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

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

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

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2008-05-05

00:00:05Bagder53 fine megs
00:01:05 Quit perrikwp (" ajax IRC Client")
00:05:00MegarainHey, any one know if the apple Firewire/usb combo cables can cause issues during restoration?
00:05:41 Quit BlakeJohnson86 (Read error: 104 (Connection reset by peer))
00:06:03krazykitMegarain, as long as they work for normal transfers, they shouldn't cause any issue
00:06:21 Join BlakeJohnson86 [0] (
00:06:36MegarainThat's just it, every time I've tried them for normal sync, after the ~256mb of data, I get error 69
00:10:02 Quit scorche|sh ("leaving")
00:11:31 Join styleism [0] (
00:11:57 Quit bughunter2 (Read error: 104 (Connection reset by peer))
00:12:13 Join bughunter2 [0] (
00:14:35 Join scorche|sh [50] (n=scorche@rockbox/administrator/scorche)
00:16:51 Join DerDome [0] (
00:17:09 Quit DerDome (Client Quit)
00:18:40 Quit faemir (Read error: 104 (Connection reset by peer))
00:19:06 Quit bluebrother ("leaving")
00:19:21 Join faemir [0] (
00:19:29 Quit gevaerts ("it's dark outside")
00:19:51 Quit faemir (Remote closed the connection)
00:20:23 Join faemir [0] (
00:22:05 Join LambdaCalculus37 [0] (n=1800dac6@gateway/web/cgi-irc/
00:22:23 Quit davina (Remote closed the connection)
00:22:30LambdaCalculus37toffe82: Ping
00:22:57LambdaCalculus37toffe82: Did you receive my beast yet?
00:23:12 Quit bertrik ("bye")
00:24:09 Quit shotofadds (Read error: 110 (Connection timed out))
00:24:13 Quit Megarain ("Leaving.")
00:24:28toffe82no, I don't know what happened , it was out for delivery on friday and they take it back to their office, I should have it on monday
00:24:51toffe82I will call them on monday morning
00:25:01LambdaCalculus37Okay, let me know as well.
00:26:08LambdaCalculus37Bye for now!
00:26:10 Quit LambdaCalculus37 (Client Quit)
00:28:15 Quit fdinel (Read error: 110 (Connection timed out))
00:37:43toffe82I have problem compiling the S, the bootloader is ok but rockbox give an error on rockbocklogo.h, no such file or directory, I just made an svn update , any idea ?
00:38:06toffe82rockboxlogo ^^
00:39:08 Quit gregzx (Connection timed out)
00:39:46toffe82jhMikeS: ping
00:40:49Nico_Ptoffe82: try cleaning your build
00:43:19 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
00:43:49 Quit ompaul (Client Quit)
00:43:52toffe82it is working
00:46:03 Join ReKleSS [0] (
00:46:44 Join shotofadds [0] (n=rob@rockbox/developer/shotofadds)
00:49:26 Join fyrestorm [0] (
00:49:27 Quit BlakeJohnson86 (Read error: 104 (Connection reset by peer))
00:50:17 Quit styleism (Read error: 110 (Connection timed out))
00:50:44 Join BlakeJohnson86 [0] (
00:52:23 Quit waldo ("Konversation terminated!")
01:01:00 Quit Mathiasdm ("Yuuw!")
01:04:59 Join spiorf [0] (
01:11:39 Join aliask [0] (n=chatzill@rockbox/developer/aliask)
01:12:10aliaskjhMikeS: Congrats on the gigabeat sound work!
01:14:23aliaskBut I have some slightly unsavoury behaviour happening. I initialised the database, enabled dircache and rebooted the player and it appears to have corrupted the filesystem.
01:15:52aliaskUnder rockbox my music folder is empty, but some files in the root show. When connecting via usb the whole media partition is empty
01:17:32 Quit Nico_P (Remote closed the connection)
01:20:52 Quit BlakeJohnson86 ("Leaving.")
01:22:55 Quit shotofadds ("Leaving")
01:24:03JdGordonXavierGr: hey, I found your patch to merge the alarm and time screens last night and thought i'd see if you wanted to finish it?
01:24:50JdGordonand maybe even redraw the screen using viewports instead of the gui_textarea stuff...
01:34:01 Quit robin0800 (Read error: 110 (Connection timed out))
01:35:58***Saving seen data "./dancer.seen"
01:51:05XavierGrJdGordon: wow that's an old patch indeed I completely forgot about it
01:51:22XavierGryou have plans for it at the moment?
01:51:56 Join Shaid [0] (
01:53:00 Join BlakeJohnson86 [0] (
01:53:33 Quit faemir (Remote closed the connection)
01:53:34 Quit spiorf (Read error: 104 (Connection reset by peer))
01:56:10XavierGrI didn't finish it mainly because I once considered to rewrite the alarm drivers to include date
01:56:54XavierGrso at the moment the screens are unified but seconds are left on the alarm screen which should be removed, at least for now that the alarm settings don't include seconds
02:06:33 Quit aliask ("ChatZilla 0.9.81 [Firefox 3.0b5/2008041514]")
02:07:04 Nick fxb is now known as fxb__ (
02:14:39 Quit ender` (" The statistics on sanity are that one out of every four Americans is suffering from some form of mental illness. Think of y")
02:18:32 Join perrikwp [0] (i=98212fe2@gateway/web/ajax/
02:28:37 Join HiddenShadow [0] (n=ce747465@gateway/web/cgi-irc/
02:30:18HiddenShadowDoes either rockbox or linux have the capability to make mp3's increase in speed? (im not talking about time stretching, just faster/slower, doesn't really matter if it changes pitch when it speeds up)
02:30:34HiddenShadowand decrease*
02:33:03ReKleSSsox can do it
02:33:16krazykitrockbox has this capability. refer to the manual for how to do it
02:42:28HiddenShadowsweet now we got 3 decks on our dj set, vinyl turntable x2, + ipod :D
02:43:23 Join Megarain [0] (n=Rainer_P@
02:43:39MegarainHey, how many paritions should an 4G grayscle ipod have?
02:44:13HiddenShadowatleast 10, lol jk, dont listen to me
02:44:35krazykitHiddenShadow, please don't do that.
02:44:57MegarainCause mine has four, and has been having issues.......wondering if that is causing part of the trouble.
02:45:08scorche|shdefine "issues"
02:46:18Megaraincorrupting with now known cause.
02:46:28MegarainWith both Rockbox and the iTunes firmware.
02:50:00Megarainso I'm trying to trace down the issues.
03:03:00 Join m0f0x_ [0] (
03:05:34 Quit m0f0x (Nick collision from services.)
03:05:42 Nick m0f0x_ is now known as m0f0x (
03:10:23 Quit merbanan (Remote closed the connection)
03:11:42 Join kb_ [0] (
03:12:36 Join JdGordon|uni [0] (i=768aa360@rockbox/developer/JdGordon)
03:13:25JdGordon|uniXavierGr: (sorry, had to run to uni...) no I dont have plans for it... but I want to remove the gui_textarea code which the clock screens use...
03:13:40kb_I'd like to dynamically generate playlists according to some specific criteria. Would it be best to alter the rockbox code to do this, or create a plugin to make these playlists?
03:14:07JdGordon|uniwhat criteria?
03:14:47kb_basically, to play songs from a certain directory for a certain amount of time, then switch to another; I wrote some python code that does it on my desktop computer, and I'm wondering how to best port it to rockbox
03:15:39kb_(more specifically, I have one language per directory, and weight the frequency I play them at, as well as how long to listen to any particular one in a row)
03:18:28 Join ryanakca [0] (n=ryan@ubuntu/member/ryanakca)
03:19:06ryanakcaIs there any way to build the database from the CLI / a linux comp? On do I need to wait for a few hours while it builds?
03:19:41JdGordon|unithere is a tool to do it under the tools/ folder... dnno if its kept up to date or not though
03:19:50JdGordon|unikb_: yeah, duno howd you go about doing that in rockbox
03:20:19kb_JdGordon: hm, ok; maybe I'll dig into how the current dynamic playlist stuff is done and do a variant on that
03:20:19ryanakcaJdGordon|uni: thanks :)
03:21:05Lloreanryanakca: It shouldn't take hours, though...
03:22:03HiddenShadowin rockbox (ipod 5g) when i set it to Charge during USB connection, it just keeps going back to disk it a known bug, or am i doing something wrong?
03:22:06ryanakcaLlorean: hmmm... *shrugs*... I had it plugged in, and it just seems to stall after a certain point... and then if I restart the iPod, it starts again at 0 and stalls at that certain point
03:23:04LloreanThat stall would indicate a problem... not that you need to wait longer.
03:23:53ryanakcaLlorean: any idea what that problem would be? The install is a daily from mid-september '07 iirc...
03:24:57ryanakca(1st gen nano)
03:25:04LloreanWell first, we really don't support old versions.
03:25:07JdGordon|uniget a new build
03:25:32LloreanBut odds are it's something in your metadata. Depends an awful lot, could be large album art, or perhaps AACs not optimized for streaming.
03:28:05kb_Llorean: by the way, any ideas/advice about the question I asked?
03:29:42ryanakcaLlorean: hmm... thanks
03:33:40 Quit Thundercloud (Remote closed the connection)
03:36:02***Saving seen data "./dancer.seen"
03:37:23ryanakcaLlorean: hmmm... fresh install worked :)
03:37:54kb_ryanakca: congrats :)
03:38:36ryanakcalol, thanks :)
03:38:55HiddenShadowanybody gonna answer my question above? :\
03:39:33kb_i would if i had any idea
03:40:53 Quit Megarain ("Leaving.")
03:41:47HiddenShadowCan't get "Charge during USB connection" to work, it keeps going back to disk mode
03:43:45JdGordon|uniyou have to hold down a button while pluggin in or it reboots
03:52:33HiddenShadowoh, i thought it was automatic, guess it was kept the same after that patch
04:05:23 Quit JdGordon|uni (" ajax IRC Client")
04:06:52 Quit HiddenShadow ("CGI:IRC (Ping timeout)")
04:07:11 Quit kb_ ("Leaving")
04:08:11 Part pixelma
04:16:29 Quit parafin (
04:16:47 Nick gnakinklp is now known as Anakin (i=0@
04:25:17 Join aliask [0] (n=chatzill@rockbox/developer/aliask)
04:32:00 Quit AndyIL ()
04:40:43 Join AndyI [0] (n=pasha_in@
04:46:44 Join cool_walking_ [0] (
04:47:53 Quit |AhIoRoS| ("Abandonando, see you")
04:48:28 Join jpt9 [0] (
04:48:39*jpt9 wants a PC port of brickmania :-)
04:48:45jpt9(Yeah, I know there are simulators.
04:48:55jpt9it'd be nice if the simulators could zoom to 3x or 4x size.
04:49:05jpt9and possibly map mouse movement to left/right
04:51:39XavierGrJdGordon: I will try to have a look on viewports tomorrow, I don't know much about them
04:54:20 Join miepchen^schlaf [0] (
04:55:38 Join homielowe [0] (n=eric_j_l@
05:04:49cool_walking_jpt9: There must be a million breakout clones for PC out there...
05:05:03scorche|shand every one is offtopic ;)
05:05:35cool_walking_Damn, that was fast. Nice, scorche.
05:09:03 Quit homielowe ()
05:09:44 Quit miepchen^schlaf_ (Read error: 110 (Connection timed out))
05:09:44 Join homielowe [0] (n=eric_j_l@
05:17:18 Quit jpt9 (Read error: 110 (Connection timed out))
05:36:07***Saving seen data "./dancer.seen"
05:37:48 Quit Horscht ("I got raided by the FBI and all i got is this lousy quit message")
05:51:45 Join simonrvn [0] (i=simon@unaffiliated/simonrvn)
05:52:48 Join parafin [0] (i=parafin@
05:53:24 Quit JdGordon ("Konversation terminated!")
05:55:07 Join JdGordon [0] (
06:12:50 Join jpt9 [0] (
06:36:33 Quit csc` ("If you can't do anything smart, do somethin right ~Serenity")
06:47:12 Quit jpt9 (Read error: 110 (Connection timed out))
07:17:15 Quit JdGordon (Read error: 104 (Connection reset by peer))
07:18:24 Quit cool_walking_ (Remote closed the connection)
07:18:38 Join cool_walking_ [0] (
07:19:45 Join JdGordon [0] (
07:29:43 Quit parafin ("So long and thanks for all the fish")
07:29:45 Join parafin [0] (
07:33:46 Quit Seed (Nick collision from services.)
07:33:54 Join Seed [0] (
07:36:08***Saving seen data "./dancer.seen"
07:37:18 Part toffe82
08:02:07 Join Rob2223 [0] (
08:03:48JdGordon:( I dont think the menu code will work too well with inline settings.. well.. not without a bit of fiddling :/
08:04:10 Quit BigBambi (Remote closed the connection)
08:08:16 Join DerDome [0] (
08:09:21 Quit DerDome (Nick collision from services.)
08:09:22 Join DerDom1 [0] (
08:09:30 Nick DerDom1 is now known as DerDome (
08:09:42 Quit miepchen^schlaf ()
08:15:40 Join grndslm [0] (
08:19:41JdGordon... this would be easier if we didnt have dynamic menus :/
08:20:20 Quit Rob2222 (Read error: 110 (Connection timed out))
08:20:37amiconnJdGordon: Why would we want inline settings now?
08:21:05JdGordonsome people do
08:21:14*amiconn thinks they make no sense on most target's LCDs
08:21:16JdGordonsome have wanted them for gaes
08:21:42amiconnOn small displays they would be confusing
08:21:59amiconnMaybe on 320x240 and higher they would work well enough
08:22:08JdGordonit wold fo course be disable-able (if not disabled by default)
08:22:28amiconnTake a look at the recording trigger screen - it's really ugly
08:22:34JdGordonI know
08:23:49amiconnImho they shouldn't be included at all on smaller displays (if we want them at all - I very much prefer the current implementation of settings)
08:25:15JdGordonI want to do it so only the value of the selected item is actually displayed
08:25:53JdGordonbut I have a strange feeling it wont happen at all because the menu code isnt perfect
08:26:20amiconnDisplayed would be okay (except on charcell), but changing is another thing
08:26:35JdGordonof course... 1 thing at a time
08:27:06 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
08:27:36amiconnAdjusting inline settings has another fundamental problem apart from the confusing screen layout on smaller screens: button assignment
08:28:26JdGordonit would work fine with left/right on all targets except the ondio
08:28:33JdGordonwhose screen is too small for it anyway
08:29:16JdGordonbut even just displaying the value is nice.. a few less presses to find out what the value is
08:30:32 Quit DerDome (Read error: 110 (Connection timed out))
08:32:09*amiconn wonders how often JdGordon needs to go through settings and find out a value
08:33:57amiconnWhat's really in need of rework is the eq screen...
08:36:25amiconn...and the recording screen
08:36:30JdGordonand fm
08:36:33JdGordonand time
08:36:56JdGordonand really.... apps/screens.c
08:38:52 Quit BHSPitMonkey (Remote closed the connection)
08:39:31amiconnWhat's wrong with the fm and time screen?
08:41:20JdGordontime isnt so bad, fm is ugly
08:41:24 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
08:41:59amiconnFM might not be perfect, but it's usable (no sysfont, for instance)
08:42:39amiconnBtw, there's a fundamental problem with many read-only info lists: they are currently useless on smaller displays
08:42:54 Quit grndslm ("Leaving")
08:43:30amiconnProblem is that the lines are cut before the relevant information, and the lines don't scroll
08:44:09JdGordonthe lines dont scroll?
08:44:23 Join grndslm [0] (
08:45:34JdGordonwhich ones? just add info.scroll_all = true; before the simplelist_init() call to fix them
08:46:23amiconn'View OS stacks', 'View partitions', 'View disk info', 'View database info'
08:47:54amiconn'Rockbox Info' does scroll on recorder, but iirc it seems to be inconsistent
08:48:01amiconnNeed to check that
08:48:31amiconn'Rockbox Info' isn't handled as read-only lists though. Same goes for the tag info viewer
08:48:59JdGordonrockbox info is handled as read only unless we arnt talking
08:52:10amiconnWhy does 'Rockbox Info' show a cursor then?
08:53:35JdGordonyou havnt disabled voice
08:54:52 Join davina [0] (
08:54:53JdGordonany ideas who to put (c) for the pitchscreen code?
08:56:19JdGordonr9146 suggests midkay, "Reworked pitch screen for easier adjustment and better looks." ?
08:57:39JdGordonor just keep zagor from screens.c?
08:58:10 Join dunder [0] (
08:59:43 Join Rob2222 [0] (
09:00:11Mouser_XSound on the S is great. Battery life not so much...
09:00:15linuxstbIt seems lots of people have changed that code over time, so I would probably just keep Zagor's name there.
09:00:45*Mouser_X realizes that the battery life will improve.
09:01:02Mouser_X(Lots of "easy" fixes to implement.)
09:03:12*JdGordon hates modifying files with Björn's name in the (c).. it never works right :p
09:04:32amiconnJdGordon: Hmm, then you actually meant the opposite...
09:04:56amiconnOf course I have voice enabled on recorder for car mode (which goes with a big font)
09:05:07 Join LinusN [0] (n=linus@gateway/web/cgi-irc/
09:05:09JdGordonthen it should show the selector
09:05:20amiconnI guess this is necessary for selecting the items to talk?
09:05:33JdGordon[16:53:50] <JdGordon> you havnt disabled voice
09:05:42amiconn[08:48:59] <JdGordon> rockbox info is handled as read only unless we arnt talking
09:05:54amiconn^ "unless we aren't" ...
09:06:00*JdGordon is a bit skitzo today :p
09:06:05amiconnKIinda double negation
09:06:28JdGordondamn that extra t
09:06:32JdGordonand n
09:07:09amiconnWith voice disabled, 'Rockbox Info' is indeed read-only, but doesn't scroll its items
09:07:35 Quit grndslm (Read error: 113 (No route to host))
09:07:45JdGordonso what your asying is.... read only lists should always scroll all items?
09:08:12amiconnI think they should (only if the items are that long of course)
09:08:20amiconnHow else would you be able to read them?
09:08:43 Join petur [50] (n=petur@rockbox/developer/petur)
09:09:45 Quit perrikwp (" ajax IRC Client")
09:10:51Mouser_XMagic, or telepathy.
09:11:32*JdGordon votes for telepathy
09:12:17*Mouser_X does as well, since magic can get messy really fast.
09:12:42amiconnKinda extreme case: Right now I only see 'Version: r' in 'Rockbox Info' on the Player
09:12:57amiconnNot very helpful...
09:13:01JdGordonfixed... wait for the next build round
09:13:11linuxstbJdGordon: You didn't use "svn cp" to move the pitchscreen code?
09:13:33JdGordonI agreed that it wasnt the nicest way and not really worth it
09:13:54linuxstbAgreed with who?
09:14:08JdGordonwith what I thought I read :p
09:14:17JdGordonthe log says where it came from if its ever needed
09:15:10linuxstbYes, but "svn annotate" now doesn't work...
09:15:17 Join perrikwp [0] (i=9821738c@gateway/web/ajax/
09:15:34amiconnHmm. I wonder whether we should try to reclaim the one (precious) column on Player which is now wasted in read-only lists
09:15:57amiconn(the one containing the arrow cursor in non-read-only lists)
09:16:08 Join OlivierBorowski [0] (
09:16:38amiconnWould mean special casing though, which isn't nice...
