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

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

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

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

#rockbox log for 2009-01-25

00:02:27LloreanIt's in RButil?
00:02:59 Join killan [0] (
00:04:20 Quit killan_ (Read error: 104 (Connection reset by peer))
00:05:18 Join solexx_ [0] (
00:08:07Bagderlow_light: two yellows to work on now then! ;-)
00:10:05 Join QuickStart [0] (
00:10:47 Quit jhMikeS (Nick collision from services.)
00:10:53 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
00:12:16Bagder95 builds
00:12:52 Quit solexx (Read error: 60 (Operation timed out))
00:13:39 Quit bertrik ("Leaving")
00:14:34 Join yhuang [0] (
00:15:48 Quit HBK (Client Quit)
00:19:20 Join HBK [0] (
00:22:07 Quit parafin (Read error: 60 (Operation timed out))
00:22:57 Quit obo (Read error: 60 (Operation timed out))
00:23:02 Join obo [0] (
00:32:27 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
00:38:25 Quit n1s ()
00:38:38 Join slyyf [0] (n=slyf@
00:39:02slyyfHey guys, I am modding the conwond2 because I foundout it actually boots on the P2
00:39:18slyyfI compiled it, it worked, well, booted and told me there was no power left..
00:39:32slyyfI changed the resolution so it would actually looks proper
00:39:36slyyfand now it wont compile
00:40:19slyyfA bunch of things tell me the lcd isnt supported, and other things tell me something_X not defined
00:40:32slyyfis there something that contains a list of resolutions the lcd supports?
00:40:53Bagderthere is things in the code that depend on the screen size, yes
00:41:32slyyfBagder: K, any idea what I should do?
00:41:53Bagderyes, edit/fix those places of the code
00:42:52slyyfcan I just not build the apps?
00:43:21Bagderyou mean the plugins?
00:43:44Bagderthose are not the only things in rockbox that need the size fine
00:43:56slyyfah ok
00:43:59Bagderbut disabling them is a start
00:44:18slyyfIs there something I should pass to make to do that?
00:44:26Bagderedit configure
00:46:45slyyfThanks for your help, I am not used to this codebase yet
00:52:29 Join Xerion [0] (
00:53:17 Join v445 [0] (
00:54:28 Quit v445 (Client Quit)
00:57:28slyyfWhat is the significance of LCD_Pixelformat?
00:57:49 Nick JdGordon|zzz is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
00:58:01slyyfOh, nevermind
00:58:21 Join tvelocity [0] (
01:00:43casainhoI am trying now to load the firmware and I would like to know if there is diferencies between bootloader and application linkers scripts files (.lds)
01:01:04casainhoBagder: is there any big diferencies?
01:01:23 Quit _lifeless (Remote closed the connection)
01:01:53kugelcasainho: I'm not sure, but I think you need to add codecs and plugin sections
01:01:58 Join _lifeless [0] (n=lifeless@
01:02:21casainhokugel: ok, I will look on that ;-)
01:02:42*kugel isn't the man to talk with about lds though
01:02:45casainhothe code mounts correctly the FAT partition ;-)
01:03:20 Join tvelocity[a] [0] (
01:03:29casainhoand I think I need to use load_raw_firmware(), right?
01:03:32slyyfThe framebuffer doesnt appear to like that resolution, even though it compiled correctly, does anybody know of any code that should be changed when changing resolution?
01:03:54kugelcasainho: no, load_firmware should be it
01:05:24kugelslyyf: have you changed LCD_WIDTH and LCD_HEIGHT?
01:05:36casainhokugel: do you know why not the load_raw_firmware?
01:05:40kugelthat should be sufficient actually, at least if the lcd driver is done right
01:05:47casainhokugel: what are the diferencoies?
01:06:03kugelcasainho: no, but almost every other target uses load_firmware
01:06:41kugelI'd guess raw doesn't handle the header, which is added to the binary by scramble
01:07:06casainhoah, okok :-)
01:07:10kugeland load_firmware does handle is (reading the model string and doing a basic checksum)
01:07:37slyyfkugel: Yes, that is what I changed, I am booting the ConwonD2 firmware on the P2, with the smaller resolution it wasnt using the screen corectly, but was readable, now its not readable, but uses the whole screen
01:08:51kugelslyyf: well, I don't know d2 code, but you should look into the lcd driver
01:09:18kugelthe other code is basically known to be independent of the resolution if LCD_HEIGHT and LCD_WIDTH are correct
01:09:57kugelit might be, that in some spots in the lcd driver the immediate nummber is used instead of LCD_WIDTH/_HEIGHT
01:11:14slyyfkugel: I dont know much about LCDs, but there is something in here setting an HTIME, and a VTIME, is that resolution dependent?
01:12:04kugelI can't really tell you
01:12:24kugelI would just look out for numbers which are equal to the D2's lcd dimensions
01:12:41kugelor 1 lower
01:12:52 Join parafin [0] (
01:13:31slyyfI looked for the actual resolution, I will look for one lower
01:13:32kugele.g. if the x resolution is 320, you should search for 320 and 319 (and replace them with LCD_WIDTH and LCD_WIDTH-1 respectively)
01:13:49kugelevery other change is probably just guess work
01:13:54Lloreankugel: Be careful.
01:13:57slyyfkugel: Yes, looking
01:14:01LloreanThere could be 320 and 319 for other reasons.
01:14:16kugelLlorean: of course
01:14:27kugelI should've been clearer, sorry
01:14:58slyyfLCDC_VTIME1 = LCDC_VTIME3 = (0<<16) | 239;
01:15:07 Quit Lynx_ (Remote closed the connection)
01:15:07 Join tvelocity[away] [0] (
01:15:12slyyfits 320x240
01:15:52kugelslyyf: see, this might be a good catch. but as Llorean said, those number doesn't necessarily have to be the lcd dimensions (it's worth a try though I'd say)
01:16:08slyyfI will try it
01:16:29 Quit tvelocity (Connection timed out)
01:16:46kugelslyyf: oh, and if you succeed, make a photo! ;)
01:17:43slyyfkugel: Its still displaying all funkly ):
01:17:51slyyfOh, should I have reconfigured and make cleaned?
01:17:58slyyfI just ran make
01:18:17kugelnot sure
01:18:25kugelbetter to play safe I guess
01:18:48slyyfOk, it'l take a few mins to build
01:18:49 Join akur [0] (
01:19:11kugelI don't think make clean is needed actually
01:19:43slyyfkugel: I didnt recompile anything, so I did so anyways, all it did was rebuild the image from the existing buid
01:19:49kugelslyyf: well, if that all doesn't work, I can't really help you. I haven't even looked once into the d2 code
01:19:56slyyfNo prob
01:20:11slyyfIs there a configure option to not bother compiling all the codecs?
01:20:17 Quit parafin (Read error: 60 (Operation timed out))
01:20:36kugelI just recall, that the lcd is the only real thing known to be different between the d2 and p2, so it might require more work
01:20:51kugelslyyf: make bin
01:22:01kugelslyyf: how "readable" was it exactly without changing the code?
01:22:13slyyfIt told me there was no battery power
01:22:28slyyfMaybe this doc I read that told me the lcd resoltion was wrong..
01:22:43slyyfnope, its right
01:23:13 Join itcheg [0] (i=62db4767@gateway/web/ajax/
01:23:16slyyfatm it looks like..youknow the old crt monitors when you would put them into an improper that
01:26:00slyyfwhats 0x3bf as an int..
01:26:24kugelgoogle "0x3bf in decimal"
01:28:43JdGordonkugel: you done any more work on the custom vp patch?
01:29:33kugelnot recently, no
01:29:37JdGordonwoo 3 more builds!
01:29:39 Quit tvelocity[a] (Success)
01:31:34QuickStartlol whens the release
01:35:21kugelJdGordon: what builds?
01:35:42JdGordonthe phillips player
01:36:05slyyfwhat does lcd_update_rect do?
01:36:27 Quit tvelocity[away] (Connection timed out)
01:37:13LloreanUpdates a portion of the LCD instead of the whole LCD
01:37:47kugelJdGordon: ah it was added to the build table, I didn't notice
01:38:19 Join Acky [0] (
01:41:27 Quit jhMikeS (Nick collision from services.)
01:41:33 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
01:52:36 Quit Acksaw (Read error: 110 (Connection timed out))
01:53:24***Saving seen data "./dancer.seen"
01:55:42 Part akur
01:56:43 Quit QuickStart (Remote closed the connection)
02:03:16 Quit casainho ("ChatZilla 0.9.84 [Firefox 3.0.5/2008121622]")
02:10:34 Quit ender` (" They could trim up the framework by treating System.DBNull the same as String.Empty, ala Oracle-style...")
02:15:37 Join parafin [0] (
02:15:37 Quit parafin (Remote closed the connection)
02:22:34slyyfkugel: It appears to need to fiddle with the bpp value, allthough I will have to digup a datasheet for the controller
02:26:10 Quit T0paz (Read error: 60 (Operation timed out))
02:32:17 Quit kugel (Read error: 110 (Connection timed out))
02:32:38 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
02:33:12mcuelenaereT0paz (logs): I'll be having exams, so I won't be around the next 3 days or so
02:36:20 Join rocko [0] (
02:37:21 Join goffa [0] (n=goffa@
02:38:10 Quit mcuelenaere ()
02:40:34 Join SirFunk [0] (
02:42:40 Quit BigBambi (Remote closed the connection)
02:43:05 Quit SirFunk (Remote closed the connection)
02:43:30 Quit culture (Read error: 110 (Connection timed out))
02:43:50 Join parafin [0] (
02:44:25slyyfI hate datasheet hunting
02:45:12 Join SirFunk [0] (
02:45:46 Quit SirFunk (Remote closed the connection)
02:47:07 Join QuickStart [0] (
02:47:23 Join SirFunk [0] (
02:47:57 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
02:52:38 Join MethoS [0] (
02:54:46 Join akur [0] (
02:54:57 Part akur
02:55:06 Join tvelocity [0] (
02:55:28 Join gromit` [0] (
02:57:31slyyfI think I found it
02:57:43 Quit efyx (Remote closed the connection)
03:00:00 Quit QuickStart (Read error: 104 (Connection reset by peer))
03:00:03 Quit MethoS (Remote closed the connection)
03:03:30slyyfOhh, this one has an onboard controller
03:03:45 Quit gromit`` (Read error: 110 (Connection timed out))
03:09:49 Join kugel [0] (
03:10:13 Quit low_light ("CGI:IRC 0.5.9 (2006/06/06)")
03:10:48kugelUnhelpful: I think I can finally test pf here
03:13:29 Join kugel__ [0] (
03:13:41 Quit kugel (Nick collision from services.)
03:13:42Unhelpfuli hope you have some tall covers like me, that will pop over the gap :D
03:13:43 Nick kugel__ is now known as kugel (
03:15:23 Join phinze [0] (
03:16:05kugelUnhelpful: well, the database doesn't initialize
03:16:40Unhelpfulgah, what? the DB worked on the sim...
03:16:43 Quit tyfoo (Read error: 104 (Connection reset by peer))
03:17:47Unhelpfulhave you used DB before on clip?
03:18:42kugelI have the feeling that my clip is broken..
03:18:47kugelI just can't get a working FS
03:20:18Unhelpfulfsck perhaps?
03:21:20 Join toffe82_ [0] (n=chatzill@
03:21:23kugelI just formatted several times
03:21:56Unhelpfulheh, if it's still hosed after a format, yes, you may have a real problem :/
03:22:28 Join tvelocity[a] [0] (
03:26:34kugelis it possible, that storage doesn't work with fat32, but only fat?
03:27:06kugelthe sansa OF uses fat, but I need to format to fat32 to make rockbox work
03:27:27kugelI'll try to enable fat16 support in rockbox, and let the OF format, later
03:28:51 Quit toffe82 (Read error: 145 (Connection timed out))
03:30:33 Join tvelocity[away] [0] (n=tony@
03:32:54 Quit BigBambi (Remote closed the connection)
03:35:16 Quit tvelocity[away] (Remote closed the connection)
03:35:50 Join Buri [0] (
03:35:57 Quit tvelocity (Connection timed out)
03:37:09 Join QuickStart [0] (
03:42:26 Quit PaulJam (".")
03:42:34 Part Buri
03:45:05 Quit tvelocity[a] (Connection timed out)
03:45:15 Join barrywardell [0] (n=barrywar@
03:48:49 Quit phinze ("leaving")
03:49:05 Join phinze [0] (
03:53:10 Quit Nico_P (Remote closed the connection)
03:53:25***Saving seen data "./dancer.seen"
03:54:16 Join Thundercloud [0] (
03:55:13 Quit fyrestorm (Read error: 104 (Connection reset by peer))
04:00:49 Quit QuickStart (Read error: 104 (Connection reset by peer))
04:06:48 Join Barahir_ [0] (
04:09:23 Quit Strife89 ("Bed.")
