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).

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2010-06-19

00:00:23dfktthe home button bounces between main menu and file list... that's not too bad
00:01:15funmanamiconn: home+select
00:01:27funman(i only found about it by reading the keymap file)
00:01:41amiconnSomeone pls disable jakorasia-cg
00:01:49bertrikI don't really mind changes to the keymap, except I really do like volume in lists and have PLAY as up, and MENU as down
00:01:56amiconnIt obviously claims to support sdl but doesn't
00:01:57funmandone already, amiconn check the backlog
00:02:35*dfkt deleted all white markings on the clip+ with acetone, so there's no play or menu... all democratic :)
00:02:39funmanbertrik: you mean UP as UP and DOWN as DOWN ;)
00:03:03bertrikyes, of course :)
00:03:59funmanhum radio.c defines specific features for the keymap
00:04:36pamauryfunman: you tried several times to get the synopsys datasheet for the usb controller right ? Who did you contact ? (I can always have a try)
00:04:54funmanpamaury: iirc i just filled their web form
00:06:05***Saving seen data "./dancer.seen"
00:06:21*funman slaps JdGordon
00:06:26togetichow do you scrobble the songs you listen to?
00:06:41togetici have it enabled, but i'm not sure how to submit to my account
00:06:47funmantogetic: check the manual
00:07:33 Quit CGL (Remote host closed the connection)
00:08:18jhMikeSamiconn: that would vary between arm7 and arm9? ?? (you read my other remark that compiling it with arm9tdmicc actually gets rid of it?)
00:08:32amiconnEver heard of gcc bugs?
00:08:55jhMikeSof course, you made it sound like a feature in your comment above. :)
00:09:01 Quit advcomp2019 (Ping timeout: 245 seconds)
00:09:02funmaniiuc gcc is the thing hackers blame when they can't see bugs in their code :)
00:10:10bertrikmaybe we're relying on some specific gcc behaviour that is really undefined
00:10:32dfkttogetic -
00:10:41amiconnHrm, 'make clean' doesn't work properly
00:10:45jhMikeSarm9tdmicc actually makes a smaller bin though (~1700 bytes smaller)
00:11:15amiconnThis libs/skin_parser/ stuff stays
00:11:27funmaneabi also triggers a bug in text_viewer, so far only on sansa ams ?
00:12:19 Quit Highlander (Quit: Quitte)
00:12:50jhMikeSfunam: btw, who does gcc blame when they can't see bugs in their code?
00:12:51 Quit einhirn (Read error: Connection reset by peer)
00:13:25amiconnThen they have a meta-bug
00:13:57funmanjhMikeS: binutils guys
00:14:08*jhMikeS thought they blamed hackers :\ (and the circle is complete)
00:14:13funmanand binutils people blame gcc in return
00:14:35funmanthere is a circle but we're not part of it, so nothing is our fault :P
00:14:48CIA-85New commit by 03funman (r26941): radio keymap: enable fuze/clip buttons ...
00:15:08funmandfkt: now center shows the list of presets so i think we can keep it that way (and home alternates between presets/scanning)
00:16:33CIA-85r26941 build result: All green
00:17:03dfktawesome, funman - now it's in line with the audio screen
00:17:22dfktdefinitely more cohesive/coherent than before
00:20:25funmanare we already beating OF battery life on AMSv2 ?
00:20:40funmanSansaRuntime has no OF benches
00:23:16togeticthanks funman and dfkt
00:24:07togetictoo bad the log wasn't present though
00:24:42dfkttogetic, once you enable scrobbling in rockbox, the log should be in your player's root
00:26:39 Join karlw [0] (cfd65326@gateway/web/freenode/ip.
00:27:26togeticdfkt: that's what i read, but it wasn't there, so i'm guessing it was never enabled :\
00:27:39togeticthough i recall enabling it, i went wrong somewhere
00:27:44*togetic shrugs
00:27:56karlwHas anyone gotten Nano 5th gen to work? I know about the encrypted firmware.
00:28:17dfktit's in settings > playback, and it's a simple on/off switch
00:28:59dfktyour files need to be tagged to work, it won't read file names
00:32:37funmanranma: *PANIC* ep1 OUT 0x80 (BNA)
00:32:59 Quit kugel (Read error: No route to host)
00:33:34 Join kugel [0] (~kugel@rockbox/developer/kugel)
00:33:35funmanafter writing rockbox.sansa and running 'sync' (sync never returned but the file was modified since i got a checksum error)
00:37:22 Part karlw
00:38:10 Quit GeekShadow (Quit: The cake is a lie !)
00:38:13CIA-85New commit by 03amiconn (r26942): Build iPod 1st/2nd Gen with EABI (main build and bootloader verified working).
00:39:52CIA-85r26942 build result: All green
00:42:34 Quit rob (Ping timeout: 276 seconds)
00:42:37CIA-85New commit by 03funman (r26943): clip recording keymap changes, thanks to dfkt
00:42:42CIA-85New commit by 03funman (r26944): ACTION_FM_QUICKSCREEN is unused
00:43:06funmanS_a_i_n_t: you tested nano1g with eabi, didn't you ?
00:43:10 Quit ender` (Quit: Women and Cats will do as they please. Men and dogs had better get used to it.-- Robert Heinlein)
00:44:06CIA-85r26943 build result: All green
00:45:56CIA-85r26944 build result: All green
00:46:03CIA-85New commit by 03amiconn (r26945): Tell the build system that iPod 1st/2nd Gen needs arm-eabi-gcc 4.4.4. Fix this for a bunch of bootloaders as well.
00:47:05amiconnscorche: Could you test EABI on your Nano 1st Gen?
00:47:43amiconnThere are only 3 ipods still on old ABI: Nano 1st, Mini 1st and iPod 4th Gen
00:48:46*amiconn summons domonoky
01:00:09 Quit kugel (Ping timeout: 240 seconds)
01:05:47funmansmall write transfers work fine on clipv1
01:09:10 Quit bertrik (Ping timeout: 248 seconds)
01:12:22 Join kugel [0] (~kugel@rockbox/developer/kugel)
01:13:08 Join bluebrother [0] (
01:13:08 Quit bluebrother (Changing host)
01:13:08 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
01:14:00funmanreading is fine: a full album at 2.14MBps, vs 0.78MBps for the OF
01:16:18 Quit bluebroth3r (Ping timeout: 240 seconds)
01:17:03 Quit dfkt (Quit: -= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.)
01:27:31 Quit kugel (Ping timeout: 240 seconds)
01:41:39scorcheamiconn: do you happen to have a build ready for me to load?
01:55:02 Quit stripwax (Quit:
02:04:24 Join kugel [0] (~kugel@rockbox/developer/kugel)
02:06:06***Saving seen data "./dancer.seen"
02:10:16 Quit DataGhost (Ping timeout: 264 seconds)
02:16:56 Quit efyx (Remote host closed the connection)
02:20:42funmanranma: have you seen the comment in otgdev_ops.c:1186 about RDE ?
02:21:18 Quit kugel (Remote host closed the connection)
02:26:39CIA-85New commit by 03funman (r26946): usb-drv-as3525: set bulk max packet size according to speed ...
02:28:55CIA-85r26946 build result: All green
02:37:50 Join CaptainKwel [0] (
02:38:31 Quit funman (Quit: free(random());)
02:48:41 Quit DerPapst (Quit: Leaving.)
03:06:04 Join fdinel [0] (
03:09:10 Join powell14ski_ [0] (
03:12:02 Quit 50UAAFOUW (Ping timeout: 272 seconds)
03:15:06 Join CGL [0] (~CGL@
03:17:56 Quit powell14ski_ (Ping timeout: 265 seconds)
03:27:00pamauryI got an error on the wiki:
03:27:01pamauryFoswiki detected an internal error - please check your Foswiki logs and webserver logs for more information.
03:27:01pamauryexec failed: Mass Storage only
03:28:54 Join powell14ski [0] (
03:30:29 Join powell14ski__ [0] (
03:31:06 Join powell14ski____ [0] (
03:31:48 Quit fdinel (Quit: Miranda IM! Smaller, Faster, Easier.
03:33:08 Quit powell14ski (Ping timeout: 252 seconds)
03:35:06 Quit powell14ski__ (Ping timeout: 265 seconds)
03:43:52 Nick powell14ski____ is now known as powell14ski_ (
03:46:06 Quit steve|m (Ping timeout: 276 seconds)
03:49:19 Quit pamaury (Remote host closed the connection)
03:58:24 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
04:00:48 Join steve|m [0] (
04:06:10***Saving seen data "./dancer.seen"
04:09:24 Join Zigtown [0] (
04:15:36 Join Barahir [0] (
04:18:35 Quit anewuser (Quit: for SELL 2 by the price of 1 now!)
04:19:37 Quit Barahir_ (Ping timeout: 276 seconds)
04:28:47 Join Guest14828 [0] (
04:29:08*Guest14828 waves at room
04:29:26CIA-85New commit by 03ranma (r26947): Comment on mps sizes
04:29:58 Nick Guest14828 is now known as AzureSky (
04:30:57AzureSkywas wondering if anybody has reported that the current build for fuze v2 isnt detecting microsd cards
04:31:26CIA-85r26947 build result: All green
04:32:40AzureSkytesting now
04:34:58CIA-85New commit by 03ranma (r26948): Add mdelay
04:35:47 Quit amiconn (Disconnected by services)
04:35:50 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:35:50 Quit pixelma (Disconnected by services)
04:35:51 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:36:06 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:36:07 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:36:11CIA-85New commit by 03ranma (r26949): Use mdelay
04:36:39CIA-85r26948 build result: All green
04:38:26CIA-85r26949 build result: 16 errors, 2 warnings (ranma committed)
04:39:14Battousaineat, it even tells you who to blame
04:39:37AzureSky47 still dosnt work with msd cards :/
04:40:07AzureSkytrying the archive build
04:41:41AzureSkyif only i didnt have to wait for the OF to refresh the db to test the new builds :P
04:44:12AzureSkyno joy, have to use the build i have from a few days ago :(
04:52:14CIA-85New commit by 03ranma (r26950): Include system-target.h for mdelay
04:53:51CIA-85r26950 build result: 16 errors, 2 warnings (ranma committed)
04:55:45ranmaHuh? Now that looks like a build system hicup? C200v2 bootloader builds fine without warnings here now.
04:56:16AzureSkyduno, i still cant get the fuzev2 to see microsd cards on any current build
04:57:28 Quit Topy44 (Ping timeout: 260 seconds)
04:57:35AzureSkythankfully i have a backup of an old build (first one i installed) that works.......
04:58:33 Quit TheSeven (Ping timeout: 245 seconds)
04:58:50 Join Topy44 [0] (
05:02:19ranmaAh, no, I'm stupid
05:02:47AzureSkyand did you break msd card support :P
05:04:00 Join TheSeven [0] (
05:04:00 Quit TheSeven (Changing host)
05:04:00 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
05:04:23CIA-85New commit by 03ranma (r26951): Of course there's no mdelay on c200(v1), so just use udelay
05:05:25 Part AzureSky
05:05:59CIA-85r26951 build result: All green
05:06:20 Join Guest14828 [0] (
05:06:30 Nick Guest14828 is now known as AzureSky (
05:06:54ranmaAzureSky: No, I don't think I broke msd card support ;)
05:07:27 Join saratoga_ [0] (9803c20d@gateway/web/freenode/ip.
