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 2010-12-22

00:00:30 Join JdGord [0] (~jonno@
00:03:24 Quit wodz (Quit: Leaving)
00:10:14 Quit freddyb (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100723085406])
00:11:52 Quit kugel (Remote host closed the connection)
00:11:59 Quit pamaury (Remote host closed the connection)
00:15:56***Saving seen data "./dancer.seen"
00:19:38 Quit sideral (Quit: Leaving.)
00:20:07 Join earcar [0] (~carmine@
00:21:58 Quit bertrik (Ping timeout: 245 seconds)
00:23:52 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
00:25:44 Join Rasi [0] (
00:25:44Rasidoes rockbox support reading files from different filesystems?
00:25:56Rasii would like to format a minisd for my sansa with something other than vfat
00:34:42 Quit earcar (Quit: zzzzzz)
00:39:56 Join Keripo [0] (
00:45:05 Quit sasquatch (Quit: WeeChat 0.3.2)
00:45:29 Join sasquatch [0] (
00:50:26 Quit Stephencc__ (Quit: Leaving)
00:53:09 Quit dfkt (Quit: -= SysReset 2.53=- Sic gorgiamus allos subjectatos nunc.)
00:57:22 Join timccc [0] (~timccc@
00:59:02 Join fdinel [0] (
01:01:32 Quit timccc (Ping timeout: 240 seconds)
01:08:11 Join timccc [0] (~timccc@
01:11:07 Quit JdGord (Quit: Bye)
01:15:25rasherRasi: explains a few of the reasons why.
01:17:56 Quit Dreamxtreme (Ping timeout: 260 seconds)
01:23:06 Quit kadoban (Ping timeout: 260 seconds)
01:29:00 Join Dreamxtreme [0] (~Dre@
01:43:05saratogascorche: please ban ab76deje
01:43:36saratogaalso fwiw that spammer somehow has a post count of 9 with only one post
01:51:08 Nick soap_ is now known as soap (~soap@rockbox/staff/soap)
01:51:38 Join telliott [0] (
01:52:04soapBuschel, yes (I'm sure it varies), but if 80 is unstable and 80 is /supposedly/ Apple OS's speed why don't I get instability in Apple OS?
01:55:33saratogatheres a lot of different ways you can setup the clocks on a chip
01:55:51Lloreansaratoga: If the script that automatically hides spam has hidden some, you'll see them as having a post count that's different from the number of posts you may actually see, it seems
01:55:52saratogajust because both run the CPU at 80 doesn't mean all the other clocks are the same, nor that the clock is generated in the same way
01:55:59saratogaah ok
01:56:35Lloreanat least, that's how it presented to me. I'm not sure if it's the same for privileged users.
02:01:26 Quit Keripo (Read error: Connection reset by peer)
02:02:51 Quit Kitar|st (Ping timeout: 240 seconds)
02:03:13 Join Keripo [0] (
02:05:17 Join CaptainKewl [0] (
02:09:39 Join Kitar|st [0] (
02:15:57***Saving seen data "./dancer.seen"
02:16:34soapsaratoga, I understand that, but I'm too ignorant at this point to know where to start testing.
02:24:39 Quit antil33t (Ping timeout: 255 seconds)
02:26:53 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
02:34:52 Quit telliott (Ping timeout: 240 seconds)
02:36:34 Quit mc2739 (Quit: leaving)
02:38:31 Join mc2739 [0] (~mc2739@
02:38:33 Quit mc2739 (Changing host)
02:38:33 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
02:38:34 Quit [Saint] (Quit: I'm only going to Heaven if it feels like Hell, I'm only going to Heaven if it tastes like caramel...)
02:40:00 Join mc2739_ [0] (~mc2739@rockbox/developer/mc2739)
02:45:22 Quit Dreamxtreme (Quit: There are two major products that come out of Berkeley: LSD and UNIX. We don't believe this to be a coincidence.)
02:47:31 Quit DerPapst (Quit: Leaving.)
03:46:02 Quit scorche (Read error: Connection reset by peer)
03:51:20 Join evilnick_ [0] (
03:54:45 Quit evilnick (Ping timeout: 264 seconds)
04:07:14 Quit TheSeven (Ping timeout: 260 seconds)
04:09:58 Join Dreamxtreme [0] (~Dre@
04:12:56 Join Topy [0] (
04:13:05 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:15:58***Saving seen data "./dancer.seen"
04:16:34 Quit T44 (Ping timeout: 240 seconds)
04:43:23 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:43:29 Quit amiconn (Disconnected by services)
04:43:40 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:44:03 Quit pixelma (Disconnected by services)
04:44:05 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:44:07 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:53:05 Quit liar (Ping timeout: 255 seconds)
04:54:52 Quit xavieran (Ping timeout: 260 seconds)
05:04:04 Join Horscht [0] (~Horschti@xbmc/user/horscht)
05:04:36 Quit fdinel (Quit: Miranda IM! Smaller, Faster, Easier.
05:07:20 Join xavieran [0] (
05:30:37 Quit Horscht (Quit: Verlassend)
05:33:42 Join scorche [0] (~scorche@rockbox/administrator/scorche)
05:43:56 Join LambdaCalculus37 [0] (~rmenes@rockbox/staff/LambdaCalculus37)
05:44:54 Quit LambdaCalculus37 (Client Quit)
05:48:53 Join rvvs89 [0] (rvvs89@pdpc/supporter/base/rvvs89)
06:16:03***Saving seen data "./dancer.seen"
07:05:30 Quit kkit|sh (Ping timeout: 250 seconds)
07:22:35 Join Feisar [0] (
07:23:01 Nick Feisar is now known as Guest93224 (
07:36:20 Quit saratoga (Ping timeout: 265 seconds)
07:41:04 Join kadoban [0] (
07:56:23 Join kugel [0] (~kugel@
07:56:24 Quit kugel (Changing host)
07:56:24 Join kugel [0] (~kugel@rockbox/developer/kugel)
08:10:49 Quit BHSPitMonkey (Remote host closed the connection)
08:11:44 Join Buschel [0] (
08:15:48 Quit JdGordon| (Quit: leaving)
08:16:05***Saving seen data "./dancer.seen"
08:18:06 Join MaxHR [0] (
08:18:39 Join JdGord [0] (~jonno@
08:24:45MaxHRHello, am getting "playlist control file is invalid" when trying to resume playback, if I try to create a new playlist, player will lock up and require 2-finger salute, any ideas on how to fix this? am considering deleting the nvram.bin, as one forum post suggests unless there are other ideas
08:25:43 Join einhirn [0] (
08:26:08crwlregardless of your player, have you tried running chkdsk/fsck to the device?
