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 2009-01-20

00:00:24amiconngcc doesn't know about the emac so it won't get in the way. This changed in 4.3.x afaik, so quite a bit of code will need review if we switch
00:01:42 Join n17ikh [0] (n=n17ikh@
00:02:38 Join solexx_ [0] (
00:02:55 Quit solexx (Read error: 54 (Connection reset by peer))
00:03:29 Quit MethoS (Connection timed out)
00:03:31Unhelpfulwill movclr from accext01 clear acc0? or will i have to do that explicitly when the scaler is finished?
00:04:52Unhelpfulor can i not use it with the accumulator extension reg?
00:05:50 Quit MethoS- (Read error: 110 (Connection timed out))
00:05:56 Join m0f0x [0] (
00:06:26saratogaholy crap that floor0 vorbis file mallocs an extra half MB of data vs. current builds
00:06:55Zagoramiconn: doesn't look like it, unfortunately. It does 1884 loops waiting for the lock bit, even if not changing the vco :-(
00:07:11 Join Shaid [0] (
00:07:20 Nick reacocar1 is now known as reacocard (i=reacocar@
00:07:44amiconnWell you don't change the vco directly. It is changed by the pll circuit trying to lock the phases
00:08:10amiconnThe only way to avoid this is to not change the pll dividers at all
00:08:19Zagoryeah I mean VCODIV and VCOOUT
00:08:29amiconnDid you try to check whether there is a wait for relock if you only change the post divider?
00:08:43 Quit gregorovius (Read error: 110 (Connection timed out))
00:09:00Zagoryes that is what I just tried
00:09:31ZagorI only changed CPUDIV and it still spent 1884 loops waiting for lock
00:09:54amiconnof course
00:10:21amiconnIf you change either VCODIV or PLLDIV, it will need to relock
00:10:36Zagorwell I didn't. I changed CPUDIV.
00:10:55amiconnI meant whether you checked your theory that if you *only* change VCOOUT (like in your patch), the relock is really not needed
00:11:09Zagormy patch changes CPUDIV, not VCOOUT
00:11:27amiconnYeah, that should work too
00:11:38*amiconn thinks there are confusingly many dividers
00:12:06 Quit Rob2222 ()
00:12:35kugelZagor: the icache is definetely active on ams, but the dcache not
00:12:44kugelI ran rockbox without icache, it was horrible
00:12:57*amiconn should also do some tests regarding the PLLCR manipulation in pcm-coldfire one of these days
00:13:00PaulJam_Zagor: not sure if this are good or bad news, but with the last patch on FS #9797 i get the same faulty behaior like with the first patch on my h300.
00:13:02Zagorkugel: :-)
00:13:29amiconnI still suspect it might be the cause of the occasional hard freezes.
00:13:34kugelZagor: also, I'm not sure if boosting is alright now. test_boost shows a difference, but it doesn't look like it was 10x faster (more like 2-3x)
00:13:36ZagorPaulJam_: I consider it good news. that means we still haven't seen any bad side effects from the fast switching.
00:13:41Unhelpfulhrm... so, it looks as if having my result data in %accext means that i will need an explicit clear after each operation?
00:13:53 Quit jhulst (Read error: 60 (Operation timed out))
00:14:06Zagorkugel: we need to fix sdram timing.
00:14:08amiconnYeas, but that shouldn't bee too bad
00:14:17 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
00:15:22 Join fdinel [0] (
00:16:03Unhelpfuli don't see where one can *do* that? just use a dummy movclr? or a move-to-accumulator?
00:16:04saratogakugel: is dcache on?
00:16:19TheSkunkManhow can i use the tools that are in the svn such as descramble.c?
00:16:21saratogawell that will explain the results with boosting
00:16:25kugelit can't, the datasheet tells it cannot be used without mmu
00:16:44saratogawithotu cache theres little sense in boosting unless you're operating out of IRAM
00:17:02BagderTheSkunkMan: you build them, like with 'make tools' or the specific ones
00:17:04kugelmakes sense
00:17:52 Quit yhuang ("Leaving.")
00:18:32 Join japc_ [0] (
00:19:46 Quit mcuelenaere (Read error: 113 (No route to host))
00:21:16Unhelpfulis junk in the top half of accext0, which "belongs" to acc1, anything i'll need to worry about?
00:22:03amiconnAs long as you don't use acc1 yourself, the top half should be zero (due to our convention)
00:22:14 Quit bertrik ("Leaving")
00:22:25 Quit Zagor ("Clint excited")
00:24:21 Join yhuang [0] (
00:25:00Unhelpfulshould i just use movclr to a dummy variable to clear acc0? or move.l #0, %%acc0?
00:26:30amiconnEither method works
00:27:04Unhelpfulthe latter doesn't clobber a register... i'm not sure how much that will matter on coldfire.
00:27:13 Quit itcheg (" ajax IRC Client")
00:27:53 Quit z35 (Read error: 110 (Connection timed out))
00:28:18 Join z35 [0] (
00:29:21Unhelpfultbh, it might be sensible on sh-1 to just use the >>24 version. anybody loading bitmaps >500x500 on that target is asking for trouble ;)
00:31:30 Quit Shaid ("*toink*")
00:32:49Unhelpfulgevaerts: if you don't mind testing again, i've re-uploaded. if it still doesn't work, i think i'll leave making it better to somebody who has one.
00:36:08gevaertsUnhelpful: still bad
00:36:56Unhelpful... actually, i see an obvious error :/
00:37:17JdGordonPaulJam_: you said v8 of the playback patch breaks wps sublines... does that mean v7 worked?
00:37:41PaulJam_JdGordon: yes, and i think iknow why...
00:37:57 Nick fxb__ is now known as fxb (
00:38:12Unhelpfuli changed the emac mode, then changed the code to work correctly in integer mode... then left it setting the emac in fractional mode.
00:39:59Unhelpfulgevaerts: once more, for great justice? ;)
00:40:04PaulJam_the changes in gwps.c look like it does a full WPS update (which resets the sublines) when "update_track" is true, but if i understand the code correctly then "update_track" is almost always set to true in the large switch statement.
00:40:18PaulJam_this was for JdGordon
00:40:28JdGordoneeek... ok
00:40:43JdGordonyeah, I have to admit that was a quick fix for the "ui responsiveness" problem
00:43:11gevaertsUnhelpful: looks OK now
00:43:40 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
00:44:16JdGordonPaulJam_: can you comment out line 639 in gwps.c and see if that fixes things? or makes things worse?
00:44:45JdGordonactually.. yeah dont bother
00:44:57JdGordonneeds a bit more sofistocated fix than that :)
00:45:40Unhelpfulfaboo, we have a coldfire-optimized version of the scaler, with the same math that the C version uses, then
00:46:21Unhelpfulthe C translates to reasonable asm on arm... sh-1 still needs some love :)
00:46:32 Quit film42 (Client Quit)
00:46:39gevaertsgcc-Unhelpful :)
00:47:09 Join film42 [0] (
00:47:21Unhelpfulit might be reasonable to use the >>24 C code there, it's still accurate to much larger scales than you want to do on... 12MHz?
00:48:16 Quit jhulst (Read error: 113 (No route to host))
00:48:37amiconn12MHz in Ondios and Player (not relevant here), 11MHz in Recorders
00:51:24amiconnAs long as no variable shifts are involved, SH1 code is pretty efficient, and RAM speed compared to CPU speed is decent (unlike on coldfire)
00:52:17amiconnIt does not yet use SDRAM, but fast page mode DRAM. Just 2 waits if the page isn't open, single-cycle for an open page. 16 bits wide.
00:52:37 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/ajax/
00:53:07 Quit tyfoo (Read error: 104 (Connection reset by peer))
00:56:12 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
00:56:28Soap_amiconn: I think you are the most (only) knowledgeable person re:1st gen iPods.
00:57:10amiconnAh, that problem :(
00:58:12 Quit ender` (" Friends come and go, but enemies accumulate. - Jones Law - ... and sometimes the real trick is telling the difference. --")
01:01:13 Join akur1 [0] (
01:03:02 Join itcheg [0] (i=62db4767@gateway/web/ajax/
01:06:47 Quit cool_walking_ (" ajax IRC Client")
01:07:06 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/ajax/
01:08:28 Quit flydutch ("/* empty */")
01:10:08 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
01:10:50 Part akur1
01:15:12 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
01:17:26 Quit itcheg (" ajax IRC Client")
01:17:57 Join PaulJam__ [0] (
01:18:52 Quit jhulst (Read error: 131 (Connection reset by peer))
01:19:15 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
01:24:14 Quit arohtar (Read error: 110 (Connection timed out))
01:24:21 Quit faemir (Read error: 110 (Connection timed out))
01:31:05 Quit yhuang (Remote closed the connection)
01:31:18 Part toffe82_
01:32:24 Quit Thundercloud (Remote closed the connection)
01:37:37 Quit DerDome ("Leaving.")
01:37:51 Quit PaulJam_ (Read error: 110 (Connection timed out))
01:39:24 Quit efyx (Read error: 110 (Connection timed out))
01:41:21 Quit Nico_P (Remote closed the connection)
01:41:49 Quit kugel (Read error: 104 (Connection reset by peer))
01:44:39 Join itcheg [0] (i=62db4767@gateway/web/ajax/
01:47:07 Quit jhulst (Remote closed the connection)
01:49:35tmztis it possible to use codesourcery arm toolchain as arm-elf-gcc?
01:50:20 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
01:50:32rashertmzt: It might be, and it might not be. If it fails, don't blame us
01:50:34***Saving seen data "./dancer.seen"
01:51:09tmztwhat is the right toolchain to use? I'm trying to find documentation in the wiki but it seems user focused
01:51:32rashertmzt: the CrossCompiler page should tell everything you need to know (and the "For developers" part of DocsIndex)
01:51:53tmztthank you
01:53:57 Quit BHSPitLappy (Remote closed the connection)
01:54:40 Nick fxb is now known as fxb__ (
02:07:10 Quit PaulJam__ (".")
02:11:42 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
02:11:44 Join mcuelenaere_ [0] (n=mcuelena@rockbox/developer/mcuelenaere)
02:11:47 Quit mcuelenaere_ (Remote closed the connection)
02:13:21 Quit m0f0x (Read error: 110 (Connection timed out))
02:16:12 Quit gregorovius_ ()
02:18:09 Join tajlero [0] (
02:24:53 Join soho [0] (
02:25:05sohohey guys, i have a serious problem with my rockbox software
02:25:11sohoi think its broken
02:26:34sohocan someone help?
02:27:09kadobansoho: just ask, don't ask to ask.
02:27:46sohoi took my ipod nano with rockbox, charged it up on my Mac
02:27:54sohoand it doesnt even load
02:28:03sohoso i formated with apple
02:28:10sohobut some of rockbox is still left
02:28:58sohokadoban can you walk me through the steps to reinstall rockbox
02:29:00sohousing linux
02:29:15kadobansoho: not really right now, the manual can though :P
02:29:44sohoah ok
02:31:14 Quit Soap_ (Read error: 110 (Connection timed out))
02:31:26 Join gregorovius [0] (
02:31:28 Quit soho ("Ex-Chat")
02:35:16 Quit jhulst (Remote closed the connection)
02:36:24 Quit jgarvey ("Leaving")
02:36:39 Join Soap_ [0] (
02:37:43 Join toffe82 [0] (
02:39:28 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
02:42:19 Quit sarixe ("Connection reset by the motherfucking peer")
02:57:04 Join MaxHR [0] (
02:58:58 Quit culture (Read error: 110 (Connection timed out))
03:00:55 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
03:02:50 Quit DataGhost (Nick collision from services.)
