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

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

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

#rockbox log for 2009-05-09

00:00:08 Join Thundercloud [0] (
00:04:44 Quit DeeboiFresh (Read error: 110 (Connection timed out))
00:05:13 Join SirFunk_ [0] (n=Sir@
00:06:49mtlinuxstb : both bswap.h are needed (one is included in bitstream.h, the other in intreadwrite.h)
00:07:53mtbut I think both could use one just one file
00:09:27 Join SirFunk [0] (
00:11:58mtlinuxstb : both files are needed just for now, later libavutil will all be dropped.
00:13:32 Quit bertrik (Read error: 113 (No route to host))
00:16:31 Join atrus [0] (
00:18:39 Quit jgarvey ("Leaving")
00:19:50 Quit SirFunk__ (Read error: 110 (Connection timed out))
00:21:32mtlinuxstb : patch sent
00:24:30 Quit SirFunk_ (Connection timed out)
00:35:36linuxstbmt: Did you send me the right file? It includes "Makefile", not "Makefile.test".
00:36:18linuxstbmt: Sorry, ignore that, I think it was me confusing the files...
00:37:04mtlinuxstb : OK , (pheww .. !)
00:42:42***Saving seen data "./dancer.seen"
00:42:54 Join CaptainKewl [0] (i=jds@
00:43:49 Quit midijunkie ("?(???~~)?")
00:47:26 Join Thundercloud_ [0] (
00:49:43 Quit Thundercloud (Read error: 104 (Connection reset by peer))
00:49:52 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
00:58:26 Part atrus ("Ex-Chat")
00:59:28 Quit _synergis (Remote closed the connection)
01:02:35 Join _synergis [0] (
01:04:30 Join fdinel [0] (
01:04:55 Quit Thundercloud_ (Read error: 54 (Connection reset by peer))
01:05:17 Join Thundercloud [0] (
01:06:45 Quit fyrestorm (Read error: 54 (Connection reset by peer))
01:09:48 Join fyrestorm [0] (
01:10:58 Join wincent_balin [0] (
01:14:58 Quit kachna|lappy (Remote closed the connection)
01:15:17 Join kachna|lappy [0] (
01:17:51linuxstbmt: I've cleaned up avcodec.h - it simply needed most of the file surrounding by a single #if 0, rather than all those commented-out lines.">
01:19:49linuxstbmt: There are still two unused vars in main() - I think they need to be surrounded by #ifdef DUMP_RAW_FRAMES (filename and fd_out)
01:24:53mtthose vars are already used and surrounded by dumo
01:25:37mtI guess the warning is due to dump_raw_frames being not defined.
01:26:22 Quit blithe (Read error: 131 (Connection reset by peer))
01:27:16linuxstbWhy are you using the bitstream.c from libwma, and not the original?
01:28:01mtbecause it was already modified for use with rockbox and compiling outside ffmpeg
01:29:45 Join Makuseru [0] (
01:30:44 Quit Makuseru (Read error: 104 (Connection reset by peer))
01:32:21 Quit robin0800 (Read error: 54 (Connection reset by peer))
01:33:09 Nick fxb is now known as fxb__ (
01:35:23 Join Darren415 [0] (
01:37:00linuxstbmt: I'm probably going to commit your code soon, but with less changes than your patch compared to ffmpeg. But we can work on it some more tomorrow, and hopefully commit the rest of your work.
01:37:36Darren415Im having a problem, I installed Rockbox and it says to restart with menu/select then select/play but when I do select/play it says OK to disconnect.
01:38:25 Quit z35 (Read error: 131 (Connection reset by peer))
01:38:34linuxstbDarren415: That's an Apple message, and it also means "OK to connect".
01:38:50mtlinuxstb : okay .. good night
01:38:50Darren415so I should connect it back to my computer?
01:39:14linuxstbYes, and fix the problem described - I'm guessing it couldn't find rockbox.ipod ?
01:39:17linuxstbmt: Goodnight.
01:39:47Darren415I couldnt get it to install the bootloader.
01:39:55 Join z35 [0] (
01:40:02Darren415Until I dled ipodpatcher.exe
01:40:11 Join Makuseru [0] (
01:42:19Darren415So what am I doing wrong?
01:43:53linuxstbWhat was the error message?
01:44:22Darren415exactly as you said, it couldnt find rockbox.ipod
01:44:43linuxstbDid you install Rockbox, or just run ipodpatcher?
01:44:58Darren415I thought I had installed Rockbox.
01:45:44linuxstbSo you just ran ipodpatcher?
01:46:40Darren415I ran ipodpatcher, and I installed rockbox and the the bootloader
01:46:57linuxstbHow did you install Rockbox?
01:47:26Darren415Im doing it again, its going alot slower then I remember.
01:47:33Darren415Its the second tab, bottom button.
01:47:51linuxstbSo you're using Rockbox Utility?
01:48:30linuxstbAre you on OS X?
01:49:03linuxstbOK, there are some known issues with ipod installs on OS X, that's why I asked. It should be fine on XP though.
01:50:46Darren415do they need to be installed in a specific order?
01:52:27linuxstbNo, but if you install the bootloader without Rockbox itself, then you'll get the error you saw. So it's suggested to install Rockbox itself first, then the bootloader.
01:53:39Darren415so should I reinstall the bootloader?
01:55:07linuxstbNo, the bootloader is working fine.
01:56:16Darren415Ok I did the install for rockbox and I get that error still. Im sorry Im so hopeless at this
01:56:19 Quit einhirn (Read error: 104 (Connection reset by peer))
02:04:53 Quit amiconn (Nick collision from services.)
02:04:55 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
02:04:56 Join pixelma_ [50] (n=pixelma@rockbox/staff/pixelma)
02:04:56 Quit pixelma (Nick collision from services.)
02:05:15 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
02:05:16 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
02:05:55 Quit froggyman ("CGI:IRC (EOF)")
02:06:00 Join froggyman [0] (n=47ba40e2@gateway/web/cgi-irc/
02:06:31 Quit Seed ("cu, Andre")
02:08:01 Part domonoky
02:12:02Darren415I keep getting the same error.
02:16:19 Part toffe82
02:21:13soapwhat .zip program do you use on XP? The built-in one or something else like WinZip or WinRAR?
02:21:33 Quit Makuseru ("Konversation terminated!")
02:21:49Darren415The automatic installation fails ever time I try.
02:21:58soapjust download the appropriate .zip file, and extract it to the root of your iPod.
02:23:45 Quit froggyman ("CGI:IRC")
02:24:57 Join Makuseru [0] (
02:26:56Darren415if I open the ipod folder there is a .rockbox folder and the rockboxutility.exe
02:27:53soap"the ipod folder" being the root directory of your iPod?
02:28:57 Quit Makuseru (Read error: 104 (Connection reset by peer))
02:30:00 Join Makuseru [0] (
02:30:56 Quit Makuseru (Read error: 104 (Connection reset by peer))
02:41:28Darren415no matter what I do I get the bad checksum error
02:42:34 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
02:42:43***Saving seen data "./dancer.seen"
03:02:38soap"the bad checksum error"?
03:03:04soapyou say that as if you had mentioned said problem before.
03:04:24CIA-38New commit by dave (r20882): Initial commit of the minimal set of ffmpeg (r18079) files required for Cook (realaudio) decoding. These are the unmodified versions from ffmpeg, ...
03:05:05Darren415yeah like linuxstb said before it cant find rockbox.ipod says bad checksum
03:05:51linuxstbDarren415: No, it won't say that. Read what it says exactly - it's important!
03:08:19Darren415ok I uninstalled, and started over. now when I menu/select then select/play it still gives me Ok to connect
03:08:44Darren415but before if I only menu/select then it would give me that cant find rockboc.ipod
03:09:14linuxstbWhat's the _exact_ error message the ipod displays?
03:10:27Darren415after the select/play it says "OK to disconnect"
03:12:37linuxstbNo, before that - when you press menu/select
03:12:38Darren415If it matters its a 1st gen nano
03:13:01Darren415Now it boots up like normal.
03:13:13linuxstbWhen you press menu/select you're rebooting your ipod, and the Rockbox bootloader runs. This is what displays the message about rockbox.ipod
03:19:33 Join hasmind [0] (
03:21:53CIA-38New commit by dave (r20883): The first part of Mohamed Tarek's Google Summer of Code work to implement RealAudio support in Rockbox. This is a self-contained Cook decoder using ...
