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 2011-12-21

00:23:21 Join thegeek_ [0] (
00:24:28 Join perrikwp_ [0] (
00:24:38 Join Guinness [0] (
00:24:45 Join tchan1 [0] (
00:25:28 Quit jhMikeS (Ping timeout: 240 seconds)
00:25:40 Join jhMikeS [0] (
00:25:40 Quit jhMikeS (Changing host)
00:25:40 Join jhMikeS [0] (~jethead71@rockbox/developer/jhMikeS)
00:26:28 Quit Guinness` (Ping timeout: 240 seconds)
00:26:28 Quit JdGordon (Ping timeout: 240 seconds)
00:26:28 Quit kkit (Ping timeout: 240 seconds)
00:26:33 Join kkit [0] (
00:26:37 Quit thegeek (Read error: Connection reset by peer)
00:26:48 Quit perrikwp (Ping timeout: 240 seconds)
00:26:48 Quit tchan (Ping timeout: 240 seconds)
00:27:03 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
00:27:42 Join limescout [0] (
00:29:29 Quit lebellium (Quit: ChatZilla 0.9.87 [Firefox 9.0/20111212185108])
00:31:20 Quit z180 (Remote host closed the connection)
00:35:14 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
00:40:05 Join paintchip [0] (
00:40:08 Quit maraz (Ping timeout: 240 seconds)
00:40:14 Join maraz [0] (
00:40:58 Nick perrikwp_ is now known as perrikwp (
00:41:21 Join stripwax [0] (
00:43:09 Join labcoattech [0] (
00:43:11labcoattechhi all
00:45:16labcoattechDoes anyone know why rockbox creates a ##MUSIC# and ##PORT# directory?
00:45:48gevaertsIt doesn't. The original firmware creates those
00:50:18labcoattechthanks, you are right it's to do with MTP (USB mode)
00:51:04 Join Rob2222 [0] (
00:54:08 Quit Rob2223 (Ping timeout: 252 seconds)
00:58:47 Nick tchan1 is now known as tchan (
00:58:58 Quit tchan (Changing host)
00:58:58 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
01:10:35 Quit stripwax (Quit:
01:13:02 Quit chkktri_ (Quit: Leaving)
01:15:44 Join kadoban_ [0] (
01:25:49 Quit zu (Read error: Operation timed out)
01:26:38 Quit Rondom (Ping timeout: 252 seconds)
01:27:45 Join zu [0] (
01:29:48***Saving seen data "./dancer.seen"
01:30:06remlapHey can I put a link to pictureflow on the favourites to make it easier to get to?
01:33:07 Join zu_ [0] (
01:37:01 Quit zu (Ping timeout: 245 seconds)
01:41:07 Quit kadoban_ (Ping timeout: 240 seconds)
02:16:41JdGordonremlap: use a shortcut, type: file, data: /.rockbox/rocks/apps/pictureflow.rock
02:19:19remlapstill learning
02:19:21remlapwill try now
02:24:58 Quit mystica555_ (Remote host closed the connection)
02:47:09 Join anewuser [0] (~anewuser@
02:47:17 Quit anewuser (Changing host)
02:47:17 Join anewuser [0] (~anewuser@unaffiliated/anewuser)
02:49:24 Join mystica555_ [0] (
03:08:58 Part paintchip
03:19:09 Quit labcoattech ()
03:29:50***Saving seen data "./dancer.seen"
03:44:55 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
03:52:08 Quit limescout (Quit: Leaving.)
04:05:38 Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
04:21:01 Quit [7] (Disconnected by services)
04:21:16 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:27:40 Quit amiconn (Disconnected by services)
04:27:41 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:28:03 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:28:49 Quit pixelma (Disconnected by services)
04:28:51 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:28:53 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:39:03 Join dys` [0] (
04:40:37 Quit dys (Ping timeout: 252 seconds)
04:54:14 Join [Saint_] [0] (~Saint]@unaffiliated/saint/x-8516940)
04:57:16funman,27766.0.html : CLIPV2 really seems f*cked up
04:57:28 Quit [Saint] (Ping timeout: 252 seconds)
04:58:00 Nick [Saint_] is now known as [Saint] (~Saint]@unaffiliated/saint/x-8516940)
04:59:53 Quit anewuser ()
05:06:05 Join perrikwp_ [0] (
05:08:15 Quit perrikwp (Read error: Operation timed out)
05:21:16 Join typn_ [0] (
05:26:17 Part typn_ (" - Chat comfortably. Anywhere.")
05:29:52***Saving seen data "./dancer.seen"
05:33:57 Quit Horscht (Quit: Verlassend)
05:46:16 Join Rob2223 [0] (
05:50:16 Quit Rob2222 (Ping timeout: 252 seconds)
05:55:47 Join busybyeski [0] (
05:56:47busybyeskihiiii - i was just reading the manual for nano first generation and it mentions tips and controls for fm radio. does the firmware access a tuner i didn't even know i had?! or is the manual just pretty general for every device?
05:57:22funmanbusybyeski: i think it's a remote tuner
05:59:32busybyeskioh some apple accessory?
06:06:06 Join Rondom [0] (~rondom@2a01:488:66:1000:b24d:4f2f:0:1)
06:09:16[Saint]Apple's tuner/remote accessory.
06:10:04[Saint]I've not had any "no-name" brand ever *quite* work properly.
06:12:12[Saint](Not to be confused with the cheap, $2, FM sender accessories that "just work" with any iPod)
06:39:25 Quit [Saint] (Ping timeout: 252 seconds)
06:50:58 Quit scorche (Ping timeout: 252 seconds)
07:05:21 Join scorche [0] (~scorche@rockbox/administrator/scorche)
07:11:52 Quit Rondom (Disconnected by services)
07:12:15 Join Rondom [0] (~rondom@2a01:488:66:1000:b24d:4f2f:0:1)
07:15:38 Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
07:29:56***Saving seen data "./dancer.seen"
07:39:15 Join typn_ [0] (
07:39:22 Part typn_ (" - Chat comfortably. Anywhere.")
07:44:38 Join kadoban_ [0] (
07:45:32 Join liar [0] (
07:50:50 Join Keripo [0] (
07:53:33*kugel has multiple database scan dirs working
07:54:25kugelI just need a UI for the setting. but I think I go with the poor text entry that autoresume uses too
07:54:28kugel(for now)
07:55:00kugelunless JdGordon finishes his tree selection work? :)
07:58:58 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
08:24:05 Join nick-p [0] (
08:30:04nick-pkugel: As the only other person interested in sleep timer menu options, I was wondering if you thought my second suggestion was acceptable?
08:30:14nick-pIt seemed like pragmatic compromise to me - not crowding out the T&D/ root settings menu, but making the options available for people that want them..
08:30:42kugelI don't like it in system
08:31:17nick-pI was thinking just the options move to system, the actual start stop action stays in T&D
08:31:42kugelsplitting related stuff over many places is even worse :)
08:32:18 Quit BHSPitMonkey (Remote host closed the connection)
08:33:04kugelI thought the t&d menu adapts itself for small screens, why is it so bad for a menu?
08:33:58nick-pIt doesn't though, on a clip there's only two lines of menu options available
08:34:45nick-pRe your suggestion of a permanent Time menu, could the current T&D setting be placed under that?
08:35:55nick-pie, all the RTC dependant stuff in there
08:37:11kugelhow many lines does the clip have again?
