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 2012-07-26

00:06:17bertrikI see a couple of malloc and realloc only at the start of the decoding, so it seems to stabilize
00:07:38gmaxwellbertrik: right, it shouldn't need to allocate more than the maximum packet + maximum page size.
00:08:06gmaxwell(at least with normal sane usage).. but it will still use a bunch more if every realloc wastes memory.
00:10:16bertrikHow many allocators do we have in rockbox? I know we have at least buflib and the naive codec_malloc stuff, and there's a TLSF allocator somewhere too.
00:13:31bertrikimplementing realloc as malloc + memcpy seems to make things work
00:16:42gmaxwelloh duh, that too.
00:17:01 Join kevku [0] (
00:18:38 Quit benedikt93 (Quit: Bye ;))
00:19:29 Quit Totalled (Quit: PETTAN PETTAN, TSURUPETTAN!)
00:20:14bertrikwell, at least I've identified on of the main problems
00:23:41 Quit TheLemonMan (Ping timeout: 244 seconds)
00:27:18bertrikI wonder if I can pre-init ov->storage for libogg (set it to twice the chunk size)
00:28:07 Join Scromple [0] (~Simon@
00:28:07bertrikgmaxwell, you said it needs max packet + max page size at most?
00:28:53bertrikAnd a page is about 64 kB (slightly less). What's a sane limit on the max page size?
00:29:05bertrik*max packet size
00:31:29 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
00:35:07 Quit ender` (Quit: The early bird may get the worm, but the second mouse gets the cheese.)
00:38:21bertrikmight be worth to create a single malloc-free optimised ogglib for rockbox, ogg is used by three codecs now: speex, vorbis and opus
00:39:50bertrikand perhaps create a single seek function for ogg too
00:48:36 Quit prof_wolfff (Read error: Operation timed out)
00:56:11gmaxwellbertrik: ogg2 basically gives you that. but it also needs codes that access it through special functions (that do the scatter-gather so there isn't any need for memory to reassemble)
00:57:02gmaxwellinterestingly, if your ogg layer can't handle big pages then it would also fail for some vorbis files (E.g. ones emitted by current ffmpeg).
00:59:24 Quit Rower85 (Quit: Hmmm...)
00:59:24bertrikyeah, I was wondering about possible current problems with vorbis too (although we rarely see problems with that in practice)
01:00:13 Quit Wardo (Read error: Connection reset by peer)
01:00:23bertrikthanks so far for you patience by the way :)
01:02:59 Join user485763 [0] (~aoeu@
01:18:39 Quit mikroflops (Read error: Operation timed out)
01:19:28 Quit user485763 (Read error: Operation timed out)
01:19:34 Quit mgottschlag (Ping timeout: 276 seconds)
01:36:07bertrikhm, for all of the AMSv2 players, we seem to make a reservation of 0x8000 bytes in the codec buffer (most targets use 1 MB)
01:37:50***Saving seen data "./dancer.seen"
01:56:06 Join mikroflops [0] (
02:13:04 Quit linuxstb (Quit: This computer has gone to sleep)
02:24:31 Join linuxstb [0] (
02:49:41 Join YungHumma [0] (
02:50:17YungHummahi using a ipod nano 2nd gen, installed latest few days ago now i can't access it via usb, says not writeable
02:50:37YungHummaI can still play music and browse the hard drive on the nano
02:51:09YungHummaany suggestions?
02:55:10 Quit lebellium (Quit: ChatZilla [Firefox 15.0/20120717110313])
03:07:57YungHummadevice shows up in my computer but its empty...I can still see files and browse on the ipod nano itself
03:08:00YungHummareally strange
03:08:18 Quit linuxstb (Quit: This computer has gone to sleep)
03:18:29YungHummaok i fixed it
03:18:53YungHummareseted with select + menu....when apple logo shows select + play to disk mode
03:18:57YungHummaand then ya works now
03:19:07YungHummarestoring ipod nano with itunes now :(
03:19:11 Quit YungHumma ()
03:19:29 Quit bertrik (Ping timeout: 246 seconds)
03:37:53***Saving seen data "./dancer.seen"
04:03:22 Join TheSphinX^ [0] (
04:04:02 Quit mikroflops (Ping timeout: 276 seconds)
04:04:41 Quit zz_TheSphinX^ (Read error: Operation timed out)
04:07:38 Quit TheSeven (Read error: Operation timed out)
04:08:03 Join kevinconnellu [0] (
04:10:34kevinconnelluguys im going to make a doom map with this: and try out the WAD ON MY IPOD NANO
04:10:51kevinconnelluSSR ABOUT CAPS
04:10:59kevinconnelluthere better
04:12:15 Join mikroflops [0] (
04:13:35kevinconnellumikroflops hey
04:14:39JdGordonkevinconnellu: this is NOT a social channel, rockbox talk only (and doom mods dont count)
04:16:07 Quit kevinconnellu (Quit: CGI:IRC (EOF))
04:38:37 Join kevinconnellu [0] (
04:39:29kevinconnelluok how do i get in #rockbox-community
04:39:32 Quit amiconn (Disconnected by services)
04:39:34 Join amiconn_ [0] (amiconn@rockbox/developer/amiconn)
04:39:38 Nick amiconn_ is now known as amiconn (amiconn@rockbox/developer/amiconn)
04:40:02 Quit pixelma (Disconnected by services)
04:40:03 Join pixelma_ [0] (pixelma@rockbox/staff/pixelma)
04:40:05 Nick pixelma_ is now known as pixelma (pixelma@rockbox/staff/pixelma)
04:41:49 Quit kevinconnellu (Client Quit)
04:42:21 Join kevinconnellu [0] (
04:43:27 Quit kevinconnellu (Client Quit)
04:45:05 Quit XavierGr ()
04:53:57 Join CaptainKewl [0] (
04:59:17 Join Totalled [0] (
05:37:56***Saving seen data "./dancer.seen"
06:31:58 Join Rower85 [0] (
06:53:22 Quit mikroflops (Read error: Operation timed out)
07:06:38 Join mikroflops [0] (
07:15:27 Quit CaptainKewl (Ping timeout: 255 seconds)
07:37:58***Saving seen data "./dancer.seen"
07:48:56 Quit dys (Ping timeout: 272 seconds)
07:57:40 Quit mystica555 (Remote host closed the connection)
07:59:49 Join mystica555 [0] (
08:01:13 Join stoffel [0] (
08:13:19 Quit derf (Ping timeout: 252 seconds)
08:13:26 Join derf [0] (
08:27:02 Quit Rower85 (Quit: Hmmm...)