09:16:49JdGordonnot do muvh
09:17:38 Quit Rob2223 (Read error: 110 (Connection timed out))
09:19:33 Quit JdGordon ("Konversation terminated!")
09:21:19 Join JdGordon [0] (
09:22:50amiconn[09:16:49] <JdGordon> not do muvh <= ?
09:24:30 Quit davina (Remote closed the connection)
09:25:21scorche|shthough it still doesnt make sense...
09:29:21linuxstbJdGordon: Red for the ZVM build (if you didn't notice...)
09:30:34amiconnWrong struct names...
09:36:10***Saving seen data "./dancer.seen"
09:40:39 Join Bagderr [0] (n=daniel@gateway/web/cgi-irc/
09:41:04 Nick Bagderr is now known as B4gder (n=daniel@gateway/web/cgi-irc/
09:42:56 Quit Llorean (Read error: 104 (Connection reset by peer))
09:43:36 Join Llorean [0] (
09:46:31 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
09:47:30 Join Synergy6 [0] (
09:50:23*linuxstb fixes the ZVM red
09:51:07B4gdercolorful delta table
09:51:34*linuxstb wonders if shotofadds minds removing the #warnings from the D2
09:52:08B4gderwould be nice to reach complete green again
09:53:56linuxstbB4gder: BTW, I don't know if the build system needs to know, but my server died a couple of weeks ago, and is still off-line.
09:54:24B4gderit doesn't need to know, although I can of course save it from trying that server
09:55:36 Quit dunder ("Leaving")
09:56:20linuxstbB4gder: Do you know if there's a reason the different svn pages on www use different timezones? i.e. is it hard to fix?
09:56:59B4gderI don't know the reason but it shouldn't be that hard I think
10:00:39linuxstbB4gder: Is the script in SVN?
10:00:49B4gderwhich one?
10:00:59JdGordonlinuxstb: nuts, thanks for fixing that
10:01:16linuxstbYou're welcome.
10:01:30 Quit Shaid (Read error: 110 (Connection timed out))
10:01:47*JdGordon heads back into windows
10:01:50JdGordonlinux even :p
10:01:54 Quit JdGordon (Read error: 104 (Connection reset by peer))
10:03:44 Join JdGordon [0] (
10:06:06JdGordonnaming conventionas ftw :p
10:20:49linuxstbB4gder: Do you have an opinion about the legal aspect of ROKR simulator build? i.e. linking to a proprietory version of the qt/embedded libs, and using reverse-engineered headers?
10:22:04B4gderI don't see any problems with that
10:22:15B4gderas long as we don't distribute dubious files
10:22:27 Quit homielowe ()
10:22:29JdGordonthere is also the speculation that the dubious files were nicked...
10:22:36linuxstbBut isn't that just the same as GPL'd apps linking with old (non-GPL'd) versions of Qt?
10:22:55B4gderbut linking with non-GPL stuff is not a problem when not distributed
10:24:12linuxstbWhat do you mean?
10:24:30 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
10:25:22B4gderI'm saying that linking GPL code with whatever is fine as long as the results isn't distributed
10:25:59B4gderso if you get the proprietary stuff from source X and rockbox from us, and link them into product Y that's fine as logn as you don't copy it to others
10:27:11linuxstbSo people could create proprietory plugins, and that's fine as long as they're distributed separately (and don't use Rockbox headers to build them)?
10:27:54B4gderyou can't build rockbox plugins without them being derivates of Rockbox so I don't think people can
10:28:29linuxstbI'm sure you could build a "rockbox-compatible" plugin independently.
10:28:51B4gdertrue, and if you do that could probably be non-GPL
10:31:24scorcheB4gder: have you approved sdoyon yet on the mentor app?
10:31:57 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
10:32:07B4gderuh, I have now!
10:32:47scorchethanks...Leslie came to me a week or so ago asking if i can vouch for him :)
10:32:52 Quit basti (Remote closed the connection)
10:34:53peturdo all slots no have a co-mentor?
10:35:45B4gderI don't think the ARM emulator does
10:36:22B4gderand in fact not Saratoga's either
10:37:04JdGordonhow on earth could someone make the compatible headers needed for plugins other than by straight copying them? imo thats a weak spot in the gpl
10:37:34ReKleSSJdGordon: reversing a plugin?
10:37:38B4gderyou could reverse engineer an existing binary and generate a blob quite similar to that
10:37:46B4gdershouldn't be very hard
10:38:00B4gderof course that is a grey area
10:38:10JdGordonthe blob is the easy bit... the api headers im talking about
10:38:33ReKleSSJdGordon: the ps2dev guys pulled it off to make the PSP sdk
10:38:48ReKleSS(it helps that they found a binary with debug info left in...)
10:38:54B4gderwell, a non-GPL plugin wouldn't use much of the plugin api I would think
10:38:57gevaertsI seem to remember that at least in some jurisdictions interfaces are not copyrightable, so the plugin api minus comments and documentation would apply
10:39:22linuxstbThere are also macros and inline functions...
10:39:38JdGordonand defines, and datatypes
10:39:43JdGordonand and...
10:40:20B4gderit would be work sure, and keeping up with changes would be painful, but it certainly wouldn't be impossible
10:40:48*gevaerts isn't actually planning to make a non-GPL pluging, so he isn't planning to work out the details
10:41:25B4gderand if _I_ were to do it, I'd write a fine GPL wrapper plugin that would load the proprietary one just like the linux drivers do their work-arounds
10:41:50B4gderand just like Neuros want to circumvent the VLC GPL...
10:42:26linuxstbA very minimal plugin would just need lcd_update(), the pointer to lcd_framebuffer, button_get and yield.
10:42:33linuxstbPlus some startup code I guess.
10:42:49linuxstbOr does the plugin loader do that?
10:42:50JdGordonwouldnt that plugins api need to be gpl also though?
10:44:10*linuxstb doesnloads the ROKR SDK - a .tar.gz file inside a .rar...
10:44:53linuxstbNo idea where that came from ;)
10:45:16scorcheso how exactly would this work out?...i mean it is a bit more than just a plugin...
10:45:20petur"to download something that doesn't work" ?
10:47:28amiconnJdGordon: If the wrapper would alsp be gpl, it wouldn't work. But I guess lgpl would work...
10:49:52linuxstbAt least in the SDK I've just downloaded, there are a lot of original Qt headers, with a "Q Public License" statement at the top of each .h They appear to date from 2004.
10:50:10B4gderwhat qt version?
10:50:49linuxstbAH, there's also a GPL statement...
10:50:57 Quit cool_walking_ (Remote closed the connection)
10:51:03linuxstbSo they're dual-licensed.
10:51:25B4gderthen we should be fine
10:51:48 Join basti [0] (i=bnc@
10:53:10linuxstbThe SDK itself has an odd home-brew license -
10:54:37B4gderI'm not so sure they can limit what the output may be used for
10:54:57Nico_PI'm restoring Zagor's copyright on apps/gui/gwps-common.[ch]. Can I make it UTF8?
10:54:57B4gderI mean paragraph 5
10:55:14gevaertsDoesn't matter, just use paragraph 4 to make a GPL version first
10:55:16B4gderNico_P: yes, I think so
10:55:32B4gdergevaerts: haha, indeed
10:55:37linuxstbgevaerts: Or even BSD...
10:55:41Nico_PB4gder: maybe I should also make his other copyrights UTF8?
10:55:48linuxstbgevaerts: Or not....
10:55:58B4gderNico_P: yes, we should make them consistent
10:56:12Nico_PI'll do that
10:56:15B4gderalthough having Zagor
10:56:25B4gder's work consistent seem a bit strange ;-)
10:56:29linuxstbWould viewvc.cgi need tweaking? I can't remember what encoding it uses.
10:56:53B4gderit might, or perhaps apache
10:57:50linuxstbHmm, there seem to be a large collection of .so files in the SDK...
11:09:41Nico_Pviewvc.cgi uses UTF8 I think
11:10:54 Join simonrvn_ [0] (i=simon@unaffiliated/simonrvn)
11:13:45*Nico_P ends up running iconv on the whole codebase
11:14:40 Join ender` [0] (
11:18:54 Quit ender` (Read error: 104 (Connection reset by peer))
11:22:03 Quit simonrvn (Read error: 113 (No route to host))
11:22:04 Nick simonrvn_ is now known as simonrvn (i=simon@unaffiliated/simonrvn)
11:22:19Nico_Phmm I don't know about things like firmware/drivers/lcd-charset-player.c... should I go all the way?
11:24:50linuxstbNico_P: Maybe ask amiconn about that one...
11:25:36linuxstbBut if we're moving all source files to utf-8, then it probably should be...
11:26:16linuxstbAre there any other non-ascii comments ?
11:26:22Nico_PI think the move to UTF8 is long overdue
11:26:27Nico_Pyes, they are a few
11:26:50Nico_Palso there is something that puzzles me in tools/creative.c
11:26:51linuxstbDo you have a patch?
11:27:01Nico_PI'll make one
11:27:29linuxstbThe device_info struct?
11:27:59Nico_Pyes. the copyright char
11:29:22Nico_Plinuxstb: here's what I've done so far:
11:30:34Nico_Phmm there's something wrong in debug_menu.c
11:30:50linuxstbWhen you commit, it might be a good idea to add "All source files use UTF-8" to docs/CONTRIBUTING
11:31:05Nico_Pwill do
11:31:35linuxstbAlso, I notice you've changed a couple of (C) headers with the same patch. Would be nicer to do it separately.
11:31:54Nico_Pprobably, yes. that's actually what got me started
11:32:08linuxstbI assume that's a "mu" character in debug_menu.c
11:32:29Nico_Pyes. it seems to be broken in svn too
11:34:01linuxstbI think I would commit everything apart from creative.c and debug_menu.c
11:36:04linuxstbI'm not sure what the "boostquota" comment is meant to say - Buschel cocmmitted it...
11:36:08linuxstb(in debug_menu.c)
11:36:11***Saving seen data "./dancer.seen"
11:37:22linuxstbAnd jdgordon did the nsec_units
11:38:46Nico_Pit seems he changed that line in r15221
11:39:55*B4gder had some progress in the tshirt department for devcon
11:42:10 Join Eden [0] (
11:42:20Edenhay guys, just installed rockbox, need a little help
11:42:48Edencan i get any help
11:42:55linuxstbSure, just ask.
11:43:34 Join ender` [0] (
11:43:34Edenum, wen i go into database on rockbox on my ipod, it says building database and takes so long, is that normal?
11:43:47Edeni cant get past it so far
11:43:56Edenive left it for max 5 mins
11:44:12linuxstbWhich ipod?
11:44:18Eden5th gen 30gb
11:44:26Edenblack :D
11:44:31Edencos the colour helps
11:44:33linuxstbCan you still feel the hard disk doing things, or has it stopped?
11:45:17linuxstbAnd then you've restarted it by holding PLAY for a few seconds to safely power off, and then turning it back on?
11:45:27Edenexaclty wat it says: Building Database... 416 found (PREV to return)
11:45:53Nico_Plinuxstb: JdGordon broke that line of debug_menu.c in r15221. I don't know if it would be ok to put the UTF8 mu char in there
11:45:55Edenyea, and then i got back into it and asks me if i want to do that again or not, and if i click no i get kicked to main menu
11:46:56linuxstbEden: I don't know - maybe someone who uses the database more than me (i.e. at all) can help...
11:47:06Edenwhat do u use?
11:47:12linuxstbI just use the file browser.
11:47:13*gevaerts uses 'you'
11:47:27B4gderme? ;-P
11:47:44gevaertssure :)
11:47:47Edenok i got to database, its says: Database is not ready, Initilize now
11:48:38linuxstbYou could try deleting all the files ending in ".tcd" in the .rockbox folder
11:48:42 Quit saratoga ("CGI:IRC (Ping timeout)")
11:49:27Nico_PBagder: what about the docs ?
11:49:50B4gderyou mean utf8-wise?
11:50:29B4gderI'm at loss there, we need to speak to the docs ninjas about that
11:50:53Nico_PI mean the files in docs/
11:51:39Edenwith rockbox, do i need to convert any of the files, or can i leave as is, eg mp3, wav, avi, mpg, etc
11:51:56markunEden: you need to convert the video files.
11:51:58 Quit ompaul (Remote closed the connection)
11:52:00Edento wat?
11:52:33Edenum, wen i go into database on rockbox on my ipod, it says building database and so far havnt got past that, is that normal?
11:53:09markunI don't know much about the database. Sometimes I had to reboot before it stopped saying that.
11:55:15*linuxstb wonders if _any_ devs apart from Slasheri use the database
11:55:29markunand does he even use it? :)
11:56:25*B4gder doesn't use it
11:56:53Edeny dont u?
11:57:01Edendoes that give the gay view?
11:57:10markunEden: I have my files in a nice folder structure
11:57:29B4gderplease use proper words
11:57:31Nico_Pis the "gay view" the folder structure?
11:58:09Edenna i mean wen u playing stuff, the gay view is the name, album, rate, and stuff like taht
11:58:29Mouser_XMy files are in a useable folder structure as well. And the tags are horrible (a very large percentige of my files don't support tags of any kind. Thus, using them in the database gets very ugly, very fast).
11:58:50markunEden: no, that's called the "while playing screen" or WPS
11:58:51*B4gder doesn't understand Eden
11:59:27Mouser_Xgay = happy?
11:59:31Edenum, just went into an app, how i get out of it?
11:59:38B4gderand gays can use both database and file browser I would say
12:00:12Edenhow i get out of app?
12:00:21markunEden: I don't own an ipod, sorry. Just try a few keys :)
12:00:25Mouser_XWell, with gay = happy, that changes it to the "happy view."
12:00:34markunor else you can find it in the menu.
12:00:43peturor the manual?
12:00:57Mouser_XThat manual is good.
12:01:08Mouser_XI sometimes read it (again) when I get bored.
12:01:20Mouser_X(I made sure to read it once before I started using Rockbox.)
12:08:47 Join faemir [0] (
12:14:13Nico_PBagder: will things (like the website) break if I touch docs/COMMITTERS?
12:15:11B4gderI think the web site's front page etc needs to be fixed to do utf8, otherwise you'll break it yes
12:15:42Nico_Pso I'd better leave it as it is for now?
12:15:52B4gderyes I think so
12:15:58Nico_Pall right
12:16:01B4gderI'm not in a mood for doing that just now
12:22:44linuxstbNico_P: Going back to the microseconds symbol in debug_menu.c, I thought Rockbox used utf-8 internally for strings - meaning it was wrong prior to 15222 (and completely wrong afterwards)...
12:23:13Nico_Pso you're saying having it in UTF8 is the correct way?
12:23:31linuxstbI think so... Maybe markun can say for sure.
12:23:36markunI think so too
12:24:21Nico_Phere's the current state:
12:24:32Nico_PI think I've done all of them except docs/COMMITTERS
12:25:44*linuxstb suspects Fergus Noone would disagree
12:25:45Nico_Papps/codecs/libfaad/syntax.c looks strange though
12:25:55*linuxstb is in the wrong channel again....
12:26:30 Join Nibbler [0] (
12:28:13*Nico_P commits
12:29:21 Quit cg (Read error: 110 (Connection timed out))
12:29:35*pixelma wonders why that has to happen so quick all of a sudden
12:30:00linuxstbNico_P: Did you commit the change to creative.c ?
12:30:09Nico_Ppixelma: IMHO it's been long overdue and I felt like doing it
12:30:18Nico_Plinuxstb: I haven't hit the button yet
12:30:21 Join nplus [0] (n=npl@141.25.Globcom.Net)
12:30:33linuxstbNico_P: OK, I don't think that creative.c change should be committed.
12:30:44 Nick fxb__ is now known as fxb (
12:30:44Nico_Pok, I'll revert it
12:31:23linuxstbAs it's inside a string, I assume the current value is correct.
12:32:57linuxstbAnd I think in rbutil/ipodpatcher/fat32format.c, the change should be to replace the odd char with an "i".
12:34:07linuxstbAnd something odd is also in rbutil/rbutilqt/zip/LICENSE.GPL
12:34:15 Quit Eden (Read error: 110 (Connection timed out))
12:34:46Nico_Phmm it seems you're right
12:35:08Nico_Pmaybe those weren't latin1, kinda like the profont doc
12:36:03 Join n1s [0] (n=nils@rockbox/developer/n1s)
12:36:08linuxstbPlus probably just use normal quotes for apps/codecs/libfaad/syntax.c
12:36:19linuxstb(or nothing at all...)
12:37:30linuxstbpixelma: Are you against moving to utf-8?
12:38:21JdGordonNico_P: arg thanks... yeah that bloody ns gets me every time.. and obviously i missed it checking the svn diff
12:38:49Nico_PJdGordon: shouldn't be a problem anymore ;)
12:38:53pixelmadon't have a strong opinion on it, but if I remember correctly there were some "issues" or at least points to consider in earlier discussions here
12:39:15JdGordonif kate wasnt so stupid it wouldnt be a problem :p
12:39:51pixelmaI'm not convinced that JdGordon's editor would do a better job now. At least he has no excuse then :P
12:40:12*amiconn would have preferred to keep the codebase latin1
12:40:18Nico_Ppixelma: kate loads files as UTF8 by default (at least if UTF8 is the system default encoding)
12:40:23JdGordonit wont.. but if I set it to force utf8 we should be ok
12:40:48*JdGordon hasnt found a better editor
12:41:05*Nico_P has found (g)vim but doesn't use it 100% of the time yet
12:41:14Nico_Pamiconn: why?
12:41:23pixelmae.g. won't the source tree to become even bigger now (a noticable amount?)
12:42:02Nico_Ppixelma: only special chars take two bytes. ASCII chars stay the same
12:42:27*n1s was woken up on time by the beast this morning :)
12:42:40amiconnIt's compatible with more editors, and utf-8 doesn't really have an advantage for source code
12:42:58 Join Eden [0] (
12:43:16*Nico_P wouldn't think much of an editor that doesn't support UTF8 by now
12:43:18Edenok, how do i play videos on rock box (5th ipod 30gb)
12:43:19preglowwoo, finally utf-8
12:43:35preglowthere's no good reason to keep the code base latin1
12:43:50*linuxstb doesn't care either way
12:43:50Nico_Pamiconn: it does have one for accented chars, which obviously we have
12:44:01preglowi just want to bloody get away from codepages and never see them again
12:44:06preglowapart from utf-8
12:44:23pixelmapreglow: and don't forget time zones ;)
12:44:27preglowpixelma: oh yes
12:44:40Mouser_XEden: Read the manual, or read PluginMpegplayer in the wiki (or is it PluginMpegPlayer ?)
12:44:47Mouser_X(I'd suggest both.)
12:45:53linuxstbEden: On the ipod video, the Apple firmware does a much better job of video playback than Rockbox, so you may be better off just using that
12:46:04pixelmaMouser_X, Eden: the wiki page is even linked from the manual... ;)
12:46:36Edenlinuxsb: u saying i shud put original ipod firmware on?
12:46:55Edenlinuxstb: u saying i shud put original ipod firmware on?
12:47:09Nico_PEden: could you please speak English?
12:47:10linuxstbIt's still there.
12:47:28linuxstbSearch for "dual-boot" in the manual.
12:49:42Edenwea in the manual wud i look?
