Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

#rockbox log for 2009-11-07

00:00:31 Quit hebz0rl (Remote closed the connection)
00:12:45 Quit freddyb (Read error: 110 (Connection timed out))
00:13:10 Quit StealthyXIIGer (Read error: 145 (Connection timed out))
00:20:53CIA-6New commit by bluebrother (r23545): Set svn:keywords property.
00:20:55 Part froggyman
00:25:46 Join Strife89 [0] (
00:32:16bluebrotherhmm. bin2c has quite a lot of similar variants.
00:33:58 Quit FlynDice (Read error: 110 (Connection timed out))
00:34:59 Quit Strife89 ("Gotta go.")
00:37:21 Quit JdGordon ("Leaving.")
00:38:51 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
00:45:54 Join Wafelijzer [0] (
00:45:58 Join Casainho [0] (
00:46:17Wafelijzerhello people
00:46:55bertrikhi Wafelijzer
00:47:25 Quit TopyMobile (Read error: 60 (Operation timed out))
00:47:40Casainhohello Wafelijzer :-)
00:47:50CasainhoWafelijzer: so, do you have a Mini2440 board also?
00:47:58Wafelijzerin the upcomming if time permits im going to try rockbox on my mini2440
00:48:08Casainhook, nice!
00:48:32Casainhoand for what would you like to use it? car player? home player? or just to test it?
00:48:57Wafelijzerit works, and i was busy making my own linux distro but had sound issues
00:49:11Wafelijzerthe nice thing about my distro is that is runs MPD
00:49:17Bob_C_What sort of sound issues?
00:49:24Wafelijzerno sound at all :-)
00:49:31Wafelijzerlinux kernel ...
00:49:46Wafelijzeri tried a hack, and then the sound was pulsed
00:50:14Wafelijzersomething with bad clock i think
00:51:42CasainhoWafelijzer: ok. Go and try the code - do you know how to get it?
00:52:06Casainhoand for example, Bob_C_ is working on a HDD interface, if you are interested
00:52:09Wafelijzerbut i dont care for the moment because i have school work to do, but i bumped to the rockbox project from the blog of mini2440
00:52:19Wafelijzerthats cool :-)
00:52:46Casainhowhat link of that bolg? on
00:54:26Casainhowe are looking also for cases....
00:54:45Wafelijzeryou mean enclosures?
00:54:46Casainhomaybe we should talk on #rockbox-community
00:55:07CasainhoWafelijzer: let's join that channel and talk there the things out of topic here ;-)
01:03:31 Join freddyb [0] (
01:05:32 Quit Casainho ("going to sleep")
01:06:56 Quit pixelma (Read error: 60 (Operation timed out))
01:06:58 Quit amiconn (Nick collision from services.)
01:06:59 Join pixelma_ [0] (i=quassel@rockbox/staff/pixelma)
01:07:01 Quit Sajber^1 (Read error: 54 (Connection reset by peer))
01:07:02 Join amiconn_ [0] (i=quassel@rockbox/developer/amiconn)
01:07:09 Nick amiconn_ is now known as amiconn (i=quassel@rockbox/developer/amiconn)
01:07:18 Nick pixelma_ is now known as pixelma (i=quassel@rockbox/staff/pixelma)
01:08:21 Join FlynDice [0] (
01:10:48 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
01:11:02***Saving seen data "./dancer.seen"
01:11:34 Join AndyIL [0] (n=pasha_in@
01:12:48 Join akur [0] (
01:13:10 Quit freddyb (Remote closed the connection)
01:17:54 Join grant__ [0] (i=439e43c9@gateway/web/freenode/x-amcfisvyzybzldph)
01:18:17grant__ive got a bug with the mpeg player
01:18:38grant__it plays really fast video and really loud static noise
01:19:00 Join AEnima1577 [0] (
01:19:34Unhelpfulwhat type of file? and what device?
01:19:54grant__its an mpeg file on an ipod nano 2nd gen
01:20:07grant__it used to play on version r23313 but now it doesent
01:20:20TheSevenwhich revision are you trying it on currently?
01:20:41grant__im tring on the latest, r23545
01:20:54grant__but everything since the older one hasnt been able to play it
01:21:18TheSevenwhat's the oldest known-broken revision then?
01:21:30grant__no idea
01:21:56grant__i havent been able to select the file until recently because of the file selection bug i was getting
01:22:19TheSeventhe one fixed in 23489?
01:22:47 Quit AndyI (Read error: 110 (Connection timed out))
01:22:51TheSevencan you try 23323 and 23324?
01:23:02grant__i got sent a file by liar that he asked me to try and it didnt have the bug in it
01:23:02TheSeveni guess it broke in between
01:23:20grant__since then, the file selection has worked but the movie hasnt
01:23:28TheSeven(and the "file selection bug" was probably introduced in 23325)
01:23:39grant__i guess
01:24:06TheSevenso 23323 should be running fine, while 23324 should be broken. can you check that?
01:24:32TheSevenif this is true, we're having a big problem.
01:24:37grant__what do you mean by broken? also, where do i get these versions
01:25:01TheSevenhm, either compile them yourself or ask somebody to do it for you
01:25:12TheSevenand with broken i mean the weird behavior you're currently observing
01:25:40grant__well i dont have a working linux so someone will need to compile them for me
01:25:45*TheSeven blames that on the r23324 pcm fix, which was needed to make normal pcmbuf audio work properly
01:26:14 Quit MaadMan ("Leaving")
01:26:36TheSevenbasically this sounds like that fix is screwing mpegplayer's timing, which is probably syncing on the audio
01:26:52TheSevenI don't really see why it should play static though
01:27:06grant__its not like white static
01:27:16grant__its loud noisy tones
01:27:36TheSevenit's probably some garbage left over in memory that's being passed to the audio driver
01:28:03*TheSeven would do the builds if he hadn't already shut his build machine down in an attempt to finally go to sleep :-)
01:28:24grant__you can email the files to me whenever you want
01:28:45TheSevenok, just PM me your email address then
01:28:46 Quit gevaerts (Nick collision from services.)
01:28:55 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
01:29:12grant__whats PM, i kinda new to irc
01:29:14saratogawhy don't you just setup the build tools and get back to him with the revision number? probably greatly improves the odds of you getting a quick fix
01:29:32grant__i dont have a build environment'
01:29:57saratogayes hence my suggestion that you install it ;)
01:30:19*TheSeven thinks he already knows at least roughly what's going on...
01:30:29grant__my linux is being really stupid right now and i dont feel like reinstalling it all over again, or installing cygwin
01:31:25saratogathen just say you don't want to do it
01:31:57grant__i already said i dont have a working linux but it doesnt matter
01:32:03TheSevengrant__: then just use a linux live cd until you get annoyed (by the constant re-setting-up the toolchain) enough to finally fix your linux :-)
01:33:13TheSevenor just tell me your email address via private message (so that it doesn't float around publicly on the web in the channel logs) and i'll mail you links to the relevant builds tomorrow
01:34:09 Part toffe82
01:34:23grant__hope that this bug gets fixes
01:35:02 Quit kadoban (Read error: 104 (Connection reset by peer))
01:36:03saratogaFWIW you can also compile in Windows if you really don't like linux
01:36:47grant__well im already getting the compiled versions
01:36:53 Quit grant__ ("Page closed")
01:37:19 Quit ender` (" The first 90% of the code accounts for the first 90% of the development time. The remaining 10% of the code accounts for th")
01:37:52 Quit bertrik ("De groeten")
01:43:00 Quit robin0800 (Connection timed out)
01:48:00 Quit flydutch ("/* empty */")
01:58:28 Quit BHSPitLappy (Remote closed the connection)
01:59:29 Quit Lynx_ (" HydraIRC -> <- It'll be on slashdot one day...")
02:02:08 Quit MethoS- (Remote closed the connection)
02:09:36 Join kadoban [0] (
02:19:55 Quit kadoban (Read error: 104 (Connection reset by peer))
02:26:24CIA-6New commit by unhelpful (r23546): Replace enum statusbar_values in global_settings with int, fixing crash with -fshort-enums.
02:28:49 Quit akur ("Leaving.")
02:34:32 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
02:36:57 Quit dmb (Read error: 104 (Connection reset by peer))
02:41:16 Quit adiroiban (Read error: 60 (Operation timed out))
02:41:56 Join AaronM [0] (
02:44:34Unhelpfulamiconn: your suggestion didn't really work out, but setting iramcopy's address *inside* .ncdata does...
