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

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

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

#rockbox log for 2010-06-13

00:11:41Buschelsaratoga: the disassemled code for the mpc dct (compiled with eabi) doesn't look too bad. nothing really obvious that could be further optimized (except maybe one or two ldm)
00:12:47saratogais it much different then what libmad uses?
00:13:52Buschelsaratoga: no. it is nearly identical. mpc uses a 64-value output (mirroring) of the calculated 32 values
00:14:16Buschelsaratoga: this is to avoid the even/odd handling in the dewindowing
00:15:29Buschelsaratoga: btw, maybe for >ARM4 it might be possible to rearrange ldr/ldm's a bit. in general the code would be hard to maintain...
00:15:36 Join Casainho [0] (
00:16:29saratogawould it be possible to efficiently reorder the coefficients before they get to synth_filter so that more of them could use ldm?
00:16:43saratogawell i guess that would mostly just help armv4, and thats not much use at this point
00:17:13BuschelI thought about this and found no solution
00:22:30saratogagoogle says mad_bit_read uses a fair amount of time, so maybe thats an option
00:31:44bertrikfunman, I got 10:59:00 runtime on my clipv1 today with a eabi+thumb build, very nearly the same as before
00:39:35 Quit r0b- (Quit: ( :: NoNameScript 4.22 :: ))
00:41:34 Quit ender` (Quit: Progress isn't made by early risers. It's made by lazy men trying to find easier ways to do something. -- Robert A. Heinlein)
00:55:35***Saving seen data "./dancer.seen"
00:56:48The_SevenZagor (for the logs): Not that it would really matter, but one could maybe suppress or change the "All green" message for the case where a build isn't even started, because only some irrelevant files were changed
01:02:09 Quit bieber (Ping timeout: 260 seconds)
01:02:26 Join bieber [0] (
01:06:57 Join r0b- [0] (
01:26:11 Join lev_ [0] (~lev@
01:27:38 Join will [0] (~chatzilla@
01:28:08will"(arm-elf-eabi-gcc) is not in your path!" −− is this new? What package do I have to install? (Ubuntu)
01:29:37Jeromwill: launch
01:29:42JeromAnd build eabi
01:30:37willwon't redo the entire installation? Can't I just install one package?
01:31:00JeromIt isn't a missing package
01:31:15JeromI just that the toolchain changed for some target
01:34:55will./ gets me this −−−− Please install "makeinfo" and re-run the script. −−- I don't see "makeinfo" in synaptic
01:35:14bertrikwill, install texinfo
01:36:41willok thanks
01:37:09Buschelsaratoga: Just took a look at mad_bit_read(). It seems like the data buffer is organized in uint-packets. mad_bit_reader() calls betoh32() to resort to little endian each time. This is inefficient. betoh32() needs to be called only once for each new uint-adress.
01:37:23 Join BHSPitLappy [0] (~BHSPitLap@unaffiliated/bhspitmonkey)
01:39:24Buschelsaratoga: We could add two variables: first holds the last uint-buffer index and the second holds the last betoh32'ed value. A similar approach was added to mpc's sv8 bitstream reader to reduce memory accesses to DRAM on Coldfire.
01:40:06 Quit S_a_i_n_t_ (Ping timeout: 240 seconds)
01:40:57Buschelsaratoga: The code would avoid to call betoh32(), if the data of the wanted address has already been converted.
01:41:46 Join S_a_i_n_t [0] (S_a_i_n_t@
01:45:49 Part Jerom
01:49:17Buschelsaratoga: quick-n-dirty approach (not compiled, not tested) ->
01:49:39saratogaBuschel: nice!
01:51:08Buschelmaybe you find time for test? I am too tired now and will better go to bed...
01:52:43saratogai'm compiling now
01:53:11 Quit domonoky (Ping timeout: 245 seconds)
01:54:06saratogaBuschel: did this compile for you?
01:55:04saratogaah last value shouldn't be a pointer i guess
01:55:54 Quit Buschel (Ping timeout: 265 seconds)
02:02:44saratogaBuschel: 27.89 MHz on PP5024 w/o COP
02:02:48saratogasorry 37.89
02:02:50 Quit lev_ (Ping timeout: 258 seconds)
02:06:18 Quit will (Quit: ChatZilla 0.9.86 [Firefox 3.6.3/20100401080539])
02:07:30saratogaBuschel: 37.60MHz without the patch
02:08:05saratoganot sure why it got slower, ARM is LE so it should save shifting
02:10:02 Quit BHSPitLappy (Ping timeout: 264 seconds)
02:11:38saratogaaduio sounds fine though
02:13:44saratogaanyway i have to go, but my random guess would be some kind of gcc weirdness, since the call to swap32 should be a lot slower then adding a single branch
02:25:07 Quit Casainho (Ping timeout: 260 seconds)
02:27:36 Quit bieber (Ping timeout: 240 seconds)
02:27:53 Join bieber [0] (
02:43:26 Join masterrogue [0] (8ea6c5b4@gateway/web/freenode/ip.
02:44:00masterroguehi room, i have a question about compatibility with rockbox and my player
02:45:17masterrogueive got a sansa e260 v2. the site has a link for the installer, but everywhere else it says its not supported. anyone in here get it to work on a v2? i'd really rather not brick my player
02:52:45mc2739masterrogue: the e260v2 is supported as a stable target
02:53:56mc2739masterrogue: where does it say it is not supported? (so that we can fix it)
02:55:37***Saving seen data "./dancer.seen"
03:05:58 Nick fxb is now known as fxb__ (
03:06:24masterrogueumm, just a second, ill post the link
03:10:52masterroguefirst line: " Rockbox doesn't function on the v2 models. The v1 hardware uses firmware named 01.xx.xx, while the v2 hardware uses firmware beginning with 03. This version number can be retrieved under Settings->Info."
03:11:55masterrogue...woops, as i look at the actual link * facepalm* my bad, ive been looking at the wrong page for months, my bad
03:15:06 Quit GeekShadow (Quit: The cake is a lie !)
03:15:29masterroguethanks for the help anyway mc
03:16:10 Quit bertrik (Quit: De groeten)
03:17:25 Quit masterrogue (Quit: Page closed)
03:21:35 Quit togetic (Ping timeout: 260 seconds)
03:35:44 Join togetic [0] (~togetic@unaffiliated/ibuffy)
03:38:40 Quit steve|m (Ping timeout: 276 seconds)
03:44:25 Join bba1973 [0] (
03:45:12 Join hebz0rl_ [0] (
03:45:51 Quit bba1973 (Client Quit)
03:45:57 Join krazykit [0] (~kkit@
03:46:35 Join moparx [0] (~moparx@unaffiliated/moparx)
03:49:04 Quit hebz0rl (Ping timeout: 276 seconds)
03:52:32 Join steve|m [0] (
03:54:05 Quit S_a_i_n_t (Ping timeout: 264 seconds)
03:54:22 Join S_a_i_n_t [0] (S_a_i_n_t@
04:19:55 Quit efyx (Remote host closed the connection)
04:25:50*JdGordon wonders why his new parse tree walker doesnt hit any TAG elements :(
04:27:15CIA-6New commit by jdgordon (r26824): count line numbers correctly, handle the VIEWPORT change so it only ever has one child
04:28:58CIA-6r26824 build result: All green
04:30:49JdGordonbieber: please be awake :D
04:33:06S_a_i_n_tJdGordon: skinupdater.exe is still valid/returns valid syntax...right?
04:33:24JdGordonsvn will yes
04:34:08S_a_i_n_tHmmmm, I have the sim bitching about lines it *really* shouldn;t be :/ No real time now, I'll get back to you later about it when I've found out more.
