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 2009-06-05

00:04:52 Quit Rondom ("Ex-Chat")
00:05:04pixelmaBagder: I also have those small PNGs now for the use in those target tables (of the Clip, Fuze and m200). In case you are already interested
00:06:20 Join bubsy [0] (n=Bubsy@
00:08:38 Join AndrewRB [0] (
00:14:30 Quit perrikwp (" ajax IRC Client")
00:14:41 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
00:15:18*funman thinks plugin.h isn't always updated like it should, especially the minimum api version
00:16:38JdGordon|it usually is... or very quickly after being broken
00:21:15CIA-38New commit by pixelma (r21191): Update the Fuze images for the manual with 'scroll fwd/back' labels -patch from FS #10262 and a little rearranging by me. Also restores the SVG's text ...
00:25:41CIA-38New commit by pixelma (r21192): Commit FS #10265 - add m200 images (complete set of SVG,PNG and PDF) to the manual source. The hard work of drawing the SVG was done by Marko Pahlke.
00:26:20 Join _lifeless [0] (n=lifeless@
00:29:07 Quit matsl (Read error: 110 (Connection timed out))
00:29:41 Quit mirak (Read error: 113 (No route to host))
00:31:04 Quit bluebrother ("leaving")
00:32:04 Quit ender` (" AAAAAA: American Association Against Acronym Abuse and Ambiguity.")
00:35:58***Saving seen data "./dancer.seen"
00:39:31 Quit domonoky (Read error: 54 (Connection reset by peer))
00:42:53 Join dstaley [0] (i=440bc91c@gateway/web/ajax/
00:44:27 Quit bertrik ("Leaving")
00:44:34 Join lymeca [0] (
00:45:56 Quit lymeca (SendQ exceeded)
00:46:27 Join lymeca [0] (
00:47:52 Quit trisiak (Read error: 110 (Connection timed out))
00:48:18 Quit hd (Client Quit)
00:48:26dstaleyHey, what are the ways to convert album art to Rockbox supported versions on OS X?
00:48:53LloreanThe same way you convert anything to a jpeg or bitmap.
00:49:44dstaleyWell, when I just resize and convert in Photoshop it doesn't work. I thought my settings were wrong for the WPS, but album art downloaded with a dedicated RB tool worked fine.
00:50:09AndrewRBwhat filename are you giving the file(s)?
00:50:24gevaertswhich version of rockbox are you using?
00:51:51dstaleyAndrewRB: cover.bmp, folder.bmp
00:51:56dstaleygevaerts: 3.2
00:52:31pixelmawhat format do you convert to? Any weird bmp settings, like compression or so?
00:52:34gevaertsdstaley: does photoshop give you options when saving bmp? You don't want rle compression
00:52:42AndrewRBdstaley: could you upload one of the images to and post the link please
00:52:42evilnickAnd they're in the same folder as your music files?
00:53:01evilnickTry opening them on target first
00:53:28 Join HellDragon [0] (
00:53:37 Quit HellDragon (Read error: 104 (Connection reset by peer))
00:55:09dstaleyAndrewRB: Let me go remake one. All of them are now from the utility.
00:55:09 Nick AndrewRB is now known as AndrwRB (
00:55:23AndrwRBdstaley: sure.
00:55:27 Join HellDragon [0] (
00:55:48 Join |ahmina| [0] (
00:56:14 Nick fxb is now known as fxb__ (
00:56:21 Quit evilnick (" ajax IRC Client")
00:56:22 Nick AndrwRB is now known as AnderwRB (
00:57:19 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.10/2009042316]")
00:58:21 Join BlakeJohnson86 [0] (
01:00:07 Join __lifeless [0] (n=lifeless@
01:02:19 Quit bagawk (Read error: 60 (Operation timed out))
01:05:29 Nick AnderwRB is now known as AndrewRB (
01:08:29 Nick AndrewRB is now known as AndrewRB_torso (
01:10:27 Nick AndrewRB_torso is now known as froggyman_bodypa (
01:11:00 Nick froggyman_bodypa is now known as AndrewRB (
01:11:37Unhelpfulamiconn: yes?
01:14:04 Quit froggyman ("CGI:IRC")
01:14:21 Quit _lifeless (Read error: 110 (Connection timed out))
01:14:57 Quit |ahmina| (Success)
01:15:22 Quit lymeca (Connection timed out)
01:16:09 Join pixelma_ [50] (n=pixelma@rockbox/staff/pixelma)
01:16:09 Quit pixelma (Nick collision from services.)
01:16:28 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
01:16:53 Quit amiconn (Nick collision from services.)
01:16:55 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
01:17:17 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
01:22:42 Quit JdGordon| (" ajax IRC Client")
01:25:31 Quit Thundercloud (Remote closed the connection)
01:25:43AndrewRBdstaley: any luck?
01:27:05dstaleyAndrewRB: Sorry for the wait. I didn't have Photoshop installed, so I had to reinstall it. Here is the image.
01:27:41dstaleyIt isn't resized, but I am under the impression that Rockbox will automatically resize it?
01:28:05AndrewRBso that file, named "cover.jpg" in your album directory does not cause the image to show beside now playing info?
01:28:40 Quit dmb ("Leaving")
01:30:21 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
01:30:27 Part pondlife
01:30:39dstaleyAndrewRB: Correct.
01:31:04AndrewRBworks fine here
01:31:10pixelmaI believe support for cover.jpg came after 3.2
01:31:17AndrewRBwhat build of rockbox is yours?
01:31:27AndrewRB(not version, build)
01:31:40saratogahe said 3.2 above
01:31:56pixelmathat's what I remember too
01:32:06AndrewRByeah, but specific buidl makes a difference, right?
01:32:10saratogaso he needs to use BMP or else update to a recent build
01:32:26LloreanAndrewRB: 3.2 IS a specific build
01:32:30UnhelpfulAndrewRB: 3.2 is *one* specific build.
01:32:31LloreanLater builds aren't 3.2 any more.
01:32:40AndrewRBah, sorry
01:33:11UnhelpfulMajorChanges shows JPEG AA to be post-3.2. 3.3 will be the first *release* to include it.
01:34:38saratogaincidentally, is there a reason we have a users mailing list instead of just using the forums? the list seems essentially redudant, and it brings out the worst in people for some reason
01:34:46dstaleyUnhelpful: I see. I guess I must have been reading the manual for the later release.
01:35:25 Join JdGordon| [0] (i=441b7a70@rockbox/developer/JdGordon)
01:35:29Lloreandstaley: The most commonly available manual is for the SVN versions, not the release.
01:35:54Lloreansaratoga: I suspect many of the people on the user ML prefer mailing lists for one reason or another
01:36:12 Quit jgarvey ("Leaving")
01:36:37JdGordon|will anyone have major complaints if I commit the patch which fixes a bunch of plugins to use the new (correct) menu api.... i have time and inclination tonight to do it...
01:37:12funmanJdGordon|: i think you should bring this to the mailing list, so everyone notices it
01:37:15UnhelpfulJdGordon: getting rid of oldmenuapi from pluginlib?
01:37:16JdGordon|its not exacty a bug fix.. but its not new stuff either...
01:37:42JdGordon|FS #10094
01:37:53JdGordon|it doesnt remove the api complelty, but it removes its use
01:38:32 Join trisiak [0] (
01:38:36JdGordon|funman: well... i'd do it before .eu wakes up so not really much to discuss....
01:39:16LloreanIs it really a fix?
01:39:20LloreanI mean, is something "broken" in some way?
01:39:30LloreanOr is it just cleanup?
01:39:34 Join lee321987 [0] (
01:40:21*JdGordon| loves using ambiguous answers
01:40:32LloreanWhat bug does it fix?
01:40:33lee321987About lua script: Are the RB "copy" and "past" commands supported?
01:41:29safetydanlee321987: no, they're not really API functions anyway
01:41:44JdGordon|Llorean: depends if you call not having all plugins which appear to use the inbuilt menu, having them act the same way
01:42:13LloreanJdGordon|: That's not really a bug. Just aesthetics.
01:42:25LloreanCan't this wait until after the freeze?
01:42:48saratogai'd say put it in now, we have the feature freeze to correct it
01:43:04 Quit gevaerts (Nick collision from services.)
01:43:10LloreanThe feature freeze has already started.
01:43:13saratogacode clean up is the kind of stuff we should be doing during freezes
01:43:15 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
01:43:20LloreanThe whole point of it is *not* to make significant changes unless they're fixing specific bugs.
01:43:24JdGordon|Llorean: yes it can.. the thing is I have time and motivation to do it now... and might not in 2 weeks
01:43:31Lloreansaratoga: It's a pretty wide ranging functionality change.
01:43:32JdGordon|its nothing major
01:43:47*Unhelpful hands JdGordon git and a manual for git branch
01:43:54*JdGordon| grumbles about getting dates wrong and thinking he had a few more days
01:44:13Lloreansaratoga: The stuff we should be doing freezes is not touching *anything* unless it fixes specific bugs.
01:44:38 Quit lee321987 ("ChatZilla 0.9.84 [Firefox 3.0.10/2009042316]")
01:45:50JdGordon|does someone want to volanteer to get the lang updates in?
01:46:18JdGordon|iirc there are 4 or 5 waiting
01:46:22LloreanShould we do a call for translators like usual?
01:46:42JdGordon|yeah probably
01:47:09LloreanI haven't seen rasher around recently. Did I miss something, or anyone know where he is?
01:47:21 Join dmb [0] (n=dmb@unaffiliated/dmb)
01:48:34JdGordon|yeah, dunno
01:51:22pixelmaJdGordon|: if you want to fix something, could you have a look at the colour choser screen statusbar problem? (yes, I know I forgot to tracker it - recipe to reproduce: go to e.g. "set background" and actually set a different colour -> on return to the menu, the statusbar will at least have the old background colour, sometimes looks blanked (as if foreground colour was the same too) until the next statusbar update
01:51:33 Part AndrewRB ("Konversation terminated!")
01:51:35pixelmacan reproduce reliably on my c200
01:51:51JdGordon|tracker it! :D
01:52:12JdGordon|but yeah, ill have a looky
01:52:28pixelmajust wanted to go get some sleep...
01:52:50JdGordon|thats an odd idea...
01:53:14JdGordon|Llorean: should we think about removeing the debug menu from release builds?
01:53:47LloreanI suggested that with the 3.0 one, but it never happened.
01:53:59 Join evilnick [0] (i=620ec27e@gateway/web/ajax/
01:54:24LloreanSince we branch for the release, we can always go back and build a copy of the release with the debug menu enabled if there's something in it we *really* want to know.
01:55:42JdGordon|should be easy enough to do... just get the script which does the release bins to set a define we can check... probably can use the same thing for version string
01:56:13LloreanYeah, didn't we release 3.1 without the proper version string?
01:56:34LloreanWe should probably automate it, or at least make the release script prompt for the version number when run or something
01:58:32JdGordon|did it say 3.0?
01:59:53LloreanOne of our releases just had the SVN revision without a version number
02:03:39 Quit dmb (Read error: 104 (Connection reset by peer))
02:05:08JdGordon|ill have a look how this is all done once i get home
02:05:31JdGordon|the seems to use svnversion.. not sure what it does with that though
02:06:17funmannothing, since has been renamed to !
02:06:22 Join lymeca [0] (
02:06:40n1syes 3.1 was missing the proper version string
02:06:47JdGordon|`svnversion` is not the same as
02:07:40 Part toffe82
02:09:27 Join trisiak_ [0] (
02:10:07 Quit trisiak (Read error: 104 (Connection reset by peer))
02:11:03 Join dmb [0] (n=dmb@unaffiliated/dmb)
02:13:44 Quit n1s ("Lämnar")
02:13:51*FrankTM_ ooh's
02:13:54FrankTM_features freeze
02:14:27 Quit JdGordon| (" ajax IRC Client")
02:14:50FrankTM_bed time
02:16:57 Quit robin0800 (Remote closed the connection)
02:19:00funmanso is it too late for FS #10274 ? I'll have to sync the patch after the branching ? :/
02:23:09LloreanYes, it should probably wait.
02:23:21LloreanAt this point, if something's working for release builds it shouldn't be touched until after the branch.
02:24:19funmanThat's what freeze means after all. So it gives a little time to check last uncertainties about this patch
02:25:06LloreanAnd an added bonus - if it has bugs after commit the release is new enough that everyone can just use it until those bugs are fixed. :)
02:27:27 Quit trisiak_ (Read error: 110 (Connection timed out))
02:30:15*JdGordon isnt sure how configure handles its args...
02:32:32 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/ajax/
02:35:20 Quit funman ("leaving")
02:36:00***Saving seen data "./dancer.seen"
02:36:06 Quit efyx_ (Remote closed the connection)
02:37:55r0b-will Rockbox tell me the CPU in my MP3 player?
