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 2006-08-07

00:02:06 Join suge [0] (
00:03:04sugehey. Does rockbox have a "remember playback position" feature so you can pause audiobooks, podcasts, and radioshows, listen to something else, and then pick up where you left off?
00:03:15Mikachulook at the 'bookmarks' feature
00:03:21linuxstbOK, fixed it.
00:03:38Mikachulinuxstb: was it #ifndef IPOD_4G\n crash(); #endif ?
00:03:45linuxstbThat was it...
00:03:52linuxstbFancy testing on your Nano?
00:04:46Mikachuvideo only?
00:05:15linuxstbYes - MPEG1 or MPEG2 raw video stream.
00:05:33Mikachuyou may have to provide me with a sample file then
00:05:52Mikachudo i stick that in apps/plugins/ and add to SUBDIRS?
00:05:54linuxstbExtract that tgz to apps/plugins and add "mpegplayer" to apps/plugins/SUBDIRS and to apps/plugins/viewers.config
00:06:48linuxstbUploading a test file for the Nano now.
00:07:00Mikachui only ever encode avi files
00:07:21linuxstbavi containing what?
00:07:39Mikachuvarious stuff, but i don't know how mpeg (de)muxing tools work
00:08:03sugeapple has ipod nano refurbs for $129. Is That a good deal?
00:08:17Mikachuwhat size?
00:08:54linuxstbYou can demux easily using mplayer - mplayer -dumpvideo -dumpfile video.m2v file.avi (if the AVI file contains mpeg video...)
00:08:58 Join polluxx2006 [0] (
00:09:18Mikachubut i would have to scale it for the nano screen too
00:09:45linuxstbYes, to a multiple of 16x16 - so 176x128
00:09:48*BHSPitMonkey 's ears perk up at talk of video playback in #rockbox
00:09:57linuxstbBut anyway, a test is here: (about 10MB)
00:10:36linuxstbBHSPitMonkey: Yes, hell's frozen over.
00:10:51Mikachuokay, testing now
00:10:56Mikachuer, soon
00:10:58Mikachudisk full
00:11:10*S0ap rather liked it when the noob question could more simply be answered "Boot Apple Firmware"
00:11:12amiconnlinuxstb: What about the greyscale targets? Does the decoder allow to skip chroma decoding completely?
00:11:20*petur thanks Bagder and says goodnight
00:11:30 Quit petur ("Zzzz")
00:11:45BHSPitMonkeylinuxstb, I'll go fetch my coat :D
00:11:46linuxstbI was thinking about that, but haven't looked at the decoder to see. Hopefully it should be easy to disable.
00:12:04BHSPitMonkeyhow does it do on your photo?
00:12:46BHSPitMonkey@ linuxstb
00:13:09Mikachulinuxstb: it just said "Loading" and dropped me back to the filebrowser, i'll try again with plain cvs
00:13:49linuxstbNot that impressive - a 224x176 MPEG-2 file is decoding at around 10fps-11fps - but there's lots of scope for optimisations.
00:14:22Mikachuactually i'll just change plugin_get_audio_buffer to plugin_get_buffer
00:14:48Mikachuokay, and now i'll try cvs
00:15:38linuxstbMikachu: Ah, you will also want to reduce the size of BUFFER_SIZE in mpegplayer.c - your test file is less than 10MB, so you can set it to that. It currently attempts to malloc 25MB.
00:16:07 Join thebug [0] (n=thebug@unaffiliated/thebug)
00:16:39Mikachuhooray, that did it
00:16:40Mikachuno cvs for me
00:18:18Mikachu20.4 if it matters
00:18:24Mikachuup and down a bit
00:18:49Mikachuwhat am i looking at?
00:19:37linuxstbStacy's Mom by Fountains of Wayne... It's the only 4:3 video I had lying around.
00:19:48 Join pike [0] (
00:19:52Mikachuand you know this already of course, but -vf field would have been nice here
00:20:03linuxstbYes, I've spotted the interlacing...
00:20:36linuxstbYou can probably try decreasing the bitrate as well - could push up the fps.
00:20:48Mikachudo you read the whole file into ram?
00:21:30Mikachuyou don't have to on the nano at least, playback started after 3 seconds
00:21:31linuxstbIt reads BUFFER_SIZE at a time. So for these test files, yes.
00:21:38Mikachuah okay
00:21:51Mikachuno exit button yet
00:21:52 Nick kclaf is now known as kclaf47 (
00:22:12linuxstbI think a proper player would use one processor for video decoding, and the other for reading data from disk and decoding the audio.
00:22:45linuxstbAnd no, no exit button...
00:22:57Mikachuhm, any reason select+play shouldn't trigger the normal softoff? (normally just play but you know me)
00:23:23linuxstbYes, because it doesn't read the events from the button queue.
00:23:41linuxstb(and then call the default-handler for SYS_POWEROFF)
00:24:21 Join arow [0] (
00:25:12arowuploading to my ipod nano is really really slow
00:25:17arow( 0,30 MB/s)
00:25:23arowthats USB 1.1 speed
00:25:38arowdoes anybody know why it is so slow?
00:26:14linuxstbApple made it that way.
00:26:28Mikachuarow: you have to boot apple's os first, it's a nano-specific bug/feature
00:26:36Mikachulinuxstb: and this was mpeg2?
00:26:42linuxstbMikachu: Yes.
00:26:47Mikachumight mpeg1 be faster?
00:26:59linuxstbI thought so, but in my tests, it was the same.
00:27:10linuxstbBut feel free to experiment with different files.
00:27:15Mikachuwhat do you use to encode mpeg?
00:27:25 Join midkay [0] (n=midkay@rockbox/developer/midkay)
00:27:25linuxstbI use mpeg2enc from mjpegtools.
00:27:38arowMikachu: when i tried the apple disc mode it is faster
00:27:48Mikachuarow: yes
00:27:55arowMikachu: why?
00:27:56linuxstbBut that's mainly because it's the best encoder I've found for generating DVD-compliant files. For Rockbox, something else may be better.
00:28:32Mikachuarow: like linuxstb said, that's the way it's built
00:28:41arowMikachu: is there a way rockbox automatic boots apples disc mode?
00:28:55Paul_The_NerdIt already does automatically boot to their emergency disk mode.
00:29:07Mikachuarow: not yet
00:29:19Paul_The_NerdYou could modify the code so that it automatically boots into the retail firmware instead, but that requires a manual reboot once you disconnect to get back into Rockbox. Not ideal.
00:29:27arowMikachu: that will solve the problem easily
00:29:32Paul_The_NerdIt won't solve the problem
00:29:36Paul_The_NerdIt just changes what the problem is.
00:29:40MikachuPaul_The_Nerd: you can hack around that
00:29:59Paul_The_NerdMikachu: You can make the Apple firmware reboot to Rockbox on disconnect?
00:30:04jhMikeSwhat's this about an MPEG player now?
00:30:06arowPaul_The_Nerd: you are right, but it is easy for nano users
00:30:21Paul_The_Nerdarow: Well, here's the deal: Rockbox on Nano doesn't have users. Only testers.
00:30:34Paul_The_NerdRight now the current method is a lot simpler for people actually doing work on it.
00:30:34MikachuPaul_The_Nerd: on linux, echo -n '{' > /dev/$IPOD will make it reboot after unmount because it thinks it has a new firmware written to it
00:30:48Paul_The_NerdMikachu: That's not a universal solution then. It depends on the computer being set up to deal with it.
00:30:50Mikachui made a 'ipodeject' script
00:31:01Paul_The_NerdI thought you meant there was a way to make the iPod itself deal with it.
00:31:10Mikachuyes, hence my use of the word 'hack' :)
00:31:20Paul_The_NerdWell, it could be an iPod-side hack. :-P
00:31:51Mikachui guess if you can find the piece of code that checks if firmware was written and reboots, and reverse the check it will have the wanted effect
00:31:58Mikachubut that is beyond me
00:32:24Paul_The_NerdI don't transfer lots of files to my Nano enough for it to matter.
00:33:05arowPaul_The_Nerd: but that also doesn't solve the problem :)
00:33:16Paul_The_Nerdarow: I didn't say it did.
00:33:24Paul_The_NerdThe solution for the problem is for Apple to fix the bugs in their firmware.
00:33:41Paul_The_NerdOr for Rockbox to have its own USB mode, which is unlikely to be soon without more information.
00:34:07arowbut i don't know why it is a fault of apple, the apple software is faster..?
00:34:19S0aparow - apple
00:34:30Paul_The_Nerdarow: You're missing the fact that they're BOTH the Apple software
00:34:33S0ap's own disk mode is the 1.1 speeds you see.
00:34:35Paul_The_Nerdarow: As I've said already.
00:34:50arowyes, now i understand, thank you
00:34:55Paul_The_Nerdarow: We boot into the emergency disk mode, because once you're done with that, it can automatically reboot you back into Rockbox. Very handy for developers.
00:34:56*S0ap grumbles about the placement of the '/" key.
00:35:03Mikachulinuxstb: i'm a bit tired, what range of bitrates is good for this res?
00:35:11Paul_The_NerdAnd on all the iPods except the Nano, it's fine.
00:35:17S0ap400 would be about SVCD equiz.
00:35:30Mikachuon some nanos it's also fine
00:35:32arowi just having bad luck :)
00:35:36linuxstbI think that test file I gave you was about 370kbps
00:35:45S0apif you were using B frames...
00:35:52Paul_The_NerdMikachu: So far, what, two of them?
00:36:11Mikachuwell, yeah
00:39:35 Join XavierGr [0] (
00:45:28 Quit sven_ (Read error: 54 (Connection reset by peer))
00:45:39linuxstbBTW, my plugin should now work on a H300 if anyone wants to test...
00:45:53Mikachuhow long did you work on this?
00:46:10linuxstbJust today.
00:47:28jhMikeSX5 or no?
00:48:14linuxstbYes, it should work.
00:48:35 Join MossImo [0] (
00:49:02Paul_The_NerdAnything with a color screen, right?
00:49:34linuxstbIt seems very slow on Coldfire at the moment. Without the yuv2rgb conversion (i.e. decoding only, no display), my h140 is only managing about 21fps on a 176x128 test file at 124MHz.
00:49:44linuxstbThis is probably because there's almost no IRAM use.
00:49:46Mikachuit doesn't like the files avidemux2 produces
00:49:54Mikachuevery other second the whole picture just blurs
00:50:06linuxstbIs it a raw MPEG video stream? What does "file" tell you?
00:50:18Mikachui did the mencoder command you said and got an identical file
00:50:25Mikachutest.mpg: MPEG sequence, v1, progressive Y'CbCr 4:2:0 video, 29.97 fps, Constrained
00:51:12Mikachuand i selected 'mpeg video' as output in avidemux
00:51:18Mikachuand it plays back correctly in mplayer
00:51:23jhMikeShmmm...where does it do the conversion? Yuv->rgb and put it in a buffer?
00:51:29linuxstbEverything plays back correctly in mplayer :)
00:51:47linuxstbjhMikeS: Yes, there's no real optimisation yet.
00:51:57jhMikeSI can see. :)
00:52:21linuxstbIt does the yuv2rgb conversion and copying to the lcd framebuffer in one function. lcd_update() then transfers the framebuffer to the lcd.
00:52:26jhMikeSWas gonna try to get it to compile then mess with some stuff
00:52:59linuxstbYou can try disabling the call to yuv_bitmap_part() and see what fps you get without that function.
00:53:08linuxstb(in video_out_rockbox.c)
00:53:31linuxstbBut as I said, it's only 21fps on my H140 - so I would expect the same on the X5.
00:53:43linuxstb...with that function disabled.
00:54:03jhMikeSGotta parse it into separate files yet or is there an easier way?
00:54:27linuxstbYes, it just accepts a raw video stream - MPEG1 or MPEG2.
00:55:35 Join sven_ [0] (
00:55:37jhMikeSlinuxstb: I gotta separate out all the files by hand? :(
00:56:14linuxstbI would use a program to do it, but if you want to use a hex editor...
00:58:11jhMikeSI got the source open in a text editor all bunched together. WinRar just opens it and show one file "mpegplayer"
00:58:47Mikachuthen i would suggest using another program
00:59:24jhMikeSOk, I'm dumb, should I just use GZIP from the Linux cl?
00:59:35linuxstbAh, you're talking about the source? "tar zxvf mpegplayer.tgz" will work.
00:59:52Mikachuheh, it probably gave you the tar file
01:00:29jhMikeSI think it was missnames just .gz and not .tar.gz or .tgz
01:00:50linuxstbThe mpegplayer directory goes in apps/plugins/. Then add "mpegplayer" to apps/plugins/SUBDIRS and an entry for mpegplayer to apps/plugins/viewers.config
01:01:06jhMikeSI just renamed the extension to "tgz" and WinRar opened it fine.
01:04:30 Part MossImo
01:05:32 Quit suge (Remote closed the connection)
01:06:43***Saving seen data "./dancer.seen"
01:07:09 Quit sven_ (Remote closed the connection)
01:07:30 Join sven_ [0] (
01:09:58Mikachulinuxstb: what input can i give mpeg2enc?
01:10:34linuxstbyuv4mpeg :)
01:11:00linuxstbMy pipeline for that test file was: mpeg2dec -s -o pgmpipe < /video1/FOW/stacy_original.vob | pgmtoy4m -r 25:1 -a 4:3 | y4mscaler -O size=176x128 | mpeg2enc -a 2 -f 8 -q 6 -D 10 -E -10 -K tmpgenc -4 1 -2 1 -o stacys-176x128.m2v
01:12:58linuxstbDo you have mpeg2dec ? That could be useful to test the files that don't work with my mpegplayer - mpeg2dec also uses libmpeg2
01:13:30 Part barrywardell
01:14:11Mikachuwhat do you do if the source is avi?
01:14:52linuxstbI never reencode AVIs - normally MPEG-2 transport streams (captured from DVB broadcasts).
01:14:55jhMikeSadd mpegplayer right by databox
01:15:14Mikachuwell, this is a special case
01:15:54linuxstbMikachu: I'm not the best person to ask about video re-encoding. I always like to keep the original, in whatever format it is.
01:16:53linuxstbjhMikeS: Personally, I've got it next to pacbox (in the same #if)
01:17:03Mikachudo you know how to do single field extraction in the above pipeline?
01:18:13linuxstbNo, sorry.
01:18:14jhMikeSNot a damn clue! =) I'm just getting familiar with linux.
01:18:24jhMikeSoh different comment, oops
01:18:45Mikachuwell, i have a progressive mpeg2 video lying around luckily :)
01:20:09jhMikeSWhat extensions? I've got m2v there, mpeg too?. Do you need the extra 6 bytes of data? What's that, just a header sig?
01:20:12Mikachulinuxstb: what if the input is already a video stream?
