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

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

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

#rockbox log for 2009-08-13

00:00:47 Join truthtaco_ [0] (
00:02:09pamaurygevearts: I have another question about usb: does the ARC OTG chipset really have only 3 endpoints ? (each endpoint being in+out) That's not much
00:02:47gevaertspamaury: you can verify it if you like. There's a register that has the number. The PP chips have only 3 I think, while the freescale chips have 7
00:03:23gevaertsThese things are really built for DAPs, and if you can do MTP or MSC, manufacturers are happy
00:03:57 Quit DataGhost (Nick collision from services.)
00:04:05 Join DataGhost [0] (i=dataghos@unaffiliated/dataghost)
00:04:41kugelis the statusbar actually configurable on the player?
00:07:28 Quit domonoky (Read error: 104 (Connection reset by peer))
00:07:43 Join jordan`` [0] (i=gromit@
00:08:00 Quit ender` (" How long a minute is depends on which side of the bathroom door you are on. -- Theory of relativity")
00:08:03 Quit jordan` (Read error: 104 (Connection reset by peer))
00:08:42 Join Juice^ [0] (
00:11:40 Quit pamaury ("Quitte")
00:15:54JdGordon|kugel: i dont think so
00:16:01JdGordon|configurable how? disbale?
00:16:38krazykitdisbale? to un-christian-bale something?
00:17:01 Quit TruthTaco (Read error: 110 (Connection timed out))
00:17:19kugelhaving a date with steven segal
00:17:41*JdGordon| wtf to 10517
00:18:25pixelmakugel: the Archos Player?
00:19:41 Quit sajes (Read error: 104 (Connection reset by peer))
00:19:51kugelJdGordon|: o.O
00:20:43 Join HellDragon [0] (
00:22:50pixelmakugel: I don't think you can disable it and it wouldn't be useful at all - the icons are not part of the charcell display so they don't take up space there and you can't do anything else with them (fixed form)
00:23:10pixelmaI don't *know* for sure though
00:23:19kugelthe setting doesn't seem to be disabled
00:28:42JdGordon|oh noes... the 2 people who actually have a working player still are losing 200bytes from the setting!
00:29:26 Quit robin0800 ("Leaving")
00:30:19kugelthat's 100bytes per user
00:30:37 Quit jgarvey ("Leaving")
00:30:42JdGordon|no.. its 200bytes per user
00:31:27 Quit linuxstb (Read error: 110 (Connection timed out))
00:32:05 Quit aaron424 ("ChatZilla 0.9.85 [Firefox 3.5.2/20090810133513]")
00:32:28 Quit jfc (Read error: 104 (Connection reset by peer))
00:32:50 Join jfc [0] (
00:34:35 Quit Strife89 ("Dinner.")
00:35:21 Quit captainkwel ("Page closed")
00:36:33 Quit Juice^ ("- nbs-irc 2.0 - -")
00:36:34kugelJdGordon|: depends how you look at it
00:36:40 Quit HellDragon (Client Quit)
00:36:47kugelanyway, the bar doesn't show up with my patch :(
00:37:05kugeland the buttonmapping of the sim is totally fucked up
00:37:05JdGordon|no... it will always be wasting X bytes on each of them
00:37:15JdGordon|which patch? customui?
00:37:17 Quit BlakeJohnson861 (Read error: 104 (Connection reset by peer))
00:37:28JdGordon|oh well :D commit anyway....
00:37:31JdGordon|... not really...
00:37:38 Quit fred_2 (Remote closed the connection)
00:37:57 Join BlakeJohnson86 [0] (
00:39:03 Join fred_2 [0] (
00:39:27pixelmakugel: which sim?
00:39:38 Quit bertrik ("De groeten")
00:40:19 Join funman [0] (n=fun@rockbox/developer/funman)
00:41:20pixelma"up/down" and "left/right" are the other way round on the player, related to the +/- and Play/Stop on the buttons
00:42:09pixelmamaybe compare to the manual ;)
00:43:22kugelthat's stupid
00:43:34kugelis the player sim supposed to play music? it doesn't here :(
00:43:35pixelmaif you say so
00:43:46kugelthat's just inconsistent
00:43:51pixelmadoes your computer have a MAS?
00:44:05kugelno, but it's not a emulator either
00:44:33JdGordon|and no.. audio doesnt work in hwsim
00:44:39pixelmafeel free to implement music playback that uses the hwcodec engine in the sim
00:44:52kugelI'd just use the sw playback engine
00:45:06JdGordon|thats as pointless and not having audio
00:45:08pixelmaand play ogg
00:45:18kugellimited to mp3 and without dsp of course....
00:45:35 Join lasser [0] (
00:48:41 Part lasser
00:49:06*funman slaps FlynDice
00:50:07FlynDicec'mom that at least deserves a wild backhand....
00:50:44funmanat least we learned that setting wide bus mode on the card has no effect
00:50:55funmanand there is now an opened task for further work ^^
00:51:40 Quit n1s ("Lämnar")
00:58:51 Quit bluebrother ("leaving")
00:59:21amiconnAudio isn't implemented in hwcodec sims so far because it was never needed
00:59:48amiconnThere once was a patch that added it, using libmad, but it had quite special requirements for building
01:01:02amiconnNow that we have libmad in our own source tree, it could probably be done using that
01:02:39 Quit Thundercloud (Read error: 60 (Operation timed out))
01:02:56kugeli think it's still not needed :)
01:04:01pixelmathe only thing that really annoys me there is the fake next track info
01:11:56 Join dfkt_ [0] (
01:13:42***Saving seen data "./dancer.seen"
01:16:03 Quit dfkt (Read error: 60 (Operation timed out))
01:23:43 Quit shaggy-h (Read error: 110 (Connection timed out))
01:25:18 Join robin0800 [0] (
01:27:30 Join StealthyXIIGer [0] (
01:27:51 Quit robin0800 (Client Quit)
01:33:38andrewRBhey, is there any way to change the state of wps_state from apps/gui/list.c?
01:35:47JdGordon|why would you want to?
01:35:55JdGordon|I mean... you're doing it wrong if you want to...
01:36:28 Quit StealthyXIIGer (Read error: 60 (Operation timed out))
01:36:49andrewRBi'm doing it wrong. is there any way i can show you the patch i am currently working on? I thought I was pretty close
01:37:36JdGordon|pastebin... but I cant help you out for another 40min or so (I'm at work)
01:37:40JdGordon|what do you want to do?
01:37:47 Join stephen__ [0] (
01:39:04andrewRBsure, well i certainly don't want to waste anyone's time, it is a personal whim anyway. just grateful for any other pointers. Amongst other things, I've added "ACTION_LIST_PLAY" and "ACTION_LIST_SKIPNEXT", i'm trying to figure how to get them to do the expectedx
01:40:16 Part toffe82
01:40:32kugelandrewRB: show me your diff
01:40:39kugelI'm not at work :>
01:40:51JdGordon|play shuold just call audio_play() or audio_resume()... skip next probably needs to call skip_<something>() in wps.c
01:41:19JdGordon|maniupalting wps_state wont help with either of them at all
01:42:24andrewRB(just to clarify, the pastebin link is to a complete patch designed to add functionality to the hardware buttons on the D2 while the hold switch is on)
01:42:26kugelandrewRB: use audio_status() & AUDIO_PAUSED instead of wps_state.paused
01:43:34kugelI'm not sure if calling wps functions outside of the wps is a good idea, but it will probably work if you entered it once
01:43:38andrewRBwhat about change_dir()?
01:44:52 Quit gevaerts (Nick collision from services.)
01:45:04 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
01:50:40 Quit Zarggg ()
01:52:40 Join rudestone [0] (
02:01:04 Join Zarggg [0] (
02:01:36 Quit Rondom (Nick collision from services.)
02:01:47 Join Rondom [0] (
02:05:18 Quit JdGordon| ("Miranda IM! Smaller, Faster, Easier.")
02:09:35 Quit mt (Read error: 104 (Connection reset by peer))
02:15:07 Join JdGordon_ [0] (i=441dc4e2@gateway/web/freenode/x-441957fac003444b)
02:15:08 Join HellDragon [0] (
02:15:20 Join hd [0] (
02:15:23 Quit hd (Read error: 54 (Connection reset by peer))
02:15:35JdGordon_kugel: an you pastebin the viewport.c file with the patch?
02:15:58JdGordon_I can look at the diff now, but i'd rather look at it post patch and dont have the rb source on this comp
02:16:14 Quit GeekShadow ("The cake is a lie !")
02:17:23kugelwrong highlightening :/
02:19:50JdGordon_eek at the whole current statusbar handling mess
02:20:12*JdGordon_ hangs his head in shame
02:21:10kugelI changed it a bit
02:23:39JdGordon_you should add some comments around ARG_STRING macro
02:23:45 Quit funman ("free(random());")
02:24:29JdGordon_it would be nice
02:24:45kugelthat's in SVN though..unrelated
02:24:51JdGordon_oh ok...
02:24:59JdGordon_it should still get a comment
02:25:08JdGordon_why its different... why it works...
02:26:16 Quit fyre^OS (Read error: 104 (Connection reset by peer))
02:28:01*JdGordon_ thinks we should set the customui vp by default for the clip so the stupid black lines dont get in the way
02:30:58kugelJdGordon_: the black line isn't a problem on the clip with the default theme, is it?
02:31:45JdGordon_umm.. cant remember ?:p
02:32:16kugelthe black line is exactly between the list title and the content, very nice acutally
02:33:05JdGordon_unless you change fonts... or disable the statusbar :)
02:33:52 Join Jay [0] (n=63ee95d1@gateway/web/cgi-irc/
02:34:02Jayhey guys.. is anyone here?
02:34:16Jayi got some questions about tockbox i would love answered
02:34:20 Nick Jay is now known as Guest81124 (n=63ee95d1@gateway/web/cgi-irc/
02:35:19Guest81124well anyways i hope someone will answer this question eventually.
02:35:34kugelJdGordon_: of course, we were talking about the defaults though
02:35:46AB3JUGuest81124: actually asking your question would probably be a good idea
02:36:14kugeldefault custom ui won't help you against a bigger font either, but against toggling the statusbar
02:36:45 Quit efyx (Remote closed the connection)
02:36:58Guest81124when I listen to a track and want to change the track on my ipod, I would think it would be intuative to just go to the song i want from the Database and click the SELECT button, and it would play immidiatly.. is there any way to achieve this?
02:37:20kugelJdGordon_: although we could use sysfont for the ui viewport, making it immune against font changes :)
02:37:27JdGordon_exactly the way you would expect to
02:37:42JdGordon_kugel: that would be a bit silly now :)
02:37:59andrewRBthanks for the help earlier guys, i've got everything working how I want now :)
02:38:09Guest81124i hope im clear about my question.. instead of it playing the song it says "Erase dynamic Playlist"
02:38:51JdGordon_so turn that setting off
02:39:04JdGordon_or press select and it will magically go away
02:39:05JdGordon_the manual is a great read...
