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 2011-10-30

00:02:58 Quit Highlander (Quit: ChatZilla 0.9.87 [Firefox 7.0.1/20110928134238])
00:04:41 Nick scorche` is now known as scorche (~scorche@rockbox/administrator/scorche)
00:04:52 Quit keyb_gr (Ping timeout: 276 seconds)
00:05:56 Join keyb_gr [0] (
00:06:44***Saving seen data "./dancer.seen"
00:25:03JdGordonlebellium: not (currently) without making the whole list use skin lists
00:25:20JdGordonbut even then, scrollbars don't work comeplty properly
00:26:08pamauryZoiah: I find out !
00:26:14Zoiahpamaury: woo!
00:27:54CIA-14New commit by pamaury (r30858): sbtools: fix handling of raw mode, have elf_write looks like elf_read, fix uninitiliazed offset of bss sections, add real key override on decrypt
00:28:13ZoiahBTW, how do you decrypt the firmware yourself now?
00:28:13pamaurythat was stupid but I'm tired and I missed it...
00:28:26JdGordonbertrik: nice clip clip pic :) pity about the colours
00:28:41bertrikI'm about to commit a fix for that
00:28:41pamauryI use the real key dumped by the tool with you and added some code, see my commit
00:29:05ZoiahAhh, ok.
00:29:35ZoiahOhh, so you don't have the original key for the device, but because of the crazy random + aes key setup this device uses, you can still decrypt it?
00:29:35CIA-14r30858 build result: All green
00:29:48pamauryyes, because you did it
00:30:06ZoiahBecause you decrypt the key on the device itself?
00:30:37ZoiahOk, I understand.
00:30:38pamaurythat basically the only use of the device: decrypt the real key with the OTP key
00:31:12lebelliumJdGordon: thank you :) I also wonder if it is possible to use the %fb tag in a conditional tag. I tried %?if(%fb, =, 192)<do that|otherwise do that> but it doesn't seem to understand "if bitrate = 192 do that, otherwise do that" hum
00:33:37pamauryZoiah: can you try with the latest commit and confirm randomness is gone ?
00:33:46Zoiahpamaury: will do, one moment.
00:34:31JdGordonlebellium: hmm, yeah I don't think that will work
00:36:17lebelliumJdGordon: Yeah the goal was to optimize the display depending if the bitrare is xxx kbps or xxxx kbps (mainly FLAC and WAV files) but I think I will make things easier lol
00:36:49CIA-14New commit by bertrik (r30859): Sansa clip zip: fix colours of icons, logos, album art ...
00:37:17lebelliumJdGordon: do you know the conditions to get a new device on this page: ?
00:37:43JdGordontalk to rasher I think
00:38:33CIA-14r30859 build result: All green
00:40:25lebelliumrasher: hello, do you know the conditions to get a new device on this page: ?
00:40:38Zoiahpamaury: randomness is gone.
00:41:21JdGordonlebellium: hmm, i don't know how to fix %if in a sane way
00:41:35JdGordonit thinks 192 is a number, but in your case it is text
00:41:43JdGordonso it does the wrong comparrison operation
00:41:55pamauryZoiah: cool
00:44:12lebelliumJdGordon: maybe this tag is problematic because of VBR files it has to support too. It cannot say if %fb = or differs from the figure I put if the bitrate is variable.
00:46:51 Quit Jerom (Quit: Leaving.)
00:46:56 Quit pamaury (Remote host closed the connection)
00:48:47 Join Topy44 [0] (
00:49:54 Quit Acorn (Quit: Page closed)
01:11:57 Join ChickeNES [0] (~ChickeNES@
01:13:50 Quit robin0800 (Quit: Leaving)
01:18:47 Join funman [0] (~fun@rockbox/developer/funman)
01:19:00funmanbertrik: ping
01:20:45funmanbertrik: clip zip OFs in mkamsboot.c are under #if 1 /* disabled for now because it's not been proven they can be patched */
01:21:26 Join robin0800 [0] (~robin0800@
01:21:30funmancould you test 1.1.12 and 1.1.15 or just disable them and wait for someone to test them (we can only care about the last version though)
01:30:44 Quit keyb_gr (Ping timeout: 260 seconds)
01:32:34 Quit domonoky (Read error: Connection reset by peer)
01:34:04 Quit ageis ()
01:35:55 Quit funman (Quit: leaving)
01:37:25 Quit dfkt (Quit: -= SysReset 2.55=- Sic gorgiamus allos subjectatos nunc.)
01:38:00 Quit ender` (Quit: It is better to stay silent and be thought a fool, than to open one’s mouth and remove all doubt. -- Abraham Lincoln)
01:42:08 Quit ChickeNES (Read error: Connection reset by peer)
01:42:26 Join ChickeNES [0] (~ChickeNES@
01:42:39JdGordonlebellium: do you do your own build? I have an idea for a quick fix
01:45:30 Quit n1s (Remote host closed the connection)
01:46:03lebelliumJdGordon: No, I don't develop at all, lorenzo is the developer for the Samsung R0 port and I'm preparing the 1st theme for this device ;)
01:51:06 Join ungali [0] (
01:52:58ungaliWhere can I find information about the format of .playlist_control?
01:53:03 Quit ChickeNES (Read error: Connection reset by peer)
01:53:10 Join ChickeNES [0] (~ChickeNES@
01:54:01JdGordonungali: apps/playlist.c
01:54:03 Quit ChickeNES (Read error: Connection reset by peer)
01:54:25ungaliThank you.
01:54:28 Join ChickeNES [0] (~ChickeNES@
01:59:04JdGordonlebellium: yeah, got a working fix for you, might think about it a bit more before puting a patch online though
01:59:20JdGordonI had to add 2 op keywords to make it work, which isnt very nice
02:00:00lebelliumJdGordon: Okay cool :) It's 2AM here in Berlin, I go to bed now, see that with you tomorrow
02:04:17kugelthe red area at the bottom is a .bmp file
02:06:47***Saving seen data "./dancer.seen"
02:07:51JdGordonthats going to make people very happy!
02:15:55kugelhope so :)
02:16:41 Quit lebellium (Quit: ChatZilla 0.9.87 [Firefox 8.0/20111026191032])
02:17:22kugeldamn lcd_bitmap_* functions. I want that name for my new function
02:17:47 Quit bertrik (Ping timeout: 240 seconds)
02:22:44kugelthat's going to work nicely for icons I think
02:30:54 Quit ChickeNES (Read error: Connection reset by peer)
02:31:16 Join ChickeNES [0] (~ChickeNES@
02:32:11 Quit markun (Ping timeout: 240 seconds)
02:42:40 Quit ChickeNES (Quit: Computer has gone to sleep.)
02:53:06 Quit ungali (Quit: CGI:IRC (Ping timeout))
02:00:00 Quit toffe82_ (Ping timeout: 240 seconds)
02:04:53 Quit ReimuHak_ (Ping timeout: 255 seconds)
02:07:00 Join ReimuHak_ [0] (
02:17:29 Join toffe82 [0] (~chatzilla@
02:24:24 Join ageis [0] (
02:30:57 Quit ageis (Ping timeout: 258 seconds)
02:46:27 Join nsx [0] (
02:47:26nsxany of you have any experience unbricking a fuze v2? I'm trying the method described in the guide to short the recovery pins, but it's simply not working −− the device continues to go into the original firmware, which hangs during the bootup animation
03:06:51***Saving seen data "./dancer.seen"
03:22:28 Join ungali [0] (
03:26:31 Quit GeekShadow (Quit: leaving)
03:26:51 Join ReimuHa__ [0] (
03:27:29 Quit ReimuHak_ (Ping timeout: 258 seconds)
03:44:03 Quit ungali (Quit: CGI:IRC (EOF))
03:47:01 Quit [7] (Disconnected by services)
03:47:13 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
03:51:16 Join bertrik [0] (
03:51:17 Quit bertrik (Changing host)
03:51:17 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
03:51:28 Quit bertrik (Read error: Connection reset by peer)
04:05:56 Join T44 [0] (
04:05:59 Quit Topy44 (Ping timeout: 260 seconds)
04:24:19 Quit pixelma (Disconnected by services)
04:24:19 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:24:24 Quit amiconn (Disconnected by services)
04:24:25 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:24:37 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:24:47 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:26:44[Saint]Someone rang?
04:31:18 Quit TheSeven (Disconnected by services)
04:31:25 Join antil33t [0] (
04:31:31 Join [7] [0] (~TheSeven@rockbox/developer/TheSeven)
04:32:57*[Saint] wonders about kugel 's example bitmap.
04:33:14[Saint]What's happening there?
04:35:48 Quit ps-auxw (Ping timeout: 240 seconds)
04:39:56 Join ps-auxw [0] (
04:51:38 Quit [Saint] (Remote host closed the connection)
04:56:13 Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
05:00:35JdGordon[Saint]: IIUC the red part at the bottom is a transparent bitmap image
05:01:16[Saint]Seems frickin pointless to me.
05:01:49JdGordonsomeone has to dissent :)
05:02:03[Saint]With no way to layer viewports consistently, its a bit of a waste.
05:02:19[Saint]Just do the transparency in the image itself...
05:02:47JdGordonyeah, but then it is 100% transparent or 100% opaque
05:03:32[Saint]No, I mean...if you want to do transparency effects do it in an image editor.
05:03:45[Saint]Then export to bitmap.
05:03:55JdGordonthats what is happening
05:04:00JdGordonsame as aa fonts
05:04:32[Saint]Oh....right. I thought the core was doing the transparency.