01:20:20Mikachulooks like a video stream, not system stream
01:20:25Mikachuit says that repeatedly
01:20:43Mikachuwait i think i got it :)
01:20:51linuxstbjhMikeS: I've just put .m2v in viewers.config. The six numbers are an icon - just copy another entry, or use zeros.
01:21:04linuxstbMikachu: Just remove the -s option.
01:21:09Mikachui tried that but no dice
01:21:27linuxstbWhat does "file" say on your input?
01:21:30jhMikeSThanks. Ends up in viewer or rocks?
01:21:31Mikachuambitious.m2v: MPEG sequence, v2, MP@ML interlaced Y'CbCr 4:2:0 video, CCIR/ITU NTSC 525, 4:3, 29.97 fps
01:21:35MikachujhMikeS: viewer
01:22:08linuxstbMikachu: Looks like an elementary stream to me... If you just type "mpeg2dec ambitious.m2v", does it play it?
01:22:43jhMikeSok, gonna try 'n' build it now...:0
01:23:08 Join MarcoPolo [0] (
01:23:10Mikachulinuxstb: ah, y4mscaler is a separate package
01:23:45jhMikeScompiled with a few "unused parameter: xxxx" but it went.
01:24:15 Quit polluxx2006 ()
01:26:27Mikachulinuxstb: that probably helped :)
01:27:40thebuganyone know if (when the bootloader is closer to ready) the irivier H10 20GB US version will support usb-mass-storage without doing stupid reset + o-on-poweron tricks?
01:28:46linuxstbRockbox won't be able to implement its own UMS mode on the H10 in the near future, so you'll have to use whatever the iriver firmware gives you.
01:29:22Paul_The_Nerdthebug: Until there's documentation or further reverse engineering of the PortalPlayer chip, it's pretty much impossible for us to implement a USB mode for Rockbox itself.
01:30:00thebugok, so here's a further question ... as long as I can use the ums that's built in, will I be able to use the normal play library stuff in rockbox w/o browsing for each file (like I have to when I use ums to load media onto it) ?
01:30:35Paul_The_NerdRockbox won't use the iRiver's library.
01:30:41linuxstbI've no idea how the H10 firmware works, but it sounds like you could use tagcache in Rockbox.
01:30:47thebugok, that's cool
01:30:51Paul_The_NerdBut Rockbox can index the files itself with a scan, and show you them by Aritst/Album/Genre
01:31:02thebugI've just not used rockbox on a player before, so I'm not 100% on what it does and does not do
01:31:23thebugexcellent, that's exactly what I'm looking for ... thanks Paul_The_Nerd
01:31:27Paul_The_NerdI'd suggest skimming the manual.
01:31:33 Join Buddy34 [0] (i=peace@
01:31:49thebugPaul_The_Nerd: will do
01:33:03thebugis the PortalPlayer chip fast enough to handle flac, when the h10 rockbox support is further along?
01:33:24 Part Buddy34
01:33:28jhMikeSplugin won't load, just quits. Should I reduce BUFFER_SIZE?
01:34:21Paul_The_Nerdthebug: FLAC requires a lot less processing power than MP3
01:34:23linuxstbthebug: Easily. FLAC is one of the fastest codecs in Rockbox.
01:34:40thebugawesome ... I really look forward to having that support then
01:34:43thebugyou guys rock :)
01:34:45linuxstbjhMikeS: Does it quit immediately, or after a few seconds of disk activity?
01:35:18thebugthanks for answering my questions, linuxstb and Paul_The_Nerd
01:35:24 Part thebug (":)")
01:36:15amiconnlinuxstb, other ipod users/devs: Is it just me or does Select not work when first pressed after boot, unless the wheel was used before?
01:36:28*linuxstb tests
01:37:36*linuxstb disables the "first keypress enables backlight only" option
01:37:40Paul_The_Nerdamiconn: I regularly have one or another issues with buttons, the hold switch, and the wheel.
01:37:53jhMikeSlinuxstb: Screen goes dark after 1/4sec, then quit almost immediately
01:38:18 Quit sven_ (Read error: 113 (No route to host))
01:38:56 Join sven_ [0] (
01:39:00jhMikeSoh yeah, "loading" stays up for a couple of secs
01:39:15linuxstbamiconn: On my Photo, SELECT seems to work only on the second press - the first doesn't work after boot.
01:39:34linuxstbjhMikeS: Are you clicking on a video file to start the plugin?
01:39:56jhMikeStried, just clicking and "open with..."
01:40:04amiconnlinuxstb: Yes, that's what I observe on my mini as well, unless I use the wheel before
01:40:14jhMikeSOn that stacy's mom m2v
01:40:53Mikachulinuxstb: there is a -vo yuv4mpeg in mplayer
01:41:02Mikachumkfifo stream.yuv
01:41:06Mikachumplayer -vc mpeg12 ambitious.m2v -vo yuv4mpeg -vf-clr -vf scale=176:128
01:41:16Mikachucat stream.yuv | mpeg2enc -F 4 -a 2 -f 8 -q 6 -D 10 -E -10 -K tmpgenc -4 1 -2 1 -o ambitious-176x128.m2v
01:42:15linuxstbMikachu: I tried that recently, and couldn't get it to work properly. But that was with a 1440x1080i H264 file that my PC can't decode in realtime.
01:42:23*amiconn 's got so solve a problem with the grayscale lib on the mini
01:42:52Mikachulinuxstb: i hope the encoder doesn't expect the video in realtime..
01:43:09linuxstbNo, but I think mplayer tries too hard to maintain a/v sync
01:43:28Mikachuyou could use -noframedrop perhaps
01:43:48amiconnThe mini's LCD resolution is _weird_
01:44:04*amiconn has to change a few design decisions
01:45:19 Join lodesi_ [0] (
01:46:21Paul_The_NerdHow is it weird?
01:46:32 Quit MarcoPolo (Remote closed the connection)
01:47:02amiconnThe pixels are packed horizontally and it's 2bpp, so 4 pixels are packed into one byte
01:47:15amiconn138 isn't a multiple of 4
01:48:51JdGordonmorning all
01:48:58JdGordoncongrats linuxstb re video
01:49:22linuxstbThanks, but it's just a first step on a long road...
01:51:08jhMikeSFails when calling mpeg2_malloc even with BUFFER_SIZE at 12MB.
01:51:42linuxstbWeird... It should still work, even with a tiny (e.g. 4KB) buffer.
01:52:02linuxstbAre you running standard CVS?
01:53:25 Join joe2 [0] (
01:54:33linuxstbjhMikeS: Ah, the X5 only has 16MB of RAM?
01:57:32 Quit pike ()
01:57:44 Join pike [0] (
01:57:53 Quit lodesi (Read error: 110 (Connection timed out))
01:57:59amiconnWee, finally :D
02:00:09jhMikeSlinusxtb: I never bothered to look. :")
02:00:45 Quit arow ("Trillian (")
02:00:48Mikachulinuxstb: i can't seem to convince mpeg2enc to use nonstandard framerates...
02:02:42 Quit akaias ("Bye. Need a pastebin?")
02:03:14linuxstbMikachu: Does the MPEG spec allow arbitrary framerates?
02:03:18Mikachuno idea
02:03:29Mikachui just thought it might be nice to try 15fps to get realtime
02:03:35Mikachuno, arbitrary ;)
02:03:50BHSPitMonkeythat was a legitimate typo
02:04:34 Quit mikearthur (Read error: 54 (Connection reset by peer))
02:04:41BHSPitMonkeywell, where I was going with this, was that "it's the end of the summer, and my vocabulary is at its peak decay"
02:04:49linuxstbMikachu: My player ignores the framerate, so if you feed mpeg2enc a 15fps file, just tell it it's 25fps.
02:04:56 Join akaias [0] (
02:05:23Mikachuyeah i suppose
02:06:14jhMikeSI arbitrarily set it to 5MB and it worked. About 10.8-11.3fps.
02:06:31 Quit YouCeyE (Read error: 110 (Connection timed out))
02:06:32 Quit Kohlrabi ("Quit")
02:07:47linuxstbjhMikeS: Is that the 176x128 test file?
02:08:02linuxstbOr the 224x176 one?
02:08:54BHSPitMonkeyhow strong is the compression you're wielding?
02:09:14linuxstbMy 176x128 (Nano-sized) test file is 370kbps.
02:09:57linuxstb(just the video stream). About 10MB for 3m 16s.
02:10:31Mikachuis it very complicated to use two cpus for one video stream?
02:10:52 Join madcow [0] (
02:10:55linuxstbOn the ipods, you'll have the cache coherency to worry about.
02:11:07linuxstbYou also need to read the data from disk and decode the audio...
02:11:27linuxstb(and everything else a video player needs to do, such as maintain a/v sync)
02:13:15 Quit madcow (Client Quit)
02:13:22 Join madcow [0] (
02:14:22linuxstbMikachu: But just optimising the lcd update and yuv2rgb conversion should get the Nano at least doing 25fps video on one CPU.
02:15:03 Join _jhMikeS_ [0] (
02:16:50_jhMikeS_linuxstb: Connection went dead. If you didn't get that, yes it's the that one.
02:17:03dan_aMikachu: Doing anything with 2 CPUs is not trivial! (My brain hurts and I've not even got sound working yet)
02:17:23linuxstb_jhMikeS_: The 176x128 one?
02:17:23Mikachudan_a: hehe, i meant just from the video decoding point of view
02:17:33Mikachuinterdepending frames and such
02:17:34_jhMikeS_linuxstb: yes
02:17:47Mikachuthis is like infinitely cool though
02:17:49_jhMikeS_A link was here earlier
02:18:14linuxstb_jhMikeS_: Have you tested commenting out the call to yuv_bitmap_part() in video_out_rockbox.c ?
02:18:54 Quit spiorf ("Read error: 104 (Connection reset by peer)")
02:18:55_jhMikeS_not yet
02:18:58Mikachuthis also means that in the distant future, when rockbox has usb host support on ipods, i can decode the live stream from my camera and record videos maybe
02:19:04Mikachuit already sends the stream in mjpeg afaik
02:19:45JdGordonforget that... how about digital usb tuners connected to the usb? tv on the go
02:20:02linuxstbA USB DVB-T tuner...
02:20:18Mikachucrystal clear 176x132 digital video
02:20:25midgey34linuxstb: I'm getting between 7.3fps and 8.1fps on my h300
02:20:40linuxstbmidgey34: Yes, the Coldfire needs IRAM...
02:20:52midgey34still, its very cool to see running
02:23:47Mikachulinuxstb: i get 200fps on 1:1 with 75mhz on my nano in test_fps
02:24:19Mikachuso if my midnight math is correct, at 20fps, 10% of cpu is just lcd updating?
02:24:43_jhMikeS_Does it decode interlaced?
02:24:59_jhMikeS_I mean does it actually write in an interlaced pattern to the buffer?
02:25:24 Quit Psilonaut ()
02:25:26linuxstbI'm not exactly sure how it handles interlaced video.. The test file I gave you was interlaced.
02:25:53Mikachuit looked like both fields were updated at the same time
02:26:06Mikachuat the framerate, not fieldrate
02:26:56_jhMikeS_I can see the field lines with motion
02:27:54JdGordonanyone know how to turn mouse gestures off in firefox? studpid thing keeps going back a page when i try scroling wtith the wheel
02:27:55 Join Arathis [0] (
02:28:48 Quit jhMikeS (Read error: 110 (Connection timed out))
02:29:09linuxstbMikachu: Have you managed to encode anything yet?
02:29:18zeJdGordon: disable the mouse gestures extension? how'd you get it installed if you didn't want it?
02:29:39JdGordonze: i dunno
02:29:53JdGordonno extensions are installed
02:31:30 Quit madcow ("CGI:IRC (EOF)")
02:31:39zemouse gestures doesn't seem like something that firefox would have without an extension, especially considering thre's a couple different extensions (afaik) that provide it
02:31:47zesorry, dunno whats up with yours then
02:32:14 Quit TomFelker ("Leaving.")
02:32:35 Join tommeyer [0] (
02:33:16 Join TomFelker [0] (
02:33:31 Quit XavierGr (Read error: 110 (Connection timed out))
02:33:41 Join safetydan [0] (
02:33:48_jhMikeS_where does the decoder actually output the vuy pixels?
02:34:52linuxstbinfo->display_fbuf->buf (in mpegplayer.c) contains the yuv data.
02:35:17linuxstb(in three separate arrays)
02:36:04 Join sharpe [0] (
02:37:11 Quit sharpe` (Read error: 104 (Connection reset by peer))
02:37:35 Quit jn ("Lost terminal")
02:39:37 Part pixelma
02:46:52 Part TomFelker
02:47:28JdGordonhaha, dont put / in the name of a folder on linux :p it fucks everything up
02:47:39JdGordonor maybe its just konquerer
02:48:56 Join BigMac [0] (
02:49:26 Quit gtkspert ("leaving")
02:50:03 Join gtkspert [0] (
02:52:28tommeyerHi, just starting rockbox dev, and trying to work out basic debugging stuff in the simulator. Why doesn't printf show up when I'm using the sim?
02:53:10tommeyerand is there some other way to log debugging info?
02:54:09linuxstbYou can use DEBUBF() in the sim - it works the same as printf.
02:54:43safetydanyou can also attach gdb if you really want to get in to debugging :)
02:55:48tommeyerI've been trying to get it working well with gdb, but how do you use configure to build with debugging info?
02:56:44linuxstbI think debugging info is turned on by default in the sim. Or you can modify the Makefile created in your build directory.
02:57:24tommeyersomehow gdb can't find line numbers, etc., when I'm running sim.
02:58:12safetydanit's on by default and works for me
02:58:24safetydanso... not sure what's going for you
02:58:29tommeyerOK. thanks. I'll poke at it a bit more.
03:01:12_jhMikeS_somewhere deep inside the decoder actually writes the YUV components when decoding the blocks but I'm kinda lost at the moment. know where?
03:02:01amiconn_jhMikeS_: I think it won't be possible to go without intermediate arrays for the yuv data
03:02:29amiconnIf mpeg works similar to jpeg wrt color components (and I think it does), the components are encoded separately,
03:03:32amiconnand so the colour components for a pixel are decoded at different times
03:03:35_jhMikeS_they'd have to be IRAM but I probably is ok if they're separate.
03:03:46_jhMikeS_do they have to be?
03:03:55_jhMikeS_or can it be interleaved
03:04:44amiconnI thought you were looking for a way to remove the intermediate storage, which I think is impossible
03:05:08_jhMikeS_I'm getting to that but it's long path. I have to understand this well first.
03:06:20_jhMikeS_It decodes zig-zaggy, right? Maybe an 8x8 block would be enough
03:06:44***Saving seen data "./dancer.seen"
03:08:40 Part BigMac
03:09:37tommeyerYeah, it decodes zig-zaggy, into Y, Cb and Cr blocks.
03:11:16linuxstbDoesn't it also need to buffer complete previous/future frames, in order to decode the P and B frames?
