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 2011-10-15

00:01:50 Quit AlexP (Remote host closed the connection)
00:13:26 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
00:16:46saratogaits weird how PP in spite of being extremely widely used never leaked any of their datasheets over the years
00:18:29jhMikeSmaybe none existed?
00:18:40Torneworryingly plausible :)
00:21:30saratogayeah actually, i suppose the code might have been the documentation
00:22:39jhMikeSthey killed the die designers
00:29:15 Quit zchs (Ping timeout: 265 seconds)
00:29:39 Join zchs [0] (
00:32:55 Join robin0800 [0] (~robin0800@
00:33:14 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
00:34:55 Quit ender` (Quit: Where python would say you can skin a cat with a knife, so adding other methods would just make the language hard to learn, perl acknowledges that sometimes you may need to skin a live tiger, so it gives you a rifle as well.)
00:39:47 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
00:41:05CIA-14New commit by fredwbauer (r30754): undef DEBUG in pcm_record.c because it's broken and breaks DEBUG builds
00:42:38CIA-14r30754 build result: 0 errors, 4 warnings (fredwbauer committed)
00:42:45 Quit Keripo (Read error: Connection reset by peer)
00:45:04 Join Keripo [0] (
00:46:18 Join freddyb [0] (
00:47:20 Quit freddyb (Client Quit)
00:49:21Tornebertrik: USBDN is set bits 0x18 of GPIOJ to 0x08 then read adc 5
00:49:39Tornebertrik: USBDP is set bits 0x18 of GPIOJ to 0x10 then read adc 5
00:50:21Tornesetting both bits to 0 selects accessory, it seems
00:50:33Tornethis is for ipodvideo
00:50:38Tornei cna't try this out right now, time for bed :)
00:54:31 Join Keripo1 [0] (
00:58:21 Quit stoffel (Remote host closed the connection)
00:58:29 Quit Keripo (Ping timeout: 252 seconds)
01:02:11 Quit Keripo1 (Quit: Leaving.)
01:28:47*KiwiCam is back (gone 18:42:10)
01:30:45*gevaerts points KiwiCam to item 7 on
01:32:38KiwiCamHi Folks. I'm about to upload a theme. Inside the .rockbox folder, I assume I only need to include a subfolder where I wish something to be uploaded. ie wps,themes are the folders I think I only need. I assume icons,fonts etc aren't required. I am using Artwiz Snap, but I don't need to include this Font do I, as the user will be requested to install the Font Pack.
01:33:46*KiwiCam Reads link. Crosses fingers.
01:36:41 Join kadoban [0] (
01:36:46 Quit bertrik (Ping timeout: 248 seconds)
01:37:59 Quit domonoky (Read error: Connection reset by peer)
01:59:26***Saving seen data "./dancer.seen"
02:23:10 Quit keyb_gr (Ping timeout: 256 seconds)
02:23:30 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
02:33:24polemonsome weird shit is happening to my server
02:33:36polemonI constantly lose SSH connections to it
02:44:38CIA-14New commit by fredwbauer (r30755): Revert 30754 because the codecs won't build correctly, anyway.
02:46:17CIA-14r30755 build result: All green
02:51:53 Join antil33t [0] (
02:54:08KiwiCamI've figured it out. My theme is uploaded now. :)
02:58:03 Part KiwiCam ("Leaving")
02:58:35 Join RiotGear [0] (
02:58:44 Part RiotGear
02:59:08 Join KiwiCam [0] (
03:12:15 Quit ruskie (Quit: ...)
03:13:07 Join ruskie [0] (ruskie@sourcemage/mage/ruskie)
03:19:04 Quit ruskie (Excess Flood)
03:22:06 Join ruskie [0] (ruskie@sourcemage/mage/ruskie)
03:59:27***Saving seen data "./dancer.seen"
04:17:16 Quit amiconn (Disconnected by services)
04:17:16 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:17:17 Quit pixelma (Disconnected by services)
04:17:18 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:17:21 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:17:43 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:26:14*KiwiCam is away: Oh look! What's that over there...