04:10:13 Quit yhuang (Read error: 60 (Operation timed out))
04:15:05 Join blkhawk- [0] (
04:17:47 Quit itcheg (" ajax IRC Client")
04:20:39 Quit kugel ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
04:22:37 Quit Barahir (Read error: 110 (Connection timed out))
04:27:28 Join Zembla [0] (n=zembla@
04:30:39ZemblaI was wondering, is it feasible to install rockbox on a non-formatted iPod?
04:31:10LloreanThe Rockbox installation requires a working iPod first.
04:31:30Zemblaright, so then that it's filled to the brim with music is not a hindrance?
04:32:12LloreanAs long as there's space for the Rockbox files.
04:32:24 Quit blkhawk (Read error: 110 (Connection timed out))
04:32:42Zemblaok, will it recognise the music files already present in my iPod?
04:33:03 Nick blkhawk- is now known as blkhawk (
04:33:15Zemblaand if so, will it arrange/sort them like the original firmware would?
04:33:24Zemblajust a matter of hitting the ground running :)
04:33:39kadobanZembla: that rather depends what format they are in. the answer you are probably looking for is that it will play all files the ipod OF plays, except DRMed ones (iiuc)
04:34:40Zemblarighto, then installing rockbox will change nothing to the usual way of using the ipod (in the sense that I can always boot into the OF), it won't disrupt itunes or any of that?
04:35:26Zemblaam I correct in understanding that it's basically a dual boot system for my ipod, and that either interfaces won't interfere with the basic system functions of the others (to put it vaguely)
04:39:08 Quit miepchen^schlaf (Read error: 101 (Network is unreachable))
04:46:49kadobanZembla: that's the theory at least. i don't have an ipod so i hesitate to answer with any kind of assurance.
04:47:06Zemblabut the theory holds up for the specific device you use it on?
04:48:33kadobanZembla: yes (e200). although i barely use the OF, they don't interfere with each other
04:49:16krazykitZembla, it is a dual-boot system. rockbox will not affect the functioning of the apple firmware
04:57:21 Quit Thundercloud (Read error: 104 (Connection reset by peer))
05:03:41Zemblaright, well, thanks for all the help, trying it right now
05:03:43Zemblac ya later
05:03:48 Quit Zembla ("For great justice!")
05:11:13 Join midkay [0] (n=midkay@rockbox/developer/midkay)
05:12:39slyyf*grumbles* telechips decided to password there downloads section
05:12:43slyyfI need the datasheet
05:16:39slyyfFound one for the TC76.may help
05:30:08 Join Darksair [0] (n=user@
05:41:46 Quit XavierGr ()
05:49:20slyyfDoes aybody have a copy of the tcc78x datasheet?
05:49:21 Join Makuseru [0] (
05:49:58 Part Aurix_Lexico
05:52:24 Join caveman26 [0] (
05:53:16 Quit Makuseru (Read error: 104 (Connection reset by peer))
05:53:28***Saving seen data "./dancer.seen"
05:53:56caveman26any info on a rockbox build for the samsung t9?
05:57:14slyyfcaveman26: none thus far
05:58:01advcomp2019caveman26, there is a thread about the t9 but it is not supported yet
06:02:46 Quit Zoxc ()
06:06:49caveman26well I was just curious.. I had a sansa c200 with rockbox.. the os was nice but the sansa had terrible audio outputs.. the t9 has truly good sound and, but a crappy OS that plays ogg files, but crashed at that crashes at the end of some of the songs
06:14:12advcomp2019caveman26, the sansa fuze and clip can play ogg vorbis with the sansa firmware and rockbox is being worked on for both of them and they have better sound quality compared the PP based sansa.. there is c200v2 and e200v2 that uses the same chip as the clip and fuze.. there is a v2 clip and v2 fuze, but i am not sure how far they are compared to the v1 clip and v1 fuze
06:21:30slyyfanybody understand this bitwise: LCDC_HTIME1 = (0x2d<<16) | 0x3bf;
06:31:57kadobanslyyf: you might want to specify which you're asking: (what does it do/why does it do it/what do the numbers mean/why isn't it doing something else)
06:32:57slyyfI figured it out
06:33:11slyyfits the problem I have been hunting down for the past four hours
06:33:20slyyfer..or more
06:33:21kadobanah, congrats :)
06:34:15slyyf3bf=((width of screen*3)/width of pixel)-1
06:36:56tmztadvcomp2019: do you know what FS bug is playback on fuze?
06:38:11advcomp2019tmzt, nope
06:38:22caveman26whats the battery life like on this new sansa's
06:39:26caveman26I know the battery in my c200 seemed like it would go on forever... if i could deal with crappy anlog section long enough
06:40:56caveman26as a matter of fact i never once ran it completely dead, unless I forgot to turn it off
06:41:49caveman26so I do give sansa credit for one hell of a good battery
06:48:58advcomp2019caveman26, the clips is about 13 hours.. i have not checked how long with rockbox for sure.. the fuze from what i heard is about 20 hours.. i do have a fuze but i have not test rockbox on it because it is v2 fuze
06:49:50tmztadvcomp2019: audio playback works on clip in rockbox?
06:51:07tmztI have V01.01.11T, after installing rockbox bootloader made with mkams using that version
06:51:16tmztI assume that means it's a v1?
06:52:01advcomp2019there is still playback issues on the clip
06:52:18advcomp2019yea you have v1 fuze
06:52:35tmztslyyf: are you serious about that last part?
06:53:33tmztmy fuze doesn't playback at all (mp3 or wav at least) and freezes when trying, current svn, the divide by zero change didn't affect it
06:56:58 Join pronto [0] (
06:59:30 Part toffe82_
07:00:18 Join yhuang [0] (
07:10:04 Join CaptainKwel [0] (
07:12:52slyyftmzt: which part?
07:13:09slyyftmzt: that math? yes, I checked the datasheet
07:14:36tmztI mean what it means, 3*the width part
07:14:51tmzt 3bf=((width of screen*3)/width of pixel)-1
07:15:17slyyftmzt: Yes, but I just now nottices thats the same as width in pixels*3
07:15:32slyyftalk about overcomplicating things
07:15:52slyyfI think...
07:17:05tmztI haven't read the datasheet for this device, it's just commonly used to set certain bits
07:22:42slyyfah, no, your right
07:22:50slyyfI was tyring to figure out why it was setting taht
07:22:58slyyfand thats what the value its setting in is
07:23:01slyyfnot the operation it preforms
07:23:54 Quit caveman26 (Remote closed the connection)
07:23:56tmztyou mean it's a packed 24-bit display?
07:25:05slyyfYa, I believe so
07:25:43slyyfNow I am just trying to figure out why it thinks its twice as wide as it is and puts everything wider then it should thus cropping it
07:27:51 Quit CaptainKewl (Read error: 110 (Connection timed out))
07:30:05 Quit CaptainKwel (Read error: 110 (Connection timed out))
07:52:40 Quit phinze (Read error: 110 (Connection timed out))
07:53:31***Saving seen data "./dancer.seen"
07:56:03 Quit JdGordon (Remote closed the connection)
07:59:08 Quit Darksair ("Emacs = ESC-Meta-Alt-Ctrl-Shift")
08:07:02 Quit suom1 (Read error: 104 (Connection reset by peer))
08:07:40 Join JdGordon [0] (
08:08:16 Join suom1 [0] (
08:16:47 Join romain_ [0] (
08:16:47 Quit romain (Read error: 104 (Connection reset by peer))
08:26:43 Quit rocko ("Leaving")
08:49:13 Join at0m [0] (n=a548c80b@gateway/web/cgi-irc/
08:58:39 Join Rob2223 [0] (
09:01:10 Part pronto
09:09:01 Join n1s [0] (n=nils@rockbox/developer/n1s)
09:10:54 Join rocko [0] (
09:13:54 Join fyrestorm [0] (
09:21:01 Quit markun (Remote closed the connection)
09:22:01 Quit slyyf (Read error: 110 (Connection timed out))
09:29:17 Quit Rob2222 (Read error: 110 (Connection timed out))
09:45:02 Join ranol [0] (
09:48:34 Part ranol
09:53:35***Saving seen data "./dancer.seen"
09:53:55 Join markun [50] (n=markun@rockbox/developer/markun)
10:21:26 Join flydutch [0] (
10:27:51 Join bertrik [0] (
10:28:40 Join Acksaw [0] (
10:42:15 Quit at0m ("CGI:IRC")
10:43:05 Quit JdGordon (Read error: 104 (Connection reset by peer))
10:43:39 Quit Acky (Connection timed out)
10:45:30 Join JdGordon [0] (
10:47:56 Quit Horschti ("Verlassend")
10:56:03 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
11:00:20 Join casainho [0] (n=chatzill@
11:00:54casainhohello :-)
11:01:06casainhoI am getting this error: /home/cas/Documentos/rockbox_player/rockbox/build/apps/action.o: In function `get_action_worker':
11:01:08casainhoaction.c:(.text+0x2b8): undefined reference to `get_context_mapping'
11:01:27JdGordonlook in the apps/keymaps/keymap-* files
11:02:08casainhoI am trying to build the firmware... and that error and others happens on linker stage, right on the end of the building...
11:02:16JdGordonlook in the apps/keymaps/keymap-* files
11:03:11casainhoare you sure? because I am getting other erros like:
11:03:33casainhorockbox_player/rockbox/build/firmware/libfirmware.a(pcm.o): In function `pcm_play_data_start':
11:03:35casainhopcm.c:(.text+0x274): undefined reference to `pcm_play_dma_start'
11:03:57JdGordonthey are two completly different problems
11:05:55casainhoplease look here:
11:06:00bertrikcasainho, link errors like that are to be expected for subsystems that are not implemented yet. What was done for the clip, was creating "stubs" for those functions.
11:07:46casainhostubs? empty functions?
11:08:01bertrika stub (at least that's how I call it) is indeed an empty function
11:10:42casainhoso, do I have to creat a keymap-*.c for my target? is tere any wiki page explaining what I have to do, on this stage?
11:10:56 Join culture [0] (
11:12:28 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
11:13:56bertrikcasainho, I don't know about the keymap-*.c file, but I think there's no wiki page to explain it. You'll probably just have to look at files for other targets to get an idea on how to implement it.
11:14:18 Join Mirra [0] (
11:14:34 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
11:14:38casainhookok - luckly someone put on source a "keymap-newtarget.c" with some explanations ;-)
11:15:00 Join gregzx [0] (
11:15:09bluebrotherBagder: the build-info file now has a release section, but that is empty. Can you check that?
11:15:14Mirrahm.. so rockbox will not work on my nano 3rd generation?!
11:15:28scorcheMirra: thats what the front page says..
11:16:42Mirrayeah, but maybe there has been some development, which made it possible..? *littlehope*
11:17:15Mirrais it actually possible?
11:17:20bluebrotherMirra: scroll down. Then: Page was last modified "Jan 9 2009"
11:17:23MirraI mean from a technical poin of view
11:17:50scorcheMirra: it sure isnt possible when no one is working on it
11:17:53bluebrotherif you figure how to break into the encryption system and figure the hardware itself, ...
11:18:40MirraI guess we're not talking about the itunes.db file (wit the SHA-hashs of the songs)
11:21:37 Quit Mirra (Read error: 60 (Operation timed out))
11:23:17 Join Mirra [0] (
11:25:09pixelmaUnhelpful: I don't fully get your reply re. action contexts in PF. My statement was just a suggestion to only have one or two generic PLA contexts at all that make sense and scrap all the others. If you need more or something else than that, let your plugin use its own button or action definitions, don't "connect" it to others with using PLA. In pf I think that such a generic one (like I described) could work, that's all.
11:25:38 Part Mirra ("Konversation terminated!")
11:27:15 Join tyfoo [0] (
11:28:19pixelmain its current form, the vertical list browsing is broken on e.g. the c200 anyways
11:36:17 Join {phoenix} [0] (
11:53:36***Saving seen data "./dancer.seen"
11:58:44 Join PaulJam [0] (
12:03:47 Quit barrywardell ()
12:12:33 Join MethoS [0] (
12:15:16 Join ender` [0] (
12:21:10 Join faemir [0] (
12:28:40 Join gregzx_ [0] (
12:29:55 Quit gregzx_ (Client Quit)
12:32:08 Quit gregzx (Read error: 104 (Connection reset by peer))
12:35:34 Join robin0800 [0] (
12:36:46 Quit robin0800 (Read error: 54 (Connection reset by peer))
12:37:32 Join robin0800 [0] (
12:39:55 Quit robin0800 (Remote closed the connection)
12:39:58 Join pyro_maniac [0] (
12:40:05 Join wpyh [0] (n=william@
12:41:22PaulJamUnhelpful: ping
12:49:31 Join robin0800 [0] (
12:50:48 Quit robin0800 (Remote closed the connection)
12:52:40 Join robin0800 [0] (
12:52:52 Quit MethoS (Remote closed the connection)
12:53:13 Join MethoS [0] (
12:57:14 Quit MethoS (Remote closed the connection)
12:58:49 Quit robin0800 (Remote closed the connection)
12:59:02 Join robin0800 [0] (
12:59:31 Join tvelocity [0] (
13:10:38 Quit casainho ("ChatZilla 0.9.84 [Firefox 3.0.5/2008121622]")
13:12:13 Quit {phoenix} (Remote closed the connection)
13:21:45 Join MethoS [0] (
13:24:39Bagderbluebrother: is now meant to get the release info
13:25:04Bagderdue to a little mistake of mine it doesn't yet, but it should get there in the next update
13:26:20 Join Thundercloud [0] (
13:32:48 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
13:33:39bluebrotherBagder: ok. I guess next update means next daily build?