03:11:25_jhMikeS_1 row of 8x8 blocks might work interleaved Y,Cb,Cr to increase locality.
03:11:57_jhMikeS_Is that the buffer it acually uses? Anyway, there should be no RGB buffer at least.
03:12:19amiconnlinuxstb: Different Q - would pacbox work with the greyscale lib on the greyscale ipods?
03:12:34linuxstbamiconn: Shouldn't be hard - it needs 16 colours.
03:12:56amiconnIt works on the H1x0...
03:12:58_jhMikeS_18 would probably be faster for the x5
03:13:19amiconn_jhMikeS_: 18bits will most probably be slower on X5
03:13:26amiconn...for everything
03:13:38linuxstb_jhMikeS_: Internally, libmpeg2 is only yuv. The code that you commented out (and still only got around 15fps) is the only yuv2rgb code in the player.
03:13:39_jhMikeS_The driver must pack it AGAIN to make it 18-bit to be sent over the bus
03:14:02amiconnBut if you want to avoid bitshifting, you would have to make the framebuffer 32bit
03:14:13_jhMikeS_that is very bad. It's no problem to just pack to 18bit
03:14:24_jhMikeS_the YUV buffer?
03:14:45tommeyeris anyone working on Java or Flash support in rockbox?
03:14:57linuxstbtommeyer: Don't think so.
03:15:43amiconnThe yuv data is usually 8bit precision
03:15:52_jhMikeS_It should pack into xxxxxxxrrrrrrggg and xxxxxxxgggbbbbbb
03:16:14amiconnYou can do that if you're going to write directly to the lcd
03:16:19_jhMikeS_8-bit per component correct.
03:16:27amiconnThe maths itself won't become very different
03:16:37_jhMikeS_amiconn: that's the plan
03:16:41 Join BigMac [0] (
03:16:54amiconnd focus on other optimisations first
03:17:14_jhMikeS_True...most of the time is spent decoding
03:17:30linuxstbIt could also be worth trying to get other mpeg decoders working, to see how they perform against libmpeg2.
03:17:52tommeyeris most of the time spent in DCT, or in other parts of the decoding, Mike?
03:17:59_jhMikeS_I got ffmpeg and core pocket in my list
03:18:09amiconnlinuxstb: Hmpf, the 0.5x scaling in pacbox doesn't fit the mini lcd :(
03:18:24_jhMikeS_tommeyer: I'm just getting to learning this...I really have no idea. Sorry.
03:19:28amiconnOh, I thought pacbox would use the grayscale lib...
03:19:38amiconn...but it doesn't seem to
03:19:50linuxstbNo, it doesn't. I think I misunderstood your question...
03:20:08linuxstbI meant that getting pacbox to work in general with the greyscale lib should be straightforward.
03:21:27linuxstbThe rotate-and-scale operation is the bottleneck on most targets though, so it benefits from assembler.
03:23:07tommeyerit seems that a set of math & geom xform libraries would be of benefit, so fast cos and sin, etc are implemented once
03:23:09amiconnThe grayscale lib eats some more cpu power...
03:29:41 Quit Arathis ("möp")
03:37:05safetydantommeyer, there's already some in various parts of Rockbox if you're talking about fixed point. The equalizer, dps, and replaygain code would be places to look.
03:37:11 Join cismo [0] (
03:42:55 Quit cismo_ (Read error: 60 (Operation timed out))
03:45:38 Quit Criamos (Read error: 110 (Connection timed out))
03:46:57*amiconn crosses fingers
03:47:28 Join bilbravo [0] (
03:48:04bilbravohi all... i have a question... the website lists Daniel Stenburg as starting the Sansa e200 port... how can I get in touch with this individual/offer help?
03:48:23bilbravoah, Badger...
03:48:31bilbravoif i had just looked for 2 seconds longer before asking... sorry :-)
03:50:46linuxstbWow... I just implemented a very quick "lcd_yuv_blit" function, and on my Photo, the fps has gone up from 15fps to 23fps for a 176x128 test file.
03:52:01linuxstbSeems we can use all those waits in the lcd driver to do the yuv conversion...
03:57:26 Join bitshift [0] (
03:57:37bitshifthey, does anyone have an iaudio x5?
03:57:47S0aprockbox porn here we come!
03:58:38amiconn/usr/local/arm/lib/gcc/arm-elf/4.0.3/../../../../arm-elf/bin/ld: address 0x20007f0 of /home/rbclient/build-ipod4gray/apps/plugins/doom/doom.elf section .bss is not within region PLUGIN_RAM <== wtf??
03:59:18linuxstbARM code is bigger than Coldfire code?
03:59:33linuxstbBut I'm not sure why it doesn't affect other ipods...
03:59:34amiconnYeah, sure.
03:59:49amiconnBut it does fit on the colour ipods, and also on the mini
03:59:57linuxstbI know Doom is very tight on the ipods.
04:00:26linuxstbAh, must be the extra size of the greyscale lib and buffers?
04:00:50linuxstbAh, it works on the mini.... (I should read, then talk)
04:00:51_jhMikeS_bitshift: yes
04:01:04amiconnBut why? No such problem on the mini - and it aslo uses the grayscale lib
04:01:55linuxstbAre the greyscale lib buffers static?
04:02:44 Quit lodesi_ (Read error: 110 (Connection timed out))
04:03:59 Join sanka [0] (
04:04:02linuxstbMaybe there are other static buffers dependent upon the lcd size. How close is the mini build to filling the plugin buffer?
04:04:58amiconn3528 bytes left
04:05:41amiconnHmm, the difference in framebuffer size is 5300 bytes...
04:07:39*amiconn wonders how this thing fits on the g5 :/
04:08:05linuxstbLooking in i_video.c, there's a static buffer (graybuffer) declared at the start. I think you could malloc that in I_InitGraphics() (at the end of i_video.c)
04:13:40 Quit dan_a (Read error: 110 (Connection timed out))
04:22:04 Quit sankao (Read error: 110 (Connection timed out))
04:23:03Mikachulinuxstb: yeah i succeeded with mpeg2enc
04:23:16Mikachulinuxstb: stuff from mencoder and avidemux2 didn't work very well though
04:23:34Mikachulinuxstb: but i think they both use ffmpeg/lavc
04:23:53linuxstbYes, they probably do. I've never found ffmpeg's mpeg encoder very reliable.
04:23:56Mikachucould be ffmpeg cvs not feeling well today
04:24:02 Quit ShadowdogMU ("You know you'll miss me :P")
04:24:10Mikachulike i said, i've never encoded mpeg1/2 before
04:24:23 Join fatherfork [0] (
04:24:53linuxstbIt should be close to 25fps on the Nano now if you want an updated patch.
04:25:17*linuxstb should have been in bed hours ago...
04:25:22Mikachume too
04:25:38Mikachui am going to the dentist in 6.5 hours
04:26:36fatherforkalrighty, so I have a freshly updated wps, and I would like to upload it to the gallery, but am having trouble understanding the upload instructions...
04:26:40*amiconn had a better idea...
04:26:50linuxstb (and a new mpegplayer.tgz in the same place)
04:26:56amiconn...regarding that buffer
04:27:04fatherforkcould someone either tell me what i need to do, or direct me to somewhere with easier instructions?
04:29:41Mikachudoes the ipod video also have a 16bit lcd?
04:30:28 Join ShadowdogMU [0] (
04:32:04amiconnMikachu: yes
04:32:20Mikachui could notice the 16bitness on one video...
04:32:24Mikachuon my nano at least
04:33:24Mikachulinuxstb: i have no idea how fast this is because the fps meter doesn't work
04:33:38bitshift_jhMikeS_,are there still bettery issues with the x5 and rockbox? i've read that there was a fix but does it fix long term battery life?
04:33:55Mikachuit sort of looks about realtime though
04:34:42Mikachulinuxstb: ah, it blinks by for half a frame every 5th second or so, i think it wants to say 24.2
04:34:55Mikachuor .3
04:35:41linuxstbThat's not bad...
04:36:47linuxstbIs that with your test file, or mine?
04:37:02Mikachuit was about the same speed as yours before
04:37:03linuxstbDo you know the bitrate?
04:37:08linuxstbAh, OK.
04:37:19Mikachui can try yours too, just wait until this one finishes
04:37:52Mikachuhave you tried other -O settings?
04:38:19linuxstbYes, they didn't seem to make much difference. But please test them again if you have time.
04:39:04Mikachui'm waiting for me to finish eating yoghurt before i go to bed
04:39:19linuxstbI definitely need to go to bed for a few hours now.
04:39:57linuxstbGoodnight all.
04:40:03Mikachugood night, your video says around 24.0 too
04:40:41*Mikachu takes care not to drop nano in yoghurt
04:41:42linuxstbFull-screen on my Photo is still only around 13.5fps - compared to 10fps before my last change... So still a lot more work needed there.
04:42:19 Quit hannesd (Read error: 104 (Connection reset by peer))
04:42:22MikachuO2 seems to not make much of a difference
04:42:33Mikachunone at all in fact
04:42:44 Join hannesd [0] (
04:42:47Mikachuif i changed the right one, that is
04:43:01Mikachu(don't let me keep you up)
04:43:01linuxstbDid you delete the files in build/apps/plugins/mpegplayer/ ?
04:43:29*linuxstb really goes to bed now.
04:43:37ShadowdogMUmpeg player?
04:43:44*ShadowdogMU is excited
04:43:52Mikachufilesize didn't change so i may have failed
04:44:07Mikachui will try more tomorrow
04:44:12Mikachumy yoghurt is nearing depletion
04:44:52Mikachui offer no support for getting it working
04:46:05ShadowdogMUHehe, I'll wait for it to go mainstream
04:48:07tommeyertrying to get it to work on my ipod video right now, but i'm still feeling my way around the whole dev't process...
04:48:32Mikachui will not be surprised if it finds its way into one of those crazy 'experimental' builds
04:49:40ShadowdogMUI want video support
04:49:49ShadowdogMUit's one of 2 things that keeps me from using Rockbox fulltime
04:49:54ShadowdogMUthe other is iTrip support
04:49:58Mikachualso let me point out it's only video, and only mpeg1/2, no audio yet
04:50:24Paul_The_NerdWhy not just get an FM transmitter that isn't locked to the iPod?
04:51:13ShadowdogMUI dunno, aesthetics mostly
04:51:50Paul_The_NerdSo, looking pretty is more important than saving money and better playback. Good stuff.
04:52:09Mikachuwhen you sing along to that 'if i was a rich man' you can skip the 'if i' part
04:58:24 Quit actionshrimp ("( :: NoNameScript 4.01 :: )")
05:01:41 Join Rob2222_ [0] (
05:04:53BigMachi all
05:05:18BigMacwould anyone be interested in downloading a parappa the rapper plugin for the 5g?
05:06:45BigMacguess not so im done working on it
05:06:46***Saving seen data "./dancer.seen"
05:08:33fatherforkI have a wps i'd like some feed back on... if you feel so inclined.
05:09:03fatherforkI just finished it.
05:12:24safetydanBigMac, if it's finished, post it to the tracker for possible inclusion
05:13:41 Part Paul_The_Nerd
05:13:58BigMacno its not but im not going to continue it if no one wants it
05:14:23safetydanDo you want it?
05:16:52safetydanYou're also probably better asking mistic or the rockbox forums if you want feedback on whether it's wanted.
05:19:13 Join fiftyfour123 [0] (
05:19:15 Quit fiftyfour123 (Client Quit)
05:19:36 Quit Rob2222 (Read error: 110 (Connection timed out))
05:20:28 Quit bitshift ("gone")
05:21:57fatherforkHello all, I am Steven. I have registered to the Wiki, and would like Wiki write permissions.
05:29:47 Join Poka64 [0] (
05:39:19 Quit secleinteer ()
05:46:58 Quit BigMac (Read error: 110 (Connection timed out))
05:52:23fatherforkI guess I'll have to wait for someone with the abilitly to add me to the list
05:57:38 Quit fatherfork ("CGI:IRC (EOF)")
05:57:56 Join BHSPitLappy [0] (
06:02:43 Join Harry [0] (
06:02:57 Nick Harry is now known as Harry| (
06:06:36 Quit Harry| (Read error: 104 (Connection reset by peer))
06:12:12 Join devoinregress [0] (
06:16:11 Quit tommeyer (Read error: 110 (Connection timed out))
06:18:07devoinregressthe text is breaking on a few of my 5g wps layouts. the alignment is formatted either right, left, or center and never spaced correctly.
06:18:07devoinregressanyone their?
06:18:07 Quit devoinregress (Remote closed the connection)
06:20:32ShadowdogMUfont issue?
06:25:48 Quit safetydan ("CGI:IRC (Ping timeout)")
06:56:48 Join tommeyer [0] (
06:56:58 Part tommeyer
06:57:00 Join tommeyer [0] (
07:06:49***Saving seen data "./dancer.seen"
07:20:16 Quit midgey34 (Read error: 110 (Connection timed out))
07:23:05 Join midgey34 [0] (
07:25:29 Quit BHSPitLappy (Remote closed the connection)
07:26:47 Join BHSPitLappy [0] (
07:44:29excitatoryi seem to get significantly faster transfer rates when dumping files onto my nano when it is fully charged. would that make sense, or is it just my perception?
07:46:32 Join hotwire_ [0] (
07:49:44 Quit BHSPitLappy ("Leaving")
07:52:09 Join BHSPitLappy [0] (
07:54:54 Join fatherfork [0] (
07:55:40 Join muesli|delhi [0] (n=muesli_t@
07:56:06muesli|delhiis there any israeli around?
07:57:19 Quit BHSPitLappy (Remote closed the connection)
07:57:33muesli|delhior does anybody know a israeli-german/english translating site?
07:58:32 Join BHSPitLappy [0] (
08:00:57 Quit fatherfork ("CGI:IRC (EOF)")
08:02:45Bg3rmuesli|delhi: ?
08:02:54Bg3rtry google
08:03:38Bg3rmorning btw
08:04:34 Quit BHSPitLappy (Remote closed the connection)
08:04:48muesli|delhiBg3r yepp yepp
08:05:01 Quit hexreel ("leaving")
08:05:31 Join BHSPitLappy [0] (
08:10:04 Quit BHSPitLappy (Remote closed the connection)
08:13:03 Join BHSPitLappy [0] (
08:19:30 Quit BHSPitLappy (Remote closed the connection)
08:23:54 Join BHSPitLappy [0] (
08:36:02 Part midgey34
08:36:22 Quit muesli|delhi ("ich will Khe!!!")
08:37:40 Join dan_a [0] (n=dan-mirc@
08:41:03*dan_a spots a grayscale lib in CVS
08:41:10dan_aWell done Amiconn!
