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 2012-03-03

00:00:22bertrikor, maybe it still was in dec 2010
00:00:29saratogafwiw i scoped my clip+ and couldn't see any sign of a DC offset
00:00:48saratogaalthough the noise floor is much higher with mute off
00:00:48jhMikeSthe mixer wasn't in until 2011
00:00:50 Quit z180 (Quit: Page closed)
00:03:36bertrikverbtim, bubbles looks fine now with your change :)
00:03:39 Quit thomasjfox (Ping timeout: 265 seconds)
00:05:10dfktbertrik, are there any news about RDS scrolling on the clip zip?
00:05:58bertrikNo. Actually it used to scroll, but it doesn't anymore for some time now
00:06:33dfkti see, just curious... i don't use the radio often, just noticed it again in a new theme i made
00:07:05dfktsince the simulator now shows a fancy "rockbox radio" station :)
00:07:11 Quit pamaury (Read error: Connection reset by peer)
00:10:52 Quit dhrasmus (Quit: Leaving)
00:12:13dfktJdGordon, it seems bar sliders break some stuff currently - they offset the bar fore- and background by 1 pixel, and they break transparency. ditching the sliders makes everything work again
00:31:58 Quit saratoga (Quit: Page closed)
00:36:38 Quit lebellium (Quit: ChatZilla 0.9.88 [Firefox 11.0/20120222074758])
00:36:56 Quit rarog (Quit: ChatZilla 0.9.88 [SeaMonkey 2.4.1/20111011102430])
00:47:05 Join CaptainKewl [0] (
00:50:25 Join enthdegree [0] (~BitchX@unaffiliated/enthdegree)
00:58:11 Quit n1s (Quit: Ex-Chat)
01:09:27 Quit DerPapst (Quit: Leaving.)
01:17:18 Quit bertrik (Quit: And That, My Liege, Is How We Know the Earth to Be Banana Shaped)
01:17:42 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
01:19:46 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
01:31:46 Quit evilnick (Ping timeout: 272 seconds)
01:34:19***Saving seen data "./dancer.seen"
01:42:21 Quit advcomp2019 (Ping timeout: 246 seconds)
01:46:53 Join advcomp2019 [0] (
01:46:53 Quit advcomp2019 (Changing host)
01:46:54 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
01:47:09 Quit dfkt (Quit: -= SysReset 2.55=- Sic gorgiamus allos subjectatos nunc.)
02:09:52 Join Easior [0] (~Easior@
02:13:11adnaphow do you put the clip zip in hold mode?
02:15:16 Part Easior ("ERC Version 5.3 (IRC client for Emacs)")
02:19:11 Join Easior [0] (~Easior@
02:20:01 Quit Easior (Remote host closed the connection)
02:21:01verbtimadnap: home + select
02:23:12adnapis home the top button?
02:23:29adnapi tried top + center
02:23:56adnapgot it
02:24:10adnapit's tfhe back buttom + select
02:24:20adnapwhich i guess is the home button
02:24:32adnapbut it has a back arrow on it
02:26:18 Quit perrikwp_ (Read error: Connection reset by peer)
02:34:56 Quit Zarggg (Ping timeout: 265 seconds)
02:36:03funmanyeah it is a home symbol on other clips
02:37:13funman , not sure how to document that one. as 'home' button since it's the exact same layout than other clips or as 'back'
02:45:37 Part verbtim ("Leaving")
02:54:29CIA-44Commit 3f82f3a in rockbox by Rafaël Carré: (Author: Dimitar Dimitrov) Sansa Clip Zip: Bubble now shows a proper background image
02:57:02CIA-443f82f3a build result: All green
03:08:22 Join perrikwp [0] (
03:15:28 Join verbtim [0] (~verbtim@
03:23:00 Quit Rower85 (Read error: Connection reset by peer)
03:32:38 Join idak [0] (
03:34:22***Saving seen data "./dancer.seen"
03:34:27idakI found that some "Daily Built Manuals" doesn't include "Contents". I checked clip+ and e200. Maybe TEX cross reference is broken (it needs "make manual" twice). Is it intentional?
03:36:41idakSorry. it was building process. now it has "contents"
03:36:55 Quit idak (Client Quit)
03:38:45 Quit advcomp2019 (Ping timeout: 252 seconds)
03:44:06 Join advcomp2019 [0] (
03:44:07 Quit advcomp2019 (Changing host)
03:44:07 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
04:16:15 Join amiconn_ [0] (amiconn@rockbox/developer/amiconn)
04:16:16 Quit amiconn (Disconnected by services)
04:16:23 Quit pixelma (Disconnected by services)
04:16:25 Join pixelma_ [0] (pixelma@rockbox/staff/pixelma)
04:16:27 Nick pixelma_ is now known as pixelma (pixelma@rockbox/staff/pixelma)
04:16:37 Nick amiconn_ is now known as amiconn (amiconn@rockbox/developer/amiconn)
04:24:03 Join dys` [0] (
04:26:07 Quit dys (Ping timeout: 265 seconds)
04:27:57 Nick dys` is now known as dys (
04:33:01 Quit anewuser (Read error: Connection reset by peer)
04:51:52remlapToday I learned Rockbox will carrying on playing FM when plugged into USB
04:55:52 Quit TheSeven (Disconnected by services)
04:56:12 Join [7] [0] (~TheSeven@rockbox/developer/TheSeven)
05:01:33 Quit verbtim (Ping timeout: 276 seconds)
05:03:39 Join fr3e [0] (
05:03:42fr3eim trying to follow a guide to install dmalloc.h. it wants me to ungzip and untar files into my source work directory. what is my "source work directory"?
05:03:42 Quit factor (Ping timeout: 260 seconds)
05:10:45 Join factor [0] (
05:12:28 Join Rob2223 [0] (
05:14:59 Quit Rob2222 (Ping timeout: 244 seconds)
05:32:15fr3ei made dmalloc, and now dmalloc.h is in /usr/include/dmalloc.h and /usr/local/include/dmalloc.h. but when i try to compile it still says dmalloc.h no such file or directory. what to do?
05:33:21 Quit ps-auxw (Ping timeout: 246 seconds)
05:34:26***Saving seen data "./dancer.seen"
05:35:31 Join ps-auxw [0] (~arneb@2001:470:c807:0:1532:4e5f:2ad3:4123)
05:39:23 Join Keripo [0] (
06:13:10fr3ei put #include <dmalloc.h> in my .c file
06:22:42 Join Keripo1 [0] (
06:23:24 Quit Keripo (Ping timeout: 252 seconds)
06:40:14 Join mortalis [0] (~mortalis@
06:58:49 Quit enthdegree (Ping timeout: 260 seconds)
07:23:05fr3ei typed cpp -v it said #include <...> search starts here: /usr/local/include /usr/lib/i386-linux-gnu/gcc/i686-linux-gnu/4.5.2/include /usr/lib/i386-linux-gnu/gcc/i686-linux-gnu/4.5.2/include-fixed /usr/include/i386-linux-gnu /usr/include . i have a /usr/include/dmalloc.h and a usr/local/include/dmalloc.h. is is supposed to be working?
07:23:44CIA-44Commit 286a4c5 in rockbox by Michael Sevakis: Revise the PCM callback system after adding multichannel audio.
07:27:13CIA-44286a4c5 build result: 9 errors, 2 warnings (Michael Sevakis committed)
07:30:12jhMikeSlol, I don't think I meant to do that yet
07:34:29***Saving seen data "./dancer.seen"
07:39:20 Join OzzieJacks [0] (
07:40:50 Quit factor (Read error: Connection reset by peer)
07:49:22CIA-44Commit 59e5a32 in rockbox by Michael Sevakis: Fix errors in unintentionally merged gerrit commit.
07:52:34CIA-4459e5a32 build result: All green
07:58:07 Join factor [0] (
08:19:16jhMikeStuners are acting wonky nowadays
08:23:38jhMikeSnot powering up properly (seen on e260 and H10 so far)
08:23:54jhMikeSfor H10, the recording screen
08:39:41 Quit nosa-j (Read error: Connection reset by peer)
08:39:44 Join stoffel [0] (
08:39:46 Join nosa [0] (
08:40:05 Nick nosa is now known as nosa-j (
09:03:03 Quit adnap (Ping timeout: 265 seconds)
09:04:43 Join adnap [0] (
09:12:16 Join captainkwel [0] (
09:15:27 Quit CaptainKewl (Ping timeout: 245 seconds)
09:18:41kugeljhMikeS: const'ifying could have better been a separate change
09:24:18jhMikeSIt was done as part of the callback form changes, which led to code that needed updating so as to not spit out warnings, etc.
09:26:28jhMikeSNoone spoke about it anyway, which I take as implicit approval
09:27:52kugelI hadn't seen the gerrit ticket
09:28:09 Join rarog [0] (
09:29:11jhMikeSI did post it, but, I also sort of hit the wrong button and put it in before I wanted too (actually true). I said, "fuck it" and just fixed the errors. :\
09:29:38 Quit GeekShadow (Ping timeout: 265 seconds)
09:30:35kugelnot questioning the change but it would be easier to look at if it wasn't inter-mixed with boring const-ifying
09:30:46 Join PopeJo [0] (
09:31:09 Part PopeJo ("aaaaand ACTION!")