02:40:08Guest81124i understand it will go away.. but i don't want to dynamically erase the playlist everytime i want to hear a different song
02:40:52Guest81124it does take quite a while to do that everytime
02:41:14Guest81124Let me know if im making sense or just talking crazy talk
02:41:30JdGordon_thats what going into the database does... you need to find the song in the playlist viewer otherwise
02:42:46Guest81124JdGordon.. yes that worked.. don't know why i dind't get that earlier
02:45:35Guest81124I'm not sure if your a developer, but just a suggestion according to my experience, the music icon should be beside playlist, and it should open up the playlist immidiatly instead of showing you ""create/ view/ save" subcategories
02:45:45Guest81124sorry my spelling not so good
02:46:03 Quit JdGordon_ (Ping timeout: 180 seconds)
02:47:54Guest81124also why isn't there an option to start directly at the playlist when you turn on rockbox?
02:49:27Guest81124i would assume ill be spending most of my time in the playlist, therefor it should be more easily accessible. if others are using rockbox in defferent fassion please share cause i love the software just don't understand why some things where done the way they were
02:50:12andrewRBthe way I use it, i never use the database or playlists, i simply choose an album and play it
02:50:28kugelGuest81124: there's such an option.
02:50:45Guest81124choose an album from where in the menu?
02:50:52andrewRBthe file browser
02:51:45Guest81124the only thing i see wrong with that is "sub directories" and maybe "lack of TAG's"
02:52:08andrewRBwell both of those are simply matters of organising your music collection
02:52:17Guest81124that is true:)
02:52:32kugelGuest66974: rockbox saves the playlist where you left off. so, hitting play or setting the start screen to wps just plays the last playlist
02:53:49Guest81124i don't believe i have a wps start screen option
02:54:18Guest81124i have a Resume playback im assuming is the same thing
02:55:47andrewRBJdGordon: well, i've got ACTION_LIST_PLAY, ACTION_LIST_SKIPPREV, ACTION_LIST_SKIPNEXT, and ACTION_LIST_NEXTDIR working, any suggestions about seeking? The wps function for seeking is pretty complicated =/
02:56:15kugelGuest81124: settings->general settings->system->start screen
02:56:20JdGordonseeking is not something that imo belongs outside of the wps
02:56:32andrewRByeah that is probably true
02:57:46kugelespecially not with the progressbar slider feature
02:57:56andrewRBsee, one of the reasons I am doing this is so the player can be used with a broken touch screen
02:59:59Guest81124my "playlist" seems to run slower then my "Database" in terms of browsing. also the database seems to be the desired way fro me to choose songs.
03:00:17andrewRBwhich iPod is this, Guest81124?
03:00:23kugelwhat's "my playlist"
03:00:43kugelthe playlist viewer is indeed a lot slower, not sure about the playlist browser
03:00:44Guest81124playlist in the menu screen
03:00:56kugelI don't mean that
03:03:26JdGordonyes, thats expected
03:03:49Guest81124About the suggestion earlier about using "Files" to choose your song. it still asks to erase Dynamic memory which is a long process if done many times throught the day
03:04:50JdGordonthe *only* way you can change songs without erasing the playlist is to use the playlist viewer, or add the song to the playlist...
03:05:00Guest81124so the only way to play a song immidiatly withought "erasing dynamic playlist" (not memore) is through "playlist"
03:05:28Guest81124lol ok were on the same track JdGordon
03:08:05 Join CaptainKwel [0] (
03:08:40andrewRBi have never seen any message about dynamic playlists
03:08:46kugelGuest81124: what target are you on?
03:09:05Guest81124im not familiar with targets
03:09:11andrewRBwhich iPod
03:09:13kugelyour dap
03:09:55kugelGuest81124: press select+down to view the current playlist
03:11:25kugelanyway, this warning is clearly not made for people using the database, I'd just turn it off if I were you
03:12:58Guest81124Well in that case Ill just put it out there that i think it would be awesome if i could select songs through the "Database" by hitting "select" on the song. Even if it addedit as a "Insert Next" but play instantly is more desirable. I say this cause (a)database runs quicker then playlist, and (B)Data base use "TAGS" and (C) it has the "music icon" :)
03:13:06Guest81124thanks kugel i definitly will
03:13:23Guest81124and thank you for the excelent support
03:13:46***Saving seen data "./dancer.seen"
03:13:48kugelGuest66974: the database *does play* the song if you hit select on it
03:14:07kugelGuest81124: ^
03:15:07 Join z35 [0] (
03:15:28kugelI still don't know what you mean with playlist, you actually mean the file browser?
03:15:46Guest81124but it needs to "Erase Dynamic Playlist" which takes lon.. no?
03:16:18Guest81124"PLaylists" located in the root menu
03:16:24kugelgo to the settings and turn it off, so that it never comes up again
03:16:53Guest81124yes but i have to wait for it to make a new playlist everytime i select a song
03:17:09kugelsure it plays the whole "folder"
03:17:57kugelalso, turn dircache on to speed up any sort of browsing
03:18:54kugelsettings->general settings->system->disk
03:19:16Guest81124will do
03:20:03Dhraakelliancould I get someone to look at a one-liner patch for Sudoku on the Fuze? I'm getting tired of doing my own builds just so that I can have a non-combo scratchpad button.
03:22:40Guest81124kugel: turning on the DirCache definitly made me like "Playlists" a lot better.
03:27:11kugelI think I never browsed by playlists
03:27:31kugelviewing the catalog shows me "no playlists" ;p
03:28:40Guest81124I really hate to bring it up. but it still baffels me why choosing songs on Database doesn't just switch to the song in the current playlist which would remove the need to constanttly "remove dynamic playlist"
03:28:47*Dhraakellian never really saw the point of premade playlists for general album listening when adding entire albums/folders to the current/dynamic playlist is so easy
03:29:36Guest81124baffels me even more now that you say that you never browse by playlist
03:29:47andrewRBi've never used playlists either
03:29:57andrewRBi really don't understand the reason to use them?
03:30:31Guest81124so every time you choose a new song to listen to you have to make a new dynamic playlist?
03:30:44Guest81124man i must be missing out on something
03:30:47andrewRBwhen i want to listen to something i simply select it, and it plays
03:30:55kugelGuest81124: every queue by the database is a dynamic playlist
03:31:00andrewRB(by select it, i mean, in the file browser)
03:31:25kugelyou always gonna destroy a dynamic one if you choose another song/playlist and the previous comes from the database
03:32:44Guest81124andrew i don't know why but before it would create a new playlist if i choose a song from file
03:32:54Guest81124but now it seems to work fine
03:33:03 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
03:33:07Guest81124and i can see why you would say that you don't know why to use playlists
03:34:13andrewRBthe way i see it is the "dynamic playlist" is just the songs which are currently playing. So when you select a file in the browser, the "dynamic playlist" changes to the contents of whatever folder you selected from.
03:34:41andrewRBi don't really see how else you would expect it to work
03:34:56Guest81124can i shuffle the songs if i choose them through file browser?
03:35:03Guest81124i thought i had shuffle turned on
03:35:10andrewRByes, the songs in the folder you are playing from
03:35:23DhraakellianRockbox's use of "dynamic playlist" confused me a bit, since I'm more used to Amarok's use of the term (which refers to auto-generated "party mode" playlists)
03:36:00ej0rgeMight be nice if you could 'play' a whole directory recursively
03:36:08ej0rgemaybe an option in the context menu
03:36:19andrewRByou can, what do you mean?
03:36:25ej0rgeoh, I see that now
03:36:41Dhraakellian <−−changes the sudoku scratchpad button on the Fuze from select+up to just select. Opinions?
03:36:44ej0rgewhat do i look like, a guy who reads manuals or something?
03:37:01Guest81124good one andrew
03:38:11*Dhraakellian has been using the one-liner patch for a while now and hasn't had a problem with it (well, other than occasional relapses to e200-trained muscle memory trying to get to the menu and toggling a number on the scratchpad instead)
03:38:23andrewRBany D2 owners online?
03:38:32Guest81124so wheres the option so i can just shuffle through all the songs in all the directories?
03:38:44Guest81124through file browser
03:39:06Guest81124well sheffel through all the songs in the sub-directories
03:39:40andrewRBsettings->playback settings->Auto-Change Directory
03:40:04andrewRB(presumably you would pick random)
03:40:47*andrewRB hopes nobody gets a hernia
03:41:14Guest81124i think i am '
03:41:28ej0rgeandrewRB: just wait until he discovers "follow playlist"
03:44:02andrewRBJdGordon: I'm a little confused, I seem to have got seeking to work in the list context, but i don't understand *why* it works. care to take a look?
03:45:04andrewRBspecifically, as ffwd_rew() is in wps.c, why can i call it from list.c without complications?
03:46:16*kugel slaps JdGordon
03:47:30kugelDhraakellian: I'll have a look (maybe, can't promise anything)
03:47:43*Guest81124 hello
03:48:40 Quit rudestone (Read error: 110 (Connection timed out))
03:49:21*andrewRB wonders if people should be forced to quote random excerpts from the manual for their target before being allowed to download rockbox
03:49:24kugelJdGordon: we have two statusbar callbacks :(
03:59:49kugelJdGordon: one in statusbar.c (gui_statusbar_changed(), directly passed to CHOICE_SETTING in settings_list.c), and one in display_menu.c (statusbar_callback()). Can you tell the difference? It seems removing the latter doesn't change anything
04:00:42JdGordonthe one in display_menu is for the setting change.. doesnt that one trigger the other one?
04:01:16kugelthe other one is alawys triggered (in quickscreen also); even twice it seems
04:01:39JdGordonandrewRB: you need to add a ffwd_rew(ACTION_WPS_SEEKSTOP); call when you want to stop seeking
04:01:44kugelI'm not sure what ACTION_EXIT_MENUITEM stands for though
04:02:01JdGordonthats what happens when the setting is exited
04:02:48JdGordonyes, it triggers the event then
04:02:49andrewRBJdGordon: do you mean ffwd_rew(ACTION_WPS_STOPSEEK); ?
04:03:00JdGordonit might be triggered elsewhere also
04:03:03JdGordonandrewRB: yes
04:03:05kugelthe other one is also called when it exited
04:03:53andrewRBJdGordon: its already there. it is tested and works, so not entirely sure what you were getting at?
04:04:14JdGordonoh ok
04:04:54JdGordonkugel: oh right.... yes the one in display_menu should be removed
04:05:34kugelany idea why the other one is called twice when changing the value (not when exiting, also not in quickscreen)?
04:05:55JdGordonyes, the list calls it each time the selection is changed
04:06:22kugelbut why is it called twice per selection change?
04:06:52JdGordonoh right... umm... thats what gdb is for :)
04:07:38kugelI tried that, same backtrace apparently
04:08:02*JdGordon is off...
04:08:07JdGordoni'm sure you'll figure it out :D
04:08:19andrewRBgoodbye Jd, thanks for the help :)
04:14:38 Join rudestone [0] (
04:27:18Guest81124hey guys. so i've been playing around with rockbox and i can't seem to get my rockbox to play random songs from all the sub directories when i chose the files from the "files" browser. I've turned on Recursively insert directories and also had a playlist with all the songs i want to shuffle through. Also i have follow playlist turned on.. whenever i click on a song from the "files" browser it would create a n
04:28:19Guest81124therefor not let it shuffle through songs in the sub directories
04:31:46andrewRBin the file browser, on the root directory, open the context menu and select playlist->insert
04:31:54andrewRBthen just have shuffle on
04:34:54Guest81124alright.. so now when i wanna choose a different song to play from the file browser i get the "Erase dynamic playlist?" option. as apposed to just changing to the song in the current playlist
04:35:32andrewRBso why are you choosing a different song from the file browser rather than the current playlist?
