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 2007-10-12

00:00:08 Join todko1 [0] (
00:00:08 Quit Kayasha-kun (Client Quit)
00:00:13 Join karl [0] (i=46518fe6@gateway/web/cgi-irc/
00:00:15alienbiker99do you guys have a list of links that you have at the ready? haha
00:00:15amiconnlinuxstb: Okay, but it works?
00:00:21karlcould i re have thiose link
00:00:29zajacattackhey, does anyone know what happened to mpegplayer?
00:00:31linuxstbamiconn: I only tested briefly, but yes. Should I test more?
00:00:32karlpop up blocker,
00:00:37karlit's kayasha here
00:00:37amiconnThe speedup might be tiny because the single-transfer port is only used for commands
00:00:52todko1hmm, everytime I try to mount my ipod recently, it just says "Okay to disconnect" so I have to boot into the iPod firmware to mount my ipod >_<
00:00:55amiconnThe bulk data transfer uses the block transfer port
00:01:13bluebrotheralienbiker99: it's rather easy to construct the link to the wiki if you know the page name ;-)
00:01:17 Join Isolinear [0] (
00:01:22amiconnH10 doesn't use that (yet), and that extra wait would cost around 20% speed
00:01:35zajacattacki updated last night to get seeking, and now my vids don't play
00:01:38amiconn(H10 svn doesn't have that wait, ipod color/nano does)
00:02:11karlmpeg 2 movie ?
00:02:14karlcould that work?
00:02:26amiconnlinuxstb: Basically it should not hang or cause graphical glitches. Testing with something gfx intensive would be best
00:02:43amiconn(mandelbrot, plasma, mpegplayer etc. Also test_fps)
00:02:45linuxstbkarl: It needs to be encoded according to the instructions on the page we linked you to.
00:02:52karlok perfect
00:03:00amiconnAny volunteers for testing on nano? preglow?
00:04:04 Quit Gnu47 (Connection timed out)
00:05:00 Quit _aLF ("Quitte")
00:05:02 Quit midkay ("Leaving")
00:08:05preglowamiconn: sure,what is it?">
00:08:58amiconnPatches the PP lcd drivers using the "color" bridge
00:08:58preglowwhat should i do?
00:09:07amiconnNeeds testing for speed and stability
00:09:43amiconnYes, plus watching for gfx glitches, ideally using several gfx intensive plugins
00:10:07 Part zajacattack
00:11:49 Nick fxb is now known as fxb__ (
00:12:34preglowman, make is starting to take quite some time on my poor old box
00:12:38 Join iamben [0] (
00:14:19linuxstbamiconn: No problems that I can see.
00:14:24 Join roolku [0] (
00:14:26amiconnNice :)
00:14:39 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
00:14:47Nico_Pamiconn: 1624 bytes less for the mob branch compared to the latest SVN
00:15:04amiconnNow if the block data port also works on H10... I'll hopefully see soon....
00:15:16linuxstbroolku: I've just tried the "new" mpegplayer for the first time. I'm not sure I like the way it takes me to a menu, rather than playing the video...
00:15:20preglownow, how the hell to increase boost ratio manually on ipods
00:15:34preglowlinuxstb: you shouldn't like that...
00:15:44amiconnYou have to scroll *left*
00:15:49amiconnIntuitive, eh?
00:16:08preglowi would expect left and right _buttons_
00:16:43roolkulinuxstb: isn't this the same with the archos video player? How else could it be done?
00:16:49 Quit n1s ()
00:17:18linuxstbroolku: Well, if there's a resume point for that video stored, it could resume it, otherwise just play from the beginning.
00:17:24linuxstb(or ask if you want to resume)
00:17:33amiconnvideo.rock asks
00:17:50roolkulinuxstb: but then you wouldn't be able to seek
00:17:57amiconn..and if no resume position is stored, plays right away
00:18:08preglowtest_fps: 80, 316.5 @ 30. 213.5, 848 @ 80. unchanged after patch
00:18:31linuxstbroolku: I would expect to be able to seek from within mpegplayer, after the video has started playing.
00:19:02linuxstb(using the normal left/right buttons)
00:19:27preglowamiconn: isn't this supposed to affect fps?
00:19:37*amiconn got a >3 times speedup in lcd update speed on H10 :D
00:19:51roolkulinuxstb: well, that was the discussion we had the last time - it doesn't seek as such, it sets the resume point
00:19:59amiconnpreglow: It's probably an immeasurable small speedup
00:20:11roolkulinuxstb: which can only be done before playback is started
00:20:15preglowamiconn: nano is already pretty fast, so
00:20:25amiconn113fps @30MHz, and 303fps @80MHz
00:20:32amiconn(that's fullscreen!)
00:20:41amiconn(small H10 though)
00:20:41 Part todko1
00:20:54linuxstbroolku: I know, but couldn't you just implement seeking by stopping playback and resuming?
00:21:05 Join SirFunk [0] (
00:21:11amiconnPlasma flies...
00:21:28*preglow wants to see cube.c with gourad shading :>
00:21:40preglowand perspective corrected texture mapping
00:21:59[IDC]Dragonand filtered textures
00:22:00roolkulinuxstb: yes, that is on the todo list, but flushing all the buffers cleanly doesn't strike me as easy
00:22:01Arathisamiconn: need someone with a big H10? ;)
00:22:03linuxstbroolku: I understand it's a work-in-progress, it's just a horrible UI at the momment.
00:22:11amiconnArathis: Not yet, but soon
00:22:26amiconnThe big H10 driver is separate, so I have to port the stuff first
00:22:27preglowamiconn: looks stable
00:22:32amiconnpreglow: goodie
00:22:46preglowamiconn: bounced through all the stuff in demos/ plus a few games
00:23:00amiconnFire is faster than on any other target I own...
00:23:14preglowamiconn: what reso is h10?
00:23:22linuxstbAbout 10x10...
00:23:23*amiconn guesses the H10 won't have any problems with peakmeters soon
00:23:30amiconnpreglow: small one is 128*128
00:23:30 Join AlexGladd [0] (
00:23:36preglowahh, ok
00:23:46preglowso transfer speed is about as fast as nano
00:23:49 Part AlexGladd
00:24:06 Quit davina (Remote closed the connection)
00:24:09amiconnpreglow: yup
00:24:13roolkulinuxstb: considering the slowness of the seeking I think it is quite useful, but I realise that there seems to be a demand for seeking after playing has started...
00:24:22 Join Fhantazm [0] (
00:24:29amiconnpreglow: Almost exactly, in fact
00:24:43amiconnHmm, but that means something interesting...
00:24:52preglowdoes it?
00:25:26amiconnThe iPod color should be twice as fast as it actually is...
00:25:28linuxstbroolku: If seeking is going to be this slow in mpeg, maybe we should consider a different container format - one that is more suited to seeking on our devices.
00:25:48 Join billytwowilly [0] (
00:25:57preglowis the ordinary mpeg one really primitive?
00:26:16 Quit chris__ (No route to host)
00:26:28amiconnpreglow, linuxstb: Could you make a dump of the 0x70008000...0x70008fff register area (on nano resp. color)?
00:26:52FhantazmSansa e260 user here and would like to offer myself as a tester.
00:26:58roolkulinuxstb: personally I am very happy with how it is now. I never suddenly get the urge to seek, but I often want to resume halfway into a video after having had to stop
00:27:23FhantazmAbsolutely LOVE whats been done on the Sansa port so far and would like to lend a hand to the developers any way possible
00:27:52roolkulinuxstb: I think introducing new and probably more exotic formats will make it harder for the average user to convert their movies
00:27:59rasherWhy does the manual use customization rather than customisation?
00:28:01amiconnThe iPod color should manage ~45fps @30MHz, and ~125fps @80MHz. There must be some speed config register...
00:28:12preglowamiconn: what, just dump the memory contents to a file?
00:28:23 Join Entames__ [0] (
00:28:37 Quit Entasis_ (Read error: 104 (Connection reset by peer))
00:28:39 Join qweru [0] (
00:28:55linuxstbroolku: I agree that's an issue, but IMO performance is important too...
00:29:10amiconnMaybe the color uses a conservative setup for the type 0 LCD, and the type 1 could go faster?
00:30:31 Join adiamas [0] (n=adiamas@
00:30:42pixelmarasher: I thought there was finally a decision to use -ize (as that can be BE too) and I also wouldn't be surprised to find both spellings in the manual
00:31:00roolkulinuxstb: probably for some - I am quite happy with the way it is. I can see possible improvements with the menus though
00:31:20amiconnThat menu is really annoying
00:31:37roolkulinuxstb: i.e. interrupt playmode to start seeking
00:31:48rasherpixelma: Ah, okay. I didn't know there had been any discussion, and also always thought that BE used -ise rather than -ize.
00:32:46roolkuanyway. let me get the fixes for jhMikes' dodgy video files committed :)
00:32:47amiconnroolku: video.rock allows seeking while displaying the video frame. Playback is paused when you start seeking. A seek bar is displayed at the bottom. When you stop seeking, playback continues at the new positioin
00:33:01 Join Toxicity999 [0] (n=bryan@unaffiliated/Toxicity999)
00:33:37linuxstbpreglow: I'm quite busy ATM, would you mind sharing your patch?
00:33:47 Quit ompaul (Client Quit)
00:33:53preglowamiconn: doesn't look like a very busy section of memory
00:34:19amiconnProbably lots of register aliases?
00:34:23preglow int lol = open("/dump", O_WRONLY|O_CREAT); \n write(lol, (void *)0x70008000, 0x1000); \n close(lol);
00:34:29preglowjust put that in app_main, before root_menu()
00:35:06 Part Fhantazm
00:35:30preglowright, copied from nano
00:36:39 Quit Arathis ("Bye, bye")
00:37:04preglowanything interesting?
00:37:22 Quit matsl (Read error: 110 (Connection timed out))
00:37:48 Quit bluebrother ("leaving")
00:38:14amiconn0x70008b00 is sparse-decoded as 0x70008bxx
00:38:46 Quit Rob2222 (Read error: 104 (Connection reset by peer))
00:39:00amiconnIn the 0x70008axx ange there are some more registers than those known from ipl
00:39:07amiconn(also heavily aliased though)
00:39:09 Join keithr [0] (
00:39:14 Quit ashes (Remote closed the connection)
00:39:40 Join ashes [0] (n=ashes@2001:5c0:8fff:ffff:0:0:0:23)
00:40:13keithrtrying to build on debian lenny −− get make error about arm-elf-gcc missing −− anyone know what to install to fix this problem?
00:40:21amiconn0x70008a00 and 0x70008a04 look interesting
00:40:31Bagderkeithr: you installed the cross compiler at all?
00:40:57keithrBadger: Nope
00:41:26keithrBagder: is there a package?
00:41:33Bagderkeithr: you on cygwin or linux?
00:41:46keithrBagder: Linux
00:42:32keithrBagder: Thanks much.
00:45:54*adiamas waves at bagder
00:46:01adiamasglad to see you're still fighting the good fight
00:46:24Bagdergood to see you alive!
00:46:33adiamasit's been a while
00:46:45adiamasoddly enough.. i _still_ running on the same jukebox recorder
00:46:59pixelmaroolku: I couldn't figure out how I'm supposed to resume by the way (and also am annoyed to have to choose play a second time)
00:47:45roolkupixelma: note sure what you mean? You just select the second option in the menu?
00:48:15roolku(from memory)
00:48:38amiconnroolku: The menu is especially annoying as it causes mpegplayer to buffer the video twice
00:49:01amiconnOnce when loading, before showing the menu. A second time when choosing playback
00:49:05jhMikeSamiconn: 3x speedup for what? Just general lcd access on H10? damn
00:49:20pixelmaroolku: well I only saw that staying at 0,0 and of course it started from the beginning again - will try with ED and run it for a bit longer as that's the longest video I have)
00:49:22amiconnjhMikeS: lcd_update() / lcd_update_rect()
00:49:40amiconnJust small H10 atm, and lcd_yuv_blit() not yet converted
00:49:42 Quit ender` (" Sometimes I lie awake at night, and I ask, "Where have I gone wrong?" Then a voice says to me, "This is going to take more t)
00:49:58amiconnI'm about to replicate the changes for big H10, will need a tester then...
00:50:04roolkuamiconn: okay, added to the todo list
00:50:18amiconnroolku: Please, get rid of that menu again...
00:50:44*jhMikeS wonders how the PP retailos can be so clunky with music playing. even using one core only rb isn't as clunky.
00:51:26amiconnjhMikeS: Exact figures for small H10: 30MHz: 35->113fps, 80MHz: 93->303fps
00:51:28|Rain|the almighty buck promotes short development times and poor QA
00:51:38pixelmaroolku: ok, just didn't expect it to be in half minute steps...
00:51:58jhMikeSamiconn: *blink* that's full, not 1/4 screen, right?
00:52:31amiconnbig H10 should reach 80% of that, inversely proportional to lcd resolution
00:52:32linuxstbamiconn: (thanks preglow)
00:52:43Nico_Pamiconn: how did you achieve such a speedup ?
00:53:05 Join Gnu47 [0] (i=Gnu47@
00:53:12*jhMikeS ought to go grab the bridge init code for that as well
00:53:50linuxstbroolku: Also, why does mpegplayer clear the button queue before bringing up the start menu? It prevents me skipping the startup menu by pressing select twice
00:54:22*linuxstb guesses he did that originally, so shouldn't be complaining...
00:54:36amiconnjhMikeS: I'm especially interested in the meaning of 0x70008a00 and 0x70008a04
00:54:41amiconnlinuxstb: Thanks.
00:55:08amiconnThe mentioned registers really look like some kind of timing setup - and they have larger values on ipod color
00:55:09jhMikeSamiconn: any effects observed from poking at them?
00:55:14amiconnNot yet
00:55:22amiconnToo busy converting stuff
00:55:23roolkuwell, removing the startmenu is definitely the next task then :)
00:55:30amiconnWant to do the yuv blit next
00:55:48amiconnNot too many steps at once
00:56:19 Quit Gnu47 (Client Quit)
00:57:00 Join midkay [0] (n=midkay@rockbox/developer/midkay)
00:57:26Nico_PBagder: that's fast bugfixing
00:57:39pixelmaroolku: wouldn't it be possible to store the exact position at least for resume (admittedly I don't know anything about how it works)
00:57:44BagderI had a look and I saw the bug immediately...
00:59:48 Quit barrywardell ()
00:59:50roolkupixelma: yes in principle, but I don't think the result warrants the required effort
01:00:13amiconnHmmm, the big H10 causes problems :(
01:00:37*jhMikeS grumbles
01:00:51 Quit scorche|w ("CGI:IRC")
01:01:14pixelmaroolku: ok, wouldn't have expected it to be that complicated but have to believe you :)
01:01:24amiconnjhMikeS: For some reason lcd_send_data() sends in the opposite order compared to the small H10
01:02:09amiconnFor command data we can keep that, but if we want to send bulk data via the block controller, this gets in the way
01:02:25amiconnWe might need to switch to the ipod's color pixel format for the big H10
01:02:42jhMikeShehe, was going to say "endian swap" :)
01:03:03amiconnHmm, interesting
01:03:16amiconnThe small H10 already uses it, but the big H10 doesn't
01:03:20kugelwhat do you guys think of ?
01:03:45jhMikeSThat just means changing the bitmap format and pixel format?
01:04:01Lloreanroolku, linuxstb: There's a post in the forums reporting that videos that worked previously are now showing up as just green splotches post resume patch
01:04:20LloreanI seem to remember people reporting a similar problem in the past in unsupported build threads where the resume patch had been included.
01:05:06amiconnjhMikeS: Yes, basically. There might be some quirks to iron out regarding command parameters
01:05:15 Quit spiorf (Remote closed the connection)
01:05:18*amiconn hates it to code blindly :|
01:05:20jhMikeSI thought there was something in the driver IC setup to swap. Maybe the bridge can swap. It wouldn't surprise me given that others can also do it.
01:05:37amiconnThe bridge can't swap as it seems
01:07:26 Join ToHellWithGA [0] (
01:07:30pixelmaLlorean: I only know the of green blocks with mpegplayer when using voice that was fixed recently... and now it's back...
01:07:46Lloreanpixelma: Ah, so it could be that?
01:08:01pixelmajust confirmed on my c200, enabled voice and voila
01:08:08LloreanI'll ask the
01:09:06***Saving seen data "./dancer.seen"
01:09:10pixelmaroolku: the patch seems to revert some changes that were done to mpegplayer recently...
01:09:17 Quit Nico_P (Remote closed the connection)
01:09:43roolkuyep, noted
01:10:01pixelmathough I havent tried latest current yet
01:10:06 Join mf0102 [0] (n=michi@
01:10:37 Quit Toxicity999 (Remote closed the connection)
01:11:07jhMikeSyeah, the thread safety stuff for the queue (because object aren't yet dual safe). Though the changes are also make it simpler and require no mutexes.
01:11:35 Join z35 [0] (
01:14:26 Join Toxicity999 [0] (n=bryan@unaffiliated/Toxicity999)
01:14:38 Join Morey [0] (
01:16:18roolkuis this a dual code issue then?
01:17:06MoreyI've gotten a lot of possitive feedback on the mpegplayer menus from users.
01:17:27pixelmano, it applied to coldfire targets as well, it was something with the voice disabling - jhMikeSshould no better
01:17:28ToHellWithGAare any of yall experts with the jpeg viewer plugin? i would like to disable its autozoom feature as i'm using my ipod to catch the images from webcomic rss feeds. with the autozooming enabled i cannot read any text
01:17:43*pixelma should give up typing for today
01:17:46amiconnjhMikeS: Found a mistake in the small H10 lcd driver
01:17:56roolkulinuxstb etc: Morey is one of the authors
01:18:02amiconnLooks like adapting the technique will be easier than I thought
01:18:11LloreanToHellWithGA: I don't even know what you mean by "autozooming". It has 1/1, 1/2, 1/4, and I think 1/8 zoom levels, and you can switch between any of them if memory permits.
01:18:19amiconn(still requiring the format switch)
01:19:04ToHellWithGALlorean: it picks a level automagically to scale the image to fit on the screen. i'd like to disable that feature in some configuration file
01:19:31 Join Gnu47 [0] (i=Gnu47@
01:20:00ToHellWithGAthe things i want to view go from illegible to hard to read to clear as i go from 1/4 to 1/2 to 1/1
01:20:31LloreanYou'd need to patch it then.
01:20:39LloreanI don't believe there's an option for such a thing
01:21:01ToHellWithGAis the code so gnarly it would make my head spin, or would that be a simple fix to build?
01:21:13 Quit jhMikeS (Nick collision from services.)
01:21:15LloreanHaven't looked at it.
01:21:19 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
01:21:39ToHellWithGAi'll download the source and see if i can find it and understand it
01:22:00roolkuMorey: any idea about the green blocks?
01:22:01ToHellWithGAthanks man
01:22:16 Quit Gnu47 (Nick collision from services.)
01:22:23linuxstbMorey: So you're happy with the startup menu?
01:22:24 Join Gnu47 [0] (i=Gnu47@
01:23:10roolkuMorey: jhMike might be able to help, see: "<jhMikeS> yeah, the thread safety stuff for the queue (because object aren't yet dual safe). Though the changes are also make it simpler and require no mutexes."
01:23:12Moreyroolku: That seemed like an issue that was resolved about a week ago. Was that an old post about the problem.
01:23:22 Quit z35 ("Leaving")
01:23:58roolkuMorey: "<pixelma> roolku: the patch seems to revert some changes that were done to mpegplayer recently..."
01:24:01linuxstbamiconn: How many resume points does video.rock remember? Is it just one?
01:24:09jhMikeSgreen blocks should be gone. that was a problem with voice menus not being disabled.
01:24:32amiconnlinuxstb: The resume point is stored *in* the .rvf. We can do that since it's our own format
01:24:36Moreylinuxstb: I think the startup menu and playback menu should be combined into one menu.
01:24:48pixelmaand reintroduced by the menu patch
01:24:57linuxstbMorey: It's just that it's annoying when I select a file to play, and it doesn't play it...
01:25:05Moreylinuxstb: Unlimited resumes stored in the .cfg file.
01:25:18roolkujhMikeS: okay, fair enough, I thought they had reappeared
01:25:28pixelmabut I'm testing once more with the latest svn
01:25:37jhMikeSroolku: not my fault if they did :P
01:26:29pixelmar15057 they are back
01:27:05jhMikeSwas the voice disabling removed along with the threading stuff?
01:27:32*jhMikeS takes a closer look
01:28:12roolkuI'm trying to find the relevant commit
01:28:19pixelmamaybe it's because the changes weren't adopted for the new menu?
01:28:30pixelma3October 3rd
01:28:31 Quit bertrik (Read error: 104 (Connection reset by peer))
01:28:42Moreylinuxstb: I do think the seek and resume should be included in the playback menu, but I like the startup menu.
01:29:09*Llorean doesn't like the idea of having to do anything more than click on a .mpg to play it
01:29:11pixelmathe startup menu annoys me
01:29:24LloreanThey should, from a user perspective, work fundamentally similar to songs.
01:30:00pixelmathe rvf player is IMO much more intuitive in that regard
01:30:20MoreyOnce the menus are combined the startup menu could be removed. Then it would be up the user to bring up the menu if they wish.
01:31:28jhMikeShmmm...talk_disable_menus() is called before the startup menu is shown
01:32:02MoreyI think that a 30 second jump by pushing left/right buttons during playback could be done but the jump would take around 4 seconds.
01:33:05roolkujhMikeS: yes, I saw that - is this a problem?
01:33:21linuxstbMorey: I would like to see the start menu go, but if there is a resume point stored for the current file, the user could be prompted with a yes/no prompt to play from the resume point. That's how the other Rockbox video player works IIUC.
01:33:30jhMikeSroolku: that's where it should go...before any menu is might be the splash
01:33:42Lloreanlinuxstb: I agree.
01:34:13jhMikeSthere's no talk_disable_splash() yet? better get that one in SVN then.
01:34:18pixelmaweird, I don't have that on the M5 where I could also see bright blocks on black before (r15074)
01:34:42jhMikeSroolku: ah, I see the difficulty
01:35:48jhMikeSwe must wait for any splash to stop voicing, then IRAM can be initialized
01:36:21linuxstbMorey: I haven't looked at your code in detail, but do you think the seeking can be made any faster? If not, what do you think about investigating other container formats, ones which are more seeking friendly?
01:36:31jhMikeSafter that, splashes must not voice
01:37:07*roolku never knew splashes were voiced...