05:08:48AzureSkyif i hotload the latist builds i can still see cards, but if u upgrade to latist builds fully it dosnt work :(
05:09:22saratoga_a few people have mentioned that
05:10:24AzureSkyif i had any programing skill i would try and fix it myself, but best i can do is edit text based config and script files :(
05:11:22saratoga_you could figure out which build broke it
05:11:25AzureSkyno older builds are in the archive, and i didnt even think to test and backup each build that worked :(
05:12:08AzureSkyif u got a link to prebuilts i got zero issue testing, Im pretty sure it was a build from today that did it :/
05:12:28saratoga_you'd have to build them yourself
05:13:01AzureSkyagain, no skills or tools to do that :(
05:13:17saratoga_then just wait for someone else to fix it
05:15:00AzureSkysomebody should make a note for people who are trying to rockbox their fuzes not to use a current build, If i check my flashdrive I have the older working version on there
05:17:54saratoga_if its not fixed someone will make a note
05:21:13AzureSkyok, i have a weird one for you
05:21:25AzureSkyit seems if you reload from a .sansa file the msd card appears
05:22:15AzureSkycould it be some kind of either display issue or an issue with load order or the like?
05:24:00 Quit saratoga_ (Quit: Page closed)
05:29:14AzureSkyok well, if anybody heres working on this, i have r26951 installed and have reloaded from its .sansa file and IT WORKS, but having to reload from a file every time i turn the player on is a bit of a pain :(
05:36:16 Quit Zigtown (Remote host closed the connection)
05:39:38 Join saratoga_ [0] (9803c20d@gateway/web/freenode/ip.
05:41:40 Join funman [0] (~fun@rockbox/developer/funman)
05:48:13 Quit Topy44 (Ping timeout: 260 seconds)
05:49:47 Join Topy44 [0] (
05:51:58CIA-85New commit by 03funman (r26952): FS #11417 by Joe Balough: fix audio/tuner on philips hdd6330
05:53:31CIA-85r26952 build result: All green
05:59:00 Quit panni_ (Quit: ( :: NoNameScript 3.81 :: ))
06:00:13funmanranma: any chance you could test last patch of fs#11297 and get a backtrace if/when it crashes ?
06:00:18funman(on clip+)
06:00:57AzureSkyhey funman :)
06:01:17funmani suppose there could be something wrong with memory when we are switching pclk and some bits are lost - no other idea atm
06:01:52AzureSkyif nobody has told you, there is a bug with fuzev2 and sd cards in newer builds
06:02:12funmandid you post about it on the forum? i just tested last build and it works fine here
06:03:02AzureSkyyeah, i did, and for some reasion the last build i tested only shows the sd card after you reload from the sansa file after the player boots
06:03:11funmanwhich sansa file ?
06:03:45AzureSkythe rockbox.sansa
06:04:00funmanah, it doesn't detect the card if i boot with it inserted
06:04:04funmanbut insertion works as intended
06:04:06AzureSkydropped it in root of player, and if i "play" it after boot the cards accessable
06:04:29AzureSkyit dosnt detect if i remove and reincert the card here,not sure why
06:04:40AzureSkyapparently 906 still worked (aib forums post)
06:04:48funmanhmm no it works fine when booting with the card inserted
06:05:45funmannow i see the card is present (in disk info) but not mounted
06:05:57funmancheck debug -> disk info -> microSD 1, is it present?
06:06:06AzureSkyif i downgrade to r26865 it works here and somebody on abi forums had it work for them with r26906,but if i go with a build from today, it just dosnt list it in files
06:06:13***Saving seen data "./dancer.seen"
06:06:24AzureSkylet me reboot to see what it says
06:07:13AzureSkymine shows 0-sd
06:07:27AzureSkywith an 8gb ridata microsd card
06:07:58funmanit shows either 'Not Found!' either 10 lines or so of details on the card
06:09:29AzureSkyduno, in view disk info it just shows m-03, 0=sd
06:10:30funmanthat's microSD '0' = internal storage; press right button
06:11:06AzureSkyon 1 sec
06:12:00AzureSkywhat value did you want, it shows it there but cant access it via file browser
06:13:02 Join hebz0rl [0] (
06:14:30funman8gb card, right?
06:14:47funmanthen the card works but it's not mounted
06:15:19funman <- perhaps the same problem
06:15:21AzureSkyand if i load that rockbox.sansa from files list it appears and is accessable
06:16:23 Quit CaptainKwel (Quit: Ex-Chat)
06:16:28funmanyou say r26906 works for sure?
06:17:15AzureSkyfor somebody else, ican tell you that for me it probbly worked(till i updated to a build from today it worked for me) for me build r26865 works if i downgrade to taht
06:17:39funmancan't spot anything obvious in the log
06:19:01AzureSkyr26926 im pretty sure worked as well (im almost serten i loaded that one and it worked)
06:20:26funmanhm perhaps r26933 or r26937
06:22:12AzureSkyif you have a copy of earlier builds i can test them, i would build my own but 1. no tools to do it, 2. no skills to do it
06:26:08 Quit Topy44 (Ping timeout: 260 seconds)
06:29:21AzureSkyinstalling cygwin
06:36:21 Join Topy44 [0] (
06:38:39funmanstorage is much faster on fuzev2 than on fuzev1 despite the lower pclk
06:39:06AzureSkyduno, never had a v1, all the fuzes i have had and gotten people have been v2
06:39:15funmanhm however i'm measuring speed with µSD on fuzev1 - but iiuc it should be faster than internal
06:40:00AzureSkyso, once this finely gets setup what files should i download to start with?
06:40:10funmanAzureSky: the wiki has all the info
06:40:43 Quit Topy44 (Ping timeout: 260 seconds)
06:42:10AzureSkyso you have a couple fuzes i take it
06:43:04 Join Topy44 [0] (
06:55:12CIA-85New commit by 03funman (r26953): sd-as3525*: handle aligned transfers without memcpy() ...
06:56:38 Part AzureSky
06:57:06CIA-85r26953 build result: All green
07:02:49 Join Guest14828 [0] (
07:10:06 Quit binaryhermit (Quit: Leaving)
07:13:24 Quit Topy44 (Ping timeout: 260 seconds)
07:14:10Guest14828906 works lestatar sent me a backup of his working install and it works fine
07:15:14funmanif r26906 works and not r26951 there's not much commits to test : and i pointed 2 which affect this player particularly
07:16:01funmangrr.. i only tested storage diff on fuzes, and it seems to fail on clipv1
07:16:26Guest14828yeah, accdently closed my irc window, and im still waiting on cygwin to fully install
07:16:49 Join Horschti [0] (~Horscht2@xbmc/user/horscht)
07:17:42 Join Topy44 [0] (
07:19:47 Quit Horscht (Ping timeout: 265 seconds)
07:22:27funmanplaying music crashes on all clips
07:23:48 Nick Guest14828 is now known as AzureSky (
07:24:24funmanhm and on fuzev2 too .. ?
07:24:25AzureSkyouch, well im letting cgywin download the 33 build
07:24:49AzureSkyno, havent tested on fuze, all the files are slowly downloading now tho
07:25:01AzureSkydespite my net being pretty fast :P
07:36:41funmangcc4.5 -flto seems not to like -mthumb
07:37:27AzureSkythat folder type ../tools/configure which will bring up a list of all possible devices
07:37:39AzureSkyjust gives me an error
07:37:45AzureSkynu such file or directory
07:38:37funmanyou built the cross compiler already?
07:39:34AzureSkyjust did what the steps to compile said to do, downloaded form the svn already......
07:40:12AzureSkyok i see, will work on that now
07:40:49 Quit Dark_Rak3r (Ping timeout: 264 seconds)
07:41:42 Join dfdf [0] (
07:43:26dfdfHi i got a problem detecting a particular micro sd card after updating rockbox
07:43:30dfdfanyone knows what to do?
07:43:31simonrvnr26945 WorksForMe
07:43:37simonrvnfunman: ^^
07:44:06dfdfthe microsd card works in OF though and other micro sd cards work with rockbox
07:44:12simonrvnplays, sees the microsd, etc.
07:44:28AzureSkydfdf, same problem here working on a fix
07:44:30simonrvnmaybe it needs to be fscked dfdf ?
07:44:38AzureSkywell tracking down where problem came from
07:44:42funmanit took me some boots to notice the problem: card is present in disk_info but not mounted
07:44:58dfdfi did a disk check but still the same
07:45:13funmandfdf: if you go in debug -> disk info -> microSD 1 the card info is listed?
07:45:13dfdfso the problem lies with the recent release?
07:45:23simonrvnhmm. that rev, 26945, it did for me.
07:45:26dfdfdebug in the player?
07:45:31funmandfdf: you didn't tell which player you have btw
07:45:36dfdfoh sorry
07:45:45funmanyes in the player (debug menu)
07:45:50dfdfver 26946
07:46:20dfdfi don't see any microsd in diskinfo
07:46:29dfdfoh wait
07:46:42dfdfthe title is microSD 0
07:46:45simonrvnbootloader i have on it is r26845
07:50:48 Quit robin0800 (Remote host closed the connection)
07:51:18dfdfthe working microsd gives abt the same thing in debug>diskinfo
07:51:33funmanpress right to see microSD 1
07:53:04 Quit Topy44 (Ping timeout: 260 seconds)
07:54:29dfdfoh right
07:54:32dfdfyeah its there
07:55:21CIA-85New commit by 03funman (r26954): fix r26953: use physical address for DMA buffer, also for IRAM ...
07:55:43 Join Topy44 [0] (
07:56:58CIA-85r26954 build result: All green
08:01:34funmangcc 4.5 -flto reduces ram usage by 18.2kB on clipv2, but i cheated a bit : the compiler adds something in ARM.exidx/ARM.extab sections, which is apparently exception tables (for C++?)
08:05:10 Join Buschel [0] (
08:06:16***Saving seen data "./dancer.seen"
08:09:59 Quit Buschel (Ping timeout: 260 seconds)
08:11:23CIA-85New commit by 03funman (r26955): make clean: delete lib/ directory
08:12:51CIA-85r26955 build result: All green
08:16:12saratogano one has figured out the dock pins on the Fuzev2, correct?
08:16:19saratogaerr dock detect pin that is
08:17:16funmansaratoga: i'm not aware if anyone did it
08:17:55funman/media/bordel/rockbox/apps/codecs/mpa.c:422:30: warning: comparison between ‘enum mad_error’ and ‘enum <anonymous>’
08:18:10funmanthe good thing with using different gcc versions is they show you new warnings ;)
08:23:07saratogasince both of those enums have the values explicitly set as unsigned integers, its a little silly
08:24:19AzureSkyfunman, after following all the directions i new got a bunch of command not found errors......think i need a copy of a working ect/profile file
08:24:42saratogais the command gcc?
08:24:43funmanAzureSky: just set your PATH to include the path to cross compilers
08:28:25AzureSkythats what is in there as the guide says....
08:34:59 Quit xavieran (Read error: Operation timed out)
08:41:50funmanah we patch gcc for no exceptions & multilib support
08:42:04 Join stoffel [0] (
08:48:34 Join xavieran [0] (
08:49:33 Join petur [0] (~petur@rockbox/developer/petur)
08:56:24 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
08:59:30 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
08:59:47funman24kB down on clipv2 with gcc 4.5 flto + binutils cvs
09:00:15funmanapparently flto generates buggy thumb asm
09:00:52funmanlinks fine without the option
09:01:14funmanbut clipv2 doesn't boot ;)
09:03:38topikfunman: my clip+ refuses to play mp3's from internal memory using r26953-100619
09:03:50topikdata abort at 3000AEE8
09:03:52funmanclipv1 neither: black screen, and clipv2 stuck at logo
09:03:56funmantopik: svn up
09:04:09topikr26954 fixes it ?
09:04:20funmanyes, the log even tells so ;)
09:04:40topikhow unlucky for me to pick the one build before then :)
09:05:22funmanout of 5 targets, i only had one which didn't show this bug, and i tested playback only on that one
09:05:52topikmurphy chuckles no doubt
09:09:20 Quit dfdf ()
09:10:17topikplayback restored. lovely
09:16:46 Join Dark_Rak3r [0] (
09:20:08 Join bimbel [0] (~Miranda@unaffiliated/bmbl)
09:20:37 Quit bmbl (Ping timeout: 258 seconds)
09:37:46 Join Kitr88 [0] (
09:39:46 Quit Kitar|st (Ping timeout: 240 seconds)
09:42:16 Quit Topy44 (Remote host closed the connection)
09:42:41 Quit Kitr88 (Ping timeout: 276 seconds)
09:47:46 Join Kitar|st [0] (
09:52:53JdGordonfunman: why the slappage?