08:37:54nick-p7 - 4 for the clock display, 1 for the title, 2 left for options
08:37:56kugelsomething like 8? should be plenty, 1 statusbar, 1 menu title, 2 t&d, that leaves 4 for the menu
08:38:34nick-pif we lost the current clock display there would't be a problem
08:38:58kugelhm, the clock display should remove the 2 empty lines for small screens
08:39:46kugelI guess the numbers need to be tweaked if the clip isn't small enough
08:40:01nick-pI can't help but think that anything not directly related to the current time doesn't need a bloody great clock at the top
08:42:24[Saint]Its a useful visual aid to setting time/date
08:42:34kugelIIRC the current time is spoken in that menu (and in no other place)
08:44:09nick-pOK, how about a new sub-menu: Timers
08:44:39kugelI don't see anything wrong with it, except that it takes space, but 4 lines for the menu seems plenty
08:45:34nick-pI mean at the same level as T&D - somewhere for everthing to go when there isn't a RTC and the T&D menu doesn't exist
08:48:38 Join herefornow [0] (
08:48:47 Quit herefornow (Client Quit)
08:49:06 Join ender` [0] (
08:50:58 Quit Scromple (Read error: Connection reset by peer)
08:51:24nick-pThis is why I brought this up in the first place - putting things in T&D that having nothing to do with the RTC, become homeless when that menu dissappears
08:58:40kugelthat's why I suggested to make it permanent
08:59:57nick-pI'll have a play and see if I can make an acceptable patch based around a permanent Time related menu
09:01:23nick-pMaybe setting the time could be rejigged so it shows a clock above it only when you're doing that
09:01:57nick-pIf people need a handy place to speak the time, maybe shortcuts is a better solution
09:03:03desowingevaerts: what TMS320DM320 your mrobe 500 has? we might be hitting silicon bug
09:04:01desowin(or my change was just plain wrong)
09:11:45 Join LinusN [0] (
09:13:15desowinI just observed similiar thing on Sansa Connect, battery_bench ended at 72%
09:13:45desowinthat's a tricky bug as it worked fine before
09:19:25 Join einhirn [0] (
09:21:16 Join petur [0] (~petur@rockbox/developer/petur)
09:30:00***Saving seen data "./dancer.seen"
09:53:41 Join ptimousse [0] (
09:54:06ptimoussebjr a tous
09:57:26 Join JdGord [0] (
10:03:49 Quit JdGord (Read error: Connection reset by peer)
10:07:58 Join keyb_gr [0] (
10:14:15 Join JdGord [0] (
10:17:26 Quit JdGord (Read error: Connection reset by peer)
10:21:11desowinZagor: can you add TMS320DM320-based to Player Type in flyspray?
10:22:36[Saint]That's a mouthful for a player type :)
10:26:52desowingevaerts: can you try battery bench with FS #12471 ?
10:26:53fs-bluebot DM320 battery bench stops at high voltage levels (patches, assigned)
10:27:11gevaertsdesowin: I will, as soon as I get home :)
10:30:27desowinTMS320DM320 in my all of my Connects is 1.1 silicon version
10:30:50desowinthe note in errata about DSP clocks mentions that it's planned to fix in 1.1
10:31:06desowinno idea if they actually fixed it, or didn't introduce any new bugs
10:31:49desowinanyway, after trying this, I think we might try to significantly lower DSP clock as it doesn't really do any heavy tasks
10:32:31desowinthe errata says that if CPBUS is more than 3x AHB then the interrupt from DSP to ARM might get lost
10:33:00desowinI didn't find what CPBUS really is though...
10:36:14 Join TomColler [0] (
10:58:055EXAAM35LNew commit by desowin (r31388): Sansa Connect: Add udelay() in sdmmc card initialization to prevent dead locks.
10:59:50Zagordesowin: which targets are TMS320DM320 based?
11:00:105EXAAM35Lr31388 build result: All green
11:01:34desowinZagor: Sansa Connect, mrobe 500, creative zvx
11:03:20desowinalso, when filing bug I cannot assing it to myself (my name doesn't appear in list), only after it's posted I can hit the "assign to me
11:04:26Zagorthat's odd
11:15:01Zagordesowin: player type added
11:30:03***Saving seen data "./dancer.seen"
11:34:53 Join rarog [0] (
11:45:085EXAAM35LNew commit by desowin (r31389): Sansa Connect: Don't enable DAC clock as it's not necessary.
11:45:38desowinno idea why bootloader sets it though, maybe they had some other screen before
11:45:56desowin(OF disables it as well)
11:46:445EXAAM35Lr31389 build result: All green
11:55:06 Join anewuser [0] (~anewuser@
11:55:06 Quit anewuser (Changing host)
11:55:06 Join anewuser [0] (~anewuser@unaffiliated/anewuser)
12:09:44 Join stoffel [0] (
12:23:04 Part TomColler
12:23:07nick-pkugel: Is this what you had in mind?
12:24:27kugelit looks like yes
12:25:31nick-pI'm still not convinced they belong there though ;)
12:27:08kugelI'm not convinced they don't belong there :)
12:27:19kugelIMO sleep and idle power timers are time related
12:28:29nick-pYeah, it's pretty tenuous though. Lots of things are kind of related to time - compressor attack?
12:32:41 Join pamaury [0] (
12:32:41 Quit pamaury (Changing host)
12:32:41 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
12:34:35rarogpamaury: I could track down the usb reinserting problem further and am now at the limit of my knowledge, why it fails. FS #12468
12:34:36fs-bluebot Hotswap fails to recognize the card after it was removed in USB mode (bugs, unconfirmed)
12:41:15pamauryr31388 seems suspicious to me, that's not a fix
12:48:23pixelmaurgh, lots of forum spam. The "Recent posts" section is unbearable (which still shows things caught by the akismet plugin otherwise)
12:48:48desowinpamaury: how come?
12:48:52 Join TheLemonMan [0] (
12:49:00*[Saint] was thinking the same thing a few minutes ago.
12:49:10pamaurydesowin: a delay to prevent a deadlock ?
12:49:23pamaurythat's seem like a nasty fix to a real issue
12:49:36desowinwell, locking inside controller
12:49:43desowinnot thread locking
12:49:56pamaurythen the comment should say so !
12:50:49pamauryrarog: thanks, I'll have a look
12:50:51desowinoh, well, unfortunate use of words
12:53:24 Quit bluefoxx (Ping timeout: 252 seconds)
12:53:46 Join bluefoxx [0] (
12:55:145EXAAM35LNew commit by desowin (r31390): Fix comment in sdmmc to reflect what it really means.
12:55:28 Quit nick-p (Quit: Leaving)
12:57:015EXAAM35Lr31390 build result: All green
13:03:43 Join TomColler [0] (
13:07:04 Join lorenzo92 [0] (~chatzilla@
13:10:52 Quit kadoban_ (Ping timeout: 240 seconds)
13:14:20 Quit swilde (Remote host closed the connection)
13:15:23 Join stoffel_ [0] (
13:17:10 Quit stoffel (Ping timeout: 240 seconds)
13:30:05***Saving seen data "./dancer.seen"
13:33:48 Quit nosa-j (Ping timeout: 240 seconds)
13:36:14 Quit Keripo (Quit: Leaving.)
13:50:47 Quit lorenzo92 (Quit: ChatZilla 0.9.87 [Firefox 8.0/20111115183813])
13:52:385EXAAM35LNew commit by desowin (r31391): Sansa Connect: Put DAC (needed for analog video output) in Powerdown mode.