08:26:19crwlie. checking the drive for errors
08:26:28Buschelsoap: as you know the PP is undocumented and it is not known what _really_ needs to be set as specified. e.g. there is some special code regarding IDE0_CFG for the nano 1G within system-pp502x. I remember this also solved some boosted clock issues (r15705)
08:27:42 Quit TheSeven (Ping timeout: 260 seconds)
08:27:53Buschelsoap: what in detail are the issues you have? freeze? noise?
08:28:44MaxHRcrwl: forgot to mention, it is a sansa e200, I have not tried running scandisk on it (anything else to try?) also, I can browse around the database fine but file view mode locks up player also
08:30:36crwlMaxHR, you should try running scandisk/chkdsk/whatever it's called these days in windows
08:31:25 Join kugel2 [0] (~kugel@
08:31:26 Quit kugel (Disconnected by services)
08:31:26 Quit kugel2 (Changing host)
08:31:26 Join kugel2 [0] (~kugel@rockbox/developer/kugel)
08:32:32 Join saratoga [0] (62e16704@gateway/web/freenode/ip.
08:33:10 Quit JesusFreak316 (Ping timeout: 240 seconds)
08:36:31MaxHRcrwl: I get "windows was unable to complete the disk check"
08:36:51 Quit JdGord (Quit: Bye)
08:39:28saratogai think that usually means you need to schedule it to run when your computer reboots or something like that
08:39:59 Join stripwax [0] (
08:41:08Buschelsoap: btw, do you know if if someone checked Apple OF for the nano clock settings? the problems which lead to the IDE0_CFG change started when raising the clock from 7x MHz to 80 MHz as well...
08:42:09*Buschel takes off for his last working day this year :)
08:42:15 Quit Buschel (Quit: ChatZilla 0.9.86 [Firefox 3.6.13/20101203075014])
08:42:33crwlMaxHR, strange, I don't know what that is about...
08:44:56 Quit ReimuHakurei_ (Quit: BRB FBI)
08:45:00 Quit stripwax (Quit: Miranda IM! Smaller, Faster, Easier.
08:46:05MaxHRIt says beginning phase 1 of scanning, no progress, then quits, then I try to browse player and it now appears empty, cant get it back w/o disconnecting, will try to delete nvram unless other ideas abound
08:48:01MaxHRhmmm not allowed to delete it :P
08:48:45 Join Judas_PhD [0] (
08:48:48 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
08:51:40MaxHRnow it deleted...
08:52:59MaxHRstill freezing
08:53:45MaxHRany other clues?
08:55:01MaxHRdo I need to be in a special mode to do the scandisk?
08:59:36MaxHRIs there a last resort solution to this?
08:59:38linuxstbMaxHR: No, just UMS mode. You could also try just reformatting it - easiest way is to use the option in the original Sandisk firmware to do that.
09:01:57MaxHRwhich option in settings? reset all?
09:03:42MaxHRlinuxstb: don't see a format option in sansa menu
09:03:53MaxHRyou mean use the sansa loader?
09:04:22linuxstbMaxHR: I don't know - I don't own an e200. But I've heard people say there's a format option somewhere there...
09:05:08MaxHRdoes UMS mode translate to MSC or MTP
09:06:30 Quit kadoban (Ping timeout: 245 seconds)
09:08:39 Join JdGordon [0] (7a6edfaa@gateway/web/freenode/ip.
09:08:52pixelmaon the c200v1 it's the last option in the settings menu, right below the USB mode one
09:08:53 Join marines [0] (
09:09:14MaxHRits not remembering my settings, I had it set originally to resume playlist for start screen, then just changed to show main menu, powered down, and then on startup it says nothing to resume
09:09:57MaxHRpixelma: mine doesn't have it there on the sansa menu
09:09:58pixelmaI mean "format" in the original firmware
09:10:20MaxHRthe original sansa menu has no option for format
09:10:29 Quit kugel2 (Ping timeout: 240 seconds)
09:11:13 Part marines
09:11:34 Join marines [0] (
09:12:23 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
09:13:14 Join kugel [0] (~kugel@
09:13:15 Quit kugel (Changing host)
09:13:15 Join kugel [0] (~kugel@rockbox/developer/kugel)
09:14:48 Join liar [0] (
09:16:25 Quit MaxHR (Quit: ChatZilla 0.9.86 [Firefox 3.6.13/20101203075014])
09:20:24 Quit kugel (Ping timeout: 265 seconds)
09:20:58 Part marines
09:21:05 Join marines [0] (
09:21:06 Part marines
09:23:35 Join marines [0] (
09:27:05 Quit bertrik (Ping timeout: 265 seconds)
09:28:45 Quit Judas_PhD (Quit: This is a quitting message)
09:32:57 Quit swilde (Remote host closed the connection)
09:34:03 Join insp_ [0] (~chatzilla@
09:34:06 Join swilde [0] (
09:34:41 Join swilde` [0] (
09:40:48 Join ender` [0] (
09:41:20 Join wodz [0] (
09:44:37 Join petur [0] (d408b802@rockbox/developer/petur)
09:45:07wodzhmm recording tags are not documented in manual, and CustomWPS is out of sync either
09:45:20wodzJdGordon: ping
09:46:12 Part swilde` ("ERC Version 5.3 (IRC client for Emacs)")
09:46:37wodzcould you expain me the line 49 of apps/radio/radio_skin.c please?
09:46:40 Join earcar [0] (~carmine@
09:48:21wodzJdGordon: could you also explain me how can I use %tr to display some graphical indicator of signal strength?
09:48:54JdGordonweb svn viewer doesnt do line numbers... paste the line?
09:49:08JdGordonpixelma might be able to help you while im eating dinner :)
09:49:12wodz%?Rr<%Sx(Time:) %Rh:%Rn:%Rs|%?St(prerecording time)<%pm|%Sx(Prerecord Time) %Rs>>
09:49:18 Quit chattr (Read error: Operation timed out)
09:49:41JdGordonwhat about it in particular?
