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 2010-06-01

00:03:50 Quit storm` (Ping timeout: 276 seconds)
00:06:04 Join jordan` [0] (
00:07:03 Join storm` [0] (
00:07:53 Quit DataGhost (Read error: Connection reset by peer)
00:11:59funmanbertrik: try throwing it in the water first. perhaps it only works if you're desperate enough
00:12:37*r0b- loves his rockbox
00:15:14 Quit halmi (Read error: Connection reset by peer)
00:16:12bertrikI'll bring it to devcon
00:16:56 Quit lpereira (Quit: Leaving.)
00:19:55funmanbertrik: want me to break it for you? :)
00:21:17bertrikmaybe you have the "magic touch" to make it work
00:21:17 Quit storm` (Read error: Connection reset by peer)
00:28:20 Join storm` [0] (
00:31:13 Join anewuser [0] (anewuser@unaffiliated/anewuser)
00:35:10CIA-98New commit by lenzone10 (r26438): Updated italian translation.
00:38:59 Quit bertrik (Quit: De groeten)
00:42:49 Quit storm` (Read error: Connection reset by peer)
00:43:28 Quit r0b- (Read error: Connection reset by peer)
00:44:58 Join oldtopman [0] (
00:45:42oldtopmanI have a iPod video (30GB 5th Gen) and would like to install rockbox...
00:45:52oldtopman...What should I partition the iPod as (in ubuntu)
00:46:04oldtopman(I deleted all partitions earlier(failed rockbox attempt))
00:46:22 Quit pbxy (Quit: Maintenance window closed)
00:47:09oldtopmanIs anyone here?
00:47:17 Join Stephen__ [0] (~S@
00:48:53oldtopmanIs anyone here?
00:49:33gevaertsoldtopman: have a look at
00:50:05 Join storm` [0] (
00:54:11 Quit anewuser (Ping timeout: 258 seconds)
00:54:24oldtopmanthanks a million
00:54:49funmanI copied the full reserved block from another fuzev1 and now my fuzev1 can boot OF. Copying only the OF file wasn't enough : rockbox booted but OF only power on the backlight and stopped
00:55:33 Quit merbanan (Read error: Operation timed out)
00:55:46funmanIt just struck me that the reserved size is exactly (OF size * 2) on all Sansa AMS with 8MB of ram
00:56:07 Quit oldtopman (Quit: Leaving)
00:56:12saratogaany idea whats in the rest of the space?
00:56:24funmanon Sansa AMS with 2MB of ram, the reserved size is exactly (OF size - header) * 2
00:56:40funmanno, it looks like random garbage at first sight. it's not a copy of the OF, strings doesn't show anything
01:00:15funmanthe extra part (beyond OF) compresses fairly well: bzip2 gives 3.3MB instead of 15MB, and compresses the whole thing to 13MB
01:00:23funmanlots of 00000000
01:00:55 Quit Stephen__ (Quit: Leaving)
01:01:36 Quit vaguerant (Ping timeout: 260 seconds)
01:02:35funmane200v2 recovery shows the same thing in dmesg: only the first SD bank is shown
01:02:54 Join Speedy2 [0] (
01:03:05 Join S_a_i_n_t [0] (S_a_i_n_t@
01:03:09funman has a link to 'of_and_padding' so it looks like the OF stores something there
01:03:26funmanRob2222: ping
01:03:37Speedy2Hey all. I'm using Rockbox on the Sansa Clip v1. Is there an easy way to use bookmarks if I'm just playing a file browsed from a folder? I checked the manual and I'm still a bit confused.
01:03:54 Quit storm` (Read error: Connection reset by peer)
01:04:11Speedy2(I'm not using a playlist)
01:04:40funmanSpeedy2: rockbox creates a playlist when you open a file from a folder: it contains all the files in this folder
01:05:14funmanpress long center -> a menu opens -> playlist is at the top
01:05:56Speedy2Can I set rockbox to always store a bookmark, so if I turn it off in the middle of a song, it will auto-resume from there?
01:06:04funmanshould e200v2 unbricking be moved to its own page so i could edit it and add info for other models?
01:07:15Speedy2I actually use rbox to listen to podcasts when I work out,and when I'm done I just want to turn the Clip off, and next time I go in, just have it resume.
01:08:26mc2739Speedy2: there is information about bookmarks in the manual
01:08:33saratogadon't we have autoresume
01:08:34 Quit efyx (Quit: Quitte)
01:08:45Speedy2saratoga: Intentionally?
01:09:36Speedy2I mean, are you avoiding this feature, or its just not there yet?
01:09:55gevaertsSpeedy2: please read the manual. It contains information about this and much morew
01:10:17Speedy2I'll re-read it
01:10:31funmanSpeedy2: i just linked you to the part about auto resume
01:11:05Speedy2funman: Thanks
01:11:08Speedy2Guys, I apprecaite it
01:11:51 Join storm` [0] (
01:12:06saratogayeah i was pretty sure we had that, never got around to trying it though :)
01:12:39funmansaratoga: it is essential with rtc wake up ;)
01:16:37funmani was not correct, the size of the recovery drive is not the size of 1 bank, it's the size reported by CSD register before we enable bank switching and ask for the CSD a second time
01:16:49 Join robin0800 [0] (
01:18:30JdGordonbieber: hey, any idea when you might get to making viewports the top level thingy?
01:18:55funmanJdGordon: you still have your clip+ ?
01:19:27funmanwant to try unbricking?
01:21:57funmanyou have it opened, access to the buttons side of the PCB ?
01:22:38 Quit Kitr88 ()
01:23:20funmanyou need to make sure it's off (unsolder the battery if you can solder it back), connect 2 pins on the PCB, and power USB
01:23:35funmanyou'll see in dmesg a 979.75MB device from "UNDEF"
01:23:45funman(well, i hope you will)à
01:24:34JdGordonwhich two pins?
01:24:35funman <- in red I circled the probable pins to short, in blue the second candidates
01:24:40JdGordonI tihnk we tried this already
01:24:51JdGordonnot those ones though
01:24:54S_a_i_n_tIs there a chance ucchan's patch for will be included to SVN?
01:25:01S_a_i_n_tIt would seem kinda silly not to.
01:25:14funmanJdGordon: did you see today i unbricked a fuzev2, fuzev1 and clipv1?
01:25:24 Quit storm` (Read error: Connection reset by peer)
01:25:24funmanS_a_i_n_t: i was hoping he would commit it
01:25:32JdGordondidnt see how.. very awesome :)
01:25:42JdGordonhow long do they need to be connected for?
01:25:47JdGordonjust while insterting usb?
01:26:15S_a_i_n_tfunman: Did you discover some random secret trick or something that works for all/most of the sansas?
01:26:26S_a_i_n_t3 targets unbricked in one day is impressive.
01:26:38funmanJdGordon: yep, better keep them 10s or so to be sure
01:27:09JdGordonany ideas how to connect them? its pretty small :)
01:27:21funmanS_a_i_n_t: the e200v2 trick seems to work on more than e200v2
01:27:32funmanJdGordon: screwdriver, knife, wire (more difficult)
01:27:32JdGordonI have to run, lemme know what to do once I get the dmesg output... ill try it later
01:27:47saratogado you really need to pull the battery?
01:27:48funmanJdGordon: dd if=/path/to/OF.bin (patched or not) of=/dev/sdX
01:28:09funmansaratoga: no, but you must make sure the dap is off when you plug it
01:28:41funmanelse boot code won't run, the CPU will just keep running the buggy code which caused the brick in the first place
01:28:43saratogamore off then holding power ?
01:28:57funmanno just off, but without feedback you never know if it's off or not
01:29:03saratogagood point
01:29:15JdGordonI'm fairly sure the batt will be dead now anyway, but the battery is easy enough to resolder
01:29:29funmanelse i would just press power quickly to power it on if it was not already
01:29:42JdGordonfunman: I'll let you know how it goes, wont be able to try for a few hours
01:29:45funmanand then press longer to make sure it's off (but you never know if it's longer enough, sometimes it doesn't work)
01:30:29JdGordondoh! blue wire just came off the battery solder point so that makes at least one thing easier :p
01:30:41JdGordonany danger with shorting the other point?
01:30:50funmanblue is thermal check i think
01:31:07funmanJdGordon: the whole thing might explode and blow your head off, i'm not responsible ;)
01:31:16JdGordonok cool!
01:32:10 Join Kitar|st [0] (
01:32:47 Join S_a_i_n_t_ [0] (S_a_i_n_t@
01:32:53 Quit S_a_i_n_t (Ping timeout: 240 seconds)
01:33:15 Join storm` [0] (
01:35:02funmanSansaAMS -> SansaAMSUnbrick looks good for a new wiki page?
01:36:22 Nick S_a_i_n_t_ is now known as S_a_i_n_t (S_a_i_n_t@
01:36:23 Join r0b- [0] (
01:39:17 Join anewuser [0] (anewuser@unaffiliated/anewuser)
01:40:05S_a_i_n_tfunman: Very glad you managed to unbrick your DAPs \o/
01:40:41S_a_i_n_tIt must be a big releif.
