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 2009-05-28

00:00:26linuxstbfunman: Generating C code in .h files isn't considered nice...
00:00:34yomamathecableguI downloaded the most recent version. How do I run it on a Mac?
00:00:45mcuelenaerehmm well not every function is ported to Lua yet, so it isn't really complete; but most seem to work
00:00:52 Nick fxb__ is now known as fxb (
00:01:01funmanlinuxstb: they're included only once
00:01:04 Quit Thundercloud (Remote closed the connection)
00:01:59funmanthink of .xpm files
00:02:30linuxstbfunman: It's still not nice style IMO. How about combining them all into one .c (and .h) file?
00:02:43AlexPrasher: there?
00:03:02funmanhum why not
00:04:13Bagderfunman: I agree with linuxstb, data such as arrays should be defined in .c files for clarity
00:04:23funmanwhat if we declare them as static?
00:04:48Bagderwhy have them in the .h file in the first place?
00:04:56funmanone less file under revision control
00:05:14kugelyou could include the .c too :)
00:05:15yomamathecableguAlexP: I downloaded the most recent version. How do I run it on a Mac?
00:05:16Bagderuh, why is a .h file less than a .c file?
00:05:33AlexPyomamathecablegu: as root, run tools/
00:06:14funmanhum we could just define the arrays in mkamsboot.c instead of using a .h
00:08:25 Quit SirFunk_ (Read error: 110 (Connection timed out))
00:09:12linuxstbfunman: Or combine them into one .c (and one .h), so there is a total of two files to commit.
00:09:46yomamathecableguAlexP: Hmm apparently I'm not root. How do I become root?
00:10:14funmanlinuxstb: or combine them into one .c , and define the arrays in mkamsboot.c ?
00:10:16LambdaCalculus37sudo is in OS X.
00:10:19kugelroot isn't needed, depending on the install dir
00:10:21AlexPyomamathecablegu: I don't know, I don't have a mac. sudo or su normally (google might help here, it isn't Rockbox specific)
00:10:32AlexPkugel: No, but for the default it is
00:11:24 Join pixelma_ [0] (
00:11:24 Join amiconn_ [0] (
00:11:28 Quit amiconn (Nick collision from services.)
00:11:28 Quit pixelma (Nick collision from services.)
00:11:37 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
00:11:47 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
00:14:44 Part wincent ("Kopete 0.12.7 :")
00:15:20 Quit HellDragon (Remote closed the connection)
00:15:25linuxstbfunman: Don't you need to know the size of them?
00:15:57funmanhum right O:-)
00:16:09funmanBagder confused me !
00:16:25yomamathecableguAlexP: Well, I can't seem to get it to work :/ I would need to get my dad to put in his password, so I'll have to wait a while. Thanks for your help so far. ^_^
00:17:14kugelyomamathecablegu: try "RBDEV_PREFIX=/path/to/install ./ " that doesn't need root/sudo if you have controll over /path/to/install
00:17:23 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.10/2009042316]")
00:20:04 Quit trisiak (Read error: 110 (Connection timed out))
00:20:46AlexPsaratoga: I've done virtually all of the other languages, I just need to check a couple of things then I'll do it tomorrow
00:21:27yomamathecablegukugel: it still says permission denied :/
00:21:43kugelcan you write to /tmp?
00:22:18kugelif not, you can also specify another tmp direcotry with RBDEV_DL and RBDEV_BUILD
00:22:23 Join HellDragon [0] (
00:22:33kugelyou should get a short introduction when you run it
00:27:56 Quit ender` (" We are surprised at our own versatility in being able to fail in so many different ways. -- Samuel McChord Crothers")
00:31:25 Join Tichodroma [0] (n=8cb402fb@gateway/web/cgi-irc/
00:31:27 Quit LambdaCalculus37 ("Fwump")
00:31:55linuxstbyomamathecablegu: What's the exact command you're typing, and what's the exact error message? (you can post to
00:32:43CIA-38New commit by MarcGuay (r21105): Set up the c100 for dual-booting.
00:33:42yomamathecablegukugel: how do I find out if I can write to /tmp?
00:34:56yomamathecablegulinuxstb: I typed RBDEV_PREFIX=/path/to/install /Users/Guess/Desktop/, and it said permission denied.
00:35:24linuxstbYou need to replace "/path/to/install" with the name of a directory you can write to
00:35:36 Quit stripwax ("")
00:35:54 Join joa [0] (
00:36:00joanow im getting confused
00:36:09joatransferred a load of music to h10
00:36:14 Join itcheg [0] (i=62db4c46@gateway/web/ajax/
00:36:16joagoto files in rockbox, nothing there.
00:36:27joai have a look at free memory info and it said 4gig is free
00:36:29linuxstbyomamathecablegu: e.g. I use $HOME/rbdev
00:36:38joathere should barely be a megabyte free, any ideas?
00:37:15linuxstbThe free memory info may not be accurate - it depends on whether your computer updates it.
00:37:32joathey files should appear in files though right?
00:37:36joawithout creating a db
00:37:50joai have a text dir in there and that's it
00:37:57linuxstbYes. Although they may be hidden - change the "show files" setting in Rockbox to "all files"
00:38:12joai think you're right linuxstb
00:38:16joathey were hidden
00:38:31pixelmaif I want to test a new bootloader on my c200, do I want the one in the flyspray task or build my own?
00:39:21TichodromaI want to run a lua file automatically on startup. I tried set_current_file("/.rockbox/lua/init.lua"); at the end of tree_gui_init(), and also tried rockbox_browse() in same place, but that didn't do it −− could somone put me straight?
00:39:43joawe have lift off, thanks linuxstb
00:40:06mcuelenaereTichodroma: look for AUTOROCK in main.c
00:40:09linuxstbTichodroma: You need to start the lua plugin, with the name of your file as a parameter.
00:40:42JdGordon|Tichodroma: you want plugin_start(APPS_DIR "/lua.rock", "/path/to/your.lua")
00:40:55JdGordon|before rockbox_broswe()
00:41:15JdGordon|maybe thats something we should tihnk about doing in svn?
00:41:16yomamathecablegulinuxstb: so what would the whole command be?
00:41:47JdGordon|seen as the disk is spining anyway, checking for the startup.lua file (or whatever) wont be a big slow down
00:43:23 Quit bmbl ("Woah!")
00:43:30*JdGordon| hasnt looked at lua so has no idea what sort of things the plugins can do, so that might not be really useful
00:44:39linuxstbyomamathecablegu: Do you have in ~/Desktop/?
00:44:57 Quit flydutch ("/* empty */")
00:45:37yomamathecablegulinuxstb: Yupp ^_^
00:45:49 Nick fxb is now known as fxb__ (
00:45:53linuxstbyomamathecablegu: And "permission denied" is the only message you see?
00:47:34TichodromaRight, thanks all. My first use was going to be to delete the empty dirs like "iPod_Control" and "Notes" and "Calendar" that the apple firmware insists on creating whenever I accidentally boot into it.
00:47:39 Quit itcheg (" ajax IRC Client")
00:48:54CIA-38New commit by mcuelenaere (r21106): Lua: add support for os library
00:48:57linuxstbTichodroma: It would be easier to just delete the Apple firmware...
00:50:01TichodromaYeah OK fair enough :) Seeing as you bring it up, what's the simplest way to do that?
00:50:38linuxstbUse "ipodpatcher -wf bootloader-xyz.ipod" to replace the Apple firmware with the Rockbox bootloader.
00:50:40kugelUnhelpful: ping
00:51:12linuxstbTichodroma: Some documentation on ipodpatcher is here -
00:53:09TichodromaCheers. What do you think of having menu items that run context-dependent lua scripts with standardised names? E.g. <run local script> looks for a file called local.lua in the current directory, and e.g. <run artist script> looks for <current artist name>.lua, a bit like searching for album art files. I'm thinking e.g. of having the latter load a playlist of similar music to current artist.
00:55:13Mikachucouldn't you just have <current artist name>.m3u in the first place?
00:56:54JdGordon|or use the db...
00:57:18 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
00:58:12yomamathecablegulinuxstb: Yep. That's all it says :/
00:59:36Tichodroma@Mikachu: <artistname>.m3u will exist already on disk; but the issue is to find a convenient way to call it up on-the-fly, say while you're listening to <artistname>
00:59:59Mikachu<artistname>-similar.m3u then?
01:00:31Mikachuoh well, i'm going to bed :)
01:00:40kugelpictureflow writes to plugin_tsr_ext, which holds the cb for resident functions
01:00:50linuxstbyomamathecablegu: Then you need to make executable - "chmod +x ~/Desktop/" (or just download the Rockbox source code properly)
01:00:50TichodromaMikachu: But where does it reside, and how do you get there conveniently?
01:00:59MikachuTichodroma: where is your .lua script?
01:01:01kugelis it possible that pictureflow overflows the stack (it only seems to happen on the sim)?
01:02:25CIA-38New commit by mcuelenaere (r21107): Fix yellow & red
01:02:26TichodromaYou mean you want to see the code, or where would I put it on disk?
01:02:38Mikachuthe latter, but i'm going to bed :)
01:04:03 Join Thundercloud [0] (
01:05:45 Join hd [0] (
01:07:44TichodromaJdGordon: I'm creating various types of playlists, for (potentially) all the artists on disk. I'm also creating .link files to similar music. The total size of all the playlists and link files is 10s of Mb. I'm concerned that trying to do this in tagnavi will be a bad idea. Otoh it works very well with .link and .m3u files on disk.
01:08:45TichodromaMikachu: It could be anywhere, e.g. /.rockbox/lua/<artistname>.lua
01:08:48 Quit MarcGuay_ ("ChatZilla 0.9.84 [Firefox 3.0.10/2009042316]")
01:09:45funmanlinuxstb: : i now build one dualboot.{h,c} which contains all dualboot.S and nrv2e_d8.S
01:10:38 Quit BHSPitMonkey (Remote closed the connection)
01:11:08mcuelenaereI suppose rb->screens[SCREEN_*]->*() is preferred over rb->lcd_*() calls?
01:12:26JdGordon|screens[i]->*() even
01:12:37 Join MarcGuay [0] (
01:12:41kugeluhm, is accessing an array with an negative index possible/defined in C?
01:12:41JdGordon|unless of course you really really dont ant multiscreen support
01:13:02mcuelenaereI think multiscreen support under Lua would be nice :)
01:13:25JdGordon|kugel: thats a good way to crash an app...
01:13:40kugelit's done in pictureflow....
01:13:44pixelmaTichodroma: I'd probably (mis)use some tag that you can use for database queries for that. E.g. stuff things like how you would call your playlists in the comment tags (even as a list) and then you can search in the comment tags for xyzplaylist
01:14:00kugelI know it works on some high level languages, but I don't know about C
01:14:20 Quit HellDragon (Success)
01:14:30kugeland pf crashes in related code, apparently writing to a completely unrelated memory location
01:14:39 Quit mirak ("Ex-Chat")
01:14:39JdGordon|that would explain it
01:14:40funmankugel: do you want to access it from the end ? [-5] = [last_item - 5] ?
01:15:02UnhelpfulPF crash report?
01:15:04kugelfunman: I don't want anything :)
01:15:57Unhelpfulthere's an array in PF which grows from the end of available space. a pointer at the last available address is used, and is indexed with a negative value.
01:16:22Unhelpfulactually, i think both the album index and track index are built this way.
01:17:01kugelUnhelpful: create_abum_art index is problematic for me
01:17:02Tichodromalinuxstb: I can see a file 'rockbox.ipod' but no bootloader*. So should I try 'sudo ipodpatcher -wf rockbox.ipod'?
01:17:24kugelI get a watch point trigger for plugin_tsr_exit() (from plugin.c) when album_count is accessed
01:17:52funmanTichodroma: did you build a _bootloader_ ? (not _normal_ build)
01:18:04Unhelpfuldo you mean create_albumart_cache, or create_album_index?
01:18:16kugelthe latter
01:18:27funmandidn't we speak about plugin_tsr_exit() being modified some days ago?
01:19:00 Quit mcuelenaere ("Gnight")
01:19:18 Quit Thundercloud (Remote closed the connection)
01:20:04kugelUnhelpful: why is that done this way? with the negative index
01:20:10Unhelpfuli seem to recall a user crash report for PF in sim that crashed at some address outside any rockbox code. the suggestion was made that plugin_tsr_exit was being modified, and that the next plugin start was calling it since it was non-NULL.
01:20:44 Join {-phoenix-} [0] (
01:20:59kugelcan it be 64bit related?
01:21:10kugelI see some long vs int mixes there
01:21:18soapHmm, Battery_bench is failing to write any records while charging via AC. Does fine when charging via USB. This is on my iPod Nano and on my iPod Video.
01:21:28funmando we use 64 bits platforms ? :)
01:21:35kugelUnhelpful: it tries to call 0x44 according to gdb
01:21:35JdGordon|in the sim yes
01:21:45funmankugel: i believe sizeof(long) == sizeof(int) on 32bits, and sizeof(long long) == 64
01:21:47Tichodromafunman: No. Am doing so now. What's the difference between replacing the apple firmware with the rockbox bootloader versus the rockbox.ipod ?
01:22:18funmanTichodroma: the rockbox bootloader will load rockbox.ipod from the .rockbox directory placed on the player (use make zip in normal build and extract this .zip on the player)
01:22:28kugelfunman: that's correct
01:22:30Unhelpfulkugel: it's the easiest way to have an array of index entries (of fixed length) and a list of variable-length strings built at the same time. the index list grows down from the top of the available buffer, the strings grow up from the bottom, and it checks before allocating each to make certain that the two tables don't run into each other - that's a failure case, since there wasn't enough memory to finish the job.
01:22:58 Join intrados [0] (
01:23:47 Quit JdGordon| (" ajax IRC Client")
01:24:46Unhelpfulwhat exactly do you mean about "mixing" long/int? is an array of integers being accessed via a long* at some point?
01:27:48kugelUnhelpful: the buf* vars are casted to long in the beginning
01:27:52kugellong* rather
01:28:34kugeland later: buf = l + buf (l is an unsinged int)
01:28:41CIA-38New commit by FlynDice (r21108): Add myself to docs/COMMITTERS and move on to the next chapter of Advanced SVN for Dummies!!
01:28:43funmanare the osx binaries universal binaries?
01:28:58funmanFlynDice: congrats :)
01:29:04kugelFlynDice: welcome :)
01:29:15Unhelpfulcongrats FlynDice
01:29:27FlynDice so far I haven't screwed up, right...
01:29:38kugelwait for the build table :)
01:29:58kugeldocs/COMMITTERS is a tricky one
01:30:08kugelit's one of the few non-utf8 files
01:30:26kugelit's iso-8859-1 or so because of the front page
01:30:31FlynDicetook me 3 text editers to get there!
01:31:56kugelfunman's name (looking at the viewvc diff) below yours seems to indicate that you didn't mess it up
01:32:17Unhelpfulkugel: those are both valid - uniqbuf takes the last UNIQBUF_SIZE * sizeof(long) bytes of the plugin buffer. the other part you mention is the album names cache, which grows from the start of the plugin buffer, and advances the plugin buffer pointer by the length of the string added, and decreases the remaining buffer size by the same amount.
01:32:54kugelUnhelpful: well, I don't understand this code there, but it's apparently broken in some environments
01:33:54Unhelpfulthe only issue i see with either of those is if the tagcache treats the uniqbuf as something *other* than an array of longs after it gets it.
01:34:53Tichodromafunman: cheers. no more apple.
01:35:53kugelUnhelpful: it casts the buffer to unsigned long* so that seems not the problem
01:36:36Unhelpfulhas anybody been able to trap if/where the tsr_exit pointer actually is modified?
01:36:41***Saving seen data "./dancer.seen"
01:36:50 Join fdinel [0] (
01:37:28 Join JdGordon| [0] (i=63cbcd3c@rockbox/developer/JdGordon)
01:37:31 Quit {phoenix} (Read error: 110 (Connection timed out))
01:38:05kugelUnhelpful: at album_count++
01:38:35Unhelpfulbut that doesn't even make sense - album_count is on the stack.
