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

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

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

#rockbox log for 2009-08-01

00:09:57 Join wincent [0] (
00:10:05 Quit aaron424 (Read error: 104 (Connection reset by peer))
00:12:05 Join togetic [0] (n=togetic@unaffiliated/ibuffy)
00:16:42 Quit ender` (" If the experiment works, you must be using the wrong equipment. -- Patrick's Theorem")
00:27:29 Join aaron424 [0] (
00:28:34 Join FOAD_ [0] (
00:33:38 Quit aaron424 (Remote closed the connection)
00:36:13 Quit bluebrother ("leaving")
00:40:15 Join dash32 [0] (
00:42:35 Quit FOAD (Read error: 110 (Connection timed out))
00:42:35 Nick FOAD_ is now known as FOAD (
00:42:42 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
00:53:08bertrikyay, I think I know how to read the touch key controller on the samsung yp-s3 now
01:01:37 Quit dmb (Read error: 104 (Connection reset by peer))
01:07:10***Saving seen data "./dancer.seen"
01:12:59 Quit DarkDefender (Remote closed the connection)
01:16:30 Join dmb [0] (n=Dmb@unaffiliated/dmb)
01:21:00 Join thx4help [0] (n=4496cf5b@gateway/web/cgi-irc/
01:22:43 Quit JdGordon| ("Miranda IM! Smaller, Faster, Easier.")
01:22:53thx4helpI noticed a couple of forum comments saying the Sansa c250 worked with rockbox. Just wondering if anyone knew off hand if this was true.
01:25:19gevaertsthe v1 works with rockbox, yes
01:25:56thx4helpalright, thank you very much
01:33:02 Join T0 [0] (
01:34:57 Join JdGordon_ [0] (i=ad814d37@gateway/web/freenode/x-pqiqlqrctpaariko)
01:35:44 Quit To (Read error: 110 (Connection timed out))
01:36:14 Quit togetic (Read error: 104 (Connection reset by peer))
01:36:34 Join togetic [0] (n=togetic@unaffiliated/ibuffy)
01:38:32 Quit thx4help ("CGI:IRC (EOF)")
01:55:51 Quit togetic (Remote closed the connection)
01:56:12 Join togetic [0] (n=togetic@unaffiliated/ibuffy)
02:00:01 Quit JdGordon_ (Ping timeout: 180 seconds)
02:06:33 Quit Rondom (Nick collision from services.)
02:06:36 Join Rondom_ [0] (
02:09:38 Join KBH [0] (
02:10:02 Join fg56lx [0] (
02:22:18 Quit Zagor ("Clint excited")
02:26:30 Quit bertrik (Read error: 113 (No route to host))
02:27:14 Quit weezerle (Remote closed the connection)
02:27:28 Join fg57lx [0] (
02:28:07 Quit HBK (Read error: 110 (Connection timed out))
02:36:02 Join evilnick [0] (i=620ec27e@gateway/web/freenode/session)
02:41:30 Quit fg56lx (Read error: 110 (Connection timed out))
03:07:13***Saving seen data "./dancer.seen"
03:07:13 Quit togetic (Remote closed the connection)
03:07:33 Join togetic [0] (n=togetic@unaffiliated/ibuffy)
03:16:03 Quit w1ll14m (Read error: 54 (Connection reset by peer))
03:29:32 Quit dash32 (Remote closed the connection)
03:47:11 Quit evilnick ("Page closed")
03:49:11 Quit saratoga ("Page closed")
04:06:26 Join xavieran [0] (
04:06:42 Quit efyx_ (Remote closed the connection)
04:13:36 Quit fbt``` (Remote closed the connection)
04:19:29 Quit togetic (Read error: 104 (Connection reset by peer))
04:19:47 Join togetic [0] (n=togetic@unaffiliated/ibuffy)
04:45:22 Join CaptainKwel [0] (
05:06:05 Join ze0 [0] (
05:07:17***Saving seen data "./dancer.seen"
05:10:12 Join webguest67 [0] (n=4b5aa4c6@gateway/web/cgi-irc/
05:11:50 Quit webguest67 (Client Quit)
05:15:44 Quit Telazorn (Read error: 104 (Connection reset by peer))
05:17:54 Quit ze (Read error: 110 (Connection timed out))
05:17:54 Nick ze0 is now known as ze (
05:19:50 Join AndyI [0] (i=AndyI@
05:24:34 Quit AndyI (Client Quit)
05:28:02 Join AndyI [0] (i=AndyI@
05:32:08 Quit AndyIL (Read error: 110 (Connection timed out))
05:36:59 Join kamlurker [0] (
05:59:11 Quit kamlurker ("ChatZilla 0.9.85 [Firefox 3.0.12/2009070818]")
05:59:14 Quit wincent (Read error: 110 (Connection timed out))
06:08:47 Join evilwombat [0] (
06:08:53evilwombatGreetings, Rockbox people.
06:09:12evilwombatIs there a single-button shortcut to go from the WPS screen to the track list on an iriver (340) ?
06:09:34evilwombatand if not, might someone point me to which file has the #defines I might need to change? I have toolchain, etc, just never went into that part of the code
06:09:41ej0rgeyeah - 'navi' if you're file browsing
06:10:06evilwombatthanks! will try
06:10:14evilwombat(asking for my friend, who actually has the iriver)
06:10:38ej0rgeI don't use the database so i don't know if that method differs
06:11:02ej0rgeoh, and that trick only works if 'follow playlist' is enabled
06:11:21 Join hd [0] (i=jd@
06:11:24evilwombatwhat does it do otherwise?
06:11:43ej0rgetakes you to the root of the filesystem instead of the file you're listening to
06:12:07evilwombatah thanks
06:12:13evilwombatit sounded like this is what he wanted.. he is checking
06:12:31evilwombatclaims he is stuck using the original FW instead of rockbox because it's "2-3 button presses to get to the song list"
06:12:44n1sy'know, things like this are in the manual ;)
06:13:00evilwombatit is true..
06:13:17evilwombatI actually said, "damnit, if it's just the buttons, we can cahnge the #defines or something"
06:17:28 Join lucas99801 [0] (
06:18:03lucas99801how do i enable timestretch? or is it not available on Ipod Video 30g
06:27:29n1slucas99801: it is described in the manual, you need to enable the setting and adjust it in the pitch screen
06:27:57lucas99801i couldn't find out how to enable it
06:28:14lucas99801do i have to edit a config file? or is it in a menu
06:28:19lucas99801i looked in almost all of them
06:28:41 Quit HellDragon (Read error: 110 (Connection timed out))
06:29:25 Join fg56lx [0] (
06:29:28n1sit should be in the menu Fir
06:33:40n1sSounds Settings->Timestretch
06:35:20lucas99801it's not there
06:42:12 Quit fg56lx (Remote closed the connection)
06:42:15 Quit fg57lx (Read error: 104 (Connection reset by peer))
06:52:59 Join Telazorn [0] (n=Miranda@
06:54:56 Join evilnick_home1 [0] (
06:56:46evilwombatI am building latest SVN and getting internal compiler error: insn does not satisfy its constraints: libfaad/ps_dec.c:72: internal compiler error: in reload_cse_simplify_operands, at postreload.c:391
06:57:09evilwombatI wanna say my toolchain is waay outdated? Has anyone got a tarfile of the m68k/coldfire toolchain for linux x86?
06:59:26ej0rgemost people get the toolchain by running tools/
07:03:12evilwombatej0rge, how odd. it appears I have ran it previously, and it built that toolchain. but, having pointed it at the right dir, I still get an internal gcc error
07:04:35 Join fg56lx [0] (
07:04:42 Join fg57lx [0] (
07:05:08evilwombatall guilds are green...
07:07:16ej0rgewhat's m68k-elf-gcc -v tell you?
07:07:19***Saving seen data "./dancer.seen"
07:07:23n1slucas99801: are you using a recent version of rockbox?
