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 2009-07-04

00:04:22 Quit K3N (Read error: 110 (Connection timed out))
00:06:12oboe200 boot builds are missing from the build page - they are currently broken
00:07:17rasherobo: I see them on the build page..
00:08:38oborasher: don't worry, I'm just going mad. Not enough sleep last night
00:11:54 Join kachna|lappy [0] (
00:16:16CIA-71New commit by 03peter (r21626): Accept FS #10244 by Wincent Balin: more pdbox work done for GSoC; also some keyword and line-ending fixes by me
00:24:27saratoga_can we please update the Sansa bootloaders
00:28:11*petur hands mcuelenaere 618 points
00:28:35mcuelenaereall yellow, cool :)
00:29:36CIA-71New commit by 03mcuelenaere (r21627): Also enable frequency switching on Onda VX767
00:33:28 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
00:35:00saratoga_how close to being supported are the Onda players?
00:39:08 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
00:39:12 Quit bmbl (Remote closed the connection)
00:39:22kugelsaratoga_: update?
00:39:41saratoga_kugel: on the clip issue?
00:40:02kugelwhat do you mean by update?
00:40:26saratoga_we still haven't released the 3.3 sansa bootloaders
00:40:45kugelah, that sansas :8
00:41:16rasherI think we should start over. Maybe create a branch for them, because the round I built have some issues
00:41:51 Nick TheSeven is now known as TheSeven|Zzz (
00:41:59rasherTheSeven|Zzz: Please don't do that here.
00:42:48 Quit Zarggg (Read error: 104 (Connection reset by peer))
00:44:05 Join Zarggg [0] (
00:44:43 Quit linuxstb (Remote closed the connection)
00:44:55 Quit saratoga_ ("Page closed")
00:45:04mcuelenaeresaratoga_: there's still some weird crashes with pictureflow, doom & genre tags on some music files
00:45:20mcuelenaerenext to that there's no rbutil support, dual boot or installation method
00:46:25mcuelenaerehmm build server stuck?
00:47:05 Quit ender` (" The reason they call it the American Dream is because you have to be asleep to believe it. -- George Carlin")
00:48:32 Quit Zarggg (Read error: 54 (Connection reset by peer))
00:50:50CIA-71New commit by 03zagor (r21628): Inverted browser check to only server multipart/mixed to Gecko browsers. ...
00:51:51***Saving seen data "./dancer.seen"
00:52:42*bertrik sees working DMA to the PCM (IIS) on the meizu M3
00:53:08rasherZagor: I'm thinking khtml can probably manage as well?
00:54:15drakonikThis is interesting.
00:54:26drakonikThe database scanner crash causes Rockbox to see a full battery as empty
00:54:27 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
00:54:41Zagorrasher: I don't think it does. Safari doesn't.
00:56:22Zagorrasher: do you have kthml installed to try?
00:57:38rasherChrome doesn't seem to like it - thinks it's a download
00:59:25Zagorsame with webkit (arora - who says "Gecko" in the user agent... grr)
00:59:41 Join Zarggg [0] (
01:00:32Zagorand so does Chrome
01:01:26drakonikEither that, or in a half of a second, it manages to expend the entire charge without melting anything.
01:02:48CIA-71New commit by 03zagor (r21629): Identify real Gecko-based browsers (Mozilla).
01:03:11CIA-71New commit by 03mcuelenaere (r21630): Lua: always expose BUTTON_TOUCHSCREEN and remove BUTTON_ constants from rocklib.c
01:20:16 Quit mcuelenaere ("Gnight")
01:23:03 Join froggyman [0] (
01:24:12 Join saratoga [0] (n=9803c264@rockbox/developer/saratoga)
01:24:26saratogaso theres no way to make Chrome or Safari work with the log viewer like FF does?
01:25:39Zagornot without changing the way it feeds data to the browser
01:26:24ZagorI didn't actually know this was a mozilla-only technique when I wrote it
01:27:30amiconnZagor: Is there a test page for this multipart thing?
01:27:32Unhelpfulon browsers that do reflow during transmission, couldn't you just never "finish" the document?
01:28:23ZagorUnhelpful: most (all?) browsers today don't render the screen until they know the boundaries, which they don't until they get all html
01:29:16Mikachusaratoga: there's not no way, but you probably have to use some fancy xmlhttp ajax javascript thing? :)
01:29:32UnhelpfulZagor: really? i've watched firefox redraw as it gets more of the page HTML....
01:29:49MikachuUnhelpful: we are saying how it only works in firefox
01:30:03 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
01:30:33Unhelpfuli was just offering a counterexample for "no draw until all html text received"... i can't say how IE and Safari do things, i don't use them.
01:30:50 Quit gevaerts (Nick collision from services.)
01:30:59ZagorUnhelpful: ok, I could be wrong. perhaps also things like wrapping can affect it. it's not a terribly reliable method anyhow.
01:30:59 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
01:31:04Unhelpfulnext time i'm on a slow link or dealing with a slow site that triggers redraw, i'll try it in konqueror too... surely this could be simulated though.
01:31:32Unhelpfulno, but if you can't use multipart, and don't want to do ajax, it may be all you have :/
01:43:11 Quit ze ("MOVING! gone for weeks, probably.")
01:45:09 Quit val3 (Read error: 104 (Connection reset by peer))
01:47:36 Quit BHSPitLappy (Read error: 104 (Connection reset by peer))
01:52:08CIA-71New commit by 03zagor (r21631): Fixed client dupe check. ...
01:53:06 Quit notlistening ("Leaving")
02:00:10 Quit petur ("Zzzzzz")
02:07:42 Quit mt (Remote closed the connection)
02:08:18CIA-71New commit by 03zagor (r21632): Exit with 22 if HELLO failed.
02:09:39Zagoramiconn: sorry, no I don't know a test page
02:13:31kugelgevaerts: new patch
02:13:41kugel −− apply with -p1 :)
02:15:33 Quit Zagor ("Leaving")
02:15:46*gevaerts builds
02:16:17 Join evilnick_bs [0] (i=620ec27e@gateway/web/freenode/x-aa235488d431a937)
02:20:55gevaertskugel: nice :)
02:21:42kugelwhich target ?
02:22:22 Quit Thundercloud (Remote closed the connection)
02:25:09 Join BdN3504 [0] (n=5ce53b8f@gateway/web/cgi-irc/
02:27:11kugelgevaerts: ^
02:27:20gevaertsgigabeat f
02:27:31BdN3504have i missed something? since when does the theme gallery support the second screenshot i uploaded to my themes (i mean the menu screenshot)? i am looking at the gigabeat f gallery, but only two of my themes show the menu upon hover
02:27:36kugeldoes going to the wps from pf work ?
02:28:19BdN3504ah no, "spartan black" also shows its menu screenshot
02:29:02 Quit Unhelpful (Read error: 113 (No route to host))
02:29:18kugelgevaerts: with the standard wps button too?
02:30:11gevaertsapparently, yes
02:30:51kugelthat's cool
02:32:29kugelgevaerts: that's using exit() in rfa btw, since the return value of main_menu() is used for other things
02:33:40 Join Unhelpful [0] (n=Militant@rockbox/developer/Unhelpful)
02:33:54BdN3504ok sorry, my bad. actually all menu screens are show if you leave your mouse long enough upon a wps. but still, since when does this work? don't we add this to the major changes?
02:35:27 Quit bertrik (Read error: 113 (No route to host))
02:40:05gevaertskugel: well, the return value indicates whether it should exit, so that bit would be fine. I think I'd rather see that exit variable become an integer instead of a boolean, so you can indicate no exit, normal exit, or error
02:40:38gevaertsThe way it is now is a bit inconsistent. I think it should either use exit() everywhere, or return everywhere
02:40:46 Part n00b81 ("Leaving")
02:40:51 Nick fxb__ is now known as fxb (
02:41:13gevaertsthose are minor things though
02:43:25 Quit BdN3504 ("CGI:IRC (EOF)")
02:47:05soapdrakonik, my idea - as Llorean suspects - was to isolate possible hardware issues from software (Rockbox) issues. I think we can conclusively say you have a non-rockbox issue. :(
02:48:40drakonikThere's a LOT of problems. One of them is that something is causing the scanner to crash. Another is that file transfers while in Rockbox mode, but not ipod-firmware-mode, cause an endless unstoppable loop of crashing and rebooting and crashing.
02:49:05drakonikOr at least, sometimes they do.
02:49:33drakonikAnd sometimes they don't, and sometimes it's a different file that causes the hangup, and sometimes there's no noticable error regarding files.
02:49:37drakonikIt's all really confusing =\
02:50:48drakonikOh, now there's the issue of a database scanner induced crash causing rockbox to stop seeing a full battery as full. Or, causing it to discharge the whole battery at once.
02:51:55***Saving seen data "./dancer.seen"
03:05:19soapdrakonik, I'm guessing there is one problem.
03:05:55soapwhatever is causing the file transfer problems is likely causing the playback (IIRC) and DB scanner problem.
03:06:31soaphmm, rockbox needs a /dev/null.
03:07:34soapThen you could test transfers when the drive is suspect (or just to isolate the wire speed w/o drive speed affecting results) by copying files from PC to /dev/null/
03:12:26 Nick fxb is now known as fxb__ (
03:12:50 Join DarkSpectrum [0] (
03:13:55JdGordon /dev/null shouldnt be too hard to imple,ment
03:15:29DarkSpectrumany word on the new bootloader yet?
03:18:14 Join r0b- [0] (
03:18:25r0b-how hard is it to install rockbox to a Fuze
03:18:29drakonikmaybe there's something in the debug tools that'll help me
03:18:48scorche|shr0b-: well considering that it isnt a supported target yet...
03:19:32r0b-oh it isnt
03:23:15 Join goffa__ [0] (n=goffa@
03:23:16DarkSpectrumok how do i install the bootloader from FS #9955 ?
03:24:24 Join topbloke [0] (
03:25:21 Quit evilnick_bs ("Page closed")
03:31:23 Quit topbloke ("bye")
03:34:25 Quit evilnick_home1 ("Leaving.")
03:34:51 Quit goffa_ (Read error: 110 (Connection timed out))
03:35:27drakonikI dunno, I'm really disheartened. A scanner crash means I have to shunt it into ipod mode so that it'll have a charge. That means 30 minutes or an hour between times I can narrow down the files being scanned. And I don't know any better way to figure out what's wrong.
03:37:23 Quit zeveroare (Remote closed the connection)
03:44:39soapdrakonik, with all the other problems you're having my first suspicion is now hardware failure, not messed-up tags. :(
03:45:26drakonikI'd think the smae thing too if it weren't for the fact that until the moment I installed rockbox, my ipod hadn't EVER done anything like this ever
03:46:02soapyou, yourself, said earlier that you didn't use the ipod before rockbox.
03:46:08drakonikNo, I did
03:46:10soapmuch, at least.
03:46:19drakonikI used it extensively
03:46:25drakonikI phrased it poorly
03:46:47soap<drakonik> Which is strange, cause I never had an iota of trouble till I started messing with Rockbox. To be fair though, I didn't really try doing anything AT ALL with it till I started dicking with Rockbox
03:47:00soap10h18m ago
03:47:05drakonikYeah, that was crappy choice of words on my part =\
03:47:24soapthe fact of the matter is rockbox has no impact on behavior of the original firmware.
03:47:46drakonikWhat I meant was I never did anything directly to it, I just let itunes/winamp/whatever sync media to it. I never directly tried to transfer this many files to it before.
03:48:31soapI only just got home, and have not read the last 12 hours of logs (except where my name was said and the context around that) - but have you checked your iPod's filesystem? If you're lucky that is the solution to all your data problems. Your battery problem is just that, a battery problem.
03:48:46drakonikYeah, I've formatted it completely clean
03:49:03soaprewriting the FAT is not a disk scan.
03:49:29drakonikNot quite sure what you mean by 'disk scan'
03:49:51soapare you on windows? (winamp)?
03:50:08soapCHKDSK on Windows XP.
03:50:23soapScandisk on earlier flavors of windows.
03:50:25drakonikOkay, I see. Sorry
03:50:28soapDo the full scan.
03:50:33soapprobe it deep.
03:51:34drakonikAlright, running it now
03:52:59drakonikhuh. Apparently, "Windows made corrections to the file system"
03:53:34drakonikFingers crossed.
03:54:18soapthat was fast.
03:54:39 Quit DarkSpectrum ()
03:54:48 Quit amiconn (Nick collision from services.)
03:54:49 Quit pixelma (Nick collision from services.)
03:54:49 Join pixelma_ [50] (n=pixelma@rockbox/staff/pixelma)
03:54:50 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
03:55:03 Join DarkSpectrum [0] (
03:55:09 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
03:55:12 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
03:55:18drakonikWell, usually, it crashes by this point.
03:55:27 Join toffe82 [0] (
03:55:27drakonikSo I'm cautiously optimistic
03:55:47drakonikAh! There is goes
04:02:31tmzthave you formatted the whole filesystem?
04:02:40tmzteven zeroed it?
04:02:57tmztfull format? not quick
04:03:03drakonikFull format, yeah
04:03:10drakonikI can do it again
04:03:31drakonikI'm not sure if Windows is all that reliable
04:08:19 Join toffe82_ [0] (
04:08:27 Quit TheSeven|Zzz (Nick collision from services.)
04:08:41 Join TheSeven|Zzz [0] (
04:16:27froggymanwhat information would I need to put up about a speaker/remote for help with ipod protocol help?
