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

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

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

#rockbox log for 2009-10-28

00:01:13 Join scorche|sh [50] (n=scorche@rockbox/administrator/scorche)
00:01:39 Quit ender` (" We're not lost. We're locationally challenged. -- John M. Ford")
00:04:09 Quit domonoky (Read error: 104 (Connection reset by peer))
00:07:01 Quit scorche|sh ("leaving")
00:09:01 Join AndyI [0] (n=pasha_in@
00:12:25 Nick Ypsy is now known as YPSY (n=ypsy@
00:21:01 Quit AndyIL (Read error: 110 (Connection timed out))
00:24:06CIA-85New commit by rasher (r23374): Copy SDL.dll from the SDL dir which was used, rather than relying on PATH.
00:27:00 Quit esperegu (Read error: 60 (Operation timed out))
00:31:57 Part linuxstb ("Leaving")
00:34:15 Join dmb [0] (n=Dmb@unaffiliated/dmb)
00:36:48 Part toffe82
00:37:19 Quit matsl (Read error: 110 (Connection timed out))
00:38:09 Join webguest07 [0] (
00:39:37 Join webgues [0] (
00:40:47CIA-85New commit by rasher (r23375): Don't rely on {} expansion since not all shells have that.
00:42:27 Quit JdGordon| ("FUCK!")
00:44:39 Quit webgues ("CGI:IRC (Ping timeout)")
00:44:39 Quit webguest07 ("CGI:IRC (Ping timeout)")
00:47:26 Quit Galois ("Leaving")
00:51:29 Quit stripwax ("")
00:53:14 Quit pcc1 (Read error: 110 (Connection timed out))
00:53:39 Join Galois [0] (
00:53:59 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
00:54:18 Part JackWinter ("Konversation terminated!")
00:56:27 Quit dmb ("Leaving")
00:57:56 Join stripwax [0] (n=Miranda@
01:01:51 Quit Zarggg ()
01:03:50 Quit StealthyXIIGer (Read error: 104 (Connection reset by peer))
01:04:22 Quit DerPapst ("Leaving.")
01:04:40 Join StealthyXIIGer [0] (
01:06:48 Join Zarggg [0] (
01:19:15 Join elinenbe_ [0] (
01:19:16 Quit elinenbe (Read error: 104 (Connection reset by peer))
01:19:18 Nick elinenbe_ is now known as elinenbe (
01:22:27 Quit Thundercloud (Remote closed the connection)
01:23:00 Quit kugel (Remote closed the connection)
01:24:11 Quit bertrik (Read error: 113 (No route to host))
01:26:59 Quit MethoS- (Remote closed the connection)
01:27:43 Quit blithe_ (Read error: 110 (Connection timed out))
01:47:22 Quit Lynx_ (" HydraIRC -> <- *I* use it, so it must be good!")
01:53:19***Saving seen data "./dancer.seen"
01:53:36 Quit JdGordon ("Leaving.")
01:57:10 Quit jordan` (Read error: 113 (No route to host))
01:58:33 Join JdGordon [0] (
02:04:42 Nick fxb is now known as fxb__ (n=felixbru@
02:06:20 Join jordan` [0] (n=jordan@
02:08:24 Quit amiconn (Nick collision from services.)
02:08:25 Join amiconn_ [0] (i=quassel@rockbox/developer/amiconn)
02:08:43 Nick amiconn_ is now known as amiconn (i=quassel@rockbox/developer/amiconn)
02:09:04 Join killan_ [0] (
02:09:15 Quit killan (Read error: 104 (Connection reset by peer))
02:10:45 Quit pixelma (Nick collision from services.)
02:10:46 Join pixelma_ [0] (i=quassel@rockbox/staff/pixelma)
02:11:06 Nick pixelma_ is now known as pixelma (i=quassel@rockbox/staff/pixelma)
02:18:07 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
02:27:22 Quit JdGordon (Read error: 104 (Connection reset by peer))
02:33:47 Quit elinenbe (Read error: 104 (Connection reset by peer))
02:39:56 Join fejfighter [0] (
02:42:09 Quit NoGare (Read error: 145 (Connection timed out))
02:44:33 Join JdGordon [0] (
02:45:58saratogawow someone actually asked about TTA support
02:46:05saratogai thought i was the only person who noticed we didn't have it
02:47:45 Quit stripwax (Read error: 145 (Connection timed out))
02:48:24linuxstbsaratoga: Quick, fix it!
02:48:36saratogai have a half working decoder if you want it :)
02:49:25linuxstbYou should post it to flyspray. You never know, someone may be bored and finish it...
02:49:26 Join xDonny [0] (
02:49:29saratogalinuxstb: while you are here i wanted to ask about the codec api
02:49:53linuxstbOK, I'll try.
02:49:55saratogafor test_codec i want to convert it to use the audio buffer as a simple ring buffer that refills when its empy
02:50:02saratogaso that the clip and such can use test_codec
02:50:10xDonnycan anybody help me with my ipod nano2g? when i try to play a video it zooms through the video, then it pauses, then plays the video with crappy sound quality
02:50:22saratogai'm not sure what that would mean for cur_position though
02:50:36saratogacan i decrease that safely without confusing codecs or do they depend on it to monotonically increase?
02:50:47 Quit GeekShado_ ("The cake is a lie !")
02:51:17linuxstbsaratoga: I wouldn't bother using it as a ring buffer, just stop when you reach a low watermark (32KB should be OK), memmove the remaining data to the start of the buffer, read some more from the file, then continue.
02:51:30saratogalinuxstb: thats basically what i mean
02:51:56saratogabut if i do that it means that when i rebuffer i either have to fake the old position values or else decrease current position so it still points to the start of the file
02:51:59saratogaerr start of the buffer
02:52:09xDonnylinuxstb: would my problem be anything to do with the cpu clocking at a different speed?
02:52:24linuxstbxDonny: Sorry, I've no idea. I haven't been keeping up with recent Nano2G work.
02:52:40xDonnywho could help me?
02:52:40saratogaprobably just a bug in whatever timer mpegplayer uses
02:53:31saratogamaybe it would be safer to just keep a seperate buffer pointer in test_codec and use it to adjust cur_position
02:54:57linuxstbsaratoga: Hmm, looking at test_codec.c I see what you mean. Yes, I think you'll need to keep a separate pointer storing the offset in the file of the first byte in the buffer.
02:55:35saratogaok thats what i was afraid of
02:55:43saratogashould still be easy but not as trivial a fix as i'd hoped
02:56:15saratogalinuxstb: why do we have two different functions for getting data from the compressed audio buffer anyway?
02:56:38saratogarequest_buffer and read_filebuf seem to do the same thing in slightly different ways
02:57:29linuxstbNo, they do different things. One copies data and advances the pointer immediately, the other gives you a pointer to the data, with up to 32KB guaranteed to be available.
02:58:01saratogabut can't one always be used in place of the other with minor changes?
02:58:02linuxstbThe latter is obviously more efficient (and the one I would always use), but the first was implemented to make things easier when porting codecs that used that API.
02:58:11saratogaah ok
02:58:24linuxstbsaratoga: Yes, I wouldn't object to the read function being deprecated and removed.
02:58:29xDonnywhat are you two working on?
03:00:27saratogalinuxstb: we seem to use request_buffer more often then not so removing it won't be fun
03:00:40linuxstbxDonny: "test_codec" is a plugin we use to test how fast Rockbox audio codecs are running. We're talking about that.
03:00:51linuxstbsaratoga: That's the one I would keep...
03:01:01 Quit liar (Remote closed the connection)
03:01:25linuxstbsaratoga: I see my reply changed the order...
03:01:49linuxstbBy "first" I meant read_filebuf.
03:02:04saratogawell then how come you used read_filebuf so much in wma.c ? :)
03:02:16 Join midgey [0] (n=tjross@rockbox/developer/midgey)
03:02:35linuxstbThere must have been a reason....
03:02:38*linuxstb looks
03:02:53saratogais the idea that it lets you easily backtrack if needed?
03:03:46linuxstbSeems I use it for the packet parsing. The actual data is accessed via a request_buffer()
03:03:51 Quit StealthyXIIGer (Read error: 104 (Connection reset by peer))
03:03:59 Join StealthyXIIGer [0] (
03:04:08saratoga if (ci->read_filebuf(data, datalen) == 0)
03:04:19saratogathats a possible place to optimize wma then?
03:05:21saratogahmm i have to go, will look into this all later
03:06:05xDonnyalright, hopefully theres a fix for my probelm soon, i hate cracky audio
03:06:09xDonnybut i'm goin to bed
03:06:11 Quit xDonny ()
03:06:25 Join Rob2223 [0] (
03:07:57linuxstbMaybe, but I doubt it's very significant.
03:08:11linuxstbEverything is just reading 1 byte, apart from the line you quoted, and I think that's just a maximum of 18 bytes (that's the size of data[])
03:08:17 Quit AlexP ("No Ping reply in 180 seconds.")
