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 2009-07-06

00:00:54webguest89Haven't used that in ages. But since it's a matter of just files, I don't think I need that. Oh, and here's something interesting: Turning on fade in prevents the light from going oon, but if fade out is also turned off, the light will switch on and off just fine. If I activate fade out, it won't turn off or on anymore (even if DISABLING the light) If I leave the light disabled and then go and touch the "fade
00:01:39webguest89And it's a clean rockbox install.
00:01:54gevaertswebguest89: rockbox is *not* rockbox utility...
00:02:06gevaertsrockbox is what runs on the player
00:02:45ZagorBagder: longest post yet? :)
00:03:05Bagderquite possibly, yes ;-)
00:03:19gevaertsI'm pretty sure that there are longer ones
00:03:50Bagder986 words wordpress says
00:03:56webguest89and where is that menu located? does it do anything to that bootloader thing that I don't ever get to see (unless installing or uninstalling rockbox completely)
00:04:20ZagorBagder: how many of those are "build"?
00:04:27 Quit Lss (Read error: 104 (Connection reset by peer))
00:04:36n1sBagder: nitpick s/a around/a round/ otherwise nice writeup! :)
00:04:46Lloreanwebguest89: What menu?
00:04:50Bagdern1s: thanks, fixed that
00:05:07linuxstbwebguest89: Main menu -> Settings -> Manage Settings -> Reset settings
00:05:36Lloreanwebguest89: When did you last update your bootloader, exactly, anyway?
00:05:43 Join barrywardell [0] (n=barrywar@
00:05:49saratogai wonder why the show fps text in mpegplayer flickers so badly on the ams targets
00:06:25linuxstbDoesn't it flicker everywhere?
00:06:38 Quit flydutch ("/* empty */")
00:06:41webguest89light's out again :) that's what teset settings did :D
00:07:56webguest89ugh, and my Icons are screwed up
00:08:27LloreanYes, they're part of "settings" which you just reset...
00:08:33LloreanAbout that bootloader?
00:09:09webguest89yep. that one DOESN'T need fixing/updating? right?
00:09:25LloreanIt needs updating occasionally. When we release new versions.
00:09:30LloreanWhich is why I'm asking.
00:09:51webguest89erm... never updated it
00:10:10Dhraakelliananyone have any advice for running on OpenSUSE 11.1?
00:10:32webguest89*ding* might be the solution to my problem, right? (or might screw it up worse)
00:10:33DhraakellianI'm running into errors with /usr/include/asm
00:10:38BagderDhraakellian: can you build "normal" things with gcc?
00:10:58Lloreanwebguest89: Seriously, why not just answer the questions you're asked instead of trying to figure out why people are asking them? It would make this go a lot faster.
00:11:11LloreanI can't suggest whether you need to update the bootloader until you tell me when you installed the one you currently have.
00:11:12Dhraakelliany'know, come to think of it, I'm not sure I've tried
00:11:19LloreanOr give me its version. Either bit of info will do.
00:11:23Dhraakellianwait, yes, I believe I have... a plasmoid or two
00:11:34BagderDhraakellian: that 'asm' dir is supposed to be there
00:11:37Dhraakellian(although those may have been builton the laptop)
00:11:45mtIn rm, when skipping a track, there's no problem if I skipped to the next track, but if I chose to skip to a previous track, it sometimes just stops. Sort of like when seeking in a codec that doesn't support seeking.
00:11:55Bagderor to be a symlink
00:12:04DhraakellianBagder: /usr/include/asm-generic and a number of /usr/include/asm-<arch> dirs
00:12:16linuxstbmt: The first skip backwards causes a seek to the start of the current track.
00:12:23mtIs this related to the 'no seeking support yet' or is it a porblem that should be fixed ?
00:12:24BagderDhraakellian: so you want the 'asm' to be a symlink to your arch's asm-<arch> dir
00:12:46CIA-71New commit by kugel (r21680): Add a possibility for plugins to go directly to the WPS after exiting. ...
00:12:54linuxstbmt: It should be easy to add support for that specific seek operation. Other codecs may do it (or may have done it in their history).
00:13:17 Quit stripwax ("Leaving.")
00:13:18BagderDhraakellian: but it does indicate a flawed install somehow
00:13:25Bagderof the kernel headers
00:13:35webguest89wait a moment... how do I do that?
00:13:41linuxstbmt: shorten.c looks like it does that.
00:13:52webguest89got it
00:13:59DhraakellianI symlinked everything from -generic and -x86 into asm/ (and renamed the sigcontext32.h symlink to sigcontext.h, but now I'm getting some errors in that file, which makes me think that maybe that wasn't the proper solution)
00:14:18mtlinuxstb: So just this specific seek operation should be handled now until full seeking is supported ?
00:14:55Dhraakellianbut that's dealing with 10.3 stuff, and this is 11.1
00:14:58linuxstbmt: Yes, it would be nice to do that. But no need to do it before a commit IMO.
00:15:11CIA-71New commit by kugel (r21681): Adapt pictureflow and random_folder_advance_config to make use of the new goto-wps exit feature. ...
00:15:19 Part domonoky
00:15:19webguest89version 2.0 . It stays on for just 1 sec so it might take a while
00:15:25BagderDhraakellian: you would make an 'asm' symlink to point to asm-x86
00:15:47Bagderor asm-i386 or whatever the name is ;-)
00:16:01Lloreanwebguest89: Try having RBUtil or the newest version of iPodpatcher update your bootloader.
00:16:15linuxstbmt: Also, have you tested that Rockbox rejects other types of .rm files cleanly?
00:16:16mtlinuxstb: Are you free for a minute to check a patch ? If yes I'll upload one so that you could review it before I commit.
00:16:30mtNo I didn't.
00:16:30linuxstbmt: I can have a quick look, yes.
00:17:26DhraakellianBagder: I think it was wanting files from both asm-x86 and asm-generic
00:17:36webguest89but will the new bootloader be compatible with that version of rockbox I got a half a year or so ago? cause if not... screw this, as soon as I get home, I'll put my old version back again
00:17:54BagderDhraakellian: no, 'asm' is the arch-specific directory
00:18:16MikachuDhraakellian: those directories are the headers for the kernel source iirc, try installing a linux-headers package or something like that?
00:19:13Dhraakellianlinux-kernel-headers is already installed
00:19:23*Dhraakellian will try a -f
00:19:25Lloreanwebguest89: No, we don't guarantee things will work with severely outdated builds or bootloaders. But since nobody else, including me (and I have a Nano) is experiencing your problem, once you're up to date you shouldn't need to use extremely outdated builds.
00:19:54Lloreanwebguest89: But, thanks for expressing your unwillingness to help figure out whether this is a real bug or not if it's likely to cause you any real inconvenience. Nice community spirit, that.
00:20:02gevaertswebguest89: the "new" bootloader is from october 2008, so it should work with any rockbox version newer than that, and probably with some older versions as well
00:21:30 Join funman [0] (n=fun@rockbox/developer/funman)
00:21:35webguest89well, I need it to work. But okay, I'll install the new bootloader. And if it won't work, you will help me put back my old one.
00:22:12Lloreanwebguest89: No, if it doesn't, we'll continue trying to figure out why it doesn't. Because that's what actually matters.
00:22:28 Quit rasher (Read error: 110 (Connection timed out))
00:22:31LloreanIf there's a bug it needs to actually be *fixed* not ignored.
00:22:41linuxstbmt: Do those DEBUGFs mean that the sim will display lots of debugging information for rm files now?
00:22:56saratogarunning an old rockbox build on the ipod doesn't make much sense given how much we've changed recently
00:23:44webguest89Uh ok. As long as it will work some day. I could tolerate a broken player for a few weeks or so... So. Doing that bootloader thingy now.
00:24:29 Quit n1s ("Lämnar")
00:24:36Dhraakellian"zypper in -f linux-kernel-headers" and then retrying hasn't crapped out yet
00:24:52webguest89@ saratoga: indeed, but at least that one _worked_ :)
00:24:52mtlinuxstb: yes.
00:24:59Tornewell, the beast flash has a configuratino area, which contains a bit which tells it what to do if the bootloader encounters a fatal error
00:25:06Torneif it's set to 1 it formats and reboots
00:25:16linuxstbmt: Hmm, you shouldn't declare those metadata variables - you're adding 1KB to the size of Rockbox. (char title[256] etc)
00:25:30Tornei've not tracked down specifically where it decides the disk is wrong, but this is probably where it ends up
00:25:34Lloreanwebguest89: The current one works too, for everyone else.
00:25:35gevaertsTorne: you mean it's possible to just tell it not to do that?
00:25:50Lloreanwebguest89: But you chose to partially updated.
00:26:07linuxstbmt: And you should disable those debugs when you commit - it would be chaos if every part of Rockbox did that.
00:26:07saratogacurrent works fine you're just doing something wrong
00:26:16Tornegevaerts: it's quite possible there are *other* places in teh code which trigger the same behaviour
00:26:24Tornegevaerts: but this is certainly one, yes,a nd it could be disabled
00:26:43Lloreanwebguest89: RButil can install an updated bootloader for you in a flash, so it shouldn't take long for you to test.
00:26:47Tornewriting to the config area in the flash seems *reasonably* safe. if its checksum is wrong it just re-initialises it from preset valus.
00:27:10Tornethe problem is its not going to make the device usable :)
00:27:27Torneif you flip the bit it skips the logic to format itself but it still goes to the infinite "crashing" loop
00:27:40Torneso the bootloader will presumably never work again then, unless you pop the drive out and fix it externally :)
00:27:45mtlinuxstb: At first I just copied the tags directly to the id3v1buffers, but then I did this thing because it looked neater. :) I could revert back to copying directly in the buffers.
00:28:04Tornegevaerts: but it seems like a relevant discovery at least.
00:28:14 Part wincent ("Kopete 0.12.7 :")
00:28:18linuxstbmt: Yes, copying directly to the buffers is better.
00:28:32webguest89I do admit I want the newer version as well. I sense a png image viewer coming up :D Yep, downloading RButil now...
00:28:51saratogaspeaking of rbutil, whats left to get it to install the AMS targets?
00:29:03saratogajust adding that list of firmware hashes to the website?
00:29:03LloreanTorne: I've had my Beast claim it needs to reformat when booting into the OF using Dual Boot.
00:29:23LloreanI rebooted into Rockbox, used it a few days, then accidentally booted the OF again later, and it didn't claim it needed to reformat
00:29:30TorneLlorean: yah. it looks like that's the fallback logic for *any* error at boot.
00:29:32Torneliterally anything
00:29:34LloreanSo there's some cases, at least, where it can mis-detect it needs to reformat
00:30:00linuxstbmt: Also, why use the id3v1 buffer? Is there a limit on tag length in RealAudio (you're limiting things to 32 characters if I understand correctly).
00:30:06*Llorean doesn't know why it stopped deciding it needed to reformat, though
00:30:08Tornei've not gotten too far into the boot process yet, not even to disk init, but already there are *loads* of places that call that error ahndler :)
00:30:44TorneLlorean: well ideally i'm going to work right through the boot process until it hands control to the nk.bin on disk, and make a list of every reason it might decide to format itself. :)
00:30:50 Quit saratoga ("CGI:IRC (EOF)")
00:30:53TorneLlorean: then we can look and make sure we never do those things ;)
00:31:05 Join saratoga [0] (i=9803c6dd@rockbox/developer/saratoga)
00:31:50TorneLlorean: also hoping i will discover why later firmwares don't seem to allow the single boot bootloader.
00:32:12LloreanThat's the most important one right now
00:32:24Torneyah. im not looking for anything in particular though
00:32:24LloreanSince it's the main thing preventing the Gigabeat S from upgrading to "supported" I believe.
00:32:28Tornei am literally trying to grok the entire boot process
00:32:32 Quit matsl ("Leaving")
00:32:34Torneit's too awkward to find anything specific :)
00:34:11webguest89Yup. version 3.0. And it boots up with less light. It's done. Checking the stuffs now...
00:34:52webguest89Um, yes it works just fine ^^;
00:35:33LloreanThis is part of why we ask you make sure everything's up to date before asking for help. ;
00:36:49mtlinuxstb: MAX_STRING is 92 bytes (the size of a id3v1 buffer). I believe this would be enough for one tag, right ?
00:37:11linuxstbmt: Sorry, I misread that as 32...
