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 2010-08-11

00:05:21 Join markun [0] (~markun@rockbox/developer/markun)
00:08:02 Quit markun (Read error: Connection reset by peer)
00:16:38 Quit petur (Quit: Zzzzz)
00:16:39 Quit Nausicaa (Read error: Connection reset by peer)
00:19:09 Quit sasquatch (Read error: Operation timed out)
00:20:43 Join Horschti [0] (~Horscht@xbmc/user/horscht)
00:22:41 Quit Horscht (Ping timeout: 276 seconds)
00:23:35 Join Mr_Sensitive [0] (~Dreamxtre@
00:23:43 Quit Mr_Sensitive (Read error: Connection reset by peer)
00:26:17 Quit Dreamxtreme (Ping timeout: 276 seconds)
00:27:30 Quit S_a_i_n_t (Read error: Connection reset by peer)
00:27:51 Quit slck ()
00:29:44 Quit domonoky1 (Ping timeout: 245 seconds)
00:31:46 Join S_a_i_n_t [0] (S_a_i_n_t@
00:32:35 Quit liar (Ping timeout: 258 seconds)
00:32:44 Join markun [0] (~markun@rockbox/developer/markun)
00:32:51 Quit bertrik (Quit: :tiuQ)
00:34:42 Join sasquatch [0] (
00:35:56 Quit S_a_i_n_t (Ping timeout: 240 seconds)
00:38:14 Join Dreamxtreme [0] (Archimedes@
00:39:17 Join Chronon [0] (
00:40:19 Join slck [0] (
00:41:00 Quit Jerom (Ping timeout: 252 seconds)
00:41:14 Join S_a_i_n_t [0] (S_a_i_n_t@
00:49:27 Quit bluebrother (Ping timeout: 260 seconds)
00:50:46 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
00:51:10 Join bluebrother [0] (
00:51:10 Quit bluebrother (Changing host)
00:51:10 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
00:53:20***Saving seen data "./dancer.seen"
01:16:32 Join dfkt_ [0] (dfkt@unaffiliated/dfkt)
01:20:40 Quit dfkt (Ping timeout: 265 seconds)
01:21:15 Join dfkt [0] (dfkt@unaffiliated/dfkt)
01:23:57 Quit dfkt_ (Ping timeout: 258 seconds)
01:25:22mc2739JdGordon: pong
01:25:26 Quit JdGordon (Ping timeout: 276 seconds)
01:26:21 Quit evilnick_B (Quit: Page closed)
01:28:54 Part toffe82
01:32:08 Join cjcopi [0] (
01:37:14 Quit n1s (Quit: Lämnar)
01:40:21 Join CaptainKwel [0] (
01:42:57 Quit kugel (Ping timeout: 265 seconds)
01:44:50 Join mal [0] (~mal@
01:49:56 Join kugel [0] (~kugel@rockbox/developer/kugel)
01:50:17 Quit kugel (Remote host closed the connection)
01:56:00 Nick slck is now known as user890104 (
01:58:05mali want install gentoo on my ipod 30gb like usb-hdd and i dont want lose any data, anybody have had expirience?
01:58:31literalthat's not really a rockbox question
01:59:31literalmore like #ipodlinux
01:59:56 Quit bmbl (Quit: Bye!)
02:07:28 Quit angelwolf71885 (Remote host closed the connection)
02:17:20 Join angelwolf71885 [0] (
02:19:29 Quit mt (Remote host closed the connection)
02:22:42 Quit mal (Quit: Leaving)
02:22:43 Join Horscht [0] (~Horscht@xbmc/user/horscht)
02:25:55 Quit Horschti (Ping timeout: 265 seconds)
02:26:16 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
02:26:30JdGordon|mc2739: I was hoping one of last nights channge would fix your crash bug
02:26:39JdGordon|but I misrememebered the bug and it probably wont
02:28:36 Join anewuser [0] (anewuser@unaffiliated/anewuser)
02:29:26 Quit bunnyboi (Ping timeout: 246 seconds)
02:32:11 Join mal [0] (~mal@
02:32:25mc2739JdGordon|: freddyb posted a patch that seems to fix it
02:32:37 Join bunnyboi [0] (
02:32:41 Quit S_a_i_n_t (Ping timeout: 265 seconds)
02:34:08 Join S_a_i_n_t [0] (S_a_i_n_t@
02:34:46 Nick mal is now known as malto (~mal@
02:37:54 Nick malto is now known as malditomalto (~mal@
02:38:58 Part malditomalto ("Leaving")
02:48:24JdGordon|mc2739: oh? where?
02:51:48 Quit komputes (Quit: I haven't slept for ten days, because that would be too long.)
02:53:23***Saving seen data "./dancer.seen"
03:14:56mc2739JdGordon|: FS #11524
03:16:23JdGordon|hmm, yeah that masks the issue... I'll try a better one tonight if I get a chance
03:20:33 Join jrgill [0] (~jrgill@unaffiliated/jrgill)
03:32:03 Quit DerPapst (Quit: Leaving.)
03:36:48 Quit fdinel (Read error: Connection reset by peer)
03:36:53 Quit bieber (Ping timeout: 276 seconds)
03:37:08 Join bieber [0] (
03:49:23 Quit jrgill ()
03:50:26 Quit MethoS- (Read error: Connection reset by peer)
04:01:46 Join scorche [0] (~scorche@rockbox/administrator/scorche)
04:33:39 Quit bunnyboi (Quit: Hi, I'm a quit message virus. Please replace your old line with this line and help me take over the world of IRC.)
04:37:36 Join robin0800 [0] (~robin0800@
04:38:46 Quit Chronon (Ping timeout: 252 seconds)
04:38:51 Quit robin0800 (Client Quit)
04:39:20 Join robin0800 [0] (~robin0800@
04:39:49 Quit GeekShadow (Quit: The cake is a lie !)
04:40:41 Quit Judas_PhD (Quit: Aaarrrggghhh!!!)