08:33:44 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
08:38:15 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
08:42:56 Quit Totalled (Quit: PETTAN PETTAN, TSURUPETTAN!)
08:55:13 Join Totalled [0] (
08:58:41 Join ender` [0] (
09:01:31 Join mgottschlag [0] (~quassel@reactos/tester/phoenix64)
09:07:54 Join wodz [0] (
09:12:47 Join poodle_ [0] (a134a605@gateway/web/freenode/ip.
09:13:10 Part poodle_
09:13:35 Join poodle_ [0] (a134a605@gateway/web/freenode/ip.
09:14:22 Quit poodle_ (Client Quit)
09:15:05 Quit gxk (Ping timeout: 264 seconds)
09:15:33 Join CaptainKewl [0] (
09:22:53 Join einhirn [0] (
09:25:59JdGordonDoes anyone have issues with using long press OK (context menu) to remove a shortcut?
09:26:12 Quit wodz (Ping timeout: 244 seconds)
09:27:05 Join leavittx [0] (~leavittx@
09:27:42 Join T44 [0] (
09:27:45 Quit Topy (Read error: Connection reset by peer)
09:30:13 Quit CaptainKewl (Read error: Operation timed out)
09:32:41 Join Poodlemastah [0] (
09:33:53 Quit Totalled (Quit: PETTAN PETTAN, TSURUPETTAN!)
09:34:29 Join Keripo [0] (
09:38:01***Saving seen data "./dancer.seen"
09:38:12 Join Totalled [0] (
09:43:24 Quit Keripo (Read error: Connection reset by peer)
09:46:36 Join Keripo [0] (
09:47:54 Quit joe_ (Ping timeout: 250 seconds)
09:48:36 Join joe_ [0] (
09:49:05JdGordonthe links to the manuals are all dead?
09:49:35JdGordonah, all the html links are dead, pdfs are available though
09:50:05Zagoroh right, gevaerts reported that the html builds fail since a while. I need to look at that.
09:52:23CIA-5Commit 4348b45 in rockbox by Jonathan Gordon: shortcuts: Remove the selected shortcut with the context menu button
09:54:47CIA-54348b45 build result: All green
09:57:23 Join CaptainKewl [0] (
09:57:44 Quit Keripo (Read error: Connection reset by peer)
10:02:43 Join linuxstb [0] (
10:08:57 Quit Scromple (Read error: Connection reset by peer)
10:18:14 Quit CaptainKewl (Ping timeout: 250 seconds)
10:22:17 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
10:23:49kugelJdGordon: what's wrong with permill
10:28:08kugelit's a proper English word isn't it (per mill actually)?
10:33:35JdGordonI've never heard of it
10:34:06bertrikopus metadata parsing works now too
10:34:24bertrik(it's basically vorbis comments with a slightly different header)
10:36:19JdGordonkugel: I think it isnt really obvious, percent (even if it is a technical lie) is more understandable imo
10:36:22*JdGordon dinner
10:36:49Zagor"The term is so rarely used in English that major dictionaries do not agree on the spelling or pronunciation" −− Wikipedia
10:37:26 Join TheLemonMan [0] (~LemonBoy@unaffiliated/thelemonman)
10:39:25 Quit linuxstb (Quit: This computer has gone to sleep)
10:55:56 Join user485763 [0] (~aoeu@
11:04:52 Join dys [0] (~user@2a01:1e8:e100:8296:21a:4dff:fe4e:273a)
11:06:54kugelso better don't use it?
11:26:16 Join lebellium [0] (
11:33:30bertrikdoes a codec get re-loaded for every track, or only when switching tracks that use different codecs?
11:37:28 Join evilnick [0] (~evilnick@rockbox/staff/evilnick)
11:38:02***Saving seen data "./dancer.seen"
11:53:47 Quit T44 (Read error: Connection reset by peer)
11:54:19 Join pamaury [0] (
11:54:20 Quit pamaury (Changing host)
11:54:20 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
11:54:39Tornebertrik: only when switching to a different one, i'm pretty sure
11:55:07 Join Topy44 [0] (
11:55:24bertrikok thanks, makes sense
11:55:44Torneiirc buffering sticks a copy of the codec in the buffer in between the tracks
11:56:52 Quit Topy44 (Read error: Connection reset by peer)
12:00:32 Join Topy44 [0] (
12:01:48 Join T44 [0] (
12:05:05 Quit Topy44 (Ping timeout: 240 seconds)
12:07:24 Quit user485763 (Quit: Leaving.)
12:11:01 Join mpescador [0] (
12:13:55mpescadorhi all - just put EmCORE/Rockbox on a iPod Classic (6th Gen - 80Gb) and it looks good and functional - but i cannot find any mention anywhere about whether RockboxUtility will support Classic/6gen anytime soon - or at all?
12:17:13 Join user485763 [0] (~aoeu@
12:20:25 Quit mikroflops (Read error: Operation timed out)
12:27:23 Join mikroflops [0] (
12:27:59 Join linuxstb [0] (
12:35:49 Quit stoffel (Ping timeout: 244 seconds)
12:39:19CIA-5Commit 132fcca in rockbox by Bertrik Sikken: Make codec_realloc retain the contents when resizing a memory segment
12:41:36CIA-5132fcca build result: All green
12:45:11 Quit mgottschlag (Remote host closed the connection)
12:48:48 Quit mpescador (Quit: CGI:IRC (EOF))
12:50:47funmanbertrik: doesn't it miss the case were new size <= old size ?