05:04:51[Saint]Right, that makes much more sense now.
05:06:52***Saving seen data "./dancer.seen"
05:30:05 Join Rob2223 [0] (
05:33:29 Quit Rob2222 (Ping timeout: 260 seconds)
05:33:53 Quit ps-auxw (Ping timeout: 260 seconds)
05:35:20 Join ps-auxw [0] (~arneb@2001:470:c807:0:1532:4e5f:2ad3:4123)
06:01:31 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
06:07:24 Quit linuxguy3 (Ping timeout: 252 seconds)
06:09:20 Join linuxguy3 [0] (
06:28:34JdGordon[Saint]: have you got bitmaps you want to use for the list scrollbar?
06:29:42[Saint]Not at this stage, no. I try to avoid doing work surrounding impossible things :)
06:31:14[Saint]Oh...*list* scrollbar, no.
06:31:33JdGordonwhich scrollbar did you think i meant?
06:32:11[Saint]Well, by lists scrollbar um assuming you're meaning skinned lists?
06:32:34[Saint]If not, I count impossible the same as "not currently possible"
06:32:39JdGordonyeah, bored, so looking into finishing them
06:34:13[Saint] sorry. I've still not looked into the nightmare that is skinned lists. Especially as if I went that route now it breaks the ability to expand padding with the recent patch.
06:34:36[Saint]...which is apparently wanted.
06:35:00[Saint]*not looked into it for these themes
06:36:12JdGordongrrrrr is all i can say to that!
06:37:38[Saint]I could possibly swap out the ui viewport based on the padding setting...but that seems rather convoluted.
06:39:00 Quit robin0800 (Quit: Leaving)
07:02:20 Part toffe82
07:06:55***Saving seen data "./dancer.seen"
07:08:45 Quit [Saint] (Quit: Quit...)
07:37:36 Quit bluefoxx (Ping timeout: 252 seconds)
07:39:43 Join bluefoxx [0] (
07:44:03 Join Topy44 [0] (
07:44:10 Quit T44 (Ping timeout: 244 seconds)
07:44:59 Quit nosa-j (Ping timeout: 240 seconds)
08:01:50 Join PurlingNayuki [0] (~yzflcyq@
08:03:47 Part PurlingNayuki
08:12:02 Join Keripo [0] (
08:19:46 Join markun [0] (
08:34:36JdGordonthere is something wrong with me if i can fall asleep with death metal screaming from the tv isnt there? :)
08:35:49 Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
08:53:39 Join pamaury [0] (~quassel@
08:53:40 Quit pamaury (Changing host)
08:53:40 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
09:06:57***Saving seen data "./dancer.seen"
09:08:55 Join bertrik [0] (
09:08:56 Quit bertrik (Changing host)
09:08:56 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
09:09:48CIA-14New commit by bertrik (r30860): Sansa clip zip: allow original firmware versions 1.1.12, 1.1.15 and 1.1.17 to be patched
09:11:50CIA-14r30860 build result: All green
09:17:31 Join lebellium [0] (
09:19:24 Join stoffel [0] (
09:20:41kugel[Saint]: hm, I was hoping you would be more excited
09:22:57lebelliumhey guys! is it possible to force an album art to use the maxwidth and maxheight of the l tag. I mean I want a 500*443 album art displaying in 240*240 on the player (cutting the edges for instance)
09:22:58[Saint]At first I thought "why?", but then I realized that it was displaying a bitmap that already contained transparency. I thought the core was applying transparency to the image.
09:23:17lebelliumof the l tag*
09:23:35[Saint]Now I'm just like "yay...pity there's no sane way to layer viewports".
09:24:33[Saint]Its still cool, just...not as cool as it could be if we could actually layer viewports/images.
09:26:57[Saint]It'd be useful for moving an image, maybe?
09:27:31 Join n1s [0] (~quassel@rockbox/developer/n1s)
09:27:37[Saint]But for now if I want transparency I just add the blending of x, y and z into the image itself.
09:27:48kugel[Saint]: the image has an alpha channel. the core is blending with the backdrop
09:28:27[Saint]Right...and, if a viewport overlaps it fucks up dramatically? Or no?
09:28:28kugelI wonder what viewport layers (what's that anyway?) has to do with it?
09:29:45 Join dfkt [0] (~dfkt@unaffiliated/dfkt)
09:29:54[Saint]A way to overlap viewports (sanely), and display the viewport behind the current one is where I'd find this useful.
09:30:04[Saint]Right now, that's not possible.
09:30:20[Saint]The skin engine has no concept of front or back.
09:30:49kugelthis is useful as to not need to hack backdrop parts into images. now you need to specially craft the images and redo them if they move only 1 pixel
09:31:04 Join Pasiasty [0] (
09:31:28[Saint]kugel: but how often do images move conditionally like this?
09:31:47[Saint]Not enough to make it part of the core I'd wager.
09:31:56kugelnot at all, because it's not possible currently :)
09:32:10[Saint]It is possible...
09:32:30[Saint]Its just done rarely.
09:32:52kugelwhat do you mean by possible?
09:33:05kugelit's possible by using a separate image per position, perhaps
09:33:35[Saint]Yes...which only bloats the theme, not the skin engine :)
09:33:56[Saint]...the time for not bloating the skin engine is long gone, though...
09:34:10kugelbut it's a hell lot of more work to create, so it's rarely done
09:34:34kugeland it's relatively recent that images don't cap the theme's ram usage anymore
09:34:46 Quit n1s (Ping timeout: 240 seconds)
09:35:11kugelI'm also thinking icons for lists
09:35:12[Saint]The reason you just mentioned is another reason I
09:35:27[Saint]m a bit "meh" about it.
09:36:17[Saint]I can load all the images I want. Ans since there's no front or back I can easily predict how it will display.
09:36:54 Part Pasiasty
09:37:00[Saint]Its still cool...ages ago I'd be damn impressed.
09:37:30kugelwell, it comes almost for free since the blending function is already there. just needs ram to store the alpha channel (if there's one)
09:37:57[Saint]How free is "almost"?
09:38:18kugeldidn't check bin size numbers yet
09:38:22[Saint]Is there a patch I can look at?
09:41:12[Saint]Done with substantially less loc than I expected. Nice.
09:42:00kugelwould be even less if the viewport/image clipping wouldnt be so stupidely copied in every funtion there :)
09:42:31[Saint]I can see myself using this for button effects...maybe.
09:43:01kugelI think it'll be nice for icons in lists
09:43:49[Saint]Actually...buttons don't move. Hmmm.
09:43:58kugelthen we can sensibly convert from tango without the post-processing needed due to transparency problems
09:44:26[Saint]I *might* find somewhere fun to use it. I've had a few drinks. Not in code-mode.
09:45:06kugelcan also use the same images for all, sbs, fms and wps with different backdrops
09:45:34[Saint]kugel: aha! Yes...the tango icon problem.
09:45:48[Saint]Very nice. Didn't think about that.
09:46:19[Saint]Anything that stops me having to do those by hand is fine by me! :p
09:46:30 Join Jerom [0] (~jerome@
09:47:47[Saint]I can finally do proper screen transition effects!
09:48:58[Saint](Might be a bit slow still, but, I should be able to pull of a fade/dither screen effect with this)
09:50:31kugelfinally some excitement! :)
09:50:37[Saint]Or fading out album art at tacks end.
09:50:55[Saint]And in again for the new track.
09:51:10[Saint]Bah! Nope...that needs a sense if layering.
09:51:20[Saint]*of layering.
09:52:55[Saint]A screen transition isn't very fun if there's nothing on the screen but the backdrop...dammit. I thought I was onto something.
09:53:38[Saint]Ill find *something* fun to do with this...
09:53:54bluebrother^fade to some random image?
09:57:59[Saint]Hummm ...we can't do conditional backdrops, bah! Well...until I think of the "fun" use for this, not having to do the tango iconsets by hand serves as a large bonus.
09:59:47 Join ender` [0] (
10:00:45[Saint]...images could fade in from the backdrop on screen change? That'd be a weird/nice effect.
10:03:11lebellium[Saint] hey, do you know the conditions to get a new device on this page: particular port status?
10:04:05[Saint]"be a supported device"
10:05:09lebelliummeans Unstable port but not Unusable port?
10:05:48lebelliumokay thanks :)
10:06:22CIA-14New commit by bertrik (r30861): Sansa clip zip: make the simulator work by adding the background bitmap
10:06:27[Saint]There's a chance it already shares a screen size with a supported model.
10:07:07bertrikthe clip zip can use album art as a kind of WPS backdrop
10:07:14bertrikthe OF does that, I mean
10:07:49[Saint]bertrik: we can too...kinda.
10:08:02[Saint]But the results are wildly unpredictable.
10:08:14lebellium[Saint]: yes sure, I'm making a Samsung R0 theme that should work with Toshiba Gigabeat (I actually use its UI simulator since none exists for R0): 240*320 resolution
10:08:14CIA-14r30861 build result: All green
10:08:48kugelwe could have that, but backdrops need to be exactly the screens resolution
10:09:09kugelso a 96x92 AA breaks it
10:09:45[Saint]lebellium: themes aren't sorted by device, but rather screen type/dimension.
10:10:32[Saint]So as soon as <whatever device> is added to the themesite, if it scares a screen size and type, it will already have themes available.
10:10:46lebelliumOf course, the device is just a "tag" right?
10:10:48[Saint]*shares rather, not scares.
10:17:03[Saint]kugel: on that thought...couldn't we scale the aa to max width, then center it and add padding to max height?