04:17:27froggymancurrently only audio works with the speakerdocks remote
04:21:43drakonikman, this chkdsk run is taking forever
04:23:50 Quit toffe82 (Read error: 110 (Connection timed out))
04:26:12soapit should - and don't format again - that won't fix the problem.
04:26:49drakonikI ran it with a new flag this second time, to recover readable data from damaged sections. It shouldn't change much, but I figured thoroughness can't hurt
04:30:25Galois badblocks takes two days (!!) on a 2TB drive
04:34:13drakonikand now it's not making any progress at all, and Ctrl+C doesn't work. Fucking awesome.
04:47:32drakonikOn the bright side, I did something that stopped the crashes from wiping the battery
04:47:54soapif your ipod fails to respond to the Menu + Select combo you have a hardware failure.
04:48:01drakonikNo, it responds to it
04:48:33soapthen I'm confused on how a crash is killing the battery...
04:49:07drakonikWhen the database scan caused a crash, on reboot, Rockbox would refuse to acknowledge even a full battery as full.
04:49:50drakonikI left it to charge overnight, and crashed it moments after unplugging it. It rebooted, and rockbox thought the battery was 100% empty. It booted up, said "Battery is empty please charge" and shut itself down.
04:49:53soapyour battery - from everything you have said - is on its way out.
04:50:27drakonikAnd I'd boot into the ipod firmware, and it'd show a solid 75 to 100% charge.
04:50:43soapa failing lithium chemistry battery develops higher internal resistance - thus high current activities such as running the hard drive pull the voltage down more than on a new battery.
04:51:09soapthere is no dipstick on a battery - you estimate remaining charge by comparing voltage to a discharge curve.
04:51:33soapwhen the voltage dips under load a lower estimation is made based on what that voltage normally means.
04:51:44soapregardless, your battery is on death's door.
04:51:56***Saving seen data "./dancer.seen"
04:52:14 Join evilnick_home [0] (
04:52:26soapI don't care what the iPod firmware says, I have no idea how they estimate remaining life - but estimate it they do.
04:52:35drakonikYeah, that's fair.
04:53:14drakonikBut, my point is, Rockbox thinks the battery is 95% full, and then one crash later, thinks it's 1% full. That's the discrepancy that concerns me.
04:53:33soapI already explained that.
04:53:41soapthe drive running is the factor.
04:53:57soaprunning while creating the database, and immediately running again upon reboot.
04:54:24soapexactly the conditions you would expect to see a failing battery start to show itself.
04:56:09soapI believe I and others have given the most likely explanation for all of the evidence you have provided so far. I also believe you don't like the answer, as it is not a good prognosis, for it appears to me you keep asking the same questions and providing evidence predicted by us.
04:56:16 Join Hillshum [0] (n=chatzill@unaffiliated/hillshum)
04:56:24soapi'm sorry about your luck, but I really do suspect hardware failure.
05:00:37drakonikYeah. It might be. I find it hard to buy and I really don't want to give up, because I've been utterly disenchanted with the actual ipod firmware and the management software that it requires. I want rockbox to work so I can be fucking done with that bullshit.
05:01:24soapyou can run iPod's diagnostic mode.
05:01:32soapApple's built-in one.
05:01:44soapdon't recall the keypresses to get it.
05:02:54drakonikI'm sure I can find it
05:04:03 Quit evilnick_home (Read error: 104 (Connection reset by peer))
05:04:35 Join evilnick_home [0] (
05:10:24 Quit kugel (Read error: 110 (Connection timed out))
05:10:34 Quit mc2739 ("ChatZilla 0.9.85 [Firefox 3.0.11/2009060215]")
05:11:14drakonikWow. This is new.
05:11:19drakonikI've never even heard of this
05:16:14drakonikShit, now it's stuck on a diagnostic
05:17:17 Join PaulPosition [0] (
05:17:19 Quit BlakeJohnson86 (Read error: 104 (Connection reset by peer))
05:21:43drakonikAlright, I dunno if I'm looking for any specific numbers, but it passes every single diagnostic I'm running
05:25:47drakonikInteresting. It seems to have populated the database pretty completely...
05:26:59drakonikHm. Crashing even when I'm not updating the database.
05:31:10HillshumI was running test_disk on my e200v2, I got a "*PANIC* flush_fat_sector() - Could not write sector 321 (error -101)" Is this of any importance?
05:34:18 Join BlakeJohnson86 [0] (
05:38:23 Quit Hillshum ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
05:52:28 Join Horschti [0] (n=Horscht2@xbmc/user/horscht)
06:02:23 Join Nausicaa [0] (
06:03:40NausicaaIs there a way to make the iPod mini 2g have less backlight intensity?
06:04:56 Join josh1238 [0] (
06:04:59 Part josh1238
06:05:30PaulPositionCan anyone explain this C-noob what the following accomplishes in plugin pictureflow code..? I'd like to get the keymap to work with my target (h10) but I can't see any logic in it..
06:06:31PaulPosition(as it is now, up and down scroll and left and right (usually std action cancel and select) do the same (ie, scroll).. :(
06:10:00 Quit Horscht (Read error: 110 (Connection timed out))
06:11:48 Quit HBK ()
06:16:05 Join HBK [0] (
06:17:09 Part toffe82_
06:23:23 Part PaulPosition
06:39:31 Quit saratoga ("CGI:IRC (EOF)")
06:48:34 Join dmb [0] (n=Dmb@unaffiliated/dmb)
06:51:59***Saving seen data "./dancer.seen"
06:52:18 Quit dmb (Remote closed the connection)
06:59:39 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
07:16:42 Join einhirn [0] (
07:18:27 Quit Nausicaa (Read error: 110 (Connection timed out))
07:18:38 Join Nausicaa [0] (
07:40:35 Join Horscht86 [0] (
07:48:49 Quit Nausicaa (Read error: 110 (Connection timed out))
07:57:02 Join petur [0] (n=peter@rockbox/developer/petur)
07:58:43 Quit Horschti (Read error: 110 (Connection timed out))
08:01:51 Quit rvvs89 (Read error: 104 (Connection reset by peer))
08:02:37 Join Rob2223 [0] (
08:03:28 Join jkl [0] (
08:15:39 Join Russel-Athletic [0] (
08:16:08 Join dmb [0] (n=Dmb@unaffiliated/dmb)
08:17:03Russel-AthleticI changed my tagnavi_custom.config, what do I need to do, to reload it?
08:18:51 Quit dmb (Read error: 104 (Connection reset by peer))
08:19:04 Join dmb [0] (n=Dmb@unaffiliated/dmb)
08:20:54 Quit Rob2222 (Read error: 110 (Connection timed out))
08:20:59 Quit dmb (Read error: 54 (Connection reset by peer))
08:24:24 Join dmb [0] (n=Dmb@unaffiliated/dmb)
08:25:01 Join stoffel [0] (
08:47:06 Quit einhirn (Read error: 104 (Connection reset by peer))
08:52:00***Saving seen data "./dancer.seen"
08:55:31 Join mt [0] (n=mt@
08:58:54 Join Grahack [0] (
09:00:12 Join Rob2222 [0] (
09:07:00 Join robin0800 [0] (
09:12:53 Nick at0m|c is now known as at0m (
09:17:22 Quit robin0800 (Read error: 104 (Connection reset by peer))
09:17:50 Quit Rob2223 (Read error: 110 (Connection timed out))
09:18:32 Join robin0800 [0] (
09:19:19 Quit stoffel (Remote closed the connection)
09:27:46 Join flydutch [0] (
09:30:04 Join daurnimator [0] (n=john@unaffiliated/daurnimator)
09:31:57 Quit Russel-Athletic ("leaving")
09:51:44 Quit robin0800 (Remote closed the connection)
09:51:47 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
09:57:14 Join robin0800 [0] (
09:57:49 Join bertrik [0] (
10:05:56 Join ender` [0] (
10:15:15 Part TheSeven|Zzz
10:25:51 Quit petur ("blub blub blub")
10:43:45 Quit mt (Remote closed the connection)
10:43:51bertrikwow, the s5l8700 apparently generates all internal clocks from a single 32768 Hz input clock
10:52:03***Saving seen data "./dancer.seen"
11:00:10 Join {phoenix} [0] (
11:21:15bertrikI'm getting confused about the number of clocks involved in s5l8700 PCM: I see mentions of PCLK, SCLK, BCLK, CODECLK and MCLK
11:21:51bertrikoh I forgot LRCK ...
11:26:12jklbertrik: google PLL
11:26:54bertrikthen what? it all becomes magically clear or something?
11:27:40jklbertrik: no, just saying that generating the various clocks for uC or SoC from a single source is common and old
11:39:01 Join mt [0] (n=mt@
11:45:05 Nick Horscht86 is now known as Horscht (
11:46:18mtCould someone check my final patch in FS #10182 ? I was going to commit it, but thought I should discuss it first.
11:46:58mtI'll be back later so I will check both the logs and the task for the comments.
11:54:42linuxstbmt: My first comment is that local functions (i.e. functions only used in the C file where you define them) should be declared "static". e.g. read_cook_extradata() and print_cook_extradata() in apps/metadata/rm.c
11:55:08linuxstbmt: Also, you seem to have set the svn:executable flag on a lot of files (e.g. apps/metadata/rm.c)
11:58:50linuxstbmt: I also notice some brackets don't look to be aligned properly - are there TABs in any of the files?
12:12:56 Quit rasher (Read error: 145 (Connection timed out))
12:23:02 Nick fxb__ is now known as fxb (
12:27:53 Join petur [0] (
12:52:07***Saving seen data "./dancer.seen"
12:55:15 Join kugel [0] (n=kugel@rockbox/developer/kugel)
12:55:27 Quit DarkSpectrum ()
12:57:08 Join DarkSpectrum [0] (
12:57:23 Join Thundercloud [0] (
13:04:18 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
13:09:24 Quit Grahack ("Leaving.")
13:09:25 Quit robin0800 (Read error: 104 (Connection reset by peer))
13:13:38 Quit Horscht ("Verlassend")
13:13:59 Quit Res1 (Read error: 60 (Operation timed out))
13:27:25 Join mcuelenaere [0] (
13:29:44mcuelenaerebertrik: what formula/how did you generate the log table for the vx747 backlight?
13:32:40 Join Minthe [0] (
13:34:54dfktis there a svn command to gradually upgrade an older revision to newer versions, or do i have to always check out a full old revision?
13:35:17Mikachuyou can check out a specific revision of just one file
13:35:28dfktmeaning, i want to go from r21550 to r21551 to r21552, etc
13:35:39Mikachujust give -r 1423
13:35:39dfktand checking out a full version takes quite some time
13:35:49bertrikmcuelenaere, markun generated that curve
13:35:51dfktover the same folder i used before?
13:36:08mcuelenaereah, markun: same question :)
13:36:15Mikachudfkt: svn update -r 3243 myfile.c
13:36:23Mikachudfkt: or a directory
13:36:24bertrikI think it was round(1+sqrt(2)^x) where x = 0..15
13:36:40dfktMikachu, thank you
13:37:45mcuelenaerethat seems to give 2, 3, 0, 1, 6, 7, 4, 5, 10, 11, 8, 9, 14, 15, 12, 13 ?
13:39:06Mikachuyou're doing something wrong then :)
13:39:21Mikachu% for a ({0..15}) {echo $(( 1 + sqrt(2)**a )) }|cut -f1 -d.|xargs
13:39:21Mikachu2 2 3 3 5 6 9 12 17 23 33 46 65 91 129 182
13:39:25bertriksqrt(2)^x - 1
13:40:01mcuelenaereMikachu: for($i=0; $i<16; $i++) echo round(sqrt(2)^$i-1).", "; echo "\n";
13:40:20Mikachumcuelenaere: i think ^ in perl is xor?
13:40:20mcuelenaereah, the ^ is probaby the wrong operator
13:40:29mcuelenaereand it's PHP btw :)
13:40:36Mikachuwell, regardless, ^ is usually xor :)
13:40:44bertrikmaybe with some fiddling for the low numbers to make it linear
13:40:50mcuelenaereyeah, that's the problem; thans
13:41:06bertrikto the power of is what I mean with that, sometimes written as **
13:41:27*mcuelenaere already tried this on his TI-83, but couldn't reproduce it on PC
13:42:00mcuelenaerebertrik: yes I know, I'm just not used to using math in scripts :)
13:43:33Tornehm. the usb mode in this beast bootloader doesn't seem to work :)
13:45:32 Part wincent ("Kopete 0.12.7 :")
13:45:46Tornewell, it doesn't show up as a disk, anyway
13:46:15Torneusb-storage: waiting for device to settle before scanning
13:46:26Tornethen nothign :(
13:47:15mcuelenaereTorne: is this just recently or .. ?
13:47:27Tornethis is the first time i've gotten it to boot at all, borrowed it from linuxsstb
13:47:32Tornethe battery was too dead for it to power on before.
13:47:46mcuelenaereah, and USB works fine in normal Rockbox?
13:47:57Torneno idea, the rockbox image is missing or something
13:48:05Torneand the OF doesn't appear to be there either
13:48:15Torneso it just reboots over and over until you plug in usb
13:50:23Torneah, weird. it mounts on the windows pc
13:50:28Tornejust not on my linux box
13:50:50daurnimatorhi mcuelenaere
13:54:29linuxstbTorne: That could be because the mbr is broken - see the GigabeatSInstallation wiki page for info on what you need to do to fix it.
13:56:37mcuelenaerehmm I don't suppose theres touchscreen support for the USBPOWER_BUTTON mechanism?