04:42:44 Quit pixelma (Disconnected by services)
04:42:45 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:42:49 Quit amiconn (Disconnected by services)
04:42:51 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:43:01 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:43:09JdGordonbieber: I dont know how you've done it, but there seems to be a bug in the parser which is giving TAG elements the vIEWPORT type, and the editor doesnt seem to mind?
04:43:11 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:46:57JdGordonbieber: p *line->children[0]->next->next->next->tag
04:46:57JdGordon$33 = {type = SKIN_TOKEN_DISABLE_THEME, name = 0x403b64 "wd", params = 0x4039f3 ""}
04:46:58JdGordon^ WTF? that should be a TAG type... (do make in utils/newparser and ./newparser cabbie<tab>)
04:49:24 Join Barahir_ [0] (
04:52:42 Quit Barahir (Ping timeout: 245 seconds)
04:55:38***Saving seen data "./dancer.seen"
05:02:32JdGordondamn, found the issue... skin_allow_element() doesnt init a value for element->type, and it happens that type == 0 here which is VIEWPORT
05:02:32 Quit krazykit (Read error: Connection reset by peer)
05:02:34JdGordonwhich is a bug
05:03:06JdGordoncurious that it doesnt effect the themeeditor though
05:03:14 Join krazykit [0] (~kkit@
05:04:45 Quit moparx (Ping timeout: 240 seconds)
05:13:03CIA-6New commit by jdgordon (r26825): initialise the element->type value so TAG types dont accidently get VIEWPORT if it element->type == 0 at alloc time
05:14:36CIA-6r26825 build result: All green
05:18:04 Quit fdinel (Quit: Miranda IM! Smaller, Faster, Easier.
05:45:36 Join brewmster [0] (
05:46:27brewmstercan rockbox on Sansa fuze v1 handle a micro SDHC card with an ext3 filesystem on it?
05:46:41brewmsterJdGordon, only vfat?
05:47:23brewmsterok thanks
05:47:41brewmsterjust installed rb yesterday−−pretty cool stuff
05:50:36*simonrvn likes rockbox a lot
06:01:07 Quit anewuser (Quit: for SELL 2 by the price of 1 now!)
06:02:38 Join FlynDice [0] (
06:02:42 Quit elcan (Ping timeout: 240 seconds)
06:03:40 Join elcan [0] (
06:04:32bieberJdGordon: Sorry, I'll be back tomorrow night
06:04:32bieberBusy weekend
06:04:55 Join JdGordon| [0] (
06:04:55 Quit JdGordon| (Changing host)
06:04:55 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
06:14:37 Quit panni_ (Read error: Connection reset by peer)
06:28:10 Join sawyer7 [0] (
06:36:20 Quit sawyer7 (Quit: Leaving)
06:37:38 Join sawyer7 [0] (
06:40:09brewmsteranyone here ever have any troubles with micro sdhc cards? mine keeps messing up when i try to transfer files to it
06:41:05brewmsteri've had "input / output error" with it twice now
06:41:53brewmsterafter the first time, I did a dd if=/dev/zero of=/dev/sde1 to it and then mkfs.vfat to make a new fs
06:42:05*S_a_i_n_t assumes this is a rockbox specific problem....?
06:42:09S_a_i_n_t...or hopes so.
06:42:57S_a_i_n_twhat I mean is, are you asking because you want to use it in rockbox? Being a rockbox dev/support channel and all...
06:43:45brewmsterah, yes
06:44:03brewmsteri've already tried it in rb
06:44:09brewmsterand it works fine
06:44:56S_a_i_n_tthen, this isn't the place to ask for support. if it works with RB, that's great, if not working elsewhere...then this isn;t really the right place for support.
06:45:55 Quit sawyer7 (Quit: Leaving)
06:55:39***Saving seen data "./dancer.seen"
07:08:36 Join Xerion_ [0] (
07:11:18 Quit Xerion (Ping timeout: 265 seconds)
07:11:19 Nick Xerion_ is now known as Xerion (
07:27:27 Quit Horscht (Ping timeout: 260 seconds)
07:29:52 Join Horscht [0] (~Horscht2@xbmc/user/horscht)
07:31:54 Quit brewmster (Quit: Leaving)
07:56:08 Join hamish_ [0] (~hamish@
07:56:37 Quit hamish_ (Client Quit)
08:18:25 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
08:45:46 Quit grndslm (Read error: Connection reset by peer)
08:47:20 Join Buschel [0] (
08:47:52saratogaBuschel: ping
08:50:00 Part toffe82
08:51:01Buschelsaratoga: I just see my last night's idea makes arm slower? maybe I can find some time today to further look into this...
08:54:09saratogaok was just checking if you saw that
08:54:41saratogado you have any idea how big the average read is?
08:54:56saratogai wonder if the common case takes most of the word, and so the cache does little
08:55:43***Saving seen data "./dancer.seen"
09:01:08BuschelI would expect an average read of (much) smaller than 16 bits. so, it should be more efficient to buffer. I will try to measure this on a small sample file
09:02:30BuschelI'll be off for a while now. see you latere
09:04:25 Join dfkt [0] (dfkt@unaffiliated/dfkt)
09:06:52 Quit Buschel (Ping timeout: 248 seconds)
09:21:16*JdGordon wonders what the heck to do about all these tags which no longer eat the \n in the skin files
09:21:26JdGordonwhich will cause problems when we move to the new parser
09:22:45 Join MethoS- [0] (~clemens@
09:25:33 Join stoffel [0] (
09:26:04*S_a_i_n_t seems to lean more toward the "Another possible option is if the line has any of a set of tags it will never cause a new line? Then we only need to get the list of such tags (pretty easy to find)..." side of things...maybe.
09:35:38JdGordonyeah, I tihnk i like that one best also
09:36:02S_a_i_n_tI have a WPS problem...
09:36:27S_a_i_n_trun skinupdater, the sim flakes on every subline.
09:36:38 Join grndslm [0] (
09:37:03S_a_i_n_thas the syntax changed for sublines?
09:37:29S_a_i_n_twhat's wrong with "%t(5.00) %Vd(a)%Vd(b);%t(5.00) %Vd(c)%?mv<%Vd(b)|%Vd(d)>"?
09:37:33*S_a_i_n_t sees nothing...
09:38:51JdGordonno thats correct
09:39:08JdGordonwhat is winging?
09:39:36S_a_i_n_tthe output doesn;t usual, but only commenting out the sublines fixes it :/
09:39:49S_a_i_n_tshall I pastebinerooney?
09:39:57JdGordondoesnt match what?
09:40:11JdGordonit parses correctly in the editor which means it is correct
09:40:16JdGordonand the tree looks correct
09:40:34JdGordonthe sooner I scrap the old code the better /me thinks
09:40:39S_a_i_n_t"ERR: Failed parsing on line 72 : ERR: Invalid parameter list for token 11: "volume" "
09:40:47S_a_i_n_tI mean, what the hell does that even mean?!?'s not even line 72.
09:41:19JdGordonignore the debug output.. it goes totally mental now :)
09:41:46S_a_i_n_tthe SIM HATES it, and apparently checkwps does too, as it fails there...well, when rasher tried to update it for me.
09:41:49JdGordonpastebin the whole skin?
09:42:02S_a_i_n_tyeppers. one sec.
09:44:56 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
09:45:05JdGordonthat isnt pastebin!
09:46:06JdGordonI jut wanted to look at the skin in the editor
09:46:36S_a_i_n_tyesterday I got all pissy trying to debug it, I attempted to get rasher to replace the zip on the themesite as the script missed it, then I handed him this theme and he said it failed checkwps on line 62, I have no idea if that was .sbs or .wps though. Interestingly, both the sbs and wps gave the same line 62 error in the sim.
