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

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

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

#rockbox log for 2010-04-29

00:01:13 Quit RadicalR (Ping timeout: 245 seconds)
00:04:15 Quit fyrestorm (Read error: Connection reset by peer)
00:08:01 Quit blithe (Read error: Operation timed out)
00:08:18 Join blithe [0] (~blithe@
00:10:52 Quit linuxstb (Ping timeout: 276 seconds)
00:11:19 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
00:11:21***Saving seen data "./dancer.seen"
00:11:27 Quit mikroflops (Ping timeout: 252 seconds)
00:11:53 Quit crwl (Ping timeout: 264 seconds)
00:16:40 Join MethoS- [0] (~clemens@
00:17:19 Join zivan56 [0] (
00:19:10 Join RadicalR [0] (
00:20:35 Join mikroflops [0] (
00:20:40 Join DoingGodsWork [0] (
00:20:43 Join crwl [0] (
00:20:56 Part DoingGodsWork
00:21:02 Quit kenguest (Read error: Connection reset by peer)
00:22:28 Join kenguest [0] (
00:25:55 Join robin0800 [0] (
00:27:52 Quit bertrik (Quit: De groeten)
00:36:02 Quit wodz (Quit: Leaving)
00:37:36 Quit DataGhost (Ping timeout: 240 seconds)
00:39:29 Quit DerPapst (Quit: Leaving.)
00:46:47 Quit TheSeven (Read error: Connection reset by peer)
00:50:01 Quit anewuser (Quit: What do you know...THE WORLD'S first NTRQ (that's for NES/FAMICOM) tracking compo. Have powerpak? Try it out! Otherwise ROM IMAGE.)
00:53:16 Quit archivator (Quit: Leaving)
00:57:24 Join mikroflops_ [0] (
00:59:16 Quit liar (Quit: Verlassend)
01:00:36 Quit mikroflops (Ping timeout: 246 seconds)
01:05:11 Quit ender` (Quit: Trying to establish voice contact ... please yell into keyboard.)
01:12:46 Quit robin0800 (Quit: Leaving)
01:15:38 Join kisak_ [0] (
01:17:55kisak_hello, I found what appears to be the wiki equivilent to lint
01:20:24kisak_on down in the Fuze v2 section there is a speculation "Presumably based on an AS353x CPU like the ClipV2." Isn't the development of the Fuse v2 to the point that we know that it contains an AS3525?
01:23:03kisak_(especially since it's referenced 3 lines later)
01:32:16 Quit MethoS- (Remote host closed the connection)
01:43:44 Join linuxguy3 [0] (
01:53:28 Part toffe82
01:57:27 Quit dfkt (Quit: -= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.)
01:58:56 Join robin0800 [0] (
01:58:58 Quit panni_ (Ping timeout: 248 seconds)
01:59:24 Join panni_ [0] (
02:01:37 Quit komputes_ubuntu (Quit: I haven't slept for ten days, because that would be too long.)
02:06:40 Join jfc^2 [0] (
02:06:40 Quit rasher (Read error: Connection reset by peer)
02:07:51 Quit Rob2223 (Quit: Rob2223)
02:08:56 Quit lifeless__ (*.net *.split)
02:08:56 Quit jfc (*.net *.split)
02:10:20 Join lifeless__ [0] (~lifeless@
02:11:22***Saving seen data "./dancer.seen"
02:11:35 Join rasher [0] (
02:11:36 Quit rasher (Changing host)
02:11:36 Join rasher [0] (~rasher@rockbox/developer/rasher)
02:16:57 Quit markun (Read error: Connection reset by peer)
02:16:58 Join markun_ [0] (~markun@rockbox/developer/markun)
02:22:09 Join Rob2222 [0] (
02:26:38 Quit robin0800 (Remote host closed the connection)
02:33:37 Quit kerwood (Quit: Leaving.)
02:43:27CIA-5New commit by funman (r25753): as3525v2: effect of CGU_PROC on fclk is instant ...
02:43:39 Join funman [0] (~fun@rockbox/developer/funman)
02:44:14funmankisak_: yup it's based on as3525, probably nobody fixed the wiki
02:47:49 Join CaptainKewl [0] (
02:48:34funmanweird: current build seems to not crash on Clip+ .. ?
02:51:29funmanah "nice" it still crashes :)
02:52:13funmanstill in set_cpu_frequency(), when storing the new frequency to memory (cpu_frequency global variable)
02:55:33kisak_well, seeing as I'm not a registered wiki user, would someone clean out that lint?
02:56:13funmanI can help you to get write rights ;)
02:57:21funmanat this point the wiki pages for each model are not particularly helpful for development (except perhaps pictures of the boards) so i have no strong will to keep them up to date with current knowledge
02:58:54tmzt_they also make good history
03:01:41kisak_oh, I just found a simple typo
03:05:18kisak_so I'm not allowed to use my alias as my wikiname?
03:07:35 Join komputes [0] (~komputes@ubuntu/member/komputes)
03:08:23 Nick fxb is now known as fxb__ (
03:08:28funmannope, you must use FirstnameLastname
03:09:58 Join Anonomizer [0] (
03:10:05 Quit Anonomizer (Remote host closed the connection)
03:15:04 Join hebz0rl_ [0] (
03:15:40 Quit w1ll14m (Read error: Connection reset by peer)
03:16:34kisak_is user registration handled manually?
03:16:59funmanregistration no, but once you're registred I have to give you write rights manually (just give me your username)
03:18:00kisak_ok, I'm waiting for a confirmation email
03:18:49 Quit hebz0rl (Ping timeout: 264 seconds)
03:20:55 Quit panni_ (Ping timeout: 265 seconds)
03:22:40 Quit n17ikh (Ping timeout: 264 seconds)
03:29:25 Join n17ikh [0] (
03:30:11 Join anewuser [0] (anewuser@unaffiliated/anewuser)
03:30:18 Join Guest23293 [0] (
03:34:12 Quit n17ikh (Ping timeout: 259 seconds)
03:34:52 Quit Guest23293 (Ping timeout: 240 seconds)
03:39:12 Join n17ikh [0] (
03:45:23 Quit adnyxo (Quit: Leaving)
03:54:50 Join phanboy4 [0] (
03:54:52 Join Strife89|PalmTX [0] (
03:56:39 Join DJVistaMan [0] (
03:57:04DJVistaManAre there any mods for the iPod Nano 3Gen available?
03:57:17DJVistaManOn any platform
03:57:21S_a_i_n_tDJVistaMan: Not here at least...
03:57:33DJVistaManok Thanks
03:58:04 Quit DJVistaMan (Client Quit)
03:58:54 Join CGL [0] (~CGL@
04:07:43 Join Barahir_ [0] (
04:08:39funmansomething fishy is going on with memory
04:08:46funmansomeone with a fuzev2 handy ?
04:11:09 Quit Barahir (Ping timeout: 264 seconds)
04:11:23***Saving seen data "./dancer.seen"
04:12:52 Quit komputes (Ping timeout: 246 seconds)
04:23:20funmancrap a full album played on Clip+ but clipv2 crashed fast
04:26:41 Quit pixelma (Disconnected by services)
04:26:41 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:26:51 Quit amiconn (Disconnected by services)
04:26:53 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:27:03 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:27:15 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:28:25funman'prefetch abort at (some instruction, not a valid address)'
04:34:49 Join antil33t [0] (
04:51:48funmanthe truth is out there...
04:54:04 Join chrisb [0] (
04:56:42chrisbi have this little cradle charger with an audio line-out for my sansa e250
04:57:21chrisbi'm running rockbox. when i plug the sansa in, there isn't any output on the audio line-out
04:59:08funmanis that a dock which plugs on the USB port?
05:03:23chrisbfound this
05:04:06chrisbmy bad, no support
05:04:21chrisbfunman: yes, it is a dock which plugs into the usb port on the bottom
05:04:56funmanyeah it will likely not work, perhaps in the future if someone has the hardware, the dev skills and the motivation
05:05:02funmanyou'd better use the jack output
05:10:14chrisbfunman: the cradle seems to cut out the jack output
05:10:40 Join RandomInsano [0] (
05:12:11 Quit Strife89|PalmTX (Quit: Bed.)
05:14:48RandomInsanoSo working on making a bootloader. Just editing the telechip's main() function. When I look at the code with objdump, I'm expecting to see some of my magic numbers (0x80003000 for PORTA) in the assembler code. Should they be in there directly or might GCC be mucking around with them?
05:15:20CIA-5New commit by funman (r25754): as3525v2: crashless cpufreq switching ...
05:16:03funmanRandomInsano: telechips is ARM, right? 0x80003000 could be calculated in 2 instructions instead of 1 load from memory
05:16:17RandomInsanoIt is arm.
05:16:44RandomInsanoThat might be entirely possible. I'm also setting everything to 0xFFFFFFFF and seeing nothing about 0xFFFF
05:16:44funmanlike, mov rX, #0x80000000 ... and rX, rX, #3000
05:17:01RandomInsanoInteresting. I'll look for something along those lines.
05:17:29funmanif you can't see the value with an hex editor, perhaps running grep on objdump output might help
05:17:30RandomInsanoI'm extremely new to real assembler. My comp sci department uses the fake LC3 to teach that stuff
05:17:49RandomInsanoI'm doing that exactly. Well, piping to less
05:17:51funman(afaik objdump sometimes use base 10, so be sure to grep for both representations)
05:18:23 Quit fdinel (Quit: Miranda IM! Smaller, Faster, Easier.
05:19:01funmanhum.. it won't help if this value is calculated on 2 instructions though. But I remember most of PP registers can be represented as an immediate value (8 bits + shift), so they show up in objdump output if mov is used
05:19:01 Quit anewuser (Quit: What do you know...THE WORLD'S first NTRQ (that's for NES/FAMICOM) tracking compo. Have powerpak? Try it out! Otherwise ROM IMAGE.)
05:19:03RandomInsanoThat makes things trickier :(
05:19:39funmanbtw are you looking at OF code ?
05:19:54RandomInsanoBit of both, but this specifically is mine
05:20:06RandomInsanoIt doesn't *seem* to be working on my DAP.
05:20:25RandomInsanoI'm putting all of PORTA high and when I test for voltages I get floating.
05:20:49RandomInsanovoid* main(void)
05:20:49RandomInsano#ifdef TCCBOOT
05:20:49DBUGEnqueued KICK RandomInsano
05:20:49RandomInsano int rc;
05:20:49RandomInsano unsigned char* loadbuffer = (unsigned char*)LOAD_ADDRESS;
05:20:49***Alert Mode level 1
05:20:49***Alert Mode level 2
05:20:49RandomInsano // LCD uses PORTA, so there's lots of pins to check
05:20:50***Alert Mode level 3
05:20:50RandomInsano uint32_t*gpio_a = (uint32_t*)0x80003000;
05:20:50***Alert Mode level 4
05:20:50RandomInsano // Set b for output
05:20:51***Alert Mode level 5
05:20:51RandomInsano *(gpio_a + 0x4)= 0xFFFFFFFF; //0x1 << 18;
05:20:51***Alert Mode level 6
05:20:51RandomInsano while (1)
05:20:52***Alert Mode level 7
05:20:52RandomInsano {
05:20:52***Alert Mode level 8
05:20:52RandomInsano *(gpio_a) = 0xFFFFFFFF; //0x1 << 18;
05:20:53***Alert Mode level 9
05:20:53RandomInsano }
05:20:56funmanhmm please use pastebins
05:20:57saratogause a pastebin
05:21:05S_a_i_n_tGAH! Pastebin, please!