01:49:09 Quit n1s (Quit: Lämnar)
01:49:56 Quit storm` (Ping timeout: 265 seconds)
01:50:47 Quit MethoS- (Remote host closed the connection)
01:53:16 Quit Strife89 (Quit: Thunderstorm, taking precautions.)
01:53:33 Join storm` [0] (
01:57:27 Quit robin0800 (Quit: - Chat comfortably. Anywhere.)
01:57:47 Join robin0800 [0] (
01:57:49***Saving seen data "./dancer.seen"
02:02:21funmanS_a_i_n_t: yeah, although i can't do anything for the 4 i had previously broken :/
02:02:43funmani just hope we can make this work on every sansa AMS
02:04:00 Join Stephen__ [0] (~S@
02:12:00 Quit storm` (Ping timeout: 258 seconds)
02:13:41 Quit DerPapst (Quit: Leaving.)
02:13:44S_a_i_n_tfunman: Only 4? be honest I would have suspected that you'd bricked far more DAPs than that by now :P
02:14:19 Quit robin0800 (Remote host closed the connection)
02:15:49 Join storm` [0] (
02:19:08 Join Transformer [0] (
02:20:56 Quit Transformer (Excess Flood)
02:22:51funmanbtw, i now have an early bootloader and i think it is much faster than the v1 bootloader
02:23:00funman(on fuzev1)
02:23:04funmancan't barely read the version number
02:23:06bieberJdGordon: Probably within four or five hours
02:23:38funmanr20603 i think
02:24:01CIA-98New commit by wincent (r26439): pdbox: Solved issue with WAV file import on big-endian targets (signed 16-bit files only); small I/O-related workaround in g_array.c
02:26:30CIA-98New commit by wincent (r26440): pdbox: Removed unnecessary code from the main include file. Reduced neede memory to 2 MB.
02:28:45funmanbertrik: the 6th pin from bottom left is 'WP#' (NOT Write Protect), according to ONFI TSOP-48 and WSOP-48 Pin assignements
02:31:52 Quit Stephen__ (Quit: Leaving)
02:32:32 Quit storm` (Ping timeout: 265 seconds)
02:34:44 Join JdGord [0] (~jd@
02:35:21 Quit JdGord (Client Quit)
02:35:41wincentQuestion to the people knowledgeable about the PortalPlayer-based targets: Would you put the most CPU-hungry thread on the main CPU or on the coprocessor?
02:37:01 Quit bieber (Ping timeout: 264 seconds)
02:37:09 Join storm` [0] (
02:37:17 Join bieber [0] (
02:42:03funman <- does it look fine?
02:42:29funmanwincent: I don't know well PP, but I would say the COP, since the CPU is used by the kernel
02:47:28 Join Strife89 [0] (
02:49:57wincentfunman: Thanks! Are there any targets with different clock frequencies for CPU and COP?
02:50:57funmanno idea
02:51:37wincentfunman: Thanks nevertheless!
02:51:38funmanset_cpu_frequency in system-pp50*.c doesn't seem to have separate settings for CPU & COP
02:58:09 Quit bieber (Ping timeout: 265 seconds)
02:58:22 Join bieber [0] (
03:02:18 Quit funman (Quit: free(random());)
03:05:10saratogawincent: on PP the COP and CPU are always the same speed
03:05:56saratogaCOP is usually idle, except for when playing MP3 where it'll have about 20MHz used by the decoder (and less commonly when playing mpeg or spc files)
03:11:59 Quit storm` (Read error: Connection reset by peer)
03:18:00 Join storm` [0] (
03:18:12 Quit steve|m (Ping timeout: 276 seconds)
03:21:25 Quit bieber (Ping timeout: 240 seconds)
03:21:38 Join bieber [0] (
03:28:05 Nick fxb is now known as fxb__ (
03:28:06wincentsaratoga: I placed main thread of pdbox on the coprocessor. Does this sound feasible?
03:29:36 Nick kisak_ is now known as kisak (
03:29:39saratogawincent: depends on how you do it
03:30:04saratogaif you can split up your work across both cores it makes sense, just putting everything on the second core probably won't make much difference though
03:30:21saratogaunless you plugin can run while audio is playing?
03:31:41 Join steve|m [0] (
03:33:06wincentsaratoga: Actually, it creates audio, hence the answer is no.
03:33:28saratogathen it won't make much difference, unless you can split the load across both CPUs
03:33:36 Quit storm` (Read error: Connection reset by peer)
03:33:37wincentI did.
03:33:51saratogawell then in theory you can double your speed :)
03:33:57wincentPD core runs on the COP and the GUI runs on the CPU.
03:34:10saratogadoes the GUI use any CPU time?
03:34:43wincentIt might be updated relatively often.
03:34:43saratogaif you're hurting for CPU cycles you probably want to split up the audio processing as well
03:35:36wincentWell, I think the processing power (at least on the H300) is rather sufficient.
03:36:18wincentWith UI simulator redesigned, how is it possible now to dump audio output into a file?
03:39:18saratogait doesn't work the way it did before?
03:39:43 Join storm` [0] (
03:39:51wincentI did not see the −−debugaudio options when running sim with -h switch
03:49:23 Join CGL [0] (~CGL@
03:51:02 Join robin0800 [0] (~quassel@
03:53:57 Join fdinel [0] (
03:54:04 Quit bieber (Ping timeout: 264 seconds)
03:54:21 Join bieber [0] (
03:56:53 Join Darkknight512 [0] (
03:57:02 Nick CGL is now known as [CGL] (~CGL@
03:57:50***Saving seen data "./dancer.seen"
04:33:39 Quit TheSeven (Ping timeout: 260 seconds)
04:37:34 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:38:09 Quit pixelma (Disconnected by services)
04:38:11 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:38:12 Quit amiconn (Disconnected by services)
04:38:14 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:38:31 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:38:36 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:42:19 Quit bieber (Ping timeout: 272 seconds)
04:42:34 Join bieber [0] (
04:45:23 Join Barahir_ [0] (
04:46:43 Quit robin0800 (Remote host closed the connection)
04:46:49 Quit moparx (Ping timeout: 264 seconds)
04:48:17 Quit panni_ (Read error: Connection reset by peer)
04:48:22 Quit elinenbe (Ping timeout: 264 seconds)
04:48:48 Quit Barahir (Ping timeout: 260 seconds)
04:54:02 Quit soap (Read error: Operation timed out)
05:10:18 Join JdGord [0] (~jd@
05:15:03 Quit JdGord (Client Quit)
05:26:50 Join togetic [0] (~togetic@unaffiliated/ibuffy)
05:27:37 Quit Darkknight512 (Quit: ChatZilla 0.9.86 [Firefox 3.6.3/20100401080539])
05:28:52 Quit Horscht (Quit: Verlassend)
05:30:31 Join JdGord [0] (~jd@
05:46:51 Quit Topy44 (Ping timeout: 248 seconds)
05:48:12 Join Topy44 [0] (
05:50:55CIA-98New commit by funman (r26441): sd-as3525v1: set up bank selection data outside of the loop
05:52:10 Quit JdGord (Read error: Connection reset by peer)
05:52:17 Join JdGord [0] (~jd@
05:57:51***Saving seen data "./dancer.seen"
05:58:41 Quit JdGord (Quit: Bye)
06:16:17 Quit anewuser (Quit: for SELL 2 by the price of 1 now!)
06:18:47 Quit simonrvn (Quit: .)
06:25:40 Quit CIA-98 (Ping timeout: 276 seconds)
06:38:48 Join CIA-6 [0] (cia@
06:44:09 Quit fdinel (Quit: Miranda IM! Smaller, Faster, Easier.
06:54:51 Quit Strife89 (Quit: Bed.)
07:04:53 Join JdGord [0] (~jd@
07:08:15 Quit S_a_i_n_t (Ping timeout: 276 seconds)
07:09:21 Join S_a_i_n_t [0] (S_a_i_n_t@
07:11:30 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
07:11:33 Quit bmbl (Read error: Connection reset by peer)
07:11:44 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
07:23:17 Quit S_a_i_n_t (Ping timeout: 240 seconds)
07:24:41 Quit JdGord (Ping timeout: 260 seconds)
07:26:07 Join JdGord [0] (~jd@
07:26:40 Join S_a_i_n_t [0] (S_a_i_n_t@
07:44:47 Join esperegu [0] (~quassel@
07:49:21bieberJdGordon: I'm starting in on the Viewport classifications, but I've got sort of a problem
07:49:53bieberWhat are we to do about %V tags that come in the middle of a line? Does that just automatically start a new line?
07:50:27bieberOkay, so a line being parsed should break if we encounter a %V, %Vi, or %Vl
07:51:05JdGordonnothing exept maybe whitespace should be on a line with those.. but that doesnt need to be enforced
07:57:53***Saving seen data "./dancer.seen"
08:00:45 Join LinusN [0] (linus@
08:00:45 Quit LinusN (Changing host)
08:00:45 Join LinusN [0] (linus@rockbox/developer/LinusN)
08:01:07 Join sbhsu [0] (
08:24:21 Join merbanan [0] (
08:28:53 Quit merbanan (Ping timeout: 252 seconds)
08:31:36 Join Zagor [0] (bjst@rockbox/developer/Zagor)
08:36:41 Join ender` [0] (
08:48:41 Quit JdGordon (Quit: Leaving.)