13:59:29mcuelenaerewhat's USBPOWER_BTN_IGNORE for?
14:03:55tmztis that related to vbus?
14:09:35 Quit perrikwp ("Leaving.")
14:11:27 Join efyx_ [0] (
14:11:30 Join Dark_Apostrophe [0] (n=tda@unaffiliated/darkapostrophe/x-0983214)
14:11:46Dark_ApostropheAny progess on the Cowon D2 port?
14:12:04Dark_ApostropheLast I checked (admittedly, long ago), there were USB problems
14:13:31Tornelinuxstb: it doesn't even come up as a storage device connected to linux
14:13:41Tornelinuxstb: let alone find any partitions.
14:13:51Tornelinuxstb: but on windows it works fien and i can browse both partitions
14:14:01Torneso i suspect my laptop is doign something stupid. it's ok :)
14:14:57bertrikDark_Apostrophe, some problems with flash access were recently fixed on the cowon d2
14:15:19Dark_Apostrophebertrik: Do you own one, or anyone else here?
14:15:30Dark_ApostropheI'd like to hear what it's like to use Rockbox on it - the default firmware is annoying as hell
14:15:53bertrikDark_Apostrophe, no I don't own one
14:16:16Dark_ApostropheWell, at least you were spared the torment of using the default fw :P
14:19:38 Join rasher [0] (
14:20:16kugeldfkt: with svn up -rXXX you can update single files/dirs, a set of files/dirs or the whole repo to that specific revision
14:20:53Mikachukugel: i already told him that :)
14:22:16kugelyea, I just saw that :<
14:26:13 Join Grahack [0] (
14:26:49CIA-71New commit by 03mcuelenaere (r21633): Consolidate all fixed point math routines in one library (FS #10400) by Jeffrey Goode
14:29:18dzhow do I install the bootloader in FS #9955 on my e200? sansapatcher -a?
14:32:08*bertrik wishes there was some kind of guide on what is expected of the pcm hardware functions
14:32:52CIA-71New commit by 03mcuelenaere (r21634): Set svn:keywords property
14:33:07*mcuelenaere wants that too
14:35:40mcuelenaerebertrik: what exactly is the issue?
14:36:17bertrikfor example, what to do in pcm_dma_apply_settings
14:36:34mcuelenaereset up the frequency
14:36:35bertrikshould I use pcm_sampr to calculate pcm rate, or fsel, or both?
14:37:05*mcuelenaere uses pcm_sampr
14:37:17mcuelenaerebut it depends on how you internally handle the frequency I suppose
14:37:58 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
14:38:40*mcuelenaere wonders about the red
14:38:49bertrikor pcm_play_dma_get_peak_buffer, does it expect bytes, or samples?
14:39:36bertrikI'm not sure if I can guarantee that the peak buffer address and size are consistent with each other, without stopping dma
14:40:31mcuelenaerebertrik: bytes >> 2 I think
14:42:27mcuelenaerehmm this doesn't seem right, FS #10400 *increases* binsize
14:43:39 Join Blue_Dude [0] (
14:44:59Blue_Dudemcuelenaere: It would increase it some. It brings in some code that existed outside the core before.
14:45:14mcuelenaerehmm then that isn't really good
14:45:30mcuelenaereperhaps that should be #ifdef'd with PLUGIN?
14:45:31rasherMaybe it could be split up?
14:45:48Blue_DudeI thought it might be a bit bloated. What's the difference in size?
14:46:12mcuelenaerearound 2,4kB
14:46:28rasherSo instead of #include "fixedpoint.h" it would be #include "fixedpoint/functionname.h"
14:47:54Blue_DudeI thought about making subset .h files. That might be a better approach.
14:49:07Blue_DudeI also wrote a coupld of brand new functions that would be exposed rather than using the actual math functions. I noticed that replaygain for instance doesn't actually need the results of exp10, it needs a gain factor.
14:50:17 Join s0u][ight [0] (
14:51:07Blue_DudeNothing right now actually uses the log10 or decibel functions. Those can be #if 0 commented out right now without affecting anything. I put them in because they might be useful.
14:51:59mcuelenaereoutput of bloat-o-meter between the two revisions:
14:52:09***Saving seen data "./dancer.seen"
14:52:45Blue_DudeTell you what, let me work more on this later today, and I'll post another patch. Can we reopen FS?
14:55:15mcuelenaereBlue_Dude: re-opened
14:56:50Blue_DudeExcellent. I'll have some free time later today to noodle it some more. Other than unneeded function bloat, how does this approach seem? Is it a decent kludge or are we better off exposing the core functions to plugins and codecs via the API's?
14:58:12mcuelenaereBlue_Dude: this approach doesn't require much change to plugins, so I'm not against it
14:59:16Blue_DudeOK. It's not very pretty but it's better than rewriting or copying code around.
14:59:20 Quit Lss (Read error: 104 (Connection reset by peer))
14:59:39mcuelenaereBlue_Dude: also, any idea about the red on the H100?
15:00:05 Join Lss [0] (
15:00:08 Quit petur ("real-life")
15:00:15mcuelenaerethe two red table cells in the upper-right
15:00:20mcuelenaere(click on them)
15:01:26rasherThat's a rather nasty delta
15:02:25Blue_DudeNope, don't know why. I'm not sure why that particular build wouldn't behave.
15:02:57pixelmathe red looks like it's in rombox which explains why it's H100s only - if the rest of the code changes don't affect hwcodec targets (or lowmem ones)
15:04:27Blue_DudeYeah. Time to rethink this one. We're probably better off revoking the commit until it behaves everywhere.
15:05:11Blue_DudeReally gotta go. Sorry to run in the middle like this.
15:05:34 Quit Blue_Dude ("ChatZilla 0.9.85 [Firefox 3.0.11/2009060215]")
15:14:33 Join n00b81 [0] (n=taylor@unaffiliated/n00b81)
15:18:00CIA-71New commit by 03mcuelenaere (r21635): Revert "Consolidate all fixed point math routines in one library (FS #10400) by Jeffrey Goode"
15:18:14 Part n00b81 ("Leaving")
15:18:59MintheExcuse me, but may I ask for the link to rm.codec?
15:23:05Dark_ApostropheAnyone here got a Cowon D2?
15:24:47kugelmcuelenaere: a comment on the reasons for reverting would've been nice
15:25:11mcuelenaereyes, forgot about that (was too busy figuring out git)
15:25:22kugelI know that problem :P
15:25:34mcuelenaereit does have a nice git revert command though :)
15:25:45kugelI actually used svn to revert my mistake yesterday
15:26:39rasherMinthe: You'd have to compile Rockbox yourself. See for the patch
15:27:13 Join Nausicaa [0] (
15:31:08 Quit Horscht ("Verlassend")
15:31:27bertrikyay, PCM on the meizu is now running at the expected rate
15:32:10MintheOK, Thank you. Hopefully I have a build environment to use usbstack(since 2008)
15:32:45 Quit s0u][ight ("Leaving")
15:34:11 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
15:35:28bertrikI think it won't take very long now before we have first sound on the meizu
15:36:37gevaertsTorne: "waiting for device to settle before scanning" and then nothing is the classic symptom of this hal/gphoto2 bug
15:36:47 Quit Minthe ("Leaving...")
15:40:37Tornegevaerts: ah, maybe this machine is all too ubuntu-y
15:40:54Torneit worked on windows anyway so it boots rockbox and charges properly now :)
15:40:58Torneand i dumped the rom
15:41:10Tornewill post a fs# with the patch to enable flash rom dumping on beast soon
15:42:44rashergevaerts: did you see we "fixed" some issues with the c200?
15:43:12 Quit kugel (Nick collision from services.)
15:43:20 Join kugel [0] (n=kugel@
15:43:53gevaertsrasher: I saw that, yes. I don't have a real opinion on which is the right id to use, so if this makes things work better for some people, why not?
15:44:28rasherYeah. It's not that it's more right than the other, but we might as well act like the OF when it makes things work better
15:48:55*Torne wonders why config-gigabeat-s.h says FLASH_SIZE is 4mb when it's clearly 2mb :)
15:49:39gevaertsTorne: is it 2mb on all gigabeats? Maybe there are different models
15:50:01Torneit's possible, i guess. the part number on the wiki is a 2mb part, and i dumped 4mb and it's mirrored
15:50:11Tornebut there might be a different one, yah
15:50:15Torneit doesn't amtter a great deal, admittedly
15:50:26TorneFLASH_SIZE is not used for anything, i was just using it for my rom dumping patch
15:50:27Unhelpful...somebody's working on beast flash? :)
15:50:45TorneUnhelpful: i'm reversing the flash code a bit, yah
15:50:54Tornenot *writing* to it :)
15:50:56Unhelpfulto what end?
15:51:13Tornetring to work out why it decides to reformat the player sometimes after rockbox has been used
15:51:36Unhelpfulah... it'd be nice if we could get the bootloader in there. :)
15:51:54Torneoh. that'd be easy, probably
15:52:26Tornein fact since it s a2mb flash you cuold prob ably put rockbox in there too :)
15:53:03Tornethe chip isn't entirely empty in this dump though so the OF might actually use some of it for purposes other htan the bootloader code
15:53:20Unhelpfulhrm... and it can execute in-place from there?
15:53:48Torneit's an intel bootflash, nor.
15:53:51gevaertsThat would be nice. The beast is quite RAM-constrained! :P
15:54:08Tornewell yah
15:54:09Unhelpfulgevaerts: it would speed boot time quite a bit. :)
15:54:15TorneUnhelpful: it probably wouldn't
15:54:32Tornenor flash is pretty slow to read on most modern platforms compared to sdram
15:54:37UnhelpfulTorne: i have the 1.3 bootflash... the one that still needs a dual-boot flash bootloader. :/
15:54:47kugelbut you still save the disk spin up
15:54:57Unhelpfuls/flash b/rockbox b/ :)
15:55:02Tornekugel: yah. you'd want to copy to ram, rather than xip'ing
15:55:13kugelprobably, yes
15:55:14TorneUnhelpful: what's the difference between the firmware versions, btw? :)
15:55:23kugelbut it would most likely improve boot time
15:55:32Tornei don't know anything about the beast besides what i've just figured out by dumping its rom and reading the wiki
15:55:49UnhelpfulTorne: i don't really know... i believe 1.3 was a fix for the leap-year bug.
15:55:53Tornenot sure which one this is in flash
15:55:57Tornebut it's not 1.2 or 1.3
15:55:59gevaertsMost boot time improvement would come from just not half-booting the OF
15:56:11Tornegevaerts: like i did on ipod, kinda ;)
15:56:29Unhelpfulexactly. ideally we'd like to be able to chain-load the OF from the RB bootloader, or from RB itself
15:56:29Torneanyway yah. it can go on the list.
15:56:30gevaertsTorne: on the ipod, it only loaded that bit :)
15:56:33Unhelpfulalso it's a bit more than half. :)
15:56:59Tornegevaerts: still took over a second :)
15:57:28gevaertsTorne: try measuring how long it takes on the beast
15:57:39Tornethis beast doesn't ahve the OF
15:57:42UnhelpfulTorne: if you have a patch for a flash dump plugin, i can get my installed 1.3 flash for you. i'm *pretty* sure it installed new flash - i used the updater tool on windows, and it acted very much like it was writing to flash, with all of the warnings that tend to go with that.
15:57:48Tornenk.bin is just the rockbox bootloader
15:57:58TorneUnhelpful: taht would be nice. i'll be uloading the patch shortly
15:59:20TorneUnhelpful: did it seem like it 3as flashing during the update, or during the subsequent reboot?
16:01:25Torneeee 701 keyboard :)
16:01:33Torneand smoking.
16:01:35Unhelpful...ah. i *believe* during the reboot, but this was some months ago.
16:03:11Torneyah, the boot log on the wiki implies that it compares the files on disk with the ones in flash at boot and reflashes if it's newer, but that may not be right
16:04:06*bertrik hears noise from his meizu \o/
16:04:10Tornewhich vesions need the dual flash bootloader? and why? :)
16:04:23bertrikwell, not mine actually, gevaerts'
16:04:25Tornedual boot :)
16:04:41Tornethis beast obviously doesn't, nk.bin is just our bootloader
16:04:45Tornethe OF is not there
16:05:32Torneso it boots quite fast :)
16:07:22 Quit Nausicaa (Read error: 110 (Connection timed out))
16:07:25UnhelpfulTorne: mine came with 1.2, i don't think it reveals a flash version if that's versioned separately at all... it worked with single-boot RB bootloader then. the official update to 1.3 broke single-boot. running a dual-boot built with the 1.3 nk.bin didn't - i did that for a bit first, and obviously booting a 1.3 nk.bin doesn't do anything to tha flash.
16:08:18Torneyah, i think to trigger the reflash you have to put the other .bin files on there
16:08:29Torneit probably deletes them after a successful flash
16:10:16Torneoh, of course, that's what else is in the flash: the recovery image
16:10:19Unhelpfulit had other .bin files after i started - perhaps some flag is set when they're uploaded, and cleared when they're flashed.
16:10:32Tornestill doesn't need 2mb though
16:11:05TorneUnhelpful: yah, normally i think the updater will copy all three bins ont ehre, and then it'll reflash the bootloader and the recovery image
16:11:16Tornethe log implies it's timestamp based
16:11:19Tornerather than a flag
16:12:25 Join _lifeless [0] (n=lifeless@
16:13:00Unhelpfulthat would work too... it doesn't seem to remove them, but it will work just fine if you do. nk.bin alone will boot it.