04:37:15andrewRBfrom the WPS, open the context menu, select playlist, select view current playlist, and select the song you want there
04:37:40Guest81124it's difficult to find a specific song in the current playlist. in playlist?
04:38:45Guest81124in the "Database" you can find songs by many different criterias like artist. in "Files" you can search by folder
04:38:55Guest81124in playlist it's just raw song after song
04:39:23andrewRBi really don't understand what you want
04:39:36andrewRBi mean, if i want to play everything randomly, that is what i do
04:39:48andrewRBif i want to play a specific song/album, i select it
04:39:54Guest81124i know and i don't want to become to much of a bother
04:40:09andrewRBno it's fine, i just can't comprehend what you are trying to do
04:42:03Guest81124when i used my ipod firmware it would play all the songs in a shuffel. if in the middle of a song i decided i wanted to play a song from the artist "TOOL" i could just search in artists. similar to "Database" and it would play instantaneously
04:42:30Guest81124while maintaining the shuffle of all the songs in my ipod
04:44:38Guest81124i really hope im making some sense
04:44:47andrewRBi think so
04:44:52andrewRBgimme a sec to test something
04:45:08andrewRB(as i don't use shuffle, i have to check)
04:50:35andrewRBcan't test it, actually, sorry. i can't find my sd card
04:50:53Guest81124thats poo poo:(
04:51:19andrewRBbut i'm pretty sure you can't do what you want unless you search within the currently playing playlist
04:51:44Guest81124ya im pretty sure too..
04:51:51andrewRBselecting anything from the file browser or database creates a new dynamic playlist, because that is usually what is expected
04:52:02andrewRBwhen you select a song, you expect the song after it to play
04:52:09Guest81124i would expect my way actualy
04:52:36Guest81124wait no you might be right
04:53:43andrewRBi think it is a difference between iPods and the majority of other music players - iPods automatically play through the entire collection, whereas most others play whatever folder or album you select
04:54:21andrewRBi mean, rockbox can do that too, auto advance to the next folder when it gets to the end of one
04:54:51andrewRBbut the dynamic playlist doesn't contain everything you have on the device
04:55:07andrewRBand shuffle only shuffles within the DPL
04:55:31Guest81124well theres one way of doign it but its a little lengthy
04:56:30Guest81124going to the context menu and clicking insert next
04:56:47 Quit rudestone ()
04:56:50andrewRBwell yeah
04:56:52Guest81124which i guess is the solution ill have to bear with
04:56:57andrewRBif you have everything in the DPL
04:57:08 Join rudestone [0] (
04:57:11 Quit rudestone (Remote closed the connection)
04:57:18andrewRBand want a specific song to play, but not erase the playlist
04:57:21andrewRBfind that song
04:57:26andrewRBopen context menu
04:57:48andrewRBplaylist -> insert
04:57:52andrewRBis that right?
04:59:02andrewRBor playlist-> play next
04:59:04andrewRBthen skip
05:00:29 Quit BHSPitMonkey (Remote closed the connection)
05:01:03andrewRBkugel: is there any way to map an action to pressing two keys simultaneously?
05:01:17Guest81124last question..
05:01:27Guest81124how do i rate my song?
05:01:37Guest81124i guess ill look in the manual for that one :)
05:02:18Guest81124under Database you have a user rating option
05:02:44andrewRBbeen using rockbox for four years, only just noticed that
05:05:09andrewRBcan you tell me what, if any, actions require you to press more than one key at once please? (i don't own an iPod)
05:06:17Guest81124except for debugging nothing i know of
05:06:56Guest81124maybe not debugging exactly but if you brick your ipod or something
05:07:34andrewRByeah, i meant within rockbox though
05:10:16Guest81124sometimes my rockbox doesn't turn on . unless i press a compination of select and another one
05:10:23Guest81124and to quit some games
05:10:50andrewRBcould you clarify that please? which plugin, and what do you press?
05:11:15andrewRB(i'm trying to find what i want in the keymap file and the manual but failing)
05:12:11Guest81124for instance "snake" under "games" need to press select and menu to quit
05:12:22andrewRBlemme check that out
05:13:47***Saving seen data "./dancer.seen"
05:13:55Guest81124i see yourt coding
05:14:08Guest81124and or i believe
05:15:11Guest81124thats code you tyed out 'BUTTON_SELECT|BUTTON_MENU'
05:16:01Guest81124ive coded before and it brings back good memories.
05:17:19Guest81124if your bored or have extra time you could check out some of my art.. .. i did all the art and player coding for the sidescroller that flashes by
05:18:01 Join Igoooorrrr [0] (
05:20:07IgoooorrrrHello, I have rockbox running on an iriver H340. How would I go about removing the original firmware entirely so that the player always boots into rockbox? I have it connected in my car, so every time I turn the engine on, it starts the original firmware charging screen.
05:20:52andrewRBmodify the bootloader
05:21:32IgoooorrrrCan you be a bit more specific? Modify it to remove the original firmware entirely?
05:22:15andrewRBwhat do you currently do to boot rockbox?
05:22:22Guest81124i gtg . thank you sincerly andrewRB.. You absolutly helped me a lot! best wishes to you.
05:22:29 Quit Guest81124 ("CGI:IRC")
05:23:01IgoooorrrrIt boots to rockbox by default when I hold down the play button, but I can still boot to the original firmware if I hold down play and circle.
05:23:33andrewRBwhat if you hold no key?
05:23:39andrewRBoh, wait, ignore that
05:23:50andrewRBthe play key is the power button, i take it
05:24:22IgoooorrrrHowever, if I plug it in, it automatically launches the original firmware charging screen, and I can't boot to rockbox from that, only to the original firmware.
05:24:49andrewRBhmmm, how do you install rockbox on it? I mean, does it require the OF upgrade mechanism?
05:26:41IgoooorrrrI'm not too familliar with the terminology you guys use, but if I remember correctly, I modified the original firmware with the utility off the RB site.
05:27:26andrewRBperhaps, patched the iriver firmware with the rockbox bootloader?
05:27:34Igoooorrrryes, that was it
05:27:44andrewRBif that is the case, i would hazard a guess there is no way to remove the iriver firmware entirely
05:29:37andrewRBof course, i've never owned or worked on the iriver, so i am just guessing, but that sounds pretty similar to my D2
05:30:20 Join evilwombat [0] (
05:31:30 Join StealthyXIIGer [0] (
05:37:08 Quit andrewRB ("goodnight all, thanks for the help")
05:43:32evilwombatdoes rockbox even have a charging screen on the h300 /
05:48:35IgoooorrrrNot really. If you try to shut it down while plugged in, it says "Battery: Charging" and doesn't shut down.
05:51:59 Quit kugel (Remote closed the connection)
05:54:41evilwombatIt's the original FW being stupid?
05:54:49evilwombatI figure enough other people must have had this issue
05:54:56Igoooorrrrit's not really an "issue"
05:55:03evilwombatanyone else here run H100 / H300 ?
05:55:12Igoooorrrrit comes up in my case because of the specifics of how i use the iriver
05:57:49evilwombatwell, i am sure other people at least have run into it
05:57:56evilwombatit's a pretty BASIC issue that needs to be fixed
05:58:00evilwombatreplace the charging screen
05:58:14evilwombatand go into the proper firmware when charger is removed
05:58:46Igoooorrrri would think the fastest fix is when the charger is inserted, boot to rockbox
06:01:00evilwombatbut you don't know who is responsible for that
06:01:12evilwombatwhich 'code' starts when charger is plugged in?
06:01:19evilwombatprobably original iriver bootstrap thingy
06:01:34evilwombatthe one that is patched to look at the Rec / hold button and decide what to load
06:01:50evilwombatwell, the charger aspect of it doesn't seem to be patched by fwpatch
06:02:09Igoooorrrri imagine it is not trivial to mess with the original firmware
06:02:30evilwombati suppose not, but they've done it
06:02:39evilwombati imagine this is a pretty weird boot situation
06:02:45Igoooorrrryeah, but i doubt they'd want to mess with it anymore
06:02:46evilwombatie, no power -> power -> what runs?
06:02:51evilwombatit's all just a guess, but prolly a decent one
06:03:07evilwombatsubmit a feature request?
06:04:06evilwombatwhat happens if you use a USB charger instead of DC In ?
06:04:27Igoooorrrrif i recall, it won't play and run at the same time
06:04:32Igoooorrrrplay and charge
06:04:36Igoooorrrrvia usb
06:04:48evilwombatoh... lame?
06:04:48Igoooorrrri remember having to wait for the car charger to arrive, and i believe that was the reason
06:05:16evilwombatif you can get your hands on the on button, i know of a way\
06:05:33evilwombatdoes attaching USB power wake it up and send it into rockbox?
06:06:46Igoooorrrr"if you can get your hands on the on button"?
06:06:53Igoooorrrrlemme find out
06:06:57Igoooorrrrbut i dont think so
06:07:10evilwombatfind out.
06:07:23Igoooorrrrit goes into "bootloader usb mode"
06:07:35evilwombatie, ROCKBOX
06:07:47evilwombatis it usb power, or usb -> PC ?
06:08:12Igoooorrrrit connects it to my pc
06:08:26evilwombatwhat if you just give it 5V via USB ?
06:08:38evilwombatie, razr charger, or any of the other ones?
06:08:54evilwombatif you give it both, it may goto rockbox
06:09:11Igoooorrrri only have one power jack in my car for it :)
06:09:18evilwombatthat's not a problem :)
06:09:23evilwombatfind out of it will work
06:09:32evilwombatie, get a plain USB 5V output
06:09:34Igoooorrrrso you want usb power
06:09:38Igoooorrrrgive me one second
06:09:41evilwombatfind one
06:09:52evilwombator reach to computer, and unplug USB green and white, leaving only the power
06:10:25evilwombatUSB needs little current - you essentially just want to wake the fucker up.
06:10:38evilwombatso you can split off the car charger and get it down to 5V easily
06:10:43evilwombatie, one part
06:11:01evilwombatjust verify that only having the power is enough
06:12:05evilwombatif it does, i will explain the rest
06:12:06Igoooorrrrstill goes into bootloader usb mode
06:12:21evilwombatso with that plugged in, attach the regular charger.
06:12:40evilwombatthen, unplug usb power.
06:13:07Igoooorrrrthat does make it boot into rockbox, yes
06:13:22Igoooorrrrthis... seems needlessly complex
06:13:25evilwombatI am thinking of a lulzy circuit.
06:13:31evilwombathop back on aim - no need this on irc
06:17:21 Quit StealthyXIIGer (Connection timed out)
06:20:50CIA-6New commit by kkurbjun (r22287): M:Robe 500: Put more interrupts in iram, reduce stack usage when ARM/DSP transfers happen.
06:21:39CIA-6New commit by kkurbjun (r22288): M:Robe 500: Forgot one.