08:51:15 Quit liar (Ping timeout: 240 seconds)
08:55:24 Join flydutch [0] (
08:55:41 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
09:02:04 Join Rob2223 [0] (
09:05:11 Quit Rob2222 (Ping timeout: 252 seconds)
09:05:11 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
09:09:17bieberJdGordon: I think I've got it all put together
09:10:10 Join petur [0] (~petur@rockbox/developer/petur)
09:11:25CIA-6New commit by bieber (r26442): Theme Editor: Made Viewport the top level parse tree element, along with a bugfix to the tag parsing function
09:11:51bieberI haven't tested it as thoroughly as I'd like, but it seems to be working well enough
09:12:29JdGordonno worries.. I want to start bringing that parser into the core so if it has bugs we'll find em soon enough
09:13:28 Join GeekShado_ [0] (
09:13:34bieberSounds good
09:14:02JdGordonMy plan is to do the parseing in two steps.. the first will use your code directly and load a skin into a small static buffer, then I'll do a second parse of the output tree which will then create the real token structs as needed
09:15:29JdGordonhmm, actually I was thinking that the tree would be throw-awayable.. but its not is it?
09:15:47bieberIf you copy off all the data you need
09:16:06 Join JdGordon1 [0] (
09:16:29 Quit GeekShadow (Read error: Connection reset by peer)
09:16:33 Quit JdGordon (Quit: Leaving.)
09:16:47 Nick JdGordon1 is now known as JdGordon (
09:17:07 Join lpereira [0] (
09:19:44 Quit JdGord (Ping timeout: 272 seconds)
09:19:51 Join einhirn [0] (
09:21:02 Join B4gder [0] (~daniel@rockbox/developer/bagder)
09:21:10JdGordonwell the tree looks correct anyway :)
09:21:10JdGordonyou're done with the parser now right?
09:21:44 Join DerPapst [0] (
09:25:31 Join DerPapst1 [0] (
09:26:08bieberYeah, I'm going to start working on GUI now
09:26:14bieberAside from any bugfixes I come across
09:27:36 Join dfkt [0] (dfkt@unaffiliated/dfkt)
09:27:40 Quit DerPapst (Ping timeout: 265 seconds)
09:35:11 Quit TheSeven (Ping timeout: 264 seconds)
09:39:23 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
09:40:04 Quit pamaury (Client Quit)
09:40:21 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
09:40:56 Join efyx [0] (
09:44:28 Join Rob2222 [0] (
09:45:08 Quit amiconn (Disconnected by services)
09:45:10 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
09:45:32 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
09:45:48 Quit pixelma (Read error: Connection reset by peer)
09:45:49 Join pixelma [0] (quassel@rockbox/staff/pixelma)
09:47:58 Quit Rob2223 (Ping timeout: 265 seconds)
09:53:55 Join TopyMobile [0] (
09:57:54***Saving seen data "./dancer.seen"
09:59:18JdGordonbieber: you allow comments inside param lists?! nice
09:59:39bieberYes, but they won't show up in the parse tree :(
09:59:44 Join Rob2223 [0] (
10:01:05JdGordonI'm sure that can be added in the app if needed.. keeping a seperate list of comments..?
10:01:16*JdGordon stumbled on a nauhty free and saw that code
10:01:52bieberThat might work. I'll have to try and figure something out :/
10:02:41bieberIn any case, at least they'll parse and render correctly
10:02:53 Quit Rob2222 (Ping timeout: 260 seconds)
10:07:09JdGordonError on line 3: Argument can not be set to default :(
10:07:15JdGordoni is defaultable int isnt it?
10:08:35bieberIt is, I probably messed something up in the tag parsing code
10:08:49biebercopy/paste me the code that errored out?
10:09:45 Join kugel [0] (~kugel@rockbox/developer/kugel)
10:10:54JdGordon%V(0,0,-,8,0,-,-)\n%wi\n%ax%Vi(0,8,-,212,1,-,-)\n is what I'm trying to parse...
10:11:16JdGordonskin_parser.c line 469+ looks out of place no?
10:12:27bieberThat is invalid
10:12:33bieberVi's first argument is a nullable string
10:12:47JdGordonoh right, sorry :)
10:12:48bieberAnd its third argument is a non-nullable int
10:13:02 Join funman [0] (~fun@rockbox/developer/funman)
10:13:20JdGordonforgot that tag changed
10:13:31bieberThat's just the code that separates out the argument types
10:16:08funman explains how spurious interrupts happen with VIC PL190
10:16:36funmani think for PCM the solution is to use what jhMikeS added to gigabeats in r26340
10:16:38JdGordonfunman: hey, I didnt have any luck with the clip+
10:16:52bieberJdGordon: Anything in particular that stood out?
10:16:58funmanJdGordon: sorry :/
10:17:14JdGordonbieber: ?
10:17:32funmani have tried what the OF does successfully but i don't understand how this work
10:17:32JdGordonfunman: I tried a few other points that looked close enough to be together, but yeah nothing.. unless im not doing it right?
10:17:34bieberYou mentioned that line in skin_parser.c looking out of place
10:18:03funmanJdGordon: well just make sure it's off, connect the pins, plug USB while keeping connected and watch dmesg output that's all
10:18:23JdGordonyeah, because it looked like you were searching for the - in the same ifelse blocks that the type is in... I see what its actually doing now though
10:19:35JdGordonfunman: once its connected i dont need to hold it anymore? I'll have another go when i get another pair of hands
10:20:03funmanjhMikeS: pcm_*_lock() just lock the callback from asking more data, not the interrupt itself (as pcm.h comment says)?
10:20:11 Join dfkt_ [0] (
10:20:11 Quit dfkt_ (Changing host)
10:20:11 Join dfkt_ [0] (dfkt@unaffiliated/dfkt)
10:20:19*S_a_i_n_t hopes this is a semi-universal product.
10:20:27funmani.e. we can finish transferring what we were doing but not ask for more until we're unlocked
10:20:47funmanJdGordon: i disconnected the pins only after seeing the device present in dmesg output
10:21:42JdGordonis it mostly instant?
10:24:13 Quit dfkt (Ping timeout: 265 seconds)
10:27:51 Join ryran [0] (
10:27:53 Join n1s [0] (~n1s@rockbox/developer/n1s)
10:29:00 Nick dfkt_ is now known as dfkt (dfkt@unaffiliated/dfkt)
10:31:36kugelfunman: did I read that correctly, you found the pins to short for the fuzev1 as well?
10:32:04kugelI still have a dead fuze somewhere which would be happy if unbricked :)
10:32:31funmanyep, i wanted to ask you this
10:32:44funmankugel: check SansaAMSUnbrick on the wiki, I've put pictures
10:33:25kugelnice, will check that asap
10:33:58kugelalthough I'm not sure if my other fuzev1 is dead due to hardware failure or not
10:38:44kugelhah, if that one can be repaired it deserves a place in some Hall Of Fame
10:44:08 Part LinusN
10:47:11JdGordonbieber: ping?
10:50:26 Quit Galois (Ping timeout: 264 seconds)
10:52:07 Join Galois [0] (
10:55:29 Join MethoS- [0] (~clemens@
10:59:30 Quit phanboy_iv (Ping timeout: 265 seconds)
11:00:21JdGordonbieber: for tomorow I guess.... I was hoping the VIEWPORT elemt would have the viewport params straight in it instead of needing to find the viewport tag... I'm going to fiddle and see if I can fix that, i'll post a patch if I get it working
11:00:41 Part ryran
11:01:14CIA-6New commit by funman (r26443): fft: use long home button to exit on fuze
11:02:36funmankey clicks don't work correctly on AMS
11:02:49funmanscrolling the wheel like furious seriously reduce playback rate
11:11:09funmansomeone with a sansa AMS can make a recording with ?
11:11:45funmani was recording on fuzev1, backlight went off and i couldn't have back - want to make sure it's not related to the patch
11:13:53funmananother freeze after 6 minutes
11:17:00funmananyone has been using on fuzev1 lately?
11:19:19 Join stripwax [0] (
11:19:29 Quit stripwax (Client Quit)
11:20:12 Join stripwax [0] (
11:20:40funmanthis could have been a SD freeze
11:22:05 Join soap__ [0] (
11:22:47 Nick soap__ is now known as soap (
11:23:02 Quit soap (Changing host)
11:23:02 Join soap [0] (~soap@rockbox/staff/soap)
11:24:53 Join voRia [0] (~voria@
11:28:45 Join ryran [0] (
11:31:31funmanyup, svn crash too with led icon
11:32:13ryran[quick question RE onplay.c:] after a modified hotkey function I've added is called, I want things to just stay where they are (in this case, in the filetree). I don't want ONPLAY_START_PLAY, and ONPLAY_OK doesn't do it either. ONPLAY_RELOAD_DIR should have the intended effect, but I'm wondering if there's a simpler way?
11:38:45ryranokay just tested my hotkey function with ONPLAY_RELOAD_DIR and after using hotkey, it still returns me to WPS. BOOO. I'm missing something else.