09:53:40funmanJdGordon: you removed SANSA_FUZE_PAD from radio code in a recent-ish commit
09:55:21funmantoo late, i slapped you already :)
09:58:39 Join Rob2222 [0] (
10:00:27 Quit kramer3d_ (Ping timeout: 276 seconds)
10:02:16 Quit Rob2223 (Ping timeout: 240 seconds)
10:04:44JdGordonbieber: I havnt checked tag_table recently but it should be 5 compulsary ints... x,y,width,height,font... colours are in a seperate tag now
10:05:58 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
10:06:20***Saving seen data "./dancer.seen"
10:06:30 Join DataGhost [0] (~dataghost@unaffiliated/dataghost)
10:06:33CIA-85New commit by 03funman (r26956): minesweeper: bugfixes / code improvements ...
10:08:11funman <- new port request, but with money promised
10:08:11CIA-85r26956 build result: All green
10:09:17AzureSkywhats with people and their addiciton to apple products then bitching about the restrictions apple products have?
10:15:48funmanAzureSky: i don't know but this is probably off-topic here, better go to #rockbox-community
10:16:24AzureSkyok funman, did a compile with help of a friend
10:16:56AzureSkybut not seeing a folder that has same contents as the .rockbox folder that comes from the download pages
10:17:23funmanyou can run 'make zip' and unzip the created on your player
10:21:34AzureSkygot it, testing 33 now
10:23:38AzureSky33 works fine!!!
10:23:43funmanyou can try also make PREFIX=F: install but i'm not sure if that works on cygwin
10:24:07funmanAzureSky: ok, next possible commit is r26937$
10:24:18AzureSkyyeah gotta svn that
10:25:25funmanis it ok to break bookmark file format ?
10:26:21JdGordonbugger, themeeditor doesnt compile in svn :(
10:28:04simonrvnit errors out
10:28:55funmanS_a_i_n_t: can you check FS #11404 on nano2g ?
10:29:01amiconnfunman: Hmm, since when does 'make PREFIX=... ' exist?
10:29:31pixelmafunman: probably not that nice, especially without warning (re. bookmark format)
10:29:35amiconnBtw, on cygwin the equvalent would be (if the target is drive F:) make PREFIX=/cygdrive/f
10:30:03pixelmafunman: also, what for?
10:30:11funmanpixelma: the format is broken : FS #11388
10:32:05 Join ender` [0] (
10:32:35pixelmaafaik ; is not forbidden in FAT32, otherwise he couldn't have a folder name that way. The others you mentioned are and you can't have these on Rockbox targets anyway
10:32:59pixelmasorry, misread
10:33:45funmani took the list of forbidden chars from VLC, i suppose it's valid
10:34:10pixelmafunman: but even if you fix this, a warning for a few days would be appropriate IMO.
10:35:08JdGordonwhich folder how?
10:35:12funmannot sure if i'll do anything else about this problem
10:35:32funmani'll let someone else make the decision
10:36:27funmanamiconn: now we just need a make bininstall so i don't copy rockbox.XXX in the player's root instead of .rockbox/ and wonder why my changes have no effect
10:36:52amiconnThat doesn't answer my question
10:37:19funmanwell, that's right.
10:37:52amiconnIf make PREFIX= just means it will use the target as a build root, it won't do any good
10:43:50CIA-85New commit by 03funman (r26957): don't parse features.txt when not needed
10:43:55CIA-85New commit by 03funman (r26958): new bininstall rule to install rockbox binary only
10:45:23CIA-85r26957 build result: All green
10:46:19CIA-85New commit by 03funman (r26959): make bininstall rule depend on the binary
10:46:59CIA-85r26958 build result: All green
10:48:29 Join DerPapst [0] (
10:48:36CIA-85r26959 build result: All green
10:52:13funmani wonder about this forum post : our rules say 'no requests' but they don't talk about paid offers
10:52:15 Quit GeekShadow (Quit: The cake is a lie !)
10:53:14funmanperhaps kugel can make some extra $ from this
10:55:27 Join Jaykay [0] (
10:59:48 Join Luca_S [0] (
11:05:29 Join piggz [0] (~piggz@
11:07:21 Join mitk [0] (
11:07:23Luca_Suhm... the as3525v2 set_cpu_frequency patch does no longer apply cleanly
11:08:01funmanLuca_S: you're still using it for all this time and without crashes?
11:11:45funmani just synced the patch, don't hesitate to add your experience on the task
11:12:10 Quit Horschti (Quit: Verlassend)
11:12:11Luca_Sthank you, i'll test it over the weekend
11:15:28Stummibluebrother: ping
11:16:18Stummihav you already tried to compile the arm-eabi-compiler for darwin?
11:16:20 Join Horscht [0] (~Horscht2@xbmc/user/horscht)
11:19:12funmanStummi: it works on 10.5.8 but not on 10.6 apparently
11:19:21 Quit funman (Quit: free(random());)
11:22:49Luca_Sjust to check: the debug > view hw info menu is not updated to reflect the changes in PLL setup?
11:23:15 Join flydutch [0] (
11:24:28Luca_Suhm.... jewels crashes on startup
11:24:41Luca_Slatest svn + cpu frequency
11:25:04Luca_Si'll try with plain svn
11:26:25AzureSkyfunman, confirmed its 37 that kills it
11:27:47 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
11:30:17Luca_Smmm jewels works fine with plain svn
11:31:05 Join thefirstM [0] (
11:31:55bertrikLuca_S, we did increase the default CPU speed a bit last night, from 240 MHz to 248 MHz (now it's exactly the same as as3525v1 and gives better playback rate error)
11:31:59AzureSkymicrosd dosnt work with 26937 and up without a live reload of the firmware(rockbox.sansa)
11:32:56AzureSkyand thats with each boot
11:33:48Luca_Sbertrik, it crashed yesterday too, i filed a bug on fs but since it wasn't the only patch applied it wasn't very useful - today's svn works fine, but with FS 11297 it panics (data abort) every time I start jewels. Also pictureflow just freezes.
11:33:50AzureSkyim dropping back to 33 now
11:34:24Luca_Smaybe this could shed some light on what's wrong with the cpu freq patch?
11:35:36AzureSkyjust so you know, (funman and the rest of you), had to get a buddy to remote in to get cygwin working properly but It seems its the changes in 26927 that are causing the microsdhc cards to not auto mount/detect :(
11:40:36bertrikSHDC still works fine for me on my clip+ with svn 26959. SVN 26927 was a change in USB, I don't know how this would affect microsdhc
11:41:54AzureSkythe card dosnt auto mount and wont mount after boot on the fuzev2
11:42:35AzureSkyu can reload firmware by "playing" the rockbox.sansa after boot and the microsdhc card works after that but thats a bit of a pain.....
11:43:29Luca_Suhm funman, I think I spotted an error in the cpu freq patch
11:44:17AzureSkythat could be the problem with my fuze and its microsdhc card detection couldnt it?
11:44:20Luca_SFCLK_FREQ is set to DRAM_FREQ / 1 in clock_target.h, i.e. 41333334, but in the patch it's tested against 40000000
11:45:15Luca_SAzureSky: I'm not a RB developer, I just happen to be able to read C and I'm trying to solve a crash with my player - not uSD related, sorry
11:45:23 Join chrissavery [0] (
11:45:39AzureSkyi never tested the games with later builds :/
11:46:20AzureSky26933 works fine, 26937 dosnt :(
11:46:43 Join Highlander [0] (
11:48:45chrissaveryhi. just checked out r26959 and it won't build. it spits out some msg about "arm-elf-eabi-gcc: not found". I guess this is some compiler switch over but there is nothing about it in the dev docs. I tried to rebuild using but it spits out a msg about needing "makeinfo". So what to do to convert working old dev environment to new one that works?
11:49:55bertrikinstall the texinfo package for makeinfo (at least that what it's called on ubuntu), run again and select 'e' (for eabi)
11:50:05AzureSkyu need to re-run the setup for cygwin and get the arm-elf-eabi-gcc off the rockbox server(we had to)
11:50:36chrissaverywill try texinfo then. thanks. I did try makeinfo package but one doesn't exist.
11:52:01chrissaveryappears to be working so far. thanks again.
11:52:21bertrikJdGordon, how hard would it be to add a radio signal strength indicator to the fms?
11:53:08bertrikSomething similar to the volume icon in the wps, but showing something like an antenna with a couple of bars next to it
11:55:40JdGordon|not hard at all
11:57:48bertrikThere is already a patch that adds a radio property for the signal strength for nearly all tuner chips. I guess we would just need a tag for it to retrieve this property and display a number or select a bitmap from a strip.
12:00:12TheSevenchrissavery: also, please remove the old toolchains before re-running
12:00:41 Quit mitk (Quit: Leaving)
12:00:45TheSevenhaving those in the PATH during compilation of the new toolchains will make things go crazy afterwards
12:01:14chrissaveryah, didn't do that. which directories need deleting for that?
12:02:13TheSevenfind /usr/local|grep arm-elf
12:03:57TheSeven(and also for the other architectures of course, in case you have those toolchains installed)
12:04:06 Quit piggz (Remote host closed the connection)
12:04:08topikis there a list of which targets have moved to eabi so far ?
12:04:20 Join mitk [0] (
12:04:25TheSeventopik: have a look at /www/buildserver/builds
12:04:56TheSevenchrissavery: that's arm-elf, arm-elf-eabi (will be caugt anyways), sh-elf, m68k-elf, mipsel-elf
12:06:03JdGordon|bertrik: as ling as there is a nice way to get a percrntage value (or some min/max) it is about 5 lines for the skin code
12:06:21***Saving seen data "./dancer.seen"
12:06:24JdGordon|im a bit hesitant to add skin stuff now though
12:06:51chrissaveryTheSeven: ok. cleared that stuff out. not sure i got it in time as it had already started compiling the new stuff. will soon see. thanks.
12:07:04topikthanks TheSeven
12:07:20*TheSeven suggests killing and re-starting just to make sure
12:07:55 Join Buschel [0] (
12:08:18TheSevenwe don't exactly know what's going on, but I'd suspect it already happens during ./configure of the new gcc/binutils
12:08:59chrissaveryrighto. will do.
12:12:42 Quit stoffel (Ping timeout: 265 seconds)
12:14:39bertrikJdGordon, the patch for RSSI (received signal strength indication) in the tuner drivers currently return dBuV (logaritmic scale based on microvolts), not a percentage, and there's no hard defined min or max either
12:16:45bertrikwe could change that of course, or maybe just assume a fixed range that works with most tuners
12:22:06S_a_i_n_tA little late...but yes, I can report that Nano1g is fine with eabi
12:22:14S_a_i_n_t(funman asked about this earlier)
12:23:59bertrikJdGordon, where are the fms tags handled?
12:26:02bertrikoh I see, in skin_tokens.c
12:26:17 Join dfkt [0] (dfkt@unaffiliated/dfkt)
12:26:49 Join stripwax [0] (
12:28:09 Join halmi [0] (
12:29:48JdGordon|bertrik: yeah. to make it work in a bar you meed to look at draw_progressbar in skin display.c
12:30:08JdGordon|ypu nees a secind tag type for thw bar
12:31:30S_a_i_n_tErrrmmm...if someone pointed me to what I needed to change, or was nice enough to make up a diff, I'd be happy to confirm/deny FS #11404
12:31:46*S_a_i_n_t apologises for being awol all day.