01:38:12 Quit billytwowilly (Read error: 104 (Connection reset by peer))
01:39:10Moreylinuxstb: In my opinion the seeking is optimized about all it can be for these processors. I don't know about other containers, maybe jhMikeS has some thought on that.
01:39:44 Quit Isolinear (Read error: 110 (Connection timed out))
01:39:51linuxstbMorey: Would that be something you would be interested in working on though?
01:40:25amiconnjhMikeS: 1st try:">
01:40:38amiconnYou need to 'make clean' and reconfigure
01:40:44Moreylinuxstb: I will certainly look into it and run it by the Gwynne (the co-author of this patch)
01:40:44amiconn(changed bitmap format)
01:40:59jhMikeSsparse buffering. Everthing seems buffered in between one point and another or else it wouldn't take so long.
01:41:32 Join billytwowilly [0] (
01:41:32 Quit Zagor ("Client exiting")
01:41:54Moreylinuxstb: I'm hoping to spend some time this weekend making an attemp at combining the menu system as well.
01:41:56 Quit Gnu47 (Read error: 110 (Connection timed out))
01:42:14 Part jbm
01:42:25preglowamiconn: i did not make clean and reconfigure...
01:42:35amiconnpreglow: *you* don't need to
01:42:38 Part toffe82
01:42:39preglowthen hooray
01:42:40 Join Ebert [0] (
01:42:43amiconnOnly big H10 needs it
01:42:54preglowwas hoping for some speedups there, for a second :)
01:42:57jhMikeSamiconn: ok. this'll take a few minutes since I need to do a full build (cleaned out all the builds in other trees other than dc)
01:43:03 Quit kugel ("ChatZilla [Firefox]")
01:43:07amiconnjhMikeS: Want a build?
01:43:20jhMikeSsure, that will be faster
01:43:34amiconnI hope just the .mi4 is sufficient?
01:43:52amiconnOr wait a few more mins
01:43:53jhMikeSas long as I play no music
01:44:24rasher is what we show users when they follow the manual for "automatic installation". Why not a page on, linking these files?
01:44:29amiconnYeah, but you might want to try plugins...
01:44:54jhMikeSyeah, they're not compatible I'll just switch over and build from a clean one
01:45:16amiconnBuild is already running...
01:45:35linuxstbrasher: Yes, those kinds of links appear in a few places in the manual. Another option would be for the manual to have direct links to files.
01:45:56roolkusorry, got to go ...
01:46:00rasherlinuxstb: that'd require the manual to change each time there was the smallest of changes
01:46:09 Quit roolku ()
01:46:17linuxstbrasher: That's no different to changing a webpage...
01:46:18jhMikeSwell, I started one too but I think you'll beat me to it
01:46:31linuxstbrasher: Or we keep the filenames the same...
01:46:31rasherlinuxstb: except the user has to download a new manual
01:46:47linuxstbrasher: True, I forget there's a PDF...
01:47:11spkyis there an ipod simulator for evilg's build?
01:47:21linuxstbAsk evilg...
01:47:29amiconnjhMikeS: Depends... if you build on linux yours might be faster...
01:47:49jhMikeSjust good ol' VMWare
01:47:59pixelmalinuxstb: and then link to all versions for the different OSs?
01:48:08linuxstbpixelma: Yes.
01:48:10jhMikeSmaking libfaad now
01:48:20amiconnHmm, the controller in the small H10 might be something else than TL1771 after all
01:48:54rasherlinuxstb: I think a few central download pages (one for rbutil, one for builds and one for install tools?) would be saner
01:48:59linuxstbAlthough maybe we could have a new "Installation" page - with links to rbutil and links to the manual.
01:49:13Moreylinuxstb: Gwynne is going to look into possible a different container. The goal would be to find one that has foward and backword indexing. avi might be an option, we'll have to look further into it.
01:49:21pixelmalinuxstb: I would think that could be a long list and would also be confusing but have no strong opinion or idea about it, quite tired about to get some sleep
01:49:32 Quit massiveH ("Leaving")
01:49:42*jhMikeS is thinking WTF little booboo he made he can't find on single-core and getting frustrated. :\
01:49:48linuxstbpixelma: Yes, I'm against my own idea now as well...
01:50:23pixelmaheh, if you say so :) good night all
01:51:15 Part pixelma
01:51:52linuxstbMorey: The problem with .avi of course is that it will confuse people into thinking Rockbox can play any .avi...
01:52:45linuxstbBut I think it's at least worth investigating other containers, to see if they are worth the effort.
01:52:49LloreanYes, people seem to think AVI is a type of video stream, rather than a container.
01:52:53amiconnlinuxstb: Doesn't mpeg have time stamps in the frames?
01:53:18linuxstbIt does.
01:53:43LloreanOf course, if we picked winff as our "official" conversion tool, we could use AVI as the container, but have it automatically name the output file something else, so the people who really have no idea won't see that it's avi anyway (but anyone who reads the docs or asks would of course be told)
01:53:45amiconnSo seeking should be possible without reading all the way through the file
01:53:56 Join Rob2222 [0] (
01:54:33amiconn(provided a resync to the stream isn't too costly - without actually decoding the video)
01:54:35jhMikeSamiconn: it's loaded with various kind of timestamps, not just PTS.
01:55:14amiconnJust a binary search through the file, and if already playing for a while, it could be supported by an estimation
01:55:22 Quit Rob2222 (Client Quit)
01:55:24jhMikeSnot to mention syncwords. it could actually decode some video while seeking for thumbnailing.
01:56:00jhMikeSMPEG 2 has 1/2 frequency D-frames just for FF/RW purposes as well
01:56:27jhMikeSamiconn: thanks
01:57:00jhMikeSlinuxstb: Those are for seeking purposes in MPEG2 so players can play double-speed and backward and such.
01:57:55amiconnIsn't that a lot of extra data just for seeking?
01:58:20jhMikeSthey're all keyframes, so don't need to reference other frames and the low-pass filtered.
01:58:38jhMikeSI don't know if they're _just_ for seeking
01:58:45 Quit krazykit ("leaving")
01:59:00 Join krazykit [0] (
01:59:42jhMikeSamiconn: it comes with test_fps too?
02:00:06jhMikeSthe rockbox logo is borked but the screen looks ok
02:00:28amiconnBorked in what way?
02:00:46jhMikeStwo small images 1/4 screen size at the top half of the display
02:01:42amiconnOh, hmm
02:01:52*amiconn wonders what's going wrong there
02:02:02jhMikeS30MHz: 1/1 129fps. 1/4 542.5 fps 80MHz: makes the screen show like the rb logo
02:02:03amiconnAre the colours correct?
02:02:24amiconnAha, so the LCD controller can't handle the bridge timing when boosted
02:02:48amiconnCould you do the 0x70008000..0x70008fff dump too?
02:02:55amiconn(with your normal build)
02:03:08jhMikeSno patch?
02:03:31jhMikeSnot the one you just sent, correct?
02:03:46amiconn[00:34:22] <preglow> int lol = open("/dump", O_WRONLY|O_CREAT); \n write(lol, (void *)0x70008000, 0x1000); \n close(lol);
02:03:46amiconn[00:34:29] <preglow> just put that in app_main, before root_menu()
02:03:58amiconn...without my patch
02:04:54amiconnSo the big H10 needs some more work - finding out how to set the bridge timing
02:05:13 Quit jhulst ("Konversation terminated!")
02:05:28amiconnI'll commit my changes then except using the block transfer on big H10
02:06:07 Quit Morey ("Ninja IRC v1.5.8.1(#1) exiting after 52m24s of use")
02:06:50amiconnHmm, perhaps a quick check would be better...
02:09:57*amiconn now remembers this half-size graphics effect">
02:10:20amiconnGot it when experimenting with dma on H300, and there were too few waitstates
02:11:44amiconn0x70008000 is zeroed - that's probably the fastest possible setting in the bridge...
02:12:22*amiconn thinks the Nano settings should work - the nano has the same lcd controller
02:13:18amiconnBut I need some sleep soon, so no experimentation regading those regs now
02:13:48amiconnjhMikeS: Could you please check a new build? Same url...
02:14:05 Quit PhinnFort (Remote closed the connection)'s ready now?
02:14:29 Join Gnu47 [0] (i=Gnu47@
02:14:30amiconnJust to make sure it doesn't break something
02:14:50amiconnI'll dig into these regs further, after getting this stuff committed
02:14:55jhMikeSH10 isn't so fragile any more
02:15:07linuxstbamiconn: Do you need anything more testing from me, I'm about to go to sleep?
02:15:10amiconnYes, but gfx glitches look ugly
02:15:34amiconnlinuxstb: No tests on color needed atm. Thanks for testing :)
02:16:01amiconnjhMikeS: gfx glitches are nasty...
02:16:20*amiconn repeats himself the colors are whacked. now just black and white look right.
02:16:34amiconnAh, yes
02:16:43linuxstbamiconn: OK, goodnight.
02:16:54amiconnjhMikeS: Now I need a swap16()
02:17:12amiconn(temporarily, until we can use the block transfer)
02:17:27jhMikeSyeah, it appears like swapping
02:17:37 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
02:17:53amiconnYeah, I know what I did wrong... swapped pixel format requires swapping, or another write fn
02:18:01amiconnI'll do the latter, as it's faster
02:19:18amiconnjhMikeS: Is video also colour swapped?
02:19:22amiconnI guess it is...
02:20:12amiconnah ok
02:20:18amiconnToo late at naight :/
02:22:21*jhMikeS just has some odd feeling the controller can swap bytes
02:23:30 Quit joshin (Read error: 110 (Connection timed out))
02:25:19 Join toffe82 [0] (
02:27:22lymecaI have a 30GB iPod video that was HFS+ formatted and I'm running GNU+Linux on my computer.
02:27:39lymecaI just used dd to copy the correct MBR file from the Rockbox wiki and copied it to /dev/sdc
02:27:49lymeca(which was the iPod HDD device file)
02:28:07lymecaIsn't that the correct thing to do?
02:28:45jhMikeSlymeca: so you have an iPod Video 30GB that boots with the latest SVN. I'm asking because of a user report on the same model.
02:29:14lymecaIt was HFS+ formatted and using the proprietary Apple firmware that was shipped with the device.
02:29:51jhMikeSah, so you're installing now
02:29:54lymecaNow it's just restarting over and over again with audible clicking noises occasionally
02:30:19lymecaYeah I was trying to copy the new partition table to the MBR of the device so I could format the partition with FAT32
02:30:36lymecaBut I perhaps have messed up since...
02:30:36jhMikeSI've heard of that but I don't have that model so I don't know what causes that problem.
02:30:42lymecaHow do I boot into disk mode?
02:30:58 Quit karl ("CGI:IRC (Ping timeout)")
02:31:22jhMikeSI'm sure those procedure must be in the iPod installation Wiki
02:32:43 Join davina [0] (
02:35:17 Quit [IDC]Dragon ("CGI:IRC")
02:37:12amiconnWhoa, -6KB for the big H10 - just due to optimising...
02:37:39amiconnAlmost -7KB in fact
02:40:06preglowthat's quite sick...
02:41:22 Join TrN [0] (
02:41:46 Join Rob2222 [0] (
02:43:07 Part TrN ("If I do not return, avenge my death.")
02:45:23 Join S1gn [0] (
02:45:50 Join Canar [0] (
02:46:21*jhMikeS is glad amiconn is making room for his latest work :)
02:47:03jhMikeSeven more than needed in fact
02:48:24amiconnI know the reason for the huge green delta... lcd_write_reg() is all inline stuff, and got simplified *a lot*
02:48:53amiconnI doubt that it *needs* to be inline, so there's room for some more green delta
02:49:12 Quit Gnu47 (Read error: 110 (Connection timed out))
02:49:28amiconnThe performance critical places could use the 2 partial functions directly (which would still be inline)
02:49:59 Part S1gn
02:51:49amiconnBut now, sleep()
02:56:33 Join JdGordon [0] (
03:00:38preglowbeware that sleep always sleeps 1 more than you ask for! :P
03:07:28jhMikeSpreglow: sort of. only if you sleep exactly on a tick boundary. otherwise it's ticks + remaining_time in reality
03:08:21 Nick vosque is now known as vosque_away (
03:09:12***Saving seen data "./dancer.seen"
03:12:11preglowi'm starting to need some sleep rather bad
03:12:23preglowthink i'll just go fix that
03:13:11jhMikeStoodles :P
03:28:22 Quit adiamas (Connection timed out)
03:34:32 Quit spky ("byte me")
03:35:53 Join Klevi [0] (
03:37:17 Quit Klevi (Client Quit)
03:37:41 Join Mouser_X [0] (
03:37:59Mouser_XIn cygwin, how do I delete a directory?
03:38:09Mouser_X(with subdirectories?)
03:38:35Mouser_XMy SVN trunk/rockbox stuff won't compile.
03:38:38jhMikeSrm -r <dir>/* ?
03:38:51 Join Klevi [0] (
03:39:15jhMikeSor just "rm -r <dir>"
03:39:23KleviQuick question, can the Sansa e200 tool be used to unbrick the sansa c200?
03:40:04Mouser_XjhMikeS: Someone told me awhile ago (2 weeks maybe?). I thought it had to be a capitol -R? (I couldn't remember the command, but I remember a capitol R. It might relate to something else though).
03:40:41jhMikeSMouser_X: Debian anyway shows -r, -R, and −−recursive as acceptable
03:41:50Mouser_XOr, is there a way to make my SVN identical to the one on the site, w/o deleting my current set?
03:42:18Mouser_XI tried to do "svn update" but that didn't work. I even tried to checkout again, and it said I already had the most recent version.
03:44:02PaulJam_i use 'svn diff | patch -p0 -R' but i think svn revert works also
03:44:23Mouser_XI tried to revert the patches I added, but it's still not working.
03:44:47Mouser_X(Or do you mean something different?)
03:45:24PaulJam_i ment the whole comand between the ' '
03:46:29Llorean"svn revert -R *" should convert all the files back to the SVN versions of it, but it won't delete any non-SVN files or anything
03:50:38 Join adiamas [0] (
03:54:11 Quit Klevi ("Quit messages. The unintentional channel spam of people closing IRC, realizing they have a life.")
03:55:48 Join Soligna [0] (i=cdc9cf3e@gateway/web/cgi-irc/
03:56:16SolignaHey, I just flashed my sansa e200r with rockbox and when I boot it says load main image failed switch to recovery mode
03:58:09Mouser_XI just remembered another question I had. When I attempt to use/view the Rockbox database (I use it for the "History" and the "User Rating"), it says "Dir buffer full" and then displays an empty/blank screen.
03:58:42Mouser_XI tried upping the limits on the directory buffer (I have it at the highest it will go), but it made no difference.
04:00:17PaulJam_have you rebooted after setting the limits?
04:02:00 Quit keithr ("Chatzilla 0.9.67+ [Iceape 1.0.9/2007050500]")
04:04:05 Quit Toxicity999 (Read error: 104 (Connection reset by peer))
04:05:18 Join Toxicity999 [0] (n=bryan@unaffiliated/Toxicity999)
04:05:45 Join eigma [0] (
04:06:12 Join psycho_maniac [0] (
04:06:40Mouser_XNo? I thought I had changed the settings before updating the database.
04:07:17Mouser_X(After updating the database, I had to reboot to view the database.)
04:08:28 Quit eigma (Client Quit)
04:08:41Mouser_XBut, checking just now, it's working. So apparently I had changed settings *after* updating the database (and then rebooting).
04:09:42 Join sdoyon [0] (
04:12:24 Quit PaulJam_ (".")
04:14:12 Quit mf0102 ("Verlassend")
04:16:15 Quit Soligna ("CGI:IRC (EOF)")
04:20:22 Join joshin [0] (n=joshin@VDSL-130-13-11-27.PHNX.QWEST.NET)
04:22:22Mouser_XSo, if I want to use the database, I have to have a high directory buffer?
04:23:05Mouser_XI've got 30,000+ (easily) files on my Gigabeat. I don't think the directory buffer goes that high...
04:24:13Ebertyou should know, if you've looked in that menu
04:24:35alienbiker99Mouser_X i don't think theres a limit as long as your change it in that menu
04:25:15Mouser_XMy recollection is 10,000 is that the directory buffer can only go to 10,000. I was in a hurry, so I didn't look very closely.
04:25:47Mouser_X(Sorry for the broken sentence. I started one thing, and then changed it, without getting rid of the first...)
04:26:48alienbiker99hmm now i'm not even sure.
04:30:49LloreanI didn't think that the database was restricted by the directory buffer limit
04:32:46 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
04:32:50 Join miepchen^schlaf [0] (
04:33:49JdGordonthere is no limit to the amount of files the db can hold (im 99% sure), but it is limited by the directory limit on the amount of files it can read in each directory
04:34:27LloreanJdGordon: I thought that it browsed by part somehow
04:35:31LloreanI seem to recall a lot of discussion on this.
04:35:41JdGordonof there being a file limit?
04:35:53LloreanOf there not needing to be a list-length limit for database
04:35:59LloreanOr maybe it was dircache.
04:36:20LloreanI remember the term "browsing by part" or something, but don't remember exactly what was meant with it
04:36:41LloreanDatabase does seem to have a file limit too, though I believe it's more a bug than an intended limit.
04:38:04Ebertwhat's the car adapter mode do?
04:39:13LloreanPauses the device when external power is cut off
04:42:22 Quit adiamas (Read error: 104 (Connection reset by peer))
04:46:07sdoyonWhen following a bookmark from the MRB, is it intentional that we don't GO_TO_WPS but rather drop back into the root menu?
04:49:46 Quit Thundercloud (Remote closed the connection)
04:52:14sdoyonjdgordon: are you there?
04:52:59 Join TNTales [0] (
04:53:59TNTalesanyone awake?
04:54:33scorcheif you have a question/concern, feel free to voice it
04:54:35TNTalesI'm in over my head with a sansa 260r that windows no longer recognizes, could someone walk me through this
04:54:50scorchewell, what did you do?
04:55:02TNTalesI have one of the Rhapsody players
04:55:22TNTalesand I was following the instructions for
04:55:26TNTalespatching the bootloader
04:55:40TNTalesthere is a part where you run e200tool recover
04:55:56TNTalesand it looks for a driver and for whatever reason it didn't find it
04:56:12TNTalesand now windows says "unknown device" and when
04:56:15scorcheplease use paragraphs instead of hitting the enter key every so makes it much easier to read and is much less spammy
04:56:29scorcheare you doing this in windows?
04:57:06TNTalesas you wish. Windows says unknown device and I am unable to manually install the driver through Device manager because it says "this does not contain information about your device" and yes, it's windows xp sp2
04:57:06 Quit Daenyth (Read error: 104 (Connection reset by peer))
04:57:55*jhMikeS wonders what the mutex testing is supposed to be for in ata.c. prevent deadlocks?
04:58:26scorcheTNTales: did you happen to see this on the instructions? " NOTE: Bootloader patching in Windows is even more difficult than in Linux. Do NOT perform these steps unless you know EXACTLY what you are doing (especially with drivers)."
04:58:52TNTalesyes. I did. The think was bricked anyway with the original firmware so I thought I'd try it. How else do you learn?
04:59:30TNTales"thing" was bricked
05:00:57scorchewhat do you mean?
05:01:00TNTalesI called SanDisk support and they sent me several files and I tried formatting and reinstalling the provided firmware
05:01:13scorchewhat was wrong with it?
05:01:20TNTalesand all it does is say "load main image faild Switch to recovery mode"
05:01:37scorchewell, what did you do originally to get it "bricked"?
05:01:48TNTalesUnhooked it from the PC
05:01:58scorchebefore that
05:02:07TNTaleswas hot swapping and it lit up real bright and poof
05:02:14TNTalesI was adding files and deleting them
05:02:28alienbiker99was it updating the firmware?
05:02:54TNTalesadding some and deleting other that is. I don't believe so, I hadn't told it to upgrade and I didn't have the update/loader installed at the time
05:03:26TNTalesI've had it less than a week. When they told me I'd have to return it I thought I'd try installing rockbox before I shipped it back
05:03:47scorcheTNTales: so did you execute the recover command?
05:04:21TNTalesyes and it goes into a countdown: searching for (driver name) 9 8 7 6 ... so on and then says "not found"
05:04:37scorcheok...lets do something different...
05:04:53scorchehold power for 15 seconds or until the blue light turns off
05:04:54*jhMikeS sees JdGordon made the commit in r11461 that added the lock test and sees JdGordon is lurking :)
05:05:12TNTalespower is now off.
05:05:28scorchethe hold switch is on, right?
05:05:39TNTalesyes, hold switch is now on
05:05:52scorcheand usb is plugged in?
05:06:16TNTalesnot yet, do I need to put in in a particular mode first (recovery, main, manafacturing)?
05:06:25scorcheno...just plug it in without touching any buttons
05:06:50scorchethen hold down record and hit the power button
05:07:01TNTalesok. It came on and said "key locked, system shutdown"
05:07:13TNTalesnow in recovery mode
05:07:31 Quit Mouser_X (Read error: 104 (Connection reset by peer))
05:07:46TNTalesIt says "USB Cable connected" but isn't saying that it's in the MSD mode
05:07:58scorcheis it showing up in your computer?
05:08:05TNTalesthis is where windows tells me in device manager as an unknown device
05:08:15scorcheit should detect it..
05:08:25TNTalesyeah "Unknown Device" let me try again
05:09:13***Saving seen data "./dancer.seen"
05:09:25TNTalesok, uninstalled, unplugged and I've put it back in recovery mode, going to see if it detects
05:09:53TNTalesOk, found usb device "Unknown Device" Your new hardware is installed and ready to use
05:10:01 Join ArchFool [0] (
05:10:05TNTalesyeah, in device manager, I uninstalled "Unknown Device"
05:10:12 Quit Toxicity999 (Read error: 104 (Connection reset by peer))
05:10:15scorcheyou dont need to do that
05:10:25TNTalesoh, ok, sorry
05:10:34scorchein the future, please do what i ask, and not anything more that you assume to could be bad ;)
05:10:40TNTalesok :)
05:10:44TNTalesdo we need to start over?
05:11:33scorchetry the process over again...(turn off, plug in, boot into recovery mode, see if it will mount)
05:12:28TNTalesI am booted into recovery mode, it says "Welcome to Recovery Mode" USB Cable Connected
05:12:32 Join Toxicity999 [0] (n=bryan@unaffiliated/Toxicity999)
05:12:49scorcheand windows...?