16:13:11Torneyup. this beast only has nk.bin
16:13:26Torne(which is our bootloader)
16:13:41Tornei love the signature check bypass thing, btw
16:13:49Torneits' just hilarious.
16:14:30Tornesuch a stupid bug in the loader :)
16:16:03Unhelpfuli've actually never looked at how we do it :)
16:16:44 Quit Horscht ("Verlassend")
16:17:03gevaertsbertrik: congratulations! "Ladies and gentlemen, we have noise!"
16:19:23TorneUnhelpful: the nk.bin consists of a series of records, with a load address and size for each
16:19:35TorneUnhelpful: but it doesn't bounds check the load addreses, and the bootloader is running from RAM by this point
16:19:47TorneUnhelpful: so you can tell it to load a record at an address that overwrites part of the bootloader
16:19:55Tornedisabling the signature check by branching past it.
16:20:09 Quit {phoenix} (Read error: 110 (Connection timed out))
16:20:58Torneso one 4 byte record with a branch instruction in it disables all the security.
16:21:09 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
16:22:50Unhelpful...and it loads *all* of these records before it checks anything for signatures?
16:23:42Torneyes. because the records collectively constitute the thing it's signature checking :)
16:24:01Torneit loads them into ram at their desired addresses to save having to shuffle it all around, i guess
16:24:15Torneit would be a waste of time to load the file and check it then just have to relocate all the data again :)
16:24:46Torneit should really check that the records aren't going to overwrite the bits of ramt he bootloader is using, but good for us taht it doesn't :)
16:25:29Torneok, the flash dump patch is FS #10410 if you wanna give it a go
16:26:53Mikachuthe devil is in the details huh
16:27:22 Quit __lifeless (Read error: 101 (Network is unreachable))
16:30:00 Join {phoenix} [0] (n=dirk@
16:31:30 Quit lilltiger (Read error: 110 (Connection timed out))
16:34:12 Join fdinel [0] (n=Miranda@
16:35:46 Join robin0800 [0] (
16:48:57 Quit robin0800 ("Leaving")
16:49:26 Join robin0800 [0] (
16:50:06 Quit efyx_ (Remote closed the connection)
16:52:00 Quit Grahack (Read error: 60 (Operation timed out))
16:52:14***Saving seen data "./dancer.seen"
16:52:39 Join Grahack [0] (
16:55:56 Join Mysterytrain [0] (
16:58:49 Part Mysterytrain
16:59:15TorneUnhelpful: cheers, i'll compare it to the other binaries i have ;)
17:01:44*Unhelpful wonders how he can even look at this... arm-*elf*-objdump obviously doesn't like a raw memory dump
17:02:11Tornejust tell it the file type is 'binary'
17:02:48 Join saratoga [0] (n=41becb3b@gateway/web/cgi-irc/
17:03:24saratogathe voltage scaling patch for AMS seems to fail completely on some Clips, i think we shoudl revert it for now
17:04:16Unhelpfuli'm looking for such an option...
17:05:13mcuelenaereUnhelpful: -b binary
17:05:29Lloreansaratoga: It also seems to not work with SD cards according to one post in the thread?
17:05:50saratogathat as well
17:08:46 Join PaulPosition [0] (
17:10:54 Quit dmb (Read error: 104 (Connection reset by peer))
17:14:24 Join stoffel [0] (
17:14:53 Quit robin0800 ("Leaving")
17:15:30 Join robin0800 [0] (
17:18:03kugelsaratoga: we should try to boost on sd transfers first (or scale voltage only), imo
17:18:22kugelLlorean: only some
17:19:05kugelunfortunately, the uSD of the devs all work just fine
17:19:50Unhelpfulkugel: i think that the big "non-scrollwheel" change in the PF keymap caused some trouble. it was previously using the scrollbar for scrolling on H10... it still does now, but the right/left buttons are now scroll as well, and there's no select/cancel.
17:19:56 Join Minthe [0] (
17:19:57kugelTorne: can we try to stop bloating debug_menu.c with target dependant code? I'd rather have it in the target tree
17:20:04PaulPositionTrying to get the pictureflow plugin working with my target's keymap (iriver H10)... The code first defines a lot of actions from ACTION_STD_* and then proceeds to overload them anyway with target-level (?) (BUTTON_LEFT, etc.) names... is there any other plugin which does things similarly and which I could use to compare and understand the logic? (I'm not a coder but I still have half a functioning brain)
17:20:41Unhelpfulthere's the fellow with the problem, in fact. :)
17:20:56PaulPositionohj, lol
17:21:16kugelUnhelpful: I think having left and right is nice, but not a requirement. People are probably expecting to use the scrollpad anyway on the h10
17:22:34 Join AndyI [0] (i=AndyI@
17:22:41Unhelpfulkugel: that's what i'm fairly sure it was doing before a huge block of CONFIG_KEYPAD got replaced with HAVE_SCROLLWHEEL... we probably need to treat the H10 as a scrollwheel target.
17:22:41PaulPosition(let it be known that the scrollpad has always been a buttonUp/buttonDown affair on the h10 (ie, it never 'scrolled').. Does the H10 identifies itself as scrollwheel device?
17:24:05 Part Dark_Apostrophe ("I tawt I taw a puddy tat")
17:24:37kugelthe problem is only about pictureflow, where the albums covers are are basically horizontal list, hence using some left/right buttons is desire-able
17:24:40UnhelpfulPaulPosition: it's not how you scroll in lists?
17:24:58Tornekugel: the rom dumping thing could easily be a generic function with target defines for where the rom[s] are..
17:25:00Unhelpfulkugel: yes, but the H10 doesn't have a d-pad layout.
17:25:06Tornewell i say easily
17:25:14Torne*fairly* generic. :)
17:25:14PaulPositionYeah, I meant to say that it's not a 'flowing scroll' ie, upper part is button up, lower part button down.. You don'T slide along...
17:25:54kugelgeneric is fine, but debug_menu is already full of highly target-dependant, non-generic stuff, and it's a nightmare
17:27:46 Quit Horscht ("Verlassend")
17:28:11Torneall the versions of the dumping functions are the same, modulo where they dump from and whether they have to turn off the memory guard
17:28:30Tornei'll maybe look at that another time ;)
17:28:48 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
17:30:03UnhelpfulPaulPosition: i think this should sort it... basically treats the H10 like a scrollwheel target:
17:30:43pixelmaUnhelpful: the standard actions pf uses... where are they taken from? I ask because there are a few contexts that e.g. remap ACTION_STD_OK
17:30:49PaulPositionunhelpful - well that's helpful.. Mind if I take some time examining it (for learning) before I compile and test? :)
17:30:55TorneUnhelpful: the start of your beast's flash is substantially similar to the unpacked 1.3 boot binary :)
17:30:59drakonikI'll just cheat my way to a full database. Tell it to update and shut it down (and thus having it commit its update) before it crashes. At least, I hope that'll work
17:31:05Unhelpfulnot at all :)
17:31:07TorneUnhelpful: but as you noticed there are quite a few flipped bits in the unused regions
17:31:31Torneand some substantial changes later on as well
17:31:35Torneso that's.. interesting ;)
17:31:52Unhelpfulpixelma: they chain to CONTEXT_TREE on everything but M3, where they chain to CONTEXT_REMOTE
17:32:07kugelUnhelpful: this has nothing to do with scroll or not scroll
17:32:15kugelthe h10 uses his scrollpad as plain buttons
17:32:57kugelleft / right is ACTION_STD_CANCEL and _OK on many targets.
17:32:58Unhelpfulkugel: but it maps the scrollpad to PREV/NEXT, and the non-D-pad layout makes the mapping that other targets use make very little sense.
17:33:24froggymanwhat information would I have to put up on a speaker that docks and has a remote to help advance the iPod accessory protocol
17:33:35kugelthat's right, but the HAVE_SCROLL define is missleading IMO
17:34:06Unhelpfulperhaps call it "USE_CORE_SCROLL" or "USE_CORE_PREVNEXT"?
17:34:12kugelthe samsungs would have the same problem
17:34:25kugelthey have up/down and left/right, but no select that's in the middle
17:34:40kugelI compiled pf for testing, and I couldn't access the track list
17:34:42Tornefroggyman: you'd probably need to be willing to dump the serial connection between it and the ipod while in the original firmware
17:34:54Torneso we can see what it's supposed to do
17:35:06 Quit AndyIL (Read error: 110 (Connection timed out))
17:35:14Tornethis would need you to buy/make suitable cables to break out the dock port and watch it from a pc :)
17:35:36kugelI don't think using up/down for scrolling through the covers is a big deal, if any other list uses up/down too
17:35:42Tornefroggyman: i presume it doesn't work at all at the moment?
17:35:53 Quit Horscht ("Verlassend")
17:35:57kugelI actually used up/down on the samsung before trying left/righ
17:35:58froggymanTorne, That is correct
17:36:10Tornefroggyman: do you not even get audio out? because that *generally* always works
17:36:28froggymanTorne, well that is the one thing that does work
17:36:36Torneright. so you just can't control the ipod with it
17:36:41froggymanTorne, correct
17:36:48Unhelpfuli think on targets with a d-pad layout for up/down/left/right, and a standard select mapping other than right, it makes sense to use l/r to scroll the album list. if *any* of those conditions changes, we should probably use the standard prev/next mappings.
17:37:36kugelI agree
17:37:52kugelbut is there an easy way to detect without target specific #ifdefs?
17:38:14Unhelpfuli rather doubt it, unless we add a define in core that is based on target specific ifdefs. :)
17:38:47Unhelpfuland i see from the UI bmps that our development samsung targets have rather, um, interesting button layout choices.
17:38:55Tornefroggyman: also
17:39:23Tornefroggyman: if you report the info those pages ask for someone may be able to figure it out, but if you really want it to work then dumping the serial data is much more informative ;)
17:39:27Unhelpfuli have a YP-Z5 that i mean to *try* to work on porting whenever i find a charger for it :)
17:39:46kugelUnhelpful: interesting as in pain
17:40:06kugelit has no plugins yet because you can't have sane button mappings, basically
17:41:37Unhelpfulindeed. the Z5 is nearly as bad... or maybe worse? there's a touchpad in the center that registers touches along the edges as directional input, but can be pressed to select. it's framed by a square frame that can be pressed on any of the sides, which are labeled with functions. and there's a rocker on the side for volume.
17:42:42drakonikHm. I think I know how I could fix this problem. All the important stuff is codec in C, I suppose?
17:42:49kugelbtw, I still need testers for
17:43:18Unhelpfuldrakonik: except for speed-critical sections that may have some assembly, yes.
17:43:36 Quit jkl (Read error: 60 (Operation timed out))
17:43:51drakonikNow I gotta learn C and figure out how to make the database commit each song as it finds it
17:44:07drakonikWhich should be pretty simple.
17:44:24Unhelpfulthat's, um, not codecs, and why do you want that?
17:44:36drakonikNot codec
17:44:52drakonikUnhelpful: The db scanner crashes for some reason.
17:45:24drakonikBut because the scanner doesn't commit till it finishes scanning or the player is shut down,the crash means whatever GOOD files it finds don't get committed
17:45:50pixelmaI don't think changing the way it stores its entries will help you find out why
17:45:55drakonikNo, it won't
17:46:32Unhelpfulit will also make DB scanning much slower for the vast majority of users for (to them) no good reason.
17:46:39drakonikOh, of course
17:46:45drakonikI was going to use it for myself alone
17:46:51kugelpixelma: the ondio and c200 have some issues with PLA, correct?
17:47:09drakonikIf I have the database fully populated except for the bad files, I'll be able to find the bad ones by comparing the full list against the database list
17:47:32Slasheridrakonik: you can't commit the db by song by song, that would be _extremely_ inefficient
17:47:41drakonikWell aware of that.
17:47:46kugelassuming the db traverses in a specific sort order, which is probably a wrong assumption
17:48:09drakonikI'm just pretty desperate to get Rockbox working.
17:48:19Slasheridrakonik: the proper thing to do would be print the store the log of the db scanner in order to find the corrupt files and fix the problem
17:48:23pixelma_maybe_ you can find the troublesome files if you throw it into a simulator and initialise the db there?
17:48:24*kugel thinks Slasheri has auto-ping if someone mentions "DB" or "database"
17:48:42drakonikSlasheri: How would I do that?
17:48:43Slasherikugel: indeed, i have "database" in my hilight :)
17:49:06Slasheridrakonik: look at the function "add_tagcache" in tagcache.c
17:49:17*Unhelpful suggests compiling a database of games that are playable in rockboy per-target ;)
17:49:23drakonikGotta brush up on my C.
17:49:32Slasherithere you could add some code that adds the current file being processed to some log.txt file
17:49:37 Join toffe82 [0] (
17:49:42kugelpixelma: is that correct? If yes, can you try out please?
17:49:54rasherWhy, after so long, do we not have a better way of debugging database crashes than binary search?
17:49:58kugelwell, actually a test would be nice no matter of existing problems
17:50:17Unhelpfulrasher: on target i don't think there *is* a better way without a performance hit...
17:50:37rasherUnhelpful: There could be a "log database creation" option, in the debug menu or something
17:50:47Slasherirasher: we could always write a logfile or something like that but that would slow down the db building
17:50:54Unhelpfulkugel: does this seem less "miss" leading?
17:51:31pixelmaisn't there an entry that should show the currently processed file in the db debug menu?
17:51:35kugelmaybe just UP_DOWN_SCROLLING?, but yea, it's less misss leading :)
17:51:55Slasheripixelma: indeed there is
17:51:55rasherSlasheri: couldn't the debug menu have a "initialize database with log" entry?