13:35:40bluebrotherok. I'll finish rbutil and test tomorrow then. Now the only thing I'd like to get a good solution for is to retrieve the latest rbutil version
13:36:07Bagdersymlink on the download server to the newest one?
13:36:13Bagderso that we have a fixed url
13:36:27Bagderpossibly a text file holding the version number too
13:36:39 Quit bertrik (Read error: 113 (No route to host))
13:37:39bluebrotheryep, I want to retrieve the latest version number. My idea is to have a menu entry Help / Check for update which displays something "new version 1.0.10 available"
13:39:04 Join miepchen^schlaf [0] (
13:39:07Bagderok, let me know how you'd like such a meta-data file to look like and I'll make it available
13:39:49bluebrotherok, I'll think about it −− we should be able to have different versions for different OS
13:40:06Bagderyes, and make sure 32bit and 64bit linux can differ too
13:41:06 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
13:51:20 Join nplus [0] (n=nplus@243.131.Globcom.Net)
13:52:52amiconnHmm. I need opinions...
13:53:20amiconnI want to unify the colour handling of the mono/greyscale target screendump and simulator ui
13:53:39***Saving seen data "./dancer.seen"
13:53:49amiconnBut there's the special case called Clip.
13:54:22amiconnHow important is the simulation of the gap between the two different coloured parts?
13:54:56amiconnSimulating the gap causes quite some extra work
13:54:56Bagdermy guess would be not terribly important, but I've never used a clip so I can't tell for sure
13:58:11 Join Seed [0] (
13:58:15 Join gregzx [0] (
13:59:43 Join Horschti [0] (n=Horscht@xbmc/user/horscht)
14:06:21 Join slyyf [0] (n=slyf@
14:14:34UnhelpfulPaulJam: what's up?
14:14:59PaulJamresizing produces weird colour on coldfire (H300)
14:15:03 Quit robin0800 (Remote closed the connection)
14:15:15 Join kugel [0] (
14:15:39Unhelpfulhrm... since the resize-uses-emac commit, probably?
14:15:56PaulJamyes. one revision earlier shows up fine
14:17:18 Quit EspeonEefi ("さよなら")
14:17:27Unhelpfulis it in pictureflow, or other places as well? PF does some funny things that core scaling doesn't
14:18:06Unhelpfulkugel: it uses 32x32->64 C multiplies on ARM. coldfire lacks an instruction for that, but the emac can provide 32x32->40, which is enough.
14:18:08 Quit Horscht (Read error: 110 (Connection timed out))
14:18:21kugelwhat's "the emac"?
14:18:26PaulJamUnhelpful: i first noticed it in sliding-puzzle, but it happens in the WPS too. i haven't tried pictureflow.
14:19:11 Join nibbler [0] (
14:21:35 Join EspeonEefi [0] (i=eefi@SAFFRONCITY.MIT.EDU)
14:21:55Unhelpfulkugel: a multiply-accumulate unit that coldfire has. it can do 40-bit integer operations, and also has some support for specific operations to accelerate high-precision fixed-point math
14:22:20 Join bertrik [0] (
14:23:21kugelah, thanks for the clarification
14:23:34 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
14:23:41 Join robin0800 [0] (
14:23:58pixelmashould I expect something weird on greyscale coldfire too?
14:25:45Unhelpfulthe scaler is essentially fixed-divisor rational math until the output stage, it's calculating values that are N times the actual pixel value. it uses 8.32 fixed-point division to scale these back to the 0-255 range.
14:25:51Unhelpfulpixelma: possibly. :/
14:26:18UnhelpfulPaulJam: is it any particular parts of images that have trouble? very bright or dark areas?
14:28:54 Quit slyyf (Read error: 110 (Connection timed out))
14:33:20PaulJamUnhelpful: i made some screendumps. on the left side is how it should look like:
14:34:13PaulJambut on most other covers the errors are less visible.
14:34:16Unhelpfulexcellent! are the color blocks in the bottom set full brightness?
14:35:24bertrikblue is missing, red and green are swapped?
14:35:40 Quit solexx_ ("leaving")
14:36:28Unhelpfulhrm, but, in the yellow block, the bad pixels are bright red, only their green is missing
14:37:05bertriksorry, I was wrong, I'll leave this to the experts
14:37:10BigBambiManual people, is there a list of the \opt{} options anywhere?
14:37:10PaulJami'm not sure what you mean with full brightness. i made the image in paint using only colours that have only 255 or 0 in the fields for RGB.
14:37:30Unhelpfulthis is all quite strange, the emac asm should be doing exactly what the 64-bit C version does
14:38:22UnhelpfulPaulJam: that's exactly what i meant. are these images scaled up, or down?
14:39:12PaulJamthe cover.bmp is 150x150 and the WPS uses 75x75.
14:44:54Unhelpfulhrm... so the downscaler will be adding up exactly four input pixels... and they're being mutiplied by 75*75, and then the final asm step divides by 22500. i really can't understand how neighboring output pixxels that should be the same color are not, unless my asm is bad, and there's junk being left in %acc0
14:46:02JdGordonamiconn: I would tihnk that simulating the gap would make it much better... but if its too much of a PITA then probably not worth the effort....
14:47:44PaulJamUnhelpful: could this happen because of the dithering?
14:49:05soapIs the Gigabeat T believed to be a flash-drive version of the Gigabeat S? Or is it likely completely different hardware?
14:49:26Unhelpfulshows what i know, first thing in the morning. didn't even think of that, but yes, the dithering could easily be responsible...
14:49:57Unhelpfulsoap: if it has the same restore process, i can think of a very quick, dirty way to find out... assuming we really mean *entirely* the same hardware.
14:53:36amiconnJdGordon: Atm the sim does that, but with quite some extra effort.
14:54:03amiconnMy unification idea would become quite a bit simpler if it wouldn't have to simulate the gap
14:54:09BigBambirasher (or other lang people) - If I want to change a string (as it is wrong), can I just edit it? i.e. LANG_DISC_FULL has "e200*,c200: "The disk is full. Press UP to continue."" but for c200 it should be "Press LEFT"
14:54:11amiconnAlso, do we want the gap in screenshots?
14:54:19BigBambior does that cause issues?
14:54:32kugelrasher: the automated invalidate doesn't seem to work in my case
14:55:06JdGordonamiconn: well.. If we can assume that its the only target that will ever have that gap then a few extra #ifdefs might not be so bad?
14:55:17JdGordonand screenshots should obviously mimick the real display as much as possible
14:55:19kugelBigBambi: meh, button translations are nasty imo
14:55:31kugelI'm encoutering a similar problem right now
14:56:08amiconnIt's not just a few extra ifdefs. The gap needs to be added in some way at output time, as the framebuffer doesn't contain it
14:56:20kugelBigBambi: but, LEFT should be PREV, the other strings to PREV as well
14:58:06amiconnEither the sdl surface needs to be blitted in two parts, or it needs to contain those extra lines, which moves the problem to the simulated lcd_update[_rect]()
14:58:46JdGordonI'd say keep it in the sdl surface part
14:59:48amiconnMy idea for unification is that screendumps and the sim will always use a 128 colour palette, regardless whether it's an 1-bit or 2-bit target, and whether the greylib is running or not
15:00:12amiconnIt will lead to larger screendump files, but less ifdefing in the code
15:00:27amiconnThe clip would then duplicate the palette
15:00:53JdGordonis there a problem with the current code? I mean it seems to work fine and there are things which could do with work which are probably more important?
15:01:11amiconnThere are several problems.
15:01:41JdGordonok :)
15:01:59amiconnScreendump colours don't match the display, lots of ifdefing in the sim, inverse display of the greylib (fixed for mr100, but still existing for the clip)
15:02:15JdGordonoh is this for the on target screenshots also?
15:02:58amiconnof course
15:03:07amiconnOtherwise it wouldn't make sense
15:03:25JdGordonoh ok... I thought only sims which is why I was asking if it was worth it
15:04:53amiconnTry a screendump on the clip... and btw, the screendump function is the same for target & sim, and is quite unrelated to the actual sim display
15:04:57 Join MethoS- [0] (
15:05:31amiconn(with the exception of the Player, which has no screendump on target)
15:05:49JdGordonI dont have a clip, but ok
15:05:59amiconnThen use a clip sim....
15:07:44*amiconn needs good macro names for the LCD colours
15:09:03amiconnThe sim uses UI_LCD_BGCOLOR[LIGHT] and UI_LCD_FGCOLOR[LIGHT], where the ..LIGHT versions mean with backlight, and FG/BG mean pixel set/not set
15:09:16JdGordonah I see what you're talking about.... pretty useless on the clip :)
15:09:42UnhelpfulPaulJam: i can band-aid the problem for now... do you have any feel for how fast scaling was on the h300 before that commit?
15:10:27amiconnBut I want to change this, so that one colour is always dark, the other is always bright, with a separate macro indicating an inverse display
15:11:22amiconnBut _DARK, _BRIGHT, _DARKLIGHT, and _BRIGHTLIGHT sound silly...
15:11:45PaulJamUnhelpful: sorry, i have no idea how fast scaling was.
15:12:59amiconnPerhaps I should put an optional BL_ part in the middle
15:13:19JdGordonyeah, something like that
15:13:25*JdGordon doesnt have any ideas on naming
15:14:00 Join Zoxc [0] (
15:14:42Unhelpfulno problem. i'll patch it up to use the same math as sh does, for now. that's more than accurate enough at reasonable sizes. i'll see about trying to fix it properly later, since there's no reason this *shouldn't* work on coldfire...
15:16:27 Quit EspeonEefi ("さよなら")
15:16:29PaulJamby the way, i just tried pictureflow, and the zoom functionality in the settings doesn't seem to produce wrong colours. does that use a different method?
15:18:58Unhelpfulthere's an option to turn resizing on or off. the "zoom" just changes the distance between slides and camera
15:21:02 Quit MethoS (Read error: 113 (No route to host))
15:23:26amiconnUnhelpful: Don't you need unsigned mode for scaling?
15:24:49Unhelpfulisn't 0 unsigned/integer?
15:24:59amiconn0 is signed integer mode
15:25:03Unhelpfulor do i have the "sense" of the sign bit backwards?
15:25:45amiconnAnd btw, you don't need to reset the mode - the next user will set it anyway
15:26:17BigBambikugel: Sorry, phone. Sure, left or previous - previous might be getting a little large for the c200 screen, but either way, can I just change the string in english.lang, or will that then feck up translations etc.?
15:26:46Unhelpfuleven in the same thread? i thought it was generally supposed to be left "ready" for the DSP's use?
15:27:26kugelBigBambi: IIUC, if you chane the source (and dest) string in english.lang, the other languages will be invalidated (since their source string doesn't match with the one in english.lang)
15:28:09kugelBigBambi: PREV is as large as LEFT, isn't it?
15:28:17amiconnWell you never know what other part of the code might have set it. All emac code blobs I know set the mode initially
15:28:25BigBambikugel: Yes, sorry - PREV is what should be used
15:28:59amiconnA context switch saves the emac status, so you can yield() without worrying about a different mode set when the yield returns
15:29:09BigBambikugel: And sorry for being confused, but do I want to change both source and dest then?
15:29:23kugelin english.lang you need to change both, yes
15:29:26Unhelpfulhrm... ok. i'll ifdef around the emac version for now, and see about fixing it later.
15:29:59pixelmakugel: that's no problem - the user won't see it as long as you don't change the ID. It will still use the wrong string but it'll show as a difference to the "master" english.lang when a translater wants to update and s/he'll notice that there's something to fix
15:30:15amiconnUnhelpful: I'd rather try unsigned mode first. It's a quick fix
15:30:33BigBambikugel, pixelma OK, cheers. I'll change both in english.lang
15:30:34 Join Aurix_Lexico [0] (
15:30:39amiconnUnfortunately I don't see the effect on my h300 - it might depend on the scaling factor
15:31:00pixelmaBigBambi: source and dest in english.lang - source will be evaluated by genlang for the translaters
15:31:42Unhelpfulamiconn: i can't do that right now, but if you'd like to take a look at it, i can hold on committing the band-aid patch for now.
15:31:46BigBambipixelma: yep, cool
15:31:46 Join arohtar [0] (
15:31:58kugelpixelma: thanks for clarification
15:32:00BigBambipixelma: Did you see my manual question above?