13:53:03desowinI dunno if they have analog output somewhere
13:54:255EXAAM35Lr31391 build result: All green
14:02:14 Join dfkt [0] (dfkt@unaffiliated/dfkt)
14:14:55TheLemonManpamaury: do you have some docs about the gpmi block beside the datasheet ?
14:27:25ukleinekTheLemonMan: there is a nand driver in linux-next for imx23/28's gpmi
14:28:30dfktfunman, just FYI - i updated the SansaAMS page with links to the newest OF versions (since those changes don't show up in the wiki diff)
14:36:51 Part TomColler
14:36:54 Join nick-p [0] (
14:40:48pamauryTheLemonMan: no
14:43:46 Quit ptimousse (Ping timeout: 252 seconds)
14:46:28TheLemonManukleinek: ill check it out, thanks
14:52:59 Quit amiconn (Disconnected by services)
14:53:00 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
14:53:22 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
14:54:01 Quit pixelma (Disconnected by services)
14:54:03 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
14:54:22 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
15:01:24 Join Thra11 [0] (~thrall@
15:02:09 Join jefffp [0] (
15:03:30 Quit stoffel_ (Ping timeout: 240 seconds)
15:08:15jefffpjust wondering, is the fuze v2 port still being worked on?
15:08:43 Quit 5EXAAM35L (Ping timeout: 252 seconds)
15:09:28 Join CIA-114 [0] (
15:10:48[Saint]jefffp: all devices are being actively maintained.
15:12:03[Saint]If its supported, there's regular, active development. Not all Rockbox code is device specific, a lot is shared and applies to most if not all targets.
15:14:40jefffpah yeah, i was asking bc of the lack of usb for fuze v2
15:23:02TheLemonManthe datasheet says gpmi clock is derived from xtal, but unless i got the meaning of 'derived' wrong i shouldnt enable it too
15:27:54desowingevaerts: FYI, I've put new patch in FS #12471
15:27:54fs-bluebot DM320 battery bench stops at high voltage levels (patches, assigned)
15:29:41pamauryTheLemonMan: gpmi is derived from xtal if you set bypass_gpmi in clkseq (if it exists on stmp3770)
15:30:06***Saving seen data "./dancer.seen"
15:36:36mc2739jefffp: a lot of work has been done in recent days on the USB support for amsv2 devices (Fuzev2 and others). You may want to try out the current development build.
15:45:07 Join WalkGood [0] (~4@unaffiliated/walkgood)
15:47:49TheLemonManpamaury: no luck still, the dma request never finishes, any hint do debug that ?
15:50:25rarogpamaury: is it possible just to readout the state of what the card reader sees at the point where "while(!send_cmd(SD_SEND_STATUS, card_info.rca, SSP_SHORT_RESP, &response) && cmd_retry > 0)" fails?
15:54:09 Quit perrikwp_ (Read error: Connection reset by peer)
15:55:20 Join perrikwp [0] (
16:11:19 Join y4n [0] (y4n@unaffiliated/y4ndexx)
16:12:06 Join lorenzo92 [0] (~chatzilla@
16:13:23jefffpmc2739: will do! i just updated to 3.10 this morning anyways.. haha..
16:17:34desowinafter updating codecs for some reason WMA files just skip
16:17:51jefffpmc2739: thanks for letting me know about that!!:D
16:18:23desowinFLAC, mp3 and mpc works fine
16:30:40dfktJdGordon, do you know what issues the %LB scroll bar tag gives? can it be used with a graphic for the bar?
16:30:57 Quit lorenzo92 (Quit: ChatZilla 0.9.87 [Firefox 8.0/20111115183813])
16:42:53pamauryrarog: not really, if the command fails, it just fails. You can just read the result of ssp command but it's probably a time out
16:43:39 Join n1s [0] (~n1s@rockbox/developer/n1s)
16:44:43pamauryTheLemonMan: perhaps you are missing some pin which enable the chip ?
16:46:21desowinWMA files lock up in Debug builds
16:46:28desowinwell, not lock, just skip
16:47:16TheLemonManpamaury: i ripped that part off the linux driver so i think its ok
16:47:23CIA-114New commit by nls (r31393): libtremor: fix memory corruption bug introduced in r30728 that could cause crashes or playback failures for some album art infested files. Closes ...
16:47:29 Quit WalkGood (Ping timeout: 276 seconds)
16:47:39pamauryTheLemonMan: perhaps there some device specific part for the power ?
16:49:10CIA-114r31393 build result: All green
16:51:13 Join WalkGood [0] (
16:51:52 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
16:52:26TheLemonManpamaury: dont think the nand needs something else than gpio/gpmi to be powered on
16:53:29pamauryare you sure the dma channel clock is not gated off ? that's it's not freezed ?
17:03:01 Part Zagor
17:05:26 Quit nick-p (Quit: Leaving)
17:06:37CIA-114New commit by desowin (r31394): TMS320DM320: Check for TIMER1 bit in udelay().
17:07:05desowinthat damned define I used wasn't bit field for IRQ0...
17:08:28CIA-114r31394 build result: All green
17:13:08 Join toffe82 [0] (
17:16:14 Nick [sanyi] is now known as [Sanyi] (~sanyi@unaffiliated/alexanderc)
17:22:10 Quit WalkGood (Ping timeout: 252 seconds)
17:25:41 Quit Thra11 (Quit: kthxbai)
17:30:07***Saving seen data "./dancer.seen"
17:31:48rarogpamaury: hm... but then why isn't setting the count to something higher helping either? perhaps some code blocking this part of code preventing from reading?
17:32:17desowingevaerts: was the battery level actually lower than last value in battery_bench?
17:37:56 Nick [Sanyi] is now known as [sanyi] (~sanyi@unaffiliated/alexanderc)
17:42:10 Quit n1s (Quit: Ex-Chat)
17:56:00 Quit petur (Quit: *plop*)
18:07:46 Join dreamlayers [0] (~bgjenero@rockbox/developer/dreamlayers)
18:09:54 Join ptimousse [0] (
18:18:20 Join kadoban_ [0] (
18:18:55 Quit liar (Remote host closed the connection)
18:21:13 Quit keyb_gr (Ping timeout: 252 seconds)
18:22:32 Join keyb_gr [0] (
18:25:02 Join Misan [0] (
18:27:26 Join nosa-j [0] (
18:35:32 Join Thra11 [0] (~thrall@
18:36:19CIA-114New commit by dreamlayers (r31395): Make more local functions static.
18:38:01CIA-114r31395 build result: All green
18:39:10funmanpamaury: thanks for your work on the driver, now i would like to know exactly which change made it work
18:40:49 Join WalkGood [0] (
18:40:49 Quit WalkGood (Changing host)
18:40:49 Join WalkGood [0] (~4@unaffiliated/walkgood)
18:42:45 Quit ptimousse (Ping timeout: 252 seconds)
18:46:31 Quit keyb_gr (Ping timeout: 268 seconds)
18:47:31 Join keyb_gr [0] (
18:48:10 Part LinusN
19:02:52CIA-114New commit by dreamlayers (r31396): Add missing includes found using -Wmissing-declarations.
19:03:43 Quit Zarggg_ (Ping timeout: 252 seconds)
19:03:57 Join lebellium [0] (
19:04:30CIA-114r31396 build result: 20 errors, 0 warnings (dreamlayers committed)
19:06:13 Quit busybyeski (Ping timeout: 252 seconds)
19:08:27CIA-114New commit by dreamlayers (r31397): Fix red.