02:45:08 Join adiroiban [0] (
02:46:47 Quit LambdaCalculus37 ("Leaving")
02:47:53 Quit killan ("( :: NoNameScript 4.22 :: )")
02:55:48 Join killan [0] (
02:58:42 Quit GeekShadow (Read error: 104 (Connection reset by peer))
03:11:05***Saving seen data "./dancer.seen"
03:15:53 Quit shai (Read error: 104 (Connection reset by peer))
03:16:10 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
03:16:14 Join shai [0] (
03:18:20 Quit Lss (Read error: 104 (Connection reset by peer))
03:25:19 Join TopyMobile [0] (
03:28:21 Quit alexbobp (Read error: 60 (Operation timed out))
03:28:59 Join Xerion_ [0] (
03:30:26 Join alexbobp [0] (n=alex@
03:37:52 Join saratoga_ [0] (i=474d0741@gateway/web/freenode/x-unkagtuinbhdidrq)
03:38:11saratoga_how was the original decision to use libmad made?
03:38:25saratoga_i wonder about ffmpeg's fixed point decoder for instance
03:38:31saratoga_did anyone ever compare them?
03:41:36Unhelpfuli'm also seeing a number of warnings with gcc-4.4.1 arm-elf-eabi, most of them related to how enum types are represented. simply put, it appears that the compiler no longer considers it valid to do signed comparisons on enums with only positive values, so things that use an all-positive enum type but then assign or test for negative values throw warnings.
03:42:05Unhelpfulshould fixing this sort of thing be something i push back to svn, or should it stay in my eabi branch until a decision is reached about moving arm targets to eabi?
03:43:27saratoga_the fixes would just be a lot of explicit casts?
03:43:31Unhelpfulit seems to me that valid fixes would be to 1) add the negative values used to the enum type, 2) use an int, with the enum only existing to define possible values.
03:44:00Unhelpfulsaratoga: um, no, there's one in mpegplayer that explicitly casts *both* sides of the compare to int and still throws a "comparison always false" warning.
03:45:12 Join FOAD_ [0] (
03:46:11 Quit Xerion (Read error: 110 (Connection timed out))
03:46:11 Nick Xerion_ is now known as Xerion (
03:46:18rashersaratoga_: did it even exist then? Also, you have the fun of extracting stuff from ffmpeg vs using a simple library.
03:46:43Unhelpfulthe issue is that STATE_INTERNAL_NORETURN is *not* a member of the enum, and has a value not representable in the range of the enum. easiest fix, remove the definition from mpeg2_internal.h and add STATE_INTERNAL_NORETURN=-1 to the enum. kills the warning. :)
03:47:13saratoga_i don't know when it was added, perhaps not 2004
03:48:55Unhelpfuldoes gcc kill the false branch when it issues one of these warnings, provided a sufficient level of optimization?
03:56:08 Quit FOAD (Read error: 110 (Connection timed out))
03:56:09 Nick FOAD_ is now known as FOAD (
03:59:52Unhelpfulguh! another set of warnings in mpegplayer is due to passing a function parameter of type mpeg2_alloc_t values defined in mpeg_malloc_reason_t, some of which are not among the values defined for mpeg2_alloc_t
04:02:17Unhelpfulthe solution for this appears to be "get rid of mpeg2_alloc_t". i think.
04:05:12 Quit TopyMobile (Read error: 113 (No route to host))
04:06:50Unhelpfulor change the *argument* to int, so that the range complaint goes away
04:08:16 Quit Rondom (Nick collision from services.)
04:08:27 Join Rondom [0] (
04:08:33 Join shai_ [0] (
04:08:56 Join TopyMobile [0] (
04:11:34Unhelpfulthe argument is only ever used to fetch a string describing the reason for a particular malloc, and then only for debugging. changing argument to int, and also making sure the two different enums that define values passed for it don't overlap, which it appears they do at present :/
04:12:42 Join KBH [0] (
04:13:55Unhelpfulit appears to me there's an *actual* bug here, since the enums as currently defined overlap, even though values from both are used in a switch to fetch the string describing the malloc reason. also a non-debug build could avoid using the malloc reason arguments entirely, since they don't actually do anything.
04:16:49Unhelpfulthis fixes the warnings, and i believe corrects an actual bug as well (see mpeg_get_reason_str in alloc.c). is it reasonable to commit to mainline, or should it stay in my branch for now?
04:17:29 Quit HBK (Read error: 60 (Operation timed out))
04:17:30 Nick KBH is now known as HBK (
04:18:08 Quit TheSeven (Nick collision from services.)
04:18:27 Join The_Seven [0] (n=theseven@rockbox/developer/TheSeven)
04:18:39 Nick The_Seven is now known as TheSeven (n=theseven@rockbox/developer/TheSeven)
04:18:48 Join Xerion_ [0] (
04:19:39Unhelpfulalso i have numerous preprocessor "FOO is undefined" warnings... probably fixable by changing to #if defined(FOO) && (other test on FOO)
04:20:46 Quit saratoga_ ("Page closed")
04:21:06 Quit shai (Read error: 110 (Connection timed out))
04:31:39 Quit BHSPitMonkey (Read error: 104 (Connection reset by peer))
04:32:49 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
04:35:19 Quit Xerion (Connection timed out)
04:35:19 Nick Xerion_ is now known as Xerion (
04:39:02 Quit phanboy4 (Read error: 54 (Connection reset by peer))
04:59:16 Join kadoban [0] (
05:11:05 Join phanboy4 [0] (
05:11:08***Saving seen data "./dancer.seen"
05:22:51 Quit AaronM (Connection reset by peer)
05:23:15 Join AaronM [0] (
05:26:03 Quit TopyMobile (Read error: 60 (Operation timed out))
05:39:19 Quit adiroiban (Read error: 110 (Connection timed out))
05:46:26 Join dmb [0] (n=Dmb@unaffiliated/dmb)
05:55:04Unhelpfulhrm, actually, for mpeg_state_t, just changing the comparison to != STATE_INTERNAL_NORETURN suffices to clear the warning. the state should never be any *other* negative value, anyway...
06:10:21 Quit AaronM ("g'night")
06:33:47 Quit shai_ (Client Quit)
06:34:55 Join shai [0] (
06:36:36 Quit shai (Client Quit)
06:36:54 Join shai [0] (
07:09:06 Join stoffel [0] (
07:10:09 Quit pushnell ()
07:11:10***Saving seen data "./dancer.seen"
07:26:42Unhelpfulsaratoga: i think i have a slightly better understanding of the problem. the reason that a cast to signed int doesn't work is that the original type is an unsigned type smaller than int, and therefore the cast will be done by zero extension, so gcc knows the result can only ever be positive... i think. :)
07:32:04 Quit stoffel (Read error: 131 (Connection reset by peer))
07:32:18 Quit bluebrother (Nick collision from services.)
07:32:20 Join bluebroth3r [0] (n=dom@rockbox/developer/bluebrother)
07:46:02 Join Horschti [0] (n=Horscht2@xbmc/user/horscht)
07:47:06 Quit antil33t ()
07:57:06 Join FOAD_ [0] (
08:04:30 Quit Horscht (Read error: 110 (Connection timed out))
08:06:05 Quit FOAD (Read error: 145 (Connection timed out))
08:06:05 Nick FOAD_ is now known as FOAD (
08:23:44 Quit phanboy4 (Read error: 104 (Connection reset by peer))
08:37:00 Join antil33t [0] (
08:45:00 Quit BHSPitMonkey (Remote closed the connection)
09:03:10 Join n1s [0] (n=n1s@rockbox/developer/n1s)
09:09:02 Join drf|laptop [0] (
09:09:47drf|laptopSo hey, guys. I just bought a Toshiba Gigabeat S on eBay... since I'm a Zune fan and heard those were the predecessors. I was looking at the Rockbox page for it and it mentions needing the original nk.bin file so people can uninstall Rockbox... perhaps I could be of help?
09:11:13***Saving seen data "./dancer.seen"
09:12:23n1sdrf|laptop: what are you asking?
09:14:12drf|laptopIf you want the file? However, I have no idea how to get the mp3 player to show up as a drive...
09:14:38drf|laptopI dual-boot Windows and Linux, in Linux it simply doesn't recognize it at all and on windows it shows up as just the MTP drive, no letter or anything.
09:16:14n1sthe original nk.bin can be downloaded from toshiba
09:16:49n1sdrf|laptop: the original firmware is mtp only as far as we know
09:17:44drf|laptopSo how does someone go about installing Rockbox? That's the part I don't get.