09:46:44S_a_i_n_tmethinks checkwps is broken?
09:46:55JdGordonshouldn't be
09:47:17JdGordonWPS fails
09:47:31S_a_i_n_tthe sim gave the exact same output he got re: line 62, but for both .sbs and wps
09:47:44S_a_i_n_tthe sbs does even have code at line 62 IIRC
09:48:10S_a_i_n_tyeah, I get the sbs to load *today*, I couldn't on yesterdays sim though.
09:48:43S_a_i_n_tthe wps is fine if I comment out the sublines around line 60.
09:49:30JdGordonboth pass the new parser so meh :)
09:49:49S_a_i_n_tbut it failed checkwps!?!
09:50:09JdGordonyeah, lame aye?
09:50:15S_a_i_n_tthis also makes the sim essentially, it;s a little more than "meh" ;)
09:50:31S_a_i_n_twell, useless for me :P
09:50:52S_a_i_n_tdebug output, useless..parsing...useless :P
09:50:55JdGordonthe sbs works, it is the wps with the error
09:51:07S_a_i_n_t*Ahem* "error"
09:51:39S_a_i_n_tI haven't actually tried it on target. I trusted the sim too much.
09:51:50S_a_i_n_tbut, I also trusted checkwps.
09:52:06JdGordonyou really want such a short timeout?
09:52:19S_a_i_n_tthe theme needs it.
09:52:29S_a_i_n_twell, as short as it can possibly be.
09:52:34 Join Rob2223 [0] (
09:52:44S_a_i_n_totherwise the animation looks awful.
09:52:44JdGordoni gtg, I'll help you in a couple hours
09:52:51S_a_i_n_tno problem.
09:55:38 Quit Rob2222 (Ping timeout: 245 seconds)
10:05:51 Join Buschel [0] (
10:08:16 Join lev_ [0] (~lev@
10:10:15 Join ender` [0] (
10:11:52S_a_i_n_tJdGordon: You said the theme editor parses the theme correctly? I figured out the problem...but, it means that the skinupdater is broken. And, possibly the theme editor as well.
10:12:00S_a_i_n_tGimme a yell when you get back.
10:12:37S_a_i_n_tActually, not possibly...if it parses it correctly, then it *is* broken.
10:20:18Buschelsaratoga: new libmad patch ->
10:20:59Buschelsaratoga: the patch contains a small printf-debugging/-measuring and another caching (loading next dword)
10:22:41Buschelsaratoga: on my sample file I get the following results: a=83k, b=2000k, c=900k (hits). that means in total there were 983k calls of betoh32() against 2000k cached accesses or against 2983k betoh() in an unpatched environment.
10:23:07Buschelsaratoga: so, in total the betoh()-calls are reduced to 1/3
10:23:48Buschelwill test on target now
10:27:31Buschelsaratoga: a test on a Coldfire target could be interesting, too. I am not sure whether the bitstream is kept in IRAM or not...
10:33:48 Join showfiles [0] (
10:34:15 Quit hebz0rl_ (Quit: Ex-Chat)
10:35:14showfilesGeneral Settings > File View > Show Path > "Off" is this setting supposed to remove the path from all screens ?
10:35:48S_a_i_n_tit's suposed to only show the last part of the patch IIUC
10:36:17S_a_i_n_tie. if it was ./rockbox/path/thing it should dhow "thing"
10:36:43showfileswell it doesn't show the path when in files, but does show path elsewhere, I thought that "Off" was global
10:37:29showfilesno diaplay of path anytime
10:37:42S_a_i_n_twhat screens doesn't it show the path in?
10:37:52showfileswhen in Files
10:38:03S_a_i_n_teverything except files?
10:39:18S_a_i_n_tHmmm...actually, what I was thinking of was 'Current Directory Only"
10:40:23 Join Kitr88 [0] (
10:40:52S_a_i_n_twait...are you considering the "Rockbox" "Settings" etc. as the "Path"?
10:40:56showfilesthe "Off" option doesn't really tie up with what really happens, what does "Off' really mean, current only or global
10:41:06S_a_i_n_tit should *only* apply to "Files"
10:41:55S_a_i_n_tthe title and the path are totally different things...if you're mixing the two up.
10:42:46showfilesperhpas i am, the directory title, is that = path
10:43:20 Quit Kitar|st (Ping timeout: 276 seconds)
10:43:41S_a_i_n_tthe rest, like when you enter settings and it says "Settings" for example...that's totally different.
10:44:01S_a_i_n_tYou can only remove that by editing the .sbs
10:44:08Buschelsaratoga: slightly slower with patched version... (37.0 MHz against 37.15 MHz). I am confused
10:44:21showfilesok, so is the "Off" option supposed to only remove the directory title from the files directory
10:44:27S_a_i_n_tadd %?Lt<> to the .sbs, and it will turn those off.
10:44:41S_a_i_n_tto your last question, yes.
10:45:00S_a_i_n_tNo problem
10:45:43 Quit showfiles (Quit: CGI:IRC (EOF))
10:46:27Buschelcan somebody with a Coldfire target please make a short test_codec run with a mp3 file. svn against patch with ?
10:49:56bertrik\o\ |o| /o/ Radio receives a station on my rockboxed clip+ with mystery radio chip :)
10:55:45***Saving seen data "./dancer.seen"
10:59:55 Quit MethoS- (Remote host closed the connection)
11:03:51JdGordon|S_a_i_n_t: oh?
11:03:53 Join tjytj [0] (~velvetaci@
11:04:57S_a_i_n_tyeppers...the sim seems to hate %t(x.xx) (it has to be "%t(x.x)"
11:05:30S_a_i_n_tand, that'll be where checkwps is failing I assume, but your parser doesn;t seem to mind it.
11:05:43 Part tjytj
11:05:43S_a_i_n_ts/your parser/theme editor/
11:06:12S_a_i_n_tit took me *so* looooong to figure that out :/
11:06:21JdGordon|i dunno why x.xx should fail
11:06:40 Join tjytj [0] (~velvetaci@
11:06:46 Part tjytj
11:06:47S_a_i_n_tI've verified it...that's why the sim is failing.
11:07:10 Join stripwax [0] (
11:07:33S_a_i_n_tI can make the whole theme bork just by adding an extra digit after the decimal point on a single subline timeout.
11:08:04S_a_i_n_tskinupdater.exe didn;t pick up on it, I assume because it isn;t supposed to not work.
11:10:07JdGordon|wierd. i dint see why it should fail
11:10:33S_a_i_n_t%t(x.xx) used to be totally valid syntax, and from your reaction I assume it still *should* be?
11:11:10S_a_i_n_tdamn it took me a long time to figure it out though... :/
11:11:53S_a_i_n_tthe (rather comical) debug output from the sim had a slight hand in that though l)
11:12:08JdGordon|ok ill look into it later...
11:12:47S_a_i_n_tsweet, just letting you know I figured it out. Hope I saved you *some* time, though perhaps I just handed you more work...
11:19:09 Join DerPapst [0] (
11:20:51S_a_i_n_tOhh! ohh! JdGordon: One more thing...and alignment tag before %mv(x.x) fails to parse also ie. "%ar%?mv(x.x)<blah|blah>"...if *this* is expected or not, I have no idea.
11:21:03S_a_i_n_tI seem to remember it working in the past also.