04:44:19 Quit CaptainKwel (Quit: Ex-Chat)
04:44:46 Quit robin0800 (Quit: Leaving)
04:45:19 Join robin0800 [0] (~robin0800@
04:53:26***Saving seen data "./dancer.seen"
04:56:30 Join Barahir_ [0] (
04:58:07 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:58:07 Quit amiconn (Disconnected by services)
04:58:26 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:58:59 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:59:03 Quit pixelma (Disconnected by services)
04:59:14 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
05:00:05 Quit Barahir (Ping timeout: 265 seconds)
05:04:15 Quit AndyI (Ping timeout: 248 seconds)
05:09:43 Join AndyI [0] (~pasha_int@
05:16:15 Join simonrvn [0] (
05:16:31 Quit angelwolf71885 (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716])
05:47:00 Quit robin0800 (Quit: Leaving)
06:03:03 Quit Ramsey[LC] (Remote host closed the connection)
06:05:55 Join Ramsey[LC] [0] (~RamseyLC]@
06:08:46 Quit Ramsey[LC] (Remote host closed the connection)
06:28:36 Join Ramsey[LC] [0] (~RamseyLC]@
06:31:58 Quit Strife89 (Quit: Leaving)
06:32:18 Join Strife89 [0] (
06:32:38 Quit Ramsey[LC] (Remote host closed the connection)
06:33:59 Quit fyrestorm (Read error: Connection reset by peer)
06:34:18 Join fyrestorm [0] (
06:42:44 Join Judas_PhD [0] (
06:43:31 Quit anewuser (Quit: FAMICOMPO 7 - August 15 Through August 31 JST -
06:53:27***Saving seen data "./dancer.seen"
07:03:03 Quit xvello (Ping timeout: 264 seconds)
07:10:29 Join shai_ [0] (
07:11:30 Join xvello [0] (
07:13:29 Quit shai (Ping timeout: 265 seconds)
07:25:46 Join Horschti [0] (~Horscht@xbmc/user/horscht)
07:29:31 Quit Horscht (Ping timeout: 276 seconds)
07:41:09 Join Chronon [0] (
07:56:02 Join LinusN [0] (linus@rockbox/developer/LinusN)
08:19:12 Quit kramer3d (Quit: Leaving)
08:30:27 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
08:30:49*bertrik found the memory leak in the vorbis codec
08:31:32 Join wodz [0] (
08:31:51 Join ender` [0] (
08:32:38bertrikit's in block.c, where it first tries to allocate the "first_pcm" buffer from iram, which fails, so it then allocaties it from the general malloc pool, but never frees that buffer
08:34:40 Nick shai_ is now known as shai (
08:36:46 Quit ender` (Read error: Connection reset by peer)
08:37:13 Join ender` [0] (
08:38:16 Join Zagor [0] (bjst@rockbox/developer/Zagor)
08:45:43 Quit Chronon (Ping timeout: 252 seconds)
08:53:28***Saving seen data "./dancer.seen"
09:13:35 Join petur [0] (d408b802@rockbox/developer/petur)
09:20:30 Quit r0b- (Read error: Connection reset by peer)
09:24:06 Quit bertrik (Quit: :tiuQ)
09:38:38 Join Rob2223 [0] (
09:42:08 Join ferret_0567 [0] (
09:42:11 Quit Rob2222 (Ping timeout: 252 seconds)
09:42:44ferret_0567has anybody tried to reverse engineer the ipod nano 3g yet?
09:50:10wodzthere are some attempts AFAIK
09:50:38wodzTheSeven is good source of information
09:52:36 Quit ferret_0567 (Quit: CGI:IRC (Ping timeout))
10:01:56 Quit rvvs89 (Ping timeout: 240 seconds)
10:05:39 Join rvvs89 [0] (
10:15:37 Join u42p [0] (
10:17:59u42phi, i ran the on a sansa clip+ with svn r27770, could someone add the table to the wiki?
10:19:49u42pit looks kinda weird, but thats what the perl script gave me :)
10:20:19u42phm, index.html snuck in there
10:33:57 Join DerPapst [0] (
10:34:38 Join n1s [0] (~n1s@rockbox/developer/n1s)
10:37:53 Quit Dhraakellian (Ping timeout: 276 seconds)
10:38:19 Quit JdGordon| (Quit: leaving)
10:46:22CIA-9New commit by nls (r27771): Delete useless #ifndef #endif pair.
10:48:07 Join Dhraakellian [0] (
10:48:25CIA-9r27771 build result: All green
10:51:13 Quit Horschti (Quit: Verlassend)
10:53:29***Saving seen data "./dancer.seen"
11:11:45CIA-9New commit by nls (r27772): Mark some local variables with 'static'.
11:13:36CIA-9r27772 build result: All green
11:19:13 Quit bieber (Ping timeout: 258 seconds)
11:19:29 Join bieber [0] (
11:19:48 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
11:29:51 Quit Rob2223 (Quit: Rob2223)
11:34:11 Join Horscht [0] (~Horscht@xbmc/user/horscht)
11:52:41 Quit markun (Read error: Connection reset by peer)
11:56:38 Join markun [0] (~markun@rockbox/developer/markun)
11:57:06 Quit markun (Read error: Connection reset by peer)
11:59:46 Part LinusN
12:03:19 Join LinusN [0] (linus@rockbox/developer/LinusN)
12:12:01 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
12:17:20 Join markun [0] (~markun@rockbox/developer/markun)
12:23:11 Quit petur (Quit: reboot)
12:26:34 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
12:28:56 Quit n1s (Quit: Lämnar)
12:32:29 Join Rob2222 [0] (
12:35:19 Join anewuser [0] (anewuser@unaffiliated/anewuser)
12:37:22 Join robin0800 [0] (
12:41:45 Join kugel [0] (~kugel@rockbox/developer/kugel)
12:43:15 Join petur [0] (d408b802@rockbox/developer/petur)
12:44:22 Quit user890104 (Read error: Connection reset by peer)
12:44:27 Join user890104 [0] (
12:53:30***Saving seen data "./dancer.seen"
13:02:14 Join angelwolf71885 [0] (
13:05:35 Quit bieber (Ping timeout: 276 seconds)
13:05:49 Join bieber [0] (
13:15:23 Join shai_ [0] (
13:16:39 Quit angelwolf71885 (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716])
13:16:51 Join angelwolf71885 [0] (
13:17:34 Join T44 [0] (
13:17:36 Quit shai (Ping timeout: 245 seconds)
13:18:25 Quit mc2739 (Ping timeout: 260 seconds)
13:20:56 Quit Topy (Ping timeout: 245 seconds)
13:22:21 Join hebz0rl [0] (
13:25:25 Join preetyboyswag [0] (
13:26:00preetyboyswaghey people i need a quick answer i gonna take back my mp3 player today , is there any mp3 player that i can buy that support rockbox ?
13:26:09preetyboyswagi mean the sensa ones are they still on sale
13:26:37preetyboyswags ansa sorry
13:29:28u42psansa clip+ is great with rockbox
13:29:35u42pand a 16gb sd card :)
13:29:39u42pi wish i had 32
13:30:18 Quit pamaury (Remote host closed the connection)
13:30:31 Join pamaury [0] (
13:30:31 Quit pamaury (Changing host)
13:30:31 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
13:30:54 Join MethoS- [0] (~clemens@
13:31:20Stummisomebody with a sansa fuze 2 here? Is there a way to get the songs which are there on default? They seem to be somewhere on the filesystem, but are not visible on pc. (Maybe a special filesystem-hack?)
13:32:00preetyboyswagthx u42p i am from germany and i going look out if i can find some good rockbox mp3 player
13:33:41u42pi would get 2 or 4 gb and a mini sd card
13:33:59preetyboyswagi try to get sansa fuze today
13:35:46Stummipreetyboyswag, sansa fuze is good. But usb isn't supportet yet
13:37:54preetyboyswag@Stummi i talked to saturn already they wont take my philips back becouse i lose the bong for the rechange -.-
13:38:14preetyboyswagwell done now i need to wait or try to sell it well cya later guys
13:38:47Stummipreetyboyswag, which model exactly is that?
13:38:47 Quit preetyboyswag (Quit: CGI:IRC (EOF))
13:39:18 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
13:47:18pixelmaStummi: you mean the ones copied over in MTP mode?
13:48:12Stummipixelma, no, there a just some songs on the sanza fuse when you buy it
13:49:25Stummiand i mean this songs
13:49:28pixelmawell, if it is similar to the Sansa c200 then the few songs that came with it were copied over (by Sandisk) in MTP mode...