05:12:51TNTaleswaiting for old computer to recognize device (I'm building a new one, let's hope that goes better than this, lol)
05:14:17TNTalesah, found the problem, still had a device manager window opened. I've closed everything except mozilla and IRC now,
05:14:29TNTalesComputer says "USB Device Not Recognized"
05:14:45 Join Mouser_X [0] (
05:14:49advcomp2019TNTales, you said old computer... what operating system
05:14:54TNTalesxp sp2
05:15:12TNTalesBy old I mean 5-6 years
05:15:22 Quit bb (Nick collision from services.)
05:15:24scorcheTNTales: have you tried plugging it into a different port?
05:15:27 Join bb_ [0] (n=bb@unaffiliated/bb)
05:15:30Mouser_XWhen deleting a directory (and its subdirectories) in cygwin, how do you do "Yes to all"?
05:15:37psycho_maniacTNTales: this might be a long shit but it works for me sometimes. do you always have the coard plugged in and you just plug in the part to the player when you want?
05:15:51TNTalesyes psycho_maniac, that is how I do it
05:15:54alienbiker99you might want to restart after switching drivers and such, but idk too much about your process here
05:16:31scorcheMouser_X: rm -rf
05:16:33TNTalesI will plug into another port and try the process again
05:16:55Mouser_Xscorche: Thanks
05:17:13TNTalesturning power off first, then unplug both ends. then does it matter which of the 2 I plug in first, device or computer?
05:18:41TNTalesok, plugged into new port. I will plug in the player with hold on and when it powers off I will hold record while turning on the power, correct?
05:19:34TNTalespowered up into recovery mode, windows has recognized the device and is thinking
05:20:12TNTalesdevice now says "Enter USB2.0 MSD Mode" and Windows says "Found New Hardware Embedded MP3/Video Player"
05:20:41scorchetell me when it mounts
05:20:42TNTales"Your new hardware is installed and ready to use" and the 16MB-Format drive has appeared download this file and copy it into the drive:
05:21:13advcomp2019but unzip too
05:21:28scorcheoh...heh...of course :)
05:21:29TNTalesok, I will unzip. Should I put it in a specific directory first?
05:21:42scorchejust the file directly onto the device (the unzipped file
05:21:52scorchethen hit safely remove, then unplug the device
05:22:01TNTalesok. page timed out.
05:22:02scorche(after it says you can now remove..)
05:22:13scorcheeh? works for me
05:23:20TNTaleshmm. I can't get the site to respond. Is it on sourceforge? this site keeps timing out on me (Could be my computer)
05:23:35alienbiker99works for me
05:23:44scorchelet me try hosting it...
05:25:19TNTalessorry, I've been neglecting replacing this computer as I rarely use it anymore except as a repository for media files. so it's only got a 40GB HDD and 512 MB of ram. I think it has a GeForce 4 (PCI) and the processor is an amazing 1.4 GHz
05:25:39scorchethat shouldnt matter
05:26:30TNTalesI have a zip file here that has that same name">
05:27:12scorchecan you access that?
05:27:16TNTalesreceived and downloaded
05:28:31scorcheunzip, copy the output file to the device, safely remove, unplug after it says you cna safely remove now, then turn hold off and start
05:29:13TNTalesI paste it in the 16MB-FORMAT device, correct?
05:29:47scorchethe directory should also have a "version.txt" file in it
05:29:55TNTalesyes, that's the one
05:30:06TNTalesI will paste it, and then do safely remove device
05:30:14scorcheyup...just as i said
05:30:44TNTalesI can't find the "Remove Device" icon in my taskbar
05:30:59scorchehit the extend arror?
05:31:11TNTalesyes, does not show up
05:31:34psycho_maniacyou could go into my computer right click on the defice and click eject also . i think
05:31:42scorchesomething like that, yes
05:31:44Mouser_XIt's the USB icon (I'm not sure it will say "Safely Remove Hardware" unless you click on it).
05:31:50TNTalesI see eject, this will work?
05:31:56scorcheit should
05:32:38TNTalesplayer says "LUN0 unlocked" "LUN0 ejected" is that bad?
05:32:47scorcheno...that is fine
05:33:04Mouser_X(Just get out of the way when it blows up...)
05:33:11Mouser_X( ;P )
05:33:18TNTalesI'm going to unhook the device from the cable, wait for it to power down and then turn off hold and resta4t
05:34:08TNTalespowering on. "Load main image failed Switch to recovery mode"
05:34:23TNTalesthat's why I wanted to use rockbox to replace the old firmware, if that would work
05:34:40scorchedid it say anything about updating the firmware?
05:35:05TNTaleswhen it restarted it said "reading main image" "writing image" and it says "2 seconds"
05:35:26TNTalesor words to that effect. I can repeat the process and get the exact messages if you want
05:35:30scorchehmph...something may be wrong with your player...
05:35:56advcomp2019are you sure it is an e200r i wonder
05:36:06TNTalesyou know, it may not be
05:36:24jhMikeSare there a bunch of counterfeit "R" players around or something?
05:36:33scorcheyou can do the same thing with
05:36:39TNTalesneither the box, nor the player bears any indication of which it is, I guess I just assumed since it was new it was the R firmware
05:36:56TNTalesI bought it direct from Amazon, so I dont' think it's counterfeit
05:37:08jhMikeSthe firmware identifies it iirc
05:37:11scorcheTNTales: do the process you just did, but with the file i just linked you
05:37:15TNTalesok, I will do the same steps with this new firmware
05:38:02scorchejhMikeS: there were a few rhapsody devices that were really vanilla in disguise iirc
05:38:15alienbiker99TNTales when you used it did it have "rhapsody channels" in the main menu?
05:38:51 Quit ArchFool ("po0f!")
05:38:51TNTalesI never looked, I've only had it for 2 days and I've spent that time filling it up with MP3s. I hadn't even had the chance to really check it out
05:39:04TNTalesI have the new file and am doing the process
05:40:17TNTalesdevice is loading and being recognized. I am pasting the file you sent me "PP5022.mi4".
05:41:02TNTalespasted, and ejected. Unplugging the player and letting it power down.
05:41:28TNTalesit says "reading main image" "writing main image" upgrade complete
05:41:34TNTalestook 2 seconds
05:41:52*jhMikeS 's H10 is all full of funky colors now :p
05:42:02jhMikeSamiconn: ping (for the logs)
05:42:05TNTalesit's on
05:42:23scorche(your device is not an R)
05:42:26TNTalesthat was it, it's working
05:42:31TNTalesok, that could explain things.
05:42:49TNTalesNow I can try to install rockbox again! without the messy driver thingy. I think I'll leave that to the professionals
05:43:08TNTalesthanks for your patience and assitance :)
05:43:29scorcheTNTales: the install for the non-R is much easier :)
05:44:03TNTalesI'm just hyped because of all the codec support. I just wish this thing supported files from audible. *sigh*
05:45:09JdGordonjhMikeS: sdoyon: sup?
05:46:09sdoyonI'd like to commit a bug fix for an oddity with GOTO_PREVIOUS_MUSIC. Does anyone know this stuff and can tell me if my solution is the right approach?
05:48:21jhMikeSJdGordon: " * jhMikeS wonders what the mutex testing is supposed to be for in ata.c. prevent deadlocks?"
05:50:33Mouser_XWho here as Wiki access? I ask, because I saw somewhere that "patch −−binary -p0 < patchfilename" doesn't need "−−binary" in it. Should this be updated in "SimpleGuideToCompiling"?
05:50:37sdoyonjdgordon: When following a bookmark from the MRB, is it intentional that we don't GO_TO_WPS but rather drop back into the root menu?
05:51:35LloreanMouser_X: The whole point of it being a wiki is that so anyone can fix it.
05:51:39TNTaleswhen I extract rockbox to my device from the zip file, should it show up as its own folder "rockbox" or should the files just be loose?
05:52:00scorcheit should be ".rockbox"
05:52:22psycho_maniaci think that guide was made when using cygin so maybe you do in cygin but i know you dont in vmware player.
05:52:34sdoyonjdgordon: isn't it kind of late in your timezone?
05:55:23Mouser_Xpsycho_maniac: Ah.
05:56:32jhulstDo bug fixes get attached to the bug report or should a new patch created for it?
05:56:37TNTalesWoohoo! I have rockbox running now!
05:56:41jhulst*be created for it
05:56:51psycho_maniacMouser_X: Ah what?
05:56:52TNTalesdo I have to do anything special to make it dual bootable?
05:57:14alienbiker99hold left when it boots or stick your usb cable in before you turn it on
05:57:40psycho_maniacand that will load the OF
05:58:25jhMikeSJdGordon: you did add the code so I thought I'd ask you
05:59:02TNTalesok, when i go to load the OF, it's calculating the binary type and then the screen just fades to white and reboots, it won't load the OF
05:59:16psycho_maniacsdoyon: do mean when your in the wps screen and hold down select and select a setting from that menu? and when you go back it goes straight to the wps?
05:59:43 Join spky [0] (
05:59:49psycho_maniacTNTales: will you be using the OF very much?
05:59:50Lloreanjhulst: You can really do it either way, though it's probably best to open a new task for it and then post a link in the bug task
06:00:24TNTaleswell, It's a bit more user friendly at first. I guess I have a couple of questions, but I should probably go read the wiki
06:00:52jhulstOkay, I've posted the patch for 7831, I can put a new task in too
06:01:46sdoyonpsycho_maniac: Sorry, what? I mean from the most recent bookmarks, follow one, it starts playing, but does not go to the WPS, rather drops you back to root menu.
06:01:55TNTalesdo I need to reinstall the OF using recovery mode?
06:01:56Lloreanjhulst: If you've already added it to the bug, no sense opening another task
06:02:32Lloreanjhulst: I've changed the existing task to "Patch"
06:02:43jhulstLlorean: great, thanks
06:03:02sdoyonpsycho_maniac: or my second issue: listen to radio, stop it, start something that starts playback but doesn't go to WPS, like a bmark from MRB, or playlist insert, or playlist catalog, then press PLAY to GO_TO_WPS, and bang you're back into the radio.
06:04:35psycho_maniaci do not have the radio on my player.
06:06:03psycho_maniacbut i do see the issue with the bookmarks.
06:07:38 Join massiveH [0] (
06:12:21JdGordonjhMikeS: ill have a look
06:12:25JdGordonsdoyon: gmt+10 here
06:12:43JdGordonand i dunno about the bokmark stuff
06:15:38jhMikeSamiconn: (more for the logs) dual-core is running stably on H10 with scaling now. The clockskipping is the key during scaling (along with dirty little tricks).
06:15:43JdGordonjhMikeS: is the change in 11461 causing problems? I remember adding it but not 100% sure why
06:16:18jhMikeSJdGordon: No problems but since I implemented a special spinlock function just to serve it, I wonder if it's really needed.
06:16:33JdGordongoing by the code, i tihnk it was put in so the ata calbacks wernt triggered more often than twice a second?
06:17:44jhMikeSok. maybe that should be checked in the routine itself?
06:18:05 Quit Toxicity999 (Remote closed the connection)
06:18:21jhMikeSof course you could miss a critical save then...hmmm
06:21:29 Quit TNTales (Read error: 104 (Connection reset by peer))
06:27:58 Join Toxicity999 [0] (n=bryan@unaffiliated/Toxicity999)
06:28:43 Quit sdoyon ("ircII EPIC4-2.4 -- Are we there yet?")
06:31:50 Quit Rick ("I don't need to be here.")
06:33:42 Quit psycho_maniac (" HydraIRC -> <- It'll be on slashdot one day...")
06:35:46spkycan someone please send me sh-elf-gcc-4.0.3-1.tar.bz2, it doesnt seem to finish downloading completely wiht the cygwin setup.exe
06:38:47spkyOR, will this file work?
06:39:40jhulstspky: #cygwin maybe?
06:41:48jhulstspky: no problem, they just might be able to help you better
06:45:55 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
06:46:47 Join Rick [0] (
07:05:17 Nick fxb__ is now known as fxb (
07:09:10 Join psycho_maniac [0] (
07:09:15***Saving seen data "./dancer.seen"
07:12:59 Join crakmnky [0] (
07:13:27crakmnkyI want to update rockbox but I dont want to loose my high score in chopper
07:13:34crakmnkyanyone happen to know where it's stored?
07:14:32Mouser_XWhen updating Rockbox, there's no need to delete the previous install (well, there is occasionally, but it's pretty rare.)
07:14:41psycho_maniaccrakmnky: .rockbox\rocks\chopper.cfg
07:15:09psycho_maniacbut as Mouser_X said you dont lose that data when you update rockbox.
07:17:32crakmnkysansa c200 support! :D
07:19:06 Quit jhulst ("Konversation terminated!")
07:23:16 Quit psycho_maniac (" I love my HydraIRC -> <-")
07:24:25 Join SkinInd95 [0] (
07:25:25 Quit crakmnky ()
07:31:16lymecathis is what fdisk tells me: /dev/sdc1 1 3648 29302528 83 Linux
07:31:23lymecafor my 30GB iPod video
07:33:09 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
07:33:17lymecaI copied the 30GB MBR from the FAT32 wiki conversion page
07:33:40lymecaused dd to copy it to /dev/sdc
07:34:07lymecaAnd I can copy data to /dev/sdc2 including having Rockbox installed and working
07:34:31lymecaThe /dev/sdc2 partition is a legitimate FAT32 partition and rockbox works on it
07:34:49lymecaBUT Rockbox says it's only ~6.8GB
07:35:01lymecaAnd fdisk obviously doesn't know what the hell is going on
07:37:21toffe82jhMikeS: ping
07:39:34 Join donutman25 [0] (n=chatzill@
07:40:08 Nick fxb is now known as fxb__ (
07:43:29 Join webguest82 [0] (i=04146273@gateway/web/cgi-irc/
07:43:48 Quit massiveH (Read error: 113 (No route to host))
07:44:05webguest82Is it possible to play a podcast on an ipod 80GB running rockbox?
07:44:36scorcheassuming it is in a format rockbox supports, yes
07:44:56spkydid the website go down?
07:49:07 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
07:49:28 Quit Nico_P (Remote closed the connection)
07:50:44jhMikeSamiconn: (screendump for the logs)">
07:51:47spkywhen downloading patches, do i download the one at the top of the page? the last one submitted in the replys? all of them?
07:52:30scorchespky: each time a task is updated, the patch goes to the bottom
07:52:59spkyk, that's what i've been downloading, just wanted to make sure
07:55:33toffe82jhMikeS: I was looking something for the lcd of the gigabeat, I found another Toshiba lcd module with the same beginning of the reference
07:55:50 Quit Mouser_X (Read error: 110 (Connection timed out))
07:56:11toffe82jhMikeS: it has a driver with serial interface
07:56:30jhMikeStoffe82: really? where is it?
07:56:45toffe82jhMikeS: check this 2 site : and
07:57:13toffe82it looks like it can be the same or close
07:58:12jhMikeSusually close enough is close enough :) I'm gonna look at the registers and see if it fits.
07:59:34toffe82I will check tomorrow if you find something, good nght
07:59:56 Part toffe82
08:03:32 Quit webguest82 ("CGI:IRC (EOF)")
08:07:07amiconnjhMikeS: ehumm. Is that with just my commit?
08:07:35amiconnmo0ning all
08:08:59jhMikeSamiconn: I didn't apply your patch to what I'm working on. It was fine before.
08:09:21amiconnDoes an svn build work for you?
08:09:34jhMikeSWill check to be totally sure
08:13:05amiconnIt almost looks like you didn't do a full rebuild
08:13:15amiconn(-> bitmaps have wrong format)
08:14:33jhMikeSI didn't do it with make clean, no.
08:16:06jhMikeSI wasn't aware you were actually changing all that yet
08:17:13amiconnOh, and reconfigure
08:17:17 Join Isolinear [0] (
08:17:27jhMikeSah, yes
08:17:49amiconnI changed that so that all of iPod color, iPod Nano, H10 big and small all now use the same pixel format
08:18:51amiconnWhat's left is just experimentation with the block transfer feature, figuring out how to change the timing and what's the fastest possible timing for each lcd controller
08:19:44jhMikeSwhat if the controller can swap and we don't need RGB565SWAPPED? wouldn't that be more efficient?
08:19:45amiconnThe latter can be taken from the datasheet for HD66879R (big H10, Nano, a fraction of iPod color)
08:20:26amiconnA bit. But then all the mentioned targets should be changed
08:21:29jhMikeSI guess figuring that one out is like poking at the IIS stuff
08:21:33amiconnIn fact performance is the same for RGB565 and RGB565SWAPPED wherever precomputed colours are used
08:22:09amiconnJust places with runtime colour computation are (a bit) slower with swapping
08:22:10jhMikeStrue, but what about video output?
08:22:40amiconnBut the ability to use the block transfer outweighs the extra swapping by far
08:23:47jhMikeSis that some special bridge function?
08:24:44amiconnYes. The colour LCD bridge has 2 sets of data ports, one for single byte (or 16 bit word) transfers, and one for block transfers
08:25:41amiconnThe latter needs some things set up beforehand, e.g. how many pixels to transfer
08:25:52jhMikeSany rectangle can be sent?
08:26:33amiconnThe rectangle is set up using the LCD controller and the single transfer port as normal
08:26:57amiconnThe block transfer is just used for pushing the bulk pixel data over
08:27:21amiconnIt can transfer a maximum of 0x10000 pixels in one go.
08:28:12jhMikeSso you can transfer and do something else in parallel. it's a DMA-like operation?
08:28:46jhMikeSthe processor is blocked?
08:28:56amiconnIt works very similar to the single transfer port. You write to a register (0x70008b00 in this case), then poll a bit in another register for completion
08:29:45jhMikeSso don't poll. decode the next video lines while it transfers the previous and wait for a completion before setting up the next transfer
08:29:47amiconnIt's faster than the single transfer port, it takes 32 bits of data at once, and as we know since yesterday the timing is programmable
08:30:15amiconnIt's *not* dma
08:30:24jhMikeShow can you poll it then?
08:31:02amiconnwhile (!(LCD2_BLOCK_CTRL & LCD2_BLOCK_TXOK)); LCD2_BLOCK_DATA = *addr++; in a loop
08:32:22amiconnAnd of course pixel calculation can be shadowed for yuv - we do that right now on c200 (with the other bridge which can only do single transfers)
08:32:25jhMikeSok, full reconfigure fixed it
08:32:38amiconnnice :)
08:34:10 Quit SirFunk (Remote closed the connection)
08:34:46 Join kubiix [0] (
08:34:54amiconnSome (rough) comparison of the PP lcd bridges: bridge 1 (intended for mono LCDs) transfers ~800KB/s, bridge 2 (intended for colour LCDs) transfers ~3MB/s with single transfers, and ~10MB/s with block transfers
08:35:05amiconnAll at 80MHz
08:35:09 Quit keanu (Read error: 110 (Connection timed out))
08:35:56 Join keanu [0] (
08:36:11amiconnThe block transfer is obviously configurable, Nano and small H10 now do ~10MB/s with svn. iPod color only does ~4MB/s
08:36:17jhMikeSbut you can send up to 64K of data in one send...then you wait for it for finish that by polling?
08:37:02jhMikeSwhich driver should I look at to see it operational?
08:37:23 Quit SkinInd95 ("ChatZilla [Firefox]")
08:37:33amiconnYou can send up to 64K *pixels* in the polling loop I mentioned. Wait for TX ready, then send 2 pixels at once (32 bits)
08:38:23 Quit TMM (Read error: 110 (Connection timed out))
08:39:21jhMikeSso which driver currently does this?
08:39:31jhMikeSbest example
08:39:33amiconnfirmware/target/arm/ipod/lcd-color_nano.c and/or firmware/target/arm/iriver/h10/lcd-h10_5gb.c
08:40:13amiconnThe ipod driver uses in in lcd_update_rect() and lcd_yuv_blit(), the small H10 driver only for lcd_update_rect() yet
08:41:11 Quit atsea-34 (Remote closed the connection)
08:41:20amiconnI just copied large parts from the color, but for H10 there's a possible optimisation. The whole LCD has less than 64K pixels, so we never need more than one transfer to set up
08:41:53amiconnThat optimisation would also be apply to the nano
08:42:08 Join Rob222241 [0] (
08:42:09jhMikeShmmm...I still see it's possible to calculate the next values instead of polling immediately
08:42:12amiconnProbably no change in execution speed, but less code
08:43:04*amiconn wonders whether he didn't make himself understood well today
08:43:29amiconn[08:32:22] <amiconn> And of course pixel calculation can be shadowed for yuv - we do that right now on c200 (with the other bridge which can only do single transfers)
08:43:40 Join lazpete [0] (i=c3b2aaca@gateway/web/cgi-irc/
08:44:15 Quit lazpete (Client Quit)
08:44:57 Quit BigBambi (Read error: 113 (No route to host))
08:45:03 Join Variable [0] (
08:46:33jhMikeSamiconn: ah, so that's what I'm talking about in your words
08:47:30 Join webby [0] (i=c27f0814@gateway/web/cgi-irc/
08:49:55webbyI find the name of the newly introduced function 'previous_music_is_wps' alittle bit disappointing. It's not clear from the name whether it tests or sets something. I think somethink like 'set_previous_music_to' (or without '_to') would be better.
08:50:57jhMikeSThis clock thing is just weird. Interrupts must be disabled when changing the clocking registers and can be reenabled before the PLL wait. Only thing is, after changing sources, some clocks must be skipped before enabling them again or it locks. Stopping the other core and doing that trick has resulted in 0 trouble playing SPC dual-core with scaling operating. Even the screendump was done with dcSPC running at the same time.
08:51:02amiconnjhMikeS: I guess the block transfer uses a fifo, so the first few writes will require no waiting, until the fifo is filled
08:51:51 Quit Isolinear (Read error: 110 (Connection timed out))
08:51:57jhMikeSanother 16-slot FIFO? they're usually memmapped and visibile in a dump.
08:53:08amiconnNot memmaped. (long)0x70008b00 is replicated throughout the 0x70008bXX range
08:53:36amiconn(that's the dump I requested last night)
08:54:09amiconn0x0aXX are the bridge control registers, 0x0b00 is the block data port
08:54:14jhMikeSI2S has exactly that property as well but FIFOs are memmapped from IIS_FIFOWR to IIS_FIFOWR+64
08:54:53 Quit webby ("CGI:IRC")
08:55:55jhMikeSThe gap between IIS_FIFOWR and IIS_FIFORD is all of IIS_FIFOWR
08:57:06 Join petur [0] (n=petur@rockbox/developer/petur)
08:57:54jhMikeSOther mirror: GPIO ports are repeated every 512 bytes (IIRC) _except_ GPIOL.
08:58:43 Quit qwm (Remote closed the connection)
08:58:47 Quit kubiix ("Miranda IM! Smaller, Faster, Easier.")