02:38:19evilnickYou could ask here
02:38:42evilnickOr it'd be in the wiki
02:39:15r0b-well what CPU and the approxamite speed of the CPU in the Sansa e250
02:39:19evilnickBut I don't think Rockbox (as in the builds) tells you what CPU is inside your Sansa E200 series player
02:40:31evilnick - it should be there
02:45:10r0b-i manged to get an original Doom 1 wad file from my friends computer how can i get RockDoom to see it properly
02:46:08r0b-would i just copy the .wad to Doom in the rockbox directory?
02:46:47JdGordonwhats the way to have gcc quote a #define?
02:46:58krazykityou follow the instructions on the wiki:
02:47:08UnhelpfulJdGordon: #STUFF ?
02:47:36kkurbjunJdGordon: USB is stable reading at 2.8 MB/s!
02:47:43JdGordoncomplained about stray # in the code
02:47:46JdGordonkkurbjun: AWESOME!~
02:48:07kkurbjunI need to test writing now, but I think it's working right now
02:48:30kkurbjunit doesn't appear to be corrupting the fs which is a plus :-D
02:49:06 Join hd [0] (
02:51:10JdGordonLlorean: does the word "Release" need to be translated in the version string?
02:51:24JdGordonshould it be there at all?
02:51:28r0b-i congratulate you guys for developing Rockbox it is a very nice piece of software
02:51:45LloreanShouldn't the version string just be Date-3.3-SVNRevision or something like that?
02:51:50JdGordonUnhelpful: thanks... turns out i was doing everything wrongly :p
02:51:52*Llorean seems to recall it was something like that.
02:52:19JdGordonisnt everything apart from the 3.x redundant?
02:52:36JdGordon"Version: 3.2" is probaly enough
02:52:58LloreanThe SVN revision is handy for a few things
02:53:04 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
02:53:27LloreanFor example, if we know feature X was added in r19970 if we see the release was r19963 we know it came after, without having to look up further details
02:53:46LloreanDate is probably redundant though
02:54:13LloreanThere probably should be no words in the version number string though.
03:00:07 Join trisiak [0] (
03:01:20 Join |ahmina| [0] (
03:02:50JdGordonnow... if only I could figure out how to get working
03:02:57 Quit HellDragon (Success)
03:05:33 Join arcticfang [0] (n=4b5904fe@gateway/web/cgi-irc/
03:06:05arcticfangHey everyone ^_^ Just wondering, can I view .gif files with rockbox?
03:06:37arcticfangAre you positive?
03:06:54arcticfangAlright then. Thanks anyways. :)
03:06:59onlysoaaSay, anyone knows where the GPIO addresses are defined in for TCC780x CPUs? I looked around and couldn't seem to find it.
03:07:18 Join _ahmina_ [0] (
03:07:28onlysoaaOh, nevermind. Found it.
03:07:41cool_walking_arcticfang: There's this patch ( ) but it looks very old/out-of-sync.
03:07:42 Quit lymeca (Connection reset by peer)
03:07:42 Quit arcticfang (Client Quit)
03:07:51 Join saratogahome [0] (n=41becb3b@gateway/web/cgi-irc/
03:08:11saratogahomeonlysoaa: you have the datasheet right?
03:08:16 Join froggyman [0] (
03:08:35 Quit _ahmina_ (SendQ exceeded)
03:08:58onlysoaasaratogahome: Indeed I do. I also found the LCD init routine in the P2's OF, so I'm trying to get that to work.
03:09:15 Join lymeca [0] (
03:09:54onlysoaaI'm setting up a target tree for the P2. How do I get it into SVN, so the configuration stays fixed?
03:10:21saratogahomeonlysoaa: you mean get your code accepted into SVN?
03:11:14onlysoaasaratogahome: Yeah. I'd like the P2 target tree to be accepted, so a new target wouldn't offset some configuration things, etc.
03:11:30 Quit r0b- (Read error: 60 (Operation timed out))
03:11:33onlysoaasaratogahome: Do I post a patch to flyspray?
03:11:53saratogahomeyes post a patch
03:12:16onlysoaaCool, thanks.
03:12:36onlysoaaI'll post one as soon as I figure this out without breaking the rest of Rockbox. :P
03:12:45saratogahomeand of course look at this if you haven't:
03:12:52onlysoaaOf course. :)
03:13:40saratogahomesince the hardware is very similar to the D2 and I7 i expect this should be straightforward
03:14:05saratogahomeat least up until you catch up with those ports, then it may become quite difficult
03:14:32onlysoaaYeah. I think the P2 uses a PMU not found in any current targets.
03:15:04saratogahomeis there a datasheet for it?
03:16:08onlysoaaI think so. Not sure though. If I find anything, I'll post it on the wiki.
03:16:32onlysoaaI recall seeing a specsheet, but I'm not sure if it has all the necessary data.
03:16:42 Join r0b- [0] (
03:16:51saratogahomethe wiki makes no mention of a PMU either
03:17:02r0b-if i hit Select when i plug my sansa in with Rockbox on it will it wtill charge or no?
03:17:02saratogahomeso adding scans and information would be a good idea if you get a chance
03:17:31onlysoaaThe P2 docs mention a PMU. I think they describe the PCB even better than a scan would.
03:17:53saratogahomeah yes i found them on google
03:18:09onlysoaaThe PMU docs?
03:20:21saratogahomethe schematics
03:20:35saratogahomethe "power block" it mentions doesn't seem to be software controlled
03:20:39onlysoaaAh, the P2's codec isn't supported by any current target either, huh?
03:21:06onlysoaaOh? I guess I could investigate later.
03:21:57onlysoaaI guess it's connected to some GPIO pin for the battery status readings.
03:23:20 Join kronflux [0] (n=8eb14dca@gateway/web/cgi-irc/
03:23:49 Join saratogahome2 [0] (n=41becb3b@gateway/web/cgi-irc/
03:23:49 Quit saratogahome ("CGI:IRC (EOF)")
03:23:58saratogahome2no i'm wrong theres another chip that looks like its CPU controlled
03:24:22kronfluxtwo questions. firstly, which one is better? the Sansa Fuze, or the e280v2? I have both, but my fuze has a broken screen, so I want to know which you guys think has better hardware(aka: eventually which is better to run rockbox on)
03:24:47kronfluxI assume with it being a v2, its probably better somehow. I dunno how that works.
03:25:56kronfluxbecause right now the only upside to having the e280v2 is that A) the screen works, and B) its an 8 gig instead of my 4?gig fuze. but I dont like the scroll wheel on the e280v2..
03:26:31cool_walking_I don't know the differences, but I don't think "newer is better" is always the case. The hardware might have just been cheaper to produce.
03:26:45 Quit dstaley (" ajax IRC Client")
03:26:48kronfluxthats what I was thinking. which is why I asked :)
03:27:57saratogahome2or mayber there isn't
03:28:07saratogahome2i'm getting cross eyed looking at these schematics
03:28:09 Quit saratogahome2 (Client Quit)
03:28:40onlysoaaxD Well he was quick to succumb to the datasheet.
03:32:45onlysoaaSay, anyone know the best way to do #ifdef something or somethingelse?
03:35:02Unhelpfulother than, um, #ifdef?
03:35:41Unhelpfuloh, perhaps you want "#if defined(SOME_MACRO) || defined(SOME_OTHER_MACRO)" ?
03:35:55onlysoaaWell there's some stuff in system-tcc780x.c that's #ifdef COWOND2.
03:36:01onlysoaa−−yeah, that's the best way?
03:36:21onlysoaaI'll keep the D2 stuff for now, and see how it works.
03:36:41Unhelpfulwell, the code inside that #if would be compiled if either of the two macros were defined.
03:37:22 Quit |ahmina| (Connection timed out)
03:37:48onlysoaaPerfect! That's just what I wanted.
03:40:07*JdGordon has a good start to better release scripts
03:40:49UnhelpfulJdGordon: editing
03:41:20JdGordonmy end goal is to have " 3.3" do everything needed
03:41:39Unhelpfulyou might want to check out some work i've done on it. it *might* be of use. ;)
03:41:54r0b-ok i wanna know who though to put Doom in Rockbox :P
03:41:56onlysoaaHelp, what the heck!
03:42:00onlysoaaIn file included from /home/soaa/rockbox/apps/bookmark.c:34: /home/soaa/rockbox/apps/recorder/icons.h:33:33: error: bitmaps/rockboxlogo.h: No such file or directory
03:46:04 Join HellDragon [0] (
03:46:28 Quit hd (Read error: 54 (Connection reset by peer))
03:51:16UnhelpfulJdGordon: here's a diff from current to my version... but my has been out tree for a while, and the diff *looks* ok to me, but this may not work 100%. additions are use of GetOptions, and options for source, build, and output directories.
03:52:08Unhelpfulthe only thing i see missing is a chdir into $srcroot before updating.
03:52:27 Quit evilnick (" ajax IRC Client")
03:55:37 Quit lymeca (Connection timed out)
03:56:46onlysoaaShoot, would anyone know why I get that error when I build Rockbox for my new P2 target?
03:56:56onlysoaaI probably missed something−− but what?
03:58:11LloreanWhat's the P2's screen size?
03:58:51Unhelpfulonlysoaa: likely you need to edit apps/bitmaps/native/SOURCES and/or add an appropriate rockboxlogo there.'d need to do both, then. :)
03:59:11onlysoaaIt was building correctly when I did it for the D2 with the modified res.
03:59:26onlysoaaAhh kk.
04:00:52onlysoaaHuh, in that SOURCES file, the condition is LCD_WIDTH >= 320, so it should still work...
04:01:30 Quit perrikwp (" ajax IRC Client")
04:02:40Unhelpfuland it has a 16-bit LCD?
04:02:49onlysoaaIndeed it does.
04:03:02Unhelpfulthat's rather odd, then.
04:03:25onlysoaaOh, I see. I flipped the width and height. Mah bad.
04:06:48JdGordonUnhelpful: ta, ill see about using some of that
04:06:48 Quit froggyman (" HydraIRC -> <- In tests, 0x09 out of 0x0A l33t h4x0rz prefer it :)")
04:09:14 Quit kronflux ("CGI:IRC")
04:09:31 Quit Zarggg ()
04:10:51 Join antil33t [0] (n=Mudkips@
04:12:29 Quit HellDragon (Read error: 54 (Connection reset by peer))
04:12:35 Join HellDragon [0] (
04:21:09 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
04:28:10 Quit __lifeless (Read error: 110 (Connection timed out))
04:30:06 Join Zarggg [0] (
04:36:04***Saving seen data "./dancer.seen"
04:39:54 Quit HellDragon (Read error: 104 (Connection reset by peer))
04:48:45onlysoaaArrrrgh. I can't figure this out.
05:01:28 Quit courtc (Read error: 113 (No route to host))
05:03:45onlysoaaThis is strange. I set all the GPIO stuff correctly, as far as I know, and still no effect on what's displayed.
05:04:07onlysoaaI also changed the LCD settings to match those from the P2's OF, still no dice.
05:05:38onlysoaaUgh, this is depressing...
05:13:35 Quit chandoo ("Leaving")
05:14:56Unhelpfulhrm... a draw hook per scroll line is probably overkill. per viewport might be a good bit cheaper. :)
05:17:14Unhelpfulwhere exactly is the multi-line gradient style drawn by lcd_gradient_rect_scroll used?
05:18:31LloreanGradient line selectors?
05:20:12UnhelpfulLlorean: the function *seems* to be for drawing a gradient that covers several lines, though, and i can't really remember ever having seen that done.
05:22:32LloreanBookmarks with a gradient line selector.
05:22:40LloreanBookmarks are a list of two-line items.
05:22:49LloreanSo the highlight covers two lines at a time
05:23:02Unhelpfulahhh... i never use bookmarks, no wonder i've never seen them.
05:41:02 Quit jordan` (Read error: 104 (Connection reset by peer))
05:41:38 Quit Horscht ("Verlassend")
05:41:40 Join jordan` [0] (i=gromit@
05:47:07Unhelpfuli'm trying to merge as much of the styled-text code as possible together, to reduce the cost of adding hooks for custom styles. i wasn't really sure exactly where/how the two gradient functions differed in terms of their use.
06:05:36 Quit FlynDice (Read error: 60 (Operation timed out))
06:12:17 Quit Lss (Read error: 104 (Connection reset by peer))
06:15:35onlysoaaYES. Finally it turns out it was some settings in the LCD controller regarding bpp that were off.
06:16:00onlysoaaNow the image that displays isn't distorted, but it's offset to the left and it seems to be missing the R channel...
06:16:40 Join _lifeless [0] (n=lifeless@
06:18:45onlysoaaHmm, no, rather, it's not missing the R channel, but it seems that the G channel is too active.
06:19:47onlysoaaSo, everything is slightly greenish...
06:23:55Unhelpfulperhaps the bpp setting is not *quite* right? rgb565 vs rgb555 or such?
06:25:17tmztonlysoaa: do you know where the lcd settings are?
06:25:50onlysoaaUnhelpful: I tried all the combinations, and none display correctly.