10:17:06dfktis it possible to show fullscreen album art via a button press, or is this only possible via timed viewport cycling?
10:17:33kugel[Saint]: is doable sure, but not implemented yet
10:17:51[Saint]dfkt: a *physical* button?
10:18:01[Saint]Or a touchscreen area?
10:18:03dfktclip zip :)
10:18:10[Saint]Then no.
10:18:14dfkt96x96 sucks big time :p
10:18:18dfktthought so, thanks
10:19:27[Saint]kugel: yeah, I know its not there yet...just thinking out loud about how aa-as-a-backdrop could work.
10:19:35dfktthe most i could squeeze out of the screen, so far, with still showing relevant info -
10:19:45[Saint]We could also scale to max height, and clip edges.
10:19:55[Saint]Or offer both.
10:20:26[Saint]Most media players seem to be doing this aa-as-backdrop thong lately.
10:20:37[Saint]*thing, whoops.
10:20:59lebellium[Saint]: how to force an album art to display max width und height? For instance forcing a 500*433 album art to display 240*240 on the device (stretched or edges cutted)
10:21:47[Saint]lebellium: I believe you want the "center" flag.
10:23:01lebelliumah did not know about it, not written here I guess
10:23:09[Saint]but realistically...its better to not even try to deal with non-square aa
10:25:14[Saint]A theme should never assume album art is a certain height. Just try its best to display what it can.
10:26:04[Saint]The center flag is the "best" was to achieve this.
10:26:47lebelliumhum doesn't seem to work
10:26:57lebelliumhere a 500*500 album art : lebellium/R0/Rockbox/dump%20111023-235428.bmp">
10:27:13lebelliumhere a 500*443 album art: lebellium/R0/Rockbox/dump%20111030-102548.bmp">
10:27:16lebelliumwith c tag
10:29:08[Saint] the correct syntax.
10:29:18bluebrother^bertrik: how much is left for the clipzip bootloader? IOW, does it make sense to add it to Rockbox Utility (the installation is the same as the clip, is it?)
10:29:26lebellium# Album Art
10:29:58lebelliumI use your syntax
10:29:59bertrikbluebrother^, it doesn't boot the OF yet when woken up by USB plugin
10:30:22bertrikand yes, the installation is the same as all the clips
10:30:55bluebrother^is that a problem?
10:31:08[Saint]lebellium: "l(0,0..." looks VERY wrong.
10:31:20[Saint]You want %Cl
10:31:21bluebrother^I'd like to get a new Rockbox Utility out soonish. Currently trying to update the build setup ...
10:31:23lebelliumno but it's the IRC here that cuts
10:31:29lebellium% Cl
10:31:37lebelliumI have exactly the syntax you said
10:31:38bluebrother^so unless that's much of a problem we could add the clipzip as disabled target.
10:31:47bluebrother^so people can actually try / use it.
10:32:13 Quit mikroflops (Ping timeout: 258 seconds)
10:32:30bluebrother^bertrik: oh, and I've tried FS #12082 but that makes my nano2g hang on boot. Haven't found the time to investigate though
10:32:30fs-bluebot Attempt to add IAP support to the nano 2g (patches, new)
10:33:25[Saint]lebellium: pass...I'm on my phone right now. Basically, ill say "use square album art and assume everyone else does too".
10:33:47[Saint]Its easier :)
10:33:54lebelliumindeed hehe
10:34:08bertrikbluebrother^, I'm excited about adding clip zip stuff, but maybe we shouldn't rush the clip zip bootloader into rbutil yet
10:34:22 Join mikroflops [0] (
10:34:28bertrikit's not in our build system yet
10:34:41[Saint]It used to work at one point...I'll look into it and (maybe) file a bug against it.
10:34:51bluebrother^iirc the bootloader file is downloaded from the webserver anyway.
10:34:58[Saint]We should be able to center non-square aa.
10:35:03bertrikand we may still have changes in the bootloader in the coming time
10:35:08[Saint]If its not working, its a bug.
10:35:14bluebrother^so I don't see a problem with adding it as disabled target −− it won't be shown by default
10:35:29bertrikwhat is the effect of it being a disabled target?
10:36:32lebellium[Saint]: it doesn't work neither with the UI simulator nor on the real device so I assume it's a bug yes
10:36:38bluebrother^it's not shown in the target list, but you can enable it.
10:36:55bluebrother^and on enabling you get a warning saying that you should be a developer when using that :)
10:37:09bertrikok, cool, so people have to make some conscious effort to decide whether they want it or not :)
10:37:42bluebrother^apart from that there is no difference. The advantage is that you can use Rockbox Utility for installing things. Bootloader is usually the thing that doesn't work, everything else is just downloading zip files and extracting them :)
10:38:07[Saint]The same way "unusable" stops people from installing, right? :p
10:39:13bluebrother^f.e. I added the SA9200 as disabled target recently. I cannot install the bootloader but I can use it for everything else which is quite useful (just think about generating voicefiles ... except that there's still a bug open with that. Fortunately that affects only Windows :P )
10:39:24dfktlebellium, you used two center tags "c,c" - try just one
10:40:10[Saint]dfkt: the center flag shouldn't be used if its not needed.
10:40:35bertrikbluebrother^, ok sounds fine
10:40:52lebelliumdfkt: I also tried just with one. Does it work for you? Actually I guess I'll give up and just hope people use normal album arts
10:41:19dfktyeh, i used centering in the clipzip theme i'm working on atm
10:41:31bluebrother^hmm, dualboot.c is compiled into libmkamsboot. But I don't think that's much of an issue for a disabled target.
10:41:33dfkt[Saint], why should it be avoided?
10:41:45bertrikregarding FS #12082, please add your comment. I hope Lalufu can debug it. He already made another serial port on the nano 2g work IIRC, so IAP should not be that far away
10:41:45fs-bluebot Attempt to add IAP support to the nano 2g (patches, new)
10:41:46bluebrother^maybe it makes sense to disable bootloader installation for the clipzip for the next release
10:42:43[Saint]dfkt: I wasn't saying that. I was saying "having the additional center flag there would do no harm"
10:43:23lebelliumdfkt: May I have a look at your wps file?
10:44:06dfktlebellium -
10:45:22dfkti'm using a negative height value to display square album art (or any album art) centered in a 96x64 area
10:48:09lebelliumdfkt: yep I see that, I was wondering^^
10:49:12 Join Stummi [0] (~Stummi@rockbox/developer/Stummi)
10:50:21 Join bluebrother [0] (
10:50:33 Quit bluebrother (Changing host)
10:50:33 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
10:50:50bluebrotherbertrik: can you add the USB IDs to the DeviceDetection wiki page?
10:51:12bluebrothergreat, thanks
10:51:57bertrikyou need the ones for MSC mode I suppose (I don't know if they are different for MTP)
10:53:10 Quit fs-bluebot (Ping timeout: 240 seconds)
10:54:04 Quit bluebrother^ (Ping timeout: 258 seconds)
10:55:10pamaurybertrik: can you add two usb ids for me at the same time ?
10:55:50 Join fs-bluebot [0] (
10:56:33pamaury066f:3780: Fuze+ recovery mode (i.Mx 233/STMP3780 recovery) [duplicate]
10:56:33pamaury0781:74e1: Fuze+ MSC
10:57:43bertrikpamaury, I'm done making the change for the clip zip, you can edit it yourself now if you want
10:59:55*[Saint] spots a pamaury
11:00:02[Saint]You rang earlier?
11:00:21pamauryshould I add recovery mode to sandisk part of the list ?
11:00:33*pamaury don't understand [Saint] statement
11:01:31[Saint]Andchat tells me you pinged and/or highlighted me earlier.
11:01:40 Quit Jerom (Quit: Leaving.)
11:01:43[Saint]But doesn't scroll back that far.
11:03:27pamauryah yes, that was for someone who didn't understand some wps tag
11:03:44*bluebrother spots that we don't have support for the fuze+ in Rockbox Utility :)
11:03:45pamauryI thought you could know so I invoked you :)
11:03:58pamaurybluebrother: yes, I'm working on it :)
11:05:09[Saint]pamaury: ah, I see. Cool.
11:05:39lebelliumdfkt argh I don't manage to do it, actually I want to the contrary of your: I want all album arts to display full screen in a 240*240 window, whatever their original size (500*500, 500*400). %Vl(a,0,21,240,240,-) and l(0,0,240,240,c) is not working, the 500*400 album arts don't display full screen
11:06:28lebelliumwhy the hell does the IRC cut my % Cl tag everytime -_-
11:07:00***Saving seen data "./dancer.seen"
11:08:29[Saint]Crap irc client? ;)
11:08:49bluebrotherthe number of Sandisk player is quite impressive these days
11:09:16[Saint]Never heard of "yes, probably".
11:11:46 Join MethoS- [0] (~clemens@
11:11:50[Saint]lebellium: your dedication is impressive. I myself used to be quite concerned about such things. Especially things lile missing metadata, translation and rtl languages. And then I just stopped caring about corner cases as it makes things very un-fun to code for.
11:12:30 Quit stoffel (Ping timeout: 248 seconds)
11:15:41 Join webguest52 [0] (
11:16:08 Quit webguest52 (Client Quit)
11:16:20dfktlebellium, i think you can just "crop" album art, like in my wps, but not "stretch", like you want to do
11:16:30lebellium[Saint]: indeed, my concerns don't make things easier and my WPS code is quite frightening lol But as it's my very first rockbox theme, I want it to be the best possible. Now I'm wondering why using %V(167,279,70,-,-)%ar%pp/%pe , %pe is not right aligned up to the last pixel possible hum
11:17:31JdGordonbecause you put a space there
11:17:45[Saint]I just caught that too :)
11:19:19lebelliumJdGordon: ahhh! Thanks!! So much time wasted for an invisible space :D
11:19:33[Saint]We saw it :)
11:19:58[Saint]Hint: use an editor that supports displaying whitespace.