09:18:10drf|laptopAnd I was just seeing the part where it says "TODO: How can a user uninstall Rockbox? We need a source for the original nk.bin...", I thought you guys needed the original
09:18:19drf|laptopand mine is stock firmware so it would surely be on there
09:19:55n1slots of the info in the wiki is outdated
09:20:19drf|laptopBut how do I get access to that "hidden" partition on the device?
09:20:56n1sdrf|laptop: you need to install rockbox or at least our bootloader to get access to the drive
09:21:29drf|laptopYeah but isn't it a catch 22? I need the drive to install Rockbox...
09:22:50n1sdrf|laptop: no you don't, the manual tells you how to do it, do the "manual installation" method as i'm not sure automatic is working in current rbutil
09:23:17drf|laptopI know automatic isn't working. I'm fairly experienced with computers, but I just didn't get how to install it on a MTP device...
09:23:29n1sour utility sends the new firmware over mtp, which the gigabeat firmware then installs
09:23:55n1sthe same way an update of the original firmware is done
09:24:00drf|laptopand ... is there a Toshiba site with the firmware and/or a flasher? This one I got from eBay with tons of music on it and such, but it's acting up - when I try to copy the music off it to my computer it freezes up, so I think something's wrong with it
09:24:07drf|laptopSo if I just wipe it that might help...
09:25:09n1sthe installation chapter in the manual has a link to download the original firmware, no "flasher" is needed, just our utility beastpatcher
09:25:25drf|laptopBut is there an official flasher, or not?
09:25:47drf|laptopI know for my Sansa, it was running a newer firmware that would auto-detect MTP and force it so I couldn't install Rockbox, so I got Sandisk's utility and down-graded it
09:26:20n1syes, i think there's an update utility provided with the firmware from toshiba
09:26:24Unhelpfulwe use MTP to install on S, as there is no known MSC firmware for it.
09:27:28drf|laptopbut Rockbox then turns it into MSC?
09:27:36 Nick fxb__ is now known as fxb (
09:27:54drf|laptopand wow. This thing is having similar problems to my Zune - some songs are showing up twice, etc. and others just freeze the player
09:28:01n1srockbox uses msc only
09:28:06drf|laptopNot sure what causes that in MTP players but it's ewird
09:28:22drf|laptopwell, I know with my Sansa, if you connect it with USB it always uses the OFW to charge and put data on... Rockbox can'tdo it
09:28:25drf|laptopor it couldn't back when I installed it
09:28:54n1srockbox has its own usb code now, and it is used on the gigabeat s
09:29:20pixelmaon the c200 too now
09:29:37n1sso we have our own usb mode, you don't need to keep the original firmware at all actually but you can if you want to dualboot
09:29:51 Join flydutch [0] (
09:30:06drf|laptopdualboot works?
09:30:17drf|laptopand heh, I don't really care, I updated Rockbox on my e200 and it completely broke all my themes
09:30:22drf|laptopso it plays music fine, I'm leaving it as-is :P
09:30:31drf|laptopWell maybe I'll try that later.
09:30:40drf|laptopWill it wipe the music when putting Rockbox on?
09:30:50 Nick fxb is now known as fxb__ (
09:30:56n1sif you boot the OF you need to reset with the battery switch to get it to actually reboot though
09:31:24drf|laptopThat battery switch is so tiny, I'd need a small screwdriver or somethign to get to it
09:31:34drf|laptopSo unlike the Zune you can't just hold power for a few seconds to force it to shut off?
09:31:43n1sdrf|laptop: no, but the dir structure mto creates is not very human readable so if you want to keep it as is i recomend using the database
09:31:50pixelmaRockbox won't play music that was put on in MTP mode
09:32:02drf|laptopAh, I see what you're saying
09:32:04drf|laptopsame thing happened with my Sansa
09:32:09n1spixelma: it won't
09:32:16drf|laptopDarn it, I can't even get the stupid music off right now
09:32:18drf|laptopWindows just freezes
09:32:28drf|laptopthat's probably why the guy was selling it, he musta infected it somehow :P
09:32:52pixelmaat least it doesn't on the Sansas and on the Gigabeat F, I doubt it does for the Gigabeat S
09:33:16pixelman1s: technically you can get them to play but the database won't find them
09:33:53pixelmaon the Sansas at least that is - they get a .dat extension which the Database doesn't care about
09:33:58drf|laptopI never use database
09:34:00n1spixelma: i seem to remember i could play files on my beast that were put there via mtp, never tried thed the database though
09:34:03drf|laptopit causes more problems than it's worth IMO
09:34:12drf|laptopBut the firmware on my e200 is really old, that could be why LOL
09:34:28Unhelpfulcauses problems? :P
09:34:48drf|laptopLike randomly crashes, takes 20 minutes to update if I put a single song on, etc etc
09:34:54pixelmaI seem to remember some .sat file "encryption" mentioned for the Gigabeat F
09:35:00drf|laptopso I just name all the files by the song title, it does the same thing relaly
09:35:05Unhelpfulnone of that should be happeninng...
09:35:50drf|laptopWell I dunno. Rockbox crashes quite a bit on my Sansa, to be honest
09:36:09drf|laptoplike it lags horribly when seeking, and since I'm used to my Zune I fly through seeking and end up freezing it, needing to hard-reset the device...
09:36:21drf|laptopthe OFW of the Sansa is slow and laggy too though, so it may just be the hardware
09:36:24n1spixelma: ok, tested and they do play on the beast but the tags seem to have been stripped :)
09:36:48 Join bertrik [0] (
09:36:59pixelman1s: do the files still have the correct extension?
09:37:06n1spixelma: yes
09:37:12drf|laptopAnyway. As I said before, I'm a Zune fan... and I was told before the reason nobody could put Rockbox on the Zune was that the firmware is encrypted or something?
09:37:36drf|laptopReally I'm just trying to get my Zune to show up in MSC mode so I can use it in Linux and such, the hardware is awesome as-is, that's the one drawback (stupid Microsoft)
09:37:40n1spixelma: but mangled names and stripped tags so not very usefull
09:37:43pixelmainteresting how different MTP mode seem to be
09:37:56drf|laptopMTP mode sucks
09:38:19n1spixelma: that's one of the good things about mtp actually, you can store things however you want
09:40:07n1sIIUC rockbox could have an MTP mode that had the same dir structure as the disk and copied files just like msc but could also add the new files to the database for example
09:40:42Unhelpfulthe MTP interface is essentially a serialized protocol for accessing an object database. what the implementing device does to *store* the information is up to implementors. :/
09:41:49Unhelpfulobjects are retrieved by an ID, "directories" and "albums" store lists of file/track IDs, etc.
09:42:14 Quit Horschti (Read error: 104 (Connection reset by peer))
09:42:32 Join Horschti [0] (n=Horscht2@xbmc/user/horscht)
09:43:29n1sUnhelpful: can you do an arbitrary dir structure?
09:44:00Unhelpfuli guess if we want to map it to a FAT filesystem without having to store the "MTP database" in a file, we'd use the address on disk of a file's first block, or perhaps of its dir entry, as the ID? i believe FAT32 lacks anything resembling an inode...
09:44:16Unhelpfuln1s: well, look at what beast does for its data store :)
09:45:43n1sUnhelpful: i mean expose an arbitrary dir structure through MTP and let the user create/move etc dirs?
09:46:54Unhelpfuln1s: yes, that should be possible. the user-visible hierarchy for MTP could be entirely virtual, or it could be mapped onto the real filesystem somehow, provided we have a way to keep track of object IDs.
09:47:31n1sthat would be pretty cool
09:48:48Unhelpfulthe ideas i like are 1) playback and transfer not being mutually exclusive 2) DB update during transfer
09:49:23 Join robin0800 [0] (
09:49:57n1syes, those are very nice features and if they can be done without all the things that makes us dislike MTP usually that would be very nice
09:49:59LloreanThose are the ones I like too, though 2 moreso than 1.
09:50:09 Join alexbobp_ [0] (n=alex@
09:50:42amiconnUnhelpful: I'd rather not change enums to ints, but fix the enum usage
09:51:49Unhelpfulamiconn: most of the enums needn't be changed to ints, that particular case *could* also be handled by merging the two enums, which would also conveniently make sure that their values are always distinct.
09:53:56amiconnNo enum should need changing to int
09:55:38 Quit alexbobp_ (Remote closed the connection)
09:56:03drf|laptopwell guys, I'm going to sleep. I'll be back to inquire some more tomorrow :)
09:56:06 Part drf|laptop
09:58:59 Quit alexbobp (Read error: 104 (Connection reset by peer))
10:01:08 Quit robin0800 (Read error: 104 (Connection reset by peer))
10:03:06Unhelpfulamiconn: i don't see any other way for the one in global_settings, since it's accessed via a pointer-to-int... but yes, i think that in pretty much any other reasonable case it should be avoidable.