06:26:44onlysoaatmzt: I set the timings as directed by the TCC7801 datasheet. I also tweaked some things according to the AUO panel datasheet, but that doesn't seem to have changed anything.
06:26:49tmztthat would make sense if the color issues were per pixel, not in fields (the channels I think onlysoaa is refering to)
06:27:47tmztthat doesn't make any sense with a framebuffer does it
06:28:29onlysoaaHuh? I don't get it...
06:28:55tmztit's not right, I read yesterday that you said only blue was on the whole display
06:29:04onlysoaaThe color issues are most definitely per pixel, and not with the fields being offset. The hues are right; the Rockbox logo is yellow.
06:29:26tmztis it packed? do you know the format?
06:30:08onlysoaaI don't think I know... I'm just trying whatever makes most sense from the datasheet. :P
06:30:11onlysoaaThe format of what?
06:30:18tmztpixel format
06:31:02onlysoaaWell the screen accepts rgb888, the P2's OF uses rgb666 though, I believe.
06:31:29tmzthow is your framebuffer set up currently?
06:31:43onlysoaaIt probably isn't. :P
06:31:53onlysoaaWhere do I set up the framebuffer?
06:31:57tmztis the full screen filled?
06:32:31onlysoaaYeah, just everything is offset to the left by around... 30 pixels? Just going by eye there.
06:32:45tmztand color is wrong?
06:32:59 Join animatorgeek [0] (n=4b53ec3e@gateway/web/cgi-irc/
06:33:00 Join Chesteta [0] (
06:33:11onlysoaaYup. Like I said, green is too strong.
06:33:39animatorgeekIs this the right place to ask for write permission on the wiki?
06:33:58animatorgeekI'd like to update the CFMod page with my results.
06:34:49onlysoaaanimatorgeek: I think you're at the right place, but I don't know how to give permission. You'll have to wait for somebody else to do it for you. :P
06:35:06tmztand you need to provide your username on wiki
06:35:21animatorgeekThat would be DavidAJohnston
06:35:41tmztsomeone might see it or you can ask later
06:35:49animatorgeekSo do I just need to come back another time?
06:36:08***Saving seen data "./dancer.seen"
06:37:08onlysoaaSorry we couldn't help, buddy. :(
06:37:08tmztonlysoaa: what driver are you using? I only see lcd-16bpp
06:37:14animatorgeekokay, thanks.
06:37:42animatorgeekno problem. Thanks for pointing me in the right direction :)
06:37:54onlysoaatmzt: I pretty much copied the whole Cowon D2 tree and modified it to the P2's screen size and settings.
06:38:09tmztnot familiar with svn layout
06:38:25onlysoaaObviously I got some settings wrong, but... at least it's displaying stuff?
06:38:49tmztit shounds like fp is wrong, but I would have to look at the registers to have any idea what to change
06:39:02tmztI have worked with pxafb and omap lcdc before
06:39:09onlysoaaWell the D2's driver is in /firmware/target/arm/tcc780x/cowond2
06:39:20onlysoaalcd-cowond2.c, I believe.
06:40:17tmztokay, I mean the LCDC_HTIME* registers
06:40:40r0b-ok Rockbox says it supports .m4a files but i cant get it to play songs my friend got me from Itunes
06:40:43tmztour usually process is to dump these things from the OF equivalent, ce operating system etc.
06:41:10tmztour not being rockbox, other projects
06:41:18safetydanr0b-: are they m4a or m4p?
06:41:55 Join saratogahome [0] (n=41becb3b@gateway/web/cgi-irc/
06:42:05onlysoaaI changed the HTIME and VTIME to match the P2's resolution, and I changed the LCDC_CTRL register to the P2's rgb666 setting.
06:43:01 Quit animatorgeek ("CGI:IRC (Ping timeout)")
06:43:07tmztit's not enough to match p2 if your device has a different lcd panel
06:43:14onlysoaaIf I set that one to rgb565, the colors go way off.
06:43:44onlysoaaI figured. I tried looking through a disassembly of the OF though, and I just couldn't piece that much together.
06:44:18tmztsee if you can find where a register is set to 0xf0000000
06:44:29onlysoaaI tried. ;)
06:45:06onlysoaaThere's a bunch of occurences, but I couldn't determine which one was actually relevant.
06:45:12onlysoaaI'll look again though,.
06:45:47r0b-m4a safetydan
06:46:07r0b-M4A File (.m4a)
06:46:28saratogahomehow were they encoded
06:47:00r0b-its probably encrypted
06:47:06r0b-they were purchased via Itunes
06:47:31onlysoaaThis is strange.
06:47:32saratogahomepurchased songs aren't m4a but m4p
06:48:03onlysoaaFrom the disassembly, in the part where it loads data for the LCD clock divider, it seems to load from a non-existent address.
06:49:37tmztwhen were the purchased?
06:49:57tmztrecent ones are not protected, most of them anyway
06:50:25tmztiTunes Plus
06:50:42r0b-ok why does my CPU frequency show like 30,000,000
06:50:49 Quit Chesteta ()
06:50:56saratogahomethats what the cpu frequency is
06:51:09r0b-how does that translate into mhz
06:51:55tmztmegaherz is one million herz
06:52:36r0b-so its only 30Mhz
06:52:44r0b-if i change it to 80mhz what can i do?
06:54:08tmztr0b-: where you able to play those files in vlc?
06:54:17r0b-yea tmzt
06:54:33tmztthen they are probably not encrypted
06:57:07r0b-im not to worried
06:57:21r0b-i can reencode to MP3 if needed but its just weird that they wont play
06:57:51tmztyou shouldn't even have to do that, just change the container
06:58:00saratogahomeyou can post a filespray entry with the sample if it doesn't play
06:58:07saratogahomechanging the container won't help
06:58:19tmztoh, sorry
06:58:23tmztso it's codec?
06:58:30r0b-actually if converted to .mp3 it plays fine :P
06:59:13saratogahomewe only support AAC in MP4/M4A container
07:00:23tmztsaratogahome: I wasn't suggesting changing to a different container, I mean the container could be corrupt
07:03:16saratogahomeif it plays in VLC its probably an ok container though
07:05:55onlysoaaWhat the heck? Why are these timings so obscure in the OF?
07:16:49onlysoaaAh hah!
07:16:54onlysoaaI got the picture centered!
07:17:15onlysoaaThat was an issue of horizontal and vertical sync inverting.
07:17:33onlysoaaColors are still off though...
07:17:58Unhelpfulhrm... hooks-per-viewport is a huge space win compared to hooks-per-scroll-line on beast and h300... but a pretty big lose on archos :/
07:21:22 Quit perrikwp (" ajax IRC Client")
07:21:45 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
07:22:04 Quit saratogahome ("CGI:IRC (EOF)")
07:28:13 Join einhirn [0] (
07:39:18 Join bertrik [0] (
07:43:23onlysoaaSo, no one knows how to fix the green channel?
07:43:46onlysoaaI think it's just a matter of shifting the R and B channels << 1, but how?
07:47:09bertrikAny i2c experts here? There seems to be something weird with the generic i2c for e200v2. The e200v2 needs a bit huge delay for fmradio i2 to work with MMU and caches enabled (500 cycles). My clip still works with a 20-cycle delay.
07:48:44onlysoaaHmm... I'll be back tomorrow to see if anyone figured out anything. A post in the forum thread would be real cool. (:
07:55:31 Quit perrikwp (" ajax IRC Client")
07:55:43 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
08:07:08 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
08:08:37bertrikmaybe the e200v2 is a little different in hardware, like no pull-up on the SDA line
08:09:19amiconnUnhelpful: Did you already have a look into that bitmask lookup instead of duplicating info in tagcache?
08:09:55bertrikthe generic i2c driver assumes we can just put a "hard" voltage on the line (normally in i2c, the data and clock lines are open drain and so can either only pull actively low, or let the pin float and let the pull-up make the signal high)
08:10:37amiconnI also wonder whether it's even necessary to optimise 1<<n in tagcache - sh doesn't have it as a single instruction, but with tagcache performing so many other things, just using the subroutine mit not even be noticeable
08:11:07amiconnThe shift-by-n subroutines have an average execution time of 20 cycles, plus call overhead
08:12:49amiconnbertrik: 20 cycles at what clock frequency? The maximum i2c frequency is 400kHz for most devices (for some it's even 100kHz)
08:14:15bertrikthe fm radio in ams sansas can run at 400 kHz, 20 cycles at 248 MHz
08:14:49bertrik20 cycles maybe overclocking it, but it's surprising that 20 cycles works fine for a clip, while an e200v2 needs 500
08:15:05bertrikthat makes me wonder if there's more to it than just some delay
08:15:27Unhelpfultagcache performed decently even with the array-scan version. the function call overhead for 1<<n appeared to be a fair bit of the binsize cost on SH, though, with the switch to char LUTs saving almost as much as the switch to bitmasks did.
08:15:31*bertrik has to go
08:15:39 Quit bertrik ("Leaving")
08:16:17amiconnbertrik: If your delay loop needs 20 cyles, and it is executed twice per i2c clock period, that makes an i2c clock frequency of ~6MHz, way more than 400kHz...
08:17:32 Join ender` [0] (
08:20:14 Join fyre^OS [0] (
08:20:54 Quit safetydan ("Leaving.")
08:21:03 Quit fyrestorm (Read error: 54 (Connection reset by peer))
08:24:41 Join Rob2223 [0] (
08:24:53*Unhelpful spots the rather stupid bug :/
08:25:54JdGordonBagder: 10278 for when you start thinking about the release....
08:36:11***Saving seen data "./dancer.seen"
08:38:54 Join Zagor [242] (n=bjorn@rockbox/developer/Zagor)
08:40:19CIA-38New commit by amiconn (r21193): Add language identifiers for Arabic. Also map 'walon' to Belgian French. It is probably not 100% correct, but at least there is a mapping, similar to ...
08:42:06 Quit kkurbjun (Read error: 110 (Connection timed out))
08:47:02 Quit ender` (Read error: 54 (Connection reset by peer))
08:49:38 Join flydutch [0] (
08:50:22 Join ender` [0] (
08:50:23Unhelpfulamiconn: shouldn't poking around in tagcache wait until post-freeze, too? or is fixing the data duplication and making the set tests a bit better unified across devices a big enough bug to fix now?
08:54:16 Quit Rob2222 (Read error: 110 (Connection timed out))
08:56:47 Quit perrikwp (" ajax IRC Client")
08:57:21 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
09:02:13 Quit BHSPitLappy (Remote closed the connection)
09:07:09Unhelpfulugh. +377B on recorder. i'm starting to think that style hooks in core are not the best, or at least not the lightest, way to do this.
09:07:21 Quit timc (Read error: 104 (Connection reset by peer))
09:08:10 Quit _lifeless (Remote closed the connection)
09:08:10 Join n1s [0] (n=n1s@rockbox/developer/n1s)
09:08:39 Join _lifeless [0] (n=lifeless@
09:08:45Unhelpfulbut the alternative on greyscale/mono targets for getting scrolling to work "right" in pictureflow seems to be importing the whole of the core scroll engine into pluginlib.
09:12:36 Join petur [50] (n=petur@rockbox/developer/petur)
09:12:58 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
09:13:45 Quit pondlife (Read error: 104 (Connection reset by peer))
09:17:40amiconnUnhelpful: I consider error-prone duplication a deisgn bug. And we're in feature freeze, not code freeze
09:18:15*amiconn might have a look in table-based 1<<n tonight (cest)
09:18:28 Quit _lifeless (Remote closed the connection)
09:19:08Unhelpfulamiconn: the "duplicate" data could also be generated via a macro. but if a table-based 1<<n is going to be generally faster and smaller than calling the 1<<n function, it seems like something we ought to do.
09:19:57Unhelpfulick. even adding one field to struct viewport, without any code, increases binsize on recorder by >200B
09:23:27 Join timc [0] (n=aoeu@
09:25:12 Quit hillshum (SendQ exceeded)
09:29:09 Join Thundercloud [0] (
09:47:49Unhelpfuli think a see a lighter way i might implement this hook in core. what would be the best place to add a new global pointer for a function, though? it would be called from the LCD driver files on all bitmap targets, but we don't really seem to have a file for something like that... and adding it to each lcd-*.c would be ugly.
09:49:18n1smake a lcd-bitmap-common.c maybe?
09:51:01 Quit Thundercloud (Remote closed the connection)
09:52:10Unhelpfulperhaps i ought to.
09:52:16 Nick fxb__ is now known as fxb (
09:53:03n1sthere is at least some other duplicated code in the bitmap drivers already, soo... ;)
09:55:02Unhelpfulyes, i noticed that, too, and was thinking about working on that... it's kind of a tangent from this, though. i'm not worried about moving all of that junk in there *yet* :)
09:56:28 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
10:04:50 Quit kachna (Read error: 113 (No route to host))
10:09:07 Join kugel [0] (n=kugel@rockbox/developer/kugel)
10:09:25kugelUnhelpful: :)
10:09:47 Join pyro_maniac [0] (
10:10:07kugela bitmap-generic.c would be a good thing too IMO
10:12:10Unhelpfuli used lcd-bitmap-common... maybe i'll look at that patch after freeze. then again, this work is going to need to wait until after freeze, anyway.