01:38:49kugelit's not
01:39:33 Quit matsl (Read error: 110 (Connection timed out))
01:39:52kugelI thought that too first (hence I asked if it was possible to stack overflow), but album_count is a glibal
01:40:35Unhelpfulah! even so... it being a global still doesn't explain how incrementing it smashes something in core.
01:41:49yomamathecableguAnyone know how I can install gcc on Mac OS?
01:41:49kugelUnhelpful: btw, I got the first watch point trigger at album_count++, not at album_count = 0 at the beginning of create_album_index()
01:41:58funmanyomamathecablegu: get the iphone sdk
01:42:32Unhelpfulkugel: now *that* really makes no sense :/
01:42:33 Quit hd (Client Quit)
01:43:05kugellet me double check
01:44:46kugelno, it's like that
01:45:22kugelUnhelpful: pf (in fact, any plugin) crashes after exiting pf for the first time
01:45:30 Join itcheg [0] (i=62db4c46@gateway/web/ajax/
01:46:02Unhelpfulkugel: i've seen that report. i've tried to reproduce it. that *does not happen* when i run it, on target or on sim.
01:46:02 Quit Tichodroma ("CGI:IRC (EOF)")
01:46:28kugelalso, I only get 1 watch point trigger, although album_art++ is in a loop. is this a gdb thing or indicates that it's only messed up in the last run?
01:47:59kugelcould the 0x44 value of plugin_tsr_exit mean anything?
01:48:01Unhelpfuli would guess that gdb is doing something wrong, because i don't really see how one particular loop through would lead to an corrupting data somewhere else while setting the value of a static global
01:48:41kugelit crashes at calling plugin_tsr_exit
01:48:42 Join HellDragon [0] (
01:48:53kugelbecause that's non-NULL after exiting pf
01:49:57funmanmy modified mkamsboot produces bit exact output with the old version :-)
01:51:05 Quit yomamathecablegu ("CGI:IRC (EOF)")
01:51:35kugelUnhelpful: I too thought that gdb is entirely correct, hence I asked weither the pointer math in there is entirely correct
01:51:56Unhelpfulto the best of my knowledge, those operations are valid
01:52:25kugelcouldn't album[-album_count] get below the plugin buffer?
01:52:31 Quit faemir ("Leaving")
01:52:32Unhelpfulhrm, what about dumping the code when the trap happens and making sure it hasn't been overwritten? :/
01:53:27kugelmeh, the plugin business is too different anyway for sims compared to target's reality
01:55:12Unhelpfulkugel: i don't see how it would go below the plugin buffer before it goes below the string table, but you could try adding a test to make sure buf_size is >0 right after sizeof(struct album_data) is subtracted from it.
02:00:29Unhelpfulyou'd have to somehow exhaust the plugin buffer before reading in sizeof(struct album_data) album titles, and those would all have to be ""
02:01:29 Quit JdGordon| (" ajax IRC Client")
02:02:15Unhelpfulhrm, does anything change if you replace plugin_get_buffer with plugin_get_audio_buffer? the audio buffer is both larger and at a different address...
02:03:57kugelUnhelpful: yes, indeed
02:04:11kugelplugin_get_audio_buffer doesn't crash
02:04:45 Join r0b- [0] (n=nnscript@
02:05:01r0b-is this MP3 player on the Rockbox compatibility
02:05:05funmankugel: i don't remember : when flashing a fuze bootloader, can it be named fuze.bin or only fuzea.bin?
02:05:25Unhelpfulis everybody who can reproduce this crash using x86-64?
02:05:45funmanr0b-: if it is a v1, yes (you have to switch it on to see the firmware version)
02:06:06r0b-well i havent ordered it yet
02:06:07Unhelpfulr0b-: look at the list on the front page.
02:06:15r0b-the picture sas its V1
02:06:19r0b-but it may be V2
02:06:20kugelUnhelpful: buf size is <0 already at the beginning
02:06:27r0b-if its V2 i wont need Rockbox :(
02:06:45*kugel did printf(%d, buf_size) after the substraction, and the first print was already negative
02:07:09kugelfunman: I think the suffix cannot be ommitted
02:07:22Unhelpfulkugel: i think i see the problem :)
02:07:29kugelbut you can use fuzet.bin or fuzef.bin or other known suffixes
02:07:37r0b-kinda sucks Rockbox wont work on the Ipod shuffle
02:07:41Unhelpfulthat uniqbuf is half of the plugin buffer if sizeof(long) == 4 ;)
02:08:30Unhelpfuldo we really need a uniqbuf of 64K longs?
02:09:07Unhelpfuli'm assuming amd64 longs are 8 bytes?
02:09:54Unhelpfulis there an AMS version of c200? if yes, then *maybe* that device will work. if no, then *yes*. :)
02:10:27funmanthere is, without LCD
02:11:48Unhelpfulso, the one r0b- linked is a pre-AMS c200, then :)
02:11:54kugelUnhelpful: and what would solve it?
02:12:12r0b-ok that one has to be V1
02:12:18r0b-Newegg sells the V2
02:12:20Unhelpfulkugel: reduce UNIQBUF_SIZE
02:12:21r0b-i need v1!!!!!
02:12:25pixelmafunman: without LCD?
02:12:50funmanpixelma: without LCD + "driver in the rockbox port for it" was implicit :)
02:13:17Lloreanfunman: The c200 has an LCD though. Implicit doesn't always work, so try for explicit please.
02:13:28kugelUnhelpful: wouldn't that have other impact?
02:13:28pixelmawhat I thought but better to say to avoid misunderstandings...
02:13:29LloreanThere's just no AMS c200 LCD driver, right?
02:14:12funmansorry for the confusion, the c200v2 isn't actively developed, and at the moment there is only support for dual booting and some buttons can be read. The LCD screen still can not be accessed.
02:15:11r0b-well if rockbox failed to install will it ruin the MP3 player?
02:15:17Unhelpfulkugel: i can't really say. isn't the album list already guaranteed unique? do we need a unique buffer at all?
02:16:03Lloreanr0b-: If you're using RButil (the suggested method) it will refuse to install if it's not a supported player
02:16:03kugelI think it is, yes
02:16:20r0b-ok sweet
02:16:25Unhelpfulkugel: let me see what the DB browser's album list does...
02:16:53r0b-if there a preview of the GUI for Rockbox
02:17:00kugelUnhelpful: tag album is in the uniq tags
02:17:09Lloreanr0b-: There are pictures in the manual, plus the themes page
02:17:30 Join Blue_Dude [0] (
02:17:41r0b-thats pretty sweet i really cant wait to try it :)
02:17:50Unhelpfulkugel: it is. the DB browser appears to use tagcache_search_set_uniqbuf unconditionally, but i suspect we can do away with it entirely if looking at the album list.
02:17:52kugelthe db browser probably collapses albums if they have the same name into 1 node
02:18:00pixelmafunman: is there anything known about the LCD used in the v2 c200, maybe the same as in the v1 (though it would be quite disappointing if Sandisk stayed with that choice)
02:18:38Blue_DudeAnyone know how to read errno from the file.c functions?
02:18:59funmanpixelma: I think bertrik has some information (he mentioned a known LCD controller on IRC) and was expecting someone to bring a c200v2 at devcon
02:19:07UnhelpfulBlue_Dude: i don't think we have an errno in rockbox...
02:19:31Blue_DudeHm. Are they just leftovers from the GNU code?
02:19:31kugelso, I was right. 64bit issue and buffer underrun (if that's the correct term for this case) :) I was more guessing than knowing though
02:19:44scorche|shBlue_Dude: what GNU code>
02:19:59kugelUnhelpful: half of the plugin buffer is only used for the album list?
02:20:04Blue_DudeThe file routines seem based on GNU file code.
02:20:45Unhelpfulkugel: no, half of the plugin buffer is temporarily used for the uniqbuf - that's freed after the album list is built.
02:21:08Unhelpfulthe album list takes however much space it takes for the strings, plus sizeof(struct album_data) per album
02:21:34kugelI see. I'm still not really used to dynamic memory management in rockbox :S
02:22:02Unhelpfuli'm not sure what else you'd call the audio buffer ;)
02:22:26kugelthat's not dynamic. that's "take whatever is left" :P
02:23:29Unhelpfulit's not really so different. pictureflow grabs the whole plugin buffer, uses as much as it needs for the album list, then turns the rest over to buflib for the cover cache.
02:23:47kugelbut there's...freeing... :)
02:23:50Blue_DudeAnyway, errno is set in the file.c functions, but there's no way to read it. There's some redundancy there because the file descriptor is negative if there's an error, but I didn't know if there was a way to get more specific by reading errno.
02:24:19Unhelpfulkugel: if you mean of the uniqbuf, it's not exactly "freeing"... we just temporarily pretend we don't have the last N bytes ;)
02:24:52Unhelpfulmy thinking is that if we actually have non-unique albums with the same name, we *should* present the name twice, anyway - they're probably different albums, right?
02:25:17kugelonly if that's also done in the core
02:25:57funmanlinuxstb: FS #10253 - mkamsboot version 1.0
02:25:59 Part toffe82
02:26:16 Quit efyx_ (Remote closed the connection)
02:26:19kugelespecially if we plan on integrating pictureflow in to the db browser we should stick to what the core db browser does
02:27:08Unhelpfulthen we need to ask somebody who knows what tagcache "should" do.
02:27:48kugelthe album list is 1 file which AFAIK cannot contain a album twice
02:28:03kugelthe master index stores the offset to the name for a track
02:28:51kugelyou couldn't retrieve a result without passing another parameter if it wasn't unique
02:29:37Unhelpfulso uniqbuf is completely unneeded with tag_album?
02:30:09kugelit seems
02:30:18 Quit funman ("leaving")
02:30:23kugelthe guy that ported pf wasn't a rockbox database guru
02:30:35Unhelpfuli'm not, either!
02:30:41kugelasking Slasheri can't hurt though, either
02:32:15Unhelpfulkugel: give this a try:
02:32:34Unhelpfuli finally managed to reproduce the failure on my system by doubling UNIQBUF_SIZE
02:34:09kugelare you on 32bit?
02:42:47kugelUnhelpful: btw, I'm already using different tracks with different artists under the same album
02:42:56kugelsame album artist though
02:43:44 Quit moos ("Rockbox rules the DAP world")
02:44:48 Quit HellDragon (Read error: 54 (Connection reset by peer))
02:44:51 Join lee321987 [0] (
02:44:52Unhelpfulkugel: yes, i have those as well. i think that tracks with same album title, and a different (if albumartist ? albumartist : artist) should be considered different albums.
02:45:02 Join HellDragon [0] ( doesn'
02:45:09lee321987oops doesn't download the RB source, it just downloads tools you need to compile RB - right?
02:47:42kugelrockbox doesn't recognize the changes made by easytag ?!
02:48:12Unhelpfuli don't think rockbox rescans tags when an existing file is modified, does it?
02:50:29 Join chandoo [0] (
02:50:35 Quit HellDragon (Client Quit)
02:50:48 Join HellDragon [0] (
02:50:50Blue_DudeAre there memory limitations with any targets that would be affected by allocating an additional PCM buffer? Looking at either a 8kb or 16kb aux buffer in addition to the crossfade and voice buffers.
02:51:13LloreanFor what?
02:51:23Blue_DudeDSP mixing.
02:51:24kugelUnhelpful: the core doesn't differentiate
02:51:34kugelI just did a test with some random file
02:51:36Blue_DudeWell, DSP operations.
02:52:09Unhelpful"one name" in core always == "one album", then?
02:53:01Unhelpfuldoes pictureflow behave any differently than core, with that patch?
02:54:07 Quit froggyman ("CGI:IRC (Ping timeout)")
02:54:27Blue_DudeSo are there any targets that would hiccup if I tried to set aside 16kb?
02:54:56 Quit perrikwp (" ajax IRC Client")
02:55:26 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
02:55:43LloreanBlue_Dude: There aren't any, really, but you should try to avoid it if at all possible
02:56:20kugelUnhelpful: not even without that patch
02:57:33Unhelpfulkugel: well, that patch removes the use of a uniqbuf from PF entirely. *without* that patch it's building the album list exactly like the core DB browser, ie, tagtree always sets up a uniqbuf regardless of which tag is used
02:58:10 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/ajax/
02:58:21Blue_DudeI can't figure another way around it though. I need to keep a certain number of samples back before they're committed. The "easy" way to do it is to set aside two PCM buffer chunks, one active and one standby. The active one would have been operated on already, using a portion of the standby buffer. Once the active one is read out, then the standby buffer becomes active and the standby...
02:58:23Blue_Dude...buffer is filled.
02:59:04 Quit kugel (Read error: 60 (Operation timed out))
02:59:47Blue_DudeAnd the two would swap roles. They would be fed to the crossfade or voice buffers if they're active, or assigned to the PCM buffer for playback if not.
03:00:40Blue_DudeIf I keep them in even chunks then the code mod would be pretty minimal.
03:02:22Blue_DudeAny ideas?
03:04:55Unhelpfulis this still for a DRC DSP?
03:05:01 Join kugel [0] (n=kugel@rockbox/developer/kugel)
03:05:11kugelUnhelpful: is that done in the tagtree or tagache?
03:05:19kugelthe tagtree doesn't know whether album artist was filled or not
03:05:31kugelit gets filled with artist during initialisation
03:05:39Blue_DudeWell, it's not as ambitious as it was. I'm writing a limiter now, not a DRC.
03:06:28Blue_DudeThe good news is that the buffer will work for any DSP routine that requires an attack time.
03:06:58Unhelpfuli still don't see what exactly we want a limiter *for*. i think that the idea of pre-amp + limiter as a DRC will make for a rather poor DRC.
03:08:20Unhelpfulkugel: tagtree doesn't even look at artist when building the album list, as far as i can tell - but it always sets up a uniqbuf regardless of which tag it uses. i suspect the results for tag_album will not be any different with a uniqbuf than without.
03:09:21Blue_DudeIt's a compromise. It's something I came up with when fooling around with filters in a wav studio program. I actually liked the results better than the DRC filter. More volume, no clipping, minimal distortion.
03:10:13Blue_DudeBut with zero attack time, the initial peaks are crushed and the distortion is pretty bad.
03:10:31kugelUnhelpful: is the uniqbuf really used too?
03:11:33Unhelpfulkugel: it's allocated and passed to tagcache. i don't know if tagcache skips using it on unique tags...
03:11:53kugelseems not to be different behavior with the patch
03:12:35Unhelpfuleven with your test "non-unique" album?
03:13:47 Quit kachna (Read error: 113 (No route to host))
03:13:54 Quit lee321987 ("ChatZilla 0.9.84 [Firefox 3.0.10/2009042316]")
03:14:58Blue_DudeAnyway, about a 10ms attack time preserves the waveform and reduces distortion to an acceptable level. I've done a number of comparisons and this one seems to work best.
03:16:11Unhelpfulthat's... 441 samples?
03:17:31Blue_DudeYeah, about. My concern isn't so much with the number of samples as making it even across chunk boundaries. The chunk size is 2048, for 8192 bytes. Active and lookahead buffers make that 16kb.
03:18:04saratogai think you'll have trouble getting people to agree to give you 16KB in SVN
03:18:08saratogathats an awful lot of memory
03:18:29kugelUnhelpful: the album is still with the other album artist
03:19:14Blue_DudeYeah. If the data weren't committed in such large chunks it would be much simpler.
03:19:47saratogawhy does the size matter?
03:19:57UnhelpfulBlue_Dude: if we're working strictly in powers of two, you only need to "see" 512 samples past what's about to be committed.
03:20:41kugelUnhelpful: add_uniqbuf seems to handle unique tags and the uniq buffer
03:20:50Unhelpfulkugel: i'll go ahead and commit - it also makes the album title list build much less likely to fail due to out-of-memory.
03:21:12Blue_DudeThe problem as I see it is that the buffer doesn't receive small doses. It gets 2048 all at once. I can't process the whole thing, only about 1500 or so.
03:21:34saratogawhy does the size of the buffer matter?
03:21:38Blue_DudeHang on, maybe a 512 element array for lookahead?