04:33:11 Quit TheSeven (Disconnected by services)
04:33:24 Join [7] [0] (~TheSeven@rockbox/developer/TheSeven)
04:59:20 Quit kadoban (Read error: Connection reset by peer)
05:02:35 Join Rob2223 [0] (
05:04:45 Join kadoban [0] (
05:06:06 Quit Rob2222 (Ping timeout: 255 seconds)
05:07:01 Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
05:08:06 Quit [Saint] (Client Quit)
05:08:21 Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
05:09:02 Quit [Saint] (Client Quit)
05:09:46 Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
05:23:00 Quit robin0800 (Ping timeout: 240 seconds)
05:33:13 Quit ps-auxw (Read error: Operation timed out)
05:35:55 Join ps-auxw [0] (~arneb@2001:470:c807:0:1532:4e5f:2ad3:4123)
05:41:11 Join robin0800 [0] (~robin0800@
05:59:28***Saving seen data "./dancer.seen"
06:04:25 Quit [Saint] (Remote host closed the connection)
06:35:24 Quit ps-auxw (Ping timeout: 240 seconds)
06:40:34 Join ps-auxw [0] (
06:42:28 Quit robin0800 (Quit: Leaving)
07:38:49 Quit hilbert (Quit: Computer has gone to sleep.)
07:59:31***Saving seen data "./dancer.seen"
08:24:42 Quit kadoban (Read error: Operation timed out)
08:53:16 Quit scorche (Ping timeout: 258 seconds)
08:59:21 Join shai [0] (
09:02:43 Join scorche [0] (~scorche@rockbox/administrator/scorche)
09:41:09 Join LarryB [0] (
09:41:52 Quit LarryB (Read error: Connection reset by peer)
09:42:12 Quit zchs (Ping timeout: 265 seconds)
09:45:58 Join LarryB [0] (
09:46:12 Quit LarryB (Read error: Connection reset by peer)
09:50:10 Join Buschel [0] (
09:50:32 Join y4n [0] (y4n@unaffiliated/y4ndexx)
09:59:35***Saving seen data "./dancer.seen"
10:13:38Buschelforget about that
10:13:48 Join bertrik [0] (
10:13:48 Quit bertrik (Changing host)
10:13:48 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
10:15:00 Quit Buschel (Quit: ChatZilla 0.9.87 [Firefox 7.0.1/20110928134238])
10:18:12*bertrik wonders if we have a wiki page with the GPIO mappings for ipods
10:18:47 Join z180 [0] (
10:20:56z180someone wondered why RaAA is compiled for ARM5TE i know the reason
10:22:24z180I have a Win Mobile phone around using a PXA270 XDA has Android ROMS for it
10:22:51bertrikHow do I use a GPIO on portalplayer? Like hat is the polarity of the GPIOx_ENABLE register bits?
10:39:34amiconnipl has a (somewhat incomplete and slightly wrong) overview
10:40:22amiconnOn PP502x the gpio module has built-in atomic bit manipulation
10:40:55bertrikI think I figured out setting a GPIO now, but if i set a few specific GPIOJ pins, the ipod seems to hang (I guess I'm setting the wrong bits for this player)
10:42:08bertrikamiconn, thanks, yes I saw the GPIO_CLEAR/SET_BITWISE macro
10:50:20 Join ender` [0] (
10:51:15 Join dfkt [0] (
10:51:15 Quit dfkt (Changing host)
10:51:15 Join dfkt [0] (dfkt@unaffiliated/dfkt)
10:55:26 Join stoffel [0] (
11:08:10 Join zchs [0] (
11:27:08desowin_where in the buildsystem do I define rockbox bootloader offset (together with entry point) in memory?