10:12:20 Quit einhirn (Connection reset by peer)
10:13:03Unhelpfulthere's also the copyright question. :/
10:14:02kugelsvn blame helps
10:17:20GodEaterdo Rockbox's EQ settings affect the sound output from MpegPlayer at all ?
10:17:31Unhelpfulbut svn blame only says who committed it. which means that the new file will "belong" entirely to whoever commits it. :/
10:17:45n1sGodEater: don't think so
10:17:45UnhelpfulGodEater: i thought there was a menu option in mpegplayer for that?
10:17:59GodEaterI've no clue
10:18:04GodEatersomeone just asked me and I realised I had no idea
10:18:10GodEaterI never use Mpegplayer myself
10:19:23kugelUnhelpful: with svn blame, you get the revision
10:19:51kugeleither the one who committed holds copyright, or the copyright holder is mentioned in the commit log
10:19:56UnhelpfulGodEater: there are settings for channel modes, crossfeed, equalizer, and and dithering, all can be either forced off or use the core setting.
10:20:11 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
10:20:49kugelor the file is based off a another file, in which case the copyright should be clearly stated in the header
10:21:04Unhelpfulkugel: right, but when a bunch of code is copied to a new file... ugh, that basically means i need to copy everything i get from svn blame into the commit message...
10:21:59kugelthe whole viewports addition is entirely done by linuxstb
10:22:01Unhelpfulkugel: right, but if the code is copied from parts of other files, the only information retained about copyrights of the individual parts is in the commit message or comments. if you make sure to put that information in one of those places.
10:26:57 Join Kupop [0] (
10:29:35Unhelpfulamiconn: i suppose bit_n_table would go in firmware/target/sh/system-sh.c ?
10:29:42n1sUsually the file copyright is set to the contributor of the majority of the code in that file or the original creator of the file or mutiple persons but this case is not much different from rewriting half of a file etc.
10:34:01 Join PaulJam [0] (
10:36:15***Saving seen data "./dancer.seen"
10:39:15 Quit perrikwp (" ajax IRC Client")
10:48:05 Join pyro_maniac1 [0] (
10:50:00 Quit bmbl (Read error: 104 (Connection reset by peer))
10:51:19 Quit cool_walking_ (" ajax IRC Client")
10:53:36 Join Kupo- [0] (
10:55:38 Join kupopo [0] (
10:55:57kupopoI got disconnected :/
11:01:42pondlifeIs there a convention in which order files should appear in SOURCES?
11:02:14pondlifeIn my case, I need to only include for SWCODEC, so should I put it into an existing #if, or append a new one?
11:05:11 Quit Kupop (Read error: 110 (Connection timed out))
11:05:13 Quit pyro_maniac (Read error: 110 (Connection timed out))
11:05:42*kugel would prefer adding it to existing #ifs
11:06:00*pondlife too
11:10:47Unhelpfulpondlife: i've not noted any specific sorting scheme. it seems generally preferred to add new files inside an existing #if when possible
11:13:55 Quit Kupo- (Read error: 110 (Connection timed out))
11:15:22 Join robin0800 [0] (
11:16:11pixelmapondlife: I never did a full make before in my manual build folders
11:16:27pondlifeWeird, maybe it was make clean I did that fixed it?
11:18:41 Join kachna [0] (
11:18:47pixelmacould be, I remember that helping me sometimes too (don't remember the details but I think it was in changes to exclude/include something). What error did you get - did building stop or did the result look wrong?
11:18:56 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
11:20:37 Quit advcomp2019 (Read error: 104 (Connection reset by peer))
11:21:01 Quit kupopo (Read error: 110 (Connection timed out))
11:26:18 Join efyx_ [0] (
11:27:44 Quit PaulJam (".")
11:29:56 Quit robin0800 (Remote closed the connection)
11:30:04 Join perfectdrug [0] (
11:30:15 Join robin0800 [0] (
11:32:43 Quit fyre^OS (Read error: 54 (Connection reset by peer))
11:34:18 Join mt [0] (n=MTee@rockbox/developer/mt)
11:34:18perfectdrughi, small question is the samsung yh920 the same to the 925 in size and button placement? is the only differnce the color around the buttons in the middle (white vs. black) and the color display? or are there any minor differences e.g. the topbuttons? I ask because of a SVG i want to make for this player.
11:39:15 Join einhirn [0] (
11:39:39 Join cool_walking_ [0] (
11:42:56kugelperfectdrug: I think so, yes
11:44:22 Quit goffa__ (
11:44:22 Quit avacore^ (
11:44:22 Quit soap (
11:44:29 Join goffa [0] (n=goffa@
11:44:51NJoinsoap [50] (n=soap@rockbox/staff/soap)
11:45:08 Join avacore [0] (
11:45:15perfectdrugkugel: ok fine, thanks. I see what i can do.
11:46:58cool_walking_Is there any barrier to replacing the Gigabeat S bootloader in flash with Rockbox code, or does it just need someone to do it?
11:47:31pixelmaperfectdrug: btw. good work with the m200 (and I just noticed that the other players use "Hold" only too, for some reason I thought they used "Hold switch" - should have checked before... :\ )
11:48:24 Join Kupo- [0] (
11:52:25perfectdrugpixelma: thanks. yeah i guess the buttonlabels are the hardest part for me too. it needs a lot of checking for consistency and nice placement. thanks for the cosmetic there. the player creation in inkscape is fun though after all, I'm getting better.
11:52:42pixelmanow I remember, it was part of the idea to label some more things (like reset or microphone hole) and be a bit more precise about it
11:53:20n1scool_walking_: i don't know for certain but i think it's possible but as always messing with the loader in flash is risky
11:53:39pixelmaperfectdrug: no problem, that's the easier part
11:53:48perfectdrugpixelma: aha ok, it's sometimes not that easy to place the labels if you don't own the player
11:54:12cool_walking_n1s: Yes, I thought it would be risky, but it's the only way I can get this 240GB HDD to be useful.
11:54:20Kupo-sorry for asking again, but my connection disconnected last time. Can I access the microphone input and alter the currently playing music output from within a plugin?
11:54:31perfectdrugpixelma: so I think that can be adjusted later most times
11:55:11cool_walking_n1s: So Rockbox doesn't rely on any initing done by the OF bootloader or anything?
11:55:13n1scool_walking_: then i suggest talking to jhMikeS, he's got the most clue about the beast around here i think, haven't seen him in a while though
11:55:37n1scool_walking_: we may very well depend on some of the OF inits but i'm not sure
11:55:54pixelmaperfectdrug: true
11:56:31cool_walking_n1s: thanks
11:59:37n1scool_walking_: i think others would be interested in that too, to get rid of the random reformats and the "file not found" weirdness too
12:00:29cool_walking_Well, I know others are interested in getting this HDD working in an S.
12:01:22pixelman1s: IIRC it was your idea to label the reset hole in the drawings... my problem was to find out where to stop labelling something - on some players it could be quite a lot (optical in/out, headphone, USB and so on) and where to put all those labels in the drawing
12:01:31pixelmado you have an opinion?
12:14:41n1sthe point about the reset hole was that many users simply didn't know it existed when they needed it
12:15:00n1sa thing like a hp jack on a DAP is pretty obvious IMO
12:15:46n1sbut i am not a graphics or layout person so...
12:16:30Unhelpfulamiconn: using #define BIT_N(n) (1LU << (n)) and replacing the tables is +180, pretty much the same as was saved between the two revisions to tagcache. the version that had shown some binsize improvement, but not as much as the LUT, put the bitwise operation inside a function. perhaps a TEST_BIT_N makes more sense?
12:20:12 Quit petur (Read error: 104 (Connection reset by peer))
12:20:29 Join petur [50] (n=petur@rockbox/developer/petur)
12:24:00pixelman1s: hmmm, ok. Usually those are all labelled on the player themselves... I wonder how important that is for other players than the H100/H300s, e.g. the M5 and X5 have a reset hole too but since they also have a hard power off with a button, you never really need it (saw one stuck X5 once where even the hard power off didn't work and had to be reset the other way, that was with the OF though)
12:25:55n1si think including as many useful things as possible is good but if it gets too cluttered i'd skip obvious things like connectors
12:34:22Unhelpfulhrm, __builtin_constant_p() doesn't return a constant value?
12:36:18***Saving seen data "./dancer.seen"
12:37:56 Join fyrestorm [0] (
12:43:07Unhelpful__builtin_constant_p(expr) returns 1 if expr can be determined by gcc to be constant. but it can't be used with __builtin_choose_expr because apparently it's return value is not constant.
12:50:22 Join ch4os_ [0] (n=ch4os@gentoo/user/ch4os)
12:50:29 Quit ch4os_ (Client Quit)
12:51:24n1sUnhelpful: the gcc manual says that you can use builtin_constant_p as a static initializer so it pretty much has to return a constant value...
12:52:35Unhelpfuln1s: and yet __builtin_choose_expr complains about its first argument being non-constant if it's __builtint_constant_p. sounds like a bug, really.
12:52:52n1swhich version of gcc is that?
12:54:05n1ssounds a bit like althoug i can not tell from that if they fixed it...
12:54:29Unhelpfulsh cross-compiler build from
12:54:36n1sso 4.0
12:56:37Unhelpfulhrm, if a macro is defined using an expression in one case, and a function in another, the function version will never be subject to common subexpression elimination, will it? :/
13:05:36Unhelpfulamiconn: the best i've come up with is this:
13:07:37Unhelpfulit's +92B vs svn. an alternative would be to define the LUT char arrays as !!((1LU << n) & TAGCACHE_<set>_TAGS) for each n, which avoids the data duplication.
13:09:08 Join wark [0] (
13:15:29 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
13:17:33 Quit kugel (Read error: 110 (Connection timed out))
13:21:08 Quit vedlith (Read error: 104 (Connection reset by peer))
13:25:25 Join petur2 [50] (n=petur@rockbox/developer/petur)
13:26:29 Quit petur (Nick collision from services.)
13:26:32 Nick petur2 is now known as petur (n=petur@rockbox/developer/petur)
13:29:03 Quit avacore (
13:29:03 Quit advcomp2019_ (
13:29:03 Quit shodanX (
13:29:03 Quit scorche|sh (
13:29:03 Quit _Auron_ (
13:29:03 Quit Kopfgeldjaeger (
13:29:03 Quit Bombe (
13:30:52NJoinavacore [0] (
13:30:52NJoinadvcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
13:30:52NJoinshodanX [0] (
13:30:52NJoin_Auron_ [0] (
13:30:52NJoinBombe [0] (n=droden@freenet/developer/Bombe)
13:30:52NJoinKopfgeldjaeger [0] (
13:30:52NJoinscorche|sh [50] (n=scorche@rockbox/administrator/scorche)
13:33:52 Quit robin0800 (Remote closed the connection)
13:38:33 Join simjoko [0] (
13:39:12 Join B4gder [241] (n=daniel@rockbox/developer/bagder)
13:40:45simjokoexcuse me if this is a dump question, but what's the difference between the ipod 4g firmwares (10/3.1.1) and (4/3.1.1)?
13:41:09B4gderyou should probably ask Apple instead
13:41:29 Join petur2 [50] (n=petur@rockbox/developer/petur)
13:41:29 Join robin0800 [0] (
13:43:05 Quit petur (Nick collision from services.)
13:43:07 Nick petur2 is now known as petur (n=petur@rockbox/developer/petur)
13:44:58 Quit Bagder ("*plopp*")
13:45:03 Quit shodanX (
13:45:03 Quit scorche|sh (
13:45:03 Quit _Auron_ (
13:45:03 Quit Kopfgeldjaeger (
13:45:03 Quit Bombe (
13:45:03 Quit advcomp2019_ (
13:45:03 Quit avacore (
13:45:58NJoinavacore [0] (
13:45:58NJoinadvcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
13:45:58NJoinshodanX [0] (
13:45:58NJoin_Auron_ [0] (
13:45:58NJoinBombe [0] (n=droden@freenet/developer/Bombe)
13:45:58NJoinKopfgeldjaeger [0] (
13:45:58NJoinscorche|sh [50] (n=scorche@rockbox/administrator/scorche)
13:48:58 Quit n1s (Read error: 104 (Connection reset by peer))
13:49:55 Join n1s [0] (n=n1s@rockbox/developer/n1s)
13:50:03 Join pondlife1 [0] (
13:50:28 Part pondlife1
13:53:26 Quit shodanX (
13:53:26 Quit scorche|sh (
13:53:26 Quit _Auron_ (
13:53:26 Quit Kopfgeldjaeger (
13:53:26 Quit Bombe (
13:53:26 Quit advcomp2019_ (
13:53:26 Quit avacore (
13:53:40 Join kugel [0] (n=kugel@rockbox/developer/kugel)
13:55:23NJoinavacore [0] (
13:55:23NJoinadvcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
13:55:23NJoinshodanX [0] (
13:55:23NJoin_Auron_ [0] (
13:55:23NJoinBombe [0] (n=droden@freenet/developer/Bombe)
13:55:23NJoinKopfgeldjaeger [0] (
13:55:23NJoinscorche|sh [50] (n=scorche@rockbox/administrator/scorche)
14:02:37 Join IuDeX [0] (n=4e83d3b2@gateway/web/cgi-irc/
14:02:55IuDeXHey. FS #10048 works now?