12:51:43 Quit linuxstb (Quit: This computer has gone to sleep)
12:52:22bertrikactually, I think that should work
12:52:41funmanmemcpy would overflow the new memory
12:52:49funmanwait no
12:53:03funmanit would overread but yeah that would work
12:53:31funmanbut if we know old size we could reuse the memory and spare a codec_malloc() call
12:54:30bertrikthe current codec_realloc is very primitive, a real realloc should also be able to malloc (if ptr is NULL and size > 0) and free (if ptr is valid and size = 0)
12:59:52 Join Topy [0] (
13:02:32 Join Topy44 [0] (
13:02:45 Quit T44 (Ping timeout: 240 seconds)
13:04:05 Quit Topy (Ping timeout: 240 seconds)
13:27:10 Quit evilnick (Quit: Leaving)
13:27:35 Quit mikroflops (Ping timeout: 250 seconds)
13:36:12 Join dfkt [0] (dfkt@unaffiliated/dfkt)
13:38:04***Saving seen data "./dancer.seen"
13:45:23TheLemonManhrm, any reason not to have the fm driver remember the current frequency ?
13:48:35 Join mgottschlag [0] (~quassel@2a00:1398:200:200:18e3:daff:dfa0:9734)
13:48:35 Quit mgottschlag (Changing host)
13:48:35 Join mgottschlag [0] (~quassel@reactos/tester/phoenix64)
13:54:14 Join XavierGr [0] (~xavier@rockbox/staff/XavierGr)
13:58:42 Join stoffel [0] (
13:59:33bertrikI think it does, depending on what you mean exactly
14:00:56TheLemonManout of boredom i wanted to attempt a radio cleanup
14:01:29TheLemonManthe preset code uses the curr_freq variable declared in radio.c
14:02:03 Quit mgottschlag (Ping timeout: 246 seconds)
14:04:15TheLemonManand i guess that next_station with argument 0 just tunes into the curr_freq
14:06:16 Join mikroflops [0] (
14:23:08 Join mgottschlag [0] (~quassel@2a00:1398:200:200:18e3:daff:dfa0:9734)
14:23:08 Quit mgottschlag (Changing host)
14:23:08 Join mgottschlag [0] (~quassel@reactos/tester/phoenix64)
14:35:11 Quit factor (Quit: Leaving)
14:42:41 Join CaptainKewl [0] (
14:49:23bertrikfunman, updated the opus patch, should work now
14:51:06gmaxwellbertrik: great!
14:51:07funmanthanks, i'm going to test on simulator
14:52:21funmangmaxwell: do you have a zip with all test vectors ?
14:57:44funmanworks fine except all the decoder features tested are incorrect (preskip, trimming, replaygain)
15:00:27gmaxwellThe gain is not "replaygain", it's opus header gain: Should be one loc to fix that. opus_decoder_ctl(st,OPUS_SET_GAIN(header.gain));
15:00:37gmaxwell(it's a 16 bit value in the first packet)
15:00:58Zagorbertrik: have you joined the ietf workgroup list?
15:01:38gmaxwellThe WG list is mostly dead because the standard is done.
15:02:22gmaxwellfunman: pretty spiffy. What simulator is it working in?
15:03:53funmani think there's no difference between them but i used clip zip
15:04:43Zagorgmaxwell: oh, is the bitstream frozen?
15:04:58gmaxwellZagor: Yes.
15:05:47funmanand valgrind is happy
15:07:31gmaxwellfunman: Can it play ? (warning: 2.2GiB file)
15:07:48funmanhm i think we don't read replaygain tags for vorbis either
15:08:25funmangmaxwell: it plays the first mb
15:08:45funmani think we use OS stack on the simulator though
15:10:30gmaxwellfunman: Most things dont; but the opus header gain is _not_ replaygain. It's a part of the codec setup header thats mandatory for correct playback. (it makes it possible to losslessly adjust gain on opus files for simiar effect to mp3gain and aacgain). At the moment every available opus decoder handles it. (firefox, gstreamer, foobar2000, the commandline tools, ffmpeg's patches)
15:11:22funmanhm so we should read the tags before initialising opus decoder?
15:11:34gmaxwellIt's not part of the tags.
15:12:01Zagor15:00 <gmaxwell> (it's a 16 bit value in the first packet)
15:12:04funmanwhere is it? if it's inside the codec then libopus should get them no?
15:12:29gmaxwellIt's part of the codec setup header, the packet that identifies the codec and provides the channel count.
15:12:44funmanok that's weird then
15:13:16gmaxwellfunman: Hm?
15:13:24funmani'm trying gain_silent_output vector
15:13:34funman"if this recording is not VERY quiet..."
15:14:07gmaxwellSure, the code just isn't calling the CTL to pass the value in.
15:15:05gmaxwellThe codec itself doesn't read the header, because you need to read it before you init the thing since it changes the memory usage. (Though for rockbox its not insane to always just allocate the worst case, but thats not the norm on a desktop player)
15:15:46funmanah so we should parse first packet and use OPUS_SET_GAIN() ?
15:15:48 Quit XavierGr (Disconnected by services)
15:15:50 Join XavierGr [0] (~xavier@rockbox/staff/XavierGr)
15:17:12gmaxwellYou also need to figure out what you're going to do with surround files. Ideally you should be downmixing them and playing them, but I don't know if your targets have enough cpu power for that.
15:18:16funmani don't know what we do for other codecs
15:18:41funmani see mention of downmixing for flac/a52
15:20:48Zagorsome targets have plenty of cpu, others don't. ape is a good example of a codec that makes some targets choke on high bitrates.
15:22:43Zagorape_c5000.ape is 2.59% realtime on ipod nano 2g :-)
15:22:58 Quit CaptainKewl (Ping timeout: 255 seconds)
15:25:37gmaxwellokay, I suggest you copy opus_header.c from opus_tools to parse the header; and since surround decoder+downmix should work on at least some targets, switch to opus_decode_multistream (same invocation, though the init needs more values from the header), and add a downmix loop.
15:26:09gmaxwell(generall all file based users of opus should be using opus_decode_multistream instead of opus_decode)
15:26:31gmaxwell(the difference between the two is being able to support surround)
15:27:26funmangmaxwell: we parse it already
15:27:52gmaxwellah! Sorry, the comment above wrt gain made me think you weren't.
15:28:09funmanwe just don't call the right CTL
15:29:44funman works fine
15:31:58gmaxwellThe pre-trim is only a bit more work, as you just need to skip the first N samples out, though you also need to make your seeking logic aware of the time offset.