00:37:16*linuxstb cleans his monitor
00:37:20 Join legendarymidget [0] (n=3ce62093@gateway/web/cgi-irc/
00:37:41linuxstbmt: Well, a comment could be longer...
00:37:53webguest89Thanks for your help! :D And please excuse one more question... theme compatibility? why won't old themes work on the new rockbox. I loved my customized BlackGlassMgrbis
00:38:07LloreanBecause certain parts of the theme language have changed.
00:38:25Lloreanit's usually a relatively simple process to update.
00:38:46linuxstbwebguest89: Were you using a custom build of Rockbox in the past? (i.e. one with patches to add new theming features?)
00:38:50 Quit legendarymidget (Client Quit)
00:38:54mtlinuxstb: Comments could be read to id3v2buf then ?
00:39:01mt300+ bytes iirc
00:39:45linuxstbmt: Well, other codecs just treat the id3v1 and id3v2 buffers as one large buffer, and put the tags there.
00:40:15linuxstb(see vorbis.c for example)
00:41:58webguest89custom themes only (and tweaked by my hand to stuff in as much useful info as possible). Never patches of the actual software. Tho I downloaded cygwin today and plan to play around with it and compile myself that png plugin...
00:43:34webguest89I wish somebody posted it as a rock file and spared me all the hassle. I only program in visual c++
00:43:47linuxstbwebguest89: If you post your theme somewhere, maybe someone will look at it and help you find the problem. No-one can say why your theme won't work without seeing it.
00:44:16krazykitwebguest89, you can't really post a rock file, as it'll quickly stop working on newer builds.
00:45:05Torneanyone got a beast they'd be able to dump the flash from for me? :) patch for flash dumping is on FS #10410
00:45:21mtlinuxstb: OK, so now I've modified the parser to write directly to id3v1buf, and disabled DEBUGF (could be re-enabled with a define). The problem now is that a comment could be bigger than 92 characters, right ?
00:45:25Tornei could do with a couple more flashes to compare
00:45:46webguest89I don't intend posting it. I just noticed it's totally screwed up with the new firmware. But I'm gonna fix it. I loved it. Ok, I understand.
00:46:55webguest89can't be that hard. where do I find some info regarding the new wps files?
00:47:25linuxstbmt: Another comment - do you need to read desc and mimetype?
00:48:26linuxstbmt: Anything could be bigger than 92 characters (although comments is most likely).
00:48:39Bagderwebguest89: you might be better off by getting a fresh theme that works, and then adapt it with the changes you want
00:49:54pixelmawebguest89: I'd suggest searching the forums a bit, there were a few threads about the changes. Or read up on the new syntax in the CustomWPS wiki and compare to yours
00:49:56 Join farthen [0] (
00:50:30webguest89yeah, i put the new default cabbiev2 for the time being. It has album arts :)
00:50:37mtlinuxstb : No, no need to. Will lseek beyond them.
00:51:04webguest89ok will google that.
00:52:52***Saving seen data "./dancer.seen"
00:54:49mtlinuxstb: Done. But about the tags, do you think it's a no-commit issue ?
00:54:56linuxstbmt: Have you measured the binary size increase in Rockbox when applying your patch? (the rockbox-info.txt file shows that)
00:55:07linuxstbmt: No, it can be improved later.
00:55:56linuxstbmt: The reason I'm looking at the metadata parser is because it's part of core Rockbox - so it needs to be "lean and mean"...
00:56:22mtI see.
00:56:57 Quit mcuelenaere ()
00:56:58mtI just thought that improvements could be done when optimizing later on.
00:58:11webguest89Bye everybody, thanks for the help :)
00:58:25mtlinuxstb: What's the difference between 'binary size' and 'actual size' ?
00:58:36mtBinary size = 611328 btw.
00:58:55linuxstbmt: And without your patch?
01:00:31mthmm .. If you have time, I'll still have to checkout a clean copy. (I currently have no source copy without my patch :/ )
01:00:37 Quit webguest89 ("CGI:IRC (EOF)")
01:00:49*Mikachu hugs git
01:00:49 Quit ender (" Everything we know about the Devil has been told us by the the friends of God.")
01:00:59linuxstbYou can just reverse your patch (patch -R)...
01:01:46Mikachu(and if you don't have a patch yet, svn diff > mypatch.patch)
01:02:11linuxstbOr svn revert...
01:02:22Mikachusvn revert is not good if you haven't saved a patch :>
01:02:39linuxstbYes, it does have that flaw...
01:03:04linuxstbmt: I think actual size includes any headers or checksums added to the main binary.
01:03:41mtlinuxstb: But actual size is smaller than bin size ?
01:04:05mtMikachu : I had the patch and was able to reverse it. (phew!)
01:04:34Tornedoes actual size not include bss?
01:04:51linuxstbmt: Then I guess it's the other way around - "actual size" is the size of the binary, and "binary size" is the binary + headers etc.
01:05:09mtlinuxstb: Increase in bin size is 3072
01:05:28linuxstbTorne: Neither includes bss
01:05:46Mikachuisn't bss all zeroes? :)
01:06:12linuxstbInitially, yes...
01:06:27Tornelinuxstb: it seems like there should be a number that does.. :)
01:06:34Tornesince bss takes up ram
01:06:52linuxstbOr "RAM usage" to be precise
01:07:38mtlinuxstb: So what is the significance of this increase in the bin size ? and when is it considered big ?
01:07:40linuxstbmt: Now you need to decrease that number as much as you can...
01:07:48Tornehow would i go about moving the rom dumping code into target-specific bits of firmware/ ?
01:08:32funmanwhich code ?
01:08:32Torne(i mean where to put it, what header to stuff itin, how to know if it's present)
01:08:43linuxstbmt: You simply need to make it as small as you can...
01:09:02 Quit Thundercloud (Remote closed the connection)
01:09:09Tornefunman: kugel rightly pointed out that the rom dump code in apps/debug_menu.c is target-specific and the file is already rather complicated and huge
01:09:25 Nick fxb is now known as fxb__ (
01:09:31linuxstbmt: More RAM used by Rockbox itself means less RAM for audio buffering and other things. So we want to pay attention to it.
01:09:33Tornefunman: and i'm adding yet another set of #ifdefs for the beast
01:10:11mtlinuxstb: If I'm allowed a stupid question - How could I decrease it ? What affects bin size in a codec?
01:10:25funmanTorne: just move it in firmware/target/sh/archos
01:10:45Tornefunman: but then am i gonna need a config-*.h flag to say whether the target has that
01:10:53linuxstbmt: I'm just talking about the metadata parser (apps/metadata/rm.c). You make it smaller by writing the code more efficiently, and making sure you've removed any unneeded code.
01:10:54funmanand create a debug-target.h file next to it, it will be included automagically
01:11:25funmanTorne: sorry i'm wrong, debug-target.h inclusion is conditional (at the top of debug_menu.c)
01:11:52Tornebut debug-target is a good place to put it at least?
01:12:23funmanother targets use debug-XXXX.c / debug-target.h
01:12:34 Join xnyhps [0] (n=xnyhps@2001:470:1f14:da:219:e3ff:fed7:c57c)
01:13:07Tornethen just a define in there for DEBUG_CAN_ROMDUMP or something?
01:13:28mtlinuxstb: Ah ok.. If it's alright, I could commit the changes now and work on that tomorrow (I really have to sleep now :) ). If not I will wait till I've decreased bin size before committing any changes.
01:13:44Tornethis is the first thing i've changed in core that wasn't localised, so not sure what is good form :)
01:15:22funmanTorne: hum what is this function's name ? (debug_menu.c is *very* cluttered)
01:15:22linuxstbmt: It's never a good idea to commit when you're tired (in case something goes wrong and you need to fix it...)
01:16:22Tornefunman: dbg_save_roms
01:16:41Tornefunman: there are already three versions, and imx31 makes four :)
01:16:54mtlinuxstb: Good point :) .. I'll just upload another patch now and leave everything else for tomorrow.
01:17:13funmanTorne: whatever you choose to do, we should do the same for other target specific options
01:17:20Tornefunman: yeah i was looking at that
01:17:24Tornesome of them are kinda done that way already
01:17:37Tornedbg_ports has code in debug_menu.c for a few targets, then just calls __dbg_port for all the others
01:17:45Tornewhich is indeed in debug-target.h
01:17:46funmanTorne: or even provide a way for targets to add their own entries in the menu (from their target specific code and not debug_menu.c)
01:18:10Tornethat could be a *lot* tidier, yes..
01:18:20Tornethough the options might end up in different orders then :)
01:18:24funmanis there any common code in this file ?
01:18:46Tornevery little
01:19:24funmanlogf/cpuboost/screendump ..
01:19:31 Join farthen_ [0] (
01:19:38Torneh, actually yah
01:19:38Torneand most of the disk stuff
01:19:39Torneand usb hid
01:19:40DBUGSent KICK Torne to server
01:19:40Kick(#rockbox Torne :No flooding!) by logbot!n=bjst@rockbox/bot/logbot
01:19:51 Join Torne [0] (
01:20:01funmanI think there some people mentioned problems with the debug menu
01:20:07Torneok i need to press enter less often.
01:20:09funmanlike items not belonging there
01:20:12funmanTorne: ^^
01:20:29Tornei have a bad habit of punctuating with carriage return instead of comma
01:20:32funmanalso the possibility to remove the debug menu from release builds (which could save a bit of binsize)
01:20:52 Join dmb [0] (n=Dmb@unaffiliated/dmb)
01:21:11Torneit does seem ripe for a reorg, yes.
01:21:24Tornein which case i might, er, *not* volunteer for that one :)
01:21:47funmanTorne: well i could help you
01:22:06funmanI did that for the timer.c file
01:22:16*linuxstb sees Torne is finding the "start to make a simple change and end up refactoring a large chunk of Rockbox" effect...
01:22:28Tornelinuxstb: that's called "software development" isn't it? @)
01:22:42funmanTorne: i think there isn't much target specific bits in apps/ now
01:23:50Tornei think having the targets add their own menu options would be a good start for quite a few of those options
01:23:52pixelmaif someone is *really* bored, he or she could go through all the button tables in the manual's tex files and change them to one readable indentation and also writing style... :\
01:24:16Tornesince it looks like most of the implementations of those are completely seperate and don't depend on anything else in debug_menu.c
01:25:06Tornealso the really target specific stuff would probably be first on the chopping block to remove stuff from release builds
01:25:20kugelTorne: isn't this for the beast
01:25:30Tornekugel: that's the one i'm trying to add, yes :)
01:25:38Tornethe patch you looked at yesterday
01:25:43kugelthat one already has debug-XX.c
01:25:53funmanTorne: timer.c went from 365 lines to 63 (almost 6x less) : much more readable .. for both common code and target code
01:26:31Tornekugel: yes, but i was gonna move the other target's dbg_save_roms as well for consistency. and now i'm looking at all the *other* bits of that file that are a big ifdef'ed mess :)
01:26:36funmankugel: but debug-XX.c only has 2 menu entries (view hardware info and view ports)
01:26:46kugela #define in debug-target.h seems sensible, and some sort of debug_rom_dump(void) { target_rom_dump() }
01:26:52Tornenot all targets use debug-XX.c for those entries either
01:27:02Tornethe older ones are implemented right in debug-menu.c
01:27:50Tornekugel: that sonuds reasonable but having the targets add menu entries would remove even more :)
01:28:25kugelfor specific stuff yes, but I thought rom dump is relatively common?
01:28:44Torneit's only implemented for three processors
01:28:47funmanthere would be no common code anyway between the targets
01:28:49Tornefour, with the imx31
01:29:46Tornethe effort of actually making it generic would be sprintf'ing filenames with address ranges and the like
01:29:59 Quit farthen (Read error: 110 (Connection timed out))
01:30:05Tornewhih probaly works out as more code than just having the targets do it themselves with no shared code :)
01:30:30kugelI think we can have targets add menu entries also
01:30:32Tornenot all the targets have only one rom, and so on
01:30:47 Part xnyhps
01:31:22funmanthe target code could even be moved progressively
01:31:32 Quit saratoga ("Page closed")
01:32:32kugelsadly, I have no real idea how simplelist works, but I think items can be hidden based on some callback
01:33:36kugelTorne: but almost all targets have rom dump as you can see
01:33:53kugelit's basically only the flash-based PPs that don't have it (and the ams sansas)
01:33:55funmanthe menu is just a structure with { "string", function } elements
01:34:37Torneyah, simplelist is too simple for this :)
01:35:15Torneso just having defines seems like a better plan for now
01:35:50Torneit'll still be way shorter
01:35:54Torneand it could always be done more later
01:36:12funmanshorter than writing the function in debug_menu.c ?