14:06:37IuDeXanyone knows?
14:06:44 Join froggyman [0] (n=187b533e@gateway/web/cgi-irc/
14:06:56FrankTM_IuDeX: i'm not quite sure what it does
14:07:04 Quit pondlife (Read error: 110 (Connection timed out))
14:08:22IuDeXSansa AMS MMU and Dcache patch (something like this :) )
14:11:32 Join gregzx [0] (
14:13:07kugelIuDeX: try it out?
14:13:29 Quit IuDeX ("CGI:IRC (EOF)")
14:13:36kugelperfectdrug: funman and pyro_maniac1 have one
14:13:56kugel(samsung yh920)
14:15:15 Join IuDeX [0] (n=4e83d3b2@gateway/web/cgi-irc/
14:15:48kugelperfectdrug: this foto looks good
14:16:32IuDeXkugel: I dont have linux on computer with internet.
14:17:15IuDeXkugel: So I can't apply the patch.
14:18:07perfectdrugkugel: I use this photo right now: but I can't figure out what this pattern is and how I should do it in the svg
14:18:35IuDeXFrankTM_: So It's possible to download uncombiled RB and apply patch?
14:19:37 Quit simjoko ("KVIrc Insomnia 4.0.0, revision: 3239, sources date: 20090520, built on: 2009/06/03 23:16:56 UTC")
14:19:56kugelperfectdrug: I don't think the pattern is critical. Just make it without for now and maybe work on it later
14:20:19FrankTM_IuDeX: should be possible :p
14:20:22kugelIuDeX: yes of course....
14:20:26FrankTM_but there's a bunch of files you need
14:20:38FrankTM_anyway.. you should be able to compile it on windows aswell
14:20:48FrankTM_as far as i know that is :p
14:20:53kugelusing cygwin or a VM, most definitely
14:21:07FrankTM_i do it in VM :P
14:21:12IuDeXcygwin ;/
14:21:15FrankTM_cygwin is a complete bitch
14:21:33kugela VM too, hence I do it on real linux
14:22:12IuDeXOk, SO i'LL TRY WITH VM
14:22:19IuDeXooops, sorry for capslock.
14:22:30FrankTM_IuDeX: may i recommend virtualbox ;)
14:23:03IuDeXFrank_TM: good?
14:23:33FrankTM_works like a charm
14:23:35FrankTM_and it's free
14:23:48IuDeXOk, I'm gonna try it :)
14:23:56IuDeXThanks for help, bye
14:23:59 Quit IuDeX ("CGI:IRC")
14:24:20FrankTM_god. does he only connect when he has a question :P
14:25:06 Quit froggyman ("CGI:IRC (EOF)")
14:26:35 Quit r0b- (Read error: 110 (Connection timed out))
14:26:56 Join r0b- [0] (
14:29:10 Join LambdaCalculus37 [0] (i=44a0430d@rockbox/staff/LambdaCalculus37)
14:32:23 Quit n1s ("Lämnar")
14:35:00pyro_maniac1perfectdrug: where did you get this picture?
14:35:52 Join n1s [0] (n=n1s@rockbox/developer/n1s)
14:35:56 Join vedlith [0] (
14:36:22***Saving seen data "./dancer.seen"
14:39:42perfectdrugpyro_maniac1: just from google is it the wrong one?
14:39:51 Join FlynDice [0] (
14:40:52 Join funman [0] (n=fun@rockbox/developer/funman)
14:42:10 Quit kugel ("exit(0);")
14:42:14pyro_maniac1perfectdrug; i've never seen this in the firmware. Selecting artists looks different to mine. But this doesn't really madders thought.
14:42:19 Join kugel [0] (n=kugel@rockbox/developer/kugel)
14:44:45perfectdrugpyro_maniac1: maybe you can scan the yh 920 for me or funman now as he has joined. what is this yepp' in the picture linked by kugel has this your player too?
14:46:32pyro_maniac1perfectdrug: yepp seems to be a edition branding. mine has a napster branding. i will try to make pictures and link them in the forum.
14:46:39funmanyes i have a 'yepp'. Perhaps this picture is for another hardware revision, or even a player not made by samsung?
14:46:51 Join Bagder [241] (n=daniel@rockbox/developer/bagder)
14:47:16 Quit kugel ("exit(0);")
14:47:25 Join kugel [0] (n=kugel@rockbox/developer/kugel)
14:48:16perfectdrugso I think i don't draw the yepp and do without, but the pattern below the display is of interest for me right now
14:50:15Torneyay, my first rockbox plugin works. fsvo works. prints a splash complaining that one of the many stub functions isn't implemented and then exits. :)
14:50:22Tornebut at least the frotz core compiles and links
14:50:28 Quit kugel (Remote closed the connection)
14:55:03 Quit funman ("leaving")
15:06:23 Join kkurbjun [0] (n=kkurbjun@rockbox/developer/kkurbjun)
15:07:41 Join pondlife1 [0] (
15:08:48 Join kugel [0] (n=kugel@rockbox/developer/kugel)
15:09:22 Join ch4os [0] (n=ch4os@gentoo/user/ch4os)
15:13:30 Quit kugel (Remote closed the connection)
15:13:37 Join kugel [0] (n=kugel@rockbox/developer/kugel)
15:15:01 Quit kugel (Remote closed the connection)
15:15:10 Join kugel [0] (n=kugel@rockbox/developer/kugel)
15:21:00 Quit kugel (Remote closed the connection)
15:21:06 Join kugel [0] (n=kugel@rockbox/developer/kugel)
15:38:23 Quit krazykit (Read error: 110 (Connection timed out))
15:39:53kugelanyone with ipod here?
15:42:43 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
15:43:01 Join toffe82 [0] (n=chatzill@
15:45:03LambdaCalculus37kugel: Any kind of iPod?
15:45:17LambdaCalculus37Got my video.
15:45:18onlysoaaOh no, I forgot my flyspray username.
15:45:26kugelI'm wondering how solitaire works on them
15:45:29perfectdrugSamsung YH-92x SVGs are up for comments FS #10279
15:46:49pixelmakugel: last time I saw it on amiconn's Mini it worked nicely
15:47:17pyro_maniac1perfectdrug: you should post in forum too
15:47:26kugelthe keymap seems quite different from e200
15:47:27LambdaCalculus37kugel: It works nicely on my video as well.
15:47:34onlysoaaAh hah, got in.
15:47:40perfectdrugpyro_maniac1: will do
15:48:13LambdaCalculus37kugel: That's the best we can get with just a click wheel and a center button. ;)
15:48:17pixelmakugel: well, the Ipods have different keys
15:48:43kugelon the e200, you go up and down using up and down, and toggle through the cards using the wheel
15:49:18kugelthat seems natural to me, I wouldn't even think of inverting that as it seems to be done on ipods
15:50:28pixelmaI would guess the Ipod keymap was first - I don't know what you mean with "toggle through the cards" though
15:51:12kugelgo left and right through the 7 stacks and the upper left stack
15:52:33pixelmadon't forget that on the clickwheel it's easier to accidentally scroll when you want to hit the button
15:53:23kugelthat happens on the fuze too
15:53:31kugelI see no problem for solitaire here
15:54:02kugelit's quite annoying in rockblox, but there should be some (non-working) software solution to prevent that
15:54:44onlysoaaSubmitted a patch for my initial work on the Samsung YP-P2 port: FS #10280
15:55:49kugelonlysoaa: have you tried bitshifting the r and b channels?
15:56:02onlysoaaI would if I knew how. :P
15:56:04kugelso that the least significat bit is dropped
15:56:38pixelmakugel: and another thing - it seems to me that on the Ipods the shortcuts were made one button press things (no combos, I'd prefer them) which can have an influence on what you chose for the other buttons as well
15:56:41kugelonlysoaa: your diff lacks some files, did you svn add them?
15:57:43kugelpixelma: up down and left right could be still swapped without breaking that
15:59:04onlysoaaOh oops!
15:59:39 Join {phoenix} [0] (
16:00:19pixelmanot breaking but maybe it wouldn't be as logical anymore. I think an Ipod owner should tell what feels more natural/intuitive altogether. I withdraw from discussion, but please don't change something that has worked for years just like that
16:01:11kugelpixelma: lol
16:01:29Tornedoes solitaire really need the vertical scroll thing?
16:01:34Tornei.e. selceting cards within a stack?
16:01:42Torneit does the right thing anyway even if you select the wrong card in the sack
16:01:44kugelLambdaCalculus37: doesn't that feel more natural?
16:01:46Torneso it seems reundant
16:02:19pixelmaTorne: it also lets you see the cards better if they are a bit too close
16:02:27kugelTorne: no idea, but it seems unlogical that the scroll wheel is doing this vertical scrolling on the ipods
16:02:36pixelmakugel: I don't see what's so funny about it
16:02:47Tornepixelma: it does?
16:03:02kugelI'm a bit surprised that you ask me that. Do you really think I go and just change that?
16:03:10Tornei don't ahve any objection to the scroll wheel being the vertical scrolling, i'm just questioning the need for vertical scrolling at all in solitaire
16:03:33pixelmakugel: you have a habit of changing things a bit to quick sometimes
16:03:38kugelI'm only talking about solitaire btw
16:03:49pixelmayes and?
16:03:56kugelthat was for Torne
16:04:24onlysoaaMkay, I updated my patch.
16:04:31LambdaCalculus37kugel: I would have to try it out to see how it feels.
16:04:39Torneactually i find selection annoying in solitaire at all because the inverted colours on the cards makes it hard to read the suit at a glance
16:04:46kugelLambdaCalculus37: hence I gave you the link to the patch :P
16:04:52onlysoaakugel: Would you know how I can bit shift the R and B channels?
16:05:06Tornei'd much prefer a way of selecting hte stack that didn't invert a card, and actually just selected which stack, and didn't bother with particular cards
16:05:12Tornebut, dunno quite how best to display that
16:05:20pixelmaTorne: when you go through the cards the current one gets a bit more space, on small displays the cards can be so close together that you wouldn't see what they are at all
16:05:59Tornepixelma: hm. i've seen it compress them down together but never change the display based on scrolling, but i have ipodvideo so my screen is huge
16:06:27Torneokay, that makes sense
16:06:43Torne(but i still find the inverted colours annoying) :)
16:07:01kugelonlysoaa: it seems you would have to do that in lcd_copy_buffer_rect()
16:07:21 Join petur2 [50] (n=petur@rockbox/developer/petur)
16:07:36 Quit petur (Nick collision from services.)
16:07:38 Nick petur2 is now known as petur (n=petur@rockbox/developer/petur)
16:07:49onlysoaakugel: Yeah, it seems that I would have to reimplement that function, and use rgb888 instead.
16:09:13onlysoaakugel: The LCD controller right now expects rgb565, but it outputs rgb565 data to the LCD, which expects rgb666.
16:10:05onlysoaaColors screw up if I change the LCD to rgb565 though.
16:10:31kugelonlysoaa: yes, and rockbox doesn't have support for rgb666
16:10:54kugelhence, if you shift r and b, you can a fake-rgb666
16:11:15onlysoaakugel: That's right, so I have to kinda hack the output into fake rgb888.
16:11:34kugelwhy rgb888?
16:11:38onlysoaaThing is, the LCD controller doesn't read rgb666.
16:11:50kugelI thought the lcd wants 666
16:11:59onlysoaaIt reads rgb565 or rgb888.
16:12:40onlysoaaWhat I'm saying is, I can't have rgb666 in the framebuffer. It has to be either rgb565 or rgb888.
16:12:55kugelI guess you want to switch the lcd to 565, or write a 24bit driver :)
16:13:30onlysoaaI tried switching the LCD to 565, but it only messed up the colors more.
16:13:49kugelwhat did it to the colors?
16:14:09kugelalso, what does the D2 do? isn't that 18bit too?
16:14:15onlysoaaGradients and stuff became all fuzzy.
16:15:18onlysoaaNah, the D2's screen accepts rgb888.
16:15:22kugelare you sure you did it correctly? also, you could look what the of does in a disassembly
16:15:36kugelyes, but rockbox can't give rgb888
16:16:34onlysoaaWell Rockbox outputs rgb565, and the LCD controller is set to output at rgb888.
16:16:54kugelso the lcd controller scales it internally?
16:17:04onlysoaaI'm not sure; I think so.