03:02:58 Join DataGhost [0] (i=dataghos@unaffiliated/dataghost)
03:14:04 Join nibbler [0] (
03:16:26 Join sarixe [0] (
03:19:23 Part toffe82
03:31:41 Part Aurix_Lexico
03:41:02 Quit gregorovius (Read error: 110 (Connection timed out))
03:41:18 Join gregorovius [0] (
03:45:03 Join MrDuck [0] (
03:46:08 Quit itcheg (" ajax IRC Client")
03:46:15 Quit kachna|lappy (Read error: 131 (Connection reset by peer))
03:48:31 Join goffa_ [0] (n=goffa@
03:50:37***Saving seen data "./dancer.seen"
03:53:28 Quit goffa (Read error: 145 (Connection timed out))
03:54:12 Quit MaxHR ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
03:55:14 Quit ajb (Read error: 104 (Connection reset by peer))
03:55:26 Quit JdGordon (Remote closed the connection)
03:56:06 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
03:57:07 Join goffa [0] (n=goffa@
04:01:18 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
04:02:11Unhelpfulok, here's the patch, asm for coldfire, >>24 with C math for sh, >>32 with C math otherwise
04:06:34 Join sbhsu [0] (
04:09:43 Quit goffa_ (Read error: 110 (Connection timed out))
04:13:10 Join Barahir_ [0] (
04:18:59 Quit gregorovius ()
04:19:43 Join kronflux [0] (
04:19:56 Quit miepchen^schlaf (Read error: 101 (Network is unreachable))
04:20:09 Join blkhawk- [0] (
04:20:49 Quit blkhawk (Read error: 60 (Operation timed out))
04:21:07 Nick blkhawk- is now known as blkhawk (
04:21:45kronfluxcan anyone give me a compiled binary of the development build for the fuze? just wanna play around with it, check for bugs, etc.
04:23:43saratogakronflux: its not really usable so we don't really need bug checking
04:24:29kronfluxwell, more curiousity than anything :p I mean, if you happen to have a binary laying around with the latest work on it.
04:25:36saratogakronflux: have you read through the install instructions?
04:25:52kronfluxI'm reading the irc logs, and looking over the forum post now, but I'm not entirely sure how to compile or install specifically.
04:26:08kronfluxnot too up to date on it. but last I heard, a lot of progress was made.
04:27:02saratogayou may wish to review the install process, compiling rockbox is not the only complicated step
04:27:31kronfluxany handy links to said instruction?
04:27:41kronfluxif its not a hassle ^_^
04:27:48saratogathey're in teh wiki
04:29:07 Quit Barahir (Read error: 113 (No route to host))
04:30:30kronfluxshould it logically not be as easy as having a built firmware file, and uploading it?
04:31:14kronfluxI mean.. yeah.. if you're compiling it yourself and such... its a lot more difficult.. but hypothetically, as I had asked, if someone had it prebuilt, could one not just grab that file and upload it, then the device would do the work?
04:31:35saratogaprovided you've reformated the device and compiled the tools to upload it, yes
04:32:33saratogaactually maybe someone could patch the firmware file and give you that now that i think about it
04:33:35kronfluxthat would be nice
04:33:51kronfluxif anyone is willing to try, I'd be grateful ^^
04:36:26 Join AndyIL [0] (i=AndyI@
04:39:50 Join efyx [0] (
04:40:01 Quit goffa ("leaving")
04:41:14 Quit film42 (Client Quit)
04:43:14 Quit intrados (Success)
04:44:02kronfluxno? i tried
04:45:05Unhelpfulmany of the people who might have fuze builds on hand aren't even here, right now.
04:45:26kronfluxah well. I'll try again in the morning if I havent built one myself by then
04:45:32 Quit kronflux ("Leaving")
04:47:37 Quit AndyI (Read error: 110 (Connection timed out))
04:52:34 Join film42 [0] (
05:00:48CtcpVersion from Bensawsome!n=Bensawso@unaffiliated/bensawsome
05:02:30 Join webguest84 [0] (n=183f507c@gateway/web/cgi-irc/
05:02:50 Quit tajlero ("Leaving")
05:02:58 Quit webguest84 (Client Quit)
05:07:40 Join tajlero [0] (
05:15:03 Join yhuang [0] (
05:18:18 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
05:18:46 Quit yhuang (Client Quit)
05:19:54 Quit jhulst (Read error: 104 (Connection reset by peer))
05:19:55 Join yhuang [0] (
05:20:28 Join jhulst [0] (
05:24:35 Join webguest60 [0] (n=95abad40@gateway/web/cgi-irc/
05:31:08 Quit nibbler (Read error: 110 (Connection timed out))
05:38:27 Quit webguest60 ("CGI:IRC (EOF)")
05:40:24 Quit Horscht ("Verlassend")
05:40:30 Quit yhuang ("Leaving")
05:42:45saratogaLear: for what its worth, playing AAC files with reasonable memory usage by shrinking the sample_byte_size table is easy and massively reduces memory usage
05:43:16saratogai think seeking can be made to work without too much trouble, but so far I keep getitng crashes, probably just need to read through the specs when i'm rested
05:43:28 Join yhuang [0] (
05:45:40 Join Henry123456789 [0] (
05:48:08 Part Henry123456789
05:48:49 Quit yhuang ("Leaving")
05:49:05 Join yhuang [0] (
05:50:40***Saving seen data "./dancer.seen"
05:57:25 Quit saratoga ("CGI:IRC (EOF)")
06:09:17 Quit z35 ("Leaving")
06:17:31 Quit tajlero ("Leaving")
06:25:36 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
06:25:41 Quit tim__ (Read error: 104 (Connection reset by peer))
06:25:46 Join timc`` [0] (n=aoeu@
06:51:26 Join _lifeless [0] (n=lifeless@
07:09:54 Quit __lifeless (Read error: 110 (Connection timed out))
07:10:46 Join nibbler [0] (
07:15:13 Join Bagderr [241] (n=daniel@rockbox/developer/bagder)
07:15:40 Nick Bagderr is now known as B4gder (n=daniel@rockbox/developer/bagder)
07:25:12 Quit Seed ("cu, Andre")
07:40:02 Quit bmbl (Client Quit)
07:42:41 Nick fxb__ is now known as fxb (
07:43:32 Quit reacocard (Remote closed the connection)
07:43:43 Quit nibbler (Read error: 110 (Connection timed out))
07:48:39 Join reacocard [0] (
07:50:41***Saving seen data "./dancer.seen"
07:50:49 Nick fxb is now known as fxb__ (
07:51:08tmztwhere does bootloader-fuze come from? it looks like it was in the old SansaV2 wiki page (google cache) and have read mkamsboot.c and dualboot.S and don't see where I get this file from
07:51:30LloreanYou compilei t.
07:52:10tmztit's not in build
07:52:23 Join kugel [0] (n=chatzill@unaffiliated/kugel)
07:52:33LloreanWhat do you mean?
07:52:42LloreanYou built a bootloader target and didn't get the bootloader?
07:52:46tmztit's not in the build folder
07:52:55LloreanDid you choose to build a bootloader?
07:53:47tmztI see, configure and choose (B)ootloader
07:55:44kugeltmzt: did you run or do you use codesourcery now?
07:56:20kugelok, that should work (even though I was curious if codesourcery worked ;) )
07:56:20tmztit's patched for multilib I see
07:56:41tmztI will be happy to try codesourcery for build but not bootloader
07:58:37kugeltmzt: as long as you don't compile mkamsboot with something experimental you're fine. a not working bootloader is just as undangerous as the build
07:59:03tmzton fuze? I thought there was no recovery
07:59:25kugelbut mkamsboot contains critical code for dual booting, and once you're in the bootloader it's already too late
07:59:47kugeland dualboot is even before the bootloader
07:59:55tmztoh, so bootloader is the compressed part?
08:00:27kugelif we had more space in the OF to put our code into we wouldn't even need a bootloader ;)
08:00:39tmztis there a filename I need to name the output of mkamsboot to flash?
08:01:34kugelit has to be named like a OF, so the fuze accepts it and does the upgrade
08:01:42kugellike, fuzet.bin
08:08:45 Quit cool_walking_ (" ajax IRC Client")
08:16:28 Quit _lifeless (Remote closed the connection)
08:17:13 Join _lifeless [0] (n=lifeless@
08:18:12tmztkugel: I'm cleaning up my nand storage now to prepare to resize the filesystem (not partition) and I'm not sure how to do it wihout changing the partition table if this really is superfloppy mode
08:19:15 Quit _lifeless (Remote closed the connection)
08:27:20 Join Willwolfe [0] (
08:29:02 Quit Willwolfe (Client Quit)
08:30:04 Quit kugel (Read error: 110 (Connection timed out))
08:32:47 Quit _Auron_ ("Infinity repeatedly denies rumours of plotting with zero to bring down the Universe.")
08:33:29 Quit CaptainKewl (Read error: 54 (Connection reset by peer))
08:33:51 Join CaptainKewl [0] (
08:35:38 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
08:35:56 Join gartral [0] (
08:39:18gartralgood morning all, I'm having a minor error in my build environment, it still tries too call the old file, though it doesn't seem too affect much of anything
08:40:10 Join ender` [0] (
08:42:02 Join _lifeless [0] (n=lifeless@
08:43:56 Quit BigBambi (Read error: 113 (No route to host))
08:44:26 Join tajlero [0] (
08:48:27 Quit BHSPitMonkey (Remote closed the connection)
08:49:23 Join _Auron_ [0] (
08:50:55 Join tvelocity[a] [0] (
08:53:01 Quit tvelocity (Read error: 60 (Operation timed out))
08:54:05 Nick fxb__ is now known as fxb (
08:54:13 Quit Llorean (Read error: 104 (Connection reset by peer))
08:54:36 Join Llorean [0] (
09:00:00amiconnjhMikeS: Is there a reason why the new usb detection is only for c200 and e200? Or do I misunderstand something?
09:01:13lucentusb detection also on fuze btw
09:01:19lucentunless I'm talking nonsense
09:05:40 Join petur [50] (n=petur@rockbox/developer/petur)
09:08:07 Join japc__ [0] (
09:09:17 Join kugel [0] (n=chatzill@unaffiliated/kugel)
09:11:52lucentoh, I am mistaken, just looked at SVN log now
09:19:27 Join gammy [0] (
09:19:51 Quit japc_ (Read error: 110 (Connection timed out))
09:20:27tmztkugel: dual-boot is working, great. svn up shows at r18000 while svn log -l 50 shows newest being r19800
09:21:49 Quit xSlack (Connection timed out)
09:22:24tmztkugel: sorry, I did read that wrong
09:28:41 Join kachna|lappy [0] (
09:29:47 Quit MrDuck (Read error: 104 (Connection reset by peer))
09:34:10 Join ucchan [0] (
09:34:41 Join nibbler [0] (n=Nibbler@
09:41:08 Part gartral
09:45:17 Quit ucchan ("Leaving...")
09:45:38 Join ucchan_ [0] (
09:47:12 Join ucchan__ [0] (
09:50:02 Part ucchan__ ("Leaving...")
09:50:43 Quit ucchan_ ("Leaving...")
09:50:45***Saving seen data "./dancer.seen"
09:52:11 Join ucchan [0] (
09:52:53 Join Thundercloud [0] (
09:58:58 Join PaulJam [0] (
09:58:59 Quit kugel (Read error: 110 (Connection timed out))
10:11:58 Join kugel [0] (n=chatzill@unaffiliated/kugel)
10:17:11 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
10:19:33 Nick Barahir_ is now known as Barahir (
10:23:23 Quit _lifeless (Remote closed the connection)
10:24:51 Join _lifeless [0] (n=lifeless@
10:25:48 Quit _lifeless (Remote closed the connection)
10:32:07 Join DerDome [0] (
10:33:40 Quit Thundercloud (Remote closed the connection)
10:35:34ucchanWhen r19794, iPod Photo's battery value changed. But it is very different FS #9072. Is the committed value correct?
10:39:54Zagorucchan: saratoga committed that. I don't know which values he used. since he didn't close FS #9072 it seems he didn't use those at least.
10:40:24Zagorucchan: are you Yoshihisa Uchida?
10:42:15 Join n1s [0] (n=nils@rockbox/developer/n1s)
10:46:35pixelman1s: remember the labelling reset holes/battery switches etc. in the player's drawings? What do you think about labelling microphone holes as well?
10:47:05n1spixelma: i don't see why not :)
10:47:08LloreanSeems like a good idea.
10:50:15pixelmaalright, I should pick up that work again then... Is "microphone" and/or "reset" enough or does it need the "hole" too? I just find "microphone hole" a tad bit long. Well, "battery switch" is to but no way to shorten it.
10:50:50pixelma*too too ;)
10:52:13B4gder"reset hole" and "microphone" would work imho
10:54:31*Llorean agrees
11:01:11pixelmaalright. Something just crossed my mind - if we start label all kinds of things, what about headphone jacks, USB etc. Where to draw the line?