09:31:57 Quit fr3e (Ping timeout: 276 seconds)
09:32:42jhMikeSmostly it just follows the change from unsigned char * to const void *
09:34:32***Saving seen data "./dancer.seen"
09:36:45 Quit captainkwel (Ping timeout: 255 seconds)
09:43:08 Join pamaury [0] (
09:43:08 Quit pamaury (Changing host)
09:43:08 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
09:54:06 Join [Saint] [0] (~Saint]@
09:54:06 Quit [Saint] (Changing host)
09:54:06 Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
09:56:57 Quit OzzieJacks (Ping timeout: 252 seconds)
10:14:13 Join ender` [0] (
10:19:47 Join verbtim [0] (~verbtim@
10:20:27 Quit kadoban (Ping timeout: 260 seconds)
10:31:44 Join bertrik [0] (
10:31:45 Quit bertrik (Changing host)
10:31:45 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
10:40:45 Join evilnick [0] (~evilnick@rockbox/staff/evilnick)
10:41:54CIA-44Commit bcfa783 in rockbox by Jonathan Gordon: (Author: Osborne Jacobs) Set %cs(Current Screen) to "Bookmark browser" when listing bookmarks from the Context Menu
10:43:58[Saint]another %cs to worry about! :P
10:44:15evilnick[Saint]: On the 480x800 version, it loads (yay!) fine but whenever the user touches the screen (even to change track) then the now playing info disappears, but the popup doesn't appear
10:44:47evilnickSame for adjusting volume, the text vanishes (good) but the popup doesn't... pop up.
10:45:16CIA-44bcfa783 build result: All green
10:45:16evilnickAnd this is using the build in your Dropbox zip
10:45:24[Saint]Ok, thanks. I'll have a look.
10:45:48evilnickOnce installed over the top of the build I was running, then to check I uninstalled, deleted the Rockbox folder and did everything from scratch
10:45:52CIA-44Commit d5e676b in rockbox by Jonathan Gordon: (Author: Osborne Jacobs) Add a %cs(current screen) state for the "Track Info" screen
10:46:29[Saint]You're not helping the "massive bitmapstrip" thing here ;)
10:47:44[Saint]thanks evilnick, that's some mighty fine feedback there. I also appreciate the time you take to make sure that its not some weird conflict in existing files.
10:47:48[Saint]Its appreciated.
10:48:19*evilnick tips his hat to [Saint] and walks off John Wayne swagger style
10:48:24CIA-44Commit f1e1ab8 in rockbox by Thomas Martitz: (Author: Jean-Louis Biasini) Plugin snow PLA integration (code + manual)
10:49:21CIA-44d5e676b build result: All green
10:50:12bertrikjhMikeS, saratoga, I think we can close FS #11812, right?
10:50:12fs-bluebot Clip+ non zero output when paused (bugs, unconfirmed)
10:50:40kugel[Saint]: can't you do without bitmaps?
10:51:00kugelalso IMO the whole titlebar thing isn't needed at all :)
10:51:08[Saint]kugel: I can, but if its not going to be translated it might as well not look like crap.
10:51:35[Saint]Ideally, I'd like it translated...but that means adding a few lang strings.
10:51:39kugelwhy does it look like crap? with anti-aliased font it should look nicely?
10:51:53[Saint]Nah, not really.
10:52:16 Quit stoffel (Ping timeout: 272 seconds)
10:52:29CIA-44f1e1ab8 build result: All green
10:52:35[Saint]I mean, it doesn't look *terrible*, but the bitmaps are certainly nicer on the eye IMO.
10:52:58pixelmamaybe if you would draw the bar over the complete width without rounded corners, so it doesn't look weird when the title line needs to scroll
10:53:00kugelis it worth the screen space and memory it takes away?
10:53:25jhMikeSbertrik: I know I commented there on something but really I'm not sure. I have no Clip+ and don't see how anything I did could affect that.
10:53:35[Saint]screen space isn't an issue, IMO, there's plenty there...its also a button, and a notification area.
10:54:00[Saint]so, if I take it away, screen space gained is negated by having to add icons elsewhere.
10:54:24bertrikjhMikeS, I think the DC issue was fixed by fading audio with the mixer framework instead through the DAC volume control
10:54:37kugelI'm not convinced it looks worse with an anti-aliased font
10:55:38[Saint]<pixelma> maybe if you would draw the bar over the complete width without rounded corners, so it doesn't look weird when the title line needs to scroll <−−unless there's some *really* convoluted language I'm unaware of, it shouldn't ever scroll.
10:55:44kugelwith the new backdrop drawing tag you can draw the bar into the backdrop buffer and write text over it just fine
10:56:13[Saint]I could do that anyway...
10:56:15 Join DerPapst [0] (
10:56:21JdGordon[Saint]: german almost laways is longer than the avilable width
10:56:28JdGordonwhatever the text or screen is!
10:56:30 Join y4n [0] (y4n@unaffiliated/y4ndexx)
10:56:48[Saint]JdGordon: I did some rough checks when I looked at it initially, and all the strings used *should* fit.
10:56:52kugelbut yea it would be nice if the thing could be translated
10:56:54JdGordon:D whats german for "shuffle"?
10:57:22JdGordonmy "plan" for this is adding translation support to the skin but outside of the usual language system
10:57:54pixelma[Saint]: I thought you were using it for list title (potentially longer than the available width), and/or title of the now playing song (potentially longer)
10:58:04JdGordonthat way people wouldnt have exploded voice clips
10:58:12kugelJdGordon: the %cs titles are fine with the normal lang system
10:58:15jhMikeSbertrik: if it seems fixed, close it of course
10:58:23[Saint]pixelma: Nope, just "current screen" title, and a few notifications.
10:58:25 Join lebellium [0] (
10:58:41[Saint]like play state other than stopped, and screen locked, etc.
10:58:51JdGordonkugel: i disgaree
10:59:29[Saint]I actually agree with kugel here...*shock*.
10:59:54kugelthe screens (and their titles) are part of rockbox core, it's not strings random joe makes up
10:59:56[Saint]perhaps the merits of your idea haven't sunk in, but I saw it as a huge PITA.
11:00:02pixelmaJdGordon: "shuffle" is a bit unfair as it's not a literal translation (it's more like "shuffled playback")
11:00:32[Saint]wow...and I agree with kugel for the same reason too...
11:00:37*[Saint] looks for falling sky
11:01:18JdGordonthey are not used anywehre else and so *are* random strings some guy invented
11:02:26[Saint]Errr...what? 98% of the strings I'd use, if I used them, are in the lang files already.
11:02:37[Saint]So, if they're not used elsewhere, they're dead-weight.
11:02:41[Saint]and should be stripped.
11:03:39kugelif the default theme uses them it's official
11:03:52kugelit's not random strings by random joe then
11:05:02[Saint]there's only the one theme specific lang string (I think), If I had my way there'd be two...perhaps three.
11:05:11kugelI'm not against adding strings just for the default theme. the default theme is part of rockbox proper just like all the .c files and should ideally be translated
11:06:59 Quit verbtim (Quit: Leaving)
11:07:26[Saint]If not tonight (probably won't be, RL stuff), I'll have a look tomorrow night and see what string(s) I might need. There's only one I can think of immediately "Screen Locked", but I'm certain there's at least one more.
11:07:57kugelwe have "Keys locked" IIRC. that doesnt work?
11:08:29[Saint]Yeah, I saw that, but no...unfortunately. As the hardware keys remain functional.
11:08:32 Join n1s [0] (
11:08:33 Quit n1s (Changing host)
11:08:33 Join n1s [0] (~n1s@rockbox/developer/n1s)
11:08:37[Saint]Its only the screen that's "locked".
11:08:56[Saint](touchscreen specific, obviously)
11:09:43kugeli see
11:10:09 Quit rarog (Quit: ChatZilla 0.9.88 [SeaMonkey 2.4.1/20111011102430])
11:10:31[Saint]Fwiw, I did consider it. But I thought it would be really misleading.
11:11:27kugelJdGordon, Zagor: is there any point in re-syncing the dynamic lcd patch or should I wait for the lcd branch?
11:12:39JdGordondo we *want* dynamic lcd? i tohught we were more inclined to go with Zagor's scaling approach?
11:12:53 Quit perrikwp (Read error: Connection reset by peer)
11:13:28*jhMikeS wanted the ability to have additional framebuffer that may be larger or smaller than the display so I would say, in some indirect way, unequivocally, "YES"!
11:13:59JdGordonjhMikeS: well that is much trickier
11:14:03 Join perrikwp [0] (
11:14:59jhMikeSmaybe, but unpossible?
11:15:09kugelJdGordon: oh hmm
11:15:39kugelI'm not convinced dumb scaling is the best approach, but that makes dynamic lcd not needed indeed
11:16:27JdGordon"dumb" scaling might not give the best output all the time, but it sure is better than having to manually scale the same skins for every single target
11:16:39JdGordonand by taregt there i mean android device
11:16:44kugelthe dynamic lcd changes aren't so massive, might be worthwhile to do it as well (to explore more approaches)
11:16:47jhMikeSthinking beyond skins might be a good idea
11:17:40*n1s has trouble decoding "It proved the system work as unintended, even if slightly prematurely."