11:39:12CIA-6New commit by funman (r26444): as3525 PCM: implement locking like done for the gigabeats in r26341 ...
11:40:37funmanpamaury: ^ you should check this if this causes a problem with usb-drv-as3525v2
11:44:22ryranGAH! After more looking around now I'm really curious why doesn't ONPLAY_RELOAD_DIR do what I want... anyone awake have experience with this? I'm trying to prevent.. e.g. hotkey_insert from returning you to the WPS after you insert some tracks, so that you can just continue inserting more.
11:44:58kugelryran: have a look at tree.c
11:45:33ryrankugel: hmmmmmm
11:47:52ryrankugel: well.. not much there. are you saying that I definitely need to change something there, versus onplay.c?
11:48:30kugeldoesn't tree.c handle starting the wps?
11:49:05ryranI wouldn't know. I'm really not at all familiar with the source.
11:49:06kugeltree.c combines the common parts of the filetree and the database browser (tagtree)
11:50:52ryranI just don't understand why.. from a user perspective: if you hold_select > playlist > insert shuffled; it adds to your current playlist without taking you to WPS. if you set tree-hotkey insert_shuffled and then do that from the filetree, it DOES take you to WPS.
11:51:03ryranthat's the behavior I'm trying to change..
11:51:10ryranbut I guess you got that
11:51:16kugellook at lines 680+ and 840+ in particular
11:51:20JdGordonkugel: it doesnt... it returns GO_TO_WPS
11:51:57*JdGordon wishes he understood bieber's parser :/
11:53:25ryranah well. guess I'm not gonna figure this one out. suppose knowing C would actually be handy here...
11:54:11 Quit MethoS- (Remote host closed the connection)
11:54:49ryranI still am curious to find out the reasoning behind the design decision of the behavior I described... I'm sure it's been covered somewhere.
11:55:24AlexPryran: I doubt there was one
11:55:37kugelryran: don't assume design decisions :)
11:55:45JdGordonryran: that is a bug... please create a task in flyspray
11:56:20JdGordonexecute_hotkey() in onplay.c is being naughty
11:56:24ryranWAIT holy crap ONPLAY_RELOAD_DIR *does* work.... I must've had my hotkey setting on the wrong function
11:57:55***Saving seen data "./dancer.seen"
11:58:57 Join swilde [0] (
12:03:05JdGordonwho knows Makefiles?
12:03:32funmanJdGordon: a bit
12:04:18JdGordonI need to compile a few files from utils/themeditor to a .o which the rockbox bin can link with...
12:04:33JdGordonI don't know where to start
12:05:56funmanJdGordon: grep for FIRMLIB, it's the libfirmware.a
12:06:53kugelJdGordon: just hack them into apps/SOURCES :)
12:07:08JdGordonthats what I'm doing.. but that is obviously not a long term solution
12:07:37kugelI guess the ling term solution would be to move the files out of utils/themeeditor...
12:07:56 Nick fxb__ is now known as fxb (
12:08:52JdGordonyes, utils/skin_parser or tool/skin_parser
12:08:56JdGordonnot apps
12:09:36kugelrockbox shouldn't link stuff from anything else besides apps and firmware
12:09:44JdGordonthose files are used by the theme editor and apps
12:09:59JdGordonits called a shared library for a reason...
12:10:15kugelshared library?
12:10:22JdGordonobviously it needs to be built seperatly for each build, but then it is no different to libfirmware.a
12:10:23kugelwhat "rockbox bin" are you talking about?
12:10:50*JdGordon starts from the begining
12:11:33funmani think r26444 should be backported as it fixes keyclick
12:11:52funmanbut it doesn't harm to let it live one more day in trunk
12:11:54JdGordonthe new parser currently in utils/themeditor is going to be used by the editor and rockbox... Ideally without any mention of .c's in the SOURCES.. just a .a dependancy
12:12:21JdGordonor a .o like sysfont.o
12:12:34kugelsvn mv apps/gui/skin_engine tools/ as well?
12:12:46JdGordonsvn rm apps/gui/skin_engine
12:13:10kugelI would be against that but I don't know what others think
12:13:23JdGordonI'm rebuilding that from scratch to work with the new parser...
12:14:08JdGordonat the very least skin_parser.c is totally useless with the new code
12:14:40pamauryfunman: How is PCM related to usb ?
12:14:56JdGordonso... I need to know how to make a Makefile for those files to build a .a or .o, then have that magically work with root.make...
12:14:59funmanpamaury: not PCM, but spurious interrupts (check the commit log)
12:15:14ryranthanks for tips kugel, JdGordon. peace ya'll.
12:15:20 Part ryran
12:15:57kugelmaybe it would not be such a bad idea to have a lib/ dir in the root dir, that would make it less hacky for plugins to link tlsf malloc for example
12:16:13CIA-6New commit by funman (r26445): as3525 ascodec: make internal functions statics and remove some details for ascodec-target.h header ...
12:16:39JdGordonkugel: yes that would be nice
12:16:52pamauryfunman: I have a doubt, I'm not sure mzy code clear the interrupt with VIC_INT_EN_CLEAR, I don't have my computer with me so I can't be sure
12:16:53JdGordonZagor: are you still the Makefile guru?
12:16:55 Join watto [0] (~watto@
12:17:49funmanpamaury: grep told me so ^^ but for now it's only done at init so it should be ok
12:20:02 Quit einhirn (Read error: Connection reset by peer)
12:22:46 Quit sbhsu (Quit: leaving)
12:24:03JdGordonbieber: you know what... ignore the message above... as long as the parser guarentees each node from the root will only have one %V[il] tag in it then it is fine how it is...
12:27:41S_a_i_n_tare they *real* tiny? A mate of mine has like 1mm long fingernails...tops.
12:28:01S_a_i_n_ton a few fingers, you can't see any nail, he bites then so damn much.
12:28:16JdGordonWRONG CHANNEL!
12:28:20S_a_i_n_t*crap*...errrr, wrong channel.
12:31:55 Join sbhsu [0] (
12:33:16 Join leavittx [0] (~leavittx@
12:38:48JdGordonAlexP: is the release happening at devcon or before?
12:39:31rasherBefore, I thought?
12:40:39JdGordonok cool
12:45:05funmanis lenzone10 on irc? he made an update to italian.lang in r26438 but it's not in branch
12:46:08 Quit stripwax (Quit:
12:51:55kugelFS #11341 - not a bug?
12:54:37JdGordonnot a bug
12:54:53JdGordonjust like you cant easily stop using a .wps
12:55:51kugelright, thought so too
12:56:11kugelsimple fix is to place a resetfms.cfg into the themes directory
12:58:14JdGordonemptytheme.cfg with -'s for everything would be better
12:58:39JdGordondoes - mean use the default (cabbie) or hardcoded?
13:02:12JdGordonok cool, then yeah resettheme.cfg would be nice to add
13:08:21 Quit funman (Quit: free(random());)
13:14:05 Join ucchan [0] (
13:14:22 Quit kugel (Ping timeout: 240 seconds)
13:18:57 Join Casainho [0] (
13:20:10ucchanS_a_i_n_t: I confirmed all cross compilers were built (on cygwin 1.5/1.7) !
13:20:45ucchanAnd updates more.
13:20:56Casainhoanyone want to join us?
13:21:06Casainhoon hacking an ARM9 @ 450MHz from SigmaTell / FreeScale?
13:21:41ucchanchanges: the problem which binutils is built using by "make" (not "make -j4") when two or more cross compilers are selected.
13:21:58Casainhoit's a SoC with analog features, such DC-DC for low power, analog input/ouput for audio, etc
13:24:00 Join stripwax [0] (
13:25:21ucchanI will check in if there is no problem in new
13:26:09JdGordonbieber: FS #11345
13:33:26AlexPJdGordon: Yes, thursday is the plan
13:36:18 Join elinenbe [0] (
13:37:49 Quit S_a_i_n_t ()
13:38:03 Join S_a_i_n_t [0] (S_a_i_n_t@
13:40:03 Quit yosafbridge (Quit: Coyote finally caught me)
13:40:11 Join yosafbridge [0] (
13:41:59 Join nls [0] (
13:44:45ZagorJdGordon: yes
13:44:59 Quit n1s (Quit: Lämnar)
13:45:01 Quit nls (Client Quit)
13:45:18 Join n1s [0] (~n1s@rockbox/developer/n1s)
13:47:58 Quit n1s (Client Quit)
13:48:34 Join fatih [0] (~Fatih@
13:53:21 Join einhirn [0] (
13:55:32 Join Szpila [0] (~sszpila@
13:57:39 Quit B4gder (Quit: It is time to say moo)
13:57:56***Saving seen data "./dancer.seen"
13:58:10 Join n1s [0] (~n1s@rockbox/developer/n1s)
13:59:40 Quit n1s (Client Quit)
13:59:59 Join n1s [0] (~n1s@rockbox/developer/n1s)
14:06:30 Quit dfkt (Quit: -= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.)
14:13:30 Join JdGordon1 [0] (~jonno@
14:16:22JdGordon1Zagor: utils/themeditor has a bunch of .c's I want to compile to a lib to share between the target build and the editor.... I dont know where to start
14:16:44 Quit shai (Read error: Connection reset by peer)
14:18:24 Quit bieber (Ping timeout: 264 seconds)
14:18:33 Quit bmbl (Quit: Bye!)