07:07:44evilwombatn1s, using what is in svn
07:08:15lucas99801n1s i used the application that installs rockbox for you
07:08:17evilwombatej0rge, claims gcc 3.4.6
07:08:18lucas998015 days ago
07:08:21 Quit fg56lx (Read error: 104 (Connection reset by peer))
07:08:21 Quit fg57lx (Read error: 104 (Connection reset by peer))
07:08:35ej0rgeevilwombat: same as I'm running then
07:08:57n1slucas99801: strange, as that setting is there in the svn code and has been for quite a while
07:09:32ej0rgeevilwombat: from what google can tell me, that bug was fixed in gcc 3.4.5
07:09:59 Join fg57lx [0] (
07:10:05 Quit fg57lx (Read error: 104 (Connection reset by peer))
07:10:15evilwombatej0rge, weird. will recheck my paths. the script just built me 3.4.6
07:10:31evilwombatit appears i have a 3.4.4 floating around also, and will rm for good measure
07:10:42ej0rgeevilwombat: some sort of path issue would be my only guess at this point, but I'm hardly an expert
07:10:48lucas99801my sound settings menu stops at dithering
07:11:21 Quit evilnick_home (Read error: 113 (No route to host))
07:11:41evilwombatej0rge, building now... we'll see
07:12:32evilwombatej0rge, ooh, it got past that file. thanks!
07:13:11lucas99801n1s, should it appear below dithering?
07:15:00evilwombatwell, that is all i needed. thx for the help, and good night all
07:15:06 Part evilwombat ("^D")
07:29:12 Join Horschti [0] (n=Horscht2@xbmc/user/horscht)
07:31:27 Quit lucas99801 ("Leaving")
07:47:21 Quit Horscht (Read error: 110 (Connection timed out))
08:08:14 Quit CaptainKwel (Remote closed the connection)
08:13:18 Join flydutch [0] (
08:29:36 Join Rob2222 [0] (
08:34:29 Join _lifeless [0] (n=lifeless@
08:34:47 Quit ze (Read error: 110 (Connection timed out))
08:36:33 Join bertrik [0] (
08:46:57 Quit Rob2223 (Read error: 110 (Connection timed out))
09:07:21***Saving seen data "./dancer.seen"
09:19:19 Quit Lss (Read error: 104 (Connection reset by peer))
09:19:45 Join w1ll14m [0] (
09:27:20 Quit amiconn (Nick collision from services.)
09:27:20 Quit pixelma (Nick collision from services.)
09:27:22 Join pixelma_ [0] (i=quassel@rockbox/staff/pixelma)
09:27:24 Join amiconn_ [0] (i=quassel@rockbox/developer/amiconn)
09:27:31 Nick amiconn_ is now known as amiconn (i=quassel@rockbox/developer/amiconn)
09:27:40 Nick pixelma_ is now known as pixelma (i=quassel@rockbox/staff/pixelma)
09:32:47pixelman1s: there is one shortcut from WPS to the playlist viewer now (and targets that had a button (combo) spare) which I don't think is in the manual yet
09:46:14bertrikI can now get stable key readings from the "touch keys" on the samsung yp-s3
09:48:25bertrikIt turned out to be more easy than I first thought. It's a simple protocol where the touch key controller pulls down a GPIO to indicate an event, then the CPU can toggle another GPIO to clock out the key presence bits.
09:51:07linuxstbbertrik: Is there a decent number keys?
09:51:29linuxstbI mean a decent number of keys
09:52:00bertrikyes, it has up/down/left/right/middle and back/menu buttons
09:52:43linuxstbNo "play", "stop" or "pause" ?
09:53:48bertrikin the OF, "play"/"pause" is just the middle button
09:54:40bertrikthe up/down/left/right/middle buttons are not marked with any indication of function
09:56:36bertrikeach button has a LED behind it (some are coupled though) I wonder if we can do something useful in rockbox with that, like lighting the LED if the button has some meaning in the current context
09:57:41linuxstbI would guess that would result in all LEDs being always lit....
09:57:57bertrikhm, possibly :)
10:00:26bertrikmarkun, some time ago you mentioned some other DFU image for s5l8700 based players IIRC, something like bluesNAND... or NANDblues
10:01:10bertrikcould you send it to me so I can have a try to see what it does?
10:01:54bertrikdo you also have a firmware upgrade / recovery utility that this DFU file belonged to possibly?
10:03:11 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
10:03:18markunbertrik: yes, one moment
10:06:35 Join FOAD_ [0] (
10:07:46 Join fg56lx [0] (
10:15:16 Quit martian67 (Read error: 104 (Connection reset by peer))
10:17:53 Join ender` [0] (
10:17:58 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
10:20:34 Join njr [0] (n=Administ@
10:23:08 Quit MrDuck (Read error: 104 (Connection reset by peer))
10:23:11njrHi ... can anyone point me to a link for compiling Rockbox on an iAudio7
10:23:49 Quit FOAD (Read error: 110 (Connection timed out))
10:23:49 Nick FOAD_ is now known as FOAD (
10:26:26 Join einhirn [0] (
10:26:32xavierannjr: You can probably just use tools/configure and specify iAudio7
10:26:56pixelmaI'd try searching the wiki or the new ports section in the Rockbox forums. Be aware that the port is in early stages and seems to have stalled, so things could be broken. By the way - I doubt you could compile Rockbox "on" the device ;)
10:30:51njrlol .. yeah... will have to download (compile ?) a cross compiler... all i have is a 10 yr old linux laptop and a windows pc ... sigh
10:31:26linuxstbnjr: The last person that tested Rockbox on the iaudio 7 reported that the LCD didn't work, so you may not have much luck.... But as pixelma said, check the New Ports forum thread for the iaudio 7, plus the wiki.
10:31:45njrpixelma, will do .... xavieran, whats tools/configure
10:32:41linuxstbnjr: The Rockbox website has lots of info on compiling Rockbox - see the "For Developers" section here -
10:32:45njrlinuxstb.. ok .. i dont know any assembly.. but dont mind fooling around with the C stuff.. and id like to be useful
10:32:48xavierannjr: when you have your cross compiler installed, you can execute tools/configure to configure your build for your device, I think it includes the iAudio7, then you would just have to execute make and make install
10:33:21njrxavieran.. ok
10:33:40linuxstbxavieran: No, there is more to it than just "make install". You first need to install a bootloader, and that process is specific to the device.
10:34:19pixelmaalso, "make install" is only needed to build a simulator </nitpicking mode>
10:34:38njrok linuxstb.. thats the same for all player?...where do i get a cross compiler from?
10:34:38xavieranlinuxstb: yes, that was just a generalized instruction...
10:34:48xavierannjr: Read the wiki page
10:36:45 Quit advcomp2019 (Read error: 60 (Operation timed out))
10:36:46pixelmaI'd look around for target specific pages first
10:37:21njrxavieran, pixelma ook thanks
10:37:45xavieranno problem :)
10:39:20njrsorry to bother ... any links for compiler?
10:41:06njrcool thanks
10:41:48pixelma ;)
10:43:43njr:)... was doing a search from the wiki .,.. was sending me to google and some irrelevent results
10:56:42 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
11:07:25***Saving seen data "./dancer.seen"
11:25:08 Quit einhirn (Read error: 104 (Connection reset by peer))
11:27:44 Join darkhamm [0] (
11:37:33 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
12:05:25 Join mcuelenaere [0] (
12:05:27 Quit KBH (Read error: 104 (Connection reset by peer))
12:07:07 Join HBK [0] (
12:13:00 Quit T0 (Remote closed the connection)
12:16:36 Quit xavieran (Remote closed the connection)
12:20:01 Join stoffel [0] (
12:31:16 Join xavieran [0] (
12:37:20 Quit darkhamm ("Sto andando via")
12:50:11 Quit Telazorn (Read error: 104 (Connection reset by peer))
12:58:50 Join njr1 [0] (n=Administ@
13:03:57 Join njr2 [0] (n=Administ@
13:04:19 Join BdN3504 [0] (n=5ce225be@gateway/web/cgi-irc/
13:06:32BdN3504hey, it's me again on the topic of accepting FS 10187 into svn. I have posted compiled patched builds for testing to the forums, but i can't add them to the official test builds section.
13:06:40BdN3504who's allowed to do that?
13:07:29***Saving seen data "./dancer.seen"
13:09:29BdN3504come on guys, i spent three hours yesterday to do this...