11:17:42kugelJdGordon: if it gives ugly output on one device we would need to work around, unless we want to keep it ugly
11:17:44JdGordonwell, dynamic lcd should be pretty trivial now
11:18:15JdGordonapart from the plugins that is
11:18:27lebelliumJdGordon : I got it!!! I can display the menu viewport full screen when the mini-player is OFF! Yes! Thank you for this skin engine improvement
11:18:29JdGordonthis is why i pushed the previous lcd change so quickly
11:18:35jhMikeSviewports with optional alternate framebuffer?
11:18:38JdGordonlebellium: :)
11:18:50jhMikeSn1s: me parse bad too
11:19:30*[Saint] *must* be missing something about lebellium's that was possible beforehand.
11:19:32JdGordonkugel: actually, i don't tihnk it would necessarily give bad output, if only the images are scaled it should work fine
11:20:13JdGordon[Saint]: no, he used an image to cover up part of the screen when the playback was off, but wanted th elist to use he full screen... it might have been doable but wouldnt have looked the same
11:20:27 Join verbtim [0] (~verbtim@
11:20:31lebelliumand now it's so awesome
11:20:38lebelliumI put the pic on background layer
11:20:38[Saint]Ah, that's right. the image being there screwed it up.
11:21:04*jhMikeS thinks a huge lcd change should be done with an eye towards possible use of hardware overlays and such
11:21:23*[Saint] sends jhMikeS to work on it then
11:21:36kugelJdGordon: images always look regressed after scaling (some times not as much but still)
11:21:57[Saint]SVG support in core!
11:21:58jhMikeS[Saint]: you implying I'm lazy and haven't done anything major? :)
11:22:01JdGordonregressed probably isnt the word you want, but yes I undertsand the sentiment
11:22:03kugeland I think our scaler isn't very high quality
11:22:38[Saint]jhMikeS: Yeah...I mean, its not like you completely reimplemented the playback engine or anything...sheesh.
11:22:42JdGordonkugel: Zagor wanted to use massive images to scale down so it wouldnt destroy the output so much
11:23:00pixelma[Saint]: svg scaled to pixels also only look good with a good algorithm which is *probably* lots of code too
11:23:12*jhMikeS never did that; it was some crazy dude that sleeps by the dumpster.
11:23:42[Saint]pixelma: Yeah, I know. jhMikeS's seperated the smiley from my comment.
11:23:47[Saint]'twas said in jest.
11:24:02evilnick[Saint]: Just FYI - using your 480x800 theme on a 1280x800 display is actually quite usable :)
11:24:15kugelmaking the source image huge doesn't really help
11:24:56kugelnot saying it will look ugly, but it will most probably don't look as nice as the pixel-perfect images we have now
11:25:03jhMikeS[Saint]: flash-based UI?
11:25:14JdGordonalso for skins, alot of the scaling shouldnt be the images at all, it shold be the viewports
11:25:26JdGordonwe dont necessarily want to resize th eplayback controls
11:25:56JdGordonso dynamic lcd size + scaling in the parser might be better again
11:25:56evilnickJdGordon: Yes, exactly that
11:26:03 Quit mshathlonxp (Ping timeout: 252 seconds)
11:26:18 Quit n1s (Read error: Connection timed out)
11:26:20[Saint]Yeah, and have sets of images for small/medium/large and pick from those groups based on the screen size.
11:26:37JdGordonthat will make the parser very coimplicated
11:27:23kugelI have another approach in my mind, define the skin a bit loose, i.e. with negative viewport values and transparent images that don't depend on the exact positioning on the backdrop, then one skin can cover a number of similar resolutions
11:27:47kugelthinking 480x800 and 480x854 (and all between) here
11:28:33[Saint]the three possible sizes in between? thats very specific.
11:28:45[Saint]Woops. I'm blind.
11:28:49[Saint]Carry on.
11:28:50kugelwe still need a number of themes (or perhaps generate from a template if the aspect ratio is similar enough), but not one for each resolution
11:29:10JdGordonthat is supported already
11:29:17JdGordonmaybe it needs to be tweked a bit
11:29:25kugelyes it is mostly supported
11:29:31 Quit evilnick (Read error: Connection reset by peer)
11:30:26kugelwhat's missing is that the same build runs on 480x854 properly on a 480x800 device (and vice versa) by adapting the frame buffer
11:31:33JdGordoneither way, at least some of the scaling work will be usable with whatever aproach we go
11:31:44JdGordonwe need the notification saying when the screen size changes
11:31:59JdGordonwe need to deal with inbuilt imaes better
11:32:06JdGordonplugins can go DIAF :)
11:34:01jhMikeS[Saint]: The problem with me is if I work on certain things, it will be too radical and this fine crowd is fairly conservative.
11:34:34***Saving seen data "./dancer.seen"
11:35:52[Saint]On application targets, yes.
11:35:55[Saint]Yes indeed.
11:36:23kugelthat's why small and iterative changes are better :)
11:36:30[Saint]I don't even know why the ones present are present.
11:37:06[Saint]kugel: You worked on "stopwatch" for RaaA I assume?
11:37:35[Saint]1: Good work with the UI, its nice. 2: its pretty broken.
11:37:53jhMikeSapocalyptic changes, it's too easy to be careless with small ones
11:38:31kugel[Saint]: it's lua. fix it! :)
11:38:33JdGordonjhMikeS: wont radical changes to the lcd system be rather wasteful on the non RaaA targets?
11:38:43[Saint]If you start the plugin, then press "stop" without actually initiating a countdown it increments an arbitrary amount to the counter.
11:38:46[Saint]kugel: ^
11:39:05[Saint]then you can keep pressing "stop" and incrementing random values to the count.
11:39:08jhMikeSJdGordon: not if they could be hidden in some way by preprocessing
11:39:18[Saint]Figure out what its adding, and call it a feature! :P
11:39:41JdGordonthough.. the mr500 does have multiple video buffers in hardwar ei tihnk
11:39:51jhMikeSon some targets, actual positionable framebuffer could allow overlays methinks
11:40:16jhMikeSpopup menus?
11:40:19JdGordonjhMikeS: well you can do that already anyway
11:40:28JdGordonhmm, maybe not
11:40:54*JdGordon should add something "copy forground to background" so popups would really work nicely
11:40:57JdGordonlike the splash
11:40:58[Saint]Not easily.
11:40:58jhMikeSbeing able to utilize graphics acceleration would be nice
11:41:10jhMikeSto do it some fairly general way, anwyay
11:41:17*JdGordon might even play with that tonight
11:42:08 Quit amithkk (Quit: Connection closed for inactivity)
11:43:06jhMikeSjust in say, mpegplayer, drawing the OSD without special crazy code would be nice
11:44:15JdGordonhmm, no need to do a massive memcpy... just switch the front buffer to the back buffer before entering splash, then draw as normal
11:44:34JdGordonthis should be nice and simple *if* we can grab a big enough buffer
11:44:47kugelI thought about that long ago, but was afraid of the additional memory needed
11:45:37JdGordoncan buflib_alloc do non-restarting playback allocs?
11:45:55 Join Zarggg [0] (
11:45:58kugelhow would that work?
11:46:14jhMikeSif it already has a proper allocation gap, I think it might
11:46:18kugelwell it can of course do that if enough memory is unallocated
11:47:49kugelaudio does maximise its buffer after stopping playback, though
11:49:45kugelbuflib hasn't changed our fundamental desire to use all memory for audio buffering :)
11:50:22kugeltaking audio buffer away while playing is probably possible but it's tricky to implement
11:51:10JdGordonkugel: so are oyu going to resync dynamic lcd then?
11:51:31jhMikeSresizing it dynamically shouldn't be the worst undertaking. trying to relocate bunches of stuff at the wrong time is ill-advised
11:51:40 Join TheLemonMan [0] (
11:59:29 Quit verbtim (Quit: Leaving)
12:06:01CIA-44Commit 10c566e in rockbox by Thomas Martitz: (Author: Jean-Louis Biasini) Plugin logo PLA integration (code + manual)
12:06:37kugelJdGordon: think so
12:08:58CIA-4410c566e build result: All green
12:09:13bertrikkugel, nice to see how the code size gets reduced with the PLA changes :)
12:10:56CIA-44Commit a92696d in rockbox by Thomas Martitz: (Author: Jean-Louis Biasini) Plugin Starfield PLA integration (code + manual)
12:11:25kugelbertrik: only the source code size :)
12:11:37kugelthe plugins code size is probably even increased due to linking pla code
12:13:03jhMikeSkeyclick_click needs to be restored to work with raw key reading
12:13:14 Quit randumbum (Ping timeout: 260 seconds)
12:13:52CIA-44a92696d build result: All green
12:18:03 Join randumbum [0] (
12:19:36lebelliumJdGordon: I have an issue though. The progress bar viewport background is white, it won't let the default background (grey pixels at the bottom on my background pic) anymore, I don't understand why
12:22:14 Join n1s [0] (
12:22:14 Quit n1s (Changing host)
12:22:15 Join n1s [0] (~n1s@rockbox/developer/n1s)
12:23:00JdGordonlebellium: you need to show me the skin
12:23:05JdGordon[Saint]: piiiing!