13:49:32 Join einhirn [0] (
13:52:05pixelmaunfortunately I don't know the file structure of the Fuze, but on the c200 I could find them in a SYSTEM subfolder as ".dat" files. I read somewhere that the Fuze OF creates other folders which are hidden differently, maybe you can find something about it in the forums - or if you have access to WMP try to get them off in MTP mode
13:52:41 Quit sasquatch (Ping timeout: 265 seconds)
14:02:05 Quit delt_zZzZ (Read error: Connection reset by peer)
14:03:11 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
14:05:14 Join sasquatch [0] (
14:08:26JdGordonkugel: glad to see you've changed your mind. Are you liking the idea of downloading on first install? I'm not sure what to make of your email
14:19:23 Quit krazykit (Quit: be back in 20)
14:19:50 Join stooo [0] (
14:20:04kugelJdGordon: well, that's the only way to avoid 20MB apks, isn't it?
14:20:18 Join Temik [0] (~Temik@
14:20:31JdGordon20MB is extreme, and on froyo we can install to sdcard, but yes
14:21:08JdGordonany objections then to me cleaning up that patch and getting it commited so we can add android to the build system?
14:21:31 Quit stooo (Client Quit)
14:21:43kugelthere's more needed for this in the build system, isn't it?
14:22:14JdGordonnope. configure and make zip need a tiny bit of fiddling, and build clients need the NDK
14:22:14JdGordonnothing else
14:22:42kugelwhat do you want to build with the build system?
14:23:05JdGordonI mean get it in the usual builds
14:23:06kugelyou can't simply only build rockbox, you need the complete apk
14:23:31JdGordonif we go with my idea/patch the apk is seperate
14:23:43kugelwhat about the java parts?
14:24:07JdGordonyes, they would be built when needed
14:24:11JdGordonnot automatically thogh
14:24:30JdGordonthey could be, but that would be extra work
14:24:31kugelI think we should build the complete package
14:24:52JdGordonthe complete package is *what* though?
14:25:00kugelthe .apk
14:25:21JdGordonsome images, the java wrapper and enough code to download the
14:25:29JdGordoni.e 90% static
14:25:58kugelI don't think we're in a state where it's almost completely static yet
14:26:04JdGordonthere is no actual need to build it frequently
14:26:13kugeland even then, it's 100% required for rockbox being functional
14:27:17 Quit anewuser (Quit: FAMICOMPO 7 - August 15 Through August 31 JST -
14:27:26 Quit Temik (Remote host closed the connection)
14:27:35JdGordonI understand that. What I'm saying is we start building now, and we/someone builds a apk which works with the nightly. people who want to test can manually install that untill we decide it is ready for the market
14:27:53JdGordonexactly like how new ports are built almost as soon as it compiles fully anyway
14:28:46 Quit angelwolf71885 (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716])
14:28:57kugelwhy are you pushing this?
14:30:31JdGordonbecause I want it to be working from the beginning, and I've mostly finished the patch anyway
14:31:17JdGordonimo the port is on par with other targets now. especially the d2
14:32:39kugelbut it has individual problems like any other port, we don't need to get something out within a day (your mail isn't even a day old)
14:33:13JdGordonI'm not saying anything about building the apk yet
14:35:05 Join Ramsey[LC] [0] (~RamseyLC]
14:37:32kugelI think we should discuss more about this stuff and get more opinions
14:37:41kugelwe're not in a hurry anyway
14:37:56 Join mal [0] (~mal@
14:38:00 Join krazykit [0] (~kkit@
14:38:26 Quit JdGordon (Ping timeout: 240 seconds)
14:41:07 Part mal
14:44:52 Quit markun (Read error: Connection reset by peer)
14:47:58kugelI think it could be worthwhile to examine what's needed to make rockbox screen resolution independant (I actually don't think it's that much). that we could ship a single for all devices, and only need to download the themes
14:53:29 Join markun [0] (~markun@rockbox/developer/markun)
14:53:31***Saving seen data "./dancer.seen"
14:57:20Zagorkugel: research is always good
14:57:30Zagor(meaning: go for it)
15:01:44 Nick shai_ is now known as shai (
15:11:14 Join fdinel [0] (
15:14:41 Quit markun (Read error: Connection reset by peer)
15:17:56 Join evilnick_B [0] (0c140464@rockbox/staff/evilnick)
15:19:11 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
15:19:25 Quit robin0800 (Quit: Leaving)
15:20:11JdGordonthe lcd size problem only fixes one issue I have with the apk
15:20:12 Join LambdaCalculus37 [0] (~LambdaCal@rockbox/staff/LambdaCalculus37)
15:21:29kugelthe other is?
15:21:41 Quit Dhraakellian (Ping timeout: 264 seconds)
15:22:24JdGordonnot being able to load any build you want (given that the build is compatable with the java wrapper)
15:22:53kugelbut you can install any apk you want?
15:23:01kugelwhere's the difference?
15:23:21JdGordonbuidling the apk is a massive PITA
15:23:33kugelnot really
15:24:04kugelI find it much more convinient actually to run the apk than to put something on the microsd when using eclipse
15:24:23JdGordonmuch more than just building the zip
15:24:30kugeland once there's a script for doing it in the command line you don't even need eclipse
15:24:47JdGordonyou do need the sdk though
15:25:00JdGordonwhich is more than just the sdk download
15:25:26JdGordonso why should we make it harder than needed?
15:25:48 Join sko [0] (~AndChat@
15:25:56JdGordonwhy the hell shuold I be forced to rebuild java to run my own build when it changes nothing on that wrapper level
15:26:13kugelbecause that's the way it works on android
15:26:19JdGordonthats bloody stupid
15:26:56JdGordonIf I want to try someone elses build I have to install their apk? and if they named it the same I have to uninstall the official apk
15:27:23JdGordonthats just moronic when the answer is so simple
15:27:26kugelexactly like on daps
15:27:32JdGordonnot at all
15:27:50JdGordonyou dont need to install a new bootloader on the ipod to load somoenes build
15:28:00JdGordonyou do need to download their though
15:28:08kugeland overwrite the official one
15:28:12LambdaCalculus37kugel: Any new updated apks I can try out on my Magic?
15:28:42 Join markun [0] (~markun@rockbox/developer/markun)
15:29:12JdGordonARG! I shouldnt need to load a new apk to try out someones (someones could be the svn build here)
15:29:36JdGordonbecause thats just stupid
15:29:45kugelthat's the only way to make sure the java stuff fits to the rockbox binary
15:29:55JdGordonno it isnt
15:30:20JdGordonwe could easily version the .so
15:31:17kugelLambdaCalculus37: I can make you one if you like
15:31:51JdGordonI'm surprised having to do builds for everyone isnt enough for you to see why the current system is stupid
15:31:56LambdaCalculus37kugel: Sure. :)
15:32:36kugelJdGordon: I'm not proposing to keep the "current system" (there's no current system)
15:32:44LambdaCalculus37JdGordon: But is it worse than, say, building for webOS or iOS?
15:33:11JdGordonLambdaCalculus37: how do you mean?
15:33:17JdGordonkugel: your apks is the current system
15:33:29kugelbut that's no system
15:34:08JdGordonyou're arguing against a system which makes sense and can be implemented *now*
15:34:21*kugel stops discussing with JdGordon for now
15:34:58LinusNfrom a user perspective, the only sane system is to have rockbox in a single apk that is updated via market
15:35:16LambdaCalculus37JdGordon: From what experience I gained from tinkering with iPhone development, Rockbox on the iPhone may be much harder to do.