11:31:01 Join Luca_S [0] (
11:33:31 Join flydutch [0] (
11:34:27 Quit stripwax (Quit:
11:37:43 Join halmi [0] (
11:39:39*Buschel is shaking head about libmad implementation
11:54:46 Join hamish_ [0] (~hamish@
11:56:14 Quit hamish_ (Client Quit)
11:57:32 Quit halmi (Read error: Connection reset by peer)
12:00:10 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
12:00:24 Join halmi [0] (
12:03:11Buschelcould someone with Coldfire target please test FS #11398?
12:09:14 Join petur [0] (~petur@rockbox/developer/petur)
12:12:17 Quit Kitr88 (Ping timeout: 258 seconds)
12:15:53S_a_i_n_tHahahah! S_a_i_n_t is on a roll...
12:16:20*S_a_i_n_t discovers that even if it's in a theme config, that the SIM won't load a .kbd file.
12:16:43S_a_i_n_tand if a .kbd file is loaded manually, that it won;y unload after the sim settings are reset.
12:17:24S_a_i_n_twhen did the sim get so....for lack of a better word, useless?
12:24:08S_a_i_n_thmmm...if a theme that doesn;t use icons is loaded in the sim, then one that *does* is loaded, the icons won't show either.
12:27:27S_a_i_n_tthe only thing that seems to get icons displaying after that is "reset settings" :?
12:27:47S_a_i_n_tI don't recall it being this broken in the past...
12:37:54 Join Kitar|st [0] (
12:41:02dfktdo i understand it correctly that %we enables an .sbs status bar, and %wi enables the default one?
12:42:54 Join Jerom [0] (~Jerom@
12:51:33S_a_i_n_tdfkt: %we enables *any* status bar, %wd disables it....%wi draws the inbuilt status bar in the "current viewport"
12:53:46S_a_i_n_tand, for the record...all status bars are .sbs now.
12:55:48***Saving seen data "./dancer.seen"
13:04:10dfktthanks, S_a_i_n_t (i also just noticed that still uses the old syntax, no wonder i can't load it...)
13:04:28S_a_i_n_tit does!?!
13:04:32S_a_i_n_twhat target...?
13:04:41dfktx5, current build, and current simulator
13:05:02S_a_i_n_twell, that just simply should't be.
13:05:10dfktah :)
13:05:21S_a_i_n_tthanks for sharing that. good to know.
13:06:10S_a_i_n_twhen you say "current sim", you built it?
13:06:18dfktno, i downloaded if from rasher's site
13:06:33S_a_i_n_taha...that's a 3.6 sim.
13:06:42S_a_i_n_tnone if that will be in current syntax.
13:07:14S_a_i_n_twell, depending which link you followed, it could even be 3.5
13:07:27dfkti upgraded the firmware it runs to r26813
13:07:39S_a_i_n_tyou can't "upgrade" a sim.
13:07:39rasherS_a_i_n_t: Only if he downloaded the 3.6 version.
13:08:11rasherS_a_i_n_t: I have both current and release sims
13:08:32dfkti didn't download the 3.6 release sim, i downloaded a current revision build
13:09:03S_a_i_n_twhat does "[23:07] <dfkt> i upgraded the firmware it runs to r26813" mean?
13:09:09S_a_i_n_tthat kinda threw me.
13:09:19The_Seven**** Target ipodnano2g is using non EABI compiler !
13:09:21dfkti just checked the system info in the sim, that's what it shows
13:09:33dfktmaybe i didn't upgrade it then
13:09:47dfktmabye that was the downloaded sim revision
13:09:58rasherS_a_i_n_t: if you unzip over the sim dir, that's in effect upgrading a sim.
13:10:53S_a_i_n_tsweet as, just made me think you'd tried to "upgrade" the sim...which, isn't possible ;)
13:11:39S_a_i_n_trasher: If you write over a sim with another sim...of course. It sounded to me like a target build had been applied to a sim, just checking.
13:13:05dfktto clarify/confuse more, this is what i see in -
13:13:13dfkti assume that's old syntax?
13:14:44S_a_i_n_tyou're correct.
13:15:04S_a_i_n_tand, this came with the sim?
13:15:18dfktit's also in the current firmware build i just downloaded, not only the sim
13:16:11dfktcurrent meaning r26825, directly from
13:16:57S_a_i_n_twow...what the hell?!? that *IS* old syntax.
13:17:04S_a_i_n_tyet, it runs on device here.
13:17:12dfkti'm afraid it's in the clip+ build too...
13:17:36S_a_i_n_tI believe you, it's in my nano 1st/2nd gen build also.
13:17:50S_a_i_n_twhy does it run on device though?
13:19:05dfktmaybe it uses the old "hardcoded"(?) status bar, not the .sbs?
13:19:20dfkt%we doesn't show anything on the sim, %wi does
13:19:34dfktyet %we works on my real x5
13:19:46S_a_i_n_tAFAIK there is no "hardcoded" .sbs anymore.
13:20:26S_a_i_n_tthat change was made a while ago, that's the reason there's "classic_statusbar"
13:21:21S_a_i_n_tNo ifs/buts/or maybes...this just simply shouldn't be.
13:21:33S_a_i_n_tnot only that, it shouldn;t *work* :/
13:21:54*S_a_i_n_t is more than a little confused.
13:22:19S_a_i_n_tJdGordon: ^^
13:26:37S_a_i_n_tI'm trying to se if I'm insane/confused or not...but I can run this old syntaxed on sim and on device.
13:29:03 Join TopyMobile [0] (
13:32:58S_a_i_n_tHmmm, ok, I am slightly insane I think. Now the SIM *is* rejecting it.
13:34:20S_a_i_n_twhich is, good? I think there must still be an inbuilt .sbs, but I was under the impression that there was no inbuilt .sbs and that was the reason there was "classic_statusbar"
13:34:48S_a_i_n_tDespite that wee messup of mine, it still shouldn;t be in the old syntax.
13:35:10dfkti checked, the built in one was in apps/gui/gwps-common.c.. which doesn't exist in current revisions
13:35:34S_a_i_n_tthat doesn't mean it's not somewhere else though ;)
13:35:46S_a_i_n_tit *has* to be getting it from somewhere...right?
13:37:05dfktah, there's statusbar.c too
13:37:26S_a_i_n_tand the sim is now failing on line 7 (which really means line 8), which makes sense...that's the first line of code, in old syntax"
13:44:00 Quit elinenbe (Ping timeout: 272 seconds)
13:46:56 Join fdinel [0] (
13:52:46 Join efyx [0] (
14:05:30 Join Jaykay [0] (
14:07:25 Quit sinthetek (Ping timeout: 260 seconds)
14:15:36 Join elinenbe [0] (
14:18:09 Quit bluebroth3r (Ping timeout: 260 seconds)
14:19:44 Join bluebrother [0] (
14:19:44 Quit bluebrother (Changing host)
14:19:44 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
14:21:54 Quit elinenbe (Ping timeout: 245 seconds)
14:23:07 Join elinenbe [0] (
14:28:52 Join kugel [0] (~kugel@rockbox/developer/kugel)
14:29:32 Quit elinenbe (Ping timeout: 258 seconds)
14:36:32 Join elinenbe [0] (
14:41:47 Quit elinenbe (Ping timeout: 252 seconds)
14:42:21 Quit Topy44 (Ping timeout: 248 seconds)
14:44:57 Quit stoffel (Remote host closed the connection)
14:45:09 Join Topy44 [0] (
14:50:35 Quit Buschel (Ping timeout: 252 seconds)
14:50:41 Quit TopyMobile (Quit: TopyMobile)
14:51:15 Quit petur (Quit: Leaving)
14:53:42 Join funman [0] (~fun@rockbox/developer/funman)
14:54:40 Quit Jaykay (Ping timeout: 260 seconds)
14:55:49***Saving seen data "./dancer.seen"
15:06:32 Join sinthetek [0] (~sinthetek@unaffiliated/sinthetek)
15:07:06 Quit sinthetek (Client Quit)
15:07:13 Join sinthetek [0] (~sinthetek@unaffiliated/sinthetek)
15:10:19JdGordonS_a_i_n_t: ok back
15:16:15 Quit Luca_S (Quit: CGI:IRC (Ping timeout))
15:16:43CIA-6New commit by jdgordon (r26826): update classic_statusbar to the new syntax
15:16:45 Quit Jerom (Remote host closed the connection)
15:18:16CIA-6r26826 build result: All green
15:23:37JdGordonbieber: (thinking loudly so ignore if you are sick of the parser)... in parse_viewport you check to see if the line will have sublines, and then if it does you do the same code to count how many are there. Why not always call parse_subline() (so dont check if it is needed) and then in that if sublines == 1 (so no ;'s) change to the parse_line instead?