03:22:58*linuxstb congratulates mt on getting code committed before SoC formally starts...
03:23:59 Quit hasmind (Client Quit)
03:25:05saratogayes very impressive work
03:25:28 Quit Thundercloud (Remote closed the connection)
03:28:40Darren415got it
03:31:52Darren415do I put files on it the same way as before?
03:37:34 Nick JdGordon is now known as JdGordon|afk (n=jonno@rockbox/developer/JdGordon)
03:44:27 Join toffe82 [0] (
03:47:44Darren415how do I research the database?
03:54:08soaphave you read the manual section on it yet?
03:54:29Darren415doing that now
03:55:36Darren415it only shows how to have it auto update.
03:57:08 Part wincent_balin ("Kopete 0.12.7 :")
03:58:08soapYou must be reading a page I haven't seen.
03:59:04Darren4154.2.3 Auto Update
03:59:05Darren415If Auto update is set to on, each time the player boots, the database will automatically be updated.
04:01:11soapDarren415> it _only_ shows...
04:01:32soapThat is the page I've never seen. Just keep reading and _then_ feel free to ask questions.
04:07:55 Part Darren415
04:31:40 Quit Lss (Read error: 104 (Connection reset by peer))
04:31:56 Join Lss [0] (
04:39:28soapscorche and/or Llorean and or (well, neither of the old world admins are around) -;u=22789
04:40:19scorchedid you just want his displayed name changed?
04:40:36soapI wanted to put it up for discussion, yes.
04:41:06 Quit miepchen^schlaf (Read error: 101 (Network is unreachable))
04:41:19scorchehow about I change it to LilFurryBunbuns and send him a PM to change it to something decent?
04:42:44***Saving seen data "./dancer.seen"
04:44:17soapfurries freak me out.
04:45:31scorchesame....but still...
04:46:10soapIn all seriousness, I think that is the best course of action.
04:52:36 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
05:08:10 Quit intrados_ (Read error: 104 (Connection reset by peer))
05:08:41 Quit _lifeless (Read error: 113 (No route to host))
05:16:53 Quit gevaerts (Nick collision from services.)
05:17:04 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
05:20:27 Join intrados_ [0] (n=intrados@
05:28:56 Join Tristan [0] (
05:35:48 Quit intrados_ (Connection reset by peer)
05:40:37 Join robin0800 [0] (
05:45:48 Quit perrikwp (" ajax IRC Client")
05:46:43 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
06:02:02 Join intrados [0] (
06:02:12 Quit KBH (Read error: 110 (Connection timed out))
06:08:31 Quit z35 (Read error: 54 (Connection reset by peer))
06:09:25 Join z35 [0] (
06:10:54 Quit perrikwp (" ajax IRC Client")
06:11:27 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
06:22:07 Quit robin0800 (Connection reset by peer)
06:42:48***Saving seen data "./dancer.seen"
06:49:22 Join Sharn [0] (
06:49:28SharnIs there anyone here who knows about compiling for a Cowon D2? I'd like to help out keeping patched original firmwares up to date
06:49:41Sharn(And yes, before anyone mentions, I know it's an unsupported platform)
06:58:35 Quit CaptainKewl (Read error: 110 (Connection timed out))
07:04:03 Join AndyIL [0] (i=AndyI@
07:16:09 Quit BHSPitLappy (Remote closed the connection)
07:17:46 Quit AndyI (Read error: 110 (Connection timed out))
07:21:31 Join blithe [0] (
07:25:15 Join Makuseru [0] (
07:26:26 Join _lifeless [0] (n=lifeless@
07:27:16 Join bertrik [0] (n=bertrik@
07:29:02 Quit z35 ("Leaving")
07:32:15 Quit Makuseru (Remote closed the connection)
07:40:31 Join __lifeless [0] (n=lifeless@
07:42:56 Quit _lifeless (Read error: 60 (Operation timed out))
07:53:36 Join Horschti [0] (n=Horscht@xbmc/user/horscht)
07:54:09 Join Traveler6 [0] (
07:54:31Traveler6unused variables are still instantiated in rockbox plugins, right?
07:56:09bertrikyes I think so
07:56:51Traveler6alright, just making sure I don't have to worry about running out of memory down the line when I start using such variables
08:01:00 Part toffe82
08:08:25 Quit Traveler6 ("Java user signed off")
08:08:49 Quit dude187 ("Leaving")
08:10:30 Quit Horscht (Read error: 110 (Connection timed out))
08:23:19 Join lucent [0] (i=lucent@unaffiliated/shadows)
08:37:19 Quit Lss (Read error: 110 (Connection timed out))
08:42:50***Saving seen data "./dancer.seen"
08:57:20 Join Rob2222 [0] (
09:12:37 Join flydutch [0] (
09:15:13 Quit Rob2223 (Read error: 110 (Connection timed out))
09:27:37 Join daurn [0] (n=daurnima@unaffiliated/daurnimator)
09:31:30CIA-38New commit by unhelpful (r20884): Split 8-bit-to-native conversion in bmp.c into a function, add support for plugging unscaled output in BMP and JPEG loaders, use output_row_8_native ...
09:33:27 Join moos [0] (i=mustapha@
09:36:28 Join homielowe [0] (n=ce748e82@gateway/web/cgi-irc/
09:41:27 Quit homielowe ("CGI:IRC (EOF)")
09:42:50 Quit daurn| (Read error: 110 (Connection timed out))
09:52:50 Join tvelocity [0] (
10:22:49 Quit martian67 (Read error: 104 (Connection reset by peer))
10:23:49 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
10:33:20Unhelpfulamiconn: hrm... i *tried* inlining output_row_8_native, and it reduces the delta to almost nothing if nothing else wants the function. i could try inlining it if neither HAVE_JPEG nor HAVE_BMP_SCALING are defined.
10:33:38 Nick fxb__ is now known as fxb (
10:38:23 Quit Horschti ("Verlassend")
10:40:39 Join Grahack [0] (
10:42:53***Saving seen data "./dancer.seen"
10:45:48mtlinuxstb , saratoga : Thanks :) (I was asleep when I received those messages)
10:47:04 Join firedix [0] (n=firedix@
11:03:55 Join schrottplatz [0] (
11:09:02bertrikI vaguely remember talk about experiments being done on hacking the WPS to use it for the radio screen
11:09:35bertrikdoes this ring a bell to anyone, or does anyone have more info (or even an experimental patch) for this?
11:13:12 Quit intrados (Connection timed out)
11:15:52 Quit Tristan (Read error: 104 (Connection reset by peer))
11:16:52Lloreanbertrik: I'm not sure it ever got past theory
11:18:03LloreanBasically, there seem to be three proposals. 1) Try to automatically use existing WPSes (use things like the Album and Artist tag for station name and station frequency, etc), 2) use the existing wps, but have a FM conditional and fm-specific tags so it's still one file, but with unique tags for the radio, or 3) Have a completely separate WPS file for FM
11:18:23LloreanOf course, 2 and 1 could be combined as well
11:18:26 Join Tristan [0] (
11:21:19bertrikLlorean, thanks, I found a thread about it on the forums
11:22:18bertrikmy current opinion is to re-use the existing WPS (option 1 or 2)
11:22:47LloreanThat's what I prefer too.
11:24:05bertrikand I see JdGordon| has been working on this, I'll ask him when he wakes up
11:33:10bertrikI'm not so familiar with the WPS code yet, I looked a bit at it last night. In a very basic form, we could just start the WPS upon entering the radio screen and use track change events with fake id3 data to display the currently tuned station
11:33:36bertrikwe would also have to handle keys in a way appropriate for radio of course
11:35:38 Quit moos (Read error: 145 (Connection timed out))
11:37:25lucenton fuze 8gb, the wheel light flickers pretty crazy with m4a playback
11:38:00lucentcould be wrong about that
11:40:49bertrikUnhelpful, some jpeg album art looks way too big now on my e200
11:41:57Lloreanlucent: Problems with the incomplete port should go in the thread.
11:42:56*lucent pokes to find the thread URL
11:43:49LloreanIt's where it's been the entire time the port's been going on, and I'm sure I've told you to post problems in it before. It's in the new ports forum.