03:08:37 Join AlexP [0] (n=alex@rockbox/staff/AlexP)
03:09:27linuxstbsaratoga: But vorbis is guilty of copying everything with a read_filebuf() - so that might be a useful optimisation.
03:09:59 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
03:10:50linuxstbBut that would probably mean writing a new Ogg parser, and discarding the one Tremor includes.
03:11:02*linuxstb leaves for bed
03:11:49 Quit HBK ()
03:11:59 Join HBK [0] (
03:21:58 Quit Rob2222 (Read error: 113 (No route to host))
03:25:12 Quit JdGordon (Read error: 104 (Connection reset by peer))
03:30:51 Quit Stephen_ ("Leaving")
03:35:19 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
03:36:15 Quit sbhsu (Read error: 104 (Connection reset by peer))
03:37:48 Join rodan [0] (
03:41:11rodanDid sansa purposely build Fuze V2 so that rockbox couldn't be used on it?
03:42:06 Join sbhsu [0] (n=a6530466@
03:43:25 Join dmb [0] (n=Dmb@unaffiliated/dmb)
03:43:47scorchehighly unlikely...
03:44:18 Quit cowgarden (Read error: 104 (Connection reset by peer))
03:44:46rodanthe difference between v1 and v2 is firmware , but is it firmware from the company or from end user updates?
03:53:22***Saving seen data "./dancer.seen"
03:59:31 Join angelwolf71885 [0] (
04:00:14 Quit angelwolf71885 (Client Quit)
04:07:49 Part rodan
04:07:57 Quit Rondom (Nick collision from services.)
04:08:08 Join Rondom [0] (
04:16:36 Join evilnick_ [0] (
04:16:36 Quit evilnick (Read error: 104 (Connection reset by peer))
04:34:57 Join FOAD_ [0] (
04:43:29 Join blithe [0] (
04:47:02 Quit adiroiban (Read error: 110 (Connection timed out))
04:51:03 Quit FOAD (Read error: 110 (Connection timed out))
04:51:03 Nick FOAD_ is now known as FOAD (
04:58:13 Quit BHSPitLappy (Remote closed the connection)
05:27:31 Quit midgey ()
05:32:49 Quit panni_ ("( :: NoNameScript 3.81 :: )")
05:36:00 Quit fejfighter (Remote closed the connection)
05:46:48 Quit Horscht ("Verlassend")
05:53:24***Saving seen data "./dancer.seen"
06:07:51CIA-85New commit by jdgordon (r23376): First go at converting the inbuilt statusbar into a skin, right now its only perfect for 176pixel wide displays (and not liked to the build system so ...
06:09:25JdGordonapparently the "volume changing" tag doesnt work in the custom statusbars
06:16:11 Join Llorean [0] (
06:18:39 Join shai [0] (
06:37:42JdGordonanyone know if you can simulate charging in the sim?
06:39:06CIA-85New commit by jdgordon (r23377): fix the charging icon so it will actually display correctly
06:51:23 Quit Llorean ("Leaving.")
06:53:41 Join fyre^OS [0] (
06:58:30 Quit sbhsu (Read error: 145 (Connection timed out))
07:06:49 Quit StealthyXIIGer (Read error: 104 (Connection reset by peer))
07:07:10 Join StealthyXIIGer [0] (
07:11:30 Quit fyrestorm (Read error: 110 (Connection timed out))
07:15:35 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
07:20:02 Join matsl [0] (n=matsl@
07:21:07 Quit fyre^OS (Read error: 60 (Operation timed out))
07:29:25 Join mt_ [0] (n=mtee@
07:31:45 Quit StealthyXIIGer (Read error: 60 (Operation timed out))
07:32:33JdGordonwould anyone madly object to me fixing the viewport parser in the skin thingy so you could have a negative value for the pixel placement/dimensions? (i.e -8 for height) which would mean X pixels away from the edge.. so -8 height would be LCD_HEIGHT-8?
07:33:37JdGordonto make sharing skins between different displays simpler
07:36:12 Join sbhsu [0] (
07:42:51*JdGordon is slightly surprised how easy that was actually
07:43:20 Quit mt (Read error: 113 (No route to host))
07:44:12 Join mt [0] (n=mtee@rockbox/developer/mt)
07:44:38CIA-85New commit by jdgordon (r23378): Allow the x and y pixel values of viewports to be a negative number.. ...
07:49:09JdGordonto go along with that commit... it would be nice if we used "u" for unsigned ints in that list parser and "d" for signed ints to make it slightly more type safe
07:49:19 Quit mt_ (Read error: 145 (Connection timed out))
07:53:25***Saving seen data "./dancer.seen"
08:02:16 Quit phanboy_iv ("Leaving")
08:08:10 Join Rob2222 [0] (
08:08:19 Join kugel [0] (n=kugel@rockbox/developer/kugel)
08:08:46amiconnsaratoga: Why does the clip require you to make test_codec imprecise? Can't you just use small enough test files?
08:09:17kugelJdGordon: %mp can do recording (and fm radio) status too
08:09:49JdGordonthe CustomWPS page needs to be updated then...
08:10:27kugeloh, I seem to have forgotten this
08:12:33JdGordonsomeone was complaing about this in the forum very recently also
08:13:36 Quit dmb (Connection timed out)
08:13:44JdGordonarg... putting the values for rec and radio in #ifdefs means the skins cant be shared easily
08:14:38 Join Grahack [0] (
08:14:45kugelJdGordon: no, it's compatible
08:15:09JdGordonoh right
08:15:17*JdGordon isnt entirely awake
08:15:24 Quit Rob2223 (Read error: 60 (Operation timed out))
08:15:46kugel%cs was a poor choice for current screen :/ all RTC tags start with %c
08:16:29 Join dmb [0] (n=Dmb@unaffiliated/dmb)
08:16:43JdGordonwe dont really have much wiggle room there... very soon we will need to think of how to go to a different label scheme
08:22:09kugelalright, I updated CustomWPS
08:23:09amiconnTwo letters, small+capital allows 2704 combinations...
08:24:06 Join esperegu [0] (n=quassel@
08:25:36JdGordonyes, but they arnt very understandable
08:26:38 Join ender` [0] (
08:27:13 Join Zagor [242] (n=bjorn@rockbox/developer/Zagor)
08:27:13JdGordonits 2756 :) 52 single letter tokens also
08:37:18amiconnAre there such?
08:37:43*amiconn doesn't remember, because he didn't change his wps'es in ages
08:37:54kugel%C and %X
08:38:30 Join flydutch [0] (
08:38:35kugeland %D, %x, %s, %t
08:42:54 Quit bubsy ("Party time, EXCELLENT! *wring wring wring!* Party on, Wayne! Party on, Garth!")
08:44:46pixelma%D ?
08:45:38pixelmaisn't %d for directory and needs a number for the level, like %d1 etc.
08:46:00pixelmabut there are %V and %T too
08:46:48kugel%D is %d for the next track, apparently
08:47:15pixelmayes, but does pure %D exeist without the number?
08:47:28pixelmaor exist
08:48:40 Join bubsy [0] (n=bubsy@
08:48:44kugelprobably not. but technically the tags are %d/%D
08:52:57pixelmaah ok, similar things apply to e.g. %t
08:53:19pixelmaexcept if you want to use the default 2 seconds
08:54:03pixelmakugel: any news on the progressbar problem, or did I miss something?
08:55:58 Join bertrik [0] (
08:56:06 Join petur [50] (n=petur@rockbox/developer/petur)
08:56:30kugelno, I haven't had a chance to look into it
08:56:47 Quit bubsy ("Party time, EXCELLENT! *wring wring wring!* Party on, Wayne! Party on, Garth!")
08:58:19 Join bubsy [0] (n=bubsy@
09:00:49 Quit bertrik (Client Quit)
09:00:49kugelI think I see the problem
09:01:24 Join mt_ [0] (n=mtee@
09:01:50 Join Overand [0] (
09:02:07OverandI have an iPod 5th gen (Video, 60 gig) - and the clock keeps getting set back exactly one year.
09:02:51OverandI'm not sure if it's a rockbox thing, or an ipod-end thing - it does occasionally get plugged into a windows machine (at work) in 'iPod mode.' (rescue mode has much slower USB transfers) - Curious if anyone's noticed anything like this.
09:03:03Overand(Incidentally, I checked the date on the work machine and it's correct)
09:03:17Overand(Also I don't think the work machine has any apple software instlaled on it)
09:03:24kugel%pb is problematic, it's the only tag which demands to know the default vp dimensions at parsing time
09:04:12 Quit HBK (Read error: 104 (Connection reset by peer))
09:05:08 Join HBK [0] (
09:05:12 Quit mt (Read error: 145 (Connection timed out))
09:05:19 Nick mt_ is now known as mt (n=mtee@
09:06:04 Join adiroiban [0] (n=adiroiba@
09:09:05 Quit adiroiban (Client Quit)
09:09:14 Join adiroiban [0] (n=adiroiba@
09:13:22 Join maruk [0] (
09:18:09 Quit Lss (Read error: 131 (Connection reset by peer))
09:18:26 Join Lss [0] (
09:19:16 Quit bubsy (Read error: 54 (Connection reset by peer))
09:20:27pixelmait doesn't only need to know the default viewport dimensions, in case you put it in an own viewport it would need to know that vp's dimensions as I think it should use that viewports width then
09:20:28 Join bubsy [0] (n=bubsy@
09:21:47pixelmaI think that was working before but am not sure
09:21:58kugelthe dimensions of the other viewports are known, that's not the problem
09:22:35 Join Thundercloud [0] (
09:23:12kugelmoved setting up the default viewport to the end of parsing (so that %wd/%we is known, and because the default vp got in the way of sbs)
09:23:42 Quit Grahack ("Leaving.")