01:36:44Tornedebug_menu.c can just have the menu item in an #ifdef DEBUG_ROM_DUMP
01:36:52Torneand call the function in debug-XX.c directly
01:37:30funmanyou could just add your targets to the list without using a DEBUG_ROM_DUMP
01:37:49Torne#ifdef huge list of cpus is kinda ugly though :)
01:38:08funmanyes but #ifdef huge list of cpus mixed with capabilities defines is even more ugly
01:38:39*funman adds debug_menu.c cleaning in his TODO list
01:38:47Torneyou wouldn't need the cpu defines any more, is the point
01:39:17funmanTorne: yes but there is a lot of other functions conditional on "list of cpus", no ?
01:39:18kugelyou want to add debug-sh.c, debug-cf.c and and debug-pp.c?
01:41:34Tornei dunno, that's why i was asking :)
01:41:47kugelyou could totally do that imo
01:42:42*Torne volunteers funman :)
01:43:11*Torne is having too much 'fun' reversing the beast bootloader
01:43:48funmanTorne: well i make no deadline promise, but i'm willing to do this
01:44:05Tornefunman: it's not particularly vital from my pov
01:44:16Tornehaving the beast dumping patch committed is not that important
01:44:33Tornei am guessing anyone willing to dump a beast rom for me is already building it themselves :)
01:46:43 Quit funman ("free(random());")
02:02:10 Join kadoban [0] (
02:06:32 Quit barrywardell ()
02:07:31CIA-71New commit by kugel (r21682): Fix returning too early (before cleanup) in RFAC, which led to not cancelling ...
02:18:53CIA-71New commit by zagor (r21683): Killed children must let go of their cores.
02:20:30 Join homielowe [0] (n=homielow@
02:21:02CIA-71New commit by zagor (r21684): Moved handoutbuilds() to last, to avoid end-of-round being ran before end-of-build. Skipped the .info file, since that's already logged in the ...
02:30:46 Quit efyx_ (Read error: 104 (Connection reset by peer))
02:32:39 Join efyx_ [0] (
02:34:07Zagorgetting closer:
02:34:40kugelZagor: uhh nice!
02:35:10Zagorthe last round was as fast with 12 server as the old build was with 22!
02:35:36kugeland the fast servers are still working for the old system, I guess?
02:37:23Zagorwell it's a mixed bunch in both the old and new server pools
02:38:16kugelI see, still impressive. The new system is the win
02:38:35kugeldoes it already do "live updates" also?
02:39:29CIA-71New commit by zagor (r21685): Bumped minimum client revision to 11.
02:40:57Zagorkugel: yes it does
02:41:12Zagorwe added that during devcon already
02:41:28kugelheh, I didn't know
02:42:30kugelthe logs aren't accessible yet though
02:48:57kugelZagor: how does the ranking of the heaviness of a build work?
02:51:31Zagorthe server simply sends the heavy builds to the heavy servers
02:52:30kugelI mean how does the server know a build is heavy?
02:52:55***Saving seen data "./dancer.seen"
02:53:37Zagorwe keep it in a table. it is currently a static table derived from gevaerts' binsize builds, but the idea is to create the table dynamically using statistics from previous builds.
02:54:17kugelHmm, I wonder if that is smart
02:55:30kugelgiving a fast client a heavy build will make it appear less heavy then. And a utterly slow client that needs 100s for a bootloader will let the build appear heavy, won't it?
02:55:53Zagornaturally the relative speed of the clients will have to be taken into account
02:56:11kugelI'm thinking the length of the logs (i.e. the number of compiled files + genlang etc) is a good messure
02:56:26kugelbut what do I know :)
02:57:15Zagoractually we plan to simply to a "calibration run" every 100 commits, where a single server builds all targets. that eliminates the client speed from the rating.
02:57:57ZagorI really need to go to bed now... see you tomorrow.
02:58:00 Quit Zagor ("Clint excited")
02:58:10 Quit kugel (Remote closed the connection)
03:08:12 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
03:13:27 Quit farthen_ ()
03:39:32 Quit Sajber^ (Read error: 54 (Connection reset by peer))
03:42:45 Quit HBK (Read error: 104 (Connection reset by peer))
03:42:58 Join HBK [0] (
03:45:47 Join rasher [0] (
03:50:43 Quit bubsy (Client Quit)
03:56:06 Quit pixelma (Nick collision from services.)
03:56:06 Join pixelma_ [50] (n=pixelma@rockbox/staff/pixelma)
03:56:24 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
03:56:59 Quit amiconn (Nick collision from services.)
03:57:03 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
03:57:12 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
04:02:18 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
04:27:49 Quit efyx_ (Remote closed the connection)
04:41:43 Quit dmb (Read error: 113 (No route to host))
04:52:58***Saving seen data "./dancer.seen"
05:11:24 Join bubsy [0] (i=Bubsy@
05:14:18 Quit bubsy (Client Quit)
05:20:56 Join bubsy [0] (i=Bubsy@
05:21:16 Quit bubsy (Remote closed the connection)
05:22:16 Join bubsy [0] (i=Bubsy@
05:28:43 Quit homielowe ()
05:33:04 Quit Horscht ("Verlassend")
05:55:08 Quit FlynDice (Remote closed the connection)
05:58:53 Join FlynDice [0] (
06:13:34 Quit JdGordon ("Leaving.")
06:17:07 Join JdGordon [0] (
06:27:16 Join ReKleSS [0] (
06:38:52 Join donutman25 [0] (n=Dagni@
06:52:24 Join homielowe [0] (n=homielow@
06:53:00***Saving seen data "./dancer.seen"
07:15:05 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
07:21:02 Join stoffel [0] (
07:23:00 Part BHSPitMonkey ("Ex-Chat")
07:31:38drakonikWhat's the difference between the sim build and the actual build you'd install onto your player?
07:31:46 Part donutman25
07:33:31drakonikI ask because whatever it is, it matters. I copied the sim onto my ipod, wtih the full contents of the disk inside the simdisk directory. And the sim didn't crash. If it really was JUST a hardware issue, wouldn't the sim crash as well?
07:33:31LloreanOne has hardware drivers and works as an OS, the other is an application that runs on the PC.
07:34:08LloreanWhy would the sim crash? The sim has no hardware drivers at all, it just uses the host computer's operating system.
07:36:22drakonikI dunno. My understanding of the causes of the crash was "Your ipod hardware is faulty". I figured that if it was true, then the sim should choke on the hardware issue the same way the real software does.
07:36:56drakonikI dunno, I'm kinda grasping at straws now. I've exhausted my options and I'm not ready to give up yet.
07:37:49LloreanThe sim doesn't even work remotely like how the real software does.
07:37:58drakonikYeah, I figured.
07:41:36drakonikHm. Is there some kind of uberlogging option/tool that I could use? Comb through everything rockbox is doing before it crashes? Having the db log the file opening isn't sufficient cause the log doesn't stop at the same file every time. Somewhere there's a common cause for the crashes and I want to know what it is, in more detail than "your hardware's fault kthbai" =\
07:44:34LloreanYou're going to need to code up your own solution if you want more logging than Rockbox provide.s
07:44:44drakonikI figured as much.
07:48:02LloreanReally though, why are you so unwilling to believe it's a hardware issue?
07:48:35LloreanWhat about it makes you think there's something else?
07:52:19 Join Blue_Dude [0] (
07:53:01drakonikBecause my ipod has never EVER exhibited any behavior like this.
07:53:05drakonikIn its entire lifespan.
07:53:17drakonikNot a single crash, not a single battery glitch
07:53:25LloreanWhat is that supposed to demonstrate exactly?
07:53:35Blue_DudeA fix for the fixed point code bloat is at FS #10411. Thanks.
07:53:37LloreanProblems *start* which means there's a point before they were there.
07:53:38 Quit Blue_Dude (Client Quit)
07:53:41drakonikOf course.
07:54:56LloreanIs there a formula for consistently reproducing the problem? Have you isolated a file that triggers it each time? Can you cause it to happen on another iPod of the same model using the same version of Rockbox?
07:55:22drakonikI don't have access to another ipod.
07:56:16LloreanThe problem is, you're the _only_ person experiencing exactly this.
07:56:19drakonikAnd yeah, I know how the evidence looks. I just think that if my ipod were prone to having these kinds of fits, it'd have started showing a long time ago, and not the exact moment I installed Rockbox and tried to do a database scan
07:56:32LloreanWhy would it have started a long time ago?
07:56:52LloreanHardware fails all the time. iPods are particularly prone to start showing problems after they've aged some.
07:56:58drakonikOf course.
07:57:40LloreanSo you're going to either need to thoroughly investigate with a systematic plan, or just accept that failure may be the case.
07:57:52drakonikBut. Well, what exactly is a database scan, from an operational point of view? Opening a file, reading its contents/metadata, then closing it.
07:58:08LloreanComing up with random things really isn't going to help you, especially without developing a good enough understanding that you think trying the sim is going to tell you something.
07:58:33drakonikThe sim was a shot in the dark, I won't deny it.
07:59:13LloreanRationalizing isn't going to do too much good. Yes, that's more or less all a database scan is. But saying "so if other file openings work, why doesn't the database" doesn't necessarily mean anything.
07:59:22LloreanThe database opens a lot more files, a lot more quickly, than most other operations, for example.
08:00:08LloreanRather than try to rationalize why it's *not* hardware, try to find out specifically what it *is* by doing some deeper investigation. Logic games aren't going to get you much if they haven't gotten you anything already.
08:00:34drakonikWhich is why I asked about logging.
08:04:00LloreanAs I said, you'll have to code up your own solution for that if the current logging offered isn't good enough.
08:06:35 Quit stoffel (Read error: 104 (Connection reset by peer))
08:08:06drakonikYeah. I'll work on it in the morning, I'm in no fit state to code.
08:16:19 Quit JdGordon (Read error: 104 (Connection reset by peer))
08:29:51 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
08:30:46 Quit homielowe ()
08:33:19 Join ender` [0] (
08:37:25 Part safetydan ("Leaving.")
08:38:54 Join r4v5 [0] (
08:40:18r4v5So I'm using the most recent daily (r21682-090706) on my sansa e280 and usb has stopped working. In both mac os and windows, the device seems to disconnect itself or time out.
08:42:07 Join stoffel [0] (
08:42:57Lloreanr4v5: Unrelated, but we don't really recommend using the daily. Why not the current build instead?
08:43:16LloreanAlso, what build did you upgrade from?
08:44:17r4v5I wish I had kept track. I noticed the reversion from somewhere in the last month.
08:44:53r4v5Time to start with 3.3 and binary search through builds, I suppose.
08:45:11LloreanAnd why use a daily build rather than the current?
08:45:59r4v5I was using the current; I apologize, I used the wrong term.
08:46:33LloreanSo you skipped over about a month of builds and USB stopped working?
08:46:58LloreanAre you sure the cable hasn't just gone bad. It's surprising how often it's something as simple as that. We haven't had anyone else report problems or anything.
08:47:18r4v5I tried with a second cable because I was suspecting that, actually.
08:47:53r4v5Transfers work with the OF but using rockbox it times out. Moving the mousewheel gives me a screen at like half brightness compared to usual.
08:53:01***Saving seen data "./dancer.seen"
08:56:23 Join petur [50] (n=petur@rockbox/developer/petur)
08:57:36 Join Rob2222 [0] (
08:59:48Dhraakellianmy brother noticed that my e260v1 wasn't mounting on his XP laptop with recentish svn
08:59:55Dhraakellianworked with the OF
08:59:59Dhraakelliansame on my linux box
09:00:04r4v5Okay, it's working fine on linux for me now
09:00:19Dhraakellianreverted to 3.3, and it worked on my linux box. didn't try it on his laptop
09:00:31Dhraakellian(or is it vista? ...I don't recall)
09:01:04LloreanAll well and good, but you could've also filed a bug report documenting the version, the hardware, and how to reproduce...