13:09:51 Join Thundercloud [0] (
13:09:55Tornevery few people are around at this time of day, expecially on a weekend
13:10:44Torneofficial test buiulds forum can only be posted to by developers
13:11:49 Join darkless [0] (
13:13:07mcuelenaereBdN3504: I can create a new thread in that forum
13:16:38 Quit hd (Read error: 104 (Connection reset by peer))
13:16:51 Join hd [0] (
13:18:05pixelmaalso, the whole project is a volunteer effort and no-one forced you to spend the time
13:18:42 Quit njr (Read error: 110 (Connection timed out))
13:19:10 Join njr [0] (n=Administ@
13:19:34 Quit njr1 (Read error: 110 (Connection timed out))
13:27:44njrHi.... am haveing truobe compiling ... did a make ... and it says apps/bitmaps/rockboxlogo.h is not a file
13:28:17 Join kugel [0] (n=kugel@rockbox/developer/kugel)
13:28:21 Quit thegeek (Read error: 104 (Connection reset by peer))
13:28:32 Join thegeek [0] (
13:29:08tmztnjr: re run configure
13:29:37njrhad done that ..sorry ... did another make clean and its compiling
13:30:23njralthough its giving a lot of #error: unsupported keypad
13:37:03 Quit njr2 (Read error: 110 (Connection timed out))
13:40:10pixelmamcuelenaere: by the way, I thought that the 3x3 grid operation on touchscreen targets was agreed on being the default because you can use it without looking (also as a blind or visual impaired person, and if I think I read that surprisingly there are some). And could you think about a better naming of "absolute point"? I translated it a bit differently in the German lang to make it a bit more descriptive (I hope)
13:41:13*gevaerts still is one of those people who think that grid mode is a bad default
13:41:31mcuelenaerepixelma: isn't a touchscreen a general bad idea for blind people?
13:41:46mcuelenaere(or visual impaired)
13:42:24mcuelenaereand no, I don't know a better alternative for "absolute point mode"; perhaps it should be renamed to 'normal' and 'grid mode' to 'emulated buttons'?
13:42:24pixelmamaybe, but people thought the same about the Ipods' touch wheel
13:43:21kugeldirect touch mode?
13:45:13pixelmamy impression of looking at some translation that I sometimes use for inspiration (nederlands, svensk) was that translators used the noun "meaning" of point and that didn't make sense to me in German. Indeed I used something like "direct touch"
13:45:22njrHi ... sorry... apps/plugins/flippit.c ...#error No keymap defined FLIPPIT_QUIT (And a whole bunch of others) undeclared.. t
13:45:33mcuelenaerenjr: what target is this?
13:46:00mcuelenaereI don't think plugins are adjusted to the iAudio 7 yet, try compiling without them
13:46:01BdN3504mcuelenaere: thanks!
13:46:18njrook how2 do i do that... edit Makefile?
13:46:25mcuelenaereeasy way: make bin
13:46:33mcuelenaerebetter way: edit tools/configure
13:46:46mcuelenaereeven better way: fix the plugins
13:47:09njrook thanks mcuelenaere
13:47:59mcuelenaerepixelma: IMHO it should be renamed so it makes clear the 'emulated buttons' is non-standard (even though it is on by default), and 'normal mode' would be absolute point
13:49:07pixelmanormal mode is even less descriptive to me
13:51:23mcuelenaerewhen you have a touchscreen, you'd expect it to activate a menu when clicking on something; hence 'normal mode'
13:51:24kugeli like "direct touch" and "grid"
13:51:39mcuelenaerewhen clicking on it*
13:51:57gevaertsI'd use "virtual keypad" or "keypad emulation" for grid
13:53:02pixelmaI wouldn't understand that, I only understood now, with the explanation of the difference to "emulated buttons" what that means :\
13:53:54kugelnormal mode alone doesn't mean anything
13:53:57xavieranhow's it going njr ?
13:54:08kugelexcept that it's normal?
13:54:16gevaertswell, as a first time user I'd understand "grid mode" to mean "something that returns coordinates", i.e. the thing it isn't
13:54:50 Join Jaykay [0] (
13:54:50pixelmavirtual keypad might work if you see symbols on the screen (maybe around the border)
13:54:52kugelI think you suck at simulating a first time user :p
13:55:44 Join dash32 [0] (
13:55:59Jaykaykugel: FS #5886 may also be a duplicate, it also changes the function of the rec button
13:56:19gevaertsall I want to say is that "grid mode" is unclear and ambiguous as well
13:56:27pixelmaalthough, maybe people would think they had to touch that directly
13:56:53kugelJaykay: I was about to close it, but I realized it also does it in menus
13:56:56xavieranWill there be support to for example scroll down in the text editor with the touchscreen?
13:57:23mcuelenaeregevaerts: that's what the manual is for ;)
13:57:37kugelgevaerts: I disagree, and nobody ever questioned the name grid mode so far?
13:57:41mcuelenaerexavieran: if someone implements it, yes
13:57:58gevaertsmcuelenaere: then why are we having this discussion? ;)
13:58:20mcuelenaerebecause users don't RTFM
13:58:38pixelmais there even a manual for one touchscreen target? ;)
13:58:58mcuelenaerenot atm I think :)
13:59:05Jaykaykugel: does it have any chance to be committed?
13:59:10kugelbecause some people think that absolute point isn't a good name nevertheless
13:59:14*gevaerts also thinks that it's reasonable to read "absolute point mode" as "the same point always means the same thing", i.e. what we call "grid mode"
14:00:21gevaertskugel: all I'm saying is that some people will get it wrong from those names, not necessarily all of them
14:00:31pixelmabut it doesn't - what if you scrolled down a bit?
14:01:11mcuelenaereanother problem is that most users I've seen don't even know of the existence of the option
14:01:19mcuelenaerethey just assume Rockbox only has grid mode
14:01:26gevaertsthat too
14:01:43mcuelenaereand mentioning it in the manual won't solve that I think
14:01:49pixelma"absolute point" just sounds like something scientific to me
14:02:10kugelyea, like -273,15°C :p
14:02:17gevaertsI really think that making grid mode the default is intentionally crippling the software for users who really shouldn't use sa touchscreen target in the first place
14:02:20 Quit njr (Read error: 104 (Connection reset by peer))
14:02:29*mcuelenaere agrees with gevaerts
14:02:51*kugel too
14:03:48 Quit dash32 (Remote closed the connection)
14:04:42pixelmapeople who want to use their e.g. D2 in the car wouldn't know about the "grid mode" option either if the other was the default... just saying
14:05:02mcuelenaereI think people who want to use their D2 in the car will read the manual
14:05:25 Join dash32 [0] (
14:05:32gevaertsI'm not convinced that people who use a touchscreen target in a car *want* grid mode
14:05:52kugelhow many navigation systems with touch screen and grid mode are there?
14:05:59mcuelenaeretrue, a bigger font could solve it too
14:06:15pixelmareally? Will a "RTFM" splash appear if they connect to aux-in or so? </sarcastic mode>
14:06:21*mcuelenaere already uses a bigger-than-normal font on his Onda VX747 and really wants multifont..
14:06:40kugelalso, you need more clicks in grid mode, which is also bad in the car if you can do things with a single click
14:07:25pixelmabut you could use it without even looking at the screen
14:07:31 Quit darkless ("Leaving")
14:07:32 Quit BdN3504 ("CGI:IRC (EOF)")
14:07:43mcuelenaerepixelma: no :) but I consider using a DAP in the car as a 'special usage case' and would read the manual looking for any special options for that
14:08:35 Quit dash32 (Remote closed the connection)
14:08:38gevaertspixelma: you have to look at the device anway
14:08:59xavieranI'm thinking, if someone can't be bothered to read the manual, then they won't find out about the option, or any of the other awesome things which are specified in the manual but not necessarily in the interface...
14:09:08gevaertsyou can't really feel your way until you find the right button, because you'd touch the screen while doing that
14:09:49kugelI think it's nearly impossible to use the grid mode accurately without looking at the screen *at all*, it's easy to accidentally touch the wrong button
14:09:56pixelmayou feel the corners of the screen, no?
14:10:14gevaertsis that good enough? You must have much more finger-control than I do then...
