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 2010-07-16

00:00:17 Quit jgarvey (Quit: Leaving)
00:00:23***Saving seen data "./dancer.seen"
00:01:09CIA-6New commit by nls (r27440): Don't duplicate byteswap code. Invent system where NEED_GENERIC_BYTESWAPS is set if the generic functions from system.h are needed.
00:02:36 Quit halmi (Read error: Connection reset by peer)
00:03:37 Quit slck (Ping timeout: 276 seconds)
00:03:59 Quit Battousai (Quit: No Ping reply in 180 seconds.)
00:04:22 Join Battousai [0] (~bryan@gentoo/developer/battousai)
00:05:57CIA-6r27440 build result: All green
00:06:27 Join funman [0] (~fun@rockbox/developer/funman)
00:06:29 Join slck [0] (Venci@Slackware.SlackPix.Com)
00:11:32saratogado we support wavpack lossy?
00:13:46 Join jHoNDoE [0] (
00:13:56 Quit captainkewlll (Quit: Page closed)
00:15:05 Quit domonoky (Read error: Connection reset by peer)
00:16:08 Quit jhMikeS (Read error: Connection reset by peer)
00:19:48 Join jhMikeS [0] (
00:19:48 Quit jhMikeS (Changing host)
00:19:48 Join jhMikeS [0] (~jethead71@rockbox/developer/jhMikeS)
00:19:51 Join dfkt_ [0] (dfkt@unaffiliated/dfkt)
00:20:24 Quit dfkt (Disconnected by services)
00:20:28 Nick dfkt_ is now known as dfkt (dfkt@unaffiliated/dfkt)
00:27:14 Quit GArik_ (Quit: Leaving)
00:39:04 Quit Jerom (Read error: Connection reset by peer)
00:40:52funmani've found the fuzev2 version check, the OF checks if the new file has a version >= .31 located at 0x500 in the file
00:43:40saratogadoes mkamsboot set the version number?
00:44:10funmanno, this part is just OF code that we replace
00:44:34funmanin all other OF, and fuzev2 OF < 2.3.31 it's also random code
00:51:32funmanin our patched OF it is dualboot.S code
00:52:01 Quit DerPapst (Read error: Connection reset by peer)
00:52:56 Join DerPapst [0] (
00:53:29 Join djzn [0] (
00:53:46djznis it true that MPC has better battery life than AAC, OGG and MP3 ?
00:58:06 Quit DerPapst (Quit: Leaving.)
00:58:42djznthat meant yes ?
00:58:53 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
01:01:28 Quit efyx (Remote host closed the connection)
01:02:56 Part djzn
01:03:26n1sdjzn: basically yes but the differences aren't exactly huge and it depends on what player you are using
01:04:41saratogadjzn: well 2.15MHz better then vorbis, depends if you care about 2Mhz
01:08:42funmani have a patch for fuzev2 ready
01:11:28n1ssaratoga: how does the performance change if you run the cpu at full speed?
01:12:57n1snevermind, i see comparison numbers on the wiki page
01:15:57kugelfunman: nice find!
01:16:14 Join riad [0] (
01:17:33funmankugel: we could've seen it by looking a bit closer at the OF
01:18:54kugelfunman: could we maybe restore the code areas dualboot.S overwrites before branching to the of?
01:19:21funmankugel: this field is checked by the OF doing the upgrade
01:19:54kugeloh, you mean it that way
01:20:29funmankugel: , i'm going to test it in skyeye before running it
01:21:38funmanbasically i just put the dualboot at 0x200 instead of 0x0 and insert a branch at 0x0
01:25:42 Quit funman (Quit: free(random());)
01:26:23 Join funman [0] (~fun@rockbox/developer/funman)
01:27:17kugelfunman: why do you generate the other vectors with mkamsboot?
01:27:22 Quit JdGordon (Ping timeout: 240 seconds)
01:27:30funmanwhy not?
01:28:04kugelah wait, I get it now :)
01:29:32kugelis there a need to change them at all?
01:30:28funmanyes, better than random vectors that the OF use
01:30:47funmanwe used to just put nothing there, but ranma changed them to infinite loops when he used JTAG
01:31:08kugelhm, ok
01:32:39kugelin my of analysis they were not random IIRC
01:33:49funmanwell, they are meaningful only in the OF, not in our dualboot code
01:35:15funmanhmm how can i specify the entry point of a .S file using as/ld ? ld -e=0xdeadbeef doesn't work, the elf file keeps using 0x8000
01:39:30 Quit bertrik (Quit: :tiuQ)
01:41:13funmannevermind i can force laoding address in skyeye. the patch doesn't seem to work
01:43:49funmanwait, skyeye doesn't seem to work, same effect when simulating a known to work patched OF
01:47:00 Quit dfkt (Quit: -= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.)
01:47:33saratogan1s: FWIW the performance unboosted is most important, since on AMSv2 few formats actually boost during normal playback
01:48:03kugelsaratoga: all or only the 24mhz ones?
01:48:32saratogawell i expect the 24MHz ones to boost more often then the 40MHz ones, but they'll still spend nearly all their time unboosted
01:49:08n1ssaratoga: i was mostly interested in how it
01:49:22n1schanges with cpu clock, is the memory speed constant?