11:20:02*bluebrother thinks that some module tests for the skin stuff would be nice
11:20:09[Saint]Save you a bunch of hurt.
11:20:20lebelliumI use Notepad++ on Win 7. Maybe I can set it like you say
11:20:26bluebrotheror configure your editor to hilighttrailing space :)
11:21:07bluebrotheruse gvim!
11:21:10*bluebrother hides
11:21:11dfktbertrik, the zip ui sim has very low screen brightness/contrast here... do you know why that is?
11:21:19*[Saint] couldn't be happier with SciTE
11:21:36JdGordonbluebrother: which parts? the onl real bit that feasibly can be unit tested is the parsr, and it doesnt really need
11:23:59bluebrotherJdGordon: as much as possible of course!
11:23:59bertrikdfkt, I don't know, but I see that brightness 9 (the brightest) is indeed very dim, while brightness 8 is fine
11:25:11bertrikI suspect a bug in the general simulator code w.r.t. brightness
11:25:14dfktah, that did it, thanks - very weird
11:25:27JdGordonbluebrother: how about someone put proper unit testing for important bits like... playback, *buflib*, etc first? :)
11:26:29 Quit Whopper (Ping timeout: 255 seconds)
11:26:31bertrikcan we make a plugin to do buflib stress testing perhaps?
11:26:57bertrikwe don't really have a framework for doing unit tests, do we?
11:27:09JdGordonbuflib shold be testable outside of rockbox
11:27:31bertrikwe can do a test with a plugin in the simulator
11:27:56JdGordonnot as easily as building a library and a seperate program
11:28:10JdGordonbuflib is used everywhere but isnt tied to rockbox core (much)
11:28:14bertrikI don't see how you can draw that conclusion
11:28:56bluebrotherJdGordon: I would welcome that!
11:29:13bluebrotherwasn't part of the SoC plan to have a test program for buflib btw?
11:29:32JdGordonbertrik: what does buflib require? just a buffer to deal with, what else in the core does it require
11:29:41JdGordonplayback and everything builds ontop of it
11:30:08JdGordonbut its pointless now, seems like the buflib bugs are gone now
11:30:21bluebrotherIMO adding unit test is never pointless
11:30:43bluebrotherI'd also welcome some tests that are automatically run by the build system / some nightly test system
11:31:05bluebrotherand just because the *known* buflib bugs are fixed doesn't mean it's completely bug-free :)
11:31:29*bluebrother would also like to see unit tests in Rockbox Utility but hasn't found the time looking into it
11:32:12JdGordonnoone is going to disagree with you, but testing is not fun :)
11:33:23bluebrotherthat heavily depends on the test system
11:33:42bluebrotherthough it's something quite different to hacking around, sure.
11:36:20 Quit Keripo (Quit: Leaving.)
11:40:17 Nick [sanyi] is now known as [Sanyi] (~sanyi@unaffiliated/alexanderc)
11:44:59kugelgot a much better approach
11:46:11 Quit [Saint] (Quit: Quit...)
11:46:25 Join [Saint] [0] (~Saint]@
11:46:26 Quit [Saint] (Changing host)
11:46:26 Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
11:46:48kugelI'm actually re-using the font drawing function as is now
11:50:38CIA-14New commit by bertrik (r30862): Sansa clip+: correct manual, clip+ has no button lights
11:52:10CIA-14r30862 build result: All green
11:56:44 Quit markun (Ping timeout: 260 seconds)
11:59:31CIA-14New commit by bluebrother (r30863): Rockbox Utility: Preliminary support for Clip Zip and Fuze+. ...
12:01:02CIA-14New commit by bluebrother (r30864): beastpatcher: check WMP version. ...
12:01:12CIA-14r30863 build result: All green
12:02:53CIA-14r30864 build result: All green
12:09:48pamaurybluebrother: thnaks
12:10:52pamaurybluebrother: what is the bootloadername field ?
12:11:08 Join y4n [0] (y4n@unaffiliated/y4ndexx)
12:14:15 Part Kuitsi
12:15:54 Join Thra11 [0] (~thrall@
12:19:29 Join Horscht [0] (~Horscht@xbmc/user/horscht)
12:22:53bluebrotherpamaury: bootloadername is the filename on the server (i.e. the path on
12:23:15pamauryok, that makes me think I need to change configure to scramble the bootloader file too
12:23:28bluebrotherbootloaderfile is the filename on the device, i.e. where it will be placed at on installation
12:25:23bertrikbluebrother, the Clip Zip does indeed show up when showing disabled targets. It shows up as "Sansa Clip Zip (Unknown)", where "Unknown" is Stable or Unstable for other targets and I guess that's OK.
12:26:15bluebrotherbertrik: yes, that's how it's supposed to work.
12:26:43bluebrotherI've disabled bootloader installation for now so it can't mess up things.
12:27:01bluebrotherhmm, might make sense to put some comments in rbutil.ini on the meaning of the fields
12:27:28bertrikbluebrother, that wasn't meant as criticism or anything, just telling what I was seeing
12:27:28 Join JUNX [0] (
12:28:20bluebrotherbertrik: no worries
12:31:09dfkti want to display square 96x96 album art in a 96x64 rectangle, and i want that album art to scroll up and down. this isn't working - %t(1)%Cl(0,-16,96,96,c);%Cl(0,-15,96,96,c);%Cl(0,-14,96,96,c);%Cl(0,-13,96,96,c);
12:31:17dfktcan anyone give a hint on how to do that?
12:31:53[Saint]that's a very amusing string. :-S
12:32:09dfktthere should be an 'etc' at the end too ;)
12:32:23[Saint]I had never thought of that before, and I've tried some amusing hacks in my time...nice.
12:32:25 Quit JUNX (Quit: CGI:IRC (Ping timeout))
12:32:25dfktnever used alternate sublines before
12:33:34[Saint]add the %t(foo) declaration to each segment.
12:34:11dfktnope, same results
12:34:57 Join JUNX [0] (
12:34:59[Saint]Humm...well, I don't suppose albumart was intended to be displayed in an alternating subline like that.
12:35:06[Saint]It can be done, though.
12:35:15JUNXhello am new here and would need some holp
12:35:31JUNXjust tryin to get some roms for my ipod video
12:35:32dfkta different approach, not sublines?
12:35:34*bluebrother goes for some holp
12:35:36 Quit CIA-14 ()
12:35:37JUNXlike pacbox
12:35:39[Saint]preload X viewports, one for each aa "frame", and load the viewports conditionally.
12:35:47dfktgood idea
12:35:57 Join CIA-75 [0] (
12:36:15JUNXbut cant find roms which work for me
12:37:11[Saint]JUNX: No one here can help you with that. The only way to get ROMs is to extract them from your own original copy of the game, and that would be off-topic. Anything else would be illegal.
12:37:35[Saint]And all illegal things are bad. Ever.
12:38:05JUNXhm ok am very sad to hear that
12:38:27JUNXso how do i make a original copy of my pacman
12:39:50bluebrotherthe MAME project has some ROMs they are allowed to distribute. No idea if any Rockbox plugin can make use of those though
12:48:33kugelbluebrother: what do I need to change to keep transparency and get a 32bbp bmp with your script?
12:53:41 Quit JUNX (Quit: CGI:IRC)
12:54:19 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
12:54:52bluebrotherkugel: changing the last 2 convert calls should do the trick. The last drops transparency and replaces it with the mask the next to last creates
12:55:06 Join [Saint_] [0] (~Saint]@
12:55:10 Quit [Saint_] (Changing host)
12:55:10 Join [Saint_] [0] (~Saint]@unaffiliated/saint/x-8516940)
12:55:15bluebrotherthe intermediate file is png so you would need to simply convert that to bmp
12:55:21bluebrotherhaven't tried that though :)
12:55:57kugelhm, there website is down
12:56:16 Quit [Saint] (Ping timeout: 252 seconds)
12:56:29 Nick [Saint_] is now known as [Saint] (~Saint]@unaffiliated/saint/x-8516940)
12:58:34CIA-75New commit by rmenes (r30865): Fix a typo spotted from r30864.
12:59:42 Quit Stummi (Quit: Bye!)
13:00:04CIA-75r30865 build result: All green
13:00:35kugelbluebrother: how to use?
13:03:23kugelbluebrother: seems a bit out of date. it fails to find 2 images
13:07:02***Saving seen data "./dancer.seen"
13:10:54kugelhm, rockbox-clef.svg is missing
13:10:58dfkt[Saint], any idea what (or how bad) i messed up with your idea? -
13:11:49 Join keyb_gr [0] (
13:12:26JdGordondfkt: what are you doing?
13:12:36JdGordon%Cl wont work more than once in a skin file
13:12:42dfkti want to display square 96x96 album art in a 96x64 rectangle, and i want that album art to scroll up and down.
13:13:07JdGordonnot doable
13:13:22dfkttoo bad, thanks
13:13:31JdGordonand would presumably look terrible :)
13:13:51dfktscrolling 1px at a time should look pretty nice, i'd think
13:14:13dfktthe clip zip's 96x96 screen is really annoyingly limited with what one can do...