15:32:03 Quit arohtar (Read error: 104 (Connection reset by peer))
15:32:11BigBambii.e. is there an opt for flash?
15:32:12pixelmaBigBambi: no
15:32:55BigBambias at the moment the recording section has \opt{swcodec}{\note{When you start a recording, the hard disk will spin up. etc.
15:33:08amiconnTest building...
15:33:09BigBambiWhich isn't true for non-hd swcodecs
15:33:53BigBambiI couldn't spot a flash disk opt, but I figure I probably missed it :)
15:34:03gevaertsThere should be one
15:34:05kugelaren't \opt{} automatically generated from features.txt=
15:34:22pixelmaI believe there is - would have to look it up for myself. It's either a "manual" UseOption in the platform files or an automatic one from the parsed features.txt (to be found in a features.tex in a manual build folder)
15:34:23kugelif yes, adding this to feature.txt should sufficient
15:34:50BigBambiOK, I'll check features.tex
15:34:54gevaertsflash_storage ?
15:34:55pixelmafeatures.txt is also used for the language building, you can't just add something
15:35:09gevaertsOr actually disk_storage in your case
15:35:35BigBambiDo any of the archos disk targets record?
15:35:39pixelmathat's why not everything is automated (e.g. the HAVE_BACKLIGHT option)
15:35:42BigBambithe recorder perhaps?
15:35:47*BigBambi slaps self
15:35:48gevaerts(i.e. you want a hypothetical flash+disk target to show this text)
15:35:59pixelmaBigBambi: there's a reasonon they are called "Recorder"s ;)
15:36:08pixelmareason too
15:36:23BigBambipixelma: Do you know if there is a reason why theydon't get this note at the moment?
15:37:13amiconnBecause they apply a special trick to not have to spin up the disk as long as possible
15:37:16pixelmathis peakmeter thing? Maybe because the effect isn't as visible in the hwcodec recording system... just a guess though
15:38:02BigBambiSo I can just nest the two opts for swcodec and disk-storage?
15:38:29BigBambiIs there a special syntax for that, or is it as simple as one after the other?
15:38:34 Join T0paz [0] (
15:38:42 Nick JdGordon is now known as JdGordon|zzz (n=jonno@rockbox/developer/JdGordon)
15:39:12amiconnThe peakmeters also don't stop on hwcodec when the disk spins up. All that happens is that they do less readings per second, in order to leave enough cpu time for the disk access
15:39:17pixelmayes, nesting should work
15:39:34BigBambiOK, cheers
15:39:42 Nick JdGordon|zzz is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
15:44:31Unhelpfulamiconn: apparently scaling 150x150->75x75 will exposed the problem, with full brightness (255) in an color comonent:
15:44:52 Nick romain_ is now known as romain (
15:45:38Unhelpfulbut... neither factor should ever have the sign bit set, at that scale :/
15:46:06 Join MethoS-- [0] (
15:49:43 Nick Barahir_ is now known as Barahir (
15:52:13rasherBigBambi: you should be able to remove c200 from that string with tools/langtool, and then add a line for c200 manually
15:52:31 Join CaptainKewl [0] (
15:53:00 Join MethoS [0] (
15:53:07BigBambirasher: Is that better than just manually removing c200 and adding it on its own line?
15:53:28rasherBigBambi: Yes, because using langtool you can easily remove it in all langs that have it there
15:53:40BigBambirasher: Ah, OK
15:53:42***Saving seen data "./dancer.seen"
15:53:56n1skugel: shouldn't FS #9662 be closed now?
15:54:11rasherI'm not sure it matters when you're adding a new one though, actually
15:54:12BigBambiand then manually add it to english.lang only, and translators can add it to others whenever they next do
15:54:13amiconnUnhelpful: Looks like setting the emac to unsigned fixes the problem (tested on pixelma's M5)
15:54:15rashermight not be worth it
15:54:22rasherBigBambi: Indeed
15:54:28kugeln1s: I wanted to do that right now, I just waited for the build table ;)
15:55:13amiconnUnhelpful: Shall I commit, or post a ptach for further testing? It's essentially a 2-line patch (one in resize.h, and one in system-target.h for coldfire)
15:56:06 Nick JdGordon is now known as JdGordon|zzz (n=jonno@rockbox/developer/JdGordon)
15:59:32BigBambiWhy does e200 have * after it? i.e. e200*
15:59:33 Quit MethoS- (Read error: 110 (Connection timed out))
15:59:52rasherBigBambi: to catch e200v2 or e200r I presume
16:00:03BigBambiThat makes sense :)
16:00:12*BigBambi is being a bit slow today
16:02:31 Join {phoenix} [0] (
16:05:13 Quit faemir (Remote closed the connection)
16:05:25 Join faemir [0] (
16:09:43 Quit Thundercloud (Remote closed the connection)
16:11:25 Quit Seed ("cu, Andre")
16:11:43 Quit MethoS-- (Read error: 110 (Connection timed out))
16:11:49 Quit MethoS (Read error: 110 (Connection timed out))
16:15:10 Quit robin0800 ()
16:15:57BigBambirasher: Does "../../tools/ −−changetarget −−from e200*,c200 −−to e200* −−id LANG_DISK_FULL −−inplace *.lang" look correct to you? As it runs, and seems to go through all of the languages, but nothing actually changes in the lang files
16:16:56rasherBigBambi: Looks correct... Odd. A bug in isn't out of the question - do other languages have that string in the first place?
16:17:06BigBambisome do
16:17:16rasherI'll have a look
16:17:17 Quit SirFunk (Read error: 60 (Operation timed out))
16:17:20BigBambiAnd I get a list of languages being printed to the terminal
16:17:50rasherThat's just printed when opening the file, so it's no guarantee that anything's being changed
16:17:51 Join robin0800 [0] (
16:20:41 Join MethoS [0] (
16:21:18 Join SirFunk [0] (n=Sir@
16:21:51 Quit T0paz (Read error: 113 (No route to host))
16:26:42 Quit fyrestorm (Read error: 104 (Connection reset by peer))
16:32:07rasherBigBambi: just committed a fix
16:32:47 Join fyrestorm [0] (
16:34:23BigBambirasher: ta very much
16:34:25*rasher sees that kugel would've hit the same bug
16:34:40rasherBut evidently he went through more trouble to do the same thing
16:34:41kugelI did it all manually :S
16:34:54 Join SirFunk_ [0] (n=Sir@
16:35:25rasherkugel: don't worry, wouldn't have worked for you anyway (until now)
16:36:34kugelrasher: I wasn't sure if it worked for my problem anyway, I didn't change strings, only which strings apply for the e200*
16:36:51PaulJamUnhelpful: could it be that your commit in r19847 causes errors when compiling the uisim (h300) in cygwin? i get this: . (the warnings were there before, but the errors at the end are new.)
16:37:01Unhelpfulamiconn: if you were able to reproduce, and fix, the problem, i don't see any reason not to commit it ;)
16:37:16rasherkugel: langtool can change targets - look up for BigBambi's commandline
16:37:30Unhelpfulhrm... that's... odd.
16:38:23kugelrasher: ah cool, I'll give it a shot the next time :)
16:38:46BigBambirasher: works nicely now, thanks
16:39:11rasherkugel: At least you actually *did* it the manual way. Most devs are too lazy for that (understandable, really), and just let the translators bother with it
16:39:26rasherWhich is why I created the script
16:39:38kugelwell, I actually thought about doing that too :p
16:40:00kugelbut this was a no-brainer, and it's sunday and I have time, so I did it :P
16:40:04rasherCan't say I blame you
16:41:19 Quit SirFunk (Read error: 145 (Connection timed out))
16:41:31kugelalso, I noticed that it would show *wrong* translations, which would directly be *wrong instructions* on the screen in this case, instead of showing no translation
16:44:25UnhelpfulPaulJam: maybe do a make reconf and make dep? i was able to build h300 sim...
16:44:28 Quit SirFunk_ (Remote closed the connection)
16:45:29PaulJamUnhelpful: i did 'make veryclean' and '../tools/configure' before.
16:46:23PaulJamdo you use cygwin?
16:49:10Unhelpfulno, i don't :/
16:50:59Unhelpfuli'm not really sure why that would even be happening, gcc-support.c includes plugin.h, which declares rb as extern... not to mention, there are piles of other pluginlib files that reference rb in the same fashion without problems :/
16:52:08PaulJamamiconn: if you have time, could you check if the h300 uisim builds for you (in cygwin)? maybe it's just a local problem on my end.
16:53:15 Quit nibbler (Read error: 110 (Connection timed out))
16:56:20*BigBambi points people at FS #9825 and FS #9826
16:57:15 Join MethoS- [0] (
16:58:31Unhelpfulpixelma: if we had one button context that offers vertical or horizontal scrolling, a select button, and a cancel or exit button, yes, i think that pictureflow could definitely get by without combining, then.
16:58:38 Quit nplus (Remote closed the connection)
17:00:20pixelmawithout combining is a must IMO, otherwise you lose control about what will happen on targets which have different actions in the combined actions on the same button (action)
17:00:35pixelmacombined contexts, I mean
17:03:10Unhelpfulit's a little tricky for things like the scrollwheel targets, which tend to use the wheel for scrolling in either direction... and confusingly, i think, for movement on one axis, while using buttons for the other
17:03:35 Join PaulJam_ [0] (i=PaulJam_@
17:06:19 Quit PaulJam (Read error: 60 (Operation timed out))
17:09:24 Join t0mas [0] (n=tomas@rockbox/developer/t0mas)
17:11:59 Join at0m [0] (n=a548c80b@gateway/web/cgi-irc/
17:13:21 Quit MethoS (Read error: 110 (Connection timed out))
17:14:41amiconnUnhelpful, PaulJam_: I get the same error. Looks like the problem is that battery_bench.c is #ifndef SIMULATOR
17:15:16amiconnBut it isn't excluded from SOURCES, only within the plugin itself, and the #ifndef SIMULATOR block includes PLUGIN_HEADER
17:15:16 Quit at0m (Client Quit)
17:15:52kugelcan anyone tell me why CONTEXT_LIST and CONTEXT_TREE are seperate?
17:15:59Unhelpfulthat's bizarre
17:16:32amiconnkugel: The tree is not simply a list, there are a few extra functions
17:16:48amiconn(like going directly to wps, stopping playback)
17:17:33kugelamiconn: which? and why should the list not have these functions?
17:18:14kugelwell, I can stop playback and go to wps from both list and tree on my e200
17:18:27kugelexcept for the id3 viewer, which drives me crazy
17:18:55amiconnYou just need to leave the tag viewer...
17:19:18kugelbut I cannot do that with the button that I use everywhere in rockbox to go to the wps
17:19:29kugelthat one does nothing in the id3 viewer
17:20:16Unhelpfulamiconn: so, battery_bench should probably be ifndef SIMULATOR in SOURCES, shouldn't it?
17:20:17amiconnEverywhere? Certainly not....
17:21:58 Join nibbler [0] (
17:22:40kugelfrom filebrowser, database browser, main menu, playlist viewer, pitchscreen, settings
17:22:47kugelok, not everywhere, but almost
17:24:40amiconnUnhelpful: There are more than just that one...
17:24:57Unhelpfulamiconn: i'll start looking, then :)
17:25:15amiconnFound firmware_flash and rockbox_flash so far
17:25:37amiconnI wonder why cygwin ld wants to link memcmp against an essentially empty object though....
17:25:54Unhelpfulwait, are you fixing these already?
17:26:07kugeldoes anything speak against adding ACTION_TREE_WPS to CONTEXT_LIST?
17:26:32kugelso that the id3 viewer is leaveable with the usual go-to-wps button?
17:27:11*Unhelpful notices that much of firmware_flash.c is also ifdef PLATFORM_ID
17:28:46 Join phinze [0] (
17:29:18amiconnPlain lists shouldn't have that
17:29:42Unhelpfulvideo.c also
17:31:48 Join T0paz [0] (
17:32:08kugelamiconn: why? what's the definition of plain lists?
17:32:25kugelfrom the user pov, everything is a plain list anyway
17:33:09amiconnIf you add that to plain lists, everything using plain lists would need to deal with the extra event. Plugins, anything
17:34:28kugelthe other posibility is to use CONTEXT_TREE for the id3 viewer
17:34:58*amiconn doesn't see everything as just a list, as the behaviour is quite different in menus, browser, settings
17:35:01 Quit robin0800 (Remote closed the connection)
17:35:03kugeleven though, I tested that, and stop didn't do stop in the id3 viewer
17:35:12amiconnAnd it has to be like this, as their purpose is different
17:35:56 Quit CaptainKewl (Read error: 60 (Operation timed out))
17:35:59kugelbut at least play exited it
17:38:19kugelamiconn: anything against that?