19:10:09CIA-114r31397 build result: All green
19:12:11 Join Buschel [0] (
19:14:28 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
19:14:55 Join lorenzo92 [0] (
19:17:51 Join mortalis [0] (~mortalis@
19:18:52 Quit rarog (Quit: ChatZilla 0.9.87 [SeaMonkey 2.4.1/20111011102430])
19:24:30 Join fml [0] (
19:25:11*fml has now completed FS #12470 and its companion FS #12473
19:25:11fs-bluebot Rename "mp3entry.embed_albumart" to "mp3entry.has_embedded_albumart" (patches, new)
19:25:12fs-bluebot Rename "mp3entry.embed_cuesheet" to "mp3entry.embedded_cuesheet" and pull out a field (patches, new)
19:26:46lorenzo92kugel: i never really took care about how firmware is exncrypted...
19:27:08lorenzo92samsung script just works so fine :D
19:27:43kugelis it open source?
19:28:01lorenzo92eh no it isn't, no clue what's the point!
19:28:21fmlIf nobody objects I'll commit them in a couple of days (if nobody does it before)
19:28:40 Join GermanMushroom [0] (
19:28:43kugellorenzo92: where did you get it from?
19:28:51lorenzo92the open source package...
19:29:09lorenzo92in the muon.tar.gz archive...
19:29:16*kugel looks
19:30:09***Saving seen data "./dancer.seen"
19:30:50kugelif it's part of their opensource package we can probably put it into svn
19:31:40kugelalthough it's so small, it could probably be re-implemented
19:32:00 Quit bluebrother (Disconnected by services)
19:32:02 Join bluebrother^ [0] (~dom@rockbox/developer/bluebrother)
19:33:55 Quit fml (Quit: ChatZilla 0.9.87 [Firefox 3.6.24/20111107173218])
19:34:07 Quit fs-bluebot (Ping timeout: 240 seconds)
19:34:19lorenzo92kugel: ah okay if it doesn't need to long...
19:35:09desowingevaerts: I've added CLKOFF functionality to FS #12471 patch, would be nice if that gets tested as well
19:35:59lorenzo92kugel: of course for the new packer, I will test the roms ;)
19:36:27 Quit WalkGood (Ping timeout: 240 seconds)
19:36:31 Join fs-bluebot [0] (
19:36:56kugellorenzo92: it doesn't look as if it does anything other than xoring really
19:37:01dfktlast post of mine disappeared again from the forums... is there a whitelist? if so, could i be put on it? (promise to only spam mailorder brides once a week :p)
19:37:30 Quit bluefoxx (Quit: 5a 65 75 73 73 2d 73 6f 64 64 69 74 2d 66 75 63 6b 2c 20 49 27 6c 6c 20 62 65 20 62 61 63 6b 2e 2e 2e)
19:37:47lorenzo92kugel: don't understand actually how it works, but yes it should be very easy (as well as useless) as encryption
19:38:16gevaertsdfkt: sorry, no whitelist unfortunately. I resurected your post though
19:38:30dfktthanks, gevaerts
19:38:52lorenzo92kugel: is it something like using the ^ operator in C, right?
19:40:49 Join WalkGood [0] (
19:41:58 Join Zarggg [0] (
19:42:17desowinlorenzo92: ^ is literally xor
19:42:28lorenzo92yeah :D
19:42:54lorenzo92it's very common as encryption as I see on google...
19:45:55 Quit WalkGood ()
19:46:52 Join ptimousse [0] (
19:49:06 Quit ptimousse (Client Quit)
19:49:07 Quit y4n (Disconnected by services)
19:49:10desowinencryption? maybe some kind of obfuscation, but definitely not encryption
19:49:14 Join y4n [0] (y4n@unaffiliated/y4ndexx)
19:55:30desowingrrr, I forgot to turn repeat on with battery_bench
19:55:33 Join Stummi [0] (~Stummi@rockbox/developer/Stummi)
19:57:21[Saint]I've done that several times...
19:57:30funmancould battery_bench turn on that setting ?
19:57:47 Join bertrik [0] (
19:57:47 Quit bertrik (Changing host)
19:57:47 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
19:58:17 Join bluefoxx [0] (
19:58:22dreamlayersfunman: I can't find "repeat" in the plugin api.
19:58:23[Saint]Sounds plausible, but only if it can restore the previous setting on exit.
19:59:48 Join Horscht [0] (
19:59:48 Quit Horscht (Changing host)
19:59:48 Join Horscht [0] (~Horscht@xbmc/user/horscht)
20:02:02funman[Saint]: there's no exit, the player powers off!
20:02:24[Saint]It'd be really cool if files in, say, /rockbox/test_files automagically launched battery_bench or test_codec, presenting a menu to choose either and applying repeat for battery bench.
20:02:38dreamlayersfunman: i think you exit it if you load another plugin.
20:04:20[Saint]funman: right, bit it does a "safe" shutdown no? It could still write to disk and save a setting if it can write the bench file out.
20:04:32[Saint]*but it
20:07:09 Join perrikwp_ [0] (
20:08:32 Quit GermanMushroom (Read error: Connection reset by peer)
20:08:36bluebrother^scorche|sh: we don't have a "Samsung - Installation/Removal" in the manual installatin section of the forums.
20:08:49 Join GermanMushroom [0] (
20:09:11bluebrother^oh, and Philips is also missing
20:09:14funmandreamlayers: [Saint]: true but since battery_bench brings the player near battery exhaustion we should be ready for a (unwanted) brutal power off, and we'd need to enable the setting in a non persistent way
20:09:20funmannot sure if it's possible at all curently
20:09:29bluebrother^since we have supported players having a section would be nice.
20:09:39bluebrother^or maybe add some "others" section?
20:10:06 Quit perrikwp (Ping timeout: 252 seconds)
20:11:31dreamlayersfunman: Yes, good point. Even a normal shutdown might not be able to write because battery level is dangerous for that.
20:12:52[Saint]Then, how does battery bench even *work* lol?
20:13:09[Saint]It's never guaranteed to save a result?
20:13:54kugeldreamlayers: global_settings is in the plugin api
20:14:41[Saint]If there's not enough juice left to write out current user settings on a safe shutdown, somethings broken IMO.
20:15:14funman[Saint]: it's saving lines one by one
20:15:36kugel[Saint]: a HDD can put significat pressure on the battery
20:16:48[Saint]Right, but if it can't save an entire config.cfg, then shouldn't that be an indication that the min % for shutdown is too low?
20:17:18funmansomething else
20:17:24funmanrockbox can crash while running battery bench
20:19:41[Saint]Well...yeah, there is that. But I can't really think of a sane way of doing a non-persistent setting.
20:20:04[Saint]Not without doing something obvious like a magic file.
20:23:42dreamlayers[Saint]: there are battery_level_dangerous values and battery_level_safe() which tells if the battery level is safe for disk writes. The dangerous level can be higher than the shutoff level.
20:25:09dreamlayersThere's a possible bug here though. The dangerous level is ignored if NO_LOW_BATTERY_SHUTDOWN is defined, and several Archos targets define that.
20:25:23 Join anewuser_ [0] (~anewuser@
20:25:55funman[Saint]: if we can read the config we could print a splash warning "you did not enable repeat"
20:28:07 Quit anewuser (Ping timeout: 240 seconds)
20:29:01dreamlayersWhy did I just get a forum notification when there's no new post in the thread?
