Revised log levels proposal: "fyi," "wtf," and "omg.")
01:08:29	kugel	I didn't know you can put negative regs into the offset for ldr/str
01:08:37	kugel	e.g. str r6, [r0, -r5, lsl #2]
01:09:07	jhMikeS	could the funny use of sp cause problems?
01:10:09	kugel	not within that code, but perhaps the host does something with sp during preemption?
01:10:22	jhMikeS	my thoughts too
01:10:30	kugel	maybe check for stack overflow?
01:11:04	jhMikeS	maybe it really expects it to point to the stack, not be a gp register
01:11:54	saratoga	do you have preemption on this target?
01:11:56	jhMikeS	sometimes registers are stacked by kernels, no?
01:12:04	kugel	saratoga: this is linux...of course
01:12:40	saratoga	linux wouldn't depend on the application to maintain the stack pointer though, that would be a pretty big security risk i think
01:12:47	saratoga	unless this is some weird embedded kernel
01:12:53	kugel	otoh there's no real other task running
01:13:23	kugel	security risk?
01:13:27		Quit tessarakt2 (Ping timeout: 240 seconds)
01:14:30	saratoga	kugel: kind of defeats the purpose of having virtual addressing if any application can make the kernel write to whatever memory it wants
01:14:38	jhMikeS	ARMs own code examples depend on USER sp being coherent when doing nested ISR (just sayin')
01:15:10	kugel	saratoga: the kernel doesn't write to the apps stack
01:15:33	saratoga	if changing sp is the problem, it must right?
01:15:50	kugel	no
01:16:14	kugel	it could just check for the stack value if overflow occured. our kernel does this too afterall
01:16:28	Torne	No, it shouldn't do
01:16:38	Torne	the kernel ABI shouldn't make any assumptions about SP
01:16:51	Torne	it should just save/restore it like all other registers
01:17:03	Torne	you detect overflows with huge guard page ranges on OSes with virtual memory :)
01:17:21	kugel	explain why the code is failing please :)
01:17:36	jhMikeS	well, one would hope it's done properly, but who knows
01:18:40	jhMikeS	do any other routines in rb do that?
01:20:03	kugel	it's a 2.6.24 kernel, I don't know how they did things back then
01:25:24	kugel	where does that #36 coe from when loading the parameter fro mthe stack?
01:27:58	saratoga	kugel: 9 registers are stacked
01:28:07	saratoga	so 36 bytes is the old top of the stack IIRC
01:28:20		Join ssxxaa [0] (~fbrucher@ip68-5-15-240.oc.oc.cox.net)
01:30:40		Part ssxxaa
01:37:22		Join tessarakt2 [0] (~jens@e180074195.adsl.alicedsl.de)
01:49:08		Quit tessarakt2 (Ping timeout: 240 seconds)
01:54:40		Quit Xerion (Read error: Connection reset by peer)
01:55:38		Join Xerion [0] (~xerion@5419F783.cm-5-2d.dynamic.ziggo.nl)
01:59:01		Quit Thra11 (Quit: kthxbai)
02:00
02:09:33		Quit tr\|nn\| (Read error: Operation timed out)
02:12:23		Quit Xerion (Read error: Connection reset by peer)
02:13:09		Join Xerion [0] (~xerion@5419F783.cm-5-2d.dynamic.ziggo.nl)
02:13:37		Quit prof_wolfff (Quit: CGI:IRC (Ping timeout))
02:32:40		Join thegeek [0] (~thegeek@87.37.34.95.customer.cdi.no)
02:36:46		Quit thegeek_ (Ping timeout: 268 seconds)
02:49:43		Quit Kiwi_CAM (Ping timeout: 240 seconds)
02:54:33		Quit CIA-88 (Ping timeout: 244 seconds)
03:00
03:00:33	***	Saving seen data "./dancer.seen"
03:09:11		Join Kiwi_CAM [0] (~Kiwi_Cam@189.60.247.184)
03:14:03		Join tessarakt2 [0] (~jens@e180074195.adsl.alicedsl.de)
03:17:27		Join CIA-15 [0] (~CIA@cia.atheme.org)
04:00
04:05:34		Quit Kiwi_CAM (Quit: Leaving)
04:12:36		Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:12:36		Quit pixelma (Disconnected by services)
04:12:36		Quit amiconn (Disconnected by services)
04:12:37		Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:12:39		Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:12:57		Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:23:52		Quit dys`` (Ping timeout: 252 seconds)
04:24:29		Join dys`` [0] (~andreas@krlh-5f713dce.pool.mediaWays.net)
04:25:17		Quit TheSeven (Disconnected by services)
04:25:42		Join [7] [0] (~TheSeven@rockbox/developer/TheSeven)
04:26:05	CIA-15	New commit by saratoga (r31643): Commit FS #12454 - Faster line scroll speed by Martin Sagmuller.
04:28:34	CIA-15	r31643 build result: 4 errors, 0 warnings (saratoga committed)
04:29:25		Quit krabador (Remote host closed the connection)
04:32:38	saratoga	sweet, jhmikes's fault
04:32:45	saratoga	well back to wrok
04:51:34		Join ungali [0] (~Samuel@S01060007e907c60a.cg.shawcable.net)
05:00
05:00:36	***	Saving seen data "./dancer.seen"
05:33:15		Nick perrikwp_ is now known as perrikwp (~quassel@cpe-024-163-024-033.triad.res.rr.com)
05:43:01		Quit curtism (Quit: Live Long and Prosper)
05:44:27		Part ungali
05:50:18		Quit Tuggers (Quit: Leaving)
05:51:07		Join Tuggers [0] (~joshbeck@pool-71-170-200-217.dllstx.fios.verizon.net)
05:53:42		Join Rob2222 [0] (~Miranda@pD9FAC1D0.dip.t-dialin.net)
05:57:25		Quit Rob2223 (Ping timeout: 248 seconds)
06:00
06:00:18	jhMikeS	erm, wtf, didn't see that
06:00:52		Quit MethoS- (Remote host closed the connection)
06:11:26	jhMikeS	and...it makes no real sense offhand
06:23:27	CIA-15	New commit by jethead71 (r31644): Fix Clip Zip bootloader build.