15:32:24funmanaw ok
15:32:47gmaxwellAlso, seeks in opus need to discard audio after the seek (otherwise you can potentially get loud chirps). This may be a moderate PITA depending on your infrastructure.
15:33:21gmaxwell(the OggOpus spec recommends seeking 80ms early and dropping it)
15:38:06***Saving seen data "./dancer.seen"
15:45:32 Join evilnick [0] (~evilnick@rockbox/staff/evilnick)
15:49:11 Join linuxstb [0] (
15:50:13funmanbertrik: how do i send you this change through gerrit? i don't know if i can push a new patchset on the same number
15:50:26gevaertsYou can
15:51:43bertrikyou can hack the changeid
15:52:31 Join SBookman [0] (
15:54:22SBookmanGentlemen, brand new to IRC, I have a Sansa Clip Zip Rockbox problem, anyone out there? Thankyou.
15:54:51ZagorSBookman: explain your problem and someone might be able to help
15:56:17 Quit user485763 (Ping timeout: 255 seconds)
15:57:37SBookmanThankyou. Installed Rockbox, and whilst reading manual, maybe I pushed a wrong button whilst connecting to USB to recharge WHILST listening to my music, now screen reads two lines in white on black; First line: Undefined instruc[truncated]; second line: bt pc: 30800000. Will do nothing: cannot switch off, cannot be read by PC USB.
15:58:36ZagorSBookman: that's a software crash. hold power off ~10 seconds and it should work
16:00:18SBookmanSir, most excellent, all is resumed! Most kind. What can I do in return for this service?
16:00:47Zagornothing, enjoy rockbox! :-)
16:01:05ukleinekSBookman: alternatively test rockbox and report problems :-)
16:01:56SBookmanNothing? A man seeks to return a favour, and you can think of, wish for nothing?
16:03:00Zagorhaha. well if you want, we have a donation button on the web page.
16:03:31SBookmanSir, how do I get to it?
16:03:41Zagorand as ukleinek says, we welcome contributors of all kinds
16:04:08Zagorand click on the "paypal donate" button in the bottom of the left menu
16:04:22SBookmanSir, contributors? How can one contribute?
16:04:30 Quit mikroflops (Ping timeout: 246 seconds)
16:05:28ukleinekSBookman: check out the source code, compile it, notice problems, fix them and send a patch
16:06:01ZagorSBookman: lots of ways, depending on your skills and motivations. if you are not a programmer you can always help testing bug reports, improve the manuals and help keep the wiki information updated
16:13:41SBookmanDonation given. ukleinek: I am now an old grey one, who once had an IQ of 150, I once knew only fortran and cobol. Zagor: I have rewritten help manuals; I once spoke fluent, reasoned coherent, read proper, and wrote very neat; how may I, specifically, improve the manuals?
16:15:57Tornecheck out the source code, modify the text, send a patch :)
16:16:08Tornethe manual is written in LaTeX and is in the same tree as our source
16:18:03ZagorSBookman: we're a project of autonomous volunteers, each choosing to work on what they find interesting. we don't have much work assignments going on. working on the manuals is a suggestion because as a newcomer you may have fresh eyes and can point out areas that seem obvious to us old hands but are not to novices.
16:19:27Zagorbasically read the manuals, make notes where they are unclear or unsufficient and suggest improvements. either by mail or ideally by patches to the latex source.
16:20:18Zagorand ask here when you need help
16:20:38bertrikvery useful, (but possible also a bit boring) would be to make screenshots of various things in the UI for the user manual, like the plugins
16:21:11SBookmanZagor: Will read wordmeal the pdf manual, make notes, polish where necessary, submit same. Sir, "patches to the latex source", escapes my understanding, but not the word mail.
16:21:44Zagorexcellent. welcome to the project!
16:27:26 Join MethoS- [0] (~clemens@
16:29:00 Quit pamaury (Remote host closed the connection)
16:31:34 Join lebellium_ [0] (
16:34:07 Quit lebellium (Ping timeout: 276 seconds)
16:34:13 Nick lebellium_ is now known as lebellium (
16:34:35 Join mikroflops [0] (
16:36:57 Join y4n [0] (~y4n@unaffiliated/y4ndexx)
16:37:24SBookmanGentlemen, is there a "The Rockbox Manual for Sansa Clip Zip" pdf, [mine, just downloaded, is dated July 26, 2012], which helpfully contains the images in place of: "WARNING! Image not found"?
16:39:16Tornepossibly nobody has made images for the zip yet :)
16:39:25Tornealternatively, it's broken.
16:39:31Tornethere's not a different download :)
16:42:12 Quit SBookman (Quit: CGI:IRC (EOF))
16:53:17 Join WalkGood [0] (~4@unaffiliated/walkgood)
17:00:56 Join hype [0] (~hype@
17:01:23 Quit Zagor (Quit: Clint excited)
17:19:44 Join prof_wolfff [0] (
17:24:28 Join pretty_function [0] (~sigBART@
17:25:15 Join anewuser [0] (~anewuser@
17:25:16 Quit anewuser (Changing host)
17:25:16 Join anewuser [0] (~anewuser@unaffiliated/anewuser)
17:26:05 Join anewuser_ [0] (~anewuser@
17:30:04 Quit anewuser (Ping timeout: 264 seconds)
17:30:40 Nick anewuser_ is now known as anewuser (~anewuser@
17:30:41 Quit anewuser (Changing host)
17:30:41 Join anewuser [0] (~anewuser@unaffiliated/anewuser)
17:35:46 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
17:38:10***Saving seen data "./dancer.seen"
17:39:09 Join amayer [0] (
17:43:37 Quit amayer (Read error: Connection reset by peer)
17:50:12 Join amayer [0] (
17:52:04 Join n1s [0] (
17:52:04 Quit n1s (Changing host)
17:52:04 Join n1s [0] (~n1s@rockbox/developer/n1s)
17:57:36 Join Strife89 [0] (~Strife89@
18:12:53amayeris rockbox supposed to stay on the "usb connected" screen after eject?
18:12:59 Quit n1s (Read error: Connection timed out)
18:13:38 Quit pretty_function (Remote host closed the connection)
18:15:53 Quit amayer (Remote host closed the connection)
18:17:08funmanno, your clip probably crashed so you should turn it off
18:17:34 Join amayer [0] (
18:24:42amayerfunman: im using an ipodclassic
18:25:15Tornewhen you say eject, you men just ejecting it from the OS, and leaving it plugged in?