13:14:28[Saint]and take up a metric fuck-tonne in buffer on the viewport declarations ;)
13:14:44kugelbluebrother: any hint what would need to be changed exactly? :(
13:15:41JdGordondfkt: it *might* be possible to modify the AA display code to draw a specific rectangle if it is bigger than YxZ, but i wouldnt hold your breath
13:17:23CIA-75New commit by bluebrother (r30866): Correct location of logo svg file.
13:18:09bluebrotherkugel: gimme a seconds, I'll have a look
13:19:00CIA-75r30866 build result: All green
13:19:38dfktmy other idea was this: %t(1)%Cl(0,-16,96,96,c);%Cl(0,-15,96,96,c);%Cl(0,-14,96,96,c);%Cl(0,-13,96,96,c); etc
13:19:49dfktbut since you said only one %Cl, yeah...
13:20:50bertrikdfkt, album art is a bit fuzzy for me on the zip, I think it would look nicer if it was sharpened up a bit
13:21:26dfktbertrik, yes... and it seems the colors are still a bit off - some usually grey/black colors have a brown-ish tinge to it on my zip
13:21:26JdGordonif the display is only 96*96 why waste space on AA?
13:21:34dfktcause i like it :p
13:22:40dfktso i just leave it cropped/centered, without scrolling -
13:27:03 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
13:28:41bertrikdfkt, we're using the colour settings of the original OF, but maybe we can fix those a bit on rockbox
13:29:56kugelbluebrother: thanks! do you know why rockbox-clef.svg went missing?
13:30:24bluebrothersure, because it was moved and that script forgotten :)
13:31:38kugelheh :)
13:37:18 Quit fs-bluebot (Quit: So long, and thanks for all the fish.)
13:37:31 Join fs-bluebot [0] (
13:38:32*JdGordon is experimenting with smaller skin buffers for each skin instead of one big one in the hop it will allow (re)-loading individual skins more quickly
13:38:47JdGordonspecifically so load a .cfg which only changes the volume will be *much* faster
13:39:15kugel :)
13:40:24bluebrotherkugel: that Rockbox logo in the bottom is wrong!
13:41:52kugelsorry! :p
13:42:04kugelcabbie isn't the best showcase though
13:48:52[Saint]It'd be a lot worse ~6 months ago, thankyouverymuchkbai
13:52:10[Saint]ReimuHa__: "cabbie isn't the best showcase though"
13:52:20[Saint]whoops, nick completion fail.
13:52:33kugelhas it changed in the last 6 month?
13:52:54bluebrotheryou're referring to the Android version?
13:52:59[Saint]I was being sarcastic, anyway. And I thought you'd be using the RaaA cabbie.
13:53:07kugelnot on the e200 sim :P
13:53:10bluebrother(or rather: touchscreen)
13:53:51kugelbut even the RaaA one hasn't changed. anyway, I meant the backdrop isn't so suitable to show off transparency
14:00:50 Quit antil33t (Read error: Connection reset by peer)
14:01:17 Join antil33t [0] (
14:04:45JdGordongot a comparrison with svn?
14:05:11kugelhm, could compare to [Saint]'s icons he uploaded months ago (svn doesnt have 24x24)
14:05:36bluebrotheryou can easily create 24x24 icons with the script :)
14:05:58bluebrother[Saint] manually improved the icons
14:06:24*[Saint] did his by hand, the good 'ol fashioned way.
14:06:31[Saint]no fancy scripts for me.
14:07:30bluebrotheriirc that was before the script existed
14:07:31kugelthe difference isnt that big because [Saint]'s icons are well done. however, it's visible
14:08:13[Saint]Oh yes, I'll bet they're better.
14:09:00[Saint]going from svg->png->bitmap, then tweaked by hand can't compare.
14:09:46bluebrotherI guess it could, if alpha channel would be the same for both. But AFAIU that's using 8bit alpha now, right?
14:10:15JdGordon\o/ loading skins into individual smaller buffers is incredibly simple and actually works first try!
14:10:36kugelbluebrother: the bmp has 8bit, but I only use 4bit
14:11:35bluebrotherthat's still way more than 1bit :)
14:11:53kugel(I uploaded a pic with [Saint]'s for comparison if you didn't notice)
14:12:03bluebrotherI did
14:12:57bluebrotherand if one looks close it pretty much looks like you're using more than 1bit alpha values :)
14:13:36 Quit fs-bluebot (Quit: So long, and thanks for all the fish.)
14:14:13JdGordonkugel: can you add this to buflib: If something asks for an unmovable handle it always allocs before the audiobuf, and they shuold provide a "free requested" callback which would be called wehenever anything else asks for an unmovable alloc, in which it can free itself (which would likely realloc later)?
14:14:19 Join fs-bluebot [0] (
14:14:45JdGordonso unmovable allocas are always alloced from the start of the whole buffer
14:16:08 Quit fs-bluebot (Client Quit)
14:16:35kugelbuflib doesn't know about the audiobuffer
14:16:44kugelI'm not sure I understand what you want
14:17:04 Join fs-bluebot [0] (
14:18:00JdGordonI want to have buflib try its best to only put unmovable allocs at the very start of the buffer
14:18:07JdGordondo whatever it needs to for that to happen
14:18:40kugelthen it would put the audio buffer at the start then
14:19:06JdGordonI want to use them for the skin engine, and in most cases they will be alloced at boot and never freed, but if it doesnt put it infront of movable data it will create wholes
14:19:23JdGordonaudio_buffer is alloced with the max call so it is distinguishable
14:19:34kugelthe same is true for many other buffers
14:19:56JdGordonok, so add a flag saying it wants to be put at the front
14:20:57JdGordonotherwise I'll just keep using slightly wasteful static buffers
14:21:03kugelis it so difficult to make the skin engine movable?
14:21:09JdGordonfor the sake of <50KB it probably isnt worth it
14:21:42JdGordonfeel free to grab my attempt off github and fix it
14:21:57JdGordonbut even then there wil still be a static buffer
14:23:48JdGordonI'm experimenting with smaller buffers and I've split it to 25KB per skin (down from 40KB total), which is slightly too small for the wps (likely) and way too big for the sbs (likely)
14:24:01JdGordonbut it means *instant* .cfg loads
14:24:19kugelwhy does splitting up make it faster?
14:24:32JdGordonit only loads skins it knows it needs to
14:25:08JdGordonI guess I could go back to one buffer, but then if any skin needs reloading they all get reloaded
14:25:36JdGordonit probably is going to be all or none in real usage anyway though so hmm
14:25:50JdGordonanyway, bed time
14:25:57[Saint]so...kugel, when do we expect your nange of iconsets to hit svn? ;)
14:26:38JdGordondont forget to commit the icons on buflib patch if you're fiddling with icons anyway
14:28:55kugel[Saint]: hehe
14:29:25dfktbertrik, i've gotten "data abort at 300528F0 FSR 0x8 (domain 0, fault address 0x8F6377)" a few times on the zip lately, when switching between themes (even 'failsafe'). tried different build revisions, did a scandisk - didn't help
14:29:36 Join _jhMikeS_ [0] (
14:29:36 Quit jhMikeS (Read error: Connection reset by peer)
14:29:36 Quit _jhMikeS_ (Changing host)
14:29:37 Join _jhMikeS_ [0] (~jethead71@rockbox/developer/jhMikeS)
14:29:37 Nick _jhMikeS_ is now known as jhMikeS (~jethead71@rockbox/developer/jhMikeS)
14:32:06 Join funman [0] (~fun@rockbox/developer/funman)
14:34:30 Join GeekShadow [0] (
14:37:26 Quit funman (Ping timeout: 260 seconds)
14:39:24 Nick [Sanyi] is now known as [sanyi] (~sanyi@unaffiliated/alexanderc)
14:39:29 Quit antil33t (Read error: Connection reset by peer)
14:39:53 Join antil33t [0] (
14:45:50 Join The_Seeker [0] (
14:56:10 Join funman [0] (~fun@rockbox/developer/funman)
15:07:05***Saving seen data "./dancer.seen"
15:08:05 Quit The_Seeker (Quit: Stay hungry, stay foolish)
15:10:08kugelthe alpha channel stuff allows for other nice things too, like see-through splash or background-dimming
15:21:46 Join liar [0] (
15:29:56 Quit benedikt93 (Quit: Bye ;))
15:32:43CIA-75New commit by jethead71 (r30867): Gigabeat S: Maintain power thread at realtime priority at all times so it may always service the watchdog on time regardless of playback priority.
15:34:23CIA-75r30867 build result: All green
16:23:38 Join fyrestorm [0] (
16:33:56bluebrotheryay! Finally managed to build a 32bit Rockbox Utility on 64bit linux :)
16:44:15rasherWait, Rockbox database scans /system/ on Android? That seems wrong.
16:44:33[Saint]it scans *everything*
16:44:46rasherI don't see any way that's reasonable
16:45:00rasherI don't want my ringtones in the db
16:45:08rasheror media of other system apps
16:45:09[Saint]Its something I've always cried about...
16:45:28[Saint]JdGordon has a WIP to select dirs for the database.
16:46:00bluebrotherwouldn't it be sensible to simply use the same folder that is set for the file browser?
16:46:20[Saint]It would, yes.
16:46:21rasherI'm not convinced that's even necessary, but scanning /system/ makes no sense at all
16:46:48[Saint]rasher: its not specific to RaaA, more of a general Rockbox thing.
16:47:09[Saint]I want X Y and Z scanned, but not A, B and C...etc.
16:47:10rasherNot really
16:47:20rasherOn regular OS-replacing Rockbox it makes sense
16:47:29rasherBut for raaa, it doesn't make sense to scan system areas
16:47:40[Saint]I didn't imply it did.