15:29:57 Join elinenbe [0] (
15:30:54 Join panni_ [0] (
15:32:43kugelThe_Seven: someone needs to check if eabi works and commit the changes to configure and the build file
15:32:58kugel(re: why no eabi on ipodnano2g)
15:34:42funman [FFmpeg-devel] [CHECKLIST] what is preventing us from switch to git
15:36:05 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
15:36:34funman [vlc-devel] Switching to git (this thread seems to not contain any flamewar, just technical details on the switch from svn to git)
15:39:25JdGordonfor the record I wouldnt really have any problems if every switched to git... the learning curve is annoyingly steep though to get the the point where it actually has a benefit over svn
15:41:01The_Sevenkugel: IIRC i confirmed that it worked during the transition
15:45:20CIA-6New commit by theseven (r26827): Switch ipodnano2g to eabi in /www/buildserver/builds
15:45:35 Quit elinenbe (Ping timeout: 252 seconds)
15:46:25CIA-6New commit by theseven (r26828): SSwitch ipodnano2g to eabi in /trunk/tools/configure
15:46:27 Join voRia [0] (
15:47:25 Join elinenbe [0] (
15:48:03CIA-6r26828 build result: All green
15:49:07CIA-6New commit by funman (r26829): Buildserver: build cowond2 with eabi
15:49:12CIA-6New commit by funman (r26830): build cowond2 and iaudio7 with EABI, they aren't stable
15:50:09*funman looks for b0hoon
15:50:42funmanamiconn: did you test all the ipods still building with gcc 4.0.3 ? what's the problem with them, stack overflow iirc ?
15:50:51CIA-6r26830 build result: 0 errors, 2 warnings (funman)
15:54:02funmannot sure what i should do here: make the code do what the comment says, at the risk of breaking cowond2 lcd ?
15:54:36CIA-6New commit by jdgordon (r26831): get the renderer working more better! "handle" sublines correctly, dont "draw" in the first viewport if we are using viewports, get rid of the idea of ...
15:54:56 Quit elinenbe (Ping timeout: 265 seconds)
15:55:30 Quit ender` (Quit: I like work: it fascinates me. I can sit and look at it for hours. -- Jerome K. Jerome)
15:56:06CIA-6r26831 build result: 0 errors, 2 warnings (jdgordon)
15:56:13kugelJdGordon: how many parsers do we need?
15:56:18funmanJdGordon: fix your yellow!
15:57:00kugelJdGordon: what is newparser for? I thought the theme editor's parser was the way to go?
15:57:23JdGordonkugel: the way it works is text -> new parser -> awesome text target independant parse tree -> middle ground to make it target specific and build "tokens" which are useful -> renderer
15:57:28JdGordonOH, no thats just my playground
15:57:34JdGordongit is too hard :p
15:57:48JdGordonbassically I'm building the new renderer outside of the core to make sure it works
15:57:56JdGordononce it is mostly done it will be copied in
15:58:12JdGordonthis is to replace skin_parser.c and skin_display.c and the .h's
15:58:59kugelso the themeeditor's parser won't be used?
15:59:05JdGordonit will be
15:59:09JdGordonI am using it directly
15:59:32JdGordonthe code in that folder is after it goes through the parser
16:00:07JdGordonit would be pretty slly if we wernt using the new parser :p
16:00:23kugeloh ok
16:00:32CIA-6New commit by funman (r26832): cowond2: make the code consistent with the comments and really clear the bits ...
16:00:56kugelJdGordon: "dont "draw" in the first viewport if we are using viewports" isn't this allowed currently (except for sbs)?
16:01:17JdGordonthis is explicitly not allowed
16:01:33JdGordonwe go through the viewport but not draw anything
16:01:57The_SevenBagder: Have a Mydoom worm?
16:02:04*kugel misremembers then
16:02:10CIA-6r26832 build result: All green
16:02:12 Join elinenbe [0] (
16:03:14JdGordonkugel: maybe misunderstood... if any viewports are defined then the default viewport isnt drawn
16:03:46kugelyea, in my memory it was drawn in the wps but not in the sbs
16:04:48kugelah, the default vp in the sbs gets the VP_NEVER_VISIBLE flag, in the wps this doesn't happen
16:05:48*kugel now thinks he remembers correctly
16:06:15JdGordonhows RaaA giong? still going th android route?
16:06:51kugelquick topic change?
16:07:23kugelit's going good I think, although I'm busy with uni-work currently. yes, android is planned
16:10:53 Quit elinenbe (Ping timeout: 252 seconds)
16:13:38funmanbertrik: can you test on your clipv1 ?
16:22:25 Join elinenbe [0] (
16:25:08JdGordonis it better to case a void* to a intptr_t and use it as an int? or alloc a int and store the pointer there?
16:25:08 Quit halmi (Read error: Connection reset by peer)
16:26:03funmanint could be too small to hold a pointer, but not intptr_t
16:26:48 Join halmi [0] (
16:26:58JdGordonthe struct member is a void*
16:27:09JdGordonI just need to store some flags there
16:28:04funmanflags in a pointer ? Oo
16:28:29JdGordonvoid* data;... most types store a struct * there, I only need an int here
16:28:31JdGordonfor this one
16:28:52funmanwhat about using an union?
16:29:30JdGordonmeh, I'll just do it safewly and add a struct, saves doing &'s later also
16:33:26kugelI'd use a union too
16:37:15 Quit elinenbe (Ping timeout: 264 seconds)
16:40:14*JdGordon has line eating tags working :)
16:40:29 Quit GeekShadow (Read error: Connection reset by peer)
16:41:05 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
16:42:11CIA-6New commit by jdgordon (r26833): Start dealing with LINE elements... setup a flag which lets tags tell the renderer to not start a new line in the viewport (i.e %we/d/i %X/x/xd etc)
16:42:49CIA-6New commit by jdgordon (r26834): duh, this is needed also
16:43:58CIA-6r26833 build result: All green
16:45:31CIA-6r26834 build result: All green
16:48:14 Join anewuser [0] (anewuser@unaffiliated/anewuser)
16:49:21 Quit perrikwp (Quit: CGI:IRC (Ping timeout))
16:49:27 Join elinenbe [0] (
16:52:16 Quit elinenbe (Read error: No route to host)
16:54:46 Join elinenbe [0] (
16:55:52***Saving seen data "./dancer.seen"
16:59:09bertrikfunman, I'll have a look, anything specific I should look at?