06:27:20evilwombatyeah, patched iriver fw to boot rockbox when DC is attached would definitely be nicer than having to make a PIC your timing bitch
06:31:27 Quit stephen__ (Read error: 110 (Connection timed out))
06:38:06 Quit DarkSpectrum (Read error: 54 (Connection reset by peer))
06:38:11 Join DarkSpectrum [0] (
06:38:43 Quit ehntoo (Read error: 110 (Connection timed out))
06:55:19kkurbjunrasher: is there a set timeframe that the simulator builds you have on your website are updated?
07:03:39 Join Horschti [0] (n=Horscht2@xbmc/user/horscht)
07:13:48***Saving seen data "./dancer.seen"
07:22:10 Quit Horscht (Read error: 110 (Connection timed out))
07:24:26 Quit CaptainKwel (Remote closed the connection)
07:24:47 Join shaggy-h [0] (
07:28:34 Quit Igoooorrrr ()
07:30:44 Join stoffel [0] (
07:33:38 Quit readabil1ty (Read error: 110 (Connection timed out))
07:38:22 Quit evilwombat ("^D")
07:42:25 Quit martian67 (Read error: 104 (Connection reset by peer))
07:42:50 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
08:07:20 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
08:16:17 Join Rob2223 [0] (
08:22:49 Join ender` [0] (
08:23:41 Part safetydan ("Leaving.")
08:33:58 Quit Rob2222 (Read error: 110 (Connection timed out))
08:34:08 Join flydutch [0] (
08:43:51 Join petur [50] (n=petur@rockbox/developer/petur)
08:46:18 Join jfc^3 [0] (
08:47:10 Join KBH [0] (
08:48:41 Quit jfc (Read error: 60 (Operation timed out))
08:48:43 Join Thundercloud [0] (
08:51:54 Quit Tristan (Nick collision from services.)
08:53:18 Join Trista247 [0] (
08:53:48 Quit HBK- (Read error: 60 (Operation timed out))
08:55:16 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
09:06:43 Join GodEater_ [0] (n=godeater@
09:06:49 Quit FlynDice (Read error: 104 (Connection reset by peer))
09:09:49 Join FlynDice [0] (
09:12:10 Join GodEater__ [0] (n=godeater@
09:13:53***Saving seen data "./dancer.seen"
09:18:50 Quit GodEater_ (Remote closed the connection)
09:20:48 Quit GodEater__ (Remote closed the connection)
09:40:35 Join n1s [0] (n=n1s@rockbox/developer/n1s)
09:47:46 Quit Thundercloud (Remote closed the connection)
09:55:51 Quit bah_ (Read error: 110 (Connection timed out))
09:55:56 Join bah_ [0] (
09:57:29Unhelpfuli think i've got most of the names attributed in lcd-bitmap-common.c, i'm not sure how detailed i ought to be about who did what... though i suppose svn blame can always be used on the original files if it comes down to needing to know.
09:59:51 Join bertrik [0] (
10:02:34CIA-6New commit by unhelpful (r22289): LCD bitmap driver code consolidation from FS #4817: ...
10:04:12 Join pamaury [0] (
10:06:13n1sI was cheering for reduction in code duplication :)
10:06:38Unhelpfulah. indeed. i touched lcd.h though, probably why the build is running long
10:07:20Unhelpfuli think that i can rename it to lcd-common.c and also roll in the scroll_fn from the charcell driver... maybe. hopefully without making too much #ifdef hell
10:08:16Unhelpfulhm, i didn't get any positive deltas on targets i test-built for :/
10:09:41 Join efyx [0] (
10:12:06bertrikUnhelpful, only iaudio m3 is positive and not a lot
10:14:32 Quit ender` (
10:14:32 Quit bmbl (
10:14:32 Quit Horschti (
10:14:32 Quit gevaerts (
10:14:32 Quit dfkt_ (
10:14:32 Quit Kohlrabi (
10:14:32 Quit ej0rge (
10:14:32 Quit AlexP (
10:14:32 Quit sharp (
10:14:32 Quit kkurbjun (
10:14:32 Quit ch4os (
10:14:32 Quit scorche|sh (
10:16:14NJoinender` [0] (
10:16:14NJoinbmbl [0] (n=Miranda@unaffiliated/bmbl)
10:16:14NJoinHorschti [0] (n=Horscht2@xbmc/user/horscht)
10:16:14NJoingevaerts [0] (n=fg@rockbox/developer/gevaerts)
10:16:14NJoindfkt_ [0] (
10:16:14NJoinkkurbjun [0] (n=kkurbjun@rockbox/developer/kkurbjun)
10:16:14NJoinKohlrabi [0] (
10:16:14NJoinsharp [0] (
10:16:14NJoinej0rge [0] (
10:16:14NJoinAlexP [0] (n=alex@rockbox/staff/AlexP)
10:16:14NJoinch4os [0] (n=ch4os@gentoo/user/ch4os)
10:16:14NJoinscorche|sh [50] (n=scorche@rockbox/administrator/scorche)
10:24:20Unhelpfuleven more odd, bloat-o-meter shows 0 difference between the two revisions on M3
10:35:31 Quit kachna (Read error: 113 (No route to host))
10:47:28 Quit ender` (Read error: 110 (Connection timed out))
10:51:20 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
11:13:58***Saving seen data "./dancer.seen"
11:21:24 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
11:37:50 Join jernejovc [0] (n=jernejov@
11:50:23 Quit yawnage ("leaving")
11:56:52 Join yawnage [0] (
12:03:08 Quit Rob2223 ()
12:04:11 Join Rob2222 [0] (
12:10:00 Quit Unhelpful (Remote closed the connection)
12:10:09 Join Unhelpful [0] (
12:10:43 Quit Unhelpful (Remote closed the connection)
12:12:05 Join Unhelpful [0] (
12:15:43 Join safetydan [0] (n=deverton@rockbox/developer/safetydan)
12:22:59 Join BdN3504 [0] (n=5ce2260d@gateway/web/cgi-irc/
12:24:57BdN3504quick question: how do i checkout 3.3 using the svn checkout command with the -r parameter? would it be "-r 2009-19-06" ?
12:25:23gevaertsno. You check out the 3.3 branch
12:26:35BdN3504how do i do that?
12:27:53BdN3504is it 21335? as mentioned here: ?
12:28:01Unhelpfulwhy exactly do you want to? if you want to work on the source, you really should be using svn trunk... otherwise any patches you generate probably won't apply.
12:28:18UnhelpfulBdN3504: no, the branch is not a particular revision.
12:28:36gevaertssee the -dev list, one or two weeks before the release. There's an announcement mail that tells you what to do
12:30:51BdN3504Unhelpful: i want to create a test build. i made one using r22090 and one person reported an error that was not due to the patch but another issue with that revision.
12:31:15BdN3504so i thought i'd go with the release and patch that, so there's less room for errors
12:32:22BdN3504thanks for the help though
12:34:10BdN3504if anyone else is looking for the info on how to check out 3.3 you'll find it here:
12:34:27*gevaerts happens to know :)
12:44:25 Quit n1s ("Lämnar")
12:56:33 Quit BdN3504 ("CGI:IRC (EOF)")
13:01:44 Join funman [0] (
13:14:00***Saving seen data "./dancer.seen"
13:14:45 Join robin0800 [0] (
13:28:27 Quit stoffel (Read error: 113 (No route to host))
13:39:13 Join einhirn [0] (
13:45:54 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
13:48:38 Join AndyIL [0] (i=AndyI@
13:59:30 Quit AndyI (Read error: 110 (Connection timed out))
14:00:57Unhelpfulgevaerts: i see that appsversion is in plugin api, iriver_flash ought to use it
14:01:30gevaertsUnhelpful: ignore me. The line in iriver_flash that uses it is commented out
14:02:09 Join fdinel [0] (
14:31:30 Join pyro_maniac [0] (
14:35:26pyro_maniaci don't know if its a bug but my e250 seems to charge to 99% in OF and only to 94% in rockbox. is this a known thing?
14:40:29pyro_maniacfunman: ping
14:41:47pyro_maniaci tried to disassemble the the yh920 bootloader more
14:42:36pyro_maniacbut i am got stucked
14:43:23 Join LambdaCalculus37 [0] (i=44a0430d@rockbox/staff/LambdaCalculus37)
14:43:50pyro_maniaci think i have found a place where DEV_INIT1 and PLL_CONTROL gets sets but i am not sure and i don't know if i am on the right way
14:45:22pyro_maniacthis is all in the codec test
14:46:02 Join kugel [0] (n=kugel@rockbox/developer/kugel)
14:47:14funmanpyro_maniac: what aren't you sure about?
14:48:46pyro_maniacare DEV_INIT1 and PLL_CONTROL hints for a codec activation? i am no hardware hacker so i have never seen this before
14:49:04pyro_maniacand in the codec code only GPIOs are mantioned
14:50:52funmanno idea, i don't know PP
14:51:03kugelneither I do
14:51:42pyro_maniacare we so addicted to low_light :-)
14:51:52tmztPLL_CONTROL would be related the clock
14:52:37kugelpyro_maniac: trial & error is successful sometimes, did you try to just run the code you see there?
14:53:17pyro_maniacno i had enough problems to get there and understand it
14:53:42pyro_maniacbut this entire test function may be helpful to run
14:54:10pyro_maniacthere are also hdd, lcd and flash test functions in the bootloader
14:55:16 Join n1s [0] (n=n1s@rockbox/developer/n1s)
15:07:10 Join BdN3504 [0] (n=5ce2260d@gateway/web/cgi-irc/
15:08:41pyro_maniackugel, funman: could you take a look at this codec test?
15:09:11kugelin the disassembly?
15:09:18kugeli don't think so :(
15:09:53BdN3504What does this error mean, when i try to compile? "structure has no member named `strlcpy' " was strlcpy introduced only a short time ago? is it part of lua or something?
15:10:36BdN3504using vmware image debian distribution
15:11:58pyro_maniackugel: no time?
15:14:02***Saving seen data "./dancer.seen"
15:14:41kugeltoo lazy :/
15:15:09kugelI'm a novice in that area anyway
15:15:12pyro_maniackugel: my hope was to find a fix for the no sound problem on the yh920
15:16:58n1sBdN3504: it means that you try to access a struct member called strlcpy that doesn't exist; it was added to the plugin api a few weeks ago
15:17:13n1sso i recommend a svn up
15:20:57BdN3504aha ok, but i am trying to include a patch into 3.3 to upload a test build. would i have to rewrite the patch using strcpy? or can i update 3.3 to only contain that plugin api change?
15:20:58 Join stoffel [0] (
15:21:55n1s3.3 doesn't have strlcpy so you'd have to add the function too, using strncpy instead should be pretty easy
15:23:06 Join pyro_maniac1 [0] (
15:26:03 Quit antil33t (Read error: 104 (Connection reset by peer))
15:26:17 Join antil33t [0] (n=Mudkips@
15:26:17 Quit robin0800 ("Leaving")
15:27:55 Part safetydan ("Leaving.")
15:29:09 Quit funman ("free(random());")
15:34:00tmztcan someone comment if rockbox uses libgcc.a or an equivalent (such as a custom one)?
15:35:50 Join Creposucre [0] (n=86f69562@gateway/web/cgi-irc/
15:36:12BdN3504essentially i can simply replace strlcpy with strncpy, right? i know that the string that has to be copied won't be too long, so this won't cause any problems. that's the only advantage of strlcpy to strncp, right?