11:01:32pixelmaoptical in/out
11:09:01 Join _lifeless [0] (n=lifeless@
11:14:16kugelI can't get it to work
11:16:02Zagorkugel: mmu?
11:17:57 Join gartral [0] (
11:18:09kugelI actually don't think I did something wrong, F/X doesn't do more
11:18:16kugelyet the main failsl to boot
11:20:01gartralok, this isnt really a bug, as it doesn't hurt anything, but i noticed that with back light fade-in oon the e200, if you hold a button or spin the wheel when its fading, it causes it too fade slower, without flickering...
11:23:07ucchanAs for iPod Photo, if the commited value is correct, the battery value might be different because of the difference of the capacity of HDD(30GB/60GB) or the model number.
11:23:57ucchanFor iPod Photo users. please coud you run the battery bench ?
11:25:08 Quit kachna|lappy (Read error: 104 (Connection reset by peer))
11:25:15 Join kachna|lappy [0] (
11:33:14 Join snether [0] (n=47854714@gateway/web/cgi-irc/
11:40:47kugelamiconn: ping
11:46:43kugelZagor: any hint what to do about the backlight fading cleanup?
11:46:56kugelI'd like to have a decision
11:47:02 Quit AndyIL ()
11:50:48***Saving seen data "./dancer.seen"
11:53:50 Join AndyI [0] (i=AndyI@
11:54:03 Join ajb [0] (
11:55:35ajbSo who is interested in playlists?
11:56:00B4gderI use them all the time, but I can't say they interest me much ;-)
11:56:15kugelsame here :)
11:59:31ajbI'm just trying to get some responses to FS #9677 and possibly get it merged
11:59:50ajbObviously I think it improves the situation :-)
12:00:05*B4gder never used a dynamic playlist...
12:00:11ajbI guess it depends on how many people build playlists on Rockbox than importing them into the system
12:00:33B4gderand even how they do it
12:00:34ajbB4gder: Do you contruct all your playlists on you PC?
12:01:11ajbB4gder: So how do you deal with them?
12:01:21B4gdercreate a playlist, then play it
12:01:38B4gderI can't say I create them very often, I usually have a few large ones
12:02:01kugelajb: does it fix that the database browser acts weird?
12:02:19kugelyou always get the warning in the database browser, no matter of the previous playlist
12:02:44ajbYou mean even if it hasn't been changed?
12:03:01kugeldoes FS #9677 fix it?
12:03:33 Join Seed [0] (
12:03:37*ajb checks
12:04:47Lloreankugel: I would expect there to need to be some logic for it to work well
12:04:59LloreanThe database doesn't have "folders"
12:05:13pixelmakugel: I don't understand your "bug" description, could you explain more?
12:05:30LloreanAnd queries can have different results each time they're invoked, so a playlist from it will never be the same as one from the filetree (which is more or less assumed to be unchanging)
12:05:51Lloreanpixelma: If you click select on a song in a folder so that you "play" the folder, you don't get the dynamic playlist warning when you try to play something else.
12:06:05LloreanYou only get the warning if you've used "insert" or similar
12:06:34LloreanIn the database, you always get it, because clicking "select" on a song actually involves basically using the "insert" behind the scenes (IIUC)
12:06:51ajbCorrect. I've just tested. Load a playlist, select a song from the DB and it jumps straight to it without warning (as the old playlist isn't modified)
12:06:57gartralas far as playlists go, i ussually just play a folder of music, or play party mode
12:07:13ajbkugel: Is that what you mean?
12:07:21pixelmaLlorean: now I'm even more confused, I would have expected it the other way round
12:07:39Lloreanpixelma: IIUC it *always* warns when your previous playlist came from the database.
12:07:58pixelmaseems I need to experiment
12:08:01ajbIf you then select another song after that you get a warning as the playlist is a fresh dynamic one.
12:08:17LloreanBasically, "Folders", if you don't modify them, aren't considered a dynamic playlist.
12:08:31LloreanBut database queries such as a whole album are considered dynamic.
12:08:36*gartral nods too Llorean, "thats what ive seen everytime i switch what im listening too
12:08:48ajbBut the menu then is "Save as /Playlists/dynamic/m38u", as... or No (Changes will be lost)
12:08:56kugelajb: no
12:09:18pixelmaI usually switch the warning off, that's why I have no idea what happens currently
12:09:21*ajb re-reads
12:10:11kugelthis was my main concern when it has been enabled by default
12:10:46LloreanPeople can still turn it off...
12:10:53ajbkugel: It might be easier if you just apply the patch and tell me what you think it should do. Certainly any loaded playlist will not trigger the warning unless it has been modified. I could change the logic to skip dynamic playlists if there are only one song and an unnamed playlist
12:11:09pixelmaLlorean: are you saying that if I still have a current playlist and then select a song in a folder, it will not warn me about creating a new playlist (and is this a folder in the file or database browser)?
12:11:37Lloreanpixelma: No. I'm saying "click select on a song in a folder. Then while listening, click select somewhere else and you won't get warned."
12:11:41kugelthis is the bug report on that one
12:11:51LloreanBecause it's just the one folder, with no modifications, it doesn't issue the warning.
12:12:42pixelmaah, in the same folder? Never tried that...
12:12:57Lloreanpixelma: Then, if you do the same thing with the database (go to database, browse to an album, click select on a song, then while playing click select on any other song) you'll always get a warning (because database queries aren't handled the same way folders are)
12:13:00LloreanAny folder.
12:13:47kugelpixelma: the link explains it. the "expected behavior" is what you get by file browsing
12:13:49LloreanThe trick is that if you don't use "Insert" (if you're just playing the contents of one folder, launched with a single click of select) it's assumed that you don't "lose" anything so there's no need for a warning - you never *really* created a dynamic playlist.
12:14:00Lloreankugel: Yes, but the problem is this is a bad expectation.
12:14:29Lloreankugel: The reason you don't get a warning with the filetree is that the filetree is considered "static". If it changes, it's because a user explicitly deleted a song from a folder. The database changes itself.
12:15:10LloreanThe playlist code will need to be able to differentiate what sort of query was used to create the playlist (was it a manual search? Based on Runtime? Or just an album) before it can reliably decide whether or not to warn.
12:15:18LloreanAlbums are not folders.
12:15:28pixelmahuh? Unless I'm still misunderstanding (which wouldn't surprise me) then I *think* that was different once but I really stop here until I tried
12:15:29ajbOk so th scenario as described in 9660 still occurs as playlist_modified(NULL) returns TRUE
12:15:30kugelit's bad expectation to expect the database do not behave different from the file browser?
12:15:49Lloreankugel: The database ISN'T a file browser
12:15:53LloreanWhy would you expect it to behave the same?
12:16:06LloreanIf you enter "Recently Played" it will be different nearly every time
12:16:18LloreanThat means a playlist from it will be different every time.
12:16:26kugelWhy should I expect (as a user), that if I plainly selected albums to play from, that this will issue the warning?
12:16:27LloreanIf you're happy with the one you've got now it should *certainly* warn before you launch other files.
12:16:43kugelsure the inner workings are different, but it's still a bug imo
12:16:48Lloreankugel: If you're selecting a single album, why can't you select it from filetree?
12:17:03LloreanIt's not a bug if it's working as intended...
12:17:11kugelI don't think it's working as intended
12:17:13LloreanYour desire for it to work differently is a feature request.
12:17:21Lloreankugel: It's working as the designer intended.
12:17:27LloreanYou don't get to define intended.
12:17:30ajbkugel: This is not really what I was trying to address in 9677, that simply tries to make the "About to erase dynamic playlist" more useful
12:17:47kugelHow do you know it's working as intended? I don't get your whole argumentation
12:17:53 Quit _lifeless (Remote closed the connection)
12:17:55ajbkugel: The logic about detecting if a playlist is modified is unchanged
12:18:04Lloreankugel: Until all the fancy logic's in to determine whether it's a static database query or a self-modifying one, it really should be safe and warn always, rather than never.
12:18:15kugelif I play an album in the database, that album doesn't change, no other albums don't change. so it should behave as if I selected a folder
12:18:20Lloreankugel: What do you mean "How do I know it's working as intended"?
12:18:30kugel"no other albums do change"*
12:18:49LloreanThe playlisting code _doesn't know_ you selected an album.
12:19:13LloreanPlaylisting is _entirely_ filename based.
12:19:29LloreanWhat you get out of the database, in the end, is a list of filenames.
12:19:55kugelyes, as you get when selecting a folder
12:20:05LloreanExcept when you select a folder, they're all "real"
12:20:09kugeland albums in the database are static
12:20:24kugelthe filenames are real in both cases
12:20:29LloreanBut the playlist code doesn't know it's an album.
12:20:39LloreanThere's no difference between an album and a genre and a recently played list.
12:20:42LloreanThey're all database queries.
12:21:57kugelthe point is, the warning is supposed to be issued when you manually changed the current playlist. and this isn't given by simply selecting songs in the database
12:21:58ajbIf I start playing an album from the File Browser and then select another one it doesn't trigger as playlist_modified(NULL) returns false.
12:22:25Lloreankugel: The warning is meant to be issued when you're erasing a dynamic playlist
12:22:28LloreanIt says NOTHING about manual changes.
12:22:36LloreanWhen you're playing a folder, you haven't really created a playlist.
12:22:44LloreanWhen you select a song in the database, a dynamic playlist is created.
12:23:42tmztwhat causes a playlist to be added to catalog? it's not automatic from what I can see
12:23:54 Quit ucchan ("Leaving...")
12:24:02Lloreantmzt: The playlist catalog isn't covered in the manual?
12:24:03ajbBut if you then "Insert" a track into a playing folder you then get a dynamic playlist
12:24:04kugelI see your point, but it's still a bug for me
12:24:10linuxstbThe problem (IMO) is that the database doesn't store the queries used to generate a playlist - it inserts all the tracks individually. The file browser does store the "query". Hopefully someone one day will be motivated to fix that...
12:24:15pixelmaI'm pretty sure this warning happened when I played a list from a folder and then "clicked" on a different song in a different folder. The playlist will be replaced, if it would "insert" the new tracks you would have a longer list consisting of tracks from both folders. Looks like it doesn't currently though
12:24:50*pixelma confused
12:24:52linuxstbI also think that from a user's point of view, the database and file browser should act the same
12:25:11Lloreanlinuxstb: It might be better to just always warn, then.
12:25:52Lloreanlinuxstb: For the database I'm not sure storing the query is good enough
12:26:01LloreanWhat do you do when the contents of the playlist don't match the query?
12:26:07tmztLlorean: I see you have to save it as a file, that requires giving it a name? I'm wondering because the automatic playlists aren't saved and the keyboard just causes them to be named .m3u8 or something like that
12:26:23 Join _lifeless [0] (n=lifeless@
12:26:26kugelLlorean: In this case, warn, I'd say
12:26:30linuxstbLlorean: What's the difference in that respect between the file browser and the database? Both will change if the user changes the files on the disk.
12:26:39Lloreankugel: But the playlist was never changed... the query results were
12:26:57Lloreanlinuxstb: The database will change even if the user *doesn't* though
12:27:08LloreanIt has all those runtime filters you can include in a query
12:27:27kugelif (last_query =! playlist_to_be_deleted) warn, else destroy
12:27:30LloreanIt's possible to create a playlist where you listen to a song, and then because you listened to it, the same query won't create that playlist again.
12:27:41ajbAside from the all these arguments about when a playlist is considered modified what about the action rockbox takes when it detects it is modified (however that might be).
12:28:11kugelthose runtime gathering based playlist are certainly a bit harder, but for simply selecting albums and stuff this should be sufficient
12:28:27Lloreankugel: Yes but they're the problem.
12:28:54kugelEvery selection issues the warning, not only the runtime gathering based ones
12:29:23kugelthose which are certainly static, albums and artist for example, shouldn't issue the warning
12:29:29LloreanYes, but you can't fix the static ones without fixing the runtime generated ones unless you're just saying "I don't care if I introduce a new bug, as long as it lets me use the queries *I* want to use without warnings"
12:30:10PaulJamkugel: but how should the playlist code know if the playlist was created from a static or a dynamic database query?
12:30:58kugelit could check if the previous query still gives the playlist which is going to be destroyed
12:31:17kugelor flag database selections, and remove the flag upon manually editting
12:31:37Lloreankugel: But that means it won't seem to behave like the file browser.