15:35:32JdGordon*ONLY* if the user wants to stick with builds which we push
15:35:48LinusNJdGordon: yes, a user
15:36:28*LambdaCalculus37 has no idea about how webOS apps are done, and if Rockbox is easy to port to it
15:36:49JdGordonyes fine, but what abuot us who can build our own? you are still forcing us to either use sanctioned apks (either always pushed or from the rockbox site) or build our own apk
15:37:07ZagorJdGordon: devs will always build their own
15:37:11LinusNusers are used to sticking with what the developer pushes at them
15:38:04JdGordonZagor: not if I have to piss around with building the apk and installing that when all I really should need to do is just make && make zip in my android folder like I do with my ipod
15:38:05 Quit markun (Read error: Connection reset by peer)
15:38:24JdGordonI'm not asking for earth shattering ideas... just to be able to load the zip if it is there
15:38:32JdGordon99% of users wont care about this feature
15:38:39JdGordonthe rest will love it
15:38:47Zagorexactly. so we shouldn't force it onto those 99%
15:38:53*kugel is also a bit worried about loading executable code from the sdcard because it's a bit of a security hole
15:38:55JdGordonwe arnt forcing anything
15:39:36ZagorJdGordon: what's wrong with make && make apk?
15:39:40JdGordonZagor: my patch sees /sdcard/rockbox/ is newer than the installed version, it asks "An updated rockbo has been detected, install it"?
15:40:56JdGordonZagor: because its then make && make apk && push to the phone && install it && launch instead of make && make zip && cp /sdcard
15:41:13JdGordonsubtle but big difference
15:41:21 Quit u42p (Quit: Leaving)
15:41:24JdGordonand you dont lose the official updates when they are pushed
15:41:28kugelmake && make apk && adb install -r rockbox.apk is dead simple
15:42:12JdGordonI dont want to have to have the sdk installed on my computer
15:42:42JdGordonyou are over complicatoing the whole thing
15:42:48ZagorI don't think it's reasonable to support android building without the android sdk
15:42:49LinusNJdGordon: if you plan to develop for android, you need the apk
15:42:50JdGordonfor absolutly no benefit
15:43:31ZagorJdGordon: I'd say it's you who is overcomplicating things, by creating a special workaround for android
15:43:40JdGordonLinusN: the apk for all intents and purposes is 100% static. it is just a simple wrapper which loads the rockbox binary. Why does that need to be reinstalled every time I want to update just the binary (unless of course the wrapper needs updating)
15:44:00Zagorhow much time does it take to build the extra apk stuff anyway?
15:44:19kugela few seconds
15:44:27Zagorso the only issue is "I don
15:44:28kugelon my slow laptop :)
15:44:31Zagor't want to download the sdk"?
15:44:33JdGordonit requires a working sdk + downloaded android system + thats not the point
15:44:40LinusNJdGordon: i still don't get it, why is updating the apk such a pain?
15:44:43ZagorJdGordon: what *is* the point?
15:45:05kugelLinusN: actually, updating the apk is the most convinient way to go
15:45:17LinusNkugel: i would think so too
15:45:32kugelbecause putting stuff on the sdcard means you need to mount it on the pc, copy over, unmount again and remount on the phone
15:45:47kugelwhile updating the apk is a single adb command
15:46:03JdGordonadp push /sdcard/rockbox
15:46:04JdGordonthats one command also
15:46:15kugelif you got the sdk yes :)
15:46:21kugelbut you don't want that
15:46:58JdGordonok, so you want the apk to be a full install right?
15:47:03JdGordonthat means it must include the fonts also
15:47:17LambdaCalculus37kugel: Thanks!
15:47:17JdGordonwasted XMB every single download and install
15:47:22JdGordonstupid stupid stupid
15:47:54LinusNJdGordon: download via wifi then
15:48:27LinusNthis is the android world, people are used to apk's and using wifi for downloading large apps
15:48:49JdGordonthose people wont ever use the feature and wont care either way
15:49:18JdGordonexactly.. so why all the angst against adding it?
15:49:39JdGordonit wont affect you AT ALL if you dont use it
15:50:11JdGordonand it makes the build system simpler
15:50:18Zagorwhy do we need two install methods? one official and one hack?
15:50:38Zagorjust because you don't like the sdk?
15:51:02JdGordonthis is EXACTLY the same as using RButil to do the initial install and then downloading the svn builds onto your ipod
15:51:18Zagorno. the svn builds will be apk files.
15:51:20kugelrbutil could be the third :)
15:51:25JdGordonwhen we want to force users to update we push a new apk to the market
15:51:55JdGordonotherwise 99% of users will never update, the other 1% will have no problem downloading a zip
15:52:13Zagorwhy would they never update?
15:52:15JdGordonif we have to maually install an apk we lose the market l;inking which means we dont get forced updates
15:52:17kugeland yours is a hack in that regard that the java wrappers might get incompatible at any time, and then you do need a new apk no matter how you turn it
15:52:41JdGordonZagor: not untill a new apk is pushed
15:52:47 Join markun [0] (~markun@rockbox/developer/markun)
15:52:52JdGordonkugel: thats nonesense. we can EASILY version the wrapper
15:53:02kugelthat doesn't make rockbox run
15:53:09ZagorJdGordon: why would we want users to update to any other version than we push to Market?
15:53:12kugelit just makes it fail with more sanety
15:53:27JdGordonZagor: for the same reason we do svn builds for every target
15:53:39JdGordonkugel: ok, we version the zip
15:53:44ZagorJdGordon: those are for devs
15:53:51kugelthat's already too much hassle IMO
15:53:54JdGordonyou're tyring to bring in a stupid pointless nonesenss arguement for which there are a hundred answer
15:53:59Zagoractually, they are mostly to check the build result
15:54:15JdGordonZagor: so? we should make it harder for us?
15:54:16kugelanother version number to take care of...
15:54:19kugelno, thanks
15:54:37kugelthe apk way is simpler, really
15:54:56ZagorJdGordon: I don't agree that it is any harder to run "make apk-install" or whatever than to mount/copy/unmount/install
15:58:14JdGordonha, no it isnt! go download an apk and try to install it on your phone. it forces you into the market app and then 3 more presses before it gets installed
15:58:28JdGordonthats assuming the key is the same and you dont have to uninstall the old one also
15:58:45Zagor"adb install -r rockbox.apk" does that?
15:59:14JdGordonand if someone uses the svn builds they will just have an apk and no sdk installed
15:59:59Zagorwho do you expect to be shocked and dissappointed that we supply our android builds as apk files?
16:00:20JdGordonnoone, I'm saying producing apks for this is stupid
16:00:45Zagoryes, you keep saying that. and I keep asking what specifically is so stupid.
16:02:38kugelI think it's sane to require the sdk if anyone wants to make rockbox-android development
16:02:47kugelbtw, the ndk requires the sdk
16:03:37JdGordonno it doesnt
16:04:22kugel"It is designed for use only in conjunction with the Android SDK, so if you have not already installed the latest Android SDK, please do so before downloading the NDK."
16:04:41JdGordonthat doesnt mean it wont compile if it isnt installed
16:05:02 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
16:06:40JdGordonand people doing their own builds arnt the only people who use them
16:07:07kugelthey can distribute apks just fine, even on the market
16:07:07 Quit markun (Read error: Connection reset by peer)
16:07:54JdGordongreat, so flood the market with multiple unoffical rockbox builds... yay support nightmare
16:08:00skotake a look at the system requirements of the ndk: "A comlete Android SDK installation (with all dependencies) is required."