20:29:35funmandreamlayers: message hold down by the spam filter?
20:29:39[Saint]Yeah...its a lot simpler than the idea I have in my head for a non-persistent setting.
20:30:07funmanthe spam filter is annoying but i guess it's still better than the spam :/
20:30:11gevaertsdesowin: let me check. I wouldn't be surprised if my benchmarks failed for the same reason...
20:30:28dreamlayersfunman: where can I find those? I don't see it in removed posts.
20:30:36funmanlook in recent posts
20:31:41[Saint]Its somewhat backwards that they still show up in recent posts after being removed. That never used to happen.
20:32:15funmanbut the posts are *not removed
20:32:21gevaerts[Saint]: as far as I know akismet has *always* done that
20:32:46dfkton the other hand, the restored-from-spam posts don't seem to show up
20:32:55dfktin the list i mean
20:33:18gevaertsIn which list?
20:33:45[Saint]gevaerts: yeah, bad wording on my part. "it" == "The Forum", I was rather vague...
20:33:47dfktforum front page.. eg. my post you restored in new ports, and it still shows gurglin's post as last
20:33:57gevaertsah, right
20:34:06*gevaerts knows that there are days with more than 20 *actual* spam messsage interceptions
20:34:45gevaertsYes, akismet intergration is a bit hacky it seems
20:35:13gevaertsdesowin: I'll have another go at the original battery benches, and then do further testing later I think
20:39:07*dreamlayers wonders what to do with FS #12403 now that there could be a hwcodec fork.
20:39:08fs-bluebot INIT_ATTR for SH Archos HWCODEC targets (patches, new)
20:40:08*funman checks last git switch status report
20:41:19gevaertsdreamlayers: that looks like it doesn't touch much outside of sh bits
20:42:18dreamlayersgevaerts: it makes changes only needed for hwcodec in buflib, core_alloc and main.
20:42:23gevaertsTo be honest, I don't think the possible forking of hwcodec is a good reason for not doing any hwcodec-related changes
20:42:37funmanTorne: is there still documentation needed for git?
20:42:57 Quit Xerion (Quit: )
20:43:34dreamlayersafter a fork, parts of those changes will have to be reverted
20:43:50gevaertswill they?
20:46:06gevaertsAnyway, my basic standpoint here is that talk of possible future changes should not influence things now. That would only lead to doing nothing at all
20:46:36dreamlayersThey won't cause problems, but they'll be pointless. For example, CONFIG_CODEC == SWCODEC is pointless, and if there's != instead, you have unused code.
20:46:55TheLemonManpamaury: what interrupt is raised on apbh channel completion ?
20:46:55dreamlayersgevaerts: sounds reasonable
20:47:20gevaertsTrue, but there are *lots* of places with CONFIG_CODEC tests. One more won't make that much difference
20:48:39 Quit GermanMushroom (Read error: Connection reset by peer)
20:51:43 Join saratoga [0] (47e22765@gateway/web/freenode/ip.
20:52:07Tornefunman: you tell me :)
20:52:08saratogaweren't we going to leave SWCODEC in the branch?
20:52:28funmanTorne: then the question is when do we switch?
20:52:30pamauryTheLemonMan: the dma interrupt (or dma error interrupt) I guess
20:52:33gevaertssaratoga: I think dreamlayers meant removing SWCODEC (and HWCODEC) ifdefs from the main branch
20:52:42dreamlayersAny objections to making buflib_buffer_in() create an empty block at the start instead of moving data. Only pictureflow uses it, and it doesn't break pictureflow.
20:53:06Tornefunman: er, let me get back to you on that
20:53:10Tornei've been deathly ill
20:53:13Tornei'm mostly better now
20:53:21Tornebut i have other stuff i need to catch up on first
20:53:29Tornei still need to work out something for
20:53:42Torneas for docs, read what i've written and comment on what is lacking
20:53:57dreamlayersyes, i meant ifdefs in the trunk, in the future.
20:54:05funmanTorne: ok, get well. can i help with ?
20:54:30Torneer, i dunno. if you can think of a sensible scheme for producing human readable version numbers, feel free to propose it
20:54:33lorenzo92kugel: in 2 words, what do we have to do to put our port to svn?
20:54:43Tornei've been experimenting with sticking special tags on a bunch of stuff and using them as input to git describe
20:54:56Torneit's possibly horrible. not convinced.
20:55:02funmanTorne: git-describe ? perhaps we can extract only number of commits since last tag
20:55:26Torneyou need quite a yucky pattern of tags to make release branches sanely versionable
20:55:32Tornee.g. you need to tag both sides after the branch point
20:55:34ukleinekfunman: that's a subset of what git-describe does
20:55:50Torneukleinek: yes, but describe is basically useless
20:55:51funmanukleinek: yes i meant to remove the sha1 from git-describe
20:56:04 Quit y4n (Quit: Today is the perfect day for a perfect day.)
20:56:07funmanhmm why both sides?
20:56:24Tornefunman: because otherwise how do you tell the difference? :)
20:56:43ukleinekTorne: yeah, if a release isn't done on master, there is no tag reachable from it.
20:56:44Torneit would suck if the third commit after the branch point, on trunk and on the release branch, had the same version
20:56:46funmannumber of commits since the tag?
20:57:06Tornealso, the commits on the release branch will mostly come *before* the "real" tags
20:57:17funmantag 3.11-pre, branch
20:57:26funmantag 3.11 when released
20:57:27ukleinekdon't use a release branch then :-)
20:57:28TorneBut then you also need to tag the next commit on trunk
20:57:30funmanbut for trunk/HEAD hmm
20:57:32Torneotherwise they are also different
20:57:32kugellorenzo92: in 2 words? :)
20:57:41TorneSo yeah, i've been playing with doing -pre and -dev
20:57:48Torneso the trunk would be 3.11-dev currently
20:57:52lorenzo92kugel: no I mean, is there something particular?
20:57:58gevaertsDo we care about version strings anywhere apart from the official builds?
20:58:07Tornegevaerts: i think we *should*
20:58:11desowingevaerts: define "original battery bench"
20:58:24gevaertsdesowin: the set I tried a few days ago
20:58:30Tornegevaerts: otherwise our merely human users are going to find the process of using a current build, or upgrading, or reporting bugs, very confusing
20:58:33gevaertsI still don't have a baseline
20:58:46gevaertshm, true
20:59:03gevaertscurrent builds can be official
20:59:13TorneOh, sorry, i see what you mean
20:59:18Torneyes, we could just number builds
20:59:24funmanTorne: or we do something "special" for master branch in ?
20:59:34Tornefunman: Incredibly difficult to do sanely/reliably
20:59:43Tornegit gives you no remotely sensibel way to work out whether you are the "main" branch or not
20:59:46kugellorenzo92: no, but I want to understand the installation procedure before
20:59:47Tornebecause that's just not a thing
20:59:57funman'git branch' ?