18:25:38amayerwhen i disconnect it rockbox resumes normal use. but if i eject it and leave it connected(to charge) the "usb connected" screen stays there
18:25:41Torneif so then, well, it's not really *supposed* to, but that is hwo it behaves :)
18:25:57Torneeject is a weird and wonky concept in usb storage and it appears that at some point we changed how we handle it
18:26:10Torneso it's not unexpected or new
18:26:15Tornebut it isn't really correct either.
18:27:46amayeri was just wondering. normally its not a problem. but it would have been more convenient today when adding alot of files and needing to keep charging
18:36:09slowdow... it'd be nice to just be able to connect it! ;)
18:36:11 Quit stoffel (Ping timeout: 272 seconds)
18:39:38Torneamayer: you can charge without going into usb mode by holding any button
18:39:49Torneso yeah, unfortunately what you need to di is eject and remove it, then hold a button and plug it back in
18:40:03Tornei dunno if there is a bug open for this regression, but it's been this way for a long time :/
18:40:08Tornenot just on the ipod classic, either.
18:40:15Torneit depends what OS you have, also
18:40:23Tornethere are multiple different ways to "eject" a usb storage device, it's not well defined
18:40:32Torneand i believe some of them do indeed kick rockbox out of usb mode
18:41:01amayeri understand that. i do the hold any button and connect thing all the time.
18:41:03amayeri was just wondering if there was a way to return to normal operation after OS eject
18:41:43Torneso yeah. no, there isn't, but there should be :)
18:41:56Tornesee if we have a bug open for it, and if not raise one
18:42:01Tornebecause otherwise nobody will remember :p
18:42:44amayerhaha ok
18:46:45*gevaerts isn't convinced that this is a bug
18:47:40Tornegevaerts: i *think* it's since we started doing HID
18:47:49Torneand yes, it may not have been considered a bug by the person changing the behaviour
18:48:01Tornebut in the absence of any other way to exit the blocking USB screen I think it's wrong
18:48:04Torneand it's annoyed me before :)
18:48:18gevaertsIt's older than HID
18:48:21 Quit mgottschlag (Ping timeout: 246 seconds)
18:48:26Tornethen what were you thinking of? :)
18:48:34 Quit Strife89 (Quit: Vamoose.)
18:48:41gevaertsTell me how to do it properly, and I'll fix it :)
18:48:55Tornei dunno.
18:49:00amiconngmaxwell: Btw, rockbox won't play >= 2 GiB files, at least on target.
18:49:01Tornethere's the different meanings of eject..
18:49:14Tornewhich kind of removable are we? :)
18:49:26gevaertsDepends on the target, actually, IIRC :)
18:49:33Tornewell okay
18:49:39Tornei mean, on targets that don't have cardslots
18:49:41gevaertsWhich doesn't exactly help
18:50:18Torneanyway. what i want is for my ipod (no slot) to drop out of usb mode when you do whatever it is windows does when you "safely remove device"
18:50:21gevaertsHm, actually, dese days we're a SCSI removable device everywhere
18:50:22Torneit doesn't right now
18:50:52gevaertsI know it doesn't, but I've never managed to figure out how exactly we can detect this
18:51:11Torneit used to work! :)
18:53:16amiconnWindows does different things for "safe removal" depending on the windows version
18:53:25gevaertsAs far as I can figure out, we never changed how things behave there
18:53:31amiconnBehaviour changed between XP and Vista
18:53:36gevaertsExcept for gigabeat S, IIRC
18:54:10gevaertsI *believe* at least some systems suspend the device on eject, but I'm not sure
18:54:16amiconnIirc iriver OF detects safe removal from XP, but not vista or higher
18:54:19Torneamiconn: that's also possible :)
18:54:29*gevaerts challenges Torne to provide the revision that changed this :)
18:54:43Tornei don't think i have an XP machine any more
18:54:46Torneto test that on
18:54:57Tornebut in any case, it should realyl work on all OSes :p
18:55:25*amiconn would like to find the difference and fix behaviour in rb (if possible, which might not be the case with hardware bridges)
18:55:53amiconnTorne: There are these nice things called VMs...
18:56:11amayergevaerts: so you think this isnt a bug?
18:56:13amayershould i post something to the bug tracker or not?
18:56:39Torneamayer: it is a bug
18:56:46amayerhaha ok
18:56:53gevaertsIt's not a bug IMHO :)
18:56:54Torneeven if it's caused by windows changing its behaviour, it's still our bug
18:57:05Tornebecause we should handle *every* reasonable thing :)
18:57:08amayerwell im on linux and it doesnt work
18:57:20amayer*takes back "doesnt work"*
18:57:29amayerit doesnt act as i would expect it to act
18:57:33gevaertsI'd say if you know *exactly* what we're not doing, file a bug
18:58:09amayeri dont know on a code level. but i plan to look at it in my spare time(some time this weekend)
18:58:12Torneyou're on linux?
18:58:22Torneagain, we should still handle all reasonable behaviour
18:58:34Tornebut on linux there is the caveat that linux does have an unmount operation that is distinct from ejecting
18:58:37gevaertsSure, so tell me what that means :)
18:58:39Torneand we shouldn't exit usb mode on that.
18:58:57Torneso just to check: you are actually doing some UI action that implies it's ejecting/disconnecting/etc the device, not just unmounting it?
18:58:58gevaertsunmount isn't something we notice
18:59:10Tornei don't remember offhand whether this works for me on linux or not :)
18:59:10 Join Thra11 [0] (~thrall@
18:59:17 Quit mikroflops (Read error: Operation timed out)
18:59:28linuxstbCouldn't we (excluding me) just log what the OS is doing?
18:59:40Torneyes, we (excluding everyone here who is too busy) could do that.
19:00:20gevaertsWell, unmount does a SCSI_ALLOW_MEDIUM_REMOVAL, but things like the partition table scanning also do that, so it doesn't mean much
19:00:34Torneyah, but that's just for a drive lock
19:00:40Tornedifferent semantics
19:00:45amayerim using Ubuntu 12.04.