15:38:05pyro_maniac1BdN3504: i remember that the strlcp explicitly replaces the strncp
15:39:00pyro_maniac1there is a lib or a include missing i think
15:39:00BdN3504pyro_maniac1: so you know which patch i'm talking about... did it cause any problems?
15:39:14 Quit pyro_maniac (Read error: 110 (Connection timed out))
15:39:20 Quit linuxstb (Read error: 110 (Connection timed out))
15:39:47 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
15:39:51pyro_maniac1there was a big strcpy and strncpy replace patch for plugins i think
15:40:01Creposucren1s: (about FS 10494) i finally managed to remove everything from the generic radio code, except handling for tuner disconnect
15:40:15linuxstbtmzt: Yes, Rockbox uses libgcc.a
15:41:22BdN3504yes, i understood that by what nls has said. nvm though, it works for the patch i am concerned with.
15:41:42CreposucreI haven't posted the patch yet, i would like to know if anyone has an idea ho/where I can refresh the root menu
15:43:16n1sCreposucre: sounds good, as little target specific code as possible in the generic app code is prefered, JdGordon is probably the man to ask about the menu
15:44:14BdN3504thanks for the help guys
15:44:21Creposucreabout the code, i still have a strange behavior when i want to change the region settings
15:44:32Creposucreand mono mode
15:44:44Creposucreit gets stuck and loops in it
15:45:31pyro_maniac1kugel: the yh925 bootloader also got this test functions but i think with some different values. i thought this may be the key.
15:45:42Creposucreis this a general bug?
15:47:00Creposucrebasically, it does that everytime that i go in a sub menu in settings mode
15:54:58 Join evilnick [0] (i=0c140464@gateway/web/freenode/x-kpnlakhzgjyrmcfv)
15:55:32 Part LinusN
15:55:56 Quit Torne ("shit")
16:01:26 Quit moos ("Rockbox rules the DAP world")
16:04:52 Quit BdN3504 ("CGI:IRC (EOF)")
16:06:52 Quit Creposucre ("CGI:IRC (EOF)")
16:06:56tmztlinuxstb: ah thanks, for the kernel components or just the .rock files?
16:14:27 Join moos [0] (i=mustapha@rockbox/staff/moos)
16:19:47 Quit DarkSpectrum (Read error: 54 (Connection reset by peer))
16:19:49 Join DarkSpectrum- [0] (
16:19:53 Nick DarkSpectrum- is now known as DarkSpectrum (
16:26:59 Quit killan (Read error: 104 (Connection reset by peer))
16:27:11CIA-6New commit by kugel (r22290): Make gui_syncquickscreen_run() static as it's not used anywhere.
16:32:31 Join robin0800 [0] (
16:37:24 Join toffe82 [0] (n=chatzill@
16:39:34CIA-6New commit by kugel (r22291): Make quickscreen items enum name lower-case, as per docs/CONTRIBUTING.
16:49:01 Join funman [0] (n=fun@rockbox/developer/funman)
16:49:39funmantmzt: i believe all object files (isn't libgcc.a automatically linked by gcc?)
16:54:54 Join ender` [0] (
16:56:45 Nick dfkt_ is now known as dfkt (
16:57:14 Join einhirn [0] (
16:58:44pamaurygevearts: did you have time to test the usb driver ?
16:58:45funmando you have advices for the phone conversation i'll have with AMS in 2 hours?
17:00:07 Quit einhirn (Client Quit)
17:00:10pyro_maniac1pamaury: you should exchange e and a in the name
17:00:36pamaurygevaerts: did you have time to test the usb driver ?
17:00:40pamaurythanks :)
17:02:33bertrikfunman, be honest and polite, but make it very clear what we want
17:02:43bertrikwhat do we want exactly? the datasheet, I suppose?
17:04:03 Quit toffe82 ("ChatZilla 0.9.85 [Firefox 3.5.2/20090729225027]")
17:07:57 Quit antil33t ()
17:08:02 Quit DataGhost (Nick collision from services.)
17:08:07 Join DataGhost [0] (
17:09:15pamauryfunman: I submit a patch to fix the const correctness of usb_serial_send, can you have a look at it ? (that's really annoying to have compiler warnings about that) (FS #10519)
17:09:38gevaertspamaury: not yet. I'll try to find some time later today
17:10:31 Join jgarvey [0] (
17:10:50pamaurygevearts: ok, I really don't see what make this driver fails. Perhaps there is a fundamental difference between ums and "normal" usb
17:11:29 Join toffe82 [0] (n=chatzill@
17:14:05***Saving seen data "./dancer.seen"
17:14:14funmanbertrik: right
17:16:49bertrikit also helps if we can find some kind of common interest, but I can't think of any right now
17:17:57bertrikas far as I know, we have been able to keep the as3525 datasheet private until they released it themselves, so you can mention that to show we can be trusted (don't emphasize it too much though)
17:19:57 Join killan [0] (
17:21:54funmani already did it in the previous email
17:22:03funmanhis answer was short, i think i'll start by asking him to elaborate on it
17:22:24funman"I think the specification for our device (AS353x series) might not give you additional information about the chip you are searching for."
17:23:18funmanperhaps there is a misunderstanding, because i contacted the linux/uboot porter for as353x (ams employee), about the specific SD controller
17:26:17 Join kachna [0] (
17:44:28 Quit Xerion (" ")
17:44:30 Join Llorean [0] (
17:54:36 Join liar_ [0] (
17:58:40 Nick jfc^3 is now known as jfc (
17:59:20 Quit bertrik ("De groeten")
18:02:12 Quit funman ("free(random());")
18:02:26amiconngevaerts: Is usb serial specified for high speed? Afair full speed has a lower packet size limit
18:05:09 Join JdGordon_ [0] (i=483e197d@gateway/web/freenode/x-ftdycyumgkpahaki)
18:16:04pamauryamiconn: the usb serial does not specify a speed iirc
18:16:36 Join funman [0] (n=fun@rockbox/developer/funman)
18:19:20 Join TheSeven [0] (
18:23:34 Quit petur ("work->home")
18:25:52 Join Torne [0] (
18:28:28 Join Ubuntuxer [0] (
18:29:20 Join Lss [0] (
18:29:58kugelit can't have a top item because the buttons behave weird when the lcd is flipped (re FS #9706)
18:31:30kugelthere's also 1 or more remotes that have weird buttons (and damn small screens anyway) so that they don't have a 4th item too
18:31:59kugelthe small screen shouldn't be a problem though, clip is also fine
18:32:19 Quit JdGordon_ (Ping timeout: 180 seconds)
18:34:02 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
18:38:48 Part pyro_maniac1 ("Leaving.")
18:39:40 Quit funman ("free(random());")
18:40:07CIA-6New commit by moos (r22292): Update the french translation.
18:43:24gibbon_i could not find any documentation about the iPod accessory bus anywhere... and i don't really know if this term really refers to the thing i am looking for
18:44:25gibbon_i am trying to find out, if equipment, connected to the "additional 4pin port" near the headphones might work with rockbox ... more specifically the remote that sometimes came with the iPod 3G
18:44:37TorneIt might
18:44:47CIA-6New commit by moos (r22293): Add Laurent Papier to credits for FS #10514 reworked and committed in r22255.
18:44:48 Join mt [0] (n=MTee@rockbox/developer/mt)
18:45:00 Join JdGordon| [0] (n=Miranda@nat/microsoft/x-hkuayydqxspwukjq)
18:45:00Tornenot sure if we have IAP support for the 3G serial port
18:45:09Tornewe do for some accessories on the dock serial port on later models.
18:45:53TorneAh, that look slike a no.
18:45:59TorneThere's no serial driver for 3G afaict
18:46:01 Join jyu [0] (i=2669ecc2@gateway/web/freenode/x-uoyughtqgqcidihw)
18:46:22 Nick jyu is now known as CaptainKwel (i=2669ecc2@gateway/web/freenode/x-uoyughtqgqcidihw)
18:46:45 Part Llorean
18:47:31gibbon_Torne: i want to use this remote with the mini 1st gen
18:47:40gibbon_the 3rg gen is long gone ... :/
18:48:24Tornei dunno what the mini has
18:48:41gibbon_the device looks a lot like this, but does not connect to the dock, but the headphones
18:48:52Tornebut ipod mini 1g in rockbox has no accessory protocol support
18:48:56Torneso i would guess it doesn't work
18:49:14gibbon_i don'T even know which protocol this extra port talks and if its the same as the IAP in the dock
18:49:23gibbon_oh... sad though.
18:49:26gibbon_thanks anyway
18:49:33 Join Xerion [0] (
18:49:38TorneI doubt we know that either :)
18:49:50TorneThe only accessory support we have on ipod is for stuff hat plugs into the dock connector
18:50:04TorneI have no idea if htat's the same protocol as the extra port on the mini/3g
18:50:06Torneit might be, might not
18:50:11Tornebut in any case nobody seems to have tried
18:50:36gibbon_i would volunteer, if i were a better programmer
18:50:46gibbon_i have the device but lack a lot of the skills ;)
18:51:15gibbon_at least, there is some "accessory power" setting on the mini.. but i guess that does not say anything at all
18:51:41TorneNo idea what pins that will be controlling power to
18:51:45Torneit may be the same pins on the dock connector
18:51:53Torneor it may be pins on the remote port
18:52:05Tornebut that's jus a GPIO we know how to toggle on and off
18:52:14Tornesupporting that doesn't mean we can support the actual communications
18:52:17gibbon_ah, ok
18:52:32gibbon_this at least toggles something on the remote port
18:52:48Torneoh? your accessorye turns on and off when you enable/disable it?
18:52:50gibbon_i have a wireless remote that flashes a led on initialisation
18:52:53TorneAh right
18:53:00gibbon_on the apple OS at least
18:53:21gibbon_this light is flashing continously when Accessory power is turned on in rockbox
18:53:29TorneYah, it's probably rtying to talk to rockbox
18:53:33Torneand we're not listening.
18:53:36gibbon_but only under some conditions that i had to reinvestigate
18:53:47Tornethere's no sign of serial being supported on the mini.
18:53:50Tornesorry :(
18:53:55gibbon_its really flashing very fast (like PWMed)
18:54:02TorneYes, it's probably resetting itself :)
18:54:06gibbon_nothing to be sorry about :)
18:54:30TorneWell I prefer when I can tell people "yes rockbox does that you just need to do such-and-such :)
18:54:58 Quit lucent_ (Read error: 104 (Connection reset by peer))
18:56:18gibbon_the last thing about that before i get too off topic: rockbox really does nearly everything i want to do with my ipod and keeps me calm on long car rides... there would be nothing in this software that could set me up after what it has done for me :)
19:00:32gevaertsamiconn: I'm not sure, but this shouldn't matter at all. pamaury tested with his own libusb-based code, and it didn't work then either, so this is not caused by weird driver interaction
19:14:08***Saving seen data "./dancer.seen"
19:16:19 Quit kugel (Remote closed the connection)
19:19:38 Quit robin0800 ("Leaving")
19:26:08gevaertsa 512 byte ringbuffer for USB serial is clearly not enough when enabling logf everywhere
19:31:23gevaertsok, the usb-drv-arc driver is clearly not fully bug-free yet. I get "no prime" errors every now and then on c200
19:42:36 Join FeyLong [0] (i=dec8818c@gateway/web/freenode/x-demfxxxtsxsgzxin)
19:50:41 Join robin0800 [0] (
19:54:50 Quit FeyLong (Ping timeout: 180 seconds)
19:56:46 Quit AB3JU (
19:57:15NJoinAB3JU [0] (
19:58:03 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
20:00:29 Join petur [50] (n=petur@rockbox/developer/petur)
20:02:41*gevaerts suspects that the ARC driver flags some things as errors when they are not
20:06:25 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
20:06:27 Quit linuxstb (Nick collision from services.)