16:17:11onlysoaaI don't have a D2 to test. ;)
16:19:02*kugel finds it weird that setting the lcd to 565 doesn't work
16:19:37onlysoaaWell actually I don't think I can set the LCD. The LCD seems to be hard set.
16:19:50onlysoaaWhat I can set is the format that the LCD controller outputs.
16:20:44kugelso the lcd wants 666, but the controller can only do 888 and 565?
16:21:10 Join evilnick [0] (i=0c140464@gateway/web/ajax/
16:21:50 Join chandoo [0] (
16:22:26kugelyou can get to fake-888 with a bit of shifting, but you'll need to change lcd_copy_buffer_rect() I guess
16:23:04kugelthough, I wonder what the lcd does when it's fed with 888
16:23:49onlysoaaWhat's strange is that the LCD's manual specifies 8 pins per channel, but the P2's sheet shows only six pins per channel attached.
16:24:46kugelonlysoaa: the shift-part shouldn't be too hard. make a struct { char r, char g, char b }, then fill with LCD_RED(*fb_data)<<3 etc,
16:24:53kugelI might be wrong though
16:25:42kugelthat's probably very slow though
16:25:51onlysoaaThe lcd_copy_buffer_rect() I'm looking at is in asm. :P
16:26:11kugelnot nice for you
16:26:30kugelbut if you need to write your own, you can do it in C to
16:27:11onlysoaaIt's alright, hacking this thing has given me fair asm experience. :P
16:27:27onlysoaaI still haven't managed to find how the OF does it though, Oh well...
16:27:56kugelyou know the addresses of the lcd controller?
16:28:16onlysoaaThat's where I've been tweaking things all along. :P
16:28:43pyro_maniac1perfectdrug: did you noticed the rec switch on the yh-920 on the side?
16:29:01kugeland you can't find where the OF writes to those?
16:29:36onlysoaaThere are several places, and they all refer to registers that load data from non-existant addresses. :|
16:30:07 Quit thegeek (Read error: 60 (Operation timed out))
16:30:33perfectdrugpyro_maniac1: no I haven't, which side?
16:31:41Torneyay! my frotz port is able to run the inform code [ Main; ];
16:31:50Torne(i.e. zmachine program that does nothing at all)
16:31:59 Quit pondlife1 (Read error: 110 (Connection timed out))
16:32:03pyro_maniac1perfectdrug: same side like the play ffwd rew buttons
16:32:12Torne"screen output" is next :)
16:32:25 Join krazykit [0] (
16:33:54perfectdrugpyro_maniac1: oh i see now, i just had a low resolution picture from the side and I thought this is just the mic
16:35:17 Join jgarvey [0] (
16:35:50pyro_maniac1perfectdrug: will you make svg for the yh-820 too?
16:36:26***Saving seen data "./dancer.seen"
16:37:16perfectdrugpyro_maniac1: i don't mean with pattern the logo (napster or yepp) but the gradientlike pattern. you see the display isn't a straight line at the bottom
16:40:43perfectdrugpyro_maniac1: yes I think I do the yh-820 next, but I don't know which port is the most necessary right now at all. are the gogears further along in devlopment than the samsung players?
16:42:25 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
16:42:53pyro_maniac1perfectdrug: the yh-820 blongs to the yh-925 and yh-920. they are made at the same time.
16:43:15perfectdrugpyro_maniac1: oops the flyspray comment was from funman and not you regarding the pattern
16:44:28 Join petur2 [50] (n=petur@rockbox/developer/petur)
16:44:48 Quit petur (Nick collision from services.)
16:44:50 Nick petur2 is now known as petur (n=petur@rockbox/developer/petur)
16:45:32 Join Lss [0] (
16:46:23 Join _lifeless [0] (n=lifeless@
16:47:04kugelhm, the 820 could make use of the button bar :)
16:47:20onlysoaaI can't figure this out... D:
16:47:26onlysoaaAnd I need to go...
16:47:35onlysoaaSee y'all some time later then. (:
16:47:53pyro_maniac1perfectdrug: on funmans second gif you can see the buttom line of the display. the 920 has some dots there
16:49:32perfectdrugpyro_maniac1: it's not that i don't see them its just that I dont see them close enough to rebuild them ;)
16:50:38pyro_maniac1perfectdrug; better?
16:51:46perfectdrugpyro_maniac1: could be enough yeah, I try
16:52:34pyro_maniac1perfectdrug: good work so far! cya
16:53:08 Quit pyro_maniac1 ("Leaving.")
16:53:23 Quit saratoga (K-lined)
16:53:23 Quit onlysoaa (K-lined)
16:53:23 Quit LambdaCalculus37 (K-lined)
16:53:23 Quit evilnick (K-lined)
16:55:50 Join LambdaCalculus37 [0] (n=44a0430d@rockbox/staff/LambdaCalculus37)
17:02:27 Quit Zagor ("Don't panic")
17:07:12 Join Ubuntuxer [0] (
17:12:50pixelmathe volume "buttons" on the m200 - which is "volume up" and which is "volume down"?
17:13:34 Quit kachna (Read error: 110 (Connection timed out))
17:15:23 Join evilnick [0] (i=0c140464@gateway/web/ajax/
17:16:08 Quit petur ("connection reset by beer (Leffe blond to be exact)")
17:16:26 Nick evilnick is now known as tinyurlnick (i=0c140464@gateway/web/ajax/
17:30:28 Nick tinyurlnick is now known as evilnick (i=0c140464@gateway/web/ajax/
17:31:50 Join barrywardell [0] (
17:38:26 Quit B4gder (Read error: 110 (Connection timed out))
17:41:45mtI'd really appreciate it if someone could explain this : :/
17:44:50domonokymt: and you are sure id3v1buf isnt overwritten later ? maybe check id3->artist directly after this code ?
17:46:46domonokyor pastebin the complete metadata code for this, so we can take a deeper look :-)
17:46:57 Quit n1s (Remote closed the connection)
17:47:07mtI'm almost pretty sure it's not overwritten later .. That's why I'm going mad ! :(
17:47:24mtdomonoky : Sure, 1 min.
17:52:49 Quit robin0800 (Remote closed the connection)
17:53:26mtdomonoky: here :
17:53:55 Quit wark (Remote closed the connection)
17:55:16domonokyyou are writing all three strings at nearly the same spot.between id3v1buf[0] and id3v1buf[1] is only one byte space :-)
17:55:23 Join wark [0] (
17:55:24domonokythey overwrite itself.
17:56:39domonokyso you need to memcpy the second string to id3v1buf+ sizeof(string1).
17:57:06domonokyalso MAX_STRING is probably not the size you want to use for a memcpy :-)
18:00:41mtdomonoky : aren't id3v1buf[0 or 1] two different character arrays of size 92 each ?
18:01:21domonokyno. id3v1buf is the character array.
18:01:49 Join moos [0] (i=mustapha@rockbox/staff/moos)
18:01:55domonokyyou can also use id3v2buf, its bigger :-)
18:03:18domonokytreat it as one char buffer for your strings and it will work :-)
18:03:26domonoky char* buffer = id3->id3v2buf;
18:03:26domonoky char* buffer_end = id3->id3v2buf + ID3V2_BUF_SIZE;
18:03:42mtdomonoky: now it shouldn't be overwritten right ?
18:04:40mtI'm just testing on id3->artist now.
18:04:43domonokyyes, but you still copy MAX_STRING bytes, even when the title string is smaller.
18:07:06r0b-how hard is it to change the speed of my sansa e250 with rockbox?
18:07:08 Join HellDragon [0] (i=jd@
18:07:53domonokyr0b-: what speed ?
18:08:18mtI'll strncpy instead of memcpy later. I'm just trying to figure out why this is not working. I tried the last code I sent (the one with only one memcpy) and it still doesn't work for buf[0] and works for buf[2].
18:09:15domonokyand did check the string directly after coping it there ? ie output id3->title after you set it.
18:09:59mtyes. Also, in my stubbed codec (the qulification task) I did this : and it worked.
18:11:13 Quit {phoenix} (Remote closed the connection)
18:12:08domonokyreally ? that would give "CMT" as title and "MT" as artist, as far as i know
18:12:10r0b-i wanna set it to 80mhz
18:12:51domonokyr0b-: you mean the cpu speed ? why do you want todo this ?
18:13:27domonokymt: how big is MAX_STRING ?
18:13:32mtdomonoky : id3v1buf[4][92] .. My understanding is that id3v1buf[0] and id3v1buf[1] are 2 different arrays, or rather id3v1buf[1] would be the address of the 93rd byte
18:13:42mtdomonoky: is 92 bytes
18:14:06mt*MAX_STRING is 92 bytes :)
18:16:24*domonoky goes checking the code.
18:16:58mtdomonoky: And no it gave 'Codec works' as the title and 'MT' as the artist. If it were CMT I wouldn't have submitted the patch as final. :P
18:17:58r0b-well id like to get the best speed :P
18:18:01domonokyoh, you are correct.. id3v1buf behaves different then id3v2buf...
18:18:01r0b-its 30Mhz now
18:18:43domonokyr0b-: rockbox will automatically switch the cpu speed as needed.. so if you set it to always 80Mhz you just waste battery.
18:19:00mtdomonoky: So I guess my computer is mad then ? :D
18:19:52domonokymt: yes, something is strange. :-)
18:21:16domonokyi would suggest to try it with id3v2buf :-)
18:22:26r0b-so it goes to 80 when i play like doom?
18:24:21domonokywhen ever it needs the power.
18:25:11mtdomonoky: I guess it would work .. I just wanted to use id3v1 since RM has only 4 tags for metadata, 3 of which are the ones we'd be interested in, and it's unlikely that a tag would be more than 90 characters long. Using id3v1 just 'looks nicer' :P.
18:26:50domonokymt: yes.. i just dont know what could be wrong. maybe try again with specifiying both indices ie id3v1buf[1][0]. it shouldnt change, but who knows :-)
18:27:07mtAnyways I'll have to leave it at that and go continue studying ..
18:27:13mtdomonoky: I tried :(
18:27:36*domonoky whished fun while studying :-)
18:27:46domonoky-d +s :-)
18:28:24mtdomonoky: Thanks, and thats for your help. :)
18:28:43mts/and thats/ and thanks .. !
18:28:55mtBye everyone !
18:29:47 Quit cool_walking_ (Remote closed the connection)
18:33:06UbuntuxerHi, I wrote a patch, which remove the strange viewport menu from star. FS #10281
18:35:30 Quit Ubuntuxer ("Leaving.")
18:36:27***Saving seen data "./dancer.seen"
18:36:59kugelhe just can't stop mixing several changes into 1 patch
18:38:11 Join funman [0] (n=fun@rockbox/developer/funman)
18:38:35 Quit barrywardell (Remote closed the connection)
18:39:19r0b-well the auto setting cpu speed could explain why the battery time alters
18:42:40pixelmakugel: what's the several changes there? Really, just curious
18:43:49kugela) removing the viewport and b) remove some (apparently unsed) #defines
18:43:57oboAre there any devs who can give me some help/guidance with ARM disassembly for the View?
18:44:35pixelmaI wonder if it also fixes the inverted icons in the star menu on greyscale (as reported in 8998 which is not entirely fixed although someone closed the report)
18:45:10funmanobo: ask your question
18:45:42pixelmathat status bar thing looks odd too (it just appears sometimes if the statusbar is redrawn, then vanishes if the star menu is redrawn again)
18:46:37kugelyea, it's sort of broken
18:47:08kugelI can't tell why though
18:47:09obofunman: okay, where to begin... :) one of the things I need to do is find the registers used for LCD and MMC/SD access, since they seem to be different. As far as the LCD goes Daniel suggested looking for related strings and following them
18:47:32obobut there are only one or two of those in the bootloader, and they haven't led me anywhere
18:48:35oboCan you suggest other ways to locate the needed registers?
18:49:21 Join {phoenix} [0] (
18:49:51 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
18:50:02 Quit perfectdrug ("Leaving.")
18:50:24funmanIf you follow the codeflow from the start (0x0) you will find a function with several (around 20 or 30) function calls to the same function with a different argument, that's the lcd init function calling lcd_write_cmd
18:51:04 Join biengo [0] (
18:51:28funmanabout MMC/SD you can look for known values, like 0x1AA , 0x40FF8000
18:51:54funmanthese values come from SD specification and are used in the identification phase
18:55:30obofunman: thanks! I'll get a copy of the SD spec and have a search.
19:03:27 Quit LambdaCalculus37 ("CGI:IRC 0.5.9 (2006/06/06)")
19:04:18 Join LambdaCalculus37 [0] (i=44a0430d@rockbox/staff/LambdaCalculus37)
19:15:18 Join bertrik [0] (
19:18:28pixelmakugel: about FS #10281 - I know it sounds a bit like applying different standards but I don't think removing an unused define at the same time would hurt much, it would be a general cleanup and "standardising" of the star menu. In this case making the removal a different patch seems a bit like overkill to me (works nicely on greyscale btw., the menu isn't "inverted" anymore but it was the only plugin that did this)
19:19:09 Join kugel_ [0] (
19:19:17 Quit kugel (Nick collision from services.)