19:00:48amayerin the file browser(nautilus) i click the eject symbol next to "ipodclassic"
19:00:50amayerive also right clicked on the launcher(hovering over the "ipodclassic") and clicked "eject"
19:01:02Torneincidentally if we are presenting as a removable device in the sense of a thing with a card slot in it, on devices that don't have one
19:01:05gevaertslinuxstb: yes and no. We can log SCSI commands, but I suspect things like USB suspend are involved in at least some of the cases, and we don't handle that
19:01:06Tornei would suggest that's wrong
19:01:22Torneusb removable is so weird.
19:01:35gevaertsTorne: I would strongly agree, but that makes just two of us I fear
19:02:34gevaertsSee e.g. 81d3a029
19:03:12amayerTorne: so to answer your question i believe i am ejecting it but what the "eject" command does as far as eject vs unmount idk
19:03:26gevaertsThat was early days though, maybe the osx issues hinted at there were actually caused by other stuff
19:03:27Torneamayer: what you believe you're doing is the significant thing to deciding if it's a bug
19:03:49Torneamayer: what it's atually doing under the hood is significant to working out how to *fix* it :)
19:04:00 Join pretty_function [0] (~sigBART@
19:04:31amayerwell im clicking eject(in both cases)
19:04:33amayeridk what going on under the hood
19:04:35amayerso to my knowledge i am ejecting the device
19:05:02amayerwith OF if would return to normal operation
19:05:04amayerso im pretty sure it is actually ejecting
19:07:50gevaertsamayer: can you run eject -v on it?
19:08:15gevaertsI'd like to see which methods it tries
19:08:18 Join anewuser_ [0] (~anewuser@
19:11:07 Quit anewuser (Ping timeout: 252 seconds)
19:11:20bertrikWhat is the problem really? Do you want to use the rockbox UI once it's been "safe removed"?
19:11:47gevaertsbasically, yes
19:12:29gevaertsWell, the *real* problem is that Torne and I can't agree on whether this is a feature request or a bug :)
19:13:06*gevaerts has lots of vague fragments of recollection about this, but he doesn't remember the full picture
19:15:03gevaertsI believe one of the tricky issues is that if you do have a card slot, you want "eject" to mean "safe to remove the card", not necessarily "disconnect everything", or you can't plug in another card anymore
19:15:58 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
19:16:26gevaertsThat's easy to fix by not misusing DEVICE_REMOVABLE of course, but then you run into other issues with people (including those involved in OSes) totally misunderstanding what that's supposed to mean
19:19:40bertrikAnyone here familiar with the lrcplayer plugin and with lyrics tags?
19:24:13 Join mikroflops [0] (
19:24:27[Saint]I just tried here with my iPod Color on Ubuntu 12.10, if I *eject* the device, it stays in the USB screen, if I *safely remove* the device, it does as expected.
19:24:35[Saint]Torne: gevaerts: ^
19:25:08gevaerts[Saint]: please don't say things like "as expected" in this context. That's *vague*!
19:25:25[Saint]where "as expected" == "drops out of the USB screen"
19:25:31[Saint]hey, I was fixing that!
19:27:53*gevaerts has no idea how this works...
19:31:01 Quit WalkGood (Quit: ♪ ♫ ♪ ♫ ♪ ♫ ♪)
19:37:45 Join advcomp2019__ [0] (
19:37:45 Quit advcomp2019__ (Changing host)
19:37:45 Join advcomp2019__ [0] (~advcomp20@unaffiliated/advcomp2019)
19:37:49 Join stoffel [0] (
19:38:13***Saving seen data "./dancer.seen"
19:38:43 Quit fs-bluebot (Ping timeout: 276 seconds)
19:40:14 Join fs-bluebot [0] (
19:40:58 Quit advcomp2019_ (Ping timeout: 260 seconds)
19:43:31 Quit Thra11 (Read error: Operation timed out)
19:45:30 Join Thra11 [0] (
19:50:29 Join Wardo [0] (
19:55:22amayergevaerts: how do i get a list of devices connected to my computer?
19:55:24amayeri tried dmesg but it just says sdb:
19:55:26amayerit doesnt give me a path such as /dev/...
19:56:30amayerok its /dev/sdb
19:56:58amayeri ran eject -v /dev/sdb and it output this:
19:57:01amayereject: device name is `/dev/sdb'
19:57:03amayereject: expanded name is `/dev/sdb'
19:57:05amayereject: `/dev/sdb' is mounted at `/media/iPodClassic'
19:57:07amayereject: unmounting device `/dev/sdb' from `/media/iPodClassic'
19:57:09amayereject: `/dev/sdb' is a multipartition device
19:57:11amayereject: unable to open `/dev/sdb'
19:58:12 Quit linuxstb (Read error: Connection reset by peer)
20:00:15 Join linuxstb [0] (
20:06:56bertrikWhat's the IRC nick name again of Teruaki Kawashima ?
20:07:36funmanbertrik: i have no warnings when building opus
20:07:40funmanbertrik: teru iirc
20:07:55funmandocs/COMMITTERS confirms
20:08:01bertrikfunman, I got some w.r.t. _MSC_VER or something like that, and also for lrint
20:08:14bertriklogbot has never seen teru
20:08:36bertrikcan't exactly remember seeing him here much either
20:08:37funmanwhen building for sim or target?
20:08:43bertrikfor target I think
20:08:53funman14:08 -NickServ(NickServ@services.)- Last seen : Mar 30 13:20:44 2012 (16 weeks, 6 days, 04:48:03 ago)
20:14:29 Join n1s [0] (~n1s@rockbox/developer/n1s)
20:15:27amee2kdoes rockbox have a feature that i'd call a soft key lock? i.e. like mobile phones without a dedicated lock switch have
20:15:44amee2kwhere you have to press a simple 2 or 3 button combination within a few seconds to unlock
20:16:19*funman adds -DDISABLE_FLOAT_API to libopus
20:16:46amee2kthe broken lock switch on this ipod is really driving me nuts when carrying
20:17:20bertrikamee2k, some targets without a hard lock have that
20:17:48amee2kis there a way to enable it for targets that would normally have a switch? :)
20:18:57bertrikI don't know, so far we have refused support requests like "I broke a key, please remap it" or "my thumb hurts when I press a specific key"
20:19:15 Quit TheLemonMan (Quit: Before this gets dirty!)