12:32:30mitkclip+ has no AA in features but has PictureFlow plugin enabled. Is it bug or by intention? PictureFlow flickers and freezes occasionally on clip+.
12:33:14S_a_i_n_tThat is intention IIUC
12:34:30S_a_i_n_tI think AA would look terrible on the clips if it was small enough to keep the rest of the WPS info oncreen also.
12:34:49S_a_i_n_tIt doesn't look great in pictureflow as is.
12:35:03mitkyeah, but PictureFlow looks terrible also
12:35:35 Quit hebz0rl (Ping timeout: 260 seconds)
12:35:36S_a_i_n_tyeah, but pictureflow is a plugin, not one of the main screens ;)
12:35:47S_a_i_n_tchoosing to use it is one thing, forcing it another.
12:36:02mitkpixelma: I've found Ondio build I made for you. gcc: sh-elf-gcc (GCC) 4.0.3 (rockbox patch #1)
12:39:26mitkS_a_i_n_t: Using PFlow on a clip* is some kind of desperacy I think.
12:40:18S_a_i_n_tthere are a couple of patches that make it vastly better IMO, which should possibly be considered for inclucion into SVN
12:40:28*S_a_i_n_t looks.
12:40:39chrissaveryFS #11380
12:40:51S_a_i_n_taha...there's the man himself. ;)
12:41:06bertrikS_a_i_n_t, the PCM driver has support for 11, 22, 44, 88 kHz. I don't know what exactly is meant with "do not work" in FS #11404
12:41:35chrissaverythat was off the top of my head so hope it's right - patch for Clip+ PF improvements.
12:42:24S_a_i_n_tbertrik: Near as I can tell, apparently sound is vastly improved by forcing 44kHz, but I have no idea what to modify to confirm this.
12:42:39S_a_i_n_tI can confirm however, that rockboy/doom sound awful.
12:43:02mitkchrissavery: No. It's FS #11380 - Keys not mapped properly in Recording Screen (Sansa Clip+)
12:43:35*chrissavery goes to look
12:44:10dfktfunman changed the recording keymap:;revision=26943
12:44:44chrissaveryah, yes, it's FS #11300
12:45:11chrissaveryhas combined patch for reflection and mono.
12:48:10pixelmacan someone tell me if funman's commit regarding parsing of features.txt coul break the automatic generation of UseOptions for the manuals from it?
12:48:36S_a_i_n_tthat'd be a big bonus for the clips IMO, can you put up a screenshot somewhere chrissavery?
12:48:43S_a_i_n_t(or do you not have the target?)
12:48:43pixelmaI'n only able to find out by testing
12:52:55chrissaveryS_a_i_n_t: I don't have any Clip variants. I just tested in USim and it doesn't even show the same as real. The person who has screen grabs from real clip is VagueRant. He has posted on FS about it and could probably provide a good screen shot. He did post one but just to show a problem.
12:53:27S_a_i_n_tI see.
12:53:43 Join pixelma_ [0] (~pixelma@rockbox/staff/pixelma)
12:55:10 Join M3DLG [0] (
12:56:49pixelma_ok, seems to be working still
13:00:42pixelma_chrissavery: for what it's worth - pictureflow was and is still not able to start playback from itself on the hwcodec players. Since no-one worked on this, the integration doesn't make much sense there. (Not blaming you but the change could have been discussed a bit before committing, I direct that at the committer mainly)
13:00:52chrissaveryTheSeven: looks like my build with new toolchain finished. Thanks again.
13:02:42chrissaverypixelma: sorry, I didn't know about that. What players are hwcodec? I suppose it should exclude the integration chgs for those targets. I only own a Fuzev2.
13:03:17pixelma_all the supported bitmap Archoses
13:10:30 Join stoffel [0] (
13:11:15 Join hebz0rl [0] (
13:15:41 Quit chrissavery (Remote host closed the connection)
13:16:33 Join chrissavery [0] (
13:20:17 Quit xavieran (Ping timeout: 240 seconds)
13:25:58mitkPictureFlow on clip+: r26959 -> and r26959 + FS #11300 pictureflow-reflection-mono-3.patch -> Both are terrible or I'm doing sth wrong.
13:27:53 Join mischasworld [0] (
13:28:45pixelma_mitk: that's not a general picturflow problem (and the ability to run it) but the greylib just doesn't work nicely on the Clips' OLED display - looks ok-ish on other monochrome screens
13:31:49pixelma_the feature is about being able to show album art in the WPS which doesn't use the greylib and hence doesn't work on monochrome targets at all (has been mentioned somewhat already but maybe this info is a bit more in detail)
13:33:18 Quit liar (Quit: ...)
13:33:55pixelma_mitk: I found your info about the Ondio build now (needed to read the logs a bit as I have some trouble with my IRC client currently). Can you give me the exact size of the ajbrec.ajz too? I wonder if I misremembered or if there is more to it
13:35:19mitkpixelma: thanks for explanation. I'll then wait for greylib improvements. Your Ondio build ->
13:36:35 Join MethoS- [0] (~clemens@
13:37:38*S_a_i_n_t suggests screendump for pictureflow screenshots...
13:37:57S_a_i_n_t(assuming the Clips have this feature of course...)
13:39:10pixelma_errr... greylib screenshots off target won't give you any helpful results
13:39:37ender`clip's oled screen has too fast refresh rate for greylib to work?
13:39:59*S_a_i_n_t facepalms.
13:40:03mitkIt is hard to show flickering on jpg
13:40:05S_a_i_n_tYes, thanks for the reminder.
13:40:06pixelma_yeah, well it works but doesn't look good
13:40:33pixelma_it doesn't work well with cheating the eye ;)
13:40:34ender`(and the screen update rate isn't fast enough?)
13:41:39ender`btw, why does clip's screen always look like it's blinking (even with maximum contrast enabled)?
13:42:29 Join fml [0] (
13:42:43 Quit amiconn (Remote host closed the connection)
13:42:43 Quit pixelma (Remote host closed the connection)
13:42:47fmlHello. Why aren't manuals built? Are they broken?
13:42:48 Join ubuntu-nathan [0] (~eeepc904@
13:42:52ubuntu-nathanHi all!
13:42:55gevaertsThe refresh rate isn't too fast (or too slow), it's just that the pixels turn on and off instantaneously, while on LCD they tend to fade in and out
13:42:58 Join pixelma [0] (quassel@rockbox/staff/pixelma)
13:43:02*ubuntu-nathan is back
13:43:02 Join amiconn [0] (quassel@rockbox/developer/amiconn)
13:49:06*fml pokes the build masters and asks them to look at the manual builds (which don't happen)
13:51:59 Join kugek [0] (
13:53:20 Part chrissavery
13:55:21 Quit kugek (Client Quit)
13:55:21 Quit stripwax (Quit:
13:58:15 Join kugel [0] (~kugel@rockbox/developer/kugel)
13:59:42 Quit flydutch (Quit: /* empty */)
14:04:45 Quit MethoS- (Remote host closed the connection)
14:06:23***Saving seen data "./dancer.seen"
14:06:42 Join pamaury [0] (
14:06:42 Quit pamaury (Changing host)
14:06:42 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
14:06:43 Join LionCub [0] (
14:09:25 Quit Buschel (Ping timeout: 272 seconds)
14:10:17 Quit fml (Quit: ChatZilla 0.9.86 [Firefox 3.6.3/20100401080539])
14:12:48 Quit kugel (Read error: Connection reset by peer)
14:14:15pixelmajust wanted to point fml to the mailing list where a similar thing about the voices was mentioned... of course he's gone already
14:16:04bertrikFunny how we have player-specific i2c addresses for some i2c devices. It's actually the i2c controller in the player that is different, not the i2c device.
14:21:14 Quit LionCub (Quit: CGI:IRC (EOF))
14:25:53 Join kugel [0] (~kugel@rockbox/developer/kugel)
14:40:21thefirstMI know this is unsupported, but I am trying to compile the EABI toolchain with GCC 4.5 instead of 4.4. But I get this error: "checking for library containing strerror... configure: error: Link tests are not allowed after GCC_NO_EXECUTABLES." Does anyone know why that happens? I have checked the config.log, but there isn't anything useful in there.
14:51:41JdGordonAlexP: S_a_i_n_t: pixelma: kugel: anyone else who cares: From what I remember from IRC conversations, and the code, and such.... if viewports are used then the default viewport should only work to enable viewports, images and text should not be drawn
14:52:10S_a_i_n_tfwiw, they are.
14:52:48JdGordonthey being what? images or text?
14:53:41kugelI cant remember that
14:53:51JdGordonbertrik: did you figure it out? I'm free now if you want help
14:54:26S_a_i_n_tJdGordon: Both IIRC, images definitely.
14:54:47kugelIt should be allowewd to draw in the default vp, except in sbs
14:54:48bertrikJdGordon, I'm watching soccer now and after that will do some shopping, so no :(
14:55:05S_a_i_n_tkugel: I'm of that opinion also.
14:55:37 Join chrissavery [0] (
14:55:39*JdGordon bases this alot on code comments like "/* don't draw anything into this one */" and skin_display.c linmes 1339-1343
14:55:55S_a_i_n_tBut, it didn;t bother me that I "couldn't" draw in the default viewport, as whatever was supposed to stop it didn't work.
14:56:10S_a_i_n_tBut, now I understand it...I have re-coded all my themes.
14:56:31S_a_i_n_t(to stop breaking in future)
14:57:16 Quit Jaykay (Remote host closed the connection)
14:57:37S_a_i_n_tJdGordon: Have you managed to figure out why drawing in the default viewport with other viewports defined *is* possible?
14:57:40chrissaveryjust updated to r26959 and now it won't recognize SD card slot. it doesn't show up in Files browser even after several restarts, unplugs etc.
14:57:52JdGordonS_a_i_n_t: broken code most likely
14:58:06S_a_i_n_tI too recall conversations about viewports which support your belief that it indeed shouldn;t work.
14:58:25S_a_i_n_tOnly, I thought the change never happened.
14:58:49pixelmaJdGordon: what you just said was what I remember being said in the beginning and I *think* I remember update problems or so if someone used the default vp this way. If you asked me I would always have told that but there are lot of working examples it seems and somehow my impression was that things changed back and forth so I'm not sure anymore (can't pinpoint any commits though)
14:58:55JdGordonin fact I remember being told off by kugel for nearly changing behaviour to outright never go into these always hidden viewports (whihc the default is one) because it will break everything
14:59:23pixelmachrissavery: known bug
14:59:41*S_a_i_n_t seems to remember that discussion also.
14:59:50chrissaveryok. just browsing irc log now.
14:59:53S_a_i_n_t(re: breaking stuff with default viewports)
15:01:12JdGordon$ svn log -r17759
15:01:22JdGordonthat is the initial conditional viewports change
15:01:53pixelmathere are also examples with pverlapping viewports which still seem to work somehow
15:02:04pixelmaoverlapping too
15:02:29S_a_i_n_tyeppers, I actually abuse that behaviour in several themes.
15:02:35kugelDrawing into the default vp always worked, thats why i needed to add a special case for the sbs. Drawing into the sbs default vp means drawing over the lists
15:03:09S_a_i_n_twas overlapping viewports not supposed to work also?
15:03:21 Quit hebz0rl (Ping timeout: 252 seconds)
15:05:16JdGordonz-order is ignored, and viewports arnt cleared except on full redraws or when becoming enabled
15:05:20kugelit wasn't supposed to work. it just did :) we never explicitely disallowed it in the code IIUC
15:05:24JdGordonso overlapping can work if you do it correctly
15:05:37pixelmaoverlapping viewports (shown at the same time) is not supposed to work or at least said to cause update problems too, that's what I remember
15:05:51S_a_i_n_tquite often my viewports only partially overlap...and that seems to work fine.
15:06:15S_a_i_n_tAlso, I often just draw one on top of the other to simulate "clearing" the viewport.