12:23:38JdGordonsoo, i started thinking about relative viewport positioning in skins, how does this sound...
12:24:36JdGordon%V(%Vr(other viewport label, 0), 10, -, -)... basically %Vr() specifies another viewport and a number, that could be used for any of the 4 position params (or all 4)
12:25:02JdGordonso that example says "put it at the same x pixel as the other viewport, 10 pixels from the top of the screen, and the rest of the screen
12:25:12JdGordonyes its messy, but would be useful
12:25:23JdGordonthen we get scaling for free almost
12:25:55JdGordonlebellium: are you drawing the progressbar onto the backdrop buffer?
12:27:38lebelliumJdGordon: nope, I draw the progressbar onto the backdrop buffer I get the right background but the progressbar doesn't display then :S
12:27:40JdGordonkugel: ^
12:27:48lebelliumif I*
12:27:52*pamaury has a working prototype for moving dualboot out of mkimxboot
12:28:13pamauryshortstat: 33 files changed, 2116 insertions(+), 1054 deletions(-) :(
12:29:18jhMikeSnice, that way you know you're getting stuff done :)
12:29:42JdGordonnot with 2:1 added lines!
12:30:02pamauryand that's doesn't include the mkamsboot changes, which will be short since I already moved all the files needed
12:31:03kugelJdGordon: %Vr would something different depending on where it is in the list?
12:31:43kugelsounds ugly
12:31:47JdGordonkugel: not really, whatever param it replaces it makes that one relative to the other viewport
12:31:53pamaurythere are ~600 lines to create a new tool in tools/ to deal with a new bootloader format (I preferred that over using zip). However there are Makefile changes :'(
12:32:00JdGordonand yes, it is ugly, any better ideas?
12:32:51jhMikeSJdGordon: is there some reserved action bit I can use instead adding another param to keyclick_click (I hate params for the most part)
12:33:06jhMikeSneed to tell it to read raw vs. action codes
12:33:21JdGordon1<<30~ probably works
12:33:31JdGordonim sure there is a spare bit in the top 16
12:34:38kugelJdGordon: what does that really give?
12:35:34kugelthe if the referenced vp has fixed values, the referencing one can just as well have fixed values
12:36:01JdGordonthe idea is that you'd make a theme with no fixed references
12:36:16JdGordonthen it would just scale to the screen automatically
12:36:28jhMikeSthose bits are removed before keyclick_click?
12:36:28JdGordonthough i think we need to support %ages also
12:36:30kugelhow would it scale?
12:37:09JdGordonmaybe it isnt necessary
12:37:50JdGordonim just trying to figure out how you'd make a themes viewports move in a sensible way when you want to use it on a much larger display
12:37:59JdGordonsome viewports you want more or less fixed to another one
12:38:04JdGordonjhMikeS: ?
12:39:12kugelI don't think having vp1 at a fixed offset of vp0 which is at a fixed position helps there
12:39:52kugelit may make sense if vp0 could move/resize dynamically
12:42:08kugelpixelma: you want g#168 committed or commit yourself?
12:42:09fs-bluebotGerrit review #168 at,168 : Manual: Fix misaligment in the 'Images' theme tags table. by pixelma (changes/68/168/1)
12:42:58 Quit Keripo1 (Quit: Leaving.)
12:43:34n1skugel: she said she wanted to do it herself, just pushing through the gerrit review stuff to get used to it
12:46:48kugelso many failed hunks
12:48:00jhMikeSJdGordon: hmm, it needs to be free for button codes and actions
12:52:59 Join jlbiasini [0] (
12:55:56*jhMikeS reckons the best solution is to just add a context parameter
13:05:24n1sis anyone against wtachi's approach for librbcodec?
13:05:49n1si'm tempted to pull the trigger and start pushing parts to see if that will get the ball rolling
13:06:29 Join Rower85 [0] (
13:08:32kugeln1s: perhaps we should review the patches one by one, and push if it's not too bad
13:09:05kugelI briefly looked at the first one. it just adds a program, like checkwps, that has lots of hackery in it to get codecs running :)
13:11:08 Join dfkt [0] (dfkt@unaffiliated/dfkt)
13:11:34 Join anewuser [0] (~anewuser@
13:11:34 Quit anewuser (Changing host)
13:11:34 Join anewuser [0] (~anewuser@unaffiliated/anewuser)
13:11:49 Join amithkk [0] (u4289@2buntu/writers/amithkk)
13:13:18CIA-44Commit f688710 in rockbox by Michael Sevakis: Change keyclick_click so that it may accept raw buttons or actions.
13:14:51jhMikeSor should I have done it the other way? damn
13:16:08jhMikeSprobably not
13:16:21n1skugel: yeah i've started looking at the first patch too, i think hackery is fine at this point as it's a test program. for a proper library that should obviously not be needed
13:16:46JdGordonjhMikeS: yuck! context is the wrong word there
13:16:49n1sif i have time later today ot tomorrow i might push that then
13:16:55kugelJdGordon: just thought the same
13:16:58CIA-44f688710 build result: All green
13:17:16JdGordonjhMikeS: a bool would have been better
13:17:27kugelwhy not have a keyclick_rawclick() or so
13:17:30kugelor a bool
13:17:50jhMikeSick, i hate bools too
13:18:04jhMikeSone whole param wasted on just a bit
13:18:05kugelwhat's wrong with bools?
13:18:57JdGordonjhMikeS: yes, but your context param is just wrong
13:19:04JdGordoncontext in actions has a special meaning
13:19:05jhMikeSthey also aren't generally obvious upon looking at usage of the function itself in the code: foo(true, true, false, false, ...);, I don't mind things like enable_foo(true); though
13:20:49jhMikeSsomehow I feel it was appropriate looking at the other usage of context bits
13:21:16JdGordonwhat were you looking at?
13:21:26JdGordonthe context stufff is not bits
13:22:18JdGordonCONTEXT_* is a mess :(
13:22:41jhMikeSwell, why are they used as bitsmasks? :)
13:25:22JdGordonok, 1) can you revert that change please, 2) what are you actually trying to do there?
13:27:00jhMikeSI want it to read raw keys OR the actions
13:27:59jhMikeSand as afaics it shouldn't disable the use of the callback
13:28:30jhMikeSI'm not certain I want to blow an extra param on something so specific.
13:28:42jhMikeSand have it forever be just that
13:29:07JdGordonyour use of context there is wronge
13:29:22jhMikeSand Rush "Tom Sawyer" is playing on the TV ... "changes aren't permanent, but change is..." :)
13:29:41jhMikeSok, ummm....goddammit
13:29:50JdGordondoes mpegplayer set a callback?
13:30:12JdGordonkeyclick_click(0, action) is just as magic as keyclick_click(false, action)
13:30:16jhMikeSnot yet, because the mech. didn't exist at the time it could use keyclicks
13:30:30JdGordonso why didny ou always just pass in ACTION_STD_OK?
13:30:33JdGordonthat would have worked fine
13:30:44JdGordonoh hmm, maybe not
13:31:27*JdGordon will change the CONTEXT_ bitmasks
13:32:20JdGordonactually no i wont, im lazy
13:34:03jhMikeSfine, I'll boolify it instead. if anything better or more needs come about, np to just change things
13:34:37***Saving seen data "./dancer.seen"
13:35:19JdGordonthank you
13:40:16 Join enthdegree [0] (~BitchX@unaffiliated/enthdegree)
13:53:54JdGordondfkt: if scrolling lines are eating the backdrop image you're not doing the viewports correctly :(
13:53:58JdGordonit is a bit of a PITA
13:54:20CIA-44Commit d05db0a in rockbox by Michael Sevakis: Correct actionable offense for misappropriation of action context.
13:57:45CIA-44d05db0a build result: All green
13:57:49*[Saint] disagrees
13:58:14[Saint]its more of a PITA to do it incorrectly, doing viewports "right" is often a lot less hassle/code.
13:59:06 Quit enthdegree (Ping timeout: 272 seconds)
13:59:59*gevaerts is of the opinion that if scrolling lines still work you're not showing enough creativity in your abuse of the skin engine :)
14:01:19dfktJdGordon, scrolling lines were my first attempt with the backdrop buffer, in the meantime i got better ;) ... but that's no use anyways, since fonts would need outlines or shading for that to work :)
14:02:07[Saint]wait...what? why are you drawing text in the backdrop's buffer?
14:02:27dfktalbum art in the backdrop, text over it
14:03:28JdGordoni'd have to see your skin, text erasing the AA image was an issue i was hitting with my test skin but i fonud a fix
14:03:51dfktthat issue was before i knew you had to have a fullscreen viewport to make it work
14:04:12JdGordonso its working now?