11:44:22pixelmabertrik: what do the keys have to do with what is displayed on the screen? IMO only reusing existing WPS code won't work completely - e.g. what would you use to let preset/scan mode display, or in case of the Archoses the current radio screen displays prerecording or recording time (because you can record directly from the radio screen and I still hope it will be on swcodec targets someday too)
11:44:37lucentwhoa, chill baby. I'm looking for it
11:45:23lucentI'm pretty pleased with rockbox on the fuze, just noticed the quirky wheel light flickering
11:45:40Lloreanpixelma: That's why a combination of 1&2 would be good. Un-adapted WPSes would work minimally, while adapted ones would simply be better
11:45:49bertrikpixelma, I'm not claiming to have thought out a fully functional impementation :P
11:47:08pixelmawell, I also was only saying what came to mind after reading this - just things to consider
11:47:11bertrikAs far as I understand, the wps code also handles keypresses that happen in the wps which cause a call to audio_next for the forward/back buttons for example
11:47:41Unhelpfulbertrik: interesting. it *should* only be taking the "don't use scaler" branch after testing that the decoder will output exactly the desired size. :/
11:48:46pixelmabertrik: really? Sounds somehow nasty - from a non-coder point of view
11:50:27pixelmathe radio screen has a few unique buttons (like the mode switching, or mute, stuff like that)
11:50:38lucentLlorean: what more does sansa fuze target need for inclusion in rockbox releases?
11:52:58bertrikpixelma, yes it looks like there are a couple of hard links to the playback code. Not sure yet how those special radio buttons would have to be handled.
11:53:35BigBambiAlso there are RDS targets to consider - so far only gigabeat S
11:54:08Lloreanlucent: Not to freeze up during playback, an installer incorporated into rbutil, a finished manual. It basically needs to be working at the level of the other supported ports.
11:54:21bertrikmy gut feeling is that radio using the WPS can be made to work in a basic form and that we can tackle the other problems as we go
11:54:25BigBambiIt'd be nice to link up station logos to presets as "album art" also
11:54:48linuxstblucent: Is Rockbox stable enough to replace the original firmware for day-to-day music playback? Plus we need safe, user-friendly installation procedures.
11:55:22lucentI'm using it for day to day playback, since the fixes for mp3 playback went int
11:55:28lucentit doesn't freeze on me
11:55:34BigBambiwhat about other codecs?
11:55:39BigBambiand all other apsects?
11:55:48lucentFLAC, m4a, mp3, vorbis, ape, are all working
11:56:05lucentno more funky weird pixels on the screen,
11:56:07linuxstblucent: So you would consider it as stable as, for example, the e200 port?
11:56:19BigBambilucent: look at the gigabeat S - that is far more advanced, yet still isn't quite finished as far as supported status goes
11:56:24lucentyeah I haven't been able to make it crash
11:56:44*Llorean knows that as of shortly after the "mp3 fix" there were still playback problems on other AMS Sansas, at least
11:57:11lucentthat's not to say that all fuze models work
11:57:52lucentthere seemed to be subtle differences between 8gb and 4gb fuze, but I never confirmed beyond that the issues I had with 8gb fuze are resolved
11:58:36BigBambilucent: even if it is as stable as the other suported targets, there still remains easy, safe installation and a manual
11:58:57bertrikBigBambi, nice idea about using the station logo as album art
11:59:18Unhelpfulbertrik: ... yes, that test is completely wrong, and it will basically always decode jpeg with at the next (2^N)/8 size up from the requested one :/
11:59:23lucentokay a manual, I could help with that... about the easy safe installation, what's the criteria for that?
11:59:43 Join moos [0] (i=mustapha@rockbox/staff/moos)
11:59:45Lloreanlucent: Basically, rbutil integration is ideal.
12:00:07LloreanThe only time it should ever be something other than that is if for some reason that's impossible.
12:01:09BigBambiSo who knows about editing english.lang?
12:01:46BigBambiI want to correct a couple of spellings - do I change just the dest and voice - or do I change source too, which then means changing all the other language files to match?
12:02:42lucent aha
12:02:55LloreanBigBambi: Rasher seems to be the best person to ask.
12:03:15pixelmaif you change the source you don't have to change all other languages, but it's needed to show up as difference when running genlang, so translators will notice
12:03:21BigBambiLlorean: Yeah, but he isn't about much - I was hoping someone else might know too :)
12:03:43BigBambipixelma: I'm only changing spelling - e.g. center to centre, so no meaning will change
12:03:46rasherBigBambi: tools/ should be able to fix source in lots of languages at once
12:03:52bertriklucent, are you still seeing issues with the RTC on your sansa fuze?
12:03:56BigBambirasher: aha :)
12:04:04rasherCheck the diff before you commit though
12:04:19lucentbertrik: not sure I had issues with RTC :/ did I?
12:04:25BigBambirasher: So it is best to change all in english.lang, then use langtool to change the source line in the other langs?
12:04:48rasherBigBambi: Yes. As long as you're not changing the meaning of the string
12:05:08BigBambiNo, no meaning change - just a couple of US - UK spellings
12:05:15bertriklucent, there was some weirdness regarding the date/time shown in the OF and the date/time shown in rockbox
12:05:17pixelmaBigBambi: yes, I just meant that if you don't change source in the other languages with it - it will show up as difference and translators will notice that source changed and can adapt (one by one)
12:05:42lucentbertrik: oh I can check for this
12:05:59BigBambipixelma: ah right - well in this case as there is no change in meaning I can do as rasher suggested and save them work
12:06:02Lloreanbertrik: Wasn't that just because the e200v1 used a different offset, and we hadn't adjusted for the offset the AMS Sansas used?
12:06:08*Llorean seems to recall that being fixed some time back.
12:06:38 Join ender` [0] (
12:07:42lucentbertrik: 4am as seen by RB <=> 8am as seen by OF
12:07:56lucentconfirming the weirdness, I suppose?
12:08:29bertrikLlorean, yes a fix was committed, but there were somehow still problems with time set in the OF not correctly transferring to rockbox or the other way around
12:09:36lucentbertrik: if I set from OF, it shows up in RB okay
12:09:52lucentbertrik: with the wrong date, I think
12:10:22lucentbertrik: what information can I provide to help?
12:10:30CIA-38New commit by unhelpful (r20885): Fix test for direct JPEG output, bump plugin API as r20884 changed struct custom_format.
12:10:56bertriklucent, can you add your finding to FS #9985?
12:10:59Unhelpfulbertrik: that one fixes it for me.
12:11:17*lucent pokes internets
12:11:27bertrikUnhelpful, ok, thanks, I'll test it right away!
12:15:22Unhelpfulgrr-enter-key. anyway, no hurry, it was a very, very obvious mistake, calculating a valued and forgetting to compare it to another, turning the test into a test for whether the jpeg output dimension was non-zero
12:15:58lucentbertrik: method to dump RTC value read from rockbox side? so I can compare against what I see from OF side?
12:16:15lucentthe patches in SVN confuse me a bit now what to report
12:17:28bertrikUnhelpful, thanks, album art seems to work fine now (scaled properly) on my e200
12:19:35bertriklucent, never mind the RTC dump, do you see the same behaviour as the comment by fragilematter from 12 april 2009, 20:23?
12:20:56lucentbertrik: I think so, I cannot be certain because he doesn't provide a test procedure. I will add a comment with a test procedure to reproduce the symptom
12:23:04bertrikI'm beginning to think that the OF doesn't really change the RTC time when changing the date and time. Maybe it just assumes the RTC runs in UTC and stores only an offset to the RTC time.
12:23:15lucentSet date "May 09 2009 04:18" from OF. Read date "Aug 09 2009 04:19" from RB. Set date "May 09 2009 04:18" from RB. Read date "Feb 06 2009 04:17" from OF.
12:23:26lucentthat's what I found now
12:31:08 Join intrados [0] (
12:32:09CIA-38New commit by unhelpful (r20886): Inline output_row_8_native when building bmp.c in core without HAVE_JPEG or HAVE_BMP_SCALING.
12:32:15CIA-38New commit by unhelpful (r20887): Don't build 16-point IDCT on greyscale targets, since it's only used for chroma components.
12:39:17lucentbertrik: I think I agree with you
12:39:24 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
12:39:31 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
12:40:06lucentbertrik: OF *must* be saving an offset, but not modifying the actual value
12:41:07Unhelpfulbertrik: that may be by design - doesn't subscription-based DRM require a non-user-modifiable, so that it can know for sure if your license file is expired?