12:49:59gevaertsEden: please use normal english
12:53:21markunEden: on the left under "manual"
12:53:56 Join simonrvn_ [0] (i=simon@unaffiliated/simonrvn)
12:54:42 Join barrywardell [0] (n=barrywar@
12:56:51 Join davina [0] (
12:58:25 Join DerDome [0] (
13:00:44 Join simonrvn1 [0] (
13:01:52 Quit JdGordon ("Konversation terminated!")
13:03:38 Join JdGordon [0] (
13:04:42Nico_PJdGordon: have you started thinking/working on an iconview?
13:04:57 Quit simonrvn (Read error: 113 (No route to host))
13:04:57 Nick simonrvn1 is now known as simonrvn (
13:11:55 Quit linuxstb (Read error: 110 (Connection timed out))
13:14:09 Quit simonrvn_ (Read error: 113 (No route to host))
13:19:58 Quit bughunter2 ("Leaving.")
13:20:18 Join bughunter2 [0] (
13:22:14Nico_PjhMikeS: wow, buffering is superfast now
13:22:44jhMikeSyeah, it perks it up a bit ;)
13:23:10JdGordonNico_P: nope
13:23:27Nico_PJdGordon: I was thinking about giving it a go
13:23:33JdGordongo for it :)
13:23:59jhMikeSis it going to be hard-coded for icon size? I want that for a scene browser for mpegplayer.
13:24:02n1sjhMikeS: I have working alarm wake up, but am not sure about some of the interrupt stuff, want to take a look before I commit?
13:24:47JdGordonjhMikeS: when your done with the beast... wanna get the mr500 talking?
13:24:57jhMikeSn1s: sure. I thought about an event registration for that stuff since it's going to get hairy after dealing with changing and other things.
13:25:05 Join ex1stenz [0] (n=existenz@
13:25:37Nico_PjhMikeS: depends on what you mean by hard-coded... I was thinking of making it as generic as possible
13:25:40jhMikeSJdGordon: If I had one. :) It might even help develop an overlay API.
13:26:06jhMikeSNico_P: so I can place frame thumbnails in a list grid but of course they shouldn't be icon-sized.
13:26:07B4gderex1stenz: spam somewhere else please
13:26:44JdGordonjhMikeS: but.. you dont really need one to do the magic.. :D
13:26:51Nico_PjhMikeS: that's the aim... I'll want to experiment with AA too
13:27:40 Quit ex1stenz ()
13:27:57jhMikeSJdGordon: the latest HW is so much more complex than previous stuff I'm pretty sure I would need one. :)
13:28:08 Quit DaCapn (Read error: 113 (No route to host))
13:28:28n1sjhMikeS: here
13:30:16 Join DaCapn [0] (
13:31:13 Join jpt9 [0] (
13:31:34 Join moos [0] (n=c40cd9ab@gateway/web/cgi-irc/
13:33:48jhMikeSn1s: why is a pending[0] check inside the pending[1] block? (if I read that correctly)
13:36:13***Saving seen data "./dancer.seen"
13:36:19n1sshould of course be pending[1] since the TODA bit is in register 1
13:37:48n1sbut i was a bit unsure if it's needed at all to handle this interrupt because all it does is turn off the alarm if it triggers while we are already awake
13:38:08 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
13:38:59n1sand on the pmic the alarm will not reoccur untill the DAY register wraps around and gets back to the same value which should take many years
13:42:26n1sso we could just poll the interrupt bit on start up like i do in rtc_init and never unmask it
13:42:53n1sum, no, that wouldn't work
13:43:08 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
13:45:48jhMikeSif the alarm wake is just a one-shot at wakeup then why wouldn't it?
13:46:11 Part jpt9
13:46:37JdGordonjhMikeS: surey the datasheet is al you need? :D :p
13:46:54jhMikeSJdGordon: haha
13:47:01n1sjhMikeS: it it's masked we wouldn't wake up
13:47:37jhMikeSDoes the masking prevent the PMU from executing the wakeup?
13:47:46*moos is frustrated, he canot have the beast runing :(
13:48:09n1sjhMikeS: that's what I gather from reading the data sheet but I could of course try it :)
13:48:10jhMikeSUsually it's set to an unreachable day value (highest value in day counter)
13:49:05n1sAlso i noticed that it wakes up on headphone plug in, dunno if I caused that
13:49:06Nico_PjhMikeS: will the disk get faster with DMA?
13:50:10linuxstbmoos: What's the problem
13:51:12jhMikeSNico_P: I think it would get a good bit faster and use less CPU during transfers. The next thing has to be to get the SDMA working (for both audio and ATA).
13:51:28mooslinuxstb: I tried several bootloader versions, but for all, the boot failed after the disk init, failed to find the .gigabeat file :(
13:51:46Nico_PjhMikeS: awesome :)
13:51:54*jhMikeS wonders what moos is doing that it can't find it.
13:52:57linuxstbmoos: Is that an S30 or S60?
13:53:01n1smoos: did you extract a regular .rockbox dir to the root of the large partition?
13:53:25moosthe fact that I canot build here doesn't help me, I canot find which version caused the hell here :(
13:53:35*Nico_P leaves
13:53:44mooslinuxstb: S60 n1s:yup
13:54:35moosin the little partition I have the .bin
13:55:08*linuxstb is tempted to play around with manipulating his S's partition table, but is unsure if it can be recovered without an external adapter...
13:55:44*n1s fixes the fat again...
13:57:09 Join Thundercloud [0] (
13:57:19 Quit moos ("CGI:IRC (EOF)")
13:57:26 Join DefineByte [0] (
13:57:37 Join moos [0] (n=c40cd9ab@gateway/web/cgi-irc/
13:57:37*n1s waits untill 14.00 to se if it wakes up
13:58:09n1sis anyone else getting lots of file system corruption on their S ?
13:59:27jhMikeSnot me
13:59:39 Quit Eden (Read error: 104 (Connection reset by peer))
13:59:52 Part LinusN
13:59:59jhMikeSwell, actually I did have an odd occurrence
14:00:08 Join Eden [0] (
14:00:48n1sit has mounted as write protected and needed to be fixed with fsck 3 times for me since yesterday morning...
14:01:17 Part Eden
14:01:50moosisn't the USB stack not yet bug free?
14:02:10moosI mean could cause data corruption
14:02:20n1sjhMikeS: seems like it doesn't wake up if the bit is left masked
14:03:39*linuxstb can report the Gigabeat S doesn't like the partition table being changed...
14:03:51moosoops :(
14:03:58jhMikeSJust because the interupt is enabled doesn't mean it must be handled explicitely. It will be cleared off by the PMIC even handler.
14:04:27 Part DefineByte ("Bye all")
14:05:02linuxstbRecovery is easy though - just enter recovery mode (hold the left of the cross and the windows key after toggling the battery switch), and then run sendfirm nk.bin
14:05:20gevaertsmoos: I would be very surprised if the USB stack causes data corruption
14:06:59moosgavaerts: it isn't built by default yet for thos target that it need it, right?
14:07:43jhMikeSthe ATA timings looked ok to me but I'm not sure what to make of Ton/off though
14:07:59gevaertsmoos: I'm not sure about the beast, but it's still not enabled on PP
14:08:21jhMikeSit's enable on beast becuase there's no nice alternative anyway
14:08:25n1sjhMikeS: so nothing bad happens if the interrupt fires and isn't handled (explicitly)?
14:08:59jhMikeSn1s: no, the event handler will clear it off whenever it gets around to it
14:10:00jhMikeSactually it should just set it in there where other things are polled at startup becuase the handler runs way before rtc code gets called
14:10:54n1sok, then we only need to set it and unmask the interrupt and if it fires while we are on, just don't care because it will take about a hundred years before it reoccurs
14:10:55 Join barrywardell_ [0] (n=barrywar@
14:11:02jhMikeSthat can work for now. it does need refinement though so it doesn't clear pendings that aren't enabled.
14:11:37*moos gives up and decide to store his beast for now, he will wait a bit before try again
14:12:00 Quit Thundercloud (Remote closed the connection)
14:12:26jhMikeSn1s: right, it will just have no event handler and that's ok
14:13:37jhMikeSactually it won't fire at startup even if left unmasked because it masks things before the GPIO pin is enabled anyway
14:14:53 Join tvelocity [0] (
14:19:37n1ssigh, corrupted again...
14:20:12n1slots of "Contains a free cluster (19181). Assuming EOF."
14:22:20jhMikeSwhat steps lead to that?
14:23:15n1si just unzipped a new, unmounted restarted the S a couple of times, mounted again -> write protected because of errors
14:23:46petursound like the FAT wasn't updated...
14:23:46 Quit XavierGr (Read error: 110 (Connection timed out))
14:23:47n1sdisconnected it after unmount too
14:24:16*jhMikeS wonders about write-behind of some sort going on
14:24:34peturunmount should have fixed that
14:24:36*jhMikeS sees the "set_features" function and reconsiders his timing implementation
14:24:48n1sjhMikeS: ok, new approach works fine :)
14:24:50 Quit barrywardell (Read error: 113 (No route to host))
14:26:15jhMikeSI wonder about the reset being improper and perhaps it should wait for the interface to be inactive before doing it
14:27:22 Join DrMoos [0] (n=c40cd9ab@gateway/web/cgi-irc/
14:28:22 Quit moos ("CGI:IRC (Ping timeout)")
14:28:38 Nick DrMoos is now known as moos (n=c40cd9ab@gateway/web/cgi-irc/
14:29:10n1sjhMikeS: what do you think of masking the headphone plugin interrupt when we shut down to prevent a plugin from waking up the player?
14:30:00jhMikeSthen I suppose an mc13783_shutdown function should be implemented and called at the proper time
14:30:20jhMikeSthough I don't mind the USB plug wakeup at all
14:31:01n1smaybe the headphone thing should even depend on the setting for resuming playback on plugin? and I agree about usb
14:31:41jhMikeSthat one I'm not so crazy about doing :)
14:32:04n1sok, should I commit the alarm stuff or do you want another look?
14:33:19jhMikeSno big deal if it works. I'll probably get to looking later anyway given I want to refine the event thing.
14:34:00n1sok, thanks for the help :)
14:34:23*linuxstb can confirm the Gigabeast bootloader is happy with only the Rockbox bootloader nk.bin on the disk, and nothing else
14:35:12linuxstbIt's _very_ sensitive to the partition table being edited though...
14:35:40linuxstbe.g. I tried to make the first partition smaller, and it went directly to recovery mode
14:38:11 Join swimmer [0] (
14:39:05jhMikeSI can throw all kinds of junk on the firmware partition without problems and delete all the cruft from the OF without problems.
14:39:14swimmerhello - I'm trying to install Rockbox on an ipod video 80gb via rbutil and it fails with this error msg:
14:39:18swimmer[INFO] Scanning disk devices...
14:39:20swimmer[ERR] FATAL: Permission denied on 2 device(s) and no ipod detected.
14:39:23swimmer[ERR] You need permissions for raw disc access for this program to work!
14:39:25swimmer/dev/sdz: No such file or directory
14:39:32swimmerI configured mountpoint + device manually
14:39:43swimmerit's mounted on /media/ipod ...
14:40:32linuxstbAre you sure it's an ipod video, and not a "classic" ?
14:42:00swimmerlinuxstb: yes - I even phoned the delivering party to make sure it's a 5th generation ipod :)
14:42:13swimmerlinuxstb: is there a way to check this?
14:42:18n1sswimmer: as the error says you need permission for raw disk acces
14:42:59n1sswimmer: google knows for sure how to tell the difference
14:43:11swimmernls: does it mean it wants to access /dev/sdc2 and not /dev/ipod?
14:43:41n1sor rather /dev/sdc
14:43:56n1sthe _disk_ not partition
14:44:08swimmern1s: ok - then I have to find out udev-rules how to change the owner of /dev/sdc ...
14:44:22preglowdidn't we agree we didn't want bootloader logos?
14:44:23n1sor just run it as root/ use sudo
14:44:31 Join styleism [0] (
14:44:39swimmern1s: I wanted to avoid running it as root ...
14:45:03swimmern1s: how that I didn't have any problems with my iriver?
14:45:17barrywardell_preglow: last time it was discussed here, people seemed generally in favor of it. it was logos loading from disk that people didn't want
14:45:22n1sswimmer: the installs wor differently
14:45:27swimmern1s: does rbutil not access the raw discs in this case?
14:45:31swimmern1s: I see
14:45:49preglowbarrywardell_: i just thought we wanted the bootloader to shut up as much as possible unless it has a reason not to
14:46:06preglowi certainly don't want a logo showing up for 0.2 seconds during the time my nano takes to boot
14:46:27swimmern1s: thx for the info and your patience :)
14:46:47preglowperhaps it makes more sense for hd based targets that need time to spin up a disk
14:46:57preglowbut for flash based targets it'll just look clumsy
14:47:17barrywardell_preglow: it also makes sense on sansa where we can replace the OF bootloader completely
14:47:42preglowhow long does the bootloader run for?
14:48:15n1sbarrywardell_: does the logo delay the booting or is stuff done in the background?
14:48:15barrywardell_it depends
14:48:41barrywardell_about 1 second when booting rockbox, several seconds when loading the of
14:49:10barrywardell_and, then when rockbox loads, the only thing that should change is the text at the bottom
14:49:14 Quit moos ("CGI:IRC")
14:49:27preglowbarrywardell_: well, i agree a logo would be nice there
14:49:28barrywardell_n1s: my timing made the delay out to be about 6ms
14:49:31preglowbut it makes no sense on ipods, for example
14:49:33preglowthere already is a logo
14:49:45preglowwe don't want booting to look messy
14:49:51barrywardell_I agree
14:49:51preglowrockbox is already more than messy enough
14:49:53linuxstbHow is the transistion with Rockbox itself? Should we not re-init the LCD on some devices?
14:50:33barrywardell_it works fairly seemless on Sansa and H10 anyway, except that the background colour changes (black->blue)
14:50:36n1sbarrywardell_: wow, how did you time that? i meant more like if it sleeps to display the logo for a time but i see that it doesn't .)
14:51:06barrywardell_n1s: checked the usec_timer before and after showing the logo
14:51:33 Join LambdaCalculus37 [0] (n=44a0430f@gateway/web/cgi-irc/
14:54:14preglowdo we actually replace the sansa bootloader completely, though? i would have thought tons of device inits would be missing, but perhaps that's only an ipod situation
14:54:30linuxstbpreglow: What's the problem (on ipods) if the logo looks the same as the logo at the start of Rockbox itself?
14:54:31barrywardell_linuxstb: the background thing is more a controversial topic. there doesn't seem to be any consensus on black vs rockbox blue.
14:54:39barrywardell_my personal preference is blck
14:55:06preglowmine too
14:55:08linuxstbI would go for black because that's the colour for the default theme.
14:55:09barrywardell_preglow: it's not the default install mode with sansapatcher, but it works perfectly for my sansa. the only thing we're missing is usb recovery mode
14:55:56preglowi think going for black makes sense for other themes as well, going from black to some other colour will pretty much always look clean
14:56:01preglowgoing from light blue to red won't
14:56:12LambdaCalculus37That'd just look nasty.
14:56:36linuxstbI know amiconn stated a preference for blue - did anyone else?
14:56:53preglowi've never liked the blue fixation rockbox seems to have
14:56:54*B4gder votes black
14:56:58pregloworange goes perfectly with black
14:57:01preglowless so with blue
14:57:08 Join K4rP4D [0] (n=KrPD@unaffiliated/krpd)
14:57:08*LambdaCalculus37 votes black as well
14:57:19LambdaCalculus37Would go well on the iPods.
14:57:41barrywardell_JdGordon: was arguing for blue too
14:57:47*gevaerts prefers freshly pressed orange
14:58:00preglowlinuxstb: if the logo overlaps perfectly with what rockbox shown on startup, it's all good
14:58:45linuxstbSo, we have 5 blacks, 2 blues, and one orange...
14:59:25*LambdaCalculus37 makes a chart on the whiteboard
14:59:30*n1s raises his commit count :/
14:59:54barrywardell_it's an easy change to make:
14:59:55 Quit nplus (Remote closed the connection)
15:00:04*n1s votes for octarine and runs.
15:00:15barrywardell_although I don't know what LCD_DEFAULT_LS is for...
15:00:36linuxstbSomething to do with the gradient selection bar...
15:00:39LambdaCalculus37n1s: Yes, we should make the bootloader screen the color of magic. :)
15:00:43*gevaerts thinks that octarine can't be expressed in RGB
15:02:25*barrywardell_ wonders if he should go with the democratic vote and commit the change?
15:03:46*jhMikeS is raising his commit count artificially by doing something, committing it, then rewriting it and commiting again. :p
15:06:31linuxstbjhMikeS: Did you read my comments in IRC yesterday about the 300KB of zeros in the middle of the beast's bootloader?
15:06:55linuxstbI think it's caused by the vectors being loaded after the bss.
15:07:31B4gderbarrywardell_: want us to vote about you going with the vote results? ;-)
15:08:12*barrywardell_ goes with the majority
15:08:42jhMikeSlinuxstb: 300KB of zeros?
15:08:56linuxstbjhMikeS: Yes - look at bootloader.bin in a hex editor...
15:09:07B4gderso austriancoder's work broke the mrobe100 sound
15:09:19linuxstbMy suspicion was raised because it's so big - about 370KB, compared to around 70KB on other targets.
15:10:20jhMikeSlinuxstb: or it's because the NOLOAD attribute isn't used
15:10:28B4gderI'm getting a bit annoyed that austriancoder seem to always break 11 things on each of his "cleanups"
15:10:52linuxstbjhMikeS: That's why I didn't fix it - I'm not comfortable with lds files.... (I don't know what NOLOAD does...)
15:11:16jhMikeSit marks section as not part of the loadable image itself
15:11:34linuxstbWhat should have NOLOAD?
15:11:40 Join DefineByte [0] (
15:11:50 Part DefineByte ("Bye all")
15:12:11LloreanB4gder: Just a small bit of mitigation, that person is testing dailies so it's not *absolutely* certain that changed cause it. Though its placing makes it likely.
15:12:39B4gderyes, and his mentioning of bass/treble also points to that
15:12:58LloreanSince it was broken elsewhere too.
15:12:59jhMikeSlinuxstb: everything that doesn't contain data that doesn't need to be loaded such as ibss, bss or other zero-init sections
15:14:00linuxstbHmm, the other files don't have that for the bss/ibss
15:14:01jhMikeSbut actually it won't help here since the address isn't compensated
15:15:05linuxstbI'm guessing just moving the vectors before .stack would work.
15:15:25JdGordonsounds like I'm too late (and changed my vote anyway), but having a different colour in the bootloadser and the main binary can be useful incase either freezes, you'll be able to tell which froze by the colour
15:15:28*jhMikeS just moves the vectors after data and the size goes to 77KB :)
15:15:43barrywardell_JdGordon: you can still tell by the text at the bottom ;)
15:16:07JdGordonit doesnt say bootloader does it?
15:16:11linuxstbYes, it could be nice for the text to show the status.
15:16:16barrywardell_it says the bootloader version
15:16:23barrywardell_which then changes to the rockbox version
15:16:50linuxstbThis will annoy the logo swappers though...
15:17:01JdGordonmeh to them :D
15:17:03B4gderand that we like! :-P
15:17:12linuxstbI didn't say that was a bad thing ;)
15:17:15*jhMikeS does some other bl changes that need doing too
15:17:49barrywardell_linuxstb: what do you mean by status? like whether we're loading rockbox or the of?