05:21:07RandomInsanoGood plan.
05:21:14RandomInsanoTime to try this 'paste bin'
05:21:45funmanbtw, wouldn't (gpioa+4) be 0x80003010 ?
05:21:58RandomInsanoOh poop, you're right.
05:22:15RandomInsanoCurse you decimal!
05:22:30funmanit's not decimal, it's int32_t :)
05:22:34RandomInsanoWait no, I should be good.
05:23:25RandomInsano+4 should be 0x800003004 I would think.
05:24:02RandomInsanoWell, in... Whatever. I know that the direction register is a 4 byte offset from the base of porta
05:24:38RandomInsanoAlso my comment is a tad outdated :P
05:26:24funmanRandomInsano: no.. +4 skips 4 elements of whatever gpio_a points to, here it's an int32_t so it skips 4*4 bytes
05:26:26RandomInsanoEverything looks okay. I have no idea if the firmware crashes before this point or what. I'm using a TCC77x chip as my base, but I assume none of it's initialization stuff runs before main()
05:26:33RandomInsanoOh crap!
05:26:36RandomInsanoYou're right
05:26:58funmanfirst thing you need is some visual output: backlight or button light if it's present
05:27:05RandomInsanoComing from a 'java for everyone' university, I'm not to great with these pointers
05:27:28funmanhope you can learn with rockbox :)
05:27:31RandomInsanoYeah, tried for backlight, turns out it's harder than I thought. So I can use my voltmeter to check for high output
05:28:14RandomInsanoThe anode and cathode seem to get power 100% of the time. I'm assuming there's a transistor nestled away somewhere.
05:30:02chrisbmaybe i can boot into sansa os before docking...
05:30:54***Alert Mode OFF
05:31:38phanboy4Does the rockbox scrobber have a built-in limit on the size the log can reach?
05:31:54RandomInsanoBah, changing 4 to 1 didn't fix my problem.
05:32:20 Quit S_a_i_n_t ()
05:33:00 Join S_a_i_n_t [0] (S_a_i_n_t@
05:33:51RandomInsanoALL GPIO TO ON! MWAHAHA! (here's hoping that works).
05:38:10RandomInsanoNope. Nothing :(
05:38:10RandomInsanoCurse you... Telechips?
05:39:00 Quit Horscht (Quit: Verlassend)
05:46:04 Quit CaptainKewl (Remote host closed the connection)
06:10:51 Join BHSPitMini [0] (
06:11:27***Saving seen data "./dancer.seen"
06:13:27 Quit funman (Quit: free(random());)
06:13:29JdGordonS_a_i_n_t: try the patch?
06:18:42 Quit CGL (Remote host closed the connection)
06:23:43S_a_i_n_tJdGordon: Well, it's built...but I have to wait until I get home to actually test it on a device.
06:23:55S_a_i_n_tI'll let you know when I have done so.
06:24:00 Quit elcan (Read error: Connection reset by peer)
06:29:56 Join elcan [0] (
06:30:30 Part Boldfilter
06:40:45 Part RandomInsano
06:41:18S_a_i_n_t <==Failed SIM build...any ideas anyone?
06:41:26S_a_i_n_tClean, brand new checkout...
06:49:15S_a_i_n_tBlarggh!!! It happened again, after a "make veryclean" as well...A 'Normal Build' builds fine, but the SIM fails with those errors.
06:55:15 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
07:04:22 Nick shai_ is now known as shai (
07:16:42crculverAny plans to have Rockbox display the album art contained within a FLAC file? Or perhaps that has been decided against? Nothing on the bug tracker.
07:17:21S_a_i_n_tcrculver: There *is* a patch on the tracker to display embedded AA
07:17:33S_a_i_n_tnot sure if it'll work with .flac though
07:18:02 Nick shaggy-h is now known as chrism (
07:18:02 Nick chrism is now known as shaggy-h (
07:20:35S_a_i_n_tcrculver: At the present point, apparently it only supports .jpg files embedded in .mp3
07:31:52chrisbmy playlists go into the root directory by can i set up the default directory to be /PLAYLISTS ?
07:42:51S_a_i_n_tchrisb: Open the config file (config.cfg) and change the path in there.
07:43:39S_a_i_n_tchrisb: Add/Edit the line so it appears as "playlist catalog directory: /Playlists"
07:44:43S_a_i_n_tyou'll need to do this on your PC (well, you *can* do it on the DAP, but it's a lot easier to do it using your PC) using a text editor, Notepad is OK to use in this case.
07:44:44LloreanCan't you also just do it from the context menu when selecting that folder or something?
07:44:52LloreanMost options have a way to change them without editing config files
07:45:15S_a_i_n_tLlorean: I *think* there's something like that, but I wasn't I went with the way I know.
07:47:31S_a_i_n_tBut, I think you're right...I believe you can just create a new .dir on the DAp, name it whatever you want, then bring up the context menu on it and select "Set as Playlist Directory" (or similar)...but I'm not 10% if that is the *exact* details, and I don't have a gadget I can check it on nearby.
07:47:53S_a_i_n_t*100% (lol)
07:48:04Lloreanchrisb: Well, check the context menu first. It could save you some effort. :)
07:51:36S_a_i_n_tHehehe...depending on your idea of "effort", I've personally always found the virtual keyboard a _nightmare_ to use (on the iPods at least) to use myself ;)
07:51:58S_a_i_n_tI tend to delegate renaming things to a PC :)
07:52:05LloreanYeah, but if you have a folder already, there's no vkeyboard use needed
07:53:10S_a_i_n_tI imagine the virtual keyboard would be a LOT easier to use on touchscreen targets, but I've never had a chance to test that theory.
07:53:59crculverS_a_i_n_t, Thanks
07:54:12CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
07:54:12*S_a_i_n_t nods
07:54:26crculverMost of the album art I have embedded is hi res, though and this requires the image be less than 96 kb.
07:54:44crculverBut maybe it could be a starting point for FLAC support
07:55:18S_a_i_n_tSetting up AA for Rockbox can be a daunting task...I'm fairly certain there are a few apps that can automate this task though.
07:57:22S_a_i_n_tI believe Mp3Tag can run through a dir and extract embedded AA recursively
07:57:38saratogatheres been some discussion of that patch, but I see none of it was actually added to the patch
07:57:47saratogaperhaps unhelpful or kugel could add their thoughts
07:59:09*Unhelpful thinks that the issue ought to be fixable with the right implementation - a jpeg decoder with a hook to read the next bit of image data, imo.
08:05:50 Quit zivan56 ()
08:09:09 Nick markun_ is now known as markun (~markun@rockbox/developer/markun)
08:11:28***Saving seen data "./dancer.seen"
08:13:10crculverS_a_i_n_t, The whole point of embedding was to avoid the clutter of extra files
08:13:49S_a_i_n_tyes, all depends on wether or not you want AA in Rockbox really...
08:14:19S_a_i_n_tthere are several different naming/placement conventions...I'd suggest looking on the wiki to find one that suits you.
08:15:25S_a_i_n_tIt is *easy* to avoid "clutter" (visually at least) by just changing the file properties to "hidden" for all your AA...then you won't see them in the filebrowser on your DAP at least ;)
08:16:01 Join B4gder [0] (~daniel@rockbox/developer/bagder)
08:22:10chrisbS_a_i_n_t: thanks
08:22:49 Join kugel [0] (~kugel@rockbox/developer/kugel)
08:24:43*JdGordon suddenly wonders if there is any point adding some "is showing a splash?" tags to the skin anguage and some extra suff so the splash viewport can be set in the .sbs
08:26:05crculverS_a_i_n_t, If I can find the time, I'll look into extending the patch to FLAC
08:26:51S_a_i_n_tJdGordon: I could dig that idea...
08:27:12S_a_i_n_tthat way a theme could have a designated "info area"
08:27:31 Join ender` [0] (
08:27:32JdGordonyeah, I'm not sure how it would work though... it might
08:27:43JdGordon(one of those awesome ideas that only happen when on the can)
08:28:00S_a_i_n_tOf which there are many...
08:28:59S_a_i_n_tI can send you some screenshots of my "fat cabbie" soonish...the battery and volume look awesome using the progressbar logic ;)
08:29:37S_a_i_n_tespecially as the reflections are there (and actually move with the battery/volume as well)
08:29:46S_a_i_n_t*a nice touch I thought.
08:36:10*JdGordon needs to see what happens when splashes block
08:36:17JdGordonusing a simple conditional might be enough for this
08:39:25 Join flydutch [0] (
08:40:10 Quit elcan (Ping timeout: 264 seconds)
08:40:33 Join esperegu [0] (~quassel@
08:40:41 Join elcan [0] (
08:42:38amiconngevaerts: The h300 *could* do MTP as well, via the isp1362
08:42:52amiconn(full speed only though)
08:49:36 Join wodz [0] (
08:49:42wodzamiconn: ping
08:55:42wodzaa I got You :-)
08:55:52wodzcan I take You few minutes?
08:58:55 Quit BHSPitMini (Read error: Connection reset by peer)
08:59:41wodzhmm I gues not :-/
09:00:14B4gderwodz: it's better if you just ask what you want and then he can reply when he reads it and gets time to answer
09:00:29B4gderjust higher latency conversation
09:00:52amiconnwodz: Don't ask to ask, just ask
09:00:56linuxstbAm I understanding correctly, that it's stealing 96KB of the codec buffer, and using that to store the album art?
09:02:24*Llorean wonders why there's a limit on the bitmap size, or if that's just the existing limit.
09:03:00linuxstbThat also seems to be stored in the codec buffer.
09:03:26LloreanIsn't there already a place album art is stored, or is that to avoid reopening the file once playback gets to it?
09:03:41wodzamiconn: I don't quite understand lcd_grey_data in lcd-as-m3.S
09:03:47kugellinuxstb: I think it fread()s the image to a seperate buffer and then applies jpeg_decode_from_mem() to it. I don't quite understand the reasoning for that. Also, I'd think the image is already in rum due to metadata buffering
09:04:31wodzwhat does numbers in comments regariding register content mean?
09:04:41wodzpixels or what?