20:06:35 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
20:10:16 Quit robin0800 ("Leaving")
20:25:26 Join Thundercloud [0] (
20:28:40 Quit moos ("Rockbox rules the DAP world")
20:29:32 Join mortti_ [0] (
20:33:16mtRegarding the fixed-point conversion of libatrac, for the imdct the lib could use the exact same files in libcook (which is actually tremor's fixed point imdct) - I'm thinking that including those files for libatrac again wouldn't be a problem as they would be there just for the test program, and they wouldn't affect the core, any opinions on that ?
20:36:52mtlinuxstb: ^^^ ?
20:43:47domonokymt: if it is the same, would it make sense to put it into apps/codec/lib ?
20:44:57mtdomonoky: There's already an mdct library in rockbox (same thing), I'm just including those files to be able to compile a standalone test program, outside rockbox.
20:45:41 Quit mortti (Read error: 110 (Connection timed out))
20:45:41 Quit Thundercloud (Read error: 104 (Connection reset by peer))
20:45:46 Join Thundercloud [0] (
20:47:45domonokycant you just link rockbox mdct lib into your standalone programm ? i think that would be better then to have multiple files with the same content around. (think of bug-fixes, which would have to be done at two place to fix rockbox and your standalone program)
20:49:33mtdomonoky: I could try that, but we already have the same situation for cook, and it hasn't been a problem thus far.
20:50:32domonokyas its a external program, it isnt really a problem. But it would still be nicer :-) ofcoure then the cook standalone program should also do this...
20:50:48CIA-6New commit by gevaerts (r22294): don't treat extremely quick transfers as errors
20:51:35 Join jks [0] (n=jks@
20:53:22jksHas the sansa bootloader been updated for the 3.3 release? If I plug in my sansa when it is off the bootloader boots to the original sansa software and not rockbox
20:55:30jksI have a daily build from sometime in between the 3.2 and 3.3 releases
20:58:48gevaertsjks: there is a new bootloader, but it's not been moved to the correct place on the download server yet
20:59:28jksok thanks
20:59:56jkswill it be announced on the home page when it is published?
21:01:07gevaertsI guess so, yes
21:02:39 Quit flydutch ("/* empty */")
21:03:49 Join saratoga [0] (i=442cd0c4@gateway/web/freenode/x-tlnoblimpjdjovzy)
21:04:41 Quit jks ("ircII EPIC4-2.6 -- Are we there yet?")
21:06:16linuxstbmt: What is the current situation for cook? That the main Rockbox codec is using the imdct from the codec lib, but the test program has its own?
21:07:36mtlinuxstb: Yes. They are both tremor's fixed point imdct but the one in rockbox has some asm optimizations.
21:08:09linuxstbIt would seem more useful for the test program to use the same code as the real codec - isn't that the point of the test program?
21:09:23 Join Zagor [242] (n=bjst@rockbox/developer/Zagor)
21:10:07mtI guess it wouldn't really be of much benefit if they both used the same imdct.
21:10:42gevaertsZagor: I think I know why ccache isn't helpful on build clients: -DAPPSVERSION=\"r22294-090813\" -DYEAR=2009 -DMONTH=08 -DDAY=13
21:10:46 Join stripwax [0] (
21:11:11mtAlthough they are actually using nearly the same imdct but with one faster than the other.
21:11:14 Quit stripwax (Client Quit)
21:11:29*linuxstb needs to leave - back later...
21:11:50saratogaif the imdct version hasnt changed much, maybe a simple header could be added that defines the right things to use the rockbox codec lib imdct with the test programs
21:12:04saratogaid like anyone using those codecs on ARM to get the otpimized version
21:13:07Zagorgevaerts: doh!
21:13:23mtsaratoga: Ah, good point.
21:14:12***Saving seen data "./dancer.seen"
21:14:29gevaertsZagor: only a few files actually use those defines : apps/main.c (APPSVERSION), firmware/drivers/fat.c, apps/menus/time_menu.c (YEAR, MONTH, and DATE)
21:14:57Zagorgevaerts: yes, we should pass those params only to those files
21:15:08saratogaah so even faster building!
21:15:24gevaertssaratoga: faster, but less predictable :)
21:15:51 Quit stoffel (Remote closed the connection)
21:15:58saratogaafter a while each client should have everything cached right?
21:17:17Zagorsaratoga: no. the new bestfit actually lowers the "randomness" of the build distribution
21:17:47 Quit Ubuntuxer ("Leaving.")
21:18:15Zagorgevaerts: are you fixing this, or should I?
21:18:37gevaertsZagor: I don't think I'm familiar enough with the build system for this
21:19:11gevaertsremoving the defines for the general case is easy. Adding them for specific files probably less so :)
21:19:47Zagoryeah you need new rules for those files
21:21:01ZagorI'll fix it
21:21:34gevaertsI'll reinstall ccache on my client :)
21:24:40 Join fdinel [0] (
21:25:48mtHow about this, to be able to link the mdct lib to external programs :
21:26:16 Join qurvel [0] (
21:27:47mtThis line (or something similar) should be added for the program's makefile/headers : -D"ROCKBOX_LITTLE_ENDIAN=1" -D"ICONST_ATTR=" -D"ICODE_ATTR="
21:28:10mt (ROCKBOX_BIG/LITTLE_ENDIAN = 1) according to the target machine of course.
21:30:20mtdomonoky, saratoga : ^^^
21:31:02saratogathats all it takes?
21:31:19mtYes, surprisingly :)
21:31:38saratogathat sounds good
21:31:59mtok, I'll commit the codeclib changes first then.
21:32:00ZagorI question the value of using build date in fat.c and time_menu.c
21:32:05saratogamaybe add a note to the readmes mentioning that they should define CPU_ARM (or whatever it is) if you are running on arm
21:32:30 Quit linuxstb (Read error: 110 (Connection timed out))
21:32:40gevaertsZagor: I don't disagree. Any valid date should be fine. How about the date of the 1.0 release?
21:32:53 Part qurvel
21:33:01Zagorgevaerts: yeah I was thinking something like that
21:33:50mtsaratoga: Where do you think such comment would go ?
21:33:57 Join qurvel [0] (
21:34:08saratogainto the readme file should be fine i think
21:34:25saratogajust say "enable arm optimizations on arm or this will be very slow"
21:35:35JdGordon|why on earth the the build date used in the code?
21:36:03 Quit amiconn (Nick collision from services.)
21:36:06 Join amiconn_ [0] (i=quassel@rockbox/developer/amiconn)
21:36:25 Nick amiconn_ is now known as amiconn (i=quassel@rockbox/developer/amiconn)
21:36:48mtsaratoga: Which readme file ? :) .. I wanted to write a general comment about linking the mdct lib to external programs. Or did you mean just adding this to atrac's & cook's readmes ?
21:37:07saratogathe individual ones
21:37:17saratogasince thats what most people will probably read
21:38:08mt"/docs/README" for example ?
21:38:20 Join pixelma_ [0] (i=quassel@rockbox/staff/pixelma)
21:38:20 Quit pixelma (Nick collision from services.)
21:38:36 Nick pixelma_ is now known as pixelma (i=quassel@rockbox/staff/pixelma)
21:40:11mtapps/README points to docs/API −− can't find any other appropriate readmes there.
21:40:51amiconnZagor, gevaerts: fat.c uses the build date on targets without rtc. Before that, it used a fixed date, but that was a rather bad solution
21:41:21Zagoramiconn: why is build date better than a fixed date?
21:41:32amiconnIt's not a big problem when changing existing files, but when creating new ones.
21:42:28amiconnNew files are created with that date. If it is fixed, it will further and further in the past over time
21:42:42Zagorwhat is the problem with that?
21:43:47amiconnYou don't even remotely know how new a file is, even if you update rockbox regularly
21:43:49JdGordon|considering the date is goign to be wrong anyway, why not just use 0 for it?
21:44:25amiconnBy using the build date it is guaranteed that a newly created file is never "older" than the rockbox build which created it
21:44:40JdGordon|why is that needed thoguh?
21:44:43JdGordon|either way, its wrong
21:44:46JdGordon|or misleading
21:44:47Zagorthat may be logical to you, but not to a user
21:45:35JdGordon|I would guess that having the date come up as jan-1-1970 is more obvious to users whats happening than a date 3 months ago...
21:45:51amiconnWell you can't make it 100% without having an actual rtc. That shouldn't stop us from making it as good as possible
21:45:58Zagorthese targets don't have a clock. hence the file dates on them will *always* be wrong. using a single fixed date for all files is thus _less_ confusing than setting different dates depending on whch version the user happened to have installled at the time
21:47:29Zagorvarious strange dates is not "as good as possible" in my opinion
21:49:32pixelmait at least gives you a hint when you recorded or saved it
21:49:42amiconn"DIR_WrtTime and DIR_WrtDate must be supported, however." from
21:49:50Zagorpixelma: does it? how?
21:50:02mtsaratoga: (question about 10 mins ago).
21:51:01pixelmawell if it has a date of July 1st, you know it can't be from the meeting you recorded on July 15th
21:51:08Zagorin fact a 0 date would be more informative, since it would say that this file was created on the player rather than on the computer. a random date in the past year or two says nearly nothing
21:51:09saratogamt: is it needed to say that we've linked the codeclib?
21:51:26amiconnpixelma: It can, if your rockbox build is from before July 1st
21:51:34Zagorpixelma: what? that's exactly what it can, with the current scheme
21:52:20pixelmahuh, I'm confused now and am not sure anymore what you are talking about
21:52:24amiconnZagor: A zero date (== 01 Jan 1980, 00:00) is not valid for the fields I menitoned. Such a date is e.g. displayed empty in the date+time column in windows explorer
21:52:26mtsaratoga: Not that. I just wanted to write "To use the mdct lib outside rockbox do .... " in somewhere reachable by someone who is looking for such information.
21:52:47amiconnI know there are a few embedded systems which do this, but it's a violation of the fat specs
21:53:00Zagoramiconn: well, a token zero then. Jan 1990 00:00 is just as obvious
21:53:16saratogayou mean someone who wanted to use the codecs?
21:53:20JdGordon|yeah, any date that is obviously wrong would be better
21:53:31JdGordon|octember 32nd :)
21:53:59pixelmaamiconn: it can but then I probably didn't record it on the player then... and I don't think I would create rec_0001.mp3s on the computer
21:54:39mtsaratoga: No, someone who wants to use the mdct lib when porting a codec, unless such comment isn't necessary (opinions ? ). For cook and atrac I'll just add some comments to their readmes.