01:49:57saratogan1s: yes, the memory is locked at the unboosted speed
01:50:21saratogaso the fuzev2 is actually a good bit faster then teh clip+ since its memory runs at 40MHz rather then 24MHz
01:50:29kugelfunman: could the vector table be generated in the same way (with bin2c) as dualboot? would make the mkamsboot code a bit less akward
01:50:30saratogai benched all the combinations
01:50:41funmankugel: it could, but i think it's simpler this way
01:50:47n1sso at boosted speed, more time will be spent waiting for memory
01:50:54kugelI think it's simple either way :)#
01:51:03saratogascaling is surprisingly poor
01:51:10funmankugel: i'm going to commit the patch, it works on both fuzev2
01:51:12saratogai think we run the boost clock much too high to be efficient
01:51:33saratogaon CF and PP the boost to unboost ratio was about 2.5
01:51:37saratoganow its 10
01:52:12funmanthe new fuzev2 can be patched, but it gives ATA error -2 at boot
01:52:28kugelsaratoga: you cannot really compare that
01:53:07kugela) both PP and CF could be clocked lower and b) neither off such a high max clock
01:53:36saratogayes that was my point
01:53:58n1ssaratoga: it would be interesting to know how power usage compares at different boost clockspeeds
01:54:01saratogathe max clock is much higher, which means efficiency is much different boosted then unboosted
01:54:52funmankugel: ok for committing the patch?
01:55:06kugelwith the ata error?
01:55:17kugelIMO that should be resolved first
01:55:24funmankugel: well the mkamsboot patch is not related
01:55:45funman1/ make it possible to patch it
01:55:50funman2/ make it possible to patch it with working code
01:55:57n1smaybe a hw difference is the reason for this version check?
01:55:58kugelyes and no, fixing mkamsboot doesn't serve anything if rockbox is completely unfunctional
01:56:07funmansure it does
01:56:12funmanpeople can help making it functional
01:56:15funmanand the LCD works
01:56:27funmann1s: very probably, now we have to find why
01:56:36saratogai think committing often is best
01:56:48funmankugel: FWIW mkamsboot worked long before storage worked on AMSv12
01:57:03funmanwell on v1 and v2 anyway
01:58:10kugelhm, that's true
01:58:33kugelI would still prefer generating the vectors with bin2c, but do whatever you want
02:00:24***Saving seen data "./dancer.seen"
02:00:34CIA-6New commit by funman (r27441): mkamsboot: fix for newer fuzev2 ...
02:02:27funmankugel: FWIW, I wrote the b 0x200 as just "b 0x200" in the .S file
02:02:35funmanno need to make fancy padding + label
02:02:45CIA-6r27441 build result: All green
02:04:17kugelfunman: what do you mean?
02:04:54 Part jHoNDoE
02:05:53CIA-6New commit by funman (r27442): mkamsboot: update outdated comment
02:06:32funman'b 0x200' is a relative branch, it will branch 0x200 bytes forward, not set pc to 0x200
02:07:40kugeli know
02:07:40CIA-6r27442 build result: All green
02:07:54funmani didn't :)
02:09:41 Quit GeekShadow (Read error: Connection reset by peer)
02:13:44 Join JdGordon [0] (3a601b9e@gateway/web/freenode/ip.
02:22:44 Join CaptainKwel [0] (
02:28:57 Quit n1s (Quit: Lämnar)
02:31:17funmanthe fuzev2 doesn't boot after an OF upgrade
02:31:50funmanah no it's ok
02:31:54*funman slaps self
02:32:59JdGordonfunman: skin_parser.c?
02:33:08JdGordon(just reading the log now)
02:33:19funmanJdGordon: nevermind, it was about the bug mentioned on the mailing list
02:34:13JdGordongevaerts: conditional viewports should be working (they work for me...).. can you comment in the tracke please?
02:38:59 Quit kugel (Remote host closed the connection)
02:43:16funmanthe fuzev2 model seems to affect SD and some power settings
02:56:12 Quit S_a_i_n_t (Ping timeout: 240 seconds)
02:56:51funmanso there is at least 2 c200v2 revisions, 2 clipv1 revisions, 2 fuzev2 revisions, 2 clip+ revisions .... you'd better make a fuzev3 Sandisk!!
02:59:38 Quit funman (Quit: free(random());)
03:02:30 Join S_a_i_n_t [0] (S_a_i_n_t@
03:05:25CIA-6New commit by jdgordon (r27443): prevent the parser crashing if a font fails to load and is refresencd in more than one viewport
03:07:26CIA-6r27443 build result: All green
03:14:16saratogafuze v2.5
03:14:28 Join [Saint] [0] (S_a_i_n_t@
03:16:02 Quit S_a_i_n_t (Ping timeout: 276 seconds)
03:17:56 Join JdGordon| [0] (3a601b9e@gateway/web/freenode/ip.