08:46:19 Join aliask [0] (
08:51:36 Join theli_ua [0] (
09:02:50 Join petur [0] (
09:03:31 Join einhirn [0] (
09:03:32 Quit einhirn (Client Quit)
09:05:18 Join fejfighter [0] (
09:06:53***Saving seen data "./dancer.seen"
09:07:48 Join safetydan [0] (
09:09:44 Join markun_ [0] (
09:10:53 Quit markun (Read error: 104 (Connection reset by peer))
09:14:34 Join Dumbguy [0] (
09:14:50Dumbguylook im realy lazy,
09:15:07Dumbguyis there a video player for ipod color or ipod nano
09:15:19Dumbguyi cant find any mention of it
09:15:39safetydanDumbguy, no there's no video player in Rockbox
09:15:57Dumbguyso when might there be?
09:16:09Dumbguyany time soon?
09:16:23 Join Zagor [0] (n=bjst@rockbox/developer/Zagor)
09:16:29safetydan"when it's done" is probably the best answer I coud give
09:16:39Dumbguyhmmm ok
09:16:43dan_aDumbguy: I think someone is working on it, and doing pretty well
09:16:44safetydanthough there's been some promising work done towards it lately
09:17:03Dumbguyis there any 3rd party software for rockbox?
09:17:24Slasheripetur: hi, wondering about the "eeprom dump in debug screen: fix return value check"
09:17:27safetydannot really no
09:17:41Dumbguyhmmm ok thanks for your help
09:17:45 Part Dumbguy
09:17:58peturSlasheri: I was wondering how it could have worked for you
09:18:03Slasheripetur: eeprom_24cxx_read returns bool type so it should have been correct
09:18:07Slasherihmm, how?
09:18:31Slasherii wonder if it works now :)
09:18:47peturit didn't work for me until i changed it
09:19:11peturlemme check
09:19:13Slasherihmm, so you "fixed" a non-working driver by inverting the return value? :D
09:20:17peturoh oh...
09:20:38Slasherihehe :)
09:20:54peturI must have confused it with the sw_i2c_xxx calls :(
09:21:10Slasheritrue, they return a different type of value..
09:21:31Slasherihmm, maybe it could be better to modify those exported functions to return int type also
09:21:36Slasheribut feel free to fix that :)
09:21:50peturno time until tonight...
09:22:17peturyou can revert my commit
09:22:37peturdon't spare me in the comment
09:22:59Slasherihmm, i will do that if i have something else to commit too
09:23:52 Quit sanka (Read error: 104 (Connection reset by peer))
09:25:02 Join RoC_MM [0] (
09:25:40peturat least it explains why I was getting unstable readings.... :/
09:26:20peturmust also check why the pcf driver vommits if I leave the i2c lines high
09:27:00Slasheriyes, the eeprom chip seems to be very strict about the correct i2c signaling
09:27:11Slasherihmm, then it sounds like a bug in the pcf driver..
09:27:17Slasherinormally the bus should float hight
09:27:59peturif I leave it high I get a keypress 'right' and battery empty indication. It locks up too
09:28:32Slasheriyou should check in what state the pcf driver leaves the bus after generating the stop signal
09:29:14peturthe normal implementation sets them high, but there's an asm one too
09:29:54peturI've been doing this inbetween other stuff and it's not a good way to work on something like that
09:29:59Slasherihmm, after setting the line high in the eeprom code, try to set the bus floating too (both lines as input)
09:30:05Slasherimaybe that could fix it
09:30:29peturah right, you were setting them as output...
09:30:46peturthat must be it then...
09:30:55Slasherilets hope so :)
09:32:13BHSPitLappycuriosity: what stands in the way of figuring out how to utilize tvout on iPods? (or is even the path unknown?)
09:32:26 Join markun [0] (
09:32:42JdGordonipod has tvout?
09:33:08RoC_MMthrough the flat connector probably...on 5G's.
09:35:04 Quit markun_ (Read error: 104 (Connection reset by peer))
09:35:08BHSPitLappyno, it comes through the headphone connector.
09:35:10aliasknope, throught he headphone jack I think...
09:35:12BHSPitLappyon 5g's and photos
09:35:57BHSPitLappyI think that'd open the door to some nifty potential plugins for rockbox (or iPL)
09:36:16aliaskAye, if we could interface with it
09:36:38BHSPitLappyimagine wowing people by giving a presentation from your iPod :P
09:36:52BHSPitLappygranted, one could do that already, with images for slides in Apple OS
09:36:56BHSPitLappybut not all powerpointy
09:37:05BHSPitLappyI mean
09:37:12BHSPitLappyOpenOffice Impress-y
09:37:12aliaskOpenoffice ;)
09:37:17BHSPitLappybeaten again :)
09:37:21aliaskI won it on my screen
09:37:26aliaskI dont care what you say.
09:39:11aliaskHrm, can't get this mpegplayer to work properly on my H300. Only about 100px x 10px of video playing...
09:40:29 Join dj-fu [0] (
09:50:53BHSPitLappythat's better than nothing.
09:57:16 Quit dan_a (Read error: 110 (Connection timed out))
10:01:03linuxstbaliask: Is that with a file you encoded yourself, or one of my test files?
10:01:39linuxstbAh, looking at the source, I broke it last night for the H300...
10:01:42aliaskI thought I got the encode wrong, but I tried that stacey's mom video, same thing
10:01:53aliaskOh ok.
10:02:28 Join mikearthur [0] (
10:02:53linuxstbAdd a call to rb->lcd_update() around line 218 of video_out_rockbox.c - after the call the yuv_bitmap_part() and before the #endif.
10:03:30Mikachugood morning
10:03:40linuxstbGood morning...
10:03:54Bg3rlinuxstb: morning ;)
10:04:12BHSPitLappygood 3am
10:06:22 Join MarcoPolo [0] (
10:08:34aliasklinuxstb: Did the trick, thanks.
10:09:37 Quit sando ("blargh")
10:11:03 Join dan_a [0] (
10:16:30BHSPitLappyis the compression decent? as far as disk usage is concerned
10:17:18 Join n00b [0] (
10:17:52n00bcan i look at PDF file on my ipod?
10:17:59n00bwith this software
10:17:59linuxstbBHSPitLappy: A 3m 16s test file was about 10MB for the Nano.
10:18:10linuxstb(just the video stream)
10:18:45linuxstbAbout 24fps on the Nano now.
10:18:49 Quit n00b (Client Quit)
10:18:57BHSPitLappyI mean, the video itself
10:19:06BHSPitLappynot how it performs on the hardware
10:19:09linuxstbIt was a PAL file, so 25fps.
10:19:29BHSPitLappythink adding an audio stream would blow it?
10:20:05linuxstbShouldn't do - we would use one CPU for video (which is what happens now), and the other for everything else - disk reading, audio decoding etc.
10:20:42Bagderjtag-attached sansa board received, details follow later
10:21:04linuxstbIs there a developer attached as well, holding documentation?
10:21:18Bagderthere's just a bare board with a cable attached
10:21:26Bagderbut I'll be back later, gotta run
10:21:36linuxstbOK. Bye.
10:23:02JdGordonhehe someone did a commit without cleaning out their tree first.. the keymaps folder is in apps :p
10:23:13 Join spiorf [0] (
10:23:23 Join barrywardell [0] (n=barrywar@
10:23:55JdGordonor im an idiot and its always been there?
10:24:25Mikachuls: apps/keymaps: No such file or directory
10:24:47linuxstbJdGordon: Checking the cvs emails, Linus added it about a week ago (31 July)
10:24:48JdGordon its there...
10:25:14Mikachuah, i use the prune empty dirs with cvs update
10:28:08 Quit RoC_MM ("Leaving")
10:28:12hotwire_what are the max fps the various lcd screens are capable of outputting?
10:28:42*petur thinks there's a wiki page about that
10:28:45linuxstbI think there's a wiki page listing them, but I forget the name of it...
10:40:33Mikachuhotwire_: i get 200fps on my nano on test_fps
10:41:56amiconnhhhhmmmm.... \-:
10:41:59*amiconn silly
10:42:44 Join unknownhost [0] (
10:42:45amiconn155fps on mini, not 60 as I remembered (wrong)
10:43:04*amiconn verifies
10:43:57linuxstb?? Does that mean that you can run the greyscale lib at the full framerate?
10:44:07unknownhostgetting tagcache not ready...?
10:46:00 Quit barrywardell ()
10:49:46unknownhostnevermind... turn it off and turn it on... go figure.
10:50:59dan_aHmmm. I've just upgraded from GCC 3.4.5 to 4.0.3 on ARM and now I get hundreds of warnings about pointer targets differing in signedness
10:51:22amiconndan_a: reconfigure...
10:51:32dan_aamiconn: Ah, thanks!
10:52:32dan_a(I needed to do the upgrade to get the greyscale lib to compile)
10:54:32 Quit Ave (Read error: 60 (Operation timed out))
10:55:37 Join shadou [0] (
10:57:46ZagorI'm updating the server, including apache. cgiirc users may need to reconnect.
10:59:02amiconnlinuxstb: Yes it does mean that...
10:59:15 Quit petur ("CGI:IRC (EOF)")
11:00:11*amiconn is looking at significantly less flickering JPEGs :)
11:00:25aliaskBecause of the FPS difference? seems to be down
11:01:44 Quit mikearthur ("Konversation terminated!")
11:01:56amiconndan_a: [10:57:08] <Zagor> I'm updating the server, including apache. cgiirc users may need to reconnect.
11:01:56dan_a(or at least the webserver)
11:01:58safetydandan_a, see a few lines up
11:03:17ZagorUpdate completed.
11:03:49 Join petur [0] (
11:03:50*amiconn likes how the grayscale lib now looks on his mini
11:04:26linuxstbIs it running at 88Hz?
11:04:45amiconnAll that's needed is to boost
11:05:14amiconnI already had that, but for some reason I thought we merely reach 60fps when boosted... but that's unboosted
11:05:19Slasherilinuxstb: btw, i have found a way to automatically detect if the rockbox version of disk is never than in flash. Then even flash auto-upgrade would be possible to do
11:05:44amiconnAuto flash would be bad imho
11:05:56amiconnThe version on disk might be a test version
11:06:01linuxstbIt would be nice if there's a prompt to confirm the upgrade.
11:06:10Slasheriamiconn: probably true. But that would be calling the flash plugin automatically if user wants to do that
11:06:16aliaskYeah, you should at least be able to put it off until you have a charger or something (if the battery is low)
11:06:17linuxstb(similar to the rolo prompt)
11:06:20 Join afrika [0] (
11:06:43Slasherilinuxstb: hmm, yes. So flash plugin should first confirm if user really wants to run it
11:06:54amiconnaliask: The battery status should be verified anyway by the plugin (@ Slasheri: does it?)
11:06:57***Saving seen data "./dancer.seen"
11:07:02amiconnThe archos flash plugins do...
11:07:40Slasheriamiconn: the current version doesn't yet interact with user at all.. but it's planned to add some verification prompts to the plugin
11:08:16Slasheriamiconn: however, it still verifies all checksums and doesn't flash bootloader unless hardcoded crc32 checksums match
11:08:24 Nick afrika is now known as afrika[a] (
11:09:33 Join barrywardell [0] (n=barrywar@
11:10:53Slasheriamiconn: there is a call rb->battery_level_safe() in the flash plugin but i am not sure if that works correctly on iriver yet
11:10:54amiconnWhy does doom rebuild the tranmaps everytime on ipod? On iriver they're saved to a file after calculating...
11:11:11Slasheriat least it seems it allows flashing at quite low battery levels
11:11:49 Quit afrika[a] ()
11:12:11 Quit dj-fu (Connection timed out)
11:13:41 Join afrika[a] [0] (
11:13:56 Join Mmmm [0] (
11:19:02 Quit hotwire_ ("Chatzilla [Firefox]")
11:23:02 Quit shadou (Read error: 60 (Operation timed out))
11:23:37safetydanShould I bother adding "inline" to any functions in plugins for performance reasons? It's really just hint to the compiler isn't it?
11:23:51 Quit unknownhost ()
11:24:29afrika[a]ask shadou
11:25:10 Join dj-fu [0] (
11:25:11safetydanWell apparently it's worth it as there's quite a few plugins that do it.
11:26:12linuxstbIIRC, gcc won't inline functions automatically unless you specify -O3 (or explicitly use -finline-functions).
11:26:34JdGordoni read somehere it can online inline functions that are in the .h or the same file?
11:26:35amiconngcc does inline functions without -O3
11:26:54amiconnSometimes even with -Os (if the function is only used in one single place)
11:27:13safetydaninteresting, I would have thought it would avoid that for -Os...
11:27:17safetydanoh, single function makes sense
11:27:46 Quit dj-fu (Client Quit)
11:27:52amiconnYes, it saves the call overhead
11:28:06amiconnOf course the function must be static for this
11:28:58 Join Arathis [0] (
11:29:46linuxstbThe manpage says that -finline-functions-called-once is enabled when -funit-at-a-time is enabled, which it is at -O2 and -Os... But -finline-functions is only enabled at -O3.
11:30:51dan_aYou can use __attribute__((always_inline)) too
11:31:17astecpyou can't always trust the gcc manpage on optimisations :)
11:32:43safetydanit's not that performance critical... I just hate thinking of the function call overhead for all these deg_to_rad/rad_to_deg calls I have in this plugin
11:39:18 Quit MarcoPolo (Read error: 110 (Connection timed out))
11:41:22 Join Kohlrabi [0] (
11:56:48 Quit afrika[a] ()
11:58:25 Join amiconn_ [0] (n=jens@rockbox/developer/amiconn)
12:05:21 Quit amiconn (Nick collision from services.)
12:05:22 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
12:05:41 Join lodesi [0] (
12:06:28 Join chendo_ [0] (
12:21:35*linuxstb wonders if he should commit mpegplayer to CVS.
12:21:42JdGordonhell ye!
12:21:48 Quit chendo (Read error: 110 (Connection timed out))
12:21:55JdGordonalthough, if you do expect heaps of bug reports
12:22:32safetydanlinuxstb, do you have grayscale support? I'm curious to see how usable it is on the h120
12:22:34JdGordonlinuxstb: if your itching to commit something... i have a few patches ready to go.. hint hint.. nudge nudge...
12:22:43dan_alinuxstb: Do it!
12:22:55aliaskI can hear them now "There's no sound on my mpeg file" "It's running slowly how do I fix it lol????????????"
12:24:25aliaskI actually already have a bug report, I tried a 1200kb stream because the quality on this file was TERRIBLE (even compared to computer rendering it) and it went rainbow on me (can provide test file)
12:25:01linuxstbFirst I need to decide where my lcd_yuv_blit() function should live. It's currently in lcd-ipod.c, but I don't think it belongs in the core.
12:25:14safetydanplugin lib maybe?
12:26:49linuxstbI think so.
12:27:09safetydanSpeaking of the plugin lib, what do people think the chances are of getting a libm port to the plugins included?
12:27:31safetydanIt's either that, or start looking at rewriting some algorithms in fixed point... which will either be interesting or horrible.
12:31:15 Quit Kohlrabi (Nick collision from services.)