11:33:51 Join Horschti [0] (~Horscht@xbmc/user/horscht)
11:36:52 Quit Horscht (Ping timeout: 252 seconds)
11:36:56JdGordonKiwiCam: after reading the last 3 days logs, you
11:37:13JdGordonyou've been warned 3 times... if anyone sees your away message again you'll cop a ban
11:42:26 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
11:43:15 Quit fs-bluebot (Read error: Connection reset by peer)
11:47:30 Join fs-bluebot [0] (
11:49:02 Join pamaury [0] (
11:49:03 Quit pamaury (Changing host)
11:49:03 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
11:58:06 Quit z180 (Ping timeout: 248 seconds)
11:59:37***Saving seen data "./dancer.seen"
12:04:12 Join domonoky1 [0] (
12:04:14 Quit domonoky (Read error: Connection reset by peer)
12:17:37 Join Jaykay [0] (
12:23:06 Quit stoffel (Ping timeout: 276 seconds)
13:01:34 Quit markun (Ping timeout: 248 seconds)
13:13:09Jaykayhi :) when an unrockboxed fuze v1 shuts down randomly a few seconds up to a few minutes after booting, how high is the chance of a hardware failure?
13:19:42 Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
13:23:20 Join TheLemonMan [0] (
13:26:09 Join JdGord [0] (~AndChat@
13:30:24 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
13:41:48 Join hilbert [0] (
13:57:29 Join T44 [0] (
13:59:41***Saving seen data "./dancer.seen"
14:01:27 Quit Topy (Ping timeout: 260 seconds)
14:08:46 Quit avacore^ (Ping timeout: 248 seconds)
14:10:49 Join avacore [0] (
14:22:22 Join keyb_gr [0] (
14:35:02 Quit JdGord (Quit: Bye)
14:46:58 Quit mystica555_ (Quit: Ekkusu Chatto)
14:47:25 Join mystica555 [0] (
15:03:34 Quit keyb_gr (Ping timeout: 258 seconds)
15:10:12 Quit Horschti (Quit: Verlassend)
15:29:17 Join dfkt_ [0] (dfkt@unaffiliated/dfkt)
15:31:05 Quit dfkt (Ping timeout: 252 seconds)
15:33:11 Join freddyb [0] (
15:35:54freddybWe do not do a dircache_disable() on a battery safe shutdown in misc.c. Does anyone know if this is intentional or an oversight?
15:42:07 Nick [Sanyi] is now known as sanyi (thg@gateway/shell/
15:42:23 Nick sanyi is now known as [Sanyi] (thg@gateway/shell/
15:43:14 Nick dfkt_ is now known as dfkt (dfkt@unaffiliated/dfkt)
15:57:56 Quit hilbert (Quit: Computer has gone to sleep.)
15:59:44***Saving seen data "./dancer.seen"
16:05:34 Join hilbert [0] (
16:06:06 Quit Jaykay (Ping timeout: 248 seconds)
16:14:41 Quit tchan (Quit: WeeChat 0.3.5)
16:19:08 Join keyb_gr [0] (
16:29:26 Quit Llorean (Read error: Connection reset by peer)
16:33:19 Join Jaykay [0] (
17:10:36 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
17:14:39 Quit antil33t (Read error: Connection reset by peer)
17:15:02 Join antil33t [0] (
17:21:43 Quit TheLemonMan (Quit: WeeChat 0.3.5)
17:26:52 Quit zchs (Ping timeout: 260 seconds)
17:33:48 Join zchs [0] (
17:42:24 Join thomasjfox [0] (~thomasjfo@rockbox/developer/thomasjfox)
17:52:59 Join kadoban [0] (
17:59:46***Saving seen data "./dancer.seen"
18:32:51 Join fyrestorm [0] (
18:46:38 Join stoffel [0] (
18:51:22pamauryfreddyb: dircache doesn't cache stuff that isn't yet on disk iirc so there is no need to disable it (ie it will no incure any data loss)
18:52:05freddybpamaury: Thanks.