16:47:45[Saint]did I?
16:47:53rasherBut it *is* specific to raaa
16:48:10 Quit fyrestorm (Ping timeout: 260 seconds)
16:48:13[Saint]I was talking of Jd's WIP solution to managing scanned dirs.
16:49:41 Join hilbert [0] (
16:50:18rasherwell, I'm not convinced that's even necessary. Just exclude system areas on raaa (and follow the OS's method, for Android that would mean recognising .nomedia, if it doesn't already)
16:51:50[Saint]It is still needed. Think of it as a .ignore/.unignore manager.
16:52:15[Saint]But you're quite right, there's no reason to scan system dirs.
16:54:40CIA-75New commit by bluebrother (r30868): Rockbox Utility: allow building for 32bit on 64bit Linux. ...
16:56:14CIA-75r30868 build result: 0 errors, 1 warnings (bluebrother committed)
17:00:32 Join Jerom [0] (~jerome@
17:04:25 Quit GeekShadow (Remote host closed the connection)
17:05:16 Quit pamaury (Ping timeout: 244 seconds)
17:06:13lebelliumwhy when some text with these characters: y, g, (, ), p is scrolling the bottom of these caracters doesn't scroll at the same time than the rest of the character?!
17:07:10***Saving seen data "./dancer.seen"
17:07:28lebelliumat least with the font 15-Adobe-Helvetica
17:08:42lebelliumit's quite funny to see the bottom of the character is late lol
17:09:34jhMikeSafaiu the scrolling is done in one list-size chunk
17:10:42lebelliumthere is no problem on the UI simulator, maybe it's an issue with the current unusable build on Samsung R0
17:21:36kugelare we going to branch 3.10 soon?
17:21:48kugelwe fixed the important bugs didn't we?
17:22:17jhMikeSI'm waiting for word on the mpegplayer OSD issue. I _looks_ fixed but I want that confirmed somewhere, by someone.
17:22:37kugelwhat mpegplayer OSD issue?
17:22:58jhMikeSIt was crashing when it was displayed. It appeared related to font rendering.
17:23:10*kugel didnt know that one
17:23:15jhMikeSFS #12300
17:23:16fs-bluebot MPEG Player - Data Abort if any button is pressed during playback (bugs, researching)
17:24:12kugel"Looks like the prophesies of the buflib beast are coming to pass. There's an antipattern present in the entire design that's going to be very difficult to repair." :(
17:24:49jhMikeSthere are some things I don't like about it and it was pissing me off a bit that day :)
17:25:46kugelyou didn't speak up during gsoc/before committing
17:26:53*jhMikeS was being a bit off the hook in that post
17:29:17 Join stoffel [0] (
17:32:07 Join wodz [0] (
17:40:13kugeljhMikeS: mpegplayer works here
17:40:38kugelwith anti-aliased font and normal
17:41:40 Join markun [0] (
17:48:10 Join GermanMushroom [0] (
17:48:55jhMikeSit looks like it get resolved...everything player that had a problem before works now. btw, is that on a rotated screen?
17:49:54jhMikeSs/everything player/every player I checked before/
17:52:35jhMikeSbtw, I did express some concerns about buflib and threading and imho I thought it should have worked more like buffering where handles are kept, not pointers, no callbacks are needed and the scheme could work under any sort of threading
17:52:52 Join toffe82 [0] (
17:55:58 Join Unhelpful [0] (~quassel@rockbox/developer/Unhelpful)
17:59:14kugeljhMikeS: that was on e200, so yes
17:59:37kugeljhMikeS: buflib works with handles
18:00:33kugelbut getting rid of pointers entirely would have been a too massive amount of work
18:01:44wodzkugel: about scrollwheel/touchstrip issues - you mentioned it is related to event suppression on screen change. Could you give me some reference?
18:02:01kugelwodz: not anymore than the several bug reports I linked
18:02:34wodzkugel: I looked at this bugs and at least scrollwheel one was purely driver issue
18:02:48kugelwhich one do you mean?
18:03:29wodzFS #5230
18:03:30fs-bluebot Keypad HOLD doesn't let go until scrolling is used (bugs, requires)
18:03:59kugelyea, I don't mean that one
18:04:17kugelI mean FS #8816 and its related tasks
18:04:18fs-bluebot Extra button press required to continue browsing menus (bugs, unconfirmed)
18:04:30wodzok thanks
18:05:50[Saint]reminds me of the iPod "can't seek/skip until changing volume via the scrollwheel"
18:06:21wodz[Saint]: thats basically duplication of the bug I linked above (5230)
18:08:12 Join pamaury [0] (~quassel@
18:08:13 Quit pamaury (Changing host)
18:08:13 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
18:10:10 Quit GermanMushroom (Ping timeout: 276 seconds)
18:15:33wodzwhoa FS #10883 <- he seems to find out the root of the problem
18:15:33fs-bluebot clickwheel stops working when its used during the bookmark setting process (patches, unconfirmed)
18:16:55wodzI'll try to understand this
18:20:08 Join mortalis [0] (
18:21:00wodzmortalis: ping
18:21:32jhMikeSkugel: I know it works with handles and so does buffering. And yeah, some things are alot of work or at least just alot of places to deal with.
18:21:46mortaliswodz: pong
18:22:29wodzmortalis: could you dump rom content on your rk2706b ?
18:22:48kugeljhMikeS: you can't go handle-only unless you want to change a couple of low level functions too (lcd_bitmap*, read/write, ...)
18:22:51jhMikeSalot of things are so "stovepiped"...better abstraction would help
18:22:59kugeland you need those duplicated because they still need to work with pointers
18:24:37jhMikeSwhat in read/write or lcd_bitmap* need pointers available at all times?
18:24:46wodzmortalis: I disassembled dump from my rk2705 and it has some addons compared to dump from Alemaxx rk2706
18:25:18mortaliswodz: How can i do that?
18:25:56kugeljhMikeS: the first can yield
18:26:44jhMikeSwouldn't be an issue
18:26:49 Quit pamaury (Ping timeout: 256 seconds)
18:27:15wodzmortalis: basically you need to disable interrupts, disable cache, unremap, copy 8kB starting from 0x0 to some save buffer, remap again, enable cache, enable interrupts, dump buffer to file
18:27:34kugeljhMikeS: I don't understand
18:27:40kugelit's sure an issue
18:27:47wodzmortalis: BTW in your 'jump to dfu' code you should disable interrupts also
18:27:55jhMikeSobtaining a pointer would make the allocation unmoveable (always) and it would need balanced unlock - no callback and no threading issue
18:28:01kugelthe buffer passed to read/write can move during the call
18:28:37kugelthat's what's being done, just that the lock is implemented on the client side
18:29:02jhMikeSwouldn't happen with what I'm considering since it would be locked since there's a pointer with a reference count
18:29:20jhMikeSget it, use it, release it
18:29:29kugelthat's what's being done, just that the lock is implemented on the client side
18:29:39jhMikeSI think simpler interfaces are better
18:29:47amiconnBuflib integration is buggy enough as it is - I'd rather concentrate on stabilizing the thing before even considering to make more stuff use it
18:30:08kugelamiconn: that only works if you put the bugs you report on FS#
18:30:28jhMikeSyou always use the same procedure without exception. if you mess it up, your fault. the debug screen can tell you about outstanding buffer references
18:31:33jhMikeSand it would working with and threading type, so doesn't exclude running on OS directly, preemption or multiprocessor
18:31:50kugeljhMikeS: I know all that
18:32:51jhMikeSthat's encouraging :)
18:32:53kugelthat's not fundamentally different to the current situation
18:33:41 Quit Zarggg (Ping timeout: 245 seconds)
18:33:46kugelthe callback has other advantages other than acting as a lock, though
18:34:28jhMikeSI hate to say it wouldn't be gone in the case of the user of the remaining space, but would limit the clutter of callbacks
18:35:10*jhMikeS doesn't like another thread goofing with another's pointers unless that's the explicit design
18:36:17 Quit Rob2223 (Quit: Rob2223)
18:36:24kugelit's by design that all threads can make allocs, and thus thread a) can cause an alloc in thread b) to move
18:36:50 Join GermanMushroom [0] (
18:37:40kugelamiconn: there's still this bug around you haven't reported
18:37:55 Join pamaury [0] (~quassel@
18:37:55 Quit pamaury (Changing host)
18:37:55 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
18:37:56kugelnobody can fix it if you don't log it somewhere
18:38:04jhMikeSyeah I know the server expects it, but it's not nice for code that shouldn't have to worry about it, which is most of it and it's buggering up the DSP
18:38:37kugelit's doing what?