06:25:29	CIA-15	r31644 build result: All green
07:00
07:00:37	***	Saving seen data "./dancer.seen"
07:02:14		Join perrikwp_ [0] (~quassel@cpe-024-163-024-033.triad.res.rr.com)
07:02:39		Nick dys`` is now known as dys (~andreas@krlh-5f713dce.pool.mediaWays.net)
07:03:28		Quit froggyman (Ping timeout: 240 seconds)
07:03:34		Quit perrikwp (Read error: Operation timed out)
07:10:55		Join froggyman [0] (~froggyman@unaffiliated/froggyman)
07:38:00		Nick perrikwp_ is now known as perrikwp (~quassel@cpe-024-163-024-033.triad.res.rr.com)
07:40:34		Join Unhelpful_ [0] (~quassel@rockbox/developer/Unhelpful)
07:41:04		Join martii_ [0] (martii@sokrates.zagrodzki.net)
07:43:17		Join advcomp2019_ [0] (~advcomp20@97-114-237-193.sxcy.qwest.net)
07:43:17		Quit advcomp2019_ (Changing host)
07:43:17		Join advcomp2019_ [0] (~advcomp20@unaffiliated/advcomp2019)
07:44:02		Join remlap1 [0] (~Patrick@190.28.169.217.in-addr.arpa)
07:46:15		Quit martii (Ping timeout: 240 seconds)
07:46:15		Quit froggyman (.net .split)
07:46:15		Quit Tuggers (.net .split)
07:46:15		Quit tessarakt2 (.net .split)
07:46:22		Quit remlap (.net .split)
07:46:22		Quit dv_ (.net .split)
07:46:22		Quit bieber (.net .split)
07:46:22		Quit tchan (.net .split)
07:46:22		Quit jhMikeS (.net .split)
07:46:22		Quit scanf (.net .split)
07:46:22		Quit funman (.net .split)
07:46:22		Quit Utchybann (.net .split)
07:46:22		Quit yosafbridge (.net .split)
07:46:22		Quit sinthetek (.net .split)
07:46:22		Quit advcomp2019 (.net .split)
07:46:22		Quit mc2739 (.net .split)
07:46:22		Quit Slasheri (.net .split)
07:46:22		Quit Unhelpful (.net .split)
07:46:22		Quit polemon (.net .split)
07:46:22		Quit lembas (.net .split)
07:46:22		Quit aevin (.net .split)
07:49:45		Join Utchybann [0] (~Utchy@rps6752.ovh.net)
07:49:45		Join froggyman [0] (~froggyman@unaffiliated/froggyman)
07:49:45		Join Tuggers [0] (~joshbeck@pool-71-170-200-217.dllstx.fios.verizon.net)
07:49:45		Join tessarakt2 [0] (~jens@e180074195.adsl.alicedsl.de)
07:49:45		Join dv_ [0] (~quassel@chello080108009040.14.11.vie.surfer.at)
07:49:45		Join bieber [0] (~quassel@162-78.97-97.tampabay.res.rr.com)
07:49:45		Join tchan [0] (~tchan@lunar-linux/developer/tchan)
07:49:45		Join jhMikeS [0] (~jethead71@rockbox/developer/jhMikeS)
07:49:45		Join funman [0] (~fun@rockbox/developer/funman)
07:49:45		Join yosafbridge [0] (~yosafbrid@li125-242.members.linode.com)
07:49:45		Join sinthetek [0] (~sinthetek@unaffiliated/sinthetek)
07:49:45		Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
07:49:45		Join lembas [0] (toastin@inafire.com)
07:49:45		Join Slasheri [0] (miipekk@rockbox/developer/Slasheri)
07:49:45		Join polemon [0] (mcp@polemon.org)
07:49:45		Join aevin [0] (eivindsy@unaffiliated/aevin)
07:50:15		Quit dv_ (Max SendQ exceeded)
07:53:17		Join dv_ [0] (~quassel@chello080108009040.14.11.vie.surfer.at)
08:00
08:04:29		Quit factor (Read error: Connection reset by peer)
08:22:35		Join factor [0] (~factor@74.197.205.204)
08:39:06		Join GodEater_ [0] (93722cd0@rockbox/staff/GodEater)
08:42:56		Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
08:44:10		Join ender` [0] (~ender@foo.eternallybored.org)
08:55:19		Quit Scromple (Read error: Connection reset by peer)
09:00
09:00:39	***	Saving seen data "./dancer.seen"
09:02:42		Join nick_p [0] (~nick@82-69-105-120.dsl.in-addr.zen.co.uk)
09:08:54		Quit bertrik (Ping timeout: 240 seconds)
09:13:32		Join einhirn [0] (~Miranda@bsod.rz.tu-clausthal.de)
09:30:31		Join stoffel [0] (~quassel@pD9E41953.dip.t-dialin.net)
09:35:20		Join Guest8744 [0] (~bjst@46.35.227.87.static.tab.siw.siwnet.net)
09:35:20		Quit Guest8744 (Changing host)
09:35:20		Join Guest8744 [242] (~bjst@rockbox/developer/Zagor)
09:36:04		Nick Guest8744 is now known as Zagor (~bjst@rockbox/developer/Zagor)
09:51:42		Quit bieber (Quit: No Ping reply in 180 seconds.)
09:55:02		Join scanf [0] (~x32@unaffiliated/scanf)
09:55:32	kugel	Zagor: what's needed to add a new target _with_ new toolchain to the build system?
09:56:20	kugel	I added support for the toolchain to rockboxdev.sh but don't know how to get it to the build system
09:56:58	kugel	well, I guess I figured out the files that need to be be changed, but perhaps there's some button(svn up?) on the server that needs to be pressed
09:57:34		Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
10:00
10:00:28		Quit [Saint] (Client Quit)
10:00:36	kugel	also, I bumped the sdk requirement for android
10:00:46		Join [Saint] [0] (~Saint]@101.98.129.152)
10:00:46		Quit [Saint] (Changing host)
10:00:46		Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
10:01:48	kugel	while the command line tools don't seem to complain (your client still makes green android builds) the target version is now api level 15 and existing and new clients should have that
10:03:57	JdGordon	far more important.... git move :)
10:04:56	kugel	yea, torne is waiting for (one of) the admins
10:25:26		Join bieber [0] (~quassel@162-78.97-97.tampabay.res.rr.com)
10:32:18		Join efyx [0] (~efyx@lap34-1-82-225-185-146.fbx.proxad.net)
10:35:03		Join webguest76 [0] (~65628198@www.haxx.se)
10:38:27		Join merbanan [0] (~benjamin@2a00:801:f5:10:221:70ff:fecb:d6f9)
10:40:02		Quit webguest76 (Quit: CGI:IRC (Ping timeout))
10:43:54		Join wodz [0] (~wodz@iwl138.internetdsl.tpnet.pl)
10:44:43		Join webguest93 [0] (~65628198@www.haxx.se)
10:45:15	wodz	kugel: it would be nice if new toolchain added by you to rockboxdev.sh checked prerequisites properly.
10:45:26		Quit webguest93 (Client Quit)
10:47:29	JdGordon	what was the FS# for the theme crashing with usb connected?
10:47:32	*	JdGordon cant find it
10:47:39	kugel	wodz: what's wrong?
10:48:28		Join [Saint_] [0] (~Saint]@unaffiliated/saint/x-8516940)
10:49:05	wodz	kugel: I selected all toolchains to build and the only one which failed due to missing tools was new RaaA toolchain
10:49:12	nick_p	JdGordon: Was it FS #12475?
10:49:13	fs-bluebot	http://www.rockbox.org/tracker/task/12475 Crash while playing audio (bugs, unconfirmed)
10:49:26		Part merbanan ("Leaving")
10:49:32	kugel	wodz: can you please be more specific? what was missing?
10:50:05	kugel	I didn't know that has more dependencies than the other toolchains
10:50:11	JdGordon	nick_p: hmm, no... though with that one, if its crashing with audio playing it means its a buflib issue
10:50:16	wodz	kugel: autoconf, texinfo, g++ packages If I recall correctly
10:51:03	wodz	ah and ncurses
10:51:21	kugel	I think we rockboxdev.sh reports only dependencies that are additional to what the toolchains need to build
10:51:30	kugel	i.e. patch and wget/curl
10:51:50	kugel	ah, ncurses might be needed for ct-ng right
10:51:57		Quit [Saint] (Ping timeout: 244 seconds)
10:52:00	wodz	curl is needed as well while rockboxdev can use wget as well
10:52:00	nick_p	JdGordon: I was just thinking of funman's comment: http://www.rockbox.org/tracker/task/12475#comment42006
10:53:32	JdGordon	nick_p: hmm, ok, I still suspect buflib issues
10:53:58	JdGordon	I thought there was a probably bug with usb not unloading fonts correctly causing crashes
10:54:12	*	JdGordon hasnt been paying much attention recently
10:54:13	wodz	kugel: it is not a big issue but since you have to delete temp dirs by hand it is a bit annoying to iteratively resolve dependencies :-)
10:55:25	wodz	and BTW ct-ng builds toolchains using all cores while rockboxdev is not
11:00
11:00:42	***	Saving seen data "./dancer.seen"
11:07:28	JdGordon	why is the fuze+ sim not being built?
11:09:22	JdGordon	GeekShadow: ?
11:09:30	JdGordon	gevaerts: even ^?
11:10:53	gevaerts	Because nobody added it to the builds file?