14:10:59mcuelenaerepixelma: feeling the corners of the screen isn't enough to handle it using grid mode
14:11:12 Quit Thundercloud (Remote closed the connection)
14:11:27*mcuelenaere thinks a visual indication of the corners would be nice
14:11:36pixelmaok, I would have to try
14:11:53mcuelenaereeh I mean the grid buttons
14:11:57kugelI think even peoples in cars are expecting the absolute mode, that's what any other car gadgets with touchscreen also use
14:12:15gevaertsyes, with a big-button theme
14:18:32*pixelma wonders if there could be a setup mode in RBUtil for first time users and if that would be something for Mr. Someone
14:19:59pixelmamcuelenaere: couldn't that be a part of the backdrop or what do you mean?
14:21:21mcuelenaerepixelma: if it would be a part of the backdrop, you couldn't see it all the time
14:22:35mcuelenaereI mean something like a raster that dissapears when the user hasn't pressed the screen in 2 seconds or so
14:22:50mcuelenaerebut it probably would annoy the user more than it would help
14:32:42 Join stripwax [0] (
14:43:38 Quit kugel (Read error: 110 (Connection timed out))
14:49:31 Quit stoffel (Remote closed the connection)
15:02:03 Join teru [0] (
15:03:14 Join mt [0] (n=MTee@rockbox/developer/mt)
15:04:50 Quit antil33t (Read error: 104 (Connection reset by peer))
15:05:04 Join antil33t [0] (n=Mudkips@
15:07:31***Saving seen data "./dancer.seen"
15:19:19 Join To [0] (
15:27:54 Quit jon-kha (Read error: 131 (Connection reset by peer))
15:30:37 Quit _lifeless (Read error: 60 (Operation timed out))
15:30:40 Join jon-kha [0] (
15:34:40 Join kugel [0] (n=kugel@rockbox/developer/kugel)
15:43:44 Join AndyIL [0] (i=AndyI@
15:46:10 Join theli_ua [0] (
15:47:10bertrikyay, can now read the yp-s3 buttons without using arbitrary delays
15:48:26*gevaerts thinks that ccache is useless on build clients
15:49:13Shaidif the client has previously build each target it probably wouldn't be
15:49:30gevaertsallocations are pretty stable
15:49:36 Quit stripwax ("")
15:52:11gevaertsI've now removed ccache from three (or four, depending on how you count) of my clients: rb*, monster, and hal. rb* is at least 30% faster now (670 p/s instead of ~500), monster is about the same (15 instead of 16, but not enough rounds to be really sure), and hal is also about the same (stays at 41 on average, although the latest build was slightly faster (43))
15:52:33amiconngevaerts: Any reason why the png viewer is colour only?
15:52:39*pixelma wonders about the coldfire binsize jumps by pyros-roolki
15:52:55pixelmarather roolku
15:53:01gevaertsamiconn: only that nobody has done the greylib work yet
15:53:17 Join suprleg [0] (
15:54:02gevaertsI thought it shouldn't be held back by that, and that it would be more likely that someone would step up and do it if it was in svn
15:54:33pixelmahmm... didn't have the table once gcc info on hovering over the cells?
15:54:58*pixelma tries to push the "have" to the right position
15:55:00 Quit AndyI (Read error: 110 (Connection timed out))
15:55:15 Part suprleg
15:57:20*kugel agrees with gevaerts
15:59:12 Join Strath [0] (
16:02:27 Join Strath_ [0] (
16:03:21 Join stoffel [0] (
16:05:42 Join _lifeless [0] (n=lifeless@
16:13:29pixelmait's fun to see though that the manual would already be prepared for the pngviewer on greyscale/greylib targets
16:14:13 Part theli_ua
16:16:20 Join efyx_ [0] (
16:18:17kugelgevaerts: any idea why jpeg and png isn't a single plugin?
16:19:03 Quit Strath (Read error: 113 (No route to host))
16:19:21 Nick Strath_ is now known as Strath (
16:22:44pixelmaI guess for the same reasons (not done yet) and I also guess it's not that easy, also - jpegviewer is for all bitmap targets, the pngviewer isn't yet which doesn't make things easier
16:23:05amiconnkugel: (1) jpeg supports greyscale (in fact it was first implemented as greyscale only, using the then-new graylib, on archos), png doesn't. (2) The zoom stratregy is quite a bit different, in that jpeg can be scaled down while decoding. (3) A combined plugin would be too big for the plugin buffer on lowmem
16:23:50kugelfor 3, couldn't the decoder loaded dynamically?
16:23:59kugelwe only need 1 at a time
16:24:24amiconnOne viewr per format does exactly that, without the need for a custom loader
16:25:04kugelhm yea, but as of now the menu & diashow code is duplicated, isn't it?
16:25:28amiconn(3) could be worked around by making it an overlay, but that would still waste some precious ram on lowmem targets (the unused decoder and the wasted plugin buffer)
16:26:01amiconnWell, the code could be unified and used for both plugins - *if* it makes sense considering the different zoom strategy
16:34:34kugelamiconn: we can't view jpgs along with pngs (if they were in the same folder) though
16:36:02PSPdemonalthough you may have already had that mcuelenaere
16:36:18kugelamiconn: also, how different is the zoom strategy?
16:36:56mcuelenaere PSPdemon: is this of an Onda VX797HD?
16:37:03PSPdemonsadly no
16:37:07PSPdemonits of the...
16:38:49PSPdemonim looking for alot of tear downs and info on the chips...but not coming up with alot
16:39:28PSPdemoni was looking for a Onda VX797HD teardown... no luck yet
16:39:47mcuelenaerePSPdemon: do you have a fw update of the VX797HD?
16:40:01PSPdemonprobably can get one
16:40:13PSPdemon( i still wait for my onda.... )
16:41:44PSPdemonand ill get the fw update for it
16:42:41mcuelenaereisn't there any other information about this chip on the net?
16:43:02PSPdemonnot that i can find....
16:43:22PSPdemonfirmware for the unit
16:45:20 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
16:45:47PSPdemonsupposedly 2.0 lets you load your own code...meaning that there has to be SOME sort of documentation somewhere
16:50:32mcuelenaerePSPdemon: does that mean there's an SDK?
16:51:07PSPdemoncould.....still looking into it
16:51:42PSPdemoni honestly got the unit as a 7" ebook reader... but i saw chinachip devices loading rockbox so i figured id also get it and tear it down
16:51:58PSPdemonand look for more info on chinachip cc1600
16:52:13PSPdemonmy guess is there might be...but im still unable to find anything on SDK wise
16:52:53PSPdemoninterestingly enough the firmware updater tool puts the device in bootloader mode
16:53:33PSPdemonoh wait my bad
16:53:34PSPdemonYou need to use the burning tool which comes with the same download, and you need to hold down some of the buttons on the player before you plug in USB to computer. This puts the player into a bootload device mode, whereby windows recognises it as a different device and tries to install drivers. These are in the package you just downloaded.
16:54:53mcuelenaeresounds exactly as it's done on Ingenic chips
16:55:19mcuelenaerethe firmware update seems to be one of Chinachip binary files
16:57:26PSPdemontheres the PDF reader homebrew
16:57:39LambdaCalculus37Someone want to look at FS #10431 for me?
16:58:39mcuelenaerehmm it definitely isn't MIPS though, probably ARM
16:58:50amiconn'make reconf' doesn't understand advanced build options :)
17:00:16mcuelenaereyep, it's ARM so this is a completely other SoC
17:01:26 Quit teru ("Quit")
17:03:21 Quit kugel (Nick collision from services.)
17:03:29 Join kugel [0] (
17:04:47 Join stripwax [0] (
17:07:32***Saving seen data "./dancer.seen"
17:11:19amiconngevaerts: Regarding ccache and build speed - is this on a box with several clients running in parallel? If so, it might be caused by the concurrent disk access
17:12:42gevaertsamiconn: I think it was slower even with a single client, but this is an 8-core box so there are lots of compilations running in parallel
17:14:02 Join barrywardell [0] (
17:26:23kugelcan someone finally repear the CIA bot? :(
17:28:02 Quit stripwax ("")
17:29:36 Join stripwax [0] (
17:35:14Mikachui re-enabled the polling option, let's see what happens
17:35:30 Quit bertrik ("reboot")
17:36:45 Join kugel_ [0] (
17:36:53 Quit kugel (Nick collision from services.)