09:27:06 Quit shaggy-h (Read error: 104 (Connection reset by peer))
09:45:50 Quit kugel (Read error: 110 (Connection timed out))
09:48:51 Join kugel [0] (n=kugel@rockbox/developer/kugel)
09:48:57 Quit Thundercloud (Remote closed the connection)
09:49:14 Join midijunkie [0] (n=Miranda@
09:53:29***Saving seen data "./dancer.seen"
10:05:05 Quit mt (Read error: 113 (No route to host))
10:19:18 Quit midijunkie ("?(???~•~)?")
10:19:39 Join midijunkie [0] (n=Miranda@
10:20:20 Quit adiroiban (Read error: 145 (Connection timed out))
10:20:44 Quit jon-kha (Read error: 148 (No route to host))
10:21:19 Part midijunkie ("I'm a happy Miranda IM user! Get it here:")
10:32:39 Join shaggy-h [0] (
10:46:52 Join Grahack [0] (
11:38:58 Quit Xerion (Read error: 104 (Connection reset by peer))
11:39:13 Join funman [0] (
11:39:59 Join Xerion_ [0] (
11:40:24 Nick Xerion_ is now known as Xerion (
11:40:29 Quit kugel (Read error: 60 (Operation timed out))
11:47:54 Join DerPapst [0] (
11:53:31***Saving seen data "./dancer.seen"
12:13:44 Nick YPSY is now known as Ypsy (n=ypsy@
12:20:00 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
12:24:57 Join adiroiban [0] (
12:25:47 Join MethoS- [0] (n=clemens@
12:57:28 Quit alexbobp (Read error: 145 (Connection timed out))
13:02:01 Quit antil33t (Read error: 104 (Connection reset by peer))
13:02:07 Join antil33t [0] (n=Mudkips@
13:09:17 Quit preglow (
13:09:17 Quit crwl (
13:09:43NJoinpreglow [0] (
13:09:43NJoincrwl [0] (
13:25:32 Join teru [0] (
13:42:16 Quit AB3JU (Read error: 60 (Operation timed out))
13:42:19 Join dz` [0] (
13:42:50 Join jon-kha [0] (i=jon-kha@
13:53:34***Saving seen data "./dancer.seen"
13:58:26 Quit DerPapst ("Leaving.")
14:08:30 Quit GodEater (Read error: 113 (No route to host))
14:08:52 Quit Bob_C_ (Read error: 104 (Connection reset by peer))
14:09:15 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
14:10:27 Join itcheg [0] (i=4117734b@gateway/web/freenode/x-ywhxusopvwwrgogw)
14:10:46 Quit MethoS- (Remote closed the connection)
14:23:26 Join evilnick_B [0] (i=0c140464@rockbox/staff/evilnick)
14:24:16 Join _zic [0] (
14:33:21 Join panni_ [0] (
14:37:24saratogaamiconn: the idea was to pause the timer while rebuffering
14:38:00saratogathe problem with using small files is that 300-400KB is only a couple seconds of lossless files, and it would mean not using the same test files on the AMS targets as existing ones
14:40:50 Quit funman ("free(random());")
14:54:06 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
15:04:34 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
15:13:32 Quit _zic (Read error: 113 (No route to host))
15:18:24 Join jasio [0] (
15:20:45 Join Crackerizer [0] (
15:21:47CrackerizerMay i ask a question, is there any math function in rockbox?
15:21:57Crackerizersuch as ceil, log, etc..
15:24:08ZagorCrackerizer: we avoid using floating point math
15:25:14 Nick fxb__ is now known as fxb (n=felixbru@
15:25:25 Join kkurbjun [0] (n=kkurbjun@rockbox/developer/kkurbjun)
15:25:33CrackerizerZagor: so, how does we implement these functions? or we just dont use it?
15:26:03ZagorCrackerizer: it depends. tell us what you want to do.
15:27:19CrackerizerZagor: I'd to find ceil(log2(A)). To me, ceil implementation is easy but not log2.
15:29:06Zagoryes, but what feature are you trying to implement?
15:31:02 Quit itcheg (Ping timeout: 180 seconds)
15:31:57TheSevenlog2 isn't hard either
15:32:02CrackerizerI'm working on diacritic character patch. There is a folk here suggests me an algorithm which need these functions.
15:32:35CrackerizerYou can have a look at patch no#10720
15:33:12TheSevenint result = 0; while (input) { input >>= 1; result++; } return result; (can be further optimized)
15:33:24TheSeventhat's ceil(log2(input))
15:35:44CrackerizerTheSeven: Oh, that's very simple..
15:40:59CrackerizerTheSeven: How do you find this code?
15:41:55TheSevenCrackerizer: i just quickly wrote it?
15:43:00 Join toffe82 [0] (n=chatzill@
15:45:40CrackerizerTheSeven: I know you did :). But what knowledge do you use to construct such a simple algorithm out of a quite complex math function?
15:46:10CrackerizerThis is very interesting to me.
15:46:26Tornelog2 is just the number of digits in a number ehen written in binary
15:46:52Tornewell, if you round it to ceil
15:47:00Tornesame as any other logarithm base
15:47:38CrackerizerTorne: Oh, I see...
15:47:40saratogalog2 is just a shift
15:47:58saratogaideally you should try to precompute the shift amount though so that the loop is unneeded
15:48:35Torneer, log2 is not a shift, it's the inverse
15:48:59CrackerizerSaratoga: Ah, true.
15:49:23saratogaoh yeah
15:49:43saratogai misread what that code was trying to do
15:50:15TheSeveni just found that function yesterday in an iBSS disassembly, they did it in quite an interesting way that only needed 2 AND instructions and then a maximum of 8 loop iterations
15:50:16CrackerizerYou can have a look at patch#10720
15:50:22saratogabut ideally you should try to avoid doing the log2 calculation if you can, many parts of rockbox simply store numbers as the power of 2 needed, so that they don't have to compute logs
15:50:58CrackerizerI'm working at diacritic character patch.
15:51:17 Join DerPapst [0] (
15:52:05Crackerizerand need suggestions.
15:53:35***Saving seen data "./dancer.seen"
15:54:52 Join itcheg [0] (i=4117734b@gateway/web/freenode/session)
15:57:18 Join b0hoon [0] (
15:57:26 Join Strife89 [0] (n=nds@
15:57:31CrackerizerSaratoga: Emm, the input is actually constant. I would use the precalculated value instead.
15:57:50CrackerizerThank you for your help.
15:59:37 Quit b0hoon (Client Quit)
16:00:27jasioHi. I just tried software keylock on an Onda VX777. It doesn't lock the touchscreen. Is that expected in the current status or did I forget to enable something?
16:08:29 Quit itcheg (Ping timeout: 180 seconds)
16:09:59 Join StealthyXIIGer [0] (
16:12:45 Quit teru ("Quit")
16:16:34 Join jgarvey [0] (
16:18:38 Quit Rob2222 ()
16:22:44 Join kugel [0] (n=kugel@rockbox/developer/kugel)
16:23:54 Part LinusN
16:25:13 Join phanboy4 [0] (
16:35:16 Join Rob2222 [0] (
16:38:33jasioAnybody knows if software hold/lock can lock the touchscreen as well as the keys?
16:39:03jasioI tried on VX777 and the one key is locked but not the screen.
16:41:40*domonoky doesnt know about vx777, but as this is a new target (and all touchscreen targets are new and in development) i would suspect this is just not correctly implemented at moment.
16:42:52jasioi thought so. looking at the source but not sure yet how the soft lock works
16:43:00kugeldomonoky: hey. did you test whether collecting 5 messures from the mini2440 is really needed?
16:43:13pixelmasoftware lock in its current state is only used on the Archoses, physical button only, targets so I guess locking the touch screen as well hasn't been implemented yet for the touch screen targets there.
16:43:13kugelmini2440's touchscreen
16:44:07domonokykugel: not really, i just copied the code from d2. But i saw variing touchscreen readings in my early experimentation.. so some filtering is probably needed.
16:44:09kugelalso, the test program from friendlyarm uses interrupts to read the touchscreen
16:45:17kugeldomonoky: I can imagine that readings are *slighly* off (a hand can't be entirely steady). IIRC the D2 had more serious wrong readings
16:45:25domonokykugel: yes, the example uses interrupts for adc readings. but i wanted to start simple, also the current adc code from gigabeatf doesnt use ints.