16:08:34kugelwe don't support unofficial builds anyway
16:08:44JdGordonusers wont know the difference
16:08:45kugeland I don't care if they floot the market or our forums, really
16:08:54kugelthat has never been a problem so far
16:09:35 Quit Ramsey[LC] (Quit: (Killed (NickServ (GHOST command used by dissident))))
16:09:58kugelI think we'll have a lot more support problems if we distribute rockbox.zips that are incompatible with the already installed java wrappers
16:10:16kugeland no, versioning the .so or the .zip is not worth it
16:10:30JdGordonyou are wrong
16:11:04JdGordonversioning the .so takes about 3 lines of c and 5 in java to make the call and display an error if it isnt compatable
16:11:11JdGordonand whent hat happens, oh my, we push a new apk
16:11:29ZagorJdGordon: again, what is the advantage?
16:11:36kugelthat's more complicated than shipping the apk in the first place
16:12:23JdGordonZagor: I've outlined them in here and the email. obviously you dont see them as advantages and you wont be able to be convinced
16:13:45JdGordonone that might have gone unnoticed is that you dont need to rework the build system to make it work with apks
16:14:09kugelexcept we don't need to rework it
16:14:11Zagorwe still want to add apk building to the build system
16:14:36JdGordonnot really
16:14:56Zagorwhy wouldn't we?
16:15:24JdGordonbecause it would only need to be rebuilt when the java changes, and that person can easily rebuild the apk
16:15:34JdGordonno different to how bootloaders are released
16:15:47Zagorumm, our build system does bootloaders
16:15:51JdGordonor we do a release, again, one exra steo in the release checklist
16:16:14JdGordonI meant the uploading part
16:16:51Zagorwell sure, we don't want anyone to upload official rockbox builds to Market
16:17:00kugelthat apk step is needed in the release checklist no matter of your suggestion
16:17:01Zagor*just anyway
16:17:11Zagor*anyone :)
16:17:29ZagorI still want "make apk" to work
16:17:48 Join markun [0] (~markun@rockbox/developer/markun)
16:17:51 Quit markun (Read error: Connection reset by peer)
16:18:17JdGordonregular users will only use the release builds (3.6, 3.7, etc). they will be pushed as normal to the market.... my suggestion doesnt change that
16:18:45JdGordonAll im saying is that apk can then load a regular zip so noone else needs to stuff around with apks
16:18:53JdGordonmost users wont know or care about that
16:19:13JdGordonthose that use the svn builds might find that a nice feature
16:19:27kugelthe java wrapper may break at any time. I want not only the problems with incompatibilities when installing but I also want the build system to catch compilation failure if it happens
16:19:27ZagorI would claim most users much rather would grab their daily rockbox build from market than download a zip and fiddle with disk mounting and copying
16:19:30JdGordonthere is no reason to not do that
16:19:57JdGordonkugel: im ignoring that argument, it is stupidly simple to work around
16:20:13JdGordonZagor: And then they would be annoyed every single day with an update request
16:20:24 Quit Ronald (Quit: Leaving)
16:20:42kugelthey only get the notification once, if they don't chose to update it won't annoy anymore
16:20:49JdGordonwe could do that also
16:20:51ZagorJdGordon: you say annoyed, I say "notified". if they install the daily build rockbox, they expect to get a daily build
16:21:13JdGordonI dont want a notification every day that a new build is ready
16:21:21JdGordonI'm positive I'm not alone there
16:21:30Zagorthen disable the auto notification
16:21:33kugelthen don't install from the market
16:21:37kugelor that
16:21:38Zagoror that
16:22:33JdGordonbugger it, im doing my own thing for android then.
16:22:44 Quit bmbl (Quit: Bye!)
16:23:00ZagorJdGordon: I seriously don't understand why this is so important to you
16:23:02 Join markun [0] (~markun@rockbox/developer/markun)
16:23:14JdGordonyou both havn given a single reason why the idea is bad
16:23:16JdGordonditto to you
16:24:07Zagorthere is a standard method to handle upgrades in android. I propose we use that, to avoid confusing people and adding extra code.
16:24:44JdGordonand the people that care abouit that wont ever know about or use this feature so thats a moot point
16:25:09gevaertsIf you package an application, you need *very* good reasons to bypass the standard methods associated with the package format you use
16:25:23kugelit would probably most confusing if they come to our website and find that we handle current build installations *completely* different
16:25:26evilnick_BIt wouldn't be bypassing them, it's an alternate way to update
16:25:36ZagorJdGordon: so we should add this duplicated upgrade method just because you think it's nicer?
16:25:55kugelbecause he doesn't like to install the sdk
16:26:02JdGordonwe dont add any code to to android apk updates at all so no it wouldnt be duplicate code
16:26:16JdGordonit *is* nicer
16:26:20ZagorJdGordon: duplicated functionality
16:26:30JdGordonand kugel you're a twat when you want to be....
16:26:40gevaertsAlso, the java bits don't change *today*, but as soon as the first native widget gets in...
16:27:08JdGordonZagor: 3 lines of java code. the unzip code is alll there in svn now
16:28:37JdGordongevaerts: the java bits are more likely to change *today* than later.. and I've already said it is simple to version the binary so it cant load incompatable builds
16:29:05 Quit user890104 (Ping timeout: 248 seconds)
16:29:08JdGordonand it would make it easier for others to make new apps which hook into the java to provide widgets or whatever if it was seperate
16:29:26*kugel fails to see the advantage of "versioning zips so it won't work if it can't work" over "ship apks that always work without need for a versioning system"
16:29:46JdGordongood for you
16:30:00JdGordonits called having a closed mind
16:30:17gevaertsWhat is? Your attitude?
16:30:53 Nick [J] is now known as DarkJuju (~juju2143@gateway/shell/
16:30:59kugelno need to insult now
16:33:57 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
16:33:58 Quit markun (Read error: Connection reset by peer)
16:36:24LambdaCalculus37Now that Rockbox works on touchscreen targets, shouldn't we consider creating a proper touchscreen interface for those targets?
16:36:39ZagorLambdaCalculus37: I think so too
16:36:50Zagorbut it's a pretty big task
16:36:57kugelI think a lot of people have considered it :)
16:37:00LambdaCalculus37Zagor: Indeed.
16:38:01 Quit JdGordon (Ping timeout: 252 seconds)
16:38:06 Join user890104 [0] (
16:38:12LambdaCalculus37Zagor: The UI we have now is great for non-touchscreen targets, but an icon-based UI for my Android phone would be dandy.
16:38:28Zagorstart drawing :)
16:38:57*LambdaCalculus37 brings forth the grimoire :)
16:39:44LambdaCalculus37I can make a very crude mockup.
16:40:04kugelamiconn: rockbox will function without sdcard, but it won't be able to save settings etc
16:40:20amiconnWhy is that?
16:40:22kugelthe alternative is having no access to config.cfg
16:40:32amiconnDo android devices not have some fat(32) formatted internal storage?
16:40:34kugel(no user access)
16:40:59amiconnHow does rockbox access music stored internally?