12:41:24 Quit firedix ("Ex-Chat")
12:41:43bertrikUnhelpful, yes, sounds plausible
12:42:55***Saving seen data "./dancer.seen"
12:43:59tmztif you look at the motorola source for razr2v8 there is an implementation of the WM DRM clock
12:46:23lucentso do we want to avoid setting the RTC? matching it with a procedure that updating the RTC from rockbox causes the OF to match?
12:46:41lucentnot sure where to go, presuming that it's an offset only in the OF
12:48:11LloreanWe do have a policy of not making Rockbox dependent on the OF.
12:48:44LloreanBesides, once you set the time/date in Rockbox, and then adjust it in the OF, it should match from there on out, right? The offset should at that point be nearly 0.
12:49:19lucentadjusting the date that way doesn't quite work, I think
12:49:44lucentI wish it did
12:49:48bertrikyeah, I'm inclined to not make any further changes to the rockbox RTC code
12:50:17 Join mirak [0] (
12:50:23bertrikrockbox cannot know the OF time offset anyway
12:51:34lucentFWIW when I "overflow" my fuze OF date, it resets to jan 1 1970
12:51:37lucentnot 1980
12:52:47lucentbeyond this it is just material for the manual, cannot set RTC using OF, big fat warning that setting date from RB might hose your DRM songs on OF
12:54:11CIA-38New commit by alex (r20888): Correct a couple of spellings (US to UK).
12:56:55amiconnThere is a similar offset in the H300 OF, which rockbox does account for
12:57:13bertriklucent, I agree, and as Llorean said the time set by RB should now quite closely match the time that the OF expects
12:57:22amiconnIf it wouldn't, the OF would see an out-of-range date in its sanity check, and reset the RTC
12:57:33Lloreanamiconn: The offset changes when the time's updated in the OF, rather than the RTC time itself?
12:58:17bertrikamiconn, I think we did see some sanity-check reset too on the sansa ams targets, before we switched to using 1970 as start-of-time for ams
12:58:35amiconnWe're not *exactly* matching that offset though, due to programmer silliness at iriver. They chose an offset of 35 years, making leap years not work
12:59:19amiconnWe use an offset of 36 years, so that OF and rockbox will disagree by one year, but leap years will work
13:00:12pixelmauh, oh - not that -ise vs. -ize thing again
13:01:16BigBambiIt is quite clear - UK english uses ise, and Rockbox uses UK english
13:01:47BigBambi"Initialize" doesn't exist in UK English
13:02:58amiconnSee firmware/common/timefuncs.c
13:03:32pixelmabecause I remember quite long discussions , someone (linuxstb?) found a quite official source which said -ize was possible too. Would need to dig through the logs but that will be hard with those "keywords"
13:03:45pixelmaI think the discussions were related to the manual
13:03:46BigBambiIt is "possible" but not standard
13:04:03BigBambiStandard UK English uses -ise
13:04:19pixelmaI'd brefer -ise too, but thought we settled for that one
13:04:24pixelmaprefer too
13:04:49BigBambiI don't see why we should be wrong for the vast majority of people just because -ize isn't strictly outlawed
13:04:52pixelmaunfotunately I don't remember the details, quite a while ago
13:06:18LloreanIf "ize" is technically legal in some areas, and the only legal one in others, while "ise" is only legal in a subset of that total area, doesn't that mean "ize" is *right* for the majority? :-P
13:06:27LloreanIn spite of that, don't really care about this one.
13:06:36BigBambiIf people really want -ize they can have a US english
13:06:38 Part Llorean
13:06:46 Join Llorean [0] (n=DarkkOne@rockbox/administrator/Llorean)
13:06:53BigBambiLlorean: That doesn't matter at all, given that is it UK English :)
13:06:55pixelmaand the previous revision - discussion must have taken place around that time
13:07:05BigBambipixelma: Yes, I didn't know about that
13:07:17LloreanBigBambi: Didn't we just establish, "ize" is also UK English?
13:07:24BigBambiStill - to my eyes and I suspect the vast majority of British people, -ize looks wrong
13:07:27pixelmaLlorean: see link
13:07:31BigBambiLlorean: It is allowed, but not standard
13:07:48LloreanSo saying it's "wrong" is incorrect, though.
13:07:58BigBambiwrong in most people's eyes
13:08:02BigBambiIt looks wrong
13:08:32LloreanLooks wrong is different from being wrong. People spell things incorrectly all the time because the right way looks wrong.
13:08:42pixelmawe can go on changing it back and forth and call that "people are still developing" when there are no other commits ;)
13:08:42LloreanI'm just saying, be careful with your reasoning for doing things.
13:08:45BigBambiYes, true - but not the same thing here
13:09:03LloreanIf it's technically correct, you're changing it "to be the more common usage" rather than "to be right"
13:09:12*domonoky thinks that if we care about "most people" we should write chinese :-)
13:09:51BigBambidomonoky: Most people in the language in question, which here is UK English :)
13:09:57 Quit mirak ("Ex-Chat")
13:10:17BigBambiLlorean: It isn't technically incorrect apparently, but it is in modern usage
13:12:00BigBambiAnyway, if people really want I can live with -ize (even though I don't like it) - center was what I really wanted to change :)
13:17:55bertriklucent, now that we have an explanation for the weird time sync behaviour between OF and RB, I think I'll close the task
13:18:36bertrikmaybe, just maybe we can figure out where the OF stores its time offset and clear it when setting the time in RB
13:19:00 Join Trista743 [0] (
13:19:03 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
13:20:16*bertrik checks the .SYS and .BIN files on his sansa
13:21:03 Quit Tristan (Read error: 110 (Connection timed out))
13:22:00bertrikargh, I forgot that the clip doesn't show time in the OF ...
13:27:30 Join hd [0] (
13:28:00 Quit HellDragon (Read error: 54 (Connection reset by peer))
13:29:17 Join Xerion [0] (
13:30:20GrahackHi all, I'd like to know how I can control the space between the characters in the Virtual Keybord. The default layout is displayed very tight, and my custom layout is spaced. Thanks.
13:33:15LloreanGrahack: If you look in the custom layout file, it's probably just using normal spaces.
13:35:11 Join pyro_maniac [0] (
13:36:38GrahackLlorean: you mean between each and every column ? No, here is my layout:
13:37:40LloreanIf you mean you're trying to design a layout with partial spacing, it's impossible.
13:38:52LloreanCharacter spacing is dependent on the font
13:49:56GrahackOK, this is explained here: but what if I prefer the system font? I never wanted to use a special font. I can I continue to use the system font for the Virtual Keyboard?
13:53:31 Quit flydutch ("/* empty */")
14:02:07gevaertsI don't like the idea of re-using existing WPS tags for the radio screen at all. It depends on the assumption that everyone uses the same basic subset of tags in the same way, which isn't true
14:03:10Lloreangevaerts: I think the only tags it'd really be making assumptions about are artist/album/title. And a combination where it can use those, but has explicit tags (that are preferred) is probably more likely anyway.
14:03:24gevaertsLlorean: my WPS usually doesn't show artist...
14:04:13LloreanI don't see how that's a real problem though. You can customize yours, and "good" WPSes should include the FM tags anyway
14:04:29LloreanIt'd mainly be a way to ensure that at least a significant subset of WPSes are "usable", at least minimally
14:05:03*gevaerts thinks that they are too different
14:07:46LloreanAlbum and Station Name are basically the same thing anyway, at least.
14:08:33gevaertsreally? I'd see Artist as more like Station Name if you want to look for similarities...
14:08:41LloreanI dunno.
14:09:11LloreanSongs on CD come on albums. The same Artist and Song can be on many different albums, the album is just the name of that instance of that media.
14:09:16LloreanWhich works the same with FM.
14:09:50LloreanBut I really don't understand the objection to it being a possibility.
14:09:53gevaertsLlorean: well, "songs" (another one of my pet peeves ;) come in programs, which come on stations...
14:10:20pixelmayou could also see station as the title tag - and album the .fmr file or so
14:10:45gevaertsI actually also think that two different files will turn out to be more practical in the end. If you use one file using conditionals, you increase the difficulty of a simple basic WPS, which will (I think) scare away some people.
14:10:51pixelmaor maybe the latter is similar to the playlist info (which can be shown in the WPS too)
14:11:16gevaertsThat also avoids the need to "support" the FM tags on non-radio targets
14:11:38gevaertsthat's a different issue though
14:11:41pixelmawith viewports you can easily get complicated WPS code even for a simple WPS
14:11:52gevaertsyou *can*, but you don't have to
14:12:21pixelmayeah, just adding to your thought about complicated WPSs
14:12:53LloreanWell, as it stands if they're separate file we'd need wfs and rwfs.