21:00:02lorenzo92kugel: okay indeed ;)
21:00:02kugeleventually we want rbutil to support the ypr0
21:00:03gevaertsI'm mainly thinking of a system where the build master tells build clients if they're doing a trunk build, a release branch build, or an experimental patch build, and that would be included in the version string
21:00:05Torne..but everyone's branch is called "master"
21:00:16gevaertsThis doesn't help for local builds though
21:00:21Tornegevaerts: sure, but i think it'd be nice if we can make local builds agree
21:00:53Tornefunman: you can't look at master, you would need to look at, say, origin/master and then check if origin is infact the right place :)
21:00:57desowingevaerts: so if it was indeed lack of repeat option, then the FS #12471 could get closed as invalid (and just the CLOCKOUT taken out of it, if it works on mrobe)
21:00:59Torneand then what if the remote is not called origin
21:01:01Torneetc etc
21:01:10funmanTorne: but it's less a problem for builds of other people i would say
21:01:16gevaertsdesowin: I'll report back as soon as I know more :)
21:01:26funmanpeople can change their remote but they can also change to cause trouble
21:01:38Tornenot by accident
21:01:44Tornei'm not talking about it being intentional, here
21:01:56Tornepeople clone things that are not the "original" repo for a project quite often
21:01:59Tornesee: all the people that use github
21:03:01Torneaanyway. i am gonna fiddle with some options maybe later, or tomorrow
21:03:08Tornebut if you want to try something feel free
21:03:14Torneright now i need to make food :)
21:05:46 Join bug2000 [0] (~bug@unaffiliated/bug2000)
21:05:47desowingevaerts: even it that was the problem, the udelay() got aware of problem with disabled interrupts... but still, this implementation although shouldn't lock anymore, but could exit before delay passes
21:06:03bug2000Just wondering. Is it me or 3.10 didn't get enough testing?
21:06:35Tornedid you test it?
21:06:50gevaertsThat's a philosophical question
21:06:55kugelcan't we just include the branch name, and count commits since branch-off?
21:07:17bug2000Torne, Only now. I haven't used nightly's lately.
21:07:24TorneThen don't complain :)
21:07:35Tornekugel: no; the build clients will not really get branches
21:07:42Tornethey will be being asked to build a SHA
21:08:02kugelwhat do the build clients have to do with ti?
21:08:06desowinif: a) interrupts were disabled, there was unserviced TIMER1 interrupt, and then called the udelay() in some cases it can return immadietely and b) similiar problem if the interrupt status changes during "probe" part of udelay()
21:08:25desowinboth of these problems apply only to disabled interrupts thing
21:08:31Tornekugel: ...the build clients are the primary consumer of teh output of
21:08:31kugelaren't we talking about rXXXX replacement anymore?
21:08:50Tornethe most important use of which is to compile into builds produced by the build system
21:09:24Torne(isn't it?)
21:09:49desowinI can think of rather crude way of clearing that prnding interrupt, in worst case that would make the udelay() 10ms (one tick) longer, but it won't be able to exit too dast
21:10:03kugelwe'll just be telling the clients to build a specific commit, i.e. check that out?
21:10:15Tornethis is the only remotely sensible way to do it
21:10:17gevaertsWe do that now
21:10:18kugelthen they're detached
21:10:20Tornebranches move
21:10:26kugelgevaerts: now?
21:10:36Torneyes, currently we tell clients the svn revno, and htey svn up -rX
21:10:41Tornegit co <sha1>
21:10:45Torneis the equivalent
21:11:08Torneif they pulled the branch head, they may not all compile the same thing.
21:11:11kugelso why didn't the build farm did release builds so far? :)
21:11:18gevaertsThe only differences are that (a) svn revision identifiers are well-ordered, and (b) it's not easy to do this cross-branch with svn
21:11:34Tornekugel: because you can't svn up to a different root directory
21:11:45Torneand thus you can't trivially pull down a branch
21:11:48kugelah yes, you need an extra svn switch
21:11:59Torneso yeah, this is a *huge* feature :)
21:12:12gevaertsIn my opinion this alone is a good reason to switch :)
21:12:15Tornesince doing it this way means we will be able to do release builds, test builds, whatever.
21:12:49funmangevaerts: x264 use git but still uses svn-like rev numbers
21:13:12funmani.e. ordered, incremental (but they don't use git merges though)
21:13:20Tornegood to know; now go find out how and tell us
21:13:23kugelwe could provide the branch/tag name instead of sha1 and make the farm build the HEAD of that
21:13:34Tornethat is a terrible idea
21:13:39Tornethe machines may not even build the same code then
21:13:40funmanTorne: git describe since first commit?
21:13:45desowinhmm, arm should be fast enough to compare if count value is lower than one from function entrance and not relay on the ticks at all
21:14:06kugel(after updating of course)
21:14:20gevaertskugel: racy
21:14:34desowinthat would solve the problem, but introduce another one when "entry count" is low
21:14:37Tornekugel: commit, trigger build, someone else commits immediately afterward, half the build clients update to a different version.
21:14:44kugelhm, right
21:15:17Tornefunman: ok, so they are just counting
21:15:24TorneThere's one big "problem" with us doing that
21:15:38Tornenamely that would mean our build numbers go down a lot when we switch :)
21:15:43bertrikdesowin, can't you use a free-running timer that counts all the way up to the 32-bit range?
21:16:19Tornefunman: i quite like the notion of having all versions include an actual version number, also, to make it clearer to users that rxxxxx is newer than 3.10
21:16:44kugelI think that's not a _big_ problem
21:17:02Torneno, but i think we can do better.
21:17:04bug2000Torne, OK, so how do I help with it? 3.10 is far from being stable on Sansa Clip - Can crash after 5 minutes of listening to podcast.
21:17:05kugelperhaps confusing for a few weeks
21:17:49gevaertsbug2000: report bugs, and next time test release candidate builds
21:18:03gevaertsUnreported bugs do not exist
21:18:07Tornealso, make sure you are not running with a corrupt filesystem, or similar
21:18:09 Quit Buschel (Quit: ChatZilla 0.9.87 [Firefox 8.0/20111104165243])
21:18:41Tornei.e. eliminate possible local causes first
21:18:41bug2000gevaerts, :) I just don't know how to fill crash reports.
21:19:05bug2000Torne, Well it worked fine just before upgrading. Think the filesystem died suddenly?
21:19:29 Join n1s [0] (
21:19:29 Quit n1s (Changing host)
21:19:29 Join n1s [0] (~n1s@rockbox/developer/n1s)
21:19:30Torneno idea, but since checking the filesystem takes all of five minutes it's a sensibel step to start with
21:19:55bug2000Torne, fsck / ckdsk or format?
21:20:46gevaertsSeriously, if you consider "format" to be a way to check a filesystem, you shouldn't be using a computer
21:20:59bug2000Data abort at 3001913E FSR 0x8 (domain 0, fault 8) address 0x7261702D −− Sample crash message,
21:21:11bug2000gevaerts, Well it fixes a broken file system :P
21:21:17gevaertsIt doesn't
21:21:23bertrikthat address looks like a part of an ascii string
21:21:24gevaertsIt makes a new filesystem
21:21:38bug2000gevaerts, That is a minor detail :P
21:21:59 Quit dreamlayers (Quit: back later)
21:22:36*gevaerts waits for bug2000 to clarify which player *exactly* this is
21:22:45saratogaclipv1 i think
21:22:58saratogaoh could be v2
21:22:58funmanTorne: why would our rnumber change after the switch?
21:23:02bug2000Oh I forgot to mention v1. You are correct.
21:23:13saratogabut probably v1 since the clip+ is working great
21:23:30saratogai'll update my clipv1 to current svn once i'm back from the holidays and see how its working
21:23:31Tornefunman: beause there are many less commits in master than there are in svn
21:23:46bug2000I am sorry for forgetting such an importent detail. I kind of managed to forget the entire v1 v2 [though I did download the correct firmware]
21:23:47funmanwe can fixup the count then
21:24:11saratogado we have anyone with a c200v2?
