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 2008-07-19

00:00:55oborasher: my c220 mentioned Audible on the back, but was a V1
00:01:14rasherobo: okay, so there's some uncertainty either way
00:02:03 Join cbr|w [0] (n=cbr@
00:02:38KeripoHm, page doesn't seem to give enough info ->
00:03:30BigBambiUnfortunately, the only sure way is to check the original firmware version
00:04:42BigBambiHowever, as agaffney said, the strong likelihood is that if it is new, it is a v2
00:05:35 Quit robin0800_ (" HydraIRC -> <- *I* use it, so it must be good!")
00:05:59KeripoReview: "Amazing price on sale. Overall I am very happy with it, but was disappointed it couldn't be rockbox'd because it is a v2" <- Well, there goes my hopes ; /
00:06:55 Quit CyBergRind|w (Read error: 60 (Operation timed out))
00:06:58KeripoWell, thanks anyway everyone for the info. Guess I'll go eBay hunting then
00:07:44 Quit Nibbl (Remote closed the connection)
00:08:07 Join nplus_ [0] (
00:08:44 Quit nplus (Read error: 104 (Connection reset by peer))
00:08:49KeripoWell, scratch that - a few other reviews discuss how they enjoy using Rockbox on it. I'll just end up keeping the receipt and hoping.
00:12:27 Quit PaulJam_ (Read error: 110 (Connection timed out))
00:13:08 Join Zarggg [0] (
00:17:25 Join fiXXXerMet [0] (
00:18:04 Nick nplus_ is now known as nplus (
00:18:38 Quit obo ("bye")
00:19:32fiXXXerMetHow do I transfer music to my device? When I connect it to my computer, it boots into the stock firmware.
00:19:48Nico_PfiXXXerMet: which device is it?
00:19:59fiXXXerMetsansa e260r
00:21:09KeripofiXXXerMet: Rockbox can operate as a file browser I believe, so you can just copy the files over directly and play them in Rockbox. The normal requires the Sansa software for it to recognize copied over music
00:21:15Nico_Pthen it's normal that it reboots into the OF when you plug USB in. Rockbox doesn't (officially) do the USB yet. You can transfer files with the OF
00:22:25fiXXXerMetI have a few hundred songs on there now and I don't see them in the drive that shows up.
00:22:28fiXXXerMetI see the .rockbox folder
00:22:41PaulJamOF = original firmware
00:22:46fiXXXerMetI'll try dragging them to that drive
00:22:48fiXXXerMetOh ok
00:22:59fiXXXerMetWell where are the songs that are on there now?
00:23:01fiXXXerMetHow can I manage them?
00:23:24 Quit ender` (" A man without religion is like a fish without a bike.")
00:23:31KeripoFrom what I've read, Sansa hides the MUSIC folder so turn on hidden folders on your computer
00:23:42fiXXXerMetduh :)
00:23:52fiXXXerMetI have window on here and haven't in so long
00:24:02Keripoyour Sansa probably came with a software CD that had the program for managing the music, content, etc
00:24:08 Join andrewbeveridge [0] (
00:24:20fiXXXerMetMusic folder was hidden
00:24:47fiXXXerMetThank you much
00:26:36andrewbeveridgeis anybody here part of the people who chatted to me a few days ago about wma file playing problems?
00:38:16 Join sitwon [0] (
00:45:51 Join danderson [0] (
00:46:09 Quit nplus (Read error: 104 (Connection reset by peer))
00:46:13 Join nplus [0] (
00:46:23 Join pedrov [0] (
00:49:39 Quit borges ("ChatZilla 0.9.83 [Firefox 3.0/2008052906]")
00:53:37 Quit Zarggg ()
00:54:24*andrewbeveridge is a little impatient
00:54:50 Part Lars_G
00:56:17 Quit sitwon ("Leaving")
00:56:21Nico_PPaulJam: I just managed to reproduce FS #9137 with an MPC album on the sim
00:56:31Nico_Pit caused early skipping
00:56:44PaulJamgood to hear
01:01:45 Part Keripo
01:02:18linuxstbandrewbeveridge: If you have a question, why not just ask?
01:05:51andrewbeveridgehmmm? sorry, I don't really have a question - I'm just hoping/waiting for one of two people to come online :)
01:06:21andrewbeveridgethey already answered my question two/three days ago
01:09:13andrewbeveridgesaratoga is really the only person who can help me, as far as I know he is the only person who has a working knowledge of his own code, so he alone can solve the problem with the wma decoding fault.
01:10:29 Quit pedrov ("leaving")
01:13:38linuxstbandrewbeveridge: If saratoga isn't here now, he may read the logs later (or others can point him to the logs), which is why it makes sense to just ask your question. But other people know the WMA codec - e.g. I wrote most of the ASF parsing code in it.
01:14:11***Saving seen data "./dancer.seen"
01:16:03Nico_PPaulJam: I think I've got it
01:16:06andrewbeveridgeAh ok I don't really have a question as such, I just wanted to know if there was any news about debugging it, or perhaps find out if there is anything I can do myself...
01:17:52Nico_PPaulJam: it's actually a very stupid one... you'll see the fix is a one-liner
01:18:11PaulJamtell me if i should test anything
01:18:17BigBambiandrewbeveridge: You will find in general that usually more than one person know parts of the code - especially when they had a large hand in it, such as linuxstb and wma
01:18:36Nico_PPaulJam: I'll commit the fix in a minute and you can make sure it's actually good :)
01:19:06PaulJamok, thank you!
01:19:08BigBambiandrewbeveridge: But debugging it suggests a problem, and I suspect others would be interested to know what that problm is
01:19:40Nico_PPaulJam: thank *you* :) It was a bad bad bug in the buffering code, good thing you found it
01:20:07PaulJamwell, actually my bugreport was a duplicate :)
01:20:52andrewbeveridgeYeah sorry I suppose I didn't think like that - the problem was that the latest ffmpeg build plays my wma files fine, but rockbox plays the same files badly, with gap every few seconds, producing a "skipping" effect.
01:20:55 Quit shotofadds ("yay for bed!")
01:21:25linuxstbandrewbeveridge: Was your problem file "doors-test.wma", or am I confusing your problem with someone else's?
01:21:38 Join n17ikh [0] (
01:21:41Nico_PPaulJam: of FS #8194?
01:23:31PaulJamNico_P: yes, i think this is the same issue
01:23:38Nico_PI think so too
01:23:44andrewbeveridgelinuxstb: someone else's - my example file was named "Biffy Clyro - Machines [Faulty] ~70kbps VBR.wma"
01:24:17PaulJamjust triggered a bit different by skipping back instead of resinding
01:24:51Nico_Pyeah, same thing
01:25:59linuxstbandrewbeveridge: Ah yes, I remember now. It has corruptions when I play it with (quite an old version of) mplayer as well...
01:26:44Nico_PPaulJam: committed :)
01:26:57linuxstbandrewbeveridge: Did you (or do you know if someone else) tested it with the very latest ffmpeg wma decoder?
01:28:30andrewbeveridgelinuxstb: no, however the ffmpeg executable on my computer ( decodes the file perfectly on my laptop and my server
01:30:40*linuxstb downloads it
01:33:31linuxstbandrewbeveridge: That's pretty new - built on 16th June 2008...
01:33:50andrewbeveridgesaratoga assured me that was the build he used though
01:35:15linuxstbAnd the file didn't decode correctly for him?
01:35:47andrewbeveridgeI'm not sure
01:36:01andrewbeveridgehold on -
01:36:47andrewbeveridgearound 2:00
01:37:39 Quit bertrik (Remote closed the connection)
01:38:29andrewbeveridgeo.O there have been two releases since that build!
01:39:35 Quit Bensawsome (Read error: 104 (Connection reset by peer))
01:39:59linuxstbThat sounds like it might be the bug fixed in r11115 of ffmpeg...
01:41:10 Quit tvelocity ("Αποχώρησε")
01:42:13 Part pixelma
01:42:45 Quit fiXXXerMet ("Leaving.")
01:43:32 Quit desowin (Read error: 104 (Connection reset by peer))
01:44:34andrewbeveridgelinuxstb: could you give me a link to more information about said bug? where are you finding that?
01:44:54Nico_PPaulJam: have you tested?
01:45:27PaulJamNico_P: yes, so far it looks good :)
01:45:44linuxstbandrewbeveridge: I'm looking at the ffmpeg svn log here -
01:45:45Nico_Pgreat :)
01:45:52linuxstbBut I tried that fix in Rockbox, and it made no difference...