14:13:23LloreanIt seems to me, at least, that FM is still just an audio listening screen, and it'd make sense if they mostly used the same code/images/etc.
14:13:52soapWhy not make them separate files with auto rollback to the wps file if there is no fmps?
14:13:58LloreanI think the WPS code is intimidating enough that adding a few more conditionals won't really raise the bar of learning it at this point.
14:14:35gevaertsAnyway, once there is a set of FM tags, converting an existing WPS to use them should be fairly easy (in both the one file and the two files case), so I don't really see a reason to support tag matching
14:15:20gevaertsMaybe we need to experiment a bit to see how complex the thing becomes.
14:17:23soapWhy not reuse existing tags as a first step and see where the requests come from?
14:19:01pixelmaanother argument for reusing album (if you want to do that) is the wish for "album/station" art, expressed earlier
14:19:16pixelmafor current station that is
14:20:19*pixelma just throws in some random thoughts
14:25:16bertrikWhat I'm afraid of is that the bar gets raised so high for a WFMS that nothing ever gets implemented, and we remain stuck with the current ugly screen
14:25:54 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
14:32:20 Quit gromit` (Read error: 60 (Operation timed out))
14:35:51 Join kugel [0] (n=kugel@rockbox/developer/kugel)
14:39:17 Quit Grahack ("Leaving.")
14:39:45 Join gromit` [0] (
14:42:59***Saving seen data "./dancer.seen"
14:53:46 Join Lss [0] (
15:11:27 Quit kugel (Read error: 110 (Connection timed out))
15:18:50amiconnsaratoga: Re your latest change to SoundCodecs - did you verify that 5.1 AC3 is now realtime on PP5002?
15:19:02amiconnLast time I tried it was not.
15:27:43 Join midijunkie [0] (
15:33:27 Join miepchen^schlaf [0] (
15:35:14 Join mirak [0] (
15:46:08 Join calman_ [0] (
15:46:55 Join Ubuntuxer [0] (
15:47:41 Join KBH [0] (
15:58:31 Join kugel [0] (n=kugel@rockbox/developer/kugel)
16:01:07 Join wpyh [0] (n=william@
16:02:19wpyhhi, I'm trying to remove the use of pluginlib actions in the bubbles plugin
16:03:30wpyhhowever, I get "undefined reference to `button_get'" when compiling
16:03:30wpyhwhat should I do? (I've re-ran "make dep" but it still errors out)
16:04:57UbuntuxerPlease paste the your current code.
16:05:36SharnDoes anyone know what processor to build for for a Cowon D2?
16:06:20gevaertsSharn: you mean the compiler? arm
16:13:27Sharngevaerts: yes, thanks
16:13:39wpyhUbuntuxer: false alarm −− I found out the reason: I didn't use rb->
16:16:18 Join taylor_ [0] (
16:18:37 Join CaptainKewl [0] (
16:18:56kugelI think we should not use album/artist/song for FM station info
16:19:58kugelthe beast has RDS support, and thus the possibility of showing real song info in the radio screen
16:20:26kugelThough, I don't think the .wps and .wfs should be seperate
16:22:11gevaertscan you change the backdrop from a conditional?
16:22:59kugelnot yet, backdrops are rather hard coded. but that should be very doable if we want that
16:23:26 Join rabbity [0] (
16:23:37 Join firedix [0] (n=firedix@
16:24:00gevaertsI think we do. I can easily imaging people wanting a different backdrop for FM. It's not the most urgent thing of course
16:24:31kugelI can imagine people wanting different backdrops for a dozens of screens :>
16:25:11kugelwe could make backdrop like fonts and put it into the viewport struct (storing an id in it)
16:25:17*gevaerts also still wants string comparison conditionals :)
16:25:24LloreanA different backdrop with and without album art would also be rather nice.
16:25:28Lloreanthe problem is, backdrop images are rather huge.
16:25:51LloreanYou'd need to reserve an awful lot of memory, or hit the disk every time the condition changes.
16:26:24 Quit midijunkie (Read error: 104 (Connection reset by peer))
16:26:28gevaertsThe FM one would be doable without extra memory if we don't use standard conditionals for it
16:27:24LloreanI don't quite understand.
16:27:32wpyhis it desirable to remove "#ifdef HAVE_LCD_BITMAP" and "#ifdef HAVE_LCD_CHARCELLS" from plugin source files? those plugins are already compiled/not compiled based on those same switches in SOURCES
16:27:34kugelinto the viewport struct would actually mean backdrop for any viewport and not necessarily fullscreen
16:27:45 Join domonoky1 [0] (
16:27:53Lloreangevaerts: Just a spinup when switching between FM / HD?
16:27:59 Part domonoky1
16:28:43kugelor we revive JdGordon's wps memory overhaul
16:28:51*Llorean doesn't think it's a bad idea to reserve memory for backdrops for a few specific screens, but universal conditional backdrop swapping is something else.
16:29:00Lloreankugel: That won't solve the issue, though.
16:29:13rabbityI feel obligated to let you guys know that two idiots from another network have decided that spamming you will be fun
16:29:16rabbity* *@
16:29:20rabbitygood day
16:29:32gevaertsLlorean: as a non-clean hack you could e.g. change %X to load <your-backdrop-name>-fm.bmp instead of <your-backdrop-name>.bmp.
16:29:47 Join sceners_n00b [0] (n=sceners_@
16:29:49 Part taylor_ ("Leaving")
16:30:09Lloreangevaerts: But the image is (usually) loaded when you select the WPS, rather than when you enter the screen that uses the WPS.
16:30:29LloreanSo it'd still need both pre-loaded with the current WPS philosophy of having it all available immediately
16:30:43gevaertshm, true
16:31:47LloreanAgain though, I don't think it'd be bad to have memory reserved for on normal backdrop, and one FM backdrop
16:32:04LloreanBut making it able to swap backdrops within general conditionals on the other hand, is probably not a good idea
16:32:21 Quit rabbity ("Leaving")
16:34:50sceners_n00bNOOO my favorite psp
16:35:11Lloreansceners_n00b: This is a technical channel. We ask that you keep discussion restricted to Rockbox development or support.
16:36:21 Join brrybnds [0] (
16:42:34 Part sceners_n00b
16:43:00 Join n1s [0] (n=n1s@rockbox/developer/n1s)
16:43:02***Saving seen data "./dancer.seen"
16:44:28 Quit domonoky (Read error: 110 (Connection timed out))
16:47:58 Quit SirFunk (Read error: 110 (Connection timed out))
16:53:04 Quit Ubuntuxer ("Leaving.")
16:53:52 Join John_coder [0] (
16:55:23 Quit John_coder (Client Quit)
16:59:08 Join toffe82 [0] (
17:04:22 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
17:04:40 Quit Lss (Read error: 104 (Connection reset by peer))
17:09:27 Quit mirak ("Ex-Chat")
17:11:25 Join webtaz [0] (
17:11:42webtazI just got a sansa fuze
17:12:06webtazand i'm willing to try out rockbox with it ( i know it's not officially running)
17:12:13webtazjust before i try:
17:12:40webtazthe bootloader -> what combination would it be to geht to the OF?
17:13:07webtazand if sth happens, is there really no known method to recover the sansa that works for all fuze?
17:16:46webtazmhh anyone seeing this?
17:16:55BigBambiyes, have patience
17:17:27webtazsry, just wanted proof irc works ^^
17:17:39BigBambiAnd please use real words as per the guidelines
17:20:19Lloreanwebtaz: have you read through the forum thread and related wiki pages?
17:20:54webtazthis, and in a german forum
17:21:53webtazI thought the buttons are described in the manuel generally, but for an device that has no build there should'nt be a manual?
17:22:33LloreanI meant the forum thread in our forums.
17:22:44LloreanThat's where the development status is discussed, including recovery methods etc.
17:22:50soapNot having seen the German forum (and not being able to read German) - it is best advised if you collect all your information regarding this work-in-progress from, not third-party sources.
17:23:37webtazwell another guy managed to run rockbox, and the infos came from kugel (who works as dev i think)
17:23:57kugelwhat infos?