09:49:51JdGordonCustomWPS on the wiki is prob a good start
09:49:55wodzI can't find documentation for tags used
09:51:17wodzCustomWPS states recording tags can be used in sbs only and it list only %Rp, %Rm, %Rf, %Re and %Rb
09:51:41JdGordongrr... needs updating
09:51:43JdGordonback in 30... must eat
09:54:08 Join kugel [0] (~kugel@
09:54:08 Quit kugel (Changing host)
09:54:08 Join kugel [0] (~kugel@rockbox/developer/kugel)
09:57:36 Quit earcar (Quit: bye)
10:00:48 Join kevku [0] (~kevku@2001:7d0:0:f000::135d)
10:07:17 Quit TheSeven (Ping timeout: 264 seconds)
10:09:37 Join n1s [0] (
10:09:38 Quit n1s (Changing host)
10:09:38 Join n1s [0] (~n1s@rockbox/developer/n1s)
10:12:44 Join DerPapst [0] (
10:16:06***Saving seen data "./dancer.seen"
10:18:23 Join insp__ [0] (~chatzilla@
10:21:09 Quit insp_ (Ping timeout: 255 seconds)
10:22:57 Quit insp__ (Ping timeout: 255 seconds)
10:33:20 Join pamaury [0] (
10:33:20 Quit pamaury (Changing host)
10:33:20 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
10:33:40JdGordonwodz: %Rr = does the target HAVE_RECORDING, %Rh,n,s = recording hours, minutes, seconds. %Rs = prerecording seconds
10:33:41JdGordoni think
10:33:51 Quit sasquatch (Quit: translocating on icy roads)
10:34:09wodzthat what I suspected
10:34:17JdGordonto make %tr a graphical bar do exactly like any of the other bars (%pb or %pv for example)
10:34:35JdGordonit depends how you want it to look determines which method to use
10:34:42 Quit DerPapst (Read error: Connection reset by peer)
10:34:45wodzok I'll try
10:34:45JdGordonthere are 3 ways you could do it
10:35:05wodzI want simple 4 -5 level indicator
10:35:23 Join DerPapst [0] (
10:35:53JdGordonthis is a discussion we had a fair bit when it was implemented.... RSSI isnt linear so 4-bars is very hard to do in a meaningful way
10:36:29wodzso maybe small bar?
10:38:16JdGordonyour best bet is probably figure out the raw values where you want the 4 bars to start then use %?if<%tr, <, ONEBAR>(%xd()| %?if<%tr, <, TWOBARS>(%xd()|%?if<%tr, <, THREEBARS>(%xd(3bars)|%xd(4bars)>>>>
10:39:10JdGordonsaint and pixelma would probably suggest %?tr<%xdaA|%xdaA|%xdaA|%xdaA|%xdaA|%xdaB|%xdaB|%xdaB|%xdaB|%xdaB|%xdaB|%xdaB|%xdaB|%xdaB|%xdaB|%xdaC|%xdaC|%xdaC|%xdaC|%xdaC|%xdaC|%xdaC|%xdaC|%xdaC|%xdaC|%xdaD|%xdaD|%xdaD|%xdaD|%xdaD|%xdaD|%xdaD|%xdaD|%xdaD|%xdaD|%xdaD> though
10:39:15JdGordonscaled as required
10:39:54JdGordonput some # line splitters in to make them eaisser to read
10:41:06JdGordonONEBAR, TWOBARS, THREEBARS need to be replaced with the raw rssi values
10:41:20 Quit wodz (Remote host closed the connection)
10:41:29JdGordonor watcher value you get from %tr
10:43:12 Join LinusN [0] (
10:43:12 Quit LinusN (Changing host)
10:43:12 Join LinusN [0] (~linus@rockbox/developer/LinusN)
10:45:52 Join wodz [0] (
10:46:45 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
10:48:16wodzshit, gimp segfaulted burying down my work :-/
11:07:14 Quit jfc^2 (Read error: Connection reset by peer)
11:07:38 Join jfc^2 [0] (
11:10:28 Quit xavieran (Ping timeout: 260 seconds)
11:19:25 Quit Guest93224 (Ping timeout: 245 seconds)
11:24:25 Join xavieran [0] (
11:29:47 Join TheLemonMan [0] (
11:32:57 Quit kugel (Remote host closed the connection)
11:38:03 Join insp_ [0] (~chatzilla@
11:38:23wodzJdGordon: how do I get next to the current fm presset? (the same as for Next track in WPS)
11:38:54JdGordonyou need to use the playlist viewer viewport
11:39:29wodz%Fn, %It are undocumented
11:40:55wodzah no they are but in rather secret way
11:41:24pixelmathe next track info ones are definitely documented
11:41:29 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
11:42:44wodzyes but they are not explicitly written so search doesn't find them
11:43:40wodzJdGordon: could you give me some hint about playlist viewer viewport?
11:44:34JdGordonsomething like %Vp(.... regular viewport stuff)\n %<preset number tag> %<preset name tag>
11:44:45JdGordonit will do that one line for as many lines as will fit in that viewport
11:46:55wodz%<preset number tag> is %Ti for current, and how to get next?
11:46:58 Join dfkt [0] (dfkt@unaffiliated/dfkt)
11:48:05wodzahm I think I understand
11:48:45JdGordonthe playlist viewer rejigs the values so that is the correct tag to use
11:49:22 Part marines
11:49:31 Join marines [0] (
11:51:12JdGordonthe %Vp() above doesnt sound right
11:52:14JdGordon%Vp(1, %pp - %?it<%it|%fn>) is the CustomWPS example
11:52:17JdGordonthat line needs to go in a viewpoty
11:52:23JdGordonthe 1 is the offset to start counting from
11:52:34JdGordonthe next block is the code to show on the line
11:55:50wodzok so if I want to print current and next station I need something like %Vp(0, %pp)\n%Vp(1,%pp) ?
11:56:54JdGordonno. that line is repeated on each line in the viewport
11:56:59JdGordonso the first one is all you need
11:58:44wodzBut if I wan't to have first line formatted differently than second I have to setup 2 viewports
12:00:52wodzthings start to be complicated :-)
12:01:58wodzso the playlist viewport has to be settled inside other viewport?
12:02:34JdGordonits a playlist viewer not viewport :)
12:02:34JdGordonmy bad :p
12:05:10soapBuschel, the issue I have is audio glitching for a few minutes before the eventual hard freeze. Audio glitching which sounds like two songs mixed together.
12:05:43soapNormally roughly 30 minutes into my use of the player.
12:14:05wodzJdGordon: %pn will be station name right? What if station has no name (from autoscan for example). Currently we display something like this %s%?Ti<Station: |>%?Tn<%Tn|%Tf>
12:15:13JdGordon%Tn is preset name
12:15:24JdGordon%pn is playlist name which isnt used for the radio
12:15:47JdGordonput that line in %pv() and it should do what you want
12:16:07***Saving seen data "./dancer.seen"
12:16:21pixelma%Tf is current frequency IIRC so that line means "use current frequency" if station name isn't known
12:16:42wodz%pv ? Current volume in decibels (dB)?
12:16:47pixelma%pv is current volume
12:19:54*JdGordon is dyslexic and lazy
12:20:13wodzok I somehow understand
12:20:15CIA-7New commit by jethead71 (r28875): MPEGPlayer: Some UI tweaking and bugfixing ...
12:20:23 Join The_Pwny [0] (
12:22:11CIA-7r28875 build result: All green
12:33:18preglowany known way to exit the android port?
12:33:25preglowsince i got 2.2, i can't kill it in any way
12:38:21 Join moos [0] (moos@rockbox/staff/moos)
12:39:10JdGordonpreglow: you dont really need to.... or get a task killer
12:39:19 Quit insp_ (Ping timeout: 255 seconds)
12:39:25JdGordonor settings > applications > running applications > rockbox > kill task
12:44:18preglowgot a task killer, but it seems to have turned a bit after i got 2.2
12:45:21preglowand btw, playback also got very flakey after, like it just stops from time to time and i need to seek around a bit to start it again. known issue?