16:59:28funmanbertrik: nope, i just wanted to know if it has a real effect on battery life
17:00:21 Quit elinenbe (Ping timeout: 276 seconds)
17:00:32bertrikok, I think I'll start a battery bench tonight and let you know tomorrow
17:01:37bertrikby the way, the clipv1 battery benchmark I did after the recent eabi + thumb changes showed no significant difference (about 1 min in 11 hours)
17:02:21funmanyeah i've seen that, perhaps memory use doesn't affect battery a lot
17:02:26bertrikquite remarkable actually that they got so close
17:02:43funmancodec speed is almost the same for thumb & arm
17:03:56bertrikI have basic tuning working now for my clip+
17:04:27BagderThe_Seven: as if those mails come from me...
17:05:33kugelThe_Seven: why not make the commit then?
17:05:42The_Sevenkugel: I've already done that
17:06:24*The_Seven still wonders where they could come from, especially how that friggin' worm could have found out about the list
17:06:34kugelThe_Seven: oh, overlooked that
17:11:08Bagdermails and worms do that all the time, I've stopped wondering. I see my email used like that every day
17:13:39*kugel wonders what this worm talk is about
17:14:45Bagdera mail that was sent with my From: adress to rockbox-dev
17:14:58Bagderit included an .exe
17:15:12Bagderinfected with a worm
17:15:50kugeloh, I didn't get that mail, only Mosner's reply
17:16:11Bagderyeah, I had the same experience
17:17:31Bagderbut the mail archive shows it
17:18:41LloreanI didn't get it either. I assume a lot of us have filters that just snagged it for us.
17:20:47 Quit halmi (Quit: halmi)
17:32:14The_Sevenyep, i saw my filter snag it
17:32:32The_Sevenhowever, i couldn't see it on the archive, it was probably already removed by some admin
17:32:35 Quit funman (Quit: free(random());)
17:32:42 Join elinenbe [0] (
17:34:00BagderI've however replaced the attachment so it isn't spreading more
17:34:09*Bagder runs off
17:36:52 Quit elinenbe (Ping timeout: 240 seconds)
17:38:42 Join elinenbe [0] (
17:40:47 Join elinenbe_ [0] (
17:43:09 Quit elinenbe (Ping timeout: 248 seconds)
17:45:25 Quit elinenbe_ (Ping timeout: 258 seconds)
17:48:29 Quit storm` (Ping timeout: 258 seconds)
17:51:30 Join storm` [0] (
17:52:13 Join kugel_ [0] (
17:52:21 Quit kugel (Ping timeout: 260 seconds)
17:52:23 Nick kugel_ is now known as kugel (
17:52:37 Quit kugel (Changing host)
17:52:37 Join kugel [0] (~kugel@rockbox/developer/kugel)
17:53:09 Quit r0b- (Read error: Connection reset by peer)
18:13:29 Join Buschel [0] (
18:14:06*Buschel again waves to Coldfire testers
18:14:16 Join Photoguy [0] (~chatzilla@
18:19:37PhotoguyDoes anybody know if it's possible to get a d2+ replacement touchscreen?
18:28:59 Quit Photoguy (Read error: Connection reset by peer)
18:36:09 Join ender` [0] (
18:47:28 Join domonoky1 [0] (
18:47:52 Quit domonoky (Ping timeout: 240 seconds)
18:48:02 Join webguest30 [0] (
18:48:09webguest30Hello lifesavers
18:48:14 Nick webguest30 is now known as m0ar (
18:48:35m0arIs there an ETA of rockbox 3.6 -> clip+ ?
18:48:37 Join elinenbe [0] (
18:48:44 Nick m0ar is now known as Guest79727 (
18:48:48AlexP3.6 has been released for all the targets it is going to be
18:49:05AlexPIf the clip+ is deemed stable at the time it will be in 3.7
18:49:19AlexPBut only if it meets the stable requirements
18:50:07Guest79727What problems exists with the current version?
18:50:40AlexPI don't know
18:50:58Guest79727Is there a wikipage or sumthing?
18:51:12AlexPThere are a couple, but they don't seem too helpful
18:51:18AlexPThey are linked fro; the front page
18:51:34LloreanAmong other things, the Clip+ doesn't yet play files back at the right speed yet.
18:51:43LloreanThere's just still hardware that isn't working right and so on.
18:52:52 Quit elinenbe (Ping timeout: 240 seconds)
18:54:53Guest79727What might be the reason to the clip having such an odd problems as a 1.1% slowdown?
18:55:02AlexPSee the bug report linked
18:55:06AlexPIt explains it
18:55:50Guest79727Trying to understand already
18:55:55***Saving seen data "./dancer.seen"
18:56:45LloreanThe bug report comment explanation is pretty straightforward, if you don't get it there's probably not much that can be done to help.
18:57:10Guest79727Well I get it technically, but I'm still unsure of what may cause such a problem
18:57:17Guest79727But thanks for the chitchat guys
18:57:20 Quit Guest79727 (Quit: CGI:IRC)
18:57:27 Join funman [0] (~fun@rockbox/developer/funman)
18:57:34dfktwhat do i have to do to get an .sbs loaded into my theme? i added a "sbs: /.rockbox/wps/" line to my theme .cfg, and enabled the status bar by either %we or %wi... but that's obviously not all
18:58:19funmanLlorean: the fact that samplerate isn't 100% accurate isn't a bug per se. Having a higher precision is merely an enhancement, not a bugfix
18:58:56Lloreanfunman: It's a bug if files are playing measurably too slow, isn't it?
18:59:11 Join stoffel [0] (
18:59:13funmanhardware limitation
18:59:19LloreanWhat would you call it if a file that's supposed to be 4 hours long completes in 3 hrs 50 minutes?
18:59:35Lloreanfunman: So it's impossible to fix?
18:59:52funmani would tell the difference is below the human ear precision, so it's ok
18:59:57AlexPThe OF (while not being perfect) is quite a bit better IIRC
19:00:01funmanit's possible to have more precision
19:00:12funmanyeah bertrik suggested that the OF resamples everything to 48kHz
19:00:52Lloreanfunman: Feel free to close the bug report as "not a bug" then.
19:01:07AlexPI always find it odd that hardware made for playing audio has such stupid limitations
19:02:24funmanwell i believe having a clock freq which is an integer multiple of every samplerate in existence is hard
19:02:42AlexPtrue, but 44.1 seems a fairly obvious one
19:02:46LloreanYeah, but when CDs are 44.1...
19:05:15LloreanAlexP: But if I recall Sandisks' response was more or less "Well, you wanted a cheap player, and cheap means not everything can be perfect"
19:05:27 Join r0b- [0] (
19:05:48kugelfunman: could we do that as well?
19:05:56AlexPYeah, and if it is small enough to not be noticeable, then never mind, but still seems slightly silly :)
19:06:06funmankugel: resample?
19:06:11LloreanAlexP: Well, people *heard* it and complained to Sandisk that it didn't sound quite right originally.
19:06:17LloreanSo some people can hear it, at least.
19:06:24funmansure but i don't think you can resample without quality loss
19:06:27AlexPah, didn't know that
19:06:37AlexPfunman: Especially not with our resampler
19:06:49LloreanI think dismissing it as "not a bug" isn't something that should happen. It should, more likely, be resigned as a "known issue" if we can't figure out how to fix it.
19:06:56funmanperhaps 22.05 -> 44.1 would be doable (just duplicate every sample)
19:07:20dfktthat was one of the biggest complaints with a whole generation of marketed as 'high quality' sound blaster cards - people complained about them resampling everything to 48khz and audibly making the sound worse...
19:07:26kugelI'd think that resampling upwards shouldn't affect quality
19:08:29kugelwhat does the pitchscreen do? I once set it to 101.2% and couldn't hear a change in quality (but I found the speed difference noticeable)
19:10:16funmani think it calls dsp_upsample / dsp_downsample
19:10:37funmanLlorean: well i closed it as notabug with an explanation
19:11:29kugelany idea how duplicating a sample once in a while would affect sq (negatively)?