16:46:23domonokykugel: i saw completly off readings.. like 512 - 500 - 0 - 490 in my early experiments... but that might be better now, i dont know.
16:47:15kugeloh ok
16:47:59domonokykugel: feel free to test and improve if you want. (uart_printf or logf over serial is very good for debugging, if you have a serial cable) :-)
16:48:13kugelI have one now
16:48:36kugelI just don't know how to use it tough, never messed with serial ;)
16:50:29domonokykugel: then learn it, it isnt hard :-)
16:51:23 Join xnyhps [0] (n=xnyhps@2001:470:1f14:da:219:e3ff:fe36:98e5)
16:52:05 Join esperegu_ [0] (n=quassel@
16:53:47jasioanother question: in the button_mapping structure how is pre_button_code used? random example, { ACTION_PS_TOGGLE_MODE, BUTTON_MENU|BUTTON_REPEAT, BUTTON_MENU } what does the last BUTTON_MENU do?
16:54:56 Join Bob_C [0] (
16:55:19 Quit Strife89 ()
16:56:37 Quit esperegu (Read error: 113 (No route to host))
16:57:53 Quit mc2739 (Read error: 131 (Connection reset by peer))
16:57:58 Join mc2739_ [0] (n=mc2739@rockbox/developer/mc2739)
17:00:53 Join GodEater [0] (
17:02:20kugelBUTTON_MENU needs to be pressed first
17:03:23 Quit Zagor ("Don't panic")
17:03:40jasioso press BUTTON_MENU once, release, press and hold ?
17:04:35kugelno, press BUTTON_MENU, and then press both without releasing BUTTON_MENU
17:05:46jasiobut the other button is also BUTTON_MENU
17:06:23domonokykugel: both ? this action triggers just on long press of BUTTON_MENU, the prereq is need in this case to distinguish from other button combinations with BUTTON_MENU i think.
17:06:41kugeloh right, it reads BUTTON_REPEAT, sorry
17:07:12 Quit shaggy-h (Read error: 110 (Connection timed out))
17:07:13*kugel somehow read BUTTON_REWIND
17:08:06kugeljasio: in that case, it mostly means that this actual will fire only once
17:08:55kugelif you start holding BUTTON_MENU, there's only a single moment where BUTTON_MENU|BUTTON_REPEAT is true and where the previous button was BUTTON_MENU only
17:09:09 Quit thegeek_ (Read error: 113 (No route to host))
17:09:11jasiook. I'll experiment a bit more. I'm trying to get a decent keymap for the VX777. There is only one button but I can at least have short press and hold
17:09:49jasioit's starting to make sense ;)
17:10:55 Join thegeek [0] (
17:11:18kugelthe button are reads every 10ms. the action will only be fired off if BUTTON_MENU|BUTTON_REPEAT is read and the previous read (i.e. 10ms ago) was BUTTON_MENU only. that's why it's called prerequisite
17:15:51 Quit DerPapst ("Leaving.")
17:16:24pixelmajasio: the touchscreen targets also have the generic touchscreen keymap - which uses the screen divided into a grid as buttons and also on some screens already you can touch things directly
17:17:46jasiopixelma: the touchscreen on the vx777 is already working pretty good both in grid and absolute modes.
17:19:16jasioone thing I'll have to look at is that the bottom part doesn't work as expected. It's like there is an area that is mapped like a button (does nothing in WPS, sort of Cancel in menus)
17:23:55 Quit StealthyXIIGer (Read error: 110 (Connection timed out))
17:25:33 Quit shai ("Leaving")
17:27:38 Join funman [0] (n=fun@rockbox/developer/funman)
17:27:50 Join Hillshum [0] (
17:29:10Bob_Cwahey, mini2440 can now emulate a very loud squirrel ;)
17:29:21 Quit dz` (
17:29:21 Quit Hadaka (
17:29:25 Join AB3JU [0] (
17:29:52 Join Naked [0] (n=naked@
17:30:05 Nick Naked is now known as Hadaka (n=naked@
17:30:55 Quit yawny (Client Quit)
17:31:47funmandionoea: any progress on Clip ?
17:32:47 Quit Xerion (Read error: 104 (Connection reset by peer))
17:33:13 Join elcan [0] (i=user36@
17:33:20 Join Xerion [0] (
17:33:47 Quit elcan (Client Quit)
17:34:03 Quit phanboy4 ("Leaving")
17:34:23 Join elcan [0] (
17:35:05JdGordon(01:23:15 AM) kugel: moved setting up the default viewport to the end of parsing (so that %wd/%we is known, and because the default vp got in the way of sbs) <- that could cause other issues.... The default viewport *should* be known before parsing starts... the %we/d tag logic are (or can be) made smart enough to fix it correctly
17:40:06kugelJdGordon: I saw that now too
17:41:03 Join n1s [0] (n=n1s@rockbox/developer/n1s)
17:41:47JdGordonok good :)
17:42:02JdGordonI had nightmares about what you might do for a possible fix :D
17:43:13*kugel gets tired of getting told wrongly how his own code works
17:43:35JdGordonthats what happens when there are no comments in the code
17:49:16kugelwell, I could've added more, but that wouldn't have changed anything (just read your reply on the ml).
17:50:16 Quit maruk ("Leaving.")
17:50:46evilnick_BJdGordon: RE: your last reply on the ml; does that mean that you're suggesting removing users being able to change fonts in favour of only using .cfg for skins?
17:51:12 Join mcuelenaere [0] (
17:53:31JdGordonyes, although only to illustrate my point... I'm not seriously suggesting that
17:53:40***Saving seen data "./dancer.seen"
17:58:54jasioGot soft lock of the touchscreen by just moving a few lines in action.c :)
17:59:11 Join bubsy_ [0] (n=bubsy@
17:59:16jasiobut my keymap is still not quite right
18:00:25 Quit bubsy_ (Client Quit)
18:02:07 Quit bubsy (Remote closed the connection)
18:02:38 Quit kugel (Remote closed the connection)
18:07:16 Join bubsy [0] (n=bubsy@
18:09:07domonokyBob_C: nice !
18:09:49mcuelenaeredomonoky, Bob_C: any reason why you guys didn't use the "generic" touchscreen calibration for the mini2440?
18:09:50 Join phanboy4 [0] (
18:10:49domonokymcuelenaere: it uses the generic touchscreen calibration.. but i use a extra correction in the driver, because the touchscreen is mounted sideways.
18:11:07*mcuelenaere checks
18:11:34 Quit petur ("*plop*")
18:11:40domonoky0,0 is bottom left on mini2440 touchscreen driver.
18:12:26 Join JdGordon1 [0] (
18:12:29domonokybut anyone should feel free to improve it, i just copied from the d2 touchdriver so that it works :-)
18:12:32mcuelenaereok, I just presumed that because you added that code you didn't
18:12:44 Quit jfc^3 (Read error: 104 (Connection reset by peer))
18:13:14 Join jfc^3 [0] (
18:15:26Bob_CI haven't looked at touchscreen so can't comment...
18:17:03 Join shai [0] (
18:27:30 Quit Grahack (Client Quit)
18:28:00 Join kugel [0] (n=kugel@rockbox/developer/kugel)
18:28:40 Join JdGordon2 [0] (
18:28:56 Quit JdGordon1 ("Leaving.")
18:36:18 Quit matsl (Read error: 145 (Connection timed out))
18:42:56*kugel wonders why parse_list doesn't simply call atoi()
18:43:28 Quit phanboy4 (Read error: 104 (Connection reset by peer))
18:44:27 Join Strife89 [0] (n=michael@
18:44:46kugelit seems parse_list is reimplementing it. linuxstb ?
18:47:35 Join DerPapst [0] (
18:52:53 Quit TheSeven ("ChatZilla 0.9.85 [Firefox 3.5.3/20090824101458]")
18:54:33 Quit thegeek (Read error: 104 (Connection reset by peer))
18:56:02CIA-85New commit by Domonoky (r23379): make the mini2440 simulator compile.
18:57:03 Nick Ypsy is now known as YPSY (n=ypsy@
18:59:11 Join TheSeven [0] (n=theseven@rockbox/developer/TheSeven)
19:00:33 Join JdGordon| [0] (n=Miranda@nat/microsoft/x-thfdwtfxqfysaeuc)
19:02:51 Quit JdGordon2 (Read error: 110 (Connection timed out))
19:04:21 Quit Horscht ("Verlassend")
19:07:37 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
19:09:20linuxstbkugel: Doesn't parse_list increment the pointer past the integer? atoi just returns the integer.
19:10:34 Nick fxb is now known as fxb__ (n=felixbru@
19:11:20 Quit Strife89 ("Switching OSes.")
19:12:49kugelit could do atoi(), and then p += strchr(p, sep), couldn't it?
19:14:19linuxstbBut then it's doing the work twice... And the semantics are not the same - e.g. it would accept something like 10a (I would hope parse_list would reject that)
19:17:27 Join Strife89 [0] (n=michael@
19:19:48 Join bertrik [0] (
19:24:14 Quit faemir ("Leaving")
19:25:02kugellinuxstb: nope
19:29:41 Join cas [0] (n=chatzill@
19:36:30 Quit Crackerizer ("CGI:IRC")
19:44:30 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
19:49:02JdGordon|do any targets not have enough buttons for up/down/left/right/select and maybe a cancel action?