09:07:03 Join petur [0] (~petur@rockbox/developer/petur)
09:07:06 Quit BHSPitMonkey (Remote host closed the connection)
09:09:41amiconnYes, those numbers refer to the pixels within that block
09:10:20wodzat line 433 You have comment d5 = ........................01234567 than You duplicate the content to the upper byte in word of d3 and if You are done it is transfered with write_word
09:10:38amiconnThe m3 asm is one of the most complex low-level greylib functions, because of the interleaved calculation and serial output
09:11:51 Join PRINCESS_FLUFF [0] (
09:12:12amiconnThe duplication is necessary since the greylib works by using fully black or fully white pixels, but the m3's lcd controller is 2bpp
09:12:55wodzamiconn: ok now I understand
09:13:03PRINCESS_FLUFFIs there a way (not too technical) to change the output volume for some plugins? When I listen to SPC music, the volume is about half as high as other songs. So I inevitably turn it up only to have my eardrums bleed from the next song if it's not an SPC file (spc = super nintendo music files)
09:13:31PRINCESS_FLUFFThis is on a Sansa E200 btw
09:13:42wodzamiconn: so basicaly I can loop this part of code right?
09:13:47amiconnLines 437..439 do exactly that.
09:14:10amiconnYes, since in your case the lcd controller is connected in parallel
09:15:15wodzamiconn: how do I test if I do things correctly? Display some jpg or is there some more analitical test included?
09:15:24amiconnIn case of the m3, I'm precalculating one word worth of pixels (lines 404..439), then outputting it bit by bit while calculating the next one in the main loop (444..543)
09:15:53amiconn...and finally outputting the last word (reusing the tail from another routine)
09:16:47amiconnDisplaying a .jpg should work as a test, however, you probably need to calibrate the lib for the hd200's lcd panel
09:16:59 Join Zagor [0] (
09:17:00 Quit Zagor (Changing host)
09:17:00 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
09:17:16wodzso I can use 404..439 to process one word, send it to the lcd and loop this until finished right?
09:17:43wodzhow do I calibrate?
09:17:43amiconnFor this purpose there are two test plugins you need to build: test_scanrate and test_grey
09:19:08amiconntest_scanrate in turn needs a working lcd_blit_mono()
09:19:33 Join efyx [0] (
09:20:03wodzand lcd_blit_mono needs working lcd_mono_data
09:21:37amiconnFirst, you run test_scanrate and increase/decrease the frequency until the black blob at the left stops, or nearly stops. Note down the frequency, then create an #elif section at the top of grey_core.c for your lcd, taking the m3 section as a basis
09:22:31amiconnChange the #define LCD_SCANRATE to the nearest HZ value of frequency you measured.
09:22:53amiconnThen rebuild the plugins, so test_grey will use the new value
09:24:37amiconnIf you run test_grey, you'll see 9 rectangles in a black frame. With left/right you can change the grey level you're adjusting, with up/down you change the value
09:24:40 Join einhirn [0] (
09:25:22amiconnThe outer 8 rectangles are use ordinary ordered dither to approximate a level of grey, only the central one uses the actual greylib mechanism
09:26:14amiconnThe goal is to adjust the center rectangle (with up/down) to match the brightness of the left, right, top and bottom rectangles at each of the levels
09:26:23 Nick fxb__ is now known as fxb (
09:26:28 Join DerPapst [0] (
09:27:21amiconnIn order to do this, watch the display in a not too bright environment, keep it away far enough so the dithering works (arm's length) and try looking exactly perpendicular (as contrast varies with angle)
09:28:05amiconnWhen finished, exit the plugin. It will write a result file to the root. This result file can then be used to create the lcdlinear[] array by using interpolation
09:28:24amiconnThis step isn't automated yet..... sorry
09:29:00kugelare all these steps documented somewhere?
09:29:51wodzso the workflow is: 1) Implement lcd_mono_data 2) rework lcd_blit_mono() 3) compile test_scanrate 4) measure scanrate 5)adjust settings in grey_core.c 6) implement lcd_grey_data 7) rework lcd_blit_grey_phase 8) fiddle with test_grey 9) implement lcdlinear[] based on data from test_grey
09:30:37 Quit DerPapst (Ping timeout: 240 seconds)
09:30:51 Join DerPapst [0] (
09:36:27 Quit kugel (Ping timeout: 276 seconds)
09:37:40 Quit solexx (Ping timeout: 260 seconds)
09:39:13 Join solexx [0] (
09:39:34 Join lpereira [0] (
09:39:47 Join kugel [0] (~kugel@rockbox/developer/kugel)
09:39:54 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
09:45:10 Quit crculver (Quit: Leaving)
09:49:32 Quit kugel (Remote host closed the connection)
09:56:03 Join kugel [0] (~kugel@rockbox/developer/kugel)
10:09:38 Quit esperegu (Remote host closed the connection)
10:11:21 Quit chrisb (Ping timeout: 265 seconds)
10:11:30***Saving seen data "./dancer.seen"
10:16:21 Join esperegu [0] (~quassel@
10:24:20wodzamiconn: I have some doubts how lcd_mono_data is supposed to work. You take first byte of data (LSB of 8 pixels strip) and duplicate it to the upper byte in the word, than You do the same with the second byte. Doesn't this mean You produce twice as much data as needed by lcd?
10:26:37 Quit einhirn (Read error: Connection reset by peer)
10:27:16wodzhmm probably not, nevermind
10:28:24S_a_i_n_t <−− Cabbie + Fat == "Flabbie"?
10:29:03 Join liar [0] (
10:30:22S_a_i_n_tComments on that theme screenshot are welcome, I have only just finished sticking it together now. ;)
10:31:29n17ikhI like it
10:31:42n17ikhgood for landscape targets
10:32:33 Quit hebz0rl_ (Quit: Ex-Chat)
10:32:38S_a_i_n_tit's 176x132 at the moment (Nano 1/2g), but I could be convinced to make versions for other targets also.
10:32:51 Join hebz0rl [0] (
10:33:30n17ikhwell, I'm using widecabbie on my sansa e200, because I like the volume level that appears when you change volume
10:34:18n17ikhI think it's something every wps should have ;)
10:35:08S_a_i_n_tNormally I'd do something like that...but in this case, the volume/battery use the logic from the progressbar, if I did that you wouldn't see the nice, smooth-scrooling volume icon ;)
10:46:49 Join shai_ [0] (
10:47:30 Quit shai (Disconnected by services)
10:47:33 Nick shai_ is now known as shai (
10:49:24 Quit wodz (Quit: Leaving)
10:57:32 Join dfkt [0] (dfkt@unaffiliated/dfkt)
11:31:33 Join wodz [0] (
11:48:25 Join Kamyk_ [0] (
11:48:28Kamyk_hello all
11:48:52Kamyk_i have Philips Go Gear SA3085/02 - is it rockbox will work on it?
11:49:36wodzlook at to figure out
11:49:53wodzonly explicitly listed devices are supported
11:50:29amiconnwodz: There is no repeated duplication, just one duplication per source byte, in order to convert mono data into lcd format
11:50:55Kamyk_i don't find my mp3, but itis any possible to upgrade my mp3 in future?
11:52:43wodzamiconn: I missread the code - I was thinking that lcd_mono_data takes data in lcd native format not b&w
11:53:13amiconnThere's a reason why it has 'mono' in its name :)
11:55:22wodzI am just uploading new firmware and test plugins to try test_scanrate
11:55:40 Quit tchan (Read error: Connection reset by peer)
11:55:44 Quit phanboy4 (Read error: Connection reset by peer)
11:56:07markunKamyk_: it's always possible. Someone (who has such a player or wants to buy one) has to volunteer to work on it.
11:57:05wodzwhat is the scanrate for M3?
11:57:53markunKamyk_: some info:
11:58:03 Join _arbingordon [0] (
11:58:31 Join phanboy4 [0] (
11:58:33 Quit _arbingordon (Client Quit)
12:00:05 Quit BlakeJohnson86 (Ping timeout: 265 seconds)
12:00:20 Quit arbingordon (Ping timeout: 276 seconds)
12:00:21 Join _arbingordon [0] (
12:00:25 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
12:01:38wodzamiconn: how should change the black blob in test_scanrate when changing frequency?
12:03:18 Join BlakeJohnson86 [0] (
12:07:36 Join einhirn [0] (
12:09:50 Nick fxb is now known as fxb__ (
12:11:32***Saving seen data "./dancer.seen"
12:12:10 Join newuser [0] (
12:12:40 Join watto [0] (~watto@
12:16:00newuseri was wondering if someone could help me with something
12:16:39markunnewuser: just ask :)
12:16:54wodzit that something is something I can help maybe I will help :-)
12:18:02 Quit phanboy4 (Ping timeout: 258 seconds)
12:18:30newuseri am trying to install rockbox on my sansa e250
12:18:39newuseri am running ubuntu 9.10
12:18:39 Join b0hoon [0] (~quassel@
12:18:56newuserit won't let me install - says permission denied when i try to install bootloader
12:19:12linuxstbAre you using Rockbox Utility?
12:19:26newusercan't figure out how to run install utility with elevated privs
12:19:27JdGordonlinuxstb: Unhelpful: (others), uuchan can in IRC a few days ago and bassically said that replies should go in the tracker because he cant come in IRC often
12:19:31wodzprobably You have to be root
12:20:01wodznewuser: sudo
12:20:18newuseri am in the directory, type "sudo RockboxUtility"
12:20:23newuserand it says that is not a command
12:20:42wodzsudo ./RockboxUtility should do
12:21:02newuseri am an idiot and just spent 2.5 hours
12:22:02newuserevery time with ubuntu i spend hours just to learn i need a period here, a slash there
12:22:06newuserthanks guys!
12:23:46newuserit works! awesome! thanks again!
12:24:49linuxstbJdGordon: Yes, I agree someone needs to say something on that task. I don't really know what to suggest to uchan, apart from saying that his patch looks to be doing things completely wrong...
12:24:53b0hoonHi, it looks like Rockbox runs well on the Vibe 500, has a complete manual and is supported by the installer :)
12:25:13b0hoonI would like to propose to jump with it to the stable section.
12:26:04JdGordonlinuxstb: yeah, I agree that stealing the codec buffer just looks wrong
12:26:15*JdGordon suggests smoeone comment in the task
12:27:20 Join JohannesSM64 [0] (
12:27:36CIA-5New commit by jdgordon (r25755): centre splashes inside the UI viewport instead of the whole screen. Fixes the flickering statusbar issue and saves alot of fullscreen clears
12:28:30pamauryJdGordon: which tasks ?
12:29:53 Join robin0800 [0] (
12:30:26linuxstbb0hoon: So all the necessary bootloader/installation binaries released ? I don't know what's been done in the past (or if people agree with this), but maybe "stable" should mean that there is an official (3.x) release for that device. So it can be made stable on the day of the next release.
12:32:08b0hoonlinuxstb: yes, there is a bootloader file on the server and the installation works through the RBUtility.
12:32:55pamauryJdGordon: I don't really know that part of the code but stealing the codec buffer is just not possible, we can't have code that keep stealing buffers this way or it will become impossible to follow. There are sufficiently many stealing hacks in the code right now :)
12:33:44b0hoonlinuxstb: don't know how the ports are accepted to the stable state, so that's why i'm asking.
12:36:02AlexPGives the criteria more or less
12:36:08linuxstbb0hoon: Did you release the bootloader? I can't see it tagged in svn.