14:05:17dfktwell, i ditched that idea since fonts aren't readable without outlines - but i can revert to that state and check
14:05:28[Saint]this reminds me I need to put up a bug report against %Tl
14:05:40[Saint]re: eating images
14:12:49 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
14:15:49 Quit perrikwp (Read error: Connection reset by peer)
14:16:54 Quit ender` (Quit: drug, n: A substance that, injected into a rat, produces a scientific paper.)
14:17:02 Join perrikwp [0] (
14:18:30dfktJdGordon, nope, still munching away on the images - what is your fix?
14:19:03 Join [Saint_] [0] (~Saint]@
14:19:04 Quit [Saint_] (Changing host)
14:19:04 Join [Saint_] [0] (~Saint]@unaffiliated/saint/x-8516940)
14:19:54 Quit [Saint] (Ping timeout: 255 seconds)
14:20:22JdGordondfkt: send me your theme
14:21:58 Join ender` [0] (
14:22:02dfktJdGordon, the theme is here -
14:22:32dfktand i just put a scrolling line in the wps for testing, here's the wps with scrolling over album art -
14:22:52dfkt(lines 64 - 66)
14:23:46 Nick [Saint_] is now known as [Saint|Zzz] (~Saint]@unaffiliated/saint/x-8516940)
14:24:39 Join enthdegree [0] (~BitchX@unaffiliated/enthdegree)
14:24:55 Quit enthdegree (Read error: Connection reset by peer)
14:38:53jlbiasinikugel: sorry for the mess. the sad part is that I do test all changes on uisimulator before commiting. I guess I should test myself even before that. The patches should be ok now!
14:51:49kugeljlbiasini: okay. will have a look in a few minutes
14:53:29 Quit lebellium (Read error: Connection reset by peer)
14:53:58 Join lebellium [0] (
14:57:17 Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
15:00:44lebelliumJdGordon: when you have could you please look at my theme? I've been working for 3hrs on it today but can't find where is the problem, I don't know if this is a skin engine bug or if I did something wrong. I can't get the right progressbar background :S
15:01:46JdGordonwhat atrget?
15:02:23lebelliumwhen you have time*
15:02:33lebelliumyou can use this simulator
15:02:42lebelliumsince there is no R0 simulator
15:03:25kugelwe should have one
15:04:33 Join verbtim [0] (~verbtim@
15:09:00JdGordonlebellium: what are you trying to do?
15:09:09JdGordonwhat should the pb's backdrop look like?
15:09:12JdGordonand in the sbs or wps?
15:09:37 Quit randumbum (Remote host closed the connection)
15:09:57lebelliumJdGordon: the WPS is fine, just look at the SBS. The background for the progress bar should be the general background pic (lebellium_Samsung-like3.bmp)
15:10:23lebelliumatm it's white for an unknown reason
15:12:01JdGordonbecause you havnt done the viewports correctly there
15:12:20JdGordonyou need a viewport on the top layer covering that whole area so the backdrop is drawn
15:12:38JdGordonbackdrop viewports are not drawn unless a front viewport draws them
15:12:53JdGordonso you shold always do %V(0,0,-,-,-) after your backdrop viewport
15:13:02*JdGordon inteneded on writing an email about this but hasnt yet
15:13:18JdGordonerr... i see you've done that
15:13:20JdGordonbugger :p
15:13:29*[Saint] doesn't get it...
15:13:46[Saint]Isn't the backdrop viewport a fullscreen viewport in and of itself?
15:14:02[Saint]What's the need to define a "dummy" fullscreen VP?
15:14:11JdGordonyes, but it replaces the backdrop image, it is never actually drawn onto the hardware
15:14:23JdGordonlebellium: why are you drawing text on the backdop viewport?
15:14:43lebelliumJdGordon: I did that?
15:14:48JdGordonah, no, those lines are commented out
15:15:03lebelliumJdGordon: sorry it's a draft so not so clean :S
15:19:05 Join randumbum [0] (
15:19:08 Quit verbtim (Quit: Leaving)
15:20:09lebelliumJdGordon: If I delete "Hide Background" the progressbar has the right background but then when the playback is stopped the mini-player still has a white background
15:21:20JdGordonyou're doing something wrong, but im a bit too sleepy to figure it out :p
15:21:30JdGordon(or possibly im doing something wrong in the code... ;p)
15:21:49*JdGordon didnt realie %Vl(k,56,293,180,20,4)%Vf(000000)%s%id %?iy<%(%iy%)|> on one line owuld work!
15:22:37lebelliumWell no problem, you can have a look at it later when you're not sleepy anymore^^
15:22:44kugelJdGordon: I think in the past %V lines needed to be on its own line
15:23:10JdGordonI suspect it migh have something to do with the %>mp line being in a backdrop viewport, though i dunno why
15:23:27[Saint]They still should be, that's a nightmare to read ;)
15:24:13*[Saint] makes it a habit of putting "declare stuff" calls on their own line.
15:25:44JdGordonlebellium: you do know that viewports can share labels right?
15:26:03lebelliumwhat do u mean with that?
15:27:03[Saint]If you have several viewports all active at once, or all active on the same conditions etc. they can share a label.
15:27:35JdGordonlebellium: your 3 id3 info viewports should all be a single viewport
15:27:36lebelliumI guess I did not know
15:27:46JdGordonor if you really want different Vp's they should share a label
15:27:57*JdGordon really needs to do a proper skinning document
15:28:42lebelliumfor the id3tags I guess I did that because if they are in the same viewport on 3 different lines, some letters (y,p etc) have the bottom cut off
15:29:10 Join [Saint_] [0] (~Saint]@unaffiliated/saint/x-8516940)
15:30:31JdGordonI tihnk i just need to redo this from scratch and let you add in the extra stuff once its working
15:31:06lebelliumthat would be nice
15:31:23lebelliumit's a bit messy now
15:31:28 Quit [Saint] (Ping timeout: 244 seconds)
15:31:28 Quit [Saint|Zzz] (Ping timeout: 255 seconds)
15:32:52lebelliumI think there is a bug in the skin engine code because after looking for 3 hrs I don't see any valid reason why the progress bar doesn't keep the right background but it's possible I misunderstood some tags.
15:33:35gevaertsIt would be nice if you try to isolate the problem, throwing away all features that aren't necessary to show the issue
15:34:02lebelliumindeed, I could delete all the status bar items
15:34:16lebelliumthey are useless for testing
15:34:40***Saving seen data "./dancer.seen"
15:34:41gevaertsOh, and three hours is *not* enough to prove it's a bug in the skin engine. Figuring out what was wrong in the previous version of Wavy took me longer than that :)
15:34:55lebelliumIt doesn't prove
15:35:00lebelliumIt's my personal guess^^
15:36:53lebelliumIt's my personal guess^^
15:37:00lebelliumoops err
15:37:55gevaertslebellium: I don't understand the viewports under "hide background" and "white backround"
15:38:12JdGordon@reboot is the correct crontab line thingy right?
15:38:57gevaertsJdGordon: it's *a* correct crontab line thingy :)
15:39:00JdGordonapparently its not working
15:39:09 Join [Saint__] [0] (~Saint]@
15:39:09 Quit [Saint__] (Changing host)
15:39:09 Join [Saint__] [0] (~Saint]@unaffiliated/saint/x-8516940)
15:39:44lebelliumgevaerts: I thought the Hide background wouldn't be necessary but if I don't put it, the mini-player keeps white background on stopped playback. The white background viewport is to draw a white background for the mini-player
15:41:13 Join fml [0] (
15:41:16*gevaerts also recommends not using one-character labels
15:41:39gevaertsThey might be a tiny bit faster when rendering, but they don't help reading
15:41:49fmlHello. gerrit admins here? I'd like to be added to the gerrit committers group.
15:41:50 Quit [Saint_] (Ping timeout: 255 seconds)
15:42:25lebelliumgevaerts: sure my theme is a pain to read :S
15:43:14gevaertslebellium: commenting out the %V line under "Hide background" doesn't seem to change anything here
15:43:47JdGordonfml: you need to hit up Zagor, Bagder or Torne
15:44:16jlbiasinilebelium are you actually talking about FS #12507 or just some close topic? deleting hide_background image solve the problem for the bug (no hang anymore)
15:44:16fs-bluebot lebellium Samsung-like Theme causing rockbox crash on usb connection (bugs, unconfirmed)
15:44:30fmlZagor, Bagder, Torne: please add me to the gerrit committers group
15:45:42gevaertslebellium: correct me if I'm wrong, but it looks like you're putting %xl in specific viewports to make then load at the right offset. That won't work
15:46:12lebelliumgevaerts: if you comment out the next string too (%xl ...) and delete %xd(K) in the next %?mp... string then when there is no music playback there is still a white background.
15:48:08lebelliumjlbiasini: no actually I'm trying to include the new skin engine improvements in my theme to be able to display the menu viewport full screen even when the mini-player is OFF
15:48:35jlbiasinilebellium: ah ok
15:48:40soapas we approach the two month mark is there any word on the SVN mailing list becoming GITpatable?
15:50:40lebelliumgevaerts: you mean that instead of %V(...) then %xl( ...) I should do %Vl(...) then %x(...) ?