03:22:00 Join cmwslw [0] (n=cmwslw@
03:22:08 Join cmwslw_ [0] (
03:22:14 Quit cmwslw (Read error: 104 (Connection reset by peer))
03:22:16 Quit cmwslw_ (Remote closed the connection)
03:22:22 Join cmwslw [0] (
03:22:38Blue_DudeIt only matters to the extent that the buffer appears to be served in largish chunks instead of individual samples.
03:22:52UnhelpfulBlue_Dude: 2KiB of extra buffer for a lookahead sounds much better than 16KiB
03:23:01 Part cmwslw ("Ex-Chat")
03:24:37kugelUnhelpful: hm, the code looks like it would use the uniqbuf for tag_album
03:24:45Blue_DudeWell the main problem is making the current chunk linger without being committed until it's processed. That means it has to stay dormant while the *next* chunk is filled. Which brings us right back to 16 KB.
03:26:17Blue_DudeWhat about reducing the chunk size? Would that affect computing overhead to a large extent?
03:26:58saratogawhy does the size of the current chunk matter?
03:26:59Unhelpfulkugel: yes, i'm not sure if this is safe. also, using the uniqbuf looks quite expensive, doesn't it?
03:27:20kugelUnhelpful: however, the comment says that the uniq buf is used to prevent duplicate entries
03:27:20 Quit itcheg (" ajax IRC Client")
03:27:45Blue_DudeIt matters because I need to set aside two chunks before committing them. That's a lot of memory.
03:28:07kugelduplicate entries as in multiple results for the same album with the *same* content
03:28:35saratogaBlue_Duke: why does the size of data you're given dictate how many samples you need to set aside?
03:28:54saratogagenerally algorithms are independent of that
03:29:10kugelThat makes me think that the uniqbuf isn't even meant to make it possible to have an album show up more than 1 time, but the opposite!
03:30:03kugeladd_uniqbuf returns false for tag_album if the same album was already found in the buffer
03:30:19l403saratoga county
03:30:28l403I used to know where that is
03:30:30Blue_DudeI have to be work with the current buffer structure. Outgoing samples are buffered in chunks. I can't operate on an entire chunk at once because I need to see part of the next.
03:30:37 Quit dmb (Read error: 113 (No route to host))
03:30:38kugelwe might get duplicated entries in pf without the uniqbuf :(
03:30:57l403oh, sry, I'll shut up with my OT
03:31:03Unhelpfulkugel: right, but it identifies it by seek offset in the album index, not by string content. and if the same string is in the album index twice, uniqbuf won't catch it.
03:31:04Blue_DudeMaybe I'm missing something. That's why I'm here.
03:31:19saratogaBlue_Dude: why can't you just operate on 100 samples at a time?
03:31:33saratogabuffer 100 samples, process the previous 100, then iterate
03:31:43saratogathe buffer size shouldn't matter at all
03:32:39kugelthat proves my theory :)
03:33:01kugelnow the question is whether albums are added multiple times
03:33:32Unhelpfulkugel: even if and album name appears twice in the album index, uniqbuf won't catch that...
03:34:10Blue_DudeI certainly can. But at what point can I "see" that data? Where do I intercept it? I have to wait for the codecs (and the EQ, and the crossfeed, etc, etc) to do their thing before I touch it. I need to interject an algorithm after they're done but before they are overmixed with voice, beeps, keyclicks, etc.
03:36:03 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
03:36:46***Saving seen data "./dancer.seen"
03:36:49saratogaBlue_Dude: I don't understand
03:37:37kugelUnhelpful: it's only once in my album index file
03:37:46Unhelpfulkugel: i don't really think that the uniqbuf helps here at all. if the same string appears twice in the album index, uniqbuf will not catch it, and we won't hit the same *entry* in the album index twice because we are basically iterating over the album index.
03:38:15Blue_DudeAt what point can I intercept outbound samples before they are committed to playback? What is the data flow?
03:38:26kugelit helps, if it's only once in the index file (and it is)
03:38:36saratogaBlue_Dude: register a function in dsp.c as all the other effects do
03:39:41Unhelpfulkugel: but it doesn't help *in pictureflow* because we're just going from one entry in the album index to the next, and uniqbuf only tests the entry offset. we won't ever get the same offset twice, anyway.
03:39:59kugelbut I doubt we get duplicates in pf, as we're searching for album only, but who knows
03:40:22kugeli.e. what you said :D
03:40:29Unhelpfulkugel: we won't get any in PF that could be caught by using uniqbuf.
03:41:42Blue_DudeI haven't figured out how to hook into dsp.c yet.
03:41:56UnhelpfulBlue_Dude: then you probably need to start there.
03:42:11Blue_DudeTrying. But it's not exactly light reading.
03:43:48Blue_DudeIs there anyone who is versed in those routines and can point me in the right direction offline?
03:44:23saratogaBlue_Dude: just add your function right after the call to EQ
03:45:00saratogayou can probably even use the same basic function with just the internals changed around
03:45:35kugelUnhelpful: btw, I think get_albumart_for_index can be speed up
03:46:15kugelwhen using load to ram, we can use tagcache_fill_tags
03:46:49Unhelpfulkugel: i'd love to hear how - apparently it's using the bulk of the cache build time on archos.
03:46:50kugelthat wouldn't involve file i/o
03:48:21Unhelpfulwould tagcache_fill_tags still "work" without load-to-ram on?
03:49:03Blue_DudeOK, I'll step through the EQ routine. Thanks for the tip!
03:49:14Unhelpfuli have to get started on work here... i'll take a look at that shortly. :)
03:50:12kugelit returns false if it's not on, so we have to fall back then
03:50:46 Quit SirFunk__ (
03:50:46 Quit joa (
03:50:46 Quit fyre^OS (
03:50:46 Quit lymeca (
03:50:46 Quit jfc (
03:50:46 Quit tmzt (
03:50:46 Quit at0m (
03:50:46 Quit Ridayah (
03:50:46 Quit J-23 (
03:50:46 Quit Chex_ (
03:50:55*Unhelpful recalls hearing that ramcache or dircache used the plugin buffer, though?
03:51:02NJoinjoa [0] (
03:51:02NJoinSirFunk__ [0] (
03:51:02NJoinfyre^OS [0] (
03:51:02NJoinlymeca [0] (
03:51:02NJoinjfc [0] (
03:51:02NJointmzt [0] (
03:51:02NJoinRidayah [0] (
03:51:02NJoinJ-23 [0] (
03:51:02NJoinChex_ [0] (
03:51:07NJoinat0m [0] (
03:51:13kugelno, they don't
03:51:30kugelthey bufalloc at startup (hence you need to reboot if you enable it)
03:51:34 Quit Llorean (Read error: 104 (Connection reset by peer))
03:51:54 Join Llorean [0] (
03:54:53kugelthere's also a few other functions for ramcache
03:57:44kugelUnhelpful: it should give a noticeable speed up on hdd, you can try it on your beast
04:00:11kugelUnhelpful: btw, what happens for <track>.<ext> album art files?
04:01:22kugelit searches for the first track and uses that album art for this case? or does it search for each track of an album?
04:01:55kugelrb->tagcache_search_add_filter(&tcs, tag_album, album[slide_index].seek); probably means it only searches for 1 track per album (?)
04:02:27kugelas the comment directly above says :/
04:03:22 Join oldrocker [0] (
04:04:42Unhelpfulthat filter causes it to only find tracks for the given album. but it only examines one result - amiconn ran into trouble with that with having two copies of an album, one without AA.
04:06:03 Join jerl92 [0] (
04:06:38kugelwhy would anyone have that?
04:07:06kugelif I was to have aa for tracks, I'd surely have a fallback cover.jpg
04:07:48 Quit oldrocker (Client Quit)
04:08:45kugelI'm wondering whether pf should prioritise <track> AA lowest anyway
04:09:16Unhelpfulthere's no way to do that, though - unless we add a flag to the core AA search for it.
04:09:40kugeli.e. search_albumart_files() could have a album mode and a track mode
04:10:46kugelit could be another parameter, or a format directive in the search string
04:17:52Unhelpfuli think a tag to skip track-specific images will suffice. we probably don't want "track" covers at all in pictureflow, and we definitely don't want to search for one for each track if none are found for the album - that will be much, much too slow.
04:21:26 Join lee321987 [0] ( is this what I need to do to run it in Ubuntu? sudo sh
04:22:07kugelUnhelpful: but then we break the magic 4-parameter border
04:22:33 Join dmb [0] (n=dmb@unaffiliated/dmb)
04:22:47kugeland with a bit of goto'ing we can still get track AA without code duplication
04:24:05kugelWell, I'm not against a 5th parameter, I don't really care
04:24:08Unhelpfulkugel: meh, i don't see any reason to try track AA next. core doesn't want that, and pictureflow can't really use it.
04:24:40kugelcore doesn't want it, and won't do it
04:24:42Unhelpfuli don't like the idea of a magic value for size_string very much... it just seems a bit *too* hackish.
04:25:04kugelthat's what C++ people say about printf :D
04:27:06lee321987(Ubuntu) I saved (from a previous Ubuntu install) the folder "usr/local/arm-elf". Can I just copy these files into the new Ubuntu, and then run If not is there any way I can use these files to avoid more downloading?
04:27:30Unhelpfulkugel: the issue is that that's not a format string - it's a string value to insert literally in the search filename
04:27:42kugelI know
04:28:00kugelbut that's not set in stone
04:28:29Unhelpfulhrm... perhaps NULL should mean "no size string", and "" should mean "no track AA search"? :P
04:29:22 Join kugel_ [0] (
04:29:25soapHmm, Battery_bench is failing to write any records while charging via AC. Does fine when charging via USB. This is on my iPod Nano and on my iPod Video.
04:29:40kugel_Unhelpful: that breaks cover<size> though
04:30:50kugel_i.e. we don't want track, but still size string
04:31:37kugel_I don't really care about either way. But I think that searching for track AA is useless and wasteful for pf
04:31:41Unhelpfulkugel: "if size_string starts with :, disable track search, then advance size_string by one char"?
04:31:52kugel_at least with highest priority
04:32:14kugel_Unhelpful: that could work
04:32:21Unhelpfulbenefits: changes *nothing* else, very quick to implement.
04:33:03kugel_and faster for the 4-parameter victims
04:36:28r0b-how long until rockbox will work on the V2 c240
04:36:43kugel_We can't predict that
04:37:08kugel_but the odds are bad because no-one is working on this target
04:39:17kugel_Unhelpful: are you going to try the fill_tags thing?
04:39:36kugel_if not, I have a HDD target now too (finally :) ), so I could test it too
04:41:12 Quit kugel_ ("exit(0);")
04:41:48 Quit miepchen^schla (Read error: 101 (Network is unreachable))
04:43:25 Quit lee321987 ("ChatZilla 0.9.84 [Firefox 3.0.10/2009042316]")
04:44:02Unhelpfulkugel: i can see no reason *not* to use fill_tags if possible, but it might be worth benchmarking it... my beast has plenty of albums on it, shall i throw together a benchmark for it?
04:44:06 Quit jerl92 ()
04:44:21CIA-38New commit by unhelpful (r21109): Remove use of uniqbuf in PictureFlow, preventing crash on 64-bit sim and freeing half of the plugin buffer on targets.
04:45:53 Quit kugel (Read error: 110 (Connection timed out))
04:51:58 Quit Blue_Dude ("ChatZilla 0.9.84 [Firefox 3.0.10/2009042316]")
04:57:29 Join Lss [0] (
05:03:49 Quit __lifeless (Read error: 101 (Network is unreachable))
05:10:48 Join pimp_3000 [0] (n=4a245322@gateway/web/cgi-irc/
05:11:34 Quit perrikwp (" ajax IRC Client")
05:11:43pimp_3000hey can someone help i have rockbox installed on my ipod nano and it wont get rid of a music folder it says its courrpted
05:11:47 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
05:12:10scorche|shdisk check?
05:12:47pimp_3000tryied it
05:19:38LloreanWhat tool did you use?
05:20:07Unhelpfulkugel: tagcache_fill_tags also needs to be exported for that.
05:25:41JdGordonanyone have a clever/hacky idea on how to handle the wps clearing the whole screen on track change? the problem here being that the screen will have the UI in some viewport which the wps shouldnt (but will) be updating....
05:25:47JdGordons updating/clearing
05:26:15 Quit l403 (No route to host)
05:27:55 Quit pimp_3000 ("CGI:IRC")
05:28:44*JdGordon is even willing to entertain not so clever... even out right bad ideas here....
05:28:56UnhelpfulJdGordon: i'd think you'll have to clear 4 rectangles surronding the UI viewport...
05:29:08JdGordontried that
05:29:40Unhelpfuloh... what exactly went wrong with that?
05:29:49JdGordonstupid WPS clears all disabled viewports also, and if its setup correctly, the UI vp is disabled
05:30:05 Join JohnTeddy [0] (i=unstable@tor/regular/sid)
05:30:24Unhelpfulso, in the "clear all disabled viewports" code, you'll need to explicitly skip the UI vp, right?
05:30:26JohnTeddyWhat is the best device to buy for rockbox, flash based. That cost less than $100? Something like 16 gigs or more. (with expansion if necessary)
05:31:19JdGordonUnhelpful: yeah, that isnt so reliable though... although it should work
05:31:27*JdGordon was hoping for a cleaner solution
05:33:27*JdGordon tries this again...
05:35:31 Quit FlynDice (Remote closed the connection)
05:36:47***Saving seen data "./dancer.seen"
05:37:58soapJohnTeddy, Sansa E280 (version 1 of the hardware) with an 8GB MicroSDHC card is the only target which fits your demands, price not withstanding.
05:38:45soapOr, on the cheap I would buy an iPod Mini from ebay with a broken HDD and replace it quick and easy with a 16GB CF card.
05:39:48 Join trisiak [0] (
05:40:42 Quit {-phoenix-} (Read error: 104 (Connection reset by peer))
05:41:13soapI guess there is also the Sandisk C2x0 series, I don't know why I never think of that first.
05:42:00JohnTeddysoap: You know a reliable place to buy the e280 v1?
05:42:14soapebay, craigslist.
05:42:27JohnTeddysoap: Which would you get, Sansa e280, or Sandisk C2x0?
05:42:59JdGordonarg, almost works... stupid statusbar fiddles with things
05:43:13UnhelpfulJdGordon: that's closer, though. :)
05:43:15soap_I_ would get the E2x0 series, as I like the larger size. The C2x0 series is too small for my hooves. That said they are nearly identical hardware.
05:43:19Unhelpfulfroobi sometimes
05:43:38scorche|shthey havent had them for a while now
05:43:43JdGordonUnhelpful: yeah :) this monstrosaty is back in the real of possibility!
05:46:45JdGordonand... working :)
05:47:56 Quit Horscht ("Verlassend")
05:49:56 Join FlynDice [0] (
05:50:31JdGordondoes the standard progressbar have update problems while seeking in the wps in svn?
05:51:27 Join Alex0008 [0] (
05:52:08Alex0008I just ran in Ubuntu. When it finished, it said to "Make your PATH include :/usr/local/arm-elf/bin". Is that something I have to do in Ubuntu?
05:52:42Alex0008what command effects PATH?
05:53:28Unhelpfulnone, it's an environment variable.
05:53:59JdGordonthats not true...
05:54:17JdGordonPATH=blaa is a command which effects PATH
05:55:12*Unhelpful disagrees with that as a "command" but suggests that the best way would be to use whatever shell profile files your system reads
05:56:43Alex0008"man path" and "path −−help" yeild nothing. If I just run "path=:/usr/local/arm-elf/bin" will the things already in PATH stay there?
05:57:18cool_walking_nope, they won't
05:58:04UnhelpfulAlex0008: environment variables are case-sensitive, and you're not "running" anything. what you want is more along the lines of this: PATH="/usr/local/arm-elf/bin:${PATH}"
05:58:09cool_walking_Also, it's case-sensitive, so it needs to be uppercase.
05:59:05Alex0008that way will keep the old variables in there?
05:59:47cool_walking_notice the value you're assigning to PATH includes "${PATH}" ?