16:41:27kugelthey have yaffs or something for internal storage, and apps have almost no access to it (if not rooted anyway)
16:41:53kugelthere's only a little folder in your apps dir which you cannot reach with a file browser or when connected to the pc
16:42:24kugelamiconn: with a file browser, if apps have permissions to open that folder
16:42:55amiconnHmm, raa uses native file system access, correct?
16:43:15*amiconn thinks raaa on n900 will be easier
16:43:44kugelit will be the same as the sdl app I guess, which has similar "problems"
16:43:53kugelbut there you have $HOME, not /sdcard
16:45:32LinusNi think much of the current rockbox UI has to go, in favor of standard android menus
16:46:07ZagorLinusN: are there really "standard android menus"? it seems to me most apps have different menus
16:46:47Zagoror is the homescreen settings menu the native model?
16:47:01LinusNi believe so
16:47:21sko there is a design guideline at
16:47:43LinusNi mean, i would use rockbox on android mostly for the nice playback features
16:48:10Zagorthose menus look like they match our needs pretty well
16:48:39Zagorbut I assume that means more mixed native/java code?
16:48:49LinusNi'm sure
16:48:59LambdaCalculus37Probably so.
16:49:21skohere it is:
16:49:26LinusNalso, most of the plugins are pretty useless as well
16:50:10kugel1 reason RaaA has no plugins yet :)
16:50:47LinusNraaa is more of a Media Player, so most of the "operating system" stuff is simply not needed
16:51:12skowell... a native ui would solve the "different screen sizes"-problem
16:51:48rasherexcept for the wps
16:52:36kugelI actually think trying to replace rockbox widgets with native ones is a neverending job
16:53:18kugelIt would probably be more easier to fork out the playback stuff and build a dedicated app around it
16:53:31LinusNmy approach would be to rip out the core stuff (playback, dsp, codecs and wps) and reimplement the gui
16:53:33***Saving seen data "./dancer.seen"
16:53:35 Join markun [0] (~markun@rockbox/developer/markun)
16:53:40LinusNkugel: :-)
16:53:45Zagorit's pretty much unavoidable since they code would diverge so much anyway
16:54:26rasherMore librockbox?
16:54:33kugeland I say fork because I personally don't really want to miss the theming possibilities, and I find the current touchscreen navigation works well enough
16:55:19kugel(absoute mode enabled of course)
16:55:20Zagorrasher: sure sounds like it
16:56:12kugelLinusN: I don't think you could keep the wps :(
16:56:13 Quit markun (Read error: Connection reset by peer)
16:56:34LinusNkugel: perhaps not in its current form
16:57:08Zagorwhy not? wps files are already resolution specific
16:57:53kugelI mean the skin engine and its drawing mainly, not the .wps files
16:58:28TheSevenwell, audio-related plugins (like fft) would actually make sense even on raaa
16:58:32Zagorstill, the wps is just a canvas that we draw to. why wouldn't that work on android?
16:59:05Zagornot the non-wps theming, of course
16:59:35kugelthe plan was to reimplement the ui in java, wasn't it? :)
16:59:40skohmm... as its possible to port rb to different platforms through raaa a separation of logic and ui would be useful? (don't know if its possible at all)
17:00:04kugelbut it may be possible, you're right
17:00:36Zagorsko: separation of logic and ui might be complex. separation of ui and playback is more feasible.
17:02:07 Part LinusN
17:04:25kugelZagor: I think that once you see it nice works when you have a proper cabbie port, then you'll realize that it's not worth the effort :)
17:05:34kugelI already like rockbox more because I have more controls from the wps compared to the intergrated media player, allowing me to change songs or playback related settings more quickly
17:06:06Zagorkugel: of course, but that doesn't go away with a proper touchscreen ui
17:06:39kugeldepends on what you mean with proper touchscreen ui
17:06:50kugelif it means native widgets I'm not so sure
17:07:48Zagorwell, having to memorize which unmarked area of the screen does what is a pretty crappy user interface
17:08:03kugelthat's what I mean with proper cabbie port
17:08:28kugelthe 320x480 one has pretty obvious touch regions which you can touch directly (no grid mode9
17:08:29Zagorwell the problem is everywhere, not just in the wps
17:09:33 Part Zagor
17:09:50kugelI'd agree that grid mode is crappy, but the absolute point mode works well most of the time
17:11:36 Join r0b- [0] (
17:12:41 Quit sko (Read error: Connection reset by peer)
17:14:02 Join sko [0] (~AndChat@
17:14:16 Quit sko (Client Quit)
17:15:48*kugel wonders if people have seen the screen shots I posted ( and
17:17:37Stummicould it be, that the most of the rockbox-community are germans? *g*
17:22:52 Join markun [0] (~markun@rockbox/developer/markun)
17:23:26 Quit antil33t (Read error: Connection reset by peer)
17:23:36 Join antil33t [0] (
17:23:53wodzgevaerts: ping
17:24:07wodzpamaury: ping
17:26:07 Join saratoga_ [0] (9803c57f@gateway/web/freenode/ip.
17:26:46saratoga_kugel: for what its worth, i think changing the menus to use icons but keeping the WPS would be pretty slick
17:26:57pamaurywodz: yes,, I'm sorry for the mrobe, I forgot it at home :( I'm the only one to blame
17:27:04saratoga_not sure if theres enough preprocessor magic on earth to make that work though
17:27:24wodzpamaury: ok
17:27:35saratoga_maybe not icons, but larger GUI elements anyway for easier clicking
17:27:55kugelpick a larger font :)
17:28:45 Quit z35 (Ping timeout: 265 seconds)
17:29:25pamaurywodz: I'm quite under pressure this week because this is the end of my internship but I promise I'll test it tonight, I've been rude by forgetting this
17:30:04wodzpamaury: ok, I'll post updated version of the patch to the tracker soon.
17:30:18pamauryok, just poke so I download it before going offline
17:34:12kugelperhaps the yesno screen could be relatively easy to do with a native dialog
17:34:13 Quit markun (Read error: Connection reset by peer)
17:34:35kugelbut I'm unsure if it's a good idea to use native widgets only partly
17:34:53kugelI think they'll look like another app pops them up, not rockbox
17:37:41 Join Kitr88 [0] (
17:39:02 Join toffe82 [0] (~chatzilla@
17:41:06 Quit Kitar|st (Ping timeout: 240 seconds)
17:42:50gevaertswodz: pong
17:44:36 Join Ramsey[LC] [0] (~RamseyLC]@
17:46:08*rasher would preferably see native menus and rockbox wps
17:55:12saratoga_bertrik: were you going to make vorbis flush the malloc buffer on track change?
17:55:31saratoga_IMO thats worth doing as well as fixing that leak
17:56:42 Quit wodz (Ping timeout: 240 seconds)
18:00:44 Quit saratoga_ (Quit: Page closed)
18:03:11 Join sko [0] (
18:04:42 Join komputes [0] (~komputes@ubuntu/member/komputes)
18:05:07 Join Stephen___ [0] (
18:05:58Stephen___S_a_i_n_t: ping
18:10:14 Quit petur (Quit: Page closed)
18:12:13Stephen___I'm abotu to update customwps wiki page wondering if anyone could have a glance over to check for mistakes? I know S_a_i_n_t had this on his todo. I had a few mins so thought i'd save him some trouble.
18:12:45 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
18:13:19bertriksaratoga, I think I found out where the memory leak in the vorbis codec occurred
18:13:34pixelmaStephen___: do you have a pastebin or so?