12:31:46LloreanYou can click select, not modify the playlist, and still get the warning because the query changed.
12:32:14LloreanIsn't it better just to tell people, up front, the database _doesn't_ work like the file browser, rather than trying to make it seem to then surprise them in odd ways when it doesn't?
12:32:29kugelI don't think so
12:32:56LloreanPeople have brains. They can understand "It's not a file browser."
12:33:01LloreanThe warning isn't going to kill them either.
12:33:38kugel"You can click select, not modify the playlist, and still get the warning because the query changed." I don't understand that
12:34:14n1sthe easiest fix would be to make it always warn for the file browser too (when the setting is enabled)
12:34:19kugelif you select something in the database, you could check if the previous query still matches for the playlist which is to be deleted, before doing stuff to it
12:34:41n1sin fact i don't see why we shouldn't just because it's only a single dir...
12:35:25Lloreann1s: I have no problem with that solution personally
12:35:29 Quit JdGordon (Read error: 104 (Connection reset by peer))
12:35:45kugelwell, that's surely possible and would band-aid it, but including removing the point of that warning imo
12:36:20Lloreankugel: My point is that it will _never_ work like the file browser. If you select a query, then listen, the query itself can change rather than the playlist. So the query won't match the playlist even though the user made no manual changes.
12:36:39LloreanSo the user didn't modify the dynamic playlist, but the playlist doesn't match.
12:36:43linuxstbSo currentlly, playing a directory isn't considered a "dynamic playlist" by the "warn when erasing dynamic playlist" option?
12:36:51Lloreanlinuxstb: Exactly
12:36:58linuxstbThen I agree, it should do.
12:37:10LloreanIf you just click select on a song it's not "dynamic" because it's assumed the folder itself serves as the playlist.
12:37:13pixelmaas I said, I think it did once
12:38:34 Join JdGordon [0] (
12:39:00kugelwhy not flag every playlist which is created by just selecting music as un-modified, then remove that flag upon every manual change, and then warn if the playlist has been modified
12:39:33kugelI haven't looked too much into the code, but from theory it sounds doable
12:40:05Lloreankugel: Because then you lose manual database queries.
12:40:25kugelthe flag would be set in the database as well
12:40:36LloreanI know...
12:40:52LloreanYou would lose manual queries where you generated a playlist by, say, searching for a certain term then clicking select on the first item
12:41:01kugelthen I didn't exactly get what you mean. As I said, I haven't looked into the respective code
12:41:10 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
12:41:37LloreanThe database isn't just "select a file in an album"
12:41:44kugelI know
12:41:49LloreanIt's also a quick and easy way to "Insert" a lot of songs that you'd have to browse around for otherwise
12:42:02LloreanIn _those_ cases it shouldn't act like it's unmodified even if you have just used "select" to start playback
12:42:56kugelhm, that's debatable. imo it shouldn't even warn in those cases
12:43:08LloreanWhy not?
12:43:11LloreanThey're not static at all.
12:43:27LloreanIf it shouldn't warn in those, it shouldn't warn if you insert several folders, because those folders aren't going anywhere...
12:43:41LloreanI understand you don't like the warning, but you CAN just turn it off.
12:43:58 Quit tajlero ("Leaving")
12:44:08kugelas I said, "and then warn if the playlist has been modified", that's not exactly what the warn on dynamic playlist is supposed to do, yes
12:44:26LloreanIt's supposed to warn you if you're about to lose a dynamic playlist
12:44:31kugelwell, I would really like to use the warning, but as of now, it's unusable for me
12:44:34LloreanIt's not called "warn on modified playlist"
12:44:42kugelYes, I know
12:45:02kugelpart of my suggestion was to warn on modified ones, not when deleting the dynamic
12:45:15LloreanSo you want to change what the feature does?
12:45:49kugelIn this regard yes. A modified one is always a dynamic one, the opposite isn't always the case
12:46:09LloreanYes, unmodified does not mean "not dynamic"
12:46:14LloreanAnd when what is modified, exactly? What is the "basis" upon which modifications happen?
12:46:52*kugel 's battery is empty
12:47:01LloreanJust saying "any time you use insert it's modified" means a lot of cases there won't be a warning when someone will lose the results of a database query. So we're making them go back through using the dynamic keyboard (which sucks) because we chose not to warn them.
12:47:01 Quit kugel ("ChatZilla 0.9.84-rdmsoft [XULRunner]")
12:47:53*Llorean really thinks "warning too often" is always preferable to "not warning enough and allowing easy data loss."
12:47:59LloreanThat's we we defaulted it to "on" in the first place
12:51:49 Join TheSphinX^ [0] (
12:51:56 Quit TheSphinX^ (Remote closed the connection)
12:52:33 Join gregzx [0] (
12:59:26*ajb ponders completion for the virtual keyboard....
13:05:21gartralajb, whats wrong with the vkb?
13:05:47kadobani like the kbd a ton more now that i switched the layout
13:05:53ajbgartral: slow, especially if you want to put a path in. It would be nice if -> cycled through completions
13:06:07Zagorit's horriby unintuitive. I'm not sure it can be intuitive, but that doesn't change the fact...
13:06:16ajbgartral: And the default keymap sucks... but thats just IMHO
13:06:20gartralahh, like tab in a terminal, yea, that would be nice
13:06:40gartrali agree there, may i post a pastebin of my default?
13:06:42LloreanZagor: I think if we just had text for "done" that you could highlight, and on-screen arrow keys it'd help tons.
13:06:42ajbI'm looking at the code now but it's a bit messy to follow
13:06:43kadobanon e200 the keymap for it already confuses me as well...i should change that
13:07:06LloreanZagor: Instead of using physical keys for backspace/navigation, onscreen virtual ones would work better for most people not familiar with the controls.
13:07:19ajbIs KBD_MORSE_INPUT used?
13:07:22ZagorLlorean: yeah. "done" and "cancel" virtual buttons would make it much better
13:07:35LloreanZagor: Done, Cancel, Left, Right, and Backspace I think
13:08:20ajbAhh I see, an iRiver feature
13:08:34Zagorpixelma: in the string
13:08:45kadobanajb: it used to be usable on different targets, i was using it on e200 for a while (that was a long time ago though)
13:09:13 Join pupuserd664a3 [0] (n=pupuserd@
13:09:16*ajb wonders why the code isn't split ... the #ifdefs make for very confusing reading
13:09:59ajbIs morse on top of the normal scroll & click vkbd actions?
13:10:02pixelmaZagor/ Llorean: and how do you access those "buttons" on the screen? (unless I'm misunderstanding again...
13:10:19Zagorpixelma: just like you access the characters.
13:10:33 Part pupuserd664a3
13:11:12Lloreanajb: There's some key combination for activating morse mode, I believe. It's probably in the manual.
13:12:41ajbhmm I think I'll leave this for now. The code could do with some TLC and refactoring to follow easily enough
13:12:53pixelmaZagor: among the characters or above? Btw. then I think left/right isn't needed if "line edit" mode is enabled everywhere, keeping the shortcuts for knowledgeable users and where possible
13:12:59 Join culture [0] (
13:13:46Zagorpixelma: above or below the characters, I guess. it probably depends on how much room there is on each target.
13:13:58ajbIs the function length so long to save stack for subroutines or just because of organic growth?
13:16:08gartralwould it be possible too display a second warning before saving changes too a text file if its a "system" file? i know it seems redundant, but it would be useful to point out the fact that "version.txt is a system file, and changing it may cause undisired results" or something?
13:17:41LloreanThe .rockbox folder's already hidden.
13:17:58gartralbut version.txt is in the root
13:18:09LloreanNot normally.
13:18:11pixelmaversion.txt is not a system file (of Rockbox)
13:18:13linuxstbpixelma: They should/could be positionable via the .kbd file. e.g. using codes like %x (%% to display a real % symbol)
13:18:27LloreanOr, I'm thinking rockbox-info.txt
13:18:37gartraland changing it does cause rbutil to not read the device correctly
13:19:32pixelmaI thought the Rockbox Utility relied on the USB IDs?
13:20:06pixelmalinuxstb: I thought the topic was the default keyboard
13:20:21Lloreanpixelma: It's more the difficulty in that screen in general
13:20:32LloreanNo matter what the keyboard is, right now you're also constrained by the keymap
13:20:49LloreanSome virtual control keys on the keyboard could help with this a lot
13:21:10 Join Darksair [0] (n=user@
13:21:12gartralit does as a secondary action if it cant find version.txt, but the current (m1.0.9) reports no sansa found if it cant find the version.txt file
13:21:24linuxstbpixelma: Isn't that displayed the same way - i.e. with a built-in default .kbd file?
13:21:28gartral\(with Sansas, obviously, :))
13:22:17pixelmalinuxstb: no idea
13:22:45pixelmaLlorean: what about line edit mode then?
13:23:19linuxstbpixelma: Looking at the code, it is - see around line 310 of apps/recorder/keyboard.c (and later for different LCD sizes)
13:23:38pixelmait would seem the same to me - finding out how to access some left/right button or how to access the input line
13:23:52Lloreanpixelma: I don't think the virtual keys are mutually exclusive.
13:24:24LloreanHaving the controls visible takes some of the guessing out of it, plus we can keep all the old controls for faster/easier use if you know what you're doing.
13:24:51pixelmathinking about screen estate
13:30:54kadobanis there some way i can tell how deep my recursion depth can go on various targets? (ballpark will probably do)
13:31:03kadoban(in a plugin, if it matters)
13:31:24pixelmaI'm just talking about left or right "virtual" buttons here, if that wasn't clear enough
13:34:03 Quit snether ("CGI:IRC")
13:37:27 Quit fredddy (Read error: 110 (Connection timed out))
13:37:32gartralanother hurtle too consider is how will the users construct there own custom vkbs with these buttons, as it is, makeing your own vkb isn't hard at all
13:38:06 Join fredddy [0] (
13:38:29kadobanif the virtual buttons were optional, even if they messed them up, they just wouldn't have virtual buttons
13:45:17gartral <- this my layout, also, it would be nice if we could have a way to "save" a custom layout over the (imho) bulky default
13:46:45PaulJamwhat do you mean by "save"? do you mean to keep the layout across reboots?
13:47:19PaulJamput it into the .rockbox directory.
13:47:44 Join robin0800 [0] (
13:50:50***Saving seen data "./dancer.seen"
13:50:54 Quit robin0800 (Remote closed the connection)
13:51:12 Join robin0800 [0] (
13:59:47gartraldoes the name matter?
14:01:12 Join blkhawk- [0] (
14:01:55 Quit blkhawk (Broken pipe)
14:02:36linuxstbgartral: No, but it probably needs to be relatively short (less than 32 chars I am guessing)
14:06:42kadobanis there some way to change the title of a flyspray entry?
14:08:32 Join kugel [0] (n=chatzill@unaffiliated/kugel)
14:08:39gartralumm, i want too test the patch in fs#8663 and post if it works in, could i have wiki right permission? my realname is Gareth Schakel
14:09:03gartraler, i guess i should say "how it works"
14:09:12 Quit robin0800 (Remote closed the connection)
14:10:30 Join massiveH [0] (
14:11:36 Join pupuserd664a3 [0] (n=pupuserd@
14:12:28 Part pupuserd664a3
14:13:39kugelLlorean: I thought a bit about the previous discussion during my downtime
14:14:57kugelbasically, if every playlist, which is formed by selecting anything but a playlist file (i.e. .m3u[8]), then yes, the file browser should indeed warn too
14:15:21*kugel adds a "is considered as dynamic" where appropriate
14:16:20 Quit massiveH ("Leaving")
14:16:57kugelbut personally, I'd like to get warned only if I made manual changes, even if it was such a dynamic playlist. So, my conclusion is that there should be an additional setting, like "warn on deleting current playlists: if dynamic; if modifed; never" (maybe a always option too)
14:17:49 Join itcheg [0] (i=41d59de2@gateway/web/ajax/
14:18:08kugeland yes, the current behavior should at least be made consistent, such as warn in the file browser too
14:18:53*Llorean thinks the database needs to be less like the filetree, not more.
14:19:04LloreanSo people don't immediately think it's a bug when "Follow Playlist" doesn't work, etc.
14:19:24 Join tyfoo [0] (
14:19:34LloreanIf there were a clearer distinction between them, the differences in behaviour would seem less like bugs and more like simple differences between the two.