20:19:47AlexPbertrik: Could he enable it for himself without too much bother though?
20:19:54amee2kfair enough
20:20:24bertrikI don't know
20:20:35funmangmaxwell: you might want the float_cast.h and opus_defines.h
20:20:55gmaxwellfunman: Got a name for me to put in the commit message?
20:21:14funmanRafaël Carré <>
20:23:15bertrikfunman, can you add a #define to opus_config.h instead of adding a -D to CFLAGS?
20:23:48bertrikThat way, we have all the defines neatly in a single place
20:24:17bertrikhmm, we have lseek in the plugin api, but no ftell?
20:24:19 Join SBookman [0] (
20:24:56 Quit Thra11 (Ping timeout: 244 seconds)
20:24:58 Quit SBookman (Client Quit)
20:25:26funmanbertrik: ah sure
20:26:22funmanbertrik: done (patchset 5)
20:30:05gmaxwellfunman: pushed, thanks.
20:30:20funmanwelcome :)
20:31:59gmaxwellI think it would be very exciting to also have "here is how you replace your player's software with rockbox" as part of the big set of opus announcements.
20:33:01bertrikw.r.t. sudoku on the clip zip, I was thinking to just move everything a bit closer together, not pick yet another smaller font
20:33:26bertrikthe white space between the sudoku board and the list of numbers already found can be compressed a bit I think
20:34:21AlexPamee2k: You might be able to just implement ACTION_STD_KEYLOCK in the keymap
20:34:40 Join mgottschlag [0] (~quassel@reactos/tester/phoenix64)
20:35:22bertrikgmaxwell, we're not always really fast with this stuff, so we might miss having it stable before the announcement
20:35:52AlexPWhen is the announcement?
20:36:33FOADThat's not been announced yet.
20:36:51AlexPThat's helpful in working out if we could be ready :)
20:37:49AlexPDoes opus kill speex by the way?
20:38:02gmaxwellbertrik: Unfortunately I don't have a hard date— the RFC publication process has its own pace, and now that the draft is approved it has wind its way through the varrious editorial queues.
20:38:22gmaxwellAlexP: Functionally? yes. It's a strict superset.
20:38:54gmaxwell (search opus)
20:39:32gmaxwellThe timeframe should be more like many-weeks rather than many months.
20:40:27gmaxwellBut this isn't bad, it means that lots of things will have time to integrate it (even if just in beta versions) in time for the bit announcements.
20:41:12AlexPwe are overdue for a release btw
20:41:28gmaxwell(Also it means that the 1.0 package we publish will have all the stupid build fixes that we're gaining from additional people building it in different enviroments)
20:41:42AlexPOther than opus, is there anything else worth waiting for on the horizon?
20:42:32funmanthe mayan apocalypse
20:43:01funmanah for a release.. sorry.
20:43:05bertrikperhaps aim for a first stable release of the clip zip
20:43:18funmanwe definitely need the manual
20:43:39bertrikit's gotten a cabbie v2 theme and nearly all plugins work
20:43:49funmanand maybe some sort of agreement that we don't need to fix subtle bugs in some games
20:43:55AlexPWhat would people think of a freeze in a month?
20:44:28AlexPI'll send a message to the list, but does that sound immediately stupid?
20:44:50funmanone month delay looks sensible to me
20:47:24amee2kAlexP: i have no idea what that is >_>
20:47:47AlexPDon't worry then :)
20:47:54amee2kjust downloaded the source and now trying to figure out how the soft lock thing works
20:48:10AlexPI think it is that keymap define I gave you
20:48:44[Saint]You'd have to use something *really* crazy on an iPod
20:48:49[Saint](limited keys)
20:49:09amee2ki bet you can do it in no time, but i started looking at the source code like 5 minutes ago :)
20:49:39amee2kthats err the 300MB worth of source code, actually
20:50:25[Saint]You'd need to use something like right+select, or left+select, or right+left (bad, stresses the clickwheel)
20:50:49[Saint]actually, I think those three are the only available combinations.
20:50:57amee2ki'd be fine with a "lock keypad" option in the menu
20:51:16amee2kand when the lock is engaged, i can use any combination for unlocking
20:51:47[Saint]that's a bit more complicated than you'd imagine.
20:52:19amee2ktechnically or conceptually?
20:52:26[Saint]not impossible, but non-trivial.
20:52:32amee2kif i can't figure out how to compile it with the soft keylock option, i might look at how to implement it as a plugin
20:52:58AlexPHow would that work?
20:53:06[Saint]just wondering that :)
20:53:13amee2ki had to hard reset it a few times because you can't return from some plugins apparently, sounds like implementing a keylock this way should be possible
20:53:35amee2klike the FFT one i think
20:53:47AlexPOh, you mean run a plugin that just sits there and only lets you quit with a single key?
20:53:48[Saint]amee2k: "can't return from some plugins" isn't the same as "didn't bother finding out how to return from some plugins" :)
20:53:50AlexPSounds annoying
20:54:43[Saint]with the ipods, because of the incredibly crowded keymap, its always a combination, but it is possible to exit every plugin.
20:54:53amee2k[Saint]: you already pointed out that there aren't too many keys on an ipod. the only way i found to get out of it was holding select and play until it shuts down
20:55:23[Saint]did you try, y'know....reading the manual?
20:55:56AlexPamee2k: On quite a lot you switch hold IIRC
20:55:56amee2kon the train?
20:56:37amee2kum.. the reason i want to recompile with the soft keylock is because the hold button on this thing is dead
20:56:37[Saint]amee2k: so, after you have a problem, instead of investigating it later at an appropriate dissmiss it entirely so you get to complain about it again? ;)
20:56:57amee2k[Saint]: where exactly did i complain about it?
20:57:28[Saint]"<amee2k> i had to hard reset it a few times because you can't return from some plugins apparently, sounds like implementing a keylock this way should be possible"
20:57:39amee2kthats a complaint exactly how?
20:57:49[Saint]Well, its a non-truth.
20:57:54AlexPamee2k: I know you do, but that doesn't mean you can't return - it just means you can't with your broken button
20:58:05amee2ki was pointing out why i think implementing it as a plugin should work
20:58:18[Saint]anyway...for most plugins on the iPods, its menu+select to exit.