08:58:56 Join qwm [0] (
08:59:43 Quit Rob2222 (Read error: 110 (Connection timed out))
08:59:56amiconnYeah, many (most?) register addresses on PP are just partially decoded
09:00:26amiconnCheck a dump of the ATA register range - lots of repetition too
09:00:42amiconnAnd it's different between PP5020 and PP5022 (!)
09:02:39 Quit qwm (Remote closed the connection)
09:03:25 Join ender` [0] (
09:03:34 Nick bb_ is now known as bb (n=bb@unaffiliated/bb)
09:03:55 Join qwm [0] (
09:06:17jhMikeSok, this DC stuff should be ready soon. the scaling with skipping needs to be cleaner though. if PP5002 doesn't like SWP, then that's no problem either.
09:09:18***Saving seen data "./dancer.seen"
09:10:23 Join nave7693 [0] (
09:10:54nave7693hi is the bootloader for the sansa safe to use now?
09:11:02nave7693like it can load OF
09:11:54jhMikeSnave7693: it should
09:12:29nave7693jhMikeS: guess I could try
09:13:16 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
09:14:03jhMikeSyou'll need to build it yourself if sansapatcher hasn't been updated
09:14:14 Join lee-qid [0] (
09:14:33nave7693is the 3.0 bootloader the lastest?
09:14:42nave7693just updated it from the location in the manual
09:15:18nave7693jhMikeS: I gave up self-compiling everything a long time ago heh
09:15:31nave7693thanks to the build server
09:16:08jhMikeSIt sounds like it was updated
09:16:23GodEater_isn't 3.0 the sansapatcher version ?
09:16:33GodEater_I didn't think we were up to version 3.0 bootloaders on the sansa yet ?
09:16:48nave7693well it says patcher 0.5 with 3.0 bootloader
09:16:54nave7693I don't know how the versioning is done
09:16:57nave7693but it works
09:17:02GodEater_ah interesting
09:17:28 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
09:17:47nave7693jhMikeS: know any good apps (non-command line) to convert videos for rockbox on the sansa?
09:18:25jhMikeSno, sorry
09:20:14 Part nave7693
09:21:20amiconnNerovision works (if you have that)
09:21:49lymecaWhen converting from HFS+ to FAT32...
09:22:26lymeca...the partition table file is written to the HDD using dd (using GNU/Linux here) with what output destination?
09:22:31GodEater_we should change the name of the process from "converting" to "reformatting"
09:22:33jhMikeSI had a demo copy of nero once
09:22:37lymecaaaa /dev/sdc is my iPod
09:22:57lymecaSo dd if=blah.bin of=/dev/sdc
09:22:57GodEater_lymeca: use /dev/sdc then
09:23:33lymecaGodEater_ I did that and then I ran mkfs.vfat /dev/sdc2 and apparently it worked because I installed ROckbox and put a couple albums on that large FAT32 partition and it worked
09:23:41lymecaBut the sdc2 partition seems to only be ~6.8GB large
09:23:51GodEater_lymeca: which ipod do you have ?
09:23:59lymeca30GB videoi
09:24:10lymecaand fdisk says this: /dev/sdc1 1 3648 29302528 83 Linux
09:24:22GodEater_pastebin please
09:24:23lymecaWhich is odd because it doesn't even report an sdc2
09:24:23Lloreanmkfs.vfat has shown some problems on iPods hasn't it?
09:24:39GodEater_Llorean: on a given partition size yes
09:24:42lymecaWhat should I use to create the filesystem then?
09:24:53LloreanGodEater_: I thought it was on both 5.5Gs? I could be wrong.
09:24:56GodEater_lymeca: did you read the whole wiki page ?
09:25:04lymecaGodEater_: Yes
09:25:09GodEater_Llorean: the wiki page I think only mentions the size on the 80GB
09:25:16LloreanAh, okay
09:25:19GodEater_lymeca: what other tool does it mention then ;)
09:25:27lymecaThe wiki page says not to use mkfs.vfat for partitions over 30GB
09:25:34LloreanIt'll be nice when iPodpatcher can handle the conversions itself
09:25:36lymecaIt mentions Mac OS X commands mostly
09:26:06GodEater_Llorean: linuxstb can't do the 80GB conversion
09:26:20LloreanOh, didn't hear about that. Why not?
09:26:20GodEater_and he's doubtful about the 5.5G 30GBs too
09:26:30GodEater_he doesn't understand large sector FAT
09:26:31*jhMikeS looks and sees he's been working on this crummy kernel for two months (on the 2nd attempt) :\. At least this one's a success. :)
09:26:38LloreanI thought we were going to use code from fat32format
09:26:41GodEater_he is
09:26:50lymecanewfs_msdos -F32 -v iPod /dev/rdiskNs2
09:26:56lymecadefinitely can't run that under GNU/Linux
09:27:02GodEater_lymeca: which page are you reading ?
09:27:25GodEater_lymeca: the bottom of the page lists another tool
09:27:25JdGordonjhMikeS: thats what you said about the last attmept also :D
09:27:46lymecaIt says to use that for 80GB 5.5 Gen iPods though
09:27:48LloreanGodEater_: Well then, not to say it's "simple" but isn't it "as simple as" figuring out which command line for fat32format works for them, and then copying whatever code is executed when those parameters are used, approximately?
09:27:55GodEater_Llorean: when last I spoke to him about it, linuxstb said he was hoping Zagor could help with the larger sector file systems
09:27:59LloreanI mean, I seem to remember people using fat32format with 80gb ones.
09:28:28Zagorwhat's the issue with large sectors? I thought we fixed that already?
09:28:44LloreanZagor: For converting HFS+ to FAT32 in preparation for Rockbox installation.
09:28:58GodEater_Zagor: linuxstb is trying do this in ipodpatcher
09:29:01Zagorah, there's no linux or windows formatting tool that uses large sectors?
09:29:19GodEater_there's a linux one which does
09:29:23GodEater_I'm not sure if fat32format does or not
09:29:32GodEater_from what linuxstb was saying, I don't think it does
09:29:32LloreanI thought it did, but there's decent odds I'm wrong. :)
09:29:36spkydoes evilg ever frequent irc?
09:29:41GodEater_or *if* it does, he doesn't understand it
09:29:42ZagorGodEater: ipodpatcher should format the ipod? sounds a bit scary to me...
09:29:45GodEater_spky: not really
09:30:00LloreanZagor: Only HFS+ iPods, which must be formatted anyway
09:30:02GodEater_Zagor: it's less scarey than getting newbs through using dd / mformat
09:30:08spkywhat would be the best way to contact him? i tried sending him an email but have gotten no response
09:30:08LloreanZagor: The idea is to no longer need to ask people to use DD
09:30:11LloreanErr dd
09:30:16*Llorean always wants to capitalize that.
09:30:19jhMikeSJdGordon: no, this is different in a big way :)
09:30:25GodEater_spky: that or the forums is the only way I know of
09:30:58jhMikeSthe demons plaguing the PP5020 are slain finally
09:31:00GodEater_the only linux tool we know which does it and works reliably is mformat
09:31:04GodEater_but the code base is ENORMOUS
09:31:07LloreanZagor: The goal is that you can hit "i" in ipodpatcher and not have to care if you have a macpod or winpod.
09:31:08GodEater_and very complicated
09:31:21GodEater_mkfs.vfat is unreliable
09:31:23JdGordonall hail jhMikeS the pp slayer!
09:31:23*Llorean cheers in jhMikeS's general direction.
09:31:28GodEater_as lymeca has found
09:31:40JdGordonjhMikeS: so dual core kernel?
09:31:40lymecaWhat can I use instead?
09:31:40jhMikeShehe *bows*
09:31:43ZagorLlorean: with lots and lots of "are you sure?" I presume
09:31:44 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
09:31:51GodEater_lymeca: again, from the wiki page, mformat
09:32:04GodEater_Zagor: of course ;)
09:32:05jhMikeSJdGordon: served up with a side of beans or salad depending
09:32:28JdGordonthat doesnt soo apetizing...
09:32:34 Nick fxb__ is now known as fxb (
09:32:34LloreanZagor: Even one "are you sure" is better than "oops, I dd'ed an iPods MBR to my system disk and didn't realize it until the next day."
09:32:37GodEater_I'll take the beans
09:32:47GodEater_Llorean: hahahahaha
09:32:51ZagorLlorean: heh, yeah
09:33:02 Nick fxb is now known as fxb__ (
09:33:15LloreanI'm not sure whether nobody's done it, or whether the people to have just have been too shamed to come complain about it.
09:33:24 Quit JdGordon ("have a good weekend all")
09:33:30GodEater_Llorean: "ps: am typing from an internet cafe. plz help. my pc is teh broken"
09:33:39jhMikeSit has beer and a 0x8a905617 in there too :p
09:34:01LloreanGodEater_: I should prepare a special badge for the first person to "brick" their PC with a Rockbox install. ;)
09:34:02 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
09:34:29GodEater_Llorean: haha - yeah
09:34:47lymeca# mformat -t 2484 -h 255 -s 63 -S 2048 -M 2048 -F x:
09:34:47lymecaargssize must be less than 6
09:34:48*jhMikeS wants to make a DAP virus that infects the host
09:35:18 Join massiveH [0] (
09:35:36GodEater_lymeca: that particular command line is for the 80GB
09:35:43GodEater_lymeca: you will have to figure out the one for the 30GB
09:39:06pixelmaisn't it important if it's a 5g or 5.5g Ipod Video too - was that said somewhere and I overlooked that?
09:39:15 Join RaZorbacK_ [0] (
09:39:42 Quit RaZorbacK_ ("bye all")
09:40:31GodEater_pixelma: yes - that's very important too
09:40:44GodEater_and no - I don't think it was mentioned
09:41:52Zagor5g is 512 bytes/sector and 5.5g is 4096 ?
09:42:01Zagoror was it 2048?
09:44:11Llorean5.5G is 2048 on the 80gb.
09:44:20GodEater_and on the 30GB too
09:44:22LloreanIsn't it 512 on the 30gb, but presents itself to the host as 2048?
09:44:28LloreanOr was it 2048 but presents itself as 512?
09:44:44GodEater_both the 5.5G's present to the host as 512
09:44:52GodEater_the difference is when you run natively
09:45:14LloreanI thought the 30gb's drive was internally 512, which was why Rockbox could be run on it without all the adaptations we had to make for the 80gb?
09:45:32GodEater_I mistyped that
09:45:36GodEater_both the 5.5G's present to the host as 2048
09:45:48GodEater_but my follow up line was still correct
09:46:29*GodEater_ still hopes someday for someone to manage to explain to him in short words how all the different sector sizes work
09:47:08GodEater_I still don't get how it's 2048 over USB, but 512 on the device, but how you have to read two at a time for it to actually work
09:47:47LloreanI clearly know nothing about such things, because my head tells me it should be impossible to format it.
09:47:48lymecaOh wait maybe I got the wrong MBR
09:48:06lymecaI got the 2048-byte sector one
09:48:13GodEater_lymeca: I thought you said you'd already DONE this and got it working ?
09:48:26lymecaI should probably get the 30GB 512-byte sector
09:48:39lymecaI get confused
09:48:44lymecaI have five iPods onmy desk
09:48:53GodEater_lymeca: you're doing this from linux right ?
09:49:00lymecaI go to college and word gets out about my ability to install Rockbox onto people's iPods
09:49:09lymecaso I get a constant influx of them now
09:49:22GodEater_so do an "fdisk -l"
09:49:29GodEater_and it'll TELL you which sector size you need
09:52:09 Quit jhMikeS (Nick collision from services.)
09:52:15 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
09:53:16jhMikeSamiconn: all the SH targets indeed use the same processor model?
09:53:24 Quit Ebert ()
09:56:50LinusNjhMikeS: yes
09:58:01 Nick fxb__ is now known as fxb (
09:58:06 Join kretender [0] (
09:58:07jhMikeSgood, then no odd suprises later :)
10:06:51 Join CaptainSquid [0] (
10:14:42 Join Sedgewick [0] (
10:14:53spkyshould make the wps layouts structured like html, would make things alot easier to manipulate
10:16:12pondlifeHmm, why doesn't configure allow selection of (V)oice until you go into (A)dvanced? I'm sure it used to...
10:18:57 Join PaulJam [0] (
10:19:00 Quit pondlife ("disconnected has pondlife")
10:19:30 Join obo [0] (n=obo@rockbox/developer/obo)
10:19:48GodEater_spky: would make the parser much more complicated
10:26:48amiconnjhMikeS: Yes, just slightly different clock frequencies
10:27:10Zagora request for all disassembler ninjas: find the interrupt used by usb on pp502x. we're currently polling on the timer interrupt and that's not good enough
10:29:04*jhMikeS wonders if not dithering on gigabeat has any point after doing this kernel thing. it no longer drops frames during buffering like before.
10:29:43amiconnSomeone might prefer undithered output
10:29:54 Quit hcs ("Leaving.")
10:30:29jhMikeSperhaps, but this code is a good deal faster and I only really added the option because of that problem
10:31:26jhMikeSI guess it can with pushing the bitrates up
10:32:46 Nick fxb is now known as fxb__ (
10:33:42 Join TMM [0] (
10:34:02linuxstbLlorean: Yes, the problem is that I incorporated the fat32format into ipodpatcher, and fat32format only handles 512-byte sectors. So once it's working reliably for 512-byte sectors, someone who knows FAT better than me will need to adapt it to 2048-byte sectors.
10:34:22TMMisn't vfat pretty trivial?
10:35:05 Join spiorf [0] (
10:36:00LinusNTMM: compared to what?
10:37:11TMMwell, pretty much any other FS I know
10:37:18TMMwell, FAT12... that's probably easier
10:37:44linuxstbTMM: I'm not saying it's complicated, just that I know nothing about FAT...
10:37:53TMMminixfs is fairly easy as well
10:38:43peturFAT12 is actually more complicated - the way they mix sectors...
10:38:52linuxstbI also don't have one of these troublesome 2048-byte sector ipods to test with.
10:39:23 Join bluebrother [0] (i=ppg1HDBX@rockbox/staff/bluebrother)
10:39:24TMMpetur, that depens on the media, really
10:39:38amiconnlinuxstb: Would a test need a hfs formatted one to start with?
10:39:53peturno, it's the way the sectors are stored in the FAT, needing 1,5 byte
10:39:53 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
10:39:55midkayamiconn: "immeasurable speedup on iPod Color" - immeasurably large or immeasurably hardly-makes-a-difference?
10:40:28TMMpetur, you can munge the sectors in any way you like. The DOS format utility does some magic to speed up floppy access though
10:40:30linuxstbamiconn: Ideally, but just implementing a working formatting function would be 95% of the work.
10:40:41Lloreanmidkay: I'm pretty sure it's "immeasurable because it's too small"
10:40:51TMMpetur, which is the sector mixing you are referring to
10:40:58midkayLlorean: hey, a guy can dream :)
10:41:00 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
10:41:33pondlifeHi Nico_P
10:41:35*linuxstb remembers his g/f has a 2nd Gen Nano, so that could be used to test (it has 2048 byte sectors).
10:41:54pondlifeHow's the skipping, I didn't see it mentioned on your git.
10:42:15Nico_Pit's working fine
10:42:32peturTMM: for each sector, there are two bytes in the FAT, one containing only 4 bits (the other 4 belong to another sector
10:42:40pondlifeI tried a git pull origin mob, but it gave me conflicts !
10:42:41Nico_Pbuffering isn't optimal, but that's just a matter of tweaking the main loop in buffering_thread
10:42:52pondlifeSo the back-skip-crash is fixed?
10:42:54Nico_Ppondlife: yeah, I'm sorry it doesn't work well when I rebase
10:43:08pondlifeOK, I'll start anew
10:43:14Nico_PI'm not sure, I didn't have crashes, but maybe you didn't test the right version
10:43:19Nico_Ppondlife: no need to clone again though
10:43:29TMMpetur, hum, really?
10:43:40amiconnmidkay: Hardly makes a difference. Only the command writes are optimized. Data transfers (the vast majority) already were
10:43:59Nico_Ppondlife: just remove your mob branch: git branch -D mob, then git checkout -b mob origin/mob
10:44:22Nico_Ppondlife: and on master you can git pull to update too
10:44:24midkayamiconn: growl.. you've shattered my hopes and dreams!
10:44:56amiconnmidkay: iPod color *will* receive a significant speedup soon - as soon as I figure out the block transfer config registers
10:45:03amiconnExpect a factor of 2....
10:45:04Zagorlinuxstb: have you considered simply integrating the rockbox fat driver?
10:45:07pondlifeNico_P: "fatal: you need to resolve your current index first"
10:45:07pixelmabluebrother: I told "John" exactly the same in reply to his post in the "Have you read the rockbox manual" thread ... but I guess it doesn't hurt to tell him another time...
10:45:14midkayamiconn: does that include iPod 5G as well?
10:45:18linuxstbZagor: I just need a formatting function.
10:45:36pondlifeNico_P: that's when I attempt to checkout
10:45:37Nico_Ppondlife: ah, that's the leftovers from the failed pull... git reset −−hard mob should do
10:45:40Zagoryeah well the relevant portions of it, I mean
10:45:50Nico_Por git reset −−hard master
10:45:51bluebrotherpixelma: ;-)
10:45:58amiconnmidkay: Nope. As said, color, not video
10:46:06amiconnThe video uses yet another hookup
10:46:33midkayamiconn: sorry, your "color" terminology is a little confusing. at least between your commit message and usage here.
10:46:34amiconnI.e. none of the PP lcd bridges
10:47:00Nico_Ppondlife: I tried your tracks... quite funny :) and no problems
10:47:08amiconnmidkay: "color" for me always refers to the iPod color. For the opposite of monochrome, I always use "colour"
10:47:19pondlifeNot my tracks. I am european :P
10:47:26midkayamiconn: haha, that's a nice system, i'll have to remember it :)
10:47:35Nico_Ppondlife: yeah, the ones you sent me
10:48:33pixelmaamiconn: to make it more obvious - maybe even use "Color" and "colour" ;)
10:48:37midkayamiconn: well, sucky about the video. do you think there's much room for optimization on the video's LCD driver?
10:48:49amiconnI have no idea (yet)
10:49:11pondlifeHmm, git reset −−hard seems to work, but all others fail in some way
10:49:16amiconnIt looks like the broadcom chip is hooked up directly to the external data bus of the PP
10:49:47pondlifeNico_P: Lots of "CONFLICT (content): Merge conflict in apps/playback.c" type errors
10:49:47Nico_Ppondlife: git checkout master; git reset −−hard; git branch -D mob; git checkout -b mob origin/mob
10:50:03*jhMikeS just killed main thread accidentally while looking at OS stacks on GB :p that won't do nice things.
10:50:08Nico_Pthat should do the job
10:50:25amiconnWe might have to use the second broadcom firmware image (and figure out *how* to do that first)
10:51:09midkayamiconn: wow, that must be quite a task..
10:51:47LloreanYeah, I thought the big bottleneck on the iPod Video was how fast we can get the broadcom chip to dance for us.
10:52:05*linuxstb wonders if jhMikeS's multi-core threading could handle two CPUs of different architectures ;)
10:52:05LloreanI seem to recall it at least being the bottleneck when we first started.
10:52:05midkaywe're not using it at all, right?
10:52:20Lloreanmidkay: I believe we're using it in a very minimal sense, since we have to use it to get to the LCD at all
10:52:47midkayLlorean: really? i didn't know that.. i thought it was left unused altogether. interesting.
10:53:06amiconnWe get ~5.5MB/sec on the G5, and ~4MB/sec on the iPoc color. The latter should allow for ~10MB/Sec, like the Nano (and now the small H10) already do
10:53:30amiconnAssuming the broadcom chip can handle ~10MB/sec too, that would be a nice speedup...
10:53:46linuxstbI didn't realise the Nano was operating much faster than the Color...
10:53:46Lloreanamiconn: Should this increase, for example, mpeg2 framerate in any measurable way with the move to ~10MB/sec?
10:53:56Lloreanlinuxstb: I think that's of his most recent change.
10:54:07midkayit sure would be a nice speedup.. i wonder if it'll ever manifest itself :)
10:54:08amiconnLlorean: Nope
10:54:15amiconnThe nano already was that fast
10:54:34LloreanIt was using the block transfers before?
10:54:44LloreanOr is that unrelated?
10:55:15amiconnYes it used block transfers before, as well as the color
10:55:33amiconnBut the config regs are set up differently, hence the speed difference
10:55:46amiconn[10:44:55] <amiconn> midkay: iPod color *will* receive a significant speedup soon - as soon as I figure out the block transfer config registers
10:55:47jhMikeSlinuxstb: thinking about it, possible
10:56:15linuxstbjhMikeS: OK, so all we need to do is figure out the architecture of the Broadcom chip...
10:56:20amiconnIf we know how they work, this will allow for extra speedups when not boosted
10:56:44jhMikeSlinuxstb: shouldn't be much trouble ;)
10:57:01Lloreanamiconn: Would that then explain why the H10 was getting worse FPS in mpegplayer than the Nano?
10:58:12jhMikeSlinuxstb: so Video would be a 3 core target then?
10:58:48linuxstbjhMikeS: 3 cores and a DSP (in the Broadcom chip)
10:59:11jhMikeSyou execute code on the broadcom chip?
11:00:21linuxstbWe don't, but Apple do. It appears to run it's own firmware, receiving commands and data from the PP.
11:05:32 Quit lee-qid (Read error: 110 (Connection timed out))
11:07:03amiconnLlorean: It doesn't anymore for quite a while... that was a different problem
11:07:39jhMikeSlinuxstb: there's 2 arm cores and 3 + DSP in the BC chip and you need to load it with a firmware yourself?
11:08:37linuxstbjhMikeS: No, the broadcom chip has 1 core and 1 DSP. 3+DSP is the whole ipod. And yes, IIUC, the host processor uploads a firmware (there is a small one stored in the flash ROM, and a larger one on disk in the firmware partition).
11:09:21***Saving seen data "./dancer.seen"
11:11:00 Quit petur ("brb")
11:12:34jhMikeSwhy would the scheduler for PP cores need to deal with another type of processor simultaneously in that case? it would be its own instance with some type of interface in place.
11:17:30 Nick fxb__ is now known as fxb (
11:17:36 Join barrywardell [0] (
11:18:05 Quit idnar (Nick collision from services.)
11:18:07 Join idnar_ [0] (i=mithrand@unaffiliated/idnar)
11:18:13 Quit Toxicity999 (Read error: 110 (Connection timed out))
11:18:17linuxstbjhMikeS: I wasn't completely serious...