12:36:39kugel also
12:37:06AlexPkugel: Ah, that's a good one
12:37:59linuxstbb0hoon: Also, did you build the bootloader with a version number, or just the SVN revision?
12:38:05CIA-5New commit by jdgordon (r25756): fix red
12:38:06*linuxstb should document a bootloader release process...
12:39:03b0hoonhmm i guess i forgot to tag the bootloader, i've just sent it to Zagor but it is some build from my svn.
12:39:37AlexPb0hoon: It needs a release version number too, not a SVN revision :)
12:39:51AlexPah, linuxstb already said that
12:41:02b0hoonlinuxstb: yeah, it would be good, cause i don't know what really what to do. I guess that i must tag it somehow as a version 1.0 and then upload it to the server again?
12:42:02linuxstbb0hoon: You just need to build it with "make VERSION=1.0" (I think...) and then tag - the UsingSVN page should tell you how to tag.
12:44:50*linuxstb adds a short comment to
12:45:41*JdGordon got in first with a shorter one :)
12:45:45 Join matsl [0] (~matsl@
12:46:44linuxstbAh flyspray doesn't have the useful "someone else has just replied" warning the forums do...
12:50:46amiconnwodz: M3 has 150Hz scanrate, as you can see at the top of test_scanrate.c
12:52:20amiconnAlso I forgot to mention something: If the scanrate is very high, and the update function is slow (like it's the case with M3), you may use a fractional update frequency in the greylib
12:53:33amiconnScanrate: 150Hz, update frequency: 50Hz; otherwise the update would draw too much cpu power (due to the serial transfer)
12:54:08amiconnOn the MPIO this is probably unnecessary (and if it also has 150Hz scanrate, greylib will probably work very well)
13:00:50 Quit kugel (Remote host closed the connection)
13:02:02wodzamiconn: Do I understand correctly that I should find frequency that black blob doesn't change colour?
13:02:20wodzI mean no moving strips or something
13:02:34 Join adnyxo [0] (
13:02:36amiconnThe black blob should never change colour. If it does, you're very far off with the frequency
13:03:03wodzso what effect should I see when f is ok
13:03:17amiconnIt should stop (or move as slow as possible, since the smallest step is 01.Hz), being fully black on fully white background
13:03:42amiconnIt shouldn't be grey or flickering
13:03:51wodzaa ok
13:03:52amiconnErr, 0.1Hz
13:04:58wodzamiconn: could You look at this:
13:05:06wodzdoes it look ok for You?
13:07:42wodzand this:
13:10:32 Join mikroflops [0] (
13:14:57 Quit mikroflops_ (Ping timeout: 264 seconds)
13:23:22amiconnwodz: Looks good
13:25:39wodzat f=145.3 the blob sloooowly dimms and comes back again
13:28:57 Quit newuser (Quit: CGI:IRC (EOF))
13:35:22 Quit robin0800 (Remote host closed the connection)
13:37:04b0hoonlinuxstb: ok, thanks.
13:41:45 Join neonesis [0] (~jkosidlo@
13:51:52 Join evilnick [0] (~evilnick@rockbox/staff/evilnick)
13:55:01wodzamicon: This is the simples form of lcd_grey_data I think, but doesn't work
13:56:59wodzand for completnes
14:09:27 Join kugel [0] (~kugel@rockbox/developer/kugel)
14:11:35***Saving seen data "./dancer.seen"
14:14:38 Join MethoS- [0] (~clemens@
14:15:39amiconnIt dims and comes back? That's not how it's supposed to be
14:17:47wodzdoes it shoud run up/down or left right when unsynced?
14:18:25amiconnThat depends on how the panel is controlled. On most targets it runs up/down. On the M3 it runs sideways
14:18:47amiconnIf that's the case on MPIO as well, you should set the appropriate #define in test_scanrate
14:18:48 Join Schmogel [0] (
14:19:18amiconn(it will change the orientation of the "blob strip" on the screen to horizontal)
14:20:41wodzhmm we will see
14:34:55*S_a_i_n_t wishes that the "line selector" used magenta as a "magic colour" for transparency also...
14:36:56S_a_i_n_tSo that I could use text colour as the sole visual reference for selection. An effect the theme I'm working on now could do with.
14:39:32 Join advcomp2019_ [0] (~advcomp20@unaffiliated/advcomp2019)
14:39:38 Join evilnick_ [0] (
14:39:52 Quit advcomp2019 (Read error: Connection reset by peer)
14:39:52 Quit evilnick (Read error: Connection reset by peer)
14:39:52 Quit Rob2222 (Read error: Connection reset by peer)
14:39:52CtcpPing from gevaerts!~fg@rockbox/developer/gevaerts
14:39:59 Join Rob2222 [0] (
14:40:13S_a_i_n_tAt the moment, it is possible to do this effect, but only if your backdrop is a solid colour...if you have a gradient effect (or anything else other than solid colour actually) in your backdrop this effect won't work. >:(
14:41:34 Part b0hoon ("Back to work.")
14:43:48pixelmacan't you use the pointer selector with line text colour?
14:44:34S_a_i_n_tpixelma: If I wanted to use the pointer...then, yes.
14:45:10S_a_i_n_tWell,, you can use the pointer *or* the line selector, and only change the text colour with the line selector.
14:45:42S_a_i_n_tthe pointer also annoyingly offsets the icons.
14:48:16S_a_i_n_tpixelma: Yes, just checked up on that now. You can't change the text colour with the pointer, only with the line selector.
14:48:49S_a_i_n_tMy desired effect would be to just use the text colour (or the change in text colour rather) as the "selector".
14:48:59 Quit liar (Quit: Verlassend)
14:49:13kugelshould be trivial to implement
14:49:39S_a_i_n_tImplement away, good Sir. ;)
14:50:22 Quit adnyxo (Ping timeout: 276 seconds)
14:50:48 Quit B4gder (Quit: It is time to say moo)
14:50:52S_a_i_n_tI figure all it needs is to accept magenta as a "magic" colour as well...I doubt many people will want a solid-bright pink line selector ;)
14:50:59S_a_i_n_t...but I may be wrong.
14:57:16 Join geertvdijk [0] (
14:58:32geertvdijkhey all, just updated rockbox to current build again on my sansa, and noticed now that when the screen goes dark after a minute, and you then plug in your headphones (without touching buttons or turning wheels), it wakes up. any plans for pause on unplug/resume on plug?
15:00:07geertvdijkit's an e200 btw, might that be of any importance
15:01:27S_a_i_n_tI'd say it will almost certainly be implemented at some stage...
15:01:46S_a_i_n_tWhen however, would be impossible to predict.
15:02:20 Quit Battousai (Remote host closed the connection)
15:02:33 Join Battousai [0] (~bryan@gentoo/developer/battousai)
15:02:33 Quit antil33t (Read error: Connection reset by peer)
15:02:40 Join antil33t [0] (
15:03:08geertvdijkS_a_i_n_t: would it be hard to implement? I mean, could a semi-experienced developer hack it in using some code from the media control found in plugins now in place of the headphone-detection wakeup?
15:04:12S_a_i_n_tNo idea sorry...
15:04:21gevaertsgeertvdijk: if the hardware can do it (and from what you say it might well do), it should simple. We do have pause on unplug on some devices
15:04:59geertvdijkgevaerts: thanks, I'm going to look into it tonight when I get home. don't expect to be able to do much, but I just might. any idea in which files I should look?
15:05:33 Quit neonesis (Remote host closed the connection)
15:06:10gevaertsgeertvdijk: start by looking for HAVE_HEADPHONE_DETECTION. Possibly some GPIO debug screen can be helpful too
15:06:56kugelgeertvdijk: I'm not aware than any of the sansas the feature you're describing
15:07:19geertvdijkkugel: neither was I, but I thoroughly tested it as it really surprised me.
15:07:23geertvdijkgevaerts: will do, thanks. if I find out anything and/or need some help i'll get right bakc here :)
15:07:33geertvdijk-but +so
15:07:33kugelwhat sansa?
15:07:42geertvdijke250v1, no fm radio
15:08:28gevaertsgeertvdijk: if you find a GPIO pin that does this, you only need to implement headphones_inserted() basically
15:09:34geertvdijkgevaerts: that does sound like something I could do, as long as there's other devices that do it already ^^
15:11:19geertvdijkthis is really odd now, it now does it only sporadically, and I'm making sure each time I check that I'm not accidently touching any buttons/sd cards
15:12:16JdGordonkugel: magenta for that might not be so trivial
15:12:48*JdGordon doesnt think using text colour as the selection thingy would work so well
15:13:17gevaertsgeertvdijk: concentrate on the GPIO debug screen, if the e200 has one (which it probably does)
15:13:24kugeljust disable selector drawing completely if the bar color is the magic color
15:13:47S_a_i_n_tBlargh! <grumble> mumble something, conflicting opinions..., mumble </grumble>
15:13:47geertvdijkyes I just went ahead and plugged and unplugged the headphones concentrating on each individual value of the i/o ports screen
15:14:14 Quit advcomp2019_ (Read error: Connection reset by peer)
15:14:23S_a_i_n_tFWIW, if you're using a solid background colour...text colour as the selector is quite effective.
15:14:39 Join advcomp2019_ [0] (~advcomp20@unaffiliated/advcomp2019)
15:14:45CIA-5New commit by wodz (r25757): HD200 - add FM support.
15:16:39JdGordonS_a_i_n_t: have you actually tried a solid selector bar with that colour?
15:17:00S_a_i_n_tIt's pink, not transparent as I'd expected.
15:18:17kugeltransparency might give some nice effects if used with gradient bars (and implemented properly)
15:18:46geertvdijkgevaerts: so far nothing, but there seem to be more values listed than fit on the screen. do you happen to know if these are all ADC-related things (as the last fully visible ones are)?
15:19:09gevaertsno idea
15:20:18JdGordonadding trancparency will slow it down a fair bit ownt it?
15:20:59kugelI doubt
15:21:36kugelit doesn't really matter if you mix two predefined patters or 1 predefined pattern with the background color
15:21:39JdGordonlcd_fillrect() cant do a simple memcpy() or memset() if that were added
15:22:25JdGordonthat is done on the area, then the text is added
15:22:31kugelit can memcpy from the background color, memset might not be possible. but you usually cannot use mem*
15:22:42kugelon color targets anyway since a pixel is more than 1byte
15:23:22JdGordonit is passed in a fb_Data* so you can
15:24:06JdGordonoh no its not
15:24:46*JdGordon is too tired to try understanding lcd driver code
15:36:19S_a_i_n_tDammit...some changes to lib/viewer.c makes FS #6697 compile witha *bunch* of errors...far beyond my comprehension. Which is a crying shame, as it is by far better than dict,rock, "by far better" is probably an underestimation...
15:39:35 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
15:39:58 Join CGL [0] (~CGL@
15:42:23S_a_i_n_tis there an SVN command to automagically remove .rej .orig files etc?