03:34:44 Quit Topy44 (Ping timeout: 260 seconds)
03:42:42 Join bunnyboi [0] (
03:42:44 Join Topy44 [0] (~topy@2a01:608:ffff:2627:2e0:4dff:fe83:a45e)
03:45:52 Quit JdGordon| (Ping timeout: 252 seconds)
03:48:28 Quit binaryhermit (Quit: Leaving)
03:50:43 Join binaryhermit [0] (
04:00:27***Saving seen data "./dancer.seen"
04:06:07 Quit pixelma (Disconnected by services)
04:06:07 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:06:12 Quit amiconn (Disconnected by services)
04:06:14 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:06:22 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:06:34 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:26:32 Quit TheSeven (Ping timeout: 240 seconds)
04:30:22 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:32:28 Quit scorche (Disconnected by services)
04:32:39 Join scorche` [0] (~scorche@rockbox/administrator/scorche)
04:44:02 Join komputes [0] (~komputes@ubuntu/member/komputes)
04:50:24 Join Barahir_ [0] (
04:52:30 Quit bunnyboi (Ping timeout: 276 seconds)
04:53:36 Quit Barahir (Ping timeout: 245 seconds)
05:00:04 Join bunnyboi [0] (
05:02:52 Nick scorche` is now known as scorche (~scorche@rockbox/administrator/scorche)
05:04:33 Quit MethoS- (Remote host closed the connection)
05:14:47 Quit bunnyboi (Ping timeout: 240 seconds)
05:23:43 Quit komputes (Remote host closed the connection)
05:39:17 Quit Horscht (Quit: Verlassend)
05:42:37 Join bunnyboi [0] (
05:42:55 Nick bunnyboi is now known as jennifur (
05:49:55 Quit Topy44 (Ping timeout: 248 seconds)
05:50:44 Join Topy44 [0] (~topy@2a01:608:ffff:2627:2e0:4dff:fe83:a45e)
05:52:26 Quit panni_ (Read error: Connection reset by peer)
06:00:30***Saving seen data "./dancer.seen"
06:03:03 Quit CaptainKwel (Quit: Ex-Chat)
06:28:55 Join NTKM [0] (
06:30:30 Part NTKM
06:32:45 Join NTKM [0] (
06:40:04 Part NTKM
06:50:20 Quit jennifur (Quit: I was raided by the FBI and all I got to keep was this lousy quit message!)
07:18:12 Join bieber_ [0] (
07:26:41 Quit JdGordon (Quit: Page closed)
07:44:12 Quit binaryhermit (Read error: Connection reset by peer)
07:50:54 Join pixelma_ [0] (~pixelma@rockbox/staff/pixelma)
07:52:52bieber_Do we have a list somewhere of features/screen resolutions of the various supported targets?
07:56:04Strife89I don't suppose there's a quick way to identify a file that's causing a Database build to lock up?
07:56:35pixelma_bieber: could help as astart, newer targets are missing though I think
07:57:29bieber_pixelma_: Thanks :)
07:58:54pixelma_maybe there is something to find in the theme page (sources) too but I don't know if there is a simple list or you have to still find things.
07:59:19pixelma_(I mean searching around in the code)
08:00:19*pixelma_ thinks she knows the screen resolutions by heart and could answer questions
08:00:34***Saving seen data "./dancer.seen"
08:02:29bieber_I just want to grab a couple to test with for now, I'll fill it all the way out later
08:05:14 Join mitk [0] (~mitk@
08:08:05 Quit pixelma_ (Quit: .)
08:10:57 Quit shai (Ping timeout: 240 seconds)
08:16:43 Join jslegers [0] (
08:16:52 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
08:19:26bieber_pixelma: Do you know the size of the remote lcds on the m:robe units?
08:20:35jslegersI am trying to update a WPS&Theme that I created for my Sansa Fuze. I can get it to work in the simulator, but it fails when I put it on the Fuze itself
08:20:57jslegersWhere could I find a log-file of the error messages that it generates?
08:25:06 Join shai [0] (
08:28:11pixelmabieber_: I believe it is something weird - 79x16 mono
08:28:40bieber_You'd think this would be easier to Google for :/
08:29:12pixelmaand they have a few fixed icons extra (like the charcell display) but I don't think those are supported by Rockbox yet
08:34:28 Quit TheSeven (Ping timeout: 265 seconds)
08:39:55 Join wodz [0] (
08:42:02wodzany objections to commit FS #11324 ?
08:44:02 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
08:50:34 Quit bieber_ (Remote host closed the connection)
08:52:42wodzamiconn: Your patch to hd200 grey bliting gives 152.0 (47% load) -> 216.0 (24% load) speedup ~42%
08:54:27 Quit JdGordon (Ping timeout: 245 seconds)
08:55:14 Quit Topy44 (Remote host closed the connection)
08:57:27amiconnwodz: That's actually a 100% speedup of the routine. lcd_blit_grey() is called from the isr, and the load tells how much time is spent there
08:58:45amiconnOf course the cpu can update the framebuffer more often if it spends less time in the isr. 153fps in 53% of the time matches very well with 216fps in 76% of the time
08:59:21amiconnI guess your statement means that the patch works properly and can be committed?
08:59:44 Join NTKM [0] (
09:00:16 Part NTKM
09:00:29wodzhave You read my comment about move.w -> move.b change?
09:04:18wodzanyway the patch is ok for commit
09:04:36amiconnThe timing is the same, so if the OF uses move.w, we should do the same.
09:04:46wodzit works with move.b
09:04:51wodzI checked
09:05:09wodzevery write to the lcd is byte actually
09:05:42amiconnI know, but then using move.w doesn't hurt either
09:06:25wodzit doesn't hurt but if someone will read this routine it gives the impression that lcd bus is 16bit wide which is not true
09:07:01wodzin OF some functions use move.b some move.w
09:07:23wodzIt think it comes from explicit casting in OF source code
09:08:41CIA-6New commit by amiconn (r27444): MPIO HD200 grey blitting: Use line reads for the pixel values as well. Almost doubles the ISR speed (47% -> 24% load), giving 42% faster greylib ...