15:52:00gevaertslebellium: I mean %xl isn't related to viewports at all
15:52:40 Part jlbiasini
15:52:50lebellium%xl doesn't display in a viewport?!
15:52:59lebelliumI mean %xd
15:53:12gevaertsSure, %xd does
15:53:19gevaertsBut %xl just loads
15:53:49lebelliumI thought it loads in a viewport and %xd calls it
15:55:32lebelliumok so maybe there is an issue there
15:56:54*gevaerts is playing around with it
15:57:40 Quit fml (Quit: ChatZilla 0.9.88 [Firefox 10.0.2/20120216080748])
15:58:37lebelliumhey gevaerts!
15:58:41lebelliumthat was the issue indeed
15:58:52lebelliumdoing that instead
15:59:45lebelliumnow the progressbar has the right background
15:59:59gevaerts%xl still doesn't *have* to be there :)
16:00:42lebelliumwell I don't know what's the rule, it seems that some people put all %xl at the beginning of the file
16:00:47lebelliumI never did that
16:01:08gevaerts%xl can be anywhere you like
16:03:18 Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
16:11:00lebelliumnow trying to understand why the menu viewport disappears when it skips to the next track and reappears when pressing any key
16:16:16 Quit randumbum (Ping timeout: 260 seconds)
16:20:18[Saint]Sounds again like viewport issues.
16:20:59[Saint]Whether its your issue or an internal one I'm not sure, but, ...yeah.
16:24:54gevaertslebellium: for the background I'd do something like
16:25:13gevaertsAnd then no other %VB
16:25:44gevaertswith back.bmp being 240x250 and the two others being 240x70 in this example
16:26:53lebelliumthanks, I try that
16:27:27gevaertsOr of course, if you don't mind wasting some RAM, just
16:27:35gevaerts(which works with the images you already have)
16:27:56[Saint]That's better, that first suggestions was far too sane for you gevaerts ;)
16:28:39gevaertsThe secret of getting skins to work is always to check if you're not trying too hard and simplify the structure :)
16:29:21gevaertsThe secret of getting skins to work is always to check if you're not trying too hard and simplify the structure :)
16:29:28gevaertsWell, it's a good first step
16:29:53gevaertsAnd if that fails, cheap
16:29:57[Saint]See, that first suggestion is the more.complicated, but "right" (IMO) solution.
16:30:31gevaertsWell yes, but both are simplifications of the current code that tries to draw a background and then draw other stuff over it
16:31:12[Saint]I haven't looked at the current code. But, done correctly, that should work fine.
16:31:25lebelliumI tried the 2nd for now and it seems to work fine
16:31:45lebelliummaybe I really think to difficult while it could be easy^^
16:32:02gevaertsAnd by "cheat" I mean things like the medieval theme having the progress bar actually in the backdrop bmp, and having the progress bar bmp be the "background" :)
16:32:05 Join randumbum [0] (
16:32:38[Saint]gevaerts: I did exactly that with iLike.
16:32:48gevaertslebellium: yes. I'd suggest using that for now, and optimising RAM later on when everything works
16:34:29[Saint]Except I had the same bar doing both the volume and progress bar, and changed out the bar viewports conditionally as they were different sizes. To make it "work" I just draw over the excess bar background with sections of image to match the backdrop.
16:34:41[Saint]It was a horrible horrible hack.
16:34:54n1sthat's the best kind!
16:35:07gevaertslebellium: the sbs looks like after some playing with it
16:36:06[Saint]Oh God! What are you using my horrible DroidSans fonts?
16:37:19gevaerts[Saint]: that doesn't sound as horrible as my first "wavy" implementation :)
16:37:30[Saint]lebellium: you'll thank me for this...
16:37:49[Saint]*much* better fonts.
16:37:50CIA-44Commit 693af46 in rockbox by Thomas Martitz: (Author: Jean-Louis Biasini) Plugin mosaique PLA integration (code + manual)
16:38:07lebelliumgevaerts: thanks, i'll compare with mine. With my SBS I still have the menu viewport disappearing when skipping to next song if you're in the menus at this moment.
16:38:54lebellium[Saint]: sorry I did know you did that, I'll give you credits if necessary. Actuallly I don't find it horrible at all
16:39:14lebelliumI did not*
16:39:44gevaertslebellium: I haven't done much yet. Mainly some cleaning up of viewports (like the backdrop), and changing some of the labels to more readable ones (and unifying some of them)
16:39:48[Saint]They're pretty bad. The ascent/descent is completely wrong between the different sizes.
16:40:28lebellium[Saint] as of today I did not know see better font for my theme, I tried all defaut ones from the RB font pack
16:40:35lebelliumI did not see*
16:40:44[Saint]My "Ubuntu" series is a lot nicer.
16:40:58[Saint](Linked above)
16:40:59lebelliumI'll have a look
16:41:12CIA-44693af46 build result: All green
16:41:50CIA-44Commit 26fc31a in rockbox by Nils Wallménius: (Author: Sean Bartell) Add the warble test program.
16:42:14CIA-44Commit 36c84f9 in rockbox by Thomas Martitz: (Author: Jean-Louis Biasini) Plugin Matrix PLA integration (code)
16:44:40CIA-4426fc31a build result: All green
16:45:23 Quit DerPapst (Quit: Leaving.)
16:46:40CIA-44Commit 8eb36cf in rockbox by Thomas Martitz: matrix: shorten statement.
16:47:16CIA-4436c84f9 build result: All green
16:50:00CIA-448eb36cf build result: All green
16:50:31 Join [Saint_] [0] (~Saint]@unaffiliated/saint/x-8516940)
16:50:56lebellium okay gevaerts, your modifications look nice, I'll work on this basis
16:51:01lebelliumthank you
16:52:35 Quit [Saint] (Ping timeout: 244 seconds)
16:52:57 Quit [Saint__] (Ping timeout: 255 seconds)
16:53:41pixelmaI wonder if some targets will use different buttons than before for those PLA-ified plugins, might be worth mentioning (and I know that this would mean lots of comparing, maybe the resulting old and new manuals would be best for that, plus some simulators)
16:56:42lebelliumgevaerts: the only visual issue I still see is this disappearing menu viewport. I don't have this issue with my clip zip theme hum
16:58:13n1sif we want to keep this warble program working i guess we should add it to the buildsystem, do we autobuild the database tool?
16:58:25kugelpixelma: you mean a PITA and lots of work?
16:58:30kugelno, thanks :)
16:59:21 Quit pamaury (Remote host closed the connection)
16:59:33gevaertsn1s: we autobuild the database tool for one HWCODEC and one SWCODEC target
17:00:16gevaertsn1s: can it build in a separate directory, like other builds?
17:00:17n1sgevaerts: i think we should do the same for warble, then
17:02:06 Join CaptainKewl [0] (
17:02:16n1si'm not sure but it doesn't look like it
17:03:08gevaertsIt doesn't look too bad
17:03:39n1smy Make fu is weak
17:04:07pixelmakugel: better than people complaining that xyz doesn't work anymore (even though it might only be on a different button now). And I always try to find that out before if I change something and would hope others do so too as this means others don't have to care about my changes. But it seems other people think differently unfortunately :\
17:04:13gevaertsIt doesn't do what I'd like it to do out of the box, but I think it doesn't need too many changes
17:05:17n1sit does get some format warnings from debug statements in buflib.c since it defines DEBUG i guess those should be fixed too, does DEBUGF end up in printf at the end or is there something homegrown somewhere?
17:07:07kugelpixelma: no it's not better in this case IMO
17:07:38kugelI don't expect anyone to complain or even notice because this were all demo plugins, with 1 or 2 buttons and which nobody uses
17:07:44 Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
17:08:09kugelcarefully examining the button changes for every single target is not worth the effort
17:14:46kugelpixelma: I understand your concern, but in this case I wouldn't worry too much. This way good work got actually done
17:18:08n1swe should just get rid of those demo plugins if noone cares about them
17:19:36[Saint]Isn't the usual workflow to let it rot until it fails to build and someone's forced to look I to it? ;)
17:19:58[Saint]*to look at it.
17:21:03n1sand then realize it draws falling snowflakes in monochrome?
17:21:46 Quit antil33t (Ping timeout: 272 seconds)
17:22:29[Saint]Even though I jest, I don't think they should be chucked out. If "not caring about foo" was a consideration for removing foo, there'd be very little Rockbox left.
17:25:49[Saint]Charcell and hwcodec are excellent examples of this in action :)
17:26:33pixelmaI also like work that makes sure a change doesn't hurt others (even if it is just a possibility of some small support issues). I find that nice towards others in a community project
17:27:02kugelpixelma: it means work doesnt get done if that is too much of a burden
17:27:58[Saint]s/doesn't get done/doesn't get done at the pace you'd like/
17:28:29kugelno, the first one was correct
17:28:37pixelmaI'm caring about hwcodec btw. [Saint]
17:28:38 Join jlbiasini [0] (
17:28:44 Join kadoban [0] (
17:29:26kugelpixelma: a number of people care about hwcodec but nobody works on them
17:29:41[Saint]pixelma: exactly, and the fact you do is awesome. My point isn't to remove them, but rather that the fact they might be annoyingto others isn't a reason to remove them.