01:46:00 Join desowin [0] (n=desowin@atheme/member/desowin)
01:46:47andrewbeveridgedoes my ffmpeg executable decode the file perfectly for you?
01:47:01linuxstbI haven't tried - I'll do that now...
01:48:34linuxstbYes, it seems fine.
01:49:04linuxstbI get the same "Multiple frames in a packet from stream 0" warning that Llorean reported though.
01:49:19linuxstbWhich may give us a clue...
01:50:01andrewbeveridgesame here, but definitley confirmed that ffmpeg build 13378 decodes the file fine. now, if the wma decoder for rockbox was taken from that same build, there must be a bug in the rockbox-created code which is causing the problem, right?
01:50:55linuxstbThe wma decoder itself was taken from ffmpeg (with major changes), but the other part of the codec (the ASF parser) was written from scratch (mostly by me), and may well be buggy.
01:51:47andrewbeveridgeah ok - um what does the asf parser do, in layman's terms?
01:52:21andrewbeveridgei thought asf was just for streaming media?
01:52:26linuxstbasf is the container format that wma files use.
01:52:53 Nick Seedy is now known as Seed (
01:52:56linuxstbSO the asf parser extracts the WMA packets from the ASF file, and passes them to the wma decoder itself.
01:53:22linuxstbEven though WMA files are generally called ".wma", they are ASF files.
01:53:41andrewbeveridgeright - so if the asf parser fails, the wma bytes are taken from the wrong parts of the file on disk? it gets the file structure wrong?
01:54:20linuxstb"multiple frames in a packet" is relatively rare (I think), so that part of the code isn't well tested.
01:54:47andrewbeveridgeright... um is there any way to test the "majorly modified" rockbox wma decoder itself? I mean, to narrow the cause of the problem don to the asf parser
01:55:25 Join D9762862 [0] (
01:56:00D9762862does rockbox work on an 80gb video? Thanks.
01:56:14scorche|shas long as it is a video and not a "classic", yes
01:56:26scorche|shas the front page should say
01:56:36D9762862hmm, mines a classic, is it a no go?
01:56:43andrewbeveridgelinuxstb: how frustrating - a quick google for "multiple frames in a packet" +wma returns nothing but the rockbox irc log
01:57:00scorche|shD9762862: if it is a classic, then i just said
01:57:01D9762862could you suggest any other programs? thanks.
01:57:06scorche|shno i cant
01:57:15D9762862thanks shitty bums.
01:57:26linuxstbYou're welcome.
01:57:38D9762862have a great day.
01:57:40 Part D9762862
01:57:40 Join toffe82_ [0] (
01:57:48scorche|shhe sure was a friendly chap
01:58:01andrewbeveridgeWhat the hell can people not read the FAQ lol... friendly
01:58:18 Quit toffe82_ (Remote closed the connection)
02:02:48 Quit Nico_P (Remote closed the connection)
02:03:07 Quit perrikwp (" ajax IRC Client")
02:07:49 Quit nplus (Remote closed the connection)
02:11:58andrewbeveridgelinuxstb: how could I help figure out the problem with the asf parser then, as it would seem to be the most likely cause of the problem. what language are we talking about here? I'm only experienced in PHP and Java, never got round to learning C.
02:12:22linuxstbandrewbeveridge: I've been looking at your file, and am starting to think the asf parser is working fine, and the problem must be in the decoder... I'm trying to debug things now, and compare our source to ffmpeg
02:12:59andrewbeveridgeok great, thanks - again, is there any way I can help, or shall I just sit back and encourage? :)
02:13:06andrewbeveridgeBigBambi: ?
02:13:21andrewbeveridgeBigBambi: oh, right. C. Thanks :)
02:13:26BigBambiandrewbeveridge: you asked what language
02:15:41 Quit toffe82 (Read error: 110 (Connection timed out))
02:26:44 Join Bensawsome [0] (
02:30:51 Quit faemir ("Leaving")
02:31:50 Part Rasi
02:33:36num1has there ever been a discussion about using git for rockbox?
02:34:25BigBambiBut the majority feel there isn't really a need, although there is a git mirror now I think
02:34:43num1do you know the address of the mirror?
02:35:41BigBambinot sure
02:36:02BigBambiYou could try checking the IRC logs, there was some discussion here
02:36:14num1alright, thanks
02:37:39BigBambi but I think there is now a more official mirror than the one mentioned there. I wasn't really following the discussion however
02:38:00soapcd ..
02:39:56num1lol @soap
02:42:32 Join Davide-NYC [0] (
02:43:31BigBambinum1: git://
02:43:42BigBambinum1: see logs on 2008/07/03
02:43:52Davide-NYCquick (irrelevant) bug report: the FAQ link does not appear on sidebar on
02:44:19LloreanY'know, one thing that never came up in the Git vs SVN debate for Rockbox, is that the SVN revision numbers are really useful when providing support.
02:45:03LloreanDavide-NYC: I think the patch tracker has a webpage category, you probably should post it there since those who can fix it aren't around.
02:45:34linuxstbIt's probably more likely to get changed if you just grab Bagder when he's around though...
02:45:39num1BigBambiL thank you for the links
02:45:50scorche|shor Zagor
02:46:14Davide-NYCllorean: no such bug category?
02:47:10Davide-NYCAlso, the sidebar contains a link to requests portion of the tracker, which I see is hidden otherwise. Good to know there's still a way to post a feature request, even though noone would see it. :)
02:47:35LloreanDavide-NYC: We don't *want* people posting feature requests...
02:47:48Davide-NYCI know, that was a joke.
02:48:14 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
02:53:53 Quit EspeonEefi ("さよなら")
03:06:17 Quit vort3x (Read error: 104 (Connection reset by peer))
03:06:19 Quit Davide-NYC ("ChatZilla 0.9.83 [Firefox]")
03:06:22 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
03:06:37 Join vort3x [0] (n=vortex@unaffiliated/dfa001)
03:10:54 Quit Thundercloud (Remote closed the connection)
03:13:08n17ikhso when the screen blanks do the subsystems that do things like scroll text and refresh the WPS and run the peak meters stop?
03:13:38 Join EspeonEefi [0] (i=espeonee@CERULEANCITY.MIT.EDU)
03:14:16***Saving seen data "./dancer.seen"
03:14:39andrewbeveridgelinuxstb: any luck? :) any particular bit of code you are looking at? I'm just skimming through wmadeci.c - I've never technically learned C but it all makes sense having spent the past few years working in Java...
03:14:56Lloreann17ikh: What do you mean "blanks"? Most players just turn the backlight off, but don't clear the screen.
03:15:04n17ikhI see.
03:15:19n17ikhactually, I'm thinking of the h10 which I haven't used in forever
03:15:33n17ikhI remember rockbox having an option about blanking or something after the backlight shuts off
03:16:25 Join ecl [0] (n=felipe@unaffiliated/ecl)
03:24:02n17ikhI've gotta say, the keymaps on the c200 series are pretty good
03:24:11n17ikhI actually like it a good deal better than the e200
03:24:18n17ikhtiny screen notwithstanding
03:24:24 Part ecl
03:24:32LloreanI think the e200 one allows much more consistent access to the functionality of Rockbox.
03:25:00LloreanHave you tried Pixelma's alternate c200 keymap yet?
03:25:05n17ikhnot yet
03:25:16n17ikhI just started playing with the c200, it's been sitting on my desk for a month
03:25:25n17ikhgetting ready for a long trip
03:25:38n17ikhmy e200's battery is either not charging or dying really quickly
03:25:45n17ikhor is reporting wrong voltages to rockbox
03:26:40andrewbeveridgeanybody else here know enough about the rockbox wma decoder to know how to bugfix?
03:29:50Lloreanandrewbeveridge: Seriously, you don't need to ask about it every 15 minutes. For one thing, a lot of our devs are European, so are on the verge of, or in the middle of, sleep right now.
03:30:14LloreanFor another, you've already more or less talked with everyone who's involved in the WMA codec, anyway.
03:30:23andrewbeveridgeyeah, good point, sorry - though it would be nice if they told me before leaving/sleeping
03:30:48n17ikhDoes anyone know if the c200 still charges when it's off?
03:30:55 Join massiveH [0] (
03:31:23Lloreanandrewbeveridge: What's the hurry?
03:32:47andrewbeveridgemy sister's birthday - I bought the ipod for her, and my/our entire music collection was converted using the same program, so unless I go to the lengthy/lossy process of converting them all, I have no tracks to put on it before wrapping it up :)
03:34:00LloreanSeems to me then, you can solve your own problem by re-encoding the tracks from the original material. It'll get fixed when it get fixed, but it seems somewhat negative to pester other people to do work sooner because you don't want to do extra work.