09:09:54*amiconn actually forgot to mention another effect
09:10:14amiconnThis version also avoids back-to-back single writes (which cause pipeline stalls)
09:10:20 Join NTKM [0] (
09:10:33CIA-6r27444 build result: All green
09:12:08 Join petur [0] (d408b802@rockbox/developer/petur)
09:13:11wodzIf you don't mind I would prefer to change move.w -> move.b for readability reasons (in whole lcd-as-hd200.S)
09:13:40 Join bor_ka [0] (
09:13:42amiconnWell, if it works, just do that if you want
09:14:00bor_karanma, ping?
09:17:54 Quit NTKM (Ping timeout: 246 seconds)
09:20:02 Quit jslegers (Quit: CGI:IRC (EOF))
09:22:16ranmabor_ka: pong, but likely no time to have a look today ^^;
09:24:09CIA-6New commit by wodz (r27445): HD200: Change move.w -> move.b in lcd-as-hd200.S when doing transfers to lcd to explicity show that lcd bus is 8bits wide.
09:25:46 Join funman [0] (~fun@rockbox/developer/funman)
09:25:49CIA-6r27445 build result: All green
09:26:07bor_karanma, ok. When you have time, take a look at ?
09:32:35funmansaratoga: damn, it seems the clip+ OF also checks the version
09:33:23 Quit bor_ka (Quit: CGI:IRC 0.5.9 (2006/06/06))
09:35:26funmanbut apparently all clip+ OF which were released are supported anyway
09:37:08TarvisI'm making a 220x176 wps into a 320x340 one
09:37:23 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
09:37:24TarvisShould all the folders still be named the same or should I append _320 or something?
09:41:11 Join Rob2222 [0] (
09:41:37 Join swilde [0] (
09:42:20swildesorry I fell on my keyboard
09:43:11 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
09:43:36 Join n1s [0] (~n1s@rockbox/developer/n1s)
09:43:56 Quit Tarvis (Quit: /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit)
09:44:12 Quit Rob2223 (Ping timeout: 240 seconds)
09:44:27wodzhmm something is broken with power thread on mpio with current svn - it doesn't update battery voltage history but from the other hand battery voltage is read correctly
09:46:17 Quit Kitar|st (Ping timeout: 264 seconds)
09:49:02 Quit wodz (Ping timeout: 245 seconds)
09:51:24 Join Kitar|st [0] (
10:00:37***Saving seen data "./dancer.seen"
10:07:12 Join DerPapst [0] (
10:24:52 Quit TheSeven (Ping timeout: 240 seconds)
10:30:39 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
10:48:54 Join mt2 [0] (~chatzilla@
10:49:24CIA-6New commit by nls (r27446): libwma has Huffman tables that are too big for the stack temp buffer. Make temp buffer static.
10:51:11CIA-6r27446 build result: 6 errors, 0 warnings (nls committed)
10:53:43n1swoha, the codec buffer is tight on those clips :(
10:54:56funmann1s: 288k
10:55:23n1saka ~5kB too small :/
10:57:11funmann1s: how big is the stack ?
10:57:35n1s0x2400 Bytes
11:05:52n1si'll poke around a little, and see if i can find something to shrink, else i suppose i'll have to use an older version of the ffmpeg code that doesn't need the temp buffer
11:08:50n1shmm, or we could store the tables sorted so we don't need the temp buffer...
11:11:12 Join einhirn [0] (
11:38:25 Join dfkt [0] (dfkt@unaffiliated/dfkt)
11:42:59 Join halmi [0] (
11:52:43 Quit TheSeven (Ping timeout: 258 seconds)
11:57:13 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
12:00:41***Saving seen data "./dancer.seen"
12:05:51CIA-6New commit by bertrik (r27447): Fix failing AMSv2 power-down when the wake-up alarm timer is set - FS #11453 by me
12:07:39CIA-6r27447 build result: 6 errors, 0 warnings (bertrik committed)
12:10:52funmanbertrik: "fix" ?
12:11:36bertrikIt goes from completely reproducible to completely non-reproducible for me
12:12:02funmanbut my tests proved that it wasn't fixed, didn't they?
12:12:17bertrikI don't know exactly what you tested
12:12:19bertrikor how
12:12:48funmanset up wake up alarm to sometime, press power button -> locked-up
12:14:49funmani should've mentioned it on the ticket though, but i'm pretty sure we talked about this on irc
12:15:04bertrikI think it's an improvement because it's fixed completely now for me, perhaps there are other problems still left.
12:15:59 Join watto [0] (~watto@
12:17:00 Quit Topy (Quit: Leaving)
12:17:11 Join Topy44 [0] (
12:17:32funmanbertrik: shouldn't alarm be a static struct?
12:18:07funmanbertrik: 1st try on fuzev2: lock up
12:18:22funmanplay something, pause, set up alarm to 00:00, power off
12:19:18bertrikhm, yeah should probably be static
12:19:40funmanalso you can group members by size
12:21:04 Quit anewuser (Ping timeout: 260 seconds)
12:23:01funmanbertrik: release_alarm can be removed i guess
12:23:43funmanhm no weirdly rtc_init() is run in a loop somewhere
12:23:55funmananyway you should just ignore it
12:24:27funmanjust assume alarm is disabled.