19:50:52bertrikmeizu m3 :P but it's not really a target yet. It has an up/down touch strip, a top and a bottom button and some kind of cancel button, but no left/right
19:51:06kugelthat should be the minimum buttons for all targets. I've been messing with that in my PLA rework
19:51:50GeekShadowbertrik, I got one if there are tests to do ;)
19:52:29 Join Sajber^ [0] (
19:53:43***Saving seen data "./dancer.seen"
19:59:29 Quit Hillshum (Read error: 104 (Connection reset by peer))
20:10:33 Quit DerPapst ("Leaving.")
20:10:55JdGordon|has anything been done about the fuzes volume issues?
20:11:25JdGordon|apart from it being generally quiter than other targets.. is the left ear quieter than the right? or am i going deaf in one ear more than the other?
20:11:56kugelI think it's you (or your headphones). my fuze has no volume issues
20:12:43 Quit jgarvey ("Leaving")
20:12:56JdGordon|it used to be noticably quiter than other targets with the same headphones... didnt it?
20:13:09JdGordon|targets being other ams ones
20:14:36kugelI can't say that, my clip is pretty much the same as my fuze in that regard
20:21:17casdomonoky: ping
20:21:33domonokycas: pong
20:22:21casdomonoky: I saw that there is a lot of code now on SVN for Mini2440. The wiki page says audio at 10%... are you working on audio or Bob_C ?
20:23:24domonokyafaik Bob_C is working on that... (he said earlier, that he already heard some squirrel sound :-) )
20:23:50casdomonoky: ok, and you?
20:25:46domonokycas: nothing at moment.. and you ? :-)
20:26:45casdomonoky: nothing :-)
20:27:28casdomonoky: I am being busy with other things for the Lyre project, not Rockbox directly related.
20:28:20kugelbtw, it's a pitty that there's no proper power off button defined
20:28:45kugelI didn't find a way to save the settings yet, shutting down would be easiest :)
20:30:06domonokykugel: how about idle shutdown ? (but mini2440 cant really shutdown at moment)
20:31:51JdGordon|I thought the idle notify stuff isnt enabled for flash targets so it just writes immediatly?
20:32:10JdGordon|so settings should get saved immediatly?
20:33:27kugelwell, it's not like that
20:35:18pixelmano, settings won't get written immediately on flash targets
20:39:46Bob_CI ran out of buttons to define a Power btn
20:39:49JdGordon|can I bribe someone to take 5 min (actually its going to be more than that now :p ) to read the current dev-ml thread and throw in their opinion?
20:40:50JdGordon|the argument has nothing to do with code.. it is entirely the logic behind it... so you only need to have an idea how the skins should work
20:40:59kugelBob_C: it doesn't need to be a separate button. actually, most targets have the power of button on a button that's almost as used as the non-power of ones
20:42:15Bob_Cwell, i defined UDLR menu + select,but we can define buttons and keymap however we like I guess
20:42:23 Quit Galois ("Leaving")
20:42:24Bob_CI built an external keypad anyway
20:42:36kugeloh really?
20:42:46 Quit TheSeven ("ChatZilla 0.9.85 [Firefox 3.5.3/20090824101458]")
20:42:47kugelcan I see a picture?
20:44:29Bob_CSure, I will have to search for my CF card first
20:45:18Bob_CIts just some stripboard and cheap buttons, nothing special
20:51:46 Quit funman ("free(random());")
20:52:18 Quit GodEater ("leaving")
20:52:35 Join GodEater [0] (
20:52:53bertrikhas anyone ever tried to make splash have rounded corners?
20:53:14gevaertsWe need WSS!
20:53:45n1sJdGordon|: i can tell you this; i have no idea how all this vieport stuff interacts with different parts of the ui or how it affects the drawing, but i sure would love to see some documentation of it like the documentation of the graphics api i the wiki
20:53:56kugelbertrik: I don't think so, but I doubt it's worth the effort
20:54:13JdGordon|n1s: documentation is for nancy boys! :D
20:54:25bertrikI think it could look slightly cooler that way, I'm just asking if anyone ever tried it
20:55:04n1sJdGordon: yes, well it's also generally good for understanding the *intended* function of the code :)
20:55:11 Join TheSeven [0] (n=theseven@rockbox/developer/TheSeven)
20:55:19kugelit's basically only 3 simple drawing function calls right now (border, filling and text). round edges need additional logic
20:55:37 Join fyrestorm [0] (
20:56:01JdGordon|gevaerts: while splash screen? its definetly doable...
20:56:03n1sbertrik: never heard of any attempts at that
20:56:36 Join Galois [0] (
20:56:57JdGordon|actually... a splash skin could be alot of fun :)
20:58:10*n1s would like the list selection bar to be skinnable beyond gradients too, imo the whole gradient thing is misplaced and should be replaced by more flexible bitmap stuff
20:58:29*bluebrother would like to have inline settings :)
20:58:57JdGordon|I probably still have my old screenshots of inline settings
20:58:57bertrikinline settings, what is that?
20:59:29JdGordon|having the setting in the menu so you dont have to go into another screen to change then
20:59:41JdGordon|and so you can see the value without entring it also
21:00:23domonokykugel: idle shutdown seems to correctly save settings on mini2440 :-)
21:01:55bertrikindeed just being able to see settings when the selection bar is over an item would be nice already
21:02:22n1sI also think that too much of the current gui is fullscreen based, if you enter a yes/no setting for example a smaller window to display the coice would look better on large displays
21:02:46bluebrothernah, popups are annoying. Just make the setting inline :)
21:02:48 Quit Utchybann (Read error: 113 (No route to host))
21:03:05bluebrotherthat original patch was really nice.
21:03:10kugelwith a ballon or roll effect please :)
21:03:15n1sbluebrother: inline has disadvantages
21:03:22JdGordon|SKINS EVERYWHERE!!!!
21:03:35*bluebrother slaps kugel with a ballon or roll effect
21:03:42n1s... with settings that have many values for example
21:03:51bluebrothern1s: true, but you don't need to use it everywhere.
21:04:11pixelmajust prepend the actual value behind the setting name
21:04:22 Join DerPapst [0] (n=DerPapst@
21:04:26JdGordon|the hardest thing about inline settings, is for the settings which to the user look like regular ones, but in the code are actually a different beast
21:04:49bluebrotherhuh? What's a "regular" setting?
21:04:51n1swell, non-fullscreen windows can be useful in other contexts than settings, such as theonplay menu
21:05:14*kugel thinks making the simple on/off ones inline for the start would be a good idea
21:05:17JdGordon|n1s: I agree... thats easily fixed though.. just pass a viewport to its do_menu() call
21:05:55n1sJdGordon: i would expect stuff in the background to keep updating, will it? :)
21:05:57JdGordon|bluebrother: regular being the ones which have all the options in a list... except there are at least one setting which look like that but arnt
21:06:01pixelmaI don't like that you can't see the amount of values easily (and e.g. if it would be easier to use wrap and go backwards - and that it wastes space the way it was done and either looked ugly on really small and really large screens
21:06:23 Join daggett [0] (n=denis@
21:06:45JdGordon|you can see the difference with the icons in the menus.. any which have the setting icon can work, any that have the one for a submenu or "function" cant
21:06:46*gitster asks pixelma to re-review his comment on FS #10680 when he has time.
21:07:04kugels/he/she/ :)
21:07:18amiconnTheSeven: log2 for 32 bits is easier than that. Pure C needs a maximum of 5 if() {shift, add} blocks
21:07:55TheSevenyes, you can of course divide and conquer it...
21:08:03JdGordon|the 3 bottom peak meter settings cant work for example
21:08:06amiconnAnd we already have a number of implementations for that. On ARMv5+ there's the clz instruction, which just needs single correction afterwards
21:09:22pixelmagitster: you were right, I confused the two. It's not an undocumented use of %St though as it is described on the CustomWPS wiki (which is very often more up to date than the manual because it seems to be easier to update)
21:09:38pixelmamore of a manual bug then
21:10:02gitsterpixelma: thanks; then there is nothing that blocks FS #10680 and FS #10681, I guess.
21:12:55bluebrotherthe original patch which I really liked:
21:13:57linuxstbkugel: What did you mean by "nope" earlier?
21:14:05kugelgitster: I think so too
21:14:09 Quit TheSeven ("ChatZilla 0.9.85 [Firefox 3.5.3/20090824101458]")
21:14:22kugellinuxstb: Looking at the code, it doesn't look like it would reject 10a
21:14:52kugelbut maybe I didn't look hard enough
21:15:23n1samiconn: isn't our ffs (!) function basically a log2?
21:15:23 Quit bluebrother (Read error: 104 (Connection reset by peer))
21:15:52linuxstbkugel: line 959 of misc.c
21:15:56JdGordon|why should parse_list reject 10a anyway?