15:42:56wodzfind is your friend I would say
15:43:16S_a_i_n_tI realise that they *shouldn't* make a difference, but it'd be nice to be able to clean up my tree without killing it and checking out again...
15:43:37gevaertssvn stat, awk and rm can help
15:43:54S_a_i_n_twodz: I was hoping for a more automated soultion...I was kinda hoping "svn revert" would take care of them actually...but, nope. :(
15:44:14gevaertsS_a_i_n_t: what's not automated about find?
15:44:32 Join evilnick_B [0] (~0c140464@rockbox/staff/evilnick)
15:45:04S_a_i_n_tgevaerts: there's no way of knowing if the .orig file is in use currently or not.
15:45:19S_a_i_n_tremoving all of then may bot be the bast of ideas.
15:45:45gevaertsS_a_i_n_t: well, use svn stat and awk then :)
15:46:10 Quit bmbl (Ping timeout: 258 seconds)
15:46:10*S_a_i_n_t has no idea of those two functions...
15:46:18S_a_i_n_tI'm a themer..gimme a break ;)
15:47:12S_a_i_n_twhat is the output of "svn stat" telling me?
15:47:27S_a_i_n_tis that the list of files *not* in the checkout?
15:47:44kugelno .orig files are in use, they're always left over from not applying patches
15:48:11S_a_i_n_tAha...good to know.
15:48:26S_a_i_n_tAs for 'svn stat'? Am I correct in that assumption?
15:48:52gevaertsS_a_i_n_t: it's the list of files that's different from the repository. The first two characters tell you what
15:48:55gevaerts's different
15:49:22S_a_i_n_tgevaerts: they all have a "?" in front...
15:49:40gevaertsThat means they're all not in the repository
15:50:15S_a_i_n_tso, if I removed them *all*, it would be a "clean" checkout again?
15:51:04S_a_i_n_t"svn diff" comes back ok...I'm just not a fan of leftover files.
15:51:08 Quit hebz0rl (Ping timeout: 268 seconds)
15:51:26 Join hebz0rl [0] (
15:51:58 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
15:52:52amiconnS_a_i_n_t: 'svn status' will tell you if you have unversioned files in your tree (marked with '?')
15:53:32amiconnYou can delete all of them if you don't need them anymore
15:54:00S_a_i_n_tIs someone able to give me a command using 'svn stat' and 'rm' that will just kill then all in one fell swoop?
15:54:06*S_a_i_n_t often bash fails...
15:54:13amiconnIf you have your build dir(s) in the same tree (like me), that will come up as unversioned too, of course
15:54:34S_a_i_n_tI do, but that is no loss if I delete those.
15:55:04 Join Gyuri [0] (~bd196e85@gateway/web/freenode/x-tzdilmmeogfyytwq)
15:55:13linuxstbS_a_i_n_t: Do any of your files have spaces in the filenames?
15:55:17pamauryamiconn: something like for file in $(svn status | awk '{if($1=='?') print $2}'); do rm $file; done (untested)
15:55:27pamaury(if you have no spaces in the filenames)
15:55:38S_a_i_n_tlinuxstb: Yep, I guess that's a problem...
15:56:04linuxstbS_a_i_n_t: No, it just makes the script a little trickier...
15:57:16linuxstbS_a_i_n_t: Maybe something like: svn st | awk '/^?/ {print substr($0,8)}' | (while read a ; do echo "$a" ; done)
15:57:27linuxstbReplace "echo" with "rm" when you're happy it's doing the right thing
15:57:36*linuxstb expects there are 50 other ways to do that
15:58:25wodzamiconn: I hanged definition in test_scanrate.c. Now I have horizontal strip. Depending on frequency it flickers more or less until around 146Hz where it stays black and very slowly changes colour but not line by line (some irregular pattern)
15:59:14S_a_i_n_tlinuxstb: THANKS!!!
15:59:31wodzamiconn: Maybe I will record video how it looks tomorrow and than You will tell if this is correct
15:59:40S_a_i_n_tit couldn;t remove the (empty) build dirs for some reason...but otherwise worked like a charm ;)
15:59:55 Part Gyuri
16:00:02*S_a_i_n_t adds this to an alias immediately.
16:00:13linuxstbS_a_i_n_t: If you want to remove empty directories, just do "rmdir *" - that will fail if any of the directories have anything in.
16:00:38linuxstb(but succeed with any empty ones)
16:00:48*S_a_i_n_t always thought that was kinda silly...
16:00:55pamauryS_a_i_n_t: learn a bit of "awk", that's really handful for that kind of stuff
16:01:05S_a_i_n_t*not being able to delete non-empty dirs
16:01:07pamaurythe man of awk is well done also
16:01:43*linuxstb drags this to -community
16:04:32wodzhow to add something to front page of rockbox?
16:05:04linuxstbwodz: Checkout the www module in svn, commit your change, and then ask Bagder or Zagor to update the site.
16:05:22wodzI see
16:06:16 Quit Hillshum (Ping timeout: 245 seconds)
16:10:41 Quit wodz (Quit: Leaving)
16:11:38***Saving seen data "./dancer.seen"
16:12:28 Join panni_ [0] (
16:12:44 Quit Zagor (Quit: Clint excited)
16:16:28*kugel wonders how to replace sdl drawing
16:18:22 Quit JohannesSM64 (Max SendQ exceeded)
16:20:09 Join JohannesSM64 [0] (
16:20:44 Quit JohannesSM64 (Max SendQ exceeded)
16:22:01 Join JohannesSM64 [0] (
16:31:17 Join Schmo [0] (
16:35:00 Quit Schmogel (Ping timeout: 245 seconds)
16:35:14 Quit flydutch (Quit: /* empty */)
16:45:12 Quit ender` (Quit: As a computer, I find your faith in technology amusing.)
16:45:25 Quit matsl (Ping timeout: 245 seconds)
16:46:17 Quit jfc^2 (Remote host closed the connection)
16:49:23 Quit kugel (Read error: Operation timed out)
16:49:24 Join liar [0] (
16:51:03 Join toffe82 [0] (~chatzilla@
16:51:17 Join ender` [0] (
16:54:29 Join kugel [0] (~kugel@rockbox/developer/kugel)
16:58:15 Quit kugel (Remote host closed the connection)
17:00:29 Nick fxb__ is now known as fxb (
17:03:18 Join jgarvey [0] (
17:13:37 Join mt [0] (~chatzilla@
17:14:53CIA-5New commit by mt (r25758): Revert r25739 which added libwmapro to apps/codecs, in preparation to commit the unmodified ffmpeg files first, for the sake of a consistent/complete ...
17:21:08 Join elinenbe [0] (
17:23:21 Quit hebz0rl (Ping timeout: 276 seconds)
17:25:03elinenbeBagder: are you here?
17:25:45 Join hebz0rl [0] (
17:29:57 Join jfc [0] (
17:37:15 Quit lpereira (Quit: Leaving.)
17:46:30 Nick fxb is now known as fxb__ (
17:49:45*linuxstb looks around the wiki for SoC pages... ;)
17:51:03 Quit mt (Ping timeout: 276 seconds)
17:52:34 Join archivator [0] (
17:54:46archivatorIs there a way for plugins to separate their headers and .c files? I'm trying to build flite under Rockbox's build system but mkdepfile seems very confused by the dir structure and insists that there should be a $(BUILDIR)/<whatever>.h file.
17:56:19linuxstbarchivator: You mean put the .c and .h in different subdirectories?
17:56:23archivatorThe original Makefile basically adds an -I$(FLITE_TOP_DIR)/include to every gcc invocation.
17:56:27archivatorlinuxstb: yes.
17:57:06linuxstbarchivator: Probably not. I expect you would need to write your own makefile
17:57:19 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
17:57:45linuxstbWhere in Rockbox are you planning to put the code though?
17:58:03archivatorlinuxstb: under plugins/ for the moment
17:58:21gevaertsfirst get it to work at all I guess?
17:58:52linuxstbI guess it will always be built as some kind of plugin though?
17:59:01linuxstb(even when integrated into the core) ?
17:59:02archivatorgevaerts: get what to work? I can't get it to build since mkdepfile is confused by #include "flite.h" statements where flite.h is not next to the .c file
17:59:37gevaertsarchivator: I mean, that's what you're trying to achieve by building it as a plugin first
17:59:38archivatorlinuxstb: part 1 is making it a plugin, so that I have something to test out and work on. Part 2 is porting it to the core
17:59:49archivatorAh, sorry, I misunderstood.
18:01:07 Quit pamaury (Quit: Quitte)
18:01:46linuxstbarchivator: Which makefile did you copy for "flite.make" ?
18:02:12archivatorlinuxstb: pdbox.make since it allows for extra compile flags
18:02:52linuxstbHave you tried adding -I there?
18:02:57 Quit petur (Quit: *plop*)
18:02:59gevaertslinuxstb: won't work
18:03:04archivatorlinuxstb: it breaks long before that - the dep file is all wrong
18:03:16gevaertsThe dependency generation doesn't use flags in the SUBDIRS makefiles
18:04:38linuxstbgevaerts: Yes, I'm just catching up.... (reading plugins.make)
18:04:42 Join komputes [0] (~komputes@ubuntu/member/komputes)
18:05:34*gevaerts found out about this when working on r24916
18:05:43*linuxstb assumes the flite source code is more than a handful of files....
18:06:20linuxstbarchivator: Maybe Zagor has some ideas - this build system is mostly his work...
18:06:26archivatorlinuxstb: you assume rightly, about 100 files total
18:07:42linuxstbarchivator: Did you also say that flite voices are compiled as code?
18:07:48S_a_i_n_tgevaerts: Did you get any further with your "pebbles" idea? Or is that just dead in the water now?
18:08:04archivatorlinuxstb: correct
18:08:38linuxstbarchivator: Does flite offer loadable voices, or are they all compiled into a big binary?
18:08:43gevaertsS_a_i_n_t: yes and no :)
18:09:21linuxstbgevaerts: Yes, you've got further, and no, it's not dead in the water?
18:09:41archivatorlinuxstb: usually, they're statically linked into a binary blob but it shouldn't be *too* difficult to compile them for a specific address and load them on demand
18:10:12gevaertsThe general idea is still there, but the particular approach is dead in the water
18:11:00S_a_i_n_tIf it were a bigger job, it would/could have made a nice GSoC project.
18:11:23linuxstbMaybe archivator will squeeze it into his project... ;)
18:11:37*S_a_i_n_t wants to be able to use a demo in his WPSes for cases where there is no AA ;)
18:11:41***Saving seen data "./dancer.seen"
18:11:48gevaertsS_a_i_n_t: it's big enough I think, and it is on the ideas list!
18:12:09*S_a_i_n_t stares at archivator then ;)
18:12:10archivatorlinuxstb: well, if anyone makes a relocatable loader, that would come in really handy.
18:12:49*archivator bows his head in shame
18:12:58*S_a_i_n_t attempts to "Jedi Mind Trick" the GSoC students...