17:36:59 Nick kugel_ is now known as kugel (
17:37:49 Join bertrik [0] (
17:39:22kugelMikachu: someone mentioned a url changed
17:39:49Mikachuthat's for someone with access to the hook script in the svn server to change
17:40:27kugelah, and polling doesn't need this change?
17:40:44Mikachuno, you set that on the cia page
17:42:09CIA-6New commit by kkurbjun (r22095): M:Robe 500/M66591: Add support for full-speed USB transfers, and fix the UART interrupt clearing.
17:42:24kugelah nice :)
17:43:08Mikachuit was set to every 5 minutes, i'm not sure if that is too high or low
17:46:49 Join GeekShado_ [0] (n=Antoine@
17:51:17 Join webguest78 [0] (n=48cf1522@gateway/web/cgi-irc/
17:51:39 Quit GeekShadow (Read error: 60 (Operation timed out))
17:51:40 Quit webguest78 (Client Quit)
17:55:07 Join KBH [0] (
17:58:29 Join Jaykay_ [0] (
17:59:58 Join yo [0] (
18:00:09 Nick yo is now known as Guest29757 (
18:00:11Guest29757is there an admin in here? i need help
18:00:30Mikachuadmin for what?
18:01:15 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
18:01:23Guest29757i installed rockbox just now and i didnt have any button function so i uninstalled
18:01:59Guest29757but now it tells me to press the menu and select button to reboot then press play and select to go into disk mode but because of rockbox i have no button function
18:02:06Guest29757so its basically bricked
18:02:14Guest29757i want to see if anyone can help
18:02:44Guest29757can you?
18:02:45Mikachuokay, turn hold on and off, then hold menu+select until it reboots, then quickly press select+play
18:03:19Guest29757it doesnt reboot
18:03:37pixelmawhich player do you have exactly?
18:03:39Guest29757i even unplugged the batt so it reboots and try to press select+play but it doesnt work
18:03:44Guest29757ipod video 5g 30gig
18:04:16 Nick Guest29757 is now known as houston (
18:04:44houstonand what is funny is that when its on hold and i try to press buttons it recognizes it but when its not on hold it doesnt
18:04:58pixelmaand for how long do you hold the button combo? I saw reports of it needing up to a minute of steady holding
18:05:25houstonive held it for that long should i try again?
18:05:45Shaidit's also somethings fiddly to get it to recognise
18:06:15ShaidI find I need to press menu before I press select
18:07:09houstonk imma try holdin for a minute then menu before select
18:07:43CIA-6New commit by mcuelenaere (r22096): Ingenic Jz4740 timer driver: fix stupid mistake
18:07:43CIA-6New commit by mcuelenaere (r22097): Doom: prevent division by zero
18:08:15houstonno restart after a minute
18:08:30gevaertswhat is the hold switch set to?
18:08:37houstonno restart after menu then select
18:08:39houstonhold is off
18:08:42 Quit GeekShado_ (Read error: 113 (No route to host))
18:08:51Shaidheey, it's building
18:09:09gevaertsAlso, are you sure that you didn't accidentally disconnect something while disconnecting the battery?
18:09:23houstonno nothing is disconnect problem was before that
18:09:35houstoni have to press the select button to turn ipod on
18:09:37houstonand it works
18:09:51houstonbut as soon as it comes on all button functions cease
18:10:03gevaertswell, the reboot on menu+select is implemented in hardware. There's nothing software like rockbox can do to make that stop working
18:10:27Shaidturn it on
18:10:35Shaidand just after the apple logo appears, turn hold on
18:11:09 Quit HBK (Read error: 110 (Connection timed out))
18:12:33houstoni just tried what you said shaid but nothing
18:12:48houstonafter i turn hold on do i press anything?
18:12:57 Join kushalone [0] (n=kushal@
18:13:07CIA-6New commit by mcuelenaere (r22098): Doom: re-add the missing modulus operation I removed in r22097
18:13:40Shaidno. turning hold on just as the ipod shows the apple (before the backlight kicks in) should cause it to boot into the original firmware
18:14:18houstonill try again
18:14:23 Quit kugel (Read error: 110 (Connection timed out))
18:14:36gevaertshouston: have you tried turning hold on and off again befoee menu+select? I think that's part of the official procedure
18:15:02houstonyes but to no avail
18:15:36gevaertsis the ipod flat on a table? I find that I often can't reset it unless it is
18:16:02 Quit Jaykay (Read error: 110 (Connection timed out))
18:16:48houstonyes ipod is on flat table
18:16:56houstonSHAID YOURE THE MAN
18:17:15houstonturn it on
18:17:15houston<Shaid> and just after the apple logo appears, turn hold on; that worked
18:17:30houstonbut now how do i prevent it from trying to boot rockbox again?
18:17:44houstonor can someone help me install rockbox that has full functionallity
18:18:09ShaidDo the buttons work correctly in the original firmware?
18:18:21houstonwhat is crazy is that while running original firmware buttons all work right
18:18:42houstonbut in rockbox they have no function at all
18:19:09Shaiddid you install using rbutil, or manually?
18:19:57houstonand when i unistalled due to the buttons misfunctioning i also used rbutil
18:20:25 Quit stoffel (Remote closed the connection)
18:20:42 Nick fxb__ is now known as fxb (
18:20:47houstonthanks all for your help by the way
18:20:56houstoni was really sad because im going on a flight tuesday
18:22:16gevaertsdid you also "Remove the bootloader" in rbutil?
18:22:28gevaertsIf you didn't, that's why it's still there
18:23:05bubsyany news on the HDD filesystem loader on the Creative ZEN Vision:M port?
18:23:19mcuelenaerebubsy: nothing new
18:23:31bubsyok :-/
18:26:29 Join lucas99801 [0] (
18:29:46 Join kugel [0] (n=kugel@rockbox/developer/kugel)
18:33:27 Quit antil33t (Read error: 104 (Connection reset by peer))
18:33:41 Join antil33t [0] (n=Mudkips@
18:35:27 Nick Jaykay_ is now known as Jaykay (
18:45:14 Join pike [0] (n=pike@xbmc/gc/pike)
18:46:10pikehi guys, I recently acquired an old Sansa c200 mp3player, what's the largest memcard this can use ? it's a v1 according to sysinfo. Does rockbox help to remove any size limitations or is it a hardware limitation ?
18:47:48gevaertswith rockbox, all microSD and microSDHC cards should work
18:48:02pixelmawith Rockbox
18:48:09pikeupto 16GB ?
18:48:18*pixelma too slow
18:48:52pikeI realize all included upto 16, just making sure
18:48:58gevaertsup to 32GB I think, but I don't think those are being sold yet
18:49:20gevaertsbut yes, people have used 16GB cards
18:49:33pikecool, thx for info
18:54:33houstoni need help once again. scrollwheel is working but select button isnt. it works on orignal firmware but not rockbox. I think i have a 5.5g ipod video as it has the search feature. its 30 gigs. please help
19:07:36***Saving seen data "./dancer.seen"
19:09:02 Quit houston ()
19:10:39 Quit kushalone (Client Quit)
19:15:33 Quit kugel (Read error: 110 (Connection timed out))
19:17:30CIA-6New commit by bertrik (r22099): Samsung YP-S3: implement headphones detection
19:19:29 Join Rob2223 [0] (
19:21:01 Quit pixelma (Nick collision from services.)
19:21:03 Join pixelma_ [0] (i=quassel@rockbox/staff/pixelma)
19:21:04 Join kugel [0] (n=kugel@rockbox/developer/kugel)
19:21:21 Nick pixelma_ is now known as pixelma (i=quassel@rockbox/staff/pixelma)
19:25:33 Join Thundercloud [0] (
19:32:43CIA-6New commit by bertrik (r22100): Samsung YP-S3: implement button lights
19:35:38 Quit Rob2222 (Read error: 110 (Connection timed out))
19:36:23bertrikamiconn, I vaguely something about a weird backlight controller that could be programmed to a certain brightness by sending it a pulse (or pulse sequence)
19:36:45bertrikdoes this ring a bell for you (I thought you mentioned it a long time ago)
19:37:14kugelbertrik: hey, backlight_off() by setting brightness to MIN_BRIGHTNESS_SETTING is wrong..