14:18:38 Join bieber [0] (
14:23:27 Part Szpila
14:29:56 Quit ucchan (Quit: Leaving...)
14:39:15 Part fatih
14:48:14GeekShado_good news, the addon FolderSync that let people sync devices with Songbird will support importing Rockbox play count :)
14:49:03AlexPSounds like time to change the format then :)
14:50:39 Join anewuser [0] (anewuser@unaffiliated/anewuser)
14:53:24 Quit efyx (Remote host closed the connection)
14:53:34 Join efyx [0] (
15:19:01 Quit DerPapst1 (Quit: Leaving.)
15:22:27 Quit flydutch (Quit: /* empty */)
15:27:34 Quit anewuser (Quit: for SELL 2 by the price of 1 now!)
15:28:17 Join kugel [0] (~kugel@rockbox/developer/kugel)
15:35:35 Join MethoS- [0] (~clemens@
15:35:56 Join grndslm [0] (
15:36:03 Join komputes [0] (~komputes@ubuntu/member/komputes)
15:41:44 Join dfkt [0] (dfkt@unaffiliated/dfkt)
15:50:00 Quit Casainho (Quit: ChatZilla 0.9.86 [Firefox 3.6.3/20100423140709])
15:50:33 Join jgarvey [0] (
15:51:39ZagorJdGordon1: which .c's should go into the lib?
15:53:38 Quit voRia (Quit: Leaving.)
15:53:45JdGordonall in utils/themeeditor/ skin_parser.c skin_debug.c skin_scan.c tag_table.c
15:54:01 Quit JdGordon1 (Quit: Leaving.)
15:54:17ZagorJdGordon: ok. could you explain the "sharing" part?
15:54:41JdGordonthe same code is being used by the target build and the theme editor
15:55:35JdGordonkugel suggested we add a libs/ folder to the tree for things like this
15:56:11 Quit pixelma (Ping timeout: 260 seconds)
15:56:26ZagorI'm not sure that makes things easier
15:56:44 Quit amiconn (Ping timeout: 276 seconds)
15:57:22JdGordonis what I'm asking not so straight forward?
15:58:00***Saving seen data "./dancer.seen"
15:58:15Zagorthis is used in the themeeditor branch, right?
15:58:50JdGordonno, trunk
16:03:47 Quit yosafbridge (Ping timeout: 240 seconds)
16:05:40JdGordonanother thing is it will need a #define so it knows if it is compiling for target or not.. I assume that shouldnt be hard to add (going by my very little understanding of the other make files....)
16:08:44Zagorskin_parser.c exists in both locations, though is not the same code
16:09:37JdGordonone can be renamed if it needs to be
16:10:57ZagorI suggest simply moving the files into apps/gui/skin_engine and then compile whichever files the editor wants, using that path
16:12:49JdGordonreally? it would be nicer if the code kept out of there
16:16:54Zagorwhat? didn't you say it should be compiled in target and in the editor?
16:17:12JdGordonyes, but the actual code belongs to neither
16:17:37 Join liar [0] (
16:19:08 Nick GeekShado_ is now known as GeekShadow (
16:19:55JdGordonif the code ends up in either of them it will eventually bloat out and cause problems like we have with checkwps where heaps of stubs and mini structs are needed for it to work
16:20:40CIA-6New commit by mc2739 (r26446): Perform radio_start, radio_pause, radio_stop in simulator. ...
16:21:03kugelnothing prevents you from compiling it as a lib in apps/gui/skin_engine and I don't think the actual folder will make any difference
16:21:13JdGordonI want those files to build to a libskinparser.a which the rockbox bin and the themeeditor can both link against (of course linking against the one built for it, with the right gcc)
16:22:10JdGordonmc2739: nice one :)
16:22:52JdGordonkugel: of course the folder doesnt really make a difference, but storing it in a logical place helps
16:27:57 Quit wincent (Ping timeout: 248 seconds)
16:30:16 Quit kugel (Ping timeout: 260 seconds)
16:31:05CIA-6New commit by mc2739 (r26447): Backport r26446 - Perform radio_start, radio_pause, radio_stop in simulator.
16:34:49 Join kugel [0] (
16:34:50 Quit kugel (Changing host)
16:34:50 Join kugel [0] (~kugel@rockbox/developer/kugel)
16:39:25 Join kugel_ [0] (
16:39:40 Quit kugel (Disconnected by services)
16:39:43 Nick kugel_ is now known as kugel (
16:39:49 Quit kugel (Changing host)
16:39:49 Join kugel [0] (~kugel@rockbox/developer/kugel)
16:44:17 Join yosafbridge [0] (
17:10:09 Quit bieber (Ping timeout: 260 seconds)
17:10:26 Join bieber [0] (
17:10:57 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
17:19:05 Part Zagor
17:27:26 Join phanboy_iv [0] (
17:32:56 Quit lpereira (Quit: Leaving.)
17:44:44 Quit bluebrother (Disconnected by services)
17:44:46 Join bluebroth3r [0] (~dom@rockbox/developer/bluebrother)
17:50:42 Join funman [0] (~fun@rockbox/developer/funman)
17:52:13 Join DerPapst [0] (
17:56:01 Quit kugel (Ping timeout: 260 seconds)
17:58:04***Saving seen data "./dancer.seen"
17:59:36 Join Zigtown [0] (
18:02:18 Quit petur (Quit: later)
18:06:08 Join stoffel [0] (
18:07:54 Join kugel [0] (~kugel@rockbox/developer/kugel)
18:13:13 Quit GeekShadow (Quit: The cake is a lie !)
18:13:51stripwaxis there a ref/doc/something similar to explain why r25938 results in a performance boost for ipod video lcd update? (changing the copy loop from 4-word ldmia/stmia to a sequence of 2-word ldmia/stmia instead)
18:15:00stripwaxin my mind, a 4-word stmia is always going to be quicker than two 2-word stmias, so interested to know what I'm missing! is it something to do with the memory controller?
18:15:26stripwax(background: wondering if this loop could/should be written in Thumb instead)
18:17:14kugelwhy in thumb?
18:17:24kugelto save 16 bytes?
18:20:28funmanstripwax: perhaps it comes from the hardware ?
18:20:43stripwaxkugel - I thought tight loops benefited from reduced code size? or is there no benefit on ipod video when running from icode?
18:21:18 Quit pamaury (Remote host closed the connection)
18:21:22stripwaxfunman - perhaps, wondering if someone knew :) Buschel, presumably, but he's not online right now (although this isn't urgent :-) )
18:21:27funmanARM System Developer's Guide says ldm/stm take 2/1 + N (== number of regs) cycles on arm7tdmi so more registers would be faster
18:21:54stripwaxexcept when it's not, presumably
18:22:01funmanstripwax: for tight loops, i think it only matters when the loop is loaded in the cache
18:22:52stripwaxfunman - that's ambiguous - you mean it only matters 'if the code is in cache', or the only benefit occurs 'when caching the code for the first time' ?
18:23:06funmanwhen caching
18:23:32funmanhave you seen fs#6734 & fs#11335 ?
18:23:34stripwaxah; in that case no benefit whatsoever from Thumb there.
18:23:38funman(about thumb)
18:24:19stripwaxyep (also saw the commits replacing mov pc,rx with bx)
18:24:20 Join DataGhost [0] (
18:24:21 Quit DataGhost (Changing host)
18:24:21 Join DataGhost [0] (~dataghost@unaffiliated/dataghost)
18:25:34funmandid you see the replacing of mov pc, lr by bx lr which bricked my 2 fuzes ? :/
18:25:56stripwaxthis patch just makes the ipod video lcd update code thumb-safe, but doesn't change any of the video asm to use thumb instructions
18:26:05stripwaxfunman - hm, actually no I didn't see that!!
18:26:07 Quit kugel (Ping timeout: 265 seconds)
18:26:20funmanah now I understand what happened:
18:26:49funmanthe code is pure thumb and uses 'mov lr, pc' + branch
18:27:03n1sstripwax: did you have time to look more at my tremor patch (FS #11268)? And what do you think of the approach of not using the iram_malloc stuff and instead using the regular ogg_malloc eveywhere when there is no iram as opposed to the easier slightly hacky approach of pretending to have enough iram to avoid the copying?
18:27:26funmanso when returning from thumb, to thumb, it relies on the fact that mov pc, lr will not update the T bit, so CPU will stay in thumb mode
18:28:11funmanaddress in lr is aligned on 2, so bx lr will either do something unpredictable if it's not aligned on 4, either switch to ARM since bit 0 of lr is 0 => boom
18:28:54funmanbut i think 'mov pc, lr \n beq X' could be bleq X ?
18:28:55 Join kugel [0] (~kugel@rockbox/developer/kugel)
18:29:33funmanah no BL can not use conditions in thumb
18:29:50stripwaxn1s - no not yet unfortunately. maybe sometime this week if I get time, but lots on :(
18:31:05n1sstripwax: no worries, i have still not installed a toolchain on my new machine :)
18:31:32stripwaxfunman - right, bx lr will not do anything unpredictable if lr not algined on 4, it will just jump to lr &~0x00000001 and set T if lr & 0x01
18:31:49funmanBagder: can you hand the JTAG'd e200v1 to Zagor so he can bring it to DevCon?