18:53:42pamaurybut that's just a possible explaination, perhaps it's just an oversight :)
18:57:30 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
18:58:45 Join mortalis [0] (
19:16:27 Join kadoban_ [0] (
19:17:08 Quit kadoban (Read error: Operation timed out)
19:25:24 Join kadoban__ [0] (
19:25:35 Join lovasoa [0] (~olojkine@2a01:e35:8a2e:8080:e2b9:a5ff:fe5b:ca7b)
19:26:51 Part lovasoa
19:27:06 Quit kadoban_ (Ping timeout: 258 seconds)
19:29:52 Quit hilbert (Quit: Computer has gone to sleep.)
19:30:30 Join hilbert [0] (
19:32:16 Quit mortalis (Quit: CGI:IRC)
19:34:50 Join markun [0] (~markun@rockbox/developer/markun)
19:42:04 Quit hilbert (Ping timeout: 255 seconds)
19:49:33 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
19:58:51 Quit thomasjfox (Remote host closed the connection)
19:59:47***Saving seen data "./dancer.seen"
20:02:19 Quit domonoky1 (Quit: Leaving.)
20:07:14ZagorJdGordon: your brace policy is inconsistent! *shakes fist*
20:10:29 Quit sinthetek (Ping timeout: 260 seconds)
20:13:44 Quit dfkt (Quit: -= SysReset 2.55=- Sic gorgiamus allos subjectatos nunc.)
20:15:46 Join robin0800 [0] (~robin0800@
20:35:10Zagorcalc_brpos() in plugins/lrcplayer.c looks really dodgy. it's using the function parameter i as loop variable in a FOR_NB_SCREENS macro. and keeps using it afterwards. either it's very wrong or it's very obfuscated.
20:36:05*Zagor gives it a wide berth
20:41:00 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
20:41:41mc2739Zagor: maybe the reason for FS #12306?
20:41:41fs-bluebot The Lyric can not wrap around if it's too long to fit the screen (lrcplayer.c) (bugs, unconfirmed)
20:43:13 Join sinthetek [0] (~sinthetek@unaffiliated/sinthetek)
20:52:29 Quit guymann (Ping timeout: 260 seconds)
20:54:13 Join guymann [0] (
20:56:04 Quit robin0800 (Quit: Leaving)
20:56:55Zagormc2739: that rather sounds like buschels r30565 commit borked something
20:56:58 Join robin0800 [0] (~robin0800@
21:02:37 Quit robin0800 (Ping timeout: 252 seconds)
21:03:52 Join shai_ [0] (
21:04:18 Join robin0800 [0] (~robin0800@
21:04:39bertrikI think it's not strictly wrong to use/modify a function parameter, but it does look a little odd
21:05:18Zagorit's not wrong per definition, but it doesn't look intentional
21:07:19 Quit shai (Ping timeout: 255 seconds)
21:09:36 Quit stoffel (Ping timeout: 276 seconds)
21:11:47bertrikTorne, I tried the exact same GPIOs for measuring USB D+/D- on my ipod nano 1g, but setting them make my ipod hang. Was worth a try :)
21:13:00bertrikAny luck with your ipod video yet?
21:14:19 Quit robin0800 (Ping timeout: 255 seconds)
21:26:29 Join robin0800 [0] (~robin0800@
21:28:13LalufuOK, I don't know what the correct way to do this is, but I'd like to continue to use a static buffer in IAP for storing RX/TX data.
21:28:53LalufuFrom my tests it looks like dynamically allocating that memory on demand takes too long, and some devices simply stop talking if the iPod does not respond in time
21:29:39Lalufuthe buffer would be 64k+512+6 bytes, with the current code.
21:30:16LalufuIIRC iPods have lots of RAM, compared to other players, so the impact should not be that bad.
21:32:25[Saint]A few seconds of audio buffer.....nooooooo! :)
21:32:47LalufuYes, that's the reason I tried the dynamic memory route
21:33:05bertrikLalufu, I think there is a accessory response to indicate that processing the command takes a bit longer
21:33:15Lalufuyes, there is.