21:16:01kugelah yea, I didn't see that before
21:16:05amiconnn1s: no
21:16:10linuxstbJdGordon|: Because "10a" isn't an integer...
21:16:11kugelI only looked at the 'd' case
21:16:38JdGordon|10 is though...
21:16:39n1samiconn: ah, right, that is first == lsb?
21:16:49linuxstbJdGordon|: Yes, and "10" would be accepted as such...
21:16:55amiconnffs finds the least significant "1" bit. log2 finds the most significant "1" bit
21:17:04JdGordon|and then it should skip past a untill the next seperator
21:17:45JdGordon|any objections to using u for unsigned and d for signed ints in that parser? we wont always want the ability to get bac negative values for ints
21:18:19 Join TheSeven [0] (n=theseven@rockbox/developer/TheSeven)
21:18:35 Join liar [0] (n=liar@
21:18:45n1s"first" is ambigous
21:20:07 Join NoGare [0] (i=NoGare@
21:21:14linuxstbJdGordon|: Why would you want to accept "10a" as an integer? Shouldn't the wps parser be as strict as possible?
21:21:16 Join bluebrother [0] (
21:22:22bertrikhm, I could have a look at the meizu NAND contents with the samsung DFU tool, to learn more about the FTL
21:24:08JdGordon|linuxstb: two different things.. that list parser is used by more than just the skin parser, and yes, 10a isnt an integer but I dont really see why the 10 couldnt be accepted.. Its not something I really care about either way.. im just saying
21:24:38 Quit TheSeven (Connection reset by peer)
21:29:05 Part xnyhps
21:29:51 Join tomers [0] (
21:30:22tomersNew update to "FS #10728 - Cowon D2: Add support for D2 in rbutil" - hope this is the last one
21:31:46*tomers Got to go. Will read your comments (if there will be any) from the log tommorrow
21:31:58 Quit tomers (Client Quit)
21:32:11 Join domonoky1 [0] (
21:32:47 Join TheSeven [0] (n=theseven@rockbox/developer/TheSeven)
21:38:47bluebrothertomers: IMO linuxstb has a point about splitting changes from moving the file. At least posting two consecutive patches (with one only moving the file) would make it much easier to review the changes.
21:39:50 Quit mcuelenaere (Read error: 60 (Operation timed out))
21:40:00*n1s very much agrees with that
21:40:20linuxstbI don't think the file move is controversial - that can just be done...
21:40:38n1si even think svn refuses to commit a move and change unless you force it
21:40:39bluebrothertomers: also, have you thought about how to handle the two-drives issue for the D2? In the current state it's something I'd call definitely broken. Users expect it to "work".
21:40:50 Join mcuelenaere [0] (
21:41:36CIA-85New commit by nls (r23380): Use array for quickscreen item settings to simplify some logic
21:41:45bluebrothern1s: afaik you can do that even without forcing it. And doing move + change in the same commit isn't something I'd consider problematic either −− it's the diff format that makes it hard to review such a patch as it doesn't know about the move operation.
21:43:03n1sbluebrother: i think such commits are harder to review, and an additional commit to move is not a big cost
21:43:58bluebrothern1s: I do agree with that, but I don't think it's that important for the final commit −− it should have been reviewd that time anyway :)
21:44:36bluebrotherbut there's no harm in doing it in two steps so I don't see a reason to force one commit
21:44:49n1sbluebrother: yes, unfortunately you sometimes see some code and wonder when it got in and why
21:45:09linuxstbWhy don't one of us just move it now?
21:45:30bluebrothersure, would be a ... solution ;-)
21:45:37*n1s has no idea about what specific file we are discussing
21:45:57linuxstbtools/mktccboot.c - move it under rbutil/ with the other mk* tools
21:46:28 Join phanboy4 [0] (
21:46:28linuxstb(as it's not needed by the build system, it's not a problem IMO)
21:46:43n1si'd say go for it
21:47:34 Quit domonoky (Read error: 110 (Connection timed out))
21:48:06*linuxstb wonders why tomers's Makefile is so complex though, compared to the single rule in tools/Makefile
21:48:37*linuxstb will just do a trivial move, and tomers can then sync his patch, so we can see what he's actually changing
21:49:08bluebrotherI guess it's his attempt to merge it with the mkamsboot Makefile. Though I'd prefer that to be a separate change too.
21:49:55 Quit parafin (Nick collision from services.)
21:50:03 Join parafin|away [0] (
21:50:19 Nick parafin|away is now known as parafin (
21:50:32 Quit NoGare ("ChatZilla 0.9.85 [Firefox 3.5.4/20091016092926]")
21:50:59bluebrotherI'm wondering if it would make sense to move all tools that are not required by the build process from tools/ to utils/
21:51:44CIA-85New commit by bluebrother (r23381): Update bootloader download location and add a note about building with VS2005.
21:51:51 Join stripwax [0] (
21:53:44***Saving seen data "./dancer.seen"
21:55:40jasioI've just submitted my first patch. It simply adds support to the VX777 keymap for the only button on the device.
21:56:59linuxstbbluebrother: Yes, it would be nice to clean that up. Although most things seem to be assimilated into rbutil eventually....
21:57:13JdGordon|jasio: morse code input? :D
21:57:27bluebrotheryeah, that tools/ vs utils/ tends to confuse me every now and then.
21:58:00bluebrotheranother way I could think of would be to move it the other way round −− and assimilated stuff goes to rbutil/
21:58:03jasioJdGordon|: that's an idea. One button is still useful as cancel or exit.
21:59:14 Quit bluebrother (Nick collision from services.)
21:59:17 Join bluebroth3r [0] (n=dom@rockbox/developer/bluebrother)
21:59:36 Nick bluebroth3r is now known as bluebrother (n=dom@rockbox/developer/bluebrother)
22:02:21 Join petur [50] (n=petur@rockbox/developer/petur)
22:02:28 Join mitk [0] (
22:02:34 Join robin0800 [0] (
22:02:55 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
22:03:39mitkJdGordon|: ping
22:03:47kugelhm, I just remember that I named it viewport_set_fullscreen
22:04:07JdGordon|mitk: pong
22:04:47mitkJdGordon|: Just FS #10717 reminder
22:08:35JdGordon|kugel: set_fullscreen() shuold return the whole screen... there is a big difference between the whole screen minus 8 pixels height, and the whole screen minus the area left behind from the sbs
22:09:49JdGordon|mitk: yeah, i forgot about it... tnoight maybe
22:09:54JdGordon|or get someone else to do it :)
22:09:55kugelthe point of set_fullscreen is easily get a viewport to draw stuff in, but when you don't want to deactive the statusbar
22:10:13JdGordon|thats why get_defaultS() is
22:10:34kugelis what?
22:11:41JdGordon|thats what get defaults shoudl return
22:11:44CIA-85New commit by dave (r23382): Move mktccboot from tools/ to rbutil/mktccboot/ - inspired by FS #10728 but with no functional changes to code.
22:14:43mitkOk, I'll try. Dear Commiters! Is there anyone who can commit FS #10717. It's insert_last_shuffled functionality to dynamic playlists. Details in FS.
22:16:29jasioI have added a patch for touchscreen targets (FS #10733) that lets software hold lock both buttons and touchscreen. No real code change, just moving a few lines.
22:16:47CIA-85New commit by mcuelenaere (r23383): The Onda VX777 has only one button, add it to the keymap. ...
22:17:02jasioTested on VX777. Don't know if there are many touchscreen targets that don't have a hard hold button
22:17:10mcuelenaerejasio: see SVN ;)
22:17:51jasiothanks :)
22:18:12mcuelenaereoh you weren't talking about FS #10732
22:18:38jasioit's a new one
22:19:05*linuxstb hopes the build system is just on a go-slow...
22:19:35 Join barrywardell [0] (n=barrywar@rockbox/developer/barrywardell)
22:19:35mcuelenaerejasio: RE FS #10733: I don't see that behaviour on my VX747
22:19:55 Join barrywardell_ [0] (n=barrywar@rockbox/developer/barrywardell)
22:20:00mcuelenaerejasio: ah wait, the VX777 doesn't have a hardware lock button?
22:20:15stripwaxmitk - does regular "insert shuffled" insert the files shuffled into the entire playlist; or only insert the selected files in a shuffled order at the current insertion point? wondering how we can make the logic consistent for "insert shuffled last" (or whether we want an equivalent "insert shuffled next")
22:20:17jasiomcuelenaere: no hardware button other than power!
22:20:19*mcuelenaere didn't know Rockbox had software lock button
22:21:16gevaertslinuxstb: I get "svn: Can't connect to host '': Connection refused" on one of my clients. If there are some more, it will be slow...
22:21:18mcuelenaerejasio: also that patch seems to have added a space after #include "lang.h"
22:21:29mcuelenaeregevaerts: I got that too just know, retrying helped
22:21:40rashergevaerts: I got that once as well
22:21:53rasherNearing completion now.