11:18:53*linuxstb needs to smile more
11:19:04jhMikeSI'm just wondering what sort of setup would be needed...and I'm really sleepy
11:19:40linuxstbIf we knew enough, I expect we could even run Rockbox on the Broadcom chip, reversing how Apple do it...
11:20:47LloreanDo we even know roughly how fast the broadcom chip is?
11:21:29linuxstbFast enough to decode 640x480 h264...
11:21:49jhMikeSit's an arm core too or something unique?
11:22:33linuxstbI think it's something unique - Broadcom don't advertise the architecture, and all I've heard about it is that it could be MIPS-based.
11:23:43jhMikeStry different disassemblers on the blob until one makes sensible output?
11:25:01linuxstbAssuming it's not encrypted in some way...
11:26:23jhMikeSI'd expect it to compress less than a typical executable if it is
11:27:08linuxstbBut I'm not really interested in working with a chip even less well documented than the PP....
11:27:36amiconnHow can something be less documented than something undocumented?
11:28:02linuxstbThe PP was partially documented by the IPL project, and has a well-documented arm7tdmi core.
11:28:16linuxstbPlus the USB we've discovered is documented.
11:28:44linuxstbI mean I didn't have to start from scratch with the PP.
11:30:36pondlifeNico_P: Yep, skipping backwards no longer crashes here.
11:30:40 Quit qwm (Remote closed the connection)
11:30:54jhMikeSsomeone did _or_ they had a leak
11:30:55 Join qwm [0] (
11:30:55Nico_Ppondlife: nice. I don't know when I fixed the prob though...
11:31:16pondlifeRewinding during the last few seconds of a track confuses the WPS though
11:33:14Nico_Pthat's probably still the same problem as the progressbar
11:33:28Nico_PI haven't found a solution for that btw
11:33:42pondlifeSimilar timing, but probably not the same
11:34:16jhMikeSNico_P: you took out the callback from the ISR?
11:34:17pondlifeBasically the WPS doesn't get told to redraw for the next track (or maybe even any subsequent tracks)
11:34:32Nico_PjhMikeS: disabled it, yes
11:35:34Nico_Ppondlife: how much is "last few seconds"
11:35:35jhMikeSah. perhaps tagging the pcm buffer somehow with timestamps and always using that.
11:35:52 Join ddalton [0] (
11:35:58pondlifeNico_P: Probably once playback has moved onto the next track
11:36:42Nico_Pso that's once the progressbar has emptied
11:36:50pondlifeNo, before that
11:37:26Nico_Phmm I've reached a state where the WPS still shows the same track but the audio is the next
11:37:29Nico_Pis that it ?
11:38:02pondlifeAlso, if you rewind after the progressbar has stopped moving you can get into a state where no playback is occurring, although nothing crashed
11:38:03Nico_Pit's solved by hiding the WPS
11:38:26jhMikeSyeah, lets get rid of that bloat :P
11:38:27Nico_Pit's the WPS not refreshing correctly
11:39:32pondlifeI can't repro the "stalled playback" state though
11:40:08Nico_PI think I've had it but never consistently
11:40:15 Quit ddalton (Client Quit)
11:40:19 Join ddalton [0] (
11:41:10ddaltonOnce I have studied my c book I want to write a few patches. What features do people want? (I am blind so I couldn't do anything with images yet):-)
11:41:28ddaltonmaybe voice playback recording plugins
11:41:37ddaltonand buttons
11:41:50pondlifeI want Viewports to be honest...
11:42:06pondlifeNot much use for the blind, but it'll simplify stuff
11:42:13ddaltonwhat are they for?
11:42:52pondlifeTo allow parts of the screen (status bar?) to be updated independently of the main screen, for example.
11:43:03ddaltonoh ok
11:43:10 Join kubiix [0] (
11:43:29ddaltonSo what exactly is one of the hardest things coding on rockbox as a new comer to programming?
11:44:45pondlifeErm, playback.c ?? :)
11:45:11pondlifeDual-core, race conditions..? lots of fun stuff
11:45:13Nico_Ppondlife: is the WPS refresh bug is present SVN too ?
11:45:38 Join spiorf_ [0] (
11:45:41tumuddalton, reading the source as there is little of source documentation :)
11:45:44pondlifeBut I was more concerned about the stalled playback
11:45:48barrywardellamiconn: nice work on the LCD bridge. anything I can do to get things working on H10?
11:45:57Nico_POK... I think I'll focus on regressions from my code before trying to fix all the old bugs
11:46:21Nico_Ppondlife: I'm very interested in a way to reproduce the stalling
11:46:26 Quit spiorf (Read error: 110 (Connection timed out))
11:46:31amiconnbarrywardell: You could experiment with 0x70008a00 and 0x70008a04 registers
11:46:54amiconnI'll do a regdump from small H10 tonight, then I have 4 dumps for comparison
11:47:04barrywardellany suggestions what I should set them to?
11:47:28pondlifeNico_P: Yes, maybe document a load of use cases on the wiki and have columns for "works on old" and "works on MoB"
11:47:37Nico_Pgood idea
11:47:51pondlifeYou can put a few known bugs in and end up with more ticks in the MoB column
11:49:23 Join petur [0] (n=petur@rockbox/developer/petur)
11:50:19 Quit kubiix (Read error: 104 (Connection reset by peer))
11:51:00pixelmapondlife: more ticks for known bugs or fixed bugs? ;P
11:51:06 Nick vosque_away is now known as vosque (
11:51:25 Part vosque
11:51:33Nico_Ppondlife: if you want to start a page, feel free ;)
11:51:36pondlifeJust a tick if the use case is tested ok
11:51:43Nico_PMetadataOnBufferStatus ?
11:51:57Nico_Por just modifying SoftwareCodecPlayback could do
11:52:02pondlifeWhy not put it on the current page? Or is that getting too big?
11:52:27Nico_PI'd like the MetadataOnBuffer page to stay about the project
11:52:43pondlifeAh, GSoC style...
11:52:45Nico_PI think the status doesn't belong in it
11:52:58pondlifeI'd say this is about the project, until it's in SVN
11:55:30 Join advcomp2019_ [0] (n=advcomp2@
11:57:43 Nick Tanuva|off is now known as Tanuva (n=tanuva@
11:57:48linuxstbHmm, seems Nero are now claiming that faad2 is GPL-compatible again -
11:57:49Nico_Pthe page is meant to be kept up to date even after I commit MoB
11:58:29linuxstbBut not everyone agrees...
12:00:05Zagoran advertising clause is definitely not gpl compatible
12:00:29Nico_Ppondlife: I feel the project page (MetadataOnBuffer) should end with "committed to SVN, see the status page..."
12:01:42linuxstbZagor: Yes, but isn't Nero's argument that programs must display a (C) notice anyway?
12:02:18ZagorI don't know. it's a false argument. programs have to do no such thing.
12:02:58linuxstbZagor: 2c) states some conditions where they must... But I agree, not all do.
12:04:24linuxstbBut yes, I don't think Nero's text change helps - I still think it's not GPL compatible.
12:04:37Zagorheh, I've never read 2c in detail it seems :)
12:05:11linuxstbI'm not quite sure what "reads commands interactively" means...
12:05:39 Quit Nico_P (Remote closed the connection)
12:05:46pondlifeNico_P: "Codec failure" and "Couldn't load codec: /.rockbox/codecs/_.codec" (where _ is a block) after all buffered tracks have played and there are more to go...
12:05:57Zagorlinuxstb: I think it means that "ls" does not have to print out copyright everytime you run it. but any program which takes interactive input does.
12:06:07pondlifee.g. with the Cassetteboy album after 44 tracks have played
12:06:13pondlife44/98 that is
12:06:36pondlifeFollowed by a hard crash
12:06:37linuxstbZagor: So how does that apply to something like Rockbox? I would call it "interactive".
12:07:04Zagorgood question
12:07:10linuxstbMaybe we should be extending the "credits" screen to be more explicit about (C) holders and licenses.
12:07:10LloreanI would call Rockbox interactive too.
12:07:26ZagorI'll meditate on it over lunch :)
12:07:31linuxstbEnjoy ;)
12:07:33LloreanWe could just have a "Copyrights" screen too.
12:08:21linuxstbOr "Legal" - I think that's how Apple's firmware calls it, and seems appropriate.
12:08:46linuxstb(it's both (C) and licenses)
12:09:32barrywardellamiconn: my dump of the 0x70008000 is different from what jhMikeS had:
12:13:29 Quit advcomp2019 (Read error: 110 (Connection timed out))
12:17:32 Join omry [0] (
12:17:44 Join jba [0] (
12:20:24omryconnected my new sansa e280 to linux, how do I mount it?
12:20:51omrythis is what I get in dmesg: usb 1-6: new high speed USB device using ehci_hcd and address 4
12:20:58omrybut no mention of a new block device
12:21:17barrywardellmake sure you put it into msc mode
12:22:24 Join Asteriskk_ [0] (
12:23:06GodEater_and that it's running the OF
12:23:12 Quit Rob222241 ()
12:23:43omrycool, thanks.
12:23:44omryworks now.
12:25:38 Nick parafin|away is now known as parafin (
12:25:41omryshould I install the e200 release for e280?
12:26:03barrywardellyes, e200 is the series
12:26:09Asteriskk_where do I find bootloaders for my sandisk? in the current builds/
12:26:26GodEater_sansapatcher installs the bootloaders
12:26:29linuxstbFrom the links in the manual
12:27:45GodEater_pondlife: which git repo is Nico_P publishing his MoB Rockbox build in ?
12:27:57GodEater_git:// <−− ?
12:28:41linuxstbWhen git changes are committed to SVN, is there any way to keep the revision history? Or is always just committed as a single revision?
12:28:58GodEater_I think it'll have to go as a single commit
12:29:06GodEater_I'm not certain though
12:29:08bluebrothergit-svn dcommit commits the original commits individually
12:29:50GodEater_that's if you git repo is connected with the svn repo
12:29:56GodEater_I'm not sure if Nico's done that or not ?
12:31:02bluebrotherRockbox.git has a "git-svn" tag at the top, so I guess it does.
12:31:31linuxstbI'm not saying that it's necessarily a good idea for Nico, just curious...
12:31:41GodEater_I must be looking at the wrong repo =/
12:32:04bluebrotherGodEater_:;h=mob (and scroll down a bit)
12:32:09Asteriskk_will the sansafixer tell me if the version of the boot loader I am running is current or will it just happily over write everything
12:32:25bluebrotherthere is no "sansafixer".
12:32:28linuxstbsansapatcher will just happily overwrite.
12:33:33pondlifeGodEater_: it must be the mob branch, not master
12:33:36barrywardellamiconn: I just did a couple of tests with the Color and Nano versions of 0x70008a00 and 0x70008a04
12:34:09bluebrotherthe mob branch has a bunch of commits after the apparently last sync to svn about a day ago.
12:34:16 Join Rob2222 [0] (
12:34:17 Join Crash91 [0] (n=evil91@
12:34:20barrywardellwith the nano version, I get only the top half of the screen showing the rockbox logo in wierd colour (more yellowish)
12:34:48barrywardellwith the color version, I get something that looks like vertical interlacing
12:34:51amiconnDid you reconfigure + rebuild fully?
12:35:16amiconnYour dump isn't very readable...
12:35:30barrywardellyes, but not between changes to 0x70008a00 and 0x70008a04
12:35:41barrywardellthat dump is just what's different to jhMikeS's version
12:35:43amiconnHmm, ok
12:35:58*amiconn would prefer the raw binary dump
12:36:12amiconnThe pastebin is a mess...
12:36:20barrywardellcan I email it to you?
12:38:04barrywardellis your address available somewhere?
12:38:15 Join jba_ [0] (
12:39:36amiconnIn the wiki, in the ml archive..
12:40:26omrydo I need to backup my sansa to an image before installing rockbox ?
12:41:03bluebrotherholding a backup is always a good idea but installing Rockbox won't delete the data on the player
12:41:18omrywhat about the boot loader?
12:41:37bluebrotherwhat should be about the bootloader?
12:42:02bluebrotherit's in the firmware partition. That doesn't affect the data itself.
12:42:39omryah, the firmware is on it's own partition?
12:42:49omrynice. so I can just create an image for that
12:42:52barrywardellamiconn: sent
12:43:40omryI guess it's the second parition? (the smaller one?)
12:44:03omry/dev/sdb2 1019 1022 20480 84 OS/2 hidden C: drive ?
12:44:53barrywardellyes, that's the firmware partition
12:45:10omryand the boot loader is in it?
12:45:37omryif it's not first, how can the device find it? or does it use the mbr at block 0?
12:45:56bluebrotherby using magic.
12:46:22Asteriskk_I wonder if this is a problem with the player, with the computer or with the firmware, but sometimes when I plug in the c200 I get no response from the computer, i.e nothing turns up, or I get, drive could not be opened due to an IO device error. this is usb 2.0 but I only have usb 1.x, is this the problem?
12:46:26omryso I you change the size of the partition the makic breaks?
12:46:54Asteriskk_the only way I find to fix this is to remove the bettery pack and replace it then replug it in
12:47:03barrywardellamiconn: I had a silly mistake in my code. now it work on my H10
12:47:14pixelmaAsteriskk_: yes I saw that on a laptop with usb 1.1 too
12:47:15omryI`ll make a full image as well, just in case :)
12:47:24barrywardellusing the color settings for 0x70008a00 and 0x70008a04
12:47:42Asteriskk_I figured it was probably something like that. this board is like six years old lol
12:47:52barrywardellamiconn: yes bootloader is in there too
12:47:59pixelmaAsteriskk_: I then started the original firmware before plugging usb - that worked reliably
12:48:04barrywardellsorry, that was meant for omry
12:48:15amiconnbarrywardell: Color settings are slow... tried nano?
12:48:40barrywardellI get 39fps at 30MHz, still faster than the wiki says
12:48:42omrybarrywardell, second level boot loader?
12:48:44barrywardelltrying nano now
12:48:46Asteriskk_is it necessary to remove rockbox to start the original firmware or is there a power setting or something
12:49:04linuxstbamiconn: Should I try the Nano settings on my Color?
12:49:29pixelmaAsteriskk_: no, you can dual boot - that is covered in the fine manual ;)
12:49:46*bluebrother points to and leaves
12:50:01 Join mf0102 [0] (n=michi@
12:50:12amiconnlinuxstb: Please do. But in order to switch to these, we'd need a tester with a type 0 lcd...
12:50:15omrydid the manufacturer of any player released api spec? or was it all reverse engineering?
12:50:24 Join karashata [0] (
12:50:25*Asteriskk_ pokes in the manulal once more
12:50:27linuxstbamiconn: Sure.
12:50:57barrywardellomry: the boot sequence on sansa is complicated
12:51:35omrybarrywardell, similar to linux + grub in any way?
12:51:56omrymbr -> second level -> firmware ?
12:52:05barrywardellamiconn: nano settings don't work. I get two copies of the startup splash 1/4 screen size and in weird colours. then no further lcd updates
12:52:11amiconnlinuxstb: It would prbably be helpful if "View HW Info" would show the lcd type on Color (maybe it already does)
12:52:22linuxstbI think it does.
12:52:25amiconnbarrywardell: Yeah, then the timing is too fast
12:52:30GodEater_omry: not really
12:52:49linuxstbbarrywardell: What are the values I should be writing?
12:52:51GodEater_omry: flash -> bootloader on firmware partition -> firmware
12:53:43omryGodEater, I see. so since the flash does not change, backing up the firmware partition should be good enough to recover.
12:53:55GodEater_omry: yep
12:54:09barrywardelllinuxstb: 0x01100003 to 0x70008a00
12:54:14 Quit jba (Read error: 110 (Connection timed out))
12:54:20barrywardelland 0x1f140000 to 0x70008a04
12:54:40omrycoz creating a mostly empty full image is slow :)
12:56:05linuxstbbarrywardell: Are you sure? Reading preglow's dump, I thought they should be 0x00030110 and 0x00001f14
12:56:09amiconnbarrywardell, linuxstb: Wrong. PP is little endian...
12:56:21amiconnlinuxstb: exactly
12:56:42barrywardelloops, good point
12:56:47barrywardelltime to try again!
12:57:44 Join n1s [0] (
12:58:15omrydoes rockbox supports fm radio?
12:58:21omry(for e280)
12:59:08GodEater_omry: is it in the manual ?
12:59:36omryyeah, so I guess it is.
12:59:38omryit doew
13:00:00barrywardellomry: you should be most worried about backing up your data
13:00:01linuxstbamiconn: Hmm, it just freezes for me - no lcd updates at all.
13:00:30omrybarrywardell, yes. I decided to make a full image anyway.
13:00:35omrybetter safe than sorry.
13:01:04linuxstbamiconn: Do you think it matters where in lcd_init_device() I set those registers?
13:01:29barrywardellamiconn: same for me - no lcd updates with the nano settings
13:08:06amiconnEither too fast then, or some other bits are wrong -> needs research
13:09:25***Saving seen data "./dancer.seen"
13:09:55barrywardellcolor settings work, but actually give a slowdown 39->36fps
13:09:59preglowi don't like the fact that nano has no init :/
13:10:17barrywardellthe LcdFrameRate page needs updating for H10
13:10:40*linuxstb passes preglow a copy of the Nano bootloader and IDA
13:11:00 Quit Rob2222 ()
13:12:41barrywardelllinuxstb: you wanna try the H10 values? 0x00000000 and 0x00001f22
13:12:58preglowi wish there was something like ida somewhere
13:13:01preglowopen source, of course...
13:13:10*linuxstb passes preglow a copy of vi ;)
13:13:21preglowi currently use objdump + vim
13:13:24preglownot optimal...
13:13:43linuxstbI wonder how the IPL people are getting on with their disassembler...
13:14:33 Join The-Compiler [0] (
13:14:39preglowwhat, they haven't admitted defeat yet and come over to our side? :P
13:15:39 Quit Crash91 ("Bye Bye!")
13:15:54barrywardellpreglow: the disassembler in svn is marginally better than objdump in ways
13:16:16preglowin which ways?
13:16:17pondlifebluebrother: Is RbUtilQt meant to support SAPI for voicing?
13:16:49amiconnbarrywardell: The Big H10 values very probably won't work
13:17:03 Join Toxicity999 [0] (n=bryan@unaffiliated/Toxicity999)
13:17:04linuxstbpreglow: It will tell you the value of [rN, #nnnn] for example in instructions like "ldr r0, [r0, #34a]"
13:17:35linuxstbI mean [pc, #0x34a]
13:17:46barrywardellbut it doesn't support thumb code yet
13:18:00 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
13:18:39barrywardellamiconn: why not?
13:18:56preglowcool, objdump only tells you the calculated offset
13:19:13*barrywardell just got a ~30% speedup on H10 lcd_update :)
13:19:57barrywardellit seems the block writing works, even without using the color values for lcd_base
13:20:00Asteriskk_seems there should be a better way to make voice files than running a whole new os to do it:-/
13:20:04barrywardellkinda anyway
13:20:26linuxstbAsteriskk_: ?
13:21:04Asteriskk_wellt he wiki says the vbs script is out of date and to install linux to do build your voice files
13:21:16barrywardellactually, I get a ~3x speedup
13:21:31barrywardellbut I get a garbled display at 80MHz
13:21:43linuxstbAsteriskk_: I don't know the details, but I think that wiki page is out of date. You can definitely build voice files in Windows.
13:22:03barrywardelland I get the weird quarter sized splash on startup, but then it fixes itself
13:22:06pondlifeAsteriskk_: The wiki is correct I think, Cygwin will allow it
13:22:30Asteriskk_like I said, seems a lotta work to install a new os just to build a 1 meg file lol
13:22:37pondlifeNew OS? Hardly
13:23:42 Quit Nico_P (Remote closed the connection)
13:23:56 Join massive_H [0] (
13:23:57 Quit jba_ (Read error: 110 (Connection timed out))
13:24:15barrywardellah, the quarter sized splash is because it runs at 80MHz on startup
13:24:40 Join desowin [0] (
13:25:12*barrywardell thinks linuxstb should try the H10 values, unless amiconn has a reason not to...
13:25:17 Quit ddalton (Read error: 110 (Connection timed out))
13:27:01 Part pixelma
13:30:49 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
13:30:57 Quit Nico_P (Remote closed the connection)
13:30:58 Quit massiveH (Connection timed out)
13:31:14 Part Asteriskk_
13:31:18linuxstbbarrywardell: OK, I'll try now. What should I be expecting?
13:31:29barrywardellfaster lcd updates
13:31:46barrywardellas in 3.5x faster
13:32:13linuxstbBut wasn't the H10 very slow to start with?
13:32:25linuxstbBut I'm testing now...
13:32:49barrywardellno, I have it running faster now
13:33:16linuxstb67.0fps fullscreen at 30MHz
13:33:29barrywardelljust changing 0x70008a00 from 0x1f10 to 0x1f22 makes a 3x speedup
13:33:51barrywardellif the wiki is accurate, that's a 3x speedup for you too?
13:34:08linuxstbYes, the UI now feels extremely fast with the CPU unboosted ;)
13:34:36omryhow do I switch to msc mode from within e200 rockbox?
13:34:40*barrywardell thinks amiconn would be interested in these results
13:34:59barrywardellomry: there is no usb support in rockbox yet, so you have to use the original firmware for that
13:35:15barrywardelllinuxstb: do you have any problems when boosted?
13:35:20omrybarrywardell, so how do I upload audio file?
13:35:42linuxstbbarrywardell: I don't think so - I've just tested mpegplayer successfully.
13:35:53barrywardellomry: turn your sansa off, then connect the usb cable and it will start the of and connect
13:36:18 Join ddalton [0] (
13:36:56linuxstbbarrywardell: Looking good.
13:37:23barrywardellmust just be the H10 lcd that has problems when boosted
13:37:31 Quit spiorf_ (Read error: 110 (Connection timed out))
13:37:58 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
13:38:07pondlifeIs it possible to build RbutilQt under Cygwin? I lack qmake...
13:38:08linuxstb180fps fullscreen when boosted!
13:38:11 Join spiorf_ [0] (
13:38:26omrybarrywardell, for some reason this changes the block device name from sdb to sdc
13:38:57barrywardellomry: that's probably a linux thing
13:39:04barrywardelllinuxstb: wow!!!
13:39:33linuxstbbarrywardell: Compared to 48.5 previously...
13:40:51barrywardellabout 3.5x again
13:40:51 Quit CaptainSquid (Remote closed the connection)
13:41:00 Quit massive_H ("Leaving")
13:42:02omrybtw: when I tried to run doom my sansa crashed.
13:42:24 Join Thundercloud [0] (
13:42:40 Quit Nico_P (Remote closed the connection)
13:42:47pondlifebluebrother: Under Cygwin I can obtain qmake-qt3.. any point in attempting to hack at this?