19:19:31 Nick kugel_ is now known as kugel (
19:19:37pixelmastatusbar is drawn now
19:22:05 Quit daurnimator (Read error: 54 (Connection reset by peer))
19:23:52 Join daurnimator [0] (n=daurnima@unaffiliated/daurnimator)
19:25:37 Quit advcomp2019_ (Read error: 110 (Connection timed out))
19:26:19 Join Thundercloud [0] (
19:28:36 Quit linuxstb (Read error: 113 (No route to host))
19:53:34 Quit biengo (Read error: 113 (No route to host))
19:57:51 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
20:10:05 Join B4gder [241] (n=daniel@rockbox/developer/bagder)
20:13:06DBUGEnqueued KICK kugel
20:15:16 Quit martian67 (Read error: 113 (No route to host))
20:16:51 Join petur [0] (
20:17:02r0b-whats with the Dict plugin in Rockbox
20:17:38Mikachuwhat about it?
20:17:42 Nick gevaerts is now known as gev (n=fg@rockbox/developer/gevaerts)
20:17:57 Nick gev is now known as gevaerts (n=fg@rockbox/developer/gevaerts)
20:17:59 Join hillshum [0] (n=hillshum@unaffiliated/hillshum)
20:18:26r0b-how do i get it working
20:20:04Mikachudid you read ?
20:21:53r0b-i couldnt locate it
20:28:39 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
20:28:56 Join einhirn [0] (
20:30:12 Quit dmb ("Leaving")
20:34:45 Quit daurnimator (Read error: 110 (Connection timed out))
20:36:06 Quit hillshum ("Lost terminal")
20:36:29***Saving seen data "./dancer.seen"
20:36:31 Join robin0800 [0] (
20:39:19 Quit tvelocity (Read error: 60 (Operation timed out))
20:39:48 Join tvelocity [0] (
20:40:27 Join saratoga [0] (i=9803c6dd@rockbox/developer/saratoga)
20:40:48 Join daurnimator [0] (n=daurnima@unaffiliated/daurnimator)
20:42:18 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
20:43:59 Join tessarakt [0] (
20:44:24 Quit Thundercloud (Remote closed the connection)
20:53:07 Join n1s [0] (n=n1s@rockbox/developer/n1s)
20:55:39 Join mcuelenaere [0] (
20:56:52 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
21:05:13 Quit kugel (Read error: 110 (Connection timed out))
21:05:27funmanFS #10271 automatically disable voice upon rockdoom startup
21:08:47 Join mamoul [0] (
21:12:26 Quit Rob2223 ()
21:13:52 Part mamoul
21:19:29 Join DarkDefender [0] (
21:32:18 Join Ubuntuxer [0] (
21:36:31 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
21:37:59 Quit bmbl ("Woah!")
21:38:08 Quit perrikwp (" ajax IRC Client")
21:38:51 Join BryanJacobs [0] (i=80974574@gateway/web/ajax/
21:41:39funmanhum I found the source of another problem in FS #10048 : mpegplayer requests a read() into uncached memory (which happens to be different from physical memory)
21:42:56funmanhandling uncached addresses in sd or dma code is possible but just adds extra code :/
21:44:36 Join dmb [0] (n=dmb@unaffiliated/dmb)
21:47:10 Quit LambdaCalculus37 (" ajax IRC Client")
21:54:58 Quit n1s (Read error: 104 (Connection reset by peer))
21:55:23 Join n1s [0] (n=n1s@rockbox/developer/n1s)
22:03:31 Join Thundercloud [0] (
22:04:06 Join Rob2222 [0] (
22:05:27 Quit wark (Read error: 60 (Operation timed out))
22:19:26 Join hillshum [0] (n=quassel@unaffiliated/hillshum)
22:26:43 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
22:27:51 Quit petur (Remote closed the connection)
22:29:17BryanJacobsthe access-denied template for the wiki is a blank page
22:29:58 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
22:30:20AlexPBryanJacobs: Hi - ignoring the access-denied problem, I take it you are creating your GSOC page?
22:30:44BryanJacobsAlexP: yes - I just made the link on the SummerofCode2009 page
22:30:51BryanJacobsdomonoky gave me write access
22:30:52AlexPgood stuff
22:31:58AlexPBryanJacobs: So now you just need a project page - WavpackHybrid maybe?
22:32:15BryanJacobsAlexP: already on it as "WavpackHybridSupport"
22:32:44evilnickHow is that going?
22:33:26BryanJacobsevilnick: right now I'm learning how the codec works; I've got an encoder I wrote myself from the spec doing joint stereo and entropy coding
22:33:37BryanJacobsnext I'm going to encode hybrid support
22:33:57BryanJacobsonce I have a hybrid encoder that works I'll be sure I understand it well enough to add support to the Rockbox decoder
22:34:18BryanJacobsthe first step to solving a problem is to know what you're up against, right?
22:34:23evilnickBryanJacobs: Thanks - if I were in the other channel I expect I'd be told to read the (new) Wiki page :p
22:34:43evilnickAbsolutely, you have to get as much info as possible and then that paves the way for how to implement it. Probably
22:34:44BryanJacobsevilnick: I'm still writing it
22:34:56evilnickHence my smiley!
22:35:29BryanJacobssorry, the Interwebs aren't the best at passing along sarcasm/joking, even with a smiley
22:36:05 Quit Ubuntuxer (Read error: 60 (Operation timed out))
22:36:31***Saving seen data "./dancer.seen"
22:37:47 Join raphi [0] (
22:38:10evilnickIt's a very interesting idea to support the hybrid version, so I'm looking forward to testing it out - best of luck with this project.
22:38:24 Join notlistening [0] (
22:38:41BryanJacobsthanks - I don't look forward to dealing with the dual-file buffering but that's the last step
22:38:48Mikachuis the support for being able to play the two parts together for lossless, or doesn't just the lossy part work either?
22:39:07Mikachui guess that answers it :)
22:39:08BryanJacobsMikachu: the lossy decoder works just fine already
22:39:24 Quit robin0800 (Remote closed the connection)
22:39:45Mikachui have no idea how wavpack worked, maybe there was a lossy format that wasn't hybrid that worked and lossy hybrid didn't :)
22:39:48 Join robin0800 [0] (
22:39:58notlisteningHey I'd like to start helping again on the Sansa v2 development but is there a quick guide about building the custom firmware somewhere or is it all self explanatory
22:40:24Mikachunotlistening: there should be info about building on the wiki
22:40:25BryanJacobsthe lossy part of a hybrid file is itself a valid Wavpack lossy file
22:40:31Mikachuokay, that makes sense
22:40:39Mikachuwhen would you want to put both files on a dap though?
22:40:46BryanJacobsif you want lossless quality
22:40:48funmannotlistening: look on SansaAMS wiki page
22:40:52notlisteningi'll try there
22:40:59BryanJacobsMikachu: no transcoding! Yay!
22:41:08funmanreally this is as simple as ./rockbox/tools/configure
22:41:08BryanJacobsif you run out of space just delete some WVC files
22:41:14Mikachui guess
22:41:22AlexPMikachu: e.g. if you have amazing earphones, or even moreso if you have e.g. a h100 with optical out
22:41:28notlisteningah thanks funman your posts on the forum have given me hope all this time
22:41:41notlisteningjust had no free time to help
22:41:43CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
22:41:43*BryanJacobs has a pair of Ultrasone HFI-750s
22:41:45Mikachui know why lossless is good, but i thought the idea with wv hybrid was you'd have both on your computer and just the lossy one on your dap
22:42:00funmannotlistening: some way has been made since :P
22:42:10*BryanJacobs also hopes to add correction-file support to ffmpeg after this
22:42:12AlexPMikachu: The idea is that you have the option
22:42:13notlisteningyeah i watch every week
22:42:21MikachuAlexP: okay :)
22:42:22evilnickI'd totally use it to have the lossy ones on the dap and would then copy the WVC across if I wanted some albums to be lossless
22:42:23notlisteningyour comming along great
22:43:02BryanJacobsevilnick: that's how I'd use it too, except in reverse (start with everything lossless and then delete the ones I don't have space for)
22:43:02notlisteningas a side note is there anyone here that helps out with the voice tagging development for rockbox?
22:44:46saratogaBryanJacobs: have you figured out how you will handle buffering correction files?
22:44:54notlisteningor anyone know who might be intrested as I know have the SAPI engine running under linux / Mac and anything else that runs WINE
22:45:33BryanJacobssaratoga: not worrying about it yet... I was thinking allowing some kind of distinct buffer regions
22:46:19evilnicknotlistening: What do you mean by "voice tagging development". I don't understand
22:46:26saratogathats something you should hopefully figure out soon, since that will be a large part of your project
22:46:37notlisteningSo nice windows based voices under linux + a raft of languages
22:46:41saratogaporting the correction code itself to rockbox will likely be quite simple in comparison
22:47:24BryanJacobssaratoga: yeah, I just don't want to put the cart before the horse
22:47:39notlisteningif you used the voice tags under rockbox, spoken menus files and navigation for people who can not see the screen or are driving you have to add voice files to the player to get it to speak them
22:47:39BryanJacobsI wrote a harness so I can test the decoder even without Rockbox support for buffering two files
22:48:16saratogado you expect porting the decoder to be very difficult? usually lossless decoders can work in rockbox with little modification
22:48:24notlisteningyou need a text to speech engine to generate these files
22:48:24saratogathough i have not looked at wavpack
22:48:25 Quit dmb (Read error: 113 (No route to host))
22:48:51notlisteningand i have made the MS vpoces accessible in linux and rockbox is an application that can benefit
22:48:55BryanJacobssaratoga: no, I don't, but I'm not just porting a decoder I'm rewriting a decoder
22:48:56saratogai got most of TTA working in an afternoon for instance while learning the codec API (though still need to fix some bugs)
22:48:59funmannotlistening: you have an e200v2, right?
22:49:17BryanJacobsthe current Rockbox code is based on the wavpack-tiny decoder which has no hybrid support
22:49:18notlisteningyeah a 280
22:49:25BryanJacobsthe Wavpack trunk code is a whole different code base
22:49:38notlisteningactually busted that one got a 250 now :P
22:49:52evilnicknotlistening: Ah, I get it. Although I thought that there was a way of generating those voice clips under linux/OSX already. Never mind.
22:49:53saratogayes but most or all of the code should be resusable right?
22:49:53BryanJacobsso I'm basically writing new support for applying the correction bitstream into the wavpack-tiny decoder
22:50:00notlisteningfunman, 240 :P
22:50:09BryanJacobssaratoga: no, they do things in completely different ways
22:50:16BryanJacobsI can reuse some code but not all
22:50:27saratogawhat cannot be reused?
22:50:31BryanJacobsthe tiny decoder is designed for a minimal memory footprint
22:50:36notlisteningyeah your right evilnick but more choice is always better right
22:50:37BryanJacobsit decodes parts of chunks
22:50:44BryanJacobsthe full decoder does a whole chunk at once
22:51:03BryanJacobsalso, seeking needs to get done at some point - the tiny decoder does an imprecise seek
22:51:09saratogais that a serious problem? the format seems to have very little state
22:51:27BryanJacobsthe format has 8 samples worth of state plus the current weight, which is block-wide
22:51:34saratogai would expect that requires some minor refactoring, not a rewrite
22:51:35BryanJacobsnot very much
22:51:44BryanJacobsI suppose you're right
22:51:47 Quit hillshum (Read error: 110 (Connection timed out))
22:52:17saratogatry to do as little as you have to in order to get playback, then improve from there
22:52:36saratogaif code is inefficient, it can be incrementally improved, just make sure you don't get lost for a month trying to do everything at once
22:52:56saratogayou'll find your work is more fun anyway when people are using it
22:53:02BryanJacobsI kind of want to understand what I'm doing though - I could just slap the whole wavpack decoder into rockbox
22:53:06BryanJacobsand it would most likely work
22:53:08 Quit moos (Read error: 110 (Connection timed out))
22:53:19BryanJacobsbut then I would lose all the work that went into optimizing the rockbox version
22:53:41BryanJacobsI'd really rather understand exactly what needs to be ported and how it works first
22:54:04saratogathats not a bad first approach actually, the two codecs can be merged in that order if you prefer
22:54:27BryanJacobsat any rate, with the progress I'm making I'll have hybrid decoding by the end of the week
22:54:39BryanJacobsI mean around next Sunday the 14th
22:54:43saratogaeven if codecs are written differently, its often not very hard to move large blocks of code between them (WMA for instance now uses large bits of the Vorbis decoder because they were already well optimized)
22:54:46BryanJacobsso I've got plenty of time to deal with the buffering
22:54:50saratogagreat :)
22:55:19BryanJacobswait wait wait...
22:55:31BryanJacobsyou mean port a wavpack-full codec?