17:29:51[Saint]Crossed wires, I think.
17:32:05 Join pamaury [0] (
17:32:06 Quit pamaury (Changing host)
17:32:06 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
17:34:15[Saint]I'd like to fork them out...but, I don't want it to happen because even though people might say development will continue, or fixes/features will be backported where possible, that's not "interesting" and people only work on "interesting things".
17:34:41***Saving seen data "./dancer.seen"
17:35:48pixelmaI think leaving things to clean up to others just so you are done earlier could also piss possible contributors off, which could also mean other work is not being done
17:36:38pixelmain this PLA case, I'd tested and I don't think that's a big "burdon". A bit time-consuming, yes, but other than that
17:41:22kugelthere's no cleanup left to others
17:41:39pixelmaof course, there never is
17:41:59kugelthe opposite is true, now the maintainance burden for these plugins is reduced and it's a lot easier to port plugins to new targets
17:42:25jlbiasinithats the spirit
17:47:47n1si still don't see the point of the demo plugins but whatever
17:48:37[Saint]History, if nothing else.
17:57:03 Quit perrikwp (Ping timeout: 244 seconds)
18:08:10 Quit benedikt93 (Quit: Bye ;))
18:09:36CIA-44Commit 71953a4 in rockbox by Frank Gevaerts: Don't redirect file IO to app_* for __PCTOOLS__ builds.
18:12:38CIA-4471953a4 build result: All green
18:15:19kugelgevaerts: I think the database tool wants the redirection
18:16:26 Join Keripo [0] (
18:17:35gevaertskugel: hm
18:17:39gevaertsit does indeed
18:18:37*gevaerts never managed to fully understand these app_ and sim_ things
18:21:06gevaertsThis doesn't make sense!
18:21:39gevaertskugel: it's not that simple
18:22:11gevaertsthat commit doesn't break database builds for native targets, and database builds for app builds seem to be broken already
18:23:38CIA-44Commit f7a4ee2 in rockbox by Frank Gevaerts: Don't redirect file IO to app_* for __PCTOOLS__ builds for dir functions.
18:24:37gevaertsOr do you mean it breaks it at runtime?
18:25:07kugelI don't mean compile failure
18:25:46kugelI think it expects to be executed in dir structure like on the dap i.e. the current folder has .rockbox and the music folder in it
18:26:07kugelit calls the tagcache function which run through "/"
18:26:32CIA-44f7a4ee2 build result: All green
18:26:56kugeland it needs to do it like that so the paths are right when the files are transferred onto the dap
18:26:56gevaertsMy commits only touch app builds though
18:27:16dfkt[Saint], did you convert the android 4.0 roboto fonts to fnt, or know if someone already did?
18:27:24 Join DerPapst [0] (
18:27:25lebellium[Saint] 18 Ubuntu [bold] is not bad but I really like the Droid fonts too :p
18:27:35gevaertsMaybe those need a different fix than what I did, but it's a bit confusing
18:27:42[Saint]dfkt: no and no.
18:27:45kugelgevaerts: I see. which APPLICATION defines __PCTOOL__?
18:27:58dfktok, i might give them a go, they're quite nice
18:28:27gevaertskugel: I'm working on integrating this warble thing into our regular build system
18:28:45kugeldoes warble define APPLICATION?
18:28:52gevaertsWell, yes and no
18:28:52 Join T44 [0] (
18:29:17kugelthat sounds wrong
18:29:20gevaertsIts makefile is a bit hackish to make things work
18:30:14gevaertsThe way our code is structured, you basically have to pick a target, and I'd like it to work when picking *any* target, which includes apps
18:31:06kugeli see
18:31:17kugelthen APPLICATION might be the best choice actually :)
18:32:04gevaertsNext one up: firmware/export/audiohw.h defines VOLUME_MIN and VOLUME_MAX for PLATFORM_SDL, except for simulators
18:32:37gevaertsThat means it doesn't define them if you set SIMULATOR for an SDL build, and setting SIMULATOR is common for those PCTOOL things, so this breaks :)
18:32:44 Quit Topy44 (Ping timeout: 276 seconds)
18:33:10 Join kevku [0] (~kevku@2001:470:28:773::)
18:34:58gevaertsThe reason they set SIMULATOR seems to be "sim_filesize"
18:37:41kugelgevaerts: VOLUME_MIN/_MAX comes from the target in sim builds
18:38:07*gevaerts nods
18:38:21gevaertsI think I have a proper solution for that one
18:38:49gevaertsMake sure sim_filesize() is called filesize() for __PCTOOL__
18:39:10gevaertsThat makes the -DSIMULATOR unnecessary (also for the database tool)
18:43:39lebelliumHere is the beta2. If someone understands why the menu viewport disappears when skipping to the next song, I would be very happy
18:46:42 Quit kevku (Quit: KVIrc 4.0.4 Insomnia
18:46:43bluebrother^pixelma: it's always this attitude of people not involved in support issues :(
18:46:53 Join kevku [0] (~kevku@2001:470:28:773::)
18:47:06bluebrother^wouldn't be much of a problem if we had proper tests ...
18:47:14kugelare you saying I'm not doing support?
18:47:29 Quit kevku (Client Quit)
18:48:19bluebrother^I'm saying that the people that are not (or only remotely) involved in support issues pretty often show this attitude.
18:48:38 Join kevku [0] (~kevku@2001:470:28:773::)
18:48:41bluebrother^and for that I just need to look who is actively answering support questions in places like the ML and forums.
18:48:55bluebrother^and as far as I can see most support questions are handled in the forums
18:49:01bluebrother^at least *user* support
18:49:16bluebrother^supporting (potential) developers is another thing
18:51:42 Join evilnick [0] (~evilnick@rockbox/staff/evilnick)
18:52:25 Quit factor (Quit: Leaving)
18:52:27kugelright, I don't usually read the forums
18:57:53CIA-44Commit c6d2010 in rockbox by Frank Gevaerts: Fix __PCTOOL__ dependencies on SIMULATOR
18:59:37 Join lebellium_ [0] (
19:00:26gevaertshm, I did break the database tool...
19:01:46CIA-44c6d2010 build result: 173 errors, 0 warnings (Frank Gevaerts committed)
19:02:43 Quit lebellium (Ping timeout: 260 seconds)
19:02:48 Nick lebellium_ is now known as lebellium (
19:04:13n1sgevaerts: very nice of you to look into this stuff
19:04:28gevaertsn1s: the warble thing is easy :)
19:04:34gevaertsIt's everything else that's messy!
19:05:34n1syeah the database tool is a bit of a hack too
19:06:12CIA-44Commit 52724c1 in rockbox by Frank Gevaerts: filesize() isn't prefixed anymore for sims.
19:06:55kugelit's a large bit of a hack :p
19:09:32CIA-4452724c1 build result: 6 errors, 3 warnings (Frank Gevaerts committed)
19:11:44jlbiasinikugel: g158 might be ok now.
19:11:46fs-bluebotGerrit review #158 at,158 : test_X plugins PLA integration by Jean-Louis Biasini (changes/58/158/7)
19:15:54n1sgevaerts: stub cpu_boost in database.c
19:16:14n1sfix it properly?!
19:16:32CIA-44Commit 105a05c in rockbox by Frank Gevaerts: One more *_filesize() to remove
19:16:33CIA-44Commit 83b4d70 in rockbox by Frank Gevaerts: Undefine HAVE_ADJUSTABLE_CPU_FREQ for __PCTOOL__
19:16:38gevaertsWell, more properly :)
19:17:06gevaertsStubbing it in database.c would mean stubbing it in warble.c, and so on
19:17:10n1syeah, the code is not well modularized
19:18:40n1sspeaking of stubbing, i think we should guard yield calls in librbcodec with #ifdef COOP_MT or something (and of course define that for rb builds)
19:19:45gevaertshm, I broke RaaA checkwps somewhere
19:19:56n1sanyway, i think it's good to let this one sink in for a while before continuing, the next patch pretty much just moves apps/codecs/ to lib/rbcodec
19:20:03CIA-4483b4d70 build result: All green
19:21:07n1shmm, no, next patch is moving metadata parsers, the one after that moves codecs
19:23:33n1smaybe tomorrow
19:24:05 Quit evilnick (Ping timeout: 276 seconds)
19:24:43 Join OzzieJacks [0] (
19:25:00 Quit Rob2223 (Quit: Rob2223)
19:29:56 Join Rob2222 [0] (
19:34:42***Saving seen data "./dancer.seen"
19:44:12 Quit mortalis (Quit: KVIrc 4.1.1 Equilibrium
19:47:06 Quit ender` (Ping timeout: 240 seconds)
19:50:03 Join ender` [0] (
19:55:54CIA-44Commit db85689 in rockbox by Thomas Martitz: (Author: Jean-Louis Biasini) test_X plugins PLA integration
19:58:12kugelpamaury: ping
19:58:34kugelpamaury: the apps part of g#152 looks alright
19:58:35fs-bluebotGerrit review #152 at,152 : Add support for HIDDEN Attribute of FAT filesystem by Akio Idehara (changes/52/152/4)
19:58:45kugelhave you tested the patch?