12:24:27bertrikyeah I wondered about that a bit too, it always seems to be called with argument true from apps/ code
12:24:43funman2 drivers call it with false from rtc_init()
12:25:28funmanand since (who knows why) rtc_init() is run again in alarm_menu.c i prefered not touching this
12:26:14funmanpregier: ping ^
12:28:08funmanroolku wrote the driver for modded H1x0
12:30:01funmanbertrik: i don't want to sound offensive. thanks for working on this
12:32:12bertrikfunman, no problem. I left the task open for further comments. I think the change I just committed improves the situation at least and removes at least one cause for the hang.
12:32:37n1shmm, i don't think i have time to fix the failing builds properly now, should i revert or leav it as it is untill i do (probably tomorrow)?
12:33:02 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
12:33:04funmann1s: i'd say leave it like it is
12:33:22funmanwe don't need dozens of 'revert, i'll fix later' in the log
12:33:44funmannor do we need to hide problems just to make the table green
12:34:04funmanone day with trunk not building won't stop development on clipv1
12:35:07 Quit dfkt (Quit: -= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.)
12:35:34 Join dfkt [0] (dfkt@unaffiliated/dfkt)
12:36:02 Quit [Saint] ()
12:41:45CIA-6New commit by bertrik (r27448): rtc_as3514.c: make alarm struct static and reorder fields to save some binsize
12:42:42n1shmm, there seems to be some runtime inits in wma, i wonder if they are entirely independent of input...
12:43:20CIA-6r27448 build result: 6 errors, 0 warnings (bertrik committed)
12:53:53n1si think i found a way to save 512B, only a couple thousand more to go!
12:56:14funmancan we use unaligned load/stores on armv4t ?
12:57:57n1sor maybe, IIRC thay return the data rotated based on the alignment so you'd have to rotate it back or something like that
13:00:01n1sfunman: were you thinking of the bitstream readers or something else?
13:03:23funmanyeah, for packing the VLCcode struct
13:04:21n1sinit_vlc isn't very time critical so making that a little slower will probably not hurt too much
13:07:51 Quit funman (Quit: free(random());)
13:11:22 Join S_a_i_n_t [0] (S_a_i_n_t@
13:34:15 Join MethoS- [0] (~clemens@
13:51:43 Quit halmi (Quit: halmi)
13:56:30 Quit mitk (Quit: Leaving)
14:00:44***Saving seen data "./dancer.seen"
14:32:12mc2739is anyone able to build win32 cross-compiled sims with current svn?
14:43:54 Quit GeekShadow (Ping timeout: 240 seconds)
14:44:25*pixelma tries
14:46:13 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
14:50:40pixelmamc2739: for me it fails with an error in filetypes.c (Clip sim)
14:51:14pixelmaI made clean and reconfigured before
14:57:46 Quit S_a_i_n_t (Ping timeout: 246 seconds)
15:01:54 Join komputes [0] (~komputes@ubuntu/member/komputes)
15:04:53 Quit antil33t (Read error: Connection reset by peer)
15:05:00 Join antil33t [0] (
15:07:51pixelmamc2739: an M5 sim seems to work though
15:07:59pixelma(or compile)
15:08:10 Join S_a_i_n_t [0] (S_a_i_n_t@
15:18:06pixelmaoh, no it doesn't... I'm blind. It fails in libwmapro though
15:22:53 Quit mc2739 (Ping timeout: 240 seconds)
15:24:46 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
15:25:10mc2739pixelma: thanks mine fails with in libwmapro also
15:27:31 Join hebz0rl [0] (
15:29:45 Quit mc2739 (Ping timeout: 260 seconds)
15:29:53 Join kugel [0] (~kugel@rockbox/developer/kugel)
15:30:17 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
15:47:46 Join jgarvey [0] (
15:48:35 Quit Farthen (Remote host closed the connection)
15:48:43 Join Farthen [0] (
15:49:43 Join fdinel [0] (
15:53:29 Join panni_ [0] (
15:55:17 Quit mc2739 (Ping timeout: 260 seconds)
15:57:12 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
16:00:48***Saving seen data "./dancer.seen"
16:02:56 Quit storm` ()
16:04:39 Join wodz [0] (
16:09:01 Join Topy [0] (~topy@2a01:608:ffff:2627:2e0:4dff:fe83:a45e)
16:13:02saratogan1s: WMA has an enormous random number table for noise generation
16:13:24saratogaif you made it 500 entries smaller I'm sure no one would notice .. .
16:13:38 Quit Topy (Ping timeout: 252 seconds)
16:14:36 Quit Farthen (Remote host closed the connection)
16:14:38 Join Farthen [0] (
16:16:36 Join Topy [0] (~topy@2a01:608:ffff:2627:2e0:4dff:fe83:a45e)
16:19:51wodzdo the battery_bench work for You with current svn (and battery_view also)?