15:06:20S_a_i_n_twhich works fine also.
15:07:20JdGordonmy view is that the default viewport should work the same in all screens, so they should do nothing but enable other viewports
15:07:39S_a_i_n_tpixelma: That may have been the case at one point, but I know I have at least one WPS with overlapping viewports, and each viewport contains scrolling/dynamic content...and display is fine.
15:07:50JdGordonwhich is probably how I'll make the new displayer work
15:07:54S_a_i_n_tI do seem to remember update issues a long while ago though.
15:08:29S_a_i_n_tJdGordon: That will break a lot of themes though...
15:08:35JdGordonmeh :)
15:08:43S_a_i_n_t(but if it has to happen, it has to happen :/)
15:09:02kugelthat complicates making themes unecessarily imo
15:09:03S_a_i_n_tsorry, but I don;t really share the "meh" sentiment :P
15:09:21JdGordonI dont see how
15:10:25kugelyou can't make themes without vps
15:10:45JdGordonyes you can
15:11:26 Quit simabeis (Quit: Lost terminal)
15:11:26S_a_i_n_tas I understand it, it will be "don't draw in any vp", or "draw in any vp except the default"
15:11:33 Join simabeis [0] (
15:11:51JdGordonall this is only relevant when any viewports are used
15:12:02JdGordonif none are then the default acts like normal
15:12:50S_a_i_n_twhat I don;t understand though, is why you *can't* draw in the default viewport if another viewport is enabled.
15:13:13S_a_i_n_tI'm not saying it shouldn;t be that way, I just don;t get why it is/
15:13:18S_a_i_n_ts/is/should be/
15:13:40JdGordonIIRC the reason was because %xl, %x, %V, %vd etc all didnt eat the \n so the text wouldnt be on the right lines
15:14:08S_a_i_n_tall "didn't", ...but now they do?
15:14:14JdGordonalso the comment... "conditionals clear the line which means if the %Vd is put into the default viewport there will be a blank line. To get around this we dont allow any actual drawing to happen in the deault vp if other vp's are defined"
15:15:02S_a_i_n_tbut, that isn't the case apparently.
15:15:33JdGordonwhich means the code fell out of sync with comments...
15:15:58kugelwho added the comments, and when?
15:16:00S_a_i_n_tindeed. so, I still don;t really get why we couldn;t draw in the default viewport.
15:16:10S_a_i_n_t(if its only based on outdated comments)
15:16:17JdGordonkugel: that comment was added in the initial conditional viewports commit
15:16:25JdGordonby me
15:17:24*JdGordon thinks he sees what happened
15:17:29kugelthat comment was wrong in the first place then
15:17:44JdGordonno it isnt
15:17:51JdGordonor wasnt at that time
15:18:17kugeliirc drawing into the default always worked
15:18:35S_a_i_n_tI don't ever remember my code not acting right, and until yesterday it always drew in the default viewport.
15:18:52JdGordonS_a_i_n_t: waay before you showed up :)
15:18:53S_a_i_n_tI only changed it now out of fear of the new parser suddenly not accepting it.
15:19:51*JdGordon trolls through 25000 commit messages
15:20:50kugeli know that at the time sbs was worked on, drawing into the default vp worked
15:20:53S_a_i_n_twell, based on a theme I have lying around, I've been able to draw in the default viewport for at least 8 months.
15:21:02S_a_i_n_t(a piss in the wind though probably)
15:22:01JdGordonkugel: It is very likely a commit changed the conditional behaviour to not blank out those lines which means yes the comment is outdated
15:22:05JdGordonI'm looking for that now
15:22:33JdGordonI dont espeically like special handling for the default viewport in the sbs which will cause slight mess in the new renderer...
15:22:48JdGordonbut if I'm really outvoted and yelled down I'll see how cleanly it can be done
15:23:57kugeli wouldn't object changing behavior iff it makes the code *a lot* cleaner
15:24:17 Quit petur (Quit: stairway to heaven)
15:24:18S_a_i_n_tI'm not into yelling, and it isn't a democracy...but if it was based on a limitation that is no longer present I don't see any reason to not allow drawing in the default viewport.
15:24:47JdGordonkugel: have a look at utils/newparser/skin_render.c .... so damn nice to walk the new skin tree :)
15:25:10JdGordonactually, I guess it will need a "never drawn" flag there also so it wont be terrible
15:25:20JdGordonbut the default viewport in the sbs IS scanned for conditionals...
15:26:10kugelthat's how it's done currently I think, only few loc for the extra handling
15:26:28JdGordonthe default and ui viewports are not the same in the sbs... the default really should have the same behaviour everywhere
15:26:56JdGordonin svn there is a hack to make sure the default is always hidden in the sbs but not in others which i dont like
15:27:32 Join piggz [0] (~piggz@
15:28:59piggzit was working earlier, but now my ipod feels sluggish, and wont play anything
15:29:16kugelwell, it could be made so that it's not a hack (e.g. a parameter for the parser)
15:30:30JdGordonkugel: also you'll be pleased to know that the skin buffer is going back to a much simpler allocation scheme :)
15:30:33*pamaury wonders if rockbox really powers off the clip+.
15:30:40chrissaverypixelma: there doesn't appear to be a bug posted for Fuze not seeing SD cards. I'm reverting to r26906 reported to work but I'm just wondering if this will get attention without a bug posted?
15:32:08 Join efyx [0] (~efyx@
15:33:37kugelJdGordon: cool!
15:34:36bertrikpamaury, what makes you think that?
15:34:39JdGordonbecause of the way it parses now, we should be able to know before loading anything (except the .wps) if there is enough room for the skin
15:35:30pamaurybertrik: I don't know. It has been several times where I leave my clip+ at night, powered off and when I wake up, the battery is nearly empty :( Perhaps the battery is dead but that would be strange
15:36:24bertrikOh weird, I have not noticed anything like that. Didn't a wakeup timer get implemented recently? Maybe that is causing trouble
15:37:53JdGordonOK, I tihnk I might see how drawing in the defalut vp breaks things... it needs a conditional which changes frequently... the comment still seems to be valid
15:46:11*S_a_i_n_t was almost in bed ;)
15:46:16S_a_i_n_t...what does it break?
15:48:21JdGordonthe line will get cleared
15:48:27JdGordonwhich may be visible or may not be
15:48:34JdGordondepending on how the rest of the skin is done
15:49:10JdGordon%?pv<||||||||||||||||> *might* show the effect
15:49:57S_a_i_n_tthis is for text I assume?
15:50:39S_a_i_n_tMost if not all of the themes I've seen which draw in the default vp draw images in the default vp, and text in other vps.
15:51:09S_a_i_n_tI've even come across people in here that thought that viewports were *only* for text :P
15:51:17JdGordonthe point is that the above would tell the renderer that it wants to update the line, which will mean that line will be cleared
15:52:09S_a_i_n_tahhhh. I honestly can't say I have seen any ill effect from drawing in the default viewport.
15:52:31S_a_i_n_teven from other peoples code which (as you may have seen in your time) can be pretty ugly.
15:52:53 Part chrissavery
15:53:05JdGordonok, so... hmm...
15:53:19S_a_i_n_tI'm not saying that there *isn't* any ill effect, but I certainly haven't seen any.
15:54:37 Quit M3DLG (Ping timeout: 260 seconds)
15:56:59 Quit mischasworld (Remote host closed the connection)
16:00:35JdGordonbieber: I've updated SkinRendering to reflect the above convo... bassically the default viewport is only handled differently in the sbs, otherwise it is the same as any other viewport. Not sure how you want to handle it in the editor... I'll probably have to add a flag for this in mine, you can probably ignore it (or just "draw" a fake list in the UI area post "sbs update")
16:01:15bieberOkay, thanks
16:01:20 Quit stoffel (Ping timeout: 265 seconds)
16:01:54JdGordonalso svn doesnt compile...
16:06:24***Saving seen data "./dancer.seen"
16:09:36bieberIt doesn't?
16:09:53 Part ubuntu-nathan
16:10:07bieberOh, wow
16:11:08CIA-85New commit by 03bieber (r26960): Theme Editor: Fixed typo that broke compile
16:12:42CIA-85r26960 build result: All green
16:12:43 Quit kugel (Quit: Bye)
16:13:09 Join kugel [0] (~kugel@rockbox/developer/kugel)
16:14:35JdGordonbieber: cool :) a thin rectangle might be better though (not filled in)
16:14:50bieberYeah, that'll definitely be changed a lot
16:15:01bieberI'm just drawing them solid for now so I can see how it's rendering
16:15:22JdGordonI guess the screen size is static at the moment?
16:15:39JdGordoncan we get a drop down to choose the dimensions?
16:16:05bieberIf there's no project open, it just defaults to 300x200. If you open a project with #screenwidth and #screenheight properties or a backdrop image it will resize to fit
16:16:25JdGordonit would be *really* cool to be able to select a viewport in the tree or code and have only that viewport highlighted in the display
16:16:37JdGordonalso clicking inside the display to get the piixel coords
16:16:39bieberBut yes, there'll definitely be controls for just about everything to do with the machine state soon
16:17:16JdGordongreat job so far :)
16:18:29bieberGood idea with the coordinates, btw...I'm thinking maybe let the user switch cursors between click-to-select, click-to-drag, and a crosshairs type view that will show coordinates and snap to objects
16:19:20kugelok, I'm going to commit the s/#ifdef SIMULATOR/#if CONFIG_PLATFORM & PLATFORM_HOSTED/ patch and document the assumptions on RockboxAsAnApplication2010, ok?
16:19:56kugelthe assumptions: RaaA has color bitmap "lcd", no recording, no charcell display
16:20:02JdGordonwhy the 2010 in the page title?
16:20:21kugelbecause there's a page already for the 2008 attempt
16:20:37JdGordonso rename that one?
16:20:48JdGordonRockboxAsAnApplication should be current and up to date
16:20:55JdGordonotherwise it is pretty useless
16:21:01kugelI cannot rename
16:21:05kugelwe can do that I don't care
16:23:37JdGordonwell if it is a worklog then 2010 is fine, but those assumptions are important for when someone wants to change or start a new RaaA target
16:27:37 Quit halmi (Read error: Connection reset by peer)
16:28:11 Join halmi [0] (
16:35:57kugelJdGordon: it's as a worklog yes
16:36:21kugela new page should be added for directions to port the application (or add it to NewPort)
16:38:27 Quit CIA-85 (Ping timeout: 264 seconds)
16:53:05 Join CIA-8 [0] (cia@
16:57:28 Quit piggz (Read error: Connection reset by peer)
17:02:52 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
17:05:38 Quit thefirstM (Remote host closed the connection)
17:09:07 Join panni_ [0] (
17:10:22 Join S_a_i_n_t_ [0] (S_a_i_n_t@
17:11:27 Quit S_a_i_n_t (Ping timeout: 264 seconds)
17:13:43*kugel has a strange issue with the patch
17:13:55 Join dfkt_ [0] (~dfkt@unaffiliated/dfkt)
17:15:03 Join Jerom [0] (~Jerom@
17:17:18pixelmawhat about it?
17:17:44kugelI have a problem with it, it's not created reliably with my patch
17:17:45 Quit dfkt (Ping timeout: 245 seconds)
17:18:12mitkpamaury, bertrik: I have same impression concerning battery drawing in clip+ after rockbox shutdown. I will ckeck this today.
17:24:54 Join Viaken [0] (
17:26:21ViakenHey. I don't have the ability to reply to the Sansa AMS topic, but I just had success with a blue FuzeV2, BH 0909 BMUK 4GB
17:26:37ViakenUsed the 2.03.33 OF.
17:26:54ViakenAnd SVN rbutilqt
17:29:33Viaken If someone wants to report it for me or something, if it matters. :) Thanks for the awesome work, guys.