11:11:05	*	gevaerts blames Zagor :)
11:11:24	JdGordon	it doesnt build currently, looking into it
11:14:18	kugel	wodz: it shouldnt. it should use MAKEFLAGS
11:15:27	wodz	kugel: ee I lost context
11:16:10	CIA-15	New commit by jdgordon (r31645): r31444 missed a file. make fuzeplus sim build again
11:18:40	CIA-15	r31645 build result: All green
11:22:51		Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
11:23:40	kugelp	wodz: about ct-ng using all cores
11:24:25		Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
11:25:03	wodz	kugel: ct-ng definitely used all cores while other builds not
11:25:24		Part kugel
11:25:41		Join kugel [0] (~kugel@rockbox/developer/kugel)
11:25:46		Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
11:26:11		Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
11:26:21	JdGordon	funman: ping?
11:27:59		Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
11:28:11		Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
11:28:33		Quit kugel (Quit: leaving)
11:28:38		Join kugel [0] (~kugel@rockbox/developer/kugel)
11:29:39		Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
11:30:43		Nick [Saint_] is now known as [Saint] (~Saint]@unaffiliated/saint/x-8516940)
11:32:07		Quit adnap (Read error: Connection reset by peer)
11:37:57		Join [Saint_] [0] (~Saint]@101.98.129.152)
11:37:57		Quit [Saint_] (Changing host)
11:37:57		Join [Saint_] [0] (~Saint]@unaffiliated/saint/x-8516940)
11:39:14		Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
11:40:08		Join adnap [0] (~adnap@rrcs-71-42-140-57.sw.biz.rr.com)
11:40:30		Quit [Saint] (Ping timeout: 276 seconds)
11:44:56		Quit [Saint_] (Quit: Quit)
11:45:20		Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
11:48:57		Quit [Saint] (Client Quit)
11:51:06		Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
11:57:22		Join [Saint_] [0] (~Saint]@101.98.129.152)
11:57:27		Quit [Saint_] (Changing host)
11:57:28		Join [Saint_] [0] (~Saint]@unaffiliated/saint/x-8516940)
12:00
12:01:02		Quit [Saint] (Ping timeout: 240 seconds)
12:07:28		Join ghallberg [0] (~gustaf@h206n2-g-kt-d1.ias.bredband.telia.com)
12:07:56	ghallberg	saratoga: Hey, I posted on the forums a while back about headphone detection on AS3525 devices.
12:11:34		Nick [Saint_] is now known as [Saint] (~Saint]@unaffiliated/saint/x-8516940)
12:17:08		Quit stoffel (Ping timeout: 248 seconds)
12:38:12		Join [Saint_] [0] (~Saint]@unaffiliated/saint/x-8516940)
12:38:33	Torne	Zagor: i think we're almost ready to go ahead with git transition; the one major thing we don't have that would be really useful is gitweb/equivalent
12:39:39	Zagor	Torne: yeah, I'll set it up. but it's not a blocker.
12:39:52		Quit [Saint] (Ping timeout: 252 seconds)
12:40:01		Nick [Saint_] is now known as [Saint] (~Saint]@unaffiliated/saint/x-8516940)
12:40:26	Torne	Well, no, but it'd be nice to get it up and integrated with gerrit
12:40:39	Zagor	yeah
12:41:02		Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
12:42:17		Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
12:52:35	Torne	anyway, what do you want to do next?
12:52:46	Torne	I don't think there's anything else that particularly needs to be done before the transition
12:53:03	Torne	version.sh returns something sensible, gerrit is all ready once the permissions get changed to allow committers to commit
12:53:17	Torne	the docs are present and nobody has complained about them, but i haven't yet changed other wiki pages to link to them
12:53:36	Torne	we need to do gitweb, build system, website in gneeral, but those can/should happen after
12:53:47	Torne	so is the next step to just shut off commit access to SVN and enable it on gerrit?
12:58:10		Join jlbiasini [0] (~metaphys@d86-32-96-55.cust.tele2.at)
12:59:53		Quit nick_p (Quit: Leaving)
13:00
13:00:45	***	Saving seen data "./dancer.seen"
13:03:22		Join dfkt [0] (dfkt@unaffiliated/dfkt)
13:07:25	JdGordon	Torne: woo! yes! go for it!
13:08:06	Torne	when do we (read: Zagor and me) want to do that? :)
13:10:11	Zagor	tonight?
13:10:46	Torne	OK :)
13:11:02		Join stoffel [0] (~quassel@pD9E41953.dip.t-dialin.net)
13:11:17		Quit [Saint] (Quit: Quit)
13:11:25	Zagor	I'll probably be here from around 20:00 CET
13:12:17	*	Torne nods
13:18:56	jlbiasini	Zagor: We need someone to update the server with the new bootloader for fuze+ since rockbox utility was updated to handle it
13:19:23	Zagor	someone == me
13:19:32	Zagor	where do I get the new file?
13:19:33	jlbiasini	hihi
13:19:55	jlbiasini	I tried to contact bagder since several days but no success
13:20:08	jlbiasini	wait I have to find the link again
13:20:35	dfkt	saratoga, thanks for committing my amazingly sophisticated line scroll speed patch :)
13:21:55	jlbiasini	Zagor: there it is http://www.mediafire.com/?pjpwe8pi8hdah3n
13:23:44		Join TheLemonMan [0] (~LemonBoy@ppp-1-10.26-151.libero.it)
13:24:33	jlbiasini	zagor: it should go there: sandisk-sansa/fuzeplus/bootloader-fuzeplus.sansa
13:24:49	Zagor	done
13:24:55	jlbiasini	thx
13:27:44		Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
13:28:34	jlbiasini	Ok this is working! we have rbutil working on fuze+ now we will need to bump it to unstable!
13:29:06	Zagor	yay
13:31:59	jlbiasini	Zagor: do you know what I have to do for this?
13:33:26	gevaerts	jlbiasini: update the home page and build.pm I think, and also the general target status wiki pages I guess
13:33:37		Join y4n [0] (y4n@unaffiliated/y4ndexx)
13:34:27	jlbiasini	gevaerts: I don't think I have wright on the home page, where is this build.pm?
13:35:03	gevaerts	in tools/
13:36:39		Nick [Saint] is now known as testtesttest (~Saint]@unaffiliated/saint/x-8516940)
13:37:53		Quit testtesttest (Quit: Quit)
13:38:19		Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
13:41:46	jlbiasini	ok but I'm not sure if all the other files are ready bluebrother was talking about some fuzeplus unstead of sansafuzeplus problem. I don't know if it also concern rockbox utility
13:42:04	jlbiasini	I'll wait for him to show up
13:49:17	Zagor	jlbiasini: the home page is http://svn.rockbox.org/viewvc.cgi/www/index.t?view=log
13:51:06	jlbiasini	ok patch for builds here FS #12526 (I just add a field for fuze+ assuming the name are consistent with other target but I don't know If we actually have those file (icon, name definition...)
13:51:07	fs-bluebot	http://www.rockbox.org/tracker/task/12526 put Fuze+ in unstable (patches, unconfirmed)
13:52:08	jlbiasini	hum in which order do I proceed? I believe that it is better to do first all the change on the server then all the wiki stuff, isn't it?
13:52:58	Zagor	it doesn't really matter
13:53:19	Zagor	just tell me when you've updated index.t because I need to manually pull it to show on the web site
13:53:32		Join lorenzo92 [0] (~chatzilla@host143-111-dynamic.21-79-r.retail.telecomitalia.it)
13:54:12	jlbiasini	well I think I will wait for bluebrother first... just to be sure I'm not doing crap
13:54:49	Zagor	whee: http://git.rockbox.org/
13:54:54	jlbiasini	I would be a pity to update everything to find something not working
13:58:03	jlbiasini	bluebrother: so now that we have the bootloader on server do we need anything more before having fuze+ in unstable?
13:59:01	wodz	Zagor: nice
14:00
14:00:45	jlbiasini	Zagor: \o/ :/ I still have to learn how to use it... (yeah I know wiki...:D)
14:02:11	GodEater_	Zagor: yeah, but where's the gerrit interface ? :D
14:02:20	Zagor	http://gerrit.rockbox.org
14:02:35	*	GodEater_ takes back his smart alec comment
14:02:39	Zagor	:)
14:03:01	GodEater_	ah, but it's still the demo one right?