22:21:54gevaertsit's bound to delay some builds
22:22:14jasiomcuelenaere: ok, got to be careful with spaces. i'll have a look
22:22:24JdGordon|stripwax: yep, insert shuffled just puts it somewhere in the playlist... insert shuffled last is so you can add a folder and have the folder shuffled, but not shuffled in with other folders
22:22:28mitkstripwax: Regular insert_shuffled inserts to entire playlist when no music is played. When music is played inesrt_shuffled inserts between currently played track and end of playlist
22:23:20mitkIsnsert_last_shuffled is doing things like JdGordon| sad
22:23:43stripwaxyeah - does it make sense to have an equivalent 'insert_next_shuffled' (equivalent to 'insert_last_shuffled' that is)
22:24:16JdGordon|na.. too many options then
22:24:25gevaertsrasher: is the binsize server running? There seems to be a backlog building up
22:24:41stripwaxreason i'm thinking is that "insert" puts the track after the last insert, but "insert shuffled" doesn't right now (insert_shuffled is really "insert_shuffled_everywhere")
22:25:02rashergevaerts: oh ah, it's not.. and hasn't been running very much lately.. I'll let it run tomorrow
22:25:06stripwaxand i'd quite like an 'insert shuffled' that was like regular 'insert', but .. shuffled ..
22:26:09 Quit Strife89 ("Going home.")
22:26:36mitkThere is no way now to insert shuffled folder to playlist shuffled to the end of playlist. Inesrt_last_ shufled can do that.
22:26:46stripwaxi.e. each of "insert", "insert next", "insert last" have an equiavent shuffled version; then we also have "insert everywhere" which behaves like the current insert shuffled
22:27:01Tornemitk: but there's also no way to do all kinds of other combinations of these things, is the point ehre ;)
22:27:07Torneisn't there also insert vs queue? :)
22:27:09stripwaxThere is no way currently to insert a shuffled folder to playlist shuffled after the current playing track either
22:27:10Tornejust to make it worse?
22:27:14stripwaxright queue vs insert too
22:27:27stripwaxTrying to work out why adding just one of those combinations makes more sense than adding the others too....
22:27:35*stripwax is playing devil's advocate as usual
22:27:43Tornequeue vs insert, next vs last vs as-previous, shuffled vs not
22:28:00Torneand then the current "insert shuffled" as well
22:28:12Tornethat's, what, 13 combinations? :)
22:28:19stripwaxMaybe, we have a config option to let you select which insert vs queue vs shuffled combos appear on that menu?
22:28:23*stripwax is only half joking
22:28:24 Quit barrywardell (Connection timed out)
22:28:39*kugel thinks the insert submenu should be divided, with just insert and queue at the top additionally
22:28:48stripwaxoooh. like.
22:28:59stripwaxInsert -> Shuffled -> Last ?
22:29:04stripwaxtoo many clicks?
22:29:22stripwaxTorne - yep.
22:29:23kugelthat's why insert and queue would be at the top
22:29:23JdGordon|not if we can map the spare button to an playlist insert action
22:29:55stripwaxJdGordon : - ? you'd still need to pick the kind of insert you want, no?
22:30:01kugelbecause those basically repeat your last action
22:30:43JdGordon|stripwax: if you could assign rec to "insert shuffled last" then its one click... or "queue next" thats also one click
22:30:55jasiomcuelenaere: new patch without the space
22:31:00stripwaxJdGordon : - er, which 'rec' button are you referring to?
22:31:10JdGordon|the spare one on e200/h300
22:31:18*stripwax has both an H120 and an ipod 5g .. with no rec button ..
22:31:19 Quit barrywardell_ (Read error: 104 (Connection reset by peer))
22:31:49*JdGordon| also has a 5g with no rec button :D
22:32:16JdGordon|stripwax: actually.. it it were split up there shuold be another option along side insert and queue which is "repeat the last insert method"
22:32:24stripwaxok so context menu could have its own 'repeat last insert/queue action'
22:32:33stripwaxso all other targets could benefit from a one-click operation too.
22:32:45kugelisn't that what insert does?
22:32:53stripwaxdoes insert repeat the last Queue operation?
22:33:07Torneiirc there's only one "last position"
22:33:12stripwaxand will it take into account the shuffled vs not
22:33:22stripwax^will it^would/should/could it
22:33:24Torneso insert inserts at the last position used by anything and queue queues at hte last position used by anything
22:33:29JdGordon|isnt there an open bug which says that the "insert" action makes no sense?
22:33:44stripwaxoh, is there? in that case, probably ignore most of this convo :)
22:33:55mitkThis is why insert_last_shuffled I made
22:34:17 Quit Lss (Read error: 104 (Connection reset by peer))
22:34:18stripwaxwe ought to fix the insert action bug at the same time
22:34:31kugelwhat bug? that it makes no sense?
22:34:58JdGordon|maybe not... cant find it
22:35:02mitkI mean this is not bug. It's strange feature
22:35:07stripwaxmitk - oh hang on - insert_last_shuffled inserts shuffled at the 'last insert point you used' or at the 'last part of the playlist i.e. the end' ?
22:35:09 Join Lss [0] (
22:35:25mitkAt the end of the playlist
22:35:41JdGordon|I think it went something like you could insert (nex maybe) so the playlist order becomes "1ABC23" if you inserted 123 then ABC
22:35:56kugelJdGordon|: I think I fixed that one ages ago
22:35:59*JdGordon| isnt cool enough to use such fancy playlist insert options
22:36:07*kugel now remembers that one too
22:36:26CIA-85New commit by mcuelenaere (r23384): Touchscreen targets: fix software hold not locking touchscreen ...
22:36:28stripwaxyeah, that makes no sense
22:36:46stripwaxso there should also be 'insert before' and 'insert after' to be explicit
22:36:58stripwaxso 123ABC or ABC123 as appropriate
22:37:04JdGordon|how about a "just play the damn track" option?
22:37:05kugelthe problem was that ABC wasn't added after 123
22:37:08stripwaxinstead of just 'Insert'
22:37:18 Quit J-23 (Read error: 104 (Connection reset by peer))
22:37:21kugelthe insertion point was properly updated
22:37:31kugelinsert before/after wasn't the issue
22:37:37stripwaxkugel - understood, the insertion point was updated to one track after the last insert
22:37:55stripwaxrather than after the end of all the tracks inserted
22:37:58kugelI think it wasn't updated at all in some cases
22:38:08mitkBut you are inserting rather something you are intended to listen.Insert last, no insert first
22:38:48stripwaxif I insert ABC then 'insert before' 123 I get 123ABC, all of which is after the current playing track
22:39:02stripwaxif I insert ABC then 'insert after' I get ABC123
22:39:09*JdGordon| would like it if it was 3 options, insert, queue, and shuffle, where for the first two you would get a playlist viewer and you can actually choose where to do the insert
22:39:29stripwaxand separately fix 'Insert' to be consistent, deterministic, and less cryptically named
22:39:35kugelstripwax: that would be a new feature then. in your case doing the same as insert next actually
22:39:47stripwaxkugel - no, not necessarily
22:39:55stripwaxe.g. Insert Last followed by Insert Before, right?
22:40:06kugelstripwax: it's totally deterministic now, isn't it? always after the insertion point?
22:40:13stripwaxChoosing where to insert is a killer feature..
22:40:30stripwaxkugel - I thought you just said it had a bug ... that's what I meant about 'fixing Insert'
22:40:35kugelbtw: r20247
22:40:57kugelstripwax: I also said I fixed that ages ago :) (see the rev I just posted)
22:41:05mitkWhat is insertion point after reshuffling playlist?
22:41:06stripwaxoh ok.
22:41:16stripwaxmitk -eeep
22:41:51stripwaxwhat is insertion point when loading a new playlist from scratch? Possibly, the same logic should apply
22:42:19kugeloh, r20241 too
22:42:30 Join J-23 [0] (
22:42:53n1sJdGordon: i've been thinking that too, this interface we have is pretty good but awfully cofusing to new users
22:42:57kugelit defaults to after current track IIRC, but it's long ago since I messed with that
22:43:14stripwaxkugel - fwiw, that default makes sense to me
22:43:23n1soh and we want multi-select in the pl viewre
22:43:39stripwaxn1s - for multiple-deletes or something else?
22:43:45*stripwax tries to visualise multiple-moves ..
22:44:03n1sstripwax: yeah, multi delete, multi move
22:44:16stripwaxthere's also that 'moving a track while track changes deselects the move' bug that I'd promised to fix .. will take a look at that
22:44:18kugelisn't moving stuff around in the playlist viewer broken enough?
22:44:27stripwaxkugel - no, it is fixed, perfectly
22:44:28*stripwax smiles
22:44:40stripwaxkugel - seriously though, broken how?
22:44:44kugelah right, you did some work on it
22:44:55kugelI can't remember if I last tried it before or after that
22:45:28 Join froggyman [0] (
22:45:39 Nick fxb__ is now known as fxb (n=felixbru@
22:45:40stripwaxn1s - still trying to visualise multi-move .. which track do you see moving around, the first of your multi-selects or the last?