12:45:22 Join chattr [0] (
12:53:50JdGordonwhat phone?
12:54:20JdGordoni went 4 hours or so yesterday with no issues
13:05:55dfktis there a current wiki page that explains how rockbox gapless playback works (and which codecs specifically are possible to be gapless)? the GaplessHowTo page is rather outdated, methinks
13:06:31dfkti.e. if FhG or xing can be gapless too, or if it only works with LAME, as far as MP3 is concerned
13:13:57wodzJdGordon: please have a look at FS #11824
13:14:01 Quit Keripo (Ping timeout: 245 seconds)
13:14:25JdGordonand are you planning on doing it for other displays? :)
13:15:21 Join Feisar [0] (
13:15:47 Nick Feisar is now known as Guest17322 (
13:15:59wodzI am not that crazy :P
13:16:12JdGordonbugger! :p
13:16:48JdGordonok, I cant actually test it, but if its works then go for it... BUT!
13:17:05JdGordonPLEASE PLEASE change the viewport labels to something more meaningful, and ditto the image labels
13:17:12 Quit factor (Ping timeout: 255 seconds)
13:17:56wodzyou mean %Vl labels?
13:18:11JdGordon%?mh<%xd(Aa)|%xd(Ab)> is messier than simply %xd(lock, %mh)
13:18:33wodzit is taken from cabbie wps file
13:19:53wodzI can write something working in reasonably designed language but that doesn't mean I am aware of all quirks (especially if documentation is out of sync)
13:19:56JdGordon%?bp<%?bc<%xd(Ba)|%xd(Bb)>|%?bl<|%xd(Bc)|%xd(Bd)|%xd(Be)|%xd(Bf)|%xd(Bg)|%xd(Bh)|%xd(Bi)|%xd(Bj)>> −−−−>>> %?bp<%xd(Battery, %bc)|%xd(Battery, %bl, 3)
13:20:24JdGordonnot 100% sure about the last bit of that 2nd line, but that is so much simpler
13:21:18JdGordonthe skin syntax is not a "reasonbly desinged language" :p
13:22:17JdGordon%?bp<%xd(Battery, %bc)|%xd(Battery, %bl, 3, 1)>
13:23:01wodzAnyway - I am not going to commit this. If you/someone else with deeper skin syntax knowledge find this useful that's fine. Fix/rewrite as you like
13:23:03JdGordonditto the volume line
13:23:35JdGordonno, you should commit it... its a good base
13:23:48JdGordonits just I really hate the cabbie code and would like it cleaner in svn
13:25:33JdGordonalso the viewports where all lines are checked (like the last two) should be rejigged so they are conditionally enabled
13:29:22 Quit The_Pwny (Quit: When the chips are down, well, the buffalo is empty)
13:33:10 Part LinusN
13:37:24 Quit JdGordon (Ping timeout: 265 seconds)
13:43:09 Quit bluebroth3r (Read error: Operation timed out)
13:46:55 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
13:53:29 Quit TheSeven (Ping timeout: 276 seconds)
13:59:12 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
14:01:24preglowJdGordon: htc legend
14:01:57preglowdfkt: it works on mp3 files where gapless encoding is supported in the encoder
14:02:09preglowit does not play gappy files gaplessly
14:02:20preglowthat requires strange tricks which never work in all cases
14:03:05preglowi'm not really sure if we support the itunes way of gapless playback
14:03:09preglowbut i guess we should
14:04:00 Quit wodz (Quit: Leaving)
14:09:13 Join insp_ [0] (~chatzilla@
14:09:15dfktthanks, preglow
14:14:06 Join kugel [0] (~kugel@
14:14:07 Quit kugel (Changing host)
14:14:07 Join kugel [0] (~kugel@rockbox/developer/kugel)
14:14:55 Join factor [0] (
14:15:40kugelpreglow: no such problems on my legend
14:16:10***Saving seen data "./dancer.seen"
14:16:13 Join WonTu [0] (
14:16:27 Part WonTu
14:17:49pixelmaAFAIK itunes gapless tags are at least supported for AAC
14:18:06pixelmain Rockbox of course
14:23:05n1skugel: does this patch to parse_testcodec.rb look ok ? (it adds the filenames of new tracks in the testset and fixes the printing of MHz needed for realtime)
14:23:15n1si've never touched ruby before...
14:25:20kugelyea, if it works as expected
14:26:27n1sit does
14:26:46n1si'll commit it then
14:27:59CIA-7New commit by nls (r28876): parse_testcodec.rb: add new 'pegase' filenames so they get a heading and fix printing of the MHz needed for realtime values.
14:29:45CIA-7r28876 build result: All green
14:35:47amiconnsoap: Either it's a thermal issue, due to us not disabling all unneeded parts of the SoC, or perhaps just due to having less efficiency than the OF. What codecs are you observing this with?
14:37:26amiconnIt could also be that our clock setup isn't optimal. For PP5020 we deduced from various OFs that the PLL frequency should be within a certain window, but this window *seems* to be larger for PP5022/5024
14:38:49amiconnThe Nano and Video SoCs have 'PP5021' printed on them, but report back as PP5022. You could try a clock setup more similar to the PP5020 way
14:44:04 Join komputes [0] (~komputes@ubuntu/member/komputes)
14:48:53 Join sasquatch [0] (~username@
15:11:01 Join casainho [0] (
15:17:00 Join evilnick_B [0] (0c140464@rockbox/staff/evilnick)
15:21:53n1sjhMikeS: ping
15:27:58jhMikeSn1s: yea?
15:28:51n1sjhMikeS: since you know about the thread stuff i wanted to ask if you can think of a clean way to work around a problem when building for coldfire with newer gcc
15:29:23jhMikeSonly if I know the problem :) even then, no guarantees
15:30:10n1sin switch_thread if profiling is enabled the saved stack pointer is wrong since the arg for profile_thread_stopped is pushed onto the stack but the stack pointer is not restored untill after the inlined store_context
15:30:49n1sit sort of works by luck right now too
15:31:18n1sand i have no desire to rewrite the entire function in asm just for this and profiling is a nifty feature
15:32:18n1si managed to hack around it by saving the stack pointer in a register before calling profile_thread_stopped
15:33:09jhMikeShmmm, those things can be a headache
15:37:09jhMikeScall and restore in some sort of asm block maybe?