06:00:06Alex0008ahhhh. Thanks.
06:00:38cool_walking_That won't be permanent, however. It'll be lost when you close the terminal.
06:01:01Alex0008wait what?
06:01:06cool_walking_If you want to keep it, add the line to the end of your ~/.bashrc file.
06:09:51 Join flx_ [0] (
06:10:06 Quit flux (Read error: 104 (Connection reset by peer))
06:10:33 Quit Alex0008 ("ChatZilla 0.9.84 [Firefox 3.0.10/2009042316]")
06:14:58CIA-38New commit by unhelpful (r21110): Suggested by Thomas Martitz, use tagcache_fill_tags to speed up PictureFlow album art search a bit.
06:20:08 Quit chandoo ("Leaving")
06:34:00 Join AndyIL [0] (i=AndyI@
06:34:09CIA-38New commit by unhelpful (r21111): Suggested by Thomas Martitz, don't check for track-specific files when searching for covers in PictureFlow.
06:46:58 Quit AndyI (Read error: 110 (Connection timed out))
06:50:10 Join itcheg [0] (i=62db4c46@gateway/web/ajax/
06:55:44 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
06:57:36lucentJohnTeddy: Sandisk Sansa Fuze does it too
06:57:49lucentI got my 8gb v1 Fuze for like $59
07:07:28 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
07:18:29 Nick fxb__ is now known as fxb (
07:19:24 Quit itcheg (" ajax IRC Client")
07:25:18 Quit advcomp2019 (Read error: 110 (Connection timed out))
07:36:02 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
07:36:50***Saving seen data "./dancer.seen"
07:38:31 Quit jmillikin (Read error: 110 (Connection timed out))
07:38:41 Quit DataGhost (Read error: 104 (Connection reset by peer))
07:48:31 Join nibbler [0] (
07:50:27 Join DataGhost [0] (i=dataghos@unaffiliated/dataghost)
07:58:57 Nick fxb is now known as fxb__ (
07:59:00 Join l403 [0] (n=l@
07:59:27 Join bertrik [0] (n=bertrik@
08:04:52 Join LinusN [0] (n=linus@gateway/web/cgi-irc/
08:08:37 Join ender` [0] (
08:12:06 Join Alex0008 [0] (
08:12:15 Join evilnick_home1 [0] (
08:13:08Alex0008(Ubuntu) Does anyone know why 'configure' would result in "permission denied"? (I tried it from a "sudo bash" too).
08:14:26 Quit perrikwp (" ajax IRC Client")
08:14:41 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
08:16:32Unhelpfulyou should not need to use superuser permissions to run tools/configure. can you copy and paste *exactly* what you typed, and the results, to pastebin?
08:17:39 Quit l403 (Read error: 113 (No route to host))
08:25:40Unhelpfulhrm, any thoughts on how best to word-wrap text? almost seems like i'd have to split it into words myself and use lcd_getstringsize...
08:26:46Alex0008it's actually building now, so I can't copy it, but I screencapped it: From my "build" dir I typed "../tools/configure" the next line is this "bash: ../tools/configure: Permission denied" (all quotes added here). I got around it: I went to the 'configure' file in Nautilus, and in the files' properties put a check next to "allow this file to be run as an executable". I may have changed ...
08:26:47Alex0008...a permission or two, but that shouldn't change anything - right? Anyway I ran this command "chmod 777 * -R" from the rockbox dir −− I got the idea from the page on the Rockbox VMware image −− that command is okay to use in Ubuntu, I guess.
08:29:37 Quit evilnick_home (Read error: 110 (Connection timed out))
08:30:58amiconnUnhelpful: Several plugins do word-wrap already . Maybe it's time for another pluginlib module?
08:31:40 Join Zagor [242] (n=bjorn@rockbox/developer/Zagor)
08:33:00Unhelpfulamiconn: i was wondering if that might be a good direction to go. i was going to knock out a quick, dirty patch to try to help this guy: but though that maybe a *nice* way to display text with word wrapping and line breaks would be generally useful as well.
08:33:34Unhelpfuli suspect that it almost certainly *did* tell him what was wrong... it's just that PF uses a 1s splashf for that, and then exits.
08:33:52 Join Rob2223 [0] (
08:35:10 Quit Alex0008 ("ChatZilla 0.9.84 [Firefox 3.0.10/2009042316]")
08:37:29 Quit safetydan ("Leaving.")
08:40:13Unhelpfuli'm a bit curious as to the impact of r21110 on PF cache build on archos
08:41:29Unhelpfuli was looking to see if i could search tracks by album without reinitializing tagcache entirely each time i change albums, but i'm not really getting anywhere with that, and i suspect some of the more cpu-intensive parts will require repeating per album, anyway :/
08:49:02 Join flydutch [0] (
08:51:03 Quit Rob2222 (Read error: 110 (Connection timed out))
08:55:40 Join einhirn [0] (n=Miranda@
09:03:30 Join Alex0008 [0] (
09:03:33 Quit Alex0008 (Client Quit)
09:04:32 Join lee321987 [0] (
09:04:40lee321987;revision=21106 (Lua: add support for os library ) −−- Does that mean lua should/does now support the RB "copy" and "paste" functions?
09:05:58 Quit bertrik (Read error: 113 (No route to host))
09:06:22 Quit HellDragon (Read error: 104 (Connection reset by peer))
09:07:36 Join kachna [0] (
09:09:57 Join petur [50] (n=petur@rockbox/developer/petur)
09:09:58Unhelpfullee321987: i don't think so.
09:10:39Unhelpfuli didn't realize rockbox even *had* any kind of a clipboard? what exactly do you want to do, anyway?
09:12:08GodEaterwell we use the "copy/paste" paragigm for file copying...
09:12:50UnhelpfulGodEater: i don't think that data survives exiting the file browser, though?
09:13:05GodEaterno indeed not
09:13:42lee321987create a script that backs-up the Bubbles game save. I'm trying to do it by reading the entire contents to a variable (size=1608 bytes), but it hangs.
09:14:44GodEaterI would have thought that's exactly the sort of OS functions the lua library *should* implement
09:15:04LloreanCan't you pick which level to start in in bubbles anyway, once you've reached that level?
09:15:31lee321987Yeah, but I think you start at zero points
09:16:29LloreanWell, yeah, a score represents your skill. It's not skill if you cheat to avoid failing.
09:16:47 Quit r0b- (Read error: 104 (Connection reset by peer))
09:16:59Lloreanif you want a high score that bad, you can just edit the high scores file
09:17:10lee321987I'm bragging to people about these scores. It just gives me a goal −− highest points ever.
09:17:23lee321987lol @ previous
09:17:42lee321987*NOT bragging.
09:18:02lee321987What language is it good to know if you want to write .rock files?
09:18:17Unhelpfullee321987: even so, it's not the same if you *start* at a high level.
09:18:20Bagderlee321987: C
09:19:24cool_walking_You're missing out, guy. Everyone digs my wicked-cool brickmania scores.
09:20:11lee321987I'm sure I'd play a lot of that game if my DAP had a scroll wheel.
09:22:38*GodEater justs wishes he could get brickmania scores like cool_walking_
09:29:56 Join Thundercloud [0] (
09:34:36 Join matsl [0] (
09:36:54***Saving seen data "./dancer.seen"
09:43:12*Mikachu remembers brickmania was a lot easier when he changed the controls to absolute instead of relative
09:44:16joastrictly on topic i know, but alot of you maybe interested in this -
09:45:45Unhelpfuljoa: if there's something people might actually be interested in, and it's not about rockbox, you know, we *do* have an off-topic channel.
09:46:40LloreanMikachu: That defeats the purpose of the game though, it's about planning ahead for a fixed maximum speed paddle movement. Though, right now, our Brickmania still has ball speed issues too, I think
09:46:42 Join l403 [0] (n=l@
09:47:18MikachuLlorean: all breakout clones for pc has absolute positioning with the mouse (more or less)
09:48:12LloreanThat essentially makes them an entirely different game.
09:48:32LloreanIt's a gameplay thing, one of the challenges is dealing with the limitations of the paddle movement. Linear, and a fixed speed.
09:48:41Mikachuokay so brickmania is a lot easier when it's a completely different game :)
09:49:32LloreanFair enough, that.
09:52:00 Quit lee321987 (Read error: 110 (Connection timed out))
09:54:46 Quit BHSPitMonkey ("Ex-Chat")
09:56:07 Part LinusN
10:03:29 Quit Thundercloud (Remote closed the connection)
10:09:04 Quit freqmod (" - Chat komfortabelt. Hvor som helst.")
10:09:27 Join freqmod [0] (
10:11:22 Join daurnimator [0] (n=daurnima@unaffiliated/daurnimator)
10:12:56 Join efyx_ [0] (
10:15:35 Join robin0800 [0] (
10:22:38*GodEater thinks this PDF / Text view thing on the forums is a bunch of horse crap
10:22:58LloreanWell, someone "confirmed" it.
10:23:04LloreanI think it's probably a genuine problem. With Foxit.
10:24:11LloreanOr possibly LaTeX
10:24:34LloreanEither way, there's nothing we can really do about it
10:24:55GodEaterI'm just going through the "working" ones that cool_walking_ posted
10:25:06GodEaterI'm wondering if it's to do with having embedded graphics
10:25:31GodEaterI mean - the problem does exist - Foxit displays an empty window when you choose this "Text viewer" thing
10:25:39GodEaterbut I've no real idea what that's supposed to do
10:26:21LloreanI can't imagine embedded graphics would be a problem for a view mode that's most likely supposed to be a way to disable graphics and view just the text?
10:26:22 Quit FrankTM (Read error: 60 (Operation timed out))
10:26:30 Join killan_ [0] (
10:26:38 Join FrankTM [0] (
10:26:39GodEaterno - that's not the issue it seems
10:26:47GodEaterone of the "working" ones has gfx too
10:27:52 Join rasher_ [0] (
10:28:49 Quit rasher (Read error: 104 (Connection reset by peer))
10:28:59 Quit killan (Read error: 54 (Connection reset by peer))
10:30:05joathere is Unhelpful?
10:30:14GodEateryes there is
10:30:16joacool, didn't know
10:30:24joawhat's it called?
10:30:40GodEater"Unhelpful" ?
10:31:05cool_walking_#rockbox-community - it's stated in the channel topic.
10:31:56cool_walking_GodEater stole my line, so I had to give a real answer.
10:32:13GodEatersorry ;)
10:37:26 Join pyro_maniac [0] (
10:38:46 Quit einhirn (Read error: 104 (Connection reset by peer))
10:42:14 Join pyro_maniac1 [0] (
10:54:25 Join joa_ [0] (
10:57:26 Join einhirn [0] (n=Miranda@
10:58:17 Quit pyro_maniac (Read error: 110 (Connection timed out))
10:59:49 Quit joa (Read error: 60 (Operation timed out))
11:07:17pixelmaUnhelpful: your error message in pictureflow seems to have problems with user font or so? At least the last (?) word on the first line is truncated when looking at it in an Ondio sim and 09-Nedore, guess it's a "Please"
11:07:52Unhelpfulperhaps i should just use sysfont?
11:08:02pixelmaand I wonder where the "Select" button is...
11:08:17 Quit cool_walking_ (" ajax IRC Client")
11:08:33 Join _lifeless [0] (n=lifeless@
11:08:56Unhelpfulpixelma: a number of system messages use the same wording (in english, anyway)
11:08:57pixelmacan't you put it in a usual splash? I thought that's doing line breaks correctly (on its own), independent of fonts
11:10:23Unhelpfuli can try that :)
11:10:31pixelmaI think "any button" would be better, core messages can have different language strings for the different targets
11:11:20pixelmaI guess plugins could do too but sounds like overkill in this case to me
11:11:21Unhelpfultrue. that's a bit of a problem if the display sleeps while the error is displayed... perhaps a 1s sleep after the button press?
11:13:06pixelmawouldn't you have the same problem now if the user accidentally already hits the "Select" button (and no first press enables backlight only)
11:13:29Unhelpfultrue, i would. i think i'll just add a sleep after the button press.
11:16:45Unhelpfulpixelma: splash word-wraps but does do newlines...
11:18:55Unhelpfulwhich would be kind of nice to have if i'm going to add a "press any key" message
11:24:34 Nick rasher_ is now known as rasher (n=rasher@rockbox/developer/rasher)
11:36:58***Saving seen data "./dancer.seen"
11:44:23 Join gregzx [0] (
11:53:48 Quit nibbler (Read error: 110 (Connection timed out))
12:01:53 Join mcuelenaere [0] (
12:02:45mcuelenaerelee321987 (logs): see regarding documentation of the os lib (however not all functions are supported in Rockbox)
12:04:01mcuelenaereand I don't think Rockbox has a file API to copy a file..
12:09:12 Nick fxb__ is now known as fxb (
12:10:30 Join kugel [0] (n=kugel@rockbox/developer/kugel)
12:13:05pixelmahow does it copy a file then?
12:13:05*Torne wonders if anyone knows a way/patch to limit playback volume
12:13:15gevaertspixelma: read a bit, then write it, and so on
12:13:29gevaertsnothing higher level
12:13:56LloreanTorne: You could use the EQ precut
12:14:10TorneLlorean: that's actual processing, though
12:14:18pixelmagevaerts: aha, thanks
12:14:24Tornei mean like apple's firmware, where you can simply set a max volume level and it won't let you turn it up louder than that :)
12:15:06Tornei keep deafening myself by trying to use the scrollwheel through the fabric of my clothes and going too far. my headphones are hueg. ;)
12:15:39 Quit robin0800 (Read error: 110 (Connection timed out))
12:16:38Unhelpfulany objections to newline support in splash?
12:17:25Unhelpfuli have a working implementation based on adding a strtok loop (on \n) around the existing one, and taking words of of the current line for the wordwrap part
12:18:43 Quit bubsy ("Don't panic.")
12:18:57 Quit killan_ ("( :: NoNameScript 4.22 :: )")
12:20:21mcuelenaereUnhelpful: I'm all for it!
12:21:35pixelmaon small displays, newlines can make it look worse sometimes (e.g. first line is already wrapped for one word so it looks like "wasting" the rest of this line)
12:22:08Unhelpfulthe main objection i see is core binsize bloat for something core doesn't use... it costs 108B on ARM
12:22:11pixelmajust my opinion
12:22:28LloreanTorne: Processing-wise it's pretty minimal
12:22:40Unhelpfulpixelma: you think it might be better if PF just did "Error message. Press any key to continue."?
12:23:34TorneLlorean: well, yes. i guess you could rephrase my question as "does anyone think it might be useful to have such a settign"
12:23:41pixelmaUnhelpful: how I imagine it now - yes (but haven't seen yet)
12:23:54 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
12:25:29 Join n1s [0] (n=n1s@rockbox/developer/n1s)
12:26:26 Join bubsy [0] (i=Bubsy@
12:27:51AlexPTorne: This has been discussed a few times in the forums, and if I remember correctly the answer was generally no
12:27:56mcuelenaereUnhelpful: hmm pictureflow still 'crashes' when trying to view the playlist of an album on an Onda VX747 here..
12:29:23Unhelpfulmcuelenaere: andy ideas about that? :/
12:29:25TorneAlexP: yah, i've just found that discussion ;)
12:29:52mcuelenaereUnhelpful: not really, you're the pictureflow expert :)
12:30:20mcuelenaereI don't even get any exceptions
12:30:33Unhelpfulit just hangs? :/
12:32:29Tornemight just write it anyway and see how many bytes it is
12:32:58Unhelpfulare we *able* to trap weird memory access, etc, on onda?
12:33:37mcuelenaereyes, weird memory accesses would trigger a MIPS exception but I didn't get one in pictureflow
12:33:50mcuelenaereand MIPS does have a watch register, but haven't used that yet
12:35:45CIA-38New commit by unhelpful (r21113): Use core splash for PictureFlow's error_wait, exit error display 2s after any keypress, remove DEBUGF committed accidentally.