03:34:38LloreanIf you've filed a bug report, and talked to the appropriate devs, then really the best thing you can do is either work on it yourself, or wait for them to request more information from you.
03:35:15 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
03:35:17n17ikhwould she notice if you converted lossy tracks to say, mp3 v0?
03:35:31 Quit DerDome (Nick collision from services.)
03:35:32 Join DerDome1 [0] (
03:35:34n17ikhand then went back and replaced them once the decoder was fixed
03:35:44 Nick DerDome1 is now known as DerDome (
03:38:42andrewbeveridgeyeah, i'm not going to do any more pestering ;) I was really hoping somebody could point me in the right direction to help myself - I understand the source code pretty well, I just lack a starting point. Sadly there is no original material.
03:38:43andrewbeveridgeIf I still had my original collection (all FLAC, on a 500GB external hard drive), I would never have asked here, I would simply have re-encoded to MP3 or Vorbis, my sister couldnt care less what format the music is in as long as the compression ratio is high enough for the entire 3000-ish tracks to fit on the 4GB ipod.
03:40:19andrewbeveridgeI think thats what I'm going to end up doing - converting them all to MP3 - in fact I've already started. I was just reluctant to do so because I've never managed to find a way to achieve transparent WMA -> MP3 conversion (obviously, they are both lossy)
03:42:58 Quit neddy ("Leaving.")
03:44:22 Nick fyre^OS is now known as fyrestorm (
03:45:17PaulJamif you want to use low bitrates then i think mp3 is not the best choice. ogg vorbis and mpc are said to give better results at lower bitrates.
03:46:19 Quit PaulJam (".")
03:46:35LloreanI thought MPC was tuned for mid-to-higher bitrates.
03:47:05andrewbeveridgeVorbis is clearly the best lossy format for several reasons, including compression ratio, but until now I hadn't found any portable media player compatible with anything other than MP3 and WMA.
03:54:52scorchenothing is "clearly" the best format, but this is getting offtopic...
03:55:25andrewbeveridgetopic aside, i can't see any disadvantages of ogg vorbis
03:57:54massiveHslightly more cpu usage
03:58:06scorchethere is no "topic aside" here
03:58:33andrewbeveridgeok end discussion, sorry.
04:16:33midgeyhmm, still not having any luck on fixing the abort issue on my mac
04:16:53midgeystrange thing is, i could build rbspeexenc fine yesterday
04:18:15midgeyi uninstalled and reinstalled xcode (plus libraries and headers), reran, and built GNU make from source
04:18:34midgeyi can build sim and target builds for gigabeat fine
04:19:16midgeyif anyone has any ideas on what I should try, i'm all ears
04:23:18 Join miepchen^schlaf_ [0] (
04:30:51 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
04:38:29 Quit DrMoos ("Rockbox rules the DAP world")
04:39:27 Quit MXIIA ("System Terminated")
04:40:31 Quit n17ikh ("Leaving")
04:42:12 Join jasutton [0] (
04:44:15 Quit andrewbeveridge ()
04:52:10 Join n17ikh [0] (
04:56:02 Nick jasutton is now known as jsutton (
04:59:06 Join troy_ [0] (n=toppy@
05:11:15 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
05:14:12 Quit Bensawsome ("The awsome is gone :(")
05:14:18***Saving seen data "./dancer.seen"
05:14:48 Join Bensawsome [0] (
05:14:59 Join n17ikh|Lappy [0] (
05:28:03*midgey figured it out
05:28:42midgeyr18002 breaks "make voice" on mac (well, mine at least)
05:29:09midgeyI'm running 10.4.11, and the abort comes from the -arch flag
05:30:04midgeyfor me, $ARCH is empty and using the -arch flag followed by nothing causes an abort
05:31:25 Quit n17ikh (Connection timed out)
05:36:24 Quit jsutton (Remote closed the connection)
05:42:19 Quit Bensawsome ("The awsome is gone :(")
05:43:45 Join Bensawsome [0] (
05:46:25 Quit Bensawsome (Client Quit)
05:46:57 Join Bensawsome [0] (
05:57:41 Quit num1 ("Leaving.")
06:08:40 Quit Bensawsome ("The awsome is gone :(")
06:12:21 Join BlakeJohnson86 [0] (
06:13:36 Quit BlakeJohnson861 (Read error: 110 (Connection timed out))
06:14:41wpyhis anyone working on a big-icon-based menu?
06:14:56wpyhI mean like, right now RB's menu is just a "list" basically
06:15:04wpyhis anyone interested in something with big icons?
06:15:12*wpyh is just suggesting an UI
06:19:03 Join coatman [0] (
06:22:19 Quit Llorean (Read error: 104 (Connection reset by peer))
06:24:09 Join Llorean [0] (
06:28:31 Quit coatman ("Leaving.")
06:28:47 Join coatman [0] (
06:30:02 Quit coatman (Client Quit)
06:30:46Lloreanwpyh: This has been requested before, but it needs to be blind friendly, not slow things down, and not make the binary huge. So the general feeling is "show us a patch and we'll think about it" I think
06:31:14wpyhhm... ok
06:31:29 Quit Llorean (Read error: 104 (Connection reset by peer))
06:31:36wpyhdoesn't go into my TODO list since I (currently) don't know how I would go about implementing it
06:31:48 Join Llorean [0] (
06:32:17wpyhfor it to be blind-friendly, doesn't it just need to have the same text as the regular menu?
06:33:03wpyhthe only differences are: 1) icons are not vertically laid out, but horizontally, and 2) the icons are bigger, and 3) there is a scroll bar
06:35:12LloreanNavigating in a line (list) is easier blind than navigating in a grid, too.
06:35:38wpyhwell, what we have in mind may be different
06:36:00wpyhwhat I have in mind is, for example there are 9 icons
06:36:07LloreanYou're navigating on two axis instead of simply one.
06:36:12wpyhwith the current menu system, it's laid out vertically
06:36:27wpyhwith the new big-icon menu system, it goes like this: 1 2 3\n4 5 6\n7 8 9
06:36:48wpyhand "forward" goes from 1 to 2 to 3 to 4 to 5 .. .and so on
06:36:54wpyhbackward goes the other way round
06:37:04LloreanIn that case, what benefit does it offer?
06:37:12wpyhthe benefit is big icons
06:37:19LloreanWhat functional benefit is that?
06:37:30LloreanIn almost every screen but the main menu, the icons are repetitive.
06:37:33wpyhno functional benefit, just looks good
06:38:09wpyhit will drive people to draw new icons
06:38:16wpyhthat won't make the binary big, would it?
06:38:17LloreanI don't know if you can easily justify a complication of the menuing system for a reason as trivial as that, though, especially since not everyone will agree it looks good.
06:38:42LloreanIt will make the binary larger than it is, and it will mean that you need an entire new version of the list drawing code for big-icon lists.
06:40:24LloreanIf it streamlined the interface by allowing improved navigation methods, that'd be something else. But Rockbox really can't accommodate every aesthetic change someone thinks is a good idea, or it'd quickly snowball into something rather large and ungainly.
06:41:15 Join iceowl [0] (
06:41:18wpyhare you saying that any menu system will be the only menu system in RB?
06:41:48wpyhwhat I'm thinking is that the current menu system can and should stay, while we add an option to switch to the new menu system
06:42:09LloreanWhich makes it even worse because then you have redundant code with no functional benefit.
06:42:54wpyhthe code is loaded dynamically, so it shouldn't increase the binsize
06:42:59iceowlis there such a thing as a Go app for rockbox?
06:43:18Lloreanwpyh: There's no dynamic memory allocation.
06:43:25Lloreaniceowl: "Go app" as in the game?
06:43:35LloreanNot included
06:43:36wpyhwhich brings the question of... is it possible to make the interface separate from the core?
06:43:39Lloreaniceowl: there might be one in the tracker
06:43:47 Quit troy_ (Read error: 104 (Connection reset by peer))
06:43:54iceowltracker? as in torrent?
06:44:00Lloreaniceowl: No, the patch tracker...
06:44:37wpyhLlorean: or make it selectable as a config option, and RB chooses which one to load on boot
06:44:42Lloreanwpyh: The real question is how much work would that take, would it lower performance, and would it make the code harder to maintain.
06:44:45wpyhthis way we don't need maloc()
06:44:58wpyhhmm.... you're being practical here
06:45:01LloreanThen you'd need to reboot to change themes.
06:45:14wpyhLlorean: no, no need to reboot to change themes
06:45:26wpyhwe just need to reboot to switch between menu items
06:45:39LloreanAnd if your theme uses small icons...