18:32:10stripwaxbut yeah, jumping from thumb to thumb and not ensuring lsb is set will do bad things
18:33:15funmanstripwax: if lr & 3 == 2, i think it's unpredictable because you will clear T bit and address will not be aligned on 3 bits
18:33:52funmanin nrv2e_d8 code the lsb is unset because they store the return address by reading pc
18:34:32funmanmov pc, lr will not work on armv7+ because it will also update the T bit (in this case clear it)
18:35:05kugelhow about manually or'ing with 0x1?
18:35:45 Quit stripwax (Read error: Connection reset by peer)
18:36:06funmanworks too but cost one more instruction
18:36:33funmanif i read ARM Architecture Reference Manual for ARMv6 i understand that mov pc, lr will update the T bit on ARMv6
18:36:55funmanIn T variants
18:36:55funmanof architecture 4 and in architecture 5 and above, the instruction BX LR must be used in place of MOV
18:36:59funmanPC, LR, as the BX instruction automatically switches back to Thumb state if appropriate
18:37:19funmanhmm, no, they say MUST
18:38:54Torneyou have to use bx lr on v6 too, no/
18:40:00 Nick fxb is now known as fxb__ (
18:40:17funmanunless you don't want to switch :o
18:40:24kugelIIRC it's relaxed a bit v6+ but of course bx lr is recommended, if only for compatibility
18:40:29TorneYeah, ARM just think that shouldn't happen :)
18:40:39Torne*all* your symbols should have lsb set if they refer to thumb functions
18:40:51 Quit slck (Ping timeout: 276 seconds)
18:42:26funmani hope to see what's wrong with e200v1 built with thumb with the JTAGed e200v1, then i will check if fs#11335 doesn't impact performance
18:42:51 Join slck [0] (Venci@Slackware.SlackPix.Com)
18:43:19funmani think it's better to always use bx rather than have a ROCKBOX_THUMB define to economize 4 bytes per return and 1 cycle on arm7tdmi
18:44:49*kugel agrees
18:44:54CIA-6New commit by bieber (r26448): Committing FS #11345 by JdGordon. Theme editor parser now includes full tag information in the skin_element struct
18:44:55funmanthe mov pc, #X remaining in s3c2440 are branches from crt0 to OF or reset
18:44:59kugelbut that patch doesn't fix all asms does it?
18:45:15kugelwhat about system-arm.h?
18:45:17 Join shai [0] (
18:45:24funmankugel: it should, i gave teh regexp i used and what the matches remaining are
18:45:54funmankugel: this patch only fixes returns so the functions can be called from thumb
18:46:08funmanthen, we can move to thumb experimentation ;)
18:46:46funmanPP crt0 has a mov pc, #0x40000000 which should be IRAM_ORIG
18:47:49Tornefunman: there isn't any risk of us having to support ARMv4 is there?
18:47:54Tornei.e. not ARMv4T ?
18:48:19funmani don't know but the linker has a workaround to replace bx by armv4 instructions (there is discussion about this on debian eabi port)
18:48:41funman −−fix-v4bx Rewrite BX rn as MOV pc, rn for ARMv4
18:48:41funman −−fix-v4bx-interworking Rewrite BX rn branch to ARMv4 interworking veneer
18:49:42funmanPP crt0 copies its own ARM code so it should be ok
18:49:45 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
18:51:40kugelfunman: weren't they sane enough to realize there's no point in supporting armv4 non-T?
18:53:28CIA-6New commit by funman (r26449): mkamsboot / nrv2e_d8.S thubm decompressor : comment the use of mov pc, lr
18:55:42funman <- should be ok ?
18:57:30 Quit shai (Quit: Leaving)
18:59:54funmannope bx needs register not label :/
19:00:54r0b-ok if i was wanting to look at the FM Radio init code where would i look?
19:04:01 Join Bob_C [0] (
19:04:23Tornekugel: Lots of machines debian is regularly run on requires support for machines that don't know bx
19:05:19Bob_CHow is EABI going? :)
19:05:45bieberI've committed FS #11345, it would be much appreciated if someone with admin access on FlySpray could close it
19:05:47kugelBob_C: good these days, I just added a related discussion idea to the devcon agenda
19:05:57Tornekugel: indeed.. they run the arm port, not armel
19:06:15kugelisn't the arm port deprecated?
19:07:29 Join kugel_ [0] (
19:07:47 Quit kugel (Disconnected by services)
19:07:51 Nick kugel_ is now known as kugel (
19:07:57 Quit kugel (Changing host)
19:07:57 Join kugel [0] (~kugel@rockbox/developer/kugel)
19:08:50Bob_Ckugel: thats good news!
19:08:57 Quit kugel (Read error: Connection reset by peer)
19:11:23 Quit phanboy_iv (Read error: Connection reset by peer)
19:11:37funmanr0b-: for clip+ / sansa ams it is in firmware/target/arm/as3525/fmradio-i2c-as3525.c
19:11:57r0b-thank you sir
19:13:39 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
19:14:53r0b-oooh i2c oughta be fun
19:17:47r0b-are there any special make flags i need to specify?
19:18:20 Join kugel [0] (~kugel@rockbox/developer/kugel)
19:18:36kugelgevaerts: ping
19:19:31 Quit DerPapst (Quit: Leaving.)
19:21:27gevaertskugel: gnip
19:21:51kugelgevaerts: seen my mail?
19:22:55gevaertslooking now
19:27:42r0b-can anyone explain why i get this when i build rockbox?
19:28:03r0b-im using the VMWare development set
19:28:34kugelgevaerts: I send another, I forgot 2 important points
19:29:15saratogar0b-: you probably have a compiler setup that does not support the CPU you're compiling to
19:29:33r0b-im using the VMware enviroment from
19:29:43saratogadoes it support the CPU you're compiling for?
19:29:50saratogamy guess would be "no"
19:29:56r0b-let me check
19:30:00r0b-how doi check?
19:30:10 Join Luca_s [0] (
19:30:15saratogatry building and see if it gives you the errors you get
19:30:32funmani guess it's a multilib problem
19:30:36saratogarun and see if that fixes the problem
19:30:43Luca_sr0b-: I had problems with the vmware env
19:30:48saratogathe wiki has instructions for setting up the cross compilers if you've never done that before
19:30:54Luca_sfixed by running
19:30:54saratogaits extremely simple
19:31:46kugelspeaking of, who shoved the -j4 into its make command?
19:32:26kugelit gives problems on cygwin, and presumably single core (or even dual core) machines as well, I think it should rather look for $CONCURRENCY_LEVEL
19:32:26funmanJdGordon: in r26078
19:32:52funmanhe probably forgot to check his diff
19:33:55funmanjust revert it (or calculate a good -j :P )
19:35:21gevaertskugel: I suspect that your mentors don't actually prefer android :)
19:35:32kugelmake -j "0$CONCURRENCY_LEVEL" ?
19:36:01kugelgevaerts: yea, n900 fanboys :P
19:36:13funmanwhat's CONCURRENCY_LEVEL ?
19:36:27gevaertskugel: I suspect shouldn't be setting -j at all, that's what MAKEFLAGS is for
19:36:49kugelfunman: dpkg-buildpackage uses that for the job count
19:36:57 Quit swilde (Remote host closed the connection)
19:37:28kugelgevaerts: you never know if setting such variables breaks the scripts
19:37:42gevaertskugel: it doesn't here
19:38:12gevaertsBut anyway, people who want -jwhatever can already set it
19:38:20kugelit may not break, but it's not too save generally
19:38:34kugelmaybe just add an -j option to
19:39:22kugelgevaerts: seriously though, you wouldn't want android?
19:39:37 Join pixelma [0] (quassel@rockbox/staff/pixelma)
19:39:40 Join amiconn [0] (quassel@rockbox/developer/amiconn)
19:41:18gevaertskugel: for myself or as the first RaaA target?
19:41:26kugelthe latter
19:41:56gevaertsIt's probably the most realistic platform right now
19:42:26evilnick_Bkugel: Excuse the possibly silly question here, but would RaaA if on Android, work with any currently released Android version, or only the newer versions?
19:42:29kugelwell, I'd expect an n900 port to be less of a challenge
19:43:01gevaertsThe choices would be windows mobile, symbian, iphone, maemo, android and webos I guess
19:43:10kugelevilnick_B: I don't know, I haven't messed with the version issue very much so far
19:43:55gevaertskugel: yes, that would be the other one. It might be less of a challenge (although I'm not sure if it really makes that much difference), but it's a lot less affordable
19:43:57kugelI think winmo is out, as far as win 7 mo is concerned
19:44:07kugelthey take the same route as apple IIUC
19:45:45kugelgevaerts: I think maemo (or meego in the future?) doesn't run on many enough devices to be really interesting right now
19:45:56gevaertsthat too
19:45:58kugelhow much is the n900?