12:35:50Unhelpful...what does it watch? :)
12:36:00 Join StevenTyler [0] (n=StevenTy@
12:36:37StevenTylerrock rock rock rock n roll highschool
12:37:15 Part StevenTyler
12:37:18 Join killan [0] (
12:37:20mcuelenaereUnhelpful: AFAIK a specific memory address you tell it to watch, and then it triggers a watch exception when it is accessed
12:38:04kugelUnhelpful: we could also have a splash that runs as long as no button is pressed
12:38:39kugel(in the core, I mean), it's just a matter of replacing sleep() with get_action_w_tmo()
12:39:24kugeland yes, I'd like to have \n-support (generally in Rockbox, not only splash)
12:40:30mcuelenaereUnhelpful: hmm the problem seems to be troubled at or before show_track_list(); gets called
12:41:25Unhelpfulmcuelenaere: i'll try to look at it tonight.
12:41:59mcuelenaereUnhelpful: yep, definitely *in* show_track_list()
12:46:24mcuelenaerehmm it seems to be located in l.2244->2246
12:48:57AlexPI mentioned this a few days ago, but not many people were around. Lua in SOURCES depends on plugin buffer size (#if (PLUGIN_BUFFER_SIZE >= 0x80000)
12:49:18AlexPWhat are people's thoughts in converting this into manual \opts
12:49:32LloreanManual opts for plugin buffer size?
12:49:41AlexPor something else to check
12:50:05LloreanUsually we go by RAM size, right?
12:50:18LloreanSince the plugin buffer (so far) has been defined as 512kb if >8mb RAM
12:50:29*Bagder came up with two new topics to discuss at devcon
12:50:35LloreanBagder: Oh?
12:50:39AlexPDo we have RAM size \opts available for the manual?
12:50:48scorche|shBagder: good!
12:50:50AlexPI can't look now a,d can't remember (at work)
12:51:20AlexPpixelma: Where to put HID will need a little thinking too :)
12:51:20LloreanAlexP: It may just land on the swcodec/hwcoded split for now, since for the moment it's the same line.
12:51:48pixelmaAlexP: my opinion depends on whether you want to make it an opt by using features.txt or not (I'm always skeptical about it because features.txt was actually invented for the language parsing)
12:51:48AlexPLlorean: Is that true for some of the potential AMS targets?
12:52:36mcuelenaereUnhelpful: the problem seems to be this: track_names = (char *)buflib_buffer_out(&buf_ctx, &out);
12:52:43LloreanAlexP: Nope. Though *hopefully* the 512kb plugin buffer for SWcodec can be preserved.
12:53:01AlexPpixelma: It'd be nice to be able to do all the >= stuff too :)
12:53:04LloreanAlexP: I believe the Clip only has 2MB of RAM
12:53:24pixelmanot sure though, maybe I have to rethink that. Does adding features "cost" anything when not used for lang building
12:53:36AlexPLlorean: This is it though - it gets measy doing \opt{swdoce}{\nopt{clip}{ etc
12:53:40pixelmaAlexP: sure, if you know how to...
12:53:51AlexPpixelma: that was my point :)
12:54:03LloreanAlexP: Well, *hopefully* the Clip *won't* be an exception, though
12:54:20AlexPLlorean: I can certainly do it like that, it would be nice to have a cleaner way though :)
12:54:23LloreanSince it's Flash, a fairly small compressed buffer ought to be possible.
12:56:01Torneoh, talking of buffers
12:56:18Tornethe background loading of the dircache and tagcache on my player seems to cause playback to skip if i shake it
12:56:21*pixelma would also really really know how to make the manual build process chose a 2 or 3 columns wide button table based on remote (or button pad as a start)
12:56:26Torneeven though there is loads of compressed data buffered
12:56:27LloreanTorne: So don't shake it...
12:56:44pixelmaI would also *like* to know
12:56:46Torneby shaking it i mean "having it in my pocket while walking"
12:56:58Bagderstop walking! ;-O
12:57:14LloreanTorne: Dircache and tagcache require time to be spend doing disk access, which returns control to playback as often as it can, but shaking the player can result in it taking longer in areas that should be relatively fast
12:57:25LloreanTorne: And last time you brought it up, you said it happened when *running*, not walking, I thought
12:57:37Torneyah, but i've experimented more since then
12:57:41Torneyou barely have to move it at all
12:57:52Tornei guess the disk's accelerometers are set for a pretty low level
12:58:26LloreanTorne: Maybe turn it on a short time before you use it, or turn off auto-update for the database.
12:58:35Torneit's not updating, just loading to ram
12:59:00LloreanThat should only take a few seconds.
12:59:07Torneit doesn't if you're playing.
12:59:12Tornei have resume on poweron
12:59:20Torneit seems to buffer audio first
12:59:26LloreanSo disable that.
12:59:33*Torne eyerolls
13:00:01LloreanTorne: What SVN version are you using, anyway?
13:00:07Torneer, recentish
13:00:12LloreanThat's not meaningful in any way
13:00:14scorche|shturn on, wait a could of seconds, hit play...doesnt sound hard
13:00:14Tornemay 5th or thereabouts
13:00:20n1spixelma: adding more features to features.txt is essentially free (except for making the file longer)
13:00:22LloreanThat's not at all recent
13:00:28Llorean23 days is quite a long time.
13:00:31*Bagder spots that the evil green zeroes finally pushed out the peaceful blue blanks from the build table
13:00:33Lloreanhundreds of revisions.
13:00:38Tornethere's not much in teh interim that's relevant
13:00:45LloreanTorne: Do you have album art?
13:00:50LloreanThen yes, there is.
13:01:02 Join Lynx_ [0] (
13:01:02LloreanThere was a problem similar to this related to album art that may be what you're experiencing
13:01:20n1spixelma: also make sure to not create duplicates and try to have them in alphabetical order, i bumped into a weird bug caused by one feature defined twice depending on different preprocessor macros :/
13:01:38LloreanTorne: Anyway, as a standard rule, update to a version less than a day old and verify it still happens before reporting.
13:01:50LloreanIt's just an issue of procedure, since we're going to ask you to double check it happens in a current build still anyway
13:02:18Tornei'm just wondering if that kind of thing is expected to work, is all
13:02:23Tornei'm already donig a new build
13:02:42Tornei can see where the problem might lie even aside from any issues with album art
13:02:52AlexPIt'd be good to find out if it does work currently before wondering if it should :)
13:02:54LloreanBut it also might not.
13:03:02 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
13:03:59pixelman1s: it still strikes me a bit odd to add features only for the manual there (although the automatic parsing was my idea (and your work))
13:04:30*Torne tries resuming something without art while he waits, though ;)
13:05:53n1smaybe but if it makes something a lot easier in the manual i think it's worth it
13:06:21AlexPpixelma: We could either have features.txt as a general (not only language) resource, or an additional manual specific one
13:06:22n1sOne thing I would like to see is a bit more automatic magic for things like battery capacity and volume limits etc.
13:07:04n1sAlexP: it is essentially a general thing now and i don't see any good reason to split it
13:07:38AlexPn1s: I agree, and as long as that is the case think we should just go ahead and use it where it makes sense
13:10:19pixelmaAlexP: then for the memory size thing, maybe add an "lowmem" feature depending on this check. Although that's no answer for the >= etc. check and won't work if we have more than one memorysize "limit" check
13:10:46pixelmafor different amounts I mean
13:10:47gevaertsI think we should find a way to not only differentiate current and release builds, but also manuals. features are added changed so fast these days that the current manual can be seriously wrong for the latest release
13:10:48mcuelenaereJdGordon: I'm having this weird WPS bug: some music files trigger it and some don't, when it gets triggered it, value at l.1405 has an invalid value which triggers a memory exception at l.1411 (gwps-common.c). However when I play the same file after I played a working song (i.e. through the playlist) it doesn't get triggered...
13:11:21Lloreangevaerts: It shouldn't be hard to just build a manual with the release (maybe even just pack the pdf in the .zip)
13:12:02AlexPpixelma: Yes, I think I'll do that for now, and try and investigate comparatives
13:12:28pixelmagevaerts: how would you determine which revision the manual is up-to-date to?
13:13:05AlexPgevaerts: yes. It'd be good to try and get a bit more manual updating when introducing features to. Not necessarily in latex, but just some text that someone can more easily manualify would be useful
13:13:07 Join Horschti [0] (n=Horscht@xbmc/user/horscht)
13:13:21gevaertspixelma: we'd have to assume that the release manual is up to date on the release date
13:13:55LloreanIt should be (or we should try to make sure it is)
13:13:56 Join MethoS [0] (n=clemens@
13:14:33 Quit flydutch (Read error: 104 (Connection reset by peer))
13:14:36gevaertsThe problems I see are plugins that aren't in the latest release, changed keymaps, and things like HID and OF vs rockbox USB
13:14:46 Quit kugel (Read error: 110 (Connection timed out))
13:15:24LloreanThere's no doubt it'd be beneficial for users, overall, if they got a manual that matched what they were actually installing
13:15:46AlexPLlorean: yeah, 'spose so :P
13:20:20n1sthe release page has release manual pdfs
13:20:36LloreanWow, we did that fast.
13:21:00*Torne sniggers
13:21:07n1swell, it could be made clearer as most people wont' see that if they use rbutil to install
13:21:24Lloreann1s: Yeah, RButil should present a "release / current draft" option for the manual
13:21:35LloreanBut we'd need HTML of the release ones too then
13:23:27 Join flydutch [0] (
13:23:36pixelmaRBUtil shows the manual? I thought it only downloaded the correct PDF for you... never actually tried as you can tell
13:24:16LloreanI think it shows the HTML one?
13:24:29*n1s has never used it so he doesn't know
13:24:44LloreanIt can open the HTML one in your default browser
13:24:54LloreanOr download pdf/html zip
13:25:35pixelmaah, well. Shouldn't be a big problem either way I think.
13:25:50LloreanYeah, we just need all forms of the manual for both release and current
13:26:22gevaertsMaybe we just need to make the "See the download page for the manuals for the latest release" on the manual page a bit more prominent
13:26:49 Quit Horscht (Read error: 113 (No route to host))
13:27:00Lloreangevaerts: The problem is, with RBUtil the user will skip right past it.
13:27:03LloreanNo visible warning.
13:27:07gevaertsexcept of course that the release download page was simplified so it doesn't actually have manuals anymore...
13:27:38LloreanThat's the exact opposite of good.
13:27:48gevaertsLlorean: sure, but if someone just goes to the website afterwards, the release manual should be reasonably easy to find. Currently it isn't
13:27:57LloreanThis is true.
13:28:10LloreanIt needs to be easy to find, but I think also accessible via rbutil
13:28:23gevaerts still has them, but that's several clicks away, some of which don't mention the manual
13:28:56 Quit perrikwp (" ajax IRC Client")
13:29:13 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
13:30:34l403gevaerts, thanx for the input on the dfu tool. I'll I think I'll have to reread it to understand more :)
13:32:27 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
13:34:14 Quit gevaerts (Nick collision from services.)
13:34:23 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
13:37:03***Saving seen data "./dancer.seen"
13:38:16 Join VytenisS [0] (
13:39:32n1sgevaerts: yeah, the top level release page should have them
13:41:01n1shmm, and the "Old releases" link only gives 2.5 and earlier releases, maybe it should have 3.0 and 3.1 in it too, Bagder?
13:41:26*Bagder whistles. I'd rather just blame Zagor.
13:41:40n1sok, that is perfectly acceptable :)
13:42:06*Zagor whistles _harder_
13:42:26n1sbtw, i have taken down my build server and am not sure if it will come back up in the forseeable future
13:43:11Bagdertoo bad, but we'll manage
13:45:34n1sIt was one of the slowest machines in the farm anyway :)
13:47:00 Quit mcuelenaere ()
13:49:45n1sIf I get a better box I'll set one up again when i get a more permanent place to live
14:05:00 Join froggyman [0] (n=187b533e@gateway/web/cgi-irc/
14:07:20 Quit froggyman (Client Quit)
14:09:55 Join froggyman [0] (n=187b533e@gateway/web/cgi-irc/
14:16:16 Quit SirFunk__ (Read error: 60 (Operation timed out))
14:19:43 Join SirFunk__ [0] (
14:24:12 Quit froggyman ("CGI:IRC (EOF)")
14:27:33 Join LambdaCalculus37 [0] (i=44a0430d@rockbox/staff/LambdaCalculus37)
14:29:56 Join SirFunk_ [0] (
14:37:34 Join wincent [0] (
14:38:56 Join robin0800 [0] (
14:41:09 Quit SirFunk__ (Read error: 110 (Connection timed out))
14:41:52 Quit robin0800 (Remote closed the connection)
14:42:46 Join itcheg [0] (i=41d59de2@gateway/web/ajax/
14:44:34 Join SirFunk__ [0] (
14:48:30 Join kugel [0] (n=kugel@rockbox/developer/kugel)
14:51:49 Join martian67_ [0] (
14:53:51 Quit SirFunk_ (Read error: 110 (Connection timed out))
14:54:37 Quit martian67 (Connection timed out)
14:59:07 Quit evilnick_home1 ("Leaving.")
15:00:44 Quit matsl (Read error: 110 (Connection timed out))
15:01:48 Join matsl [0] (
15:05:11 Quit _lifeless (Remote closed the connection)
15:05:27 Join _lifeless [0] (n=lifeless@
15:13:54MarcGuayLooks like the Sansa c100 NAND is formatted to FAT16...?
15:14:25MarcGuay"No partition found" error...
15:15:00MarcGuayLambdaCalculus37: Care to check the m200 if it's the same?
15:15:15LambdaCalculus37MarcGuay: Haven't got it with me right now, but yeah, I'll try tonight.
15:15:29tmztbased on partition type?
15:15:35LambdaCalculus37But my m200 is a v3, with the SD bridge chip.
15:16:59MarcGuayLambdaCalculus37: Okay, those seem to be FAT32 (I've got the c100 SD as well).
15:18:26*MarcGuay contemplates a reformat...
15:18:49 Join cool_walking_ [0] (
15:18:57 Join teru [0] (
15:19:03LambdaCalculus37MarcGuay: IIRC mine also uses FAT16, but I'll make sure of that tonight.
15:19:29 Quit perrikwp (" ajax IRC Client")
15:19:36 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
15:22:01MarcGuayThat solved the "no partition found" error. Now to get "file not found"....
15:23:30LambdaCalculus37MarcGuay: Cool. :)
15:23:42MarcGuayThe OF still works... Strange, no?
15:23:42 Quit MethoS (Read error: 113 (No route to host))
15:26:06LambdaCalculus37MarcGuay: You formatted the volume on your c100?
15:26:18MarcGuayLambdaCalculus37: Yep.
15:27:06LambdaCalculus37MarcGuay: That confirms that no part of the OF is in the volume. I think it just keeps configs and the database there and nothing else.
15:27:22 Join funman [0] (n=fun@rockbox/developer/funman)
15:27:28LambdaCalculus37I did the same to my m200 and it still worked as well, after rebuilding all of the configs.
15:29:02amiconnMarcGuay: For small volumes, usign FAT16 makes sense. Such targets should use FAT16-enabled rockbox (several do already)
15:29:18amiconnIn fact everything that accepts memory cards should support FAT16
15:30:13tmztMarcGuay: are you basing it being FAT16 on the partition type in the partition table?
15:31:01LambdaCalculus37amiconn: Which targets already use FAT16?
15:31:28JohnTeddye280 firmware version 03.01.11A, ... that is a v2 right?
15:31:54funmanJohnTeddy: right
15:32:23MarcGuayamiconn: I see. tmzt: Basing it on Windows properties "Partition type: FAT"
15:32:58tmztMarcGuay: okay, I think that detects it based on the FAT
15:33:30amiconnQuite a number. Do grep -r −−include=config-*.h FAT16 [your-rockbox-working-copy]
15:33:54LambdaCalculus37amiconn: Will do.
15:35:10amiconnYou'll see a whole number of wrong comments there...
15:35:48amiconn/* Some Sansa E200s seem to be FAT16 formatted */ for many others than E200s
15:37:05***Saving seen data "./dancer.seen"
15:39:25FrankTMis there a default way of closing programs?