18:12:59archivatorI'm afraid that sorta stuff is out of my competence
18:13:09gevaertsTo work well it also needs the buflib work
18:13:28gevaertsS_a_i_n_t: you should have done that earlier! Projects have been selected now
18:13:52linuxstbMaybe I should have asked this earlier, but what's the point in TTS in the core? Is the intention to replace the voice files and talk clips?
18:14:02S_a_i_n_t"These are not the droids you're loo...I mean, write pebbles.rock"
18:15:01S_a_i_n_tlinuxstb: I was wondering that also...I *believe* that is the goal?
18:15:14S_a_i_n_t*removal of .voice .talk files
18:15:18archivatorlinuxstb: ultimately, yes. In practice, desktop engines offer far wider language support.
18:15:20 Quit geertvdijk (Quit: ChatZilla 0.9.86 [Firefox 3.6.3/20100401080539])
18:15:35linuxstbarchivator: And better quality...
18:15:44archivatorFor now, removing .talk files for English, adding voicing to plugins/database, that kinda thing.
18:15:55archivatorAnnouncing the artist/track, too :)
18:16:07gevaertsBasically voicing everything that currently isn't
18:16:12linuxstbThat could also be achieved with talk clips though. (the database)
18:16:36linuxstbOr maybe not easily...
18:16:44archivatorlinuxstb: right, and the moment you change a single tag, you'd need to re-voice everything
18:17:09archivatorWell, not *everything*, but it'd be rather hard to maintain
18:17:14gevaertsIt's a tradeoff I think. Regenerating talk clips every time isn't that wonderful
18:17:24S_a_i_n_tAnnouncing track info in the wps would be cool, also volume, anything that changes really...awesome.
18:17:42*S_a_i_n_t sends good vibes archivator's way.
18:17:52gevaertsIf I'd use voice, I guess I'd prefer voice clips for the menus and TTS for track info
18:18:28*linuxstb wonders what happened to midgey's work on voicing plugins
18:18:41gevaertslinuxstb: wasn't that just translating?
18:18:48gevaertsOr was it both?
18:18:52*gevaerts doesn't remember
18:19:02linuxstbgevaerts: Hmm, I'm not sure...
18:19:06*S_a_i_n_t *thinks* it was both...
18:19:15S_a_i_n_t*but, isn't sure either.
18:19:15archivatorIn any case, the general consensus is that my current situation with the dependencies can't be fixed quickly, correct?
18:19:40linuxstbarchivator: I think the current consensus is that no-one who is here can fix it quickly...
18:20:06gevaertsarchivator: the easiest way to get it working would be to use full relative paths in the include statements I think
18:20:17gevaertsNot good in the long term of course
18:20:37archivatorgevaerts: can't I just add an EXTRADEPINCLUDES in the Makefile and add that to mkdepfile?
18:20:56 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
18:21:18 Join lifeless_ [0] (~lifeless@
18:25:44 Quit lifeless__ (Ping timeout: 265 seconds)
18:27:08linuxstbarchivator: Just in case you don't know it already, "make V=1" is helpful.
18:28:35archivatorlinuxstb: didn't know about that but it doesn't really help in any way. Thanks, though.
18:37:30 Join abc [0] (
18:39:03 Quit abc (Client Quit)
18:42:59 Join pamaury [0] (~c2c7a50a@rockbox/developer/pamaury)
18:45:01 Quit Schmo (Read error: Connection reset by peer)
18:48:20 Join Luca_S [0] (
18:48:42gevaertsarchivator: the problem I think is that all dependencies are done in one go, so that would add those includes to all C files
18:50:23archivatorgevaerts: would it? from what I can see, mkdepfile calls gcc with the -MM and -MG parameters. Building on that, if I add an -I<path-to-flite-include> to the call, it should resolve the header files properly. Or am I missing something fundamental?
18:50:46 Join kramer3d [0] (~kramer@unaffiliated/kramer3d)
18:51:38 Quit elinenbe (Quit: CGI:IRC (EOF))
18:52:50gevaertsarchivator: you could try. I suspect it's not that simple though
18:58:20 Quit Luca_S (Quit: CGI:IRC (EOF))
19:01:43 Join ATTIKIT [0] (
19:04:21archivatorgevaerts: turns out, it *is* that simple. Good news is it works. Bad news is it's ugly and hacky.
19:04:49gevaertsarchivator: you're sure it doesn't influence other code at all?
19:04:50 Quit ATTIKIT (Client Quit)
19:05:15 Join Boldfilter [0] (
19:05:46pixelmaI thought miidgey added a patch to the tracker with as far as he came. I don't know what the question "wasn't it just translating" means
19:06:27archivatorgevaerts: if by influence you mean "add flite headers to every single file's dependencies list", then no. That's as far as it goes right now since it doesn't really compile :)
19:06:55gevaertsok then. That's reasonable I think
19:07:51pixelmain this context, I mean. I sure can translate it literally
19:11:21 Quit shai (Read error: Connection reset by peer)
19:11:33 Join shai [0] (
19:11:53 Quit Utchybann (Ping timeout: 276 seconds)
19:12:27 Join Utchybann [0] (
19:13:24 Quit DerPapst (Quit: Leaving.)
19:17:37 Quit JohannesSM64 (Ping timeout: 240 seconds)
19:19:18 Join JohannesSM64 [0] (
19:21:10 Join kugel [0] (~kugel@rockbox/developer/kugel)
19:21:58kugelgevaerts: do you think I should paste what I wrote in my application to the wiki page?
19:24:18kugelI'm a tad bit worried I might fail through the final evaluations if I happen to not manage a complete port (without SDL) to a mobile device, even though that's not the primary objective
19:25:14 Join Horscht [0] (~Horscht2@xbmc/user/horscht)
19:27:07archivatorUpon further investigation, the build fails exactly where it should - at the first call to malloc. Awesome.
19:27:43kugelyou could link tlsf malloc from the codec lib
19:27:57kugel(for now, malloc is a nogo for the core)
19:28:21 Join Silverwolf [0] (
19:29:16archivatorkugel: there's also dbestfit in pdbox..
19:29:35SilverwolfI just registered on the Wiki, under the name RichardNease, and I'd like to have write permissions so I can add to the Wiki
19:29:40kugeldidn't pdbox switch to tlsf as well because there were problems with dbestfit?
19:29:53kugelthat's what I seem to remember from last year
19:31:13archivatorkugel: no idea, I only saw that it used dbestfit at some point in the past..
19:33:59kugelI might work on relocatable plugins for the next year's gsoc
19:34:04 Quit Silverwolf (Quit: CGI:IRC)
19:37:06 Join panni__ [0] (
19:37:08 Quit panni_ (Read error: Connection reset by peer)
19:38:53gevaertskugel: maybe copy the actual technical content but leave out the milestones and dates for now? They should be there in a few weeks of course, but I think it's important to be able to start technical discussions as soon as possible
19:39:47 Join panni_ [0] (
19:40:06 Join Kitr88 [0] (
19:40:34 Quit panni__ (Read error: Connection reset by peer)
19:41:26kugelgevaerts: right, I'll do so very soon
19:41:47kugelgevaerts: what I wondered, if anyone has tried to use rockbox' own threads when running the sim on an arm device
19:42:08 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
19:42:08gevaertsprobably not I guess
19:42:29gevaertskugel: when you tried the sim on your mini2440, was that with X on it?
19:42:42kugelit really bugs me that I can't manage to get it running on my mini, I highly suspect I'm simply OOM even though I think 50MB free ram should be plenty
19:42:58kugelI will buy a bigger sd card and setup a swap partition
19:43:18kugelgevaerts: no, I also can't get X to run :\
19:43:25 Quit Kitar|st (Ping timeout: 246 seconds)
19:44:39gevaertskugel: my latest plain unmodified svn sim crashes on my phone somewhere in unicode.c...
19:44:50 Quit Kitr88 (Ping timeout: 268 seconds)
19:45:37kugelvery strange
19:46:16 Quit Xerion (Ping timeout: 252 seconds)
19:46:17 Join Kitar|st [0] (Kitar_st@
19:46:50gevaertsI'll do some debugging
19:48:34kugelgevaerts: I meant to ask you if you have some tips on getting some useful qemu image(s)
19:49:57gevaertshm, not really. My qemu setup is a scratchbox thing, probably not ideal for what you're doing
19:56:19 Quit archivator (Quit: Leaving)
20:00:59kugelI have no experience with qemu at all :\
20:01:36*gevaerts decides to have a go at
20:02:00 Join DerPapst [0] (
20:02:01 Join RichardNease [0] (
20:02:48RichardNeaseCould I be added to the WikiUsersGroup so I can edit some pages?
20:03:41kugelRichardNease: what would you like to edit (just curious)?
20:04:38RichardNeaseThe pages referring to the Sansa line of products
20:04:41 Join Xerion [0] (~xerion@
20:04:47 Join dfkt_ [0] (~dfkt@unaffiliated/dfkt)
20:05:07 Join phanboy4 [0] (
20:06:09 Quit dfkt (Disconnected by services)
20:06:15 Nick dfkt_ is now known as dfkt (~dfkt@unaffiliated/dfkt)
20:06:49kugelRichardNease: done, but promise to be nice :)
20:07:05RichardNeaseI will be nice. Thank you
20:08:53gevaertskugel: should be useful I suspect
20:09:05kugelyes, I was looking through it as well
20:09:10kugelit features X at least
20:09:16 Quit RichardNease (Quit: CGI:IRC)
20:09:39gevaertsI'm now downloading the lenny stuff there
20:10:36 Join fyrestorm [0] (
20:11:44***Saving seen data "./dancer.seen"
20:13:06 Join robin0800 [0] (~quassel@
20:15:26saratogaqemu is really easy, i tested the tremor stuff in it
20:15:35saratogathe problem is its also really, really slow
20:19:07 Join Moltare [0] (
20:22:18 Quit kramer3d (Quit: Leaving)
20:22:27 Join Courage [0] (
20:22:28 Quit Moltare (Client Quit)
20:22:35 Nick Courage is now known as Moltare (
20:24:03MoltareGood evening, ladies and gentlemen. I seem to have managed to entirely freeze my rockboxed iPod Video 60GB, such that it won't accept the standard reset solution and Windows claims the device has malfunctioned when plugged into USB.
20:24:11MoltareCould anyone offer advice on this, please?
20:24:19 Join Luca_S [0] (
20:25:15saratogatry holding the buttons for even longer
20:27:40Moltareaha! Thank you
20:27:58MoltareTwo minutes is rather more than the eight seconds claimed, but whatever works :)
20:28:07Moltarecheers and g'day to you all
20:28:13 Quit Moltare ()
20:30:10gevaertskugel: have you ever looked at libao? (
20:30:42kugelnever heard of it
20:34:18 Quit Luca_S (Quit: CGI:IRC (Ping timeout))
20:40:09 Join lyngaas [0] (
20:43:54 Join wodz [0] (
20:44:01kugelgevaerts: not sure if that's useful, that's basically sdl with audio only, isn't it?