17:52:11drakonikThat would be awesome
17:52:22rasherWhich enables a bit of logging code
17:52:50kugelSlasheri: btw, we were wondering what the count at building db actually counts. is that the total number of files and dirs traversed, or the found tags?
17:52:50Unhelpfulwe could also offer a pluginlib import of the tag scanner, with some defines that insert logging code, and a plugin to run it. zero cost to a non-debug run in core, that way...
17:53:02drakonikWhat I will need help with is compiling the code for my ipod.
17:53:21pixelmakugel: yes, some plugin controls don't work correctly on the Ondio and on the c200 due to PLA (and context "overlapping") and I might try the patch this weekend - but don't have much time (and energy) currently because of RL things
17:53:27Slasheripixelma: and entering "view db info" slows done the building a lot because the engine will wait for the screen to update the current file being processed. So that should already help if used
17:53:37Mikachudrakonik: there is a page or three on the wiki about setting up a compiling environment
17:54:04drakonikC first, wiki later, success eventually
17:54:07drakonikI refuse to give up
17:54:09Slasherikugel: it's the total number of files and dirs
17:54:16Unhelpfuldrakonik: i would suggest you try building the DB in a sim, presumably the problem is not in target-specific assembly.
17:54:29rasherdrakonik: okay, try starting database initialising and going into System > Debug menu (keep out!) > View database info
17:54:35drakonikUnhelpful: Alright. I just have no idea what that means.
17:54:38Unhelpfula crash in the sim can be trapped by a debugger...
17:54:49*pixelma wants some points for that ;)
17:55:17drakonikProgress: -1% (0 entries)
17:55:25drakonikI don't think that's quite right...
17:55:52drakonikNope. 100% wrong. I've got 700 songs in the db so far.
17:55:52rasherdrakonik: wait for a while, database creation is slow when you're in that screen
17:56:02drakonikNah, I checked it jsut now
17:56:12drakonikI'll clear out the db and reinitialize
17:56:21rasherI told you to initialize...
17:56:28drakonikAnd I'm doing it now.
17:56:43rasherWhy did you not do as I told you?
17:57:07drakonikI wanted to see what the db info had to say for a good database, before I wiped it.
17:57:10 Quit bmbl ("Bye!")
17:57:13rasherStart "Initialize Now", go to System > Debug menu (keep out!) > View database info and wait!
17:57:32kugelSlasheri: ok. it shows some 3k-4k on my e200, I don't think I have that many files and folders
17:57:37rasherProgress doesn't make sense if you're not in the middle of a scan...
17:57:41kugelbut you know it better :>
17:58:00Torneanyone got the gigabeat s updater version 1.1?
17:58:17Slasherikugel: it might count . and .. as files also, i don't remember
17:58:41kugelah, that would make more sense
17:58:43rasherSlasheri: err, does Curfile: not scroll?
17:58:59UnhelpfulTorne: i rather doubt it... i think the only one seen in the wild as an official updater is 1.3?
17:59:05saratogaSlasheri: having a log feature would be really nice, then we'd get more bug reports and could look into codec bugs that cause them
17:59:08Tornei have 1.2 and 1.3
17:59:12 Quit stoffel (Read error: 60 (Operation timed out))
17:59:25Tornethe toshiba page linked from the wiki has 1.2 and a different bit of toshiba's site has 1.3
18:00:01kugelI actually didn't know the database info screen shows that
18:00:04drakonikSome goddamn beatles song causes thecrash
18:00:16drakonikI've been looking for an excuse to get rid of that stuff
18:00:21rasherdrakonik: DONT
18:00:26drakonikOh, right.
18:00:29rasherFor the love of god. We want the file
18:00:34drakonikI'm not sure which one it is.
18:00:50drakonikThe debug page only shows like 15 chars and most of that was the directory info
18:01:03rasherdrakonik: try a smaller font
18:01:07rasherThere are some really tiny ones
18:01:19drakonikI just bumped my font size up, too
18:01:45rasherHum, it actually cuts off the display
18:01:50rasherSo that won't help much
18:02:03rasherMaybe it'll be enough though
18:02:03drakonikOnly 10 possible files
18:02:10drakonikLet's see...
18:02:10rasherSo remove half, try again
18:02:14UnhelpfulTorne: i never knew we had an official 1.2 updater... before the 1.3 one was found i'd never heard of an updater in the wild, part of the trouble with dual-boot was that the only nk.bin we had came from dumping from the HDD.
18:02:24rasher(and maybe remove or put a database.ignore to ignore the rest of your files)
18:02:48Torne has 1.2
18:02:56Torneyou have to go through a flash applet to get the real download link :)
18:03:07Unhelpfulor just make a "removed" dir with a database.ignore in it and move files in and out of it to "remove" them from scanning
18:03:19drakonikThat was my plan
18:03:29drakonikExcept explorer is locked up
18:03:31kugelmarkun: ping
18:03:33drakonikGo Windows!
18:03:57Tornesee now i'm unsure which binary to reverse :)
18:04:11 Quit saratoga ("CGI:IRC (Ping timeout)")
18:04:13Tornethe copies from the updaters are 'neater' in that they don't have garbage in unused areas etc
18:04:22Tornebut neither updater is the version that's in flash on the player I have :)
18:04:45Unhelpfulinteresting. is the 1.3 updater similar to what was in my ROM?
18:04:57Tornethe beginning is almost identical except in the unused areas
18:05:03Tornebut nearer the end it is not the same
18:05:25Tornethe certificates etc seem to start at a different address
18:05:36drakonikcrap, the DB info debugger didn't show any curfile info
18:05:39drakonikBut it still crashed.
18:06:05TorneUnhelpful: yours has the same checksum in the header as the 1.3 updater, at least
18:06:25Tornebut i'm not sure which bits that's a checksum *of*
18:06:34Torneit can't be the whole lot because yours doesn't match :)
18:06:51kugelI don't actually know why the simplelist is croping the entry there anyway
18:07:20Unhelpfulpresumable a checksum of the ranges of records in the .bin file?
18:07:43rasherdrakonik: drakonik maybe you weren't fast enough
18:07:45Torneanyway, yours looks like a different build, i have to say :)
18:07:52drakonikI'm gonna see what I can do
18:08:01drakonikUpdated the build, which should wipe the database and give me a clean start
18:08:04rasherdrakonik: try adding back some "irrelevant" files, and moving the suspect ones into a shorter path (a.mp3 b.mp3 etc)
18:08:05Tornethe ethernet bootloader's strings have a different date in
18:08:22Torneyours says "built jan 15 2007" and the 1.3 updater says feb 3 2009
18:08:23kugeldrakonik: updating doesn't wipe the db
18:08:34kugeldeleting the *.tcd files in .rockbox does
18:08:36drakonikoh, duh
18:08:39Torneupdating doesn't wipe anything if we can help it
18:08:42rasherkugel: there's no reason to do that
18:08:45pixelmaor initialise
18:08:48rasherJust do Initialise Now
18:09:02drakonikThat makes sense. Mine crashed, it never committed, there was no db ever initialized
18:10:36TorneUnhelpful: Ooh. Actually, now i have my brain turned on, yours is *much* more similar to the 1.2 updater
18:10:37drakonikOkay, battery is dead. Gotta let it charge up in ipod firmware mode for a lttle while
18:10:52UnhelpfulTorne: i wonder if the update always contains a complete ROM, or if they only put records in the bin file for specific ranges, and it's treated as a sort of patch...
18:10:53Torneeverything between 0x1000 and 0x30b20 is identical
18:12:04Tornein fact the only difference between yours and 1.2 is that some areas are data in yours that are 0 in the updater
18:12:15Tornepresumably because there was something bigger in your flash before? i dunno
18:12:25Torneso i think you actually have the 1.2 eboot in your flash.
18:12:51Torne1.2 -> 1.3 have very few changes in the bootstrap code, but more in the nk.exe
18:13:12pixelmais my suggestion to try and find the files in the sim that bad?
18:13:27drakonikpixelma: No, but it requires more setup than this
18:13:49drakonikOnce I've exhausted this, I'll move on to that. And once I've exhausted that, I'll move on to coding up a debug log.
18:14:11pixelmawell, you probably won't need to wait for some charging etc.
18:14:24UnhelpfulTorne: it did something suspiciously like writing to flash during the 1.3 update, and since then won't accept a rb-only nk.bin :/
18:14:45TorneUnhelpful: well, i can't explain that :)
18:15:14*Torne checks the ranges in 1.2's pmcboot_secure.bin to see if it's really all the same
18:15:40Unhelpfulinstalling an RB-only nk.bin triggers format->ask-for-firmware-upload
18:17:06Torneyup. every range of bytes that's present in pmcboot_secure.bin from the 1.2 update is exactly what's in your flash
18:17:20Unhelpfulvery odd. :/
18:17:23Torneand it doesn't seem to just be a patch
18:17:32Torneit seems complete as far as the CE dumping tools are concerned
18:17:39Torneit's happy to believe that's a bootstrap and nk.exe
18:17:50drakonikAwesome. The db initializes and crashes, but the debug screen only shows "−−-" for curfile.
18:18:18rasherdrakonik: As I said, add some unrelated files to delay the crash a bit
18:18:29rasherIt's probably crashed before you even reach the debug screen
18:18:32drakonikWell, it takes 30 or 45 seconds to crash
18:19:06drakonikAnd during that time, 'curfile' never changed to an actual file, it was just stuck at '−−-'
18:19:55rasherAnd yet, it actually did what we want it to the first time you tried.
18:20:07drakonikWhich is why I'm so weirded out.
18:20:20TorneUnhelpful: i think i'm going to try and reverse the 1.2 version for now, from the updater
18:20:32TorneUnhelpful: it doesn't match what's on the beast i have but hey.
18:20:49Unhelpfuldoes the beast you have do spurious reformats at all?
18:20:54Tornehaven't a clue
18:20:55rasherdrakonik: And yet you don't want to try anything else?
18:20:58Torneonly booted it twice
18:20:59drakonikI do.
18:21:04 Quit jon-kha (Read error: 145 (Connection timed out))
18:21:08drakonikI'm just working on them from least to most difficult
18:21:23TorneUnhelpful: it's not my dap, borrowed it from linuxstb
18:21:31drakonikI'm looking up how to do a sim like pixelma suggested now
18:21:40drakoniksince for whatever reason, the db debug isn't working
18:21:51rasherdrakonik: It is working. You saw it worke.
18:21:59drakonikYeah,but then it stopped working.
18:22:08*rasher sighs
18:22:16drakonikI dunno why, and I can't fix it.
18:22:19rasherIt is the same code. It did not "stop working"
18:22:28drakonikOf course.
18:22:41drakonikBut for whatever reason, curfile was no longer being updated.
18:22:57rasherAnd I say, probably because it was already done
18:22:59Unhelpfulor it was, after you removed the non-offending files, crashing before you could look at debug...
18:23:43TorneUnhelpful: hm, or maybe 1.3 since that way anything discovered is indirectly applicable to everyone ;)
18:24:52drakonikI'm going to start this install 100% fresh. latest build, no previous settings or db from before. All the non-relevant files are in a folder with a database.ignore file in it.
18:25:38rasherdrakonik: NO!
18:26:01rasherdrakonik: Do *not* put all the irrelevant files in a folder with a database.ignore file in it. Will you please listen to what we're telling you?
18:26:41 Quit MarcGuay_ (Read error: 110 (Connection timed out))
18:27:44drakonikThe first run I did watching the debug page, the scanner crashed on some Beatles song. I've taken all my beatles files and put them where the db can scan them. I'll eventually be able to pare it down to the one file causing trouble. Or multiple files.
18:27:45 Join MarcGuay_ [0] (
18:28:12 Join Nausicaa [0] (
18:28:12drakonikIf that's not the approach I should be taking, I misunderstood.
18:28:28rasherdrakonik: 1) It could be crashing on the file *after* the beatles songs 2) You've just told us that that approach means you did not get to see the debug screen while it was building.
18:28:39rasherdrakonik: Add some irrelevant files. A few albums or such
18:28:50rasherTo make sure you reach the debug screen before it's done
18:30:17drakonikYeah. Makes sense.
18:36:53kugelgevaerts: do you think anything speaks against committing the plugin-goto-wps work?
18:37:13 Quit Nausicaa (Read error: 104 (Connection reset by peer))
18:37:29 Join stoffel [0] (
18:38:47gevaertskugel: there's this bit in RFA that I don't really like that I mentioned last night, but I can fix that later. I seem to recall that some people didn't like the idea in general though, and I'd like to know why :)
18:47:06PaulPositionunhelpful - Thanks, that patch from made it all work, thanks. Would you mind posting it to the tracker ( FS #10404 ) in case there is need for further discussion and so it may some day make its way into 'current build' ?
18:49:01UnhelpfulPaulPosition: if it works, i don't think it really needs discussion.
18:50:39PaulPositionheh.. dunno about the inclusion in svn and closing task process so you may well be right. :)
18:52:15***Saving seen data "./dancer.seen"
18:52:36pixelmalot of keymap stuff can already be tried in a sim (at least the software part, not the electrical or mechanical limitations or awkward button placement though)
18:53:07 Join lee321987 [0] (n=chatzill@
18:53:10drakonikugh, I put in filler files and did the scan, but curfile isn't showing anything other than '−−-'
18:53:29drakonikHow fast does the db scanner work? Would it be able to do about 40 files in under 3 seconds?