06:45:49LloreanThemes already include iconsets of varying sizes.
06:45:50iceowlhuzzah, looks like there
06:46:00iceowlthere's lots of stuff in the patches
06:46:04wpyhwell, the themes can support just one menu system, or both
06:46:29wpyhI think it would be a *lot* of work, which is why I'm not volunteering, but I'd certainly be interested in it
06:46:52Lloreanwpyh: I can't see this being practical.
06:47:04wpyhit might lower the performance somewhat because of another layer of indirection, but it's negligible
06:47:14LloreanThese are not fast devices.
06:47:21wpyhand it will make the code easier to maintain by cleanly separating the interface from the core
06:47:27LloreanAnd every wasted cycle means lowered battery life.
06:47:37Llorean"Negligible" doesn't really apply on battery powered devices.
06:47:46wpyhLlorean: yeah, it will be negligible; it will mostly be an extra pointer
06:48:13wpyhLlorean: I'm not saying you see this as being practical ;) just that you're considering the practical considerations here :)
06:48:31LloreanThe practical considerations are the ones that actually matter.
06:48:41wpyhyes, they are the ones that actually matter
06:49:27LloreanHaving swappable UIs means that any time a new UI screen is added, it needs to be adapted to fit both (all) of them.
06:49:42LloreanOr any time a new port happens ,both UIs need to be ported, and made to deal with the new buttons.
06:50:24wpyhwhat are you referring to "it" when you say "it needs to be adapted..."?
06:51:07iceowlhmm.. I found a go app. hopefully it's not limited to playing with another person
06:51:23Lloreanwpyh: any new UI screen
06:51:43Lloreaniceowl: These players have slow processors. It's unlikely a decent Go AI can run on them at any reasonable speed
06:51:52scorchego AI is not terribly capable in the first place...only 6 or 7 kyu and those are the decent ones
06:52:05iceowlI know
06:52:18iceowlI just want to play Go on the bus to work :/
06:52:54scorcheiceowl: you could probably modify the plugin to set up a series of life-and-death problems
06:53:11iceowlprobably. if I knew how to code at all.
06:53:22wpyhI'm thinking of a generic core-to-ui layer, so anytime a new UI screen is added, it only needs to be compatible with the current layer
06:53:36wpyhjust like how a plugin accesses RB's api
06:54:09wpyhwhen a new port appears, both or one UI needs to be ported in the beginning
06:54:34Lloreanwpyh: When a new feature needs a new screen, each UI will need a version of that screen though.
06:54:42wpyhand the idea of big icons actually comes to me a few days ago, when I saw the Onda VX747 port
06:55:09wpyhon such devices (with touchscreens), it is way easier to click on big icons
06:55:21LloreanYou're basically adding a work-multiplier for those interested in developing new ports, for those providing support for ports, and for those who add new features.
06:55:45wpyhLlorean: hm... I'm not sure I fully understand what you mean by a new UI screen. any examples?
06:55:50*Llorean doesn't see how big icons will be any easier than simply having a large menu font.
06:56:13wpyhsince basically all the UI screens are lists anyway
06:56:36Lloreanwpyh: For example, the Bookmarks list is a unique sort of screen. If it were to be icon based, someone would have to convert that screen separately. If a new feature that used a customized list went in, for the UI to be consistent, a customized version would have to be done for each type of UI available
06:56:51wpyhLlorean: do both or all menu systems have to be supported in a port?
06:57:20LloreanRockbox is one piece of software, we only leave out features if they're features the hardware is incapable of.
06:57:32LloreanSeparate UIs for separate ports is contrary to what Rockbox is generally about
06:58:04wpyhhm... I wonder where the bookmarks screen is
06:58:30LloreanThere's also Quickscreen, the color and equalizer settings, etc.
06:58:49LloreanWhile they could simply be used as they are, they wouldn't fit the "feel" of a big icons theme.
06:59:01wpyhLlorean: ok, then we have to work hard to port both menu systems then
06:59:19wpyhLlorean: that's what I was thinking...
06:59:33wpyhfor example the WPS: it should be the same regardless of menu system
06:59:51wpyhthe big-icon-system would only affect lists
07:00:05 Quit iceowl ("leaving")
07:00:13Lloreanwpyh: But that's just big-icon.
07:00:17wpyhinstead of scroll up/down lists, now we have touchable or horizontal-scrollable lists
07:00:21wpyhyes, big-icon
07:00:29LloreanIf it's going to be abstracted like that, there's no reason other ones won't be proposed.
07:00:39LloreanAs well, you still have to reserve memory for the largest one.
07:00:44wpyhmuch like the iphone's ui (hope I'm not making LambdaCalculus37 mad)
07:00:45LloreanWhich means the binsize issue is still a concern.
07:01:19wpyhcan't the largest size be set as a property?
07:01:27wpyhso that the memory is allocated only at boot
07:02:07LloreanPossible, but I'm rather against a mere theme option requiring reboots or restarting playback.
07:02:09wpyhLlorean: yes, others will probably be proposed, which is good IMO
07:02:38wpyhas I said, reloading themes won't require a reboot
07:02:46LloreanYes it will, if they require different UIs.
07:02:46wpyhuh, I mean changing themes
07:02:58LloreanYour big-list one is strictly compatible with normal lists.
07:03:04LloreanWhich IMHO makes it more or less pointless anyway.
07:04:15wpyhwhat I mean is, when using one UI, loading a theme of a different UI will give an error message, something like "this one is for xxx UI only, please change to that ui and reboot first before loading this theme"
07:04:17LloreanBut if every UI is just a different way of displaying the same old lists, then there's no functional benefit at all.
07:04:29Lloreanwpyh: Then changing themes requires a reboot...
07:04:34LloreanIt just doesn't automatically reboot.
07:04:52wpyhLlorean: you're thinking of someone who changes themes between UI's
07:05:09wpyha theme can support multiple UIs, and a theme can support only one of those UIs
07:05:42LloreanYou've never met our users.
07:05:45wpyhif for example,I'm in the small-icon UI, then changing from "cabbiev2" to "icatcher" works, but I'm still using the small-icon ui
07:05:48LloreanSome admit to changing themes hourly
07:05:58LloreanIf changing themes may require a reboot, it's very not user-friendly
07:06:24wpyhif I want the big icons, then I must change a config option to "big icons"
07:06:34LloreanI didn't mean to say every theme will require a reboot, I mean to say I'm against any theme ever requiring a reboot.
07:06:44wpyhLlorean: if you don't want to expose the incompatible themes, we can hide them
07:07:02LloreanIt's very simple: This idea doesn't add any functionality, is only subjectively aesthetically pleasing, and will complicate support and development.
07:07:05wpyha theme will be much like a firefox plugin
07:07:06LloreanBesides "it looks good" what justifies it?
07:07:15LloreanThemes are .cfg files
07:07:30LloreanAre you suggesting scanning every .cfg in the themes folder to determine which ones require big or little icons?
07:07:32wpyhwhat I mean by a firefox plugin is its versioning
07:07:43wpyhyes, that's what I'm suggesting
07:08:06LloreanThat means an entirely new directory browser feature just for themes.
07:08:08 Quit massiveH (Read error: 104 (Connection reset by peer))
07:08:08wpyhhave a different theme folder for each UI, which you would probably reject anyway
07:08:10LloreanThis gets less and less "low overhead"
07:08:28 Join massiveH [0] (
07:08:28Lloreanwpyh: How would that work for multi-UI themes, then?
07:08:40wpyhLlorean: which one, the first one or the second one?
07:09:08LloreanThe second one. The first one is kinda obvious that you could just list all compatible UIs.
07:09:18scorcheis there any point in continuing this? the end, wpyh can code his ideas...and this in particular will almost assuredly be rejected...the end
07:09:25LloreanThe second one, separate folders, would either require one folder for every permutation of UIs, or theme duplication.
07:09:39wpyhah, the second one... the theme author must write multiple cfg files, which means it will most probably be rejected
07:09:59wpyhscorche: are you saying this will be a NoDo?
07:10:17LloreanI'm saying that unless this can overcome most of the objections I've raised, it's unlikely to be accepted.
07:10:35LloreanNot a "NoDo" so much as a "Don't bother unless you can do it well enough that you can answer most objections"
07:10:35scorcheno...i am saying it doesnt really serve much of a purpose and will likely be rejected unless you can solve the issues that have been brought up
07:10:39*wpyh thinks Llorean is reasonable in this case −− better than a NoDo
07:10:51scorchethat said, feel free to code it up and we can discuss this again then
07:11:22wpyhlet me collect the objections here:
07:11:27wpyh1. binsize
07:11:28wpyh2. overhead
07:11:34wpyh3. code complexity
07:11:37scorchedont make a list and spam the channel...