14:04:45	Zagor	no it's the one. but we've kept the sandbox playground.
14:06:37	GodEater_	so we can commit there now?
14:06:45	GodEater_	s/commit/push/
14:06:55	GodEater_	and the code review bits all work?
14:08:02	Zagor	i'll let torne answer that :)
14:14:43		Part jlbiasini
14:21:14		Join desowin_ [0] (~desowin@atheme/member/desowin)
14:23:11		Quit wodz (Quit: Leaving)
14:23:13		Quit tessarakt2 (Ping timeout: 240 seconds)
14:23:31	Torne	Zagor: there's some foo or other in gerrit's config to tell it where gitweb is, so hyperlinks work
14:23:51	Torne	GodEater_: you can't commit there yet
14:23:57	Torne	GodEater_: we're gonna do it this evening :)
14:24:01	GodEater_	\o/
14:29:49	Torne	i love how gitweb's main page for a repo shows you something that looks like the history, but is just nonsense.
14:30:08	Torne	Well, right now for us it's suspiciously similar
14:30:16	Torne	but in general.. not so much
14:30:55		Quit lorenzo92 (Quit: ChatZilla 0.9.88 [Firefox 9.0.1/20111228084940])
14:31:06	Torne	("recent commits in this repository in chronological order" is an.. interesting thing to display in a system that encourages branching)
14:33:46	Torne	Also, I'm pretty sure we have more than 32 active committer
14:34:04	Torne	Everyone: if you are not yet in the Rockbox Committers group on gerrit, now is the time to ask :p
14:36:08	*	GodEater_ just checked, and is definitely in it.
14:36:32	Torne	All the people that immediately come to mind (i.e. people whose names i recognise) are on it
14:36:38	GodEater_	did anyone get around to playing with the CSS for it yet to get it into the Rockbox theme?
14:36:45	Torne	but since that boils down to more or less "most of you that i have physically met" then that's not really complete
14:36:52	Torne	I don't think anyone actually did
14:37:03	Torne	there were some complaints about the limited flexibility of its html insertion
14:37:03	Torne	:)
14:37:16	Torne	feel free
14:37:16	GodEater_	people here are never happy :D
14:37:26	Torne	docs for it are accessible in gerrit
14:37:31	GodEater_	what I know about CSS can be written on a partially torn postage stamp
14:37:39	Torne	http://gerrit.rockbox.org/r/Documentation/config-headerfooter.html
14:37:45	GodEater_	and I've not the inclination to learn to be honest ;)
14:40:14		Join LinusN [0] (~linus@giant.haxx.se)
14:40:35	Torne	Zagor: looks like we just need to set gitweb.url in gerrit config to get it to generate links
14:40:45	Zagor	yeah, just did that
14:40:48		Join jlbiasini [0] (~metaphys@d86-32-96-55.cust.tele2.at)
14:41:11	Torne	also can we kill Woummamains off the forum?
14:41:17	Torne	they've been spamming for a while :/
14:41:35	*	GodEater_ goes to kill
14:42:06	Torne	i think akismet is killing their posts, but it still bumps the threads
14:42:44	GodEater_	point me at a post of theirs?
14:42:51	Torne	i don't think any still exist
14:42:52	GodEater_	I can't see that idea immediately obvious as a spammer
14:43:44	Torne	look on http://forums.rockbox.org/index.php?action=unread;all;start=0
14:43:49	Torne	if you haven't already read everything :)
14:44:07	Torne	they've posted to dozens of existing threads in the last 24 hours, but their posts appear to have all been killed by the antispam system
14:44:13	GodEater_	he's gone
14:44:15	Torne	so it just uselessly bumps them all in unread
14:44:20	Torne	oh
14:44:22	Torne	ok then :)
14:44:29	*	GodEater_ blows smoke from his shotgun
14:48:06		Join MethoS- [0] (~clemens@134.102.106.250)
14:57:49		Quit stoffel (Remote host closed the connection)
14:59:43		Quit godzirra (Ping timeout: 240 seconds)
15:00
15:00:48	***	Saving seen data "./dancer.seen"
15:00:57		Join godzirra [0] (~shawn@184.7.68.9)
15:18:36		Join liar [0] (~liar@clnet-p09-185.ikbnet.co.at)
15:25:01		Quit remlap1 (Quit: Leaving.)
15:25:14		Join remlap [0] (~Patrick@190.28.169.217.in-addr.arpa)
15:27:15	kugel	saratoga: the other asm in libmad works just fine, including the two other functions in that file
15:27:30	kugel	is it really the re-use of sp as gp reg that breaks it?
15:27:41		Quit anewuser ()
15:40:17	kugel	how do I push onto stack on arm? str rN, sp? or do I need to increment before?
15:40:58		Quit T44 (Read error: Connection reset by peer)
15:41:28		Quit Lonely_Dragon (Ping timeout: 252 seconds)
15:41:48		Join Topy44 [0] (~Topy44@f048041199.adsl.alicedsl.de)
15:48:46		Join Thra11 [0] (~thrall@87.115.59.177)
15:49:52	Torne	kugel: str rN, [sp], #-4
15:49:56	Torne	to save one register
15:50:11	Torne	stmfd sp!, {r0-r3,r11} or similar to save more than one
15:50:32	Torne	ARM stacks are full-descending, so you need to post-decrement sp
15:51:42	desowin_	I thought you can do FA, FD, EA, ED on ARM
15:51:50	Torne	You can, but you don't
15:51:54	kugel	"The stack pointer can either point to the last item in the stack (a full stack)"
15:51:59	Torne	EABI requires fulldescending, and Thumb code only understands fulldescending
15:52:02	Torne	Nobody uses anything else
15:52:05	kugel	doesn't that mean I need pre-increment?
15:52:13	Torne	no, the stack grows downward
15:52:20	Torne	Wait
15:52:21	kugel	pre-decrement*
15:52:23	Torne	Yeah
15:52:24	Torne	Sorry :)
15:52:32	Torne	str rN, [sp, #-4]!
15:52:44	Torne	i'm doing too many things at once :)
15:53:00	kugel	okay
15:54:56	kugel	meh
15:55:22	kugel	I cant replace sp with r0 in this code. there's an ldm that needs the data in sp
15:55:56	Torne	what do you mean?
15:56:17	kugel	ldmia r1!, {r11, r12, r0, lr} doesnt work
15:56:37	Torne	It doesn't work because the arguments to ldm have to be in order
15:56:48	kugel	yep
15:56:56	Torne	So rearrange all the registers so you can load them in order :)
15:57:21	Torne	i.e. use r0, r11, r12, lr, and fix all references to r11-13 elsewhere
15:57:43	kugel	easy!
15:57:52	Torne	:)
15:59:11	kugel	saratoga: ^ :)
15:59:38	Torne	if you just want to test this, then you could just break up the ldm instruction
15:59:44	Torne	to see if the problem goes away
15:59:49	Torne	then fix it properly later for performance
16:00
16:00:00	kugel	good plan
16:00:24	Torne	ldmia r1!, {r11, r12}; ldr r0, [r1], #4, ldr lr, [r1], #4
16:05:07		Quit hurdman (Quit: leaving)
16:11:16	kugel	aha. with the sp use changed it doesnt crash anymore
16:11:27	Torne	oh dear.
16:11:45	kugel	but...why?
16:11:58	Torne	a very good question ;/
16:12:03	Torne	I'm pretty sure the kernel shouldn't care
16:12:21	kugel	perhaps freescale added it to the kernel and/or enabled some config
16:12:57	jhMikeS	it was the sp thing?
16:14:15	jhMikeS	o_O
16:14:33	kugel	looks like it
16:18:05	kugel	perhaps the memory location where the sp is stored at is corrupted?
16:18:56		Join anewuser [0] (~anewuser@186.93.165.98)
16:18:56		Quit anewuser (Changing host)
16:18:56		Join anewuser [0] (~anewuser@unaffiliated/anewuser)
16:20:28	jhMikeS	make two copies and compare them after the reload?