18:39:15jhMikeSmaking things a more complex than they ought to be, to worry about a callback and a pointer changing during a processing run
18:40:12jhMikeSit got my attention the most at this time because that's what I was digging into regarding issues with tdspeed
18:40:15kugelthen remove the yields
18:40:18 Quit Unhelpful (Remote host closed the connection)
18:40:38 Join Unhelpful [0] (~quassel@rockbox/developer/Unhelpful)
18:40:45 Join ageis [0] (
18:40:58jhMikeSyeah, that's where things become unsavory for me −− relying on yields, which can change without notice in remote code or between targets
18:41:00kugelI don't find the callbacks as horrible as you do
18:41:53jhMikeSwe have a moveable allocator that doesn't require them at all and could be tweaked slightly to be completely safe on the client side no matter what's going on
18:42:31*jhMikeS might change his name to "Concurrency Nazi" :P
18:43:03kugelright. it doesn't require them at all
18:43:14kugelyou can write code perfectly fine wihtout using any of the callbacks
18:43:31*jhMikeS 's gone through enough hell in the past about these yield-y things
18:43:35kugelbut you need either locks or temp buffers
18:44:06kugel(okay, locks need a tiny callback)
18:44:13jhMikeSI'd prefer a reference count requiring balanced calls, which I sort of did in buffering.c
18:44:42kugelthat is implemented numerous times
18:44:52jhMikeSptr = get_pointer(handle); use it cause it can't move ; unlock(handle); can move again;
18:45:05kugelbut you often don't need locks, and sometimes the release calls are awkward to place
18:45:24kugeljhMikeS: I got your point
18:46:02 Join T44 [0] (
18:46:19kugelJdGordon also wants this. if you all feel so strong about it just change it
18:47:11jhMikeSand so fine, any code I implement will operate as I've described. I've largely abandoned yield consideration unless the code is under tight control cause it was always like whack a mole when code change in other spots.
18:48:56 Quit Topy44 (Read error: Connection reset by peer)
18:53:29jhMikeSYes, I guess Jd and I are in agreement afaict. But I also witness enough of what transpired around it (was I mistaken) and felt that it wasn't worth further discussion.
18:54:01 Join ageis00 [0] (
18:54:09 Quit ageis00 (Remote host closed the connection)
18:55:57jhMikeSShouldn't the user of the remaining audio buffer not be the one receiving notifications, not always the playback engine?
18:56:29 Quit ageis (Ping timeout: 260 seconds)
18:58:04kugeljhMikeS: who would that be?
18:59:03jhMikeSwho would what be?
19:00:14 Join criegod [0] (
19:01:10jhMikeSif you mean "the user of the remaining audio buffer", would that not be anyone calling audio_get_buffer?
19:03:35 Quit GermanMushroom (Ping timeout: 258 seconds)
19:04:03kugeljhMikeS: those callers where never notified
19:04:45kugelthey could be. it also wouldnt be bad if they needed to explicitely release the audio buffer again
19:04:47jhMikeSI know. It possibly could be an issue now.
19:05:20bertrikdfkt, that could be some generic bug, probably not related to the specific target
19:06:42 Quit pamaury (Read error: Operation timed out)
19:07:11***Saving seen data "./dancer.seen"
19:07:18criegodQuestion: I recently bought a Fuze 2. I thought i read somewhere there might have been a hardware change with these last models. I tried to install rockbox with the automatic installer method but it said the firmware is "unknown", even though i have V02.03.33A. Can anyone tell me if there is any realization of a hardware change causing this latest hardware not to be supported yet? Or is... possibly some more trivial problem?
19:07:20 Join Rob2222 [0] (
19:10:51jhMikeSthe major drawback with handles and references is backing buflib with malloc on a host is more complex
19:14:14 Join ageis [0] (
19:23:34 Join GermanMushroom [0] (
19:31:40lebelliumhow to code that in a theme: "if numeric battery display do that but if graphic battery display do that" ?
19:33:09lebelliumI'd like to display different things according to what the user choose in theme settings>Status/scrollabar>battery display
19:37:23 Quit [Saint] (Ping timeout: 240 seconds)
19:39:25 Join robin0800 [0] (~robin0800@
19:39:33 Join ungali [0] (
19:41:04 Quit ungali (Client Quit)
19:45:11dfktcan the %lh tag for virtual LED be used with an image, or can it only display "h"?
19:47:17 Quit robin0800 (Ping timeout: 258 seconds)
19:47:48kugeldfkt: use it as conditional
19:48:14dfktthanks kugel, will try
19:56:11 Quit mgue (Ping timeout: 245 seconds)
19:57:57 Join mgue [0] (
19:58:35 Quit Rob2222 (Ping timeout: 240 seconds)
20:00:10 Join robin0800 [0] (~robin0800@
20:03:14 Join Rob2222 [0] (
20:03:21 Quit GermanMushroom (Read error: Connection reset by peer)
20:14:26dfktthere seems to be an issue that the first image displayed in a theme on the clip zip doesn't show the actual image, but a garbled version of the yellow rockbox logo, like in the bootup screen
20:14:56dfkti created a 1x1px dummy file which gets loaded and displayed first to work around that... is that a known behavior?
20:18:00bertrikit seems the clip zip uses a swapped RGB565 pixel format, when you display it as non-swapped, the colors will be wrong
20:19:01dfktearlier, i had the battery icons in the sbs loaded first, and it became garbled - then i put the disk activity icon before that, it became garbled - now i have that dummy file loaded first, and all is fine
20:25:12funmancriegod: double check you downloaded the correct firmware
20:25:28funmancriegod: also open the .bin file with rockbox utility, not the .zip file
20:26:19kugelFS #12358
20:26:19fs-bluebot Support for transparency in 32bit Bitmaps (patches, new)
20:39:09bertrikIs the minimum display brightness required to be equal to 1? Can it start at 0?
20:39:59 Quit Thra11 (Ping timeout: 260 seconds)
20:41:09 Quit ageis (Remote host closed the connection)
20:41:49 Join ageis [0] (
20:43:40bertrikIt looks like the simulator calculation for brightness is incorrect, it gives a brightness of 111% when min=0 and max=9
20:45:15bertrikreturn ((val - MIN_BRIGHTNESS_SETTING + 1) * 100)/MAX_BRIGHTNESS_SETTING;
20:45:57 Quit robin0800 (Quit: Leaving)
20:46:32 Join robin0800 [0] (~robin0800@
20:47:11 Quit ageis (Remote host closed the connection)
20:47:23kugel0 doesn't work I think
20:47:53 Join ageis [0] (
20:48:15 Quit ageis (Remote host closed the connection)
20:48:28 Quit keyb_gr (Ping timeout: 244 seconds)
20:52:23 Join Thra11 [0] (~thrall@
20:55:37CIA-75New commit by bluebrother (r30869): Remove buildserver_modelname value. ...
20:55:51 Join keyb_gr [0] (
20:56:41 Quit dfkt (Quit: -= SysReset 2.55=- Sic gorgiamus allos subjectatos nunc.)
20:57:04bertrikkugel, I'll fix it 0 also works
20:57:19CIA-75r30869 build result: 0 errors, 1 warnings (bluebrother committed)
21:04:30 Quit tchan (Quit: WeeChat 0.3.6)
21:06:28 Join ageis [0] (
21:06:31CIA-75New commit by bertrik (r30870): Fix simulator brightness calculation in case MIN_BRIGHTNESS_SETTING != 1
21:07:12***Saving seen data "./dancer.seen"
21:07:56 Join Horschti [0] (~Horscht@xbmc/user/horscht)
21:08:05CIA-75r30870 build result: All green
21:11:29 Quit Horscht (Ping timeout: 260 seconds)
21:12:36 Quit MethoS- (Read error: Connection reset by peer)
21:13:55lebelliumis it possible to display text over a picture (non background) ?
21:14:04lebelliumeverytime the picture displays over the text
21:15:03 Join lorenzo92 [0] (
21:15:09CIA-75New commit by bertrik (r30871): Sansa clip zip: use a state machine for scanning buttons to avoid busy-waits
21:15:47 Join MethoS- [0] (~clemens@
21:16:57CIA-75r30871 build result: All green
21:20:10 Join nosa-j [0] (
21:26:45 Quit mortalis (Quit: CGI:IRC (Ping timeout))
21:29:49wodzwhy scrollwheel doesn't generate BUTTON_REL ?
21:31:32bertrikwodz, looking into the "extra button press needed" problem? :)
21:34:11wodzLooking at lowlevel driver I don't understand why there is any difference at all. When you scroll there is a stream of events generated just like long press.
21:35:04CIA-75New commit by bertrik (r30872): Sansa clip zip: implement lcd_write_data function
21:35:33 Quit lorenzo92 (Quit: ChatZilla 0.9.87 [Firefox 7.0.1/20111008085652])
21:35:46wodzwell there IS the difference - scrollwheel driver posts directly to button_queue
21:35:52kugelwodz: you cannot "hold" the wheel
21:36:05kugeltherefore there's no good point to generate release
21:36:43CIA-75r30872 build result: All green
21:36:55wodzyou stop messing with wheel -> release
21:39:19bertrikLast time I looked at this problem, it seemed to be caused by a wait for a release between button context switches (ignoring buttons until a release is seen)
21:41:35wodzbertrik: yes, but it comes from the fact that wheel driver posts directly to button_queue and BUTTON_REL logic is omitted
21:43:34 Quit y4n (Quit: PANTS OFF!)
21:46:38lebelliumplease does someone know it it's possible to display UI viewport or text over a normal picture (non background) ?
21:47:48 Quit Jerom (Quit: Leaving.)
21:57:59 Quit stoffel (Remote host closed the connection)
22:06:23 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
22:17:23bluebrotheryou can very well "hold" the wheel, on Ipods with scrollwheel at least.
22:17:41bluebrotherthe debug menu even shows that: position and if the wheel is touched or not
22:22:51wodzthe question is more if this is desirable to fix scrollwheel driver to generate BUTTON_REL
22:24:51bluebrotherno idea. It's also that not all wheels can do this. I can't see how f.e. the e200 could do that.