07:11:55wpyhok, maybe I'll put it in my wikipage
07:12:02scorcheyou can fit plenty of numbers on one line
07:12:24wpyhah, sorr
07:14:21***Saving seen data "./dancer.seen"
07:14:33wpyh1. binsize, 2. overhead, 3. code complexity, 4. needs to be blind friendly, 5. "what functional benefits?", 6. maintenance nightmare, 6. porting nightmare, 7. reboot after every theme change, 8. reboot after every UI change, 9. scan all themes or have multiple folders
07:14:38wpyhdid I miss something?
07:15:41scorcheeven if you did, we can addres it after you solve all those ;)
07:19:12wpyhOK, you can see the summary here:
07:27:21*wpyh wonders why everyone is so quiet again
07:27:35scorchelikely because no one has anything to say...
07:31:03 Quit vort3x (Read error: 54 (Connection reset by peer))
07:31:38 Join vort3x [0] (n=vortex@unaffiliated/dfa001)
07:35:17 Quit jhulst (Read error: 60 (Operation timed out))
07:45:26 Join num1 [0] (
07:59:45MU{lappy}if i want a plugin to possibly display an error and then exit when it's acknowledged, is there something in the plugin API already for that? the closest thing that i see is splash
08:01:54 Quit massiveH ("Leaving")
08:03:14amiconnThat's what plugins normally use when displaying specific errors
08:04:05amiconnIf you just need to error out, it's sufficient to return PLUGIN_ERROR instead of PLUGIN_OK. This makes the core show a generic splash ("plugin returned error")
08:06:52midgeyregarding plugin languages, should plugin strings have their own id range or should they start at zero
08:07:33midgeyfor example, should the first plugin string be number 0x16000 or 0
08:07:38 Quit gevaerts (Nick collision from services.)
08:07:49 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
08:08:01LloreanI think they should start at zero
08:08:28amiconnRight now the language ids are 16 bit, and the msb is already used to distinguish LAN_ from VOICE_
08:08:30LloreanOtherwise plugins will be dependent on the ranges of other plugins (if you add strings to jewels, you might end up restricted by the reserved range because the next plugin in the list already starts at 0xSomething)
08:08:54midgeythat's how they currently are implemented, but i'm not sure how to use those ids for voice ids
08:08:58 Quit scorche|sh (Remote closed the connection)
08:09:32 Join massiveH [0] (
08:10:30midgeyright now to access a core string (LANG_ or VOICE_) you can call talk_id via the plugin api
08:10:32amiconnYes. For plugin localisation and voice, we need a way to segment .lng and .voice files, so that each part (core, plugin) can load the segment it needs
08:11:02midgeyand to access plugin strings, you call plugin_talk_id via the plugin lib
08:11:06amiconnThere needs to be some identifier in the source (the .lang file). Iirc this is already prepared
08:13:11midgeyi've been experimenting with loading a string only if the user matches the caller
08:14:05amiconnHmm. On flash targets this might work, but on HDD targets it would cause unacceptable delays
08:15:22midgeyin lang_load, the entire lng is loaded into memory via one giant read
08:15:35amiconnI think that each lng/voice user should load its segment on startup. .lng can be permanent for the core, the plugins would load it into a local buffer
08:15:53midgeyi experimented with byte-by-byte but not on target. it was probably slow
08:16:21amiconnVoice should have a method to share the core's voice buffer. Of course this means that the core needs to reload its voice segment on plugin exit
08:17:08midgeyplugin_language_builtin is a giant array of all the english strings for plugins and is linking to the plugin_lib. i want to change this so only the need strings are linked to each plugin
08:17:21 Join Rob2223 [0] (
08:18:40midgeywhen you mention loading segments, are you thinking of sorting the lng by user (blackjack at the front, followed by chessbox, followed by chopper, etc) and then storing the id offset, file offset and size?
08:21:03amiconnI was thinking more like "use the existing .lng file format for each user, but instead of having separate files, concatenate them all, and put an offset table inb front of it"
08:22:21midgeywould there only be one english.lang file or english-blackjack.lang english-chopper.lang
08:22:33amiconnOnly one .lng per language
08:22:34 Quit massiveH ("Leaving")
08:23:03amiconnOtherwise the number of files would explode, which means wasting space on FAT (each file would occupy one cluster)
08:23:24midgeyi meant the text file before running through genlang. genlang would still output one lng file
08:24:44amiconnThat should probably also be just one file. Imagine 25 languges * (say) 20 localised plugins. That's 500 files with one file per user ...
08:25:31amiconnIf you're going for one source file per user, they could be put in subfolders though
08:26:01 Quit BHSPitLappy (Read error: 110 (Connection timed out))
08:26:24*amiconn just had an idea
08:27:48amiconnI'm not sure whether it's good or bad: What about putting all translations into one source file, i.e. no more <language>.lang
08:28:27amiconnHmm, that would probably cause problems with detecting which strings need attention for a certain language...
08:29:47midgeyit might be possible, i'm going to look into the concatenation method tomorrow and get working on it
08:29:59midgeybut right now, i'm heading to bed
08:32:03 Quit midgey ()
08:35:17 Quit Rob2222 (Read error: 110 (Connection timed out))
08:48:40 Join Vorador [0] (n=tomas@
08:48:58 Quit Vorador (Client Quit)
08:55:16 Join ender` [0] (
09:12:40 Quit advcomp2019 (Read error: 104 (Connection reset by peer))
09:12:54 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
09:14:22***Saving seen data "./dancer.seen"
09:26:02 Quit scorche (Nick collision from services.)
09:26:17 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
09:26:52 Join scorche [0] (i=Blah@rockbox/administrator/scorche)
09:27:18 Join sarkar112 [0] (n=sarkar11@unaffiliated/sarkar112)
09:27:52 Part sarkar112 ("Leaving")
09:28:16 Join fml [0] (n=4fd3d1c2@gateway/web/cgi-irc/
09:29:16 Join swears [0] (
09:30:51fmlI wonder why the WPS code hasn't caused rather bad crashes. Many tags (e.g. %mm) access the int at the address 0. Is it ok? I partly fixed this in FS #8965. But how didn't it strike before?
09:32:15amiconnReading from address 0 doesn't cause a crash by itself. On SH1 and coldfire you can enable a debugging aid that will fire an exception when this happens though
09:32:32amiconnIt's nevertheless bad behaviour, and needs to be fixed
09:36:40 Join fyre^OS [0] (
09:36:45 Join thegeek_ [0] (
09:36:48 Join Lambdumb [0] (n=Lambda@
09:36:56 Join CyBergRind|w [0] (n=cbr@
09:37:03 Join midkay_ [0] (
09:37:24 Join pixelma_ [50] (i=pixelma@rockbox/staff/pixelma)
09:37:46 Quit pixelma (Nick collision from services.)
09:37:46 Nick pixelma_ is now known as pixelma (i=pixelma@rockbox/staff/pixelma)
09:39:53 Quit svens (
09:39:53 Quit perrikwp (
09:39:53 Quit cbr|w (
09:39:53 Quit fyrestorm (
09:39:53 Quit soap (
09:39:53 Quit tarbo (
09:39:53 Quit thegeek (
09:39:53 Quit Lambduh (
09:39:53 Quit erram (
09:39:53 Quit midkay (
09:39:53 Quit dionoea (
09:39:53 Quit basti (
09:39:53 Quit cpbills (
09:40:14NJoinperrikwp [0] (i=4aa794a0@gateway/web/ajax/
09:40:14NJoinsoap [50] (n=soap@rockbox/staff/soap)
09:40:14NJointarbo [0] (n=me@unaffiliated/tarbo)
09:40:14NJointhegeek [0] (
09:40:14NJoinerram [0] (
09:40:14NJoinmidkay [0] (n=midkay@rockbox/developer/midkay)
09:40:14NJoindionoea [0] (
09:40:14NJoinbasti [0] (i=bnc@
09:40:14NJoinsvens [0] (
09:40:14NJoincpbills [0] (
09:44:15 Join davina [0] (
09:48:35 Part num1
09:51:11 Join scorche|sh [50] (n=scorche@rockbox/administrator/scorche)
09:53:45 Quit thegeek (Read error: 110 (Connection timed out))
09:54:41 Quit fml ("CGI:IRC")
09:54:43 Quit midkay (Connection timed out)
10:05:56 Quit GodEater (Read error: 104 (Connection reset by peer))
10:06:00 Join GodEater__ [50] (n=bryan@rockbox/staff/GodEater)
10:09:18 Nick JdGordon|afk is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
10:11:11 Quit swears (Read error: 104 (Connection reset by peer))
10:12:18 Join swears [0] (
10:12:23 Quit avis ("Leaving")
10:18:40 Join Horschti [0] (
10:19:24 Quit Horscht (Nick collision from services.)