16:20:32		Join LambdaCalculus37 [0] (~rmenes@rockbox/staff/LambdaCalculus37)
16:20:39	jhMikeS	or move it somewhere else
16:26:22	kugel	moving from .bss to .data (verified in the .map) => still crashing
16:27:14	Torne	i haven't looked at the code at all, but are you sure you are not inadvertantly manipulating sp somewhere else? :)
16:28:19	kugel	it's pretty safe. it stores sp to some location in the prologue, and loads from that location just before return
16:28:28	kugel	in the meantime it of course messes wildly with spo
16:30:58	kugel	apps/codecs/libmad/synth_full_arm.S
16:31:15		Quit bieber (Read error: Connection reset by peer)
16:32:53		Nick Jack87 is now known as Jack87\|Away (Jack87@nasadmin/admin/jack87)
16:33:45		Join bieber [0] (~quassel@cpe-72-185-229-148.tampabay.res.rr.com)
16:35:06	kugel	Torne: this code works flawlessly on a couple dozen existing targets so an error on the code should have shown up earlier
16:38:13	Torne	and it appears to die when you are being preempted by the kernel on a reasonably modern ARM linux?
16:39:20	kugel	it's a guess. can't say it "appears to"
16:39:26	Torne	ok
16:39:29	*	Torne has a peek
16:39:40	kugel	actually I'd be surprised if we
16:40:42	kugel	're preempted at all. there're aren't really other tasks running and we normally don't use our time slice fully
16:40:50	kugel	I expect
16:40:58	Torne	but changing it to someting other than sp makes it work, in any case? :)
16:41:24	kugel	yea
16:41:47	kugel	I changed it to r0 because r0 is only used at the end of the function
16:44:33	Torne	well, if it's possible to spill r0 instead of sp at no extra cost of instructions, i would encourage doing that anyway
16:44:47	Torne	just because using sp as a general purpose register looks scary
16:45:03	Torne	i really can't see why this would break though
16:45:30	Torne	Ah. Uhm. Unless you're taking signals.
16:45:39	Torne	Haha. That might be it :p
16:45:46	Torne	Are there any signal handlers in the process?
16:46:03	Torne	because by default your signal handler will run on the user stack :):)
16:46:10	Torne	unless you point it somewhere else with sigaltstack
16:46:28	Torne	so if you're using sp as a GPR and you take a signal the handler will promptly crap itself
16:49:18	jhMikeS	doesn't alsa do signals?
16:50:16	Torne	try blocking all signals during the code :p
16:50:31	Torne	if that also fixes it then you have a proof :p
16:52:50		Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
16:53:06	kugelp	Torne: ah genius
16:53:09	kugelp	!
16:54:04	Torne	did that work?
16:54:21		Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org)
16:55:05	kugelp	changed alsa to not use signals and it works now
16:55:23	jhMikeS	didn't it work on other targets or was it crapping on others?
16:55:44	Torne	Okay, so that's a nice big warning sign to slap on things then
16:56:04	Torne	"Don't fuck with the stack pointer on RaaA unless you know your thread can't take signals, or have set a signal stack"
16:56:53	Torne	sigaltstack in threaded code has some horrible caveats as a bonus.
16:57:10	Torne	who invented this api
16:57:35	Torne	sigaltstack is thread specific (good) but if you set one then threads you create inherit the same one from you (VERY VERY BAD)
16:57:59	Torne	that seems suspiciously like it might create an unavoidable race
16:58:43	Torne	Huh. Except pthread_create(3) disagrees with pthreads(7)
16:58:45	Torne	ohwell
16:59:13		Quit anewuser ()
16:59:23	Torne	aha. fixed in kernel 2.6.16 :)
16:59:27	Torne	lololol
16:59:33		Part Zagor
16:59:36	Torne	kernel version dependent concurrent-program correctness!
16:59:50		Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
17:00
17:00:04	jhMikeS	send them out in the desert on a mule for it
17:00:25	Torne	kugel: anyway, my original statement stands: since the program is saving/restoring sp anyway, if it's possible to change the register assignments such that sp is left alone and it just saves/restores r0, without making the code slower, then please do so :)
17:00:52	***	Saving seen data "./dancer.seen"
17:03:03	Torne	i dunno what alsa uses signals for, but if it's the default behaviour then maybe it's desirable? :)
17:03:29	jhMikeS	for the audio callback, if you want it that way
17:03:31		Join anewuser [0] (~anewuser@186.93.165.98)
17:03:32		Quit anewuser (Changing host)
17:03:32		Join anewuser [0] (~anewuser@unaffiliated/anewuser)
17:04:04		Join wodz [0] (~wodz@89-76-160-35.dynamic.chello.pl)
17:04:04	Torne	for feeding it our PCM buffer?
17:04:13	Torne	i suspect we probably do want that delivered in a signal, if possible
17:04:25	Torne	by analogy to low watermark interupts on hardware? :)
17:04:28	jhMikeS	yeah, whenever audio is in need of more data
17:05:30	wodz	Who is in charge of ircviewer?
17:09:06	Torne	also, using sp as a general purpose register is forbidden in later verisons of ARM, i think
17:09:32	Torne	allowing it causes too many problems for things like branch return stack prediction
17:10:03	Torne	even if not forbidden, advised against for performance reasons
17:12:02		Join Lonely_Dragon\|2 [0] (~kvirc@128-70-24-84.broadband.corbina.ru)
17:17:23	mc2739	wodz: code is here: http://svn.rockbox.org/viewvc.cgi/www/irc/reader.pl?view=log
17:17:33		Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
17:19:10		Quit nosa-j (Ping timeout: 252 seconds)
17:19:29		Part ghallberg
17:19:52		Join nosa-j [0] (~m00k@adsl-74-235-79-157.clt.bellsouth.net)
17:20:35	kugelp	Torne: alsa recommends against signals
17:20:59	Torne	..they do?
17:21:00	Torne	why?
17:23:38	kugelp	can't remember
17:24:41	kugelp	my pcm-alsa implementation can also work via polling (e.g. in tick task)
17:25:30	kugelp	one using a separate thread is doable
17:26:07	Torne	i suspect that most of the arguments for not using signals to do alsa are about how difficult that kind of thing is
17:26:21	Torne	which is logic that doesn't really apply to us since presumably our PCM system is already interrupt-capable
17:26:24	Torne	and that's basically the same thing
17:26:51	kugelp	Torne: i think their docs don't even mention the async callback from a signal handler methodanymore
17:28:42	kugelp	well, its not capable of signals on user stack :p
17:29:46	kugelp	well, the asm code isn't
17:30:08	kugelp	I wonder if we have more asmcode like this
17:30:57	jhMikeS	Not much if any other
17:36:35		Join n1s [0] (~n1s@nl118-175-223.student.uu.se)
17:36:35		Quit n1s (Changing host)
17:36:35		Join n1s [0] (~n1s@rockbox/developer/n1s)
17:36:42	jhMikeS	libspeex might have some trouble spots
17:37:16	jhMikeS	apps/codecs/libspeex/filters_arm4.S comment starting line 212
17:38:54	*	kugelp doesn't feel like auditing all asm code
17:39:32	kugelp	could just install an altstack
17:40:25	jhMikeS	at least preglow thought to comment on something funny going on :)
17:41:13	Torne	kugelp: bear in mind the horrible caveats about sigaltstack with threads i just mentioned :)
17:45:46		Quit liar (Ping timeout: 252 seconds)
17:47:03		Join liar [0] (~liar@clnet-p09-185.ikbnet.co.at)
17:55:39		Quit n1s (Read error: Connection timed out)
18:00
18:01:54		Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
18:05:53		Join n1s [0] (~n1s@nl118-175-223.student.uu.se)
18:05:53		Quit n1s (Changing host)
18:05:53		Join n1s [0] (~n1s@rockbox/developer/n1s)
18:08:49		Part jlbiasini
18:32:48		Join Floweraddict [0] (~Flowerpix@c-98-244-27-34.hsd1.ca.comcast.net)
18:33:30	Floweraddict	Hi hi?