19:46:50gevaertson, 489 euro right now, which they say is a temporary low price
19:47:15kugelI see, the google money won't be sufficient for that
19:48:09gevaertsYes, it's a bit too much to expect someone to spend if there are nearly-as-good (if the "nearly" is even true) options for a fraction of that
19:48:54gevaertskugel: anyway, about your other question, I think there are a few loose threads that should be tied up at some point. There's mainly the system vs rockbox libc issue that I think should be either a configure switch or an advanced option
19:48:57Bob_CI would be interested in android because I can run it on mini2440 board
19:49:47kugelBob_C: it's very slow though, and I think it's unable to run 2.x androids
19:50:04Bob_Coh, ok
19:50:23kugelgevaerts: there hasn't been a demand for it lately
19:51:02gevaertskugel: well, people don't keep repeating it, but I don't think the demand is gone
19:53:35 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
19:54:03kugelgevaerts: ok, I'll have a shot at it. shouldn't be too hard
19:55:44kugelgevaerts: another thing: I'm well in schedule. my plans were to be done with the mid-term work by july 1st. I think can afford (and need a bit too) concentrating on my uni a bit more. Is that OK?
19:56:15 Quit stoffel (Remote host closed the connection)
19:56:25saratogaandroid is nice because theres a million new cheap android devices coming out
19:56:48kugelit also looks like my schedule for the exam period was a bit off, I doubt I'll be completely unavailable for 3 weeks
19:57:06saratoganot just $300 phones that require long contracts (at least in the US)
19:57:06 Join toffe82 [0] (~chatzilla@
19:57:28gevaertskugel: I'd expect you to prioritise uni things anyway
19:58:07***Saving seen data "./dancer.seen"
19:58:29gevaertskugel: I'd say we discuss things next weekend :)
19:58:44kugelfine with me :)
19:59:44kugelsaratoga: it also runs on a lot DIY systems
20:03:14 Join Horscht [0] (~Horscht2@xbmc/user/horscht)
20:04:27r0b-ok i built rockbox :)
20:05:29kugelevilnick_B: android 1.x might be a bit tricky, I haven't looked into it really, but I think the NDK (for writing apps in C/C++) is less complete or even in-existent for 1.x
20:10:48 Join panni_ [0] (
20:11:31 Join Whendoze [0] (~mIRC@
20:11:59funmankugel: android ndk only requires android sdk 1.5 or later (
20:12:45Whendozei have a question that you all have been asked
20:13:00Whendozeyou guys gonna release a build for the new iPod Nano?
20:13:30n1sWhendoze: if someone ports rockbox to it we will release a build
20:13:32AlexPThe ipod nano has been in releases for ages
20:13:39AlexPoh sorry, new
20:13:43*AlexP should read
20:13:44Whendozeyes :)
20:13:49Whendozeits ok
20:13:57AlexPWhendoze: See
20:15:05 Join einhirn [0] (
20:15:07Whendozeif it involves coding anything like C C++ then i will just wait haha
20:15:15n1sit does
20:15:16AlexPThat is the least of your problems
20:15:26AlexPAnd the wait will be long
20:16:33n1syes, there are working exploits and even some drivers ade for the (3G and?) 4G nanos but still, no rockbox port has been started
20:16:49n1son the 5G there is no known exploit afaik
20:19:36evilnick_Bkugel: Thanks for the info
20:19:42 Quit Whendoze ()
20:19:48 Quit shaggy-h (Ping timeout: 240 seconds)
20:23:22 Quit [CGL] (Ping timeout: 260 seconds)
20:23:23 Quit gevaerts (Disconnected by services)
20:23:33 Join gevaerts [0] (~fg@rockbox/developer/gevaerts)
20:26:24 Join jae_ [0] (
20:26:37 Join ehntoo_ [0] (
20:26:45 Quit Horscht (*.net *.split)
20:26:45 Quit sbhsu (*.net *.split)
20:26:45 Quit Don_Roberto (*.net *.split)
20:26:45 Quit ehntoo (*.net *.split)
20:26:45 Quit jae (*.net *.split)
20:26:57 Quit bertrik (Remote host closed the connection)
20:27:16 Join Don_Roberto [0] (
20:27:35 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
20:29:18 Quit esperegu (Read error: Connection reset by peer)
20:29:18 Quit bertrik (Read error: Connection reset by peer)
20:29:21ranmafunman: Yeah, the -g is for gdb
20:29:40 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
20:29:49 Join sbhsu [0] (
20:29:55bieberJdGordon: Whenever you get back, I went ahead and implemented your suggestion to separate out the Viewport tag at the beginning of a VIEWPORT section in the parse tree.
20:30:01funmanranma: check r26444 log - btw i think the problem only happened on AMSv2
20:30:13 Join esperegu [0] (~quassel@
20:30:27ranmaOk, I'm currently testing rockbox with gcc 4.4.4 and forced boost and so far it's not happened here.
20:30:29funmanthat, or it was much less frequent and i couldn't reproduce it (forum reports mentioned fuzev1)
20:30:50bieberNow, a VIEWPORT element will have 1 child as usual if it's the default viewport, but if it's started with a viewport tag then that tag will be child 0 and everything else will be child 1
20:31:02 Join Horscht [0] (~Horscht2@xbmc/user/horscht)
20:31:15funmandid you see how the OF disables interrupts ?
20:32:00CIA-6New commit by bieber (r26450): Theme Editor: Made the viewport tag a top-level child under VIEWPORT elements in the parse tree
20:32:00funmanint save = VIC_INT_ENABLE & ~(interrupt_to_disable); VIC_INT_EN_CLEAR = 0xffffffff; VIC_INT_ENABLE = save;
20:32:03ranmaYeah, but I don't think it's necessary to do it that way.
20:33:13funmanif you're interested, check the .au link in the commit log it's the most detailed
20:33:42ranmaAh, so it's likely a spurious irq problem as I thought.
20:33:49funmannot sure if that scenario can happen for INT_AUDIO though
20:34:03funmanif the interrupt is asserted until we read the i2c registers, that would be a very long time
20:34:52funmanwe could try making several adc_read() requests and plugging/unplugging usb
20:35:38funmanif there is a race condition between the interrupt firing and writing VIC_INT_EN_CLEAR the timing would be short since it's the first thing we do in the isr
20:36:25funmanoh well, in this case we could just make UIRQ() return - it didn't work for DMA, I think the interrupt is asserted until we clear it from the DMA peripheral (does this make sense?)
20:38:28 Join lpereira [0] (
20:38:36 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
20:39:09funmanbtw I tried building a linux kernel for a machine using the VIC but I still don't know what the address '32' is
20:43:07ranmaHmm, reading the explanation on, the problem here _could_ be that the interrupt is taken while the write to the vic_en_clear is in the pipeline.
20:43:31ranmaSo we really should disable irqs before writing to vic_en_clear and re-enable afterwards.
20:44:16funmanno it doesn't work
20:44:51funmani understand that the problem is when the peripheral asserts the interrupt towards the VIC when vic_int_en_clear is being written
20:45:09funmanso the interrupt is forwarded to the CPU but with the default ISR
20:47:10ranmaDid you try that already? I think it might work. In that case even if an interrupt comes in right at that time, the cpu should ignore it because interrupts are disabled.
20:47:21funmani did
20:47:35funmanthe cpu will get it when interrupts are enabled afterwards no?
20:48:18ranmaIt shouldn't because when they are enabled afterwards the irq line is no longer asserted.
20:49:11 Join spike_ [0] (
20:49:16spike_hey guys
20:50:12spike_i just got a sansa fuze v1 and I want to put rockbox on it, i have the rockboxUtility running on my ubuntu install
20:50:26ranmaThe 'real spurious interrupt' referred to on should only happen if the peripheral is broken or interference is coupled in on the periphal irq line or if you generate it on purpose as in the example
20:50:38spike_I'm still just worried about bricking the player, what should I do to make that less likely to happen, just charge the thing up all the way first?
20:51:02funmanranma: what's with transient interrupts?
20:51:22funmanspike_: if you want to make it less likely to happen don't do it
20:51:29 Join DerPapst [0] (
20:51:38spike_i figured someone would say that
20:52:12spike_but the whole reason i bought the thing was to put rockbox on it =\
20:54:06funmanIIUC, DMA shouldn't cause transient interrupts, the interrupt should stay asserted
20:54:08ranmafunman: Normally it shouldn't be transient because you have to explicitly clear it at the periphal.
20:54:59ranmaOne option could be to just ignore the spurious irq.
20:55:08funmandoesn't work for DMA
20:56:40 Join shai [0] (
20:56:52 Quit Luca_s (Quit: CGI:IRC (EOF))
20:56:53ranmaArgh, I hate that arm site. /me would like to read "What happens if an interrupt occurs as it is being disabled?", but...
20:58:19ranmaOk, works in firefox, but not in chrome...
20:59:19 Nick [keanu] is now known as keanu (~keanu@unaffiliated/keanu)
20:59:20kugelnormally irqs are post-poned to after when interrupts are re-enabled, no?
21:00:40ranmaI don't think so. Since they are level triggered, if a transient happens while they are disabled the transient should be ignored then.
21:00:57 Join moparx [0] (~moparx@unaffiliated/moparx)
21:01:05ranmaNormally because the irq is not transient it's taken when interrupts are re-enabled since the line is still asserted.