10:23:25 Quit amiconn (" reboooot")
10:26:28 Join avis [0] (n=ident@pdpc/supporter/student/avis)
10:26:47 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
10:27:30 Join bertrik [0] (
10:36:48 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
10:38:42 Quit swears (Read error: 104 (Connection reset by peer))
10:39:18 Join swears [0] (
10:43:27*bertrik got 14:41 runtime on his c200 :)
10:44:51pixelmawow, hadn't measured recently. I just remember getting about 8 hours in a runtime test somewhen before Buschel's power saving works
11:04:13 Quit DaCapn (Remote closed the connection)
11:07:20bertrikpixelma, I'm now testing a patch that completely disables the display controller if the screen is off
11:08:02bertrikno sorry, actually it turns off the display and puts the controller in standby mode
11:09:25pixelmawonder how much difference that'll make. It's already become quite ok recently (just my impression though)
11:11:08bertrikI don't expect it to be a lot of difference, but every little bit helps
11:14:26***Saving seen data "./dancer.seen"
11:18:38 Quit pixelma (" bbl")
11:44:10 Join Thundercloud [0] (
11:57:08*amiconn wonders whether we should add large file support to rockbox
11:57:32amiconnIt would mean going 64 bit for the file api... is it worth it?
11:58:27amiconn64 bit for ssize_t, that is
11:58:35linuxstbI can't recall many complaints about the fact that we don't...
11:58:41amiconn(and friends)
11:59:08LloreanWhat happens if you see a >2GB-1 file in the browser and/or select properties or something on it?
11:59:38amiconnUmm, actually I don't know
11:59:50amiconnEither it will show a negative value, or throw an error
12:00:03LloreanI can't imagine people wanting to open giant files on any of our devices, and recording split cap uses a safe upper limit anyway I think, but I'm sure there are people who might user their DAPs to store a few large files sometimes
12:00:23LloreanI don't see that we need to "support" them as long as there's no "bad" behaviour resulting from having 'em around
12:00:41amiconnRecording cares about this, yes. But someone could put really large files on it, be it music or video
12:00:57domonokyisnt it difficult to get > 2gig files on a fat32 partition ?
12:01:16amiconnWindows should have no problem putting such files on it
12:02:34LloreanI think a lack of large file support is more likely to cause some strange behaviour in the future when nobody's thinking about it (or when nobody around knows about it)
12:03:59amiconnFor .wav, it needs 3 hours 39 to exceed the 2GB limit (standard CD quality)
12:04:31bertrikI think we should at least make sure we fail gracefully if we don't support it (i.e. refuse to work with large files instead of crashing)
12:10:10linuxstbshotofadds: (for the logs) I've just tried the NAND driver on the DAX, and it fails to mount the partition, but I'm not sure what I should do to help debug...
12:13:09 Quit Thundercloud (Remote closed the connection)
12:24:27 Join obo [0] (n=robert@rockbox/developer/obo)
12:32:18 Join nplus [0] (
12:38:08 Join pixelma [0] (i=pixelma@
12:38:30 Quit AndyI (Read error: 110 (Connection timed out))
12:51:48 Join moos [0] (
12:55:26 Join TaffinFoxcroft [0] (i=79d9d79c@gateway/web/ajax/
12:56:15TaffinFoxcroftanyone know how to get a creative zen v plus working with the MTP plugin?
12:57:03TaffinFoxcroftoops wrong channel
12:57:05 Part TaffinFoxcroft
12:57:12 Join petur [50] (n=petur@rockbox/developer/petur)
12:59:14 Join aliask [0] (
13:14:30***Saving seen data "./dancer.seen"
13:19:17 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
13:24:09 Quit Horschti ("")
13:28:22 Quit swears (Read error: 104 (Connection reset by peer))
13:39:44 Join Thundercloud [0] (
13:41:02 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
13:45:29 Join Lear [0] (i=chatzill@rockbox/developer/lear)
13:56:34 Quit amiconn (Nick collision from services.)
13:56:40 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
13:58:23 Quit GodEater__ (Read error: 113 (No route to host))
14:04:37 Quit k4y_ (Read error: 110 (Connection timed out))
14:05:09 Join MethoS [0] (n=clemens@
14:12:46 Join k4y [0] (
14:26:06 Join Rob2222 [0] (
14:30:14amiconnLlorean: I have created a ~2.4GB WAV file and put it on my H340
14:30:40amiconn'properties' shows a very large negative size (in bytes)
14:31:37amiconnPlaying it doesn't work - it's skipped immediately
14:31:56amiconnThat's probably because negative return values from open() mean error
14:32:55amiconntest_codec splashes "Cannot read metadata", probably due to the same reason
14:33:20 Quit Rob2223 (Read error: 60 (Operation timed out))
14:33:32 Part pixelma
14:33:43*linuxstb senses amiconn is tempted to implement large file support...
14:34:02amiconnThe question is whether it'd make sense
14:34:19amiconnAt least rockbox doesn't do something nasty when encountering such files
14:35:10linuxstbIt would be nice to fully support the capabilities of FAT, but I don't know if it's worth introducing lots of 64-bit ints to do it...
14:35:16 Join LambdaCalculus37 [0] (i=1800dac6@gateway/web/ajax/
14:35:45 Join Mathiasdm [0] (
14:35:49 Join PaulJam [0] (
14:36:13amiconnThe question is how many places would need to use 64 bit
14:36:45*petur remembers getting called back when he added support to properties for large files
14:36:50linuxstbCodecs would need them - it could be significant for codecs that buffer the seektable.
14:37:03amiconnAll plain filesize values can still be stored as 32 bit (unsigned)
14:37:20linuxstbHmm, that's true, we're still limiting to 4GB...
14:37:32amiconnJust the special cases where negative values mean 'error' need to be 64 bit
14:37:32linuxstbSo no problem for seektables
14:37:51amiconnAnd that's mainly the file system api
14:38:00amiconn(because it's posix)
14:38:26*petur likes having funtion results returned separate from retuned data
14:43:02 Quit EspeonEefi ("さよなら")
14:43:05bertrikwhich functions of the file system api are we talking about really?
14:43:26linuxstbopen, lseek, read, write...
14:43:54linuxstbBut maybe we don't need to be able to read/write more than 2GB...
14:44:57 Join EspeonEefi [0] (i=espeonee@CERULEANCITY.MIT.EDU)
14:45:51*amiconn thinks it's strange that read() and write() expect the count as size_t, but return ssize_t
14:46:13amiconnIn the special (v)fat32 case, only ssize_t would need to be 64 bit
14:46:24amiconn(and off_t)
14:46:59peturdirent specifies size as long, that could change to unsigned long
14:48:06 Join Mathiasdm2 [0] (
14:49:02amiconnHmm, open() itself doesn't have problems with >=2GB, because it only returns the file descriptor
14:49:23 Join pedrov [0] (
14:49:23 Quit LambdaCalculus37 (" ajax IRC Client")
14:51:05amiconnLooks like functions in file.c needing attention are read(), write(), lseek(), filesize(), ftruncate() (and related static functions)
14:51:34amiconnfat.c may also need attention, as well as dircache
14:54:01bertrikso the options are 1) do nothing 2) extend max filesize to 4 GB by using unsigneds 3) go full blown 64-bit for size_t, ssize_t and off_t
14:54:47amiconnThe H300 OF does handle my 2.4GB WAV file...
14:55:05bertrikI prefer option 2 if possible
14:55:21amiconn(2) also requires a few places being converted to 64 bit
14:56:19amiconn.. ssize_t and off_t. size_t could stay 32 bit, but somehow that sounds dirty...
14:57:17amiconnOuch, OF seeking is sloo-oooww
14:57:40amiconnIt basically just does 3 second steps, which is a bit fine-grained for a 4-hour file
15:00:28 Join Nibbl [0] (n=Nibbler@
15:00:40bertrikamiconn, you mean that using 32-bit unsigneds will be a bit of a hack, so it's basically all-or-nothing w.r.t. 64-bit file file lengths and offsets?
15:01:10amiconnIt is a hack. Not sure whether it's good or bad.