10:03:39amiconnThen the access need to be changed to use a pointer to the correct enum type
10:04:00 Join alexbobp [0] (n=alex@
10:05:47Unhelpfulhow should we handle that? reset_setting can't reasonably have a case for every enum type, and if we just added short and byte sizes to it we'd have to guess at correct enum-type sizes.
10:07:41TheSevenUnhelpful: hm, can't we let some sizeof magic do the guessing?
10:09:24UnhelpfulTheSeven: i'm not sure, can we use sizeof in an expression to initialize runtime data? enums might also be represented as signed or unsigned based on whether any enumerated values are negative. :/
10:10:03TheSevenUnhelpful: yes, sizeof is evaluated at compilation time, so we can use it in e.g. initializers
10:10:18TheSevenand signed vs. unsigned shouldn't make a difference if you're just copying things around
10:10:58Unhelpfulright, should only matter for comparisons... or i suppose if we're converting any of these integer values to text?
10:11:16TheSevenyes, but is this an issue for reset_setting?
10:11:44 Quit reid05 (Read error: 110 (Connection timed out))
10:12:10 Join reid05 [0] (
10:12:32Unhelpfuli don't *think* so? where else do we use struct settings_list to extract settings data? would it matter in any of those places?
10:13:15 Quit amiconn (Read error: 54 (Connection reset by peer))
10:13:15 Quit pixelma (Read error: 131 (Connection reset by peer))
10:13:28 Join pixelma [0] (i=quassel@rockbox/staff/pixelma)
10:13:30 Join amiconn [0] (i=quassel@rockbox/developer/amiconn)
10:13:45TheSevenas long as stuff is generic (and we only have that problem for generic stuff), I can't see any sane reason to any more involved comparison than equals/doesn't equal, or to convert the raw enum value to a string
10:14:13TheSevento do any*
10:14:22Unhelpfuland each of these cases would also be gaining new branches in their switches for handing uint8, int8, uint16, and int16 values. it seems to me that using an int for things that will be in settings is more sane...
10:14:58TheSevenhm, it would only be uint8/16/32
10:15:16TheSevenbut yes, this *may* have a performance impact
10:15:38TheSevenis there any penalty on arm for non-word-aligned byte/halfword reads?
10:17:12Unhelpfulthere'd be a code-size hit. and cfg_int_to_string cares about signedness.
10:18:42Unhelpfulis there a way to tell the compiler that a particular enum type, or a particular instance of it, should be int-sized, even though it *could* be smaller? i couldn't find any, you can make an enum small without enabling short enums by giving it a packed attribute, but there doesn't seem to be an "unpacked" :/
10:20:07amiconnEach enum is a specific type, so there are no differently sized 'instancec
10:21:44 Quit AEnima1577 ("Leaving.")
10:22:34Unhelpfuland even if we could pad the struct field, things that set it directly instead of via-int-pointer wouldn't set the extra bytes (in the case of a signed enum) correctly. :/
10:22:55n1si don't see the problem with using an int, it should always be safe and is simple
10:24:29Unhelpfuln1s: you lose certain types of compile-time error-checking if the compiler doesn't know the variable is enum... really only helps if it's going to be in a switch statement, i beileve.
10:48:14 Join robin0800 [0] (
10:48:30 Join pamaury [0] (
10:58:51 Join TopyMobile [0] (
11:05:27pixelmalooks like someone broke early USB (at least on Archos) *again*... just when I wanted to recommend a current build to the one reporting problems with that in r3.4 (which was fixed later)
11:09:59pixelmaon my Ondio the USB screen is also slightly broken when something in the screen you are in before entering the USB screen is scrolling, it stays on the screen
11:11:17***Saving seen data "./dancer.seen"
11:13:22 Join ender` [0] (
11:13:42pixelmathe same happens on my c200
11:20:04*pixelma thinks that there either should be a Rockbox 3.4.1 for early USB targets or some more daily builds available again (maybe some marked as "good ones)
11:20:59pixelmaat least if it fixes something that is broken in the most recent release
11:21:17pixelmaa known problem I mean
11:23:41n1swhen did we switch to only archiving 3 daily buids, and why?
11:25:37n1sb ut yeah, a 3.4.1 might be a good idea
11:28:19Unhelpfulif a feature was broken at release, it seems to me the right thing to do...
11:31:52 Join gb_master [0] (i=82c03212@gateway/web/freenode/x-mhqiogppodujdjgc)
11:47:22TheSeven[10:18]<Unhelpful>is there a way to tell the compiler that a particular enum type, or a particular instance of it, should be int-sized, even though it *could* be smaller? i couldn't find any, you can make an enum small without enabling short enums by giving it a packed attribute, but there doesn't seem to be an "unpacked" :/
11:47:41TheSevenwhat about adding some dummy entry with a specified value that doesn't fit into 16 bits?
11:48:08CIA-6New commit by nls (r23547): Introduce new config header 'config-sim.h' that takes care of undefining things that shouldn't be defined for sims so that the target configs no ...
11:48:19UnhelpfulTheSeven: that may cause warnings if the enum type is used in a case statement :/
11:49:06TheSevenjust use big-enough values for actual entries then? :-P
11:50:17 Join stripwax [0] (
11:50:26Unhelpfulwhat, just always set the last value to 65537 or something? i still think i like making the field in global_settings an int better than that. :P
11:51:48*Bob_C_ thinks there is no sneaky way around enum size issues
11:52:36UnhelpfulBob_C_: -fno-short-enums would make all enums everywhere int-sized :)
11:53:01n1sUnhelpful: is the gain from short enums large?
11:53:54Bob_C_I think in most projects I have worked on, enums are always ints. there is a size penalty, but it is safer
11:54:06Unhelpfuln1s: i've not checked. they're specified as the default for eabi, though...
11:55:03TheSevenUnhelpful: I didn't say that I'd like it, I'm just digging around to see if there is any sane solution
11:55:38TheSevenbut you're probably right, int enums will probably be less size penalty than a switch statement + additional argument to some functions
11:56:54Unhelpfuli personally think the sane solutions are either fno-short-enums (or making it the default for our toolchain, an easy patch) or making variables used in some generic context via a pointer, as with settings, ints.
11:58:32TheSeveni would think the best solution is only making the enums that need it an int, but if that isn't possible, I'd probably just go for no-short-enums
11:58:59TheSevenis there maybe some __attribute__((...)) to get rid of them for a specific enum?
12:05:30 Quit stripwax ("")
12:06:50 Join stripwax [0] (
12:11:32Unhelpfulyou can make an enum short with a packed attribute, i believe
12:12:07 Quit stripwax ("")
12:14:33Unhelpfulsomehow rockbox-info.txt is telling me that -fno-short-enums makes ramsize for an eabi build on e200 144B *smaller*. with no difference in binsize, which is where i'd expect any savings for that case (long enums can be transferred to/from memory using store/load-multiple instructions, short enums can't)
12:16:27seaniMorning, can I ask if I understand something about powermgmt.c correctly? It looks like the main thread functions are invoked every HZ/2 ticks - half a second.
12:16:51seaniThis means if you want to test if a timeout matures, you could be as much as HZ/2 ticks out from the target time? So in general, "high level" timeout features have a resolution of around HZ/2 ?
12:17:57seaniAnd when I say "morning", I mean it's morning in those parts of the world that *really* matter. God Save the Queen etc.
12:20:09 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
12:26:16Bob_C_seani: yes, anything in power thread would have that resolution
12:30:52 Quit killan (Read error: 131 (Connection reset by peer))
12:31:24 Quit TheSeven ("ChatZilla 0.9.85 [Firefox 3.5.4/20091016092926]")
12:34:09seaniBob_C: Thanks. Are there other places where timeout behaviour is typically added? This is timeout behaviour apropos of nothing, i.e. *Just* as a result of the passage of time, not in relation to keypresses etc.
12:35:37Bob_C_I'm not sure what you mean by added?
12:37:41seaniWell if I want to check if another timeout - like the sleep timeout for instance that just gradually counts down irrespective of other activity - has matured.
12:38:20seaniIIRC some timeouts are checked to see of they've matured in response to UI events like play / pause etc.
12:38:57 Join teru [0] (
12:39:57 Quit Horschti ("Verlassend")
12:39:59 Quit yosafbridge ("Coyote finally caught me")
12:40:10 Join yosafbridge [0] (
12:42:43Bob_C_I think I see what you mean, but I don't know
12:47:21seaniThanks, the powermgmt stuff would give me what I want, just a check I wasn't missing something obvious.