17:38:57kugelwell, it doesn't sound entirely correct, but it does what I want it to do
17:39:04amiconnUnhelpful: Hmm, iiuc the clever multiplication-saving trick doesn't work for RGB888? Or at least not using one 32-bit int
17:39:28amiconnkugel: You can't just change the context, you also need to actually handle the additional events
17:40:02kugelthere's no additional events, except TREE_WPS an d TREE_STOP
17:40:11Unhelpfulamiconn: multiplication-saving? you mean the one to do undither output with values scaled directly to 0..31/0..63?
17:41:00Unhelpfulalso, does this look right? i cleaned up some other unneeded ifdefs in source for things that are already handled in SOURCES, as well.">
17:41:01 Quit {phoenix} (Read error: 104 (Connection reset by peer))
17:42:19Unhelpfulif you mean the trick that pictureflow uses, then it should work *for pictureflow*, where we can (pretty much) assume that max(in_w,out_w) * max(in_h,out_h) > 63
17:42:20amiconnUnhelpful: I mean the one that scales an RGB value without (full) unpacking
17:42:52Unhelpfulamiconn: oh! that... yes, and know.
17:42:54amiconnI am thinking about using such a trick for calculating the bmp palette for screenshots
17:43:30*gevaerts tries to make sense of testing the latest patch in FS #8663
17:43:56amiconnIiuc it should be possible if I just unpack into G and 0R0B
17:44:13Unhelpfulyou can do that, and save one multiplication, yes
17:45:07Unhelpfulbyte-order etc might make it hard to determine which bytes you want, if you're talking about RGB888 and not RGBX8888
17:45:57amiconnWhat BMP uses in its palette
17:46:54 Nick fxb__ is now known as fxb (
17:48:18Unhelpfulwell, assuming we don't really care what ends up in the X byte, you can use two multiplies, after splitting a 32-bit color value c into c1=(c >> 8) & 0xff00ff and c2=c & 0xff00ff
17:48:40amiconnX must be 0
17:48:40 Join itcheg [0] (i=62db4767@gateway/web/ajax/
17:49:01Unhelpfulwell, if it's already 0, you should be ok
17:49:52amiconnAnd I can store the input values already split, so I just have to multiply, mask and combine
17:49:55Unhelpfulit's best if your scale factor is in range 0..(n^2), rather than 0..((n^2)-1), because then you can shift to divide. doing anything else while maintaining precision will get complicated.
17:49:56 Quit nibbler (Read error: 110 (Connection timed out))
17:50:18 Join slyyf [0] (n=slyf@
17:51:04UnhelpfulSWAR is not real SIMD, so even things that temporarily need >unit precision get ugly
17:51:32amiconnInput values are decided compile time. I think it's better to calculate the palette at runtime instead of wasting 512 bytes of constant data
17:52:04slyyfDoes anybody have a copy of the tcc78xx datasheet lying around your harddrive from back when it was available?
17:52:10amiconnOf ocurse only if this calculation (+ required input values) is smaller than 512 bytes...
17:53:30amiconnOn the clip it would ben 1024 bytes
17:53:46***Saving seen data "./dancer.seen"
17:56:07Unhelpfulamiconn: this is for sim, right?
18:00:49Unhelpfulamiconn, PaulJam: can one of you try this on cygwin? it builds for me on linux, but i don't have a cygwin environment:">
18:01:44 Quit Xerion (" ")
18:07:10 Quit T0paz (Read error: 113 (No route to host))
18:09:55 Quit axionix (Remote closed the connection)
18:11:09Unhelpfulit's a shame MMX lacks an 8x8->16 version of pmulhw/pmullw
18:11:52PaulJam_gah, pastebin sripped the @@ at start of the lines
18:12:14UnhelpfulPaulJam: copy from the edit box... it's the only way :/
18:12:22Zoxcr = (r * rvalue) >> 5; g = (g * gvalue) >> 6; b = (b * bvalue) >> 5;
18:12:38Zoxccan you combine the multiplies with a SWAR here? :)
18:12:38PaulJam_i did. there they are missing too.
18:12:40 Quit itcheg (" ajax IRC Client")
18:13:15UnhelpfulPaulJam: holy cow... that's insane. i'd think it would preserve them *there*, so that you still have the same highlights if you edit it... :/
18:13:18 Join {phoenix} [0] (
18:13:45UnhelpfulZoxc: if rvalue==bvalue==gvalue
18:14:21UnhelpfulSWAR is not true SIMD. if you have true SIMD with separate fields, things like that are possible.
18:15:26Zoxcbetter get some AVR32 DAP :D
18:17:54Unhelpfultrue SIMD would be troublesome for this case, also, though... in my experience, available field sizes are 8, 16, 32 or 64 bits
18:18:07 Join toffe82 [0] (
18:18:45ZoxcYou would use RGB888 with an OLED screen in the ideal AVR32 world :)
18:20:46Unhelpfuland mixing SIMD with non-SIMD operations on the same data *murders* your efficiency
18:22:17 Join Darksair [0] (n=user@
18:25:40slyyfAny idea how I could tell if an LCD is TFT or not?
18:30:50 Join BigBambi_ [0] (
18:30:52 Join Aristad [0] (
18:33:35 Quit BigBambi ("Please insert girder")
18:45:45soapdoes it look better or worse in the sun?
18:46:17 Quit faemir ("Lost terminal")
18:49:48 Join gartral1 [0] (
18:49:59 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
18:52:55 Join TheSphinX^ [0] (
18:53:56PaulJam_Unhelpful: the uisim compiles with the patch.
18:54:15 Quit TheSphinX^ (Remote closed the connection)
18:54:34 Join TheSphinX^ [0] (
18:56:12gartral1ok, this problem with the 000090D0 (0) abort is not just my player, my moms player has starting to have the same issue, with both the "modified" and "stock" builds, please, will someone help me with this?
18:57:03 Join MethoS-- [0] (
18:59:36mcuelenaeregartral1: you'll really need to tell which player this is..
19:01:57gartral1sorry, there e200s,
19:07:25 Quit phinze (Read error: 110 (Connection timed out))
19:08:30gartral1the symptoms are: if you play anything, the buffers all fill under 30 seconds, then it aborts at address 000090D0 (0), this is with OGG, MP3, WAV, WavePack, and ACC, i haven't tested anything else, though its most likely to reproduce it, and it seems to be related to who last edited the trunk., though this is only speculation
19:09:48amiconnUnhelpful: Both target and sim
19:10:34gartral1no, cant reproduce in sim
19:10:59gartral1but the fact that both players are doing it, at the same address is what gets me
19:11:03 Quit MethoS- (Read error: 110 (Connection timed out))
19:11:20mcuelenaeregartral1: have you filed a bug report?
19:12:00*gevaerts would try to reproduce this, but his e200 is currently busy trying to reproduce another issue
19:13:08amiconnPaulJam_: Does r19853 fix the wrong colours on scaled bitmaps for you?
19:13:33PaulJam_yes, thanks for fixing.
19:13:42gartral1not yet, ive asked here before, but i can't get enough info to file a bug report, ide really like to know what goes into that address, but my system wont produce a readable map file, and less cuts it short over tty
19:14:04gevaerts000090D0 isn't in the map file
19:14:37gartral1ok, then whats going on here?
19:15:19mcuelenaereinvalid pointer?
19:15:50 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
19:15:58 Part LinusN
19:16:03gevaertshm, I seem to misremember things...
19:17:23gevaerts000090D0 is somewhere in buffering.c if I read things right
19:18:31UnhelpfulPaulJam_: thanks, i'll commit it, then.
19:19:19krazykitgartral1, this happens with latest svn? i can try on my e200
19:19:48gartral1hmm, ive been having this issue since just before the new year... my moms player hasnt been updated since about november of last year, and i just brough both up too 19852
19:20:04krazykitso "yes"?
19:20:15gevaertsgartral1: is the address always the same?
19:20:35gartral1yes, across all codecs, both DAPs
19:21:04gevaertsAlso across revisions and builds?
19:21:09gartral1and sometimes it does, sometimes it dont
19:21:21gartral1only specific builds do it
19:22:09gartral1last one was 19844, before that, 19832
19:22:10gevaertsWhat's the full message? Data abort?
19:22:44*gevaerts would like to have a rockbox.elf and matching exact message
19:22:46gartral1Data Abort at 000090D0 (0)
19:23:02gartral1how do i produce that?
19:23:21gartral1ohh, wait, its right in the build dir, one moment
19:23:34 Quit mcuelenaere ()
19:25:01 Quit kugel (Read error: 60 (Operation timed out))
19:25:30gartral1that file is huge >.> it's on JD's system, unless i can have him FTP it, were gonna have to wait
19:25:59gevaertsgartral1: how big is it?
19:26:17*gartral1 needs too remember scp is in BYTES not KB!
19:26:35gevaertsThere's also a next to it. Maybe get that as well
19:27:01*krazykit cannot reproduce that behavior in r19853, but will check 19844
19:27:09gartral1im looking at "1,045,063" but it dosent designate byte/kb
19:27:15 Join kugel [0] (
19:27:31gartral1well, 19852 does it, as well
19:28:19krazykitwell, i'm playing both mp3 and vorbis just fine
19:28:33gevaertsgartral1: do you use the eq?
19:28:47gartral1yes i do, and dithering
19:29:20 Quit kugel (Read error: 60 (Operation timed out))
19:29:23*gevaerts waits for the files before he asks further questions
19:29:49gevaertsgartral1: maybe open a bug report and attach those files, together with the exact address you get with that same build
19:30:34gartral1wait, i was only aware i needed rockbox.elf... what other file do you need?
19:30:53gevaertsyes. If possible the that's in the same directory
19:31:20gartral1its alot smaller than the ELF, np
19:31:30 Join kugel [0] (n=kugel@
19:32:09soapgevaerts, speaking of SD - what is your take on the latest comments to FS #8663?
19:32:26gevaertssoap: testing that as we speak
19:33:15soapI mean in theory. What he did is out of my depth. I was just curious on why that would have an effect.
19:33:41 Join Barahir_ [0] (
19:34:25 Join nibbler [0] (
19:34:42gevaertssoap: this UNKNOWN register is unknown, so it could well have undiscovered results :)
19:38:53gartral1ok, where waiting on my slow upstream to give the files too FS
19:39:41 Quit Darksair ("Use the Force, Luke!")
19:40:40gartral1please excuse my poor garmmar
19:40:49Unhelpfulthat's a curious yellow on clip sim... looks like fnarfbargle might be responsible?
19:44:36 Quit Barahir (Read error: 110 (Connection timed out))
19:47:11gartral1gevaerts: please ping me if/when you find anything
19:47:46gevaertsgartral1: I think it's definitely in buffering.c, but I'll have to leave it to people who are actually at home in there
19:49:26gartral1well, whats a quick fix? last time, i cleaned the database and it fixed the issue till i updated today, i tryed the database clear again, and nada
19:49:35 Nick Barahir_ is now known as Barahir (
19:50:24kugelcan someone test a quickscreen patch?
19:53:45 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
19:53:50***Saving seen data "./dancer.seen"
19:54:23gartral1kugel: where is, whats it do?
19:55:13 Join MethoS- [0] (
19:58:26 Join faemir [0] (
20:00:34 Quit nibbler ("Ex-Chat")
20:01:20 Join Xerion [0] (
20:03:07 Quit rocko ("Leaving")
20:03:17 Quit BigBambi_ (Remote closed the connection)
20:04:57 Join tex [0] (n=4cfae06c@gateway/web/cgi-irc/
20:05:25gartral1patch says that the files to be patched dont exist...
20:07:15kugelgartral1: I don't think so
20:07:38n1sgartral1: you're probably doing it wrong then
20:07:54texhello room. I'm having trouble with version 3.1 on an 80GB 5.5 and couldn't find any bug reports on the web
20:08:08 Join akur [0] (
20:08:12 Part akur
20:09:30texI upgraded from an older version, put a new ipodwidgets theme on it
20:09:58kugeltex: just saying, it would be helpful if you told what your problem is
20:10:06texand now it starts to play and then pauses for a second or two after playing only a couple seconds
20:10:18texsorry, was looking up the theme name
20:11:02 Quit wpyh (Read error: 110 (Connection timed out))
20:12:44slyyft/win 4
20:13:24texI saw one reference to "resume playback on startup" and ChanServ playback resume on hard drive scan bug 9719
20:13:30kugelslyyf: hey! how's your progress?
20:13:52texdo you think I should reset my cfg to defaults and see if that helps?
20:13:58 Join BigBambi [0] (
20:13:58slyyfkugel: its goin...soso
20:14:10slyyfkugel: I have it displaying stuff on the screen better
20:14:26slyyfstuff is shifted, so one of my values is wrong
20:14:38kugelhow much?
20:14:48slyyfmaybe less
20:15:12slyyfI think I am just not sending a correct value for the bpp or something
20:15:37slyyfBitwise whats &~?