16:26:36 Quit Farthen (Remote host closed the connection)
16:26:39 Join Farthen [0] (
16:30:36 Quit Farthen (Remote host closed the connection)
16:30:37 Join Farthen [0] (
16:45:10 Nick S_a_i_n_t is now known as [Saint] (S_a_i_n_t@
16:49:26 Quit [Saint] (Ping timeout: 246 seconds)
16:50:46 Join hmmmmmm [0] (
16:56:00 Quit petur (Quit: later)
16:56:13 Quit hmmmmmm (Quit: CGI:IRC (Ping timeout))
17:04:33 Quit Topy44 (Quit: Leaving)
17:10:31 Quit dfkt (Quit: -= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.)
17:10:34 Quit wodz (Ping timeout: 240 seconds)
17:13:46gevaertsmc2739: which mr500 buttons would you propose for soft hold? :)
17:15:45pixelmaisn't there only one physical - which is probably used for other things already but could be "freed" for the WPS I guess)
17:16:00gevaertsyes, the power button
17:16:45pixelmaalso, editing the keymap file alone doesn't help, you need to enable the functionality somewhere (I guess in the config somewhere)
17:18:01gevaertsyes, that too
17:19:39gevaertsin button-target.h
17:20:08gevaertshm, wait
17:21:29gevaertsHAS_BUTTON_HOLD has to be undefined
17:22:55 Quit cjcopi (Ping timeout: 264 seconds)
17:26:46n1ssaratoga: ah, i'll look into that, i have made another size optimization by statically initializing the lsp_pow_m_table[12] tables but the test files i have don't use that so i need to find such a file for testing
17:32:13n1sor I think i can see an even better approach actually
17:34:12n1sthe buffers for these tables are only used if vlc3buf is not used and vice versa
17:34:40n1sso putting them in there will save some ram and not make for a very large hack :)
17:35:22 Join cjcopi [0] (
17:37:14 Quit Farthen (Remote host closed the connection)
17:38:21 Join Farthen [0] (
17:42:33 Quit Farthen (Remote host closed the connection)
17:43:43 Join Farthen [0] (
17:54:23 Join stoffel [0] (
18:00:50***Saving seen data "./dancer.seen"
18:01:25 Quit bieber (Ping timeout: 260 seconds)
18:01:42 Join bieber [0] (
18:06:04 Join dfkt [0] (dfkt@unaffiliated/dfkt)
18:09:25 Join freddyb [0] (
18:11:05 Join milz [0] (
18:12:14 Quit TheSeven (Ping timeout: 258 seconds)
18:12:33 Quit Kitar|st ()
18:13:50 Quit DerPapst (Quit: Leaving.)
18:15:57freddybCould I get someone to fix the category for FS #11483, thanks.
18:16:42kugelgevaerts: would you consider anything important missing for sdl-RaaA besides the paths?
18:17:05*gevaerts thinks for a minute
18:17:59gevaertsI suspect not
18:20:13bertrikfreddyb, what do you want the category to be?
18:20:32 Join Kitar|st [0] (
18:21:47bertrikok, done
18:22:41 Join bunnyboi [0] (
18:24:49 Quit swilde (Quit: ERC Version 5.3 (IRC client for Emacs))
18:26:30kugelgevaerts: that'd suits my schedule :)
18:26:49gevaertskugel: now get busy on the Android bits! :)
18:26:59kugelI can't think of anything important too, I also have nothing on the wiki page
18:28:39kugelgevaerts: I have had a quick look, things are going to get tricky. especially building rockbox as a library in a way android accepts it
18:29:28gevaertsThat's why you have another month for it :)
18:31:19kugelI set up eclipse for the java part, and it works if I use the android build tools for the ndk sample apps, but if try to use my tools (or even mimicing the ndk's build process) wasn't successful so far
18:31:55gevaertssounds like fun
18:32:12 Quit juju2143 (Ping timeout: 276 seconds)
18:32:27kugelthere's some magic involved that the makes emulator launched from eclipse find the .so file
18:33:18kugelalso, I fear it might require non-target-tree changes to make playing music in the background work
18:33:43gevaertswhere? The threading code?
18:33:56kugelplayback engine
18:34:20kugelthreading maybe too now you mention it :\
18:34:39 Part pregier
18:35:30kugelyou need to establish a "service" for all background stuff, rockbox in a single library would need to run under an "activity" (and there can only be one activity active at a time)
18:36:24gevaertsso you basically have to decouple playback and UI?
18:37:11kugelI think so
18:37:32kugelprobably not a bad thing, just a lot of work
18:42:45kugelI think I'll start with a sample app for sound playback
18:44:52gevaertssounds good.
18:45:46gevaertsAfter that, I guess getting everything to run (as one foreground app) would be a good step.
18:47:58 Join Jerom [0] (~heidi@
18:49:59 Join merbanan [0] (
18:50:45 Join juju2143 [0] (juju2143@gateway/shell/
18:54:05saratogan1s: LSP is only used for 16kbps and below WMA files
18:54:11saratoganoise coding is 32k and below
18:54:57n1ssaratoga: i found a static buffer that could be reused for the lsp tables, saving 1k
18:56:01n1sbut i'd like to test and see if it works correctly so do you know offhand where i can find such a file?
18:58:43n1sam i right thinking that the mdct window tables depend on the input file (frame_len_bits) and cannot be pregenerated?