19:37:24bertrikI'm seeing some weird code in the yp-s3 OF that seems to send a variable number of pulses to the backlight
19:37:54bertrikkugel, makes perfect sense to me
19:38:19kugelMIN_BRIGHTNESS_LEVEL is the lowest brightness level, not "no brightness"
19:38:21bertrikwhy would it be wrong?
19:39:29Unhelpfulamiconn: if you wouldn't mind looking at FS #4817, i've down some reworking there and markun had said you'd had an objection to such a merge previously.
19:40:24bertrikkugel, I guess the brightness level curve in the yp-s3 is wrong then too because it contains brightness 0
19:40:51kugelseems so, yes
19:42:04kugelbertrik: seems you've done it this way on the meizus as well
19:42:53kugelit's almost the same driver anyway, apparently?
19:46:33 Join fdinel [0] (
19:48:01amiconnbertrik: iPod Video/Nano... I implemented the driver after RE'ing
19:49:16bertrikamiconn, ok thanks I'll have a look. PWM seemed to work, but not very reliably.
19:49:55amiconnMIN_BRIGHTNESS_SETTING should be the lowest non-zero brightness. Backlight on/off is separate functionality
19:50:24kugelthat's what I meant
19:52:26CIA-6New commit by lowlight (r22101): Philips SA9200: initialize i2c, i2s, and as3514 as the OF does. Fixes clicks/pops heard when reading the adc.
19:52:46 Join CaptainKwel [0] (
19:52:57amiconnHow this is implemented internally depends on the target of course; but backlight_off() followed by backlight_on() should return to the previous brightness, and the light should never be completely off when backlight_on() is in effect
19:55:26 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
19:55:42bertrikok that's a choice that wasn't immediately obvious to me
19:57:25amiconnHmm. More CPU seconds spend on cancelled builds than on finished builds can't be good....
19:57:40*kugel isn't sure if a _backlight_init() should do a _backlight_on() as well
19:58:22amiconnNot in general; it depends on the target
19:58:41kugelan explicit call to _backlight_on() in this case
19:59:01amiconnOn ipod 1st and 2nd Gen it would be a very bad idea to enable the backlight on boot
19:59:24amiconnIt all depends on what is the best behaviour on boot
20:00:30kugelIMO, even if the best behavior is to have backlight on, then the backlight_on() should be seperate
20:02:28amiconnThere is probably no single best method for all targets. Don't forget that most of these inits are also used in the bootloaders, and that the hardware comes up in different states depending on target
20:02:49amiconn(e.g. the archoses come up with backlight on by default)
20:03:41amiconnCorrection: this does apply to the Player and the Ondios (when modded)
20:03:55 Nick hd is now known as HellDragon (i=jd@Wikipedia/HellDragon)
20:04:14kugelI don't see how that affects my opinion
20:05:30amiconn_backlight_on() switches the hardware, and is meant for use by the backlight driver and the backlight thread *only*. backlight_on() signals the backlight thread, which does not exist in most bootloaders
20:06:01kugelyea, I forgot the _, I meant _backlight_on()
20:06:24 Quit lucas99801 ("Leaving")
20:07:00gevaertsamiconn: not all CPU seconds are equal
20:08:54kugelamiconn: IMO, _backlight_init() should only make _backlight_on() working, not do a _backlight_on(). I don't see how the initial state of the hardware counts in here
20:09:18amiconnWhy does that matter at all?
20:10:43Unhelpfulgevaerts: even so, ideally every host will be working during the entire build cycle on non-cancelled work. less ideally, the work is broken up for us into chunk, and the sizes of those chunks aren't even fixed or nicely predictable. :/
20:11:13gevaertsUnhelpful: yes, but that's the points/second numbers. CPU seconds are not very helpful
20:11:29 Join AsaelReiter [0] (n=d44c6395@gateway/web/cgi-irc/
20:12:35 Join lucas99801 [0] (
20:12:40*kugel waits for his uber-client to be back :p
20:12:44Unhelpfulgevaerts: right, we have the points to make it more clear which CPUs' seconds have the most value... but all time on cancelled builds is wasted work. have you and zagor not solved knapsack yet? ;)
20:16:25PSPdemonwish i could take one right about now -_-
20:17:02gevaertsUnhelpful: no. We're counting on the number of clients growing to match the number of builds :)
20:17:08PSPdemonmcuelenaere, what version arm is chinachip?
20:17:15PSPdemonwell the cc1600
20:17:37Unhelpfulgevaerts: that will only make the wasted time problem worse :/
20:17:41mcuelenaerePSPdemon: I don't know, I haven't looked very closely at the disassembly
20:17:57bertrikI'm now relying on global variable backlight_brightness to remember the previous backlight brightness, this is correct right?
20:18:09gevaertsUnhelpful: yes, but the round should be fast
20:19:15PSPdemonARM Cortex A8
20:19:43Unhelpfulgevaerts: maybe we should just be using distcc ;)
20:19:52gevaertsUnhelpful: slow networks...
20:19:56PSPdemonopps wrong model ( looking at the wrong thing )
20:20:26Unhelpfultrue. and needing a super-machine to be the build host since it needs to do all preprocessing locally.
20:21:19 Quit FlynDice (Remote closed the connection)
20:21:43Bagderwe tried distcc
20:21:52Bagderit really doesn't help with these slow networks
20:22:04Bagderand we have fairly small C files
20:23:08Unhelpfulit's a shame, as it *would* let you break up the work more evenly
20:23:36gevaertsBagder: I think it's worth experimenting with −−no-ccache on multicore clients
20:27:12amiconnBagder, gevaerts: Since not all CPU seconds are equal, perhaps listing the number of wasted CPU points and their percentage would be more useful
20:27:25 Quit stripwax ("")
20:27:33Bagderwould indeed
20:27:47gevaertsamiconn: actually, there is the total client speed and the effective round speed
20:28:10gevaertsI think the "45% efficient" line at the end is what you
20:28:13 Quit lucas99801 ("Leaving")
20:28:14amiconnHmm, it is already shown
20:28:14gevaerts're looking for
20:29:29Bagderisn't that just the "optmimum time" vs the "time it took" ?
20:29:56CIA-6New commit by bertrik (r22102): Meizu: update backlight brightness curve so backlight is still on at the minimum brightness setting
20:30:07gevaertsyes, but that translates pretty directly to wasted cycles
20:30:32 Join Zagor [242] (
20:30:48*gevaerts wants to know why some of todays build rounds were so slow
20:31:12Shaidgevaerts: I can disable ccache on mine if you'd like, it's a quad core.
20:31:27Shaidthough it just got told to build about 8 targets at once
20:32:02gevaertsShaid: has it done enough builds to get a clear idea of its "normal" speed?
20:32:10Shaidprobably not
20:32:16Shaidstill doing the raid scan too
20:33:50Shaidis there a way to reduce the number of targets the client gets told to build at once?
20:34:54amiconnHmm, still no progress display :\
20:34:55ZagorShaid: which client is this?
20:36:03pikegevaerts do you know if Sansa c240 *all* includes SDHC ?
20:36:58gevaertspike: can you rephrase that? I have no idea what you mean
20:37:44 Quit AsaelReiter ("CGI:IRC")
20:38:30pikegevaerts do you know if this card would work with Sansa c240 ? (there's different types or classes of micro cards it seems)
20:38:44pikebut Im not sure if the older type of cards even exist at 16GB size
20:39:10gevaertspike: ah. All SDHC cards should work
20:39:21pikethanks for help
20:39:22Zagormajor server bugginess in the last few rounds
20:39:33bertrikpike, rockbox supports SDHC but we can't give absolute guarantees for each individual card
20:39:50ShaidZagor: the rockbox buildclient
20:39:50pikebertrik no compatible list somewhere ?
20:40:00ShaidMine got asked to build about 10 different targets at once
20:40:15bertrikI would tell you if I knew of one
20:40:29ZagorShaid: yeah, everyone did. for some reason it barfed out all builds at once.