12:31:25 Join Kohlriba [0] (
12:32:08 Nick Kohlriba is now known as Kohlrabi (
12:32:33amiconnlinuxstb: The general rule is that plugins shouldn't bang the hardware directly. Now direct blitting doesn't really belong in the core either...
12:35:11amiconnThis rule is the reason why lcd_blit() exists (for the monochrome and greysacle targets)
12:38:20 Join rconan [0] (n=richard@
12:41:44barrywardellsafetydan: are you working on an astronomy plugin? what will it do?
12:42:59safetydanAt the moment, sun related calculations like this:
12:43:09safetydanbut I hope to extend it beyond being a clone of that sunangle program
12:44:33JdGordonhmm... has anyone else had a problem with the script? it doesnt look like it actually installed m68k-gcc... i did run it as sudo
12:44:59JdGordonarm-elf-gcc and sh-elf-gcc are both there
12:45:15 Join Rudy4Pez [0] (
12:45:16linuxstbamiconn: So do you think such a function should be in the core?
12:45:45 Quit Kohlrabi ("Quit")
12:45:54amiconnI'm not sure
12:46:06linuxstbChecking the size, it's 1412 bytes for the ipod Photo.
12:47:21amiconnI wouldn't want it in the core because it's quite specialised, but otoh plugins shouldn't bang the hardware
12:48:14amiconnAnd I can't think of a way to split this in 2 (hardware access in core, conversion in plugin lib) and keep it fast at the same time
12:48:33amiconnDoom would also profit from a specialised blit funtion
12:49:18amiconn(blit indexed colour with palette)
12:49:36amiconnDoom is breaking the rules on H300
12:53:59 Join barrywardell_ [0] (n=barrywar@
12:55:16barrywardell_safetydan: looks good. what else are you planning on adding?
12:56:49safetydanbarrywardell, ephemeris for planets at least, definitely moon information and so forth
12:57:25Mikachuhow does the "point at sun" thing work on palms? it has a sensor of some sort?
12:57:46Bg3rMikachu: i suppose it uses the local time and your geographic location
12:58:06JdGordonjonno@jonno:~/build-rbdev/gcc-3.4.6$ sudo make install
12:58:06JdGordonmake: *** No rule to make target `install'. Stop.
12:59:35linuxstbSounds like there is no Makefile in that directory.
12:59:45 Join lightyear [0] (
13:00:10barrywardell_cool. i look forward to trying it out when it's done
13:00:12safetydanMikachu, take the local time and location and you can figure out where the sun is on the compass.
13:00:43 Quit barrywardell (Read error: 110 (Connection timed out))
13:00:51Mikachusure, but why does the page say "point your palm at the sun"?
13:01:04 Join uski [0] (
13:01:11linuxstbJust so they can laugh at you...
13:01:18uskihi all :)
13:01:20safetydanit doesn't mean directly :)
13:01:24Mikachui see
13:01:34Mikachuthey have all sorts of fancy ir sensors on those things, you never know
13:01:37safetydanthough there is an IR sensor on the top of most plams :)
13:01:42safetydanpalms even
13:05:33amiconnMikachu: " "mov r0, #0 \n"
13:05:33amiconn "mov %[rx], %[rx], lsr #1 \n"
13:05:33amiconn "adc r0, r0, r0 \n"
13:05:33DBUGEnqueued KICK amiconn
13:05:33amiconn "mov r8, r8, lsr #1 \n"
13:05:33amiconn "adc r0, r0, r0 \n"
13:05:34***Alert Mode level 1
13:05:34amiconn "mov r7, r7, lsr #1 \n"
13:05:36amiconn "adc r0, r0, r0 \n"
13:05:38amiconn "mov r6, r6, lsr #1 \n"
13:05:40amiconn "adc r0, r0, r0 \n"
13:05:42amiconn "mov r5, r5, lsr #1 \n"
13:06:09peturassembling the channel?
13:06:17Bg3rMikachu: did you understand what amiconn wanted to tell u ? :D
13:06:39amiconnMikachu: "point your palm at the sun" is meant literally.
13:07:00***Saving seen data "./dancer.seen"
13:07:32nudelynI found that code very moving.
13:07:36amiconnYou can find out the north/south direction this way
13:08:01amiconnnudelyn: It's even a non-working version I still had in my clipboard...
13:08:43amiconnShould have been movs instead of mov
13:12:06Mikachuright, to relate the directions after you input all the data
13:12:30 Quit chendo_ (Remote closed the connection)
13:12:38 Join chendo [0] (
13:13:04 Quit lightyear (Remote closed the connection)
13:15:35***Alert Mode OFF
13:17:51linuxstbamiconn: Could a yuv_blit function (if it's kept general) be useful for the jpeg viewer? e.g. to speed up panning?
13:19:41Mikachudoesn't that pan a decoded rgb image?
13:20:42amiconnThe jpeg decoder only decodes to yuv.
13:21:19amiconnDisplay decoding is done "live". Requires less memory this way (at least if chroma is subsampled)
13:22:25 Quit Mmmm (Read error: 110 (Connection timed out))
13:22:26Mikachuwould it be too slow to do the whole jpg decoding "live"?
13:22:46Mikachuassuming 1:1 zoom then...
13:24:14Mikachuwould save some time if you zoom in to part of a picture and never intend to view the rest, like a detailed map of the city
13:25:26amiconnIt's not possible with the current implementation. The jpeg data is huffman coded, and that requires decoding from the start
13:25:54amiconnThe huffman decode is a major part of the whole decoding
13:26:11 Join lightyear [0] (
13:26:15Mikachuso when you lossless-crop a jpeg image, it does the huffman decode and encode for the whole file
13:27:13amiconnBut the huffman step is lossless
13:36:08 Join webguest40 [0] (
13:37:36webguest40hi, is there any dev that would be ineterested in committing this patch please, its the "insert random folders" patch,
13:39:38CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
13:39:38*webguest40 looks around the room for life
13:41:14*webguest40 offers free beer
13:41:26peturwhat kind?
13:41:30aliaskDo I have to do anything for it?
13:42:38JdGordonbeer... mmm...
13:42:55webguest40beer for a patch, thats a fair swap :)
13:43:27webguest40free beer for patch 5338
13:43:32JdGordonwebguest40: id actually rather that patch wait untill someone commits my dirwalker patch, becuase this will break it
13:43:55JdGordonof course.. its not up to me, and i wouldnt complain if either/both were commited
13:44:05webguest40ok, dir walker patch, then 5338, offer still stands
13:44:18JdGordonwebguest40 = Falco?
13:44:39webguest40no i'm not falco
13:45:06peturyou are now, JdGordon just assigned your nick
13:45:18webguest40hehe, just someone who would like to have random directory play
13:45:21peturhe should have used ==
13:47:22 Quit lightyear (Remote closed the connection)
13:48:21webguest40I like the dutch Heineken beer advert, send a rover to Mars to look for life, pnce there it puts out a little bar counter, pours a beer and lights up a neon sign saying "beer anyone ?"
13:49:29peturdon't ask my opinion on heineken... how they even dare to call it beer...
13:50:35 Join lightyear [0] (
13:51:08webguest40well, touring a third world country the choice is heineken, local beer or cholora,
13:51:42 Join t0mas [0] (n=Tomas@rockbox/developer/t0mas)
13:51:47JdGordonmmmm.. cholera
13:51:59 Join Mmmm [0] (
13:54:23JdGordonhmm... looking at the linked patch.. i tinhk it might not break the dirwalker patch.. so someone commit it!!!
13:55:38webguest40there you go, come on dev's a worthwhile patch,
13:56:20webguest40it'll bring peace to mankind
13:57:02JdGordoni doubt that... but its worth it anyway
13:57:27JdGordonanyone is aus trying to get rid of a 1.8" hdd for an ipod?
13:58:11aliaskAre iPod hdd's different in any way?
13:58:32JdGordonim after a 1.8" disk tho
13:58:47MmmmJdGordon: I like this patch (5338) but I am not keen on the 20 sec or so freeze that happens when it is first run
13:59:20JdGordon:'( oh, that remidns me.. im supposed to look at the dircache code and see if its useable
13:59:37Mmmmthat'd be good! :)
13:59:42*JdGordon doubts it will be
14:00:51 Quit safetydan (Read error: 110 (Connection timed out))
14:00:56webguest40thank you, laters
14:01:13JdGordonwhere the heck is the dircache code? i cant find it :p
14:01:17 Part webguest40
14:01:25 Join JJure [0] (
14:01:35 Join Kohlrabi [0] (
14:05:27Mmmmpetur: are you still chugging through the AGC patch?
14:05:44Mmmmpetur: seen my latest additions?
14:06:09peturI've only seen that you're working on it
14:06:20peturhave little time
14:06:41peturbut I've been thinking of a better way to handle the recording screen
14:07:00Mmmmsounds exciting!
14:08:20*linuxstb needs to go back to ipod recording...
14:08:50Mmmmpetur: Anyway, I put my latest changes here: hows it going with getting rid of the assumed constants?
14:09:15peturstill need to do that
14:09:44JdGordonMiika Pekkarinen is slasheri ye?
14:09:56JdGordonSlasheri: you around?
14:10:45peturMmmm: it looks like I'll have to make some time (night work) to clear up my stack
14:11:22*Mmmm comes over all dizzy at the thought of night work
14:11:53 Quit fejfighter ()
14:12:49JdGordonMmmm: the major reason for the slow down at the start is because i have no idea how many directories are on the disk, using DC i _could_ use the entry_count, but the problem with that is its files and folders which isnt reallyy that good... if i could change that too 2 seperate entries then it shuold be much faster if DC is enabled.. but i dont know if that is acceptable
14:13:38JdGordonalthough, just looking now, i tihnk that change would be minute!
14:14:17linuxstbBut don't you also want folders that contain music? Or do you pick a random folder, then check it for music content?
14:14:27linuxstb^only want folders that contain music
14:14:41JdGordonye, pick one and check at random
14:15:44linuxstbBut if dircache is enabled, why does it take 20 seconds to scan for all folders?
14:15:57linuxstb(assuming that is what is going on)
14:16:02JdGordonbecause it copies their name to a buffer...
14:16:13*JdGordon is confused.. im not sure which patch im talking about
14:17:36 Quit uski (Read error: 110 (Connection timed out))
14:17:52JdGordonye i am confused.... there should be no 20sec pause in that patch
14:18:14linuxstbMmmm: Do you have dircache enabled?
14:18:18JdGordonits the other one which pauses (the jump to random folder next)
14:18:21 Nick Arathis is now known as Arathis-afk (
14:19:09Mmmmhang im confused...Im talking about the patch that is enabled by choosing random in 'continue to next directory'
14:19:26JdGordonhaha, ye that a different one
14:19:37*JdGordon hopes both will eventually be commited
14:20:25Mmmmsorry...silly me!
14:20:26JdGordonye, that one
14:21:22MmmmThis is my favourite! just that wait....
14:21:57JdGordoncommit 5338 and ill work on 2808 :D
14:22:23MmmmA bribe?
14:24:05MmmmGet 2808 sorted and I'll look at 5338 :D
14:24:09 Quit barrywardell_ ()
14:24:34JdGordonhmm... would this work? could each directory name be hashed and if its within X of current_tick then use that folder?
14:24:43JdGordonwith a suitably quick hash function?
14:27:22 Join lodesi_ [0] (
14:27:37 Join JazzBone [0] (
14:27:42 Join Febs [0] (
14:28:13JdGordonanyone think that could work?
14:29:41JdGordonany other ideas how to grab a directory randomly?
14:30:18MmmmPut themn all in a big hat, shake them around a bit, close your eyes and...
14:30:41JdGordonif only... and thats pretty much what im doing now...
14:32:33 Quit aliask ("sleep")
14:33:34 Join fiftyfour123 [0] (
14:34:00 Quit fiftyfour123 (Client Quit)
14:34:15 Quit JJure ()
14:39:30 Quit lodesi (Read error: 110 (Connection timed out))
14:43:29 Join cismo_ [0] (
14:45:10JdGordonho humm...
14:45:36 Join cq [0] (
14:46:37MmmmJdGordon: are you still doing all that button optimising stuff?
14:46:58JdGordonits bassically ready to go as it is
14:47:13Mmmmare you still being annoyed by all the remotes?
14:47:48Mmmmor did you get that sorted?
14:47:53cqhow 'safe' are dailies for the h140? I'm always a little wary of daily builds...
14:47:53JdGordonif by annoyed you mean not having mapping for any of them except the h100 lcd remote then yes
14:48:03JdGordonperfectly fine
14:49:11MmmmFor the H3xx remotes all they need is to be mapped directly onto the main unit buttons as the buttons are almost identical
14:49:27 Quit cismo (Read error: 104 (Connection reset by peer))
14:49:38peturcq: quite safe, the bootloader is the most critical part and that is not build every day but a fixed and tested download
14:49:52JdGordonalmost, but there are differences.. we have the mechanism to switch between then done, just need the correct mappings for all of them
14:51:05SlasheriJdGordon: are you using dircache in your dirwalker? You should use DIRCACHED instead of DIR and opendir_cached, readdir_cached and closedir_cached
14:51:11Slasherithen it should be fast
14:51:47Slasherii haven't yet replaced those entries globally so that dircache would be used automatically because that would confuse the simulator
14:51:51JdGordonim not, but i can change that
14:52:05*JdGordon assumed they were called automatically...
14:52:46Slasheribut there is no reason why not to use those dircache specific names all over the code
14:52:55Slasherithose works automatically on all targets
14:55:00 Join mooch [0] (n=data@
14:55:20moochwondering if there is any showstopper in the latest ipod video builds..
14:55:40Mikachutry it and let us know
14:55:42 Join Arathis_ [0] (
14:56:05moochand rick loosing my music collection? ;)
14:56:15moochMikachu: i will
14:56:21 Quit Arathis-afk (Read error: 110 (Connection timed out))
14:58:39moochany major features added latelly?
14:58:43cqis there a way to select the default os or is it always rockbox if installed (and the dual boot option)?
14:58:47moochjust out of curiosity...
14:59:01moochcq: in ipod, pressing menu during boot
14:59:44peturon iriver, hold rec when turning it on
14:59:58peturso yes, dual boot
15:00:17 Nick Arathis_ is now known as Arathis (
15:00:33*petur runs off
15:00:41 Quit petur ("later")
15:01:21cqthe smaller font rocks, finally you can see something on the player...
15:01:31 Join barrywardell [0] (n=barrywar@
15:02:20JdGordonSlasheri: is the (struct entry*) tc->dircache always the same value for the same directory? and if yes, is it also the same if DC is off?
15:03:49MmmmJdGordon: The only differences in the remotes are the extra +10/ -10 buttons on the LCD remote. These should be mapped the same as Bitrate/Source on the H100 remote. Everything else the same as the main unit and it works perfectly. I know..I've tried it! :)
15:04:21JdGordonMmmm: isnt the menu/on buttons reversed on one of them also?