18:14:03Stephen___I have the editor still open I could copy to pastebin ?
18:14:09pixelmaor maybe just do the changes and people can look at it
18:14:27bertriksaratoga, an attempt was made to allocate a PCM buffer from IRAM, when that fails, it is allocated with ogg_malloc (and never freed)
18:14:34Stephen___ok cool. I'll submit it now.
18:15:30 Join stoffel [0] (
18:15:49bertrikit looks like it only happens with the specific 48 kbps ogg vorbis file and not with my own vorbis files (-q5, around 150 kbps)
18:16:32bertrikReinitialising the ogg_malloc system for every file seems to fix it
18:17:22 Join [Saint] [0] (S_a_i_n_t@
18:17:40bertrikI think it's not so nice that the player just hangs with a data abort when running out of memory, IMO a panic indicating the cause would be nicer
18:18:54Stephen___Saint the very person
18:18:57 Quit S_a_i_n_t (Ping timeout: 248 seconds)
18:24:15pixelmaStephen___: looks ok as far as I can see. The only thing I'd add (but could do so myself) is the %( and %) in the "other tag" section
18:24:31pixelmaI wonder if a comma needs escaping too now?
18:24:39 Quit DerPapst (Quit: Leaving.)
18:25:40Stephen___Commas do need escping, what part is that ? do you want me to change the other tag section whilee I'm at it ?
18:26:24Stephen___I was a bit unsure about the playlist viewer part, teh syntax of it now
18:28:28pixelmaI don't know that part well, can't help you there
18:29:05Stephen___i'll wait till S_a_i_n_t or JDGordon is around and pester them.
18:30:15Stephen___anyway back to work for now, thanks for the help.
18:30:19pixelmaI guess the example.wps is outdated now. Should it be kept? Updated?
18:30:21 Quit Stephen___ (Quit: CGI:IRC 0.5.9 (2006/06/06))
18:30:28 Join wodz [0] (
18:30:41wodzgevaerts: Thanks for testing. Gigabeat F seem to have something called HARDWARE_BEEP - does it work with patch?
18:30:51*kugel has test_fps running with a quick hack
18:34:51 Join GeekShad0w [0] (
18:35:01 Quit GeekShad0w (Read error: Connection reset by peer)
18:35:39wodzAbout jdG idea of loadable - this is serious security flaw. Even if user IS asked
18:37:26wodzpamaury: I've uploaded new version of the patch to the FS #11531
18:38:10 Quit LambdaCalculus37 (Quit: Bye)
18:38:34 Quit GeekShadow (Ping timeout: 276 seconds)
18:39:49 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
18:45:04gevaertswodz: the new patch?
18:45:55wodzgevaerts: I've uploaded slightly changed version but I was asking You about v01
18:46:04gevaertsHow would I know?
18:46:14gevaertsWhat does this HARDWARE_BEEP thing do?
18:46:38wodzI was thinking that You know as gigabeat f user :-)
18:46:45wodzI'll check in manual
18:52:27wodzgevaerts: hmm it looks like it disables rockbox's keyclick framework
18:53:02wodzdoes gigabeat f have keyclick?
18:53:34***Saving seen data "./dancer.seen"
18:54:01 Quit hebz0rl (Ping timeout: 260 seconds)
18:54:16*gevaerts never used keyclick
18:55:14gevaertswodz: if I enable keyclick with your patch, I hear clicks when I press keys. Is that what you need to know?
18:56:01kugelok, we currently have 271fps
18:56:02wodzhmm looks like this is long forgotten ifdef
18:56:23kugelbut only because they happen async. synchronizing it brings it down to 64fps
18:58:20 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
18:58:50 Quit t0rc (Remote host closed the connection)
18:59:09kugelwodz: I thought so too (re security flaw). I could put a which only implements Java_org_rockbox_RockboxService_main(), then the phone would be yours (well, still limited to native code but still)
19:00:23kugelthe phone would be mine of course :)
19:00:30 Quit GeekShadow (Quit: The cake is a lie !)
19:02:28 Quit antil33t (Read error: Connection reset by peer)
19:02:34 Join antil33t [0] (
19:02:52wodzI read somewhere about some worm targeting android presenting itself as a media player so this is not so unlikely scenario to use to attack phone
19:03:33 Quit T44 (Ping timeout: 265 seconds)
19:03:47kugelyou cannot do that with native code IIUC
19:03:56 Join DerPapst [0] (
19:04:27kugellike sending expensive sms
19:04:50wodzI see but DoS is possible
19:06:32 Join Topy44 [0] (
19:08:12 Quit DerPapst (Ping timeout: 240 seconds)
19:12:45 Join DerPapst [0] (
19:14:41wodzgevaerts: It looks like one of the timer can be used to generate beeps ( but I don't see this to be used in rb
19:20:27wodzgevaerts: aaa I see it IS used. If You can hear keyclicks that mens everything is ok
19:22:02bertrikThis is the fix I'm planning to commit for the vorbis malloc problem: it basically causes the ogg_malloc system to be reinitialised for each file when playing multiple vorbis in sequence (as opposed to resuming from the previous state earlier)
19:24:18wodzbertrik: what about the leak You have revealed?
19:24:44bertrikthat leak is worked around with this patch
19:25:05wodzbut isn't it better to fix the leak?
19:27:16kugelbertrik: don't feel like freeing the leak?
19:27:21bertrikthat's quite complicated, we have basically two allocators for ogg, a fancy TLSF one for main DRAM and a completely naive (without a free) one for IRAM
19:28:27bertrikThe leak occurs when ogg tries to malloc from IRAM using the naive allocator and finds it has no memory, it then does a malloc from main DRAM, this malloc is never freed.
19:28:58bertrikkugel, there's a hack in place now that I rather don't touch
19:29:15kugelso I guess it should be reinitialized anyway, to reclaim iram
19:31:33gevaertsI think it should be reinitialised, even if the other issues are fixed
19:31:37kugelbut I wonder if it's a good idea to have a sort-of-heap in iram
19:31:49gevaertsIt costs nearly nothing, and it's a good safeguard against other bugs
19:32:26wodzreinitialization is ok but why leave known leak? If some day someone implement more sophisticated allocator for iram and removes reinitialization the issue will popup again
19:32:49bertrikas far as I can see, the naive IRAM allocator is already re-initialised for every file, the DRAM allocator is not yet
19:36:05wodzbertrik: I would still add free() (if you know where leak is). Reinitialization maybe still good idea to prevent fragmentation and friends
19:38:11bertrikwodz, the problem is that there is not really a point in the code where deinitialisation is done already (as far as I can see), so I can't just add a "missing free".
19:38:34 Join stripwax [0] (
19:38:47 Quit stripwax (Client Quit)
19:38:47 Join Buckwheat [0] (~chatzilla@
19:39:08BuckwheatCould someone link me to the svn for rockbox
19:41:03 Quit Buckwheat (Client Quit)
19:42:48 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
19:46:07 Quit pamaury (Ping timeout: 246 seconds)
19:46:55bertrikwow, we have only 20.5 kB IRAM available as IRAM for codecs !?