15:18:45linuxstbYes, I was thinking of a "loading xxxx" message, followed by "starting xxx" - so if anything fails, that can show the status. But maybe that's doing too much....
15:18:58linuxstbjhMikeS: I'm guessing the untar code can go now?
15:19:45 Join Synergy66 [0] (
15:19:59barrywardell_linuxstb: hmmm. I'll have a think about that
15:21:20linuxstbjhMikeS: Also, something I noticed today was that if the bootloader powers off the S, the hard disk makes a nasty click - I'm guessing the disk doesn't get shutdown cleanly.
15:22:28linuxstbAnyone know anything about where the S's bootloader is stored? Is it a normal NOR flash chip, or some kind of secure ROM in the imx31?
15:22:42jhMikeSlinuxstb: 1) I'm not sure to jump on removing the untar code because a testing mess up could make it useful 2) It will click if power is cut suddenly
15:22:52 Quit tvelocity (Remote closed the connection)
15:23:13 Join nplus [0] (n=npl@141.25.Globcom.Net)
15:24:10JdGordonHEY! when was the main binary default colour up for debate? wer'nt we talking about the bootloader colours?
15:24:16JdGordon:'( /me liked the blue
15:24:28linuxstbDuring the last hour...
15:24:29 Quit Synergy6 (Read error: 104 (Connection reset by peer))
15:24:45linuxstbAnd your vote was counted as blue...
15:24:47JdGordondidnt scroll up to the start of the convo :P
15:24:57JdGordonoh well.. shit happens :p
15:25:09LloreanI like the blue, but with the new theme I'd have voted for black too if i were around
15:26:06linuxstbI'm persuaded most by preglow's argument - using black for the splash screen should look reasonable no matter what the user's theme is.
15:26:27JdGordonnot it its a dark on light one...
15:26:33JdGordonbut yeah, that sounds reasonable
15:26:43 Join tvelocity [0] (
15:26:54barrywardell_the only thing that it changes is the background of the startup splash
15:27:20linuxstbDo any plugins revert to the default colours?
15:27:41barrywardell_and and maze and snow, which i'm fixing now ;)
15:28:07barrywardell_actually, maze and rocklife
15:29:31 Quit JdGordon (Read error: 104 (Connection reset by peer))
15:30:56 Quit ReKleSS ("Leaving")
15:31:47 Join JdGordon [0] (
15:36:17***Saving seen data "./dancer.seen"
15:39:47*amiconn *hates* light-on-dark display :(
15:40:24 Quit DaCapn ()
15:42:29 Quit K4rP4D ("Leaving")
15:44:16JdGordonany objections in moving the id3 browser code from screens.c to metadata.c ?
15:44:56linuxstbWhy would you want to do that?
15:45:01barrywardell_amiconn: fortunately, it's easy to change themes ;)
15:45:14JdGordonor just out of screens.c
15:45:53 Join DaCapn [0] (
15:47:11amiconnbarrywardell_: But not built-in defaults....
15:47:56barrywardell_but if you change the theme, you should never see light-on-dark
15:48:09*preglow enjoys light on dark displays
15:48:32linuxstbBut isn't it just the startup splash screen we're talking about?
15:50:34jhMikeSamiconn: in case you were still wondering about the switch_thread calls with interrupts disabled - that's what is part of the design with list-changing calls like block_thread, block_thread_w_tmo and sleep_thread and occasionally wakeup_thread (high-priority thread woken than current).
15:51:29jhMikeSthough the last isn't currently done
15:51:32linuxstbHmm, the Rockbox bootloader doesn't start Rockbox any more for me.... Same problem as moos was reporting.
15:51:49barrywardell_linuxstb: to change the startup splash, it's easiest to change the default bg and fg. the only place I found them used was in the startup splash and in a couple of plugins which are fixed to use the old rockbox colours
15:51:50amiconnUnfortunately I seem to be unable to understand the current scheduler
15:51:51jhMikeSLlorean: had it magically start working again
15:52:01jhMikeSthat was to linuxstb not Llorean :)
15:52:13linuxstbbarrywardell_: I know - I'm agreeing with what you've done...
15:53:13jhMikeSamiconn: not much is ususual about it as far as schedulers go though some aspects of inheritance implementation is unique to make it complete without object ownership lists being maintained.
15:54:48jhMikeSSomething points to something other than an interrupt enable issue being the root cause since I'm rather certain that should plague every target at all time
15:55:24barrywardell_actually, reset colours also uses the default fg/bg, but it probably shouldn't. it should probably revert to the theme's defaults
15:56:31 Quit parafin (
15:57:04*petur arrives late, scrolls up and votes black
15:57:20peturmaybe something to discuss on devcon ;)
15:57:56linuxstbIsn't there too much democracy going on today?
15:58:13barrywardell_or for a steering board...
15:59:57*linuxstb can't make his S work any more...
16:00:28jhMikeSfyi of the day: some schedulers don't have complete priority inheritance protocols because it's considered "hard" to do and cheat by leaving threads running at their most heightened priority until all resources are released rather than dropping to the priority of the highest still-owned resource greater than the current base (too esoteric? :p). usually (and any example I came across) the full release priotocol is done by maintaining a list of
16:01:23jhMikeSlinuxstb: does going back to an earlier revision do it or is it just now refusing anything?
16:01:45linuxstbI haven't tried an earlier revision... Does anyone have an older than current SVN nk.bin handy?
16:02:05jhMikeSseveral are already online, one moment
16:02:05 Quit barrywardell_ ()
16:03:12*linuxstb thinks white-on-black in the gigabeast bootloader looks nicer though"> (v8, v9, v11 are up, change the file name)
16:03:23linuxstbAnd current is v12?
16:03:31jhMikeSthose are plain btw
16:03:36jhMikeScan put up v12 too
16:03:46*linuxstb tries with v11 first
16:04:12jhMikeSv12 is up
16:04:28linuxstbNo luck with v11...
16:04:55jhMikeScan't make it work by batt switch reset?
16:06:05linuxstbNo, that doesn't seem to help.
16:06:58 Quit Ragnor (Nick collision from services.)
16:07:08 Join Ragnor [0] (
16:07:54*linuxstb decides to give his S a charge and try again later...
16:08:23*jhMikeS wonders if disabling the .tar search would change things
16:09:00*LambdaCalculus37 hopes that linuxstb isn't experiencing the same thing he did with his own Gigabeast
16:09:35*gevaerts decides that nicks startint with l are not goot for beasts
16:10:23LambdaCalculus37gevaerts: :P
16:10:36jhMikeSLambdaCalculus37: I don't think it's that.
16:11:29linuxstbLambdaCalculus37: What was your problem?
16:12:04LambdaCalculus37linuxstb: I left mine on the charger after going an update and copying music to it, and I couldn't get it to turn back on after that.
16:12:15LambdaCalculus37toffe82 should be receiving it today.
16:12:38linuxstbv8 doesn't display "file not found", but pauses for a while, and then powers off...
16:13:23krazykitLambdaCalculus37, mine did the same thing. charged it, left it, it never turned on again
16:13:29 Quit Synergy66 ("Adios")
16:13:57jhMikeSthe error string was displayed properly in older revisions
16:14:15*linuxstb tries reformatting the main partition
16:14:47LambdaCalculus37I wonder if it's a problem with the battery or the charging hardware.
16:14:48NJoinparafin [0] (
16:17:12jhMikeSthere is a fuse in series with the battery
16:17:37jhMikeSalso one in series with the charger line
16:17:57 Join crope` [0] (
16:21:26pixelmahmm... barrywardell left...
16:25:44 Quit crope`` (Read error: 113 (No route to host))
16:29:14 Join dabujo [0] (
16:33:14pixelmacan one assume that he won't be fixing the broken greyscale target builds anytime soon then?
16:33:53B4gdermaybe if we taunt him properly until he does it?
16:35:31B4gderproperly means involving hamsters and elderberry
16:36:26*gevaerts hands B4gder a coconut
16:36:41B4gderan african one?
16:38:06*linuxstb senses B4gder isn't taking this violation seriously ;)
16:38:06*gevaerts assumes so. The origin of the coconut was never fully clear
16:39:07*B4gder just got side-tracked thinking of tauntings
16:41:57 Quit crope` (Read error: 113 (No route to host))
16:42:37swimmerhmmm - I installed rockbox now and some music files (40G ogg ;-) ) on it and I can start playing music but after 1- seconds or somerthing rockbox crashes and I can see this error on the display:
16:43:12B4gderyou have "some" 40G files on your dap?
16:43:13swimmer"Data abort at 00530070 <0>"
16:43:34 Nick JdGordon is now known as JdGordon|zzz (i=jonno@rockbox/developer/JdGordon)
16:43:37swimmerB4gder: sorry - I got an Ipod Video 80G today ...
16:43:50swimmerinstallation of rockbox went fine so far
16:44:01swimmerI could copy the music files as well
16:44:01B4gderoh you mean you have a total of 40GB...
16:44:12swimmerB4gder: yes
16:44:22*gevaerts wondered what kind of unit the gigaogg was
16:44:28swimmerB4gder: actually not very relevant - sorry for the confusion
16:44:51B4gderswimmer: does it crash every time on the same ogg file?
16:44:53swimmerwhat *can* be relevant that it is in ogg-format
16:45:09swimmerB4gder: did not test that thoroughly enough
16:45:28swimmerB4gder: once I started from a playlist and once by selecting a file directly
16:45:53linuxstbDo any music files work?
16:46:08linuxstbAnd how did you install Rockbox? With the official instructions in the manual?
16:47:23swimmerlinuxstb: I installed it with rbutil following the official instructions for automatical installation ...
16:47:39swimmerlinuxstb: I first have to check now if it crashes with mp3 files as well
16:47:56swimmerB4gder: I did a reset and tried the same file and it crashed again ...
16:47:58linuxstbOK, because using an old bootloader could cause those symptoms. But that doesn't seem to be your problem.
16:48:34 Join toffe82 [0] (
16:48:56pixelmacould imagine one of two things - wrong tag format (though I don't know whether Rockbox crashes because of that - or one of the rare 80GB Ipods with 32MB RAM
16:50:10swimmerhow can I check the RAM?
16:50:28linuxstbYou could try installing a build for the 30GB ipod video
16:50:57swimmerfirst let me try a mp3 file
16:51:15 Join Thundercloud [0] (
16:51:59 Quit z35 (Read error: 110 (Connection timed out))
16:52:22swimmercrashes as well with "Prefetch abort at <DIGIT> .."
16:52:26 Join z35 [0] (
16:53:02swimmerlinuxstb: what will installing a build for 30GB do different?
16:53:08linuxstbYou could check your disk for errors - e.g, with chkdsk or fsck
16:53:21swimmerit's a new ipod ...
16:53:25linuxstbThe 30GB ipods only have 32MB of RAM - that's the difference between the two builds
16:53:32swimmerlinuxstb: I see
16:54:15linuxstbDo you always safely eject your ipod when disconnecting?
16:54:16swimmerno way to check how much RAM it has?
16:54:29swimmerlinuxstb: yes I did (pumount though)
16:55:09swimmerthe filesystem on the desktop is utf-8 so are the tags ... could this be a reason?
16:55:19swimmerthough it works on my iriver without any problem
16:55:36linuxstbNo, that shouldn't be a problem
16:56:02swimmerso let's try the 30GB build ...
16:58:12swimmerdoes it have a different bootloader as well?
16:58:23 Join K4rP4D [0] (n=KrPD@unaffiliated/krpd)
16:58:24linuxstbNo, the bootloader is the same
16:58:33 Join Robust2 [0] (
16:58:48linuxstbThe bootloader only uses the first 32MB of RAM, so works on both.
16:59:48linuxstbDo any of the plugins work?
17:01:12Robust2hey all
17:01:18markunhi Robust2
17:01:21Robust2i've got quite a problem with Rockbox on my M3
17:01:30LambdaCalculus37Do share with us.
17:01:34Robust2been trying to get some assistance on the forums, but alas
17:01:46Robust2the remote display dies on me after 5 seconds of playback
17:01:53swimmerlinuxstb: it seems to work - played already 1 minute and did not crash :)
17:02:01Robust2on an M3, this means there is no display at all left
17:02:31Robust2it's been discussed a bit on flyspray, but seems like there are no other M3 users out there
17:03:06markunRobust2: amiconn did the M3 port
17:03:16markunI hope he knows what it could be
17:03:24Robust2is he afk at the moment ?
17:03:31LambdaCalculus37Robust2: You can ask him.
17:04:23linuxstbswimmer: I don't know, but may diagnostics mode will show the amount of RAM your ipod has - hold SELECT+LEFT immediately after your ipod turns on to enter it
17:04:35Robust2ok I will, i've tried mailing him a few days ago but nothing yet
17:05:11 Join knittl [0] (n=knittl@
17:05:13markunRobust2: he's here every day from time to time, so if you just keep on IRC you can talk to him eventually
17:05:17swimmerlinuxstb: will do ..
17:05:45Robust2ok, great to hera markun, thanks for your help so far :)
17:06:01swimmerwhen I ever manage to get out of the mandelbrot demo ;-)
17:06:18linuxstbHolding MENU+SELECT exits most plugins
17:06:25linuxstb(for a short time...)
17:06:34swimmeryeah - found out already :)
17:07:00 Join Kage [0] (n=chatzill@
17:07:09swimmerok - I'm in the diag mode now
17:07:19swimmerI guess I need "Auto Test"?
17:07:43linuxstbI've no idea, just browse the options...
17:08:10swimmerit's just 2 options: manual or auto ;-)
17:08:21markunKage: just ask
17:08:42linuxstbswimmer: That's odd - there should be more options than that...
17:09:22swimmerhihi - it runs through 8 tests and shows "Completed" at the end ;-)
17:09:38swimmer"M25 Diagnostics 0.7" ...
17:12:02swimmerthe manual test stops at the firewire test ,,,
17:12:24swimmerto be more precise: the accessorize test
17:13:58Kageumm can the apple os and the rockbox os can i run a duel boot off both im on a 1gen nano
17:14:36Kageive looked everywhere and i cant find anything off it
17:14:45LloreanjhMikeS: Well "file not found" didn't display until you added the one line of output for the attempt to open /.rockbox/rockbox.gigabeat, so I assume it's not related to the tar search
17:14:49*gevaerts points to the manual
17:15:09Lloreanlinuxstb: What fixed it for me was testing v9 (which worked), then upgrading to v11 again (which worked at that point).
17:15:18*Llorean doesn't have a clue as to why.
17:15:38 Quit knittl (Remote closed the connection)
17:16:04Kageso can someone help me with this
17:16:04linuxstbKage: If you search for "dual boot" in the manual, all will become clear.
17:16:20Kageokie thank you
17:16:59Kageumm how do i serch the manual
17:17:13 Quit MU{lappy} (Read error: 113 (No route to host))
17:18:55linuxstbEasiest is to download the PDF and then use your pdf viewer's search features
17:19:18gevaertsOr read it all. It's full of useful information ;)
17:19:44LambdaCalculus37Use your eyes to search.
17:20:00LambdaCalculus37The manual is everyone's best friend. :)
17:20:44Kageits still very vaugue, like i tried setting a custom loader, but it dosent show up, it just straight boots rockbox
17:21:08LambdaCalculus37Custom bootloaders are unsupported.
17:21:27LambdaCalculus37The manual covers only the official bootloaders and builds.
17:22:00LambdaCalculus37You have to speak to whoever released the custom bootloader for support.
17:22:18Kageim so confused
17:22:26Kagei just got the rockbox off the site
17:23:08linuxstbYou can't find the dual boot section in the manual?
17:23:28markunKage: well, welcome to the wonderful world of rockbox :)
17:23:39 Join mf0102 [0] (n=michi@
17:24:03LambdaCalculus37Kage: Right here:
17:24:16Kageya but i wanna know if there is a way i can set it so i dont have to go through the pressing of buttons precicly
17:24:27Kagejust so i have a menu
17:24:35LloreanNot with any official software.
17:24:44LambdaCalculus37Write your own bootloader if you want a menu.
17:24:53B4gdereven a menu will force you to press at least one button I think
17:25:00LloreanYou'll have to use Loader2, which we do not provide support for, as changes to how Rockbox works can conflict with it.
17:25:07swimmerok - I'm very happy for the moment :) Thanks to all you you helped me!!!
17:25:34Kageya ya the menu is fine
17:25:39Kagei just dont know how to get one
17:26:29*LambdaCalculus37 wonders if he was just speaking in some obscure language right now =/
17:26:46Kagei did that and when i try to open rockbox it says its not found
17:27:03LambdaCalculus37Kage: If you use Loader2, you ask the iPodLinux guys for help, not us.
17:27:14markunLambdaCalculus37: are there still any ipodlinux guys left?
17:27:23Kagea few...
17:27:29Kagei dont have it now though
17:27:30Lloreanmarkun: It really doesn't matter if there are or aren't though.
17:27:40LambdaCalculus37markun: I haven't followed the project for a long time to really keep track.
17:28:00markunLlorean: it does matter for Kage, maybe not for you
17:28:06 Quit aliask ("ChatZilla 0.9.81 [Firefox 3.0b5/2008041514]")
17:28:52Lloreanmarkun: My point is that even if there's nobody to support it, it doesn't make it on-topic here. So there's no point in asking if there are any of them left, since it won't change the fact that we won't support it.
17:29:17markunLlorean: ok, next time I'll ask in #rockbox-commnity
17:29:28markunand tell Kage there results in a PM
17:30:09Lloreanmarkun: And please, make it clear to him not to ask about bugs unless he switches back to our loader first, since I'm not sure he got that point.
17:30:15pixelmamarkun: are you able to solve Kage's problem with the ipodlinux's bootloader?
17:30:32markunpixelma: no, I've never used loader2
17:30:35Kageim not using ipodlinux boot loader :(
17:30:39markunI don't even have an ipod
17:31:00LambdaCalculus37I do, but my iPod is strictly Rockbox.
17:31:08LloreanKage: Well, the Rockbox one does not have a menu. You have to use it with the buttons like the manual says.
17:31:30LambdaCalculus37It's a matter of timing.
17:31:44LambdaCalculus37MENU+SELECT, then immediately reach up and flick the hold switch on.
17:32:17 Join MU{lappy} [0] (
17:32:29 Quit nplus (Remote closed the connection)
17:33:30pixelmamarkun: my point, I couldn't too that's why I wouldn't even be able to give support, so chances are someone will find better support for that one over at ipodlinux......
17:34:01LambdaCalculus37pixelma: That's why I said that the iPodLinux guys are the ones to turn to in this sort of case.
17:34:16markunand that's why I asked if there are any ipodlinux guys left
17:34:47scorche|shthere are
17:34:55LambdaCalculus37Kage: Go to #ipodlinux and ask them for Loader2 help.
17:35:12gevaertsunless you're happy without a menu...
17:36:21***Saving seen data "./dancer.seen"
17:44:56B4gderbtw, I mailed the sansalinux guy to ask him to acknowledge Rockbox on his site
17:45:34B4gderthere's nothing on there that says he's using a lot of Rockbox, he's mostly just saying it is based on ipodlinux
17:47:03 Join desowin [0] (n=desowin@atheme/member/desowin)
17:51:26 Quit K4rP4D ("Leaving")
17:52:28 Join K4rP4D [0] (n=KrPD@unaffiliated/krpd)
18:03:56 Quit davina (Remote closed the connection)
18:06:34 Join davina [0] (
18:08:28*Llorean cannot seem to make his gigabeast dual boot.