18:33:36		Quit guymann (Ping timeout: 252 seconds)
18:33:39		Join guymann_ [0] (~charles@64-252-123-137.adsl.snet.net)
18:34:36	Floweraddict	Anyone alive here?
18:37:03	[Saint]	Many.
18:38:39	[Saint]	Just as your (Rockbox related) question.
18:39:13	[Saint]	If someone is available, and can answer, they'll do so.
18:40:53		Quit [7] (Disconnected by services)
18:41:08		Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
18:43:35		Quit TheSeven (Disconnected by services)
18:43:53		Join [7] [0] (~TheSeven@rockbox/developer/TheSeven)
18:46:41		Quit [7] (Disconnected by services)
18:47:04		Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
18:59:40		Join Martus [0] (~chatzilla@afoj168.neoplus.adsl.tpnet.pl)
19:00
19:00:55	***	Saving seen data "./dancer.seen"
19:01:24	n1s	TheSeven: well, it did happen, i tried twice but now i've installed the new emcore and it's reformatted the drive
19:02:00		Quit liar (Read error: Connection timed out)
19:02:35	TheSeven	hm, interesting
19:03:33	Martus	it's great that finally rockbox on fuze v2 can handle usb connection, buts is there any way to switch to "only charge" mode?
19:03:55	n1s	Martus: hold a button pressed when inserting the usb plug
19:03:56	[Saint]	Yes.
19:04:16	Martus	nls which button?
19:04:41	TheSeven	any one should work
19:04:44	n1s	any regular button should do it
19:04:53	[Saint]	except power, iiuc.
19:04:59	Martus	ok I'll check it
19:05:20	[Saint]	The manual describes this in detail. Or should.
19:06:20	Martus	wow! it works thank you very much. I cannot find any clue how to switch only charge mode
19:06:30	n1s	[Saint]: it does
19:06:41	n1s	http://download.rockbox.org/daily/manual/rockbox-sansafuzev2/rockbox-buildch3.html#x5-370003.3
19:07:22	Martus	I haven't seen that. Will be more carefull next time
19:08:30	[Saint]	To be fair, the search function isn't that useful unless you manage to guess the exact right search term.
19:08:56	[Saint]	"charge" should've found it though.
19:12:05	Martus	I promise to look better next time. By the wat in charge mode windows 7 is detecting my fuze as "rockbox media player" but can't find driver for that. On rockbox.com can I find right driver? I have x64.
19:12:53		Quit remlap (Ping timeout: 240 seconds)
19:13:16		Join remlap [0] (~Patrick@190.28.169.217.in-addr.arpa)
19:14:14		Quit desowin_ (Read error: Connection reset by peer)
19:26:05	Martus	I've tried with other computer win 7 x86 the same situation. Can't find driver using google.
19:28:36		Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
19:28:50		Join Zagor [242] (~bjst@rockbox/developer/Zagor)
19:28:54		Part kugelp
19:31:01	saratoga	damn, new c200 is a v1
19:34:49	[Saint]	bummer.
19:35:30	[Saint]	Still cool, like, yay c200...but, bummer its a v1
19:36:26		Quit TheLemonMan (Quit: WeeChat 0.3.6)
19:36:27	n1s	Martus: you don't need a driver
19:36:55	Martus	My two windows 7 can't perform instalation
19:37:32		Join TheLemonMan [0] (~LemonBoy@ppp-1-10.26-151.libero.it)
19:38:37	Martus	but it's a trifle. The most important: usb and charge mode work : )
19:41:02		Quit bluebrother (Disconnected by services)
19:41:03		Join bluebrother^ [0] (~dom@rockbox/developer/bluebrother)
19:42:10		Quit Floweraddict (Remote host closed the connection)
19:42:31		Join Floweraddict [0] (~Flowerpix@c-98-244-27-34.hsd1.ca.comcast.net)
19:42:54		Quit fs-bluebot (Ping timeout: 240 seconds)
19:44:18		Join lebellium [0] (~chatzilla@91-65-137-216-dynip.superkabel.de)
19:44:30		Join fs-bluebot [0] (~fs-bluebo@g225254122.adsl.alicedsl.de)
19:47:23		Join Keripo [0] (~Keripo@CPE0022b0d4bdb7-CM001a6680d4fe.cpe.net.cable.rogers.com)
19:50:26		Join prof_wolfff [0] (~d525d9d6@www.haxx.se)
19:51:27		Quit Martus (Quit: ChatZilla 0.9.88 [Firefox 9.0.1/20111220165912])
19:58:31		Quit Xerion (Quit: )
19:58:35		Join Xerion [0] (~xerion@5419F783.cm-5-2d.dynamic.ziggo.nl)
20:00
20:02:20	saratoga	dfkt: no problem, i would have done it sooner but i forgot about that patch
20:04:02		Nick Unhelpful_ is now known as Unhelpful (~quassel@rockbox/developer/Unhelpful)
20:05:14		Join tessarakt2 [0] (~jens@e180074195.adsl.alicedsl.de)
20:07:38		Join cereal_killer [0] (~b2bfced8@www.haxx.se)
20:08:11		Quit Zarggg (Ping timeout: 268 seconds)
20:08:55	cereal_killer	Hello rockbox community, have there recently been reports about "playlist buffer full" problems in 3.10?
20:10:17	cereal_killer	I incresed the limits, but I am still not able to play a new file selected from the file browser, whe I was just saving a playlist before
20:11:47		Join Zarggg [0] (~zarggg@24.229.139.169.res-cmts.sm.ptd.net)
20:12:56	cereal_killer	I even reinstalled rockbox, because of a maybe corrupted .playlist_control file
20:14:18	gevaerts	cereal_killer: just reinstalling doesn't delete files that aren't part of the installation, so wouldn't clear a bad .playlist_control file
20:15:40	cereal_killer	ah right, this file is created after installation, so I try to delete it then and have it recreated
20:18:48	cereal_killer	ok, deleted this file, and retried, but no luck, still get the warning, I also checked the filsystem of my clipv2 a few days ago, but windows gave me no error messages
20:19:44		Quit Keripo (Quit: Leaving.)
20:23:32	cereal_killer	still can't select a file in the file browser to play, when playing a saved playlist, any ideas what could cause this?
20:23:43	n1s	bugs?
20:25:55	cereal_killer	alright, I'll create a bug report then...
20:28:24		Quit Floweraddict (Remote host closed the connection)
20:28:53		Join Floweraddict [0] (~Flowerpix@c-98-244-27-34.hsd1.ca.comcast.net)
20:29:56		Quit factor (Read error: Connection reset by peer)
20:32:54		Join factor [0] (~factor@74.197.205.204)
20:33:33		Quit tessarakt2 (Ping timeout: 240 seconds)
20:36:17		Join FlynDice [0] (~jack@c-24-19-225-90.hsd1.wa.comcast.net)
20:37:59		Quit cereal_killer (Quit: CGI:IRC)
20:46:11		Join tessarakt2 [0] (~jens@e180068152.adsl.alicedsl.de)
20:56:54		Quit ze (Ping timeout: 268 seconds)
20:58:45		Quit Floweraddict (Remote host closed the connection)
20:59:04		Join Floweraddict [0] (~Flowerpix@c-98-244-27-34.hsd1.ca.comcast.net)
21:00
21:00:13		Quit Zarggg (Quit: Rebooting client...)
21:00:57	***	Saving seen data "./dancer.seen"
21:01:11		Join ze [0] (ze@ze.yi.org)
21:02:56		Join liar [0] (~liar@clnet-p09-185.ikbnet.co.at)
21:04:54		Join Zarggg [0] (~zarggg@24.229.139.169.res-cmts.sm.ptd.net)
21:05:12		Join Keripo [0] (~Keripo@CPE0022b0d4bdb7-CM001a6680d4fe.cpe.net.cable.rogers.com)
21:12:55		Quit Floweraddict (Remote host closed the connection)
21:13:08		Join Floweraddict [0] (~Flowerpix@c-98-244-27-34.hsd1.ca.comcast.net)
21:18:47		Quit y4n (Quit: PANTS OFF!)