17:35:06bertrikthis issue with fuzes refusing to get flashed is quite mysterious indeed
17:41:18ViakenHm... I'm having some playback skips. UI showed 2:50 left in that song and it was ending.
17:42:28ViakenThere was a noticable skip at about 0:50.
17:42:41ViakenI'll poke around, see if it's related to a specific option or something.
17:43:04 Join hebz0rl [0] (
17:48:40 Join kramer3d_ [0] (~kramer@unaffiliated/kramer3d)
17:51:43 Join einhirn [0] (
17:51:45 Quit einhirn (Client Quit)
17:52:11 Join einhirn [0] (
17:54:23 Join einhirn_ [0] (
17:58:18 Quit einhirn (Ping timeout: 265 seconds)
18:02:19 Join r0b- [0] (
18:03:28 Quit pixelma_ (Quit: .)
18:03:32 Nick dfkt_ is now known as dfkt (~dfkt@unaffiliated/dfkt)
18:06:28***Saving seen data "./dancer.seen"
18:07:34 Quit halmi (Quit: halmi)
18:07:59 Join dfkt_ [0] (dfkt@unaffiliated/dfkt)
18:09:09 Quit dfkt (Disconnected by services)
18:09:16 Nick dfkt_ is now known as dfkt (dfkt@unaffiliated/dfkt)
18:10:17 Quit t0rc (Quit: Leaving)
18:10:36 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
18:10:40 Quit Jerom (Remote host closed the connection)
18:11:34 Join Jerom [0] (~Jerom@
18:11:50 Part Jerom
18:11:52 Join Jerom [0] (~Jerom@
18:31:46 Quit Jerom (Remote host closed the connection)
18:32:24 Quit BradC (Quit: Tune in next week or I'll come through your set and rip yer bloody arms off)
18:32:42 Join Jerom [0] (~Jerom@
18:34:05 Join flydutch [0] (
18:39:29 Join pamaury_ [0] (
18:43:00 Quit pamaury (Ping timeout: 258 seconds)
18:46:43 Join funman [0] (~fun@rockbox/developer/funman)
18:49:07funmanpixelma: my commit shouldn't break anything; the features were not needed in these rules, i only left them for voices; manual makefile was not touched by this commit
18:50:06kugelfunman: ACTION_FM_QUICKSCREEN should have been used, I guess it also got lost in the fms commit
18:50:33funmani didn't check the history
18:51:58funmanalso i wondered what a FM quickscreen would look like
18:53:44kugelfunman: like the normal quickscreen
18:54:24kugelI added quickscreen access from the radio screen not too long ago but I also recently noticed it doesn't work anymore
18:55:47 Join stoffel [0] (
18:57:51pixelmawhat settings would you change in the radio from a quickscreen?
18:58:05kugelfunman: the bininstall doesn't work for the sim
18:58:43bertrikpixelma, wondered about that too. repeat, shuffle ? :P
18:58:50funmanclip+ with PLLA@384MHz doesn't work fine
18:59:15kugelpixelma: whatever you have in the quickscreen
18:59:32*kugel doesn't understand why the "where you come from" matters at all
19:00:39pixelmawell, because you usually change settings when you need to change them, when listening to the radio I would never think "oh, forgot to change back my repeat settings"
19:01:00S_a_i_n_t_funman: Now you're here...(and I am as well) 1: Yes, I can confirm that there is no undesirable effect running Nano 1g with arm-elf-eabi. 2: I would be happy to either confirm/deny FS #11404 if I were pointed to the exact feilds I needed to edit in the source to force 44kHz sound for plugins or if some kind soul could make a diff to that effect.
19:01:08kugelwhat if you have settings in the quickscreen that make sense everywhere, like backlight brightness?
19:01:33funmanS_a_i_n_t_: just run them unmodified or play a 22.05kHz mp3 file
19:02:01kugelyou could add some radio specific settings to the quickscreen too
19:02:14pixelmathe things I could imagine wanting to change easily in the radio would be preset mode or force mono which are easily accessed sometimes directly or through the radio context menu
19:02:17funmanbertrik: did you test playback on clip+ with PLLB@384MHz ?
19:02:23bertrikno not yet
19:02:35S_a_i_n_t_"run them unmodified"? sound in doom/rockboy for example is awful, but I thought this was known?
19:02:43funmanno it works fine for me
19:03:22pixelmaI don't, so it's hard to imagine (besides I almost never use the quickscreen, usually just because entering by accident and then trying to leave as soon as possible)
19:03:43funmanthe driver should support 11.025 22.05 44.1 and 88.2 kHz
19:04:35S_a_i_n_t_Well, then yes...rockboy is possibly the best example of this. the sound on Nano2g is horrible enough to force someone to turn the sound off or violently snatch the IEMs out of ones ears.
19:05:01S_a_i_n_t_(IMO), but I thought this was well known.
19:05:11bertrikWhat's so horrible about it? is it choppy? noisy?
19:05:43S_a_i_n_t_it sounds like it's being played through an internal PC speaker...from the 1980s
19:06:14S_a_i_n_t_IIRC, Nano1g isn't any better though.
19:06:15funmancan you record it and share the recording?
19:06:31S_a_i_n_t_unfortunately, no, I can't. :/
19:06:46kugelS_a_i_n_t_: that's how the gameboy sounded back then
19:07:12kugelin fact, the first game buy was from the 80s and only had a small speaker
19:07:26funmanffmpeg -i file.mp3 -r 22050 out.mp3 # should give you a 22.05kHz file
19:07:31S_a_i_n_t_kugel: No, it's worse than sounds fragmented/choppy/scratchy
19:07:46kugelprobably because of frame skipping
19:07:50S_a_i_n_t_it sounds...sick.
19:08:42S_a_i_n_t_And the 2g can handle the lowest frameskip quite well, it doesn't seem to effect the sound.
19:08:44funmankugel: what's wrong with bininstall in sim ?
19:08:59kugelthe binary isn't in .rockbox
19:09:54funmanso the rule shouldn't exist?
19:10:19kugelthe binary is in the build dir so nobody would run it anyway
19:10:51CIA-8New commit by 03funman (r26961): bininstall: works for PREFIX directories with spaces
19:12:28CIA-8r26961 build result: All green
19:13:07*kugel finds the all green message annoying
19:14:38funman384MHz PLLB for playback works fine
19:18:26bertrikfunman, sorry, I somehow misread that you were talking about PLLA
19:20:25funmanour macro for calculating fclk divider overflows with 384MHz
19:21:24funman(*(volatile unsigned long *)(0xC80F0000 + 0x10)) = ((((((384000000*(8-0)/8) + 192000000 - 1) / 192000000) - 1) << 4)
19:23:02funmanfclk@192MHz boots fine
19:25:31funman"integer overflow in expression"
19:27:48funmansupposedly 120MHz fclk boots fine: PLLA@384MHz, fclk prediv @ 5/8
19:36:49 Quit kugel (Ping timeout: 240 seconds)
19:36:56 Quit hebz0rl (Ping timeout: 245 seconds)
19:39:38funmanbertrik: apparently changed pclk causes problems on fuzev2
19:40:34funmanand my test shows that predivider is correctly applied to cpu freq
19:45:33bertrikfunman, you got 120MHz instead of 240MHz with PLLA@384MHz and prediv of 5/8?
19:45:54funmanno, i got 120MHz with PLLA@384MHz, prediv 5/8, postdiv 2
19:46:04 Join binaryhermit [0] (
19:46:06 Quit Strife89 (Ping timeout: 260 seconds)
19:46:15funmanif i try to get 240MHz with postdiv 1 the clip+ doesn't boot and crashes randomly
19:46:53funmani'll try with PLLA@384MHz, prediv=1 (8/8), postdiv=2 => fclk@192MHz
19:47:54funmanworks fine
19:48:23funman(all tests with pclk@24MHz on clip+)
19:48:31 Nick detaos_ is now known as detaos (
19:50:06funmanPLL@384 FCLK@240 PCLK@64 = same problems
19:51:00funmanAzureSky: could you try to modify clock-target.h to use 62MHz PCLK on fuzev2? (instead of 41333333...334
19:52:24funmana bit more battery life on clipv1 perhaps: 8h18, i need to test if it's my patch
19:52:30 Quit funman (Quit: free(random());)
19:52:52bertrikI'm not completely sure how this prediv works. I can imagine it just passes 5 ticks and then skips 3 ticks. The 5 ticks could be fast ticks at 384 MHz.
19:57:43*TheSeven has a bit of time during the next weeks
19:58:07TheSevenI'm still wondering what to do about the nano2g bootloader/FTL issue
19:58:49S_a_i_n_t_yelling at it didn;t work then obviously?
19:59:23pamaury_TheSeven: what is the problem ?
19:59:26 Nick pamaury_ is now known as pamaury (
19:59:42 Quit pamaury (Changing host)
19:59:42 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
19:59:48 Nick S_a_i_n_t_ is now known as S_a_i_n_t (S_a_i_n_t@
20:00:33 Quit einhirn_ (Read error: Connection reset by peer)
20:00:37TheSevenpamaury: things are just broken
20:01:19pamaurybasically everyone uses the iLoader no ?
20:01:42S_a_i_n_tTheSeven: You seem to hit more issues than I do it seems?
20:02:07TheSevenI, personally, am not hitting any issues (I use iLoader)
20:02:25S_a_i_n_tMy 2G Nanos (using RB bootloader) have been solid for a while now.
20:02:30TheSevenhowever, other people keep having trouble with the iloader installation/uninstallation
20:03:07S_a_i_n_tI haven't seen any of the really annoying issues the 2G used to have in a loooong time now.
20:03:12*S_a_i_n_t touches wood.
20:03:15TheSevenand iLoader is intended to be a very flexible solution for power users, not the default way of booting rockbox
20:03:39 Join CaptainKwel [0] (
20:04:04TheSevenS_a_i_n_t: Still, the possibility of storage corruption is blocking the target's promotion to "stable" IMO
20:05:00 Join Strife89 [0] (
20:05:31S_a_i_n_tBecause of the sturdyness it seems to have lately, I have even started to give the 2G preference over the 1G...but, I do see what you mean. Although I have been updating it regularly, and transferring large ammounts of data without a single hiccup sine I don't know when.
20:06:09TheSevenSadly I had some iloader-unrelated hiccups not too long ago
20:06:26TheSevenand i have no idea what was going on
20:06:32***Saving seen data "./dancer.seen"
20:06:34TheSeven(and can't reproduce it either)
20:07:02S_a_i_n_tMy 2G (one in particular) used to WSOD daily, now...seems solid.
20:07:24S_a_i_n_tbut I still don;t really trust it.
20:07:36 Quit mitk (Quit: Leaving)
20:07:53TheSeventhose wsods aren't fixed, just hidden (most of them were stkovs which were "fixed" by increasing the stack size)
20:08:12TheSevenIMHO the storage/file system layer needs some serious work
20:08:18S_a_i_n_tAh, yes...I keep forgetting about that.
20:08:22TheSevenstill way too many buffers all over the place
20:08:43pamauryThere aren't that many buffers
20:08:57TheSevenIMO there should never be *any* of those on the stack
20:09:00S_a_i_n_tI've noticed that it takes a lot longer for the 2G to start now.
20:09:05S_a_i_n_t(using RB bootloader)
20:09:12pamauryAh, on the stack, yes
20:09:27S_a_i_n_tit used to blitz the Nano1G, now it's about a second behind it.
20:09:29pamauryTheSeven: where are those famous buffers on stack ?
20:09:44TheSevenif i knew...