20:40:52ShaidI've not seen a load average of 49.7 before
20:40:56bertrikpike, the original firmware might get confused by SDHC cards and take a really long time to boot or even fail to boot
20:42:12pikeyeah I noticed if the unit is off and I plug it into USB, the original fw loads. wonder what will happen if this card is inserted then
20:42:57 Quit Thundercloud (Remote closed the connection)
20:48:25 Quit kugel (Read error: 110 (Connection timed out))
20:49:16PSPdemonits a ARM9
20:49:52mcuelenaerethat's a broad range of processors
20:51:04PSPdemonHua Fei CC1600 core ARM9 processor used as a host processor, if a good guess, then, CC1600 also used in Korea Chips & media companies RMVB IP core, IP core
20:51:15 Join FlynDice [0] (
20:51:27PSPdemonbad translation by google
20:51:39PSPdemon( everything is broken engrish )
20:54:50CIA-6New commit by bertrik (r22103): Samsung YP-S3: update backlight brightness curve so backlight is still on at the minimum brightness setting
20:54:53 Join saratoga [0] (i=9803c6dd@gateway/web/freenode/x-muqmjzyokhdauapt)
20:57:53PSPdemondoes it help that its MT7506
21:01:37 Quit advcomp2019 (Read error: 104 (Connection reset by peer))
21:03:40PSPdemondont know if it helps ( trying my best to find all i can )
21:06:08PSPdemonso far all i can figure is that its a ARM9 32B-RISC
21:07:40***Saving seen data "./dancer.seen"
21:08:06 Quit bmbl ("Bye!")
21:08:35 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
21:10:54saratogaall ARM9 chips are 32 bit and RISC
21:17:06mcuelenaereIsn't id3_is_genre_string() in apps/metadata/mp3.c at line 98 invalid? It determines whether a string is in an array of string pointers by comparing it with the first and last pointer of the array..
21:18:00Mikachudepends on if the string it compares always comes from that array or not
21:18:18Mikachuor how the array is allocated, i mean
21:18:48mcuelenaereMikachu: the point of the function is to determine whether the given string comes from the array or not
21:18:56Mikachuthe genres array is defined a few lines up
21:19:02mcuelenaerethe array is static const char* const genres[]
21:19:33Mikachushould work, i think
21:19:57mcuelenaereanyhow, that crashes on my target with an invalid genre_string value for an id3 struct (has to do with how adjust_mp3entry() handles it)
21:20:56mcuelenaereMikachu: can't the values of genres[] be pointers allocated everywhere not close to each other? (meaning id3_is_genre_string() isn't correct)
21:21:14mcuelenaerenot next to each other*
21:21:28mcuelenaereie the last value could be physically before the first one
21:22:04*mcuelenaere doesn't want to solve this with a for() check though
21:22:26Mikachuyou're probably right
21:22:48Mikachusince it's an array of char*, not a double array
21:23:32mcuelenaerea double array?
21:23:53*mcuelenaere wonders why this didn't data abort on ARM
21:23:57Mikachui think i just made that up :P
21:24:08mcuelenaereyou mean a multidimensional array? :)
21:24:09Mikachui mean if it was a genres[][] then all the strings would be there in ram
21:24:23Mikachubut you can't do that, they're not the same length, and the syntax is wrong
21:24:29saratogai don't really understand how that function is expected to work
21:24:48Mikachuit will probably usually work, because the strings will be allocated in the order they are written, next to eachother
21:24:51mcuelenaeresaratoga: neither do I
21:25:09Mikachu(i'm not an expert on this kind of stuff though)
21:25:15 Join einhirn [0] (
21:25:25mcuelenaereis there an equivalent to static const char[] const genres[] ?
21:27:12Mikachuthe strings would have to be padded then, or you can't index by string
21:27:34Mikachuis nicolasp still here?
21:27:37mcuelenaerewhich means binsize increase :(
21:27:57mcuelenaereI think I'll just solve it with looping the whole array and checking each pointer
21:27:57Mikachuhe added the code in question in oct 2007, r15346
21:30:58Mikachuit would be nice if there was a comment explaining what adjust_mp3entry() actually does
21:32:36mcuelenaeresee copy_mp3entry()
21:33:15saratogaactually looking at it, shouldn't that always work?
21:33:23saratogasince entries in the static array have to be sequential
21:33:29mcuelenaerewhy would they?
21:33:38mcuelenaere(btw it doesn't work)
21:33:46Mikachusaratoga: it's just an array of pointers, they can point anywhere
21:34:06 Join fg57lx [0] (
21:34:16saratogayes but its checking if the entry was originally taken from that array
21:34:31Mikachumcuelenaere: i looked, but it didn't help much :)
21:34:39Mikachumcuelenaere: a struct mp3entry has pointers into itself?
21:34:52mcuelenaereit looks like it
21:36:36saratogaits basically just checking if the entry was allocated by "id3_get_num_genre" or custom defined elsewhere
21:38:13Mikachuinstead of checking if the string is in the genres[] array
21:38:18Mikachuwhy not check if the string is inside the mp3entry?
21:39:08Mikachuit is the char id3v2buf[ID3V2_BUF_SIZE] member that has all the strings?
21:41:00Mikachuif (entry->genre_string > orig && entry->genre_string < orig + sizeof(struct mp3entry)) or so
21:42:17mcuelenaereisn't that orig->id3v2buf then?
21:42:35Mikachuit's not likely to be somewhere else in the mp3entry, is it? :)
21:42:41mcuelenaereand sizeof(orig->id3v2buf) (or ID3V2_BUF_SIZE)
21:42:44*mcuelenaere doesn't know
21:42:55Mikachuthe id3v2buf is not a pointer
21:43:11Mikachu(obviously, since copy_mp3entry just does one memcpy)
21:43:39Mikachuand if it is somewhere else in the mp3entry, it will still work
21:43:51 Join Lss [0] (
21:44:28CIA-6New commit by mcuelenaere (r22104): MP3 metadata parser: ...
21:45:27*mcuelenaere spots Zagor
21:45:31Mikachucan you check if my method works too?
21:45:38*Zagor waves
21:45:52MikachuZagor: have you fixed the cia hook script?
21:46:06mcuelenaereZagor: have you seen <mmadia> if not, the post-commit script needs to be updated to use '' instead of http://cia.navi.* ?
21:46:23Zagormcuelenaere: umm, we get announcements don't we?
21:46:33MikachuZagor: i changed the bot to poll the svn server in the meantime
21:46:33mcuelenaerethat's because we switched to polling mode
21:47:30 Quit fg56lx (Read error: 110 (Connection timed out))
21:47:51Zagoraha. yes, I changed it three days ago
21:48:03Mikachuit's odd that it didn't work then, i changed to polling today
21:49:10Zagorblah, I changed it wrong... instead of fixed now.
21:49:17Mikachuokay, i'll turn off the polling
21:49:21mcuelenaereMikachu: your method works too
21:49:23Mikachuor you'll probably get double notices
21:49:26Mikachumcuelenaere: woo :)
21:52:57CIA-6New commit by mcuelenaere (r22105): Metadata parser: make adjust_mp3entry() not depend on id3_is_genre_string(), so there's no need to compare the string to the whole genre array (thanks ...
21:53:04*mcuelenaere predicts a small binsize decrease
21:55:31 Quit LambdaCalculus37 ("Fwump")
21:57:45Mikachuah, and obviously the hook works fine too
21:58:11Mikachumcuelenaere: you were right :)
22:16:48mcuelenaereheh, /me spots a firmware->apps call :)
22:17:45*gevaerts gets torches and pitchforks
22:20:53Zagormcuelenaere: where?
22:21:16mcuelenaerefirmware/sound.c calls dsp_callback() at l.235
22:21:37Zagorcallbacks are ok
22:22:10Zagorit's the whole point of them :)
22:22:18Mikachushouldn't it have to be set via some firmware_sound_set_dsp_callback(dsp_callback); or so?
22:22:21Mikachu(from apps)
22:22:33Zagoryes. isn't it?
22:22:39Mikachui don't know
22:22:59*Mikachu is watching tv4
22:23:22Mikachuno room for code windows :)
22:23:49mcuelenaereMikachu: you mean apps/dsp.c should set a pointer in firmware/sound.c with a callback to its apps/ function?