18:09:56Nico_Pwow, a slew of commits!
18:10:45*linuxstb can't even make it single boot...
18:11:57LloreanAny time I construct a dual boot bootloader, it boots just the OF with hold off, and nothing with hold on (freezes with the progress bar full)
18:11:57LloreanIf I install single-boot bootloader, it works fine.
18:11:57Lloreanlinuxstb: Have you tried rolling back to v9?
18:12:26*Llorean thought that hold on was supposed to be OF boot anyway
18:12:43jhMikeSLlorean: I had that when using the larger of the firmware images (20MB) but never the 12MB one
18:13:02LloreanI'm using the 12mb one.
18:13:42LloreanOr at least *a* 12mb one.
18:13:57linuxstbLlorean: I've tried various...
18:14:20Lloreanlinuxstb: I only saw mention of v10 in the logs, so I wasn't sure how far back you went or anything
18:15:12 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
18:15:41*Llorean only wants dual boot so he can charge anyway.
18:16:18 Quit faemir (Broken pipe)
18:17:08 Join crope` [0] (
18:17:32 Quit gevaerts ("time to go home")
18:17:38Lloreanlinuxstb: The only other thing that I did that you may not have, was actually booting once without a .rockbox folder at all (for a genuine 'file not found') then reinstalling rockbox after that.
18:17:50*Llorean doesn't see why that'd make a difference, but this is a weird behaviour anyway.
18:20:56*jhMikeS suggests building a bootloader with the .tar scan disabled just to see
18:21:31pixelmawould putting the new LCD_RGBPACK stuff under a HAVE_LCD_COLOR and the reintroducing the old code under LCD_DEPTH > 1 in the two plugins be the correct fix for the broken greyscale build?
18:21:33LloreanjhMikeS: For my problem too, or for his?
18:22:00jhMikeSboth - I'm curious if the disk activity does something
18:22:21 Join faemir [0] (
18:22:53LloreanOn a vaguely related note, would it make any sort of sense to look for .rockbox on both partitions, so people could keep it even more out of the way if they wanted?
18:23:37pixelmaor better - rearrange it a bit
18:23:51jhMikeSLlorean: then we need multivolume I think
18:24:13LloreanWell, I mean, I can think of a few arguments against it (you can't browse into it if you ever for some reason need to)
18:24:14 Join waldo [0] (
18:24:18jhMikeSI did once get a filesystem corruption when I didn't make sure retailos was put to sleep before cutting power.
18:24:18Lloreanbut some people like a cleaner root. :)
18:24:36 Join miepchen^schlaf [0] (
18:25:34 Quit Nibbler (Read error: 113 (No route to host))
18:28:04LloreanjhMikeS: For mknkboot do I want nk.bin or bootloader.bin from my build directory?
18:28:36*Llorean wonders if he's just doing something stupid.
18:28:49linuxstbNo idea why it's working, but I went through recovery mode and did a restore of the OF, then installed the bootloader again, followed by a Rockbox build, and it's working...
18:29:03linuxstbLlorean: bootloader.bin
18:29:07Lloreanlinuxstb: That's more or less what happened to me.
18:29:27 Join bertrik [0] (
18:29:40linuxstbThe only thing I can think of is that I had deleted the contents of the main FAT32 partition previously, and then restored .rockbox.
18:29:50linuxstbThis time, the OF files are still there.
18:30:38 Quit swimmer (
18:31:23 Quit crope` (
18:31:23 Quit tvelocity (
18:31:23 Quit advcomp2019 (
18:31:23 Quit ali_as_ (
18:31:23 Quit brent113_ (
18:31:23 Quit wpyh (
18:31:36LloreanjhMikeS: Isn't boot order supposed to be "Rockbox with hold off" nowadays?
18:31:41linuxstbpixelma: I'm not sure why barrywardell wanted to keep the old colours - that doesn't seem the right thing to do...
18:32:31NJoinswimmer [0] (
18:32:36NJoincrope` [0] (
18:32:36NJointvelocity [0] (
18:32:36NJoinadvcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
18:32:36NJoinali_as_ [0] (
18:32:36NJoinbrent113_ [0] (n=brent113@
18:32:36NJoinwpyh [0] (
18:32:38preglowjhMikeS: anything new to add to 8918?
18:32:59pixelmalinuxstb: so just revert that commit?
18:33:22 Quit petur ("work->home")
18:33:31linuxstbpixelma: That won't work either - I think it used black on LCD_DEFAULT_BG before - which is now black on black...
18:34:01LloreanjhMikeS: Well, I have working dual boot. But it seems like it was just because I was doing something stupid. But I did remove the tar-search in my local tree, so we'll see if the strangeness stays gone with future bootloader updates, I suppose.
18:36:42 Quit faemir (Remote closed the connection)
18:36:42pixelmalinuxstb: revert the default colour commit too, then? ;) </joke> What would be the best - handle greyscale and colour displays seperately but chose a different value for LCD_RGBPACK? Or something else? I just don't like seeing the builds broken for even more commits...
18:36:53 Part swimmer
18:37:12linuxstbpixelma: I'm just looking now...
18:38:13 Part B4gder
18:39:34pixelmalinuxstb: my local M5 build is just done but I'd rather leave that to someone who knows more than me... :)
18:40:38linuxstbI think what you've done is the easiest fix - but maybe someone who knows those two plugins should adapt them to work with the default theme's colours.
18:41:24 Join faemir [0] (
18:41:25linuxstbi.e. handle the greyscale and colour displays separately, and use the old Rockbox blue as the background...
18:41:26*Llorean didn't realize boot order depended on the version of mknkboot rather than bootloader version.
18:41:53 Join Kage_ [0] (
18:43:24 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
18:43:54 Quit Kage (Read error: 113 (No route to host))
18:44:05 Nick Kage_ is now known as Kage (
18:46:00linuxstbpixelma: This is what I think would work -
18:47:19bertrikNico_P: you were looking for playback bugs?
18:47:33Nico_Pbertrik: I'm always looking :p
18:47:47 Quit faemir (Remote closed the connection)
18:48:43Nico_Plinuxstb: wouldn't using HAVE_LCD_COLOR be cleaner (I know the old code uses LCD_DEPTH, but it might be a good opportunity to clean up)
18:48:44 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
18:48:48 Quit Kage ("Dreaming Of Xian...")
18:49:02pixelmalinuxstb: almost the same as I have, I only used HAVE_LCD_COLOR in rocklife.c
18:49:19linuxstbNico_P: Yes, I was thinking about that...
18:49:35pixelma(for the new #if)
18:50:07 Join nplus [0] (
18:50:18Nico_Plooking at lcd.h, LCD_RGBPACK is defined iff HAVE_LCD_COLOR is
18:50:31pixelmalinuxstb: :)
18:51:22pixelmaand you used LCD_BLACK... /me learns
18:51:41Nico_Ppixelma: it's #ifdef HAVE_LCD_COLOR
18:51:44Nico_Pnot #if
18:51:52pixelmaah ok
18:52:37pixelmayes, well my c200 test build failed now...
18:54:21Nico_Ppixelma: after doing what?
18:54:35 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
18:54:41 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
18:56:43pixelmaNico_P: what you said - the '#if' only gave me "error: #if with no expression"
18:57:38Nico_Preplacing #if by #ifdef broke it?
18:57:57pixelmano, that fixed it
18:58:25bertrikNico_P: I'm seeing a problem where the cpu stays boosted during the last song of an album with repeat turned off
18:59:44Nico_Pwell that's strange
19:01:19bertrikNico_P: I do have logf enabled and USB serial enabled. I'll try it with an official build and open a bug in the tracker if I can still reproduce it, ok?
19:01:38Nico_PI'm fine with that
19:04:10pixelmalinuxstb: who will commit the fix now? I don't care as long as it happens in the not too far future and if you want to clean up some more...
19:04:57 Join Mathiasdm [0] (
19:05:00linuxstbpixelma: I'm happy for you to do it - at least you've compiled your version...
19:09:29pixelmaok, improvements can always happen later, I just want the builds fixed...
19:09:54pixelmaand thank you both for looking over it :)
19:10:19 Join moos [0] (n=c40cd9a5@gateway/web/cgi-irc/
19:10:42*preglow notices broken h120 build
19:11:57mooslinuxstb: Just read you encountred the strange bug too. I will try to fix it here the way you made, let's see....
19:12:17linuxstbmoos: Had you reformatted the main partition, or deleted the OF's files?
19:12:53*LambdaCalculus37 also notices that the greyscale targets are in the red right now
19:13:09mooslinuxstb: I didn't reformat yet
19:13:19BigBambiLambdaCalculus37: Just like me after buying a new big disk for the beast :(
19:13:22moosneither deleted OF files...
19:13:54linuxstbmoos: OK - I was wondering if that caused my problem for some reason...
19:14:01LambdaCalculus37BigBambi: That sucks. :(
19:14:13BigBambiIt'll be worth it :)
19:15:23 Quit Nico_P (Remote closed the connection)
19:15:37 Join barrywardell [0] (
19:16:13 Join DerPapst [0] (
19:16:35BigBambiLlorean: Ah, boot order depends on the version of mknboot eh? I've been wondering for a while why my boot order has stayed reversed after the change, but forgot to mention it
19:16:44barrywardelloops, sorry about the broken builds. I was sure I checked the builds were all green before I left!
19:17:00pixelmajust committed a fix...
19:17:05*LambdaCalculus37 bonks barrywardell with his squeaky hammer :)
19:17:07DerPapstBigBambi: it's an ifdef in mknboot.c
19:17:26*BigBambi goes to build new mknboot
19:18:08 Part LinusN
19:18:24*LambdaCalculus37 watches the new build to check for red
19:18:26 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
19:18:59*barrywardell thanks pixelma and apologies again
19:20:31 Quit simonrvn (Read error: 113 (No route to host))
19:20:52pixelmawe'll see... it's a bit of new ground for me these kind of fixes...
19:21:18LambdaCalculus37pixelma: All green, save for the D2.
19:21:42DerPapstsomeday you'll port rockbox to a new target on your own i bet :-P
19:22:06LambdaCalculus37DerPapst: It takes time, learning, and hard work, of course. :)
19:22:31LambdaCalculus37I hope to add the Dell Digital Jukebox to that build table one day. :)
19:22:32 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
19:22:52*pixelma wonders if a bit of TurboPascal knowledge will help there :P
19:23:42*DerPapst hands pixelma lousy 3 points :-)
19:26:52 Join simonrvn [0] (i=simon@unaffiliated/simonrvn)
19:31:12barrywardellthe colour settings in the themes menu seems wrong. entering the colour select screen and accepting the same colour (ie. not changing it) gives and invalid colour error splash
19:31:27barrywardellsurely the splash is unnecessary, isn't it?
19:36:23***Saving seen data "./dancer.seen"
19:36:45 Join simonrvn_ [0] (i=simon@unaffiliated/simonrvn)
19:37:27 Quit simonrvn (Nick collision from services.)
19:37:32 Nick simonrvn_ is now known as simonrvn (i=simon@unaffiliated/simonrvn)
19:37:52 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
19:38:01 Join bluebrother [0] (n=dom@rockbox/staff/bluebrother)
19:40:34DerPapst*PANIC* Unhandled IRQ -1 in irq_hander: <unknown> with the beast bootloader r17364 (i think v0.0[...]011)
19:40:46DerPapsteverytime i connect usb
19:41:38 Quit linuxstb (Nick collision from services.)
19:41:40 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
19:42:06 Join MrDOS [0] (
19:42:07jhMikeSDerPapst: I've gotten that one very occasionally but never every time with anything. :\
19:43:14MrDOSquestion - what is the best way I can, as a non-developer, aid the development of the port to the iPod Nano 3rd-gen?
19:43:32BigBambiMrDOS: I don't think there is much
19:43:50DerPapsti'm trying rockbox' usb mode now.
19:44:04DerPapstand btw: thank you for sound :-D
19:44:14BigBambiMrDOS: Someone needs to find out how to crack the encryption to run third party code, and after that start trying to work out the completely undocumented hardware
19:45:10*linuxstb would never suggest breaking into Apple's engineering HQ and stealing the blueprints and encryption keys...
19:45:12MrDOSBigBambi: Ah, I wasn't aware of the encryption problem... I thought it was "simply" a matter of getting the bootloader (and then the rest of the code) to function on the new hardware. That's no fun
19:45:32LambdaCalculus37That's also going to apply to the iPod Classic and 2nd gen nano as well.
19:45:43MrDOSyeah, I thought it did
19:45:50 Join petur [50] (n=petur@rockbox/developer/petur)
19:46:03MrDOSso Classic/{2|3}gen Nanos use the same base hardware then?
19:46:11BigBambiMrDOS: Even if there were no encryption, writing drivers for completely undocumented hardware is anything but simple (I realise you used quotation marks)
19:46:39LambdaCalculus37MrDOS: The SoC on the three devices appears to be similar.
19:47:03MrDOSYeah, the quotation marks were there for a reason - I'm amazed that Rockbox can even run on iPods
19:47:08LambdaCalculus37But there's no information about it available publicly.
19:47:37 Join gregzx [0] (
19:47:48LambdaCalculus37MrDOS: The older iPods all used the same Soc: PortalPlayer.
19:48:10BigBambiBut different variations of that
19:48:21LambdaCalculus37But it still took reverse engineering to get code onto them, didn't it?
19:48:38MrDOSyeah, that's what I'm commenting on - that had to be a lot of work
19:49:02MrDOSanyway, can anyone recommend a good place to get a 2gb 1st gen nano then?
19:49:05BigBambiAnyone have a pic of the beast running Rockbox lying about, or do I need to dig my camera out?
19:49:11BigBambiMrDOS: ebay I guess
19:49:16MrDOSthought that'd be the response
19:49:34MrDOSthanks for the help, guys
19:49:56 Quit MrDOS ("Leaving.")
19:50:00linuxstbMrDOS: Why do you want a Nano? ....
19:50:02markunMrDOS: you could also get a sandisk sansa e200 for example
19:50:10markun(1st version)
19:50:11linuxstbWe missed him...
19:50:24markunwe're slow :)
19:50:38linuxstbOr he could get a nice Meizu...
19:53:16 Join kugel [0] (n=chatzill@unaffiliated/kugel)
19:54:04*kugel is glad to see, that the situation is cleared now (
19:54:35kugelHopefully it can be committed soon (it needs some bug fixing)
19:57:15 Join webguest39 [0] (n=44a1e097@gateway/web/cgi-irc/
19:58:30 Quit webguest39 (Client Quit)
19:59:34 Quit Horscht (Nick collision from services.)
19:59:52 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
20:00:36bluebrotherpetur: can you tell me if the H10 pure needs to have the mi4 file named H10EMP or does simply H10.mi4 work (for uninstallation)?
20:00:43bluebrotherI guess the latter ...
20:07:15 Join Buschel [0] (
20:09:01bertrikgevaerts: how does USB serial work? I just enabled it in the debug screen, but I still get the mass storage.
20:09:05 Join cg [0] (n=cromos@
20:09:32bertrikdid you create two USB configurations? or multiple interfaces?
20:09:42gevaertsbertrik: you get a composite device then. If you only want serial you have to disable USB_STORAGE in usb_core.h
20:11:05*petur wonders why his irc client didn't highlight
20:11:59*DerPapst tests...
20:12:35moospetur: does it highlight for *beer* ? :)
20:12:36peturbluebrother: I have only H10.mi4 and OF.mi4 in my system dir
20:12:57peturmeh... my highlights are gone :/
20:13:20*amiconn thinks about forking rockbox
20:13:26 Join Shof [0] (
20:13:34bluebrotherwhy that? (forking)
20:13:42 Quit barrywardell (Remote closed the connection)
20:13:47Shofi have a quick question
20:13:54BigBambiask away
20:13:55DerPapstthe fork would be a rockbox light version
20:14:38Shofi just updated to r13790 on my h320, and now when i turn it on. the main page that shows the rockbox logo and ver. number is black when it used to be white..anyway to chance that?
20:15:20LambdaCalculus37Do you mean r17390?
20:15:37Shofr 17390
20:15:44BigBambiShof: Only by editing the code
20:15:49Shofoh ok
20:16:06Shofjust a bit strange that in the other releases that screen is white while now it is black
20:16:18BigBambiSee the front page
20:16:23BigBambiIt lists recent commits
20:16:39LambdaCalculus37A commit was just made to change the boot and splash screens.
20:16:48BigBambiAnd one of those was changing this. Also check the IRC log from today for reasoning
20:17:09Shoflend a helping a hand a link or something
20:17:11 Quit petur ("switching...")
20:17:15 Join petur [50] (n=petur@rockbox/developer/petur)
20:17:20amiconnbluebrother: In order to again have something that is fun to work on, without being pissed about useless feature creep, code bloat and slowdowns...
20:17:51Shofi see the irc log now
20:18:51 Join AceNik [0] (n=AceNik@
20:18:52 Join spiorf [0] (
20:19:14AceNikhey guys i dont understand what this patch does 7977
20:19:34BigBambiproper link?
20:19:41*BigBambi can't be bothered to do it manually
20:19:44Shofit is normal that when i turn on my h320, i get a white screen that shows the battery info in white and it changes to the rockbox splash screen which is black?
20:20:12moosamiconn: long time that we didn't have a feature freeze period with organisation of bug hunting...
20:20:26bertrikamiconn: what parts do you consider bloated?
20:20:30Shofok then
20:21:10BigBambiShof: I haven't looked at the new changes
20:21:26BigBambibut a switch like that sounds not the prettiest
20:21:40*bluebrother would consider antialiased fonts bloat on a dap
20:21:47*BigBambi too
20:21:54*bertrik agrees
20:22:07BigBambiBut they are just a patch
20:22:09*jhMikeS disagrees after having seen them
20:22:09amiconnbertrik: I would remove a whole lot of things, mainly stuff I never use but which adds large chunks to the binary
20:22:17*scorche|sh wonders if he should add a few more ifdefs
20:22:43BigBambiTo be honest, I haven't looked at the AA on target
20:22:46AceNikguys what does this patch do, 7977
20:22:58BigBambiAceNik: If you give me a proper link I'll go and look
20:23:02jhMikeSthe beast OF uses true type AA font rendering
20:23:18Nico_PI think the beast can afford it :)
20:23:27bluebrotherI never understood why this should be useful on a dap. You don't look hours on it anyway.
20:23:34BigBambiThe beast is a law unto itself
20:23:35jhMikeSI think others could too
20:23:37linuxstb12 hours isn't a very good runtime though (in the OF)...
20:23:54LambdaCalculus37AceNik: That was just committed today.
20:24:09LambdaCalculus37It displays the Rockbox logo while your player is booting up.
20:24:12BigBambiAceNik: It shows the logo during the bootloader
20:24:31AceNiklambda: i know but whats the use, the rockbox logo always shows, with the build date at the bottom ?
20:24:42BigBambiAceNik: That is the build not the bootloader
20:24:49BigBambiThe bootloader comes first
20:24:55AceNikwhat logo rockbox logo right ? but that was always shown?
20:25:06Nico_Pamiconn: you could build custom versions with some features disables (I would guess DB, album art and things like that)
20:25:08jhMikeSbefore the egg or the chicken?
20:25:08linuxstbAceNik: Which player do you run Rockbox on?