15:38:50n1ssounds like it could work, i just have to figure out how to call c functions from asm :)
15:40:10jhMikeSnot hard, but I forget if there's anything weird for cf. I guess just look at what the compiler does. :)
15:40:35 Quit sasquatch (Ping timeout: 276 seconds)
15:41:31 Join sasquatch [0] (~username@
15:45:11 Join kkit|sh [0] (
15:45:38n1sjhMikeS: i'll try that, thanks
15:48:59 Quit kugel (Ping timeout: 250 seconds)
15:51:09 Join kugel [0] (~kugel@
15:51:10 Quit kugel (Changing host)
15:51:10 Join kugel [0] (~kugel@rockbox/developer/kugel)
15:51:31 Quit GeekShadow (Read error: Operation timed out)
15:55:03 Quit 5EXABW64X (Ping timeout: 276 seconds)
15:57:36 Join fred_2 [0] (
16:05:49 Quit sasquatch (Ping timeout: 240 seconds)
16:07:21soapamiconn, I have only tested with mp3
16:08:34soapand unless the nano has really weird thermal dissipation issues I don't see how it could be directly thermal issues. Let me explain as that seems counter to me saying it happens less when cold:
16:09:35soapI was jogging in -10C weather the other day with it on my shoulder (exposed to the elements) with stock Rockbox and it happened within 20 minutes.
16:09:43 Quit mortalscan (Remote host closed the connection)
16:09:58soap(outside my coat)
16:09:59 Join mortalscan [0] (~mortalsca@
16:11:34soapThe next day it was against bare skin while running on a treadmill in a warm room, ran well over an hour w/o issue using my 76Mhz build ((19/3)*24)/2
16:12:49 Quit {phoenix} (Read error: Operation timed out)
16:13:53soapAnd with the OF I can run it indefinitely on the heater vent of my dash w/o issue (except I'm probably cooking the battery testing like that).
16:16:12***Saving seen data "./dancer.seen"
16:16:54Lloreansoap: The old 80mhz issue also used to be agitated by heat.
16:17:46soapyes, I can make it crash later by storing it in the freezer before use.
16:18:17soapLet me take that back. I used to be able to delay crashing by storing it in the freezer. I have not directly tested that since the original issue.
16:19:16 Quit kugel (Quit: Yaaic - Yet another Android IRC client -
16:23:50 Join kugel [0] (~kugel@rockbox/developer/kugel)
16:25:24 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
16:31:41 Join sasquatch [0] (~username@
16:36:06 Join toffe82 [0] (
16:37:14moosHello all. I want to add filesize information to the tag viewer. The needed changes are trivial, but it will add few bits because of the added english.lang string.
16:37:49mooshere is the patch:
16:38:00 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
16:38:18moosit could be easily swcodec specific if needed...
16:40:09 Quit insp_ (Ping timeout: 255 seconds)
16:40:47pixelmawhy should it only be swcodec, and - why in the tag viewer? The "properties" show filesize already
16:43:11moospixelma: hi, the "if needed" part was in reference to binsize added
16:44:59moosyes properties show this info indeed, but in the file browser. Here that's for track infos in WPS, and imho filsize could be a usufull info to have here, without having to go back to file, bring the context menu for properties
16:45:00pixelmait would introduce a seemingly arbitrary difference and the hwcodec targets aren't even the ones with the smallest usable RAM
16:45:37moosand in the properties screen I will have "again" info about file tag
16:45:54moospixelma: remove the swcodec part then :)
16:46:58pixelmaisn't there already a "File size" string or is the thing in "properties" hardcoded?
16:47:12LloreanWouldn't it be simpler just to also add "properties" to the context menu?
16:47:26LloreanEr, to the WPS context menu, I mean
16:48:07moospixelma: yup there is, I didn't want to add complexity
16:48:11pixelmaand the properties also uses the nice byte/kB/MB adaptation
16:48:21moosLlorean: indeed it could
16:48:32moospixelma, yes I saw that
16:49:14LloreanI don't know what's in the tag viewer, but I wouldn't go into it looking for filesize in the first place, so I'd rather have an option that people are likely to "guess" would have the information.
16:49:27LloreanI would expect it only to have information from the tags.
16:49:35moosI surely could reuse this here too, but I just wanted to make easy, since we already have id3->filesize info
16:49:41pixelmaso why a new lang string then? I'd expect reusing the existing to reduce complexity
16:49:44moosjust reused it
16:50:36moospixelma: because all items are translated here
16:50:49moosLlorean: justa look at! :)
16:50:59n1spixelma: properties is a plugin
16:51:17moosthere are infos about the current playing track, and imho filesize fit well here
16:51:24pixelmayes, but then moos' answer was misleading
16:52:11moosindeed properties is an exception plugin called from core
16:52:43LloreanIf it's not just metadata, we should probably change the name from "tag viewer" to something else.
16:52:58moostrack informations?
16:53:18pixelmastill, the kB/MB etc. adaptation is a core feature AFAIK and used in "Rockbox Info", isn't it easier to use this instead of hardcoding /1024?
16:53:25LloreanI'd say "track info" or "track information" is good, yeah.
16:53:49n1spixelma: it is a core feature, yes and agree
16:53:54n1si agree, even
16:54:12moosI could change the menu name too in the mean time
16:54:36n1si also agree with llorean that the two ways to show info about tracks shuld have the same info instead of a partial overlap
16:54:42moospixelma, n1s: yup indeed, but again I didn't want to add code here, just reuse infos :)
16:55:03pixelma-s ;)
16:55:39Lloreann1s: Either the same info, or explicitly different. "Properties" could be about the file, and "Tag viewer" could just contain things in metadata.
16:55:41n1sit's just one call to output_dyn_value
16:55:54moosn1s: properties show id3 tags too... for me the 2 ways are differents since this is for 2 contexts. File browsing vs WPS
16:56:29n1smoos: yes, the contexts are different but i see no reason why the info shown should differ
16:57:41n1sreusing the plugin would be ok i think but then we lose the translation
16:57:49moosindeed :(
16:58:05n1si still think that is better though
16:58:37moospitty we still don't have yet plugin localisations :(
16:58:46*pixelma wonders were midgey has gone
16:58:55moosthe patch is rotting for long now :(
16:59:36n1si never quite grasped how it should work so never tried to pick it up, maybe that's something for the christmas break :)
17:01:15mooshehe, why not, and could be a nice gift too :)
17:01:37moosWhat should I do then?
17:02:46 Join krabador [0] (
17:03:19CIA-7New commit by kugel (r28877): Redo raising the priority of the codec (and voice) thread to fix audio dropouts under heavy UI load, such as during kinetic scrolling or pictureflow. ...
17:03:38moosI personnaly prefers the way I made to have the string translated and reuse the id3-filesize info too
17:05:04moosthe added string to english.lang could always be deprecated in the future if someone (n1s :) get the plugin localization in...