18:54:28lee321987is there any possible way to update RB when disconnected from the computer (from an update on the external MicroSD card) on a c200v1?
18:56:27pixelmayes, if you have the unzipped .rockbox folder on the card you should be able to copy it over the other one on the internal memory (note: haven't tried on the c200 and my try on the Ondio is some years ago...)
18:56:46 Quit stoffel (Remote closed the connection)
18:57:28drakonikI think I found it.
18:57:36pixelmasome people also tried patched builds that can run off the microSD, if I remember correctly
18:57:47drakonikTried to play Lucy in the Sky and I got a crash
18:58:52drakonikExcept it didn't crash the second time I tried to play it.
18:59:52CIA-71New commit by 03unhelpful (r21636): Replace HAVE_SCROLLWHEEL in PictureFlow with USE_CORE_PREVNEXT, defined on targets where special horizontal-scroll mappings are broken or make no ...
19:00:34UnhelpfulPaulPosition: thanks for testing that, it should be in the regular builds shortly. :)
19:01:13PaulPositionThanks a lot mate :)
19:01:48AlexPUnhelpful: Was that to fix FS #10404 ?
19:02:50PaulPositionAlexP - it is.
19:02:52 Quit scorche (" rawr...that is all...rawr")
19:03:18AlexPUnhelpful: It is nice practice to mention the FS# you fix in the commit message :)
19:03:36UnhelpfulAlexP: i'm on that now :)
19:04:10AlexPIf it has already been committed?
19:04:25 Quit Grahack ("Leaving.")
19:04:32 Join doink12121 [0] (n=nick@
19:04:34Unhelpfuloh... yes, i see what you meant. i was addressing the open FS task. :/
19:04:55AlexPUnhelpful: Yeah, it is just useful to have blah blah blah - fixes FS#xxxx or whatever
19:05:21doink12121i am trying to get rockbox working on my sansa clip using ubuntu but it wont reconize the mount point and it keeps giving me an error
19:05:35AlexPUnhelpful: and if you do it as FS#xxxx it gets automatically parsed on the website so you can click on it and go straight to the task :)
19:06:23Unhelpfuland i try to do that, but forgot that there was a related FS task. thanks for the reminder. :)
19:07:01lee321987pixelma: It just worked on my c200. I thought maybe some of the executables couldn't be overwritten while not in USB mode. Thanks.
19:07:53AlexPUnhelpful: Hehe, just me being officious :)
19:08:16AlexPlee321987: No, Rockbox is copied to RAM then run
19:13:46drakonikalright, time to see what I can do in the sim
19:14:21 Quit robin0800 ("Leaving")
19:14:44doink12121how do i know if rockbox is working, i installed it but my player hasn't changed at all
19:15:58Tornethen it's not working
19:16:08Unhelpfuldoink12121: then it's not running, and you're in what we often call the "OF" (Original Firmware)
19:17:37 Quit HellDragon (Connection timed out)
19:17:38doink12121will rockbox ever work with the original sansa clip
19:17:56doink12121because i just google the models numbers from the sansa menu, and its looking quite grim
19:19:12 Join scorche [50] (n=scorche@rockbox/administrator/scorche)
19:22:56bertrikdoink12121, yes it will eventually work
19:27:04 Quit daurnimator ("Ex-Chat")
19:28:32 Quit fyrestorm ("lamers envy me like they envy bill g -- main boot xp, just the way it should be!")
19:32:13drakonikokay, copying music over to the simulated disk. shouldn't take more than six or seven hours :|
19:33:02Unhelpfuldrakonik: what OS are you using?
19:33:06drakonikWindows XP
19:33:21 Quit flydutch ("/* empty */")
19:33:31Unhelpfulugh... that makes things a bit more tricky, i was going to suggest you symlink it. :)
19:33:38drakonikI considered it
19:33:48drakonikBut I wasn't sure if that'd actually work
19:35:00drakonikI was exaggerating
19:35:06Unhelpfulit should work fine, but be very very very careful about what you do to the directory after that
19:35:17drakonikBut it takes an hour or more to copy my 7 gigs or so of music
19:35:22Unhelpfula recursive delete of the simdisk would descend into your player, i believe.
19:35:42drakonikDon't worry, I've abandoned the command line here
19:35:53drakonikWIndows is...not good for command line work.
19:37:58 Quit DarkSpectrum (Read error: 60 (Operation timed out))
19:38:24 Join DarkSpectrum [0] (
19:39:27 Quit linuxstb (Read error: 104 (Connection reset by peer))
19:40:23 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
19:40:47 Join HellDragon [0] (
19:41:12drakonikokay, symlink didn't work
19:41:38drakonikBut instead of copying the music into the disk folder
19:41:46drakonikI'll just move the disk folder to the music folder could do that. :)
19:42:01Unhelpfuli guess.
19:42:07drakonikIt's not a great solution
19:42:24drakonikbut better than duping 7 gigs of music on my hard drive
19:42:36Unhelpfulsymlink works fine on linux... and i'd hope Vista since that has "real" symlink. i'm surprised a junction on XP doesn't do the job.
19:42:54drakonikWell, a lot of things you'd expect to work in Windows don't
19:45:29Tornea junction would work
19:45:54Tornebut i am guessing drakonik isn't aware of them :)
19:46:05Tornewhich is not a surprise given that windows contains no tool to create them
19:46:14drakonikAha, there we go
19:46:18drakonikNow I'm getting useful data
19:47:11drakonikHow do I reboot the sim? If I close it and rerun the executable, that counts as a reboot right?
19:47:23 Part doink12121
19:47:56drakonikWell. Shit.
19:47:58drakonikThis is bad.
19:48:11drakonikThe sim loaded up everything flawlessly.
19:48:15 Quit flux (Read error: 60 (Operation timed out))
19:49:32drakonikWhat are the odds that I could copy .rockbox from the sim, into my ipod?
19:49:34drakonikand ahve it work?
19:49:46scorchewhy would you want to do that?
19:50:02drakonikBecause the shit on my ipod oesn't work for whatever reason
19:50:07drakonikand the shit in the sim does.
19:50:25drakonikSo, hopefully, if I copy the sim data into the ipod, it will work. would be better if you figure out why "shit" doesnt work on your ipod...
19:51:03Torneif metadata parsing crashes on target but not on the sim my first suspect is gonna be alignment, pretty much automatically ;)
19:51:16drakonikTorne: Alignment?
19:51:32Tornedrakonik: not something you can do anything about. a bug in one of the metadata parsers.
19:51:35drakonikScorche: I tried, but crashes are inconsistent and the debug tools don't work when I need them to
19:52:14Torneand no, the files in .rockbox are not even relevant to this, the metadata parsing is all in the main rockbox binary (and the ones from the sim wouldn't work anyway)
19:52:24drakonikI figured as much.
19:52:29drakonikBut I hoped
19:52:52drakonikSo alignment. What is this problem?
19:53:33Tornethe details are not very interesting unless you are going to fix the code yourself :)
19:53:45drakonikI couldn't
19:53:47Tornewe basically can't do anything unless you can track down which file(s) cause it to crash.
19:53:54drakonikThat's the problem
19:54:00drakonikit's not a single file
19:54:09drakonikSometimes it's one, sometimes it's another.
19:54:28drakonikAt least as far as playback goes.
19:54:42Torneit crashes on playback as well? i must've missed you saying that
19:54:45drakonikI can't get an accurate fucking readout of what the db is scanning when it crashes
19:54:59drakonikOnly sometimes.
19:55:26Unhelpful"a real OS" and "run under gdb" would be smart.
19:55:31drakonikLike, Lucy in the Sky With DIamonds caused a crash, but I rebooted and played it through twice
19:55:36scorchedrakonik: that is why we suggest doing a bunary search method to figure out which file(s) cause the DB scanning to have issues
19:55:36drakonikAnd no crash
19:56:07Torneit sounds like it's crashing in metadata parsing, at which point it's not the file that's currently playing which will be causing the crash, but the one that's currently being read into the buffer
19:56:11drakonikscorche: Yeah. I tried to do that, but because windows explorer is shit, it takes 15 minutes to SELECT half of the music in my collection, much less operate onit
19:56:13Tornei.e. a file several ahead of the one you are actually hearing
19:56:21scorchethen dont use windows explorer?
19:57:42 Join Hillshum [0] (n=chatzill@unaffiliated/hillshum)
19:58:18kugeldrakonik: have you weird tags in that files, such as embedded AA or long commments (weird as in not suitable for rockbox)
19:58:29drakonikI'm not sure
19:58:42drakonikI don't really know the limits on 'weird'.
19:58:50drakonikBut it's possible
19:59:07drakonikokay, I've got an idea for a test.
19:59:30drakonikNo. Nevermind, that's dubm
20:00:17CIA-71New commit by 03bluebrother (r21637): When changing TTS settings from the talkfile dialog make sure to not reset the currently selected folder if its valid. Fixes FS #10409.
20:03:30rasherdrakonik: does it crash in the simulator?
20:03:44drakonikLoads up 100% perfectly
20:04:07rasherdrakonik: With all your files, or just the beatles one you suspect to be the cause?
20:04:13drakonikAll of them
20:04:16drakonikMy entire music collection
20:04:22 Quit lee321987 ("ChatZilla 0.9.85 [Firefox 3.0.11/2009060215]")
20:04:32drakonikWhich is why I said 'shit'
20:05:38drakonikAlright. That option didn't work. Time to move on to the next level of difficulty: implement a db scan log.
20:08:42rasherShould be simple
20:09:03drakonikYeah, but i'm going to have to go to a lot of work to set up a compile environment
20:10:12HillshumI'm detailing a bug on my e200v2 in the forum thread and mentioning the data abort addresses I'm getting, where should I put my rockbox.elf? (I'm not up to disassembling it
20:10:23drakonikfucking reinstall cygwin with all the necessary components and I don't even knwo what I have to do. That's later. Right now, I've got to re-learn C, find a copy of the source that I can use, find the function to edit, edit it. THEN I can worry about setting up the damn compiler
20:10:40Mikachudrakonik: please don't use profanity in here
20:10:57 Join flux [0] (
20:17:10 Part PaulPosition
20:17:57 Quit kugel (Nick collision from services.)
20:18:03 Join kugel [0] (
20:19:04kugelgevaerts: what "bit in RFA"? Also, I can only remember JdGordon having some doubts, and he wasn't able to explain why
20:21:24 Join desowin [0] (n=desowin@atheme/member/desowin)
20:22:25rasherdrakonik: which target do you have?
20:22:41drakonikipod video. Which is what you're asking, I think
20:25:53drakonikWhy do you ask?
20:26:03rasherDoing a build
20:26:14rasherWith some very dumb logging
20:26:28drakonikDumb logging is what I need. I am very dumb and it's gotta be on my level
20:27:46rasherdrakonik: try this:"> (patch here:
20:27:57rashershould create a database.log file in the root
20:28:11rasherWhere it writes the name of each file as it's about to open it
20:28:21drakonikThat's perfect
20:28:35rasherWell, hopefully
20:29:04rasherIt's going to be a great deal slower
20:29:20drakonikI have nothing but time.
20:32:01 Quit Hillshum (Read error: 110 (Connection timed out))
20:35:47kugellet's add database.hardignore
20:35:54 Quit tvelocity (Read error: 104 (Connection reset by peer))
20:37:51rasherHow about no
20:38:07 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
20:38:40saratogarasher: you should consider making that patch sim only, and then commiting it
20:38:47saratogalogging on the sim would be a very useful feature
20:38:59Mikachuit didn't crash on the sim
20:39:10drakonikYeah,but logging on the sim is still useful.
20:39:20rashersaratoga: you can always enable logf on the sim. The database does a lot of logf
20:40:13kugelrasher: why that
20:40:29rasherkugel: Because it's a hack on top of a hack
20:40:33saratogarasher: yeah but something a little easier for non-developers to use would be nice
20:40:35drakonikand we've got our log.
20:40:41drakonikTime to see the culprit.
20:40:52saratogaideally we'd be able to point people at the EXE and tell them to give us the resulting log
20:41:08rashersaratoga: Problem is it didn't crash in the sim at all
20:41:14kugelrasher: then let's remove the first hack. Although markun apparently wants that
20:41:29saratogarasher: I saw, which makes me assume its a hardware problem on whatever target
20:42:20rashersaratoga: so it wouldn't really have helped us
20:42:36drakonikhahahaha wow
20:42:41drakonikThat's a very short log
20:42:52rasherdrakonik: Oh?
20:42:58drakonikyeah, 11 lines long
20:43:02rasherWhatever you do, remember not to delete the file..
20:43:15rasherSo, you have a suspect file now?
20:43:15drakonikIt's not a music file at the end of the log, though
20:43:25drakonikI think it was put there by one of the chkdsks I ran
20:43:44rasherIt must look like a music file then
20:44:08drakonikLet's look at the file
20:44:15kugelI really think we should offer some speed for more helpful diagnostics
20:44:16rasherWhich is not unlikely, if it's a chkdisk file
20:44:39kugelat least until those diagnostics helped us to kill common failure issues so that we can go for speed again
20:44:43drakonikIt's called 'clock'
20:44:53drakonikSo I'm pretty sure it's part of the old ipod firmware files
20:44:54saratogarasher: most of the time its due to an actual problem in the tags though
20:44:59kugelchkdisk doesn't add this
20:45:06saratogain which case the sim and target should be equivilent
20:45:08rasherdrakonik: can you share that file somewhere?