14:20:46gartralok, fullspeed USB too a 4gig transcend mSDHC class 6 card works great, yet too test a USB highspeed build
14:21:48kugelJdGordon: ping
14:25:06linuxstbLlorean: I would want the opposite - things like "follow playlist" to work in the database...
14:25:16Lloreanlinuxstb: That's more or less impossible though.
14:25:26LloreanMost songs don't even have a unique database entry.
14:25:32LloreanThey usually show up several times.
14:25:59kugelwell, depends on how you save the query
14:26:04Lloreankugel: Not really
14:26:12LloreanThe query your playlist was generated from may not even exist any more
14:26:26LloreanFor example if it came from "unplayed songs", by the time you try to "follow" it it may not exist in unplayed songs any more
14:26:27 Join robin0800 [0] (
14:26:52kugelI don't say safe the query in the playlist. just save it, keep it in memory or whatever
14:26:58LloreanThe most consistent way you could do it is "even if started from the database, follow playlist takes you to the real file" since that's always unique
14:27:08Lloreankugel: It doesn't matter where you save it.
14:27:19LloreanThe results on a query won't always be the same every time that query is run.
14:27:57kugelI just want to say it's basically possible, sure it needs refinement for such (rare?) conditions
14:28:06LloreanIt's impossible to always work.
14:28:13LloreanThe best you can get is "usually"
14:28:24kugelwell, that would be enough imo
14:28:44LloreanBecause unexpected behaviour is fine?
14:28:55LloreanWouldn't it be better to always _actually_ follow the playlist, so there's never any surprise?
14:29:02 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
14:29:25 Nick blkhawk- is now known as blkhawk (
14:29:46LloreanI don't see the sense in doing it a way that is sometimes broken and sometime fine, when there's a perfectly viable "always fine" alternative.
14:29:55kugelI don't think it's necessarily unexpected if the query isn't possible anymore, and if, just print a warning that because of blah following the playlist failed
14:30:16kugelalways broken, rather
14:30:26 Join TheSphinX^ [0] (
14:30:34LloreanAnd do you plan to write the database queries in the file, or will "follow playlist" only work on the database if dynamic, but if it's a saved playlist, silently go back to following files?
14:31:32linuxstbLlorean: You simply take the user to the list in the database that the track was added from
14:31:32kugelit's all depends of how much of a database-lover one is. I am such database-lover, which means I'm very happy if a feature that works perfectly for the file browser would work in 95% of the time in the database too
14:31:45 Quit TheSphinX^ (Remote closed the connection)
14:32:08Lloreanlinuxstb: So if a track was added from a list, but isn't in the list any more, we show them the list that doesn't have their track in it?
14:32:30 Join TheSphinX^ [0] (
14:32:33kugelthe node above would be sufficient
14:32:42Lloreankugel: How is that "following" then?
14:32:43kugelplus the warning, that following failed
14:32:51LloreanThe whole point of "follow playlist" is to take you to the song playing
14:32:55LloreanIf you get somewhere else, it's broken
14:33:01linuxstbLlorean: We simply get rid of dynamic tags ;)
14:33:03LloreanFollowing _never_ needs to fail
14:33:08Lloreanjust take them to the file
14:33:42kugelYou don't use the database much, do you?
14:34:21LloreanI don't understand your line of logic. "I like the database, so we should use it even when it's unnecessary and creates unpredictable behaviour."
14:34:40kugelit's surely not unpredictable
14:34:45jhMikeSamiconn: I was under the impression only c200/e200 needed that method and then made aware it wasn't as simple as that. So, I'm working on just making standard for the USB stack.
14:34:53Lloreankugel: It is for the user until they've seen it fail a few times in different ways
14:35:16kugelif you browse and select something in "never played", it's very predictable that following won't quite work in this case
14:35:46 Quit TheSphinX^ (Remote closed the connection)
14:35:58Lloreankugel: But it could be predictable
14:36:00 Join TheSphinX^ [0] (
14:36:02LloreanWhat's _wrong_ with showing them the file?
14:36:29linuxstbjhMikeS: Also, do I understand correctly that the new check disabled in the bootloaders? I thought the e200/c200 did USB detection in the bootloaders.
14:36:31kugelwell, simply that it leads to the filebrowser, for people which probably expect to land in the database
14:36:47*kugel would expect that too
14:37:11Lloreankugel: But it'd be 100% consistent. You'd only need to learn a single quirk, instead of several.
14:37:16kugelbut that's not a feature that I'm looking forward to, since I imagine it will introduce some complexity, which isn't worth it
14:37:20jhMikeSlinuxstb: It is disabled there since the full driver isn't available it only ever checked the pin state and rebooted.
14:37:50kugelyes, I learned to live with it, and turned follow playlist off
14:41:53 Part B4gder
14:52:35 Join fragilematter [0] (n=barbu_do@
14:52:58JdGordonkugel: sup?
14:53:29kugelJdGordon: I'd like to finalize custom list vp soonish
14:53:46JdGordonas would I
14:54:04*kugel thinks it rotted long enough :p
14:54:51kugelI have two questions: how to handle screens that aren't exactly lists, but would probably wanted to have this parent too?
14:55:35kugeland, how to parse the custom list, either via your viewport manager thing (such as viewportmanager_set_customlist), or via a parameter in viewport_set_defaults
14:56:06 Quit culture (Read error: 110 (Connection timed out))
14:56:24 Quit AndyI (Read error: 104 (Connection reset by peer))
14:57:23JdGordonwell... firstly, I've changed my mind about this... (well changed ages ago but we ahvnt talked about it for a while...) I'm more inclined towards the WPS/theme setting aside a viewport which the rest of the UI uses, so its not the lists which are getting the viewport.. its the whole "screen"
14:57:32kugelin it's current form, it's basically easily possible to have arbitary custom parents, such as "quickscreen vp" in the settings, and use that for the quickscreen
14:58:06 Join AndyI [0] (i=AndyI@
14:58:07kugelor, fall back to list parent for every screen that shows the backdrop (for simplicity)
14:58:39kugelwhich would be the same as a "rest of UI parent"
14:59:05JdGordonI dont see the need to ever specify more than 1 viewport outside of the WPS
14:59:11kugelJdGordon: so, you basically say now what I always said?
14:59:35JdGordonthe only possible exception is if we take things further and have context menus/screens a seperate one
14:59:35 Join LambdaCalculus37 [0] (i=44a04303@rockbox/staff/LambdaCalculus37)
14:59:44*JdGordon doesnt remember what you've always been saying
15:00:03 Part LinusN
15:00:38kugelwhen we talked about it the last time you said every screen should be able to have its own parent, and I said that would be unneeded complexity, the list parent would be suitable for all
15:01:53kugelanyway, I'm fine with having only 1 parent for anything but wps
15:02:26kugelthat's what the current version already does, it would basically only need sync if we keep it like that
15:02:43JdGordonwait a sec... I'm not saying that screens shouldnt be setting up a parent viewport to work in...
15:03:50kugelscreens as in "rec screen, quickscreen etc" or as in "main and remote"
15:03:57 Join Rondom [0] (
15:04:40JdGordonthe first
15:04:43kugelhow did you exactly mean "a viewport which the rest of the UI uses, so its not the lists which are getting the viewport.. its the whole "screen""
15:04:49 Part fragilematter
15:05:42kugelthat sounds to me like the e.g. quickscreen should use the "list" viewport too
15:06:50JdGordonmy rough idea is bassically this.... the theme/wps specifies a viewport which is stored in viewport.c and gets returned by viewport_set_Defaults() which EVERTHING MUST use
15:07:09JdGordonso yeah, a bit of svn needs changing to ignore parents and just use this
15:07:30JdGordonlist is the exception because screens like time/date would need to specify where the list goes
15:07:37kugeloh, well, that's possible too
15:08:07kugelbut I imagine that wouldn't work so well for plugins, or keyboard screen
15:08:27JdGordonplugins are a different beast
15:08:54JdGordonand it depends on the theme if it would actually "work" or not
15:09:05LloreanThe vkeyboard should probably define its own viewport in the .kbd file
15:09:17LloreanSo that the keyboard can basically say "this is how much space I _must_ have"
15:09:20JdGordoni.e if the theme bassically just replaces the top statusbar with something fancy and uses the remaining height for the list there would be no change
15:09:30JdGordonbut if it uses say the AA rectangle, then things could look crap
15:09:59*JdGordon doesn't like the vkeyboard at all
15:10:12*Llorean is not the vkeyboard's biggest fan.
15:10:26kugelJdGordon: so we could maybe have a bool force_fullscreen in viewport_set_defaults for screens (and plugins!) which are meant to be fullscreen?
15:10:30*gartral likes his custom one
15:10:48JdGordongartral: I dont like the system... not the layout
15:10:51kugeland the wps of course, if that's using viewport_set_defaults
15:10:59JdGordonkugel: well yeah... bassically whats done now to disable the statusbar
15:11:22kugelI'd rather have it as a parameter though ;)
15:11:29JdGordonthis is why I wanted to get that stuff done and working before talking about custom list again
15:11:47kugelI'm back in a few minutes
15:11:52 Quit kugel (Remote closed the connection)
15:19:54 Join nplus [0] (n=nplus@243.131.Globcom.Net)
15:21:27 Join archivator [0] (i=foobar@
15:21:29 Quit itcheg (" ajax IRC Client")
15:21:36 Quit yhuang (Read error: 60 (Operation timed out))
15:21:42 Join itcheg [0] (i=41d59de2@gateway/web/ajax/
15:22:50 Quit Llorean ("Leaving.")
15:23:33scorchea new emulator seems to possibly be coming to fruit...a snippet of text from an ipodlinux channel:
15:27:04 Join kugel [0] (n=chatzill@unaffiliated/kugel)
15:27:11*LambdaCalculus37 thinks that information should be kept by a few individuals here, in case that pastebin expires
15:27:25*kugel is back
15:28:03 Quit MTee (Read error: 110 (Connection timed out))
15:28:28kugelJdGordon: so would rather have it like viewportmanager_set_customlist(bool yes) instead of a force_fullscreen parameter for viewport_set_defaults?
15:30:07JdGordonkugel: no, for 2 reasons... 1) it should be a PITA for screens to not use the specified viewport, and 2) compiled-code-wise.. thats basically wasting a param which 90% of the time would be false
15:30:19JdGordonso, by no.. i mean yes...
15:31:08kugelok, I think i can do it
15:31:54kugelthat means fullscreen screens are responsible for setting the custom viewport again upon leaving, or every screen makes sure on its own which parent to use?
15:33:35JdGordonyes, screens should always assume the "fullscreen" viewport they get is the one specified by whatever... not literally the full lcd
15:34:02kugelbtw, I think there should be 1 global viewport which holds the custom dimensions, and every screen draws into this one (via display->set_viewport() in viewportmanager_set_customlist()), instead of having it to reparse again and again
15:35:12kugelwhich would render viewport_set_defaults rather useless
15:37:34kugelJdGordon: opinion on that?
15:37:36JdGordonwhat reparseing? viewport_set_defaults does copies... of course we dont store a string and reparse that every time....
15:38:46kugelwell, but we could also set viewportmanager_set_customlist do set_viewport() globally (either the custom or the fullscreen one)
15:39:15kugelhm, no, that wouldn't work so well I suppose
15:39:40 Quit CaptainKewl (Remote closed the connection)
15:44:12 Quit TheSphinX^ ("XChat@Linux")
15:44:32 Join TheSphinX^ [0] (
15:45:38 Quit Seed ("cu, Andre")
15:48:29 Quit ajb (Read error: 110 (Connection timed out))
15:50:53***Saving seen data "./dancer.seen"
15:52:26 Join {phoenix} [0] (
15:54:39 Nick JdGordon is now known as JdGordon|zzz (n=jonno@rockbox/developer/JdGordon)
15:56:16scorcheLambdaCalculus37: it is set to expire in a year, but it is in mine and likely others' logs forever...
15:56:18 Join MTee [0] (n=mtee@
16:04:23 Join Llorean [0] (n=llorean@rockbox/administrator/Llorean)
16:11:36 Quit TheSphinX^ ("XChat@Linux")
16:11:51 Join TheSphinX^ [0] (
16:11:54 Quit TheSphinX^ (Read error: 54 (Connection reset by peer))
16:16:16 Join TheSphinX^ [0] (
16:19:21 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
16:19:44 Quit Rondom (Nick collision from services.)