12:49:39 Quit Xerion (Read error: 104 (Connection reset by peer))
13:05:01 Quit gb_master ("Page closed")
13:07:13 Join Paleskin [0] (
13:07:49Paleskinis there a way to flash rockbox into fuze's newest firmware ?
13:10:20 Join Xerion [0] (
13:11:19***Saving seen data "./dancer.seen"
13:13:25 Join ender [0] (
13:14:26 Join DerPapst [0] (
13:24:51 Join MethoS- [0] (n=clemens@
13:26:15 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
13:27:49 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
13:28:48 Quit ender` (Read error: 110 (Connection timed out))
13:29:29 Part Paleskin
13:29:54 Quit MethoS- (Remote closed the connection)
13:31:34CIA-6New commit by alex (r23548): Make it clear in the beast manual that you need to play Russian dolls with the OF you download from Toshiba - it is inside an exe file which is ...
13:32:35 Quit pamaury (Read error: 113 (No route to host))
13:38:43 Join kugel [0] (n=kugel@rockbox/developer/kugel)
13:39:05kugelsizeof() can also be evaluated at runtime AFAIK
13:39:07 Join dfkt_ [0] (n=dfkt@unaffiliated/dfkt)
13:43:49 Quit dfkt (Read error: 60 (Operation timed out))
13:44:52Bob_C_hey kugel, are your audio problems a hardware fault do you think?
13:45:20kugel"my audio problems"
13:45:34Bob_C_on the mini2440?
13:46:25 Nick dfkt_ is now known as dfkt (n=dfkt@unaffiliated/dfkt)
13:50:08kugelBob_C_: the only problem is the bad sound quality.
13:50:56Bob_C_right, so could be faulty hardware on your board? It seems that everyone else is ok
13:52:29kugelI don't really mind that, I have no plans to listen to audio from it regularly
13:54:33Bob_C_Mine sounds the same when the same file is played on my PC, but maybe you have better hearing than me
13:54:42kugelI doubt that
13:55:16kugelif yours sound fine, I probably just have bad luck and got a bad one
13:55:34Bob_C_Hearing does decline with age...
13:57:30kugelI'm no audiophile. when I say it sounds bad I don't mean supbar, I mean horrible. It lacks bass completely for example
14:00:55pixelmahmm... can't make much sense of the address the error message for the broken early USB gives me with a current build. Looking at the map file - functions in cuesheet.c and talk.c are closest (cuesheet was already turned off before and now I also turned off voice)
14:01:31pixelmaon my Ondio that is
14:04:06 Join pamaury [0] (
14:08:08Bob_C_kugel: that could happen if the decoupling capacitors are the wrong values I think
14:19:09 Quit robin0800 (Remote closed the connection)
14:21:52 Quit DerPapst ("Leaving.")
14:21:56 Quit pamaury (Read error: 148 (No route to host))
14:23:46 Join stripwax [0] (
14:27:35 Join killan [0] (
14:30:46n1spixelma: the addresses in the map are start addresses for the functions, an exception can and usually does happen inside the function 8well, depending on what exception it is of course) but the addresses in the map will not match the error exactly
14:31:11n1ss/will not/will usually not/
14:47:03bertrikthe meizu m3 *also* has two sdram init sequences, different from the ones in the m6sp :)
14:50:11markunbertrik: yay :)
14:50:24 Join freddyb [0] (
14:50:41 Quit stripwax (Read error: 104 (Connection reset by peer))
14:51:44bertrikthe actual init sequence is nearly identical to the one of the m6sp, mostly just different numbers written to the registers
14:52:52 Join stripwax [0] (
14:52:58bertrikso it should be easy now to get meizu m3 sdram working too
14:54:18 Nick fxb__ is now known as fxb (
15:02:04bertriknow we *just* need an USB bootloader for the meizus to load a main rockbox to sdram :P
15:03:28gevaertsbertrik: can't be done. You need a notes buffer overflow for that
15:04:28 Join chellomere [0] (n=alex@
15:07:01 Quit chellomere (Client Quit)
15:11:21***Saving seen data "./dancer.seen"
15:11:44 Quit stripwax (Read error: 104 (Connection reset by peer))
15:13:40CIA-6New commit by teru (r23549): theme remove: add sbs to remove list.
15:14:09 Join stripwax [0] (
15:21:01 Quit stripwax (Read error: 104 (Connection reset by peer))
15:22:34 Join einhirn [0] (
15:23:09 Join stripwax [0] (
15:23:37 Quit Xerion (
15:23:37 Quit lyngaas (
15:23:37 Quit Torne (
15:26:54 Nick fxb is now known as fxb__ (
15:27:34NJoinXerion [0] (
15:27:34NJoinlyngaas [0] (
15:27:34 Join Torne [0] (i=torne@rockbox/developer/Torne)
15:29:46seaniDoes the "Reported Version" in the add new patch screen have a purpose when adding proposed features, or is it just to record where a bug was first spotted?
15:30:04gevaertsit's irrelevant for patches
15:36:31 Quit Torne (
15:36:31 Quit Xerion (
15:36:31 Quit lyngaas (
15:36:54 Quit stripwax ("Miranda IM! Smaller, Faster, Easier.")
15:38:31 Quit Horscht ("Verlassend")
15:39:50NJoinXerion [0] (
15:41:01NJoinlyngaas [0] (
15:41:01NJoinTorne [0] (i=torne@rockbox/developer/Torne)
15:43:22 Join TorneWuff [0] (
15:44:09 Quit Torne (Remote closed the connection)
15:44:14 Quit seani ("Leaving")
15:46:51 Join funman [0] (n=fun@rockbox/developer/funman)
15:50:53domonokycan somebody tell me what should happen on usb after i get the first reset interrupt ? i would expect a ep0 interrupt for the first request from the host, but nothing happens.. does the device driver have todo something special after the reset request ? (at moment i reset all ep packet sizes to recommended sizes, set FUNC_ADDR_REG to 0, enable all ep interrupts, and call usb_core_bus_reset() )
15:53:30 Join TheSeven [0] (n=theseven@rockbox/developer/TheSeven)
15:54:03gevaertsmaybe there's some setup needed to get it to actually start looking for data?
15:55:27domonokyso its correct that the host should issue some command after reset ? if so,i am probably missing some init.
15:56:34gevaertsafter the bus reset, it should ask for the device descriptor
15:57:15domonokyoki, so i am back to hunting what i am missing in the init..:-)
15:58:43TheSevendomonoky: I would guess that you need to actively start receiving data on ep0
15:59:06TheSevenuntil then, everything arriving on the bus will probably just be NAKed
16:00:51domonokyah, i would have expected to get a interrupt when a packet is in the fifo... lets hunt the datasheet hints :-)
16:01:13TheSevendo you have a datasheet of whatever controller you are implementing right now?
16:02:05domonokyi only have the datasheet for the s3c2440 itself, it contains all registers for its usb controller, but unfortnunatly very little explanations
16:02:16TheSeventhat's fine
16:03:05TheSevenonly usb1? :-/
16:03:25domonokyyes, unfortunatly.. it will be slow..
16:05:41TheSeven...and a pretty lowlevel controller
16:06:23TheSeventhe synopsys otg really was fun compared to that :-P
16:07:00funmanTheSeven: which target has a synopsys otg ?
16:07:33freddybfunman: have you look at my keyboard patch for Fuze?
16:07:34TheSeventhe apple thingys
16:07:44funmanfreddyb: nope, where is it ?
16:07:51funmanTheSeven: afaiu the sansa ams as well
16:08:23freddybfunman: FS #10763
16:08:24TheSeveni called the driver usb-s3c6400x when i wrote it, because that datasheet was fitting but i hadn't found out who designed that one yet
16:10:23freddybfunman: it changes the way you enter text. you scroll fwd to get the right column and scroll back to move down to the right char. scrolling fwd again inserts the char and resets to the top corner. it's pretty quick. try it out and see what you think
16:11:06funmanTheSeven: interesting, i spot similarities with a quick look ..
16:11:10TheSevenfunman: do you know any details on the ams otg? any reg meanings etc? I guess you don't have a datasheet for that one...