20:45:14kugelit adds a folder FOUND.00X with dozens of files
20:45:31rashersaratoga: I've no problem with adding that patch in #ifdef SIMULATOR
20:45:48drakonikrasher: yup
20:45:56drakonikuploading it now
20:45:57kugelor we go for that plugin
20:46:11kugelthat does the stock db scan but more verbosely
20:49:11drakonikAnd I misread the log, radio was the last one opened
20:52:17 Join petur [0] (
20:52:20***Saving seen data "./dancer.seen"
20:52:24Slasheriplease check the last music file
20:52:24 Quit DarkSpectrum (Read error: 60 (Operation timed out))
20:52:37Slasherirasher: or please put that logging after the probe_file_format() :)
20:52:55drakonikSlasheri: There were no music files
20:52:57rasherSlasheri: Hm, don't we want to catch as much info as possible?
20:53:02saratogafiles without extensions don't even make it into the metadata code if I understand correctly
20:53:18saratogaso they should be harmless
20:53:19Slasheridrakonik: oh, so you didn't have music files at all?
20:53:28drakonikI do.
20:53:32drakonikBut the scanner didn't get to them.
20:53:35rashersaratoga: There's no guarantee that it's a bug in the metadata code drakonik is hitting
20:53:46drakonikIt crashed when it tried to scan FOUND.001\radio
20:53:48Slasherihmm.. really interesting
20:53:57saratogathis is on the sim?
20:54:07saratogabad hardware
20:54:12rasherdrakonik: humour me. Try doing the same again
20:54:16rasherDelete the log first
20:54:17drakonikrasher: Way ahead of you
20:54:26rasherI wouldn't be surprised if it was a different file this time
20:54:35Slasherihmm.. maybe reading that file crashed the FS layer of rockbox
20:54:42kugeldrakonik: thats a file created by chkdisk
20:55:27Slasheribecause database shouldn't process files without known extension
20:55:30rasherAnd yeah, I'm leaning towards bad hardware
20:55:47drakoniksame file
20:56:16kugelif chkdisk already creates files, the fs or hardware seems to be broken
20:56:30rasherkugel: could be simple fs corruption
20:56:42kugelthat's what I said
20:57:05rasherBut it shouldn't be crashing... Plus you've been saying how explorer is very slow to list files on the disk. I think your disk is bad
20:57:45kugelSlasheri: can we change the counter to only count recognized files?
20:58:13kugelthat would be awesome. that would make much better progress informationo
20:59:36Slasherikugel: yes, i have thought about doing that many times already
20:59:55drakonikrasher: It's everywhere
21:00:01drakonikon my hard drive, on my thumb drive, everywhere
21:00:07saratogarasher: so you're going to add an option for that in the sim? or at least post a sim build that has it enabled?
21:00:14kugelSlasheri: And why didn't you just do it?
21:00:14Slasheriinitially the reason for this obscure counter was that with dircache it would know the number of files and show the actual progress in percents
21:00:38Slasherikugel: i just haven't been so active lately :) too much work to do
21:01:00saratogaSlasheri: did you work on the disk cache stuff?
21:02:25drakonikI'm going to make a copy of the file to my hard drive, remove it from the ipod, and try the scan again.
21:02:43drakonikIs this a bad idea?
21:03:09Slasherisaratoga: the dircache? i wrote that
21:03:54saratogai've wanted to add a way to merge the microsd file directory and interneal file directories into one tree for a while, could you speculate on how hard that would be to do?
21:05:19drakonikIf it is, it's too late to stop. Scan's going, log is building up.
21:05:20rashersaratoga: I don't see why it even needs to be an option in the sim. Just always do it
21:05:34kugelyea, it's logf
21:05:51pixelmasaratoga: I remember amiconn having a good reason for not doing that... wasn't it a problem with what would happen if you have the same file (in the same dir structure) on both volumes?
21:05:57kugelsaratoga: a nice feature, definitely, but clashes are sort of very likely
21:06:43saratogakugel, pixelma: XBMC already supports it as Llorean pointed out to me, and it works very well in my testing
21:06:52saratogaduplicate files are simply ignored
21:07:14saratogano just one copy
21:07:24pixelmaand what would be the preferred one, e.g. when you want to copy/paste etc.?
21:07:43kugelI'd actually be fine with that
21:07:49rasherJust by convention
21:07:56rasherOr whichever drive holds the directory you're in
21:08:01saratogaor just don't allow copy and paste while this is enabled
21:08:04saratogaeither is fine really
21:08:07kugelit needs to be internal I guess. what if you have .rockbox on the µSD
21:08:11pixelmahow to get access to the other one then?
21:08:14linuxstbsaratoga: You mean files with the same name, or duplicated files? And how do you specify where to save a new file to?
21:08:19rasherpixelma: disable this feature
21:08:23drakonikthis time, it crashed on reading database.log
21:08:45kugeldrakonik: it really seems to be a bad FS or even HW actually
21:08:48rasherlinuxstb: It gets saved to the drive holding the directory you save in
21:09:02linuxstbrasher: Where is the root?
21:09:06kugelwhat if the dir exists in both?
21:09:06saratogain xbmc I think it just disables writing to directories that have been merged, at least I couldn't figure out a way to
21:09:07rasherlinuxstb: Internal.
21:09:16rasherlinuxstb: As per the made-up convention
21:09:18linuxstbrasher: Then how do you access the root of the external?
21:09:23kugeldrakonik: hardware
21:09:35saratogalinuxstb: just unmerge them
21:09:35rasherlinuxstb: You don't. You can access the files there though
21:09:36pixelmatoo much complication for a fwature I'd probably turn off anyways. If you want one list you could already do that now - with the database
21:09:42pixelmafeature too
21:09:48drakonikI dunno, I think it it was bad hardware, it'd have shown symptoms before I installed rockbox
21:10:00saratogaits actually much simplier to use then you'd think
21:10:16*rasher doesn't think it'd be terribly complicated either
21:10:25linuxstbsaratoga: So this would be an option, rather than a change in behaviour?
21:10:31drakonikUntil I did that, it wasn't crashing ever.
21:10:41kugelsaratoga: to use, yea. but how complex is the code?
21:10:57rasherWith hybrid wavpack, you could have the correction-files on an sd card. A special losslessify-card!
21:10:59saratogaoriginally i just wanted to hack up my own build, but Llorean said he thought it could be commitable
21:11:09kugelalso, I think that would also be only usable with dircache at all
21:11:29saratogakugel: I'm not sure, i haven't looked into it, just the rockbox file tree code a little
21:11:44rasherI'm not convinced it'd be worth the delta, but the feature itself seems fairly straightforward to me
21:12:04kugelthat wouldn't be happening in the apps/ level I think, but in the firmware/
21:12:21saratogawell it only impacts targets with flash memory slots, so the delta is only on the sansas and maybe the onda [?]
21:12:22linuxstbI don't use any targets with *sd cards, but can see how it would be convenient for people who do.
21:12:32pixelmaI don't enable dircache, even on my c200 - no use for it on a flash target (and not a real "user" of the db)
21:12:35kugelwindows 7 implements it too, seems to be a new trend
21:12:51rasherlinuxstb: Or people with multi-partition archoses
21:12:54kugelpixelma: dircache has nothing to do with the db
21:13:01linuxstbrasher: Both of them?
21:13:03rasherThough that's probably a rather small userbase
21:13:09saratogapixelma: dircache is still useful with on flash targets, skimming through flash memory is quite slow compared to ram
21:13:30pixelmakugel: it has - for speed reasons when updating
21:13:44kugelyes, but not for browsing
21:14:01saratoganot to mention you don't get the sansa electrical noise when browsing the file tree so often
21:14:07kugeldircache is a feature on it's own. the database might make use if it, of course
21:14:41pixelmasince I don't update but only initialise, it's no use... and I did not feel access to dirs is slow
21:14:52kugeland it does speed up file browsing. I notice it on my e200. And I miss it on my fuze
21:15:19saratogahmm Slasheri escaped
21:15:44kugelnot to mention it saves waking up the SD controller, yielding in actual battery life
21:16:06pixelmaumm... but then it takes RAM away...
21:16:37pixelmaso that's a loss again if you don't browse too often and just listen to music ;)
21:16:44kugel500k of 30M is neglible
21:17:11kugeland you won't even reach 500k on a c200
21:17:17Tornegah, why does the gigabeat system diagram not number the GPIO lines?
21:17:23Tornethey;'re all labelled "GPIO"
21:17:27Tornebut it doesn't say which one is which :0
21:18:07kugelfor the fuze, it *might* be a loss, but almost certainly not for the c200v1
21:18:11Tornethe rom is looking at bits 10 and 11 in GPIO3, which is not labelled in any of the diagrams and isn't used in rockbox afaict
21:18:26rashersaratoga: So you think I should commit it (for the sim)?
21:19:20saratogarasher: yes absolutely, though I wonder if theres any consequence to having that log generated everytime?
21:19:24kugel"for the sim"?
21:19:30saratogathough i suppose not since most people won't have the database on in the sim anyway
21:19:34pixelma311k with an 8GB microSD and my current files/folder structure
21:19:45kugelwhat's the problem with using logf and have it always enabled
21:19:55saratogatheres really no cost to having a smaller compressed audio buffer on the sansas
21:20:13rasherkugel: you mean always enable logf in the sim?
21:20:16saratogathe flash memory spends the same amount of time energized with a 1MB buffer as it does with a 30MB buffer
21:20:29kugelrasher: no
21:20:31saratogawith logf you get a lot more then just the database thought right?
21:20:44kugelyou enable it per-file
21:21:02kugelI think it should be a plugin, always available, actually
21:21:06rasherkugel: Right, so always enable logf, and enable it for tagcache.c for the sim?
21:21:10 Join matsl [0] (
21:21:26rasherkugel: sure, but that takes someone to write it, which hasn't happened for 3 years now
21:22:02kugelSlasheri: ping
21:22:22kugelrasher: meh, I don't know
21:23:35 Join notlistening [0] (
21:24:43kugelUnhelpful: have you seen my plugin-goto-wps patch?
21:25:04Unhelpfulkugel: i've not really looked at it...
21:25:10kugelrasher: anyway, if you commit it, do some error checking :)
21:25:20rasheroh er
21:25:57kugelthe log should probably also stored in .rockbox/
21:26:41rasherYeah, I already did that
21:26:52kugelUnhelpful: please do so :>
21:28:08Unhelpfuli probably can't do so in depth until tonight. we're going to a picnic shortly. i'll try to remember to look then.
21:28:50CIA-71New commit by 03rasher (r21638): Crude logging for the sim in database creation/updating - to aid in debugging
21:29:03*Unhelpful still wonders what exactly happens when the plugin loader is called from somewhere unusual... shouldn't a plugin started from the WPS put you back in the WPS on exit?
21:29:25rasherUnhelpful: there's credits.rock
21:30:42Unhelpfulrasher: i can't start that from WPS any way that i know of, though...
21:31:22Unhelpful...but i can open rockpaint. :)
21:31:29rasherBut you can start it from a "weird" location
21:32:00rasherOr is that not weird enough?
21:32:26Unhelpfulkugel: i'm not sure we need an "exit to wps" hook. if you use the context menu in WPS, you can open the current file with a viewer. doesn't matter if it works... open a currently-playing file with rockpaint, it'll try to open it, fail, and give you a blank canvas.
21:33:11Unhelpfulwhen you quit rockpaint, you're back in WPS. i think that being able to launch a plugin from WPS will suffice to take the user back to WPS on exit.
21:34:05Unhelpfulthe same works with rocklife
21:38:46kugelUnhelpful: You're getting it wrong I think
21:38:59kugelit's not about launching a plugin from the wps
21:40:02Unhelpfulkugel: it at least partly is. the ability to "go to WPS" is a non-issue if we assume that the plugin is started from there...
21:40:11kugelit's about launching from everywhere, end exit to the wps (there's also really no hook involved, it's all managed by the root menu with existing code). May it be a plugin used in the core for consistency, or a music-orientated plugin
21:41:29kugelUnhelpful: this assuming doesn't work
21:41:50rasherWhy do you want a plugin to exit to the wps?
21:42:14Unhelpfuli'm assuming the main case would be if it's going to start playback?
21:42:36kugelyes, in the case of pf and rfa which I adapted for now
21:42:54kugelbut I'd also like properties and credits to be adapted, for consitency
21:43:42kugelUnhelpful: I gave you the wrong diff, btw
21:43:49rasherkugel: Why would those want to exit to the wps?
21:44:14kugelWhy not?
21:44:15rasherYou can open properties from the wps, but then exiting it will bring you back anyway, with the current system..