21:33:20bertrikand why is a 64k buffer needed?
21:33:36Lalufubecause the device could send 64k of data. the spec allows that.
21:34:26[Saint]Wouldn't that only be for things we wouldn't support, like logo display?
21:34:53Lalufuwho says we don't support that?
21:35:06Lalufuwell, we currently don't, but it's on my to-do list
21:35:06CIA-14New commit by zagor (r30756): Changed the FOR_NB_SCREENS macro to always be a for loop that declares its own loop variable. This removes the need to declare this variable in the ...
21:37:40CIA-14r30756 build result: 0 errors, 8 warnings (zagor committed)
21:38:22 Quit simonlnu (Quit: brb)
21:39:28bertrikLalufu, my general feeling about iap is that iap stuff has been implemented not really following any design idea, so we may need to rethink/refactor some stuff
21:40:43amiconnZagor: Checked the delta table? This is exactly the reason why FOR_NB_SCREENS should *not* be a loop for single screen
21:41:36Zagorthat's odd. I did test compiles before commit, and got the same code.
21:41:58amiconnI doubt you checked all locations
21:41:59Zagorgcc -O sees that the loop is not a loop, and removes it
21:42:21amiconnGcc is weird wrt optimisations like these - rather don't rely on them
21:43:08amiconnBtw, FOR_NB_SCREENS could be kept like before, and still save the extra declaration
21:43:36amiconnIt would make use of a C99 feature though
21:43:55amiconn(mixed code and declarations)
21:44:29amiconnCould even say 'const int i' in that case
21:44:46Zagoryou need to create a new block
21:45:06bertrikLalufu, I think a reasonable way is to lazily claim a full 64k buffer upon detection of an accessory
21:45:23amiconnThe pre-r30756 way is cleaner though
21:45:40Zagorbecause we have multiple FOR_NB_SCREENS in the same functions. having each be a new declaration causes warnings
21:45:59Zagorabsolutely not cleaner. I removed 143 "int i"!
21:46:24amiconnThe old way made it explicit which variable to use
21:46:40Zagorthat did not change
21:46:45amiconnThe new one might shadow another variable without notice
21:47:17amiconnIf you have 'i' in the outer scope and then use FOR_NB_SCREENS(i), the outer i gets shadowed
21:47:47Zagorif we had code that relied on this, it would be very buggy.
21:47:48amiconnBtw, you overlooked a few 'int i'
21:47:55Zagoryes, I saw
21:48:00CIA-14New commit by zagor (r30757): Missed two spots.
21:48:26ZagorI enabled -Wshadow during this work to ensure we don't have this problem
21:48:55Zagorthe only questionable case is lrcplayer as I mentioned earlier
21:50:08CIA-14r30757 build result: All green
21:51:55 Join simonlnu [0] (3RIpgWQEVY@unaffiliated/simonrvn)
21:51:56 Quit [Saint] (Remote host closed the connection)
21:54:11*amiconn wonders whether misusing 'if' might work
21:54:48bertrikhm, calculator has a function which ends in return 0, but then has some code after that ...
21:55:18 Quit shai_ (Quit: Leaving)
21:55:52bertrikI think I'm just going to remove the extra code
21:59:48***Saving seen data "./dancer.seen"
21:59:53 Quit mystica555 (Remote host closed the connection)
22:06:20Zagoryikes, -Os gives a 37272 bytes smaller fuze+ binary
22:08:59 Join mystica555 [0] (
22:09:04 Join petur [0] (~petur@rockbox/developer/petur)
22:10:10 Quit mystica555 (Client Quit)
22:10:22Lalufubertrik: I currently do that, but allocating that memory takes a few seconds. and some devices simply do not repeat their initial message, which is missed because the buffer is being allocated
22:10:39 Join mystica555 [0] (
22:10:39amiconnZagor: I have a somewhat odd solution:
22:11:56amiconnIf you use #define FOR_NB_SCREENS(i) for(int i = 0; !i; i = 1) if NB_SCREENS = 1, the binary for mini2g becomes 272 bytes smaller
22:13:47amiconnThis is still relying on gcc optimisation of course
22:14:30amiconnHrmm, but it makes it larger on sh1
22:14:38amiconngcc is weird...