16:19:48saratogasscorche: has that guy working on the emulator seen the rockbox emulator Toni started?
16:20:00 Join Rondom [0] (
16:20:06saratogai believe he'd implemented some of the PP hardware already
16:23:07scorche|shsaratoga: no idea...i shall have to point him to it though
16:25:19saratogaToni's could actually run the good bit of the Sansa bootloader, so I figure he must have implemented some of the hardware
16:30:42 Join toffe82 [0] (
16:32:16jhMikeSwhat are these #ifdef STUB things in crt0-pp.S? It is it some atavistic IPL thing? :p
16:32:51 Quit Llorean ("Leaving.")
16:35:20 Join Llorean [0] (
16:38:25 Quit linuxstb (Remote closed the connection)
16:39:44 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
16:43:05 Join evilnick [0] (i=0c140464@gateway/web/ajax/
16:45:09 Quit nplus (Remote closed the connection)
16:45:26 Join webguest06 [0] (n=48be7505@gateway/web/cgi-irc/
16:46:09 Quit webguest06 (Client Quit)
16:46:11 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
16:46:37 Join KennW [0] (n=48be7505@gateway/web/cgi-irc/
16:47:15KennWHey guys,so what's going on atm?
16:48:55 Join webguest97 [0] (n=5898ab6d@gateway/web/cgi-irc/
16:48:56scorche|shKennW: i am not sure if you are aware, but this is a technical channel -not a social might want to read the topic
16:51:23 Join yhuang [0] (
16:52:56 Join rolox [0] (n=rolox@
16:53:22 Join jgarvey [0] (
16:56:36 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
16:57:25KennWWhere can the topic be read?
16:58:35LambdaCalculus37Try typing /topic.
16:58:54 Quit webguest97 ("CGI:IRC")
17:01:39 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
17:02:34 Join edrz [0] (
17:02:44 Join culture [0] (
17:03:41 Quit Zagor ("Client exiting")
17:05:37 Quit KennW ("CGI:IRC (EOF)")
17:06:38 Quit kugel (Remote closed the connection)
17:07:00 Join flydutch [0] (
17:10:23 Quit FlynDice (Remote closed the connection)
17:13:19 Quit robin0800 (Remote closed the connection)
17:14:58 Quit rolox (Read error: 110 (Connection timed out))
17:17:25 Quit nibbler (Read error: 113 (No route to host))
17:20:00 Part gartral
17:25:35 Quit bmbl ("Woah!")
17:30:53 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
17:39:02 Quit Nico_P (Remote closed the connection)
17:44:27 Quit tyfoo (Read error: 104 (Connection reset by peer))
17:47:47 Join FlynDice [0] (
17:49:14 Quit jgarvey ("Leaving")
17:50:25 Quit saratoga ("CGI:IRC (Ping timeout)")
17:50:56***Saving seen data "./dancer.seen"
17:52:27 Join nibbler [0] (
17:57:38 Join z35 [0] (
17:58:56 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
18:04:53 Nick japc__ is now known as japc (
18:05:03 Quit japc ("Ex-Chat")
18:05:25 Join japc [0] (
18:07:29 Join Willwolfe [0] (
18:14:27 Join kugel [0] (n=chatzill@unaffiliated/kugel)
18:15:48WillwolfeDoes the bootloader change the workings of rockbox or simply load it?
18:15:51 Quit petur ("now sports")
18:17:54 Join faemir [0] (
18:18:09LloreanWillwolfe: The bootloader initializes hardware.
18:18:16LloreanThis can have an effect on things.
18:19:00 Join Comokanu [0] (n=5c0ce2b3@gateway/web/cgi-irc/
18:21:08WillwolfeSo a rebuild of the bootloader may or may not be necessary in subsequent builds depending on whether commits have refered to hardware? I am guessing it should be built every time as well to ensure compatibility?
18:21:09 Quit Comokanu (Client Quit)
18:21:39LloreanWe release binary bootloader builds
18:22:09LloreanThere should generally be no reason to custom build your own bootloader unless you've personally made a change to the bootloader, or unless we mention in the changelog new bootloaders will be necessary.
18:25:02 Join sarixe` [0] (
18:25:15WillwolfeBy release do you mean through svn or through the rockbox website?
18:25:46WillwolfeI was referring to build using the configure tool for the current revision?
18:26:01WillwolfeCertainly not making my own changes.
18:26:48 Quit sarixe (Nick collision from services.)
18:26:55 Nick sarixe` is now known as sarixe (
18:27:21Unhelpfulthe bootloader doesn't change in substantial ways very often
18:28:09LloreanWillwolfe: Through the website / utility.
18:28:17LloreanYou should really never compile your own bootloader
18:32:09 Join QuickStart [0] (
18:32:23QuickStartcan I ask a quick question?
18:33:06LloreanQuickStart: We actually ask that you go ahead and ask, rather than asking to ask, as long as it's a Rockbox question
18:33:39QuickStartwhat is the bmp restricitions for creating a theme for a sansa e200 series
18:33:47QuickStartother then resolution
18:34:24LloreanA total number of pixels approximately equal to a bit more than twice the total number of pixels on the screen
18:35:04QuickStartso resizing a png and saving it as a bmp won't do it?
18:35:47LloreanI'm not sure I see your reasoning, unless the small bitmap is still much larger than the screen...
18:37:02QuickStarthere is the situtation
18:37:20QuickStartthe png that the bmp is based of is 1200X1600
18:37:34LloreanThe only file that matters is the actual bitmap.
18:37:38QuickStartI resized it to 176x220 the size of the screen and it won't show
18:37:48QuickStarteven after I saved it as a bmp
18:38:02LloreanAre you using it as a backdrop or an actual image in the WPS?
18:38:18QuickStartbackdrop for the theme
18:38:23QuickStartwps hasn't been done yet
18:38:42QuickStartbackdrop coding is as follows
18:38:46 Join einhirn [0] (
18:38:56LloreanTriple check that it's actually 176x220. Some programs occasionally save them one pixel too large/small, or attempt to maintain the aspect ratio giving you an imperfect size
18:38:56QuickStartbackdrop /.rockbox/backdrops/morning.bmp
18:39:18QuickStarttwo things with that
18:39:37QuickStartone the picture is landscape and saves it 176x180ish
18:39:44QuickStartso I flip it and nothing
18:40:00PaulJamthere is a ":" missing in the config entry
18:40:01QuickStartit does show up I tried a jpg to test
18:40:02LloreanIt needs to be _exactly_ 176x220
18:40:14QuickStartah man
18:40:17LloreanYou'll need to add in extra padding data.
18:40:31QuickStartits just that the image is alittle weird
18:40:40QuickStartit looks better in landscape
18:41:38QuickStartand to extend it if I were to cut it it would be difficult to recreate
18:43:09LloreanHonestly, none of that is really relevant. The restrictions on backdrops are that they're exactly, pixel-for-pixel the shape of the screen.
18:43:29QuickStartso one slip and no good I see
18:43:34QuickStartman what a waste
18:44:06kadobanQuickStart: i'm not sure what you'd like rockbox to do in that case...when it displays it would need to add borders or stretch the image anyway
18:44:12rasherI don't understand. Why wouldn't you resize&crop on your computer where you can control what happens
18:44:38QuickStartyeah its just that I want to make a theme base for a os that I represent
18:45:12QuickStartand to maintain the official wallpaper in its entirety instead of cutting it and doodling to fill in the rest
18:45:19LloreanQuickStart: What they're saying is "to display properly it would need padding anyway, because it doesn't match the size of the screen, so why not pad it on your computer so the padding is a color, and position, that you like"?
18:45:34rasherQuickStart: To do that you'd have to change the physical size of the display...
18:45:47QuickStartoh ok got you
18:46:25 Part Llorean
18:47:09QuickStartjust to be clear on one thing where can I get a list of codes acceptable by the theme .cfg file
18:47:22QuickStartI want to see to what extent I can go
18:50:04QuickStartis their a link I know where I can get the WPS coding but the cfg I have no idea
18:51:12 Join bertrik [0] (
18:51:55dionoeaQuickStart: the same place as the WPS would be a good place to start looking
18:52:04rasherQuickStart: ThemeGuidelines lists some examples
18:52:34QuickStartahh ok I understand
18:52:37QuickStartone more question
18:52:45rasherAnd the manual has a list of all cfg options I believe, obvoiusly not all of those apply to a theme
18:52:54QuickStartis is possible to make an xml to change the backdrop a different times of the day
18:53:09QuickStartI'll check the manual then
18:53:27dionoeatime dependant backdrops would be cool :) (but totaly useless)
18:53:39 Join xSlack [0] (
18:54:08QuickStartits just that my OS has the feature for the backdrop
18:54:20rasherI guess you could make a TSR plugin to do it, but eh..
18:55:17QuickStartum can you elaborate on it
18:55:51rasherYou'd need to write a plugin that would run in the background and change the backdrop.
18:56:11QuickStartso it is possible theoretically?
18:56:27 Join Thundercloud [0] (
18:56:28rasherI think so
18:56:29kadobanQuickStart: little is impossible if you're willing to spend the time
18:56:46rasherBut it'd break if the user started another plugin
18:57:08QuickStartI guess it has to be written into the coding for the next release
18:58:02 Quit Horscht ("Verlassend")
18:59:16rasherDon't hold your breath
19:01:09 Join miepchen^schlaf [0] (
19:02:28 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
19:02:33QuickStartits just I have a couple of suggestions that would make the system more home owner friendly
19:04:10linuxstb"home owner friendly" ?
19:05:20*Unhelpful has thought that at least a "make backdrop" plugin that could scale, etc, might have some users
19:08:58QuickStartwell an alarm clock function and of course the tsr plugin and visualization funtion
19:10:43 Join robin0800 [0] (
19:10:43BigBambiWhat has home owning to do with that?
19:10:50rasherRockbox has an alarm feature already.
19:10:56BigBambiAlso, some targets have an alarm
19:11:33 Join Seed [0] (
19:13:24PaulJamsome of the demo plugins could be considered visualisations
19:13:48Unhelpfulindeed, we could almost make a category for them.
19:15:41 Join kkurbjun [0] (n=karlk@
19:16:20kkurbjunkugel: you are having trouble getting the arm9 mmu running?
19:16:36Unhelpfulamiconn: if you have the time, at some point, i think that DIV255 might be faster using umull on arm than the shift-add sequence
19:20:05amiconnSH1? That's not yet used, is it?
19:20:55Unhelpfulonly in test_greylib_bitmap_scale
19:21:25QuickStartbut could they be used in tandum with other plugins?
19:21:59Unhelpfuli'm hoping that at least the 32x32->32 multiply is reasonably speedy on sh-1? i don't think i can go lower without reworking a lot of other things, possibly in ways that are slower than now
19:22:02amiconnOh, and CPU_SH doesn't exist, as there's only a single CPU of that family
19:22:25Unhelpfulah! what should i test against?
19:22:58UnhelpfulCONFIG_CPU == SH7034?
19:23:26amiconnMaybe we should introdice CPU_SH for completeness...
19:24:05 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
19:24:21Unhelpfulit wouldn't hurt, but i've committed the correction :)
19:24:28amiconnAlthough that'd be a bit too generic; there is SH1, SH2, SH-DSP, SH3, SH4 and SH5 afaik.
19:25:13amiconnSH2 already has a better multiplier (offering 32*32->32, as well as a full 64 bit accumulator for MAC)
19:25:19 Join gartral [0] (
19:25:24kugelkkurbjun: indeed
19:25:50amiconnUnhelpful: Regarding DIV255 - did you check the asm?
19:25:55kugelkkurbjun: may I request some assistance? :)
19:26:22kkurbjunsure, do you have a diff of what you are currently doing? You are just getting stuck loading the main build from the bootloader?
19:26:37gartralthe RoLo page in Wiki contains 4 links too itself, i think this is slightly confusing
19:26:42 Join Llorean [0] (
19:26:51kkurbjunthe gigabeat F/X has the MMU enabled for both the bootloader and main biuld
19:26:53kugelkkurbjun: correct
19:27:26kugelI tried enabling in both and only the main, both didn't boot
19:27:32 Join Ewidiusz [0] (
19:27:44kkurbjunthe crt0.s code relies on being started from the non-virtual mapping.