17:24:20webtazthat rockbox is able to run quite stable in normal use
17:24:32webtazi know that is not official
17:24:57 Quit thegeek (Read error: 60 (Operation timed out))
17:24:57 Part calman_
17:25:02webtazso i'm going to search for recover methods in the rockbox forum
17:25:10webtazthat's gonna be some work
17:25:57kugelthere's no recovery method
17:26:17webtazah, that's an answer :)
17:26:39webtazand to get from the Original Firmware?
17:26:55 Join calman_ [0] (
17:27:11webtazcrap to get to the OF
17:27:17kugelbut so far, nobody has bricked a ams sansa by just installing rockbox after dual boot has been stabilized
17:27:36webtazthat's good to hear :)
17:31:07 Join Lss [0] (
17:38:11 Join RoC_MasterMind [0] (
17:38:57 Join robin0800 [0] (
17:40:37 Quit tessarakt (Read error: 104 (Connection reset by peer))
17:40:58 Join tessarakt [0] (
17:45:12 Join thegeek [0] (
17:47:53linuxstbBasically, "ize" is more "correct" in UK English, but people incorrectly associate it with US English, so use "ise"...
17:48:23BigBambiWell, in the end I'm not bothered
17:48:31linuxstbIn the past, when arguing about language, we've decided to use the OED as the definitive source.
17:48:36BigBambiThat was a surprise for sure
17:49:02 Join stoffel [0] (
17:49:13linuxstbIt surprised me as well - I only found that out recently (when discussing things in Rockbox).
17:49:25BigBambiOK, I'll change initialise back - center/centre was what I actually went in to change - I just spotted inialize whilst there
17:54:50*webtaz is quite happy to have Rockbox running on his Sansa Fuze :)
18:02:20 Quit thegeek (Remote closed the connection)
18:03:07 Join thegeek [0] (
18:03:12 Quit RoC_MasterMind ("Leaving")
18:04:58 Quit robin0800 (Read error: 110 (Connection timed out))
18:06:04 Quit bmbl ("Woah!")
18:09:05CIA-38New commit by alex (r20889): I defer to the OED -ise/-ize.
18:09:42 Quit thegeek (Read error: 60 (Operation timed out))
18:09:47 Join domonoky1 [0] (
18:11:07 Join Thundercloud [0] (
18:12:40kugellinuxstb: what about the sentence case patch?
18:16:04 Join thegeek [0] (
18:20:42 Join flydutch [0] (
18:26:12 Quit thegeek (Remote closed the connection)
18:26:57 Join saratogahome [0] (n=41becb3b@gateway/web/cgi-irc/
18:27:45saratogahomeamiconn: I only tested a stereo file, I didn't realize 5.1 was much slower, I just assumed we dropped the extra channels but if you're commenting I guess we mix them
18:28:11 Join thegeek [0] (
18:40:19 Quit calman_ (Connection timed out)
18:42:09 Quit saratogahome ("CGI:IRC (EOF)")
18:42:27 Quit thegeek (Remote closed the connection)
18:43:05***Saving seen data "./dancer.seen"
18:43:18 Join mirak [0] (
18:56:43 Quit flydutch ("/* empty */")
18:59:16CIA-38New commit by Domonoky (r20890): rbutil: make RbSettings a static class. (FS #10183 with improvements)
19:00:03linuxstbsaratoga: Yes, the ac3 decoder downmixes.
19:02:22 Join thegeek [0] (
19:06:04 Quit mt (Read error: 110 (Connection timed out))
19:16:03 Quit faemir ("Leaving")
19:17:44 Quit Thundercloud (Remote closed the connection)
19:18:34 Quit webtaz ("ChatZilla 0.9.84 [Firefox 3.0.10/2009042316]")
19:20:57 Join saratogahome [0] (n=41becb3b@gateway/web/cgi-irc/
19:29:23 Join faemir [0] (
19:32:44 Join seancho [0] (
19:33:46 Quit stoffel (Read error: 113 (No route to host))
19:43:50seanchonewbie question: My player has 15k tracks in the main music dir. The largest playlist I can 'insert shuffled' is 10k tracks. Does this playlist select randomly from all sub-dirs, or are there 5k tracks that will never be selected?
19:44:52gevaertsseancho: increase the "Max playlist size" settings/general/system/limits
19:45:58 Join Seed [0] (
19:46:20seanchoah, thanks..
19:46:31 Join dmb [0] (n=dmb@unaffiliated/dmb)
19:53:27 Join stoffel [0] (
20:07:27 Quit saratogahome ("CGI:IRC (EOF)")
20:09:17CIA-38New commit by bertrik (r20891): Remove unneeded #include "backdrop.h" ...
20:14:39 Quit miepchen^schlaf ()
20:17:08CIA-38New commit by Domonoky (r20892): rbutil: enforce parents for bootloader install classes. and rework OF handling a bit.
20:18:21CtcpPing from gevaerts!n=fg@rockbox/developer/gevaerts
20:18:21CtcpPing from gevaerts!n=fg@rockbox/developer/gevaerts
20:18:22CtcpPing from gevaerts!n=fg@rockbox/developer/gevaerts
20:20:52 Join Mastery [0] (
20:27:28CIA-38New commit by bertrik (r20893): Remove unused function audio_has_changed_track from apps/mpeg.c
20:33:20seanchofollow up newbie question: if I have 15k tracks in a dir and then I set 'max playlist size' to, say, 500, and then 'insert shuffled' that dir, will those 500 tracks be randomly selected from all 15k?
20:34:10gevaertsno idea, but I wouldn't count on it
20:34:55seanchoyeah, thats why Im asking...
20:35:13seanchoseems like it might not be random
20:35:35BagderI doubt it, since it'll limit the buffer used for reading the dir
20:35:55Bagderit'll probably be the first 500 tracks it reads from disk
20:36:23seanchoI figured that may be the case
20:36:42soapAssuming you don't listen to even 500 tracks at one sitting, what is the difference between having a 500 track random playlist and having a 1500 track random playlist?
20:37:18soapie - why would you set the max playlist size lower than the number of tracks you have on your player?
20:37:18seanchoexactly. I just want the 500 to be pulled from the whole collection
20:37:48Bagderseancho: why not just make a playlist out of all 1500 instead?
20:37:50seanchobecause it spins the disk for about 5 minutes loading them up
20:37:57soapIt saves memory, a bit, but there are other ways to scrape up more RAM savings if your are on a low-mem target.
20:38:09seanchoIve got 15000 tracks
20:38:15Bagdera playlist loads instantly
20:38:30Bagderand you won't notice any speed diff from a 500 vs 15000 playlist much
20:38:34soapyes, but it only buffers 16ish or 32ish or 64MBish at a time.
20:39:03soapAre you on an iPod video 80GB?
20:39:09seanchoIf I insert shuffled the whole dir, it takes a few minutes
20:39:21soapis dircache on?
20:39:22BagderI said playlist, not insert shuffled though
20:39:37Bagderthat's a much slower operation
20:39:41seanchook, I could make a playlist of the whole thing
20:40:00seanchobut then I have to re-make the playlist every time I add stuff
20:41:01LloreanOr just insert the new stuff into the playlist and re-save it.
20:41:02seanchoI was wondering if I decreased playlist max size if I could add shuffled and still get a selection from the whole dir
20:41:16soapthe 5G 80GB is (was?) the slowest disk performer by far, but 5 minutes seems like a long time to insert that many tracks shuffled.
20:41:20seanchosave a bit of effort
20:41:28LloreanIf you limit the playlist size to 500, I'm pretty sure it'll only use the first 500 songs in the folder.
20:41:39LloreanYou can of course test this to find out, and probably could've by now. :-P
20:41:39seanchoyeah, its not a huge problem
20:42:10seanchobut its so much easier to bug you guys :)
20:42:36seanchoI appreciate the effort you all put in
20:42:39seanchogreat player
20:43:09***Saving seen data "./dancer.seen"
20:43:14 Quit __lifeless (Remote closed the connection)
20:43:30 Join __lifeless [0] (n=lifeless@
20:46:02 Join _lifeless [0] (n=lifeless@
20:46:15 Join Bobthebuilder [0] (n=bob@
20:48:22 Quit Bobthebuilder (Client Quit)
20:49:52CIA-38New commit by Domonoky (r20894): rbutil: show logger windows earlier in bootloader install, to catch all errors.
20:53:22 Quit KBH ("ZNC")
20:54:22 Quit firedix ("Ex-Chat")
20:55:16 Quit pyro_maniac ("Leaving.")