22:15:13amiconnSo definitely the loop does not (always) get optimized away
22:15:19Zagorsh is using the oldest gcc
22:15:40amiconnYes, but the increase in r30756 was on arm
22:15:44bertriktime to upgrade it ? :)
22:15:54Zagoramiconn: yes
22:16:10amiconnbertrik: Sure, but first rockbox needs to become stable again (for all architectures)
22:16:57Lalufuamiconn: I've stared at the ARM assembly for the current FOR_NB_SCREENS macros for some time, and the for loop for i=1 does get optimized away (there is no loop in the assembler any more, but the result is structured differently
22:17:56ZagorLalufu: where did you look? in my separate test code I got identical result.
22:18:31LalufuI also get the same result for a test case function.
22:18:48LalufuI looked at... let me dig..
22:19:59amiconnI think gcc is too complex for test cases to be useful
22:21:01amiconnThe same source often produces different assembler code (not just changing registers) depending on the surrounding code
22:22:19Zagorregarding getting stable, I'd like to have flyspray bugs for all release-critical bugs.
22:24:41LalufuZagor: list_init_viewports(), from apps/gui/list.c
22:26:56 Join kadoban [0] (
22:27:34 Quit kadoban__ (Ping timeout: 260 seconds)
22:28:46Lalufubertrik: I have thought about using a small (512RX. 512TX) static buffer to get IAP going and to allocate a larger buffer in the background, to hopefully be ready when the first large packet arrives, but that would make the scheme even more complex than it is now, requiring additional locking to make sure you don't swap buffers while something is still using them
22:29:54bertrikI'm a bit surprised that allocation is that slow to be honest
22:30:23 Join MethoS- [0] (~clemens@
22:31:36freddybI'm trying to get the glyphfiles working and I get duplicate copies of the file in the same directory. If I put a sleep(HZ/10) in between glyph saves it seems to fix the problem. Do we have a global fsync or anything better than a fixed delay?
22:32:53Lalufubertrik: if the player is playing (which it is in my tests) buflib has to shrink the audio buffer
22:32:58 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
22:33:00Lalufuin order to get the memory I want
22:33:32 Quit Jaykay (Ping timeout: 256 seconds)
22:33:39amiconnfreddyb: It's not just the glyphcase that's broken. Resume position is also not saved
22:33:42Zagorah, sh has -Os while arm has -O
22:34:17amiconnOn coldfire (H1x0) a shutdown after playing music causes a hard freeze (requires paperclip to reset)
22:34:43amiconnThis is with voice enabled
22:35:39bertrikPut it on flyspray. Maybe we should make a release wiki page with links to bugs like these to be resolved before release.
22:35:53amiconnAlso talk clips sometimes don't work
22:36:07amiconn(on hwcodec: s/sometimes/always/)
22:36:33 Quit kadoban (Remote host closed the connection)
22:36:38ZagorI think we could just tag them critical on flyspray
22:36:41 Join Rob2222 [0] (
22:37:25Zagoror prio "urgent", since the default view sorts on prio
22:38:02freddybShould I go ahead and submit the glyphcache fix with the delay until the i/o system is figured out? Glyphcache save is completely broken in svn right now.
22:38:10CIA-14New commit by bertrik (r30758): plugins: another round of making local functions static
22:38:39 Quit y4n (Quit: PANTS OFF!)
22:39:57CIA-14r30758 build result: 0 errors, 78 warnings (bertrik committed)
22:40:37KiwiCamAre there any Clip owners here today? I'm wondering if someone might check my theme for me. My testing shows that the .zip file is fine, and it installs using the Utility, but I'm getting overly paranoid about it and I'd like someone to just confirm that the theme works on someone else's Clip. Thanks.