21:07:56 Join wincent [0] (
21:08:44 Part watto
21:10:52funmanlinux copies the vectors at #0xFFFF0000
21:12:12ranmaThat's an alternate offset for arm irq vectors I think. But linux doesn't use the vic vectors at all AFAICS.
21:12:24ranmaIt just reads the VIC irq_status register to determine the source ranges from c0004000 to c028xxxx, no trace of 32
21:12:52ranmaNo Idea why it writes 32 to the default vector register.
21:13:54funmanah right
21:16:44ranmaThe write to VIC_INT_EN_CLEAR from INT_AUDIO is probably safe because it's in the irq handler and irqs are disabled.
21:17:03funmanwriting INT_EN_CLEAR with irqs disabled didn't work for pcm DMA
21:17:24funmanranma: linux 2.6.22 doesn't use the vectors but 31.5 (at least) does
21:18:01ranmaOh, ok, I only looked at the 2.6.22 patch
21:19:24funmanas353x irq handler only looks at VIC_IRQ_STATUS though
21:20:12funmanso the vectors code is available but not used, so address 32 doesn't mean anything
21:20:16 Quit lpereira (Quit: Leaving.)
21:20:26 Join lpereira [0] (
21:21:28ranmaEven if wrapping VIC_INT_EN_CLEAR in disable/restore_irq didn't help with the pcm DMA issue, I still think it should be done for correctness.
21:21:30funman <- as3525v2
21:22:07evilnick_Bspike_: There will always be a (very small) risk in flashing a new firmware image, but Rockbox on the Fuze v1 is a relatively mature port
21:22:08funmanranma: it's only needed in usb-drv-* now
21:22:35funmantimer-*.c disables the interrupt source in the peripheral before writing VIC_INT_EN_CLEAR and ascodec does it in the isr so interrupts are disabled
21:23:37funmanperhaps in usb-drv-as3525.c it's the same thing than with timer?
21:25:14 Quit funman (Quit: free(random());)
21:26:47 Quit sbhsu (Quit: leaving)
21:29:04CIA-6New commit by ranma (r26451): Prevent a race condition when writing VIC_INT_EN_CLEAR. Should be very difficult to hit in this case though...
21:30:13ranmaNo, I wasn't masking the interrupts before the write (now fixed)
21:30:27 Part Bob_C
21:32:11 Join sbhsu [0] (
21:34:07 Quit Zigtown (Ping timeout: 240 seconds)
21:40:21spike_oh well, turns out the firmware is v2
21:40:31spike_i'll just wait till its supported by the official installer
21:43:46 Quit esperegu (Remote host closed the connection)
21:45:51 Join esperegu [0] (~quassel@
21:46:48 Join kugel_ [0] (
21:47:08 Quit kugel (Disconnected by services)
21:47:11 Nick kugel_ is now known as kugel (
21:47:16 Quit kugel (Changing host)
21:47:16 Join kugel [0] (~kugel@rockbox/developer/kugel)
21:47:44 Quit einhirn (Read error: Connection reset by peer)
21:54:17 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
21:55:21CIA-6New commit by bieber (r26452): Theme Editor: Enabled editing tag parameters from a treeview
21:55:42 Quit esperegu (Remote host closed the connection)
21:58:09***Saving seen data "./dancer.seen"
22:03:11spike_blah, i want to try it anyway
22:13:18pixelmais there a specific schedule for the release on thursday?
22:13:41Rob2223spike_: whts you model number of your fuze? read side? a 1004?
22:15:50pixelmaI'd like to update the German translation before (and maybe review related patches if there are some) but I am not sure if I can get around doing so tomorrow. It should be possible on thursday - it's a public holiday in the federal state I'm currently work in
22:19:54CIA-6New commit by bieber (r26453): Theme Editor: Made text and comments editable from a treeview
22:21:39 Join esperegu [0] (~quassel@
22:24:37 Join kugel_ [0] (
22:24:57 Quit kugel (Disconnected by services)
22:25:01 Nick kugel_ is now known as kugel (
22:25:05 Quit kugel (Changing host)
22:25:05 Join kugel [0] (~kugel@rockbox/developer/kugel)
22:25:11 Join merbanan [0] (
22:28:32 Join kugel_ [0] (
22:28:46 Quit kugel (Disconnected by services)
22:28:50 Nick kugel_ is now known as kugel (
22:29:05 Quit kugel (Changing host)
22:29:05 Join kugel [0] (~kugel@rockbox/developer/kugel)
22:31:37Rob2223spike_ : a BI1004CAHK?
22:33:29Rob2223spike_: lucky guy. youre not from germany, or?
22:33:46spike_nah, usa
22:34:11Rob2223ok according my theories that player should accept the rockbox firmware
22:34:33Rob2223do you install rockbox?
22:34:54spike_i haven't tried yet, i'm having trouble getting the rockboxUtility to recognize it
22:35:18Rob2223i think you have to manually install it
22:35:39spike_arg =\
22:36:02 Quit kugel (Ping timeout: 240 seconds)
22:37:04AlexPpixelma: I'd guess evening
22:37:19 Join kugel [0] (~kugel@rockbox/developer/kugel)
22:37:41CIA-6New commit by wincent (r26454): pdbox: Code cleanup, optimizations. ...
22:39:04 Join kugel_ [0] (
22:40:53 Join kugel__ [0] (
22:41:06 Quit kugel_ (Read error: Connection reset by peer)
22:41:13 Quit kugel (Disconnected by services)
22:41:17 Nick kugel__ is now known as kugel (
22:41:23 Quit kugel (Changing host)
22:41:23 Join kugel [0] (~kugel@rockbox/developer/kugel)
22:41:37 Quit wincent (Changing host)
22:41:37 Join wincent [0] (~wincent@rockbox/developer/wincent)
22:42:43wincentI would like to add recording capability to pdbox. Which plugin can I look at to see how the reading of audio input is done?
22:43:42 Quit S_a_i_n_t (Ping timeout: 260 seconds)
22:45:44 Quit bmbl (Quit: Bye!)
22:47:05n1si think the pitch_detector is a pretty good example now
22:47:06 Quit kugel (Ping timeout: 260 seconds)
22:47:27pixelmaAlexP: would be good (and maybe someone will remember my question now)
22:52:21 Join S_a_i_n_t [0] (S_a_i_n_t@
22:58:11 Quit lpereira (Quit: Leaving.)
23:00:37Rob2223which CPU settings i have to choose for firmware disassembling of the fuze2 firmware? is it arm?
23:02:58 Join kugel [0] (~kugel@rockbox/developer/kugel)
23:05:35wincentn1s: Good example, thanks.
23:07:53 Quit kugel (Ping timeout: 265 seconds)
23:10:17 Join tom` [0] (
23:16:27 Join CGL [0] (~CGL@
23:17:29 Join kugel [0] (~kugel@rockbox/developer/kugel)
23:17:35kugelmy hybrid sim-raaa actually builds and starts up \☺/
23:17:37kugelit doesn't define SIMULATOR but it loosly based of the cowond2 sim
23:18:41 Quit spike_ (Quit: leaving)
23:20:05kugelthe changes needed in addition to the SIMULATOR => CONFIG_PLATFORM are relatively minimal
23:21:20evilnick_BDo we have a policy on swearing in the forums?
23:25:05CIA-6New commit by bieber (r26455): Theme Editor: Put together a simple GUI to test going back and forth between a tree view and a text edit box
23:25:09 Join kugel_ [0] (
23:25:17 Quit kugel (Disconnected by services)
23:25:21 Nick kugel_ is now known as kugel (
23:25:27 Quit kugel (Changing host)
23:25:27 Join kugel [0] (~kugel@rockbox/developer/kugel)
23:29:09tom`I recently had Rockbox on the Fuze V2 hang/crash while copying files (black screen, I couldn't get the device to turn on again).
23:29:25 Quit bertrik (Quit: De groeten)
23:29:47tom`The FAQ says that (and I read elsewhere online that) holding the power switch for upwards of 30 seconds should reset the player, but this didn't work
23:29:48 Quit merbanan (Ping timeout: 240 seconds)
23:30:16tom`I was afraid I'd managed to brick it until a few minutes ago, when it started back up - apparently the battery just had to run down
23:30:29 Quit GeekShadow (Quit: The cake is a lie !)
23:30:33 Quit evilnick_B (Quit: Page closed)
23:30:43tom`Is this a known issue, and is there any way to resolve it in the future without waiting around 4 days for the battery to die before restarting?
23:31:58CIA-6New commit by wincent (r26456): pdbox: Updated CPU boosting API
23:32:08 Join oldtopman [0] (
23:33:42kugelmeh, it doesn't play sound :(
23:38:40kugelfailed to load ./_temp_codec0.dll
23:38:40kugeldlopen(./_temp_codec0.dll): ./_temp_codec0.dll: undefined symbol: plugin_end_addr
23:39:44 Join fdinel [0] (
23:43:35 Join Strife89 [0] (
23:44:57CIA-6New commit by wincent (r26457): Added myself and Pure Data team to credits. ...
23:58:11***Saving seen data "./dancer.seen"
23:58:20 Quit n1s (Quit: Lämnar)

Previous day | Next day