16:11:19TheSevennice to hear
16:11:20 Quit flydutch (Read error: 104 (Connection reset by peer))
16:11:33funmanTheSeven: we have the list of registers from the AS3525 datasheet, all the registers are in as3525/usb-drv-as3525.c
16:11:50funmanDCFG = USB_DEV_CFG , the following regs correspond
16:12:05 Join flydutch [0] (
16:13:17funmanfreddyb: i don't know much keyboard, you should talk with FlynDice since he tested it and likes it ;)
16:13:33TheSevenhm, yes, there are similarities, but the overall layout is entirely different
16:14:27funmanit gives another source of info for the as3525 driver anyway, perhaps time will tell what can be merged
16:14:28TheSevenand if that file is meant to be complete, it's having far less regs than the one i implemented
16:14:47freddybfunman: FlynDice says he doesn't use it very much. Thanks anyway.
16:14:52funmanit's not working in any way
16:14:59funmanfreddyb: try asking kugel
16:16:32 Join AEnima1577 [0] (
16:17:21freddybfunman: ok, i'm out
16:17:28 Quit freddyb ("Konversation terminated!")
16:19:03 Join robin0800 [0] (
16:21:06 Join Strife89 [0] (
16:39:07TheSevendomonoky: this usb controller is weird
16:39:46domonokyTheSeven: how is it weird ?
16:40:13TheSevenlooks like it's pushing as much work as possible towards the driver
16:40:29TheSevenfrom what i can tell, it should be setting OUT_PKT_RDY when something has come in
16:41:08 Join adiroiban [0] (
16:41:09domonokyjup, thats what i also think, put nothing happens.. do i have to somehow react with the pullup to reset ?
16:41:58TheSevenas far as i can tell from the datasheet, you'll only need to reset the address to zero
16:42:21TheSevenare you polling the EP0 OUT_CSR1_REG or waiting for an int?
16:42:37TheSevendid you accidentally set OUT_DMA_INT_MASK?
16:42:49domonokyat the moment i just wait for ints, and the only one it get is reset.
16:43:14domonokynope, i played with that, but it didnt change anything.
16:44:11TheSevenand EP0_INT_EN is set? (did you try reading it back and checking?)
16:48:36 Join stripwax [0] (
16:49:15domonokyyes, it surely is enabled... (i just read it back to make sure)
16:51:13 Join adiroiban1 [0] (
16:51:37 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
16:51:51 Join freddyb [0] (
16:53:58Bob_C_domonoky: if it helps, there may be useful clues in the USB driver for mini2440-uboot
16:54:20domonokyah, that could be another sample to look at :-)
16:57:03 Join FOAD_ [0] (
16:58:43Bob_C_in that driver the driver sets EP0 out_pkt_rdy when a reset int is received, does that make sense?
16:59:35 Quit einhirn (Read error: 104 (Connection reset by peer))
16:59:57 Join kugel_ [0] (
17:00:11 Quit kugel (Nick collision from services.)
17:00:19 Nick kugel_ is now known as kugel (
17:00:28 Join mirak [0] (
17:00:34 Join pamaury [0] (
17:02:32domonokyBob_C_: it doesnt really make sense to me, but maybe thats what i might miss :-) can you point me to where it does that ?
17:03:13Bob_C_oh wait, the driver is clearing that bit
17:05:34Bob_C_ is the relevant bit
17:05:56domonokybut this driver calls PrepareEp1Fifo in the reset handling.. which seems to send something on ep1 ?
17:06:26 Quit Strife89 ("Switching machines,")
17:06:37domonokyuh, which driver is that ? the linux one ?
17:07:19Bob_C_that was from the uboot driver
17:10:22 Quit adiroiban (Read error: 110 (Connection timed out))
17:11:22***Saving seen data "./dancer.seen"
17:11:27*domonoky still strugles to find the file with this driver.
17:13:39Bob_C_I got my version here i think
17:13:43 Quit FOAD (Read error: 110 (Connection timed out))
17:13:43 Nick FOAD_ is now known as FOAD (
17:14:29 Quit funman ("free(random());")
17:15:40 Join dfkt_ [0] (i=dfkt@unaffiliated/dfkt)
17:16:16domonokyah, thats probably a newer uboot version, then i got... lets get it take a look.
17:17:29Bob_C_It appears to just implement DFU, so maybe it's fairly simple and self-contained
17:23:05domonokyand its coded by Harald Welte (, this has to be a good sample to get usb running :-)
17:23:54gevaertsdomonoky: make sure you check the license properly before touching it :)
17:25:10Bob_C_The openmoko project may have died, but it has left some useful bits
17:26:39freddybAnybody out there use the keyboard?
17:27:12 Quit pamaury ("exit(*(int *)0 / 0);")
17:27:25 Join pamaury [0] (
17:28:41teruI want opinions for FS #10338: Add "use morse input " entry to settings menu and commit it if possible.
17:31:54 Quit robin0800 (Remote closed the connection)
17:33:42mc2739teru: regarding r23549 manual update, should the first \fname{.rsbs} actually be \fname{.sbs} ?
17:34:04 Join darkham [0] (
17:34:11pixelmateru: maybe it would even male it possible to use morse input on targets that don't have a spare button combo for entering the morse input screen from the virtual keyboard
17:34:58 Quit stripwax (Read error: 131 (Connection reset by peer))
17:35:55 Part adiroiban1
17:36:05terumc2739: ah, yes. thanks.
17:37:39pixelmalast time I looked that button mappings of the morse input screen and the usual virtual keyboard screen were quite tied, so I guess there is more to making morse input work on those targets is not only done after making entering the morse input screen a setting
17:40:36CIA-6New commit by teru (r23550): fix typo in manual. thanks to mc2739 for noticing.
17:40:39 Quit dfkt_ (Read error: 110 (Connection timed out))
17:41:03 Join dfkt_ [0] (n=dfkt@unaffiliated/dfkt)
17:41:25pixelmamc2739: did you have a look at the pictureflow manual patch (posted a new version)?
17:42:05 Quit pamaury ("exit(*(int *)0 / 0);")
17:42:29 Quit dfkt (Read error: 110 (Connection timed out))
17:43:29mc2739pixelma: sorry, no, I have not had much time lately. I will try to look at it this weekend.
17:45:48pixelmahoped you could help with the "issues" outside the button table. That could be a seperate patch though
17:46:11 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
17:51:08 Join DerPapst [0] (
17:54:26 Join fyrestorm [0] (
17:59:37*domonoky gives up on usb for now..
18:12:39 Join Rob2222 [0] (
18:14:11 Nick fxb__ is now known as fxb (
18:15:10 Quit freddyb (Read error: 110 (Connection timed out))
18:18:32 Part Wafelijzer
18:31:24 Quit teru ("Quit")
18:32:58 Quit AEnima1577 ("Leaving.")
18:39:01 Join Grahack [0] (
18:39:40 Join crazyrabbit [0] (
18:42:49 Nick YPSY is now known as Ypsy (
18:44:10 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
18:48:09 Quit darkham ("Sto andando via")
18:49:00 Quit Grahack ("Leaving.")
18:52:24 Nick crazyrabbit is now known as nawks (
18:52:32nawkshai Dhraakellian
18:52:42nawksand everyone else :)
18:53:43nawksi can't see my sd card on usb bus in my e200
18:54:03nawkse200 shows it in rockbox and sansa firmware
18:54:26nawksmaybe something is missing in my kernel usb?
18:54:54nawksis sd card supposed to show as another usb mass storage dev?
18:55:00nawksor an mmc reader?
18:55:08*nawks checks that
18:55:49nawks# CONFIG_SCSI_MULTI_LUN is not set
18:55:53nawksty gevaerts
18:58:03gevaertsout of curiosity, what distribution is this? CONFIG_SCSI_MULTI_LUN is needed by lots of USB things
19:00:55 Join phanboy4 [0] (
19:00:55nawksbut would probably happen to me on other distros too
19:01:01nawkswhen i roll my own kernel
19:01:24nawkssince kernel help says unhelpful things such as
19:01:26nawksThe vast majority of SCSI devices have only one LUN, and │
19:01:28nawks │ so most people can say N here. The max_luns boot/module parameter │
19:01:29nawks │ allows to override this setting.
19:01:39nawks"vast majority"
19:01:46nawksmeaningless term
19:01:58nawksshould never be used
19:01:59gevaertsthat's true for old-style SCSI devices, but nearly all USB card readers need this.
19:02:13nawksdocumentation is last thing to get fixed :)
19:03:06nawksi guess i am supposed to know that a DAP is a "cd jukebox"
19:04:10nawksthis is how i learn what parts of the kernel are needed for various functions
19:04:18 Join AEnima1577 [0] (
19:05:28gevaertsyou could use the max_luns parameter
19:07:05 Join freddyb [0] (
19:10:23 Quit freddyb (Client Quit)
19:10:33 Join freddyb [0] (
19:11:09nawksin "files" menu of rockbox, it shows PHOTO RECORD tmp VIDEO folders
19:11:14nawksbut does not show MUSIC folder
19:11:23nawksthis is by design?