13:45:25The-CompilerI need some info about the Sansa c200. What in Rockbox does work, what not?
13:46:39Zagormost things work
13:47:32Zagorin fact I can't off the top of my head recall anything that doesn't
13:48:40omrywhat is the normal transfer speed when sending files to sansa e280?
13:49:59 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
13:51:19 Join costasc [0] (i=591a9148@gateway/web/cgi-irc/
13:51:43Zagoromry: are you getting very slow speeds?
13:52:04omryZagor, I am not sure if it's very slow. 2 - 3 mb/sec
13:52:46 Quit ddalton ("leaving")
13:53:39 Quit costasc (Client Quit)
13:56:48Zagorhdparm -t gives 4.5 MB/s read speed on my c200
13:56:52linuxstbZagor: Regarding your question about the USB interrupt, are you saying it's not going to work without it, or just that it will be better if we can find that interrupt?
13:56:58Zagorso 2-3 sounds about normal
13:57:06omryZagor, thanks.
13:58:01Zagorlinuxstb: I think it will work, but it will be inefficient and also some things (like device reset) doesn't work since we don't react to it quick enough. but at least so far it's not a show stopper.
13:58:42 Join Daenyth [0] (
13:58:42omryI am playing a podcast, and I don't get any time feedback on the screen (how long am I into the podcast, and how long is the podcast)
13:59:16omryis that a theme thing, a problem with the file or just the way rockbox works?
13:59:19linuxstbDo you see that when you're playing other kinds of files?
13:59:52 Join costasc [0] (i=591a9148@gateway/web/cgi-irc/
14:00:41linuxstbRockbox can show that information, but it's up to the theme whether it's included or not.
14:01:02The-Compileromry: What format is the podcast?
14:01:23omryxmms shows the time fine.
14:01:31omry(for the same file)
14:01:46costaschi - there were some questions on my patch 7898 - "Insert Next" bug?
14:01:58The-CompilerAnd wich themes are you using?
14:02:06 Join jba_ [0] (
14:02:28omryblackglass - anemone
14:04:10omryyeah, I think it's the theme
14:04:16The-CompilerHmm... Do you use the newest build?
14:04:32omryI used whatever the automated installer downloaded
14:04:41omry(just now)
14:04:55The-CompilerThe theme has a progress bar just under the artist info. Do you see the "Time: ..." info (the 4th line)
14:05:06omryI only see it on other themes
14:05:58 Join webguest58 [0] (i=52474cfe@gateway/web/cgi-irc/
14:06:06 Nick webguest58 is now known as njd (i=52474cfe@gateway/web/cgi-irc/
14:06:18The-CompilerHi njd
14:06:22amiconnbarrywardell: The quarter sized splash is due to too fast transfers. The LCD controller drops every other pixel as it can't follow
14:06:26 Join marc___ [0] (n=marc@
14:06:27omryah, it has a small hidden progress bar without numeric informaiton
14:06:43amiconnThe cpu is boosted during boot, and switches back to 30MHz afterwards
14:06:51njdI seem to have a problem with my ihp-140 on party mode
14:07:01amiconnIn the end we will dynamically adjust the timing when changing cpu clock
14:07:02 Quit petur ("reboot")
14:07:05The-CompilerI need some info about the Sansa c200. What in Rockbox does work, what not?
14:07:20barrywardellamiconn: that's what I figured
14:07:24amiconnlinuxstb: On Color I'd expact around 2.5x speedup
14:07:26linuxstbThe-Compiler: We answered that question.
14:07:53The-CompilerOOOOPS! Sorry....
14:07:58barrywardellLCD_BASE is wht affects the timing?
14:08:31amiconnThere is no LCD_BASE
14:08:53barrywardelli mean LCD2_BASE
14:09:33 Join petur [0] (n=petur@rockbox/developer/petur)
14:09:49linuxstbamiconn: I seem to get more than that - over 3.5 at 80MHz.
14:10:07linuxstb(and no visible problems)
14:10:21amiconnI think that both 0x70008a00 and 0x70008a04 affect timing. In what way needs to be researched
14:11:55 Quit spiorf_ (Remote closed the connection)
14:12:45njdIf I install with blank firmware then go to Playlists: create new playlist to create a blank playlist, then turn on party mode and add some songs, when I go to the playlist and try and play the first song it hangs solid
14:14:29linuxstbnjd: What do you mean by "blank firmware" ?
14:15:27njdlinuxstb: I mean, I deleted the .rockbox directory and copied in a new one from the r15085-071012 build + fonts
14:18:39linuxstbAnd it works fine without party mode enabled?
14:18:44amiconnlinuxstb: Around 175fps??
14:20:23amiconnIf I don't miss something, that's outside the controller specs
14:20:36njdlinuxstb: yes it works fine without party mode
14:20:54 Quit jba_ (Connection timed out)
14:21:22njdlinuxstb: I've just reproduced it again: 1. Delete .rockbox directory, copy in fresh one + fonts 2. Turn on party mode 3. Queue a file (says "queue last") 4. try to view playlist - this also crashes
14:21:36njdlinuxstb: it's locked up again, I will have to reset
14:22:02Daenythdoes anyone know which revision is safe to use with a 5g ipod video?
14:22:08njdlinuxstb: to me it looks like the party mode doesn't play very well with the fact that I haven't got any playlists stored yet
14:22:10Daenythall the ones I've tried so far make it lock at the splash screen
14:23:43barrywardellamiconn: is the H10 controller the same as the color? I can get 194.5fps boosted without problems
14:24:16amiconnYes, but that's not out of specs
14:24:29 Quit ze (Remote closed the connection)
14:24:35 Join ze [0] (
14:24:35barrywardellah, the specs are in bytes per second?
14:24:37amiconnThe data transfer rate is what counts
14:24:55amiconnAnd the Color has more than twice as many pixels as the big H10
14:25:13 Quit Entames__ (Read error: 104 (Connection reset by peer))
14:25:15amiconnNot, more, but 1.9*
14:29:05linuxstbnjd: Sorry, I don't know much about party mode or the playlist code - I was just asking you questions for the benefit of someone who does...
14:30:00n1sDaenyth: are you using a current bootloader?
14:30:05 Join Entasis [0] (
14:30:15DaenythI just updated it a sec ago, right now I have the 9/11/07 build
14:30:23njdlinuxstb: it looks like a fairly straightforward bug but I'm not familiar with the code to have a look at it: is it worth trying to raise a bug?
14:30:32Daenythlemme try again with current
14:30:47n1sDaenyth: just updated the bootloader?
14:31:03DaenythI had an old build
14:31:07Daenythsometime from 2006
14:31:15linuxstbnjd: Maybe wait a short time to see if someone here replies to you, but yes, if they don't a bug report would be helpful.
14:31:16n1sok, try renaming your .rockbox dir and unzip a clean one to the player
14:31:17Daenytheither that or from very early this year
14:31:29Daenythdid the config file change?
14:31:40njdlinuxstb: OK I will leave this open for a while
14:31:49n1sDaenyth: yes, among lots of other things
14:31:57barrywardellamiconn: this works well for me
14:31:58linuxstbDaenyth: Just to confirm, you've run the current version of ipodpatcher to install a new bootloader
14:32:12Daenythjust downloaded it from the manual page today
14:32:27Daenythlet me try as you suggested
14:33:27Daenythok, done
14:34:45Daenythbootloader says can't load rockbox.ipod
14:34:54Daenythhang on, let me make sure I did everything ok..
14:37:13Daenythlatest bootloader + 09/11/07 build loads
14:37:26Daenython the 5g video
14:37:28Daenyththanks much!
14:38:30linuxstbCan you also try the current build?
14:38:47DaenythI'll try it with a clean .rockbox
14:38:51Daenythlet me back up that one
14:39:25amiconnbarrywardell: bah
14:39:40amiconnWe want to get rid of that weird outl() and inl() stuff...
14:40:02Daenythit works also
14:40:14Daenythlinuxstb: I suspect that the lock was coming from old bootloader + new firmware
14:40:26linuxstbYes, it would be.
14:40:32barrywardellamiconn: I know, but it was easiest to use for testing
14:41:04*amiconn doesn't deem inl() and outl() easy - they're confusing
14:41:19Daenythlinuxstb: is it ok to just copy my old config file?
14:41:20preglowthey are indeed
14:41:37Daenythold as in from early this year, maybe even last
14:41:44preglowthe only use i can see for inl/outl is to allow easy switching to user mode execution in the future, but we don't want that
14:41:56linuxstbDaenyth: It may work, but I would just recreate the settings again manually.
14:42:09Daenyththanks for the help :)
14:45:34njdanyone any idea about my party mode issue?
14:46:02 Nick idnar_ is now known as idnar (i=mithrand@unaffiliated/idnar)
14:48:23barrywardellamiconn: do you think I should commit the change (with the inl/outl fixed)?
14:48:30 Quit donutman25 ("ChatZilla [Firefox]")
14:48:36barrywardellor do you want to do some more testing first?
14:52:07amiconnI'd rather do some research regarding the registers first
14:52:27 Join miepchen^schlaf [0] (
14:53:16Daenythnow to find a theme :]
14:53:38 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
14:53:59 Join miepchen^schlaf [0] (
14:56:19desowinis it playable? :D
14:56:35desowinheh, music player instead of sword
14:56:50DaenythI'm gonna try it
14:57:08Daenyththis may be stupid, but I don't see in the manual where it explains how to install themes?
14:58:13linuxstbYou install them like anything else - you extract the zip file directly to the root of your device, and the zip file should contain the correct folder structure.
14:58:27Daenythahh nifty
14:58:28linuxstbYou then select them from the Settings -> Browse Themes menu option.
14:58:44Daenyththanks for all the help :)
14:59:38 Quit miepchen^schlaf (Connection reset by peer)
14:59:59 Join miepchen^schlaf [0] (
15:01:00DaenythI like how every theme on has stairway to heaven on the WPS :D
15:01:41 Join spiorf [0] (
15:02:38costascI read the earlier irc stuff on my patch and added a comment to try and clear it up:
15:09:03njdlinuxstb: Thanks for your help - I've got to get on really. I raised a bug about the issue and will have a look at the source if I get some time later
15:09:28***Saving seen data "./dancer.seen"
15:10:29Daenython some of the themes I'm getting they don't have the correct directory structure
15:10:39 Quit njd ("Bye bye")
15:11:06TMMhas anyone ever seen this? :
15:11:37TMMyou can buy 10 USB product id's for 30 dollars, if rockbox is to get software USB support, I'd be very willing to buy a set and donate them to the project
15:12:39bluebrotherpondlife: rbutilqt won't work with qt3 −− it's quite different to qt4.
15:13:09bluebrotherand Domonoky did the voicing stuff −− I _think_ it should support the sapi script. But I never tried (I only changed the settings dialog)
15:13:23 Quit elinenbe (Read error: 104 (Connection reset by peer))
15:13:59ZagorTMM: nice!
15:14:50linuxstbTMM: Although it's 30UKP - 60 USD at current exchange rates.
15:14:58linuxstbBut still very affordable
15:15:02TMMwe could write some integration plugins for the likes of rhytmbox and whatnot, would be nice if we had some integration between OSS desktops and OSS firmwares
15:15:31linuxstbAnd I'm guessing we would need more than 10.
15:15:32TMMlinuxstb, I still wouldn't mind coughing that up :)
15:15:57TMMlinuxstb, why? we could just use the same product ID for all players, from the outside world they'd look the same anyway
15:16:15linuxstbrbutil for example wants to be able to distinguish different models.
15:16:31linuxstbThe GUI installer app.
15:16:46TMMpersonally, I'd say that's an implementation detail of rockbox
15:16:55 Quit The-Compiler ("Connection reseted by god")
15:17:06TMMie: we just need a magic file or command on the USB stack that will give out the 'real' hardware it's running under
15:17:27TMMfrom a technical point of view, it makes no sense to distinguish between players
15:17:34TMMfrom the outside, that is
15:17:49bluebrotherit definitely makes sense.
15:17:50TMMrockbox is an OS, the user shouldn't have to care what hardware it's running on
15:18:04bluebrotherjust imagine the Ipod users who want to keep using Itunes.
15:18:27TMMwe'll make it configurable
15:18:45linuxstbbluebrother: They would need to either use Apple's USB mode, or hope that someone implements the extra proprietory Apple USB commands in Rockbox.
15:19:06bluebrotherdo they have proprietary usb commands? ouch.
15:19:17GodEater_what commands do we believe itunes uses ?
15:19:26TMMhaving a real product ID for rockbox would make it possible to 'auto detect' rockbox from OSS audio programs, like rhytmbox, banshee or amarok, and get some real integration going. like updating the tagDB from software
15:19:29amiconnTMM: It makes very much sense to distinguish players, because of their different features
15:19:32linuxstbIt sets the time on the ipod. Plus there's the XML file it can retrieve with hardware info.
15:19:46GodEater_ah yes
15:19:57TMMamiconn, from client's USB ports POV, what's the difference??
15:20:00linuxstbamiconn: But maybe we could expose a description of the features - so apps don't need to know.
15:20:33linuxstbGodEater_: And I'm sure there are more...
15:20:34TMMamiconn, NOW it doesn't really distingish either, it's just an USB mass storage device
15:21:04linuxstbTMM: Yes, but it can know what hardware Rockbox is running on.
15:21:11linuxstb(based on the manufacturer's USB IDs)
15:21:27TMMbut then it can't know that rockbox is running on it in the first place
15:21:35 Part LinusN
15:21:56Zagorit's not exactly a huge benefit to use our own IDs either, for the same reasons
15:22:23linuxstbTMM: I agree with you, one USB IDs with a proprietory (but obviously openly documented) command to retrieve the hardware capabilities could be a nice solution.
15:22:25Zagorwe can use stock ids and our own strings, for example. so it reads as "Rockbox device" but still uses stock ids
15:22:25 Quit homielowe (Read error: 110 (Connection timed out))
15:22:40TMMZagor, how so? we'd have a rockbox ID, so applications would know they are talking to a rockbox device, it could automagically update the rockbox tagDB
15:23:07ZagorTMM: well the question is if any current programs do the opposite, and fail if we switch id?
15:23:25Zagorit's easier for us to write new software that works with old IDs than the other way around
15:23:41TMMZagor, I disagree strongly
15:23:42linuxstbAnd I assume we can't modify the IDs for the hardware USB bridges?
15:24:23GodEater_I'm inclined to agree with Zagor
15:24:27TMMZagor, normally devices are detected based on their vendor/product ID, rockbox would be VERY different if we just change the vendor strings send by the device
15:25:14TMMZagor, plus, I believe that with libusb the vendor name and product name are read from a file, like with PCI, so itwouldn't work at all there
15:25:17ZagorTMM: well what if, for example, itunes only work with certain ids. what have we gained?
15:25:19GodEater_we can't change the vendor string though
15:25:28ZagorGodEater: yes we can
15:25:34GodEater_sorry - we can, but we shouldn't
15:25:42GodEater_unless we own the vendor ID we change it to
15:25:49TMMZagor, what use is itunes if you don't use the apple firmware?
15:26:06GodEater_TMM: lots of users still use iTunes to sync their ipods
15:26:06ZagorI don't think the vendor string is linked to the vendor ID in any way. it's just a string the device sends.
15:26:23TMMGodEater_, with OF, yes
15:26:25Zagormany bridge solutions send "false" vendor and product strings.
15:26:27bluebrotherTMM: I bet a lot of Ipod users still want to be able to use AppleOS.
15:26:34GodEater_TMM, no, with Rockbox too
15:26:51TMMGodEater_, they sync with OF loaded, then play from rockbox
15:26:58 Join Domonoky [0] (
15:26:58TMMbluebrother, this won't stop that at all
15:27:11preglowwhat's the point in getting our own ids?
15:27:14GodEater_it'll be something they'll complain about though
15:27:21GodEater_preglow: I can't see a point
15:27:24TMMif we have our own usb/product ID for rockbox, we could do itunes like integration from desktop apps, the same level of integration apple now enjoys
15:27:28bluebrotherTMM: well, if Itunes stops working due to a changed ID the synced music won't work using AppleOS
15:27:44Zagorpreglow: from my point of view it's mostly fun
15:27:44TMMbluebrother, we can't change it in OF
15:28:00GodEater_supplying our own usb/product ID doesn't improve the chances of people making their app work with Rockbox
15:28:05GodEater_it's not required AT ALL
15:28:06bluebrotherTMM: that's overcomplicated. We already can detect the Rockbox build using rockbox-info.txt
15:28:07TMMI'm pretty sure I'm making quite a good case for this idea here, but noone seems to be listening, or noone cares
15:28:14bluebrotherwhich amarok f.e. already uses.
15:28:19preglowwhy do these sell the pids so cheaply?
15:28:41Zagorpreglow: because the've got 64k of them ? :)
15:28:41TMMpreglow, you don't get a vendor ID, just a product ID
15:28:43GodEater_there is no requirement I can see for our own PID here
15:28:49bluebrotherTMM: we can't change that in the OF but users will expect to be able syncing using Itunes regardless if they have booted RB or the OF.
15:29:05TMMbluebrother, and, right now, we can't even do that on apple devices
15:29:06Zagordo we know itunes check usb id?
15:29:08amiconnTMM: One of the major reasons for me to use rockbox is that it does *not* need a silly "integration" a la itunes
15:29:20ZagorI was just being hypotethical
15:29:20GodEater_Zagor: no we don't, but we don't no it doesn't either
15:29:24TMMamiconn, it won't NEED the integration
15:29:25preglowbut i've gotta run, later
15:29:29TMMamiconn, it'll be an option
15:29:38linuxstbGodEater_: I'm almost certain people won't be able to use itunes with our usb mode - not unless we implement those features I mentioned.
15:29:39TMMamiconn, a lot of people LIKE that level of integration
15:29:45GodEater_TMM, but that option exists already without the PID
15:30:00bluebrotherI really don't see any advantage in using our own ID. We could simply use our own device description string instead.
15:30:12GodEater_linuxstb: you're probably right
15:30:13TMMGodEater_, which is a horrible unstandard way of detecting such things, and does NOT work with hotplug or any other part of the utopia stack
15:30:22 Join moos [0] (
15:30:29ZagorTMM: there's a technical disadvantage with new IDs. not with old IDs. that's hard to argue around,
15:30:35TMMbluebrother, that won't work! that's resolved using usb.ids
15:31:04TMMZagor, what's the disadvantage? for all intends and purposes a device running RB is a completely different beast than the same device running OF
15:31:06Zagorwhat do you mean that won't work? crap software that doesn't read usb decriptors are to be ignored.
15:31:15bluebrotherTMM: well, I have a small serial-usb converter demo application lying around. And I'm able setting any usb product string I like to use.
15:31:22linuxstbZagor: But a device running Rockbox isn't the same as a device running the OF IMO, so you could argue for new IDs for that reason.
15:31:47GodEater_linuxstb: but the point is having them solves no problems
15:31:48bluebrotherso even without knowing the details about usb I know it's possible to set an arbitrary string as product description, regardless of any file on the host.
15:31:52Zagorsure, but that's purely philosophically.
15:32:10TMMZagor, still, what is the disadvantage?
15:32:42linuxstbGodEater_: Problems might happen with things like itunes (or other proprietory syncing software) that expects to be talking proprietory USB extensions with the device.
15:32:46TMMZagor, and, linux's hal and udev do NOT read the vendor strings, you can only make udev rules based on USB ids, I'm fairly sure of that
15:33:05Zagoryou can use whatever you like
15:33:21bluebrotherTMM: still, you can use rockbox-info.txt for detection.
15:33:23*Domonoky thinks we should postphone this discussion until we got a working usbstack and usbdrivers.. :-)
15:33:24TMMZagor, plus: I'd say that being another device just by changing the vendor string is broken behaviour
15:33:24GodEater_linuxstb: that's hypothetical though - you don't know that yet...
15:33:34linuxstbGodEater_: I accept that ;)
15:33:37Zagorthe point is that linux will create device entries for every endpoint the usb device presents. no matter which usb id it has.
15:33:40TMMbluebrother, that's HORRIBLE way to go about it!
15:33:45GodEater_in which case, Domonoky makes a good point
15:33:56bluebrotherTMM: and why?
15:33:58GodEater_we should stop bothering to talk about it until we KNOW what RB-USB breaks ;)
15:34:18GodEater_I don't get why that's horrible either
15:34:19bluebrothernot being able to distinguising the device based on the ID is horrible too.
15:34:42bluebrotherand as I said, amarok still uses this technique.
15:34:47TMMbluebrother, that would mean that to add support for rockbox to, say, HAL, it would have to check each and every usb mass storage device that's plugged in for a certain file... something NO OTHER device requires, and I doubt that'll happen.
15:34:52bluebrother(rbutil too, as fallback)
15:34:53GodEater_and it's availabe for anyone else to use as well
15:34:54*Zagor goes back to coding
15:35:10bluebrotherwhy do I want to add Rockbox support to HAL?
15:35:13TMMbluebrother, I don't want my desktop app to do the detecting, I want my OS to do that, and take appropriate action when I plug in my device
15:35:30ZagorTMM: take what action, exactly?
15:35:43 Quit mf0102 ("Verlassend")
15:35:45bluebrotherand you do NOT need to check every UMS device −− you match the device first against the ID and look if it's a rockboxable player at all
15:36:31*GodEater_ fails to understand what special action an OS would need to take with a Rockboxable DAP that's different to any other UMS device
15:36:35TMMZagor, I don't know, make coffee for all I care, but I can imagine it would like to start my desktop application, perhaps some yet-to-be-written script/desktop tool that auto-syncs music from a certain folder, who knows? the sky's the limit if we have a device with OSS firmware
15:37:13TMMbluebrother, so, we update a list somewhere, and distributions that are released before we add a new target do... what exactly?
15:37:36TMMI just dream of linux desktop + rockbox device == seamless experience (or CAN be for default users without any hassle)
15:37:39linuxstbTMM: I think in order to convince us, you need a concrete example of something that can only be done if we use our own IDs, or can't be done if we use the original vendor's IDs.
15:38:02linuxstbTMM: And remember some Rockbox devices have hardware USB where we can't change IDs.
15:38:09bluebrotherwell, ID lists are _always_ out of date.
15:38:28linuxstbs/some/about 50% of/
15:38:48TMMbluebrother, and they don't need to be if we have a separate ID for rockbox... it would be always up-to-date as soon as this is done
15:39:10TMMO Well, I'll start this discussion again once rb-usb actually works
15:39:22TMMI don't think I'm getting my point across very well anyway
15:39:33*bluebrother is not convinced and will probably never :)
15:40:35TMMbluebrother, I'll try one more time: what's the point in rockbox saying it's an apple ipod if it is not capable of actually integrating with itunes? itunes will probably just thing this is a broken player and helpfully try and erase it, and recreate the database stuffs...