22:45:46 Quit phanboy4 (Read error: 54 (Connection reset by peer))
22:46:05stripwaxI think it could work reasonably well if it's a contiguous block you're moving, but if it's discrete items being moved at the same time .. <brain explodes>
22:46:51gevaertsstripwax: easy. They keep the same number of tracks between them, so if you move e.g. tracks 17 and 19 to before 2, you get tracks 2 and 4, with ex-2 now being 3
22:47:24 Quit jhulst (Remote closed the connection)
22:47:36n1sstripwax: i was thinking of a continous block
22:47:47stripwaxgevaerts - oh no, I realise that. I'm just wondering what you see while you're moving. say you selected tracks 6 and 66, .. what point has the green array, the entry point for track 6, or the entry point for track 66 .. ?
22:47:51*stripwax thinks probably track 6
22:47:59*stripwax thinks contiguous is much easier in any case
22:48:06gevaertsstripwax: weighted average
22:48:22stripwaxoops, meant 'arrow' not 'array'. but yeah.
22:51:15 Quit TheSeven ("ChatZilla 0.9.85 [Firefox 3.5.3/20090824101458]")
22:52:01 Part froggyman
22:52:02stripwaxanyway where were we? I have no objections to an insert_shuffled_last function.
22:52:24mitkMe too :)
22:54:00mcuelenaeredo the MTP people still hang out here regularly? Looks like the last commit in the MTP tree was at 17/10
22:54:09stripwaxanyone else - any objections? if not, I'll commit it ..
22:54:19stripwaxbut we should continue the discussion of rationalising all of this..
22:54:25JdGordon|I onder if it should be called "append shuffled" though
22:54:50stripwaxonly if we rename 'insert last' to 'append' ...
22:54:57mitkAppend is new action!
22:55:19Torneappend doesn't distinguish queue/insert..
22:55:43stripwaxqueue last could be renamed to 'queue append', or somesuch
22:55:55stripwaxblah. no.
22:56:16*Torne doesn't like queue/insert anyway tbh
22:56:41Torne(entirely aside from the fact tht i've never wanted to use queue at all, the names don't really explain the difference)
22:57:02*stripwax agrees (although does use queue..)
22:57:09 Join TheSeven [0] (n=theseven@rockbox/developer/TheSeven)
22:57:35 Join phanboy4 [0] (n=benji@
22:57:51*kugel agrees (although hardly uses any isnert/queue)
22:57:54mitkBut append will be much confusing. For all users because doesn't distinguis insert/queue
22:58:12*gevaerts thinks that they should be merged, with a "play N times" option, with "infinite" being equivalent to "insert" now, "1" being what "queue" does, and the rest being new and improved features
22:58:47kugeland we offer all values from 1 to infinite? :)
22:59:00*kugel also wants NaN then!
22:59:02gevaertsyes :)
23:00:39n1si think those two dirty things are floating point constructions, the max number of playtimes should be INT_MAX, should be enough for anyone really ;)
23:02:12JdGordon|what about quappend?
23:02:20JdGordon|and quinsert?
23:03:36kugeland quincy
23:03:36mitkMaybe rand( queue or append or insert)?
23:03:44n1seh so quinsert is "insert queued"?
23:08:08 Join matsl [0] (
23:10:11 Join saLOUt [0] (
23:10:31 Quit n1s ("Lämnar")
23:16:17 Quit daggett ("Ex-Chat")
23:17:03saLOUthi everyone. i have the recent build on my e240 v2. I get everytime i want to play a video a "data abort" + restart. the same videos worked for some month.. any ideas?
23:19:43JdGordon|did you update the whole zip?
23:20:16saLOUtJdGordon|: i just copied the whole zip over the existing installation
23:20:23stripwaxmitk - so looking at that patch, what should the behaviour be if audio is not currently playing? looks like insert shuffled last will do nothing, is that right?
23:20:40saLOUtdo you think it it better to start from a clean .rockbox ?
23:22:18 Quit domonoky1 (Read error: 104 (Connection reset by peer))
23:22:34Torneit should be fine to just overwrite the existing files
23:22:43Torneas long as you extract the entire zip
23:23:23saLOUtTorne: i used Ark -> extract all here
23:23:45mitkstripwax: Yep. When audio is not curently playing rockbox creates new playlist from scratch
23:24:52mitkstripwax: Then insert_last_shuffled doesn't make sense. Just insert_shuffled
23:25:05mcuelenaerehmm I got make working but I'm getting '/home/mcuelenaere/rockbox_git/rockbox/apps/plugins/plugins.make:46: target `/home/mcuelenaere/rockbox_git/build/apps/plugins/boomshine.lua' doesn't match the target pattern'
23:25:14stripwaxmitk - fair enough
23:26:29kugelmcuelenaere: it doesn't need to be preprocessed/compiled right? you could simply copy it in make install/zip/7z/etc
23:26:48mcuelenaerekugel: well it needs #ifdef TOUCHSCREEN #endif in SOURCES
23:27:13kugeloh, well
23:27:33mcuelenaereis make install/zip/7z/etc handled by one script?
23:27:40CIA-85New commit by stripwax (r23385): Committing FS #10717 by Tomasz Kowalyczyk: add PLAYLIST_INSERT_LAST_SHUFFLED (appends the inserted folder into shuffled order to end of playlist)
23:27:59 Nick mc2739_ is now known as mc2739 (n=mc2739@rockbox/developer/mc2739)
23:28:51mitkstripwax: Thanks. My life will be much easier. I'm using it hardly :)
23:29:36 Quit parafin ("So long and thanks for all the fish")
23:29:39stripwaxhardly? as in 'using it very hard' or 'barely using it' ?
23:29:57 Join parafin [0] (
23:31:10mitkI mean I'm using it very often so life without it was hard:)
23:31:43 Join TopyMobile__ [0] (
23:32:09stripwaxmitk :)
23:32:25 Join Topy [0] (
23:33:07saLOUtcan you give me the link to the newest bootloader for e200 v2 ?
23:33:08stripwaxJdGordon: , kugel etc, shall we continue that discussion on ml ?
23:33:15 Quit TopyMobile_ (Read error: 60 (Operation timed out))
23:33:47JdGordon|stripwax: hrm?
23:33:49mitkAll green. Going to bed. cya
23:34:18 Part toffe82
23:34:23stripwaxJdGordon : - you know, inqueue versus quinsert ..
23:34:34 Quit mitk ("CGI:IRC")
23:34:45JdGordon|oh please no :)
23:35:18kugelmaybe tomorrow, I'm soon getting bloody fingers
23:35:21stripwaxwell, at least the serious parts .. :) submenus rather than one giant list of all permutations ; playlist browser for insert points ; multiple select etc...
23:35:32stripwaxkugel - i'll start a thread anyway ..
23:35:57saLOUti solved the problem with "data abot on ........." by delete and copy .rockbox
23:36:01JdGordon|oh, sure.. serious parts are fine
23:36:07JdGordon|but lets finish the current fight first!
23:36:23 Quit matsl (Read error: 110 (Connection timed out))
23:37:09*kugel would like a playlist viewer item somewhere in the quinsert forest
23:38:25kugeland possibly a preview where after which song the first song of the current selection would be inserted (unless you insert shufflled that is)
23:38:47kugelbut otoh I'm hardly using this insert features anyway ;)
23:42:58 Join Hillshum [0] (
23:43:02 Quit evilnick_B ("Page closed")
23:45:11stripwaxkugel - there's already a playlist viewer there, in the context menu. or do you mean 'to view a different playlist than the current one' ?
23:45:43stripwaxkugel - oh you mean the thing where you can browse to the insertion point. yep. agreed. i want that too.
23:46:34kugelI mean "View current playlist" next to insert,queue
23:47:52stripwaxthat's already there, no?
23:48:24CIA-85New commit by mcuelenaere (r23386): Add Lua support to plugins Makefile
23:48:25kugelnot here
23:48:27CIA-85New commit by mcuelenaere (r23387): Add new Lua touchscreen-only game: Boomshine ...
23:49:10 Quit T44 (Connection timed out)
23:49:36kugelbertrik: any idea why your latest patch causes causes blue pixels and shifts?
23:49:36 Join uflops [0] (
23:49:50kugelI mean, it's basically what I expected to happen, but you said it shouldn't happen in theory
23:50:48bertrikI don't know, we wait for FIFO empty before switching from write to read, but maybe there is still a write in progress when it indicates FIFO-empty
23:51:23kugelwell, the obvious thing is that we're writing to DBOP_DOUT possibly during and lcd update
23:51:55kugelI suspected it would confuse the controller, but you said it should be fine
23:51:57 Quit ender` (" All general statements are false. -- The Ultimate Law")
23:53:08stripwaxkugel - hm, indeed. might as well merge a 'view playlist' option with an 'insert dynamically somewhere into the playlist' option, since you can always invoke the latter, view the playlist, and cancel
23:53:47***Saving seen data "./dancer.seen"
23:54:30 Join froggyman [0] (
23:54:57kugelbertrik: maybe if we don't set enw?
23:55:22 Part froggyman
23:55:25 Join thegeek [0] (
23:57:34bertrikI don't know, I won't look into it tonight anymore

Previous day | Next day