19:11:27***Saving seen data "./dancer.seen"
19:12:16mc2739it is because the Sansa firmware makes the MUSIC folder hidden
19:13:20mc2739you can remove the hidden attribute, but the next time you boot the OF it will hide it again
19:14:42nawksi can only see this attribute from windows OS?
19:14:56nawksit is vfat filesystem attribute?
19:19:05bertrikTheSeven, did you work on the USB controller as described in the s5l8700 data sheet?
19:19:46TheSevenbertrik: no
19:19:55TheSevenI think you said that one would be similar to tcc?
19:29:45 Quit AEnima1577 (Read error: 104 (Connection reset by peer))
19:30:25 Join AEnima1577 [0] (
19:36:10 Join AEnima15771 [0] (
19:38:48CIA-6New commit by kkurbjun (r23551): Add optional viewport clipping, can be enabled with HAVE_VIEWPORT_CLIP. A simulator check is also added to set_viewport that will show an error/note ...
19:41:11 Quit AEnima1577 (Read error: 60 (Operation timed out))
19:42:41nawks# mattrib c:/MUSIC
19:42:43nawks H C:/MUSIC
19:42:53nawks# mattrib -h c:/MUSIC
19:43:12nawksmtools pkg
19:45:48 Quit AEnima15771 ("Leaving.")
19:46:37AlexPnawks: a) What was that in aid of? and b)
19:48:37CIA-6New commit by kkurbjun (r23552): Fix two width calculations
19:49:18nawksksry alexp. it was solution to removing hidden attribute from sansa MUSIC directory using Linux OS
19:49:44AlexPnawks: I suspect people here can already handle that
19:49:54AlexPAlso, the OF will just put in back whenever it boots
19:50:28nawksmc2739 mentioned that
19:50:37nawksbut no-one replied to:
19:50:39nawks[13:14] <nawks> i can only see this attribute from windows OS?
19:50:47AlexPNo, it is a standard FAT flag
19:51:32nawkssry for excess noise. this is mainly dev channel?
19:51:38AlexPdev and support
19:51:49AlexPbut Rockbox support :)
19:52:57nawkswell thanks for the support :) and for rockbox :)
19:53:01nawksit is very cool
19:53:53nawksrbutil is excellent
19:54:40AlexPglad to hear it :)
19:55:52CIA-6New commit by kkurbjun (r23553): Fix Red
19:59:27 Join MarcGuay [0] (
20:00:33MarcGuayHi all. Taking another shot at installing the single-bootloader on the gigabeat S. Should beastpatcher work while the unit is attached in one of the rockbox USB modes?
20:00:50AlexPI don't think so
20:00:57AlexPIt looks for MTP devices
20:01:18 Join TopyMobile_ [0] (
20:01:19MarcGuayIt's reporting "Found device: Ahead - Nero Fast CD-Burning Plug-in"
20:01:56MarcGuaySo replacing an existing RB bootloader involves overwriting the existing nk.bin?
20:01:58AlexPFor RB USB, you would normally just copy over the new bootloader to the firmware partition
20:02:05CIA-6New commit by kkurbjun (r23554): Just a test to make sure all targets build with viewport clipping enabled - will disable once all builds pass
20:02:27MarcGuayI'll give her another shot - here's to not formatting 60GB of music...
20:02:33AlexPfingers crossed
20:02:53 Quit TopyMobile (Read error: 113 (No route to host))
20:03:30 Quit phanboy4 (Read error: 54 (Connection reset by peer))
20:05:38MarcGuayAlexP: The nk.bin here: is the most up-to-date single bootloader?
20:06:07AlexPyes, should be the 1.0 release
20:07:34UtchybannTheSeven: placing a return at the beginning of nand_power_down() function fixes the usb write problem.
20:07:36AlexPMarcGuay: What OF do you have?
20:07:53AlexPMarcGuay: The theory is that anything after 1.1 won't accept the single boot IIRC
20:08:28MarcGuayAlexP: I think it was 1.2... But the testing I did before made it seem like that wasn't the problem... ROLO worked, here's to full reboot working...
20:08:50MarcGuayError 3. Farknuts.
20:08:51 Join AaronM [0] (
20:08:58AlexPI don't think ROLO goes via the bootloader
20:09:09TheSevenUtchybann: can you try only commenting the code that actually deals with the nand, but not the mutex lock/unlock and setting powered to 0?
20:09:15MarcGuayAlexP: Right.
20:09:56CIA-6New commit by kkurbjun (r23555): Everything passed - disable viewport clipping.
20:11:05 Join robin0800 [0] (
20:12:40 Quit domonoky (Read error: 104 (Connection reset by peer))
20:14:26 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
20:17:45MarcGuayAlexP: Installing the dual-bootloader with beastpatcher doesn't work for me either... Is there a copy of the 1.0 OF available online somewhere?
20:18:12AlexPno idea I'm afraid
20:18:22 Nick fxb is now known as fxb__ (
20:18:54MarcGuayAlexP: Thanks anyways.. Going back to my ancient dual-booter..
20:19:01 Quit kugel (Remote closed the connection)
20:19:12MarcGuayNo format, praise Jah.
20:19:14 Join kugel [0] (n=kugel@rockbox/developer/kugel)
20:20:19 Nick Ypsy is now known as YPSY (
20:20:45 Quit kugel (Client Quit)
20:21:38 Join kugel [0] (n=kugel@rockbox/developer/kugel)
20:22:33TorneWuffMarcGuay: Not that I can find anywhere
20:22:35 Nick TorneWuff is now known as Torne (
20:23:11AlexPMarcGuay: When using beastpatcher, you gave it an OF nk.bin right?
20:23:28MarcGuayAlexP: Yes, I passed it v1.2.
20:24:18AlexPI don't know if it was tested with 1.2, but it should work I would think
20:24:24MarcGuayAnd beastpatcher copies it successfully, but when the beast reboots I'm back to Error1,Error2..
20:25:15Torneyou're having the beast problem? ooh )
20:25:26Tornecan you send me a dump of your flash once you get into rockbox? :)
20:25:54 Quit kugel (Remote closed the connection)
20:26:03MarcGuayTorne: Sure, hold on.
20:26:08 Quit AaronM (Remote closed the connection)
20:26:11Torneno rush
20:26:13Tornejust as a data point
20:26:16AlexPTorne: So have you made big disks work yet?
20:26:25AlexPTorne: And how about solving world hunger :)
20:26:31TorneAlexP: I *think* i've established that the FAT code is standard in the bootloader
20:26:38Tornei.e. the whole TFAT thing is probably a red herring
20:26:48Torneand thus it's probably something more subtle and awkward causing the formats :)
20:27:05MarcGuayTorne: System->Debug->Dump ROM?
20:27:06Tornei found a bunch of button combiatnions that do crazy things, also
20:27:08TorneMarcGuay: yup
20:27:10AlexPSo that is bad then?
20:27:20TorneAlexP: well it means i need to read even more of the rom
20:27:54AlexPAh, so bad for you then :)
20:28:14Torneand no big disks won't ever work unless we do a rombox for the beast :)
20:28:33AlexPbugger :/
20:29:35Tornewell, big disks like the ipod one, i man
20:29:45Tornedisks which work with 512 byte accesses should be fine, actually
20:30:33 Join petur [50] (n=petur@rockbox/developer/petur)
20:30:42AlexPah right
20:30:58Torneit doesn't care what the size of hte second partition is, only that it's formatted with particular parameters. if it repartitions, it makes the first partition 150MB and the second partition "all the rest"
20:31:18Tornehm, it might not do LBA48 either
20:31:18AlexPTorne: So I guess the Toshiba 240GB one is bigger sectors
20:31:35TorneAlexP: i dunno. remember i have not *tried* this, only proven it to be true ;)
20:32:02AlexPTorne: If it didn't do LBA48, I guess that'd make it think the disk was buggered if it was a bigger partition?
20:32:41TorneAlexP: probably. i forget. the page on the wiki says the things i've found
20:33:22Torneso yeah, you might find the only option that wold actally work would be a 120gb disk with 512 byte sectors (or emulation of same)
20:33:52 Quit MarcGuay ("ChatZilla 0.9.85 [Firefox 3.5.4/20091016092926]")
20:34:08*pixelma wonders why she bothers test building here while others seem to use SVN for that :\\
20:34:46TorneAlexP: if i remember i'll check whether lba48 works and post that as well, as i seem to have forgotten to state either way :)
20:35:02AlexPTorne: That's what I have now, but 120GB is too small :)
20:35:07AlexPTorne: Thanks :)
20:35:10TorneOhr ight
20:35:15Torneyou already upgraded it that much?