15:41:09bluebrotherwell, that's a point, but in that case we need to figure out why it doesn't integrate with itunes.
15:41:10 Quit costasc ("CGI:IRC (EOF)")
15:41:14GodEater_the key word there is "probably". We still don't know it's true.
15:41:40Zagorbluebrother: because we don't support their weird non-ums mode.
15:41:46linuxstbbluebrother: But I don't think we do want to integrate with itunes. If a user wants that, they can sync with Apple's USB mode.
15:42:03bluebrotherand what are the Apple specific extensions? Setting the clock on the Ipod and retrieving the player info xml.
15:42:04Zagorit's a multi-faceted issue, for sure
15:42:13TMMand even then, we still wouldn't be able to write ANY specific desktop software for use with rockbox, we could come up with anything we damn well please to do, it's all open source. let the creative juices flow :)
15:42:14linuxstbbluebrother: Those are the two I know about...
15:42:19bluebrotherand I bet Itunes will work even if this isn't supported.
15:42:27linuxstbbluebrother: How much? ;)
15:42:51GodEater_TMM: you keep saying that, and yet there are already apps out there that deal with a rockboxed dap.
15:43:00linuxstbBut yes, as has been said, lets wait until we hit problems before solving them...
15:43:47TMMGodEater_, I know but you have to START amarok yourself for it to recognize :"Hey, this is a rockboxed device" while I want to just plug my player into my OS and have the OS know it's a rockboxed device, and do whatever magic we may come up with as being useful
15:44:26GodEater_TMM: then you're not talking about apps.
15:44:26bluebrotherTMM: do you really think it's good to allow the OS starting GUI applications? That should only be done be a gui handler imo.
15:44:27ZagorTMM: still, we don't really need new ids for that
15:44:31pondlifebluebrother: How do I select SAPI in RbUtilQt? The last built one on the wiki (1.0.1 r14595) only has espeak or flite
15:44:46bluebrotherpondlife: you need an updated build, that was added later.
15:44:59pondlifeOK, do you have one lying around...
15:45:10 Join donutman25 [0] (n=chatzill@
15:45:22Domonokythe usb ids is not really a technical problem.. we can just choose an id.. we could even make a (compile time) option..
15:45:28donutman25hi guys
15:45:32TMMbluebrother, very well, 'gnome-device-manager' whatever you like. point remains
15:45:42donutman25whats up with that new logo on the home page?
15:45:50ZagorDomonoky: it can be a technical problem, both keeping the old and using new ones
15:46:01GodEater_donutman25: what new logo ?
15:46:05Zagordonutman25: there's no new logo
15:46:16bluebrotherTMM: no, as gnome-device-manager is not HAL ;-)
15:46:34donutman25on my pc it have pandemic studio's and electronic arts logo
15:46:49bluebrotherdonutman25: try clearing the browser cache.
15:46:51Zagordonutman25: then your pc is broken
15:47:05GodEater_donutman25: or you're looking at the wrong web page
15:47:06TMMbluebrother, no, but I might want to start a non-gui syncer app from hal, or something... there's just no point in preventing apps like that to be written imho
15:47:25TMMbluebrother, udev, not hal
15:47:30*GodEater_ can't remember the last time anyone expressed an interest in writing such an app
15:47:39donutman25bluebrother: thanks that fixed it
15:47:52linuxstbTMM: Couldn't you do that with the original firmware's IDs? Isn't it up to you how you configure your OS with devices?
15:48:07amiconnTMM: You forget that around half of the rockbox targets has hardware usb where we can't change the id or strings
15:48:24amiconnAnd also, that an own usb ID costs money
15:48:44linuxstbamiconn: But not very much -
15:48:45GodEater_amiconn: he's only talking product IDs, not a vendor ID
15:50:53GodEater_in any case, yes, having custom behaviour on your PC when some device is plugged in HAS to be done by the user
15:51:11GodEater_be it a udev, or HAL rule
15:51:26pondlifebluebrother: Any idea what I should set as the TTS executable for SAPI5?
15:51:45*amiconn prefers that the PC does nothing special when plugging an usb msd apart from automounting
15:51:56*linuxstb prefers that it doesn't even automount
15:51:58*GodEater_ agrees with amiconn for once :)
15:52:00TMMGodEater_, but it would be possible to pre-configure some sane rules, like some of the end-user distro's do
15:52:07pondlifeSAPI default gives me an "Path to TTS is wrong!" error.
15:52:10GodEater_TMM: such as ?
15:52:12Domonokypondlife : point it to the rocbox sapi_voice.vbs script..
15:52:24TMMGodEater_, ubuntu does for instance
15:52:36ZagorTMM: as I keep saying, udev can string match. separate id is not technically necesary for udev handling.
15:52:47pondlifeDomonoky: OK... could that be built-in somehow?
15:52:51Zagorit's still worth discussing, but don't focus udev:ability on this
15:52:58GodEater_and I still don't want my OS doing anything beyond mounting it, if that.
15:53:26Domonokypondlife: yes, thats possible, but i didnt founf a good way to do it..
15:53:40TMMGodEater_, this is not taking away that option, this is ADDING an option for users that want it, end-users like seamless experiences
15:53:44 Join Crash91 [0] (n=evil91@
15:53:49TMMZagor, very well, I didn't know that :)
15:54:06bluebrotherDomonoky: adding it as resource and extract it to %TEMP% before running the script?
15:54:19GodEater_TMM: but if you're talking of including this in "end user distros", then you have to decide what the action is that makes the most sense to ALL users
15:54:23GodEater_TMM what action is that ?
15:54:28 Part Daenyth ("Leaving")
15:54:39TMMGodEater_, starting rhytmbox, and importing the playlists
15:54:58TMMGodEater_, that is what end-users would expect to happen
15:55:19Domonokybluebrother: thats possible, you only have to make sure, that you onlay do this on windows :-)
15:55:52bluebrotherand what is with end-users expecting amarok to open?
15:56:11pondlifeDomonoky: Is it possible to generate directory .talk files, but not file .talk files?
15:56:17GodEater_does anyone actually *use* Rhythmbox ? :)
15:56:26*bluebrother doesn't :P
15:56:31TMMwhy would you have to start in manually, if the only thing you are going to do is start it, then plug in the device?
15:56:36Crash91GodEater_: i used to
15:56:39TMM*I* use rhytmbox
15:56:47linuxstbTMM: So you're talking about starting applications that will _only_ work with a Rockbox'ed DAP, and not with a DAP's original firmware?
15:56:50pondlifeWhat's rhythmbox?? ;)
15:56:51Domonokypondlife: if there is no option for it.. no.. but it would be easy to add..
15:56:52GodEater_I guess someone has to :)
15:57:08pondlifeDomonoky: OK, shall I FlySpray it? or is this conversation enough
15:57:11Crash91GodEater_:Ive switched to banshee now =)
15:57:11TMMlinuxstb, apps that can use specific funtionality yet-to-be-developed for rockbox, yes
15:57:14GodEater_it's like, the suckiest music management thing ever IMO, but I guess whatever blows your hair back ;)
15:57:35linuxstbTMM: What I mean is, what is the harm in those apps being started if the device isn't running Rockbox?
15:57:42Crash91not was the only app with mp3 support bundeled with Fedora 7
15:57:45Domonokypondlife: put it on flyspray or on the rbutil wiki page.. under task for milestone 2 :-) +
15:57:52pondlifeOK, wiki it is
15:58:04Domonokyor else i will forget it.. :-)
15:58:04Crash91except VLC which is a bit too lieghtweight
15:58:09*GodEater_ wonders why any linux user wouldn't use amarok
15:58:19 Quit donutman25 ("ChatZilla [Firefox]")
15:58:24TMMGodEater_, because it's interface is so messy
15:58:34TMMGodEater_, I like simple things :)
15:58:34*Crash91 wonders why GodEater doesnt use Banshee
15:58:46GodEater_Crash91: never heard of it till 30 seconds ago ;)
15:58:48TMMlinuxstb, that's not a big problem, but the app NOT starting is :)
15:58:51linuxstbGodEater_: I've never used amarok (I've also never seen it...)
15:59:06GodEater_I quite like amarok :)
15:59:11Domonokyanother good extension for the talk file generation would be an option to only voice certain file types..
15:59:14TMMand completely back off-topic :)
15:59:22karashataAmarok is a nice app, IMO, but right now I'm in Windows...
15:59:28GodEater_amarok is not offtopic
15:59:31GodEater_it supports rockbox daps
15:59:40 Quit Zagor ("Client exiting")
15:59:41TMMscrew you guys!
15:59:53*GodEater_ goes to install banshee to see what it's like
16:00:16*linuxstb goes to install amarok to see what it's like
16:00:33GodEater_linuxstb: it's KDE based - don't know how much of that you have installed already...
16:00:37*TMM goes to install windows vista to see what it's like
16:00:38*bluebrother likes the interface of amarok
16:00:40Crash91GodEater_: you need the gstreamer-ffmpeg-dirty or something like that to work
16:00:47bluebrotherTMM: don't do it!
16:01:00TMMhahaha :)
16:01:02TMMI'd never :P
16:01:07linuxstbGodEater: apt-get only wants 28MB, so it's not too bad.
16:01:08GodEater_Crash91: oh god.
16:01:17TMMCrash91, there's -good -bad and -ugly :)
16:01:22*GodEater_ doesn't understand the logic behind the gstreamer mess
16:01:27Crash91ah yes, ugly =P
16:01:28*karashata feels left out of all the linux discussion, but watches it anyway
16:01:56*Crash91 is using DSL on his worthlesss windows XP homo edition
16:02:05TMMGodEater_, 'good' plugins are neatly coded, and have no patent issues. 'bad' plugins have patent issues but are well-coded, 'ugly' plugins have patent issues AND are not well-coded
16:02:18*Crash91 feels XP homo sp2 is too gay..
16:02:34*TMM kindly points Crash91 to
16:02:44TMMnot .org :)
16:02:46DomonokyCrash91 and others: please stay on topic ..
16:02:56GodEater_TMM: no - I meant why is gstreamer SO over engineered...
16:02:57Crash91TMM: i ordered a free CD =P
16:03:02*bluebrother doesn't like ubuntu
16:03:08TMMDomonoky, what was the topic again? ;)
16:03:08Crash91to #rockbox-community!
16:03:09*karashata wishes he could back his files up and move to Gentoo, but doesn't have a backup hard drive or money to get one
16:03:23*Crash91 is a proud Fedora user
16:04:01pondlifeDomonoky: Do you think it sensible that one day RBUtil could build a voice file, given a lang file?
16:04:24Domonokypondlife: that would be really really nice.. but its not easy..
16:04:25pondlifei.e. Download the .lang file, not the .voice file.
16:04:32karashataI liked Gentoo until something broke and it wouldn't work for me anymore, but I had it as a dual-boot install with Vista, since then I gained too much stuff and don't have enough space to make a new partition anymore...
16:05:00pondlifeWell, we'll see if you (or someone) can get the SAPI script stuff incorporated first then.
16:05:00GodEater_karashata: like Domonoky says - take it to -community now - we've veered wildly offtopic
16:05:08Domonokythe problem is, that voice file generation now needs the features file.. so we would need to proviced this file along to the builds..
16:05:10karashataahh, sorry...
16:05:18 Join scorche|w [0] (n=8dc5049d@rockbox/administrator/scorche)
16:05:27pondlifeDomonoky: I thought that had been eliminated for some reason..
16:05:47GodEater_karashata: it's not that bad - there's some of us there already ;)
16:06:03karashata*nods a bit* kay
16:08:06 Join jba_ [0] (
16:08:26Domonokybut voicefile generation with rbutil would be nice, because the user can easily have nice sapi voices, which we cant provide on because of copyright issues..
16:08:36 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
16:10:05 Quit marc___ ()
16:15:31 Quit barrywardell (Remote closed the connection)
16:16:23pondlifeDomonoky: It's also nice to have the same voice for menus and .talk files.
16:16:56GodEater_pondlife: consistency?!?! What else do you want? Blood ?
16:17:07pondlifeHmm, beer would do
16:17:14Domonokyyes.. so i will take a look into this, if i find time for it...
16:18:50 Quit pill (Nick collision from services.)
16:19:11 Join pill [0] (
16:21:06 Quit pill (Nick collision from services.)
16:21:26 Join _pill [0] (
16:26:32 Join costas_c [0] (
16:29:40 Part costas_c
16:30:14 Join costas_c [0] (
16:32:51 Join haemmy [0] (n=stefan@
16:32:55 Part Crash91
16:33:44 Quit Nico_P (Remote closed the connection)
16:34:49 Join cendres [0] (n=ashes@2001:5c0:8fff:ffff:0:0:0:b)
16:35:08 Quit ashes (Read error: 113 (No route to host))
16:35:55 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
16:36:12 Quit jba_ (Read error: 110 (Connection timed out))
16:37:11 Join barrywardell [0] (
16:39:20 Quit jhulst ("Konversation terminated!")
16:40:55 Join kugel [0] (
16:41:03 Quit barrywardell (Remote closed the connection)
16:44:09 Join outside [0] (i=nat-021a@
16:45:50 Join SirFunk [0] (
16:45:57 Join barrywardell [0] (
17:00:11 Join darksaboteur [0] (
17:00:15TMMscrew you guys! I'm going to the pub :P
17:00:56GodEater_bit early
17:01:02GodEater_some of us still have work
17:01:17karashatait's only 1 after 11 am here
17:01:35karashataway too early for the pubs, or bars, or whatever ya wanna call them...
17:01:44GodEater_they open at 11 here
17:01:52TMM5am here, and my current employer is a government body
17:01:55linuxstbI thought they never closed...
17:02:11GodEater_linuxstb: that's not the pub, that's your living room :)
17:02:12TMMso, I'm actually the last here... at 4 on a friday it's deserted
17:02:20karashataI wouldn't be surprised to find them open here, but to sit at them for the sake of being there, is early, I think...
17:02:26 Part Entasis ("Leaving")
17:02:33GodEater_karashata: clearly you're not from a drinking nation :)
17:02:48*TMM off
17:02:49 Quit petur ("connection reset by beer")
17:02:50 Quit TMM ("Ik ga weg")
17:03:01karashataCanada, I have no idea about it since I'm not a heavy drinker myself
17:03:12karashatahas partly to do with lack of free money...
17:03:29karashataso far what little I have in reserve is for bills
17:03:30GodEater_that is a decisive factor...
17:03:33*linuxstb wants some free money
17:03:41karashatawouldn't we all love it?
17:04:03*scorche|w thought this was -community for a bit there...
17:04:50*karashata drags this to -community
17:04:55 Quit haemmy ()
17:05:19 Join haemmy [0] (n=stefan@
17:09:29***Saving seen data "./dancer.seen"
17:10:54 Quit giantbicycle ("暫離")
17:11:02 Join nicktastique [0] (n=nick@unaffiliated/nicktastic)
17:16:34 Join bluey- [0] (
17:16:52 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@
17:18:19 Join Arathis [0] (
17:20:05 Quit nicktastic (Read error: 110 (Connection timed out))
17:25:08Arathisamiconn: how can one recognize lcd speedup with your last commit?
17:33:34GodEater_Arathis: I would imagine by running MpegPlayer
17:33:39GodEater_and turning on "Show FPS"
17:33:42barrywardellArathis: use the test_fps plugin. the speedup is only on H10 small so far
17:33:48GodEater_or that....
17:33:56barrywardellspeedup for 20 will come soon
17:34:28 Quit billytwowilly (Remote closed the connection)
17:34:29Arathisis the test_fps plugin a standard plugin?
17:34:41barrywardellit's not built by default
17:35:01barrywardellyou have to add test_fps.c to apps/plugins/SOURCES for it to get built
17:35:27 Join Gnu47 [0] (i=Gnu47@
17:36:23amiconnbarrywardell: The big H10 should also see a bit of speedup without the multiple write, due to the removed cruft
17:37:31barrywardellyes, but it's only small compared to what's possible
17:38:00barrywardelland compared to the speedup on small H10
17:38:15Arathisin mpegplayer I'd need to disable fps limitation, right?
17:38:17 Join toffe82 [0] (
17:39:06 Join billytwowilly [0] (
17:39:29amiconnmpegplayer will see no significant speedup on small H10 either
17:39:36amiconnlcd_yuv_blit() is not converted
17:40:52*Arathis is not in the mood to build rockbox ..
17:41:56 Join Crash91 [0] (n=evil91@
17:48:07 Quit Crash91 ("Bye Bye!")
17:55:24 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
17:57:02 Join marc___ [0] (
17:59:10 Part toffe82
17:59:21*Domonoky just changed the talkfile creation of rbutil, so you can select to create files only for Folders or only for Files, or for both.. :-)
17:59:31 Quit SirFunk (Remote closed the connection)
18:04:02 Quit darksaboteur ("ChatZilla [Firefox]")
18:04:46*bluebrother should release a new rbutil before free time is gone ...
18:06:30 Join toffe82 [0] (
18:07:44bluebrotherwill most likely happen quite soon −− I recently got a job offer
18:10:28 Quit jhulst ("Konversation terminated!")
18:10:48 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
18:20:40 Part bluey- ("Leaving")
18:20:46 Quit costas_c ("Leaving.")
18:21:19 Quit barrywardell ("Konversation terminated!")
18:23:16 Quit obo ("bye")
18:23:58Aveanyone, anyone? ipod nano status
18:24:17Avethe 2 line diff with clock speed and setup code changes, is it getting applied?
18:24:39bluebrothernoone, noone.
18:24:42Aveseems to fix the glitch problems. I have yet to test
18:25:10n1sAve: where is this diff you talk about?
18:25:36Nico_PDomonoky, bluebrother: Benjamin Mako Hill is going to do some rockbox installs at an install party ( I suggested he use rbutil
18:25:57Avebasically it sets the max clock to 78 MHz instead of 80
18:26:12amiconnI wouldn't call that a fix
18:26:33Aveeven if the problem is somewhere else, like the IO code, if it gets fixed by 2 MHz change, I'd be content
18:26:52bluebrotherNico_P: interesting.
18:27:22amiconnIt just tries to work around the issue. It will become less likely, but won't go away. Whether that helps or not depends on the individual nano
18:27:57bluebrotheroh, that's tomorrow. Bringing out a new release might be good ...
18:28:04amiconnAnd it sacrifices speed. Someone needs to go ahead and properly analyse the problem and fix it
18:28:18Avebut oes 2MHz really make a difference?
18:28:26Avein Real World
18:28:31 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
18:28:52amiconnIt can decide whether a track skips or not
18:28:59n1sAve: you are of course free to use any patches you want for yourself
18:29:12Aveyes I know
18:29:34Avebut for the community id would be beneficial if this thing got sorted out for good
18:29:52desowinif someone will fix it...
18:30:07AveI'm sure they would if they knew what was the problem
18:30:15Aveor is
18:30:20desowinthey? you mean 'we'?
18:30:39Avethe ones with the knowledge of rb internals
18:30:42Aveand nano
18:30:54amiconnNo one would magically know what the problem is. It needs to be analysed
18:31:05Aveeven so there remains the (big) unknown of apple firmware functionality
18:31:06pondlifeDomonoky: Thanks, I await the new version... :)
18:31:33pondlifeamiconn: You're no closer to holding a duff Nano then?
18:31:53desowinAve: it's portalplayer unknown functionality, atleast from what I know
18:31:59pondlifepity, it's an annoying issue
18:32:06pondlifeFrom a support POV I mean
18:32:21Avenobody has pp documentation?
18:32:28Aveprolly under nda or so
18:32:44pondlifeWe SO need an insider at Apple....
18:32:51amiconnI don't think PP releases docs to anyone. Not even apple
18:33:03amiconnI think they just sell their SDK
18:33:05marcosourcestrange, my rockboxed Sandisk Sansa e270 won't reg as a usb disk anymore on my mac.. but on the gentoo machine it's no problem.. I've seen this with usb memory sticks to, as far as I rememember, it only helped by reformating them... I'd rather not do that to my e270 :/
18:33:16amiconn...which would be of no use for us, even if someone had it
18:33:39Avebut if apple can make it work under the sdk why wouldnt rb
18:33:47desowinmarcosource: have you tried first booting into OF and then plugging in the cable?
18:33:59 Join Falen [0] (
18:34:01|Rain|rb is GPL, the SDK is almost certainly not
18:34:06pondlifeThe SDK would likely include closed source components
18:34:18marcosourcedesowin: yes, (that's the only place it'll reg anyway so ;) )
18:34:57marcosourcedesowin: I'm guessing it has something to do with the FS itselfe
18:35:02Aveyeah so the only road is reverse engineering apple firmware
18:35:05n1s|Rain|: up for another midi test?
18:35:13pondlifeNico_P: Been using your code all day, with only the occasional problem...
18:35:16 Part pondlife ("disconnected has pondlife")
18:35:29|Rain|n1s: I'm slightly distracted right now, but I'll try to work one in, sure
18:35:49marcosourceoh, is USB transfer working in rb yet ?
18:36:00n1s|Rain|: thanks, I'll post a link in here when I have a patch ready
18:36:10Nico_Pmarcosource: for portalplayer ?
18:36:22marcosourceonly ? no other ?
18:36:49Nico_Pmarcosource: it works on most targets execpt portalplayer (ipods are amont those)
18:37:03marcosourceand sansa ?
18:37:10Nico_Psansa is portalplayer too
18:37:26Nico_PZagor is working on an USB implementation
18:37:47krazykithe's building off of austriancoder's work, right?
18:38:11marcosourcehum, my build is a "stable" build, might not have any currently new features
18:38:37desowinis there stable for sansa?
18:38:47marcosourceerr, okay, non cvs ;)
18:39:06marcosourcecan't get the damn build env right..
18:39:07desowinit's svn dailybuild anyway
18:40:00Nico_PjhMikeS: you were asking about the ISR pcmbuf position callback... any idea how the same thing could be achieved another way ?
18:40:29krazykitmarcosource, the easy way would be to just use the virtual machine. all the work has been done for you :P
18:41:35marcosourceI'm talking about compiling rb itself, it'd be nice to get that going so I could fiddle around a bit with some plugin ideas
18:42:11n1smarcosource: well, waht is the problem?