15:07:03***Saving seen data "./dancer.seen"
15:07:50 Join Ave [0] (
15:09:05 Quit tommeyer (Read error: 104 (Connection reset by peer))
15:10:32JdGordonSlasheri: the dirwalker uses tc_load and tree_context, is that wrong?
15:11:40MmmmHmm, I haven't looked at your code so I dont really know what you've done but yes, the 'Navi' button acts as 'on' on the LCD remote but, (as the code stands now) mapping this to BUTTON_SELECT does the trick nicely. (It still remains the 'on' button as this is handled separately)
15:12:17JdGordonk, so that needs handling, which isnt a problem
15:13:20Mmmmbut for this remote the NAVI button is by far the best choice for the on button. The play button is one of those plasticky rubbish rocker buttons that could easily be pressed by accident.
15:13:37 Nick Arathis is now known as Arathis_ (
15:14:06JdGordonk, but its defined wrong in the code then? its a simple fix to support all the different remotes
15:14:20 Nick Arathis_ is now known as Arathis (
15:15:25MmmmIt is pretty simple yes, but you still need to be sure it is a remote key you are pressing and not the main unit, so it cant be mapped in button.c straight to the main unit keys
15:16:12JdGordonsimple fix i mean the patch handles it very easily.. and if as you say its 2 buttons which need changing its about 3 min work
15:16:27MmmmAh..ok great! :)
15:17:12Bg3rMmmm: really, h300 lcd remote mappings aren't good atm
15:17:12MmmmMaybe if you get it committed and then I can sort out the other remotes (as I have one)
15:17:22 Join actionshrimp [0] (
15:18:15MmmmBg3r: I know...only because the code is all designed for the H1xx remotes. The H3xx remote keys are the same as the main unit
15:18:32Bg3rJdGordon: so, what's the status of the action patch ?
15:18:40JdGordonMmmm: there is a wiki page for the patch which has the latest version, if you would like to... the file youd need to fix is keymaps/keymap-h1x0_h3x0.c, it is fairly simple
15:19:08JdGordonBg3r: ready to go, except tiny stuff like the F* keys and not all screens are abck to how they were
15:19:35Bg3ris the patch on the wiki the last one ?
15:19:59 Quit Febs (Read error: 110 (Connection timed out))
15:20:21Bg3rbtw, does anyone here have grado SR-XX/Alessandro headphones ?
15:20:59 Nick Arathis is now known as Arathis_ (
15:21:03 Nick Arathis_ is now known as Arathis (
15:21:57JdGordonamiconn: is a simple hash faster than a memcopy on a string?
15:22:24Bg3rJdGordon: simple hash as ?
15:22:43JdGordoni dunno, conveting the string to an int somehow
15:23:07amiconnMost probably not
15:23:34Bg3rJdGordon: really, the memcopy is ... just a memcopy
15:23:37 Join JoeBorn [0] (
15:24:03*JdGordon doesnt think it is possible to have randomness withot the 20s lag
15:24:11amiconnA hash could be faster if memcpy() wouldn't use tricks to speed up operation
15:24:49 Join Criamos [0] (
15:25:23 Quit JazzBone ("Chatzilla 0.9.75 [Firefox]")
15:28:41*linuxstb wonders if earHertz published his ARM memcpy anywhere...
15:30:15cqhm, the iriver interface has the enqueue set to the a/b button, here I have to go through menus to enqueue the next son, is there a way to simplify that?
15:31:51 Quit Pyromancer (Read error: 104 (Connection reset by peer))
15:31:52 Join Pyromancer [0] (
15:35:41JdGordonwhat the heck?? from line 315 filetree.c (in ft_load())
15:35:48JdGordon if (dptr->attr & ATTR_DIRECTORY) /* count the remaining dirs */
15:35:48JdGordon c->dirsindir++;
15:35:48JdGordon }
15:35:49DBUGEnqueued KICK JdGordon
15:35:49JdGordon c->filesindir = i;
15:35:49JdGordon c->dirlength = i;
15:36:20JdGordonoh woops, im asleep
15:36:22JdGordonsoz, ignore me
15:38:59Bg3rcq: long navi press over the track, after that playlist->insert
15:39:49cqbg3r will that playlist just keep growing, or does it die when I turn off the player?
15:40:10Bg3rit will grow up
15:40:25Bg3runless u use the "queue", not "insert" function
15:40:42cqis there a way to change that menu, to have queue at the top then?
15:40:54Bg3ru can move it
15:41:04Bg3rwhen u go to "view current playlist"
15:41:09Bg3rand with long navi
15:42:45cqno, what I mean is change the option so that when I hit longnavi navi I get queue as the first menu option...
15:43:08Bg3r <= look at this
15:43:30cqlooking at the pdf version right now, haven't got that far yet, i'm still inteh instructions
15:44:14Bg3rthere aren't customisable buttons yet
15:44:41Bg3rbut soon it could be possible (at least it will be easier to redefine buttons' functions)
15:45:53cqthe menus are hardwired into the os?
15:46:23linuxstbThe source code is available though...
15:47:03cqyeah, but as I see it, i need to get a crosscompiler compiled under cygwin...
15:47:22linuxstbThere are packages available for cygwin with the crosscompilers
15:47:31 Quit rconan (Client Quit)
15:47:53cqoh? lemme see if I Can find them,,,
15:48:10linuxstbThey're on the Rockbox wiki somewhere - look for pages about Cygwin.
15:51:33JdGordonok, well i changed how next random dir works, about to test it.. its not going to work well with the dirwalker unless that patch is chaneed also
15:52:52cqgot it
15:52:56JdGordonoh goody, the changes dont even work :p
15:58:30 Join crash3m|wrk [0] (n=crash3m|@unaffiliated/crash3m)
15:59:17crash3m|wrkI have an archos jukebox studio20, it plays the first few seconds of each song fine, but then I get 0 output and the time counter keeps incrementing
16:00:09crash3m|wrkany ideas what might be wrong?
16:01:22linuxstbI don't, but if you wait, someone who does have an idea might read your message in the logs.
16:02:01linuxstbHas it just started doing this, or is this the first time you've tried Rockbox on it?
16:02:47crash3m|wrkafaik it just started, but I havent used it in over a year
16:02:55crash3m|wrkinteresting, appears it doesnt happen on every song
16:03:19cqreinstall all the latest stuff maybe? codec issues?
16:03:22crash3m|wrkit might be some of the horrid rips that I have being problematic, when I find another one that does it I'll check it on my laptop
16:15:16 Join secleinteer [0] (
16:16:00crash3m|wrk'dir buffer is full' wouldnt cause this would it?
16:19:17cqshouldn't it be relatively easy to make the menus into some kind of structured text file that can be read at bootup?
16:19:51linuxstbYou would then need a way to associate functions to call in Rockbox with those menu entries. Plus ways to translate them.
16:19:58linuxstb(and voice them)
16:20:39 Quit Mmmm ("Byeee")
16:21:00 Join barrywardell_ [0] (
16:23:06cqwell, if you have the menu functions and names predefined, you could have a config file that structures them...
16:23:11 Join klrspz [0] (n=klrSpz@
16:24:07linuxstbAnd then what happens when menu functions get deleted from Rockbox, or new ones added?
16:24:40cqthat item is ignored from the config file...
16:26:15 Join daurn|laptop [0] (n=quae@unaffiliated/daurnimator)
16:29:06cqor ignore the entire config file and use a default menu
16:29:25cquntil the custommenu parses 100% correctly and is matched up to functions
16:33:05linuxstbFeel free to implement it... But you may just find it easier to cut and paste a few lines in the relevant source file and recompile.
16:33:14cqprobably ;)
16:33:38cqhow do patches wor inthe project? make them and submit somewhere and hope someone includes them?
16:33:50 Quit eGen (Read error: 110 (Connection timed out))
16:34:02linuxstbcq: Basically, yes.
16:35:17linuxstbGenerally it helps getting your patch committed if you discuss your idea here (or the rockbox-dev mailing list) with the other devs - so you're going in the same general direction as everyone else.
16:35:26 Quit barrywardell (Read error: 113 (No route to host))
16:38:00 Join Febs [0] (
16:51:58 Join rconan [0] (n=richard@
16:55:54 Join daurnimator [0] (n=quae@
16:56:07 Quit daurn|laptop (Read error: 104 (Connection reset by peer))
16:59:19 Nick daurnimator is now known as daurn|afk (n=quae@unaffiliated/daurnimator)
17:00:43 Quit Zagor ("Client exiting")
17:03:52 Quit theli_ua ("Leaving")
17:07:06***Saving seen data "./dancer.seen"
17:07:16 Quit Daishi (Remote closed the connection)
17:07:29 Join eGen_ [0] (
17:15:31 Join barrywardell [0] (
17:17:23 Join ziphad [0] (
17:17:37ziphadhello everyone
17:18:06ziphadim having problem with my rockbox
17:18:15ziphadits on a 3g ipod
17:18:36dan_aziphad: What's the problem?
17:18:48ziphad i imported music via itunes and i cant seeit i know thier hidden
17:19:04ziphadso how can i enable tag cache or whatever its called
17:19:07markunziphad: did you read the faq?
17:20:21ziphadok if i understand correctly on the same menu that notes contacts and claender is there is supposed to be music?
17:20:59Mikachusort of, but you have to change the mode first
17:21:17Mikachuhold down menu for a second to get to the quickscreen, then play until it says id3 database
17:21:21Mikachui think
17:21:21ziphadhow do i enable tag cahce browser ?
17:22:04dan_aDo what Mikachu just said
17:22:19amiconnThe database needs to be generated first
17:23:27 Quit barrywardell_ (Read error: 113 (No route to host))
17:24:56 Join ender` [0] (i=null@
17:26:03ziphadok i now have a list: artists albums ect
17:26:14ziphadbut apparently i need to initailize now
17:26:28ziphadfrom the tag chache menu
17:27:23ziphadeverything i do results in tag not ready
17:27:53dan_aziphad: Have you done the "initialize now"?
17:27:59linuxstbYou need to shut down your ipod and turn it back on again after initialising tagcache.
17:28:14ziphadhow do i initialize?
17:28:48ziphadi cant get to the tag cahce menu
17:29:53ziphadim on the 23rd page of the manual and it only tells me the options not how to get there
17:30:16markunziphad: General Settings -> File View -> Tag Cache
17:30:28linuxstbPress (and release) the MENU button to go to the main menu.
17:30:41linuxstbThen what markun said.
17:34:25ziphadok thanks guys
17:34:45ziphadthis stuff should be added to the manuak :)
17:34:47 Part ziphad
17:34:55FebsIt's already in there.
17:34:59 Quit barrywardell (Read error: 113 (No route to host))
17:35:31FebsWhat Markun wrote is exactly what is described in section 2.3.
17:35:57markunAh yes, in 2.3.4
17:37:57CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
17:37:57*dan_a wonders if it would be worth automatically resetting after initializing tagcache
17:38:12amiconndan_a: nope.
17:38:35dan_aWhat would be the problem with that?
17:38:35amiconnThere is already some thought about a reallocation mechanism
17:38:45amiconnThen no reset is necessary
17:39:11dan_aAh... even better!
17:43:04 Join barrywardell [0] (n=barrywar@
17:46:48 Quit thegeek (Read error: 113 (No route to host))
17:47:11 Quit _Veseliq_ (Remote closed the connection)
17:51:34 Join thegeek [0] (
17:59:58 Join Jazon [0] (
18:00:21Jazonso, I have rockbox on my 5g 30GB video ipod - liking it!
18:00:41Jazonquestion though: is there a way to get a bit better gui running on it?
18:01:38Jazoni am using ipl2 boot loader btw... ipodlinux is also installed, but i cant make GPD or whatever its called work, so I am focussing on rockbox......
18:02:10 Quit Febs ("CGI:IRC (Ping timeout)")
18:02:14 Join fatherfork [0] (
18:02:17dan_aJazon: what do you mean by a better gui? Have you looked at the different themes?
18:02:31fatherforkhey everyone
18:03:02Jazondan_a: hi. ummmmm a bit, that affects just the screen while the music is playing?
18:03:22 Join barrywardell_ [0] (
18:04:11dan_aJazon: And the fonts, I think (I've only got a grayscale iPod, so i don't know what else it might do with backgrounds etc.)
18:04:13Jazoni have 4 options dan_a : boxes, iCatcher, rockbox_default, and UniCatcher
18:04:43dan_aWhat do you want to be better about it?
18:04:56fatherforkyou can download themes, but they don't really affect the menu iten layout, only the font and background
18:07:47 Quit barrywardell (Read error: 113 (No route to host))
18:09:15 Join pixelma [0] (i=pixelma@
18:14:01 Join einhirn [0] (
18:14:01 Quit einhirn (Client Quit)
18:14:28 Join ender [0] (i=null@
18:15:44fatherforkcan anyone here grant write permissions in the wiji?
18:17:05 Join bluey- [0] (
18:17:30linuxstbfatherfork: What's your Twiki username?
18:18:49linuxstbfatherfork: OK, you should be good to go.
18:18:58fatherforkfantastic, thanks much
18:19:06fatherforkI have a new theme to upload
18:22:01 Join ender1 [0] (i=null@
18:22:28 Quit JoeBorn ("Leaving")
18:28:23klrspzBagder, what's that about?
18:28:38Bagdera Sansa "dev board" I got from SanDisk
18:28:43Bagderjtag cable attache
18:29:15Bagderyeah, it could mean fun things
18:30:19 Quit ender` (Read error: 110 (Connection timed out))
18:30:28barrywardell_have you tried it out yet?
18:30:33 Join JoeBorn [0] (
18:30:49BagderI didn't get any pinout and I don't have the arm jtag equipment
18:30:54BagderLinus does
18:31:34barrywardell_so how does the jtag thing work? what will it give you?
18:32:00Bagderif things are fine, we might be able to run the existing code, break-point and read memory/registers
18:32:02barrywardell_will it allow you to figure more about the pp chip for example?
18:33:35barrywardell_or is it too early to know?
18:34:04Bagderif we're lucky we can use it to figure out chip details, yes
18:35:04barrywardell_i suppose it will just take a lot of experimentation and trial and error
18:36:02 Quit ender (Read error: 110 (Connection timed out))
18:36:48 Quit bluey- (Remote closed the connection)
18:37:13BagderI don't even know what model of Sansa this is yet
18:38:03Bagderthe flash sizes are covered with tape and sticker
18:38:31 Join bluey- [0] (
18:38:56 Quit bluey- (Remote closed the connection)
18:38:57barrywardell_can't you just hook it up to a computer and find out?