20:25:23AceNiki see ill compile a new build n try it out then thanks
20:25:44*Nico_P was thinking about possibly adding customization to tools/configure
20:25:52*BigBambi thinks forking over not using some features that others do would be a little extreme
20:25:56LambdaCalculus37Does it show an iriver logo when you start it up?
20:26:00amiconnDatabase, album art, cue sheet support, logging, horizontal list scrolling, crossfeed, crossfade, graphical eq, this weird study most, directory skipping of the currently existing features, perhaps also most wps customisation options.
20:26:03 Join Hillshum [0] (
20:26:13amiconns/study most/study mode/
20:26:29gevaertsNico_P: the problem with that is that you actually need to build all combinations for every revision
20:26:29AceNiklambda: yes it does then it proceeds to a rockbox logo
20:26:42*BigBambi doesn't use lots of those but uses a few, and thinks that you can never please everyone
20:26:43LambdaCalculus37Then the iriver logo would be the bootloader starting up.
20:26:43Nico_Pamiconn: wow, that's quite a list. presonnally, there are plenty of features I don't use but I don't mind them
20:26:55*amiconn would rather try to concentrate on bug fixing and performance optimisation, plus new ports
20:27:07Nico_Pgevaerts: hmm, I'd build the all features build by default
20:27:20amiconnNico_P: I probably even forgot a few.
20:27:25gevaertsNico_P: that will lead to ifdef errors
20:27:38*bertrik uses database and album art but could live without the rest
20:28:07*BigBambi uses album art,, scrolling, and some WPS features
20:28:20*bluebrother would like to see more bugfixing too
20:28:24*linuxstb uses none of them
20:28:26*Nico_P agrees
20:28:28amiconnBigBambi: I'm not talking about scrolling text in general, btw
20:28:43*gevaerts wonders if a more advanced plugin system would help, i.e. have wps plugins, dsp plugins, playlist control plugins,...
20:28:51*LambdaCalculus37 uses Album Art,, the database, and the EQ
20:28:53*jhMikeS is largely a "hunt and peck" listener
20:29:00*bertrik was thinking the same as gevaerts
20:29:04BigBambiamiconn: I suspect I have misunderstood - which scrolling do you mean?
20:29:06domonokybluebrother: can you reproduce FS #8948 on Linux ?? for me it works fine (on windows) ..
20:29:07Nico_Pamiconn: like it or not, the userbase (and some devs too) likes eyecandy and fancy features
20:29:28*markun uses crossfeed and replaygain
20:29:33amiconnBigBambi: The one that allows you to scroll the whole list left or right manually
20:29:40BigBambiI like being able to show Rockbox to people and not have the immediate reaction "That looks shit"
20:29:40bluebrotherdomonoky: haven't tried −− will do later.
20:29:41AceNikguys will the committed patch 7797 conflict with this
20:29:43*jhMikeS likes nice eye candy and ear candy
20:29:46BigBambiamiconn: ah, I do use that :)
20:30:04scorche|shamiconn: i dont use any of those features either, but i cant see that rockbox should take those out
20:30:17BigBambiI have some badly named songs that would take forever to scroll otherwise
20:30:25amiconnscorche|sh: Well that's why I'm thinking about a fork
20:30:43*jhMikeS doesn't use voice either, finds it a development pita but doesn't think that should go
20:30:48LambdaCalculus37AceNik: FS #7538 hasn't been updated since January.
20:30:48BigBambiThat would be unfortunate
20:31:19markunamiconn: do you think it would be an option to add some ifdefs to leave out some features in a custom build instead of forking?
20:31:34markunor does a fork have other advantages?
20:31:50AceNiklambda: its working fine with the current svn revisions, i would actually likt this path t be committed btu it seems its not writtent he proper way
20:31:59amiconnWhile I can see that some of these features seem to be wanted by a number of people, I dislike them, and their gradual accumulation often eats my motivation to do some rockbox coding
20:32:23LambdaCalculus37AceNik: I don't think that patch will be accepted. It would also cause support nightmares.
20:32:40 Join jgarvey [0] (
20:32:52*Nico_P thinks adding customization to tools/configure would be a good solution
20:32:56LambdaCalculus37Because then people will come screaming that something is broken because of it, but we'll never be able to pinpoint what broke what or who broke what.
20:33:07markunNico_P: me too
20:33:12AceNiklambda: yes i know that, i jus wanted to know whether the patch would conflict with current svn ill try it out
20:33:25*Buschel also likes eye candy (if it doesn't affect battery life ;)
20:33:30amiconnmarkun: A fork would have the substantial advantage that I wouldn't have to care about those features (not breaking them etc)
20:33:34linuxstbAceNik: That patch has been formally rejected... (7538)
20:33:36Buschelare there any screenshots with AA fonts?
20:34:07linuxstbThe patch had some
20:34:09 Quit AndyI ()
20:34:11Nico_Pamiconn: I think a fork would be really bad for the project as a whole (and honestly, not very popular either)
20:34:36toffe82markun: Nico_P : so the current build would be a full feature and if you want a custom you have to build it yourself ?
20:34:59*jhMikeS thinks breaking up the rockbox DAP monopoly is long overdue :p
20:35:02Nico_Ptoffe82: I guess. I haven't thought about it much
20:35:09amiconnNico_P: That might be the case, but it might not be avoidable
20:35:15gevaertsI think this should be talked about at devcon
20:35:19Nico_Pamiconn: I'm sure it's avoidable
20:35:28*LambdaCalculus37 agrees with Nico_P's idea of adding customization to tools/configure
20:35:39*BigBambi thinks it would be a bit of a disaster
20:35:41Nico_Pamiconn: but I guess you know that better than me
20:35:44bertrikbut wouldn't that bring even more #ifdef madness?
20:35:46AceNiklinuxstb: you told me to et familiar with the rockbox svn i want to start with custmising the interface for rockbox, firstly by extending th customising of the main menu with themes
20:35:51markunI'd rather have amiconn start a fork than to stop working on rockbox
20:36:08Nico_Pbertrik: I like a few more #ifdef more than a fork
20:36:51amiconnBtw, the statement that made me think about a fork was kugel's mention of font anti-aliasing (imho the most useless thing on a dap or other embedded device)
20:37:10*BigBambi blaims kugel for discord in the ranks
20:37:20 Quit Hillshum ("ChatZilla 0.9.81 [Firefox]")
20:37:41bluebrotherI was quite ... surprised by that "soon" too.
20:37:43kugelI don't understand
20:37:46*markun likes anti-aliasing, also on a DAP
20:38:03*jhMikeS too (did /me mention that?) :)
20:38:24amiconnI'm even disabling normal anti-aliasing on windows. It only makes the text blurry and harder to read. The only anti-aliasing I do use is the lcd sub-pixel rendering (on *fast* PCs)
20:38:27jhMikeSI suppose it helps the bigger screens more than the small ones
20:38:43*linuxstb would also quite like a far simpler Rockbox, but accepts Rockbox will always grow in size...
20:39:00kugelamiconn: What did you mean?
20:39:08Nico_Pamiconn: what I don't really understand is how features you don't use can become a problem if they don't negatively interfere with regular use (binsize isn't a strong argument in my eyes)
20:39:29linuxstbNico_P: Code complication and bloat...
20:39:30jhMikeSI'm sure tighter UI code and such would take quite a chunk out. Things like lists seem more complicated that they have to be.
20:39:43*Nico_P agreees with jhMikeS
20:39:44BigBambiThere are plenty of features I don't use, but I don't see how I can say they are less important than others, when other people think different things
20:39:53pixelmaseeing what full blown pc graphic programs do when it comes to antialiasing of small fonts, leads me to thinking it's useless on a DAP
20:40:50*kugel remains confused
20:40:51bluebrotherkugel: don't you think it somewhat to early talking about "committing soon" if you tell the same time that there are issues?
20:41:15bluebrotheralso, quite some people (including me) consider such a feature useless and bloat.
20:41:15amiconnNico_P: They negatively interfere with coding.
20:41:19BigBambikugel: You should check out the whole discussion in the logs
20:41:35jhMikeSI did write a whole custom control for a windows program and it's 5k lines but it also does quite a bit of fancy stuff so 5k lines isn't so much.
20:41:47kugelbluebrother: No? Not if I said "it can hopefilly committed soon", which just expresses my wish
20:42:32jhMikeSI don't know how many lines of list code exist but it's sort of in that category
20:42:33kugelBigBambi: where to start? I was quite a long time absent from this channel
20:42:46bluebrotherwell, I don't think it's a good idea expressing a whish like that, especially given that's a controversial feature.
20:43:06BigBambikugel: 30 minutes ago
20:43:27BigBambibluebrother: Thought control? :P
20:43:32Shofthe only features that i dont use on my rockbox that much is the most of all the games and demos. It will be nice if there was a a sperate release for that since it can make the size of the rockbox download smaller and use like 1-2mb less
20:43:50 Quit simonrvn (Remote closed the connection)
20:43:54BigBambiShof: Plugins are largely irrelevant though
20:43:55bertrikthe plugins don't add to the binsize AFAIK
20:44:09*linuxstb doesn't think binsize is the issue
20:44:29jhMikeSthe plugin buffer cuts away from the possible audio buffer though
20:44:52 Join AndyI [0] (n=pasha_in@
20:45:09BigBambitrue, but once the buffer is there adding more plugins (bar linking to the core) is largely free
20:45:09 Quit gregzx ("ChatZilla 0.9.81 [Firefox]")
20:45:13preglowi'd like a tighter rockbox for sure, but i do see that won't happen according to my specific likes and dislikes anyway
20:45:33linuxstbpreglow: Stop coding features you don't use then ;)
20:45:33Nico_Pthat's where feature customizations comes in
20:45:42BigBambiI think everyone would like that
20:45:59BigBambi(tighter Rockbox that is)
20:46:09BigBambiBut everyone uses different features
20:46:12*linuxstb thinks feature customisation would just make things more complex...
20:46:16preglowlinuxstb: what, you mean like my piece of art eq? and deprive people of valuable features???
20:46:16*gevaerts thinks that feature customizations need a lot of code reorganisation first. #ifdefs scattered around the code are not good
20:46:34jhMikeSwhat about having a master config (more master than config.h) that specifically disables certain parts?
20:46:37preglowifdeffing out features is out of the questin, if you ask me
20:46:46preglowthat'll just make it harder for devs to make sure they don't break things
20:47:07preglowamiconn: i think "study mode" should be renamed, if nothing else, but i also don't see the use of it
20:47:10amiconnSeeing that I'm obviously a minority and cannot dictate rockbox development, I have 4 options. (a) download some svn from end-2007, apply fixes for known bugs, build that as "rockbox 2.9" and install it on the daps I want to keep for real use. Sell the others. (b) fork rockbox as mentioned, and work on stuff I want to. (c) keep my coding to a bare minimum (fixing stuff I wrote). (d) somehow arrange with the ton of features, and continue as before
20:47:29amiconnSomehow (c) and especially (d) don't sound very tempting....
20:47:39amiconnAnd I agree, more ifdefing is not an option
20:48:09linuxstbamiconn: Is it just apps/ code you don't like, or things in firmware/ too?
20:48:42*jhMikeS is quite sure he hates the kernel with a passion :)
20:48:56Nico_Pamiconn: what would you code if there are no new features you want to add? how would you exploit new powerful targets? in short: where would the fun be?
20:48:59amiconnVery few things in firmware too
20:49:20Nico_Pthere are very few things you like?
20:49:28preglowjhMikeS: fs #8918 status?
20:50:10*amiconn isn't tempted by any of the new targets
20:50:23pixelmaNico_P: adding new features isn't fun when the already existing code doesn't work nicely (enough). For a user too... (thinking of the resume last 2 seconds problem etc.)
20:50:55kugelMeh, /me starts getting it
20:51:00BigBambiThat is very personal
20:51:01Nico_Ppixelma: the resume last 2 seconds bug isn't fixed?
20:51:03amiconnThey have useless large screens, nasty touch controls, and the DAC of every new dap generation seems to sound worse than the previous ones...
20:51:29jhMikeSNico_P: I don't even mention playback bugs anymore. I can always make it do something weird. :\
20:51:33bluebrotherhmm, why is this strange "study" feature called "tracklock" in the tracker task but not anymore?
20:51:54BigBambiI don't see how it relates to studying in the slightest
20:52:00pixelmaNico_P: nope
20:52:02*linuxstb didn't notice a "study" feature and goes to check the logs...
20:52:14amiconnI'd want a large screen if I'd want a video player (as its *main* purpose, aka PMP) - but that's something I don't need at all
20:52:21jhMikeSstudying a small section of audio?
20:52:26bluebrotherBigBambi: me neither.
20:52:27Nico_PjhMikeS: depends if it's a new different weird thing each time
20:52:36Nico_Ppixelma: damn, I really thought it was
20:52:37*kugel was suprised when that study mode was comitted, that it was comitted
20:52:38preglowlinuxstb: the confusingly named and also useless feature called "study mode"
20:52:59Shofi noticed that feature and im still looking up what it does
20:53:07jhMikeSNico_P: no, just metadata out of sync. wrong play order, weird repetitions of songs, certain cases don't unmute audio hardware, etc.
20:53:11*BigBambi neither understands the name, nor sees the use in the feature
20:53:21bluebrotherShof: search the tracker for "tracklock" (make sure to search closed tasks too)
20:53:24preglowas far as i can see, it just makes rev/ffwd skip around in a track instead of skipping tracks
20:53:32preglowthis to avoid accidentally avoid skipping a long track
20:53:32kugelit basically locks trackskipping in favor of jumping a set time into the track
20:53:40jhMikeSit's especially vulnerable if you induce race conditions by banging the controls rapidly
20:53:47preglowbut if you don't want to accidentally control you dap, why don't use the hold switch?
20:53:48bluebrothermaybe we should implement some kind of voting process before new features get in?
20:53:59preglowbluebrother: or the rockbox steering committe
20:54:09markunpreglow: I vote for that
20:54:16preglowme too
20:54:19bluebrotherpreglow: good point. I think we really need such a thing
20:54:26LambdaCalculus37Me three.
20:54:29preglowwell, i've always thought so
20:54:32mooswas the study mode debated here?
20:54:35*domonoky forth.. :-)
20:54:36preglowand i'm glad to see some thought going into it
20:54:38amiconnpreglow: Not all targets have a hold switch... but all targets have a hold feature at least in the wps
20:54:50*scorche|sh wonders why LambdaCalculus37 is voting when he hasnt heard the proposal ;)
20:54:51bluebrotherwe could make it a two steps process: let devs vote and the committe decide.
20:54:55preglowamiconn: and i think that should be enough to avoid accidental keypresses of the kind described in the study mode feature entry
20:55:01 Join midgey [0] (
20:55:07BigBambiSteering committee?
20:55:08preglowas far as i'm concerned, i'd like that feature out of the tree again
20:55:23scorche|shBigBambi: something that was proposed a bit back on committers list
20:55:41BigBambiscorche: To have a few devs make executive decisions?
20:55:42amiconnpreglow: Definitely. And the other thing it does, skipping around in small steps, can be done by seeking as well
20:55:51preglowamiconn: exactly, it just adds more complexity
20:55:51scorche|shBigBambi: something like that
20:55:57BigBambiSounds a reasonable idea
20:56:08 Quit XavierGr (Connection reset by peer)
20:56:24*BigBambi would remove study mode too (not that it counts :))
20:56:41*moos too
20:56:47*bluebrother too
20:56:48amiconnpreglow: Actually, I would perhaps not mind it if it were less code. This is not only from the binsize point of view, but also from the complexity point
20:56:50*LambdaCalculus37 doesn't even know what the hell study mode does; he's never used it at all :)
20:57:01*scorche|sh joins the people in wondering what the hell stufy mode is
20:57:05amiconnEspecially since A-B repeat is doing something *very* similar
20:57:07scorche| mode
20:57:12BigBambiLambdaCalculus37: Rest assured, it does sweet FA related to studying
20:57:33LambdaCalculus37BigBambi: That's what my textbooks are for. :)
20:57:42preglowamiconn: yeah, i noticed it added a big amount of code for what it did
20:58:05linuxstbReading the FS task - - LinusN said he thought it "looks useful".
20:58:20markunI think complexity can also be reduced in some cases without reducing the code (like the target tree)
20:58:23linuxstbAnd Llorean seems to blame for calling it study mode - he referred to it being similar to an iriver feature.
20:58:23LambdaCalculus37I figure if people want a study mode, we can tell them to set repeat mode to A>B and let them go from there.
20:58:49Nico_Ppixelma: I just tested with r17143 on my ipod. I know it's not the latest but I don't think there were any major changes since then. I'll test with a more recent version later
20:59:02amiconnNico_P: The 2..3 seconds resume-at-end is definitely unfixed. It only affects swcodec, as on hwcodec, I get the 'End of playlist' splash+announcement everytime. Never so on swcodec
20:59:29 Part Shof
20:59:34 Join Synergy6 [0] (
20:59:58Nico_Pamiconn: see what I said above. maybe I'm not talking about the same bug though... would you mind reminding me the repro recipe?
21:00:41Nico_Plet me also remind you that the tracker is useful for those things. feel free to assign playback tasks to me
21:01:51amiconnNico_P: Just play an arbitrary folder, starting on any file *except the last one* (second-last is easiest for reproducing the bug). Stop playback after some time, then resume and let it play to the end
21:02:48amiconnYou won't observe the bug if you play a folder from start to end without intermediate stop/resume cycle(s), and you also won't observe it when starting on the last file
21:03:42Nico_PI need to stop at some point in the last track... let it play to the end and then resume?
21:03:46 Join simonrvn [0] (i=simon@unaffiliated/simonrvn)
21:04:00Nico_Pthat's not what I remembered, so I guess it's why I thought it was fixed
21:04:19amiconnNo, you need to stop at some point in the second-last (or earlier) track
21:04:29Nico_Phence the point of using the tracker
21:04:39amiconnI get this all the time the way I'm listening normally
21:04:41Nico_Poh, then I think I was doing it right
21:05:59amiconnThis explains why the 3-second resume only "works" once. It resumes into the last track, and hence it won't happen a second time. But even then the 'End of playlist' splash+announcement doesn't work iirc
21:06:06 Quit AceNik ("Trillian (")
21:06:52Nico_PI don't see what the "end of playlist" splash is. I only get "nothing to resume"
21:07:32amiconnThat slpash is supposed to appear when playback ends at the end of playlist
21:07:33 Join roolku [0] (
21:07:53amiconnIt works on hwcodec, and it used to work on swcodec too
21:08:28 Join lee-qid [0] (
21:10:05*Llorean doesn't think he should be blamed for the "study mode" name, he didn't come up with it but rather just knew iriver called it that.
21:10:14roolkuthe problem with the study mode is the small range (250s) and the tiny increments (1s) which make it useless for my purpose (skipping in audiobooks); even the original iriver version was better
21:11:34amiconnapps/gui/gwps-common.c line 291
21:11:54 Quit Seed (Nick collision from services.)
21:12:11LloreanBeing able to do coarse skipping within a large file is useful with audiobooks (for example 10 or 15 minute increments in a 8-hour long program) and can get you to a specific range much faster than seeking if you want to keep acceleration low for fine seeking later.
21:12:20 Join Seed [0] (
21:12:42LloreanBut I don't think it's worth it if it's bad in terms of complexity (I have not looked at all at what the patch added)
21:12:46Nico_Pamiconn: what's with that line?