21:23:50		Quit Zarggg (Ping timeout: 244 seconds)
21:24:45		Join Zarggg [0] (~zarggg@24.229.139.169.res-cmts.sm.ptd.net)
21:32:20		Quit [Saint] (Ping timeout: 240 seconds)
21:39:53	FlynDice	Torne: Would you please add me to the Rockbox Committers group on gerrit.
21:41:47		Quit FlynDice (Remote host closed the connection)
21:42:15		Quit liar (Ping timeout: 252 seconds)
21:43:06		Join liar [0] (~liar@clnet-p09-185.ikbnet.co.at)
21:48:22	saratoga	TheSeven: did you see the new ipod6g patches?
21:48:40	TheSeven	you mean the crash fix + voltage scaling?
21:48:58	TheSeven	didn't have time to look at that, but it seems to me like it needs some rework of the clickwheel driver that might affect other targets as well
21:49:19	TheSeven	so this should probably be reviewed by some more people
21:51:35	wodz	TheSeven: could you elaborate?
21:52:09	TheSeven	reading the hold switch involves I2C on the classic, which uses mutexes => is not safe to be run from a tick task (which is run from an IRQ handler)
21:52:29	TheSeven	this seems to be the reason for all of those weird random lockups we've been observing
21:53:11	bertrik	sometimes I wonder if we should have a sanity check on calling conditions of the the kernel primitives
21:53:41	TheSeven	at least some panic'ing assertions when compiled as a debug build would be great
21:54:03	wodz	bertrik: what is the meaning of debug menu for tea fm chip?
21:54:43	wodz	TheSeven: what about making i2c driver irq safe?
21:54:54	bertrik	wodz, I don't remember exactly, but I think it helps to see the raw values
21:55:13	wodz	bertrik: I mean does read and write should match?
21:55:20	TheSeven	that means locking out IRQs completely during rather lengthy I2C transactions
21:55:56	TheSeven	this eats a lot of CPU power and might even be sufficient to cause audio stuttering
21:56:11	TheSeven	I'd rather say move the hold switch readout from a tick task to a thread
21:56:22	TheSeven	or the tick tasks altogether
21:56:32		Quit prof_wolfff (Quit: CGI:IRC (Ping timeout))
21:56:42	TheSeven	there is just no good reason for that to be done inside an IRQ handler
21:57:01	wodz	TheSeven: can't you just setup async i2c transfer?
21:57:25	TheSeven	I can't do an I2C transfer while another one is active
21:57:27	bertrik	wodz, I don't remember
21:57:49	TheSeven	so if an I2C-doing IRQ comes in while something else accesses I2C => bang!
21:58:03	TheSeven	that's what happened until now
21:58:35	bertrik	good to know we know now :)
21:58:57	wodz	TheSeven: for hold switch you can simply ignore and read next time when i2c is idle
21:59:10	TheSeven	yeah, but that's an awful hack
21:59:22	TheSeven	I just see no sane reason for this to be done from the timer IRQ
21:59:22		Quit Floweraddict (Remote host closed the connection)
21:59:27	TheSeven	get it out of there and things will be fine
21:59:37		Join Floweraddict [0] (~Flowerpix@c-98-244-27-34.hsd1.ca.comcast.net)
22:00
22:00:11	TheSeven	I'm wondering if most of the tick tasks should be executed from a "tick thread" that gets woken up once per tick instead of the IRQ handler
22:01:17	bertrik	I think it's OK to run it from interrupt context for button handlers that just read GPIO for buttons
22:03:09		Quit liar (Ping timeout: 252 seconds)
22:04:42		Join liar [0] (~liar@clnet-p09-185.ikbnet.co.at)
22:08:23	wodz	TheSeven: anyway this is plain weird to utilize i2c to just read hold status :-)
22:09:24	TheSeven	bertrik: is there any advantage of running ticks from the IRQ itself except for locking out other IRQs during that time?
22:11:03	TheSeven	wodz: It seems like it's connected to a GPIO of the power manager which we can only access that way
22:11:46	bertrik	TheSeven, not much I guess, it saves a thread
22:13:49		Quit Floweraddict (Remote host closed the connection)
22:14:31		Join Floweraddict [0] (~Flowerpix@c-98-244-27-34.hsd1.ca.comcast.net)
22:15:22		Join prof_wolfff [0] (~d525d9d6@www.haxx.se)
22:16:38		Quit LambdaCalculus37 (Quit: Fwump)
22:18:00	funman	JdGordon: pong
22:19:18	CIA-15	New commit by bertrik (r31646): FS #12511 - Polish translation by Artur Juraszek
22:21:10		Quit TheLemonMan (Quit: WeeChat 0.3.6)
22:21:54	CIA-15	r31646 build result: All green
22:24:01	Zagor	svn now set to read-only
22:24:19	Torne	yay
22:24:25	bertrik	lebellium, want me to make a patched zip now (with the bgr test)?
22:24:38	*	Torne stops his migration loop
22:25:20	Torne	So, do we want to turn pushing to git on now? Or do we want to poke the build system first?
22:25:23	Torne	and see what falls over? :)
22:25:56	funman	Torne: you can push a small patch for a start
22:25:57	Zagor	nah, let's enable pushing. we'll probably prod the build system for a number of days.
22:26:04	Torne	okay. one sec.
22:26:16	Torne	let me take my maagic permissions away first
22:26:26	Torne	i needed to be able to create crazy stuff and forge my identity to run the migration :)
22:28:05	Torne	Done
22:28:27	bertrik	lebellium, http://wikisend.com/download/789758/rockbox.zip
22:28:29	Torne	Anyone in the group can now push to refs/heads/* and can forge the author of patches (to enable you to properly commit stuff written by other people)
22:28:37	Torne	i haven't enabled any of the code review stuff right now
22:28:39		Quit Unhelpful (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.)
22:28:46	Torne	will get to that after we sort out the basics
22:29:30	Topic	"Rockbox is now using git! \| Please read before speaking: http://www.rockbox.org/wiki/IrcGuidelines \| Please direct offtopic/social chat to #rockbox-community \| This channel is logged at http://www.rockbox.org/irc" by ChanServ (ChanServ@services.)
22:29:39	Torne	so who has something to commit? :p
22:30:30	Torne	and who wants to start updating all the docs on the wiki and in docs/*
22:30:36	Torne	:)
22:32:33	lebellium	bertrik: thanks, I try that now. The only issue is that I did not really understand what's the issue. I set the lowest brightness possible and I don't find the display to be reddish...
22:32:34		Join Unhelpful [0] (~quassel@rockbox/developer/Unhelpful)
22:32:53	*	Torne adds the link to gitweb to UsingGit
22:33:19	bertrik	lebellium, I though I saw several people complain about the colour balance of the clip zip
22:34:08	jhMikeS	just like that were moved over?
22:34:34	Torne	jhMikeS: just like that and then updating all the docs/wiki/buildbot/website/CIA/etc, yes :)
22:34:49	Torne	and solving all the exciting problems that people only discover at this late stage.