22:24:07Mikachui think so
22:24:18mcuelenaereah, well it isn't the case then :)
22:24:36Zagormcuelenaere: yes, callbacks must be registered to be a callback. otherwise it's just a call to a function named *callback :-)
22:25:07Zagorjudging by the name, it might be a "I'll do it properly later" thing :)
22:25:27 Quit togetic (Read error: 104 (Connection reset by peer))
22:26:44mcuelenaereZagor: I'm not sure what you mean, but sound.c does a 'hard' function call to an apps/ function which sets some variables; I don't see any callbacks there :)
22:27:25 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
22:27:30ZagorI just mean the name dsp_callback() implies someone at least had the idea of making it a real callback
22:28:30mcuelenaereor they just used it to disguise a dsp_set_value() ;)
22:28:38mcuelenaere(as it looks like dsp_callback() does that)
22:29:42amiconnsound.c is indeed a bit tricky. It is firmware layer because it needs to deal with the interdependencies of the various sound settings, but it also needs to call dsp functions for software treble/bass if the audio codec doesn't have that in hardware, and some other things
22:30:25mcuelenaereamiconn: do you know of a way to do software volume control? (and if possible, only in a -73 -> 0 dB range or so)
22:31:20*mcuelenaere tried converting the PCM data on-the-fly but ran into a too slow processor problem (or too inefficient code)
22:31:40Zagoramiconn: the proper way is to simply register the dsp functions it needs to use
22:32:04saratogamcuelenaere: how were you adjusting volume?
22:32:19mcuelenaeresaratoga: just dividing every sample by a value
22:32:35saratogawell multiplying by a faction would be faster :)
22:32:43Mikachuwhy is dsp in apps/?
22:33:12mcuelenaeresaratoga: multiplying with a fraction without floating point?
22:33:24saratogamcuelenaere: fixed point
22:33:32*mcuelenaere is confused
22:33:40saratogagrab one of our fixed point mul functions from somewhere and use that
22:33:50ZagorMikachu: I don't remember :)
22:33:55mcuelenaeresaratoga: you can't do val *= 1/2; with an int, can you? or do you mean something else?
22:34:00Mikachuheh, okay
22:34:07saratogathey'll basically just do a 32x32->64 mul, then shift the result down to 32 bits again
22:34:36saratogamcuelenaere: you can do that with ints, it just takes at least 3 instructions on most CPUs
22:35:22LambdaCalculus37lowlight: (for the logs) I remember you telling me that you modified the keymap defines for the SA9200, but I forgot what the modifications were. I need to fix up the final version of my plugin keymap patch.
22:35:29 Quit LambdaCalculus37 ("Fwump")
22:36:16mcuelenaeresaratoga: is FRACMUL_SHL (or FRACMUL) in fracmul.h one of those fixed point mul functions?
22:36:35saratogayou're basically just defining some value, say 65536 to be 1, then 32768 becomes 1/2, etc, so to mulitply by 1/2 you would do (x*32768)>>16 which gives you 1/2 the original value (or any other fraction)
22:36:43saratogamcuelenaere: i'm not sure i've only used the ones in codecs
22:36:51saratogabut i'm sure dsp.c has the right functions
22:39:14saratogaour output samples signed?
22:39:16 Quit fg57lx (Read error: 110 (Connection timed out))
22:39:33mcuelenaeresaratoga: still it would be too slow, I tried doing a simple *ptr++ = *ptr; and that gave me stuttering audio
22:39:46saratogais this in an interrupt or something?
22:39:58mcuelenaeresaratoga: I think so, when converting it to unsigned it sounded awefull
22:40:01mcuelenaeresaratoga: DMA
22:40:05saratogaone fixed point mul per sample is extremely little overhead
22:40:08mcuelenaere(with 16 byte burst size)
22:40:16saratogamaybe processing needs to be done earlier in the pipeline?
22:40:43mcuelenaerewhere? that's why I thought of doing it in the DSP
22:40:43mcuelenaerebut that would mean plugins won't have software volume control (except 0 -> 6 dB)
22:40:51amiconnmcuelenaere: Imo software volume should be done in the dsp code, where stereo width, treble/bass (on some targets), eq etc are done as well
22:41:22saratogaif you've got a single cycle multiplier, scaling only costs you about 3 cycles per sample, which is just ~250kHz overhead (ignoring memory access delays)
22:41:26amiconnWhy would that affect plugins at all?
22:41:38mcuelenaereamiconn: sure, what part should I use? The apply_gain function?
22:42:19mcuelenaereamiconn: I was doing it before sending out samples to the DAC
22:42:50Mikachuisn't there already a replaygain thing that would do something similar?
22:43:02*mcuelenaere uses the equalizer precut for now
22:43:23amiconnThere is a reason why dsp effects are applied in larger chunks in a separate thread. If you apply in an isr or similar, it will have a bad effect due to cpu frequency scaling
22:43:48mcuelenaereyes, I noticed that
22:45:43 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
22:46:02 Join lucas99801 [0] (
22:48:54 Join fg56lx [0] (
22:58:04mcuelenaereZagor: ah, you're right; I hadn't seen the static int (*dsp_callback)(int, intptr_t)..
22:59:07Mikachuit is perhaps a bit unfortunate that the public function and the callback parameter has the same exact name
23:00:06Zagorslightly confusing, yes
23:03:54mcuelenaerecouldn't we place the DSP_CALLBACK_* enum in say sound.h ? That way there's no need to duplicate them in dsp.h & sound.c
23:05:05 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
23:07:16 Part lucas99801 ("Leaving")
23:07:41***Saving seen data "./dancer.seen"
23:19:09 Join einhirn [0] (
23:30:46 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
23:31:00 Join mcuelenaere_ [0] (
23:31:24 Nick mcuelenaere_ is now known as mcuelenaere (n=mcuelena@rockbox/developer/mcuelenaere)
23:38:22CIA-6New commit by mcuelenaere (r22106): * Move DSP_CALLBACK_* enum to sound.h ...
23:39:38saratogais rockbox-info.txt not being updated right on the build servers?
23:39:48saratogai downloaded one and got "Version: r21884-090715"
23:41:06Zagorsaratoga: which did you download?
23:41:39Zagorwhere did you find that link?
23:41:43Zagorit's old
23:41:56saratogaits in the wiki
23:42:03saratogadid we move around the build server stuff?
23:42:09Zagoryes we did
23:42:27Bagderoh, time to symlink and kill old files I guess
23:42:27saratogaah ok i'll fix it
23:50:22CIA-6New commit by mcuelenaere (r22107): Fix red: only use SW_VOLUME_MIN if HAVE_SW_VOLUME_CONTROL is defined
23:51:08saratogadoes the ingenic really not have any volume control hardware?
23:51:51mcuelenaeresaratoga: it only has 0 -> 6 dB
23:51:56saratogahow odd
23:51:58mcuelenaerebut it has volume control for the line in
23:52:11mcuelenaereyes, I don't understand why they did that
23:52:15saratogadoesn't that degrade quality when you adjust the volume?
23:52:39mcuelenaereyou mean when I go over 0 dB?
23:52:45saratogawell under
23:52:50bluebrotherZagor: the new build system puts the current builds in a different location than the old one, breaking Rockbox Utility :/
23:53:02saratogasince you're basically just throwing away signal but keeping noise constant
23:53:11bluebrothercan you provide compatibility symlinks for now?
23:53:48mcuelenaeresaratoga: so the software version is different from the hardware then?
23:53:51Zagorbluebrother: yes
23:54:32bluebrotherplus, something to keep the old files persistent until the new ones are around? I just got a 404
23:54:32saratogamcuelenaere: most players have a register that variest the gain on the output stage, so decreasing the volume decreases both signal and noise
23:54:42saratogaif you do it in software you decrease signal but at constant noise
23:55:00mcuelenaerehmm that's not very good, isn't it possible to decrease both signal & noise in sw?
23:55:42saratoganot without changing some hardware
23:55:46Bagderbluebrother: really? the new ones are supposed to be mv'ed to the upload place...
23:56:00Bagderuh, mv'ed from the upload to the download place I should say
23:58:07saratogai'm actually really surprised they'd make it like that
23:58:25Zagorthe server deleted the zips to avoid different versions being there at the same time. I've removed that now though.

Previous day | Next day