17:05:16CIA-7r28877 build result: 0 errors, 34 warnings (kugel committed)
17:05:23CIA-7New commit by kugel (r28878): Fix yellow
17:07:01CIA-7r28878 build result: All green
17:07:26moosfor the human_size_prefix made on properties, I don't know if it's really needed, since in most cases music files aren't that big vs file browsing and properties use.
17:08:05n1sin the task midgey stated that the patch was "nowhere ready to commit" but the only thing he explicitely said didn't work well was the build magic
17:09:07pixelmamoos: it's not true for me, .flac or .wav make a huge difference to .mp3 or even some chiptune codec files
17:09:26moosnjust a matter of makefile guru love then? :)s:
17:09:39moosduh n1s
17:09:53pixelmaand it's just the better way than hardcoding something
17:10:31moospixelma we have the infos in bits, that's not that hardcoded
17:10:37n1smoos: sounds like it (and syncing since the last patch is a year and a half)
17:11:14pixelmayou hardcode that it is shown in kB if I understand correctly
17:11:43moospixelma: just to covert the information we have in bits
17:12:26n1sthat will show 0 for files smaller than a kB too
17:12:51kugeljhMikeS: this is what makes the main thread run next:
17:12:56kugelI feel it's too hacky
17:13:28moosn1s, pixelma: I could reuse the petur's human_size_prefix function
17:13:57moosI really just didn't want to touch much
17:14:40n1smoos: <n1s> it's just one call to output_dyn_value
17:14:42pixelmaas been said there already is one core function that could take care of it, n1s even mentioned its name
17:14:53*pixelma too slow
17:14:59moosok I will look at it then
17:16:28 Quit casainho (Remote host closed the connection)
17:18:26 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
17:19:54pixelmaI think the form used in the plugin is not as nice as the core one. IIRC it switches e.g from 9,234 kB to 10MB and that's a huge gap in precision, the core one also has .xx values and at least gives the user the impression to be more accurate
17:20:37moosmore precision indeed
17:20:46pixelmaat least this is how I remember it currently, I believe it puzzled me once
17:22:04n1spixelma: the code agrees with your memory
17:22:20pixelmagreat :)
17:29:58 Quit factor (Ping timeout: 250 seconds)
17:33:14 Join factor [0] (
17:37:26 Quit shai (Quit: Leaving)
17:41:49 Quit n1s (Quit: Lämnar)
17:44:52rasherUsing the same code makes sense anyway. We already got it right one place, let's not reinvent the wheel
17:53:13 Quit mc2739_ (Quit: leaving)
17:56:19 Quit petur (Quit: Page closed)
18:00:07moosrasher, n1s, pixelma: I'm finishing works, then I'll look at it...
18:02:13 Join wodz [0] (
18:02:39*wodz just got archos studio 20 and nano 2g with broken screen
18:11:39 Join JesusFreak316 [0] (
18:13:55 Quit TheSeven (Ping timeout: 240 seconds)
18:14:39 Quit swilde (Quit: ERC Version 5.3 (IRC client for Emacs))
18:16:16***Saving seen data "./dancer.seen"
18:19:11 Join kadoban [0] (
18:22:56 Join shai [0] (
18:23:50 Join thomasjfox [0] (
18:26:11wodzI managed to start playback blindly on nano 2g, so this has 'only' broken screen
18:34:48 Join eWill [0] (
18:36:37Bushmillsthat's where a voice file may come in handy
18:37:32eWillI trying to prevent dir change when I skip then immediately fast-forward/rewind. I thought commenting-out these two lines would do it: /apps/gui/wps.c line:833 and line:859 It's still skipping dirs though...?
18:38:08eWill*not the lines, but the "else" statements there
18:39:58eWillmaybe I need to clean build
18:41:43 Part Dustri ("WeeChat 0.3.3")
18:42:36 Join bmbl [0] (
18:42:36 Quit bmbl (Changing host)
18:42:36 Join bmbl [0] (~bmbl@unaffiliated/bmbl)
18:44:51 Quit krabador (Ping timeout: 240 seconds)
18:48:01eWillnope. Still skips.
18:50:10 Quit moos (Ping timeout: 260 seconds)
18:52:39 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
18:55:19 Quit cjcopi (Ping timeout: 264 seconds)
18:56:14 Join domonoky1 [0] (
18:58:12 Quit domonoky (Ping timeout: 245 seconds)
19:00:11 Quit kugel (Remote host closed the connection)
19:00:19eWillHere's the diff. Why don't it work?
19:04:32 Quit thomasjfox (Ping timeout: 255 seconds)
19:07:41 Quit kadoban (Ping timeout: 255 seconds)
19:08:00 Join kugel [0] (~kugel@
19:08:01 Quit kugel (Changing host)
19:08:01 Join kugel [0] (~kugel@rockbox/developer/kugel)
19:13:47 Quit kugel (Quit: Yaaic - Yet another Android IRC client -
19:14:05 Join kugel [0] (~kugel@
19:14:06 Quit kugel (Changing host)
19:14:06 Join kugel [0] (~kugel@rockbox/developer/kugel)
19:20:04 Join Keripo [0] (
19:20:55 Nick YPSY is now known as Ypsy (
19:29:39 Quit Keripo (Read error: Connection reset by peer)
19:31:35 Join Keripo [0] (
19:32:08 Quit Keripo (Client Quit)
19:45:08 Part marines
19:45:46 Join Zinc [0] (
19:46:13 Join Zinc_Alloy [0] (
19:46:13 Quit Zinc (Client Quit)
19:49:55 Quit kugel (Ping timeout: 240 seconds)
19:50:00 Join Horscht [0] (~Horschti@xbmc/user/horscht)
19:51:28 Quit Zinc_Alloy (Quit: CGI:IRC)
19:51:41 Join zinc_alloy [0] (
19:55:03 Join kugel [0] (~kugel@
19:55:04 Quit kugel (Changing host)
19:55:05 Join kugel [0] (~kugel@rockbox/developer/kugel)
19:56:53 Quit zinc_alloy (Quit: CGI:IRC)
19:57:06 Join ZincAlloy [0] (
19:58:52 Join Strife89TX [0] (~cstrife89@
19:59:16Strife89TXDoes Pictureflow create a log file for listing the album art it reports as "bad"?
20:04:14Strife89TXHmm, evidently not .... Dammit.
20:07:03 Quit ZincAlloy (Quit: CGI:IRC (Ping timeout))
20:12:02kugelStrife89TX: no but it shows the ? for each missing so it should be pretty obvious
20:13:40Strife89TXkugel: Some albums I outright don't have AA for, so actually not-so-obvious. :/
20:16:19***Saving seen data "./dancer.seen"
20:20:12 Nick Ypsy is now known as YPSY (
20:38:20 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
20:45:30 Join marines [0] (
20:45:43 Quit kugel (Ping timeout: 264 seconds)
20:46:03 Quit JesusFreak316 (Remote host closed the connection)
20:46:50 Join kugel [0] (
20:46:51 Quit kugel (Changing host)
20:46:51 Join kugel [0] (~kugel@rockbox/developer/kugel)
20:47:46 Quit Strife89TX (Quit: Shutting down Palm.)