18:59:29n1sah, i see that can have 3 different values so pregenerated tables would take more ram
18:59:45 Quit MethoS- (Read error: Connection reset by peer)
19:01:12kugelgevaerts: I too think that trying to make the background thing work before having a solid basis is a recipe for failure
19:01:20saratogan1s: theres 5 different mdct window sizes, but our new mdct only stores the table once and reuses it for different sizes
19:02:03n1ssaratoga: yeah, was a bit unclear, i meant the sine tables in wma /* init MDCT windows : simple sinus window */
19:02:41saratogayeah theres 5 of those
19:03:25n1sbut could they be pregenerated? the generating function depends on the frame length...
19:05:24 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
19:06:52n1sit seems there are 3*5 possible sine windows so pregenerating them would take more memory
19:09:15 Join fudge [0] (ad442595@gateway/web/freenode/ip.
19:09:21saratogaoh those windows
19:09:54saratogatheres 5 possible windows, but low bitrate files don't use all of them
19:10:05saratoga64kbps and above typically use all
19:12:27saratogai bet the windows could be shared with the IMDCT trig tables
19:13:31 Join GeekShad0w [0] (
19:13:41 Join anewuser [0] (anewuser@unaffiliated/anewuser)
19:14:03 Quit GeekShadow (Ping timeout: 252 seconds)
19:14:11n1ssaratoga: i'll take a look tomorrow, if that fails i'll try reducing the random table
19:15:54n1sbtw, from a quick look it seems that table is using about 17 bits for values, do you think it would be better to shift them down by 1 and store as 16 bit (and compensate by shifting up by one in the noisecoding)
19:18:33saratogan1s: what would that help with?
19:19:34 Quit antil33t (Read error: Connection reset by peer)
19:19:41 Join antil33t [0] (
19:24:58 Join stripwax [0] (
19:26:45 Quit pamaury (Remote host closed the connection)
19:30:41 Quit fudge (Quit: Page closed)
19:44:18 Join dlenski [0] (
19:44:35 Part dlenski
19:46:34 Quit juju2143 (Ping timeout: 276 seconds)
19:48:28 Join joeyg [0] (
19:49:46 Join mt [0] (~mtee@rockbox/developer/mt)
19:52:01 Join Horscht [0] (~Horscht2@xbmc/user/horscht)
19:53:58 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
19:58:33 Join Muchix [0] (
20:00:15 Join juju2143 [0] (juju2143@gateway/shell/
20:00:54***Saving seen data "./dancer.seen"
20:16:50 Quit dfkt (Read error: Connection reset by peer)
20:19:54 Quit juju2143 (Remote host closed the connection)
20:20:41 Part watto
20:27:13*kugel wonders if he could use opengl for the lcd driver
20:33:33 Quit Horscht (Quit: Verlassend)
20:36:46 Join Horscht [0] (~Horscht2@xbmc/user/horscht)
20:41:33 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
20:42:11pixelmamt: you could try to find out why crosscompiling sims currently fails in libwmapro ;)
20:42:29 Join juju2143 [0] (juju2143@gateway/shell/
20:42:29 Quit juju2143 (Changing host)
20:42:30 Join juju2143 [0] (juju2143@unaffiliated/juju2143)
20:42:30 Quit juju2143 (Changing host)
20:42:30 Join juju2143 [0] (juju2143@gateway/shell/
20:42:39pixelmamy guess it'll also fail under cygwin then
20:42:56*mt never knew about this
20:43:25pixelmaI'll get you the error message but I need a few minutes
20:43:29mtpixelma : Was it discussed here so I could check the logs ?
20:43:44 Join esperegu [0] (~quassel@
20:43:47mtAlright thanks !
20:43:53pixelmanot discussed, just found out today
20:46:23 Join binaryhermit [0] (
20:46:35 Quit kugel (Ping timeout: 240 seconds)
20:49:56pixelmaurgh... there is a codepage issue
20:52:19 Quit stoffel (Remote host closed the connection)
20:52:31mtHow much time does "Enabling win32 cross compiling" normally take ? it's been stuck at this for like 5 minutes.
20:53:19 Nick evilnick_B is now known as evilnick_ (0c140464@rockbox/staff/evilnick)
20:53:25*mt presses a key - problem solved !
20:54:11mtpixelma: thanks
20:54:19pixelmaah, you meant the advanced build options... yes they need an additional keypress to allow for multiple inputs
20:54:51pixelmaand to prove you are an advanced user who know what he does ;)
20:58:18 Nick evilnick_ is now known as evilnick_B (0c140464@rockbox/staff/evilnick)
21:01:33mtpixelma: Can't reproduce here
21:01:42*mt tries again to make sure
21:02:33 Quit freddyb (Quit: ChatZilla 0.9.86 [Firefox 3.5.9/20100401213457])
21:02:44pixelmamc2739 asked first about it
21:05:36 Join halmi [0] (
21:11:18mtYeah I made a mistake and chose a normal cross compile instead of a sim one .. anyway it's now complaining about not finding sdl-config though it's installed and in PATH
21:11:35 Quit Muchix (Ping timeout: 265 seconds)
21:13:56 Quit M3DLG (Ping timeout: 260 seconds)
21:14:23 Quit stripwax (Quit:
21:16:35 Join Tarvis [0] (
21:16:47pixelmamaybe your system still needs some packages but I can't help you with this as I didn't set it up myself. There is this guide but I'm not sure how up-to-date it is
21:18:30pixelmaI wonder though what else than a sim you could crosscompile - RaaA or what?