21:24:21saratogai'm wondering how that build is working these days
21:24:22Tornemy baisc point here is i think we can *better* than a meaningless monotonically increasing number
21:24:22bug2000I only have Clipv1 and Fuzev2.
21:24:33gevaertsThat seems to be in evaluate_conditional()
21:24:36bertrikI think ranma or funman has a c200v2
21:24:39Tornesomething more meaningful to humans
21:24:40funmansaratoga: ranma
21:24:52funmanbertrik: you're right i have a c200v2 (in france) but iirc it is broken
21:24:56saratogahaven't heard from him in ages
21:25:03saratogai wonder if the c200v2 could be stable now
21:25:03funmanat least some buttons are, perhaps the storage is too
21:25:12kugelTorne: hopefully crafting that out doesnt delay the git switch too much
21:25:14saratogawe've fixed a lot of the buffering stuff since ranma was around
21:25:35Tornekugel: i am intending to give up if i can't find a decent solution easily ;)
21:25:41Tornei just haven't actually looked at it much yet
21:25:49funmansaratoga: and −−Wl,−−gc-sections could help a bit
21:26:26bertrikyet another idea: could the build master create a version number?
21:26:38Tornebertrik: yes, we covered that :)
21:26:44bertrikoh, ok, sorry :)
21:26:49Tornebertrik: that would be easy, but somewhat disappointing that other people can't reproduce it
21:27:01Tornewell, easy for master
21:27:08kugel[Saint]: might be interested in FS #12474
21:27:08Tornestill somewhat debatable how to handle release branch builds
21:27:09fs-bluebot [PATCH] Database: Support for multiple search roots. (patches, new)
21:28:02gevaertsI think having build-system builds tagged in some way is useful whatever the rest of the version number looks like
21:29:25funmanjust make sure it fits on 96 pixels wide display :)
21:29:40 Join Xerion [0] (
21:29:46gevaertsbug2000: please file a bug report. Since this seems to be in the skin engine, it's probably useful to mention which theme you use
21:30:09funmanbug2000: do you know how to use ?
21:30:12***Saving seen data "./dancer.seen"
21:30:23bug2000gevaerts, I'm still trying to figure what is first FAT / second FAT.
21:30:29bug2000funman, Nope :\
21:30:42gevaertsfunman: you can do that as well :)
21:30:54gevaertself and map files for the release are available
21:31:03funmangevaerts: yeah but i'm lazy
21:32:11gevaertsHm, we could keep a database with symbol tables around, and have a bot scan flyspray for pairs of panic messages and revision strings
21:33:07bug2000funman, fsck says the filesystem is fine.
21:37:36 Quit n1s (Read error: Connection timed out)
21:41:32desowincan compiler be trusted to init count before initializing prev_tick? - function udelay
21:44:00 Join toffe82_ [0] (
21:44:25kugeldesowin: not sure, but perhaps it helps: as3525/v2 also has implemented udelay via the tick counter
21:45:39kugelwe also faced some problems initially, so svn log might be of help also
21:47:15 Quit toffe82 (Ping timeout: 252 seconds)
21:47:27 Nick toffe82_ is now known as toffe82 (
21:49:04bug2000funman / gevaerts; So how to file the bug report?
21:49:06kugeland it also implements the "NEVER less" principle. it's inaccurate for some values of usec
21:50:45saratogabug2000: bug link on the front page, describe the problem and how to reproduce it
21:50:51saratogaif possible try and figure out what release created it
21:51:15bug2000saratoga, Crash errors sometimes end up changing for me.
21:51:41*bug2000 goes to fill a crash report.
21:51:47saratogakeep in mind that if other people can't reproduce it, you'll pretty much have to fix it yourself
21:52:06bug2000saratoga, Do you have clip v1?
21:52:15saratogayes but not with me today
21:52:25 Quit mortalis (Quit: KVIrc 4.1.1 Equilibrium
21:54:42bug2000saratoga, Bugtracker doesn't let me choose 3.10?
21:55:06saratogayou should probably test the current build
21:55:35gevaertsbug2000: then just clarify in the text
21:56:27bug2000gevaerts, Mark it as 3.9 or nightly?
21:56:39gevaertsDoesn't matter much
22:00:47bug2000gevaerts, I find it funny that on the bus I want the crashes to be as far as possible from each other and now I want 'em quickly.
22:01:14bug2000gevaerts, Checking to make sure track lentgh doesn't play a role in this issue.
22:03:10 Quit Stummi (Quit: Bye!)
22:03:12bug2000gevaerts, Does it seem ok?
22:04:22dfktuser name doesn't seem ok ;)
22:04:32gevaertsI think so. If you find more information later, don't forget to add it
22:05:18bug2000dfkt, Just because it lacks real name?
22:05:31bug2000dfkt, Awesome theme for the clip. Thanks. Though I use it with a different font.
22:05:51gevaertsreal names aren't important for bug reports
22:06:31bug2000dfkt, I'd like to ask you a personal question. Which theme do _you_ use on your sansa clip?
22:06:52dfkti only use the themes i make myself on any rockboxed player
22:07:41bug2000dfkt, What are the others? For which players are they?
22:09:01bug2000dfkt, Can you check if your clip crash as well with 3.10?
22:09:09kugellorenzo92: ping
22:10:24dfktbug2000, my clip doesn't crash with any revision, it's perfectly stable. i made some themes for the cowon x5, clip/clip+, and raaa. they're all on the themes site.
22:11:40CIA-114New commit by desowin (r31398): TMS320DM320: Ensure udelay() will not end before specified amount of time.
22:12:19desowinif I find some time tomorrow I'll do some probing with Agilent scope for any faults
22:13:06bertrikdesowin, can't you use a free-running counter to simplify things?
22:13:25CIA-114r31398 build result: All green
22:13:56desowinbertrik: it does freerun, just it generates ticks as well
22:14:11desowinbertrik: and the problem is there aren't more timers...
22:15:01desowinthat is, I don't want to block the mrobe 500 port from being able to support camera
22:15:25 Join Strife89 [0] (~Strife89@
22:17:28 Quit perrikwp_ (Read error: Connection reset by peer)
22:17:55lorenzo92kugel: here I am
22:18:39 Join perrikwp [0] (
22:20:02kugellorenzo92: I'm wondering what you changed in your rom
22:20:04desowinanyway, I hope those comments in this function will be useful to anyone facing same problem with different hardware sometime later
22:20:23kugelI found the mods folder, but you probably also changed some other scripts?
22:20:39lorenzo92profile script
22:20:44lorenzo92that launches
22:20:50lorenzo92and so forth :)
22:21:45kugellorenzo92: another question: what is NEO-GHOST?
22:22:30kugelis cputricks needed with ondemand governor?
22:22:56lorenzo92kugel: hehe don't ask me why this name, but it's a small portion of nand (/dev/stl? atm I don't remember) mounted as /media1 that stores OF settings (about 60 mb)
22:23:02lorenzo92cputricks isn't needed
22:23:10lorenzo92it's just another mod to be used with OF
22:25:14lorenzo92you can see all the features in the readme of the modded firmw, wait
22:27:49kugellorenzo92: okay, so do I understand correctly? profile->>>>rockbox?
22:29:14lorenzo92yes :)
22:29:36lorenzo92of course I kept the loader out of the rom to ease the initial debug...
22:33:52kugellorenzo92: what is rescue_kit/DRK?
22:35:17kugelthe default rom cat start r0 from the sdcard if present?