19:12:02funmanwell you're just outputting a different sound
19:13:59ranmadfkt: The problemwith the sound blaster emu chip is that it does hw resampling for lots of channels and they had to cut corners so the resampler isn't quite that good.
19:14:19 Join kramer3d [0] (~kramer@unaffiliated/kramer3d)
19:14:29ranmaAnd doing high quality resampling in software would probably mean less runtime for rockbox...
19:15:09funmanwhat's the samplerate precision of other rockbox targets (not the AMS) ?
19:15:25dfktiirc, one can find out if a resample does its job well with a simple sine sweep
19:15:32 Quit stoffel (Ping timeout: 265 seconds)
19:20:53 Quit lev_ (Ping timeout: 258 seconds)
19:29:09funmanabout svn properties i'll wait a bit and propose change of the conventions:
19:29:31funman- make Id keyword optional (not mention it on the wiki)
19:29:46funman- make eol-style optional as well (and mention that text files should have unix line endings)
19:30:10funman- give image files their correct mime type (not octet-stream)
19:31:46funmanhm well nobody (but gevaerts on irc) said Id was useful so I'll propose to remove it completely
19:32:51 Join hebz0rl [0] (
19:34:02AlexPfunman: For Id I'd rather have it one way or the other - use it everywhere or not at all, rather than a mishmash on files
19:34:26AlexPPersonally, I don't really care if it is used or not :)
19:34:45bertrikWait, I don't see the problem really, just seems to be that some people can't or won't want to use subversion and try to find reasons to use git instead, which is IMO a separate discussion
19:35:00funmanbertrik: then you didn't read the thread
19:35:28funmanreally it's not about "some" people using git-svn and not svn
19:35:35AlexPbertrik: There are two discussions IMO - one is whether guidelines should be followed (yes, absolutely IMO), and the second is what those guidelines should be
19:36:15AlexPIf the guidelines are there, then they should be followed regardless on whether some people don't want to, whether through thinking them pointless, or choice, of tools, or whatever
19:36:34AlexPHowever, what the guidelines actually are can be changed of course :)
19:37:06bertrikSo, we're discussing now a change in the subversion guidelines here, right?
19:37:18AlexPI think so :)
19:37:55 Quit kugel (Remote host closed the connection)
19:38:15funmanI wait a bit to see if more people have something to say about these properties and i'll send an email about removal of some of these properties
19:41:00bertrikChange 3) looks fine to me, 1) I find very reasonable, but I wonder if change 2) might give problems for people still using windows
19:41:53AlexP1) Just bin them, not have a half-way house, 2) no idea) 3) sure :)
19:42:39funmanon windows we still use unix tools which expect unix line endings
19:43:27funmanand the plan is for new files, unless someone wants to go through all existing files and remove properties
19:43:44 Join saratoga_ [0] (463f90ed@gateway/web/freenode/ip.
19:44:25saratoga_why is likely the OF resamples? Couldn't they just use a different PLL clock?
19:44:52bertrikPeople on a cygwin setup might still want to edit files with a windows text editor, although that's probably a rather small percentage of people compiling rockbox
19:46:04saratoga_Buschel: isn't Coldfire BE? so this would probably be slower since you have the overhead from the branch
19:47:04Buschelsaratoga: the access to DRAM is much more expensive than the branching (from what I experienced so far)
19:48:00Buschelsaratoga: I am wondering why the patch is slightly smaller on arm7tdmi −− which is LE
19:48:11 Quit kramer3d (Read error: Connection reset by peer)
19:48:15funmansaratoga_: there's only 2 PLL settings (that i've seen) in the OF : 160 & 240MHz
19:48:39 Join kramer3d [0] (~kramer@unaffiliated/kramer3d)
19:48:56funmanbertrik: windows text editors handle unix line endings (just not the original notepad.exe, at least not on XP)
19:52:02saratoga_its possible to test for resampling by playing a tone really to 22khz and recording it (ideally at 48khz) and seeing if it comes out at some other frequency
19:52:11 Part ocin
19:52:26saratoga_but dfkt's rmaa results don't look like resampling
19:52:35saratoga_at least not very fast resampling
19:52:50bertriksaratoga, AFAIK the resampling option is just a guess now and is not really supported by experiments yet
19:53:22saratoga_the frequency response looks identical to RB even out to 22khz, which would require a very good resampler and a lot of CPU time
19:57:09funmanranma: what's left for AMSv1 usb driver ?
19:57:22funmani just see a TODO: switch max packet size depending on speed
19:59:50bieberJdGordon: Makes sense, I'll take a look at it
19:59:51bieberShouldn't be a hard change to make
20:00:06 Join Blue_Dude [0] (~chatzilla@rockbox/developer/Blue-Dude)
20:01:11Blue_DudeHi. Been gone a week, and now I don't have the right toolchain anymore. How do I get eabi on Cygwin?
20:01:28funmanecho e|./tools/
20:01:58saratoga_you'll probably need to install some other packages too since i don't think we ship cygwin with all the dependencies
20:02:40funmanthat'd be texinfo
20:02:54saratoga_why is that needed anyway?
20:03:24Blue_Dudetexinfo appears to be installed already.
20:04:36Blue_Dude"echo e|./tools/" ? Is that all?
20:05:01Blue_DudeWhat does the "echo e|" do for me?
20:05:58Blue_DudeNever mind, figured it out.
20:06:52funmansaratoga_: some docs which can't be disabled without tweaking makefile i think
20:07:14Blue_DudeROCKBOXDEV: Please install curl or wget and re-run the script
20:10:21funmanhm the SD change seems to get 1h30 of battery on fuzev2 (now 21h)
20:10:42funmani started benchmark on clipv2/clip+ today
20:10:54saratoga_which changes?
20:11:23funmandisable the 4 clocks in sd_enable
20:12:54saratoga_ah thought you meant this was already committed
20:13:14funmani'll make sure to run test_codec before committing (because that kind of change broke on AMSv1)
20:13:46 Quit Buschel (Ping timeout: 260 seconds)
20:13:52 Join Luca_S [0] (
20:16:01Blue_DudeHoly crap! That script is... busy...
20:17:13 Quit scorche (Ping timeout: 264 seconds)
20:22:06Luca_SHello! Is frequency switching still unstable on fuzev2? I just completed an entire battery cycle without crashes (not a bench, just a few days of casual listening) in my custom build that includes the latest patch on the tracker
20:23:25funmanLuca_S: i don't remember how it crashed on fuzev2
20:23:39funmanBut i just remember that with svn the battery life should be ~14h30, not ~19h30
20:23:49funmanso with the SD patch, 14h30 -> 21h is quite awesome
20:24:10bertrikwow, indeed
20:24:34funmanLuca_S: you have made a bench of unpatched SVN recently?
20:24:59Luca_Sin the early days of the port it simply froze, no panic message and no reactions to button presses
20:26:09Luca_Sand it happened randomly but usually once in half an hour
20:26:15funmanLuca_S: reading FS #11297 again, I think it didn't crash on fuzev2, but I didn't want to enable it because the same settings crashed on clips
20:27:36 Join scorche [0] (~scorche@rockbox/administrator/scorche)
20:28:06 Quit bieber (Ping timeout: 272 seconds)
20:28:21 Join bieber [0] (
20:28:28Luca_Scouldn't it be selectively committed? I'm asking since I'm unable to create a recent custom build, still can't build the eabi toolchain with the vmware image :(
20:29:14 Join petur [0] (~petur@rockbox/developer/petur)
20:29:53Luca_Si.e. enabled only on fuzev2
20:30:11funmansure it could
20:31:57 Quit storm` ()
20:35:00 Quit Luca_S (Quit: CGI:IRC (Ping timeout))
20:37:41 Join Luca_S [0] (
20:41:09Luca_Sahh, damn cgi, I couldn't read your replies :D
20:41:37Luca_Sfunman Luca_S: you have made a bench of unpatched SVN recently? << no, but I could start it now if you need it
20:41:46funman <- 14hours is probably for old builds
20:42:34funmanLuca_S: yeah please, i want to measure how affects battery life
20:42:54Luca_Sok, just plain latest svn then
20:43:16Blue_DudeI think the Cygwin config wiki needs updating in regards to required install packages. In addition to the listed packages, we need python, curl, and so far, libiconv...