20:35:30Tornewell that confirms part of my research at least :)
20:35:37AlexPHence wanting the 240 GB one :)
20:35:55Tornewell, i don't recall *which* drives it is that fail to do 512 byte sector emulation
20:36:02Torneother than the ones in the ipodvideo
20:36:32AlexPI might do a bit of searching
20:36:53Tornei pondered if you could trick the beast into working on such a drive. it might be possible with some major difficulty
20:37:11Torne(it checks the sector size in the fat BP is 512, but we could lie about that)
20:37:11 Join StealthyXIIGer [0] (
20:37:29Tornebut whenever it decided to format itself, it would temporarily brick itself by failing to be able to write
20:37:41Torneand you'd have to pull the drive out and reformat it suitably outside the beast before it would work again
20:37:57Torneso that might be considered inadequate even if it was possible by some torturous process ;)
20:38:01AlexPI've never had it format itself, and have a 1.8" caddy
20:38:10AlexPSo that *might* work :)
20:38:19Torneyou would probably need a custom rockbox build with nasty hacks
20:38:29AlexPWell, let's forget about that
20:38:38AlexPIt sounds like more trouble than it is worth
20:38:49Tornelikely :)
20:40:34 Join AEnima1577 [0] (
20:46:34 Quit TopyMobile_ (Read error: 113 (No route to host))
20:48:41CIA-6New commit by bluebrother (r23556): Update portugues-brasileiro lang file (FS #10765 by Adilson Xavier).
20:55:22 Join grant__ [0] (i=d83bf855@gateway/web/freenode/x-xbyyjkcjjcupbxvv)
21:03:33CIA-6New commit by bluebrother (r23557): Allow the bin2c output filename stem to contain a path.
21:03:38CIA-6New commit by bluebrother (r23558): Fix a compile warning.
21:08:28 Quit grant__ ("Page closed")
21:11:31***Saving seen data "./dancer.seen"
21:11:43 Join Bob_C__ [0] (
21:13:54 Quit freddyb ("Konversation terminated!")
21:27:23 Join TopyMobile [0] (
21:41:51 Join phanboy4 [0] (
21:48:42 Quit Zambezi ("2weeks idletime.")
21:52:34 Quit TopyMobile (Read error: 60 (Operation timed out))
21:57:53bertrikmarkun, since I'm a bit stuck with getting past the LCD initialisation for the meizus, I think I'll do some of the simpler stuff first, like buttons and touchstrip for the meizu M6SP
21:58:57GeekShadowbertrik, will the M6 work will be useful for the m3 one .?
21:59:24bertrikprobably not, the m3 is already ahead of the other meizus
22:01:02 Join Strife89 [0] (
22:01:22bertrikthe init sequence of the sdram on the meizu m3 is something that does benefit from m6 experiments though
22:02:12GeekShadowok cool
22:02:36GeekShadowI'm not sure how I can help with rockbox
22:02:53GeekShadowis there l10n work ?
22:08:56 Join knittl [0] (n=knittl@unaffiliated/knittl)
22:09:25knittlhi … i am experiencing strange problems with the rockbox repository
22:09:46knittli made a git clone, and also kept rebasing svn
22:10:01knittlbut since 300 revisions git and svn history are in two different branches
22:10:07knittlwith one commit being different
22:10:39knittlmaybe someone can enlighten me on the subject :)
22:11:47 Quit Zarggg ()
22:14:26bertrikmarkun, what shall we call the buttons on the m6?
22:15:34bertriklike, should the "M" button be menu or up? and should the "<-" button be enter or down?
22:17:11bertrikI'll try menu and enter (instead of up and down) and see if anything fails because of the lack of up/down buttons
22:18:30 Join dfkt [0] (
22:21:23 Quit GeekShadow (Read error: 54 (Connection reset by peer))
22:23:04 Quit dfkt_ (Nick collision from services.)
22:27:04 Join Zarggg [0] (
22:27:09 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
22:27:40 Quit Xerion (Read error: 54 (Connection reset by peer))
22:31:39 Join saratoga_lab [0] (i=9803c264@gateway/web/freenode/x-mhjgbktkenwazcnz)
22:37:31CIA-6New commit by bertrik (r23559): Meizu M6SP: initial button driver (adapted from Meizu M3)
22:40:24Unhelpfulbertrik: cancel/back seems the obvious meaning for <-
22:57:17 Join pamaury [0] (n=pamaury@
23:00:46 Join tchan1 [0] (
23:00:48 Quit tchan (Read error: 60 (Operation timed out))
23:01:03 Join Strife1989 [0] (
23:02:43 Quit Bob_C__ ("ChatZilla 0.9.85 [Firefox 3.5.4/20091016092926]")
23:05:06bertrikUnhelpful, the schematic of the M6 calls it "enter"
23:06:25Unhelpfulhrm. well, i don't have an M6. the similarly-labeled button on my beast is back, and that makes sense based on the symbol and the OF's use of it...
23:11:27CIA-6New commit by FlynDice (r23560): sd.h: No functional changes to present code. Added some additional SD CMD's and a #define for card errors flagged in an R1 response. ...
23:11:35***Saving seen data "./dancer.seen"
23:12:50 Quit petur (Read error: 110 (Connection timed out))
23:16:01 Join toffe82 [0] (
23:17:29 Quit Strife89 (Read error: 110 (Connection timed out))
23:21:19 Join KBH [0] (
23:25:03 Join hiiiiiiiiiiiiiii [0] (
23:25:41 Quit hiiiiiiiiiiiiiii (Client Quit)
23:27:37 Quit HBK (Read error: 60 (Operation timed out))
23:27:38 Nick KBH is now known as HBK (
23:30:08Unhelpfuldo we have a set of redistributable codec test/benchmark files? if not, do we want one?
23:30:47UtchybannTheSeven: keeping mutex and nand_powered = 0 are enough to fire the problem. 'sync' flushes the buffers to disk but never finish.
23:31:22TheSevenis there anything special about a sync?
23:31:36TheSevenas far as i can see, it's just a bunch of ordinary writes at the storage layer
23:33:03UtchybannTheSeven: I just sync before unmounting (unmouting also flushes buffers to disk).
23:33:18TheSevenso does unmounting also fail?
23:33:31UtchybannTheSeven: not tried, but I guess yes.
23:33:45*Utchybann need to restore is ipod.
23:35:20*TheSeven can't see the connection between a sync and nand power management
23:35:45TheSevenletting it sit for some seconds and then transferring some data again should be the same from that point of view
23:36:28 Quit pamaury (Read error: 60 (Operation timed out))
23:37:39UtchybannTheSeven: true, I will try an umount without manual sync. but it could be specific to the nand type. My other nano2g does not have this problem.
23:40:04 Quit Utchybann ("I like core dumps")
23:41:53TheSevenwell, *either* there is something specific about writing to the superblock (or an arbitrary block multiple times?) which points towards an ftl bug, *or* it only happens if nand power management is enabled => nand pm bug
23:42:01TheSevenbut I can't see any connection between those
23:42:10TheSevenso you say you can copy files just fine
23:42:38 Join BlakeJohnson861 [0] (
23:43:24TheSeventhen try copying ~100MB, letting it sit for a minute, copy another ~100MB, and then immediately unmount, (cp ... /media/ipod && umount /media/ipod), and have a look what happens
23:43:59TheSevenI have no explanation why the unmount could fail in that condition
23:47:32 Quit dmb ("Leaving")
23:47:47 Join Utchybann [0] (
23:49:36CIA-6New commit by kkurbjun (r23561): M:Robe 500 - Add some more debug information.
23:49:51UtchybannTheSeven: I will try. just make a copy of the mbr and 1st partition to be able to do a manual restore.
23:52:43 Join TopyMobile [0] (
23:54:24TheSevenUtchybann: I have images of that lying around if you need one
23:54:39TheSevenjust tell me your flash size and if you want the rb bootloader or iloader
23:55:05CIA-6New commit by kkurbjun (r23562): M:Robe 500: Menu button is now power for JPEG viewer
23:55:22 Quit BlakeJohnson86 (Read error: 110 (Connection timed out))
23:56:01CIA-6New commit by kkurbjun (r23563): Test codec: Add exit button for touchscreens
23:56:29TheSevenwith these you'll just need to dd it, unplug/replug, extract to the already formatted data partition, and unplug it again

Previous day | Next day