22:41:04 Quit Rob2223 (Ping timeout: 256 seconds)
22:41:32UnhelpfulKiwiCam: if it works for you, post it to the theme site (which validates themes)
22:41:35 Join wodz [0] (
22:42:31bertrikmeh, why are those functions called main, when they're not really the entry function
22:42:39wodzIs there any fundamental reason why we build with -O on arm? I thought it was for historical reasons.
22:43:14bertrik-O instead of what?
22:43:41freddyb-O2 rulz!
22:43:55wodzfreddyb: not in embedded world :-)
22:44:33Zagorwodz: I don't know. codecs use their own optimization flags, so I don't think it affects performance much
22:45:18wodzZagor: thats the point - it shouldn't affect performance and should save quite some binsize
22:45:43Zagor<Zagor> yikes, -Os gives a 37272 bytes smaller fuze+ binary
22:45:55wodzBTW. any comments about FS #12331?
22:45:56fs-bluebot move PortalPlayer stuff to subdir (patches, new)
22:46:07amiconnThe core has some performance critical code as well
22:46:35Zagoramiconn: sure, but we run Os on clip and c200v2 too
22:46:37amiconnThat said, -Os should usually perform better than -O1
22:46:54wodzamiconn: -O is unlikely to produce better (in terms of performance) then -Os
22:46:57amiconnZagor: Yes - these are lowmem targets
22:47:18 Quit benedikt93 (Read error: Connection reset by peer)
22:47:53 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
22:48:59Zagorwodz: I'm not sure I like adding yet another directory level
22:49:34Zagorthough I haven't thought very much about it yet
22:49:34wodzZagor: Do you prefer current mess?
22:49:56bertrikWhat is the problem really?
22:50:03wodzand not to say about misuse manufacturer as soc name
22:50:31wodzbertrik: all this pp files laying in target/arm
22:51:01wodzI guess this is relict from the days PP was the only arm supported
22:51:14CIA-14New commit by bertrik (r30759): Fix warnings
22:51:17amiconnWe don't need an extra level
22:51:40wodzamiconn: what do you propose?
22:52:05amiconnJust restructure the portalplayer targets the same way as the later arm targets
22:52:38wodzamiconn: this is not feasible - I looked into it
22:53:06 Join kadoban [0] (
22:53:11amiconnWhy not?
22:53:18CIA-14r30759 build result: All green
22:53:35wodzamiconn: ipods share different bits of code with virtually random scheme
22:53:46amiconnJust move target/arm/ipod -> target/arm/portalplayer/ipod
22:54:05amiconnThe few files in subdirs of 'ipod' could just be moved one level up
22:54:13wodzamiconn: thats exactly what I did
22:54:48wodzamiconn: and this subdirs in ipods are actual models with different code
22:55:10amiconnYes, but the files have different names and there aren't many
22:55:37amiconnMaybe the .h files need some extra attention
22:56:27amiconnWe could split pp502x and pp5002 (instead of having just 'portalplayer)
22:56:42*amiconn wonders what's the status of the sansa view
22:57:37amiconnThe diff listing in flyspray seems to be buggy
22:57:41amiconn^ Zagor
22:58:13wodzamiconn: hmm this might be an option
22:58:35Zagoramiconn: nope, that's how the diff looks
22:58:44amiconnHmm, then the diff is broken
22:59:08 Join Kiwi_Cam [0] (
22:59:30 Part Kiwi_Cam
22:59:43amiconnwodz: Sansa view is pp6100 iirc (and I don't know the status of the port)
22:59:54BagderI believe that port is pretty much dead
23:00:06Zagoractually no, it's a diffstat bug
23:00:48 Join hilbert [0] (
23:00:59wodzsansa view is in unusable state AFAIK
23:01:27Bagderafair there's no working lcd driver
23:01:40Bagderwhich makes it fairly unusable =)
23:02:11 Join Buschel [0] (
23:02:24 Join Jaykay [0] (
23:03:00wodzamiconn: does ipod 1g and 2g share the same SoC?