22:26:38wodzas I said - for me wheel should generate stream of button events and _REL when stopped
22:27:43kugelwodz: that's what it doing when repeat is generated
22:28:04wodzkugel: no - it never output _REL
22:28:33wodzdriver posts directly to button_queue
22:28:40kugeli know
22:29:41kugelsorry, I thought it does _REL after _REPEAT
22:30:05wodzif it posts slower than once per 10ms then solution is easy
22:30:30kugelanyway, you need to poll to see you stopped moving the wheel, I suspect that's why it's not done
22:30:59kugelmost wheels are interrupt driven, so you don't see directly when it's stopping to move
22:31:48 Quit T44 (Ping timeout: 244 seconds)
22:31:49wodzkugel: you pool in button_tick anyway
22:33:32wodzI'll try to hack this during this week
22:34:14 Join lorenzo92 [0] (
22:35:06lorenzo92I have problems with implementing the hold button functionality...can someone explain me what further steps I need apart setting in the keymaps ACTION_STD_KEYLOCK?
22:35:40wodzwe could post wheel events to intermediate queue instead and combine this with other readings in button_read_device()
22:35:50funmanlorenzo92: afaik STD_KEYLOCK only works in wps
22:36:04lorenzo92uhm okay I'll test that
22:36:12wodzthis way _REL will be automagicaly generated
22:38:35criegodfunman: i did the obvious things right. Only tried auto install method though.
22:39:15 Join saratoga [0] (9803c57f@gateway/web/freenode/ip.
22:39:21saratogawodz: did you send that email to the list, I never got it?
22:39:31funmancriegod: run md5sum on the file you're trying to open
22:40:00 Join Keripo [0] (
22:40:12lorenzo92funman: uhm I cannot make it work...ah ops maybe I forgot the options in the target config (sorry hehe)
22:40:30saratogacriegod: also check that you have a newer version of rbutil, very old ones won't know how to patch a fuzev2
22:41:15wodzsaratoga: yes I did. Maybe it was filtered out as I am not subscribed to this secret ML
22:41:42saratogawodz: i believe you must be subscribed to send anything to it, but you should have been added when you got svn access
22:41:54lorenzo92is it to be enabled #define HAS_BUTTON_HOLD if I have hold key that the same as power key (no slider or what)
22:42:10saratogamight be a good idea to ask bagder or zagor to add you
22:42:26saratogain the mean time i can forward the email if you like
22:43:40criegodI think it just worked. I was giving it the name of the file i downloaded with the sansa firmware in it.
22:43:57funmanwodz: did you get gevaerts emails about RSB ?
22:44:00criegodDid it say to unzip that file and extract the bin file, and _that_ is what you use? If it did i must have missed that part.
22:44:08nsxfunman, when you got fuze v2 recovery working, was it with the switch in hold mode?
22:44:20saratogayes you must give it the firmware file, not a zip
22:44:21funmannsx: don't remember
22:44:56criegodDoes it mean anything when it says: "could not write firmware file" right after "Patching firmware" ?
22:45:10wodzfunman: I only got mail that election was finished (I guess it was from -dev ML)
22:45:14nsxthe only time I've been able to get my fuze to not boot up into the OF by shorting those recovery pins, I got this in dmesg:
22:45:27nsxso, no usable usb device
22:46:10funmancriegod: your device is mounted readonly perhaps
22:46:17nsxI'm thinking that from the start, this device wasn't "bricked" by anything I did, but is suffering hardware failure, but I'm sure by now I've voided my warranty
22:46:39funmannsx: hardware failure is not unlikely, fuzes are very weak
22:47:02criegodfunman: no, it's writable (because i can, and it lists that way)
22:47:33criegodfunman: that's a real error, though?
22:48:04nsxfunman, it seemed like one of the best music players supported by rockbox that are still in production
22:48:26saratogathe fuze is nice but it also costs about $30 so you get what you pay for :)
22:48:26 Join Teasp00n [0] (
22:48:27funmancriegod: your drive is full
22:48:50nsxmore like $70, iirc
22:48:54nsxI just got it
22:49:03criegodfunman: You're right! I filled it with music. Only left something like 4MB left on it. That's bad, huh?
22:49:18saratogayou will need enough space on the disk to store the firmware file
22:49:30nsxk it was $80
22:49:35saratogaand probably rockbox as well assuming you want the player to boot
22:49:58nsxdoes the rockbox boot loader have any recovery method of its own? it boots fine, but says "firmware not found"
22:50:03saratogansx: i think i paid 20 for my last 4GB clip and maybe 30 for my fuzev1 back in the day (on sale)
22:50:07criegodsaratoga: right. Thanks.
22:50:24saratogansx: no, we don't even have working USB on the fuzev2 anyway
22:50:42nsxI've noticed
22:51:06nsxbut perhaps it could induce this SoC into its "special recovery mode" somehow :-)
22:52:00funmannsx: OF doesnt work?
22:52:12nsxfunman, it freezes up during bootup animation
22:52:22nsxthis all happened when I loaded some addons
22:52:28nsxiconset, fonts, and a few themes
22:52:35criegodOne other thing... i noticed that the version of rockbox on the sansa page is not the same as the version listed on the page that has all the different devices on one page (i could be more specific if you don't know what i mean).
22:52:48nsxI umount'd, disconnected, and the OF started acting weird as soon as it did its synchronization thing and never booted since
22:53:04criegodThe one on the page with all the different supported models is
22:53:32nsxthat's the stable release
22:53:34criegodThe one on the sansaAMS page is ... and they are different.
22:53:51saratogathats because they're different versions ;)
22:53:52criegodThe latter seems a tiny bit older.
22:54:12criegodFrom context, it seemed to me they should both be the same (latest version).
22:54:20criegodMaybe a website updating error.
22:54:55criegodNow i have to decide which one to use.
22:55:26saratogaanyone looked at the D2+ patches on the tracker?
22:55:36saratogai don't have that player but the power management one looks important
22:57:15nsxfunman, from your memory, how long does it take for the fuze to reach this "recovery mode" after plugging in?
22:58:09Bagderwodz: me added you to committers list right now
22:58:20nsxI'm thinking something is amiss. seriously, the picture you uploaded slightly differs from mine.
22:58:24wodzBagder: thx
22:58:36nsxmine has like 4 resistors above the pins sandwiched between the two resistors
22:58:42nsxyours has two
22:59:20 Quit criegod (Quit: ChatZilla 0.9.87 [Iceweasel 7.0.1/20111004070756])
22:59:23funmannsx: different fuzev2 revisions
22:59:40funmannsx: reading what happened to your fuzev2 i'd just say Hardware error
22:59:59funmanbertrik: ping, did you try to identify the ARM cpu in clipzip ?
23:04:06bertrikfunman, no
23:05:00saratogai guess someone could try out test codec and see if its any faster
23:06:07funmanlet me give you a patch
23:06:08bertrikfunman, how can one find out
23:06:47funmansomething in cp15
23:07:15***Saving seen data "./dancer.seen"
23:07:23bertrikCP15 = 0x0005107D for the clip zip
23:07:45bertriksame as clip+
23:09:02funmanbertrik: that's config register
23:10:52lorenzo92uhm still issues with my hold key...
23:11:03lorenzo92the thing is that I cannot see any change
23:11:15lorenzo92I've set this both in menu + wps ACTION_STD_KEYLOCK
23:15:43 Join mystica555_ [0] (
23:16:35funmanbertrik: ID code register is CP15 register 0
23:21:38 Quit ageis (Remote host closed the connection)
23:22:21 Join ageis [0] (
23:23:48lorenzo92okay I managed to get keylock working. I warn you all that in SDL button define there is a subtle #define HAS_HOLD_BUTTon. It should not be there...
23:26:07bertrikfunman, do you have any reason to believe the CPU is different?
23:26:21funmanhope it's faster
23:26:49funmanlike if stocks for clip+ cpu are over
23:28:27 Join Scromple [0] (
23:33:00bertrikdoesn't compile
23:34:41wodzbertrik: move return in read_arm_info out of asm volatile ()
23:35:13wodznevermind I read it wrong
23:38:28bertrikfunman, the cpu id thing is 0x41069265
23:43:43nsxomg I finally got it to work
23:43:47*nsx hugs and kisses funman
23:44:07nsxnow let me assemble this thing back together
23:44:24nsxit presented a 8 GB usb device though, not a 1024 MB one
23:46:07 Join The_Seeker [0] (
23:46:53funmanbertrik arm926 layout rev 5
23:47:06funmannsx: doesn't look correct then, it showed the normal storage space
23:47:10 Quit The_Seeker (Client Quit)
23:47:12bertrikso, is it faster or better?
23:47:20funmanit's the same
23:47:28nsxthe of boots normally
23:47:49 Join The_Seeker [0] (
23:47:52funmanfrequency could be different though but we don't know. OF use same speed than clip+ though
23:48:06bertrikI see it supports single cycle MACs, do we use those in our codecs?
23:48:45saratogabertrik: yeah
23:48:57saratoga1 cycle 32x16 bit MAC
23:49:02saratoga32x32 is still 2 cycles
23:49:28saratogaso far its mainly ATRAC3, mpegplayer, APE that takes advantage of this
23:49:35saratogabut i have a half finished patch for mp3 on the tracker
23:50:03saratogaalso, heh:
23:51:40saratogawell i guess all codecs take some advantage of it, since most multiplies are faster due to the 16 bit wide multiplier instead of 8 bit on arm9tdmi
23:52:59 Quit markun (Ping timeout: 240 seconds)
23:53:09saratogabut for best results you want to use the armv5 multiply instructions when possible
23:55:07 Quit ReimuHa__ (Ping timeout: 260 seconds)
23:57:47 Quit The_Seeker (Ping timeout: 258 seconds)

Previous day | Next day