19:59:00pamauryok thanks, no, but I will
19:59:05CIA-44db85689 build result: All green
20:02:28jlbiasinikugel: some more plugins that could be turned into PLA have define to display the keymaps of the plugin in the help menu. Do you have any idea on how we could have something like that in PLA?
20:03:14CIA-44Commit fcb9fda in rockbox by Thomas Martitz: android: drop copied lcd_blit_yuv() and use generic one.
20:05:34kugeljlbiasini: add the key names to PLA
20:06:05jlbiasinilike for all devices isn't it?
20:06:11CIA-44fcb9fda build result: All green
20:07:57jlbiasiniI thought about it but I was hoping that there would be a less painfull solution...
20:08:24kugelNo i don't think there is
20:08:42jlbiasinion the other hand this allow to have automatic menu in all pla plugins with keymaps
20:08:53jlbiasinianother question what is pdbox?
20:09:21kugelsome music making plugin IIRC
20:16:48funmanpd = puredata
20:19:14CIA-44Commit 466a7c6 in rockbox by Frank Gevaerts: Integrate the warble tool in the regular build system.
20:19:16CIA-44Commit 13a7c9a in rockbox by Frank Gevaerts: Don't set CONFIG_CPU for __PCTOOL__, to avoid wrong asm
20:23:03kugelhm, now I remember I had plans to auto detect CONFIG_CPU
20:23:07CIA-44466a7c6 build result: All green
20:24:39CIA-44Commit efd25bf in rockbox by Frank Gevaerts: Disable some more sim_ and app_ redirection for __PCTOOL__.
20:25:17gevaertskugel: sound nice. It appears that that commit doesn't fully fix it for me
20:25:28gevaertswarble still gets confused for non-app builds
20:28:03kugelgevaerts: the build system does target cpu detection already and uses it for firmware/asm
20:28:07CIA-44efd25bf build result: All green
20:30:33OzzieJackskugel: would you mind looking at my replies to your inline comments for I1fa1bf0064dfd22b897eca7c56acc1ce8956f14c: Add new actions to %Tp (Touchscreen areas)
20:31:31kugelOzzieJacks: you can use gerrit to request a review, I get a mail then
20:32:00kugelI wasnt notified for your replies
20:32:20kugelwait, there are no replies
20:32:34kugelI guess you need to push a button to make them happen
20:33:08OzzieJacksI didn't know that. I'm new to all this. Thanks
20:33:47kugelnow, great
20:36:42 Join z180 [0] (6d2ddb47@gateway/web/freenode/ip.
20:39:30CIA-44Commit 1b216e8 in www by Frank Gevaerts: Add "warble" build for sdl application
20:40:03CIA-44Commit b9ce249 in rockbox by Frank Gevaerts: Don't define CPU_* for __PCTOOL__
20:40:03CIA-44Commit 8d41b23 in rockbox by Frank Gevaerts: Move includes of other makefiles around to make GCCFLAGS work properly.
20:42:41 Join factor [0] (
20:43:49CIA-448d41b23 build result: All green
20:46:09CIA-44Commit 0cc6ba0 in rockbox by Frank Gevaerts: Remove standalone makefile. Use configure to build
20:47:47z180i had a very small lisp interpeter from somewhere would that be a good idea for replacing the theming scripts?
20:48:23funmanupload it to gerrit
20:48:26bertrikhow many people know lisp?
20:49:24CIA-440cc6ba0 build result: 1 errors, 0 warnings (Frank Gevaerts committed)
20:49:44z180but it was not my interpreter so need to write a new
20:50:18z180i believe it was from a coding contest
20:50:30CIA-44Commit a217996 in www by Frank Gevaerts: Make warble build expect the right binary
20:54:11 Part jlbiasini
20:55:49kugelgevaerts: shouldn't warble be a target on its own?
20:57:00gevaertsI don't know
20:57:25gevaertsIn the long term, I think so. In the short term, I think it doesn't make much sense
20:58:11gevaertsThe infrastructure isn't really there for it I think
20:58:58kugelI think you could, say press 300 for warble, then set up an sdl based build internally
21:03:14gevaertsHow would that work in practice? Leave the makefiles alone, add a file in firmware/export/config, and hack configure to generate a warble build for "normal"?
21:03:54gevaertsThe way things are structured now, a warble build is technically similar to checkwps or the database tool, so it makes some sense to treat it that way in configure
21:04:29kugelI didn't mean an extra target in this way
21:05:02kugelI meant in the configure selection, so that you don't select a target then warble, but only warble
21:05:03 Quit TheLemonMan (Quit: WeeChat 0.3.7)
21:05:24kugelinternally you would do an sdl app build with warble extras, or whatever is best
21:05:45gevaertsIsn't that what I was saying?
21:06:08kugelI don't think so
21:06:31kugelI don't want to suggest to add a warble.h in firmware/export/config
21:08:42gevaertsThat's minor I think. You need a file in firmware/export/config, you can reuse the one from the sdl app, but I think in the long term reusing isn't a good idea
21:10:13kugelI just find it awkward to select a target, then warble, when warble is really meant as a tool on its own
21:11:39gevaertsI don't disagree, but the way the code is structured today I don't see any other way to do it differently, apart from hiding that fact in configure
21:12:13kugelhiding the fact was what I was trying to suggest :)
21:14:54gevaertsI think this should be revisited as soon as lib/rbcodec is something real
21:15:38gevaertsThat might be soon of course :)
21:26:54 Quit kevku (Quit: KVIrc 4.0.4 Insomnia
21:27:54n1sthe following patches touch the now deleted makefile so they need some trivial updates, i'll try to look into it tomorrow
21:29:38*gevaerts looks
21:30:35gevaertsThat shouldn't be too terrible to change
21:30:47n1spretty much calling the libs makefile and linking to it
21:31:52gevaertsIf you don't manage to integrate that in warble.make, feel free to restore the old makefile so you can go on
21:32:19 Join saratoga [0] (980329e4@gateway/web/freenode/ip.
21:32:44n1si think it should be pretty straight forward
21:33:11saratogagevaerts: how would that work if you wanted to debug a codec on a specific target (e.g. due to bad ASM on armv6)?
21:33:25 Join enthdegree [0] (~BitchX@unaffiliated/enthdegree)
21:33:33 Quit randumbum (Quit: ChatZilla 0.9.88 [Firefox 10.0.2/20120216101208])
21:33:38saratogaif its a separate target would it prompt you for the device flags or get them from the host?
21:33:56gevaertssaratoga: I don't know. kugel wanted a separate target :)
21:34:21saratogai guess a separate target works, since you'll probably have to deal with worth setting up a cross compiler for an arm linux box or whatever
21:34:33gevaertsI don't know yet what the best way is in the long term
21:34:34saratogaor QEMU if you're working in that
21:34:43***Saving seen data "./dancer.seen"
21:44:27 Quit y4n (Quit: PANTS OFF!)
22:00:17 Join wodz [0] (
22:00:54 Join Kiwi_Cam [0] (
22:01:01 Quit anewuser (Read error: Connection reset by peer)
22:01:46wodzsaratoga: I managed to build and run sdl RaaA in qemu arm emulator when I was testing unwarminder. It is rather trivial.
22:02:10saratogacross compiled or built in the emulator?
22:02:20wodzbuild natively on emulated arm
22:02:44 Join anewuser [0] (~anewuser@
22:02:45 Quit anewuser (Changing host)
22:02:45 Join anewuser [0] (~anewuser@unaffiliated/anewuser)
22:04:15 Quit saratoga (Quit: Page closed)
22:04:52 Quit KiwiCam (Ping timeout: 272 seconds)
22:11:18 Quit z180 (Quit: Page closed)
22:14:07 Join bluebrother [0] (
22:14:07 Quit bluebrother (Changing host)
22:14:07 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
22:14:37 Join tufur [0] (
22:17:10 Quit tufur (Remote host closed the connection)
22:17:21 Quit bluebrother^ (Ping timeout: 260 seconds)
22:17:32 Quit fs-bluebot (Ping timeout: 272 seconds)
22:18:51 Join fs-bluebot [0] (
22:22:06 Quit enthdegree (Ping timeout: 240 seconds)
22:29:07 Join dhrasmus [0] (
22:32:20 Quit pamaury (Remote host closed the connection)
22:58:52 Quit dhrasmus (Quit: Leaving)
23:15:46 Join perrikwp [0] (
23:22:17 Quit Keripo (Read error: Connection reset by peer)
23:34:26 Quit Kiwi_Cam (Ping timeout: 244 seconds)
23:34:44***Saving seen data "./dancer.seen"
23:39:00 Join enthdegree [0] (~BitchX@unaffiliated/enthdegree)
23:40:48 Quit wodz (Quit: Leaving)
23:47:00 Join Kiwi_Cam [0] (
23:55:53 Quit OzzieJacks (Quit: Leaving)
23:56:08 Join OzzieJacks [0] (

Previous day | Next day