09:01:08 Join flydutch [0] (
09:02:19*Dhraakellian wonders where to find version information in the zip file
09:02:32LloreanIn the rockbox-info.txt file
09:02:49DhraakellianI need to go to bed. I didn't even see that
09:02:50LloreanIt also can be viewed from the menus in Rockbox.
09:02:56 Join wincent [0] (
09:03:10DhraakellianLlorean: yeah, but my brother has my e260 atm
09:03:10LloreanAnd since you'd need to be in Rockbox to confirm the exact behaviour, including what your host OS says when attempting to connect, I don't know why you'd be looking in the .zip anyway
09:03:31DhraakellianVersion: r21593-090701
09:03:57LloreanI didn't ask for your version number. I asked you to file a proper bug report documenting as many useful details as possible.
09:04:13DhraakellianLlorean: I'll look into it tomorrow
09:04:42LloreanI'm not someone who can fix this, and bug reports are the proper way to make sure information is posted for the people who can look into it more deeply
09:04:53LloreanBut they should always be filed against the current version, with instructions to reproduce if possible
09:04:57GodEatermorning everybody
09:06:34Dhraakellianmy first instinct was to put it back to 3.3, since it'll probably be heading back down south with him when he goes home.
09:07:05DhraakellianLlorean: test with current version, describe behavior, how to reproduce, and possibly track down when it started?
09:07:36LloreanDhraakellian: Exactly.
09:08:24Dhraakellianshould probably also touch base with r4v5, since I only mentioned it right now because I saw what he said as I was passing by
09:09:34Dhraakellianoh, and just a general thanks/kudos to the devs specifically working on the SansaAMS ports
09:11:42Dhraakellian(not that the rest of the devs don't deserve thanks as well, of course)
09:12:07Dhraakellianand now I must head to bed before I make a total idiot of myself or veer off-topic
09:13:04r4v5Unfortunately, in my attempt to figure out steps to reproduce, it's working fine for me.
09:13:12r4v5Which I guess is fortunate.
09:16:06 Join n1s [0] (n=n1s@rockbox/developer/n1s)
09:16:10 Quit Rob2223 (Read error: 110 (Connection timed out))
09:16:42Dhraakellianwell, I'll try to see if I can reproduce it tomorrow
09:24:35 Join Thundercloud [0] (
09:31:04 Nick fxb__ is now known as fxb (
09:32:35 Join kugel [0] (n=kugel@rockbox/developer/kugel)
09:45:21 Quit stoffel (Remote closed the connection)
09:47:34 Quit Thundercloud (Remote closed the connection)
09:50:10 Quit Llorean ("Leaving.")
09:51:22 Nick fxb is now known as fxb__ (
09:51:24 Nick fxb__ is now known as fxb (
09:52:01 Join Llorean [0] (
09:52:16 Quit Llorean (Client Quit)
09:53:34 Join Llorean [0] (
09:57:59 Join _Auron_ [0] (
10:00:05 Part Llorean
10:00:34 Join Sajber^ [0] (
10:04:23 Join Llorean [0] (
10:07:39 Quit Llorean (Client Quit)
10:13:45 Nick fxb is now known as fxb__ (
10:13:57 Join Llorean [0] (
10:18:23CIA-71New commit by lenzone10 (r21686): Updated italian translation.
10:28:38 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
10:29:44 Quit XavierGr (Client Quit)
10:32:52 Join efyx_ [0] (
10:39:42 Quit flydutch ("/* empty */")
10:39:58 Join flydutch [0] (
10:46:42 Join DarkDefender [0] (
10:50:34 Join einhirn [0] (
10:53:03***Saving seen data "./dancer.seen"
11:16:46 Join stoffel [0] (
11:20:53 Join barrywardell [0] (
11:23:05 Join DarkSpectrum [0] (
11:27:44 Quit kugel (Nick collision from services.)
11:27:52 Join kugel [0] (n=kugel@
11:52:15 Quit perrikwp ("Leaving.")
12:01:35 Join robin0800 [0] (
12:16:52 Quit mt (Remote closed the connection)
12:24:55GodEateris the idea that the build client should clear up the build-nnnnn directories it creates ?
12:48:34 Quit Sajber^ (Read error: 54 (Connection reset by peer))
12:49:09 Join Sajber^ [0] (
12:53:05***Saving seen data "./dancer.seen"
13:02:33 Quit kugel (Remote closed the connection)
13:04:35 Join mcuelenaere [0] (
13:09:39 Join funman [0] (n=fun@rockbox/developer/funman)
13:28:07 Nick fxb__ is now known as fxb (
13:30:22 Join chrysaetos [0] (n=9d52e746@gateway/web/cgi-irc/
13:31:22 Quit chrysaetos (Client Quit)
13:32:43 Join mt [0] (n=mt@
13:43:27 Join linuxguy4 [0] (n=timj@
13:47:47 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
13:56:26 Quit linuxguy3 (Read error: 110 (Connection timed out))
14:01:31 Join paulk [0] (
14:02:33paulkHello ! Are the SDHC microSD cards supported by Rockbox (for a sansa 250) ?
14:03:14paulkmicroSDHC cards*
14:05:24 Join linuxguy3 [0] (n=timj@
14:06:02paulkgreat !
14:07:28paulkthank-you !
14:07:34GodEaterwe like the easy questions :D
14:09:28 Join kugel [0] (n=kugel@rockbox/developer/kugel)
14:17:24 Quit linuxguy4 (Read error: 113 (No route to host))
14:19:16mtlinuxstb: Shouldn't I consider RAM usage not Binary size ? Whatever modifications I make to metadata/rm.c only affect RAM usage.
14:20:15kugelfunman: so, after FlynDice's post, I think we can disable voltage scaling and make the forum post?
14:20:22funmanboth in fact, binary size is the size of the rockbox.XXX file, while RAM usage is the size filled by it (rockbox.XXX file + stack + bss)
14:20:40funmankugel: yeah, are the bootloaders and mkamsboot binaries online ?
14:27:23mtDoes using lseek() instead of read() make it possible to increase any of both ?
14:29:11funmanit's not related
14:29:35funmanthe binsize depends of the size of your code (functions), and the static variables you declare (they go in bss)
14:30:21mcuelenaeremt: did you read my Flyspray comment?
14:30:44mtmcuelenaere: No, will check now.
14:31:05kugelfunman: yes
14:32:19funmanthe fuze bootloader is 11kB bigger than e200v2 !
14:35:24mtmcuelenaere: Are you talking about metadata/rm.c ?
14:35:29 Quit paulk ("Ex-Chat")
14:35:41mcuelenaeremt: I think so yes, lemme check the patch
14:37:41mtmcuelenaere: read_uint* functions are already defined in metadata_common.[c/h] ..
14:38:08*mcuelenaere still thinks they are inefficient
14:38:16 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
14:38:53mtWell, they could be altered, but this will be sort of a big change. :)
14:39:13mcuelenaereyes ok, but that's unrelated to your work so forget I said it :)
14:39:23 Join perrikwp [0] (n=Keith@
14:41:21mtI think maybe later (after soc ends :) ) new functions could be written, with the current ones intact, until all metadata parsers are using the new ones.
14:41:32 Quit robin0800 ("Leaving")
14:42:32kugelfunman: the logo is bigger on the fuze
14:43:36 Join robin0800 [0] (n=robin080@
14:43:51funmanhow come since they use the same screen size (only orientation is different) ?
14:44:29kugelthe logo is full width
14:44:37kugelnot full <whatever side is larger>
14:44:51 Join MethoS [0] (n=clemens@
14:45:50kugelthe e200v2 one is 176x54, the fuze's one 220x68
14:49:28kugelfunman: your mkamsboot binaries are somaller than mine too
14:51:17funmanupx packed ;)
14:51:47funmanand stripped
14:51:53kugelwhat is upx packed?
14:52:10funmanupx is a compressor for executables
14:53:06***Saving seen data "./dancer.seen"
14:53:38 Join Blue_Dude [0] (
14:54:45Blue_DudeSaratoga was right. The exp function wasn't broken as long as you keep the input within bounds. My test was bad.
14:54:58Blue_DudeSo the bloat is gone. FS #10411
15:00:04 Join Nico_P [0] (i=39428a0e@rockbox/developer/NicoP)
15:00:51mcuelenaereBlue_Dude: I'm willing to commit this, but I'm not sure about the changes you did; is this the original code or did you optimise it a bit?
15:02:31Blue_DudeIt's mostly converting back to 32 bit variables. It is modified from the original to allow for something other than 12 fractional bits (notice the constants are larger), but it's mostly the same. fp_factor is a single line now.
15:02:32 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
15:03:20 Quit perrikwp ("Leaving.")
15:03:24mcuelenaerewell I can't verify whether this breaks something, so I'll wait till someone who knows the code verifies it doesn't
15:03:27kugelBlue_Dude: you seem to have the svn executable property always set, what platform are you on?
15:03:44mcuelenaerekugel: I have that too as a default option if I don't change it
15:03:55mcuelenaere(when creating new files, that is)
15:04:11kugelhm, I see
15:04:15Blue_Dudekugel: that's something generated by svn diff. I didn't know if it's significant so I left it in. I'm on win XP running cygwin.
15:04:30 Join michael__ [0] (n=michael@
15:04:32Nico_Phas anyone considered giving the recently released "milepost GCC" a try?
15:05:10Blue_Dudemcuelenaere: how to change svn so I don't get svn:executable anymore?
15:05:11 Quit michael__ (Client Quit)
15:05:21mcuelenaereBlue_Dude: svn propdel svn:executable [FILE]
15:05:39Blue_Dudeok, I'll make them go away then.
15:05:41mcuelenaerebut don't worry, I'll remove those (if they even get set, I don't think 'patch' sets these parameters)
15:05:42kugelfunman: I'd also still would like reboot on usb insert :(
15:13:20funmankugel: a bit of cleanup of the usb code is needed for that
15:13:50CIA-71New commit by kugel (r21687): Sansa AMS: Disable voltage scaling for now until we found a way to make it reliable for everyone, it's causing problems with storage for many people.
15:13:59funmancurrently, bits of target usb code are needed for that (and that is incorrect)
15:14:40kugelit just needs a little hack :>
15:14:51funmanwell when things can be done properly it's better
15:14:58kugelyes sure
15:14:59funmani'll look at this today
15:15:22funmani don't think this is a big work
15:15:30funmangevaerts: do you have an idea ?
15:20:38gevaertsfunman: as far as I know the *only* bits of target usb code that are needed are the actual plugin detection bits. Is that really too much?
15:21:42funmanusb.c try_reboot() is #ifdef HAVE_USBSTACK , which implies a bit more code
15:22:12 Join Lynx_ [0] (
15:23:06Nico_PBagder: here?
15:23:24gevaertspossibly, but the rest can be just stubs
15:23:34funmanwhy needing stubs at all?
15:24:21funmanusb plug detection is not an "usb stack"
15:24:33GodEaterNico_P: Bagder is on holiday
15:25:02 Join kachna|lappy [0] (n=kachna@
15:25:04Nico_Poh, good for him :)
15:25:14Nico_PI wasn't sure whether he had left already
15:25:38gevaertsfunman: rebooting on plugin is meant as a temporary hack until a full usb driver is ready
15:25:40 Quit MethoS (Read error: 110 (Connection timed out))
15:29:48funmanok i'll look at it then
15:30:59 Quit MarcGuay_ (Remote closed the connection)
15:32:38GodEaterNico_P: what were you after from him ?
15:33:38GodEaterLambdaCalculus37: did you ever manage to get your 2G into DFU mode *without* scrapping the firmware partition ?
15:34:02LambdaCalculus37GodEater: No, I haven't tried. I can now, if you'd like.
15:34:24GodEaterit seems there are some Nanos which will do it, and some which won't
15:34:34Nico_PGodEater: just a few questions about the new build system
15:34:35GodEaterI tried with Mrs. GodEater's Nano a lot, and can't make it work
15:34:43GodEaterNico_P: are you running it already ?
15:35:00Nico_Pno, I have no server
15:35:14GodEaterLambdaCalculus37: however the trick to get the Nano 3G into DFU mode works like a charm
15:35:26GodEaterNico_P: just a laptop or something ?
15:35:46LambdaCalculus37GodEater: The trick is holding MENU+SELECT until the screen goes black, right?