20:15:45 Quit TheSphinX^ ("XChat@Linux")
20:15:50gevaertstex: that's always worth trying
20:16:04texbut, then if I pause playback for a few seconds and resume, it won't continue to skip every two seconds
20:16:12 Quit MethoS-- (Read error: 113 (No route to host))
20:16:13 Join dockimble [0] (n=dockimbl@
20:16:15texk, I'll give it a whirl and see what happens
20:16:28slyyfkugel: also apparently the power management is different too, because it always comes up that the battery is dead, no matter what Mv value I set for Critical
20:16:30mcuelenaereslyyf: & ~ is bit clear
20:16:39 Join __lifeless [0] (n=lifeless@
20:16:51slyyfmcuelenaere: thanks
20:17:03 Quit dockimble (Client Quit)
20:18:46 Join dockimble [0] (n=dockimbl@
20:19:03kugelslyyf: I looked very shortly into the driver
20:19:13 Quit mcuelenaere ()
20:19:13 Quit miepchen^schlaf ()
20:19:22kugelthere's 1 which I'm curious about
20:19:26kugel1 number
20:19:49kugel0x103 (which is 259)
20:20:10 Quit HBK (Read error: 54 (Connection reset by peer))
20:20:23slyyfWhat line specificically did you see that on?
20:20:51 Join HBK [0] (
20:20:54kugelwhat I know from the fuze lcd driver, there's a xoffset, which is by default 20, and in one command you write LCD_WIDTH + xoffset
20:21:20slyyfOhhh...Maybe..where is that though?
20:21:23kugelline 190
20:21:54slyyfof lcd-cowond2.d?
20:21:54 Quit tex ("CGI:IRC (EOF)")
20:22:06kugelyou may try changing that line to LCD_HEIGHT-1 + 30 (or LCD_WIDTH-1 - 50)
20:22:14kugelslyyf: yes
20:22:50kugelyou can also take a look into the lcd driver of the fuze, there's some general similarities it seems
20:22:55slyyflemme grab the svn copy...
20:25:19slyyfwhat makes you think 259 has any baring on it?
20:28:27slyyfdarn, nope, didnt change anything
20:28:39kugelslyyf: well, you were talking of things being shifted. I know that behaivor from the fuze, when you didn't apply the xoffset
20:29:06slyyfah but, how do you know which of those lines is the xoffset?
20:31:27 Quit _lifeless (Read error: 110 (Connection timed out))
20:32:06 Quit BigBambi (Read error: 113 (No route to host))
20:33:50slyyfohh, I see, lemme take a closer look at what the fuse is doing there
20:37:23 Join rocko [0] (
20:39:11kugelslyyf: btw, maybe this will be useful: The e200v2 and fuze driver are quite similar, but the e200v2 one doesn't always put the lcd dimensions with LCD_WIDHT/_HEIGHT, but with immediate numbers (like the d2 one)
20:41:52kugelalso, the e200v2 doesn't have this xoffset, while the fuze one does (and the e200v2 display is basically just rotated vs the fuze's display
20:42:35gartral1whats "caption backlight" do? its not documented...
20:43:21 Quit rocko ("Leaving")
20:43:26n1slooks documented to me
20:43:28 Quit MethoS- (Remote closed the connection)
20:49:39 Join miepchen^schlaf [0] (
20:56:15 Join BigBambi [0] (
20:59:56 Quit dockimble ("Ex-Chat")
21:05:37 Join T0paz [0] (
21:10:38 Quit Aurix_Lexico (Remote closed the connection)
21:10:59jhMikeSamiconn: Yeah, I'd need the __div0 stub but not an entry in the API structures. In system-arm.c an swi_handler which some formal method of using it should be defined if it proves useful for anything else later.
21:11:50 Join Aurix_Lexico [0] (
21:19:36 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
21:21:22kugellol, can we disallow sid user to file bug reports? :)
21:21:31kugel(debian sid, that is)
21:23:03rasherWhy would that be?
21:23:49gevaertskugel: why shouldn't I be able to file bug reports? :)
21:24:01*rasher would dislike that as well
21:24:36kugel1) I was kidding 2) see latest comments at
21:25:23rasherAt that rate, we'd have banned Windows users years ago
21:26:32*bluebrother doesn't understand what's so special about the comments in that task
21:27:15kugelbluebrother: the one has reported USB problems, however the last comment is "The USB failure I'm experiencing would be connecting to a box running Debian Sid with a 2.6.26 kernel."
21:27:40gevaertskugel: so? You should be glad to get details
21:27:51kugelI am :)
21:28:15Unhelpfulsurely you'd ban gentoo users first?
21:28:39kugelseems like some people took this joke too serious :/
21:29:31slyyfOk, so the normal P2 firmware displays everything upright...
21:29:38slyyflike, tall rather then wide
21:29:41rphillipsanybody know if the apple line-out dock works w/ rockbox?
21:29:42rasherkugel: I think you misunderstood what he's saying. He isn't saying (afaict) "This is just Debian's fault" - he's saying "These are the details of how the error happens"
21:29:42Lloreanslyyf: Please use whole, real words. I have no idea what "qn" means.
21:29:45bluebrotherkugel: well, at least he gives details which is a *great* thing. The wording might be a bit weird, but well ... plus, stating the OS he's trying to connect to might give a hint about an OS issue (if there was some known. As that is not the case this is still a Rockbox issue unfortunately)
21:30:24Lloreanrphillips: There's a page on the wiki showing known iPod accessories and how they work or don't. I don't know the exact page though, you'll have to search.
21:30:33 Join Tim3128 [0] (
21:30:54kugelrasher: I read that as "the usb problems could be connected (related) to my debian sid system"
21:31:00slyyfOk, guys I have a question, The normal P2 firmware displays everything vertical rather then horizontal (Unless your playing movies)...however the Rockbox is displaying Hozizontally..could that be the source of my problem, possibly? (seems to be shifted off screen almost)
21:31:32slyyfif so, how can I make rockbox display vertically
21:32:16 Join tvelocity[a] [0] (n=tony@
21:32:50 Quit tvelocity[a] (Client Quit)
21:33:04kugelbluebrother: sure, I agree
21:34:20kugelslyyf: well, rockbox should use the same display as the OF, right?
21:35:10slyyfkugel: No, apparently, Samsung must have rotated all of there images 90deg..but that makes no sence because they must render text at some point...
21:35:13kugelslyyf: what did you define exactly? LCD_HEIGHT should be the bigger resolution I suppose
21:35:31slyyfI do that and it wont compile
21:35:32Lloreankugel: Rockbox, generally, will expect the LCD to not have been rotated.
21:35:46LloreanSometimes the LCD is physically rotated. Basically, the OF displays at 90 degrees from the LCD's actual orientation.
21:35:57slyyfya, exactly
21:36:17slyyfI am trying to identify the source of this darned issue and I cant come up with anything
21:36:18kugel"it wont compile"? how that?
21:36:32slyyfSomething about rockbox logo not being defined
21:36:34 Quit faemir ("Lost terminal")
21:37:03slyyfIt must try to get a logo that fits the resolution
21:37:13kugelyes, that's right
21:37:23 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
21:37:44 Quit tvelocity (Read error: 60 (Operation timed out))
21:37:45 Join faemir [0] (
21:37:49mcuelenaereT0paz: have you read the logs?
21:38:10Unhelpfulkugel: out of curiosity, which targets had you tested pictureflow on? all of mine are portrait...
21:38:16domonokyslyyf: for the logo see: apps\bitmaps\native\SOURCES
21:38:24 Quit Xerion (Read error: 110 (Connection timed out))
21:38:50 Join Seed [0] (
21:38:52kugelUnhelpful: I haven't tested it recently
21:39:10kugelbut sometime ago, it ran on my fuze and e200 (fuze: landscape, e200: portrait)
21:39:20slyyfeither way, setting the height larger then the widthmakes it fail
21:39:24T0pazmcuelenaere, aha, hi
21:39:29T0pazi got the link to the BSP
21:39:37mcuelenaereok good
21:39:42T0pazi've been having fun messing about
21:39:43domonokyand to get the rotated display correct, you probably have to change the lcd-driver and swap x and y. ( ithe lcd hardware doesnt have a rotate 90° option)
21:39:45 Join rocko [0] (
21:39:51mcuelenaerewell, the other thing I said was that I won't be available the next days due to exams
21:39:56kugelslyyf: well, for now, you can probably just take a existing logo, renamed to fit your resolution
21:40:06T0pazah, hope they go well
21:40:52T0pazi've been having fun messing about
21:41:02mcuelenaerethe BSP you mean?
21:41:08T0pazfound the code which updates the volume control register (DACVOLUME)
21:41:18mcuelenaereyeah, I read something like that
21:41:20T0pazin the firmware, using the BSP header files as a guide
21:41:23mcuelenaereso you got code working?
21:41:37T0pazwell, i modified the firmware to stop the fade-out in one channel ;)
21:41:39 Join pinkey [0] (
21:41:44T0pazslightly pointless, but demonstrates it working
21:42:00T0pazi've been function-hunting in the assembly
21:42:08T0pazfound a few of the basics, memcpy, memset, abs
21:42:18mcuelenaereah, about that
21:42:25mcuelenaerethe kernel Creative uses is Nucleus
21:42:30T0pazah yes, i read that
21:42:31mcuelenaerethere's some source code on the internet, google
21:42:36T0pazi was pondering blagging the evaluation version
21:42:47mcuelenaereand I've documented the init function a bit
21:42:47T0pazheader files containing structures would be extremely useful
21:42:56mcuelenaerewell documented, I've talked about it with somebody
21:43:09mcuelenaerebut unfortunately it's in dutch, so I'll have to redo that :)
21:43:13T0pazheh ;)
21:43:13Unhelpfulkugel: ok... some of the changes we've talked about with regard to picking a size for images have worked out not-great for c200, apparently... it's a bit *wider* than some other landscape targets, though...
21:43:17mcuelenaerethe problem is that's it been a while..
21:43:39T0paza few other things confuse me
21:43:58T0pazthe code makes occasional BLs to addresses in 0x3fffxxxx
21:44:06T0pazwhich is below where the main binary sits
21:44:46mcuelenaerehmm perhaps there's a memory relocation switch?
21:44:57mcuelenaerehave you seen anything like that in the BSP headers,
21:45:16T0pazwell, a few things i did discover
21:45:28T0pazyou were looking for IO accesses in 0xF0060000 (the LCD registers)
21:45:43T0pazthere are no accesses anywhere to anything in 0xFxxxxxxx
21:45:48T0pazbut plenty to 0x8xxxxxxx
21:46:07pinkeyHi, in the FAQ list of available supported hardware, it appears "none" are currently in production. Is the FAQ out of date, or is true there no real option for me becoming a rockbox user if I'm just now entering the market?
21:46:14T0pazand according to the BSP headers, 0xf... is for virtual mode, 0x8... for real mode
21:46:44T0pazso i found various things like i2c init functions
21:46:57PaulJam_pinkey: there is still ebay...
21:47:01dionoeapinkey: there are some in progress ports with hardware currently in production
21:47:09mcuelenaerevirtual/real mode? you mean physical and virtual adresses?
21:47:15T0pazah yes, that
21:47:17 Join CaptainKewl [0] (
21:47:18dionoeathe sansa fuze/clip/e200 series
21:47:27domonokypinkey: one option is to buy used/refreshed players
21:47:29T0pazthis is my first foray into ARM assembly
21:47:39*mcuelenaere can't seem to find that chat he was talking about
21:47:49kugelUnhelpful: how do you mean that?
21:48:04mcuelenaereah in that case have you seen the ARM register instruction sheet?
21:48:08mcuelenaereit's quite handy at times
21:48:24T0pazi've got the intro lecture from the 'ARM university program'
21:48:24mcuelenaere(and if you're using IDA, it has a function to explain instructions)
21:48:28T0pazwhich is actually really good
21:48:39T0paz(summarises the entirety of ARM assembly pretty well in 46 slides)
21:49:22T0pazi'll take a look
21:49:30T0pazregarding memory relocation, do you mean something using the MMU?
21:49:38mcuelenaereyes, probably
21:49:52mcuelenaereah no, that's what I meant with the virtual/physical addresses
21:50:24T0pazi guess there's a possibility that it turns on the MMU and maps stuff info 0x3f...
21:50:25mcuelenaereabout memory relocation, I know at some SoC's there's a switch where some part of the physical addresses can get different virtual addresses
21:50:34mcuelenaerewithout using the MMU AFAIK
21:50:39Unhelpfulkugel: pixelma had some issues with how it looks on c200 these days... and i'm inclined to agree
21:51:07kugelUnhelpful: well, remeber that I just proposed LCD_HEIGHT/2
21:51:09T0pazwhat purpose would that serve? making the addresses nearer?