21:54:51saratogai dont think that is neccessary
21:55:11saratogait should be clear to anyone porting
21:55:29Zagorpixelma: did you mean to write june somewhere?
21:55:55pixelmayes, the last one
21:56:18Zagorin any case, the scheme assumes very frequent updates for any kind of usability
21:56:55JdGordon|a very bad idea could be to use the date+X of the newest file in the folder for new files...
21:56:58ZagorI maintain an obviously incorrect "symbolic" date is more useful and less confusing
21:57:15*Zagor throws pebbles at JdGordon
21:57:21pixelmawell, even if you only updare once a month you have a *hint*
21:57:28JdGordon|it would solve everyones peroblems though :)
21:57:37Zagorpixelma: only? once a month is very frequent
21:57:40amiconnZagor: It doesn't. Frequent updates just improve the behaviour vs. using a fixed date
21:57:56amiconnIf you rarely update, both approaches are essentially the same
21:57:59JdGordon|pixelma: really the proper way to do it is rename your files ocne the recording is finished..
21:58:09pixelmaZagor: very?
21:58:21 Quit LambdaCalculus37 ()
21:58:24Zagoramiconn: no, because a date in the past year says nothing. a 1990 date contains at least some information.
21:58:39 Quit Lss (Read error: 110 (Connection timed out))
21:59:24Zagorpixelma: yes. I believe the average user updates far less frequent than that.
21:59:28saratogahow many targets lack a clock?
21:59:39amiconnZagor: The timestamps are more complex than that. Using a single, fixed timestamp would be bad because it would cause situations where the filetime jumps backwards, which would be very bad e.g. for sync tools
21:59:59amiconnWe need to consider new files and writes to existing files
22:00:26JdGordon|non-rtc targets should have read only file systems!
22:00:29pixelmaJdGordon|: yes, with the virtual keyboard? Although... I do that sometimes, but I don't see the need to change something there. I even look at that sometimes, e.g. when deciding if I can throw away a config file
22:00:38Zagoramiconn: that is true for the current scheme too.create a playlist with your pc, then overwrite it on the archos. boom, no sync.
22:00:44amiconnThe latter should monotonely increase the timestamp, so that a sync tool notices something has changed, and sync into the correct direction. Same goes for manual syncing by looking at timestamps
22:00:51pixelmaZagor: maybe we have different understandings of "very" then
22:01:30amiconnZagor: The current scheme does what I just explained. It takes the timestamp of the existing file, and increases it (by a fixed amount: 1h 15 min)
22:01:42JdGordon|pixelma: I was repliying to your need for the hint in the date.... with non rtc you really shouldnt be going by the file dates
22:01:46pixelmaZagor: the Archos Recorders have an RTC
22:01:53amiconnNew files get a time stamp of <build date, 01:15>
22:02:21Zagoramiconn: so changing the build date to 1990 would thus not change this.
22:02:34CIA-6New commit by mt (r22295): Modify the mdct library to enable using it outside rockbox - No functional changes.
22:02:36Zagorpixelma: I know that
22:04:01amiconnZagor: Using 01 Jan 1990 as a marker doesn't work because of this. It will increase over time if this file is written to, and eventually lose its significance as a marker
22:04:08pixelmaJdGordon|: as I explained I do and find it useful in cases, yes it's coarse but better than nothing...s nice to learn that I "really shouldn't" now
22:05:22pixelmawere there reports of confused people?
22:05:23Zagoramiconn: uh, it takes two million writes to advance a single year... that's not really a problem. besides we could always use 1981 or even 01:00 1980
22:05:31 Join ball [0] (
22:06:03amiconnWhy make things worse than they currently are? By using the build date, the timestamps are at least roughly in the correct range
22:06:21ballIs there an easy way to ask Rockbox to rescan the filesystem? I have files on my iPod Mini that don't show up in the Rockbox menus.
22:06:41Zagorno they are not. they are still *completely* wrong. they just try to fool the user that they are correct!
22:06:52gevaertsamiconn: because it's not at all clear that it would be worse, and the build system would be simpler
22:07:05amiconnNo it wouldn't
22:07:34amiconnYou need to handle APPSVERSION individually, so handling build date as well is essentially the same
22:08:02ZagorA: what date is it? B: I don't know. A: Well, what is the latest date you've heard of? B: <answer> A: ok so let's claim that is the current date
22:08:07amiconnYou could just write a .h that defines these things instead of putting it onto the gcc commandline, and then include the .h where it's needed
22:08:30 Join moos [0] (i=mustapha@rockbox/staff/moos)
22:08:52Zagorit's simply a crappy concept, build system or not. we don't know the date! don't act as if we did.
22:09:00amiconnZagor: Certainly closer to the truth than just saying "X is the current date"
22:09:15amiconnThat's what your suggestion essentially means
22:09:49*amiconn had an idea how to add better signalling to a timestamp that it isn't necessarily correct
22:09:49ballI hate doing dishes without my podcasts :-(
22:09:52ZagorI claim that "close to the truth, but wrong" is actually worse than "so far off the truth that's it's very obviously not the real date"
22:10:35ballReally don't want to nuke the filesystem and start from scratch.
22:10:43pixelmaI prefer the first one
22:11:01Zagorball: rockbox does not have a filesystem scan. use your computer.
22:11:35ballI tried two computers, both show the files.
22:11:38ballRockbox does not.
22:12:38pixelmaball: how did you copy the files over (and to which player)? What is your "show files" setting set to?
22:12:41amiconnZagor: Imo "close to the truth" is better. It's possible to add some signalling "attention, this timestamp isn't necessarily correct" without giving up the close-to-the-truth approach
22:13:05ballI dragged the files in whatever Xubuntu's file browser thing is called... Thunar?
22:13:06Zagoramiconn: how would you suggest to do that?
22:13:14amiconnRight now we're always incrementing the filetime by 75 minutes when writing to a file
22:13:17ballThey show up on the Xubuntu box and also on a NetBSD box
22:13:29ballI'll have a dig in the menus for a "show files" setting
22:13:32Zagorball: what are they called?
22:13:33pixelmawhat player do you have?
22:13:37balliPod Mini
22:13:41ball1st gen
22:13:48Zagorball: the files...
22:14:07balle.g. -rwxrwxr-x 1 root wheel 3373183 Aug 13 19:52 G6_lineup_3ZEQ.mp3
22:14:13amiconnIf we would instead round it before so that the time part of a file always ends up as a multiple of 01:11:11 when written to by rockbox, this is probably a rather good hint
22:14:57amiconnHmm, that won't work after 05:55:55.
22:15:37amiconnBut the idea could certainly be modified so that it works without too uneven increments
22:16:05amiconnE.g. always setting minutes and seconds to :22:22, , and increment by an hour at a time
22:16:15ballfile(1) reports G6_lineup_3ZEQ.mp3: Audio file with ID3 version 23.0 tag, MP3 encodin
22:16:39ballPerhaps I just need to reinstall rockbox
22:16:52amiconn(not :11:11 because I forgot that wrttime seconds on fat must be even)
22:16:58Zagoramiconn: how about subtracting 20 years? that's a pretty clear flag, but you still have the date info
22:17:49Zagoranyway, that would mean keeping the need to rebuild fat.c every time, so it doesn't really change anything as far as I'm concerned
22:17:51 Quit KBH ()
22:18:05ZagorI'm just going to leave fat.c as it is for now.
22:18:27amiconnYeah, but if you're sorting by date (I know that can't be 100% correct anyway tough) it would look quite odd, as those files would sort totally different)
22:18:40amiconnIt also leaves the problem that writing to a file woud set the time backwards
22:18:45ballWell that's strange
22:18:46amiconn(first-time write)
22:18:59Zagoramiconn: only new files
22:19:08ballSomehow Rockbox was set to only view Playlists
22:19:12Zagorwe naturally don't change the increment concept
22:19:17ballI know I didn't do that deliberately
22:19:21ball...very odd.
22:19:27ballThanks for your help Zagor
22:19:36ball...and pixelma
22:19:36CIA-6New commit by mt (r22296): Also modify mdct_lookup.c for compiling outside rockbox.
22:19:40amiconnZagor: But then you have no indication that the timestamp of a file modified by rockbox isn't correct
22:19:54Zagorball: actually it was pixelma who told you about the setting :)
22:20:12ballZagor: well you helped too, so thanks.
22:20:24Zagoramiconn: true. well, as I said I'm not changing this right now.
22:20:42ballPerhaps back later, after I've washed some dishes
22:20:42amiconnThe :22:22 idea would set this on all files rockbox is writing to, and increment the hour by 1. It would also increment the date to the build date if it's before that
22:20:44 Quit ball ("leaving")
22:20:49 Quit saratoga (Ping timeout: 180 seconds)
22:21:05*amiconn thinks this is at least worth a try
22:21:24amiconnIt could also be explained in a simple way in the manual
22:21:34amiconngevaerts: What do you think about my .h file idea?
22:21:46pixelmamm... ball left
22:22:50 Join saratoga [0] (i=442cd0c4@gateway/web/freenode/session)
22:23:11Zagoramiconn: moving it to a .h file just adds complexity. it won't change the ccache penalty.
22:23:36amiconnIt will, as only those files actually needing the values would depend on that .h
22:24:17gevaertsamiconn: I think it's a bit more complex. You suddenly have a file that needs to be rewritten without dependencies saying so
22:24:49amiconnThe .h file would always be written at build time, directly, without using -D on a gcc commandline
22:25:29Zagorit still means ccache cannot cache fat.c. same as if I just add a new Makefile rule for it.
22:25:48ZagorI prefer two lines in the makefile over creating new files in the build dir
22:25:54JdGordon|what about not being so agressive with deleting the build dir each round?
22:26:13amiconnHmm, maybe I don't understand the problem correctly?
22:26:31ZagorJdGordon|: too dangerous. we want a completely clean build every time, in case there is a dependency fault
22:26:34CIA-6New commit by mt (r22297): Modify libcook to use rockbox's mdct library in its standalone test program.
22:26:36amiconnIiuc gevaerts means that those -D[efines] lead to ccache not caching anything?
22:26:40 Join HBK [0] (
22:26:50amiconnBut then I wonder why ccache helps at all
22:27:05gevaertsdoes it?
22:28:04JdGordon|Zagor: yeah, but how about.. unless some specified files change (configure, config*.h, english.lang, etc) reuse the directory, then if that ails do a clean build... surely that would still be faster on average than a full rebuild every time?
22:28:04amiconnWell if I do configure && time make && make clean && time make with primed ccache, the second 'make' is much faster
22:28:37amiconn(except on cygwin, but that's another matter)
22:28:43gevaertsamiconn: yes, but the revision is the same
22:28:55ZagorJdGordon|: that would assume we know which files trigger dependency bugs. we don't!
22:29:07Zagorwe don't even know of such bugs
22:29:29JdGordon|well, yeah...
22:29:32amiconngevaerts: Hmm, that should be tested across a daychange then?
22:30:07Zagoramiconn: two different revisions is enough, since all compiles get APPVERSION
22:30:42amiconnYes, but different revisions will cause valid recompiles
22:31:19Zagoryes but you should still see a difference if ccache works (and if the revision isn't huge)
22:31:23 Quit saratoga (Ping timeout: 180 seconds)
22:31:30Zagor*revision diff
22:31:48amiconnHmm, could potentially use a commit in www/ or rbutil/ or similar
22:32:34Zagorgood idea
22:32:59amiconnWhat was r22277 and r22278?