20:44:18gevaertsI haven't looked in detail
20:44:22kugelI thought we wanted to use os capabilities (or libs that are as close to the os as possible)
20:45:12kugelor maybe we just wanted alternatives for sdl so that we don't rely on sdl being available on the target plattform
20:45:32gevaertson n900 you basically want either gstreamer or pulseaudio
20:46:46kugelI think my alsa-sim work actualy uses pulseaudio, there's some strange interactions between them I haven't figured out completely yet
20:47:03kugelbut I read that alsa wraps around pulseaudio which wraps around alsa
20:47:14gevaertsyes, it's an interesting setup
20:48:15kugelI'm quite happy with the alsa thing, as we basically only need plain pcm playback. the only drawback seems the volume changing, but that's software volume changing is acceptable IMO
20:48:35kugelalso alsa should be available on every linux setup
20:48:53gevaertsI'm not sure what volume changing behaviour in an app should be anyway
20:49:02 Join fml [0] (
20:49:38 Join tchan1 [0] (
20:49:44kugelwell it's a definitely a nice to have, but I expect you'd rather use hardware volume buttons (which affect the whole system) or some OS widget for that on mobile devices
20:50:04kugelbut I think a desktop app should have that
20:50:44kugelall the other linux programs I looked at also do sw volume changing, but in a later step (not in the dsp) as I currently implemented
20:51:00kugelthe only difference is the 3s delay when changing volume :P
20:51:07LloreanWell, I mean, isn't that as simple as allowing software volume adjustment in Rockbox, but disabling the keys on targets where good hardware volume is considered the "better" solution (and setting the default volume to line leve)
20:51:09 Quit tchan (Read error: Connection reset by peer)
20:51:15fmlpixelma: domonoky: amiconn: bluebroth3r: hello. Please compare FS #11186 and FS #11228 (both are German translations). If I had noticed FS #11186 I wouldn't have done FS #11228. But now it's there and there are somedifferences. What do you prefer?
20:51:26LloreanThen people can still adjust it through the settings if they want it set at a specific level relative to their other apps, but won't accidentally adjust it all the time.
20:52:03LloreanI mean "as simple" in terms of the disabling of the buttons, not necessarily the implementation of software volume.
20:52:34kugeldisabling the buttons would be complicating it a slight bit :) I don't think there's a need to do that
20:52:58LloreanWell, if we're depending on hardware volume, there's no real reason there should be separate Rockbox volume controls in the WPS too
20:53:05gevaertss/disabling the buttons/not mapping the buttons in the first place/ ?
20:53:09LloreanA player doesn't need what appear to be two different volume buttons.
20:53:13Lloreangevaerts: Yes
20:53:36 Quit robin0800 (Quit: No Ping reply in 180 seconds.)
20:53:58kugelin multitasking environments having separate volume levels for each app is appreciated, and mobile phones do multi tasking nowadays
20:54:00 Join robin0800 [0] (~quassel@
20:54:41kugeldisabling the buttons would mean an artificial limit I cannot see a rationale behind
20:54:55 Nick fxb__ is now known as fxb (
20:55:21LloreanIt reduces complication - two sets of volume controls are complicated.
20:55:46LloreanIn cases where we cannot 'steal' the hardware volume buttons for use as Rockbox volume when it's in the foreground, I think we should just treat them as *the* volume buttons and not have our own.
20:56:15kugelbut it's not confusing, on normal PCs you have a much wider variety of abilities to change volume. people will just use what suits them best
20:56:17domonokyfml: maybe you can combine those language patchs ? I also like "Tastenkürzel" better then "spezialtaste".
20:56:33gevaertsAt least with pulseaudio you can set the volume per stream
20:56:46Lloreankugel: On normal PCs you also have a speaker knob, which is the primary thing you use to adjust volume after you've got the apps set up relative to one another.
20:56:58gevaertsi.e. the rockbox volume controls could be rockbox only while still leaving it to the environment
20:57:14LloreanI'm not saying disable the ability to adjust the volume
20:57:22LloreanI'm just saying, not map keys to it in the WPS / List.
20:57:24LloreanUse the setting.
20:57:31kugelgevaerts: with pulseaudio you even get a third way if I understood it correctly
20:58:00fmldomonoky: But then it's weird. We do not assign somethingto the "Tastenkürzel" (but the phrase says so). We either assign Tastenkürzel to a function or we assign function to the Spezialtaste.
20:58:45kugelgevaerts: do you think I should rather look into pulseaudio again? I thought using alsa directly is good enough
20:58:53wodzhmm I just fired up test_scanrate.c on my mini 1G and I have scanrate 92.3Hz while in grey_core.c there is 87Hz defined for mini
21:00:05gevaertskugel: it doesn't matter. You use the one that's most appropriate for the device you're working on
21:00:21pixelmawodz: oh, you also have a Mini 1G? Seems to be rare
21:00:38wodzI do
21:00:42kugelgevaerts: I'm not working a device currently, but exploring possiblities :)
21:01:36pixelmafml: Spezialtaste sounds a bit like it would give you an extra button somehow and Tastenkürzel is the usual translation for this kind of thing (although I don't have a strong preference yet)
21:01:38kugelbut I think alsa is pretty appropriate, I think it is the one with the least overhead and perfectly fits our needs
21:01:48 Part watto
21:02:06gevaertsour needs are defined by the environment :)
21:02:22kugelbut that only applies for linux targets yes, on android I'm going to need jni calls
21:02:51 Part Boldfilter
21:02:52gevaertsI'd assume so
21:03:14domonokywodz: as a hint: if you want your new MPIO target be added to the automated buildsystem, you have to add it to www/buildserver/builds in svn :-)
21:03:53kugelI've shortly looked into it, and there's a pcm-buffer-callback-like jni call so that might not be a problem
21:04:48fmlpixelma: but the English original also doesn#t use the usual term "Shortcut". And that's OK since the "Shortcut" is const but the function assigned to it varies. Normally, when "Shortcut" is used, it's the other way around: you have a function and you assign one or another shortcut to it.
21:05:29kugelgevaerts: well, w.r.t. to audio we only really need pcm passthrough, not any other fancy audio feature that an os might provide
21:06:13wodzdomonoky: thanks for the tip
21:06:32pixelmawodz: about the scanrate thing - I know that it varies a bit at different temperatures and probably slight differences in the panels (I mean same type but build somewhere else or so). It could also be that the difference is in 1G vs. 2G Mini and the former hasn't been tested yet. Not sure how important this difference you measured is
21:06:38kugelespecially not audio format conversion
21:07:32gevaertskugel: definitely
21:07:40wodzpixelma: I'm just gonna compare two bilds with current scanrate 87 vs. measured 92
21:08:03 Join Boldfilter [0] (
21:08:59domonokywodz: also trunk/tools/ and www/index.t could need updating for the new target.
21:10:18kugelindex.t too?
21:11:51domonokyafaik index.t contains the targets on the front page
21:12:40kugelah right
21:14:51 Quit robin0800 (Remote host closed the connection)
21:16:08pixelmafml: hmm... maybe we can find something different? Spezialtaste also doesn't cover button combos (which it is in some cases). I also don't like "Spezialtaste-Einstellungen", if at all then "Spezialtasten..."
21:17:33fmlpixelma: the point is: it's *the* Taste, there are not many of them :-/
21:18:26pixelmaand it would be really nice if we could find something better for "zurückspulen"... although it's the exact translation of rewind... but you're not winding your MP3 or whatever audio file
21:18:46kugelI suggested Tastenkürzel as well
21:19:26fmlkugel: but I already explained why I don't quite like it.
21:19:30 Quit Boldfilter (Quit: Boldfilter)
21:20:15kugelsometimes it's not a matter of preference, but what the usual/common translation is
21:20:51pixelmafml: btw., putting "we pe es" in thevoice string is unclean. Whether your TTS can deal with the abreviation correctly depends on your TTS and there is a script in Rockbox which does the correction for some TTSs
21:21:35*pixelma actually doesn't want to sound so negative... sorry
21:23:43pixelmaTastenbelegung? (for one key or key combo only though)
21:24:37fmlHow about: Tastenkürzel belegen mit ...? But what Tastenkürzel? I give up for now. Maybe tomorrow it will come to my mind by itself.
21:25:32wodzwhy there are no links to ports pages in "unusable ports" section on frontpage?
21:25:45wodzis it "by design"?
21:26:44kugelI don't think so
21:27:21kugelthe actual question is why unstable ports are clickable links
21:28:23 Join Luca_S [0] (
21:28:46wodzI didn't noticed that stable targets aren't
21:30:17 Join Boldfilter [0] (
21:31:59LloreanI think "stable" doesn't really need to be clickable, but "unstable" helps users trying to decide just how "unstable" something is.
21:32:06LloreanMeanwhile, "unusable" is a pretty explicit description.
21:32:45 Join Schmogel [0] (
21:33:06*kugel join #gsoc-de
21:33:08 Quit Boldfilter (Client Quit)
21:33:24 Quit Luca_S (Quit: CGI:IRC (Ping timeout))
21:34:39wodzLlorean: I found it counterintuitive - the stable target should have clickable links redirecting to download section IMO
21:34:45 Quit fml (Quit: ChatZilla 0.9.86 [Firefox 3.6.3/20100401080539])
21:35:13Lloreanwodz: But people should be using rbutil
21:35:15LloreanNot the download page.
21:35:27LloreanThat's why theres a single link to the installer, instead.
21:35:30LloreanRight at the top of the stable list
21:35:52LloreanIf you add a whole bunch more links, people are more likely to click the target one (which is less helpful) than the installer one
21:35:57 Join w1ll14m [0] (
21:36:14 Quit JohannesSM64 (Ping timeout: 264 seconds)
21:36:18wodzso make the installer link more obvious (aka BIGGER)
21:36:23LloreanAnyone with a fuze experiencing 1-5 minute boot times like Overlord Nathan is reporting in the forums?
21:37:00wodzwhat is the syntax of www/buildserver/builds ?
21:37:00 Quit hebz0rl (Read error: Connection reset by peer)
21:37:00topiknot my fuze v1 on current svn
21:37:09topikcouple of seconds max
21:37:11Lloreanwodz: But the fact that it's not bigger isn't a reason to add confusing links that could misdirect users...
21:37:46 Quit phanboy4 (Quit: Leaving)
21:37:49FlynDiceLlorean: where is that fuze report?
21:39:11wodzLlorean: ok it is going to be more philosophical dispute
21:41:44Lloreanwodz: The front page is mostly user oriented, so ideally the links should be the ones users need to find, rather than developers
21:41:52LloreanOr, more specifically, the ones we want users to find.
21:42:41topikbiggest thing that influenced my fuze's boot time was a big font, but never 1-5 minutes
21:42:56wodzLlorean: I am not going to redesign frontpage don't be afraid :-)
21:42:58linuxstbwodz: I think it would be confusing if the stable targets linked to the downloads, and the unstable targets linked to the port pages. I agree the unusable targets should be links, and the "installer" link much more prominent.
21:43:28*Llorean can agree that the installer link should be more prominent.
21:44:13Lloreanlinuxstb: I think that the "status summary" link is really where people interested in the unusable targets should be directed first, though.