15:39:55FrankTMi have troubling closing some on Fuze
15:40:30funmanthere is no default way, this is dependant on each plugin and model
15:40:39 Quit kugel (Read error: 110 (Connection timed out))
15:40:59funmanthis should be documented in the manual, but plugins part of the fuze manual still isn't written. Do you want to help on this?
15:41:28FrankTMah.. well..
15:41:37FrankTMmost plugins close by holding home button
15:41:45FrankTMbut some with the off button
15:41:52FrankTMand other.. no clue
15:42:07FrankTMi was wondering if, in some cases, this are bugs
15:42:38 Quit timc (Remote closed the connection)
15:43:35 Quit LambdaCalculus37 (" ajax IRC Client")
15:44:03funmanFrankTM: yes it's a bug, if you can make a list of plugins which close by off button, i could fix that
15:45:04funmanwell they should all close with the home button (either single press either long press, depending of the plugin). So if you see a plugin which doesn't meet these conditions please tell
15:45:27FrankTMit's mostly holding the home button
15:46:42 Join LambdaCalculus37 [0] (i=44a0430d@rockbox/staff/LambdaCalculus37)
15:46:44FrankTMbounce says: [off] to stop
15:46:50FrankTMbut actually closes on home button :P
15:47:16funmanwell home button is considered to be the off button I think
15:47:32FrankTMmight be confusing
15:48:16n1sthose help texts in the plugins need to be adjusted for each port too... or if the plugin localizations were finished it would come as a free feature without the #ifdef hell
15:48:30 Join timc [0] (n=aoeu@
15:49:50FrankTMfunman: i'll create a list
16:00:21 Quit Slasheri (Remote closed the connection)
16:08:50FrankTMfunman: are you supposed to control both sides in pong?
16:10:07n1sanyone with a gigabeaat f/x around who could check if fs#10249 is correct ?
16:10:13FrankTMrockbox1d is slightly weird
16:11:36evilnickFrankTM: The controls are weird or the game itself is weird?
16:18:01 Join Slasheri [0] (i=miipekk@rockbox/developer/Slasheri)
16:18:57 Join obo [0] (
16:21:28FrankTMthe game has no purpose ? :P
16:21:37FrankTMor am i missing something
16:23:45 Quit cool_walking_ (Remote closed the connection)
16:24:33evilnickWell, if it's a game then it doesn't need a purpose. It's a one-dimensional version of tetris, so does pretty much what you'd expect, no?
16:24:54LambdaCalculus37FrankTM: It's mostly for laughs.
16:25:05FrankTMevilnick: true
16:25:25 Quit sajes (Read error: 104 (Connection reset by peer))
16:25:42FrankTManyway. i created a list of plugins that _I_ think need a little work
16:25:46FrankTMmainly on Sansa Fuze
16:26:09FrankTMshould i open a bug report/
16:26:10LambdaCalculus37FrankTM: Feel free to work on them and send us a patch. ;)
16:26:28FrankTMi'm fine with that actually
16:26:58LambdaCalculus37Once you have a new patch, open a new task on the tracker, then come back and tell us what the FS# is so someone can check it out.
16:27:20 Part LinusN
16:27:27LambdaCalculus37n1s: saratoga and I were looking at that task yesterday, but I don't have my Gigabeat F with me (again!) today.
16:27:58FrankTMLambdaCalculus37: preferbly against SVN I suppose?
16:28:25n1sLambdaCalculus37: yeah, it should be easy to check for any f/x user and a 5 second commit :)
16:29:08 Join cool_walking_ [0] (
16:29:43LambdaCalculus37FrankTM: Always against SVN. Makes it easy to patch and then add and commit.
16:30:02LambdaCalculus37n1s: Even on the sim?
16:30:06FrankTMjust double checking :P
16:32:15 Quit Lss (Read error: 104 (Connection reset by peer))
16:32:25 Join toffe82 [0] (n=chatzill@
16:32:32 Join Lss [0] (
16:36:22n1sLambdaCalculus37: now that you mention it, i think so
16:36:50n1sdo you want the honours? :)
16:38:08 Join kugel [0] (n=kugel@rockbox/developer/kugel)
16:38:18 Join chandoo [0] (
16:40:02 Join SirFunk_ [0] (
16:47:40LambdaCalculus37n1s: Sure, I'll do that on my lunch hour.
16:50:53 Quit perrikwp (" ajax IRC Client")
16:56:41funmanFrankTM: for me just send a list where the Sansa Fuze _keymaps_ need a bit of work
16:57:34FrankTMfunman: most issues i found are due to controls
16:58:06tmztis dbop working now?
16:58:54funmantmzt: what do you mean?
16:59:18tmztare the scrollwheel issues with accuracy still there?
17:00:04tmztthe one that causes the menu selection to go to the top when scrolling though the menu, skips back to the beginning
17:00:22tmzton Fuze
17:00:52FrankTMtmzt: haven't noticed that yet
17:01:03funmantmzt: i think these have been fixed yes
17:01:48funmanin fact i think we are now looking for problems on the fuze and e200v2. Clip deadlocks while playback, so it's a bit behind Fuze and Clip.
17:01:49tmztokay, thank you. previously plugins had to be exited by sliding the power switch down I believe, but I understand home is working now
17:02:14FrankTMup to be complete ;)
17:02:15tmztare all amss still using the small memory layout in rockbox?
17:03:02LambdaCalculus37tmzt: No, we got past that now AFAIK.
17:03:04funmanno, only the Clip (and the less popular m200v4)
17:04:15MarcGuayshotofadds, LambdaCalculus37: I'm getting a "File not found" error even though I placed a file named "rockbox.c100" in the root. Any idea why it wouldn't be able to see it?
17:04:31funmanlinuxstb: about mkamsboot, I wonder if I should add Clipv2 support before or after 1.0 release. The problem is that the model number (in config-clipv2.h) is not known since rockbox doesn't support Clipv2 yet.
17:05:13AlexPn1s: If it is the chqnging the keys, then yes
17:05:18LambdaCalculus37MarcGuay: As far as I know, the TCC-based devices don't have a very reliable NAND driver yet.
17:05:41AlexPn1s: I nearly committed yesterday, but then I went through and changed all the other languages too
17:05:53AlexPn1s: I just have a slight question mark over one or two
17:06:08MarcGuayLambdaCalculus37: It should be able to read at least... The D2 and iAudio can load RB.
17:06:18AlexPLambdaCalculus37: My last couple of lines were at you too :)
17:06:43 Quit SirFunk__ (Read error: 110 (Connection timed out))
17:07:43LambdaCalculus37AlexP: I won't attempt to commit until everyone's settled on it, then. :)
17:08:05LambdaCalculus37MarcGuay: The m200 didn't, last time I tried.
17:09:06LambdaCalculus37MarcGuay: The c100 (and I guess the m200) can upgrade their firmware via dropping a new firmware file into the root directory and letting the device upgrade, right?
17:09:21MarcGuayLambdaCalculus37: Yep. "player.rom" for the c100.
17:09:34LambdaCalculus37MarcGuay: IIRC same on the m200.
17:09:50*LambdaCalculus37 should try making a proper bootloader and build for the m200 again
17:10:07LambdaCalculus37MarcGuay: And use mktccboot to put the two together, right?
17:10:10MarcGuayLambdaCalculus37: BTW, adding HAVEFAT16SUPPORT to the config file is enough to get it working.
17:10:18MarcGuayLambdaCalculus37: Yep.
17:11:49*LambdaCalculus37 really, really hopes we can have a new port release this year ;)
17:11:53linuxstbfunman: It may be nice to include the clipv2, assuming you're close to implementing it.
17:13:27funmanlinuxstb: we already have working code for lcd and buttons, the principal question is if we merge it into as3525/ (renamed to as35xx) or create a new architecture.
17:13:42FrankTMfunman: i created a patch against logo.c :P
17:13:51funmanBoth look very similar, but we will have a precise answer when we will have SD transfers
17:14:45 Quit kugel (Read error: 60 (Operation timed out))
17:14:55tmztis the cpu capable of accessing sd and usb client?
17:15:55 Nick fxb is now known as fxb__ (
17:16:12 Quit teru ("Quit")
17:16:55 Quit Zagor ("Don't panic")
17:17:45linuxstbfunman: How similar do you think they are? e.g. what percentage of registers appear to be identical?
17:19:51funman90-95% : some registers are at the same address but bits signification change, and some registers were added since as3525
17:20:02 Join kugel [0] (n=kugel@rockbox/developer/kugel)
17:20:19LambdaCalculus37MarcGuay: Any luck getting anything else to work on the c100?
17:20:50MarcGuayLambdaCalculus37: Troubleshooting the "file not found" error.
17:21:09MarcGuayCan't do much if it can't even read a file.
17:22:27linuxstbfunman: Then a merge and rename to as35xx seems reasonable.
17:23:54 Quit matsl (Read error: 110 (Connection timed out))
17:24:12funmani'll add a dummy entry for clipv2 in the build system, before i really add the target
17:26:57LambdaCalculus37Speaking of adding entries to the build system, any objections to adding the GoGear HDD6330 to the build system?
17:27:58LambdaCalculus37Since it's primarily using the HDD1630 build with the LCD resolution changed, it'll make sense to just create a new target with the values set.
17:28:04 Quit flydutch (Read error: 110 (Connection timed out))
17:28:44LambdaCalculus37Also (at least according to the wiki) the HDD6330 has 64MB of RAM, so it'll allow us to have the larger buffer.
17:30:56 Quit kugel (Read error: 60 (Operation timed out))
17:32:52VytenisSHello, i made a Lithuanian rockbox translation And i would be really thankful if someone check that out and commit if everything is all right
17:33:14LambdaCalculus37VytenisS: Sure thing!
17:34:00 Join kugel [0] (n=kugel@rockbox/developer/kugel)
17:34:28 Join jmillikin [0] (
17:35:25LambdaCalculus37VytenisS: It looks good so far. Even if you weren't able to completely translate the whole thing at once, you can always submit patches to add in missing bits, or fix existing ones.
17:36:37AlexPLambdaCalculus37: The change is fine, I was just going to do the other languages too as apart from five, they all use POWER = local string or whatever, so you can change them all without knowing the language
17:36:52AlexPLambdaCalculus37: In fact I've done this at home, I'll add it to the task tonight
17:37:07***Saving seen data "./dancer.seen"
17:37:09VytenisSLambdaCalculus37 all right then, i don't know, maybe i should look into some other multimedia software translations, just to figure those missing ones
17:38:02LambdaCalculus37VytenisS: No problem. I'll commit this later on today.
17:40:23LambdaCalculus37Keep an eye on the front page.
17:42:33MarcGuayAny suggestions on what to look for when the rockbox binary fails checksum? (c100 bootloader)
17:43:04linuxstbDo you know what the two checksums are?
17:43:24MarcGuayChecksum is 6e616e6f, Sum is 3a82a02
17:44:07linuxstbCould simply be the NAND driver not working...
17:44:47linuxstbWas the c100 empty (i.e. newly formatted) when you copied .rockbox across? IIRC, that helps on telechips targets.
17:44:49MarcGuayRight. It's not using the same one as D2/iaudio7 (FTL_v1)
17:45:10MarcGuaylinuxstb: I think so but I'll try again.
17:45:23MarcGuayDoes it matter that I'm just using a dummy rockbox.c100 file?
17:45:53linuxstbWhat do you mean by dummy? Was it created with scramble and the correct "-add" option?
17:46:33funmanFrankTM: thanks, by the way you can use 'svn diff dir/file.c'
17:46:50FrankTMi thought it would be easier for you to have one patch?
17:47:08funmani mean for generating patches against a single file (since you used diff)
17:47:21FrankTMdiff took a loot of .svn crap
17:47:41 Join rmull [0] (
17:48:32rmullHi, I have an old iPod color, is there a way to make this thing act as a USB host using Rockbox?
17:48:46 Join {phoenix} [0] (
17:49:28n1srmull: no, (not yet at least)
17:50:09rmulln1s: Thank you
17:50:13FrankTMfunman: would you rather use diff or svn diff
17:50:18funmansvn diff definitely
17:50:47FrankTMok. next time i wont suplly standard diff files then
17:52:01funmanrockbox git server doesn't answer (connection refused)
17:52:22toffe82nt kernel include in linux kernel
17:52:45Tornermull: not ever, without external hardware, in fact.
17:52:57Tornermull: the usb controller can only be a target
17:53:53gevaertswell, the controller can be host, but it needs more wiring which I doubt is there can?
17:54:45Tornedoes it just have a host controller as well on teh soc, or is it OTG?
17:54:59 Quit preglow ("leaving")
17:55:15 Join preglow [0] (i=thomj@rockbox/developer/preglow)
17:55:50gevaertsI think it claims to be OTG. That doesn't mean it works of course
17:57:24CIA-38New commit by funman (r21114): scramble : add c2v2 (c200v2) in help message
17:57:32CIA-38New commit by funman (r21115): FS #10256 by Frank Oosterhuis : use HOME button to quit plugins ...
17:57:50funmanFrankTM: thanks :)
17:57:56FrankTMwoot :P
17:58:39FrankTMwell.. dinner time
17:58:50FrankTMfunman: i might create some more later
17:59:32funmanFrankTM: you know the process now ;)
17:59:48FrankTMthat was my main reason to create these simple patches :p
18:00:51 Quit petur ("work->home")
18:03:51l403gevaerts, but after a little longer time I ran it again and get_cpu() was successfull so I'll try make it Wait 2 secs and then try again
18:04:33funmanNow that we have 10 sandisk targets (5{0-9}), how should I add the 11th : 5{00-10} or 5{0-9} and 500 ?
18:05:00 Quit martian67_ (Remote closed the connection)
18:05:20 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
18:09:38funmanhum I could use 60-69 range and give another number to Tatung Elio TPJ-1022
18:10:19 Part JohnTeddy
18:10:21linuxstbfunman: Yes, I was about to suggest that - go into the 60s for Sandisk, and move the Elio elsewhere
18:11:33funmaninserting new devices into this table without breaking the layout is a bit difficult
18:12:10 Join marc_ [0] (
18:12:17 Nick marc_ is now known as MarcGuay_ (
18:13:49MarcGuay_linuxstb: I'm getting this error when trying to roll a full build for the c100: Is there a way to cheat and cut off certain features just to see if the checksum works?
18:14:09 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
18:14:48 Join ceatinge [0] (
18:14:49linuxstbMarcGuay_: You'll need to fix those errors at some point...
18:15:03LambdaCalculus37MarcGuay_: You can disable plugins.
18:15:08MarcGuay_linuxstb: True enough. Looking into them now.
18:15:35linuxstbBut "#if 0 ...... #endif" can be helpful...
18:16:52pixelmafunman: group the Elio with the LogikDax and put it under a "Ports started by linuxstb" headline ;)
18:16:57*pixelma runs
18:22:11CIA-38New commit by funman (r21116): Add Clipv2 to the configure script and tools/scramble to reserve a target id ...
18:24:01n1sor maybe a "Comming soon..." :)
18:24:48 Quit n17ikh (Remote closed the connection)
18:25:10 Join n17ikh [0] (
18:26:13 Join AndyI [0] (n=pasha_in@
18:27:13*amiconn still thinks the Elio is a somewhat interesting target
18:28:45LambdaCalculus37amiconn: Ask linuxstb if he'll send it to you so you can work on it. ;)
18:29:03 Quit cool_walking_ (Remote closed the connection)
18:30:08gevaertsl403: what exactly are you trying to do?
18:30:59 Join flydutch [0] (
18:31:23 Quit l403 ("Leaving channel")
18:36:37 Quit AndyIL (Read error: 110 (Connection timed out))
18:37:31AlexPrasher: I have another lang question for you if you are around :)
18:37:39AlexPOr indeed anyone else :)
18:37:45 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
18:38:03n1sAlexP: shoot
18:38:39AlexPr.e. - it changes both source and dest in english.lang as for the gigabeat for one option it says A = blah when it should say POWER = blah. I can very easily change all other languages too (apart from ~ 5) as they don't translate button names
18:38:57AlexPSo should I also change source and dest in all those?