18:44:52 Quit jhulst (Connection timed out)
18:46:46 Join ArchFool [0] (
18:46:56ArchFoolW00t! :)
18:47:11 Join atsea-34 [0] (i=atsea-@gateway/tor/x-cdd6ad3bc300af5b)
18:47:36marcosourcen1s: just gentoo giving me a hard time, not to worry ;)
18:48:06marcosourceI'm more wonderingabout why my player won't show up in osx
18:48:45ArchFoolHow do I get rockbox to start playback upon powerup? Right now it takes me to the main menu, and I have to go to Resume Playback manually.
18:49:01n1s|Rain|: here's the patch the ususal before and after numbers would be most appreciated :-)
18:49:31rasherArchFool: set the startup screen to "now playing", or something like that..
18:49:47n1sArchFool: look for the "Start Screen" setting and set it to "Resume Playback"
18:50:06*n1s haers an echo
18:51:17ArchFoolI just love the fact that RockBox takes 4sec to book up. Sansa firmware takes 10-11sec!
18:52:20*ArchFool masticates with a chicken leg.
18:53:29 Join homielowe [0] (
18:54:44 Quit Toxicity999 (Remote closed the connection)
18:56:25 Join costas_c [0] (
18:57:00|Rain|anyone have any sansa e2xx OF images besides 01.00.12[AE], 01.01.11A, 01.02.15A, and 01.02.18A that they'd be willing to accidentally send my way?
18:58:04*ArchFool only has a c250. :(
18:58:35desowin|Rain|: I do have 1.02.15E
18:58:39 Quit Sedgewick ("Money can't buy friends, but they can get you a better class of enemies.")
19:02:04 Quit Falen ("ChatZilla [Firefox]")
19:07:32 Quit karashata ("Leaving.")
19:08:17 Join karashata [0] (
19:09:33***Saving seen data "./dancer.seen"
19:12:39 Join Toxicity999 [0] (n=bryan@unaffiliated/Toxicity999)
19:15:30 Join ilgufo [0] (
19:18:28 Quit ArchFool ("p00f!")
19:21:39 Quit homielowe (Read error: 110 (Connection timed out))
19:22:08 Join SirFunk [0] (
19:25:08 Join lee-qid [0] (
19:25:30 Quit lee-qid (Client Quit)
19:25:35 Join lee-qid [0] (
19:25:53 Quit haemmy ()
19:30:12 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
19:30:33 Join miepchen^schlaf [0] (
19:31:03 Join Robin0800 [0] (
19:33:00Robin0800how do you install QT for qtrbutil in windows ? the wiki page has no details
19:33:41BigBambiRobin0800: Are you trying to build it?
19:34:02BigBambiIf not, just download the compiled version
19:34:04BigBambiAh, OK
19:34:15BigBambiIn that case, don't know
19:34:27BigBambiQT website no help?
19:36:11Robin0800not realy can check again though did not think instructions were applicable to rockbox
19:36:16 Join ompaul [0] (n=ompaul@freenode/staff/gnewsense.ompaul)
19:36:41Robin0800am runing thevmware version
19:37:46Robin0800I am using the VMWare (linux)
19:38:26desowinRobin0800: for building qt apps on win32 you don't need any extra things
19:38:42desowindownload qt opensource win32 edition from trolltech website
19:39:08desowinduring install it'll ask if you have mingw32 installed, if not - it'll offer option to download and install
19:39:26desowinthen from start menu->applications->qt->qt command prompt
19:39:33desowingo into rbutilqt directory
19:40:05desowinqmake && make, and you're done
19:40:33Domonokydesowin: is right.. thats the way i use it, rbutil development..
19:40:48Domonokyarg.. +for
19:47:32 Quit moos (Read error: 104 (Connection reset by peer))
19:48:01 Join [AFX] [0] (
19:50:08 Join Ebert [0] (
19:56:14 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
20:00:31 Quit nicktastique ("Leaving")
20:00:36 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
20:02:56 Quit lee-qid ("aufwiederbyebientotsayonara")
20:06:36 Quit Canar (Read error: 110 (Connection timed out))
20:06:48|Rain|any other takers on e200 OF?
20:07:49 Quit costas_c (Read error: 110 (Connection timed out))
20:19:18 Join Zagor [0] (
20:21:01advcomp2019|Rain|, do you want a few r series ones
20:23:51|Rain|they won't do me any good, but thanks
20:24:23 Quit miepchen^schlaf ("Verlassend")
20:26:40 Quit cendres (Read error: 104 (Connection reset by peer))
20:29:23 Join miepchen^schlaf [0] (
20:30:45 Join obo [0] (n=obo@rockbox/developer/obo)
20:32:21n1s|Rain|: are you building a collection of sansa OF images?
20:32:50|Rain|I'm updating/fixing the OF database rebuild code
20:33:08|Rain|the OF DB rebuild is really, really annoying :)
20:34:06n1saha, so how's it going?
20:35:21|Rain|I've reworked the code slightly, fixed the existing firmware versions to work for me, and added 3 more versions (so far, every version I've looked at has used the same offset for each region)
20:35:30 Join barrywardell [0] (
20:37:50desowin|Rain|: keep up the good work!!
20:39:10 Join Redbreva [0] (
20:42:53 Join Wiwie [0] (
20:46:33 Quit delYsid (Connection timed out)
20:47:07 Join costas_c [0] (
20:48:32 Quit Wiwie ("Miranda IM! Smaller, Faster, Easier.")
20:52:45lymecaI ran ipodpatcher on my 512-byte sector 30GB 5th gen but only after I copied the incorrect 2048-byte sector MBR to the device. I have the correct 512-byte MBR on it now, but the partitions still seem to be messed up.
20:52:56lymecafdisk can't open the device because of an incorrect disk label or somesuch
20:53:49 Join linuxstb_ [0] (
20:55:49linuxstb_lymeca: Have you unplugged and reattached your ipod after writing the correct MBR?
20:55:59lymecalinuxstb_: Yes
20:56:08lymecaThe ROckbox bootloader is installed
20:56:18lymecaAnd it complains that there is no partition
20:56:32lymecaSo I have to reset it and press select+play to boot into disk mode
20:56:52linuxstb_Does ipodpatcher still recognise your ipod?
20:58:32linuxstb_And did you format the fat32 partition?
20:58:53lymecaCheck that link for ipodpatcher output
20:59:12lymecaYes I ran mkfs.vfat /dev/sdc2 more than once since copying the correct MBR
20:59:35linuxstb_And can you mount it from Linux?
21:00:10lymeca# mount /dev/sdc2 /mnt/ipod
21:00:11lymecamount: you must specify the filesystem type
21:01:14 Quit w0rd54 (Remote closed the connection)
21:01:42linuxstb_Specify it... - mount -t vfat /dev....
21:02:33 Join bertrik [0] (
21:03:54lymecalinuxstb_: It's a bad FS
21:03:58lymecaOr something
21:04:10 Join webguest44 [0] (i=4a469ecd@gateway/web/cgi-irc/
21:04:26linuxstb_Try formatting again. I normally do "mkfs.vfat -F 32 /dev/sdc2"
21:04:31webguest44does sansapatcher 0.5 work on rhapsody?
21:04:38linuxstb_webguest44: No.
21:05:21 Part costas_c
21:08:27lymecalinuxstb_: It mounts now but:
21:08:29lymeca# fdisk /dev/sdc
21:08:29lymecaError: Unable to open /dev/sdc - unrecognised disk label.
21:09:12 Join w0rd54 [0] (
21:09:37***Saving seen data "./dancer.seen"
21:10:13 Quit desowin ("use linux")
21:11:22 Join mf0102 [0] (n=michi@
21:12:10 Quit Ebert ()
21:12:35webguest44so that is what the developers are working on now.
21:14:39 Quit chisel (Read error: 104 (Connection reset by peer))
21:14:53 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
21:14:59linuxstb_lymeca: Does Rockbox boot?
21:15:19linuxstb_lymeca: So does fdisk actually run?
21:17:46 Quit spiorf (Remote closed the connection)
21:18:09webguest44can i use the gparted live cd?
21:19:37lymecalinuxstb_: I got it working now
21:21:31 Part Robin0800
21:22:57 Nick fxb is now known as fxb__ (
21:23:27webguest44anyone know what gparted is?
21:23:41webguest44am i able to use it
21:23:47webguest44instead of ubuntu
21:24:22iambenfor what, installing to e200r?
21:25:44iambeni'd assume so, afaik it has usb support
21:27:05scorche|wgparted is a partition editor...although i dont know how their livecd comes..
21:28:03iambenit is based on a gentoo livecd (sorta)
21:28:23webguest44but i'm still able to use it, right
21:28:26scorche|wso i see..
21:28:56scorche|wbut is it a true linux livecd, or just a very very stripped down version?
21:29:14iambenwebguest44: it should, but you'll have to try it to find out
21:34:16*amiconn found a bug
21:34:27amiconnStar freezes when starting gameplay
21:34:33amiconn(on H10)
21:36:42 Join Robin0800 [0] (
21:36:59webguest44i'm just going to use ubuntu but do i need server or desktop edition
21:37:41iambeni just booted my gparted livecd and it does have full usb support btw
21:38:13iambenbut do whatever you like
21:38:44linuxstb_amiconn: I _think_ there is a flyspray report about that.
21:40:49webguest44sorry, my comp. timed out so i have to repeat the question. i'm just going to use ubuntu but do i need server or desktop edition
21:41:28*amiconn points webguest44 to the irc logs
21:41:28 Quit webguest44 ("CGI:IRC (EOF)")
21:42:15 Quit scorche|w ("CGI:IRC (EOF)")
21:42:20 Join webguest44 [0] (i=4a469ecd@gateway/web/cgi-irc/
21:43:04webguest44sorry, my computer lost connection. So i have to repeat the question because i didn't get your replies. i'm just going to use ubuntu but do i need server or desktop edition?
21:43:08iambenwebguest44: it makes no difference, its a simple tool you run from command line
21:43:19webguest44okay thanks
21:43:44Lloreandesktop edition
21:43:50LloreanServer doesn't have a LiveCD as far as I know
21:44:08amiconnlinuxstb: Hmm, I have a suspicion
21:44:18*amiconn points webguest44 to the irc logs *again*
21:45:44amiconnlinuxstb: star uses a transition effect where it updates from center using larger and larger rectangles
21:46:21amiconnSo if either height == 0 or width == 0, *some* lcd_update_rect() implementations will act up
21:46:39amiconnThat explains why it doesn't happen on all targets
21:46:48jhMikeSamiconn: I was just trying star from lastest SVN but DC patched and it's not freezing :\
21:47:28jhMikeSI'm playing it now
21:47:33amiconnHmm, the 20GB doesn't use the block transfer controller...
21:47:49 Part Redbreva ("Out for the day...!")
21:48:05jhMikeSah, just on the small H10?
21:48:17jhMikeSthe freeze i mean
21:48:33amiconnThat's where I observed it. Need to check other targets
21:49:08amiconnAnyway, guarding against zero width or height sounds like a good idea, as well as clipping to the lcd area
21:49:25amiconnBut that will have to wait; I'm about to extend test_fps
21:50:57*jhMikeS makes a dc ipod color build and would very much appreciate a test there...will put up on ftp
21:51:13 Quit webguest44 ("CGI:IRC (EOF)")
21:52:10 Quit marc___ (Excess Flood)
21:53:35Nico_PjhMikeS: you were asking about the ISR pcmbuf position callback... any idea how the same thing could be achieved another way ?
21:53:43amiconnZero height should do no harm on H10, but zero width would.
21:54:18amiconnIt would set the transfer size to 0xffff (numpixels - 1), and then put zero pixels
21:54:37*amiconn puts a DEBUGF() call into the sim implementation
21:55:14jhMikeSNico_P: I was thinking like mpegplayer does. Timestamp the codec buffer and transfer the stamps to the pcm buffer.
21:57:07jhMikeSwhat's a good ipod color button for "quit"?
21:57:34amiconnDepends on whether it's a simple plugin or needs many buttons
21:57:34Nico_PjhMikeS: I'll have a look
21:57:50amiconnSimple plugins just use MENU for quit on ipods
21:57:55jhMikeSamiconn: just test_queue.c
21:57:57amiconnIf that's taken, MENU+SELECT
21:58:15 Join midgey [0] (
21:58:40LloreanjhMikeS: Long hold of Menu is often quit for the plugins, I thought.
21:58:47midgeyamiconn: i know star froze on the gigabeat f/x too
21:59:19midgeyiirc it's in star_transition which is ifdef'd out on gigabeat as a work-around
22:00:14*pixelma read that as "effects"... ;)
22:00:49Nico_PjhMikeS: I haven't looked yet, but is it a big job ? if not, would you consider doing it in svn ?
22:01:53amiconnmidgey: hack-around
22:01:56jhMikeSNico_P: I'm not sure how big a job it is. Maybe not too bad. I suppose I could look closer after this other work is settled.
22:02:12amiconnThe transition effect starts indeed with a series of updates with width == 0 or height == 0
22:02:29amiconnSome lcd_update_rect() implementations are susceptible to this, some are not
22:02:43jhMikeSsome just assume valid params?
22:02:53amiconn(just figured from looking at the code so far, sim still building..)
22:03:01Nico_PjhMikeS: thatks... I'm not at all familiar with the PCM code so I would probably spend way more time than necessary on it
22:04:00amiconnjhMikeS: Many do, but some don't break when width and/or height are zero
22:04:17midgeyamiconn: a hack-around indeed
22:04:30midgeyim preety sure the crash only occurred on target for gigabeat
22:04:35jhMikeSNico_P: it can probably just go in the pcm buffer descriptor structure and be accurate enough
22:04:43amiconnmidgey: It's not needed if the implementation for gigabeat gets properly safe-guarded
22:05:43amiconnA temporary workaround (and a sane idea) would be to let the updates not start at zero (which doesn't make sense anyway)
22:05:51midgeyright, i'd like to see the issue fixed on gigabeat too
22:05:59 Join costas_c [0] (
22:06:30jhMikeSwhy would gigabeat be susceptible? It checks the width and height.
22:07:32 Quit SirFunk (Remote closed the connection)
22:08:39 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
22:08:57midgeyjhMikeS: no idea, im testing a build with the transition code added back in
22:10:26midgeyno crash on my gigabeat
22:10:32jhMikeSI did put clipping in there myself awhile ago
22:11:57 Quit amiconn (Nick collision from services.)
22:12:03 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
22:16:13 Quit midgey ()
22:17:26 Quit jhulst ("Konversation terminated!")
22:17:35 Join midgey [0] (
22:18:15 Join haemmy [0] (n=stefan@
22:18:25 Join BigBambi_ [0] (n=alex@rockbox/staff/BigBambi)
22:18:27 Join haemmy2 [0] (n=stefan@
22:18:34 Quit BigBambi (Read error: 113 (No route to host))
22:18:49 Quit haemmy (Client Quit)
22:18:56 Quit ToHellWithGA ("You know you'll miss me a lot.")
22:18:56 Quit ilgufo (Connection timed out)
22:19:06 Nick BigBambi_ is now known as BigBambi (n=alex@rockbox/staff/BigBambi)
22:20:04 Quit nicktastic ("Leaving")
22:20:09 Join |AhIoRoS| [0] (n=ahioros@
22:24:19 Quit omry (Remote closed the connection)
22:28:30jhMikeScomes with test plugins and public-domain SPC files for testing:"> and
22:28:36 Quit barrywardell ()
22:28:47 Quit kretender ("Leaving.")
22:29:32jhMikeSlinuxstb: ping (see links) :)
22:31:44linuxstb_jhMikeS: What do I need to do? Just play some SPC files?
22:32:09jhMikeSlinuxstb_: yeah, it puts it through all the playback paces using COP
22:32:39 Join scorche|w [0] (
22:32:49jhMikeSH10 has no problem so I'm pretty confident here. There's also some test plugins to test cache and queue in rocks
22:33:46jhMikeSoh, make sure abolutely no voice is enabled...including no file present.
22:34:05jhMikeSswapping codecs won't operate here like this
22:34:07 Join barrywardell [0] (
22:35:47jhMikeSwhat is?
22:36:00amiconnThat voice doesn't work...
22:36:12jhMikeSit's only because of the multithreaded codec
22:36:25jhMikeSit will otherwise
22:36:48 Quit FOAD ("I'll be back")
22:37:06jhMikeSI didn't implement any interface to swap any such thing safely
22:38:37 Join hannesd [0] (
22:38:41Bagder"BTW, I trust everybody already knows enough about the interrupt contollers, right? No need to RE anything in there?" / MrH ?
22:38:59Bagder(ignore my trailing ?)
22:39:13amiconnand the missing r ;)
22:39:15Bagderhe pointed out the usb interrupt
22:39:32linuxstb_jhMikeS: Everything seems to work fine.
22:39:36Bagderthe missing r was MrH's bad ;-)
22:39:51jhMikeSlinuxstb_: heh...awsome. thanks.
22:40:13amiconnjhMikeS: I assume you already tested on PP5022?
22:42:01jhMikeSamiconn: Others have I _think_. I will wait for more tests though. Only 5020 ever gave trouble.
22:42:03n1sif anyone in here uses the stopwatch plugin you might want to take a look at a patch I posted in FS #7877 that moves the buttons around a bit
22:42:33amiconnjhMikeS: Are you planning to commit that multi-threaded codec at some time?
22:42:59amiconnImho the non-working voice would be nasty, even if it wouldn't crash
22:45:03 Quit amiconn (Nick collision from services.)
22:45:09 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
22:45:11 Join matsl [0] (
22:46:20jhMikeSamiconn: perhaps but not before dealing with that issue. Right now it's just there for proving things.
22:48:01pixelman1s: I find the definitions for the Ondio pad very unintuitive (just reading the patch) - maybe you could use button combos (menu+up/down for example) to control the volume?
22:48:45n1spixelma: sure, I'll fix it, thanks for looking
22:49:30jhMikeSamiconn: any mt codec would need notification of the swap and need to park its threads safely inside the kernel (block them all on some object).
22:49:56pixelman1s: but short "off" for start/stop and long "off" to quit the plugin is not nice - and it makes it easy to accidentally shut the player off...
22:50:26n1sso it's a good thing I asked before committing :-)
22:50:38pixelmabut I know that there are similar things in other plugins on other targets
22:52:05n1sthis button mess is pretty annoying, I almost just added buttons for the requested target and threw an ifdef around the case, fiddeling with the buttonmaps took much longer time than the code :-/
22:53:27pixelmathe button mess? It's just a lot of different targets with different keypads (and restrictions with them)
22:53:52n1sAnd inconsistent mappings and stupid button names
22:54:52n1slike why is BUTTON_SCROLL_DOWN bound to increasing volume in wps for e200? I know it's the clockwise turning of the wheel but why call it down
22:55:03pixelmastupid button names - I agree, especially the scroll up/down in the e200 map...
22:55:20n1sand ipods, which way is forward on a circle that spins?
22:56:13pixelmaI guess that these definitions where just taken from the H10 map... ehmm barrywardell? I think
22:56:21 Quit Domonoky (Read error: 104 (Connection reset by peer))
22:56:52pixelmaI already asked about that quite some time ago
22:56:54n1sis off+menu usable as a quit combination for ondio?
22:58:26n1slike this
22:59:16 Quit BigBambi ("Leaving")
22:59:34pixelmait should work but I still think "off" shouldn't be used for start/stop
23:00:23pixelmatrying to think of a better suggestion...
23:01:03n1sso right for start, off for reset and left for lap?
23:01:30pixelmathat sounds better
23:02:57 Quit barrywardell ()
23:03:38 Join bertrik_ [0] (
23:03:44pixelmahope you also plan to update the documentation ;) because menu+off is not used anywhere else (can't remember) so that at least I could read about it
23:04:37n1spixelma: sure, when I commit it, I'll let it sit in the tracker a while so people can comment
23:08:30n1sspeaking of the manual, it's starting to get pretty out of date...
23:09:01pixelmaI know :\
23:09:39***Saving seen data "./dancer.seen"
23:09:55 Quit haemmy2 ()
23:11:08 Join lee-qid [0] (
23:12:48pixelmaI already prepared some more button tables for the c200, will commit it tomorrow probably, while on it I even saw remnants of the old manual (in wormlet.tex) for example... and there were quite a few feature additions that aren't documented at all
23:14:36 Nick parafin is now known as parafin|away (
23:18:03 Join BigBambi [0] (n=alex@rockbox/staff/BigBambi)
23:18:48 Quit bertrik (Read error: 110 (Connection timed out))
23:21:55 Quit obo (Remote closed the connection)
23:31:03 Part costas_c
23:33:45 Quit mf0102 ("Verlassend")
23:36:17 Quit |AhIoRoS| ("Abandonando, see you")
23:38:12 Quit outside (Read error: 104 (Connection reset by peer))
23:40:23Nico_PGodEater_, GodEater: I've committed a debug screen
23:41:20 Join barrywardell [0] (
23:43:57barrywardellpixelma: I can't remember why that was chosen. probably just copied from H10 like you said. feel free to change it to something more intuitive
23:46:54lymecaHow do I boot into APple's firmware once I am running Rockbox?
23:47:10Lloreanlymeca: Turn off the player, turn on the player, immediately turn and leave on the hold switch.
23:47:12jepleron e200 should charging work "the same" as in the OF? Reading the source code, it doesn't look like rockbox does much besides change the indicator to show a "plugged in" indicator... but the battery reports "full" long before it actually is..
23:48:10Lloreanjepler: There is no proper charging status
23:48:21LloreanIt shows "full" because on AC power it seems to the player the same as having a full battery.
23:48:45jeplerLlorean: OK, that makes sense −− I was just noticing that the percent_to_volt_charge table is the same as the percent_to_volt_discharge table
23:49:02LloreanYes, there's still a lot of work that needs to be done with charging code on portalplayer targets, I think
23:49:08jeplerbut maybe _to_volt_charge isn't even used because of the CONFIG_CHARGING setting
23:49:19 Quit ompaul (Client Quit)
23:49:27jeplerbut as long as I leave it plugged in, it *is* charging?
23:49:34LloreanIt is, yes
23:49:38jeplerit's just the status that could be better..
23:49:41LloreanUnfortunately, Rockbox charges much slower than the original firmware
23:49:58LloreanSo unless you want/need to be using Rockbox while charging, the recommended method is still to reboot into the original firmware
23:50:24jeplerI see
23:54:57 Quit Jon-Kha ("leaving")
23:55:28 Join Jon-Kha [0] (
23:55:53LloreanIn theory at least, that should be improved when our improved USB code is finished
23:57:39jepleroh, is it slow because you can only draw a small amount of current before identifying as a usb device with a 500mA current request?
23:57:53LloreanPretty much
23:58:18LloreanThat combined with the fact that Rockbox currently uses more power than the original firmware makes things slow.

Previous day | Next day