19:47:02bertrik(on the clipv1)
19:47:26kugelbertrik: we should have 0 iram available
19:47:46kugelthe entire codec buffer is in iram, 32k resevered for the core
19:48:49 Join LambdaCalculus37 [0] (~LambdaCal@rockbox/staff/LambdaCalculus37)
19:50:26bertrikkugel, the IRAM allocator uses static unsigned char iram_buff[IRAM_IBSS_SIZE] IBSS_ATTR __attribute__ ((aligned (16))); as its pool (from oggmalloc.c)
19:50:51bertrikthe 20.5 kB is what I got from the clipv1 simulator by the way, not sure if this is representative for the target
19:50:55kugelconfig.h does #undef IBSS_ATTR IIRC
19:54:12 Quit dfkt (Read error: Connection reset by peer)
20:06:59 Join Xerion [0] (~xerion@
20:07:12 Join bunnyboi [0] (
20:07:26amiconnbertrik: Regarding running out of ram mid-playback - wouldn't it be better to just fail instead of panic?
20:24:25 Join liar [0] (
20:34:03 Quit fyrestorm (Quit: Ur skills' fireproof like a wooden panel -- U got feds talking leet on your IRC channel!)
20:40:54bertrikamiconn, it ran out of ram just after loading the next ogg vorbis file in a sequence. Maybe just bailing out instead of panicing is better indeed (anything else than crashing really)
20:43:35bertrikhm, there already seems to be something in place to handle out-of-memory, doing a longjmp
20:48:49 Join Lear [0] (chatzilla@rockbox/developer/lear)
20:48:50bertrikI think I'll put my basic patch on the tracker and let someone with longjmp/setjmp knowledge look at it
20:49:18*bertrik wonders if this longjmp/setjmp stuff in vorbis was ever tested
20:49:25LearSure it was.
20:50:55 Quit stoffel (Remote host closed the connection)
20:51:06LearBut that was some time ago (before TLSF and the "dynamic" IRAM changes too, I think).
20:52:57kugelLear: I seem to remember that you implemented the longjmp to recover (skip track) from a bad malloc
20:53:34kugelbut I don't know whether it was the old malloc or tlsf-malloc
20:53:35***Saving seen data "./dancer.seen"
20:55:16LearYep (with "bad" being "too big"). For the old malloc. I did tlsf later.
20:55:17 Quit lestatar (Ping timeout: 250 seconds)
20:55:20 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
20:55:20 Quit amiconn (Write error: Broken pipe)
20:55:35 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
21:01:36 Join Ludo6431 [0] (
21:01:52CIA-9New commit by learman (r27773): Fix FS #11175. playlist_peek() wasn't thread safe (due to a static filename buffer), so frequent calls from the main thread would cause the audio ...
21:04:00CIA-9r27773 build result: All green
21:05:21LambdaCalculus37kugel: Have you seen this?
21:07:55 Join Jennifur [0] (
21:10:42kugelLambdaCalculus37: yes, error 404 is not new to me
21:10:55 Quit bunnyboi (Ping timeout: 276 seconds)
21:11:02kugelgevaerts: did you sent in my evaluation already?
21:11:15kugel(if there is one for you?)
21:12:11gevaertskugel: not yet. As far as I know, that starts next Monday
21:12:15 Quit Dreamxtreme (Quit: There are two major products that come out of Berkeley: LSD and UNIX. We don't believe this to be a coincidence.)
21:12:31 Join Dreamxtreme [0] (Dreamxtrem@
21:12:46kugelI just wanted to say that I want still to write 2 wiki pages, one about the android port specifically and one about porting RaaA in general (what's needed at minimum, what rockbox expects, etc)
21:12:59gevaertsGood idea
21:13:19gevaertsThat also fits nicely into the gsoc soft pencils down time
21:13:32skokugel: try this link:
21:13:36kugeloh yes, final evaluations are *from* august 16, I read until
21:15:40 Quit kugel (Remote host closed the connection)
21:19:22 Quit Dreamxtreme (Read error: Connection reset by peer)
21:27:33 Quit sko (Quit: Leaving.)
21:33:18 Join Dreamxtreme [0] (Dreamxtrem@
21:36:27 Quit Jennifur (Quit: Ex-Chat)
21:41:15 Join ved [0] (
21:42:15*gevaerts congratulates Lear for fixing the bug with the weirdest effect of the last year
21:42:37 Join cd [0] (
21:46:43cdkugel: I tried to use your latest apk on Galaxy S but could not start music. Error accessing playlist control file! Do you have a hint for me?
21:46:47 Join bunnyboi [0] (
21:51:54wodzCan we close FS #7138? I think this has been commited long ago
22:04:43 Quit parafin (Ping timeout: 246 seconds)
22:05:14 Join petur [0] (~petur@rockbox/developer/petur)
22:06:26 Quit LambdaCalculus37 (Quit: Bye)
22:06:33 Join parafin [0] (
22:10:41 Quit cd (Remote host closed the connection)
22:22:32 Join Dhraakellian [0] (
22:27:17 Quit wodz (Quit: Leaving)
22:27:22 Join Kiddyshaq [0] (
22:29:42 Quit bunnyboi (Ping timeout: 260 seconds)
22:30:59 Part Kiddyshaq
22:32:03 Join bunnyboi [0] (
22:34:19 Quit advcomp2019 (Ping timeout: 258 seconds)
22:34:34 Quit Lear (Quit: ChatZilla 0.9.86 [Firefox 4.0b4pre/20100810030848])
22:36:41 Join Kiddyshaq [0] (
22:36:47 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
22:43:11pixelmawelcome. If you have a Rockbox related question just ask
22:44:52KiddyshaqSorry if you already know this but I wonder if there is a Zune port
22:47:31Bagderthere is none
22:47:41Bagderyou can start it!
22:48:08 Quit petur (Quit: here today, gone tomorrow)
22:49:21Kiddyshaqwhat about ipod nano 2g?
22:50:55BagderI believe rockbox works rather good on it
22:51:11Bagder(as linked from the front page)
22:53:37***Saving seen data "./dancer.seen"
22:56:18Kiddyshaqwhat does "Stkov nand" mean?
23:00:38 Quit Kitr88 ()
23:02:36KiddyshaqBye ;)
23:02:43 Part Kiddyshaq ("Leaving")
23:02:51 Join robin0800 [0] (
23:04:42 Join Brandonrd7 [0] (
23:04:48 Part Brandonrd7
23:11:42bertrikI've put my analysis of the ogg vorbis malloc problem on clip v1 on the tracker:
23:12:16 Join Kitar|st [0] (
23:32:26 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
23:34:01 Join stripwax [0] (
23:34:14 Quit robin0800 (Read error: Connection timed out)
23:35:07 Join robin0800 [0] (
23:38:13 Join shai_ [0] (
23:38:29 Quit shai (Read error: Connection reset by peer)
23:46:16 Quit krazykit (Quit: bbl)
23:46:57 Quit tchan (Ping timeout: 246 seconds)
23:48:21 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
23:50:37 Quit xvello (*.net *.split)
23:50:38 Quit xavieran (*.net *.split)
23:52:41saratogabertrik: in the meantime can you change the codec to reinit the malloc buffer on track change?
23:54:12 Join xvello [0] (
23:57:06stripwaxhrm, the tlsf buffer should be getting freed.
23:57:22stripwax[and/or only being alloc'd once]
23:58:00stripwaxI seem to recall n1s had a patch for simplifying the pcm buffer alloc (which I still never got a chance to take a look at) but wondering if it addresses some of this

Previous day | Next day