22:34:52	Torne	and so on :)
22:35:17	jhMikeS	I'm just wondering if I'd better get off my bum and finally update everything here
22:35:34	lebellium	bertrik: I did, I told you several times that the green doesn't look green as it should, particularly the volume bar of my theme. But reddish... Oo
22:35:56	*	Torne adds a big warning at the top of UsingSVN also :)
22:35:56	Zagor	btw, I got a request for the remoclone code the other day. hands up if you remember it :-)
22:36:08	Torne	my dinner will be ready shortly, so i will be idle a while
22:36:18	Torne	anyone who wants to go fix wiki pages, feel free :)
22:36:35	Torne	if you search the wiki for "svn" you find plenty of places
22:37:32	*	jhMikeS wonders if any straggling CVS references still exist
22:37:39	Torne	probably ;)
22:37:51	Torne	Zagor: sometime when we aren't busy talk to me about what you want to do with the dynamic-lcd branch, also
22:37:59	Torne	i have a git-svn mirror of it on my machine
22:38:03	Torne	but i haven't pushed it to gerrit
22:38:11	Torne	you can have it in any form that's handy for you :)
22:38:25	Zagor	thanks :)
22:38:26	Torne	everything else from the svn repo is in there.
22:39:41		Join pamaury [0] (~quassel@valentinblot.org)
22:39:41		Quit pamaury (Changing host)
22:39:41		Join pamaury [0] (~quassel@rockbox/developer/pamaury)
22:42:04		Quit n1s (Quit: Ex-Chat)
22:44:10	bluebrother^	urgh.
22:44:41	bluebrother^	looks like the TTS issue is caused by SAPI sometimes failing to speak a string and Rockbox Utility not knowing about it.
22:45:44	lebellium	bertrik: sorry but I don't really notice any difference :S
22:46:24	bertrik	lebellium, oh weird, no improvement or deterioration of colour balance when you change the OLED brightness?
22:47:47		Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
22:52:05	lebellium	bertrik: well they might be one but then barely visible. To be confirmed by another WiseChip user. All I can tell you is that the green is still far from what you can see here: http://themes.rockbox.org/themes/96x96/lebellium_samsung-like_v1.10/1-WPS_no_album_art2.png
22:52:39		Quit wodz (Quit: Wychodzi)
22:54:39	lebellium	and the yellow still looks orange...
22:55:24		Join fml [0] (~chatzilla@manz-5f74a70e.pool.mediaWays.net)
22:55:46	fml	How is it possible to see what directory is set as the recording directory without looking into the .cfg?
22:56:04		Quit Lonely_Dragon\|2 (Ping timeout: 248 seconds)
22:56:07	jhMikeS	I don't think there is a way
22:59:43	bertrik	Torne, what should I do to convert an existing tree based on rockbox-old into a new and shiny git tree?
23:00
23:00:21		Quit Floweraddict (Remote host closed the connection)
23:00:25		Join Floweraddict [0] (~Flowerpix@c-98-244-27-34.hsd1.ca.comcast.net)
23:01:02	***	Saving seen data "./dancer.seen"
23:02:07	fml	jhMikeS: it's a pity. Shouldn't we add an entry to the recording settings −− just for displaying the dir?
23:04:20	fml	Torne: is it possible to see the annotated code (like blame) via the new git web viewer?
23:07:20		Quit ps-auxw (Ping timeout: 240 seconds)
23:08:24		Join ps-auxw [0] (~arneb@2001:470:c807:0:1532:4e5f:2ad3:4123)
23:11:42	Zagor	anyone want to write a svn2git conversion tool for the build clients?
23:12:45	gevaerts	What should that do?
23:13:37	Zagor	auto convert a build client svn checkout to git
23:13:40		Quit anewuser ()
23:13:50	Zagor	(remove and clone)
23:13:57		Join anewuser [0] (~anewuser@186.93.165.98)
23:13:57		Quit anewuser (Changing host)
23:13:57		Join anewuser [0] (~anewuser@unaffiliated/anewuser)
23:14:16	Zagor	so we can migrate all build clients in one go
23:14:36	gevaerts	Remove is easy. Do a "special" commit to svn and tell them to update ;)
23:15:20	gevaerts	Actually, you could conceivably clone the git reposiroty and commit that to svn, then checkout that on the clients, and you're done
23:16:12	Zagor	you can just say "no" :-)
23:16:56	gevaerts	We can make the clients update rbclient.pl, right?
23:17:02	Zagor	yes
23:17:12		Quit fml (Quit: ChatZilla 0.9.88 [Firefox 9.0.1/20111220165912])
23:17:51		Join Scromple [0] (~Simon@119.225.209.134)
23:18:43	JdGordon	its probably better to manually update th elcients thouh
23:19:29	Zagor	roolku practically begged for a script :-)
23:25:31		Quit prof_wolfff (Quit: CGI:IRC (Ping timeout))
23:26:12	gevaerts	hm
23:26:31	Bagder	http://daniel.haxx.se/rockbox-git/authors.html#commits_per_author
23:27:23	Zagor	nice
23:27:29	Bagder	commits: 30089 (average 9.2 commits per active day, 8.3 per all days)
23:27:38	gevaerts	Zagor: I assume you want something safer than rm -rf android apps backdrops bootloader docs firmware flash fonts gdb icons lib manual rbutil tools uisimulator utils wps?
23:28:16	Zagor	gevaerts: is there anything safer? we have to remove those.
23:28:36	Zagor	perhaps we should just move them to a to-be-removed subdir or something
23:29:04	Zagor	then clone into a subdir and move all items from that to current dir
23:29:30		Quit liar (Ping timeout: 252 seconds)
23:29:40	gevaerts	svn up -r 1 removes nearly everything, but it leaves tools/ unless you do a make veryclean first
23:30:00	Zagor	neat trick
23:30:50		Quit Floweraddict (Remote host closed the connection)
23:30:51	gevaerts	hm, make veryclean isn't good enough
23:30:54		Join Floweraddict [0] (~Flowerpix@c-98-244-27-34.hsd1.ca.comcast.net)
23:31:26	Zagor	I think hard (re)move is just as well. we're not keeping anything.
23:31:32	*	gevaerts nods
23:32:29	Bagder	btw, a nice little git shorthand to remember: [hash]^1 is the hash of the commit immediately before
23:32:44	gevaerts	I'd say we add −−svn2git to rbclient.pl but don't run it automatically
23:32:53	Zagor	Bagder: did amiconn really add 6M lines in 2008?
23:32:54	Bagder	like if you "git blame" code you can check the version just before a change with that
23:33:14	Bagder	Zagor: well, with branches and code imports, who knows!
23:33:27	Zagor	still, 6M is a lot...
23:33:32	Bagder	is indeed
23:33:44	Bagder	I find it interesting that Fridays are so slow
23:34:20	Bagder	and the month distribution looks really wrong
23:37:37		Quit froggyman (Quit: People who think they know everything are really annoying to those of us who actually do.)
23:38:08		Join froggyman [0] (~froggyman@unaffiliated/froggyman)
23:38:47	gevaerts	Zagor: can you make a specific build client run a specific build?
23:39:06	Zagor	with some server hacking, yes
23:39:28	Torne	bertrik: add the new one as a remote, rebase your branches onto it
23:39:31	bluebrother^	nice stats.
23:39:57	*	bluebrother^ wonders if there's a way to make those graphs better readable
23:40:08	gevaerts	So we could (again, thinking aloud here) add the conversion script as a special build and push that to clients
23:41:34	*	Torne returns
23:42:41	Zagor	gevaerts: we can add a special migrate command to clients. no need to fake a build.
23:44:00	*	gevaerts makes a proof of concept migrate shellscript
23:44:37	bluebrother^	so first commit crisis was in 2008, last in 2011? :D
23:46:02		Part domonoky
23:46:23	bluebrother^	urgh. That SAPI TTS multithreaded code looks broken :(
23:46:37		Quit Floweraddict (Quit: AndroIRC - Android IRC Client ( http://www.androirc.com ))
23:46:48	bluebrother^	reusing the encoder engine from multiple threads for sure calls for problems.
23:50:21	bluebrother^	same for the encoder
23:51:10	bluebrother^	how is that supposed to speed up encoding after all?
23:51:14	*	bluebrother^ looks again
23:51:18	*	Torne writes up email to the list about the git migration
23:55:22		Join perrikwp_ [0] (~quassel@cpe-024-163-024-033.triad.res.rr.com)
23:57:48		Quit perrikwp (Ping timeout: 244 seconds)

Previous day | Next day