15:02:55 Join mugedeliorman [0] (n=mugedeli@
15:03:02 Quit Mathiasdm (Read error: 110 (Connection timed out))
15:04:59 Quit mugedeliorman (Client Quit)
15:06:14bertrikindeed, off_t has to be a signed, to make it possible to seek x bytes back from the end for example
15:10:59amiconnThere is another detail which needs attention. If the api goes 64 bit, the fat driver needs to ensure the max. file size
15:13:59*linuxstb wonders what twiki name to use for a page trying to document the PP502x
15:14:23bertrikotoh, things may just magically work themselves out because of how overflows work
15:14:32***Saving seen data "./dancer.seen"
15:15:12amiconnDon't rely on that here.
15:15:44rasherlinuxstb: PortalPlayer502x?
15:15:47amiconnThe cluster chain can easily grow beyond 4GB, but the size field in the dir entry is only 32 bit
15:16:06linuxstbrasher: Just what I had in mind...
15:16:32amiconnSo if you write a >= 4G file, the dir entry will be modulo 4GB, but the diskspace will be taken for the whole amount
15:17:01amiconnI'm not even sure whether chkdsk would detect such a situation
15:18:23bertrikso don't write files larger than 4 GB?
15:21:34bertrikif the dir entry supports only 32-bit, that's a limit of the file system anyway
15:22:02amiconnYes, the 4GB limit
15:22:13amiconnAnd the fat driver has to make sure it isn't exceeded
15:22:40PaulJam<amiconn> Ouch, OF seeking is sloo-oooww <= you should have used the Study Mode :)
15:24:01bertrikamiconn, ok, I misunderstood what you were talking about
15:24:28 Quit amiconn (Nick collision from services.)
15:24:35 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
15:25:27 Join MethoS- [0] (n=clemens@
15:30:42*linuxstb creates PortalPlayer502x and invites anyone that knows anything to edit it...
15:32:27 Join fdinel [0] (
15:36:57 Quit Nibbl ("Leaving")
15:40:48 Quit MethoS (Read error: 110 (Connection timed out))
15:41:08 Join faemir [0] (
15:43:15 Join Arathis [0] (
15:43:56 Join cbr|w [0] (n=cbr@
15:47:04 Quit Arathis (Client Quit)
15:47:43 Join MethoS_ [0] (
15:50:13 Join Arathis [0] (
15:50:43 Quit CyBergRind|w (Read error: 104 (Connection reset by peer))
15:51:07 Quit cbr|w (Read error: 104 (Connection reset by peer))
15:51:08 Join HTMLCODER [0] (
15:54:06HTMLCODERHello, may I ask if Rockbox does support .png images?
15:54:38Nico_Pit doesn't
15:55:14 Part HTMLCODER
15:55:14 Join toffe82 [0] (
15:56:17 Part wpyh
15:57:19 Quit aliask ("ChatZilla 0.9.83 [Firefox 3.0.1/2008070208]")
16:04:02 Quit MethoS- (Read error: 110 (Connection timed out))
16:07:43 Quit Horscht ("User was distributing pornography on server; system seized by FBI")
16:10:22 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
16:25:40 Join Vorador [0] (n=tomas@
16:36:19 Nick JdGordon is now known as JdGordon|zzz (n=jonno@rockbox/developer/JdGordon)
17:09:05 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
17:14:33***Saving seen data "./dancer.seen"
17:33:15 Quit Mathiasdm2 ("Invisible Internet Project:")
17:42:04 Join wpyh [0] (
17:42:07 Join GodEater__ [50] (n=bryan@rockbox/staff/GodEater)
17:49:37 Quit nplus (Read error: 104 (Connection reset by peer))
17:50:23 Quit BlakeJohnson86 (Read error: 110 (Connection timed out))
17:50:40 Join BlakeJohnson86 [0] (
18:05:07 Join noapic [0] (n=noapic@
18:05:28noapichello ... I am interested in porting libwma to Sony PSP.
18:05:37noapicany ideas on where to begin :D ?
18:07:28 Join midgey [0] (
18:09:40 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
18:18:30 Quit noapic ("WeeChat 0.2.7-dev")
18:18:31linuxstbnoapic: That would seem to be a question for a Sony PSP channel - libwma is simply C code...
18:18:36 Quit GodEater__ ("Going!")
18:22:50 Join MethoS [0] (n=clemens@
18:28:54 Quit ze (Read error: 104 (Connection reset by peer))
18:29:14 Join massiveH [0] (
18:38:18 Quit MethoS_ (Read error: 113 (No route to host))
18:44:54 Quit Lear ("ChatZilla 0.9.83 [Firefox 3.0.1/2008070208]")
18:48:38 Join MethoS- [0] (n=clemens@
18:54:13 Join massive_H [0] (
18:56:02 Quit massiveH (Nick collision from services.)
18:56:05 Nick massive_H is now known as massiveH (
19:04:14 Quit MethoS (Read error: 110 (Connection timed out))
19:12:30*amiconn wonders why linuxstb kept known outdated information when adding the ipl PP502x stuff
19:14:34***Saving seen data "./dancer.seen"
19:18:26Nico_Pbecause he didn't know it was outdated?
19:32:12soapIs the plan to use the PP USB stack for iPods even if not Sansas for 3.0?
19:32:48amiconnIt's still not working properly in some environments
19:33:19amiconn(front ports with the obviously common signal quality problems, hubs)
19:33:56amiconnI wouldn't enable it for the release unless this problem is solved by then
19:34:00soapIs there any need/desire to encourage more widespread testing, or is developer time and or technical analysis needed more than user case-studies?
19:34:17amiconnThe latter
19:34:29amiconnThere must be some setup detail we're missing
19:34:56*amiconn does have the usb stack enabled on all his PP502x targets, but tends to not use it
19:35:04soapis bench testing (scope) needed more or simply knowledgeable eyeballs?
19:35:25amiconnEven when connected to a root port there are occasional resets, stalling the transfer for a while
19:37:19 Join merbanan [0] (n=banan@
19:37:47linuxstbamiconn: I simply wanted to make a verbatim copy as the first "commit"
19:39:13amiconnDo you want to copy the PP5002 page as well?
19:39:41linuxstbI will do, but I'm about to walk out of the door to go out for the night. I can do it tomorrow if no-one else has done it.
19:40:13*amiconn should probably add his knowledge regarding clock setup
19:40:30amiconn(both PP5002 and PP502x)
19:40:34*linuxstb walks out of the door
19:51:53 Join bugjuice [0] (
19:52:42bugjuicecould someone offer some help with a small problem I have?
19:54:32 Join bluebrother [0] (n=Dom@rockbox/staff/bluebrother)
19:56:07bugjuiceupon fresh install of rockbox, my wps graphics won't load
19:58:29PaulJamthe theme is probably broken due to recent syntax changes. see if you find an updated version in the WpsGallery in the wiki or fix the theme. details here:
19:59:58bugjuicethx. I'll investigate.
20:00:54bugjuiceso if i go back to an earlier build, it should fix?
20:01:39 Join swears [0] (
20:03:25 Join Bensawsome [0] (
20:04:16bertrikNo, I think it's better to just fix the wps. Going back to an earlier version is a dead end.
20:04:23 Quit BigBambi (Read error: 113 (No route to host))
20:05:58bugjuiceby "dead end" you mean, counterproductive
20:16:59 Join saratoga [0] (n=41becb3b@gateway/web/cgi-irc/
20:17:01 Join MethoS [0] (n=clemens@
20:17:27saratogaif that noapic guy comes back tell him to contact me about the wma codec
20:17:59bugjuicethx for the help and well done to all developers who helped with the latest build. rockbox continues to improve!
20:18:03 Quit bugjuice ()
20:28:49 Quit massiveH ("Leaving")
20:31:50 Quit MethoS- (Read error: 113 (No route to host))
20:34:41 Join Strife89 [0] (n=michael@
20:35:30Strife89Is there, by chance, any way to make Rockbox show a theme even if it runs off the screen?
20:37:23Strife89I'm attempting to port a theme to the c200; that's why I'm wondering.
20:38:14PaulJami don't think that is possible
20:38:42Strife89Hmmm, darn.
20:38:47 Quit Thundercloud (Remote closed the connection)
20:40:06 Join coatman [0] (
20:41:25PaulJamwhat you could try is using an uisimulator for a device with a larger screen and modify tha WPS until it fits in an area in tha upper left corner of the size of the target screen. but i'm not sure if this will make it easier.
20:42:09Strife89I did (iPod Video) and the WPS refused to display.
20:42:26bluebrothersounds like the wps itself is broken.
20:42:36Strife89More like I broke it. :(
20:42:45PaulJamthe console output should give you a hint where it fails
20:42:48bluebrotherthat's an option too ;-)
20:42:56bluebrotheryou could also try to use checkwps.