18:39:10Bagderyes I can
18:39:39Bagderjust haven't had the time
18:40:27 Join bondolo [0] (
18:41:04 Join bluey- [0] (
18:43:09 Join Harry| [0] (
18:43:39 Part Harry|
18:44:41 Nick ender1 is now known as ender` (i=null@
18:46:32fatherforkalright, I've uploaded my new 5G iPod theme to the gallery
18:46:52fatherforkthanks again linuxstb
18:49:18ShadowdogMUfatherfork, looks like a large version of the default
18:50:29fatherforkit was my aim to bring is closer to the original iPods look
18:50:41 Quit bluey- (Remote closed the connection)
18:51:38ShadowdogMUYou wanted an exact replica?
18:52:01 Join lamed [0] (
18:52:32fatherforknot exactly, obviously there are things on there that the original firmware did not support
18:53:07ShadowdogMUI know, but your graphics seem too large
18:53:24fatherforktoo large?
18:53:44*ShadowdogMU opens photoshop
18:53:47 Join bluey- [0] (
18:54:18lamedif i'm not wrong, ON, REC & MODE h300 buttons are alt buttons (capabale of a few buttons at the same time) can anyone confirm?
18:54:40fatherforkthe hold icon could be smaller, but the rest are pretty much the same size as iRock's
19:00:11 Join petur [0] (n=petur@rockbox/developer/petur)
19:00:16 Quit dan_a (Read error: 110 (Connection timed out))
19:05:21 Quit rconan (Client Quit)
19:06:14fatherforkGot anything yet shadow?
19:07:09***Saving seen data "./dancer.seen"
19:09:29lamedwhat that means?
19:09:31lamedcvs commit: Using deprecated info format strings. Convert your scripts to use the new argument format and remove '1's from your info file format strings.
19:10:03Bagderlamed: we all get that and we've forgotten why and how ;-)
19:10:15amiconnlamed: It's a cvs configuration thing on the server. Just ignore it
19:10:27ShadowdogMUfatherfork, I do
19:10:34ShadowdogMUa quick mockup
19:10:38amiconnBagder: You should know iiuc...
19:10:40fatherforkalright, lets see it
19:10:59Bagdersince I've forgot I no longer know
19:11:02lamedyou too couldn't have told me if i'm right or wrong on the H300...? never mind. bagder I made some changes to pong
19:11:29fatherforkthat's too minimal, I wanted it to display the extra stuff
19:11:36ShadowdogMUI can add
19:11:38ShadowdogMUwhat do you want?
19:11:53Bagderlamed: no I couldn't
19:11:56fatherforkwell I could add it.
19:12:07fatherforkjust tell me what's wrong with what I have
19:12:29fatherforklike I said, it's based on iRock by Baston, so it's bound to look similar
19:12:38ShadowdogMUeverything looks un-anti-aliased
19:12:52ShadowdogMUand the battery and play logos are too big, imo
19:12:55fatherforkWhen I want a completely new skin, I'll start from scratch
19:12:57lamedbagder: I think I'm right because of button.c:1013, I don't own it as well
19:13:28*amiconn now has some asm optimisations for the grayscale lib on ipods
19:13:49amiconn..after fixing a really silly bug, a missing '!'
19:13:50ShadowdogMUI made that thing real fast to look just like the iPod
19:14:09fatherforkthey re a little large, but it's just because this insn't a complete new skin
19:14:18amiconnarm asm can be confusing...
19:15:08fatherforkok here's what your's is missing:
19:15:19 Join dan_a [0] (n=dan-mirc@
19:15:56fatherforknaxt song is a big one; volume level; playlist position; virtual hard drive activity icon
19:16:42fatherforkhowever, I do agree it would look better with the progress bar above the track time
19:17:40ShadowdogMUdo volume like the iPod does, or is that not possible with themes, and I have playlist position, HDD icon can be a simple add, and flash battery life where Now Player is
19:17:42ShadowdogMUhave them alternate
19:17:48ShadowdogMUcluttered looks bad
19:18:13fatherforkI dono't like alternating text
19:18:42fatherforksure, it's alittle cluttered, but I'm not advertising this to be the cleanest theme ever
19:18:51ShadowdogMUWho are you making the skin for, you or the public
19:19:17ShadowdogMUBecause look at the popular skins, and see what they have that makes them that way
19:19:31fatherforkit's not supposed to look like the original iPod firmware, just have close formatting
19:19:43fatherforklook at the iRock skin.
19:19:55ShadowdogMUStill, are those XP icons as well?
19:20:04fatherforkXP icons?
19:20:15ShadowdogMUthe battery and one of the 2 volume icons
19:20:27ShadowdogMUand the shortcut icon for the next song
19:20:56fatherforkI have no idea. they aren't original.
19:21:13fatherforkthe shortcut icon is from the jBlackGlass theme
19:22:43ShadowdogMUDid you steal all the graphics from that iRock thing?
19:23:11fatherforklike I said, it's a rework
19:23:12ShadowdogMUHowever he made them, they suck
19:23:21ShadowdogMUlook like something from the 1990s, no AA
19:23:25fatherforkI didn't say they didn't
19:23:46ShadowdogMUOh well, if you ever want help, I'm here
19:24:34fatherforkthanks. now that you've beat me up about this... I think I will make my own icons
19:24:44 Quit secleinteer (Connection reset by peer)
19:24:50fatherforkyou've inspired me to make it better, not just move it around
19:25:18fatherforkI didn't do all of that in about 2 hours yesterday
19:25:29 Join secleinteer [0] (
19:26:30 Quit secleinteer (Read error: 104 (Connection reset by peer))
19:26:33ShadowdogMUhave fun, glad I motivated you
19:26:45fatherforkyep. anyway, gotta go, thanks again.
19:27:13fatherforkoops, it should say "I did do all of that in about 2 hours"
19:27:16 Join secleinteer [0] (
19:27:39 Join netmasta10bt [0] (
19:28:21 Quit secleinteer (Read error: 104 (Connection reset by peer))
19:29:05 Join secleinteer [0] (
19:29:14 Quit secleinteer (Connection reset by peer)
19:29:53 Join secleinteer [0] (
19:30:19 Quit secleinteer (Read error: 104 (Connection reset by peer))
19:31:01 Join secleinteer [0] (
19:31:24 Quit secleinteer (Connection reset by peer)
19:32:05 Join secleinteer [0] (
19:32:27 Quit fatherfork ("CGI:IRC (Ping timeout)")
19:32:40 Quit secleinteer (Read error: 104 (Connection reset by peer))
19:33:20 Join secleinteer [0] (
19:41:48 Join mkey [0] (
19:42:38lameddoes " Lars van de Klomp (larsivar) " sound like a real name?
19:43:33lamedare you serious? (i'm not an english native speaker)
19:43:55peturmust be a Dutch guy by the sound of it
19:44:55 Join YouCeyE [0] (n=YouCeyE@unaffiliated/youceye)
19:45:26BagderI don't think we can force people to have names that we think sound proper
19:45:39Mikachushouldn't it be van der? :)
19:46:06Mikachuprobably not, but it sounds more dutch to me
19:46:27*linuxstb looks towards markun
19:46:43linuxstb(and HCl, but he hasn't yawned here for a while....)
19:46:52markunyes, it's dutch
19:46:59markunand t0mas
19:47:13lamedI'll just put "Lars van de Klomp" in the credit list, hope this is correct :)
19:47:38markunit means "lars of the wooden shoe"
19:47:49Mikachuan honorable name
19:48:08markunt0mas: you are dutch
19:48:13 Join Febs [0] (
19:49:00t0masvan de Klomp is indeed odd...
19:49:08t0masvan der would be more "usefull"
19:49:26markunto me it both sounds normal
19:49:36t0masyou're from the east ;)
19:50:01t0masah ok
19:50:10t0masEindhoven university right?
19:50:38markunNo, I live in the east (enschede)
19:53:26 Join ne0futur [0] (n=neofutur@pdpc/supporter/student/ne0futur)
19:53:37ne0futurhi all
19:53:44markunhi ne0futur
19:54:12ne0futurI had a look at your website , what you is great
19:54:59ne0futuris there somewhere I could ask for a firmware for zicplay minikey ?
19:55:23Bagderbut feel free to port Rockbox to it
19:55:41markunne0futur: well, a good start is to collect as much info as you can
19:55:42ne0futurhum not sure I would be able to do it . . .
19:55:53ne0futuris it C language ?
19:57:08ne0futurFace it, if you want Rockbox ported, be prepared to do most of the ground work yourself. You will need skills in electronics and embedded programming in C and assembler.
19:57:10markunthe first steps don't require any programming
19:57:21ne0futurI m good with C but asssembler . . .
19:57:26 Quit joe2 (Read error: 110 (Connection timed out))
19:57:37Bagderso gather your friends
19:57:46markunne0futur: is there a use forum for that player?
19:57:53linuxstbne0futur: IMO, the first thing is to identify the CPU (and other chips). Then find out if documentation is available. That will tell you if a port will be easy (relatively) or hard.
19:57:57ne0futur"First, don't ask Rockbox developers when they are going to port Rockbox to your platform." ;( I were sure i would find something like this ;(
20:00:07*ne0futur wonder how he could identify CPU and other chips . . .
20:00:22markunne0futur: just open it up and take a picture
20:00:24Bagderne0futur: you read the first steps on that page
20:00:24linuxstbOpen it up and have a look
20:00:29markunthen others can help you to indentify
20:00:48Bagderripping apart is always mandatory
20:02:30 Join mikearthur [0] (
20:07:41*ne0futur reached
20:08:01ne0futurbut nothing concerning mine ;(
20:08:21linuxstbThere are far too many MP3 players in the world...
20:08:48linuxstbAlthough yours might turn out to be similar to one of those if you look inside.
20:09:27ne0futuryes for sure
20:10:06ne0futurbut its brand new and i can get back my money if bringing it back within 15 days
20:10:22ne0futurwhich i will probably do
20:10:49 Quit bluey- (Remote closed the connection)
20:13:22 Quit mikearthur (Read error: 54 (Connection reset by peer))
20:13:33 Join bluey- [0] (
20:13:36 Join mikearthur [0] (
20:13:50*dan_a sighs
20:14:09dan_aIt looks like queues don't work across different cores
20:14:31lamedbagder: the cvs server is still on a 2mbit server?
20:14:35*Jazon is away: Away at the moment
20:14:48Bagderlamed: yeps
20:15:16Bagderperhaps we should at least first start providing a read-only one from a faster host
20:15:57lamedand I guess the daily are at another place. that's a wise suggestion, I can't manage a proper repository in a proper time here :/
20:16:24Bagdermost days it isn't a lot to get with updates
20:16:33ne0futurI got the firmware, could the name ( at85413a_FM.bin ) or firmware be useful to see if its near an existing port ?
20:16:47lamedit's about 5-10 minuts now that i'm trying to get a fresh /apps/lang folder
20:17:14Bagderne0futur: most likely not
20:17:27lamedne0futur: not any more then your nickname
20:17:35 Quit Jazon (Remote closed the connection)
20:18:02Bagderne0futur: if the .bin file is plain (unscrambled) code, it _could_ be used to guess the main architecture
20:18:09markunne0futur: if you are going to return it why bother with the port..
20:18:13Bagderbut not any accurately
20:18:17ne0futurstrings on it give things like FWSCF6410 BRECF6410 SD_FF644.drv . . .
20:18:53BagderSD_FF644.drv gets lots of google hits
20:18:53ne0futurmarkun: I will probably but if its near an existing one, i prefer to keep it . . . just dont want ( cant ) spend too muc time on it
20:19:20markuna port takes a lot of time..
20:19:36lamedne0futur: you should know rockbox has had a few abandond ports by now!
20:20:25ne0futurBagder: so this could be useful ( SD_FF644.drv SD_FF321.drv ... ) or really not ? what keywords coul I search with strings on the firware ?
20:20:51Bagderfinding strings will only be a very vague hint anyway
20:21:00Bagderreading the markings off the chip is the only proper way
20:21:13peturone of those strings brought up this:
20:21:15 Join Lear [0] (
20:21:23ne0futurI also have SDK-ITEM/ACTIONSGongee.ZH Wilson Co. Ltd S1 Mp3 Player
20:21:29*ne0futur goes reading
20:22:41 Quit bluey- (Remote closed the connection)
20:24:00 Join bluey- [0] (
20:24:10 Quit ender` (" HydraIRC -> <- 100,000+ downloads can't be wrong")
20:24:18 Join ender` [0] (i=null@
20:25:51 Join _Veseliq_ [0] (n=veseliq@
20:26:37ne0futurthere semms to be no way for opening it without breaking it . . .
20:27:08 Join mirak [0] (
20:27:54markunne0futur: you could talk to the s1mp3 guys in #s1mp3 on this server
20:28:40ne0futurok thanks, I try to understand their wiki first but its not easy
20:29:46mirakmarkun: hello
20:29:52markunhi mirak
20:30:01markunmirak: linuxstb is working on a mpeg player
20:30:04mirakwhat's the frame rate on doom ?
20:30:08mirakmarkun: ah ?
20:30:29mirakmarkun: but not the codecs ?
20:30:39markunask him
20:30:59mirakhe is around ?
20:31:06*linuxstb is around
20:31:20miraklinuxstb: hi
20:31:52miraklinuxstb: you got some new challenge I see
20:32:13linuxstbHi. I took libmpeg2, (the original source, not your port), and have a basic player working - just playing raw video streams. It seems much better on ipods than Coldfire targets though.
20:32:34linuxstbOn the Nano, it can do around 24fps for a full-size (176x128) MPEG-2 file.
20:32:52linuxstbBut I think it's only around 8-9fps on Coldfire.
20:33:32lamed.... can I tell 'patch' to output rejected into a form of a patchfile?
20:33:58miraklinuxstb: how did you ported ? you needed to change a lot of code ?
20:34:11linuxstbSo I'm confident it can work at 25fps on ipods (with audio playback) using one CPU for the video, and the other CPU for audio and disk reading. But Coldfire looks much harder - mainly due to the need to use IRAM as efficiently as possible (for both video and audio).
20:34:14miraklinuxstb: do you use assembly or iram ?
20:34:45miraklinuxstb: well we could live with uncompressed audio
20:35:01linuxstbNeither. There's a tiny bit of IRAM usage, but no assembly. There wasn't that much that needed to be changed in libmpeg2 itself.
20:35:01mirakI would in fact
20:35:27mirakI think the boost could be huge on the idct in assemblt
20:35:35mirakNice work indeed
20:36:08mirakI didn't knew ipods where that powerfull, and especially not they had two cpus
20:36:19linuxstbIf you check the logs (search for mpegplayer.tgz) you'll see a link to it and a patch.
20:36:29linuxstb(today's logs)
20:36:59mirakin fact the 8-9 fps is what I obtained though I didn't knew I could boost the cpu
20:37:36mirakbut well, the 25fps seems a bit out of reach :(
20:38:04miraklinuxstb: did you tried with a mpeg1 stream ? libmpeg2 can decode it, and it should be less cpu intensive
20:38:28linuxstbYes, but the framerate seemed almost identical.