22:55:48saratogayou're going to port a lot of the full wavepack decoder regardless
22:55:58BryanJacobslike, milestones: 1) make a "wavpack-full" that's only used for hybrid files with correction info present 2) implement two-file buffering 3) port things into the tiny wavpack decoder
22:56:03domonokynotlistening: you can try rbutil in wine, and see if sapi voice generation works in this combination :-)
22:56:24saratogawell you should ask your mentor about this, but that approach has its merits in my opinion
22:56:34BryanJacobslinuxstb: thoughts?
22:57:07BryanJacobsI'm not sure it's accurate that I'd be porting "a lot" of the full decoder
22:57:15saratogai haven't looked at the code though
22:57:17BryanJacobsit has support for things like five-channel audio and 24-bit storage types
22:57:31saratogaif the full encoder unsuitable (perhaps includes a lot of malloc or fp) then it may not be worthwhile
22:57:44BryanJacobsit has no fp at all actually
22:57:52BryanJacobsWavpack is an integer only format
22:58:07notlisteningdomonoky, of course i ahve the perfect setup for that
22:59:32BryanJacobssaratoga: hmm, bears some thought
22:59:55BryanJacobsI'll see what happens if I port libwavpack and slap a wrapper around it
23:00:10notlisteningdomonoky, Will come back to you in a bit after installing and testing
23:02:26*mcuelenaere wonders what __BACKLIGHT_INIT does
23:04:19bluebrotherrbutil in wine?
23:04:22 Join Chesteta [0] (
23:04:32*bluebrother bets that SAPI won't work in wine
23:05:39Chestetahello, I was wondering about patch 10185, does that modify the bootloader code at all or is it just the installer?
23:06:33domonokyChesteta: thats just for the userfriendly install.
23:09:14saratogaBryanJacobs: diffing the full and tiny decoders, they seem very similar
23:09:23saratogamost of the functions in the tiny decoder are also in the full decoder
23:09:51funmanChesteta: installer
23:10:19Chestetaok, thanks domonoky: i was going to test out 10048, should I rebuild the bootloader with that or not :/ it was mentioned above but said its "optional"... could it affect the blue lines when boosted? (had blue lines on the last patch (yesterday) but there is a new one now)
23:11:20BryanJacobssaratoga: yes, the full decoder is a superset of the functionality of the tiny
23:11:39BryanJacobsI should be able to do the full-decoder port you suggested
23:11:50BryanJacobsit sounds like a good thing to do as an experiment at the least
23:12:52 Quit evilnick (" ajax IRC Client")
23:13:06saratogaanother way to go about it would be to start with the standalone full decoder and merge back in all the tiny decoder code, then port them to rockbox
23:13:11domonokyChesteta: its probably better to also rebuild the bootloader, to make sure everything is ok.
23:13:15saratogasorry i'm not much help deciding how to do it
23:13:23BryanJacobssaratoga: eww, that sounds icky
23:13:33BryanJacobsI think I like starting with tiny and porting the hybrid support better
23:13:34 Quit perrikwp (" ajax IRC Client")
23:13:47BryanJacobsI mean, no Rockbox target can even DO five-channel audio, right?
23:13:54BryanJacobsso why would we want that junk in there?
23:13:54 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
23:14:12saratogathough some formats (AC3) actually decode 5 channel audio and downmix to 2 channel
23:14:22BryanJacobsditto 24-bit sound and IEEE floating point
23:14:46saratogaare those really difficult to support?
23:14:49ChestetaFunman: are the lines (when boosted) a part of/issue with 10048 or is that due to somehting else?
23:14:51saratogadecoding 24 bit files would be nice
23:14:57BryanJacobsthey're more code in the fastpath
23:15:05BryanJacobswhat do we do, dither into 16 bits?
23:15:07funmanChesteta: i don't know
23:15:38BryanJacobshmm. Maybe the full decoder will be realtime on modern targets anyhow
23:15:44Chestetaok, ill test on e280v2 in a min
23:15:54saratogadithering can be enabled, but the DSP pipeline actually operates on 32 bit ints anyway
23:16:03saratogaso if you can output 24 bit samples it makes sense to do so
23:16:04 Quit Thundercloud (Remote closed the connection)
23:16:25BryanJacobsbtw, about AC3: it makes sense to have 5.1 in it b/c DVDs are muxed that way; nobody has a Wavpack source that comes with multichannel audio
23:16:47AlexPBryanJacobs: That was quick! :)
23:16:52BryanJacobsI didn't know the DSP took more than 16 bits/channel
23:17:10BryanJacobsAlexP: I didn't have to go look anything up to know where I stood :-P
23:17:18AlexPgood news :)
23:17:57domonokyBryanJacobs: about the buffering thing: what happend to the idea of interleaving (on load) those two files ?
23:18:10saratogathe native integer type for CF/ARM is 32 bit so it'd actually be slower to use 16 bit ints
23:18:22saratogaplus having more precision is useful for DSP effects
23:18:24BryanJacobsdomonoky: I found out we don't decode whole blocks at once
23:18:52domonokyah, so interleaving is not really possible ?
23:18:59BryanJacobsdomonoky: it's possible
23:19:04mcuelenaerehmm anyone able to explain what _backlight_on_isr() & friends do? (part of HW PWM backlight fading) Most targets seem to #define it to _set_backlight_on()..
23:19:13BryanJacobsit's just that the block boundaries are part of the wavpack format, right?
23:19:26BryanJacobsso if we do interleaving on load the buffer code has to understand wavpack
23:19:42BryanJacobswhich means it's not generic enough to be used for, say, MIDI patchfiles
23:20:22 Join Thundercloud [0] (
23:20:23*domonoky images something like a format specific loader.
23:20:30BryanJacobsthe "generic" way to do that interleaving would be to have a callback in the codec to say "switch files here" when it hits some byte sequence/trigger
23:20:39*BryanJacobs is beaten to the punch by domonoky
23:21:20BryanJacobswhat are people's thoughts on per-codec buffering code?
23:22:14saratogaBryanJacobs: does the tiny decoder not do 24 bit files? the spec implies that interally everything is 24 bit and 16 bit files are zero padded
23:22:33pixelmadomonoky: you own an m200, right? Could you have a look at the recently added m200 drawings and tell me which part of the volume button is down and which up?
23:22:57BryanJacobssaratoga: there are "optimizations" which say they only work on 16-bit stereo code, according to the comments in the full-decoder source
23:23:29saratogaah ok that makes sense
23:23:55BryanJacobssaratoga: see function decorr_stereo_pass_i
23:24:25amiconnUnhelpful: I was thinking about adding the BIT_N() thing to firmware/ and use it wherever possible in core code at least
23:24:27domonokypixelma: right side of the switch is +
23:24:37BryanJacobsunpack.c line 1018
23:25:04amiconnThe implicit TEST_* would make this more difficult. Not all places using 1<<n use it for (immediate) tests
23:25:34pixelmadomonoky: thanks
23:26:14amiconnmcuelenaere: The _isr() variant only switches part of the backlight circuit - the part that needs to be switched fast for software pwm
23:27:07BryanJacobssaratoga: oh also, you can have 32-bit integer wavpack files
23:27:13mcuelenaerebut these parts only get used upon *hardware* pwm?
23:27:14amiconnThis is needed on ipod video and nano, which have both a software controlled hardware pwm (for brightness), and use software pwm for fading (because the hardware pwd only has a few steps)
23:27:29mcuelenaereohh ok
23:27:50amiconnThe _isr() variants only switch the actual backlight power, but leave the hardware pwm circuit running
23:28:11 Quit HellDragon (Read error: 104 (Connection reset by peer))
23:28:15amiconnSwitching the whole thing takes time - way too much time for use in an isr
23:29:14mcuelenaereamiconn: hmm so BACKLIGHT_FADING_PWM isn't really suitable pure hardware PWM then?
23:29:22mcuelenaeresuitable for*
23:29:44funmanobo: why did you remove first 64 bytes of view bootloader before disassembling ?
23:30:12amiconnNo, it means software pwm
23:30:44notlisteningurgh have sansa never heard of using a standard usb connector
23:30:49mcuelenaerequoting config.h: /* Backlight fading is controlled using a hardware PWM mechanism */
23:31:04notlisteningspent thirty mins searching for that bugger
23:31:13funmannotlistening: don't they sell accessories plugged to this port?
23:31:18 Join hillshum [0] (n=quassel@unaffiliated/hillshum)
23:31:40notlisteningyeah but not sure they are available anymore
23:31:52AlexPyeah, 'tis more than just USB#
23:32:04saratoganotlistening: the clip uses standard usb but of course has no line out, etc
23:32:06amiconnmcuelenaere: That comment is clearly wrong
23:32:06AlexPBut I agree, I'd like both
23:32:28 Quit BryanJacobs (" going home now")
23:32:38mcuelenaereamiconn: hmm what should I chose for HW PWM then? BACKLIGHT_FADING_TARGET or BACKLIGHT_FADING_SW_HW_REG?
23:32:46notlisteningThe users know best
23:33:46 Join wincent [0] (
23:34:06notlisteningany particular firmware to use?
23:34:31funmanfor patching AMS bootloader? no.
23:34:44amiconnI'm not sure. The only method I know quite well (since I refined its implementation) is software pwm. I thought about the sw setting method, but didn't want to implement it because the brightness would need really many steps to make it look good, and all targets back then had <= 32 steps
23:35:21amiconnWhile this might sound a lot, not all steps are usable for fading unless the user has his backlight set to max. brightness (which is rather unlikely)
23:35:32notlisteningok fingers crossed, any work being done on USB?
23:36:01amiconnI don't know which targets use one of the last two methods
23:36:10funmanthere is doc for a maybe similar chip but noone worked on it yet
23:36:15*hillshum only sets brightness to max when outside
23:37:09*amiconn usally sets brihtness to less than half on the targets with brightness control (e.g. H300, X5)
23:37:42notlisteningright and a normal build am i wanting?
23:37:47hillshumon the OF i kept it all the way down, but RB goes lower
23:43:19notlisteningok getting the tool chain :P
23:44:53obofunman: good question. I did it several months ago when I first started looking at the View, but it doesn't make any sense - I may have been getting it confused with other mi4 headers?
23:45:48domonokyhello wincent. how is your gsoc project going ? :-)
23:46:05wincentWorking on it, of course.
23:46:25wincentBut the way became very interconnected.
23:46:46wincentThe scheduler uses audio subsystem to synchronize.
23:46:54 Join pano [0] (
23:47:04wincentThat means, it counts time in samples.
23:47:16panois here anyone, that "feels responsible" for rbutil?
23:47:38*domonoky waves and points at blubrother and himself :-)
23:48:52wincentBecause of that, I'll need to port everything at once, following the whole process of starting of Pure Data.
23:49:26domonokyah, aslong as it isnt a big problem :-)
23:49:35 Quit {phoenix} (Remote closed the connection)
23:49:48notlisteningdomonoky, I assume the SAPI thing is not really or interest then, in native linux?
23:49:58CIA-38New commit by bluebrother (r21194): Show Qt version number in Sysinfo dialog and in startup debug output.
23:50:04*amiconn wonders whether it would make sense to allow SAPI 64 bit voice building on x64 windows systems
23:50:22panodomonoky: you're talking to me? :-D It's about the German translation of rbutil
23:50:30domonokynotlistening: not really, there are other TTS systems for linux available..
23:50:51domonokypano: the waveing was for you, everything else not :-)
23:51:31amiconnIt would be possible, but it would need to be a separate option in configure. It wouldn't be too difficult to implement: it just requires starting the sapi_voice.vbs "server" via the 64 bit version of cscript.exe
23:51:42panodomonoky: :-) k at least you now know, what my "problem" is about :-D
23:51:52amiconnThis is possible from a 32 bit process (i.e. cygwin) via the 'sysnative' virtual directory
23:52:06panoThe current German translation is, let's say "not so good"
23:52:21panoIt's like a robot is talking to you
23:52:26notlisteningdomonoky, I understand the quality is not great on them but your call
23:52:35domonokypano: feel free to help-out and improve it. patches are welcome :-)
23:52:45notlisteningdomonoky, think your busy now so laters
23:52:48panoyeah that was what I wanted to ask
23:52:56AlexPpano: I will point out that both domonoky and bluebrother are German :)
23:53:01 Quit bertrik ("Leaving")
23:53:08amiconnThe question is whether it makes sense - most third party sapi voices are probably 32 bit, and the 64 bit sapi voice from microsoft (on Vista x64) has a 32 bit equivalent
23:53:10hillshumdo we have any docs for translating rbutil?
23:53:24panoif I finished "fixing" the ts. file, where do I send it or the output of svn diff ?
23:53:48hillshumsend the patch to the tracker
23:54:05domonokyhillshum: there is something on the rbutilDev page. but its essentially just using qlinguist on the language file.
23:54:38panohillshum: yeah, right^^ Thanks!
23:55:00notlisteningwith sapi voices 32 over 64 bit only in terms of production speed?
23:55:15 Quit jgarvey ("Leaving")
23:55:53notlistening*the advantage i meant to say

Previous day | Next day