18:39:26AlexPAnd for the five that remain, will the fact that the source is different from english.lang poke translators to look at it, or is all this wrong :)
18:40:06 Join r0b- [0] (
18:40:56 Join miepchen^schla [0] (
18:42:35CIA-38New commit by Domonoky (r21117): pdbox: new patch from wincent balin. make m_memory.c work in rockbox.
18:42:56 Join cmwslw [0] (
18:43:22 Join robin0800 [0] (
18:44:27*domonoky gets a "post-commit hook failed" ... "git-push: command not found". whats up with our svn server ?
18:44:38 Part cmwslw ("Ex-Chat")
18:45:10funmandomonoky: git server is down
18:49:58 Join stoffel [0] (
18:51:17domonokyah.. :-)
18:51:26LambdaCalculus37The keymaps file for the m200 appears to have a lot of missing information, and it's crapping out the make process.
18:53:23domonokyLambdaCalculus37: whats missing in this file ? and the build servers dont have a problem using it.. (the m200 keymap is used for all m200 variants)
18:54:14LambdaCalculus37domonoky: Hang on, let me try to pastebin this.
18:54:53 Quit kugel (Read error: 113 (No route to host))
18:54:55 Quit flydutch (Read error: 110 (Connection timed out))
18:57:46 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
19:00:05LambdaCalculus37Also, here's another oddity I'm getting whenever I try to launch the sim:
19:03:43funmanlinuxstb: please have a look at FS #10253 (mkamsboot) : for me it's ready
19:04:30 Quit miepchen^schla (Read error: 113 (No route to host))
19:05:39linuxstbfunman: What is "CC?=gcc" in the Makefile? (google fails me...)
19:06:02funmanlinuxstb: so we can use CC=i586-mingw-gcc make (CC will be set to gcc if previously unset)
19:06:34funmanyou can use += also (for CFLAGS for example)
19:07:56 Join HellDragon [0] (
19:09:13 Join mewshi [0] (
19:09:20mewshihi :)
19:09:44mewshiIs 5th generation iPod supported?
19:10:07Mikachuit's in the big box at the top of the web page
19:10:36mewshithank you, mr abrasive >.>
19:10:41gevaertsHave there been reports about charging problems on sansa recently? If not, I think we can probably release 3.3 with usb on the sansas. FS #9957 still needs to be fixed properly, but I don't think this is a show stopper if we document it in the release notes
19:10:51Mikachui'm having some trouble getting my plugin to work and i'm hungry ;)
19:10:58gevaertsmewshi: you don't like people helping you?
19:11:17mewshijust a yes would've worked
19:11:19 Join flydutch [0] (
19:11:25LambdaCalculus37I had a slightly strange notion...
19:11:58LambdaCalculus37Monitoring USB traffic during a firmware upgrade of the Gigabeat T. Dunno if it'll reveal anything, but it doesn't hurt to try.
19:13:08gevaertsyou're probably just going to see more or less normal mtp traffic I guess
19:13:37LambdaCalculus37gevaerts: In other words, "squat". :)
19:14:23gevaertswell, on the S you can install the OF with sendfirm, so there isn't anything special going on on the wire
19:14:41gevaertsAll magic is going to be in the firmware file itself and in the rom
19:15:02evilnickgevaerts: I've had no problems whatsoever charging my e280 over the past few weeks exclusively through Rockbox
19:15:13LambdaCalculus37gevaerts: Should I try sending a vanilla T firmware via sendfirm?
19:15:23*LambdaCalculus37 hasn't tried that yet
19:15:31gevaertsLambdaCalculus37: if that works, it proves that sniffing won't be useful I think
19:17:31ceatingeDoes anyone know where I can get the latest bootloader for the Sansa e200?
19:18:12LambdaCalculus37Through Rockbox Utility.
19:18:34linuxstbfunman: If you look in the sansapatcher or ipodpatcher Makefile, there's some magic for dealing with cross-compiling. But that can obviously wait...
19:19:01linuxstbfunman: All seems good to me though.
19:19:26gevaertsceatinge: the latest released bootloader can be installed using Rockbox Utility, or if you want to do things by hand, sansapatcher. There's a newer bootloader for testing on the tracker, and if you want the very latest you have to build yourself
19:19:45linuxstbfunman: Will you be able to commit the cosmetic stuff separately? (not a big deal if not)
19:20:40 Nick martian67 is now known as m67|away (n=martian6@about/linux/regular/martian67)
19:20:45pondlifeHi all, is there anyone here with a PortalPlayer (or, indeed, ARM) target able to try the latest patch at ..?
19:22:13LambdaCalculus37pondlife: Got my iPod video with me; let me try it.
19:22:28gevaertsI'll try on gigabeat f
19:22:37evilnickpondlife: I can test on an E280 but only if you can send me a build (I'm at work)
19:22:44pondlifeI fully expect Data Aborts
19:22:46evilnicks/you/some kind person/
19:22:52LambdaCalculus37evilnick: I'll make a build for you.
19:22:58pondlifeevilnick: I can't (still at work too).
19:23:05gevaertspondlife: the patch is full of CRs! ;)
19:23:29pondlifeI'm sure you can dos2unix it too ;)
19:23:49pondlifeWill upload a fix anyway
19:23:50evilnickLambdaCalculus37: Thanks, send it to my yahoo address, as my gmail one NEVER works for you?!
19:23:57gevaertswell, patch just complains a bit, it still applies
19:23:58LambdaCalculus37I haven't got dos2unix on OS X.
19:24:07LambdaCalculus37evilnick: No problem.
19:24:55pondlifeOK, how's that?
19:26:04gevaertspondlife: BUTTON_RC_RW and BUTTON_RC_FF don't exist on gigabeat. I'm trying left and right instead
19:26:33*LambdaCalculus37 gets the two builds building
19:27:36 Join JdGordon| [0] (i=836b0070@rockbox/developer/JdGordon)
19:28:11 Join BryanJacobs [0] (
19:28:57pondlifegevaerts: Updated for Gigabeat...
19:30:14gevaertshm, how does this work?
19:30:19 Join perfectdrug [0] (n=5b0ed723@gateway/web/cgi-irc/
19:30:30pondlifeGo into Sound Settings and enable Timestretch
19:30:33pondlifeThen reboot
19:31:01gevaertsah, it won't crash unless you tell it to? ;)
19:31:06pondlifeGo into the pitch screen, and press the mode button (whatever that is) a couple of times until you see both a pitch and speed displayed.
19:31:12domonokyLambdaCalculus37: sorry, i was on the phone. for which m200 did you build ? (normal or m200v4 ?)
19:31:21pondlifeThen push right to speed up, left to slow down, any key to crash ;)
19:31:32perfectdrughi is there any sansa clip owner? i just want to know if the bootup screen still lokks like this picture:
19:32:04 Join calman_ [0] (
19:32:12gevaertspondlife: no crash, but no work either. It stops playing, and after that the wps shows random metadata
19:32:27pondlifeHmm, that's basically a crash...
19:32:29perfectdrughope no own has made a svn image of the clip yet :D
19:32:48perfectdrugeh svg
19:33:15gevaertsit is, yes. Doesn't the gigabeat refuse to give proper data aborts anyway?
19:34:03domonokyLambdaCalculus37: it looks like you need to update button-target.h for the normal m200. (take a look at the button-target.h of m200v4)
19:34:35 Quit LambdaCalculus37 (" ajax IRC Client")
19:34:36pixelmaperfectdrug: why?
19:34:56 Join LambdaCalculus37 [0] (i=44a0430d@rockbox/staff/LambdaCalculus37)
19:35:07 Join moos [0] (i=mustapha@rockbox/staff/moos)
19:35:18LambdaCalculus37domonoky: Normal.
19:35:31 Quit pyro_maniac1 ("Leaving.")
19:35:33perfectdrugpixelma: because mine is almost finished
19:36:06domonokyLambdaCalculus37: and ofcourse you need to update button-m200 too.
19:36:58pixelmaby the way, I *think* the Rockbox logo was now pushed down into the blue part and the version string went into the top yellow part (but isn't shown in other svgs)
19:37:01LambdaCalculus37domonoky: Figured. :P
19:37:09***Saving seen data "./dancer.seen"
19:37:16LambdaCalculus37pixelma: Yes, that's correct.
19:37:17 Quit HellDragon (Client Quit)
19:37:39LambdaCalculus37evilnick: Sent you a build. Check your Yahoo mail.
19:37:47pixelmaperfectdrug: got a preview somewhere?
19:37:51perfectdrugdoes someone have a picture? i would need it to finish the clip svg
19:38:00pondlifegevaerts: I struggle to work out what data-type hell is going on - all I know is it seems fine on H300 target and all sims here.. :/
19:38:10 Quit itcheg (Read error: 110 (Connection timed out))
19:38:26gevaertspondlife: let's wait for the PP people. They should get a nice data abort with an address
19:38:26perfectdrugpixelma: i make one
19:38:35 Join HellDragon [0] (
19:39:18pondlifegevaerts: Yes, but will they know what to do with that address..? (I certainly don't.)
19:40:28*LambdaCalculus37 is updating his iPod video's build... give him time! :)
19:40:38gevaertspondlife: look it up in the map file, or if it's a static function, look it up in the disassembled .elf. That will at least tell you in what function it goes wrong
19:40:51*gevaerts hopes that LambdaCalculus37 will keep the .elf files
19:41:32LambdaCalculus37gevaerts: I have them still. But right now is "no taking out lappy" time. :(
19:41:52 Join Thundercloud [0] (
19:42:01gevaertsLambdaCalculus37: no hurry
19:42:11gevaertsit needs to crash first anyway
19:43:11LambdaCalculus37All right... now how to crash it? ;)
19:44:03pixelmaperfectdrug: looks nice. :) Did you reuse some of the existing things (for the button "labels" or so?. Also, does it use transparencies in something (as I remember some tools for making the pdf/pngs for the manual has problems with that)
19:44:25gevaertsLambdaCalculus37: enable it, and then use it :)
19:45:56LambdaCalculus37gevaerts: Timestretch?
19:46:11perfectdrugpixelma: yes i tried to adapt the colours from the e200 yes i use transparencies for the shadows and lightning
19:46:19evilnickpondlife: Hmmm, does this timestretch only work with the + or - 2% buttons then?
19:46:28pondlifeNo, it's a 3rd mode
19:46:42pondlifeabsolute > semitones > speed
19:47:02pondlifeNot all targets have keymaps though....
19:47:20 Join bertrik [0] (
19:47:31Bagder"2009-05-28 disabled due to 1000+ seconds builds"
19:47:44evilnickI have timestretch enabled (checked my config file to be sure) but there's no change on the pitch screen
19:48:02pondlifeevilnick: You need to reboot with it enabled?
19:48:06evilnicki.e. If I change the pitch up or down then the speed changes too
19:48:08pondlifeIt allocates some extra buffers
19:48:17evilnickpondlife: I have done. I'll do it again to make sure
19:48:24gevaerts1000+? That's long
19:48:27LambdaCalculus37pondlife: I can't make it crash. Although it makes the music sound hilarious. :)
19:48:34pondlifeLambdaCalculus37: Indeed it does
19:48:58pondlifeevilnick: e200?
19:49:13pixelmaperfectdrug: can you try to do without? It usually works the same, it would only make it harder if you want to change the basic colours later. As I said it is troublesome for the pdf/pngs for the manual
19:49:15perfectdrugpixelma: can't we make the pdf and png files manual so that the tools don't get confused?
19:49:18evilnickpondlife: Yes
19:49:21pondlifeevilnick: Try pressing the record button twice to get the speed mode
19:49:43evilnickpondlife: Oooooh!
19:49:45FrankTMdo voice menus on Fuze work?
19:49:51pondlifeThen up/down should change pitch but maintain speed and left/righ should just change speed.
19:50:04evilnickpondlife: Data abort at 000097C0 (0)
19:50:26LambdaCalculus37pondlife: No data abort, but my iPod video froze.
19:50:26pondlifeLambdaCalculus37: You able to look that one up/
19:51:09LambdaCalculus37pondlife: Can't right now.
19:51:29pondlifeAh, ok. I'll do an e200 build here but I suppose it's unlikely to match maps.
19:51:29 Quit BryanJacobs ("null")
19:51:48gevaertsI'll try here
19:51:50LambdaCalculus37Data abort!
19:51:54 Join mirak [0] (
19:51:55 Join BryanJacobs [0] (
19:52:05*LambdaCalculus37 got a Data Abort at 00009CC0 (0)
19:52:11evilnickpondlife: If I press >>| to increase the speed then as soon as it gets to 101% then audio stops, but it doesn't crash
19:52:25evilnickAnd the wps becomes nicely garbled
19:52:49pondlifeSounds like what gevaerts is getting
19:52:56perfectdrugpixelma: i try to check for transparency and fix it
19:52:58pondlifeYet works on my H300
19:53:23pondlifeI wonder if ir's config related?
19:53:28perfectdrugso can someone please make a picture of the clip bootscreen, please?
19:53:40LambdaCalculus37pondlife: Got another Data Abort at 00009C74 (0).
19:53:50LambdaCalculus37You know what the magic song is that's killing it?
19:54:04pixelmaperfectdrug: I wouldn't know how to make the pdf differently
19:54:58LambdaCalculus37evilnick: Want to try the "magic data aborting" song on your e200 to see what it does?
19:55:39evilnickLambdaCalculus37: Is that with a normal build? (unrelated to the timestretch)
19:56:35*gevaerts gets a data abort on his e200 as well
19:57:03evilnickpondlife: My wps is now showing three rows of a strange box character, and the time remaining is -8524:56!
19:57:33pondlifeAnyone know how to interpret a .map file??
19:57:44gevaertsinteresting. That looks like exactly the same remaining time that I had
19:57:51LambdaCalculus37evilnick: No, it's with the timestretch.
19:58:25evilnickLambdaCalculus37: I can force a data abort by using the wheel (rather than the >>| key)
19:58:28pondlifeSeems to be in buffering.o, but that makes little sense.
19:58:31gevaertspondlife: find_handle
19:58:52pondlifeHmm, my .map file doesn't even mention find_handle
19:59:07pondlifeThis is an e200 build
19:59:08gevaertsarm-elf-objdump -d rockbox.elf|less is more detailed
19:59:13pondlifeAh, ok
20:00:43perfectdrugpixelma: where is the conversion tool so that i can check if the output works?
20:00:51pondlifeSo it's probably just scribbled all over buffering's data...
20:01:21gevaertsthat's what it looks like to me, yes. No wonder things sound funny :)
20:01:32pondlifeNah, they sound funny when it's working too....
20:01:58funmanlinuxstb: i'll look at other Makefiles for cross-compilation
20:01:59pondlifeCould someone also try the 090309 patch? Allegedly that works fine.
20:02:25pondlifeBut this now shares exactly the same core code, I believe...
20:02:39pondlifeJust the pitch screen rather than a setting.
20:02:45funmanlinuxstb: i can't easily commit cosmetics separately, but that's not a big deal since a lot of code changed/moved already
20:02:50 Quit r0b- (Read error: 110 (Connection timed out))
20:02:59evilnickpondlife: This time I tested moving the pitch down, it worked up to about 96% and then crashed: Data abort at 00009488 (0)
20:03:16 Join r0b- [0] (
20:03:26pondlifeIt probably starts mangling stuff immediately the extra code is used (i.e. non-100%)
20:03:40pondlifeThen it's just luck how long things survive ;)
20:04:11pondlifeIf someone could confirm that 090309 works, I'll get the finer toothcomb out.
20:05:23gevaertstrying now
20:07:22gevaertshow do you control this?
20:07:31pondlifeSettings menu
20:07:42pondlife(Under Sound Settings, I think.)
20:08:03pondlifeJust lets you set a speed
20:08:14gevaertsalso crashes
20:09:56LambdaCalculus37Someone should try this on the beast as well.
20:10:26pondlifeI doubt it'll build, but perhaps you could try an older one - 090220, maybe
20:10:54pondlifeNot much point in running it on many targets - I already know it's broken...