22:35:51kugelor from NEO-GHOST?
22:40:30 Quit TheLemonMan (Quit: WeeChat 0.3.6)
22:44:20 Quit keyb_gr (Ping timeout: 248 seconds)
22:44:35 Join fml [0] (
22:46:10fmlHello. I've seen a forum thread about an ancient gigabeat theme. I also like and use a theme on my H120 which is not on the theme site. IT#s this one: It is marked as "license by cc-sa". Can it brought to the theme site?
22:47:54fmlIt's also available for sansa c200
22:49:09lorenzo92kugel: sorry here I DRK is the rescue kit useful for people to unbrick easily the device. Yes you can start R0 OF throug sd card just type
22:50:48 Join tchan1 [0] (
22:51:25 Quit tchan (Ping timeout: 252 seconds)
22:51:46 Nick tchan1 is now known as tchan (
22:51:50 Quit tchan (Changing host)
22:51:50 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
22:52:23kugellorenzo92: what do you mean by type?
22:52:37kugelI saw some references to /mnt/media1/r0
22:52:41 Quit fml (Quit: ChatZilla 0.9.87 [Firefox 8.0/20111104165243])
22:53:03lorenzo92kugel: yeah sorry, I mean, you have to run a script containing usr/local/bin/ro to start OF application...
22:53:19lorenzo92yes these references where used during developement I guess
22:54:31remlapis there an option for RDS to scroll?
22:54:36kugelso it isnt actually possible in the release firmware?
22:55:14lorenzo92kugel:'s impossible to start something with it
22:56:20lorenzo92kugel: moreover, the OF has a BAD bug during library update: if it occurs, without safe mode etc, the player is bricked
23:02:29lorenzo92kugel: if you see, in every forum about R0 there is someone who had problems with that lol
23:03:26dfktremlap, doesn't seem to be implemented yet
23:03:48remlapthank dfkt
23:04:04remlaplove your work especially your minimal theme
23:04:09dfktthanks :)
23:04:37remlapcurrent build crashed on me though
23:04:40remlapso tried your patch
23:04:50dfktrds and non-rds info is in the same scrolling lines, so rds should start scrolling automagically with my theme, once it works
23:05:02remlapwouldn't boot, lucky easy access to stock firmware
23:06:30remlapRadio 2 sounds fantastic on the zip though best I've ever heard off a portable radio
23:07:07remlapbetter than any of the phones I've had or my other mp3 players
23:13:29 Join keyb_gr [0] (
23:15:00remlapwill rockbox charge on the zip with my phones charger since its only micro usb
23:15:10remlapits also micro*
23:15:45funmanyes it should work
23:16:10 Quit parafin (Quit: So long and thanks for all the fish)
23:17:16kugellorenzo92: what bug?
23:18:59lorenzo92kugel: oh well. It's the most annoying thing for R0 2 words: you copy lots of song in the memory, then the library update is run. It happens, quite randomly, that it crashes. If that crashes, you're lost :)
23:19:18lorenzo92bricked device -> samsung center. But not with the modded firmware -> safe mode
23:19:25 Part Strife89 ("Leaving")
23:23:22kugellorenzo92: Am I correct that we essentially only need to change the rootfs to boot rockbox?
23:23:51 Join parafin [0] (parafin@
23:24:02lorenzo92just a small launcher, but I don't understand why not including the safe mode too
23:24:28 Quit parafin (Client Quit)
23:24:35 Join parafin [0] (
23:26:57kugellorenzo92: I think we want the safe mode, as it prevents the database upgrade and "checking filesystem" thing if you force poweroff
23:28:43lorenzo92indeed ;)
23:30:02kugelautomatic reboot is a nice bonus as well
23:30:13***Saving seen data "./dancer.seen"
23:30:15 Quit dfkt (Quit: -= SysReset 2.55=- Sic gorgiamus allos subjectatos nunc.)
23:30:36kugelideally we implement the usb mode in rockbox in someway, but rebooting into save mode would be nice already
23:32:19lorenzo92exactly ;)
23:33:48lorenzo92it should not be too difficult to implement it in rb...since cable detection will be useful to read the charger state too
23:36:00kugellorenzo92: can we implement reboot into safe mode?
23:36:16lorenzo92uhm what do you mean?
23:36:43lorenzo92from RB to safe mode directly?
23:37:33kugelissue a reboot and have the safe mode start automatically (assuming that's easier than doing usb mode within rockbox)
23:38:12lorenzo92eh the problem is that you cannot reboot the r0 in the real meaning of invoking a shutdown + auto startup again...
23:40:07kugelperhaps suspend the rockbox process and bringing the usb screen one into foreground (like pressing ctrl+z in the terminal, then fg/bg)
23:40:19lorenzo92uhm interesting as idea :)
23:40:20kugelaaaanyway, that can be implemented later
23:40:34lorenzo92yep indeed, before the battery indicator :p
23:42:32kugelnot sure what the expected runtime is or how many battery is left, but my r0 just played for 12h
23:42:56lorenzo92uhm interesting result. Same hrs as the OF having activated the DNSe
23:43:07lorenzo92the OF uses HW decoding...
23:43:15lorenzo92DNSe is CPU thing
23:43:20lorenzo92maybe that's why
23:43:43kugelgah, no battery bench on that thing
23:43:59kugellorenzo92: no, it's still playing :)
23:44:41lorenzo92uh quite interesting :)
23:44:52lebelliumkugel: new or second-hand player? What format and bitrate?
23:45:07kugelit'd probably make anotgher few hours. can't tell without battery indicator or battery bench
23:45:14lorenzo92go to
23:45:16kugellebellium: second hand, ~1 year old
23:45:39lorenzo92sys/devices/platform/afe.0/adc, open it with the text viewer
23:45:49lorenzo92look at the first value
23:46:01lorenzo92if it's near 3.450 it's time to shut down the player...
23:46:25kugelso not much left I guess
23:47:28lorenzo92yeah I think it's better to charge it, OF shuts down more or less at 3.450
23:47:38lorenzo92there should be HW protection of course
23:48:14lorenzo92maybe activating the ARM optimizations wecould achieve better results ;)
23:48:53kugelthe 12h was just since I turned it on
23:49:00kugelI used it for a few hours before too
23:49:41lorenzo92maybe I could do a bench while sleeping tonight :D
23:51:35lorenzo92kugel: now you tested the player with a single song in repeat mode or what?
23:51:50kugelforgot to turn it off, actually
23:54:04kugelso yea, we should make a rom which can load rockbox and includes the safe mode, but the required tools into svn, and either a ready-made script for the actual generation or at least exact instructions on the wikik
23:54:13kugelthat should be fine enough to get this into svn
23:54:47kugelwhat's stock rom is your rom based on?
23:54:57kugelnot the one in the open source image is it?
23:54:58lorenzo92latest of course...1.25
23:55:08lorenzo92nono that was very old ;)
23:55:52lorenzo92I unpacked the 1.25 rom and worked on it...
23:57:41kugellorenzo92: btw, do you know about "sourcing" shell scripts? you seem to execute scripts all over
23:57:55kugelthat starts a new process, while sourcing executes the script in the same process
23:58:14lorenzo92kugel: uhm not really :D I began to write script for the R0 hihi
23:58:20lorenzo92uhm interesting
23:58:30lorenzo92ah ok maybe I understood
23:58:36lorenzo92the command: source
23:59:01 Join Topy [0] (

Previous day | Next day