20:52:04 Join bertrik_ [0] (~bertrik@rockbox/developer/bertrik)
20:52:33 Quit bertrik (Read error: No route to host)
20:52:36 Nick bertrik_ is now known as bertrik (~bertrik@rockbox/developer/bertrik)
20:55:41 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
21:06:28wodzpixelma: ping
21:07:21 Join ZincAlloy [0] (
21:27:39bertrikhah, the clock change on AMSv1 broke uSD apparently
21:28:21pixelmadidn't that happen before?
21:28:26pixelmawodz: pong
21:30:25 Join Rob2222 [0] (
21:36:46wodzpixelma: is there a way to ident wps files?
21:37:27pixelmanot that I know of
21:38:40wodzthis is serious flaw of the syntax
21:38:44pixelmayou could break overly long lines but would need to start at position 1 in the next line again
21:38:51pixelmaI don't think so
21:39:41 Join kadoban [0] (
21:41:23pixelmaserious flaw is much of an exageration I think. The way it is you don't have to use any other means of notation to insert spaces etc. and adding more complexity to the syntax just for this isn't necessary IMO, it's not a full programming language
21:41:24 Join baptiste [0] (
21:43:45 Join JesusFreak316 [0] (
21:44:06 Quit JesusFreak316 (Client Quit)
21:47:01gevaertskugel: is your database symlink work somewhere?
21:47:35kugeli hope so
21:49:13gevaertsIs it somewhere I can get at?
21:49:23kugelgive me a second
21:49:57 Join JesusFreak316 [0] (
21:51:33kugelnot a single hunk applies, let me fix that
21:51:36 Join [Saint] [0] (S_a_i_n_t@
21:52:18 Join zenlunatic [0] (
21:53:05zenlunatici just installed rb on a sansa clip+. when I attach to my laptop, lsusb doesnt detect the device
21:58:00gevaertskugel: thanks! Any chance of that appearing on flyspray at some point?
21:58:24kugelyea, I should do that
22:00:14zenlunaticrebooted device now mounts...
22:00:35pixelmazenlunatic: you currently still have to reboot into the OF on the Clip+ for a USB connection
22:01:08zenlunaticpixelma: thanks
22:01:18 Quit bmbl (Quit: Verlassend)
22:02:31zenlunaticpixelma: how do i designate booting into rb on reboot
22:05:00pixelmajust turn it on? Booting into the OF should need you to hold a button during the boot (or maybe already being connected to USB), if you installed official bootloader and Rockbox version
22:07:22zenlunaticyeah it was connected to usb, thats why
22:07:56 Quit baptiste (Remote host closed the connection)
22:09:14 Quit benedikt93 (Quit: Bye ;))
22:09:22zenlunaticwow. this is nice... thanks pixelma
22:16:22***Saving seen data "./dancer.seen"
22:17:12kugelgevaerts: does that patch work?
22:17:23kugel(the RaaA for n900 one I mean)
22:17:44kugel( - for the reference)
22:18:39gevaertsmore or less
22:18:53gevaertsWell, it works, but it doesn't feel much better than unmodified svn to me
22:19:28gevaertsThat's why I want the next one :)
22:21:35 Quit [Saint] (Quit: I'm only going to Heaven if it feels like Hell, I'm only going to Heaven if it tastes like caramel...)
22:21:44kugelthe sdl graphics part is supposed to be much faster
22:22:13 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
22:22:52gevaertsYes, but while that's certainly useful that's not where the really noticeable problems are
22:23:13 Quit domonoky1 (Ping timeout: 245 seconds)
22:23:50gevaertsIt "just" makes the thing use a bit less CPU. It doesn't solve the stuttering that sometimes happens (mainly when the LCD switches off)
22:24:17gevaertsI'm hoping that the more recent work improves that, since it should do sound a lot better
22:24:55kugelso do I
22:26:01 Join b1uebrother [0] (~dom@
22:26:58gevaertsright now just running rockbox makes pulseaudio jump to 25% CPU or so
22:27:37 Quit kadoban (Read error: Operation timed out)
22:28:14 Join moos [0] (moos@rockbox/staff/moos)
22:30:01kugelwithout audio?
22:31:26gevaertsyes, immediately after "boot"
22:31:56gevaertsThe CPU might not be running at full speed though, but still
22:32:28gevaertsActually, it's nearer 30%
22:32:50kugelthat's strange, pulseaudio does really only audio, doesn't it?
22:33:16kugelit seems the sdl port has a lot of performance problems we didn't know about yet
22:33:16gevaertsI hope so :)
22:37:42 Quit tchan (Quit: WeeChat 0.3.3-dev)
22:38:48 Quit Topy (Read error: Connection reset by peer)
22:39:35gevaertsYes. The numbers I see on n900 make the android port quite impressive
22:43:47 Quit moos (Ping timeout: 255 seconds)
22:43:55 Join moos [0] (moos@rockbox/staff/moos)
22:45:29eWillI trying to prevent dir change when I skip then immediately fast-forward/rewind. Here's my .diff. Anyone know why it isn't working?
22:51:09 Part marines
22:51:09jhMikeSkugel: I have to agree that's a bit painful to look at :o
22:52:11jhMikeSthing is though, how's that effectively different from just calling thread_set_priority and restoring?
22:53:17 Join Topy44 [0] (
22:53:42kugeljhMikeS: restoring immediately you mean?
22:54:55jhMikeSyes, like what you tried before
22:55:13 Join Buschel [0] (
22:55:19kugeli haven't tried that yet
22:55:50 Join JdGordon [0] (7a6edfaa@gateway/web/freenode/ip.
22:56:03jhMikeSI thought you tried calling set_thread_priority from the ISR and restoring when it runs...I thought...hmmm
22:56:22kugelno, that's what the above patch does
22:57:11Buschelsoap: I propose that you might add a change to svn which uses 76MHz for nano 1G. it is not the first time I have heard about the need for using a lower clock
22:57:36 Join {phoenix} [0] (
22:58:14kugeljhMikeS: what did you think the patch does? it does what you have thought would have been my other try
22:59:42 Quit b1uebrother (Quit: leaving)
23:00:47 Join dfkt_ [0] (dfkt@unaffiliated/dfkt)
23:00:47eWillNevermind −− it works. I forgot to reboot after I made those changes.
23:00:51jhMikeSkugel: that's what I though it does, with extra steps to boot! :) still, it doesn't guarantee it runs next if its stuck else where
23:01:15 Quit dfkt (Disconnected by services)
23:01:16soapBuschel, I don't have that power, nor do I feel right with something going through for what appears to be, today, one person's issue.