20:12:36pondlifeor 090225 maybe
20:13:04 Nick fxb__ is now known as fxb (
20:14:10pondlifeHmm, actually 090303 is reported as working on iPod mini, so that would be a good endpoint. Yes - I should have read the lot before posting!
20:14:29gevaertsI just installed 09020 :)
20:14:48pondlifeWhatever ;)
20:15:03funmantools/ucl/src/Makefile says "we don't use $(AR) and $(RANLIB) since we want the _native_ tools not the cross-compiler tools", but that's not true :/
20:15:15perfectdrugpixelma: it doesn't seem to work :/ as it is right now
20:15:41perfectdrugi try to tweak it
20:15:49gevaerts090220 seems to work
20:16:34pondlifeBinary chop time..?
20:17:13gevaertsapparently. Any suggestions?
20:18:04pondlifeI hope you're writing these numbers down.... though I guess logbot never forgets.
20:18:40pondlifeSadly I will have to run off again shortly. but will stay as long as poss.
20:20:34pondlifeI expect I broke something between 090306 and 090306b, but it's not obvious to these tired eyes.
20:20:50gevaerts090305 works
20:21:08pondlifeMaybe try 090306d next?
20:21:38 Nick Horschti is now known as Horscht (n=Horscht@xbmc/user/horscht)
20:21:45gevaertson its way
20:21:56pondlifeheh, many thanks for this
20:22:14pondlifeIt's frustrating when your target won't fail.
20:22:14 Nick m67|away is now known as martian67 (n=martian6@about/linux/regular/martian67)
20:22:36perfectdrugpixelma: is "unschärfe" not allowed neither? i think it would be really hard to do this without, because it's hard to get the 3D effect only with gradients at the curved top
20:22:47 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
20:23:11 Quit thegeek (Read error: 104 (Connection reset by peer))
20:23:23pixelmasounds like it could be a culprit
20:23:24 Quit flydutch ("/* empty */")
20:24:19perfectdrugah thats a shame
20:24:19gevaertspondlife: 090306d works
20:24:26 Join petur [0] (n=peter@rockbox/developer/petur)
20:24:37gevaertsso there's only 08 left
20:25:07 Quit AndyI (Remote closed the connection)
20:25:17pondlifeOK - "...and a bit of const-ing on source-only parameters" would break it?
20:25:22 Join AndyI [0] (n=pasha_in@
20:26:11 Quit Thundercloud (Remote closed the connection)
20:27:18CIA-38New commit by funman (r21118): FS #10253 : mkamsboot v1.0 ...
20:27:21 Quit MarcGuay_ (Remote closed the connection)
20:30:00gevaertshm, 090308 seems to work as well...
20:30:29pondlifeI wonder if you're playing the wrong song?
20:30:37pondlifeTry 090309 again, for sanity?
20:31:07CIA-38New commit by funman (r21119): Tag release 1.0 of mkamsboot
20:31:57funmanhum .. yes there is something wrong in svn post commit hook : "git-push: command not found" : git don't use git-xx but "git xx" for several releases now
20:32:24funmanperhaps this is why the git server isn't functioning
20:33:01pondlifeIs casting (int32_t *src[]) to (const int32_t *src[]) going to break things? I'm surprised the cast is needed.
20:34:40 Join thegeek [0] (
20:34:54gevaerts090309 works now. I'm wondering if the first time I tried that patch didn't mess up over a non-properly-removed file
20:35:24pondlifeSome of the patches were missing tdspeed.c/.h...
20:35:33pondlifeThat old svn add chestnut.
20:36:23pondlifesvn diff and compare with 090309 ?
20:36:54gevaertsa bit late now. This time I was more careful, so I'm sure it's correct now
20:36:56pondlifeI'm going to have to go shortly, I'm afraid.
20:37:42 Quit barrywardell (Remote closed the connection)
20:37:47pondlifeBut any news on which patch broke PP first would be most useful.
20:39:51LambdaCalculus37pondlife: 090528 was the first one I used.
20:40:06LambdaCalculus37That's also the one I gave to evilnick.
20:41:10gevaerts090316 doesn't apply cleanly :(
20:43:35pondlifeThe menu.c/.h mods were already committed
20:43:49pondlifesettings stuff might need resync though.
20:44:10gevaertsthe problem is in pitchscreen.c
20:44:40*gevaerts reverts r21013 for that file
20:44:56 Quit stoffel ("leaving")
20:45:31pondlifeI suspect it broke before 090316, is that not what you're thinking?
20:46:22*pondlife realises that 090309 worked...
20:47:40pondlifeThe changes after 090309 are (meant to be) UI-only stuff, nothing major.
20:48:04gevaertsit has to break *somewhere*
20:48:34pondlifehehe, no it doesn't
20:49:11gevaertsso why am I testing? ;)
20:49:37 Part ceatinge
20:50:15gevaertsimmediate data abort with 090316
20:54:44gevaertsnow also with 090309
20:59:38rmullIs there a way to shut off my rockbox ipod without waiting for the idle powerdown?
21:00:23AlexPrmull: The manual has full keymap info
21:01:07 Join pyro_maniac [0] (n=pyro@
21:01:23rmullAlexP: Thank you
21:01:35 Join SirFunk__ [0] (
21:01:47AlexPrmull: I would just tell you, but I don't have an ipod and we can both read :)
21:01:57AlexPrmull: I *think* it is long long play though
21:02:42pondlifegevaerts: Keep going - I have to go now, but am moeally supporting you from afar
21:02:54 Quit SirFunk_ (Read error: 110 (Connection timed out))
21:02:59gevaertspondlife: I have to take a break to...
21:03:55 Quit pondlife ("Leaving.")
21:04:26rmullAlexP: The manual keeps referring to a "Power" button, which the ipod doesn't have. I'll try holding down the play/pause button, I think that's what it was for Apple firmware
21:05:03evilnickrmull: That's correct
21:05:36gevaertsrmull: 3.1.2 says "long play" pretty clearly :)
21:05:36MarcGuayrmull: Where do you see the "power" button referenced?
21:05:41AlexPrmull: Could you point me at where it says power? That should be fixed
21:06:17rmullOh, baloney. I'm looking at the manual specific to a Sansa model :\
21:06:28AlexPhehe :)
21:06:33rmullRegardless, holding down play/pause does nothing, and idling does not turn it off
21:06:40rmullSo I must be missing something
21:06:40AlexPThe correct manual for your player is usually useful
21:06:49AlexPrmull: Long play it is
21:06:58AlexPHow long are you holding it for?
21:07:08rmull30s or so
21:07:14AlexPAnd (forgive the unsulting question) are you sure hold isn't on?
21:07:23MarcGuayrmull: Is the player frozen?
21:07:46rmullOne sec, it just worked.
21:07:56rmullI'm really sorry, guys.
21:08:02AlexPNo problem :)
21:08:44 Quit perfectdrug ("CGI:IRC (Ping timeout)")
21:09:56rmullMaybe I have a borked clickwheel or something - this thing has been in pieces a number of times
21:10:29AlexPCould well be
21:11:32rmullIs it too late to comment on how nice your fine product is? :D
21:12:35gevaertsit is, yes. The deadline was yesterday
21:13:37 Join Mathiasdm [0] (
21:14:18AlexPrmull: Ignore him, I'm always willing to hear how great Rockbox is (and by very very loose association I am) :)
21:14:28 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
21:15:05funmanBagder: I have some mkamsboot binaries to release, what can I do with them ?
21:16:11LambdaCalculus37funman: Need an OS X binary?
21:16:20 Join lee321987 [0] (
21:16:27funmanLambdaCalculus37: hum i have made one but now I see that sansapatcher is distributed as a .dmg
21:16:44LambdaCalculus37funman: My doing. :)
21:16:59 Quit lee321987 (Client Quit)
21:17:37funmanLambdaCalculus37: what is the advantage of a .dmg over a simple binary ? (I haven't used OSX in months)
21:19:10funmani would understand if there was several files, but for one isn't it simpler?
21:19:14linuxstbMainly so the permissions are correct.
21:19:22funmanoh perhaps we need to put the README with it
21:19:39linuxstbBut for sansapatcher and ipodpatcher, the default way to run it is to just "double-click" - no parameters or external files are needed.
21:19:55LambdaCalculus37funman: A .dmg is a disk image. We just pop the binary into one to make distribution easier.
21:20:01linuxstbSo a dmg allows Mac users to easily run it by double-clicking.
21:21:12LambdaCalculus37funman: Although that is a good idea to maybe include a README with the binary.
21:24:16linuxstbfunman: You could just email the files to Bagder, or put them on the web somewhere and post the URL here.
21:24:32 Quit moos (Read error: 54 (Connection reset by peer))
21:24:39 Join moos [0] (i=mustapha@rockbox/staff/moos)
21:25:57 Join Jaykay [0] (
21:27:46 Join Lynx0 [0] (
21:31:39 Quit Lynx_ (Read error: 60 (Operation timed out))
21:31:39 Nick Lynx0 is now known as Lynx_ (
21:35:12funmanLambdaCalculus37: can you build PPC binaries ?
21:35:42LambdaCalculus37funman: I should still be able to build universal binaries.
21:36:19funmanLambdaCalculus37: if you can build a dmg with the README and an universal binary for mkamsboot, that would help :)
21:37:11***Saving seen data "./dancer.seen"
21:37:23LambdaCalculus37funman: Sure thing. I'll do that soon and will post it to Flyspray.
21:37:49funmanLambdaCalculus37: you can just email it to me, so i'll give all the binaries to Bagder
21:38:06LambdaCalculus37funman: No problem.
21:40:41funmanby the way, does OSX need README to be named README.txt to open it with the text editor ?
21:42:36LambdaCalculus37funman: No, OS X will open a file named README in TextEdit.
21:45:22 Quit {phoenix} (Remote closed the connection)
21:48:11 Quit ender` (" Be nice to the USA, or they'll bring you democracy.")
21:48:12 Quit BryanJacobs ("null")
21:54:34 Join froggyman [0] (n=47ba0b80@gateway/web/cgi-irc/
21:54:48 Join calman__ [0] (
21:54:50 Quit calman__ (Read error: 104 (Connection reset by peer))
21:56:04 Join calman__ [0] (
21:56:22 Quit calman__ (Read error: 104 (Connection reset by peer))
21:57:12 Join calman__ [0] (
21:57:15 Quit calman__ (Read error: 104 (Connection reset by peer))
21:57:57 Join webguest77 [0] (n=3e0cdf13@gateway/web/cgi-irc/
21:58:06 Quit webguest77 (Client Quit)
21:58:51 Quit LambdaCalculus37 (" ajax IRC Client")
21:59:42 Join calman__ [0] (
21:59:46 Quit calman__ (Read error: 104 (Connection reset by peer))
22:01:48funmansomeone feels like updating SansaM200V4Port ?
22:02:28funmanhum wiki pages for C200v2 E200v2 Fuze say rockbox doesn't run as well
22:05:35CIA-38New commit by alle (r21120): Describe the magic of search_albumart_files (still not complete)
22:07:45 Join Thundercloud [0] (
22:08:55funmani have written on all of them that 'Rockbox somehow runs, but still isn't released'
22:09:28 Join midgey|w [0] (i=836b0053@gateway/web/ajax/
22:10:26 Quit calman_ (Read error: 110 (Connection timed out))
22:12:46 Part rmull
22:23:14 Quit Jaykay ("ChatZilla 0.9.84 [Firefox 3.0.10/2009042316]")
22:27:35CIA-38New commit by FlynDice (r21121): AMSSansa: Display AMBA Bus mode and mmu, icache, & dcache status on View HW info page
22:32:05 Quit perrikwp (" ajax IRC Client")
22:32:43 Join matsl [0] (
22:36:11 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
22:38:38 Join stripwax [0] (
22:38:57 Quit Thundercloud (Remote closed the connection)
22:39:18 Quit JdGordon| (" ajax IRC Client")
22:50:09 Join ender` [0] (
22:50:41 Quit perrikwp (" ajax IRC Client")
22:51:15 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
22:51:20 Quit stripwax ("")
22:51:57 Join matsl_ [0] (
22:52:19 Quit matsl (Read error: 110 (Connection timed out))
22:52:27 Quit matsl_ (Client Quit)
22:52:34 Quit HellDragon (Client Quit)
22:52:38 Join matsl_ [0] (
22:54:27 Join joa [0] (
22:55:20 Join tessarakt [0] (
22:56:56 Quit ender` (" Some people like my advice so much that they frame it upon the wall instead of using it. -- Gordon R. Dickson")
22:58:56 Nick Zarggg_ is now known as Zarggg (
22:59:05 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
23:04:22linuxstbfunman: You didn't "svn mv" all the files? Also, sorry for not noticing when looking at your patch, but some of your bracket changes aren't in line with docs/CONTRIBUTING (i.e. brackets for start of functions are on their own line)...
23:05:07linuxstbPlus svn:keywords...
23:07:59LambdaCalculus37funman: I'm going to send you the OS X binary of mkamsboot now (no README included yet).
23:10:09 Join mcuelenaere [0] (
23:10:44 Quit pyro_maniac ("Leaving.")
23:12:00funmanlinuxstb: i used the same style in the whole file (i think it it the style you use)
23:12:18LambdaCalculus37funman: Want to see that it works on PPC?
23:12:30funmanlinuxstb: i git mv the files but somehow the move was lost for some files, and I can't set svn:keywords with git-svn ..
23:12:34 Quit midgey|w (" ajax IRC Client")
23:13:07 Quit joa_ (Read error: 110 (Connection timed out))
23:13:29funmanLambdaCalculus37: well why not, i ran mkamsboot on all supported targets to check md5sums of produced bootloaders
23:13:40Mikachugit mv is the same as just moving the files yourself, it doesn't do svn mv for you when you dcommit
23:13:52linuxstbfunman: I'm very inconsistent when it comes to brackets, but docs/CONTRIBUTIING says brackets at the start of functions must be on their own line.
23:13:52funmanLambdaCalculus37: i'll upload binaries and scripts for you to test if you want
23:14:29funmanlinuxstb: ah for the functions with 8 or more arguments I had trouble with using something limited to 80 columns and looking well..
23:14:47 Quit bmbl ("Woah!")
23:15:17LambdaCalculus37funman: Sure thing.
23:15:43linuxstbfunman: I'm talking about the "{" at the start of a function declaration - you moved lots of them to the same line as the function name in mkamsboot.c
23:16:10 Join ender` [0] (
23:16:15funmanhum .. sorry for that
23:17:18 Join Thundercloud [0] (
23:18:49linuxstbMikachu: Is there any way to do "svn mv" with git? It's annoying to lose the history...
23:19:03Mikachunot that i know
23:19:18Mikachuin native git it generates that info on demand, so if you import the svn repo you should get it back :)
23:19:26CIA-38New commit by funman (r21122): mkamsboot : brackets at the start of functions must be on their line - thanks to linuxstb for noticing
23:19:31Mikachui don't use git-svn a lot for committing so there might be something i don't know about
23:22:33funmanlinuxstb: well some files were moved no ?
23:23:23funmanextract_fw.c and nrv2e_d8.S
23:24:02funmanI moved dualboot.S and bin2c.c but somehow the move was lost (perhaps because too much changed into these files)
23:26:49funmanLambdaCalculus37: (or mkamsboot_test.tar if you want to download a single file). Just run ./ and compare with the content of 'sums' file
23:32:45 Quit VytenisS (Remote closed the connection)
23:37:14***Saving seen data "./dancer.seen"
23:37:22 Nick fxb is now known as fxb__ (
23:39:08LambdaCalculus37funman: It's not allowing me in.
23:45:52 Quit ender` (" The problem with political jokes is they get elected. -- Henry Cate, VII")
23:46:35funmanLambdaCalculus37: now it should be ok
23:48:49 Quit evilnick (" ajax IRC Client")
23:50:37 Quit mewshi ("Leaving")
23:51:46 Join safetydan [0] (n=deverton@rockbox/developer/safetydan)
23:57:41 Quit mirak ("Ex-Chat")
23:58:02 Quit Thundercloud (Remote closed the connection)

Previous day | Next day