15:36:01GodEaterNico_P: I run it on my laptop when I'm using it, on the off chance a build comes through ;)
15:36:08GodEaterLambdaCalculus37: correct
15:36:32LambdaCalculus37GodEater: I have a 3G nano here. Let me see that trick.
15:36:33Nico_PGodEater: well, I'll think about doing that with the new build system
15:37:03Nico_Pthe current one made it too much of a hassle and my comp isn't powerful enough to be worth it
15:37:25GodEaterthe new system makes it very easy to join and leave, and it's designed to take advantage of everyone
15:37:36GodEateras far as I understand it anyway
15:38:13LambdaCalculus37GodEater: I think I need to time it better, but every time the screen goes black on mine, it immediately resets again when I release the buttons.
15:38:22rasherNico_P: What questions do you have? Maybe some of us know
15:38:30GodEaterLambdaCalculus37: do it whilst it's disconnected ?
15:38:35GodEaterI think
15:38:39LambdaCalculus37It is disconnected.
15:38:49GodEaterLambdaCalculus37: curious, I got it to work first time
15:38:56Nico_Prasher: actually I found the answer on BuildServerRemake :)
15:39:21Nico_Pit was about the client ranking by speed
15:40:29CIA-71New commit by kkurbjun (r21688): DM320: Add the same fix from r21647
15:40:31LambdaCalculus37GodEater: I got into DFU mode on the 2nd gen nano without trashing the firmware.
15:40:43GodEaterLambdaCalculus37: that's SO annoying
15:40:45 Quit Lynx_ (Read error: 104 (Connection reset by peer))
15:40:46GodEaterI cannot do it :(
15:41:08GodEaterit's menu+select till reboot, then immediately |<< + Play right ?
15:41:27GodEaterhas no affect on mine at all - just reboots to the OF
15:41:44LambdaCalculus37Are you holding BACK+PLAY?
15:41:49Nico_Phas anything intersting come out of the recent developments at linux4nano?
15:41:50 Join Lynx_ [0] (
15:42:36GodEaterLambdaCalculus37: of course
15:42:49GodEaterNico_P: not *hugely* interesting
15:42:56GodEaterthey think they can dump the decrypted firmware
15:43:07GodEaterbut they need to work out how to write to the dock adaptor to do that
15:43:14GodEaterwhich so far they have epic failed at
15:43:25GodEaterthey also refuse to write anything down :(
15:43:38GodEaterwhich is probably more annoying
15:45:05Nico_Pif they do manage to dump the decrypted firmware, it means a port can officially start, doesn't it?
15:45:41GodEaterNico_P: well it would help immensely of course
15:45:42funmanthey also need to run a custom firmware
15:45:56GodEaterbut we'd still need to disassemble the OF to work out a lot of the hardware I'd guess
15:46:16Torneyah, the manuals available are not exactly complete/final
15:46:26Torneeven for the soc itself
15:46:30GodEaterI don't think it'd be quite as bad as the original ipods
15:46:37GodEatersince there was virtually no docs at all for that
15:46:42GodEaterbut still pretty hard work
15:47:16 Join Blue_Dude_ [0] (
15:47:44LambdaCalculus37GodEater: The SOC is very similar to that in the Meizus, so we have a basis to work on, at least.
15:48:06GodEateryes, but as Torne says, it's still not perfect
15:48:46Torne(and christ that's an awful manual)
15:49:55 Quit kugel (Read error: 110 (Connection timed out))
15:52:32 Join kugel [0] (n=kugel@rockbox/developer/kugel)
15:52:57 Quit Blue_Dude (Read error: 60 (Operation timed out))
15:53:12 Nick Blue_Dude_ is now known as Blue_Dude (
15:54:54 Join perrikwp [0] (
15:54:54CIA-71New commit by funman (r21689): Samsa AMS: start of an USB driver (nothing working atm) ...
15:56:05*LambdaCalculus37 gives funman a beer just for starting the USB driver :)
15:56:09funmanwell, there is really nothing to applause
15:56:24funmanI just had looked at getting some interaction with the USB registers but failed
15:56:34funmankugel: oops
15:56:43funmansAMSa just looks a nice shortcut for Sansa AMS ;)
15:57:07kugelah, yea that's clever :)
15:57:27 Quit n1s ("Lämnar")
15:57:46kugelfunman: do we have reboot now?
15:58:00funmanlike the commit log says, no
15:58:28LambdaCalculus37kugel: Instructions for OS X can be similar to those for Linux (from your forum post).
15:58:46kugelLambdaCalculus37: I figured, but we have no binaries for that yet :)
15:59:13LambdaCalculus37kugel: What do you need a binary for?
15:59:14kugelfunman: oh, I didn't read the full log
15:59:23kugelLambdaCalculus37: OS X :P
15:59:42LambdaCalculus37kugel: I can do that now if you'd like. :)
15:59:49kugelwould be cool
16:00:21kugelLambdaCalculus37: build it from this branch please:
16:01:21 Quit GodEater ("Terminated with extreme prejudice - dircproxy 1.0.5")
16:01:36 Join GodEater [0] (n=bibble@rockbox/staff/GodEater)
16:02:12*LambdaCalculus37 checks out that branch
16:03:33 Join aaron424 [0] (n=chatzill@
16:05:51*kugel wonders if he should post in the official sandisk forum also
16:05:56 Join mc2739 [0] (
16:06:01funmani wouldn't :)
16:06:21kugelmany admins and mods posted there without showing any hate against rockbox
16:06:46funmani mean i think this would cause extra work for moderating the rockbox thread
16:06:55LambdaCalculus37kugel: Probably best not to whack the beehive, though.
16:07:39mc2739funman, kugel - do either of you see the virtual led on your fuze during disk access?
16:08:01funmanwhere is the code for the virtual led?
16:08:18kugelmc2739: indeed, I already wondered why that's not there
16:08:59kugelwe define VIRTUAL_LED though
16:09:08 Quit Blue_Dude ("ChatZilla 0.9.85 [Firefox 3.0.11/2009060215]")
16:10:59ReKleSSkind of obscure question... where does the coldfire CRT run?
16:11:09ReKleSSfrom the datasheet it looks like the ram starts out unmapped
16:11:18mc2739funman: I think the disk activity timer is in ata_sd_as3525.c
16:12:55funmanisn't this led meant to appear in the wps / status bar ?
16:13:13kugelfunman: Any idea why it doesn't reboot? I see a #if 0'd usb acknowledgement in the sd driver
16:13:25funmanlooking at it
16:14:20mc2739funman: yes, it should be to the right of the time on the status bar
16:14:42funmanmc2739: perhaps the icon is not here in the theme
16:14:55 Quit robin0800 (Read error: 110 (Connection timed out))
16:15:20mc2739the status bar is separate from themes
16:15:28 Quit GodEater ("Terminated with extreme prejudice - dircproxy 1.0.5")
16:15:39 Join GodEater [0] (n=bibble@rockbox/staff/GodEater)
16:16:05mc2739besides, I use the same theme on my v1 and v2
16:16:11funmankugel: that's it, the code needs to be enabled and include usb.h
16:16:44funmanmc2739: ok then there is a bug (I only had seen this icon on a samsung YH920, I thought it was ATA-specific)
16:17:49funmanmc2739: what if you cjhange MIN_YIELD_PERIOD to 1 ?
16:20:06mc2739funman: I'll test later - leaving for work now
16:20:54CIA-71New commit by funman (r21690): Samsa SD driver : acknowledge USB events, now reboots on USB insertion
16:21:22funmankugel: your forum post has too much [/list] :)
16:22:13 Quit mc2739 ("ChatZilla 0.9.85 [Firefox 3.0.11/2009060215]")
16:23:00 Join notlistening [0] (
16:24:26CIA-71New commit by funman (r21691): Samsa: fix red for clip/m200v4/c200v2 (typo)
16:26:42kugelfunman: \o/
16:27:50kugelhmm, it doesn't work for me
16:30:30 Join GodEater_ [0] (
16:34:32 Quit GodEater (Client Quit)
16:34:43 Join Grahack [0] (
16:34:50 Join robin0800 [0] (n=robin080@
16:46:22CIA-71New commit by kugel (r21692): sAMSa: Turn the backlight off before rebooting to avoid irritating lcd flash.
16:48:51 Quit notlistening (Remote closed the connection)
16:49:18 Join jgarvey [0] (
16:50:58 Quit LambdaCalculus37 ("Fwump")
16:53:10***Saving seen data "./dancer.seen"
16:57:31 Join notlistening [0] (
16:58:37notlisteningkugel, did i ned to make any other changes for booting from the SD card like a new bootloader?
16:58:56kugelof course
16:59:50notlisteningok ;)
17:00:24 Quit kugel (Nick collision from services.)
17:00:28 Join kugel [0] (
17:00:40 Quit kugel (Nick collision from services.)
17:00:48 Join kugel [0] (
17:00:56 Quit kugel (Nick collision from services.)
17:01:02 Join kugel [0] (
17:01:15 Join kugel_ [0] (
17:01:27 Quit kugel (Nick collision from services.)
17:01:31 Nick kugel_ is now known as kugel (
17:01:50 Quit kugel (Nick collision from services.)
17:02:03 Join kugel [0] (n=kugel@rockbox/developer/kugel)
17:06:33 Join toffe82 [0] (n=chatzill@
17:11:41amiconnReKleSS: crt0 runs from IRAM. SDRAM does not only start out unmapped, but unconfigured.
17:12:06ReKleSSok, thanks
17:21:13 Quit kugel (Read error: 110 (Connection timed out))
17:22:58 Join kugel [0] (n=kugel@rockbox/developer/kugel)
17:34:35 Quit ReKleSS ("Leaving")
17:38:13 Quit flux (Remote closed the connection)
17:44:19 Join evilnick [0] (i=0c140464@gateway/web/freenode/x-4c662f274b13591f)
17:44:52mtlinuxstb: change in bin size could be decreased (roughly from 3072 to 2048). But then we wouldn't be able to display debugging info in the sim except for those in RMContext. But I've seen vorbis.c and flac.c don't display as much debugging info so I think this is okay ?
17:45:56funmanyou can make debugging info display conditional
17:46:20mtfunaman: Yes but then it will look really ugly. :)
17:47:27 Quit Dhraakellian (Read error: 110 (Connection timed out))
17:47:35funmanwe don't need debugging info on target
17:49:44 Join mc2739 [0] (
17:50:01mtThe problem isn't with displaying the debugging info, it's with reading them. Cutting down the number of read()'s that has to be done makes it possible to reduce bin size (and some ram usage too), but will also hide some header parameters. (which aren't used anyway, at least currently)
17:51:02mc2739funman: changing MIN_YIELD_PERIOD to 1 did not help with virtual led
17:51:42funmani have no idea how to use the led now
17:56:43funmantry hacking the statusbar code i think
17:59:31 Join n00b81 [0] (n=taylor@unaffiliated/n00b81)
18:00:15mtI know I should minimize the bin size change as much as I could, but how much is small enough ? I mean, 2048 is < 0.34% increase ..?
18:00:36funmanit's never small enough ;)
18:02:16Mikachuthere are several changes every day, if every is 0.34% it quickly grows :)
18:02:25 Join JdGordon| [0] (i=ae90a306@gateway/web/freenode/x-3878c42a3510838f)
18:04:20mtyes, but daily changes are not always as big as the addition of a new plugin or a new codec. :)
18:05:01 Quit petur ("work->home")
18:05:10mtAnyway I'll look into how I could further reduce it ..
18:05:31 Quit thegeek ("( :: NoNameScript 4.2 :: )")
18:06:55Tornemt: you could compare it to the metadata implementation for other codecs
18:07:05Torneif you compare favourably to them that might be considered reasonable
18:07:09Tornethough obviously smaller is always better
18:08:02funmanthat's what she said
18:08:22 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
18:11:08 Join Asa_Thor [0] (
18:11:32 Quit at0m ("it is now safe to turn of your machine.")
18:11:54 Quit Nico_P ()
18:16:20mc2739funman: virtual led fix -
18:17:05mc2739duplicated code from ata-sd-pp.c
18:17:13linuxstbmt: You could replace those individual reads with a single read, and then just extract the data you need from what you've read.
18:17:36funmanmc2739: if it's duplicated you could put it into the new firmware/driver/sd.c so every sd driver use it
18:17:52funmanPP / Samsa / Ingenic
18:19:01bertrikI prefer AMS Sansa or Sansa AMS over Samsa
18:19:10funmansAMSa ?