21:21:26 Quit evilnick_B (*.net *.split)
21:21:26 Quit saratoga (*.net *.split)
21:21:38pixelmaI mean crosscompiling other Windiws executables under Linux
21:23:55pixelmahmm... maybe checkwps or the theme editor (the former doesn't build I believe)
21:25:07 Join mf0102 [0] (
21:26:00 Join evilnick_B [0] (0c140464@rockbox/staff/evilnick)
21:38:51 Quit bunnyboi (Ping timeout: 258 seconds)
21:42:43 Join M3DLG [0] (
21:50:26 Quit halmi (Read error: Connection reset by peer)
22:00:20 Quit esperegu (Ping timeout: 245 seconds)
22:00:57***Saving seen data "./dancer.seen"
22:10:21 Join skfunnyboy_ [0] (
22:11:07 Part skfunnyboy_
22:12:16mtDoes our sys/types.h need to be called that ? win32 sim cross compile sees only the system's sys/types.h because of the (ifndef APP_TYPE .. INCLUDES += -I$(FIRMDIR)/libc/include) in firmware.make
22:12:39mtwhile file.h needs our sys/types.h in that case for the typedefs
22:12:56gevaertshm, try to catch kugel for that one...
22:14:46*mt now thinks he might be wrong in something and goes back to recheck
22:15:11 Join kugel [0] (
22:15:55mtof course forcing file.h to include "../libc/sys/types.h" solves the problem but it looks like hack rather than a fix to me
22:18:06pixelmathe text viewer screenshots in the manual have no mime-type at all. Since they are png, there are some cryptic characters in the diff which confuses some text editors. They are not even application/octet-stream
22:19:33 Join einhirn [0] (
22:19:39pixelma"confuses some text editors", e.g. gedit won't display the diff at all complaining about unknown codepage
22:21:59kugelmt: what is lacking from the systems types.h?
22:22:50 Join wodz [0] (
22:23:31pixelmaone reason to set mime-type. Am I allowed to do this for the text viewer screenshots?
22:25:30 Join dfkt [0] (dfkt@unaffiliated/dfkt)
22:25:57mtkugel: I've looked into this rather quickly, so I might be missing something .. but one example is ssize_t which isn't defined in the system's sys/types.h
22:26:44mtalthough I'm now thinking it maybe something specific to codecs since metadata compile fine.
22:26:51gevaertsssize_t is a bit special
22:29:04 Join esperegu [0] (~quassel@
22:29:53 Quit wodz (Ping timeout: 265 seconds)
22:30:59*pixelma takes that as a "yes" although can't remember the final outcome of the mime-type discussion
22:35:01kugelmt: unix conform systems should have ssize_t. I will have a look when at home
22:37:36pixelmareally no objection?
22:37:54gevaertspixelma: I see no reason for not setting it :)
22:38:22pixelmathanks :)
22:38:38 Quit esperegu (Read error: Connection reset by peer)
22:44:54CIA-6New commit by pixelma (r27449): Set mime-type correctly to image/png for all text viewer screenshots. Mime-type wasn't set at all before - leading to textual comparison of these ...
22:45:46 Quit Tarvis (Quit: /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit)
22:46:31CIA-6r27449 build result: 6 errors, 0 warnings (pixelma committed)
22:47:25CIA-6New commit by bieber (r27450): Theme Editor: Added target database, now populates combo box in new project dialog but otherwise not used yet
22:48:56CIA-6r27450 build result: 6 errors, 0 warnings (bieber committed)
22:52:23 Quit Topy (Ping timeout: 248 seconds)
22:56:56mtpixelma: This works for me :
22:57:15mtBut I don't know how to squeeze those conditions into one line in a makefile :(
22:58:04kugelmt: that's a hack unfortunately
22:58:35kugelif you can wait a bit I'll have a look
22:58:44mtyeah no rush
23:02:50 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
23:05:44 Join petur [0] (~petur@rockbox/developer/petur)
23:06:04 Quit dfkt (Quit: -= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.)
23:08:39 Quit evilnick_B (Quit: Page closed)
23:10:01 Quit bluebroth3r (Ping timeout: 260 seconds)
23:11:17 Join bluebrother [0] (
23:11:18 Quit bluebrother (Changing host)
23:11:18 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
23:12:33 Join einhirn [0] (
23:12:58 Quit joeyg (Quit: lions and tigers and bears, oh my!)
23:13:49 Quit anewuser (Quit: Sixth edition of 25 samples + 48 hours + detroit party (july 31) (maybe streamed online!))
23:15:11 Join Muhd [0] (
23:23:06 Join Topy44 [0] (~topy@2a01:608:ffff:2627:2e0:4dff:fe83:a45e)
23:27:35 Quit Topy44 (Ping timeout: 248 seconds)
23:29:29 Join Topy44 [0] (~topy@2a01:608:ffff:2627:2e0:4dff:fe83:a45e)
23:43:07 Join Tarvis [0] (
23:44:44TarvisTime for you guys to tell me which font you think looks best here:
23:45:24TarvisI still think Nimbus looks best
23:46:41mtNimbus looks nice yes
23:46:50mtand Sazanami too
23:52:51 Quit n1s (Ping timeout: 240 seconds)
23:57:29 Quit M3DLG (Quit: RAGE QUIT)

Previous day | Next day