21:44:18CIA-5New commit by wodz (r25759): Add MPIO HD200 to the frontpage
21:44:24LloreanIt's a lot more user oriented of a page.
21:44:41linuxstbLlorean: The same for "unstable" as well?
21:45:26LloreanFor unstable, we've already told the user "this should at least work and you should be at least able to listen to audio" from the summary, so the individual links should probably actually be to either a warning anchor (for what's broken) or an install anchor on the individual target page
21:45:28linuxstbLlorean: Or the port pages should have a user-friendly status summary at the top.
21:45:35LloreanSince in those cases, people clicking the links *should* be people wishing to install
21:45:46domonokywodz: about the syntacs of the builds file. its; buildtype:configure targetname: Userreadable targetname:outoutfile:configure options needed:<special number saying how complicated the target is to build>
21:45:47LloreanUser friendly summary on the status pages would be great.
21:45:55LloreanBut I don't know how likely it is to happen. :)
21:46:11linuxstbLlorean: I'm not sure I agree with that - we're calling them "unstable", so people may want to know what that means before installing.
21:46:26linuxstb^what that means specifically for that player
21:46:30LloreanI just think, from one perspective the current links make a lot of sense: Stable - installer. Unstable - install instructions as user friendly as we have. Unusable - Our most user friendly description if *why* they're unusable.
21:46:33wodzcan I ask for update of frontpage?
21:46:53linuxstbwodz: You need to ping Bagder or Zagor (I guess I just have...)
21:47:01LloreanBut I do agree that it'd be nice if we could mix "explaining why they're unstable" and "easy to find install instructions"
21:47:21domonokywodz: you need to poke Bagder or Zagor for the index.t update. all other changes are updated automatically.
21:48:05linuxstbLlorean: I just think it's more "intuitive" if the name of a player in a list of targets links to a description of that player, rather than a download. Otherwise, the list should have the heading "Downloads".
21:49:17 Join JohannesSM64 [0] (
21:50:16wodzdomonoky:−−target=xxxx - xxxx is the configure name or what?
21:50:19topikfunman (logs): after your latest cpu freq changes my clip+ got very useful again. before it often didn't turn on without holding power 20 secs. now it starts always and it hasn't crashed for several mp3 albums
21:50:51Lloreanlinuxstb: Well, currently that's basically what they do.
21:50:53domonokywodz: thats the options you need to pass to configure to build your target.
21:51:26kugelwodz: actually it's what ever your switch case accapts for the mpio. i.e. also the number
21:51:39Bagderthe unusable list of targets look funny with one <li> only
21:52:13kugelthat's only for indentation :p
21:52:32Bagderfunny indentation
21:52:44 Join anewuser [0] (anewuser@
21:52:44wodzand the final number? How do I get score?
21:52:45 Quit anewuser (Changing host)
21:52:45 Join anewuser [0] (anewuser@unaffiliated/anewuser)
21:52:49Bagderit indents even without <li> actually
21:53:35Bagderwodz: the easiest is to just copy from a similar target and let it get calibrated correctly later on
21:53:49 Join Boldfilter [0] (
21:58:32 Quit JohannesSM64 (Quit: WeeChat 0.3.2-dev)
21:59:55FlynDicetopik: Can you startup with uSD inserted?
22:01:15CIA-5New commit by wodz (r25760): Add MPIO HD200 to automatic build system
22:08:24 Join bluebrother [0] (
22:08:24 Quit bluebrother (Changing host)
22:08:24 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
22:09:21CIA-5New commit by wodz (r25761): Add MPIO HD200 to
22:11:47***Saving seen data "./dancer.seen"
22:12:19 Quit bluebroth3r (Ping timeout: 276 seconds)
22:17:02LloreanFlynDice: It shouldn't take long for him to uninstall? The fact that he's "still uninstalling" might also be a sign.
22:17:32wodzanyone remembers FS number of the patch that changes ipod charging current from 100mA to some more?
22:20:06FlynDiceLlorean: Yes, it's pretty easy to flash an unpatched OF and delete the .rockbox dir. He's going to be upset when he does that and it still takes 5 mins to boot due to file system problems though I bet...
22:37:12amiconnwodz: If your blob is not moving, but isn't pure black on pure white, you didn't hit the proper scanrate, but a (simple) ratio of it, e.g. 2/3
22:37:54wodzamiconn: I fired up test_scanrate on my mini 1G so now I know how it should look
22:38:11amiconnThis often happens for new targets because the scanrate varies a lot between different targets
22:38:36wodzbut is it possible that scanrate is obove 150Hz?
22:39:17amiconnIt might be. Most controllers use something between 70 and 80 Hz. M3 has 150 Hz
22:40:12wodzis there some simple test to see if blit_mono is working correctly?
22:40:49amiconnIf I'm setting twice the real scanrate on my mini (176Hz instead of 88Hz), I'm seeing two (smaller) blobs
22:41:34wodzis scanrate specific to controller or to the lcd panel?
22:42:11amiconnIt's specific to the controller, its programming, and the clock source for the controller
22:42:51amiconnAt 2/3 (58.7Hz) I'm also seeing two blobs, but dark grey on light grey, not black on white
22:43:26wodzhave You looked at pastebins of lcd_grey_data? Do this look ok for You?
22:43:33amiconn1/3 gives one dark grey on light grey blob
22:43:50amiconn(and a bit of flickering since that's only 29.2Hz)
22:44:42amiconnI've seen those for mono blitting, but not for grey
22:45:39ranmaFlynDice: I have a crazy idea for possibly making your Clip+ 'usable'
22:45:55amiconnHmm, the blob changing colour but not position sounds weird
22:46:02*amiconn would like to see that
22:46:06 Join merbanan [0] (
22:46:38ranmaAttach an attiny to the JTAG port, which on poweron halts the CPU, pokes instructions into memory and resumes a new microbootloader and goes to sleep.
22:47:27*amiconn looks
22:47:29ranmaThe new bootloader would then read a few sectors from microsd to load the rockbox loader.
22:47:45wodzamiconn: Now I can say I saw no blob but stripe
22:48:09ranmaThen I'd just have to patch rockbox so that it always uses the external SD in place of the internal one :)
22:48:20amiconnHmm. Did you try 292Hz?
22:48:50wodzamiconn: no, I didn't go above 200Hz
22:49:07*amiconn wonders whether it's really that high, but I think that may even be the case for the TL0350
22:49:20FlynDiceranma: I was just going to suggest using the external but current svn is rather sketchy as far as the uSD goes
22:50:29 Quit fyrestorm (Quit: Ur skills' fireproof like a wooden panel -- U got feds talking leet on your IRC channel!)
22:50:53FlynDiceThe earlier builds worked pretty well but for awhile now the clip+ uSD function has been pretty iffy. I've actually been investigating that today.
22:51:43amiconnwodz: Your lcd_grey_data has several problems.
22:52:34amiconn(1) There must be 8 pixels per block, not 4
22:53:09wodzI am not surprised :-) I don quite understand what this should do
22:54:43amiconn(2) %a2 is a length (rather: a width, in pixels), not an end address. If you want to use it as such, you have to do two things to convert it: (a) multiply by the number of pixels in a block, because one pixel == one phase byte (b) add the start address
22:55:35wodz2) is easy
22:55:50wodzI don't understand (1)
22:56:13amiconnWell, you have a TL0350. The TL0350 uses 2 bit vertical interleaved pixels
22:56:21 Join M3DLG [0] (
22:56:58wodzI know
22:57:15amiconnThis means there are 8 pixels per word. The greylib uses two buffers of one byte per pixel, one buffer for the pixel values (only possible values are 0..128) and one for the phases
22:57:20wodzI need one word of data for single vertical strip of pixels
22:57:38 Nick fxb is now known as fxb__ (
22:58:14 Join dantje [0] (
22:58:24amiconnHowever, these buffers are *not* laid out like an ordniary byte packed image, but they are in the controller's pixel order, in order to speed up lcd_grey_data()
22:59:34amiconnHence, you need to calculate 8 pixel values for every word written to the lcd controller. YOu only calculate 4.
23:00:14amiconnOh, and you also forgot the final duplication, before writing to the controller
23:00:50wodzamiconn: but move.w to the lcd is actually single byte
23:01:08wodzlcd is wired to the lower byte only
23:01:12amiconnOk, then you need to transfer the same byte twice
23:01:23amiconnLike in lcd_mono_data()
23:01:39amiconnBut you also need to calculate a whole byte, not a half byte
23:01:55 Quit pamaury (Quit: Page closed)
23:03:09amiconnThe greylib always works with mono pixels, even if the controller does 2bpp natively, for two reasons
23:03:50amiconn(1) The interaction between native shades and temporal dithering would be hard calculate
23:04:21kugelgevaerts: I get a kernel panic :(
23:04:32gevaertsGive it back!
23:05:01kugelah, older kernel works it seems
23:05:43kugelactually I had that kernel panic too with qemu from the mini2440-qemu repo
23:06:15amiconn(2) Even though calculation would still be manageable (I even did it for a quick test, back in the days of the old graylib), doing this creates nasty interference, due to the way native shades work: the controller does basically the same as we do in software, but we can never match the frequency 100%
23:06:19kugelhm, no, it's not doing anything now instead of giving the panic
23:07:30gevaertskugel: the setup from works for me. Lenny and 2.6.26
23:07:40kugelah, there we go
23:08:12kugeli used the images from here now
23:09:10kugelmaybe abi incompatibilities
23:09:48 Quit domonoky (Read error: Connection reset by peer)
23:11:07 Quit evilnick_B (Quit: Page closed)
23:11:10wodzamiconn: Thank You.
23:13:44amiconnLike this: (untested, so no guarantees...)
23:13:56 Quit M3DLG (Read error: Connection reset by peer)
23:14:25 Join M3DLG [0] (
23:15:15amiconnIf back-to-back writes are too fast for this controller, you can spread out the two writes a bit. The first could be moved up two instructions, directly after lsr.l #7, %d1
23:16:14*amiconn likes using lea.l for a register arithmetics :)
23:17:51wodzI'll test tomorrow
23:23:36wodzhmm. Maybe problem with scanrate comes from lcd_mono_data doing back-to-back writes to the controller?
23:27:05 Quit kugel (Remote host closed the connection)
23:27:10 Quit shai (Quit: Leaving)
23:27:25 Quit w1ll14m (Read error: Connection reset by peer)
23:28:28 Quit wodz (Quit: Leaving)
23:29:12 Quit bmbl (Quit: Bye!)
23:34:23 Quit CGL (Quit: A trabajarrrrrrrrrrrrrrrrrr)
23:34:24 Join kugel [0] (~kugel@rockbox/developer/kugel)
23:40:02 Quit bertrik (Quit: De groeten)
23:42:08 Join adnyxo [0] (
23:46:54 Quit adnyxo (Ping timeout: 260 seconds)
23:49:35 Quit esperegu (Remote host closed the connection)
23:51:35 Quit ender` (Quit: Top reason why compilers are like women: Miss a period and they go crazy)

Previous day | Next day