20:43:09Strife89How do I bring up the console output (Linux binary)
20:43:10bluebrotherbut running the sim (using −−debugwps) gives a similar output
20:43:23bluebrotherrun it from a console ;-)
20:43:28bertrikdoes anyone have PCB photo or scans of a c200 that are more detailed than the ones at anythingbutipod?
20:43:58Strife89I could photograph mine.....
20:44:49Strife89Anyway, thanks Paul and blue, I'll try those.
20:44:53 Part Strife89
20:45:24bertrikI have one too that I can open and photograph, but if there's already a collection of high-res photos available, I'd rather use those
20:45:30soapsaratoga, congrats (again?) on the WMA improvements.
20:46:34saratogasoap: thanks
20:47:16saratogaregarding wma, i don't mind talking about porting it to other projects so if anyone else comes in asking about it i can talk to them
20:50:21 Join MethoS- [0] (
20:51:07merbanansaratoga: do you know what kind of fft the tremor transform uses ?
20:55:12 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
20:55:49 Quit bluebrother ("Leaving")
20:56:38 Join MethoS_ [0] (
20:58:48 Quit swears (Read error: 104 (Connection reset by peer))
20:59:47 Quit n17ikh|Lappy (Read error: 104 (Connection reset by peer))
21:00:30 Quit fyre^OS (Read error: 104 (Connection reset by peer))
21:00:37 Join swears [0] (
21:01:02 Join fyrestorm [0] (
21:03:07 Quit MethoS (Read error: 110 (Connection timed out))
21:09:39 Join BigBambi_ [0] (n=Alex@rockbox/staff/BigBambi)
21:10:41 Quit MethoS- (Read error: 110 (Connection timed out))
21:14:03 Join MethoS [0] (
21:14:14 Quit BigBambi (Read error: 113 (No route to host))
21:14:36***Saving seen data "./dancer.seen"
21:16:11saratogamerbanan: you mean like radix-2 or whatever?
21:16:18saratogai'm not sure what the specific transform used is called
21:21:08 Quit Arathis ("Bye, bye")
21:24:48 Quit saratoga ("CGI:IRC (EOF)")
21:30:55 Quit MethoS_ (Read error: 113 (No route to host))
21:36:50 Quit coatman ("Leaving.")
21:37:10 Join coatman [0] (
21:39:29 Join massiveH [0] (
21:41:41 Join MethoS- [0] (
21:42:34 Quit Bensawsome ("The awsome is gone :(")
21:45:41 Join n17ikh [0] (
21:46:22 Quit swears (Read error: 104 (Connection reset by peer))
21:46:59 Join swears [0] (
21:50:52 Quit n17ikh (Read error: 104 (Connection reset by peer))
21:51:31 Join n17ikh [0] (
21:53:28 Quit n17ikh (Read error: 104 (Connection reset by peer))
21:55:40 Join SirFunk [0] (
21:57:29 Quit MethoS (Read error: 110 (Connection timed out))
21:59:06 Join Bensawsome [0] (
22:13:03 Quit swears (Read error: 104 (Connection reset by peer))
22:13:40 Join swears [0] (
22:21:46 Join nuonguy [0] (
22:26:22 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
22:27:10saratogahow do I make the flac test program?
22:28:15 Join bluebrother [0] (n=dom@rockbox/staff/bluebrother)
22:28:17saratoganever mind i can't read
22:40:55 Quit coatman ("Leaving.")
22:40:59 Join coatman [0] (
22:43:29midgeydomonoky: did you see what i wrote yesterday about you changes in r18002?
22:48:21 Quit swears (Read error: 104 (Connection reset by peer))
22:49:09 Quit bluebrother (Nick collision from services.)
22:49:14 Join bluebrother [0] (n=dom@rockbox/staff/bluebrother)
22:50:30 Join swears [0] (
22:55:21 Join ze [0] (
23:02:05domonokynope, will check the logs :-)
23:05:03domonokyah, i see the problem...
23:08:22 Quit massiveH ("Leaving")
23:09:22bluebrotherdomonoky: why not build the objects for ppc and i386? Then you can build the universal lib in one go
23:09:32bluebrother*into subfolders
23:09:45bluebrother(insert at correct position ;-)
23:11:59domonokybluebrother: feel free to improve, i just needed to work around the problem, that ar doesnt handle universal binarys..
23:12:09saratogahow do I tell a makefile to make a file one level up in the tree?
23:12:23saratogatrying to do file.o: ..file.c doesn't seem to work
23:12:37saratoga file.o: ../file.c
23:14:02domonokymidgey: i commted a fix, i hope it works (it wont be universal)
23:14:06 Quit ender` (" Just because I have a short attention span doesn't mean I")
23:14:40***Saving seen data "./dancer.seen"
23:15:23 Join massiveH [0] (
23:19:55bluebrothersaratoga: make a file one level up? What do you want to do?
23:23:24 Join coatman1 [0] (
23:23:44 Quit coatman1 (Client Quit)
23:24:01 Join coatman1 [0] (
23:24:28 Quit coatman (Read error: 104 (Connection reset by peer))
23:25:54 Quit coatman1 (Client Quit)
23:26:11 Join coatman [0] (
23:26:13bluebrotherdomonoky: do we want to build the universal binary in all cases on mac? I guess not.
23:28:18domonokyhm, it wouldnt hurt to build it always universally, if it wasnt so complicated to build, damn ar ..
23:30:53saratogabluebrother: I want to make a standalone codec like flac does, but to do that I need to have the makefile reach down one level from the codec's folder to the codec's parser
23:30:54bluebrotherdoes osx allow building as universal in all cases? Or do you need to install something special for that to work?
23:31:38 Join no_BRAINerd [0] (n=4aebee43@gateway/web/cgi-irc/
23:31:41bluebrothersaratoga: wouldn't it be better to have a makefile a level up and let that do the work?
23:32:30saratogabluebrother: it would get reverted pretty quickly if I put a makefile in apps/codecs
23:32:52bluebrotherhmm, right. But doesn't using relative paths not work?
23:32:57domonokyi think you dont need anything special to build universal binarys, you just pass two "-arch" commands to gcc, but then ar cant handle the .o files correctly :-) (and also some minor makefile tweaks are needed) ..
23:33:55saratogaso I just need to put an absolute path?
23:34:47bluebrothersaratoga: no. For example, the librbspeex Makefile simply has "../rbspeexenc".
23:35:07bluebrotherShould work similarly for you (or I haven't got your problem correctly ;-)
23:36:33saratogabluebrother: where is that makefile?
23:37:47 Quit massiveH (Read error: 104 (Connection reset by peer))
23:38:44midgeydomonoky: seems to work great, thanks
23:39:54saratogagives "make: *** No rule to make target `wma.o', needed by `test'. Stop. "
23:40:20bluebrothersaratoga: you don't provide rules for the other source files as well.
23:40:23 Quit jhulst (Read error: 113 (No route to host))
23:40:44saratogabluebrother: needed to change "wma.o" to "../wma.o"
23:41:57 Join ZincAlloy [0] (n=d9eefd8a@gateway/web/cgi-irc/
23:42:31bluebrotherhow did you produce wma.o? using an implicit rule?
23:43:28saratogabluebrother: I'm not sure what you mean
23:43:40 Join BRAINerd [0] (
23:44:16 Quit no_BRAINerd ("CGI:IRC (EOF)")
23:45:16bluebrotherwell, you need a rule to compile wma.o from wma.c. The makefile you posted only lists wma.c as prerequisite for wma.o, but doesn't contain a rule
23:45:18 Quit SirFunk (Remote closed the connection)
23:46:43saratogai don't know enough about makefiles to say
23:46:48saratogai've always just done it this way
23:48:05bluebrothertry something like this:
23:48:12bluebrothercouldn't test it though.
23:48:49bluebrotheroh, and you might want to add $(CFLAGS) to the gcc call.
23:49:02 Quit Vorador ("Ex-Chat")
23:51:02bertrikbluebrother, there are implicit rules to go from .c to .o, but that is still a bit of black magic to me too
23:51:32 Quit merbanan (Remote closed the connection)
23:51:55bluebrotherbertrik: true, but do those implicit rules still get applied if you provide a new rule just adding prerequisites?
23:52:54bluebrotherbut it looks so ...
23:53:24 Join SoulSeeker [0] (
23:53:50 Quit Bensawsome ("The awsome is gone :(")
23:58:12 Quit coatman (Read error: 60 (Operation timed out))
23:58:25*bluebrother wonders if saratoga had success

Previous day | Next day