18:19:12mc2739funman: I'll try that now
18:21:07 Quit flydutch ("/* empty */")
18:27:09 Quit robin0800 ("Leaving")
18:27:25mtlinuxstb: I was thinking of replacing the unneeded reads with a single lseek
18:29:42mtDo you mean for keeping the debugging info for the sim ?
18:30:54Asa_ThorHello. I'm looking for a 32gb compact flash card for my iPod. Any recommendations?
18:31:24 Join robin0800 [0] (
18:31:33 Quit JdGordon| ("Page closed")
18:31:38linuxstbmt: Yes, but I was also thinking that before you mentioned that problem. i.e. instead of a sequence of 10 reads of 4 bytes, do a single read of 40 bytes and then extract anything you need from memory. But it probably doesn't make much difference...
18:31:55 Join mc2739_ [0] (
18:32:17 Quit mc2739 (Nick collision from services.)
18:32:19 Nick mc2739_ is now known as mc2739 (
18:40:52 Quit robin0800 ("Leaving")
18:41:11 Join robin0800 [0] (
18:41:21funmanmc2739: we could inline led() and not use a custom sd_led()
18:42:56funmanthe led icon doesn't look like a led on the fuze ;)
18:43:04 Join lee321987 [0] (
18:43:48 Join JdGordon| [0] (n=Miranda@nat/microsoft/x-0684b15fca8cf803)
18:44:09 Part Asa_Thor
18:45:01CIA-71New commit by funman (r21693): Sansa AMS: display the virtual led icon on disk transfers ...
18:45:13lee321987On Ubuntu "sudo ./e200rpatcher.linux32" gives me, "sudo: /e200rpatcher.linux32: command not found"
18:45:17 Quit stoffel (Read error: 60 (Operation timed out))
18:45:51 Quit HBK (Read error: 54 (Connection reset by peer))
18:45:57funmanlee321987: run "file e200rpatcher.linux32"
18:46:35funmandid you run chmod +x e200rpatcher.linux32 ?
18:46:42 Quit funman ("free(random());")
18:46:54 Join HBK [0] (
18:46:55lee321987e200rpatcher.linux32: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), for GNU/Linux 2.6.9, dynamically linked (uses shared libs), stripped
18:47:59lee321987I have to build this myself, if I'm on an AMD, right?
18:48:37 Join thegeek [0] (
18:48:58lee321987does bootloader.bin need to be in the same folder?
18:51:35lee321987 −−−−- what are the files in the folder "0.1"?
18:52:03 Quit mc2739 ("ChatZilla 0.9.85 [Firefox 3.0.11/2009060215]")
18:52:42mtlinuxstb: This is rm.c with sequences of read()'s replaced with 1 lseek
18:53:14***Saving seen data "./dancer.seen"
18:53:16mtthere are still unused variables that have to be cleaned.
18:54:12 Join perrikwp1 [0] (
18:55:15mtBefore that change, increase in bin size is 3072, now it's 2048. Given that asf.c has an increase of 4096 on bin size, Is this change needed ?
18:56:35 Join sajes [0] (n=sajes@
18:57:05domonokymt: saving binsize in the core is always good :-)
18:57:24sajesAt this point of the AMS sansa development, would it still be considered a stupid idea to try and install rockbox an e200v2? It didn't say which device was having the corruption issues.
18:57:39sajesIt as in development status.
18:58:09domonokysajes: no, take a look at the testing section in the forums :-)
18:59:01domonokyso its not a stupid idea to try it, but its ofcourse at your own risk :-)
18:59:41 Quit robin0800 ("Leaving")
19:00:00 Join robin0800 [0] (
19:00:10sajesdomonoky: You mean the rockbox development section?
19:00:12mtdomonoky: Yes I know, I've been getting that a lot :). It's just that without this change it still has less impact than parsers like mp4 and asf on bin size, and we still get full debugging info in the sim. That's why I'm asking.
19:00:48Mikachuthe goal isn't to make every change as large as the biggest one
19:00:51Torneifdef it to be only for the sim
19:01:01Torneit's not messy to do so, really, it's fairly standard :)
19:01:53bertrikgevaerts, has someone ever thought about what a 'real' bootloader for the meizus would look like?
19:01:55mtokay, will give it a shot.
19:02:00 Quit perrikwp (Read error: 110 (Connection timed out))
19:02:19domonokymt: that only means that the parsers for mp4 and asf probably could be improved :-) using a ifdef to enable more debugging info is a good idea.
19:02:46gevaertsbertrik: I don't think so. Lots of people probably want dual boot, but I don't think anyone has ever looked into that either
19:03:09bertrikthe OF bootloader just seems to extract a RAR image into sdram, the code for that seem quite small (< 0x2000 bytes)
19:03:34bertrikwe could imitate that (and choose between the OF image or the RB bootloader)
19:03:45bertrikdepending on a keypress or something similar
19:04:45gevaertsah yes. They use that weird rar format
19:04:46bertrikwe could use something completely different (and possibly faster) than RAR of course, maybe UCL like we do for the ams sansas
19:05:05 Join AndyIL [0] (i=AndyI@
19:05:07gevaertsI'm not sure if the m6sl was the same, but we can check that
19:06:31 Quit robin0800 ("Leaving")
19:06:48 Join Dhraakellian [0] (
19:06:49 Join robin0800 [0] (
19:07:51 Quit bmbl ("Woah!")
19:09:19bertrikthe meizu has 1MB NOR flash, the compressed OF takes 600 kB, so a RB bootloader should fit easily besides it I think
19:11:47 Part sajes
19:12:02amiconnmt: As long as your debug output is pure output, just use DEBUGF(). This collapses to nothing on target for non-debug builds
19:12:53 Nick GodEater_ is now known as GodEater (n=bibble@rockbox/staff/GodEater)
19:13:17amiconnbertrik: So you could probably decompress the OF and have an of.meizu which could be roloed
19:13:49GodEatergevaerts: I based that forum post off something from Bagder's blog. So it's not all my fault!
19:13:53bertrikamiconn, yes we can decompress it, not sure was rolo-ing involves
19:14:22amiconnLoad it from disk, set up the hardware like the OF loader does, then branch to it
19:14:30gevaertsGodEater: Good idea! We can always blame Bagder!
19:15:37bertrikamiconn, so the bootloader would in that case just boot rockbox, right?
19:16:11 Quit AndyI (Read error: 110 (Connection timed out))
19:16:44amiconnThe bootloader could also offer the choice, loading either ockbox or the OF from disk (or flash, if possible)
19:17:26 Join Dauron [0] (
19:17:45mtamiconn : It's not pure output, that's the problem. :)
19:18:26 Quit robin0800 ("Leaving")
19:18:41amiconnOn archos we have dual boot from flash rom, although OF and rockbox don't fit anymore together, so we have bootbox (basically a cut-down rockbox which loads the default on-disk firmware) and rockbox in flash
19:19:28amiconnBoth firmwares can be ucl compressed, in which case they're first decompressed into ram and then executed, or uncompressed, in which case they're executed directly
19:20:49bertrikok, thanks for the info. There's quite a few possible ways of booting.
19:20:56amiconnBootbox is always compressed to save space. Uncompressed rockbox does only fit on Player and Ondio SP, unfortunately, or if you've upgraded the flash to 512KB
19:21:33bertrikA nice intermediate step would be to have a dualboot loader that can boot either the OF or the RB bootloader from flash (we can't access NAND yet)
19:22:09amiconnRunning from ROM obviously leaves more free RAM, but is a little slower (ROM data bus is 8 bit but RAM data bus is 16 bit)
19:23:12bertrikI hope the NAND filesystem is simple enough that we can do reads at least
19:23:24amiconnNah, it's not that bad. RAM in turn needs two extra states unless the access happens within the current page (fast page mode)
19:23:25 Quit Lynx_ (" Try HydraIRC -> <-")
19:24:42 Join perrikwp [0] (
19:24:46amiconnLater targets often have 16 bit ROM data bus, i.e. same bus width as the SDRAM
19:28:25linuxstbmt: Shouldn't you check the return value from rm_parse_header?
19:29:23linuxstbmt: Plus also check that the codec type is COOK, and return "false" from get_rm_metadata() if not?
19:30:53 Join Thundercloud [0] (
19:31:06mcuelenaereBlue_Dude (logs, if you read them): don't worry about the bloat; I just wasn't sure about those changes and as I'm not familiar with that part of Rockbox I wasn't sure if no behaviour would get changed when it gets committed. If that's the case, I'll commit your patch.
19:33:02 Quit lee321987 ("ChatZilla 0.9.85 [Firefox 3.0.11/2009060215]")
19:33:15 Quit perrikwp1 (Read error: 110 (Connection timed out))
19:33:18 Quit _Auron_ (Read error: 110 (Connection timed out))
19:33:18 Nick Dauron is now known as _Auron_ (
19:38:05 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
19:44:01 Quit kugel (Remote closed the connection)
19:48:06*linuxstb agrees with bertrik about "samsa"...
19:49:32 Join kugel [0] (n=kugel@rockbox/developer/kugel)
19:57:17 Join flydutch [0] (
19:57:23 Join at0m [0] (
19:59:18 Quit aaron424 (Remote closed the connection)
20:00:21 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
20:02:02kugeltypos are independent of any agreement :)
20:04:09 Quit HBK (Read error: 104 (Connection reset by peer))
20:04:28kugelsamsa is perfet
20:04:41 Join robin0800 [0] (
20:04:46JdGordon|samsa soudns retarted :p
20:04:54*JdGordon| curses his stupid fingers
20:05:09GodEaterJdGordon|: I thought you did it on purpose to be funny
20:05:24JdGordon|i mean... yes of course :)
20:05:25Dhraakelliansamsa sammich?
20:06:11 Join petur [50] (n=petur@rockbox/developer/petur)
20:06:40kugelJdGordon: just like sansa does :p
20:06:44linuxstbkugel: It's a typo?
20:07:04kugelsee r21689 and irc logs around that time
20:07:07linuxstbkugel: You made the same typo...
20:07:28kugelwell, it was initially a typo
20:07:32 Join HBK [0] (
20:08:12Mikachu"samsa" means "get along" in swedish, fwiw
20:08:31*JdGordon| calls Mikachu a lier!
20:08:52JdGordon|samsa isnt a derivative of "bork"
20:09:06JdGordon|ergo its not swedish!
20:11:42Dhraakellianwhen using the new official test builds of rockbox, should I format from the OF first just to be thorough if I've used previous revisions
20:11:59kugelwouldn't be the worst thing
20:12:15kugelbut we actually want to get updating to work without reformatting :S
20:13:01Dhraakellianthen again, it's not like I was keeping music on the internal storage
20:14:08 Quit perrikwp (Read error: 110 (Connection timed out))
20:17:28*Bagder welcomes committer #76, teru
20:18:45*kugel doesn't how Bagder counts. Are so many duplicates in docs/COMMITTERS?
20:18:55kugelalso: \o/
20:18:59*Bagder has so many secrets
20:19:17rasherkugel: There are a bunch of duplicates I believe
20:19:35Bagdermost importantly, the number 76 is actual svn accounts
20:19:42Bagdernot all were moved from cvs
20:19:49kugelahh, that makes more sense
20:20:05Bagderso we have >76 total committers, but only 76 can commit atm
20:20:19Bagder"only" ;-)
20:30:48 Quit robin0800 (Remote closed the connection)
20:31:35 Join tessarakt [0] (
20:32:22*kugel waits for his first commit...