21:12:53*amiconn would just do a coarse seek followed by a second, fine seek afterwards
21:13:07amiconnNico_P: That's the line that shows mentioned splash - if it works
21:13:36Nico_Pamiconn: I don't see a splash call there
21:13:56amiconngui_syncsplash(HZ, ID2P(LANG_END_PLAYLIST));
21:14:12amiconnThat splash is also voiced
21:14:51Lloreanamiconn: Which involves either waiting a large amount of time for acceleration to kick in, having high acceleration making finer seeking more imprecise, or changing the values between the two.
21:15:07Nico_Pamiconn: that's on line 326 here
21:15:26roolkuamiconn: I tried, but it is a lot faster with several 10min skips than seeking, especially if you need to recognise the place you want to seek to acustically
21:15:37amiconnNico_P: Oh, maybe it's because I didn't svn up yet
21:16:02 Quit Horscht ("IRC is just multiplayer notepad")
21:16:38*roolku would like the study mode reversed, so he can re-apply his own version. :)
21:16:51Nico_Pamiconn: I don't remember ever seeing that splash, but I'll get it to work again
21:16:55preglowbut what would be a sane way of implemeting the study mode functionality?
21:16:58amiconnroolku: What's your own version?
21:17:40 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
21:17:43roolkuamiconn: just a 10min seel if there is enough time left. otherwise skip to the next file
21:17:50Nico_Pamiconn: but as you didn't acknowledge my comment, let me reiterate: please use the tracker
21:18:10roolkuamiconn: but I realise that this is my personal preference and wouldn't want to bother others with it
21:18:26*amiconn doesn't really like the tracker, as its mail feature somehow isn't reliable for me
21:18:34jhMikeSpreglow: did I forget to response to you about FS #8918?
21:18:34preglowworks fine for me
21:18:39preglowjhMikeS: think so :>
21:18:52preglowjhMikeS: at least you didn't highlight me, and if not, answers tend to get lost if i vanish for a bit
21:19:00*jhMikeS has an allergy migraine so probably isn't thinking too good
21:19:01roolkuamiconn: for music, there is almost no interference because very few tracks are longer than 10min
21:19:06linuxstbamiconn: Do you mean individual emails, or emails sent to the rockbox-sf mailing list?
21:19:26Lloreanroolku: It needs to be able to be disabled. I have several audiobooks that are 1.5 hours files, and I want to skip to say the 5th one.
21:19:26amiconnlinuxstb: Individual mails. I'm not subscribed to rockbox-sf
21:19:38*preglow thinks all devs should be subscribed to the latter
21:19:53Nico_PI find it to have to much traffic
21:19:56amiconnroolku: Add a disable to it, and it should be fine.
21:20:00jhMikeSpreglow: The particular use of the mutex and volatile puzzles me so I really need to apply it and check
21:20:10LloreanI use the Flyspray RSS for new tasks, and the individual mails for watched tasks, and it works great for me.
21:20:24*Nico_P does the same as Llorean
21:20:28preglowLlorean: that does sound quite clever...
21:20:30*preglow never got into rss
21:20:56amiconnI have one single Cd that would qualify as audio book, but it's actually split into several tracks (13 iirc), and I have one single track that's >60 minutes long, but is not an audio book
21:21:13preglowrss 1, rss 2, atom... what to choose?
21:21:25amiconnLlorean: That watching is what doesn't work for me
21:21:34Nico_Ppreglow: either ;)
21:21:37 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
21:21:51DerPapstfun roloing rockbox.gigabeat (s) gave another panic :-)
21:22:01*linuxstb wonders if the cuesheet feature could be (ab)used for study mode - e.g. something like "auto-cue" which generates fake cue points
21:22:05 Join merbanan [0] (n=banan@
21:22:08Lloreanamiconn: If I recall my past experience, flyspray just refused to send emails to one of my email addresses entirely, but then I switched it to another one and it worked great. It's a big buggy in that area, I guess?
21:22:21amiconnLooks like it then
21:22:22Lloreanlinuxstb: That sounds like a good idea actually.
21:22:23n1spreglow, et all: I think the major point of the study mode for sdoyon was to be able to seek but not accidentially skip, I have no oppinion about the complexity though
21:23:04n1sand therefore using the hold switch would not help, but rather disabling track skipping would be almost there
21:23:16Lloreann1s: I think that's not a strong reason for it, though. If it were a method of preventing accidental skipping, one could just have an on/off for next track entirely, so that only seeking could be performed.
21:23:23roolkulinuxstb: I thought about it, but it is slightly less flexible, as it is not relative to the position (could be an advantage for other use cases though)
21:23:37Lloreann1s: Coarse seeking should be treated as "coarse seeking" rather than "a solution to fumbling fingers"
21:23:42preglown1s: i think the functionality should be reduced to some kind of "skip button behaviour" with options "skip track, skip 5 secs, skip 10 secs, skip 30 secs, skip 1 min, skip 5 mins, skip 10 mins"
21:23:42saratogai think study mode should never have been added
21:23:57saratogahonestly i think we have too many playback features given the trouble in getting playback working well
21:24:01Lloreanpreglow: I'd say 1, 5, 10, 15, and 30 minutes personally.
21:24:06DerPapstunhandel IRQ 1454384 in irq_handler: 3 squares (incase anyone cares ;-))
21:24:06roolkupreglow: that would work for me
21:24:16bertrikNico_P: I can also reproduce for the official build the problem where the CPU stays boosted for the last track of an album with repeat off:
21:24:17preglowLlorean: and i'd be all for it
21:24:23*Llorean thinks less than 1 minute will be acceptably short time for normal seeking.
21:24:24preglowsaratoga: hear hear
21:24:25n1sTo clarify I'm just stating what my impression was, I don't care for the feature at all
21:24:38Nico_Pbertrik: ok, I'll take a look. thanks
21:24:41linuxstbIt doesn't touch playback.c though...
21:25:00 Join webguest42 [0] (n=4ab4c9eb@gateway/web/cgi-irc/
21:25:08amiconnNico_P: Btw, I tested on H10 and didn't get the 3-seconds resume, but didn't get the 'End of song list' (as it'
21:25:17webguest42whats another program that like rockbox that does teh new ipod mini?
21:25:21amiconns now called in english.lang) splash either
21:25:33linuxstbwebguest42: I assume you mean "new ipod nano" ?
21:25:41amiconnNeed to verify on H180 - maybe it's a coldfire vs. PP inconsistency
21:25:47BigBambiLlorean: Is there a feed for opened tasks only?
21:25:59LloreanBigBambi: The main flyspray RSS feed just covers newly added tasks
21:26:15BigBambiWhich is that (I'm being useless finding it :))
21:26:33LloreanWell, in Firefox it's just a little orange RSS icon in the URL bar.
21:26:37amiconnThe pcm code seems to behave inconsistently
21:26:39n1samiconn: I think it's unfortunate you feel the desire to fork or quit but I understand, we really should consider new features more before they are committed
21:26:40*DerPapst is glad that the beast has a battery switch :-)
21:26:44 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
21:27:00jhMikeSpreglow: it's just that none of the code the mutex protects even potientially yields to other threads so how could it make any difference to its operation?
21:27:31 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
21:27:56n1swe should ask 1) do we want this 2) is it done in a good way or could it be done simpler/nicer
21:27:59 Nick Mouser_X is now known as Winamp_Slayer (
21:28:25saratogafor instance, do we really need crossfade on swcodec now that nearly all codecs are gapless?
21:28:50linuxstbcrossfade has nothing to do with gapless IMO
21:28:53n1sand I agree with preglow about that the steering committee (if one would be created) should decide on features that there are no consensus about
21:28:55Lloreansaratoga: There are a lot of people *really* attached to crossfade.
21:28:56*jhMikeS gets out the axe with "crossfade" stamped into it
21:29:09LloreanWhile I'd like to see the feature axed, you'd alienate a pretty hefty chunk of users.
21:29:09saratogaisn't it mostly just to cover up gaps?
21:29:25amiconnJust got a hard freeze on H180
21:29:35Lloreansaratoga: No. It also appeals to some unfathomable group of people who think music should fade into and out of each other
21:29:50saratogai mean if the album is gapless, I don't see the use, its not like i've ever thought "i'd like to hear two songs at once and break gapless while doing it"
21:29:50roolkusaratoga: it is nice for pleasant sounding skipping and random play
21:30:10Lloreansaratoga: People use it on shuffle, I guess.
21:30:13amiconnUrgh, and that freeze is even reproducable...
21:30:22jhMikeSLlorean: well, let us consider eliminating that user base ;)
21:30:24LloreanMight I recommend that all people with comments on the steering committee, who are committers, respond to the email?
21:30:36roolkusaratoga: I rarely listen to an album in order
21:30:50LloreanjhMikeS: Hey, if you'd get those people out of my hair, I'd be happy. :)
21:31:14saratogamaybe we could fork the project and give them the version with crossfade
21:31:35saratogaor at least I wouldn't mind cutting features like that in the near term until playback was completely stable, even if they're eventually readmitted
21:31:36LloreanIt'll show up in an unsupported build anyway.
21:31:58n1sIMHO we should think more before committing new features because it is a lot easier to never have a feature than to have it for a while and then remove it...
21:32:16*preglow agrees
21:32:17jhMikeSLlorean: exactly unsuppored and we'd have no responsibility to it any more
21:32:23 Quit OlivierBorowski (Read error: 104 (Connection reset by peer))
21:32:36*amiconn found that the pcm inconsistency was obviously fixed at some point
21:32:50jhMikeSpcm inconsistency?
21:32:52amiconnI wonder what that reproducable hard freeze comes from though
21:33:09linuxstbn1s: I agree, but the problem is that it's already happened - i.e. the playback engine got far too many features hacked in too early.
21:33:18bertrikwell a reproducible bug is better than a heisenbug :)
21:33:20saratogawell lets cut some
21:33:26saratogathink of it as a rolling feature freeze :)
21:33:45amiconnjhMikeS: At some earlier point, coldfire and PP behaved differently concerning peak calculation.
21:34:10jhMikeSamiconn: I unified all that code awhile back CF, PP, and sim
21:34:11roolkuyes, lets cut gapless playback ;p
21:34:19amiconnI don't remember which one did what. One used to freeze the peakmeters when pausing, the other let it fall down to zero
21:34:25webguest42is a rockbox program that works on the new nanos?
21:34:28*Nico_P thinks that rewriting the playback code at this point (although it hasn't been explicitely suggested yet) would be making the same mistake that made rockbox 3.0 fail
21:34:28preglowamiconn: well, better with a reproducible freeze than an inreproducible one
21:34:29Lloreanwebguest42: No.
21:34:30linuxstbroolku: gapless isn't a feature, it's the norm...
21:34:41Robust2amiconn, do you have some time to talk with me about my M3 problem ?
21:35:03jhMikeSroolku: ummm...I think I have another axe and not one for the feature. :) It's the main reason I ever started using rockbox.
21:35:12Lloreansaratoga: I wouldn't object to trimming away buggy features when working toward a 3.0, then after the release, slowly reimplementing (or reintroducing with bug fixes) what we can.
21:35:17 Join shotofadds [0] (n=rob@rockbox/developer/shotofadds)
21:35:30amiconnpreglow: Problem is that I have no idea what might cause it, although I was able to reproduce it 3 times (each time requiring the infamous paper clip)
21:35:36Robust2amiconn: it's about the remote display problem
21:35:47linuxstbshotofadds: Hi. Have you noticed that your three #warnings are the only thing preventing a green build table? ;)
21:35:54saratogaLlorean: that idea gets my vote, provided people were interested in actually working on it
21:36:15*Nico_P would definitely be interested in working towards a release
21:36:22 Join OlivierBorowski [0] (
21:36:25***Saving seen data "./dancer.seen"
21:36:26Lloreansaratoga: I actually more or less suggested something like that some time ago. I've been a long proponent of "3.0 doesn't have to have all the features the bleeding edge has right now"
21:36:26 Quit webguest42 ("CGI:IRC (EOF)")
21:36:36shotofaddslinuxstb: yeah, I spotted (and mentioned!) that yesterday. I'll kill them in a bit (they're the user timer functions)
21:36:57amiconnI had voice enable (car config) and a resume point stored. Dircache is enabled. Now I booted, entered "manage settings", and loaded my "hifi" config (no voice, smaller font etc). Then I wanted to resume (dircache was still scanning in the background) -> hard freeze
21:37:07linuxstbshotofadds: Any interesting D2 developments?
21:37:31saratogaNico_P: I don't want to throw out the current playback engine, but surely it could be simplified somewhat by cutting features until it was more managable?
21:37:36amiconnReset, reboot (old config was active again as there was no chance for rb to save the new one), repeated the same steps - same effect
21:37:36shotofaddsnah, I've been catching some rays. very rare around here :)
21:37:53saratogathe current system seems completely inaccessible to all but a few developers, and progress on fixing it has been gradual
21:38:09linuxstbThat sums up most of Rockbox...
21:38:13preglowi don't think the playback system will ever be fully manageable by all devs
21:38:14saratogaand many of the fixes seem to create new problems
21:38:28preglowit is to some degree complex because it needs to be
21:38:44saratogawell how complex does it need to be?
21:38:50jhMikeSplayback really needs to be "designed" and not "evolved"
21:38:59preglowjhMikeS: yeah, it is very evolved...
21:39:10linuxstbA job for devcon?
21:39:11preglowpreferrably intelligent design!
21:39:14amiconnRobust2: Problem is that I don't really understand why that happens. I have an M3 and an X5, each with a remote (they're interchangeable). I have no problems on both, with either remote. In fact I already slowed down the code somewhat from the point where it was getting unstable during development
21:39:17preglowa job for jhMikeS :>
21:39:25*shotofadds wonders when the sort order changed on the build and delta tables. confusing or what?
21:39:42linuxstbpreglow: We're in danger of wearing him out...
21:39:42Robust2amiconn, so there's nothing I could try myself ?
21:39:52jhMikeSpreglow: I don't think it needs to be nearly that complex (as in many distictions without a difference code paths).
21:40:04saratogai'd love to a breakdown of how much playback code actually is actually spent on what feature
21:40:13saratogado we spend a lot of crossfade? How about gapless?
21:40:15n1sNico_P: I think we should avoid risky changes to playback (maybe until DCE) and try to put out the infamous Release (tm) first.
21:40:41preglowi think crossfade does require quite some code here and there
21:40:44amiconnSo obviously some M3's, X5's etc can't go as fast as mine. Since I can't test, the only thing I can do is offering a bunch of test builds, which slow down one or another step of the data transfer, and hope that one of them fixes the problem
21:40:46preglownot to mention the enormous buffers it has
21:40:51jhMikeScrossfade is just an awful memory hog since it must store enough PCM for the entire fade duration.
21:41:12linuxstbDo you know how much RAM it uses?
21:41:22Robust2amiconn, anything is good, as i have no other option (original fw doesn't work anymore)
21:41:22amiconnUnfortunately I'm currently working on another problem atm. I didn't forget about that speed problem though
21:41:26Lloreanamiconn: iRivers are also shwowing a similar problem (if it's the remote LCD going blank)
21:41:43Robust2yes that's it Llorean
21:41:51jhMikeScode RAM? no. audiobuffer, just calculate it based on n seconds of 16-bit pcm.
21:42:04amiconnLlorean: I know. Btw, that's the task that doesn't want to notice me even though I enabled notification :/
21:42:12linuxstbjhMikeS: So the RAM is allocated based on the crossfade settings?
21:42:14Lloreann1s: I'd rather the release have a vastly reduced featureset and *work* than be somewhat buggy as we try to slap patches on the existing features to get it out the door.
21:42:14saratogabut thats only used if its enabled right? if i disable crossfade i get that ram back
21:42:27linuxstbjhMikeS: I mean dynamically?
21:42:36jhMikeSlinuxstb: yes, with a minimum of 3s of pcm buffer
21:42:42Lloreanamiconn: Okay, just making sure. I wasn't 100% sure if it was in flyspray yet, I have a hard time keeping straight all the sources of user feedback. =/
21:43:00saratogajhMikeS: the memory cost is small if I keep that feature disabled right?
21:43:24Robust2amiconn: does this mean you'll look at this after you finished the other problem ?
21:43:33amiconnRobust2: I need to...
21:43:33 Quit roolku ()
21:43:41jhMikeSsaratoga: then there's no cost over normal playback buffer-ahead
21:43:59saratogaaside from the (tiny?) binsize increase?
21:44:02Robust2amiconn: okay well, i'd be glad to help you testing it once you get there
21:44:37amiconnBut I can't say how long that other problem (dropouts on disk spinup on ipod mini) will keep me busy, especially as not being able to understand the scheduler isn't going to help :(
21:44:55n1sLlorean: my point is that playback seems to work pretty well now and risky changes, risk breaking it.
21:45:00jhMikeSsaratoga: 529200 bytes of PCM buffer with no crossfade
21:45:11preglowtoo much for my liking :/
21:45:23linuxstbCan it be disabled completely with a #define?
21:45:24Robust2amiconn: you're free to send me any test builds, i can report the results to you if you want
21:45:46preglowjhMikeS: have we tried lowering that somewhat lately, though?
21:45:50amiconnRobust2: I will probably upload the test builds to my webspace and link from the tracker task when I get there
21:46:49 Quit styleism (Connection timed out)
21:46:50n1sLlorean: also which features will improve stability if they are removed?
21:46:54jhMikeSpreglow: I never tried. mpegplayer can get along with a fraction of a second sized buffer though. I expect playback needs more because of varying loads and you don't want to kick in priority too aggressively.
21:46:56amiconn(especially as I intend to "bend" the gpl a bit, but not publishing the source code changes for the various builds before the test, or at least not making obvious which test build contains which change)
21:47:13Lloreann1s: Not something i can comment on personally, I'm not familiar with the code. But people who've actually worked on the code complain often enough of crossfade.
21:47:15 Join fml [0] (n=4fd3fa74@gateway/web/cgi-irc/
21:47:15amiconnThis is because I want to conduct something like a blind test
21:47:34Lloreann1s: Besides, if removing the features doesn't improve playback stability, then we leave 'em in. :)
21:47:50Robust2amiconn: sounds good, i guess i'll get notified once you get there
21:48:02fmlNico_P: Hello. I have something for you to revise: FS #8965
21:48:13Robust2my M3 has been dead for almost 2 years, a little bit longer won't hurt :p
21:48:15*Nico_P is on the phone right now
21:48:50saratogai would support picking a reasonably stable build, leaving it up on the download site, and then cutting down the playback engine to a minimum complexity and feature set
21:48:54*amiconn wonders if something like those blind tests is covered by the gpl, if I am publishing the patched source afterwards
21:49:06DerPapstRobust2: you can use voice until it's fixed :-P
21:49:11saratogaamiconn: you have to provide the source, but you can wait a while
21:49:25Lloreanamiconn: Under the GPL, you just have to give notice that you will provide the source if requested, anyway.
21:49:27saratogai believe you're actually allowed to mail source code, so a few days delay seems reasonable as long as its released eventually
21:49:58*jhMikeS wonders if who the GPL-enforcement team is...and if they come bust a cap in you
21:50:05DerPapsti do't think we need to be that exact with GPL in this case.
21:50:15LloreanHonestly, *we're* the GPL enforcement team for Rockbox violations. It's "our" copyright.
21:51:01jhMikeSand in the tradition of law-enforcement, we can enforce the law by breaking it </political commentary>
21:51:41*BigBambi tasers amiconn
21:52:37amiconnMaybe I should switch the order of those 2 rpoblems....