19:28:04kkurbjunand it then enables the MMU and copies itself so that it is located at address 0
19:28:31 Quit archivator ("<>")
19:28:37amiconnIf gcc is clever, this needs just 3 instructions, and takes less cycles than umull
19:29:20kugelkkurbjun: from what I've noticed, the F/X procedure is supposed to be the same
19:29:47kkurbjunone thing to make sure of is how your linker script is setup
19:30:03kkurbjunbecuase the F/X does it different than the other targets
19:30:13kugelthe differences between ams and fx are: we use the IRAM, the bootloader is actually completely in IRAM (and iram is mapped to 0x0), and we don't map SDRAM to 0x0
19:30:13kkurbjunyou have to have the vectors at address 0
19:30:20QuickStartwell thanks for the help guys
19:30:36QuickStartI'll be working hard and let you know if I can or will release the theme
19:30:50kkurbjunthat may be some of the problem, I never could figure out iram on the FX, after the MMU was enabled I couldn't read or write to it
19:31:34 Nick Ewidiusz is now known as VVlodarrius (
19:31:39kugelwe basically must use iram
19:32:18amiconnkkurbjun: I'd suspect a mistake in the MMU setup
19:32:20kugelthe ROM bootloader of the ams3525 copies the bootloader into the iram (I don't know whether this can be changed), plus on lowmem codecs are in iram
19:34:11 Join zicho [0] (
19:34:22kugelso, unless I got the mapping thing wrong, I don't actually see where I'm doing things wrong
19:34:51kugelI flushed&invalidated caches before trying to boot the main binary
19:35:53zichois ipod Video the player referred to as ipod 5.5?
19:36:09kugelkkurbjun: does SDRAM have to be mapped to 0x0?
19:36:12krazykitzicho, the ipod video is the 5g and 5.5g
19:36:37kkurbjunkugel, I don't have access to my code tree at the moment, but my current suspicion is a problem with the linker setup and the vector code location/what/where it was linked to, I should be able to check it in more detail later.. It doesn't have to depending on how your linker is setup.
19:36:50zichookay, good. does rockbox run well on that machine? is there some sort of list of what features of rockbox that are available on it?
19:37:26kkurbjunamiconn: that could be possible, on the F/X it was so small that it didn't really drive me to figure it out
19:37:31kugelkkurbjun: that could be, I pasted what I did to the linker scripts (mostly only - TTB_SIZE), linker scripts are still some kind of black magic to me
19:37:39 Quit Darksair ("Emacs = ESC-Meta-Alt-Ctrl-Shift")
19:39:20amiconnkkurbjun: Yeah, IRAM is smaller than the cache on that Soc (4KB vs. 2*16KB). On AS3525 it's the opposite: 320KB IRAM, 2*8KB cache
19:39:57amiconnHmm, or was it 16KB?
19:40:08kugelno, you're right
19:41:41kkurbjunkugel, I'll try to take a look at that diff with respect to the full linker script and the code in crt0.s later today.. What address is the iram located at?
19:41:55 Quit VVlodarrius ("Ex-Chat")
19:42:12kkurbjunis that a register that you set to remap it, or is it always there?
19:42:22kugelsdram is at 0x30000000 (as per as3525 datasheet)
19:42:51kugelno, the iram isn't remapped explicitly
19:43:13kugelso, it's either the physical address, or the ROM bootloader remaps it
19:43:19kkurbjunthe other thing that you might want to check is an objdump on rockbox.elf to make sure that the stack is not used when calling the mmu functions from crt0.s
19:43:22zichois there any list of what rockbox features that works on a specific machine?
19:43:22kugel(I'm not exactly sure)
19:43:26 Quit MTee (Read error: 104 (Connection reset by peer))
19:44:02kkurbjunI made sure that they wern't when compiled for the FX with our recommended compilers, and I don't see why it would necessarily change, but it could be a problem if it is using the stack
19:44:05kugelkkurbjun: I also tried using C code in system_init()
19:44:14 Join tyfoo [0] (
19:44:21kkurbjunyeah, you definately can't call that C code in crt0.s
19:44:36kkurbjunsince it has more than 1 level of function calls it uses the stack
19:44:41kugeldoesn't the fx do that?
19:44:53kugelsuch as bl enable_mmu?
19:45:26kkurbjunyeah, each of those functions are setup so that they only use the scratch registers
19:46:06kugelwell, I used those too
19:46:07kkurbjunthat's why one of those calls defines at least one variable later
19:46:30kkurbjunyeah, I wouldn't expect it to change, but just something to verify
19:46:31PaulJamzicho: this page lists some features:
19:47:17kkurbjunlike I said though, I'll try and take a look at it later, what target is this for?
19:47:22PaulJamzicho: more detailed information can probably be found in the manual for the specific target.
19:47:27zichoPaulJam, thanks. But doesnt ipod video support USB-connection?
19:47:31edrzi'm attempting to install rockbox on a sansa e250r. can someone confirm that this should be possible?
19:47:36kugelmy target is fuze
19:47:46kkurbjunok, cool
19:47:56kugelbut it should apply to all ams sansas (clip, m200v4, e200v2)
19:49:45PaulJamzicho: i thnink you still need the OF for USB transfers, but i'm not 100% sure.
19:50:04kugelkkurbjun: thanks for your help, I'm basically clueless
19:50:35zichoat "TargetStatus" it says "USB: NO" on the video, but in device chart it says "USB: 2.0"
19:50:58***Saving seen data "./dancer.seen"
19:51:26kugelkkurbjun: if it matters, enabling icache definitely works
19:51:50 Quit faemir (Remote closed the connection)
19:51:51kadobani believe TargetStatus is talking about what rockbox can do, while the device chart is listing what the hardware is capable of
19:52:04gevaertszicho: DeviceChart talks about the hardwars. TargetStatus talks about rockbox status
19:52:06 Join faemir [0] (
19:52:33zichogevaerts, so in other words, the video doesnt support USB-support with rockbox?
19:53:42gevaertszicho: that's more or less it, yes. It's not a big loss though, on USB connection rockbox reboots to the Apple ROM USB mode, and when you disconnect it boots back to rockbox
19:54:22zichoah, well then. so video is a good player for running rockbox?
19:54:59gevaertsIt's certainly not a bad player, although mpeg video performance in rockbox isn't great on the video
19:55:36PaulJamgevaerts: but for that you can always boot the OF.
19:55:40zichohow come?
19:55:48gevaertsPaulJam: true
19:56:38*BigBambi would pick a few players first over the video
19:56:54gevaertsBecause the ipod video has a big screen in comparison to its CPU performance. It has a video accelerator chip to overcome that (which is why the Apple firmware handles video fine), but that's totally undocumented and unsupported by rockbox
19:56:57BigBambiNot to say it is bad, just that there are better (IMO)
19:57:53 Quit japc (Read error: 110 (Connection timed out))
19:58:17zichoso which would you say is the best player to run rockbox? i had an F60 before, but it doesnt work anymore so i need to get a new one
19:58:31 Quit TheSphinX^ ("XChat@Linux")
19:58:33BigBambithat depends on what you want
19:58:37gevaertsI'd vote the F60 :)
19:58:45BigBambirecording, fm, size, hd, ....
19:58:53 Join TheSphinX^ [0] (
19:58:54LloreanIf you liked the F60, it's really currently the best at what it was good at.
19:58:56BigBambiF60 is a very good choice
19:59:08zichoyes, i liked it.
19:59:13zichomaybe i'll get a new F
19:59:29zichoand dont think i can fix my old one, i have tried
19:59:41*gevaerts won't recommend an X. He doesn't want competition when he sees one for sale
20:00:09 Quit jhulst (Remote closed the connection)
20:00:39zichothe thing about my F is that i can read and write to it, but the ".rockbox"-directory just gives me an I/O-error when i try to do something
20:01:55 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
20:02:49gevaertsThat sounds like either a corrupted filesystem or a bad disk
20:03:24zichoim betting corrupted
20:04:48gevaertsHave you checked it?
20:05:19zichohow do i do that?
20:05:29zichoi tired with dsck but i dont really know if i did it right
20:06:34zichoyeah, typo
20:06:57gevaertsI use "fsck.vfat /dev/sdb1" (of course that sdb1 could be different for you)
20:07:27edrzit looks like I need to patch the firmware first ... according to
20:07:45BigBambibefore what?
20:07:56edrzis not found on the server.
20:08:21rasheredrz: where do you see this link?
20:08:47zichogevaerts, after i went through that it just says "Leaving file system unchanged."
20:08:49rasheredrz: append 32 or 64 to the URL, depending on whether you run 32 or 64-bit Linux
20:08:58BigBambiedrz: And select the one you want
20:09:00LambdaCalculus37edrz: Here's a working link:
20:09:08gevaertszicho: did it complain about errors?
20:09:34rasherfsck isn't really awesome at fat iirc
20:09:40rasherI'd try with windows chkdisk
20:09:40zicho"There are differences between boot sector and its backup." was the first thing i got
20:09:52BigBambirasher: I too have heard chkdsk is better
20:09:56edrzgreat. found this:
20:10:00edrzwill try
20:10:11BigBambiedrz: Yes, that is what three of us just told you :)
20:10:24edrzrasher: the broken link is under Step 1: Linux Installer
20:10:27BigBambiIf you want the 64 bit one, then that is correct
20:10:33rasheredrz: in the manual?
20:10:46edrzon the wiki
20:10:55rasherah right, I see you told already, sorry
20:11:32zichogevaerts, do you know anything else i can try?
20:12:22zichoif i backup the necessary files and format the player, will work do you think?
20:12:30gevaertszicho: can you put the full output on a pastebin somewhere (I'd suggest if you don't have a favourite yet)?
20:12:47zichoyeah sure
20:13:04gevaertsI'm not sure. The F may be a bit sensitive to reformatting, but that could as well be me misremembering
20:14:51gevaertszicho: try fsck.vfat -a /dev/sdb1 (or whatever your device is)
20:15:28zichohey that worked!!! gevaerts i fucking love you :D
20:16:17 Quit Seed ("cu, Andre")
20:17:53edrz[ERR] usb_set_configuration failed (-1)
20:17:59edrzhrmm. :-/
20:18:02 Quit TheSphinX^ ("XChat@Linux")
20:18:08 Part kkurbjun
20:18:10 Quit jhulst (Read error: 60 (Operation timed out))
20:18:25edrzthe light on the scroll ring is stuck on and it appears to be dead otherwise.
20:23:39 Join Lear [0] (n=chatzill@rockbox/developer/lear)
20:25:03edrzany suggestions?
20:26:55rasheredrz: That sounds like the manufacturer or rescue mode - I forget which
20:27:46kugelsounds like manufacturing mode, but then again I haven't used a R series ever
20:28:26edrzlsusb says: Bus 005 Device 012: ID 0781:0720 SanDisk Corp. Sansa C200 series in recovery mode
20:29:24 Quit yhuang (Read error: 110 (Connection timed out))
20:29:50edrzi'm not concerned if it doesn't come back ... it was a gift a few years ago from a clueless in-law. it's been in the box untill recently when I wondered if rockbox supported it.
20:30:12edrzbut, /me tries to unbrick.
20:30:36 Quit J-23 ("ZNC -")
20:31:59edrzok. holding down power and |<< brought it back.
20:34:03 Quit tyfoo (Read error: 104 (Connection reset by peer))
20:35:16 Quit kugel (Nick collision from services.)
20:35:19 Join kugel_ [0] (
20:35:29 Nick kugel_ is now known as kugel (
20:36:28 Quit fredddy (Remote closed the connection)
20:36:50 Join tyfoo [0] (
20:36:55edrzbut, still e200rpatcher can't see it.
20:37:40gevaertsedrz: are you sure it's an e200r and not a mislabeled e200?
20:38:01 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
20:38:02 Quit robin0800 (Remote closed the connection)
20:38:30 Quit grndslm (Remote closed the connection)
20:38:56edrzno, i'm not sure. lsusb says it's a c200.
20:39:14kugelwhat usb modes do you have?
20:40:02edrzrhapsody and "plays for sure", which I gather means it's an r.
20:40:42gevaertsIt is, yes
20:40:46 Join Zagor [242] (
20:41:39edrzso, it's set to that.
20:41:42edrzi power off.