23:06:54 Quit Jaykay (Ping timeout: 248 seconds)
23:06:54amiconn1g, 2g and 3g all have PP5002
23:07:18amiconn1g and 2g even use the same build (as they are even undistinguishable by ipodpatcher)
23:07:23Zagordiffstat has no clean way to visualize moved files
23:08:24wodzamiconn: although this might be the route it will be a bit odd to treat all ipods as single target in sourcetree
23:09:16 Quit MethoS- (Remote host closed the connection)
23:11:27Zagornano2g and ipod6g are in s5l8700/ and s5l8702/, respectively
23:12:39Zagorif we are to follow the current pattern, we should probably call the soc dir pp50xx
23:13:00Zagorand then have pp50xx/ipodvideo/
23:13:22wodzZagor: or split it into pp5002 and pp502x
23:13:32wodzand remove part of ifdef hell
23:14:17Zagorwon't that require quite a bit of duplication?
23:14:19Buschelfreddyb: seen my comment on FS?
23:14:30ZagorI see a lot of *-pp.c files
23:14:39 Quit petur (Quit: Leaving)
23:14:54freddybI do now. :)
23:15:26wodzZagor: I am not sure but this pp files are HEAVY ifdefed so at the end it may not be that much in common
23:15:47Buschelfreddyb: it is a very short comment though :)
23:16:10freddybI'll look it over and commit soon. Thanks for your patience...
23:16:46Buschelfreddyb: do you have any idea what causes this? i have a bad feeling of adding a sleep() to get it working
23:17:31Zagorwodz: the big ones are ata-sd-pp.c, pcm-pp.c and thread-pp.c. while they have some soc ifdefs, they also share a lot of code.
23:19:46wodzUnfortunately I see the best option is not to touch historical code. Our project tend to like the mess
23:20:11freddybI don't know. Amiconn said resume position isn't saved either. Threading weirdness?
23:20:33Buschelamiconn: what target does not save resume position?
23:20:36Zagorwodz: yes, it's probably best to work in small steps. make a pp50xx dir first. we can always split it later.
23:20:40amiconnBuschel: H1x0
23:21:12amiconnI guess most (all?) status saving is broken after playback, at least if voice is enabled
23:21:49BuschelI do not have voice enabled... (and have no issues with resume or other status)
23:22:52Buschelour trunk does not seem to be in the best shape these days
23:22:52Zagorwodz: but don't add a new manufacturer level, use ipodvideo rather than ipod/video.
23:23:01amiconnWhen using the H1x0 in the car (which is my most common use case for it atm) I currently have to hard reset at shutdown and have to memorise resume position
23:23:34Buschelany idea since when this is happening?
23:24:28amiconnI don't have an exact revision, as I didn't update often during summer. Also several revisions were broken in different ways
23:24:53amiconnAfaik it started when buflib has been added to the core
23:25:00Zagorwhich revision are you running now?
23:26:51amiconnr30738 on hwcodec (no freeze observed so far, but non-working .talk clips)
23:28:13*amiconn rebuilds
23:32:11 Quit tjb0607 (Read error: Operation timed out)
23:36:21 Join tjb0607 [0] (
23:38:33 Join ChickeNES [0] (~ChickeNES@
23:40:57 Quit Buschel (Quit: ChatZilla 0.9.87 [Firefox 7.0.1/20110928134238])
23:41:27ZagorLalufu: I think a working IAP is worth 64 KB
23:44:50 Quit tjb0607 (Remote host closed the connection)
23:45:00 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
23:45:00 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
23:47:05 Join tjb0607 [0] (
23:47:35 Quit wodz (Quit: Leaving)
23:47:56 Join petur [0] (~petur@rockbox/developer/petur)

Previous day | Next day