21:00:44 Quit __lifeless (Read error: 110 (Connection timed out))
21:03:42 Join petur [50] (n=petur@rockbox/developer/petur)
21:03:42 Join motionman95 [0] (i=4597cd8a@gateway/web/ajax/
21:04:44motionman95How does a viewer get the file name of the file it's sent?
21:07:04CIA-38New commit by bertrik (r20895): rbutil: fix spelling, enginge -> engine
21:07:20 Join Blue_Dude [0] (n=chatzill@
21:09:34 Join KBH [0] (
21:10:06Blue_DudeA question about patches. Is it worthwhile to submit small patches of developing code as a new feature is being developed, or try to present it all at once? I'm concerned about the patch being developed on really old code. It's going to take awhile for the thing to be ready.
21:10:50LloreanBlue_Dude: It's a good idea to svn up fairly regularly so that when you do submit a patch it's in-sync
21:11:28Blue_DudeWhen I do svn, do I do a normal checkout? I don't want to overwrite my changes.
21:11:53LloreanAre you working on an SVN checkout already
21:12:00motionman95I think you'd use "svn update
21:12:04LloreanAn "svn update" will update without overwriting changes.
21:12:14LloreanAs for submitting patches as you go, the first patch should probably be somewhere around "first draft" quality. Something that's at least somewhat usable (that allows people to see the idea in action and comment on further development, or problems that may prevent its inclusion)
21:13:29Blue_DudeIf I do update, will that update every file except the changed ones? Or will it update the files I'm working on also, but preserving my changes?
21:14:03*domonoky1 made new patches for FS #10185. now rbutil can install on my sansa e200 v2 :-)
21:14:03 Quit bagawk_ (Read error: 54 (Connection reset by peer))
21:14:24gevaertsit will update all files, trying to preserve changes. If the update changes bits in the file you also touched, you'll have to fix them, but all needed information should be there for that
21:15:12bertrikBlue_Dude, if you're afraid of svn update you can do a "svn diff >my_changes.patches" first to keep a snapshot of your current changes
21:15:39Blue_DudeOK, I'll fool around with it for a minute...
21:16:40kugeldomonoky1: for the makefile thing, I'd say preprocess the makefile for __WIN32__ (and cygwin/mingw if needed) and use the pre-build one in that case
21:16:43Lloreandomonoky1: Should it work on all "supported" AMS sansas now?
21:17:22motionman95Has anyone heard of Dev C++?
21:17:25kugelah wait
21:17:25domonoky1Llorean: it could, if you add the other targets into rbutil.ini
21:17:54kugelit wasn't about windows, it was about having arm toolchains, which applies to all OSes
21:17:57domonoky1motionman95: please stay ontopic, thats not rockbox related.
21:18:23 Join firedix [0] (n=firedix@
21:18:27 Quit motionman95 (" ajax IRC Client")
21:18:53kugeldomonoky1: tools/configure has magic for finding whether arm toolchain is installed. maybe just c&p that magic?
21:18:54 Quit firedix (Read error: 104 (Connection reset by peer))
21:19:37 Join bagawk [0] (
21:21:39domonoky1kugel: first we need to decide if we want prebuild dualboot-xxx.o files in svn, or how we would like to handle that. if we have them in svn, its just a issue of not building them, if there are already there.
21:22:28kugelwell, isn't that the only possible way without requiring the arm toolchain?
21:22:34kugelI tought we already agreed on doing that
21:23:01Blue_DudeThat was interesting. I just did a "svn up" and it did go out and update the source, with a "U" in front of most files. But one file I did changed had a "G" in front of it, and my changes were preserved. Thanks!
21:23:26domonoky1another way could be to provide the dualboot files from the download server, but that needs changes in mkamsboot.
21:23:54 Quit petur ("installing linux...")
21:26:14kugelthat sounds too compilcated imo
21:26:33kugelhaving the prebuilt in svn is just fine to me
21:27:19domonoky1maybe for you, what do others think ? :-)
21:28:34domonoky1another issue with libmkamsboot.a and libucl.a: mac os X. Both need a bit Makefile love to build as multiplatform libs (see rbspeex). Makefile gurus, please help out :-)
21:29:55Blue_DudeA question about the system settings struct: I added an element to it in settings.h and put in a handler for reading that element in settings.c. Now, if I run this on an existing machine, will it screw up when it tries to pull the settings in the old format?
21:31:06 Join miepchen^schlaf [0] (
21:33:38 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
21:34:19linuxstbdomonoky1: I've always planned to have "dualboot-xxx.bin" files in SVN (once they became stable, which I'm assuming they are by now).
21:34:44linuxstbPerhaps moving the source to them to a subdir under mkamsboot.
21:35:01domonoky1linuxstb: that sounds good.
21:37:09domonoky1linuxstb: care to take a look at my mkamsboot rework ? (FS #10185) :-)
21:38:27 Join froggyman [0] (n=187b533e@gateway/web/cgi-irc/
21:38:36Blue_DudeOr is my only alternative to add my element to the end of the struct to avoid first time read errors?
21:42:59linuxstbdomonoky1: Will rbutil have md5 checking, similar to the h1x0/h3x0 bootloaders?
21:43:22domonoky1at moment it doesnt, but it could be done.
21:47:45saratogadomonoky1: the dual boot file should basically never change so I think SVN will be fine
21:48:43saratogakugel: do you have any idea if the MMU changes will require new bootloaders? last I heard the idea was to keep MMU disabled in the bootloader but I'm not sure if thats still the case or if other changes are needed
21:49:32kugelit's not set in stone, but we (me, FlynDice and funman) agreed on not messing with it in the bootloader
21:50:00saratogaso we could begin testing official bootloaders now? or do you think its too soon for that
21:50:51bluebrotherdomonoky1: just commented on that task. You use QtGui from base/. *urgh*
21:51:11domonoky1bluebrother: please read :-)
21:51:19bluebrotherwhich is a bad thing ...
21:51:34domonoky1its just temporary because no bootloader on server
21:51:51bluebrotherwell, you could simply had code it for now
21:52:04domonoky1i even wrote that in the comment
21:52:19bluebrotherthat's at least better (IMO) than introducing a non-wanted dependency. Stuff like this tends to get forgotten about later ...
21:52:38bluebrothers/had code/hardcode/
21:52:54FlynDiceIf you want to remap memory there might be an advantage to doing it in the bootloader at some point but for right now it's much simpler to leave the mmu out of the bootloader for ams sansa
21:53:20saratogano other changes to the bootloader are expected once the MMU is enabled?
21:53:37saratogalast i heard there was some thought that changes might be needed to how things are clocked, would the old bootloader be ok with that?
21:54:03bluebrotherbtw, I think it would be better to not #include something in the header and then rely in the cpp file on that include. If the file needs QtGui it should include QtGui by itself
21:54:16bluebrotheras that makes it clearer that it needs QtGui
21:54:47domonoky1bluebrother: did you hear the word "temporary" ?
21:55:09domonoky1it doesnt matter, if it surely will be changes before commit.
21:55:10kugelFlynDice: which advantage to you imagine?
21:55:35kugelthe booting process is already insanely fast (compared to my e200v1) so that isn't worth complicating the bootloader imo
21:55:59domonoky1bluebrother: and the ofHint() thing is already in svn. :-) i think this doesnt belong int rbutilqt.cpp, its install method specific.
21:56:26FlynDiceIf we enable the mmu in the bootloader we could set up your no long calls memory map before we load the firmware
21:56:43saratogaeventually we might need MMU if we're going enable USB in the booloader
21:56:57saratogabut that can be delt with later
21:57:21kugelalso, we can always update the .bin/.o when they're in svn too
21:57:23FlynDiceI think the clocking issues are all outside the bootloader also
21:57:41kugellong calls require no set up
21:58:27 Quit BHSPitLappy (Read error: 110 (Connection timed out))
21:58:32saratogahas anyone ever seen a fast FFT for ARM or Coldfire?