20:41:48kugelif it says rhapsody and play-for-sure it's ae200R, otherwise (i.e.MSC and MTP as usb modes) e200
20:41:49edrzturn on hold
20:41:49kugeledrz: I assume you can differentiate between e200 and c200?
20:42:13edrzon the back it says e250r.
20:42:41kugeldid you make sure to run e200rpatcher with admin/root privileges (e.g. sudo ./e200rpatcher)?
20:44:25edrzgrr. sorry. worked as root.
20:45:17edrzand thanks.
20:45:20edrzon to step 2
20:46:15 Join kugel_ [0] (
20:47:29 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
20:47:36 Join Jaykay [0] (
20:48:30 Join SndBstrd [0] (
20:52:54 Quit zicho (
20:52:54 Quit kadoban (
20:52:54 Quit krazykit (
20:52:54 Quit jbenj (
20:52:54 Quit shadearg (
20:52:54 Quit Dieterbe (
20:52:54 Quit alexbobp (
20:52:54 Quit ChanServ (
20:52:57 Quit kugel (Read error: 104 (Connection reset by peer))
20:58:34 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
21:01:08NJoinkadoban [0] (
21:01:22 Quit SndBstrd ()
21:04:11 Join matsl [0] (
21:09:08 Nick kugel_ is now known as kugel (
21:10:03 Join J-23 [0] (
21:10:45NJoinzicho [0] (
21:10:45NJoinkrazykit [0] (
21:10:45NJoinjbenj [0] (n=jongle@
21:10:45NJoinshadearg [0] (
21:10:45NJoinalexbobp [0] (n=alex@
21:10:45NJoinDieterbe [0] (
21:10:52NJoinChanServ [0] (ChanServ@services.)
21:10:52Mode"#rockbox +o ChanServ " by
21:10:53 Join robin0800 [0] (
21:31:40 Join Aurix_Lexico [0] (
21:38:34gartraljust wondering, but what happend the logbot?
21:40:27Zagorhe's still here
21:40:34kadobangartral: what do you mean? it seems to be there...
21:41:36pixelmait lost op rights in one of the netsplits probably though
21:42:39gartralahh, its hiding as a normal user, im sorry, i was being lazy and not looking >.>
21:46:47 Quit Horscht ("Verlassend")
21:46:48 Join jgarvey [0] (
21:49:26QuickStartis the Ubuntu theme on official or unofficial
21:50:00n1sQuickStart: i think you may need to be a bit more specific
21:51:03***Saving seen data "./dancer.seen"
21:52:16QuickStartfor e200 sansa series
21:52:25QuickStarto and thank you guys for the help it did it
21:54:04QuickStartwell aside from my previous question
21:54:27QuickStartis it possible to align the menus similar to how the zune main menu appears
21:54:37QuickStartcenter aligned with a preview icon in the back
21:55:08 Quit Lear ("ChatZilla 0.9.84 [Firefox 3.1b3pre/20090120034652]")
21:55:13n1si'm not sure what you mean by "official" any themes for rockbox not included in the builds are simply done by people and put in the wiki
21:55:26QuickStartoh ok
21:55:40gartralQuickStart: back of what? you mean icons on the right-hand side?
21:56:05Unhelpfulperhaps he means behind the menu? i don't know, i've never seen a zune :/
21:56:32QuickStartthe menu basically is vertical and center aligned
21:56:43QuickStartthe icon is behind the selected item
21:56:56QuickStartand the text is helvetica
21:57:02 Quit LambdaCalculus37 (" ajax IRC Client")
21:57:03QuickStartI can mess with the fonts but
21:57:08QuickStartthe rest I don't know how
21:57:37Unhelpfullinuxstb: i rather doubt it has anything to do with rockbox doing anything "wrong", it's just how gcc does -save-temps - they go in the directory it's run from, disregarding any path data that is included in the input or output filenames
21:58:02Unhelpfulthough, it's clear that it's taking those filenames into account, because the temps are all <basename>.{i,s}
21:58:49 Join japc [0] (
21:58:56gartralthe screencaps i found dont even show icons, what series zune are you referring too?
21:59:16QuickStartfirst gen 65gb+
22:00:59gartrali still cant find a screen cap with icons at all...
22:02:28 Quit tyfoo ("Carpe diem")
22:03:04linuxstbQuickStart: But the short answer is no - you can't change the positions of anything in the menus. Just change the font and iconset.
22:04:17n1sand the long answer is "yes, it just takes a bit of coding" :)
22:04:30linuxstbThat's a shorter answer...
22:05:28QuickStartdo you know the coding for it n1s
22:05:44gartrali still wanna see a zune with icons at all, as far as i knew, it didnt have any
22:06:03 Join BigBambi_ [0] (n=Alex@rockbox/staff/BigBambi)
22:06:10 Quit BigBambi (Read error: 104 (Connection reset by peer))
22:06:13kadobanQuickStart: since you're the one that wants it, it seems fair for you to do the coding
22:06:44 Join yhuang [0] (
22:06:45n1sQuickStart: no, haven't messed much with the gui code
22:07:13linuxstbQuickStart: "the coding" is adding more theming features to Rockbox (Rockbox is written in C)
22:07:22 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
22:07:24 Join petur [50] (n=petur@rockbox/developer/petur)
22:07:40QuickStartso basically I would have to debug the shell and then make a new definition such as alignment right?
22:07:58linuxstb"debug the shell" ?
22:08:03 Quit jfc (Read error: 104 (Connection reset by peer))
22:08:29QuickStartwell not the correct terminology I mean read it get to know it
22:08:34kadobanQuickStart: you would have to either know or learn how to program, analyze the current menu drawing code, and change it to your liking
22:08:58QuickStartah ic one question is C very similar to visual basic?
22:09:07kadobannot at all
22:09:19gartralim not saying i would dislike the function of moving the icons/lists around, but as far as i can tell, it would almost certainly involve a near complete rewriting of the GUI code...
22:10:05 Join jfc [0] (
22:10:08QuickStartoh ic then nevermind lol
22:10:38Unhelpfulamiconn: the asm is interesting... DIV255 on ARM is 1 RSB and 5 ADDs, aside from the load and store... assuming those take 1 cycle each, that's about what a umull would take, isn't it?
22:10:46Unhelpfulthe whole thing is here:
22:10:57Unhelpfulgcc seems to make very good use of those shifted operands
22:12:15*amiconn wonders what all this adding of 127 is good for
22:13:05Unhelpfuli wondered about that as well... the value actually being calculated is (n*31+127)/255
22:13:51Unhelpfuli haven't a clue why it keeps recalculating the +127
22:14:23amiconnDoes it use your macro, or do you use literal '/'?
22:14:45amiconnIn the latter case I would understand why it keeps adding 127
22:15:06Unhelpfulyes, it's adding it separately at each shift position
22:15:22Unhelpfulit's using the macro - a literal / would just use the libgcc division, wouldn't it?
22:19:36QuickStartwhats the requirements for icon sets?
22:19:46QuickStartdo they get resized or must one do it
22:21:37 Quit Jaykay ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
22:23:40 Join gromit`` [0] (
22:24:52QuickStartthanks PaulJam
22:27:10 Quit z35 ("Leaving")
22:32:02 Quit gromit` (Read error: 110 (Connection timed out))
22:32:22 Quit Willwolfe ("ChatZilla 0.9.84 [Firefox 3.0.1/2008070208]")
22:36:24 Quit XavierGr (Nick collision from services.)
22:36:35 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
22:38:23 Quit zicho (Remote closed the connection)
22:38:47gartrali belive i asked before, but i would like access to edit the wiki, my real name is gareth schakel, i have a new theme to post for the e200s
22:39:50 Quit kugel (Read error: 60 (Operation timed out))
22:39:53Bagdergartral: and did you register such a user?
22:40:21gartraloop, i didnt know i had too, i thought you did all that, ok, ill do that now
22:41:04amiconnUnhelpful: Btw, the 'rsb' is not part of the DIV255
22:41:18Unhelpfulamiconn: yes, that's the *31
22:41:27 Join kugel [0] (
22:41:46Unhelpfulif i'm right? it's turning n*31 into (n<<5)-n
22:42:06 Quit TheSkunkMan ("Leaving")
22:43:08Zagorhmm, both h100 and h300 uses the 140 MHz capable BGA-packaged 5249
22:43:21gartralboth periods from the AC, or just the one between words and numbers?
22:43:54*edrz has rockbox :)
22:43:57edrzthanks for help earlier
22:45:30Unhelpfulit does the same thing for the *63. any idea why it re-adds the 127 at each shift, instead of adding it once before doing the shift-add?
22:45:36amiconnZagor: The 5249 in the irivers is marked "120" (by hand) in at least a number of units
22:46:03amiconnThere is a reason why we're only using 120MHz (or just a little more); it would overheat at 140MHz.
22:46:24amiconnThis is oooold knowledge; Linus tested it in the early days of the iriver port
22:46:57amiconnUnhelpful: gcc stupidity, as usual...
22:46:59gartrali have my activation code, do i need both periods, or only the first one?
22:47:04QuickStartone more question does the rocbox icon need to remain or can I change it
22:47:08*Bagder remembers that about the 140mhz as well
22:47:09amiconnTry using an intermediate
22:47:43gartralQuickStart: it can be changed
22:47:55gartralas well as the boot splash
22:48:20Unhelpfulhave to use an inline function for that, i'd think? i'm guessing that a macro that contains a code block will be bad to use as an initializer
22:49:04QuickStartummm I could do that but still don't want to mess with it just yet maybe later when I make my own original
22:49:53amiconnA macro containing a code block should work for auto vars
22:51:00*amiconn uses a similar thing to divide by (roughly) 255 and 257 in the greylib
22:51:22gartralBagder: all registered, and may i suggest editing the mail confirmation system to not include an "end" period?
22:52:08 Join toffe82_ [0] (
22:52:19Unhelpfulok then... just have the "return" value by itself at the end, as the SC_MUL one does
22:52:29Bagdergartral: permission granted!
22:53:40*Bagder guesses PoipolPoipol is not anybody's real name...
22:55:12 Join kugel_ [0] (
22:55:30 Quit kugel (Nick collision from services.)
22:55:44 Nick kugel_ is now known as kugel (
22:56:31 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
22:56:31 Quit amiconn (Nick collision from services.)
22:56:47 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
22:57:33 Quit n1s ()
22:59:25 Join tyfoo [0] (
23:01:24gartralBagder: sounds like a rear "one name" person who had too put two names in, at least, thats what i would assume
23:02:32Bagderwith no country specified and no other info either, I'd rather guess at a weirdo just filling in a nickname of some sorts
23:02:49gartralrockbox seems too not want too charge the battery to %100 its
23:03:12gartral"stuck" at 99 (sorry, hit enter trying to hit shift)
23:03:39 Quit itcheg (" ajax IRC Client")
23:04:03 Quit kugel (Read error: 60 (Operation timed out))
23:07:11 Quit {phoenix} (Remote closed the connection)
23:07:39 Join kugel [0] (n=kugel@unaffiliated/kugel)
23:07:45 Join itcheg [0] (i=41d59de2@gateway/web/ajax/
23:10:17 Quit evilnick (" ajax IRC Client")
23:10:54 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
23:12:08 Quit toffe82 (Read error: 110 (Connection timed out))
23:12:47 Quit Rondom ("Ex-Chat")
23:12:48 Quit tvelocity[a] (Read error: 113 (No route to host))
23:13:05 Quit pixelma (Read error: 110 (Connection timed out))
23:13:49 Join pixelma [0] (n=pixelma@rockbox/staff/pixelma)
23:15:59 Quit Horscht ("Verlassend")
23:16:30 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
23:18:19 Join tvelocity [0] (
23:19:03 Quit bmbl ("Woah!")
23:19:16 Quit yhuang ("Leaving")
23:34:51 Join arohtar [0] (
23:35:08 Quit matsl ("Leaving")
23:36:50 Quit itcheg (" ajax IRC Client")
23:40:53 Join Seed [0] (
23:51:06***Saving seen data "./dancer.seen"
23:55:45 Join akur [0] (
23:55:48 Part akur
23:57:12Zagormy first "boost reform" test is a dissappointment. I removed all boosts in the playback code and instead just added boost around mad_frame_decode(). this caused average cpu speed during mp3 playback to rise from ~27 to ~34 MHz.

Previous day | Next day