21:51:12kugelyou changed it further ;)
21:51:39Unhelpfulkugel: not really, that's what we were still using. and 40x40 is *darn* tiny on c200
21:51:54kugelI think LCD_HEIGHT/2 looks just about perfect (at least on my targets)
21:52:21T0pazmcuelenaere, btw, do you know where the entry point is? it would seem not to be 0x40000000
21:52:35kugelUnhelpful: haven your made it bigger? like LCD_HEIGHT-REFLECTION_HEIGHT
21:52:40mcuelenaerenope, that's SoC/architecture specific
21:52:45mcuelenaerealthough, on ARM it should be on 0x0
21:52:52mcuelenaerethat's the reset vector
21:53:04Unhelpfulkugel: that's didn't change anything for square covers, only tall ones
21:53:07T0pazah ok, so the bootstrap code at the start must be responsible
21:53:30mcuelenaerewell, the code at 0x0 is almost always a branch to some other place
21:53:36mcuelenaereas there's not much space in that spot
21:53:38kugelmcuelenaere: 0x30000000 works good on ams sansas, and watchdog resets properly
21:53:51***Saving seen data "./dancer.seen"
21:53:56mcuelenaerekugel: you mean 0x30000000 is the reset vector on AMS?
21:54:09T0pazit loads the pc from further down and hops off to 0xdbd4
21:54:32 Quit yhuang ("Leaving")
21:54:33T0pazis this all separate from the rescue mode?
21:54:43kugelmcuelenaere: not sure, but the datasheet specifies to map SDRAM to 0x30000000
21:54:46 Join mcuelenaere_ [0] (n=mcuelena@rockbox/developer/mcuelenaere)
21:54:53kugelmcuelenaere: not sure, but the datasheet specifies to map SDRAM to 0x30000000
21:55:06kugeloh, I thought you disced
21:55:10 Join Xerion [0] (
21:55:14mcuelenaere_T0paz: yes, this is all separate from the rescue mode
21:55:28T0pazcool, so there's another bootloader above this?
21:55:48mcuelenaere_kugel: but what's the initial memory address the CPU looks at when it starts?
21:56:06kugelthe rom bootloader loads 0x0
21:56:10mcuelenaere_T0paz: yes, the booting chain (on the ZVM) is like this: FBOOT -> FRESCUE/main FW
21:56:21mcuelenaere_with FBOOT in NOR and FRESCUE/main FW on HDD
21:56:27mcuelenaere_and I suspect this is similar on the Zen
21:56:29kugelthat's where our bootloader is (that's iram)
21:56:33mcuelenaere_I know it has FBOOT somewhere
21:56:43 Join MunkieMan [0] (n=29b132e2@gateway/web/cgi-irc/
21:57:12mcuelenaere_but it isn't in the nk.bin package
21:57:40slyyfWhat the.....How is this even working..this code is connecting to the LTV250QV..which is NOT the lcd that the P2 has
21:57:53slyyfdo they often connect the same way?
21:57:54 Quit miepchen^schlaf (Connection timed out)
21:58:28mcuelenaere_slyyf: what do you mean with 'the code connecting to the LCD'?
21:59:12slyyfmcuelenaere_: I meen the code that sends the raw commands to the lcd "lcd_write_reg"
21:59:44slyyfits not for the lcd I am working with, and yet, somehow it worksish
21:59:58mcuelenaere_well, it could have the same LCD IC
22:00:07mcuelenaere_or it could have a similar instruction set
22:00:42MunkieMani have a suggestion. i own the ipod video 30gig and i obviously have rockbox on it. i just find it slightly irritating when im listening to music and i want to change the song. when you press the menu button on the ipod, it takes you back to the main menu. i think it would be better if it took you to back to the last screen you were at before you started listening to your music. ie. your playlist.
22:00:49slyyfthe datasheet doesnt have any content about that
22:01:16kugelMunkieMan: what does select do? it's doing what you want on my sansa e200
22:01:26*mcuelenaere_ wonders which datasheet is 'the'
22:01:27bluebrotherMunkieMan: that's not the way Rockbox works. Use Select to return to the browser.
22:01:55bluebrotherthe Menu button does what its writing says: go to the menu.
22:03:00MunkieManoh wow i feel stupid now. thanks for the help
22:03:17slyyfmcuelenaere_: the datasheet for this lcd
22:03:55mcuelenaere_slyyf: then if you have the datasheet, why send it the LTV250QV commands?
22:03:58kugelslyyf: you need to seperate the lcd from the "logic that connects the lcd to the soc" though
22:04:25rphillipsLlorean: thanks for the tip...
22:04:33slyyfOk so, at the moment all that is there is the code from the D2
22:04:36mcuelenaere_yes, in generally it's like this: SoC lcd driver -> LCD IC -> actual LCD
22:05:02slyyfRight now it is connecting to three pins
22:05:18mcuelenaere_T0paz: it seems as that Nucleus source code is offline, perhaps I can upload my downloaded version
22:05:29T0pazah, that would be very useful
22:06:04slyyfmcuelenaere_: Right now it is connecting to three pins called "/* GPIO A pins for LCD panel SDI interface */"
22:06:42mcuelenaere_T0paz: the code I have is a bit less than the one I was referring to, I think I don't have that one anymore
22:06:48mcuelenaere_but it's still the Nucleus kerne
22:08:11 Quit pinkey ("Leaving")
22:08:59slyyfmcuelenaere_: whats "GPIO A pins LCD panel SDI interface"
22:09:17mcuelenaere_GPIO means General Purpose Input/Output
22:09:26slyyfand SDI?
22:09:44mcuelenaere_I don't know, does the P2 also run the TMS320DM320?
22:09:51MunkieManis it possible to control the lcd while in diskmode? (on the ipod)
22:10:16slyyfmcuelenaere_: I couldnt find any docs to say if it did or did not
22:10:51mcuelenaere_slyyf: aren't there any pictures taken of the inner of the P2?
22:11:08slyyfnone good quality that I know of
22:11:38mcuelenaere_oh wait, is this the samsung P2 port?
22:12:03slyyfits a modded cowond codebase
22:12:06slyyfto work on the p2
22:12:15slyyfthat at the moment almost has the lcd screen working
22:12:15*mcuelenaere_ wonders why there's no wikipage of the P2
22:12:23slyyfthere is
22:12:31 Quit mcuelenaere (Read error: 110 (Connection timed out))
22:12:39mcuelenaere_does it have its own forum thread?
22:12:42 Nick mcuelenaere_ is now known as mcuelenaere (n=mcuelena@rockbox/developer/mcuelenaere)
22:12:46slyyfSOmewhere..a bad one
22:13:59mcuelenaereoh, I was a bit confused
22:14:17mcuelenaereI thought, as the D2 has the same LCD as the ZVM, that it runs the same chipset; but these are TeleChips
22:14:17slyyfI found something
22:14:43mcuelenaereslyyf: then SDI should be explained in the datasheet
22:14:58slyyfwill it refer to it as an SDI?
22:15:45mcuelenaerelook for LCD (interface)
22:16:37slyyfIt has a list of registers
22:16:42 Join n3hima [0] (n=n3hima@unaffiliated/n3hima)
22:16:59n3himawhere might I find documentation for plugin.h?
22:17:28Unhelpfuln3hima: plugin.h *is* the documentation, to a certain extent... but what exactly are you trying to do?
22:17:55n1sn3hima: the source is the documentation for the various functions that are exported to plugins basically
22:18:13mcuelenaereslyyf: SDI seems to be specific to the LCD, not the SoC
22:18:20n3himaso the ticks argument of the sleep function is what, exactly?
22:18:28slyyfmcuelenaere: oh wow..this datasheet is a draft..
22:18:34T0pazmcuelenaere, do you know whether nucleus starts at 0x00000000 (well, after the jump), or whether they tacked on some code before nucleus starts?
22:18:46mcuelenaereT0paz: yes, that was what I was looking for
22:18:54mcuelenaereI have it written down *somewhere*
22:19:08T0paznucleus starts by setting a state register to 1 ;)
22:19:09mcuelenaereI do remember that the actual init has about 10-13 BL's
22:19:12T0pazthat'll be easy to find...
22:19:18mcuelenaereoh and look for strings
22:19:22T0pazah yes
22:19:31mcuelenaereone of those BL's only reference a string
22:19:48T0pazi've been making an HTML version of the disassembly (since i don't have IDA)
22:19:55T0pazyou can click the links to see where they go ;)
22:20:08mcuelenaeresee inc.c
22:20:34mcuelenaereI believe someone did the same with a script, I think it's committed to SVN
22:20:44T0pazah yeah
22:20:53mcuelenaereyou'll recognize INC_Initialize()
22:20:56T0pazi'm considering making it colour-code interesting bits (like hardware accesses)
22:21:03mcuelenaerelook for a license string, that's LIC_License_Information()
22:21:45T0pazah yes
22:22:44 Join Thundercloud [0] (
22:23:17 Join EspeonEefi [0] (i=eefi@STRATTON-SIX-FOURTY-SIX.MIT.EDU)
22:24:10mcuelenaereT0paz: whenever you've found INC_Initialize(), Application_Initialize() is the Creative custom code
22:24:16mcuelenaerethat's when the real RE starts :)
22:24:16T0pazah, cool
22:24:22T0pazah yes ;)
22:24:26T0pazi'll get the hang of this one day
22:24:28mcuelenaereone of the first calls in that one, is drivers_init()
22:25:15mcuelenaereI once had quite a lot of all this disassembled, but lost it during a crash..
22:25:19 Part n3hima
22:25:22T0pazah, doh
22:25:39mcuelenaereT0paz: the string is '"Accelerated Technology Internal Use Only - Serial Number: NP0000'
22:25:47T0pazah yes, i just found that
22:25:48Unhelpfulpixelma: when you get a chance, would you mind removing your .rockbox/rocks/demos/pictureflow.cfg and trying this patch?
22:25:58mcuelenaere(I found a Zen X-Fi IDA database I had laying around)
22:26:00T0pazi haven't disassembled the 40200000 block
22:26:02T0pazmaybe i should
22:26:10T0paz(it's in there)
22:26:30Unhelpfuli think it improves things greatly on c200 sim... and it looks nice on several other portrait and landscape sims that i tried.
22:27:59mcuelenaereT0paz: ah yes, look for lcd0
22:28:00 Quit MunkieMan ("CGI:IRC (EOF)")
22:28:05mcuelenaerethat's the internal LCD driver name
22:28:11mcuelenaerethere are more of these names
22:28:14mcuelenaerelike codec0 etc
22:28:21kugelBigBambi: your patch could potentially apply to the e200 too
22:28:39kugelI'm gonna test disk full cases later and report back
22:28:52BigBambikugel: The PREV not UP one you mean?
22:29:02mcuelenaerehmm this doesn't look good, it seems as the Zen also has mcu0
22:29:04BigBambiI can test too, one mo
22:29:12BigBambiI just need to find my e200 :)
22:29:36kugelah :)
22:30:44kugelBigBambi: no matter of that, I think "PLAY" is the "naming convention" of the up button (the database says play)
22:31:02BigBambikugel: yes, that's true
22:31:25mcuelenaereT0paz: ahh, the main thread is added in Application_Initialize()
22:31:33mcuelenaereso there won't be a branch in it
22:31:49pixelmaUnhelpful: just came back, will test in a bit - on c200 I assume?
22:32:47BigBambikugel: That appears to be the only instance of UP
22:32:58BigBambiI'll verify it then change it accordingly
22:33:09kugelyea, rockbox-wide even
22:33:11T0pazok, i've found the license function
22:33:40 Quit t0mas ("good night")
22:34:05T0pazi've found all the BLs
22:35:14kugelBigBambi: hm, I actually think we shouldn't even try to translate buttons. just say what's printed (i.e. not PREV, but |<< in case of the sansas)
22:36:08kugeltranslating buttons is nasty, since it's basically never what's actually printed on the buttons, and is always difficult if the button mapping changes
22:36:13BigBambikugel: I'm not so sure
22:36:21BigBambiThat could get difficult though
22:36:31Unhelpfulpixelma: yes, i've kept the reflections the same height, but allowed images to be wider - they should be about 50x50 for square covers now, on c200
22:36:34BigBambiAnd << does mean PREV pretty much
22:37:17BigBambikugel: I think that'd need wider consultation - this is just getting it correct/consisytent to what is there now - a wider shange can then be discussed
22:37:28kugelpossibly, but I don't see what's wrong with just saying the same that's printed on the button
22:37:39rasherkugel: How would you say what's printed on the lower button on an e200?
22:37:50rasher"a long horizontal line above 3 shorter ones"
22:37:56BigBambikugel: It i sometimes not clear what is printed
22:38:09kugelrasher: yea, that's what I just noticed to be problematic too :/
22:38:14BigBambihmmm, my e200 is refusing to connect
22:38:16rasherkugel: Let alone the center button ""
22:38:35kugelwell, obviously, if there's nothing printed we need to give it a name
22:39:27kugelBigBambi: Don't get me wrong, I didn't mean to change that just now (especially not in your patch), I just wanted to mention it
22:39:39BigBambikugel: sure
22:39:40kugelsure it needs wider consultation
22:40:02BigBambiAs soon as my sodding e200 decides to connect to my laptop, I can test it
22:41:03kugelBigBambi: are you using the quickscreen? :>
22:42:08mcuelenaereT0paz: do you also have at about 0x400BE0A4 a string 'system'?