23:01:18 Nick dfkt_ is now known as dfkt (dfkt@unaffiliated/dfkt)
23:02:29kugeljhMikeS: I assume it will run next once it becomes runnable. Or is set_priority() undefined for non-running threads?
23:03:19jhMikeSkugel: it's defined. the cases where it would be like would be exceptions, like being 3rd waiting for a mutex or something.
23:03:34kugelthat's what I thought the huge'ish else case in set_priority() was for
23:03:51Buschelsoap: do we loose the realtime ability for a codec? if not (and I do no think so), a change of -5% clock is not that relevant.
23:04:07Buschel(from my point of view)
23:04:19jhMikeSkugel: that's to boost the owner of a lock when the owner has lower priority than the waiter, likewise for a queue owner
23:04:45jhMikeSkugel: the inheritance is transitive across a chain of blocking threads
23:04:56Buschelsoap: can you provide a patch?
23:05:23kugeldoes that mean the owner threads don't get the priority lowered?
23:05:26soapBuschel, would likely make APE C2000 unusable.
23:06:43jhMikeSkugel: if you raise prio on a thread waiting and it's the higher priority waiter, then all thread blocking it get boosted too, till they no longer own the resources the highest prio is requesting
23:07:29Buschelsoap: c2000 is tight :/
23:09:38soapis all I did.
23:09:45jhMikeSkugel: says it better than I did I guess: :)
23:09:55 Quit ender` (Quit: Shortly after Dr. Samuel Johnson had published his Dictonary of the English Language (in 1755), a splendid British lady approached him and congratulated him on omitting all improper and coarse words from his dictionary. "Ah," replied Dr. Johnson, "so yo)
23:10:07soapFor all I know this breaks everything. All I know is I got a few hours of playback with this change whereas I could not do so before.
23:10:20kugeljhMikeS: I think I understand what it does. I'm wondering if the blocker threads get their priority raised?
23:12:03kugelif yes, then it should be save, shouldn't it?
23:12:13jhMikeSkugel: yes, of course, if the object involves PI (mutexes and owned queues when calling queue_send)
23:13:06soapwhat are amiconn's thoughts on making SVN 76Mhz for the Nano 1G?
23:13:58jhMikeSkugel: Is this "run next" really that helpful?
23:14:42kugelit's indeed a bit smoother yes
23:15:42kugelit's not about having the main thread more often, but about doing the lcd_update-after-moving-list with a (nearly) constant time difference
23:16:39kugelit makes especially the smoothness not depend so much on the cpu load (e.g. music playing or not)
23:16:43Buschelsoap: let's see what amiconn will say
23:17:24soapIs the timing of anything else dependent upon the boosted speed?
23:17:25jhMikeSkugel: heh, that needs preemption to *really* do :o)
23:17:26 Quit JdGordon (Ping timeout: 265 seconds)
23:18:05kugeljhMikeS: ah yes, can you do me the favor if implementing preemption then? :)
23:18:09jhMikeSkugel: that doesn't result in re-starving the codec thread since it effectively makes it way higher than even realtime
23:18:50kugelis that a question?
23:19:04jhMikeSkugel: there's other reason I'd want it too that have to do with audio output
23:19:06Buschelsoap: in terms of registers -> IDE0_CFG and DEV_TIMING1.
23:19:17Buschelsoap: maybe also RAM
23:19:58soapI did not test with any writing to the drive, except for the log (which I haven't looked at).
23:20:14Buschelsoap: from what I have seen so far I think the RAM speed is directly connected to the CPU clock
23:20:17soapI should go get it and test USB function at the least.
23:21:56Buschelsoap: does your nano also freeze (or even freeze faster) when keeping it boosted?
23:22:31soapI can test that. Let me go get it. Are you thinking perhaps the switching itself is doing something?
23:22:56Buschelnot sure about this, would just be good to know
23:23:01jhMikeSkugel: that was. I'd expect any codec boost to be made null and void
23:23:02 Quit kevku (Ping timeout: 272 seconds)
23:23:26kugelit doesn't starve other threads since it's back at default priority immedialtely when it runs again, it'll only put the other threads behind a bit
23:24:18jhMikeSyou never tried the skip_count change?
23:24:28kugelI'd be more than happy to have a way to make it run next without messing with the priorities though
23:24:39kugelno I haven't yet
23:25:06 Quit evilnick_B (Quit: Page closed)
23:26:26jhMikeSkugel: I was think about being a bit more gentle. what if it doesn't *have* to run next but just soon enough or won't be skipped next time it's seen? most tasks don't do much before sleeping again
23:26:59kugelthat won't change anything I think
23:27:36kugelmy goal is to make the time between two updates constant for maximum smoothness
23:30:15kugelit's the best we can do to meet this realtime need, isn't it?
23:31:53soapBuschel, ok, just started a long track at a forced 80Mhz.
23:32:24Buschelsoap: will get some sleep now as I am totally tired. I will read the logs tomorrow
23:32:28Buschelsee you
23:32:33 Quit Buschel (Quit: ChatZilla 0.9.86 [Firefox 3.6.13/20101203075014])
23:33:51jhMikeSkugel: could schedule by shortest job first instead, which has minimal latencies...or...raise the skip count and move it to the head of whatever list it is currency on, even if blocked elsewhere
23:35:22 Join Keripo [0] (~Keripo@
23:37:55 Quit mortalscan (Ping timeout: 240 seconds)
23:38:56 Quit ZincAlloy (Quit: CGI:IRC)
23:42:22 Quit bertrik (Quit: :tiuQ)
23:43:20 Join marines [0] (
23:45:20 Join mortalscan [0] (~mortalsca@
23:47:42kugeljhMikeS: I tried moving it around lists, but that only crashed on me
23:48:27kugelshortest job first probably isn't what we want, updating the lcd isn't exactly short
23:50:44jhMikeSthere's a variant of that one apparently, but, moving it in lists should be workable also. if it crashed, something got corrupted
23:51:41jhMikeScrashes can be quite interesting if the list gets done wrong :)
23:53:21kugelhard to debug also
23:53:49 Quit CaptainKewl (Read error: Operation timed out)
23:55:41jhMikeSyou're telling me!
23:56:02jhMikeSI had another idea though that involves no priority or list moving
23:56:16 Quit domonoky (Read error: Connection reset by peer)
23:56:48jhMikeSif a thread can be marked as "urgent", and the scheduler know an urgent thread exists, it can search those first
23:57:22 Join telliott [0] (
23:57:39kugelI had that idea too, it also crashed, but I probably did it wrong
23:58:51kugelI think I didn't handle it when the urgend one yields again, so the scheduler thought it wasn't an urgend one and thought it was in some list
23:58:57jhMikeSit is the least expesive way to go about it. tell the core it has them, then mark the threads. among those, schedule normally over and above those not marked

Previous day | Next day