21:44:23rasherAnd credits.. is opened from the menu
21:44:26*rasher fails to understand
21:44:29kugelThose are used in the core, where you basically can go from everywhere to the wps
21:44:50rasherAhh, with the go-to-wps button. Right
21:45:06kugelYou can go to the wps from the context menu, but not 1 menu item farther (properties)
21:45:09rasherSee, that was a better answer than "Why not"
21:45:32kugelit's really meant to be optional
21:46:32kugelUnhelpful: that's the correct:
21:46:40kugelthe code added to the core is minimal
21:46:53 Join AndyIL [0] (n=pasha_in@
21:47:56kugelUnhelpful: btw, I get audio skips when scrolling in pf on my fuze
21:48:20kugelalso, I see the empty slide when scrolling quickly. I never get either of those on my e200
21:51:46gevaertsmostly nitpicking I guess
21:52:44kugelwe can do that
21:53:14kugelI don't actually know why main_menu is called in a loop, instead of having the loop in main_menu
21:53:38gevaertsrasher: this "play shuffled" option in RFA starts playback and exits the plugin. In cases like that, I think going to the WPS is a good idea
21:53:44kugelit seems people don't really like exit() :)
21:55:04gevaertskugel: I don't mind exit(), but I don't like this single function using exit() half the time and return the other half. That will just make things hard to read
21:55:23kugelhm, using exit seems not a good idea, if we want to stick to standard behavior of exit() (if there's a defined standard)
21:55:57kugelwith my change, exit(!=0) may not result in an error
21:56:14kugelespecially not exit(1)
21:56:32kugelgevaerts: I agree with you
21:56:48kugelI just wanted a quick way to get it to work without rewriting the whole plugin
21:57:26 Quit AndyI (Read error: 110 (Connection timed out))
22:00:07gevaertskugel: :)
22:00:22 Join robin0800 [0] (
22:02:32kugelgevaerts: alright. our exit has nothing todo with the standard anyway :p
22:03:11kugelalso, they don't define what EXIT_FAILURE is, so we can get away with defining to a value that fits us
22:04:37saratogado we mind that guy posting links to patched Sandisk firmwares on the forums?
22:04:46Mikachuif the value was defined, the name wouldn't be needed :)
22:04:50saratogai don't personally but if he doesn't have permission to distribute them that seems like a bad thing to do
22:05:50kugelMikachu: well, EXIT_SUCCESS is basically defined though
22:06:12kugelnot exactly, but why would you define EXIT_SUCCESS not as zero
22:07:38 Join bluebrother [0] (
22:08:24rashersaratoga: That's not good.
22:13:51 Join p3tur [50] (n=petur@rockbox/developer/petur)
22:21:03 Join DarkDefender [0] (
22:21:45kugelgevaerts: OK, I have a exit()-less version now
22:28:17saratogadatabase logging seems to work nicely in the sim for me
22:29:57kugelwhat the hell
22:30:13kugelgcc tells me about unused variable, but its clearly used
22:30:21kugeloh, I see
22:30:21saratogamake clean
22:32:27 Join Ubuntuxer [0] (
22:34:27rashersaratoga: You should be able to point people at"> within 24 hours to get a build with that
22:34:36saratogaok cool
22:35:00rasherOr I could rebuild manually..
22:35:05rasherMight as well.
22:37:53 Quit {phoenix} (Remote closed the connection)
22:38:10gevaertskugel: actually, EXIT_FAILURE and EXIT_SUCCESS are defined elsewhere. "expand to integer constant expressions that can be used as the argument to theexit function to return unsuccessful or successful termination status, respectively, to the host environment"
22:38:20gevaertsNot that it matters :)
22:39:57kugelsomehow it doesn't surprise me that you tell me that now :>
22:42:28kugelgevaerts: new patch, even smaller
22:44:21 Join __lifeless [0] (n=lifeless@
22:45:02gevaertskugel: looks good as far as I'm concerned
22:45:29kugelmost of it is changing the indentation due to rewriting main_menu() a bit
22:48:07Mikachu-w for diff ignores whitespace changes, if you want to look at only the functional changes
22:48:12Mikachu(which you probably already know)
22:49:19kugelI heard of it yes, but I wouldn't say that I *know* it :p
22:50:51GodEaterseems the linux4nano guys think they have the decryption keys now
22:51:27 Quit r0b- (Read error: 110 (Connection timed out))
22:51:43 Join r0b- [0] (n=rob@
22:51:49notlisteningdoes the mean the encryption key for all ipods 6th gen or just nanos?
22:51:50kugelI'm not in the channel, but the guys have already claimed much
22:52:22***Saving seen data "./dancer.seen"
22:52:31 Join {phoenix} [0] (
22:52:58GodEaterkugel: you don't believe them ?
22:53:06 Quit matsl (Read error: 110 (Connection timed out))
22:53:52kugelI have my doubts, but I'm not really following the whole thing anyway
22:54:18GodEatergiven planetbeing's involvement, I trust they have genuine progress
22:57:03*scorche points GSoC mentors towards -gsoc
22:57:12 Quit saratoga ("CGI:IRC")
22:57:24scorchekugel: they arent all like Taylor...
22:57:27 Join webguest37 [0] (n=9803c6dd@gateway/web/cgi-irc/
22:57:29 Quit webguest37 (Read error: 104 (Connection reset by peer))
22:57:35 Join saratoga [0] (n=9803c6dd@rockbox/developer/saratoga)
22:57:52kugelthere's more than taylor? :>
22:58:15GodEaterkugel: read their logs from last night for a taster
22:58:21GodEaterthey pulled the bootrom to bits
22:58:37saratogacan they decrypt the NOR then?
22:59:06GodEatersaratoga: that's what they seem to think
22:59:13GodEaterthey won't REALLY know until they can dump it
22:59:20GodEaterwhich is what they're nattering about at the moment
22:59:29 Quit _lifeless (Read error: 110 (Connection timed out))
22:59:36saratogathey already dumped the NOR though last year
22:59:42kugelGodEater: where are the logs?
22:59:51GodEaterkugel: one sec
22:59:55kugelI saw their site recently, where you can get no information about the project
23:00:47AlexPGodEater: Those logs are password protected
23:00:52AlexPThey are a bit paranoid
23:01:03GodEaterAlexP: you want the username / pass too ?
23:02:46bluebrothersounds like "open" development :)
23:03:02 Join n00b81 [0] (n=taylor@unaffiliated/n00b81)
23:03:37rasherbluebrother: Yeah, you basically have to wait for a black box to appear, after which you can pick it apart and create something useful for Rockbox, I guess
23:04:05rasher isn't very useful either..
23:04:24GodEaterthey're being very lazy about documenting it
23:04:44 Quit yosafbridge ("Coyote finally caught me")
23:04:54 Join yosafbridge [0] (
23:05:12*scorche pushes Bagder into -gsoc
23:05:35*kugel join #rockbox-gsoc
23:05:52 Join slyyx [0] (n=slyyx@
23:06:11kugelGodEater: no wonder that they made progress. Many rockers around there too
23:06:28slyyx!seen onlysoaa
23:06:48slyyxIs there a !seen?
23:06:54GodEaterkugel: indeed :)
23:07:07GodEaterwho the hell is onlysoaa ?
23:07:19*p3tur walks inside.. getting too dark :/
23:07:42slyyxGodEater: This guy:;u=2222
23:08:32GodEaterI don't think he's ever been in here.
23:08:48slyyxMonths ago he was very active. Working on a p2 port.
23:09:00slyyxTwo months ago or so.
23:10:03GodEaterin here though ?
23:10:15slyyxYes, patch number #10280
23:10:19GodEatermust be in a different time zone to me
23:10:50slyyxAs I recall, usuaually from...3-5hrs from now he would get on.
23:11:00pixelmaslyyx: ask logbot... "logbot seen nick"
23:11:02kugelGodEater: I admit, impressive
23:11:11slyyxlogbot: seen onlysoaa
23:11:21gevaertsslyyx: /msg will work
23:11:24AlexP/msg logbot :)
23:11:48 Join Zagor [242] (
23:12:00slyyxAlright, 26 days ago....*shrugs*
23:12:06 Part slyyx ("Reality is that which, when you stop believing in it, doesn't go away")
23:12:35pixelmaI don't remember /msg being necessary (but working)
23:13:06kugel!seen onlysoaa
23:13:08Bagderonly recognized users at some level can do it in public
23:13:13Bagderand not with !
23:13:16kugelalso doesn't work
23:13:23Bagderit never did
23:13:32pixelmaah, that explains it
23:13:50kugelI never tried seen before actually
23:15:16kugelJdGordon: ping
23:16:06 Quit robin0800 (Success)
23:16:30kugelGodEater: I'll be laughing if we get rockbox on the newer nanos before they get linux :p
23:16:52GodEaterkugel: well they're rooting for us too
23:16:56saratogai'd assume thats more likely given how many more people we have then ipod linux
23:17:10rasherIt's not terribly unlikely. They have some hurdles in the whole linux department that we don't
23:17:19n00b81You probably will. Most of the people at ipodlinux are busy :)
23:17:21GodEaterwe have to tip our hats to the work they've done there though
23:17:36GodEaterI'm so pleased planetbeing spared us his time
23:17:41Bagderbusy is another word for dead these days? B)
23:17:50n00b81Bagder: yeah, kinda :)
23:18:07n00b81Hey, extracting a bootloader through a notes vuln is not the easiest thing on this planet :)
23:18:36saratogaat any rate maybe this will ram the meizu port through too
23:18:45saratogathough i guess the FTL issue is still looming
23:18:49BagderI don't think you need to tell this channel about how reverse engineering can be hard;-)
23:18:55ZagorBagder: which script creates the current build table?
23:19:04kugelGodEater: we should team up for sure, I was kidding
23:19:06GodEaterdo any of our meizu guys have any idea how to get the UART working
23:19:21n00b81That would be a big help :-)
23:19:31saratogai doubt it since its not something we'd actually need, but i assume its in teh datasheet?
23:19:54GodEateryeah, I suspect our chaps have spent longer looking at the datasheet than they have
23:20:15notlisteninglol @ Badger, my head is still hurting from even looking at the SansaAMS and i believe that was not the biggest challenge
23:20:57Bagderthe sansaAMSes was what I would call medium hard
23:21:15Bagderthe PP ones were harder
23:21:21kugelthe PP were definitely harder
23:21:30Bagderand these new ipods are on the top of the scale
23:21:31kugelno datasheets at all.
23:21:49BagderI figure perhaps zunes are up there too
23:22:09Bagderdarn, where's those zunelinux guys when we need them! ;-P
23:22:55saratogachapter 25: UART
23:23:00saratogasounds promising :)
23:26:05 Quit p3tur ("back inside....")
23:28:54 Join CaptainKwel [0] (
23:33:12pixelmabluebrother: for the texts in the manual that mention actions, I'd like to have a macro that - depending on HAVE_REMOTE_KEYMAP defined expands to either "\ActionSomething" or "\ActionSomething (\ActionRCSomething) where Something is passed through the new macro and then the actual button for this action is looked up in the keymap files - is this possible?
23:34:17mtlinuxstb: I totally forgot about declaring the local functions static, will do. About the executable property, it was added by mistake, but will be removed in the next patch. I never use TABs, so it might just be a misalignment, I'd appreciate it if you could point me to the specific part you're talking about :)
23:36:40bluebrotherpixelma: can't tell that off the top of my head −− never needed something like that. Will check it
23:37:48linuxstbmt: Searching your patch, I found some tabs, but I forget exactly where. Some of them are in SVN already though I think...
23:37:53AlexPpixelma: HAVEREMOTEKEYMAP :)
23:38:10linuxstbmt: Do you have any big-endian targets to test it on?
23:38:30pixelmabluebrother: well I thought the wikilinks etc. are a bit similar as you can feed them with the wiki page name and it will make a proper link out of it, but couldn't figure out how that works or where the limits are
23:39:13pixelmaAlexP: sorry, was off the top of my head and for the explanation, would have looked it up in real use
23:39:25mtlinuxstb: No I don't. About the tabs, are they in files other than makefiles ?
23:39:35AlexPpixelma: Hehe, it wasn't serious :)
23:40:55pixelmabluebrother: although the wikilinkks are "one step" whereas my idea would be two steps (combine a string and use it for the lookup in the keymap files)
23:40:56bluebrotherpixelma: well, the wikilinks macro constructs an argument for another macro while you want to construct a macro that should get evaluated, so it's different.
23:42:48bluebrotherI'm wondering how hard it would be to create a new LaTeX macro package that reads a keymaps file directly. Unfortunately it's unlikely I'll find time working on it (and I doubt I'd succeed anyway)
23:43:16pixelmawould be the nicest solution I could think of until now for the problem that you sometimes need both and sometimes one each (in the tables) - and without preparing a lot of combined actions or opting on every occurance in the text
23:43:39pixelmaI mean RC action and non-RC action
23:46:07 Join barrywardell [0] (
23:48:06 Quit notlistening ("Leaving")
23:49:50linuxstbmt: Yes, of course.
23:50:11ZagorBagder: did you have a reason for putting handoutbuilds() so high up in COMPLETED? it's a bit inconvenient since that clears $buildround on the last build
23:50:57 Quit {phoenix} (Remote closed the connection)
23:51:00*pixelma is reminded of the start of an M3 manual waiting for a commit, even though it doesn't compile cleanly yet
23:51:19rasherpixelma: It doesn't compile anyway, does it?
23:51:22BagderZagor: I think because handing out builds is a priority
23:52:10pixelmarasher: hmm?
23:52:39pixelmaI don't understand the question
23:52:48rasherpixelma: Just saying that it's still a step in the right direction
23:52:54ZagorBagder: ok. I'll simply copy $buildround to a local var and leave handoutbuilds() where it is.
23:53:55pixelmaah yes... that's why I wanted to commit it but forgot due to being occupied with other things
23:54:46mtlinuxstb: I'll chase them tomorrow then :) .. I want to finish cleaning this patch and committing it tomorrow.
23:56:09mtFor BE targets, First thing that comes to mind would be modifying the get_uint* functions according to the endianess of the target (like in metadata.c/h) ..
23:56:10Zagorhmm, no I'll move it below the file moves. we need to clean out upload/ in endround.
23:56:31BagderZagor: good point
23:57:36Zagoralso I'm changing the filenames in data/ to $revision-$id.* to more easily tie them to the build table
23:59:00rasherConsidering that there's now an "abort build" command, maybe the build round should start immediately on commit, and then restart if a commit is detected within the grace period?

Previous day | Next day