22:43:33T0pazi do at 0x4004FB20
22:43:52BigBambikugel: On the e200? I'm not doing anything except swearing at it
22:44:20T0pazwell, lots of times infact
22:44:28T0pazbut not near there
22:44:45mcuelenaereT0paz: yes, I know but at that position in the zen x-fi file is the drivers_init() func
22:44:56mcuelenaerehow does the function look like above directly that string?
22:45:18mcuelenaeres/(above) (directly)/\2 \1/
22:46:31kugelBigBambi: no, any target
22:46:33 Join QuickStart [0] (
22:46:46BigBambikugel: Yes, from time to time
22:47:12T0paza couple of bl, a couple of blx
22:47:41mcuelenaerejust that?
22:47:54kugelBigBambi: I have 2 patches which could need testing: and
22:47:57mcuelenaereand an ADR Rx, aSystem?
22:48:00 Join miepchen^schlaf [0] (
22:48:10mcuelenaereabout 8 BL's
22:48:29T0pazah, the one above that has more BLs
22:48:37T0pazi really should make it easier to find string literals
22:48:44T0pazat the moment i'm finding them in a text editor first
22:48:49T0pazer, hex editor
22:48:53mcuelenaereare there more BL's in the first BL? :)
22:49:26mcuelenaereah and one of the BL's reference a 'flash' string
22:49:41BigBambikugel: I think that lang file must have got changed - I just tried on my e200 running a very old version (r18something) and it says press left
22:49:44mcuelenaereI have about 7 very close to each other
22:49:47T0pazheh, i'm lost now
22:50:07BigBambiUnfortunately, as I cannot currently connect my e200 to my laptop, I can't update and check now
22:50:31mcuelenaerewell the 'flash' reference is in a BL under the main function (the one with 8 BL's)
22:50:36 Join MethoS- [0] (
22:50:46kugelBigBambi: hm, but the key hasn't changed?
22:50:54kugellooks like a buggy commit to me then
22:51:09BigBambikugel: I'll have a look in the keymap, but can't check on target
22:51:34mcuelenaereseems like one of those functions is also a ATA_DMA driver init
22:53:00mcuelenaerehmm it seems like it's easy to spot register_irq() calls
22:53:16mcuelenaerethey all start with setting a string in R0 containg the name of the IRQ
22:53:23BigBambikugel: ACTION_STD_CANCEL is BUTTON_LEFT still
22:54:00BigBambiit looks like in fact that I shouldn't be splitting the c200 out, but changing both c200 and e200 to PREV
22:54:24BigBambikugel: Do you know what the e200v2 uses for cancel?
22:54:43T0pazmcuelenaere, the name being things like 'lcd0' ?
22:54:52mcuelenaerenope, that's something else
22:54:57*kugel asks svn blame
22:55:02mcuelenaerethis seems like STMP37xx specific
22:55:06mcuelenaereor it's SDK
22:55:21BigBambiOK, so I'll change that patch to replace UP with PREV instead
22:55:23 Quit bmbl ("Woah!")
22:55:34mcuelenaerestrings like ATA_DMA, AUDIOOUT, MCU_HISR, SSP2_DMA, SSP1_DMA
22:56:40mcuelenaerethose with lcd0 are mostly either driver_register() or driver_query() calls
22:56:58mcuelenaere(those names are invented)
22:57:24BigBambirasher: So if I am in fact changing a string, rather than changing target then adding another one, do I just edit english.lang then leave the others?
22:58:41BigBambialthough, a lot of langs seem to leave the "UP" or "POWER" untranslated, so I could just change them
22:59:46BigBambinot all however
23:00:06mcuelenaereT0paz: about those weird 0x4xxx references, Creative mapped/copied quite some data before calling the Nucleus init on the ZVM
23:00:14BigBambikugel: In reply to the patche, I can't right now but I'll try and have a look tomorrow
23:00:14mcuelenaereI presume this is also true for the Zen, only SoC specific
23:00:28T0pazhmm, neither ATA_DMA or MCU_HISR are directly referenced by address
23:00:38T0paz(from anywhere)
23:00:42mcuelenaereso you could try looking what happens at 0x0 (lookup how the ARM MMU works)
23:00:53kugelBigBambi: no problem, no need to hurry :)
23:01:40mcuelenaereoh or these addresses could also be set up by FBOOT
23:01:57mcuelenaerelet's hope they aren't, because it isn't in nk.bin
23:02:06 Quit {phoenix} (Remote closed the connection)
23:02:16mcuelenaere+ FBOOT is partly encrypted by RSA
23:02:20T0pazthey were at 0x3fff.... (just below the main code)
23:02:22T0pazwhich seemed strange
23:02:38T0pazby some sigmatel-proprietary key?
23:02:55mcuelenaereno, I'm talking about the ZVM
23:03:00mcuelenaerewith a Creative-key :)
23:03:10T0pazthough is the key in firmware?
23:03:13mcuelenaereCreative likes to do everything as they wish
23:03:15mcuelenaereyes it is
23:03:33mcuelenaereso they've probably also rewritten the actual bootloader
23:04:12mcuelenaerebut you shouldn't worry about this RSA-stuff, it's in FBOOT and currently there's no way to access it :)
23:04:24kugelBigBambi: actually, c200 is more interesting for me anyway, since i can test on the e200 myself
23:04:27T0pazi wonder why they bother
23:04:37mcuelenaereto do all this encryption?
23:04:43T0pazunless that has the WMA DRM keys in it or something
23:04:53mcuelenaereyes, it's probably something like that
23:05:14mcuelenaerealthough there are several layers of encryption
23:05:30T0pazi guess it's just a deterrent
23:06:01T0paz(unless there is RO memory inside the SoC)
23:06:27 Quit MTee (Connection timed out)
23:07:17 Join PaulJam [0] (
23:07:33rasherBigBambi: Yeah, just editing english.lang (both source and dest) should do it
23:07:38 Join AndyI [0] (i=AndyI@
23:08:04BigBambirasher: Thanks - will it then get marked as needing attention for the other languages by your page?
23:08:13 Quit QuickStart (Remote closed the connection)
23:08:23*mcuelenaere wonders as Sigmatel seems to be taken over by Freescale whether they would provide more information
23:08:54rasherBigBambi: It should
23:09:03BigBambiCool, thanks
23:09:16rasherBigBambi: it depends on whether the <source> part of english lang is different from the <source> in the translation (which it will be)
23:09:42BigBambirasher: Ah, OK. I'm slowly starting to get it :)
23:10:54kugelrasher: but until a translater bothers enough, the langs will be wrong or show the source string?
23:11:43BigBambiShow the old string I imagine
23:12:03 Quit Zoxc ()
23:13:57T0pazmcuelenaere, do you know what the 'AHB to APBX Bridge' does?
23:14:09 Quit PaulJam_ (Read error: 145 (Connection timed out))
23:14:12mcuelenaereI recognize AHB
23:14:31mcuelenaereit's probably a generic term, it's also used in TMS320DM320
23:14:42T0pazi think that's ARM related
23:14:55T0pazah, it is AMBA
23:15:00 Quit pixelma (Read error: 110 (Connection timed out))
23:15:06BigBambiOK, 9826 updated. Look at the complexity of that patch! :P
23:15:24T0pazso presumably allows the core to talk to APBX, whatever that is
23:15:51 Join pixelma [0] (n=pixelma@rockbox/staff/pixelma)
23:16:01mcuelenaereyes, probably
23:16:02T0pazthe second call in Application_Initialize sets/clears SFTRST and then clears CLKGATE (of the APBX control register)
23:16:10mcuelenaerean ARM926-EJS has an AHB
23:16:24 Quit pyro_maniac ("Leaving.")
23:16:28kugelBigBambi: impressive! :)
23:16:37 Quit amiconn (Read error: 110 (Connection timed out))
23:16:40mcuelenaereT0paz: are you deriving this from the BSP?
23:17:02 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
23:17:03mcuelenaere(clearing CLKGATE probably means disabling the clock for APBX, so disabling it)
23:17:18mcuelenaereor it could be the opposite
23:17:25 Join tvelocity [0] (
23:17:34BigBambikugel: seeing as a lot of the other langs doen't translate UP etc, I'm going to change those too
23:17:47T0pazyeah, the BSP
23:18:57rasherkugel: The translation will show the english string, which is better than a wrong string
23:19:06mcuelenaerearen't there any drivers in that? (on which you can base yourself)
23:19:08T0pazthe call after does the same thing to the APBH interface
23:19:10kugelrasher: I agree
23:19:19rasherAt least if I understand the question
23:19:25T0pazi can't find any SMTP drivers in the BSP except for the IDE one
23:19:27T0pazwhich isn't very helpful
23:19:40T0pazer, STMP
23:19:46mcuelenaerehmm that's weird
23:20:22T0pazoh, and a serial debug driver
23:21:19mcuelenaerekernel2.6.16\linux-2.6\arch\arm\mach-stmp36xx\stmp36xx.c handles the interrupts
23:21:57 Join MTee [0] (n=mtee@
23:22:28mcuelenaerehmm it'll probably be easier to do a diff against linux 2.6.16
23:23:38 Quit bluebrother (Nick collision from services.)
23:23:43 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
23:24:43UnhelpfulSTMP progress? i have one of those... but there are so very many devices :/
23:25:04kugelBigBambi: that's what I did too today. I think if it's easy, that one should change as many langs as possible
23:25:10T0pazstill stabbing in the dark...
23:25:54BigBambikugel: Yes, I've just finished. If the lang hadn't previously translated UP, POWER, PREV etc then I have just swapped UP and PREV by hand
23:29:07mcuelenaereseems like has a backup of linux.sigmatel:*/
23:29:45 Join MethoS-- [0] (
23:30:48 Quit MethoS-- (Client Quit)
23:31:00 Join MethoS-- [0] (
23:31:03 Quit MethoS-- (Remote closed the connection)
23:31:23 Join MethoS-- [0] (
23:31:23 Quit MethoS-- (Remote closed the connection)
23:34:59*mcuelenaere is off to bed
23:36:28T0pazi think i'll improve my disassembler
23:38:45mcuelenaereT0paz: you know of ?
23:39:01T0pazi did spot that
23:39:22T0pazi'm using objdump for the actual disassembly
23:39:25kugelUnhelpful: ping
23:39:46T0pazjust wrapping it in HTML with clickable addresses and demarcations of subroutines and things
23:40:42mcuelenaereyes, I think vitja has done the same
23:41:06 Join PaulJam_ [0] (
23:42:16kugelUnhelpful: <−− I'm inclinded to close this, as I'm fairly sure it's "not a bug" (and the reporter doesn't answer our questions). what would you say?
23:42:27 Quit bluebrother ("leaving")
23:42:34kugelI think it's the usual stop shortly to resize album art
23:43:21gevaertsIt seems like the sansa SD bug may be gone soon
23:44:24kugelgevaerts: should I test too? I guess more tests are appreciated
23:44:48BigBambigevaerts: awesome - has someone worked some magic?
23:44:53rashergevaerts: woo
23:44:55gevaertskugel: I guess more tests are certainly welcome. See FS #8663
23:45:21kugelgevaerts: how would I test? copy some gigabytes and diff or md5 the files?
23:45:31 Quit mcuelenaere ()
23:45:56gevaertskugel: that's what I did. Unmount in between to make sure the reading back actually comes from the device and not from an OS cache
23:46:27kugelgevaerts: did you use full or highspeed usb?
23:46:40 Quit MethoS- (Read error: 113 (No route to host))
23:46:55gevaertsI always use highspeed. It shouldn't matter though
23:47:10*gevaerts can't imagine filling up his e280 on full speed
23:48:14 Join PaulJam__ [0] (
23:50:41 Quit n1s ()
23:51:22kugelgevaerts: ok, how to enable usb again?
23:51:42gevaertskugel: define USE_ROCKBOX_USB somewhere.
23:52:00slyyfHey guys, I am going to take a break from the LCD for a bit, any idea how I would turn on/off the three LED rings on the P2?
23:52:08*gevaerts sets EXTRA_DEFINES=-DUSE_ROCKBOX_USB
23:52:25slyyfIs the one way of figuring that out dissasembly?
23:53:34amiconngevaerts: The patch in #8663 is more of a workaround at the cost of speed than a true fix
23:53:54***Saving seen data "./dancer.seen"
23:54:00Bagderslyyf: yes, or read the data sheet if you have one
23:54:49amiconnThere's some PP register we don't know yet. The transfer involves a fifo, which should allow reading and writing at maximum speed as long as we don't over-/underflow it.
23:56:02amiconnI have an optimised write transfer routine that makes the bug hit immediately. This is quite strange and poits to some missing write timing setting, because the problem doesn't happen on read, even though reading is fully optimised.
23:58:07*gevaerts is confused
23:58:27gevaertsThe description doesn't really match the patch...

Previous day | Next day