20:09:59TheSevensome of them are certainly in plugins, especially batterybench seems to have trouble with them
20:10:21TheSevenand I have the impression that the FAT code will sometimes use stack buffers
20:10:44pamauryI thought I had cured the FAT code from this illness but perhaps I'm wrong
20:11:01pamauryGive a minute to check this assertion
20:11:05TheSeveni haven't checked that recently, at least there were some a few months ago
20:11:33TheSevenlet me check some things
20:11:53pamauryThere is one is fat_mount :(
20:11:55*TheSeven will add a panic on misaligned buffers (this is also causing quite some performance penalty)
20:11:55 Quit stoffel (Read error: Connection reset by peer)
20:12:00pamauryOne sector
20:12:04TheSevenwell, the mount one won't hurt
20:12:11TheSeventhat's only done once in the main thread
20:12:41pamauryYep but I'm pretty sure it could use another one, the remaining buffers are unused at mount time so could be used perhaps
20:13:08pamauryHum, but I guess all the buffers are in the file.c code anyway (or something like that)
20:13:49pamauryThere is one in update_fsinfo
20:14:11pamauryThis one is not great
20:14:31TheSeventhat might well be the one that's causing the stkovs everywhere
20:15:29pamauryYep, it can be invoked on lots of call actually: create/remove dir, close file, recalc free space, ...
20:15:45 Quit flydutch (Quit: /* empty */)
20:16:24pamauryThere is one in write_long_name, not good at all
20:16:47pamauryWell, in fact there are lots of them, but there are not so easy to get rid off
20:18:20pamauryI'll have a look at this. Some of them can probably be removed but I need to seriously think about it, the code is quite tricky already
20:18:24 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
20:19:40 Quit CaptainKwel (Quit: Ex-Chat)
20:19:42*S_a_i_n_t would be happy to test any patches on his Nano2Gs if need be...
20:20:43pamauryWell, I can test the fat code myself (and I'll don't want to be responsible for a massive fs destruction in RB :))
20:21:28pamauryThe problem is always the same, the 2G has a big sector size...
20:23:11TheSeventhat won't be a problem as long as there aren't any buffers where they shouldn't be
20:24:15pamauryindeed :)
20:26:12TheSevenhm, stack dumping code would be helpful
20:26:46pamauryto trace back overflows ?
20:28:05bertrikI thought we already have something like that in the "View OS stacks" debug menu
20:29:02bertrikOr is every thread affected by this?
20:35:24 Join Topy44 [0] (
20:35:31pamauryI think he was more talking about displaying the stack on overflow or something like that. The view OS stacks debug menu only list threads, status and stack use
20:39:07TheSevenactually I'd just like to see where a certain misaligned disk access is coming from
20:40:11 Quit Topy44 (Ping timeout: 260 seconds)
20:40:24pamauryIs this causing issues ?
20:41:37TheSevenit's an easy way of finding most of the on-stack buffers, and misaligned accesses need copying-around in the NAND code
20:42:28TheSevenso we could get rid of a static sector buffer in the NAND driver if we get rid of all misaligned accesses
20:42:49TheSeventhis one is probably the mount one you talked about
20:43:00TheSeven(accessing sector 0 and only requesting that single sector)
20:44:20 Join Topy44 [0] (
20:45:39 Join saratoga__ [0] (463f90ed@gateway/web/freenode/ip.
20:46:10*TheSeven wonders if that access could be done through the FAT cache to avoid the need for that buffer
20:47:14TheSevenoh, that's only for the FAT itself
20:47:24pamauryThe thing is that the FAT code hasn't got a "FAT cache" in itself so it's not that simple but I promise I'll have a look in a few hours. When i'm bored trying to make usb work.
20:48:17TheSevenit actually seems to have one
20:48:36pamauryI'm not sure of its actual usage, except for the FAT itself
20:49:06pamauryBut perhaps you're right
20:49:22pamauryAnyway, at mount time, the cache is unused so it's safe to use it :)
20:49:23TheSevenit's only for the FAT itself, but we can certainly abuse it
20:49:28TheSevenno, it isn't
20:49:37TheSevenat least not if there are multiple volumes
20:49:59TheSevenhowever, we can just lock the mutex, flush page 0 if it is in use, and use that for the boot sector
20:50:00pamauryThen we can probably evict one entry from it
20:50:38TheSeveni'd just keep the mutex locked until we're done, or we'll be running into trouble if someone else tries to flush our buffer
20:50:52pamauryFor sure
21:01:28TheSevenno, it wasn't that access
21:01:52TheSevenbut I have a suspicion what it could have been
21:02:57TheSevenit's the FS type determination
21:03:01TheSevenbut where was that code?
21:03:34pamauryI didn't know we had FS type determination code...
21:03:47TheSevenor rather partition table parsing code
21:03:56TheSeveni.e. looking for the start of the fat partition
21:04:06pamauryah, I think we have this code :)
21:04:18pamauryI woudl go for disk.c but that's a guess
21:05:02pamauryThe disk_init function does that I think, check common/disk.c
21:05:51 Quit Topy44 (Ping timeout: 260 seconds)
21:06:20AlexPJdGordon|: Yep, so if you have no need for viewports then it is perfectly fine to not use them at all :)
21:06:22TheSevenhmm, damn, no buffer around that could be abused
21:07:40pamauryhm, you're right, that's seems quite hard in this case
21:07:43 Join Topy44 [0] (
21:07:59TheSeveni'll use an aligned stack buffer then
21:08:33pamauryHow do you do that ?
21:08:46pamauryDoes gcc enforces that on stack ?
21:09:13TheSevenno, but we can just add STORAGE_ALIGN_MASK bytes to the size and ceil the pointer
21:09:48TheSeventhat may waste up to 16 bytes of stack in this case, but who cares
21:10:14*pamaury feared that TheSeven would say that
21:12:55saratoga__pamaury: you were working on the amsv1 usb right?
21:13:07pamauryI'm working on amsv2 usb
21:13:17pamauryOn my clip+ right now
21:13:21saratoga__ah ok
21:13:30saratoga__was going to ask how similar they are
21:13:42mc2739saratoga__: ranma was working on amsv1 usb
21:14:00bertrikpamaury, do you want someone to review your code? I'm volunteering to have a look and help.
21:14:06pamauryThey are different. the clip+ controller is the same as the nano2g except it's a newer version (2.6 vs 2.2)
21:15:08pamaurybertrik: I'll appreciate some help, I'm must be missing the problem. I'm currently cleaning up the code because it's a mess and I'll commit that soon. Then you are welcom :)
21:15:46Jerompamaury: I'm looking at your code and trying to get a USB_GINTMSK_outepintr but I never get one, do you have any idea why ?
21:16:19saratoga__so they're completely unrelated controllers?
21:16:36pamauryJerom: that has been my problem for two weeks now
21:17:21Jerompamaury: did you remove the usb_disable_global_interrupts(); at the end of handle_usb_reset ?
21:17:49pamauryThe svn code is quite outdated, I changed it a old
21:18:13 Quit saratoga__ (Quit: Page closed)
21:26:16JeromI hope you will get it to work pamaury because Sansa's OF is a pain when handling USB
21:27:32TheSevenhm, STORAGE_ALIGN_DOWN seems to be broken
21:31:59pixelmawhy does "make manual" care that I have not set up arm-elf-eabi-gcc yet? :\
21:37:00TheSevengrr, there are misaligned non-stack bufferes in the fat.c code somewhere
21:37:34TheSevenat least they're somewhere in here:
21:37:34TheSeven.bss 0x0810e638 0x1016c /data/rockbox-trunk/build/nano2-app/firmware/libfirmware.a(fat.o)
21:37:47TheSevenhow can i find out what this consists of?
21:38:58pamauryobjdump on the .o ?
21:39:20TheSevenhm, no easier way? (some more detailed listing)
21:43:22bertrikarm-elf-eabi-addr2line perhaps?
21:43:55bertrikoh, it's in .bss
21:48:40bertrikpamaury, we changed the PLLA frequency to 248 MHz on as3525v2. I hope that's not interfering with your USB work.
21:49:56pamauryAs long as it does not prevent bus reset and enumeration, I guess it's ok. I think my current svn head has old plla though
21:50:29 Quit Battousai (Read error: Connection reset by peer)
21:50:51 Join Battousai [0] (~bryan@gentoo/developer/battousai)
21:55:52AzureSkyfunman sorry i was in bed when you requested i edit that, I have no clue how to edit the clock seettings.....have you done a testbuild with those settings?
21:57:39TheSevendamn, fat_cache_sectors is misaligned
22:06:36***Saving seen data "./dancer.seen"
22:07:48TheSevenpamaury: oh well, misaligned buffers all over the place
22:08:01TheSevenSTORAGE_WANTS_ALIGN seems to have been ignored widely
22:08:16pamauryhehe, time to fix it
22:08:16 Quit Topy44 (Ping timeout: 260 seconds)
22:08:46AzureSkycan somebody give me a little help doing what funman wanted done earlier?(when i was asleep)
22:09:39CIA-8New commit by 03jethead71 (r26962): iPod 3G: Implement wheel acceleration and repeats. Chosen settings feel pretty decent but can be tweaked.
22:10:25 Join Topy44 [0] (
22:11:14CIA-8r26962 build result: All green
22:13:43TheSevenpamaury: I'm running into a problem
22:14:19TheSevenhow can i efficiently make sure that a buffer thats inside three levels of nested structs is aligned properly?
22:17:37 Join Dekkard [0] (
22:20:04Dekkardim having a problem with wps working..
22:23:21AlexPMore details please :)
22:23:39AlexPBut at a guess - the WPS syntax changed between 3.6 and the current builds
22:23:50AlexPI'd guess you are trying to use a wps for one on the other
22:24:32TheSevensector buffers on the usb thread's stack
22:25:41sharphey there. is it possible to write a plugin that scans the FM spectrum and makes a graph based on signal strength?
22:26:02TheSevenpamaury: could the fsinfo updater always steal a sector buffer from the fat cache?
22:26:42saratogasharp: yes, you should be able to tune the radio and measure the power from a plugin
22:28:03sharpsaratoga: cool. i can just reference functions in the radio "driver" from the plugin?
22:28:37saratogayou'll probably have to add them to plugin.h unless theres already some plugin that records from the radio
22:29:37*TheSeven is getting grumpy
22:32:57pamauryTheSeven: inside three levels ? Align each structure and align the field into the structure
22:33:21pamauryTheSeven: what do you mean by always ?
22:33:23TheSevenwhat happens if those structs have an odd size and are in an array?
22:33:40pamauryIf you ask the that the structure be align, it will workj
22:33:51pamauryI think
22:33:57TheSeventhe struct definition itself?
22:34:31TheSevenre fsinfo updating: never allocate the buffer on the stack, steal it from the fat cache instead
22:35:14pamauryOk, five me a few minute to finish have a coffee and commit my usb stuff, so I can work on this buffer problem
22:36:01CIA-8New commit by 03kkurbjun (r26963): iPod Mini 1G - Build with EABI.
22:36:58CIA-8New commit by 03jethead71 (r26964): iPod 3G: Add a small check to ensure backlight never gets stuck off if wheel is inactive for too long (because it limits the number of backlight_on ...
22:37:35CIA-8r26963 build result: All green
22:38:47CIA-8New commit by 03pamaury (r26965): as3525v2-usb: don't disable interrupts on bus reset (that was for debug purpose)
22:38:48saratogafunman: playback on the current build is somewhat broken on the clipv2, lossless files pretty quckly die with codec failure or simply static
22:38:53pamauryI should have commit usb stuff before
22:39:00saratogai'm guessing this is an SD driver issue
22:39:13AzureSkybah, i cant checkout code for a build i need to test(to be 100% sure 37 is the cause of the loss of microsd card support) bah
22:39:19CIA-8r26964 build result: All green
22:39:29Dekkardwow.. when i run like arbox widgets.. i dont get the wps.. just the theme for the rest of the screens.. sorry..