22:33:54amiconnah, so r22276 vs. 22277 would work
22:33:56*amiconn tries
22:39:03CIA-6New commit by mt (r22298): Modify libatrac to use fixed-point arithmetic.
22:49:02 Join saratoga [0] (i=442cd0c4@gateway/web/freenode/x-dfdxxnlncdhdgggh)
22:49:22saratogagreat work with libatrac
22:49:31saratogathe ffmpeg people should be impressed
22:51:06CIA-6New commit by mt (r22299): Remove mallocs from libatrac.
22:51:16mtsaratoga: Thanks :)
22:53:11 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
22:53:33*domonoky applauds to mt for his really impressiv gsoc work !
22:53:58mtdomonoky: Thank you !
22:54:18saratogaare there any RM codecs remaining?
22:54:51mtra1 and ra2
22:55:29 Quit qurvel (Connection timed out)
22:56:26CIA-6New commit by mt (r22300): Remove a forgotten DEBUGF from librm/rm.c - No functional changes.
22:59:12 Quit evilnick ("Page closed")
23:00:11saratogamt: better add the ASM versions of fixmul before someone tries to play a file ")
23:01:00mtsaratoga: Sure ;)
23:02:02UtchybannI just auto-magically sync the rtc of my ipod with udev (linux).
23:02:42UtchybannDo you think I should write a little 'howto' for linux users ?
23:02:54saratogathe codec is so light on memory you might as well just ICONST everything in atrac3data_fixed.h
23:02:59saratogashould make coldfire a LOT faster
23:03:09BagderUtchybann: sounds like a good idea!
23:03:44mtsaratoga: channel_units is 32KB, is that ok ?
23:04:22amiconnIt seems there is a bug in 'configure' regarding ccache handling
23:04:32UtchybannBagder: ok. should be ready before 3.4 and official support in rockbox ;-)
23:04:36saratogamt: thats not in that file is it?
23:04:52saratogaoh i see
23:04:54JdGordon|Utchybann: add it to the wiki....
23:04:59saratogalets wait on that, its probably too big
23:05:12*domonoky calls again for perl hackers to fix genlang.cgi for releases :-)
23:05:21JdGordon|no point adding it to the manual... linux users dont read manuals :)
23:06:04domonokyUtchybann: i would think a wikipage about this would be nice.
23:06:36*amiconn blames r2228, i.e. kugel
23:06:50CIA-6New commit by zagor (r22301): Added event logging to database.
23:07:07JdGordon|amiconn: you sure about that one? :D
23:07:07 Quit DarkSpectrum (Read error: 104 (Connection reset by peer))
23:07:10amiconn'configure' doesn't auto-enable ccache. Only −−ccache would do that. Wrong condition in elif.
23:07:23amiconnJdGordon|: Eh, r22228
23:08:25UtchybannJdGordon, domonoky : a new wiki page in 'For Users' section seems the right place for me.
23:09:13CIA-6New commit by amiconn (r22302): Fix ccache auto-detection.
23:09:50saratogamt: we have more buffers then we have space in IRAM for them on most targets, but maybe some of the buffers in ATRAC3Context and channel_unit could be shared to save memory
23:10:49*domonoky tries to think of a Price for the perl hacker who fixes genlang.cgi for releases. Maybe Beer ? (but shipping is expensive) :-)
23:11:18 Join hd [0] (
23:11:20Zagoramiconn: wow. that would be another reason why ccache doesn't work :-)
23:11:21 Quit hd (Read error: 104 (Connection reset by peer))
23:11:35amiconnZagor: yup
23:11:47amiconnI wondered why my primed ccache doesn't fill at all....
23:14:15***Saving seen data "./dancer.seen"
23:17:40saratogai have to run, but i bet tempbuf can be shared with one of the other buffers
23:17:45 Quit saratoga ("Page closed")
23:17:51*pixelma wonders about the 2 red dots in the green build table - same error but on two different servers?
23:18:26pixelmanot exactly the same but similar
23:18:34CIA-6New commit by mt (r22303): Call the asm definitions of fixmul functions when possible.
23:18:41amiconnI.e. you don't need to do anything about those -D
23:19:10mtsaratoga (logs) : Yes I was just thinking of that today.
23:19:17amiconn(this is a recorder build on 'triton', without using make -j)
23:19:21JdGordon|2s difference.... thats big enough to spend an hour on :D
23:20:27amiconnJdGordon|: Those 2s different are a mix of (a) measuring uncertainty (varying load) and (2) recompiling the few files actually using those defines
23:20:30 Join einhirn [0] (
23:21:02amiconnIf they would cause what gevaerts suspected, the 3rd build would have been in the 2 minutes range
23:21:10Zagorthis means the ccache manual is wrong: it's not comparing the commandline, only the preprocessor output
23:21:24gevaertsZagor: maybe they ignore -D?
23:21:39Zagorwell, I would call -D part of the commandline :)
23:21:48Zagorso leaving that out is still a manual bug
23:24:10pamaurygevaerts: did you test it ? I saw you comment about the usb driver flagging something as error while it's not: what do you mean ?
23:26:21gevaertspamaury: I didn't do tests yet to see if the 96 bytes thing is the same for all targets. I did verify that serial now works properly though. While doing that I found this bug in the driver
23:26:44amiconngevaerts: Did you try serial on the beast?
23:26:51gevaertsamiconn: not yet
23:28:16n1sif anyone is serious about chasing build speed, i think optimizing genlang could give some nice speedups as iirc it takes about ~1/4 of the build time, can't remember how to build only langs though...
23:28:25gevaertsWhat happened is that the driver waits for the ENDPT_PRIME bit to be set, and then looks at REG_ENDPTSTATUS, where the endpoint-specific bit should be 1. However, in some cases the transfer is fast enough to complete before that check actually happens, so the bit is 0 again. I fixed that by also checking ENDPT_COMPLETE
23:28:42n1sthat's with a primed ccache of course
23:29:22rasherSo all this time, we haven't really been using ccache much, if at all?
23:29:35Zagorseems so
23:29:43pamaurygevearts: I did some test on usb_serial and when I submit a >=97 transfer, it proceeds normally, prepare a td, tells the hardware to proceed and then there is not feedback anymore, no interrupt, nothing
23:30:09pamaurywhat is ccache ?
23:31:31rashern1s: Yeah, I think genlang could be smarter
23:32:11 Join ehntoo [0] (
23:32:27n1spamaury: a cache for preprocessed c files that speeds up compiles significantly in many cases
23:32:42Zagorrasher: not since 22228
23:33:11n1srasher: too bad i can't hack perl to save my life then :/
23:33:15pamauryand what is the problem with genlang ? Is it too slow ?
23:33:43n1swell, it could likely be faster, not really a problem
23:33:44rashern1s: Yeah, I'm not going to touch it either
23:33:48JdGordon|there might be a very small speed up if all the deprecated strings were removed...
23:34:02rasherJdGordon|: that's not going to be significant at all
23:34:16JdGordon|I said very small :)
23:34:18n1srasher: btw, what happend to your lang xml-ification?
23:35:07n1sJdGordon: dropping all the useless features will help more1 :D
23:35:49rashern1s: I think I've all but abandoned it
23:36:16pixelmathat weird Iaudio M3 delta seems to be caused be the build client - pyros-roolku
23:36:30rasherI don't think the benefits are important enough, and the preliminary result seemed slower
23:38:56pamaurygevarts: I submit a patch on FS to fix the const correctness of usb_serial, can you look and it and eventually commit it if it's correct ? It's not really important altough
23:40:09pixelmahmm, the previous build table gave some info about the server when hovering, gcc and ld version if I remember correctly. Is it possible to bring that back with the client system?
23:40:24Zagorpixelma: it's in the log
23:40:59 Quit einhirn (Read error: 104 (Connection reset by peer))
23:41:26rasherZagor: any idea about the negative upload predictions?
23:41:30pixelmait's not the same when trying to compare builds though (to find reasons for weird failure/warnungs or deltas)
23:41:43Zagorrasher: yes, it's ... crappy code :)
23:41:44pixelmawarnings too
23:42:01domonokyZagor: you wrote genlang.cgi. Is there someway to convince you to fix it for releases ? (pulling the lang file from the correct release branch, if the revision is something like 3.3) :-)
23:42:21*domonoky offers beer ! :-)
23:42:46amiconnn1s: Would it be possible to let ccache cache genlang output as well?
23:43:22Zagoramiconn: no, it only understands compiler (only gcc even?) command lines
23:43:42rasherMaybe genlang could do its own caching?
23:43:43amiconnAh, hmm
23:44:29Zagorrasher: I was under the impression the slow speed is due to poor code. is that incorrect?
23:45:18*pixelma feels ignored otherwise
23:45:28Zagordomonoky: uh, I guess. it's not at the top of my list right now though.
23:45:30rasherZagor: Somewhat. It's at least not as efficient as it could be
23:45:59 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
23:46:28Zagorpixelma: I don't really understand why you want it. Adding it means I have to load and parse 2160 log files every time that table is produced. no fun.
23:50:12amiconnrasher: One thing it could perhaps do is storing the parse result of english.lang on disk in some binary form, so that subsequent calls could just fetch that without further processing
23:50:46*amiconn doesn't know how to do this in perl
23:50:54pixelmait's not the only point I thought was worth bringing up, actually I stumbled upon this because I wanted to see if there is anything obvious why pyros-roolku produces this weird delta on the Iaudio M3 buiild - just remembered that this function had been there before and proved useful
23:51:11rasheramiconn: Indeed. It should do the same for the results of the other languages and store that with a hash of some sort
23:51:55JdGordon|are all the languages actually dependant on english.lang (in the make dependancies i mean)
23:51:57amiconnI'm not talking about automatic caching. Just parse english.lang once per build, and use the result for processing all the other languages
23:52:47ZagorI think fixing the code should be done before adding caching. caching is for inherently hard problems, not poor code.
23:53:49pixelmaZagor: that build table one was a question because I wasn't sure if I didn't overlooked something - I overlooked a bit that now there are now quite a few more builds but... thought asking wouldn't hurt
23:54:03rasheramiconn: I'm just saying there's no reason not to do the other thing also
23:54:30 Nick evilnick_home1 is now known as evilnick_home (
23:54:33*rasher predicts genlang will not be fixed in a year
23:54:45amiconnJdGordon|: If they aren't, that would be a bug in the build system
23:55:11n1sJdGordon: yes, all langs depend on the order of the strings in english.lang, and also the english.lang strings where that string is missing in the translation
23:55:11Zagorpixelma: asking doesn't hurt. I just wanted a really good reason before spending a bunch of time on it.
23:55:18JdGordon|well no... doesnt the genlang script actually let you use any .lang as the base lang?
23:55:19amiconnCreating a .lng needs english.lang for reference, so all .lng should depend on their respective .lang and english.lang
23:56:48n1sJdGordon: yes, you *can* specify any other lang as the base lang to pass to genlang but why do you want to?
23:57:14pixelmaZagor: sure, it was also meant more general - no-one replied to me 2 previous questions

Previous day | Next day