20:38:48 Join robin0800 [0] (
20:39:56 Join Rondom [0] (
20:45:06 Join t0pGuN [0] (i=266c416d@gateway/web/freenode/x-afd167b6f6928724)
20:45:09t0pGuNi just downloaded the .zip file to my desktop, unzipped it, then replated my old .rockbox folder to .OLDrockbox and copied over the new .rockbox folder but now my gigabeat is not recognized by the computer :(
20:45:10 Quit Dhraakellian (Read error: 104 (Connection reset by peer))
20:46:21 Quit notlistening (Remote closed the connection)
20:47:33 Join saratoga [0] (n=9803c6dd@rockbox/developer/saratoga)
20:47:49t0pGuNsomeone pleeeese :D
20:48:00pixelmabe patient
20:48:26saratogai don't think you even need a .rockbox folder to get USB on the gigabeat
20:48:31saratogaits got a bootloader USB mode
20:48:42 Join Dhraakellian [0] (
20:49:01 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
20:49:32pixelmaalso, could you give a bit more info, e.g. which Gigabeat is this, which OS are you running, what was the build before that (from when)
20:50:12lilltigermaybe the rockbox version for gigabeat dosent do usb interaction like the sansa ams, so he needs to turn it of or boot into the OF to enable usb transfears
20:50:19t0pGuNits F10
20:50:25t0pGuNWindows XP
20:50:31t0pGuNit was working 3 mins ago
20:50:43t0pGuNthen i did what i said, and it stopped working
20:51:28t0pGuNthe build before, i dont remeber, but must be at least a year old
20:53:08 Quit Dhraakellian (Read error: 60 (Operation timed out))
20:53:16***Saving seen data "./dancer.seen"
20:55:53 Quit robin0800 ("Leaving")
20:56:13 Join robin0800 [0] (
20:56:59 Quit n00b81 ("Leaving")
20:57:30pixelmacan you see the .rockbox folder in Rockbox's file browser itself? (You might have to change the file view settings to "all") I don't know much about Gigabeats and just try to gather as much info so that someone else would be able to help
20:58:56t0pGuNnice i see both
20:59:01 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
20:59:02t0pGuN.rockbox and OLD.rockbox
20:59:50 Join aaron424 [0] (
21:00:23 Quit flydutch ("/* empty */")
21:01:16 Join ej0rge [0] (
21:01:54Unhelpfulsaratoga: you really really want to install rockbox proper and use its usb mode, though. bootloader usb doesn't charge.
21:02:34pixelmaUnhelpful: does the F use Rockbox USB?
21:03:40gevaertspixelma: no. Hardware
21:03:55aaron424I formatted my sansa e280 in rockbox usb mode and it deleted the OF. What can I do to get it back?
21:04:14t0pGuNmmm can i just rename .rockbox to .temprockbox and rename the OLD.rockbox to .rockbox ?
21:04:22kugelaaron424: "deleted te OF"?
21:04:25gevaertst0pGuN: so, rockbox works, just not usb?
21:04:45kugelanyway, can you access the recovery mode?
21:05:11aaron424should I put the mi4 there?
21:06:15gevaertst0pGuN: are you using a dock, or only a cable?
21:06:16Unhelpfuli had thought we were speaking of the beast... does the F/X also have bootloader USB?
21:06:21t0pGuNjust a cable
21:07:03kugelaaron424: put the Sansa original firmware there, then reinstall rockbox via rbutil
21:07:42gevaertswhat happens when you plug it in? Does it show the USB logo?
21:08:19aaron424kugel: thanks
21:08:44kugelaaron424: You're welcome
21:09:04aaron424I needed to upgrade to 3.3 anyway
21:09:53t0pGuNbut then it says USB Device Not Recognized
21:10:02t0pGuNcan i just rename .rockbox to .temprockbox and rename the OLD.rockbox to .rockbox ?
21:10:30t0pGuNor will i break everything by renaming .rockbox
21:10:51gevaertsYou can do that. Make sure you end up with a .rockbox though
21:10:53aaron424kugel: what patches are needed to use most custom themes?
21:11:01kugelt0pGuN: you could've also just overwrite the old rockbox
21:11:23kugelaaron424: multifont and customlist
21:11:27t0pGuNbut wont the device stop working at all if i rename the .rockbox?
21:11:47gevaertst0pGuN: as long as .rockbox is there when booting, it will work
21:11:54t0pGuNah ok
21:12:14gevaertsPlugins and music playback won't work if .rockbox is not there, but the file browser will
21:12:18 Join Dhraakellian [0] (
21:12:22t0pGuNok great
21:12:24t0pGuNlemme try that
21:12:54aaron424kugel: thanks again, ill get those
21:12:55 Join BryanJacobs1 [0] (
21:13:04pixelmagevaerts: hmm? rockbox.gigabeat (or so) is in .rockbox, isn't it needed?
21:13:37gevaertspixelma: it needs to be loaded, but as soon as it's loaded to RAM, you don't need the actual file on disk
21:15:07t0pGuNSWEEET :)
21:15:10t0pGuNit works again
21:15:33t0pGuNor else 10 gigs of good sorted music would have been lost :)
21:15:33pixelmawould be nice to know why it didn't
21:16:04t0pGuNi prolly did something wrong
21:16:40t0pGuNthis is what i did -> i downloaded the .zip file to my desktop (for manual installation), unzipped it, then replaced my old .rockbox folder to .OLDrockbox and copied over the new .rockbox folder
21:17:59kugelthat should work actually
21:20:02toffe82execpt if the bootloader is too old
21:20:55kugelthere wasn't a new bootloader in ages I think
21:21:17pixelmatoffe82: what does the bootloader have to do with it?
21:21:39t0pGuNoh hehe
21:21:43t0pGuNi dunno
21:21:47t0pGuNbut im happy its back to normal :)
21:22:26toffe82pixelma: if you put a new version of rockbox and an old bootloader, it doesn't work , no ?
21:22:51gevaertst0pGuN: when did you first install rockbox?
21:23:20pixelmahmm... the only thing I could imagine is if it searches rockbox.gigabeat in the root and not the .rockbox folder (but that change is longer ago, and I would have thought that the problem then would be broken playback etc.)
21:23:20t0pGuNive had rockbox running for a few+ years now
21:23:30t0pGuNbut the last install was about a year ago maybe a little more
21:23:41gevaertsdid you update the bootloader then?
21:23:46pixelmatoffe82: depends on how old and what changed...
21:24:13t0pGuNlast time i installed everything, i did a simple update from what i remember
21:24:30gevaertspixelma: we had someone here yesterday with a nano and backlight problems that were solved by a new bootloader
21:24:45t0pGuNi remember the reason why i updated was cuz when you switched songs in the playlists, it would lag between switching
21:24:49t0pGuNthere was a delay
21:24:55t0pGuNand after the update the delay was gone
21:26:05gevaertst0pGuN: can you check the exact size of the GBSYSTEM/FWIMG/FWIMG01.DAT file?
21:27:09t0pGuNone sec
21:28:25t0pGuNhrmm the system is lagging, i cant access folders
21:28:28t0pGuNbrb reboot
21:28:32 Quit t0pGuN ("Page closed")
21:32:32 Quit rob (Read error: 110 (Connection timed out))
21:35:38LambdaCalculus37kugel: Building mkamsboot just involves running make from rbutil/mkamsboot, correct?
21:36:10LambdaCalculus37kugel: Let me make a disk image of mkamsboot and I'll send it to you.
21:37:27 Join t0pGuN [0] (i=266c416d@gateway/web/freenode/x-749dd66f45bb14bb)
21:37:35t0pGuNwhat file was i supposed to look up?
21:38:35LambdaCalculus37kugel: PM me your email address.
21:38:47kugelLambdaCalculus37: just mail it to bagder
21:39:03 Join funman [0] (n=fun@rockbox/developer/funman)
21:39:19 Join perrikwp [0] (
21:39:34funmanLambdaCalculus37: can you make an universal binary of mkamsboot?
21:40:08t0pGuNSize: 49.5KB (50,784 bytes) and Size on disk: 56KB (57,344 bytes)
21:40:27t0pGuNi can just tell you the version of the rockbox if you want hehe
21:40:27LambdaCalculus37funman: They're compiled automatically as universal binaries.
21:40:29gevaertst0pGuN: The current bootloader is 53304 bytes, the old one was 50784 bytes. I'd suggest upgrading the bootloader
21:40:44t0pGuNah ok
21:40:48t0pGuNill do that
21:40:49LambdaCalculus37funman: I already found that out when I still had my PPC-based PowerBook.
21:40:53t0pGuNim backing up the hd
21:41:27LambdaCalculus37kugel: And it's off to Bagder's mailbox.
21:45:03kugelLambdaCalculus37: very nice
21:46:19kugelI like the way we're installing rockbox. Our bootloader can be damaged as hell without braking dualboot. And due to patching the OF instead of hacking onto the device our installer is nicely plattform independent :)
21:48:13 Quit BryanJacobs1 ("Java user signed off")
21:49:51t0pGuNare you guys planning on expanding rockbox on the iphone and stuff and making it look super fancy like iphone's OS ? :)
21:49:52 Join Hillshum [0] (n=chatzill@unaffiliated/hillshum)
21:50:20kugelNo, we aren't planning (generally spoken)
21:50:42t0pGuNok i was just curious
21:53:51kugelfunman: someone installed Rockbox on his fuze. It was his first Rockbox installation ever. It seems the installation isn't actually too complicated
21:54:17 Join stoffel [0] (
21:54:52funmannice ;)
21:55:35LloreanWhat's with all the "samsa" in commit messages?
21:55:55Unhelpful"look super fancy" is not really a goal of the project. oddly enough, "play music well" comes first.
21:56:46funmanLlorean: check the irc logs
21:57:13HillshumLlorean: sAMSa
21:57:16pixelmaLlorean: first a typo and then it became short for "AMS Sansa"
21:57:38kugelNice, isn't it? :)
21:57:38gevaertsLlorean: don't ask...
21:57:43*Llorean thinks it's pretty horrible.
21:58:01*gevaerts thinks that commit messages should be clear to people who don't know the current irc fads
21:58:09Lloreangevaerts: Exactly what I was typing.
21:58:36LloreanThe three or four extra characters for SansaAMS or Sansa AMS isn't going to kill anyone once per commit.
21:58:38ej0rgeor a man who awoke from unsettling dreams to find himself changed into a monsterous vermin
21:58:46Lloreanej0rge: That's what I thought every time I see it.
21:59:00*Unhelpful thinks "beast" is wonderful, and we should be using "beaft" and "beaxt" for the other gigabeats. and "beavt" if it's ever supported. :)
21:59:33GodEaterhas "beast" actually ever been used in a commit message though ?
21:59:38 Quit desowin ("KVIrc Insomnia 4.0.0, revision: 3303, sources date: 20090703, built on: 2009/07/05 07:40:45 UTC 3303")
21:59:42GodEateroh dear
21:59:43Hillshumand plain old beat for the T?
21:59:54funmanGodEater: 18 times
21:59:55kugelno, butt
22:00:24*GodEater shares llorean's distaste for using silly names for commit messages that only IRC users will understand
22:00:47*gevaerts thinks that only one of the "beast" commits was reasonable, since it says "beast's (Gigabeat S) manual"
22:00:55funmanthe list of changed files makes it pretty clear that it is for as3525 SoC
22:01:12*Unhelpful also found "beastie" in a commit message, not in reference to gigabeat s
22:01:12kugelI think I can agree for commits
22:01:40LambdaCalculus37Using beast for commits? I have no problems with that. :)
22:02:18*Llorean thinks all commit messages should be held to approximately the same standards as manual text - be technical but be precise, no "conversational" language.
22:02:28funmanUnhelpful: perhaps "Ladies & Gentlemen : first song played from the beastie boys" ?
22:03:10LambdaCalculus37funman: No one's used the Beastie Boys for first song played on any target, AFAIK.
22:03:39*funman is looking towards playing music in the Clipv2 bootloader
22:04:37Hillshumfunman: in the bootloader?
22:04:49 Part Grahack
22:04:49Unhelpfulfunman: not so much, an SPC codec commit message
22:04:50funmanHillshum: Clipv2 has no storage support yet
22:04:56LambdaCalculus37Hillshum: bertrik managed to play the first music on the Meizu M3 that way.
22:05:07 Quit perrikwp (Read error: 110 (Connection timed out))
22:06:48 Join raphi [0] (
22:07:40t0pGuNcya all thanks
22:07:42 Part t0pGuN
22:09:26bertrikmarkun, have you (or anyone else) ever worked on a link script for a non-bootloader meizu build?
22:11:38 Quit LambdaCalculus37 ("Fwump")
22:11:54HillshumWill mkamsboot work on osx?
22:12:28kugelself compiled, yes it should. And prebuild binaries are coming soon
22:12:34*JdGordon| very muchly disagrees with Llorean from 10min ago
22:13:07LloreanJdGordon|: That we should strive for clarity in commit messages?
22:13:33LloreanThey're very much a log, used by people trying to look up information about old changes. Possibly people who weren't even part of the project then.