22:39:30CIA-8New commit by 03pamaury (r26966): as3525v2-usb: reorganize thing, don't renable ep0 on enum because it's already done
22:39:34CIA-8New commit by 03pamaury (r26967): as3525v2-usb: rework thing, simplify
22:39:41CIA-8New commit by 03pamaury (r26968): as3525v2-usb: tweaks things but still doesn't work.
22:39:47CIA-8New commit by 03pamaury (r26969): as3525v2-usb: simplify register definitions
22:39:54CIA-8New commit by 03pamaury (r26970): as2525v2-usb: coherence fix
22:39:55saratogaDekkard: you're using an outdated theme, update it
22:39:59CIA-8New commit by 03pamaury (r26971): as3525v2-usb: major code renaming
22:40:06saratogaassuming you got it from the theme site you should just be able to redownload it
22:40:10CIA-8New commit by 03pamaury (r26972): as3525v2-usb: end of massive renaming
22:40:18CIA-8New commit by 03pamaury (r26973): as3525v2-usb: remove useless macro
22:40:19Dekkardok.. just installed with the utility,, on a currnet build
22:40:44pamauryTheSeven: ok, now I can work on fat
22:41:00saratogahow do you commit so quickly in sequence? is that a git thing?
22:41:18CIA-8r26965 build result: All green
22:41:27pamauryI'm a serial-committer ;)
22:41:28AlexPdo them all as local commmits, then push them to svn
22:41:29gevaertsGit is the perfect tool to try to get interesting revision numbers
22:42:32saratogai need to figure out git someday
22:42:50TheSevengevaerts: do you happen to know why a sector buffer might be lingering on the usb stack?
22:42:51pamauryTheSeven: did you modify some thing in the fat code already ?
22:42:52CIA-8r26973 build result: 0 errors, 36 warnings (pamaury committed)
22:43:10TheSevenpamaury: yes, I modified fat_mount to not use an on-stack sector buffer
22:43:17pamauryhuh ? What is this build order ????
22:43:22 Join anewuser [0] (anewuser@unaffiliated/anewuser)
22:43:34pamauryTheSeven: ok, I'll have a look at the other buffers
22:43:43pamauryDid you fix the disk.c one ?
22:43:53saratogais there a make switch to build all the test plugins?
22:43:54CIA-8New commit by 03gevaerts (r26974): Also tell the build system that we now build the iPod Mini 1G with eabi
22:44:13TheSevensaratoga: afaik no
22:44:19TheSevenpamaury: yes, i just aligned that one
22:44:34AlexPsaratoga: In debug
22:44:44saratogaso its in configure?
22:45:13AlexPSorry, I meant advanced, not debug
22:45:19AlexPBut yes, when you configure
22:45:31gevaertsTheSeven: check_disk_present() uses one sector worth of stack space
22:45:40TheSevenwhere is that?
22:45:49*gevaerts checks if that's needed
22:46:48 Quit Dekkard (Quit: Going!)
22:47:29gevaertsThe call during usb_storage_init_connection() could clearly use the transfer buffer, but the one during usb_storage_notify_hotswap() isn't as straightforward
22:48:10pixelmais someone around who gets users' ml mails?
22:48:23pixelmaand can reply to
22:48:26CIA-8New commit by 03kkurbjun (r26975): iPod Nano 1G - Build with EABI.
22:49:08CIA-8New commit by 03pamaury (r26976): as3525v2-usb: fix yellow about unused functions (they can be resurrected from an older revision if needed)
22:49:17*pixelma didn't know kkurbjun had a Mini1G and a Nano1G
22:49:34gevaertspamaury: the build system doesn't do all revisions if it's running behind
22:49:55CIA-8r26975 build result: 0 errors, 36 warnings (kkurbjun committed)
22:50:06 Join kkurbjun [0] (
22:50:13 Quit kkurbjun (Changing host)
22:50:13 Join kkurbjun [0] (~kkurbjun@rockbox/developer/kkurbjun)
22:50:38AzureSkykkurbjun if ur wondering CIA-8r26975 build result: 0 errors, 36 warnings (kkurbjun committed)
22:51:12pamauryTheSeven: you removed the buffer in update_fsinfo ?
22:51:17CIA-8r26976 build result: All green
22:51:18TheSevennot yet
22:51:33gevaertsAzureSky: that information is availaible on a web page that I'm sure he knows about. Also, please use real words. Ur is an ancient city, bot a word
22:51:35saratogahow does test disk work? does it just run forever until it gets an error?
22:51:46TheSevengevaerts: OK, it was definitely check_disk_present that was causing the error
22:51:46pamauryOk, I'll try to remove the one from write_long_name. Perhaps I can use the file's cache
22:52:30pixelmakkurbjun: were those eabi builds tested on target? I'm wondering because I didn't know you had these two
22:52:33CIA-8New commit by 03gevaerts (r26977): Also tell the build system that we now build the iPod Nano 1G with eabi
22:52:41AzureSkynice getting on my arse about spelling and you cant even do it your self.
22:52:59kkurbjunpixelma: yes, they were tested on target - I have both
22:53:06gevaertsAzureSky: there's a slight difference between a typo and a deliberate decision not to care
22:53:21AlexPAzureSky: The guidelines do specify real words - this is for people using screen readers and non-native speakers. Please respect them.
22:54:00jhMikeSgevearts: though it used to show all the revisions included in a particular build when clicking the rxxxxx link
22:54:39kkurbjunthanks for the build system updates gevaerts
22:54:40gevaertshm, yes, I seem to remember that vaguely
22:55:01pixelmakkurbjun: ok, good to know, also for the next time testers are needed (especially the Mini1G as it is more rare)
22:56:02gevaertsThree targets left: ipod 4g, d2 and vibe500.
22:56:04kkurbjunI thought there was a wiki somewhere with that information.. I'll have to update it if its not there
22:56:51amiconngevaerts: domonoky has a G4
22:57:05amiconnThat's why I iwas summoning him... didn't work so far
22:58:17TheSevendamn, apparently there's another sector buffer on the usb stack
22:58:28TheSevenat least it stkov's if I remove the additional space for it
22:59:40*jhMikeS wants all ARM to get eabi'd ASAP so he can remove some long call constructs from thread asm code
23:00:41TheSevendamn, it's the one I just aligned
23:00:42gevaertsb0hoon hasn't been online since a while now
23:00:49TheSeventhanks for the misleading comment in usb.c
23:01:08*jhMikeS sees all PP getting eabi'd will be sufficient for that
23:01:48gevaertsjhMikeS: two of the three remaining ones are PP, so that doesn't help you much :)
23:04:09TheSevenwhat do we consider better (or rather less ugly)? having another static sector buffer just for reading the partition table, or increasing the USB thread's stack by the size of a sector buffer to allocate it from there?
23:04:10jhMikeShas anything failed yet (besides the pp5002 thread glitch)?
23:04:25gevaertsD2 i2c code
23:05:08*TheSeven still likes the idea of a global sector buffer pool
23:05:17jhMikeSbut that's not PP, only PP makes long calls in the thread support routines (which are dual-core related). it's inconsequential to getting rid of that.
23:05:24pixelmagevaerts: b0hoon has been active in the forums recently (seems he has a GoGear too)
23:05:29amiconnAlso this weird gradiuent in PF on arm7
23:05:59gevaertsProbably a mail to the ML would be useful
23:06:49*jhMikeS looked at that and can't see readily anything but a gcc bug being the reason (given compiling as arm9tdmicc makes it go away)
23:07:22pixelmasomeone on the users' ml who could also add "daily manuals broken too" to the English voice thread?
23:08:32pixelmaI'd like to get Zagor's attention since he's not around and we are only keeping 3 or 4 dailies now and will soon lose them (tomorrow or the day after)
23:09:04pixelmaand since he answered that thread...
23:09:32*pixelma still thinks 3 or 4 dailies is a bit too few
23:12:07TheSevenupdate_fsinfo, write_long_name, add_dir_entry, update_short_entry, fat_create_dir, free_direntries, fat_rename, ...
23:12:13TheSevenstill loads of on-stack sector buffers
23:12:37*TheSeven starts to like the idea of a global sector cache even more
23:13:54pamaurythe write_long_name one is impossible to get rid off without using the fat cache basically
23:15:56TheSeveni'm thinking of removing all the stack buffers and also the ones from the structs and having one big pool instead
23:16:28TheSeventhat way we can use otherwise unused e.g. file or dir buffers as read cache
23:17:06*TheSeven suggests a fully-associative cache with a LRU replacement strategy
23:17:31 Join M3DLG [0] (
23:19:09TheSevenpamaury: can you think of anything that could cause problems with that approach?
23:20:04pamauryCurrently no, but I'm thinking about it. I'm trying to determine if we really should get rid of all the buffers in the structure or not and how much work that would be
23:20:29TheSevenpamaury: they're effectively dead memory
23:20:44amiconnTheSeven: You'd need very careful locking to do that
23:20:51TheSevenyes, of course
23:22:07TheSevenpamaury: basically we can just allocate a buffer from the cache as soon as the file/dir is opened to replace the one in the struct
23:22:17TheSevenbut as long as the struct is unused, we don't need to waste that memory
23:22:54pamaurythat's true
23:23:13TheSeventhat shouldn't be too much work either
23:23:19pamauryAnyway, the buffer in fat_dir is only used when seeking in the file I think
23:23:39TheSeventhat can of course be optimized then
23:23:58TheSeven(once the rest of it works ;-)
23:24:26pamauryWe need to check outside use of the fat structure (like in file.c, dircache.c)
23:25:14pamauryOk no problem, the sectorcache buffer does not seem to be used
23:25:38pamaurySo we *just* need to write the sector buffer pool :)
23:25:39TheSevenand even if it is, that will just give a compilation error which can be tracked down easily ;-)
23:26:23*TheSeven might have a go at that tomorrorw
23:27:45CIA-8New commit by 03bluebrother (r26978): Add error messages to a few more failure cases to beastpatcher.
23:29:11CIA-8r26978 build result: All green
23:33:41Jerompamaury: I see you removed every USB_ prefix, where you tired of typing it ? ^^
23:35:02pamauryThat was stupid, the header is not used by any other file so it's obviously usb related :)
23:35:35pamauryand it's for readability. Where you write USB_regname & USB_regname_fieldname, each character counts !
23:38:45JeromNoob question: does logf write the log to a file ?
23:39:34pamauryno, you need to dump it via the debug menu
23:40:35 Quit Highlander (Ping timeout: 276 seconds)
23:42:44pamaurybertrik: if you want to have a look at the usb code, see if you see something wrong
23:43:07pamauryI'm going to check that all registers have a correct address and there are no wrong bits
23:49:58AzureSkyI have a problem, When I Type /tools/configure from inside the rockbox/build folder it says "bash: /tools/configure: No such file of directory"
23:51:06Bagderyou probably want ../tools/configure
23:51:16Bagderie with two dots first
23:52:38pixelmaBagder: do you know something about the daily builds breakage mentioned in - alternatively when Zagor will be around?
23:52:56AzureSkytesting one build back once this finishes compiling, wana be 100% sure its 26937 thats causing microsd cards to not work
23:52:59Bagderand nope
23:53:45pixelmadaily manuals haven't been build either (last on the 17th) and since we only keep 3 or 4 dailies recently, there won't be any soon
23:54:21bluebrothermaybe it would make sense to add the manuals to the build system?
23:55:03pixelmaso manuals after all commits too?
23:55:30bluebrotherwhy not? Shouldn't take too long. The only problem is that there is nothing like ccache for TeX
23:56:18bluebrotheror maybe as separate build round to not slow the builds itself down?
23:56:42pixelmanothing against that really, thinking of recent breakage and fixes which needed you to wait until the next day to see if all have been fixed
23:57:50pixelmamy question was more about how that could help now (and I first thought you just meant build to see if it still builds as done with sims, which wouldn't help in this case)
23:58:38bluebrotherwell, not sure if it would make sense to publish current manuals. Would be an option of course.

Previous day | Next day