21:52:53saratogahas anyone noticed that playback.c is too big for the ViewVC web interface
21:52:55Robust2DerPapst: i listen while riding my motorcycle, so that will be quite impossible :d
21:53:01saratogait runs off the bottom of the page
21:53:12n1samiconn: IIUC you need to provide the source upon request at no cost or the cost of media + shipment so charge 20€ and post a package of floppies :)
21:54:07DerPapstRobust2: and you can read the remote while bikeing?
21:54:28BigBambiPosting a load of floppies woul dbe nice and expensive :)
21:54:38Robust2no, but i can operate it when I stand still, I just can shout into the mike with a helmet :p
21:55:00BigBambiYou don't need to
21:55:07BigBambiVoice reads the display to you
21:55:18BigBambiFor e.g. if you are blind
21:55:20DerPapsterr.. you don't navigate it by "speaking" to rockbox. it speakt to you, i.e. reads the interface.
21:55:25DerPapstlike if you were blind
21:55:27Robust2oh wait lol, i thought you operated it by speaking :d
21:55:33Robust2haha how silly of me
21:56:03 Quit LambdaCalculus37 ("CGI:IRC 0.5.9 (2006/06/06)")
21:56:05Robust2hmm I could try that then
21:56:26Robust2but might as well use my small phillips mp3 player, which is rubbish compared to the M3 though
21:56:34DerPapstthus if you have a rockbox target with a broken screen you can use almost everything except plugins.
21:56:43DerPapstbut that might change for some soonish
21:56:58Robust2how do you mean, change ?
21:57:27jhMikeSa video plugin seems an odd choice for voicing though
21:57:39DerPapstthere is a gsoc project that aims to localize the plugins and maybe add voice to the
21:57:46gevaertsjhMikeS: why ? Just use a TTS engine on the subtitles :)
21:58:01*gevaerts prepares to be hit over the head
21:58:20n1sjhMikeS: it should of course tell you what happens in the movie ;)
21:58:27*DerPapst saves that for devcon
21:58:34jhMikeSgevaerts: well, hmmm...
21:59:26jhMikeSTTS subtitles + the dialogue? ;)
21:59:39 Join styleism [0] (
22:00:32amiconnsaratoga: I guess you're using firefox...
22:00:34jhMikeSyes, a movie watcher feature that watches the movie for you then tells you what happened
22:00:52saratogaamiconn: yes is there a better option?
22:00:59linuxstbjhMikeS: Or rather, tells you if it's any good...
22:01:22amiconnsaratoga: There is no problem with playback.c in ViewVC. It's firefox that often goofs on very long pages with css
22:01:38pixelmaamiconn: did you have a look yourself?
22:01:49*DerPapst wonders how often he'll see gevaerts running away at devcon
22:01:56jhMikeSlinuxstb: so it needs a critic feature too?
22:02:05amiconnI checked in ff and got a big black block at the bottom. Pasted the link into IE7, and the page is perfectly readable
22:02:15pixelmaI know about the firefox problem but this rather looks like a ViewVC problem (because you see the footer...)
22:02:16n1splayback.c seems fine in viewvc using ff 3 beta 5
22:02:37pixelmano big black block here (seen it before)
22:02:41saratogan1s: I'm using Beta 5
22:03:14 Quit waldo (Remote closed the connection)
22:03:26n1ssaratoga: well, it's displaying correctly here
22:03:48linuxstbSeems fine for me - ff (Ubuntu)
22:03:57amiconnI am on ff, and that does have the big black block effect. IE7 shows the file all the way to the end
22:04:14saratogainterstingly it works in ubuntu but not Windows when using the same version of FF beta
22:04:24n1sI also got the corruption on ff 2.x
22:04:41*n1s 's on ubuntu
22:05:53 Join waldo [0] (
22:05:55*pixelma can see the file to the end in ff using "annotate" - was just surprised by the "simple" end
22:06:06amiconnWorks fine in iceweasel (debian-amd64)
22:06:55amiconnThese css related bugs in firefox are very inconsistent
22:08:24pixelmaah... "view" only gives me the big black block
22:08:42 Join barrywardell [0] (n=barrywar@
22:09:18 Quit desowin ("KVIrc 4.0.0 Insomnia")
22:11:37amiconn'View' produces the black block in ff on windows. Footer is okay as that obviously uses different css. 'As text' and 'annotate' work. IE7 has no problem with either view, but but needs quite a while to render 'annotate'
22:12:17amiconnSummarized under 'browser quirks'...
22:12:40 Join simonrvn_ [0] (i=simon@unaffiliated/simonrvn)
22:12:50gevaertsI've _never_ seen a webbrowser that I was really happy with
22:14:03 Join webguest67 [0] (n=4aaa3bc7@gateway/web/cgi-irc/
22:15:25webguest67Testing, using firefox so don't know if it will work
22:16:38webguest67right. Hi all, was reading the rockbox player v1 wiki page, have a suggestion: could the player have maybe 6 gb at a minum? This may require a har harddrive
22:17:31Lloreanwebguest67: You're best off voicing in the related forum thread, as nobody here is really involved with that.
22:17:47*Nico_P has finished on the phone
22:18:00webguest67ah ok, sorry.
22:18:11Nico_Pfml: given how study mode is causing controversy, I'll wait a bit before committing if you don't mind ;)
22:19:05Nico_Pn1s: I've been considering playback.c as being in feature freeze for some time. I really do my best to avoid breaking playback (but often still fail anyway)
22:19:52*bertrik spots something suspicious in wma.c
22:20:02 Quit K4rP4D ("Leaving")
22:20:03XavierGrRockbox forums doesn't allow guest posts right? How come I can see a couple of posts without a registered username?
22:20:13saratogaNico_P: sure you don't want to break a few things?
22:20:22bluebrotherI think those are posts of users that got deleted.
22:20:22Nico_Pthe main feature which would reduce the complexity of the playback code is auto dir change, but I guess that's not going away
22:20:29saratogabertrik: what?
22:21:08Nico_Psaratoga: hehe, I try to avoid it but I wouldn't mind starting from scratch with a better design
22:21:10bertrikasf_read_packet can fail (line 558) in which case packetlength may not have been filled with a valid value, but is used anyway
22:21:14LloreanXavierGr: If they delete their account after posting, you have a Username and Guest beneath it
22:22:02 Quit styleism (Connection timed out)
22:22:03bertriksaratoga: also we could easily get rid of some of the goto's
22:22:21XavierGrstrange, how does he expect me to answer a pm if he deleted his account? (I assume that he deleted his account and not a moderator?)
22:22:21webguest67ah while we're on the subject of playback, I've heard from some rockbox for the blind podcasts that when you pause a song the voice won't say anything.
22:22:24fmlNico_P: it's in controversy? I thought something that has been committed isn't in controversy anymore
22:22:40fmlNico_P: but of course wait if it's needed
22:22:50Nico_Pfml: oh that's not the case... just read the logs
22:23:25 Quit simonrvn (Read error: 113 (No route to host))
22:23:27 Nick simonrvn_ is now known as simonrvn (i=simon@unaffiliated/simonrvn)
22:23:29fmlNico_P: where/when?
22:23:38saratogathe last hour
22:23:40Nico_Ptoday's have talk about the feature
22:24:03Lloreanwebguest67: This is true. Due to the way things work right now, if you need voice you should stop, rather than pause
22:24:52saratogabertrik: ah if you want to make those changes go for it
22:25:11webguest67ah. Just figured i'd mention it while on the redesigning playback thing.
22:27:54Nico_Pit's a more lowlevel limitation (pcm code)
22:28:26webguest67Also do the voice options work in the virtual keyboard? Noone's demonstrated that, in the three podcasts I've found
22:28:55 Quit fml ("CGI:IRC (Ping timeout)")
22:29:49 Nick Winamp_Slayer is now known as Mouser_X (
22:30:09linuxstbNico_P: I wouldn't miss auto-dir change...
22:30:54*jhMikeS always uses auto-dir change
22:31:05saratogawhat does auto-dir change do?
22:31:19webguest67The reason I ask, I don't have a rockbox compatible player at the moment or I would try it.
22:31:23XavierGrI have to say I would miss it
22:31:33XavierGrsaratoga: change dir when the playlist ends
22:31:50linuxstbAnd that's in playback.c ?
22:32:01XavierGrvery handy when you play a track from a folder and expect to move to the next one as soon as the current folder ends
22:32:21 Quit mf0102 ("Ex-Chat")
22:32:32 Quit Buschel ()
22:32:44saratogathat does seem like an odd thing to put in the playback engine
22:32:53Nico_Plinuxstb: it's in playlist.c mostly but has effects on playback.c
22:33:07Nico_Pdir skip has too
22:33:51saratogais there some documentation about how the playback engine works or do i have to read the source code to get an idea?
22:35:04markunany settings_list.c guru around?
22:35:22Nico_Psaratoga: I have a diagram at
22:35:51Nico_Papart from that there's the wiki page, but it's a bit outdated. hopefully the code has quite some comments, and you can ask any questions you want
22:36:39markunI want to have a setting which can have the values 3,4,5,6,7 and for each display a string: very slow, slow, normal, fast, very fast
22:37:27webguest67But why start at 3? And what do these settings do?
22:38:31markunwebguest67: it's for smooth seek acceleration, and these values were determined by experimenting
22:38:42DerPapstshouldn't matter for the implementation
22:38:48Nico_Psaratoga: need a link to the wiki page?
22:39:13saratogai'm sure i can find it when i get a chanec to look through this
22:39:49Nico_Pwell, here you are:
22:40:22linuxstbmarkun: Aren't those values in seconds?
22:40:33webguest67right. I'm off all, will check out the forum, thanks.
22:40:52markunlinuxstb: they are in seconds/seconds^2
22:41:15markunso in 1/seconds I guess
22:41:16 Quit webguest67 ("CGI:IRC (EOF)")
22:42:06 Quit davina (Remote closed the connection)
22:44:08n1smarkun: Hz !
22:44:18markunyes :)
22:44:40markunwhich wouldn't tell our poor users (including myself) a lot
22:45:12linuxstbI generally prefer numbers to things like "very slow", "slow", "normal" etc though...
22:45:35pixelmaisn't there a seeking accelaration already?
22:45:45markunme too. We could state the *2 time in seconds like it is now
22:45:55markunpixelma: there is
22:46:19bertriksaratoga: if asf_read_packet fails, should we call advance_buffer at all?
22:46:29markunI changed the implementation to be continuous instead of steps of *2
22:46:50markunehm, that didn't make any sense
22:46:52 Quit barrywardell ()
22:47:33markunpixelma: would you like to try a patch?
22:48:13bertriksaratoga: also, do you have some broken wma files to test with?
22:48:18saratogabertrik: the idea is to skip past broken packets
22:48:25saratogai've got a WMA file it fixes
22:48:45saratogasince WMA files have fixed length packets, you can skip past any broken packet and find good ones later
22:48:59bertrikah ok, so without the advance buffer calls it will keep on failing on the same packet?
22:49:14saratogabertrik: yes
22:49:20markunsaratoga: btw, isn't it time to revert my falty imdct 'optimization'?
22:49:46saratogamarkun: yes I have the patch ready to do it
22:49:55saratogai just haven't given the idea much thought lately
22:50:27markunok, thanks. Did you look at the difference in output with and without my change?
22:50:29saratogai've been meaning to look over IRAM use in the WMA codec and try to get the output buffer into IRAM
22:50:36saratogamarkun: yes about a year ago
22:50:43Nico_Pmcuelenaere: here?
22:50:47markunwas it big?
22:50:53saratogait reduced the SNR on my test sample by about 40dB
22:51:12saratogai think the RMS error verses the MS decoder was something like -60dB
22:51:23saratogawhich is tolerable but not ideal
22:51:52mcuelenaereNico_P: yes
22:52:00saratogathe problem is that WMA nees a 32kb output buffer
22:52:13Robust2amiconn, i've been tetsing it ignoring the display problem, but it freezes and locks up all the time now
22:52:18saratogawhich makes it difficult to put it all into IRAM on 48k only targets
22:52:21Nico_Pmcuelenaere: about the libnjb guy you couldn't get a hold of: what's his name? also, did you try their ML?
22:52:29markunpixelma: anyway:
22:52:41mcuelenaereNico_P: quetzalcoatl. Do they have a ML?
22:52:47mcuelenaereah sorry, I meant
22:52:51saratogabut its tempting to keep your optmization and as well as shrink the window coefficients to 16 bits, and then try to fit in the output buffer in IRAM
22:52:52Nico_Pmcuelenaere: also, they are mostly the same guys as libmtp.
22:53:10mcuelenaereyes, but the guy I need to contact is/was at
22:53:32mcuelenaereI just referred to that one from libnjb because he had contact with quetzalcoatl
22:53:56mcuelenaerebut I mailed to him and he said he doesn't know him besides on the (ex-)forums
22:54:01Nico_Pah, alright. I misunderstood then
22:54:03saratogai guess it depends on what accuracy we really want from our lossy codecs
22:54:22Robust2anybody know if replacing a player's HD with a CompactFlash adapter can cause problems with rockbox ?
22:54:31Robust2it does with the OF
22:54:33linuxstbsaratoga: I think the general approach is to favour accuracy over speed.
22:54:57linuxstb(as far as is reasonable...)
22:55:24saratogawell whats a good target to aim for?
22:55:51Nico_Pamiconn, pixelma: I'm not able to repro the "last few seconds of last tracks" bugs
22:55:54mcuelenaereit seems even google cache removed the last remains of :(
22:56:06Nico_PI'll look into fixing the splash though
22:56:14saratogathe RMS error in most lossy codecs verses lossless is something like 30 dB, and none of our DACs do much above 80dB, so thats a wide range of possible values
22:57:19pixelmaNico_P: I couldn't either but I'm sure I saw it a few times during the weekend (normal use), running some late 172xx build (after austriancoder's changes). I'd like to be able to reproduce myself... :\
22:57:35pixelmanothing you can do currently, so sorry
22:59:29 Quit mcuelenaere ()
23:00:01 Quit tvelocity (Remote closed the connection)
23:02:05markunlinuxstb: would you like to try the patch?
23:02:54 Quit XavierGr (Nick collision from services.)
23:03:05 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
23:05:19saratogai think the ideal way to speed up the wma codec (particularly on cold fire) would be to figure out if the entire 32kb buffer is actually needed for overlapping MDCT windows, but its difficult for me to figure that out since theres so many window sizes in the format
23:09:37linuxstbmarkun: Sure.
23:10:26markunlinuxstb: try seeking through a file before and after the patch. I think you will see the difference.
23:11:14markunI probably shouldn't reuse the same setting since the values now have a different meaning
23:12:59 Quit moos ("CGI:IRC (Ping timeout)")
23:13:41markunlinuxstb: new patch (same url)
23:16:15 Quit Robust2 ()
23:16:25Nico_Pwhy do I keep getting SIGTRAP in gdb with viewport related funcs?
23:17:46 Join moos [0] (n=c40cd9a5@gateway/web/cgi-irc/
23:22:34linuxstbNico_P: I don't know... What viewport-related functions?
23:23:03 Quit petur ("stkov")
23:23:11Nico_Pactually it's plenty of driver funcs, not just viewports. googling seems to indicate it might actually be a bug in the kernel
23:23:24Nico_P(linux kernel, not ours)
23:23:55n1sNico_P: whaen do you get them?
23:24:37linuxstbmarkun: It seems to work well.
23:24:39 Quit miepchen^schlaf ()
23:24:40Nico_Pn1s: I had some breakpoints set, and after reaching one of them, gdb started telling me about loads of SIGTRAP signals
23:24:58 Join miepchen^schlaf [0] (
23:25:01markunlinuxstb: great. The only remaining issue is the names of the settings I think
23:26:31n1sNico_P: breakpoints where?
23:26:40Nico_Pn1s: in wps display code
23:27:37 Join mewshi [0] (
23:27:50mewshihow do I reset my sansa c200?
23:27:57mewshii accidentally formatted it... :\
23:28:00mewshihow can I fix it?
23:28:26 Quit XavierGr (Nick collision from services.)
23:28:38 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
23:28:47 Quit bughunter2 ("Leaving.")
23:29:58n1sNico_P: I can set breakpoints in gwps.c in several places and continue with no problems here, 2.6.24 on amd64
23:30:10n1subuntu hardy bte
23:30:30Nico_Pn1s: I'm on hardy too
23:30:33 Quit OlivierBorowski (Remote closed the connection)
23:30:36Nico_PI'll see if it occurs again
23:31:04 Join csc` [0] (n=csc@archlinux/user/csc)
23:32:01bluebrothermewshi: what do you mean by "formatted"? The data partition?
23:32:31 Quit moos ("CGI:IRC")
23:36:27***Saving seen data "./dancer.seen"
23:40:41 Join shotofadds_ [0] (
23:42:02 Quit domonoky (Read error: 104 (Connection reset by peer))
23:44:53 Quit lee-qid (Read error: 110 (Connection timed out))
23:48:30amiconnNico_P: Listening for real on H180 now; let's see whether the 3-seconds resume bug is gone.
23:48:32 Quit XavierGr (Read error: 110 (Connection timed out))
23:48:43linuxstbmewshi: Do you know about recovery mode? (I think the same works on the c200)
23:48:53*amiconn meanwhile reminds Nico_P of the "wrong index during track transition" bug
23:49:20Nico_Pamiconn: that's a tough one. all my attempts to improve that have resulted in other breakage so far
23:49:44 Quit dabujo ("( :: NoNameScript 4.2 :: )")
23:49:49Nico_Pand I'm pretty sure it's not new to any of my changes
23:50:22amiconnWell, it worked correctly before MoB iirc
23:50:32*amiconn should probably verify that
23:51:10Nico_PI think I tested the rev just before MoB and it bahaved the same, but of course confirmation is welcome
23:51:29amiconnIt's also a nasty one for me - I often tend to check the wps during transition, especially at home
23:51:51Nico_Palso, the fundamental reason why this "bug" exists isn't new to MoB at all
23:51:56DerPapstshotofadds: is there a reason why you included grey.h and helper.h in the touchpad test plugin?
23:52:38shotofadds_DerPapst: probably because they were in the plugin I copied. Should probably remove those :g
23:52:46DerPapstok ;-)
23:53:15merbanansaratoga: all the mdct windows are needed :/ unless you generate the on runtime
23:53:29 Quit amiconn (Nick collision from services.)
23:53:36 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
23:53:45 Quit ompaul (Client Quit)
23:53:50 Quit shotofadds (Read error: 110 (Connection timed out))
23:53:50amiconn(yes, I'm a using 'caption backlight' all the time when listening at home)
23:54:08 Nick shotofadds_ is now known as shotofadds (
23:55:32amiconnNico_P: I wonder why the playlist index is so special - all the other tags are properly delayed until the trackchange happens at the pcm layer
23:56:01saratogamerbanan: on most files yes
23:56:21saratogai believe only some very low bitrate files will leave off some of the window sizes, typically those below 32kbps
23:58:03Nico_Pamiconn: the WPS asks the playlist code for the value instead of using something that's in the struct mp3entry.
23:58:48 Quit n1s ()
23:58:50Nico_Pstroring the value in the mp3entry and making the wps access that was one of my leads, but it caused some problems with repeat and such
23:58:54 Nick basti is now known as RIP|Hannes (i=bnc@

Previous day | Next day