20:44:54funmanBlue_Dude: something else than needs python?
20:45:05Blue_DudeAlso, the dev script terminated with an error because I didn't have libiconv. Now I can't rerun, because build-binutils already exists. Is there a workaround?
20:45:21Blue_Dudefunman: analysis tools need it.
20:45:30Blue_DudeNot required I suppose, but useful.
20:45:47funmanBlue_Dude: if binutils is installed you can comment the "build ...." line for it in; else just remove the folder to start building it agfain
20:46:07Blue_DudeI was afraid of that.
20:46:13Blue_DudeI'll start over.
20:46:19funmancan you run arm-elf-eabi-as ?
20:47:17 Quit Luca_S (Quit: CGI:IRC)
20:47:26 Join Luca_S [0] (
20:47:45Blue_DudeI already started over...
20:48:09Blue_DudeTakes a long whole to run. I hope I have all the dependencies I need this time.
20:48:30 Quit dfkt (Quit: -= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.)
20:49:37Blue_DudeI'm so dumping Cygwin in a few weeks. Is there any particular flavor of Linux that works well with the Rockbox dev tools? What would you install if starting from scratch?
20:50:52AlexPany really
20:51:10AlexPThere are people using debian, ubuntu, fedora, arch, ....
20:51:19AlexPProbably a convo for community :)
20:55:23Luca_Sbattery bench started, default settings, mp3 mixed bitrates, volume 0db
20:55:35bieberBlue_Dude: Just about any distro with a package manager should have all the crap you need in its repositories. I've always found Ubuntu the easiest to set up and use, but anything you can get running on your machine should work fine
20:55:58***Saving seen data "./dancer.seen"
20:55:58AlexPYeah, let's not start a distro war in here - any will do it :)
20:56:35Blue_DudeI moved the question to the other forum...
20:58:06 Join stripwax [0] (
21:00:19 Quit Luca_S (Quit: CGI:IRC)
21:00:26 Join Luca_S [0] (
21:04:18Luca_Sdid anybody notice that in recentish svn for fuzev2 any time OF boots after a RB session it triggers OF's 'rebuilding database'?
21:04:48funmanyes, look on flyspray
21:07:11amiconnfunman: PP5002 sees immediate stkov panic on COP
21:07:25amiconnBut increasing that stack (the idle stack) doesn't help
21:07:30funmanall the ipods listed in configure are PP5002 ?
21:07:36amiconnSo something else is wrong
21:08:21funmandid you try printing the stack pointer in the panic message to check it was really overflowing ?
21:08:25amiconn1g2g and 3g are
21:08:35amiconnThe others are just untested
21:08:46amiconndomonoky1: Could you please test 4g?
21:10:30*amiconn suspects that something in thread.c is going wrong
21:12:54funmantry changing C/ASM code line 298 of thread-pp.c ?
21:13:08funman(just throwing the idea)
21:21:23 Quit flydutch (Quit: /* empty */)
21:22:20 Quit funman (Quit: free(random());)
21:46:08 Join lev_ [0] (~lev@
21:58:26 Quit saratoga_ (Quit: Page closed)
21:59:50 Quit shai (Ping timeout: 245 seconds)
22:01:42 Join MethoS- [0] (~clemens@
22:03:03 Join shai [0] (
22:10:36Blue_Dude*Still building* eabi tools in Cygwin...
22:11:37Blue_DudeAnd here I was thinking only a few minutes.
22:12:51 Quit Blue_Dude (Quit: ChatZilla 0.9.86 [Firefox 3.6.3/20100401080539])
22:13:56 Join m0ar [0] (
22:14:41m0arWhen trying to set the YellowBarDynamic WPS I just get reverted back to the original theme, any ideas?
22:15:15mc2739m0ar: which version of Rockbox are you using?
22:15:41m0armc2739: Todays
22:15:51m0arJust upgraded, thought it might have been that
22:16:40m0arSame thing with dfkt minimum clip
22:16:43m0arSo that's wierd
22:17:39mc2739the syntax for the theme files was changed after Rockbox 3.6 was released. You need to download a new version of the theme from the theme site.
22:17:53m0arJust did
22:18:04m0arThe RBU isn't using the new files? :)
22:18:29mc2739not unless you have a recent svn build
22:18:38m0arThen there's the problem
22:18:45m0arThanks a bunch
22:18:55m0arTime to update the RBU then?
22:19:03mc2739it is in the works
22:19:25*m0ar sends some e-love to all the devs of RB
22:20:01mc2739in the meantime, go to and download an updated theme and unzip it to your device
22:20:43m0arYeah, already on it
22:22:35m0arRunning svn-stuff in windows, which I barely visits, isn't worth it
22:22:59m0arAnd I spilled a glass of water in my linuxlaptop last night, so now I'm stuck hehe
22:23:33m0arNow the themes works just fine, thanks dude
22:24:30mc2739you're welcome, FYI, you can find a link to recent Rockbox Utility builds here:
22:27:05m0arNot a nessicity, thanks anyways
22:28:05 Quit m0ar (Quit: CGI:IRC)
22:41:31 Quit petur (Quit: Zzzzz)
22:48:22 Quit stripwax (Quit:
22:56:00***Saving seen data "./dancer.seen"
23:02:11 Join ender1 [0] (
23:05:12 Quit grndslm (Read error: Connection reset by peer)
23:05:31 Quit ender` (Ping timeout: 258 seconds)
23:06:16 Join Buschel [0] (
23:10:55 Quit Luca_S (Quit: CGI:IRC (EOF))
23:14:17 Quit The_Seven (Remote host closed the connection)
23:17:26 Quit Buschel (Ping timeout: 240 seconds)
23:18:18 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
23:23:29 Join kramer3d_ [0] (~kramer@unaffiliated/kramer3d)
23:25:05 Join stripwax [0] (
23:28:35 Join stoffel [0] (
23:29:41 Quit Guest43341 (Read error: Connection reset by peer)
23:33:00 Quit stoffel (Remote host closed the connection)
23:42:43 Quit hebz0rl (Ping timeout: 276 seconds)
23:48:16literalhm, just hit a bug. I turned on the display on my Clip+ and it shows the (paused) track that I was previously playing
23:48:55literalnot the one it's currently playing, which is even in another playlist/album
23:49:27 Join moparx [0] (~moparx@unaffiliated/moparx)
23:49:41literaleven if I navigate the menus, or pause/play, it still shows the old paused track
23:49:47literalin the WPS
23:50:05CIA-6New commit by gevaerts (r26835): Build D2 with the old compiler again (builds)
23:50:11CIA-6New commit by gevaerts (r26836): Build D2 with the old compiler again (configure)
23:50:14literalthe problem "fixed" itself when the current track ended
23:51:54CIA-6r26836 build result: All green
23:51:57 Nick CGL is now known as CGL_emo_WTF (~CGL@
23:54:47 Quit ender1 (Quit: To err is human; to moo bovine.)
23:56:27 Quit stripwax (Quit:

Previous day | Next day