22:14:02funmanI use logs very often and I think they must be as descriptive as possible
22:14:17rasherSee, this is the same reason I dislike "fix red", "fix yellow" and "oops"
22:14:20funmanand never "fix red"
22:14:32 Join efyx__ [0] (
22:14:32rasherThey're useless if you see them more than 15 minutes after the fact
22:14:33JdGordon|considering how quickly everything changes, the old logs are really no more than a very quick comment on what might have changed.. which is very likely to bnot be relevant anymroe anyway
22:14:50funmanJdGordon|: not when you are looking at all revisions of a file for example
22:15:02JdGordon|rasher: then look at the actual change... if the comment isnt clear the 2 line change will be
22:15:02kugelalso, it can't work with JdTypo anyway
22:15:50rasherJdGordon|: Is it so hard to say "move definition of variable foo into the right scope" rather than "fix red"?
22:16:00LloreanCommits like "S5L8700: implement kernel timer " are useful information though
22:16:06JdGordon|no, buts its totally pointless
22:16:15gevaertsJdGordon|: it isn't
22:16:17LloreanI mean, no matter when later, if someone's curious when it was added that commit makes it clear.
22:16:31pixelmaif I look for a recently introduced bug, I read the commit look to limit the possibilities. A good commit message helps a lot here
22:17:02 Join perrikwp [0] (
22:17:14JdGordon|sure, but "fix red/yellow" will be clear from the change what it actually means
22:17:30pixelmait can also help other people to understand why you made a certain change
22:17:40rasherJdGordon|: It *might* be clear. And it requires extra work when looking back through the log
22:17:51rasherJdGordon|: It's just laziness. There's really no excuse
22:17:51gevaertsJdGordon|: fix colour is a bit less strong, but I still like to see some detail
22:17:58 Quit efyx_ (Read error: 110 (Connection timed out))
22:18:15*JdGordon| is going to write a 50000 word sermon next time he does a woops commit
22:18:45rasherJdGordon|: How very reasonable of you
22:18:45JdGordon|with html to break the web svn reader :)
22:19:21gevaertsJdGordon|: if it's relevant to the commit, please do
22:20:12funmanyeah long commit logs don't harm
22:22:30 Quit barrywardell (Remote closed the connection)
22:24:51 Join fyrestorm [0] (
22:25:00 Join webguest63 [0] (n=d8870b32@gateway/web/cgi-irc/
22:27:06 Quit webguest63 (Client Quit)
22:27:36MikachuJdGordon|: if it was clear what was wrong, why did you make the mistake in the first place? :)
22:28:13*JdGordon| throws a LED pig a Mikachu
22:28:30LloreanMikachu: You say the "JdTypo" comment earlier, didn't you? :-P
22:28:33HillshumJdGordon|: *at?
22:29:02JdGordon|blame the seattle "summer".... im fucking freezing here...
22:29:19JdGordon|.... this is an on topic channel... STFU
22:29:22 Quit jgarvey ("Leaving")
22:30:07 Quit perrikwp (Read error: 60 (Operation timed out))
22:30:56bertrikhmm, some of the register names in the drivers from the s3c2440 (gigabeat f) seem be a bit similar to ones for the meizus
22:31:23bertrikperhaps I could have a saved some time with that knowledge ... :|
22:33:29 Quit aaron424 ("ChatZilla 0.9.85 [Firefox 3.0.11/2009060308]")
22:35:46 Join perrikwp [0] (
22:37:13 Quit Thundercloud (Remote closed the connection)
22:47:49mtThis what I've been able to do so far to decrease bin size and ram usage :
22:48:52mtlinuxstb ^^^
22:49:26Mikachumt: does initializing skipped to 4 and removing the += 4; in the beginning change anything? :)
22:49:32Mikachu(probably not)
22:49:53linuxstbmt: You can just do "if (rm_parse_header(fd, &rmctx, id3) < 0) return false;" - there is no need for an extra variable.
22:50:11linuxstb(although it probably doesn't make any difference to the generated code)
22:50:25linuxstbmt: And do you check for codecs other than COOK?
22:50:38Mikachui don't think you need 3 unknown variables, just passing the 32 one to everything should work the same
22:51:32 Quit raphi ("leaving...")
22:51:38mtlinuxstb: I check for other codecs yes. in real_read_audio_stream
22:52:10Mikachuhm, the lines 241-249 look a bit strange
22:52:22Mikachufirst you return if obj.fourcc is a particular thing, and then the same if it isn't something else
22:52:28Mikachujust the second check should be enough
22:52:28mtMikachu: I just did that to use them accroding to the type of the unknown.
22:52:44linuxstbmt: Maybe it's just my personal taste, but I don't like mixing variable definitions and code - e.g. the start of rm_parse_header()
22:52:53Mikachumt: it wastes a little stack space :) you could use a union if you want :)
22:53:21***Saving seen data "./dancer.seen"
22:53:37Mikachubut i guess you could change those to lseek too?
22:53:53mtMikachu : I'll just remove all three and lseek instead.
22:54:07linuxstbmt: Also, should line 124 have a "return -1" ?
22:54:51linuxstbmt: But do you check the return value of real_read_audio_stream?
22:55:22linuxstb(I mean real_read_audio_stream_info)
22:56:43 Join Zagor [242] (
22:57:18mtlinuxstb: memset() at the start of rm_parse_header has been pushed down, and added the missing 'return -1'.
23:00:09mtlinuxstb: Added a check for the return value of real_read_*. (oops :/ )
23:03:38 Quit fyrestorm ("lamers envy me like they envy bill g -- main boot xp, just the way it should be!")
23:05:10 Quit stoffel (Remote closed the connection)
23:06:56linuxstbmt: I think using sizeof(id3->id3v1buf[0]) (and [1], [2], [3]) would be nicer than using MAX_STRING in lines 304-310
23:08:58mtYep, definitely better, and done.
23:09:35linuxstbmt: Do you have a new version? I was about to comment on something Mikachu already mentioned...
23:09:42 Join barrywardell [0] (n=barrywar@
23:10:33linuxstbmt: Also, you use "DEBUG_RM" to enable the DEBUGF() lines - shouldn't you use that instead of SIMULATOR?
23:11:41 Join dmb [0] (n=Dmb@unaffiliated/dmb)
23:12:55mtlinuxstb: Just checked something, return -1 isn't needed at line 124.
23:13:17kugelJdGordon|: there's a slight bug with the statusbars
23:13:46*JdGordon| stick fingers in his ears
23:13:55mtlinuxstb: here's the new version but still with the SIMULATOR thing :
23:14:31MikachuJdGordon|: unfortunately this is irc, you'd have to stick the fingers in your eyes :P
23:14:39 Join stripwax [0] (
23:14:51 Quit funman ("free(random());")
23:15:02kugelJdGordon|: at least on my fuze the first line of the statusbar is cropped in the pitchscreen
23:15:09kugelon my e200 not it seems
23:15:24kugelprobably wrong math of mine when doing pitchscreen vp work
23:15:47mtlinuxstb: I just used both (debug_rm and simulator) so that it could be build for the simulator with debugf() enabled or disabled.
23:16:09 Quit evilnick ("Page closed")
23:16:24saratogathe new PP bootloaders seem well recieved
23:16:29saratogaany idea when we'll release them
23:16:40linuxstbmt: Do you need bit_rate, sample_rate and duration in rmctx? They're already in the id3 struct that your codec has access to.
23:16:51JdGordon|kugel: it could be my mistake.... the bar at the bottom looks like its one pixel too high... maybe thats what you're seeing?
23:17:06kugelJdGordon|: that could cause that, yes
23:17:36linuxstbmt: And do you need that switch() in copy_metadata()? Won't the codec always be cook if that function has been reached? (and core Rockbox has already set id3->codectype to that value).
23:18:20gevaertssaratoga: sansa and mr100 are fine I think. I'd just like to see confirmation that the H10 ones work properly
23:19:30kugeldidn't barrywardell own one?
23:19:34linuxstbmt: And you could just pass the address ofo id3v2buf (cast to "RMContext*") as the second parameter to rm_parse_header() - this avoids the memcpy()....
23:19:42*linuxstb thinks he should probably stop reading and let mt commit ;)
23:20:16barrywardellkugel: I did, but the hdd died :(
23:20:17kugellinuxstb: nah, just two more complaints and it's having more code of yours than from mt :D
23:20:38kugelbarrywardell: unfortunate :(
23:20:39saratogadoes anyone own an H10 these days?
23:20:44gevaertskugel: it works on my H10/5 UMS, but there's also MTP and 20GB
23:20:52mtlinuxstb: the variables you mentioned could be written directly to the id3 struct (the it-looks-nicer attitude) .. but this could be changed.
23:21:58mtlinuxstb: the switch is not needed for cook, will remove the 'case cook' part, but later on, this function will be reached even if the codec isn't cook, so the switch will be needed, just not now. :)
23:23:06linuxstbmt: I would get rid of the copy_metadata() function completely, and put anything you need to do in the main function. You could set id3->title etc at the point where you actually read the title.
23:23:23linuxstb(you're not copying the tags, just setting pointers to them)
23:23:40Unhelpfulamiconn: text scrolling in PF is either next on my list or next after 16-point ARM IDCT. i got the impression that greylib doesn't know about viewports... is that correct?
23:24:16amiconnViewports came after greylib, and so far there was no need to introduce them there
23:25:04mtlinuxstb: hmm, actually the variables you mentioned are needed in RMContext, since cook_decode_init() needs them.
23:25:33Unhelpfulok. the text scrolling is going to need some coordinate fixups, then... styled scroll hooks should have access to the viewport, so all should be fine.
23:25:33linuxstbmt: You could pass cook_decode_init() the id3 struct instead of rmcontext.
23:26:13mtlinuxstb: back to the problem of standalone program vs rockbox :)
23:26:17linuxstbmt: But I guess you can start ignoring me now, and just commit it as it is...
23:26:44linuxstbmt: Then get rid of the standalone program - it shouldn't dictate how you write the codec...
23:26:56 Quit bmbl ("Bye!")
23:27:09pixelmaI guess it's not possible to let SVN commit ignore changes in trailing spaces? My editor was so kind as to remove some but there are other white space changes I don't want to have ignored... :\
23:28:01 Quit perrikwp (Success)
23:28:03mtlinuxstb: I could wait a little longer for the commit, those comments aren't easily received everyday ;)
23:28:07linuxstbpixelma: You could perhaps create a patch, then manually edit the patch. (then apply that patch to a clean checkout...)
23:29:12pixelmaundoing them one by one is a bit much I guess (too many files touched), looking at one file each they are not too many and understandable in a diff I think
23:29:21mtlinuxstb: I suppose I could do something condtional, the program is compiling fine currently, I don't think I'd throw it away for such probelm.
23:29:28kugelgevaerts: sansapatcher -bl worked fine here
23:31:32 Join perrikwp [0] (
23:31:49pixelmaI somehow lean towards just committing it still, already put a bit of work in there and I don't think it's too distracting... :/
23:33:16mtlinuxstb: question ; those 3 variables take up 10 bytes of space, is this worth changing whatever functions that will be changed to cope with that ?
23:34:03*pixelma now found a setting in the editor at least
23:34:22mtlinuxstb: And I'm really asking for your opinion not to get to commit quicker :)
23:34:29linuxstbmt: Plus the code to copy them.
23:34:42linuxstbmt: But I agree, it's not a big deal, just the fact that I don't like to see things duplicated.
23:36:40mtSince it isn't abig deal, what about writing the values directly to the id3 struct in the parser and then copying those 3 variables back in the codec before calling cook_decode_init() ?
23:39:13 Quit robin0800 ("Leaving")
23:39:26linuxstbmt: Well, that still duplicates them... I would either leave it as it is, or remove them completely.
23:39:33 Join robin0800 [0] (
23:41:26 Quit Hillshum ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
23:42:53 Quit tessarakt ("Client exiting")
23:45:57 Join stripwax5443 [0] (
23:47:06 Quit archstech (Remote closed the connection)
23:49:23 Join archstech [0] (n=archstec@
23:49:59 Join brndyhite [0] (n=brndyhit@
23:50:08 Join tucsbgns [0] (n=tucsbgns@
23:52:22 Join Thundercloud [0] (
23:56:11 Quit kugel (Remote closed the connection)

Previous day | Next day