21:58:35bluebrotherremoving the file filter on searching for the OF file isn't a good thing either IMO −− then you need to make the bootloader class aware of the extension (or retrieve it from the settings)
21:58:48saratogai'd really like to explore alternative transforms for our codecs, but without a good FFT I doubt I'll beat what we already use
21:59:06*bluebrother can't remember him agreeing with moving the ofHint() stuff
21:59:15domonoky1bluebrother: or just relay on the user to find the right file :-)
21:59:21Lloreansaratoga: Simply googling ARM FFT turns up some results, at least
22:00:22bluebrotherwell, as the user will see possibly 100+ additional unrelated files it's not really nice
22:00:25domonoky1bluebrother: it was already there in the earlier patch, you didnt complain about it. so i commited it :-)
22:00:43domonoky1and if you think you can improve it, go for it.
22:00:57bluebrotherI complained about moving the post install hints. You really think I'd have a much different opinion about that then?
22:02:53domonoky1yes, it was in rbutilqt.cpp where it surely doesnt belong. there is already too much logic there. so i think its better now.
22:03:25*bluebrother feels like this wasn't the answer to the question
22:04:11CIA-38New commit by Domonoky (r20896): rbutil: make sure the voice creation window updates its display on startup.
22:04:57saratogaLlorean: yeah I've spent some time googling over the years but never found anything that looked better then what we use, though FFTs are often burried deep in more complicated programs so you never know
22:05:33*domonoky1 s brain reading device is broken. i think "yes" is a valid answer.
22:14:10 Quit stoffel ("leaving")
22:18:01 Join mt_ [0] (n=MTee@
22:18:40 Nick mt_ is now known as mt (n=MTee@
22:22:39saratogaI found ARM's FFT benchmarks, and they work out to doing all the FFTs + rotations needed for a WMA/AAC style MDCT codec in about 12 MHz on ARM7TDMI
22:22:56saratogai can't find my notes but I think our MDCT library does it in about 16MHz, so we're actually not too far behind
22:28:32 Nick hd is now known as HellDragon (i=jd@Wikipedia/HellDragon)
22:33:54 Quit Blue_Dude (Read error: 145 (Connection timed out))
22:34:01 Quit bluebrother (Read error: 104 (Connection reset by peer))
22:34:09 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
22:39:25kugelsaratoga: just 33% slower ;)
22:42:15saratogayeah but good luck catching up to what ARM can do
22:43:13***Saving seen data "./dancer.seen"
22:44:10 Join matsl [0] (
22:44:13Bagderno, it's 25% slower
22:44:42*Bagder shut up and goes away
22:44:48saratogawell my numbers are +/- a lot anyway :)
22:45:23kugelBagder: :D
22:47:01amiconnsaratoga: liba52 does the downmix, but even if we would drop the extra channels, decoding 5.1 would be slower than decoding 2.0, because of the higher total bitrate
22:47:41amiconnWe need to decode everything at least partially (similar to how jpeg on greyscale targets at least needs to huffdecode chroma)
22:48:26saratogai think most of the slow down is calling the IMDCT on all those extra channels
22:48:53saratogathat would consume a good 20-30MHz on PP5022, no idea how much that is on 5002
22:49:32amiconnI need to find my test files again. It should be possible to speed up things on PP5002 by proper iram usage
22:49:43 Quit froggyman ("CGI:IRC (EOF)")
22:49:52saratogawe could use the faster imdct library and probably get real time
22:50:08amiconnThe interesting thing is that with proper use of iram, PP5002 can be as fast as PP5022, which is a few percent faster than PP5020
22:50:38saratogabut its weird because the spec says to use the ffmpeg style IMDCT, not the Tremor style one, so we'd have to deinterleave coefficients for short blocks
22:50:45amiconnIf you look at the APE performance figures you see what I mean
22:50:53saratogaamiconn: actually does AC3 use a lot of IRAM already?
22:52:15saratogaif we don't mind wasting a couple KB of IRAM, we could just use the fast imdct library for long blocks and the built in one for short blocks, and not need to reorder anything
22:52:36amiconnI don't know, need to check
22:52:40 Join Blue_Dude [0] (
22:52:41 Nick Blue_Dude is now known as Blue (
22:53:07saratogahmm the decoder does windowing during the transform, so we'd need to add that too . . .
22:53:21amiconnIt might be a tiny thing that causes the slowdown. In libdemac it was one single array - a few dozen bytes - which wasn't in iram but was used a lot
22:53:34saratogahard to be motivated for a format that no one uses :)
22:53:47 Quit Blue (Client Quit)
23:01:33 Join CaptainKwel [0] (
23:10:48 Quit CaptainKewl (Read error: 110 (Connection timed out))
23:14:07 Nick BigBambi is now known as AlexP (n=alex@rockbox/staff/BigBambi)
23:16:08linuxstbsaratoga: Supporting ac3 for video could be useful though, so people can keep original audio tracks. So a very fast ac3 decoder would be nice to have for that.
23:16:38*Llorean would personally love to have AC3 for video
23:17:49linuxstbLlorean: It shouldn't be that hard to change mpegplayer to use ac3 instead of mpeg audio. I'm not sure how to support both...
23:18:18saratogalinuxstb: thats a good idea
23:18:46saratogaAC3 has zero arm optimizations right now and its about as fast as AAC (for stereo), so it could be made quite fast
23:19:05Lloreanlinuxstb: Well, supporting both is the first step in a direction mpegplayer's probably going to go *eventually* anyway.
23:21:37saratogawell if someone adds AC3 support, I'll make it really fast
23:21:51saratogauntil then theres more interesting things to work on
23:22:26 Nick _synergis is now known as synergist (
23:23:03*Llorean wonders if the Gigabeast @ full speed could handle DVD rips.
23:23:14 Nick AlexP is now known as BigBambi (n=alex@rockbox/staff/BigBambi)
23:23:56linuxstbLlorean: I guess that depends on the resizing speed. I would expect it to handle the decoding.
23:24:23 Nick BigBambi is now known as AlexP (n=alex@rockbox/staff/BigBambi)
23:24:44saratogaits got a hardware resizer i think
23:25:46LloreanAh, yeah, DVDs are always going to need aspect ratio correction too.
23:31:32saratogain case anyone was wondering, MDCT for WMA 192k uses 16.6MHz on PP5024
23:32:23*linuxstb thinks a full dvd-player plugin for the beast (with menu support etc) would be fun...
23:32:57gevaertslinuxstb: extra points for USB host support so you can actually play real dvds
23:33:52Lloreanlinuxstb: Just except valid DVD ISOs?
23:34:40linuxstbISOs could be tricky on FAT, but VIDEO_TS folders would be fine.
23:34:50LloreanAh, right.
23:35:13 Part seancho
23:38:48saratogaVideo_TS files can have MPEG2 Video and MP2 audio, which I believe we already support in mpegplayer
23:39:03saratogawould be quite easy to get a few of them working as is, just need a TS parser
23:39:18LloreanDo we support non-square pixels?
23:40:10linuxstbsaratoga: "TS" in mpeg terms means "transport stream". DVDs use MPEG program streams, which Rockbox already supports - so you should be able to simply copy a .vob to Rockbox and play it.
23:40:19mtlinuxstb : patch2 is in FS #10182 - I think the next patch would replace cook.c with the older revision, and apply the necessary modifications to cook.h and the rest of the files in preparation for applying ffmpeg's patch.
23:40:28saratogalinuxstb: I didn't realize that
23:41:22linuxstbDVDs are basically a subset of the MPEG-2 standards, with extra features (navigation, menus etc) added.
23:46:10 Quit schrottplatz ("o.O")
23:46:21saratogathe assemblerized parts of the MDCT library only account for 50% of the total run time, perhaps we could save some more time in there
23:52:11linuxstbmt: That patch looks good to me - I'll try and commit it later tonight. But in future, can you start giving your patches unique names - e.g. add a version number to them (something like libcook.patch2.v3)
23:52:52linuxstbmt: How old is the revision of cook.c you plan to revert to? Is it very different to the version currently used?
23:53:55mtlinuxstb : yes .. it's from 6 mar 2007 iirc (as opposed to 20 mar 2009 for the current one)
23:55:25linuxstbmt: Hmm, it seems there were a lot of improvements to cook.c in that time...
23:56:01saratogalinuxstb: the changes are almost entirely for multichannel cook though
23:56:31linuxstbsaratoga: No, those are the changes after the 20 march 2009 version.
23:56:39linuxstbI'm talking about the 2 years previous to that.
23:57:06saratogai looked through them and most seemed irrelevent to us, aside from one or two by Michael
23:58:14 Nick fxb is now known as fxb__ (
23:58:15linuxstbStill, it seems bad to use such an old version of the decoder, and not work to update those patches.

Previous day | Next day