20:58:32[Saint]doom is the only thing that expects hold to work.
20:58:34amee2ki definitely tried that
20:58:51AlexPah, OK
20:58:54amee2ki tried select and all other keys together
20:59:02 Quit mikroflops (Ping timeout: 244 seconds)
20:59:06AlexPI knew hold came in on the ipods somewhere (for something other than hold)
20:59:08amee2k(in turn, not at once of course)
20:59:19[Saint]amee2k: I just tried FFT, unsurprisingly, menu+select exits it.
20:59:30[Saint]Do you have a case that behaves as you describe I can test?
21:00:49bertrikbut seriously, am I overlooking something in the plugin api that tells me how far I am in a file (like ftell) ?
21:01:18bertrikI think I can lseek with an offset of SEEK_CUR + 0 and get the current position from that, but that feels like abuse
21:02:41 Join anewuser [0] (~anewuser@
21:02:41 Quit anewuser (Changing host)
21:02:41 Join anewuser [0] (~anewuser@unaffiliated/anewuser)
21:02:58[Saint]amee2k: I'm struggling here...I haven't found a plugin (barring doom) that doesn't exit via either emnu, or menu+select yet.
21:03:07[Saint]If you find one, let me know.
21:03:07 Quit anewuser_ (Ping timeout: 272 seconds)
21:03:08gevaertsbertrik: isn't that the traditional way of doing that?
21:03:21amee2kstart playback, start fft plugin, try to get out of it? i'm not aware that i did anything special about that
21:03:44funmanbertrik: that's what zxbox does for 'my_ftell()'
21:03:55bertrikhm, ok!
21:04:01[Saint]amee2k: done.
21:04:10[Saint]As stated above, menu+select exits.
21:04:12 Quit stoffel (Ping timeout: 244 seconds)
21:04:27[Saint]Sorry...but, it does.
21:05:00[Saint]iPod keymap is universal too (for those with clickwheels) afaik. So, yeah...
21:05:58[Saint]the only plugin you should have any difficulty with is doom, which without a functional hold key you won't even be able to start, or exit.
21:07:05funmanbertrik: should not be hard to add ftell() though
21:07:26bertrikif lseek(fd, SEEK_CUR, 0) does, that's fine too
21:08:31*[Saint] adds that its not specifically menu+select, either. There's no implied order for this key combination. The only thing that does that is Hotkey afaik.
21:08:58[Saint]Both keys just need to fire at the same time, one can come slightly before the other.
21:09:35[Saint]unlike Hotkey combos, where the non-select key has to fire first (and be held) while the secondary key is pressed.
21:10:13 Join mikroflops [0] (
21:11:45 Join Thra11 [0] (~thrall@
21:12:59 Quit pretty_function (Ping timeout: 244 seconds)
21:13:39[Saint]amee2k: I just re-read the last few minutes and (as usual) I came off as quite some measure of asshole.
21:14:15[Saint]This wasn't my intention. Its not "Hahaha! You're wrong!" but rather "I'm sorry, I cannot reproduce, this is what I'm seeing, and this is how it is"
21:15:02 Quit leavittx (Read error: Connection reset by peer)
21:23:35 Join CaptainKewl [0] (
21:25:11 Join Thra11_ [0] (~thrall@
21:25:15gmaxwellbertrik: Would it be helpful if everwhere in opus there were also #ifdef HAVE_OPUS_CONFIG_H just like the normal config.h ones?
21:26:02 Quit Thra11 (Ping timeout: 252 seconds)
21:27:02bertrikthe reason I used opus_config.h is because we already have a config.h in rockbox and I wanted to make really sure we're not confusing the two. The exact define (HAVE_OPUS_CONFIG_H or HAVE_CONFIG_H) doesn't really matter.
21:27:16bertrik(in my opinion)
21:30:51 Join stoffel [0] (
21:33:31 Quit Wardo (Ping timeout: 272 seconds)
21:34:26 Join leavittx [0] (~leavittx@
21:36:56 Join Wardo [0] (
21:38:17***Saving seen data "./dancer.seen"
21:38:30 Join Thra11 [0] (
21:39:04 Quit Thra11_ (Ping timeout: 264 seconds)
21:42:42 Quit stoffel (Remote host closed the connection)
21:45:28 Quit dys (Remote host closed the connection)
21:47:19 Join Thra11_ [0] (~thrall@
21:47:35 Quit Thra11 (Read error: Operation timed out)
21:53:05 Quit Thra11_ (Read error: Operation timed out)
21:54:20 Join Thra11 [0] (~thrall@
22:04:43 Quit Thra11 (Quit: No Ping reply in 180 seconds.)
22:08:36 Join Thra11 [0] (~thrall@
22:10:03 Join factor [0] (
22:12:13 Quit [Saint] (Read error: Operation timed out)
22:13:24 Join [Saint] [0] (~sinner@unaffiliated/saint/x-8516940)
22:22:30 Join stoffel [0] (
22:26:21 Quit CaptainKewl (Ping timeout: 244 seconds)
22:29:25 Quit MethoS- (Ping timeout: 248 seconds)
22:40:01 Quit evilnick (Ping timeout: 272 seconds)
22:40:01 Quit kevku (Ping timeout: 272 seconds)
22:40:56 Quit n1s (Quit: Ex-Chat)
22:45:43 Quit Thra11 (Ping timeout: 272 seconds)
22:47:30 Quit stoffel (Read error: Connection reset by peer)
22:59:39 Quit mikroflops (Ping timeout: 272 seconds)
23:04:23 Join mikroflops [0] (
23:06:21 Quit amayer (Quit: Going ~/)
23:08:00 Quit perrikwp (Read error: Connection reset by peer)
23:09:12 Join perrikwp [0] (
23:21:35 Join pamaury [0] (
23:21:35 Quit pamaury (Changing host)
23:21:35 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
23:24:15 Quit slowdow ()
23:29:17 Quit mikroflops (Ping timeout: 246 seconds)
23:38:21***Saving seen data "./dancer.seen"
23:57:01 Join user485763 [0] (~aoeu@
23:58:52 Quit pamaury (Remote host closed the connection)

Previous day | Next day