20:40:10miraklinuxstb: the iriver screen is twice the ipod surface
20:40:15dan_alinuxstb: (since you're likely to be one of the first ones to use them) how do the threading API changes I've suggested look to you?
20:40:18miraklinuxstb: you tried wth the same stream ?
20:41:53merbananlinuxstb: ffmpeg should have arm assembly for the idct
20:44:28 Join Madcow [0] (
20:46:03 Quit Madcow (Client Quit)
20:46:29 Join ep0ch_ [0] (
20:46:32 Quit mikearthur (Read error: 54 (Connection reset by peer))
20:46:47 Join mikearthur [0] (
20:48:28Slasheripetur: hi
20:49:11petureeprom_24cxx_write tries 10 times without checking return values of the calls, is that intended?
20:49:29 Join rconan [0] (n=richard@
20:49:31 Join Madcow [0] (
20:50:11xorAxAx20:49:33 -!- Irssi: New peak in #rockbox@freenode : 130
20:50:24peturand write byte itself tries already 100 times
20:51:00ep0ch_linuxstb: nice work with libmpeg2. Seems really wierd how you can get video running quite efficiently on arm and yet audio codecs are dog slow :s
20:51:08Bagderwow, 130
20:51:41Slasheripetur: it does check, by verifying the eeprom contents
20:52:11Slasheriand 100 times because eeprom might be busy some time while writing
20:52:20Slasheriso we will try plenty of times until write succees
20:52:26peturI know. I'll leave that for now as I have no use for writing atm. gonna fix my mistake first ;)
20:52:58rconanhey peeps this is completely off topic but does anyone know a good computer support channel
20:53:14peturSlasheri: I'm making the calls return int as that's what the other drivers seem to do
20:53:23Slasheriok, that sounds good
20:54:49netmasta10btbagder: can u verify that pin1 from e200 dev board is conncted to both pin 19 & 20 of the connector (there is a solder bridge)
20:55:29 Quit rconan (Client Quit)
21:01:31linuxstbmirak: Running a full-screen file on my ipod Photo (same size LCD as H300), it gets about 13.5fps. But I have done one optimisation for the ipods - merging the yuv2rgb conversion and writing to the LCD into one function (lcd_yuv_blit). On the H300, one function does the conversion to rgb (and writes to the lcd_framebuffer), and then lcd_update() is called.
21:01:50linuxstbBefore that optimisation, the same file was about 10fps I think.
21:02:23linuxstbmerbanan: Yes, a suitable ARM idct function will probably be the next thing I'll try.
21:03:04MadcowHi! is there an admin in here? I would like to upload a wps to the gallery.
21:03:48linuxstbdan_a: I haven't thought about your proposals yet.... But you're right, I'll need to soon.
21:04:02netmasta10btBagder: I mean pin 1 and 2
21:04:19peturMadcow: you mean you need wiki write access?
21:07:13***Saving seen data "./dancer.seen"
21:07:51 Join midgey34 [0] (
21:10:24 Quit ep0ch_ (Connection timed out)
21:14:15markunne0futur: your mp3 player has a 60MHz Z80 core with a DSP
21:20:04linuxstb"This product includes technology owned by Microsoft Corporation and cannot be used or distributed without a license from Microsoft Licensing, GP."
21:20:33 Quit spiorf (Read error: 60 (Operation timed out))
21:26:17 Quit Poka64 ("nite all")
21:26:58ne0futuri found a photo of my zicplay :
21:27:21ne0futurmarkun: how did you foud this ?
21:28:09markunwell, I assume your player is one of the s1mp3 player which uses that CPU. I googled to find the datasheet
21:28:15 Join apo` [0] (
21:30:54 Quit ismo (Read error: 110 (Connection timed out))
21:31:09ne0futurmy main problem to gather people is that it seems its sold only in french speaking countries
21:31:33Madcow@petur: yes write acces.
21:32:17 Join ismo [0] (
21:32:39markunne0futur: I think it's better to help the s1mp3 guys with their firmware than porting rockbox
21:33:01ne0futuri posted on their channel
21:33:45ne0futurso you're sure its a s1mp3 ? any url / specifications ? datasheet url ?
21:35:00ne0futur( )
21:37:38markuncan't read what's one the CPU. Only that is starts with AT
21:38:07 Quit Madcow ("CGI:IRC")
21:38:15barrywardell_Bagder: did you see my newest H10 patch?
21:38:50 Nick barrywardell_ is now known as barrywardell (
21:39:07 Join Daishi [0] (
21:39:09barrywardellI moved the ata stuff out of the h10 dir
21:40:00 Quit Lear ("Chatzilla 0.9.74 [Firefox]")
21:41:53 Quit lamed ("CGI:IRC")
21:50:54ne0futurmarkun: i could open it its TEA5767
21:51:43ne0futurTEA 5767 CF 6840
21:52:18 Quit Ribs (".")
21:52:48 Join Paul_The_Nerd [0] (
21:53:32dan_aHmmm... as soon as I enable the timer on the coprocessor, Rockbox freezes :(
21:53:40 Quit Criamos ("( :: NoNameScript 4.02 :: )")
21:57:54miraklinuxstb: ipod video screen is 65000 colors or 240000 ?
21:58:37markunne0futur: that's the FM chip
21:58:48markunnot the CPU
22:00:43miraklinuxstb: how long does it took you to port the codec ?
22:03:05 Join Bger [0] (n=bager@rockbox/developer/Bger)
22:03:52dan_aamiconn: are you here?
22:04:37dan_aI think there's a tiny problem in the grayscale lib
22:05:15dan_aYou've put in a define for NEED_BOOST, which can conflict with HAVE_ADJUSTABLE_CPU_FREQ
22:05:53amiconnoh? In what way?
22:06:04amiconnNEED_BOOST is only defined for portalplayer
22:06:35dan_aIf you undefine HAVE_ADJUSTABLE_CPU_FREQ, it won't compile
22:06:51amiconnAh, yes, but who would do that?
22:07:21dan_aSomeone testing to see if CPU scaling was causing a problem...
22:07:30dan_a(like I just did!)
22:07:49amiconnhmm, I'll fix it
22:08:00amiconnI am currently working on the screendump hook
22:08:05dan_aAs I say, it's nothing major
22:11:09 Quit bluey- ("Leaving")
22:11:29amiconnBtw, does the grayscale lib work properly on the g3?
22:12:05amiconnI could send you the lcd refresh rate test plugin; it's not really suitable for cvs inclusion (aka quick hack)
22:12:15dan_aI've not tested yet, I'm busy making it crash in 100 interesting ways
22:12:28dan_aBut send me the plugin and I'll try it
22:16:21dan_aI hate programs that delete things because they are potentially harmful!
22:17:21linuxstbdan_a: Reminds me of the first time I plugged my brand new iriver H140 into a Windows PC attached to a client's corporate network - the virus scanner immediately deleted the music database file...
22:18:26dan_alinuxstb: Ewww!
22:18:40linuxstbmirak: I had looked at libmpeg2 a little previously, but I started working on it seriously (i.e. making it compile in Rockbox, writing an output driver) yesterday morning.
22:18:49dan_aamiconn: I *think* I've made sense of mIRC's dialog box this time
22:21:07BHSPitMonkeyboo, mIRC.
22:21:10BHSPitMonkey"Hooray Beer!"
22:21:23miraklinuxstb: ok. it took me less time than xvid. xvid was the hardest, and it was still very big binary
22:21:44miraklinuxstb: so it's a codec or a plugin ?
22:22:28linuxstbIt's a plugin. It's far too early to worry about making it a codec.
22:27:43BHSPitMonkeystill cool
22:29:22Bagderthe devboard claims to be a "Sansa e260R"
22:29:51netmasta10btbagder: in windows?
22:30:00Bagderno in dmesg output
22:30:14Bagder Vendor: SanDisk Model: Sansa e260R Rev:
22:30:14Bagder Type: Direct-Access ANSI SCSI revision: 02
22:32:02Bagderit also says 'Rhapsody' (logo) when I start it
22:32:38Paul_The_NerdA development codename?
22:32:41Bagderwhere start means inserting usb cable, as it has no battery
22:33:26netmasta10bthmm rip that BL off there and throw it in yer e200 dir ?
22:33:40Bagderthere's no BL file
22:34:27miraklinuxstb: what's the next step ?
22:34:41miraklinuxstb: assembly, iram ?
22:34:49Bagderbut there's an sda2 in the same way as on my 260
22:34:50 Quit mikearthur (Read error: 54 (Connection reset by peer))
22:34:54*Bagder dds
22:35:07 Join mikearthur [0] (
22:36:27linuxstbmirak: I'm currently cleaning it up, and (unless anyone objects), want to commit it to CVS. Then more optimisations until at least 25fps video plays reliably. Then after that, making use of the ipod's second CPU to do audio decoding at the same time. I'm not planning on doing anything on Coldfire...
22:36:46linuxstbBut once something is in CVS, others can play.
22:36:58miraklinuxstb: you don't like the iriver ?
22:37:07mirakI have one I gues
22:37:13linuxstbI don't own a colour Coldfire...
22:37:27miraklinuxstb: only H100 ok
22:38:06miraklinuxstb: the H100 is not shade of gray, it's just blakc or white right ?
22:38:14netmasta10btbagder: pin1 on the sansa is connected to pins 1 and 2 on the connctor via solder bridge right?
22:38:21linuxstbmirak: It's many shades of grey thanks to amiconn.
22:38:32 Join lamed [0] (
22:38:36 Join uski [0] (
22:38:38linuxstb(and the greyscale lib). Natively it's 2-bit greyscale.
22:38:39amiconnmirak: The H1x0 is 4-grey natively, and up to 33-grey with the grayscale lib
22:39:00Bagdernetmasta10bt: yes, pins 1-5 are one to one
22:39:01mirakamiconn: the lib consist in fast refresh ?
22:39:15Rudy4PezSomeone please explain what "lib" is?
22:39:22netmasta10btBagder: did u see my post on the forum −− let me know what u think of that cable
22:39:36Bagderright, I saw it
22:39:43Bagderand yes it does indeed look like this one
22:39:51miraklinuxstb: will the iram optimisations portable to H300 ?
22:39:52netmasta10bti wonder if there is anything special under the paralel box
22:39:56markunRudy4Pez: library
22:40:04 Join senab [0] (
22:40:38amiconnI guess using the grayscale lib for video will cause more flicker than its other uses, as the pattern change happens often, and isn't synchronised. I'm not sure whether we could apply the render algoritm in haltone.c in realtime...
22:40:43linuxstbmirak: ipods don't seem to benefit much from iram, so no, I think everything should be done specificially for the Coldfire targets.
22:40:45Rudy4PezWell I knew that much, that it was an abbreviation for library, but in this context, what is the definition of library?
22:41:23BagderRudy4Pez: a collection of functions
22:45:11barrywardellBagder: is that 'Rhapsody' as in Real's music store?
22:45:46linuxstbamiconn: Unless you feel strongly against it, I'm planning to commit lcd_yuv_blit() to lcd-ipod.c and export it via the plugin API - it keeps the low-level LCD stuff together, and means I don't need to duplicate the helper functions such as lcd_cmd_data() and friends.
22:46:10senabJust noticed that on the 5g you can now see the LCD flicker if you look close enough
22:46:40Bagderbarrywardell: I don't know, I guess it can very well be
22:47:20dan_aamiconn: The lib doesn't seem to work on the 3g - your test program gave me a blank screen, and Doom gave me flashing vertical lines
22:47:30amiconnlinuxstb: Hmm :/
22:48:01amiconndan_a: Oh? Does the user timer work?
22:48:19amiconnThe test plugin doesn't use the lib
22:48:19dan_aThe metronome does
22:48:33dan_aWhat else uses the lib?
22:49:01lamedanyone seen mmmm latelty?
22:49:15amiconnThe grayscale is used (on ipod) by: doom, fire, plasma, grayscale (if you enable it in SOURCES), and the jpeg viewer
22:50:13dan_aI don't think fire is working, but it could be
22:51:08dan_aPlasma works
22:51:58amiconnNow that is odd. Plasma and fire use the library in the same mode (in fact all plugins enabled in cvs builds do)
22:52:21dan_aI wasn't sure what fire was supposed to be doing
22:52:23amiconnOnly grayscale.c uses the buffered mode (and cube.c on archos)
22:52:24 Join PyromancerX [0] (
22:52:28 Quit Pyromancer (Read error: 104 (Connection reset by peer))
22:55:04dan_aJPEG viewer works. I now have a nice image of an iPod photo on my 3G...
22:55:19amiconnfire shows a fire animation (somewhat slow). You can try it in the sim
22:55:34amiconnMaybe there's some odd #ifdef that needs fixing
22:56:12*amiconn wonders why test_gray2.c would only show a white screen
22:59:37 Quit Arathis ("möp")
22:59:48*linuxstb wonders how to interpret amiconn's "Hmm"
23:00:26 Join Arathis [0] (
23:01:05*dan_a slaps himself in the forehead
23:01:15dan_aThe grayscale.c is working
23:01:35Paul_The_NerdWhat was wrong then?
23:01:37 Quit Bger ("nite")
23:02:44dan_aI was testing too many different things at once
23:07:07 Quit Febs ("CGI:IRC (EOF)")
23:07:14***Saving seen data "./dancer.seen"
23:08:50 Join bytie [0] (
23:09:05 Quit bytie (Remote closed the connection)
23:10:52Paul_The_NerdWe use ffmpeg for MP3, or something else?
23:11:18markunPaul_The_Nerd: we use mad
23:13:12linuxstbffmpeg is used for FLAC and Shorten.
23:14:16 Join EbErT [0] (
23:14:58Paul_The_NerdI recalled someone saying we used it for multiple things, but didn't know what they were.
23:15:16Paul_The_NerdI just knew that it wasn't Helix as someone was claiming.
23:15:33linuxstbNo, the GPL doesn't like Helix's license...
23:15:42Bagderipl uses helix afair
23:15:46linuxstbThey do.
23:16:34linuxstbFor both MP3 and AAC I think.
23:17:06senabwhat does rockbox use for aac decoding?
23:17:21markunsenab: faad
23:17:45markunbut I hope we can switch to the aac decoder from ffmpeg soon
23:17:47Paul_The_NerdWe're hoping for the Summer of Code AAC bits in ffmpeg to be usable though, I believe.
23:18:14senabi thought so, its just im sure someone said it used faad and i didn't beleive them
23:18:17linuxstbAlthough I think that's only going to be floating point - so still a lot of work to use it in Rockbox.
23:18:26bluebrotherwhat would be the advantage of ffmpeg over faad?
23:18:39Paul_The_Nerdlinuxstb: I thought part of what they requested was a fixed point implementation? *can't remember*
23:18:54senabisn't faad no longer developed
23:19:06senabmenno & ivan are at nero iirc
23:19:37markunthere is a license problem with the newer versions of faad iirc
23:20:20markunand we hope that the ffmpeg implementation will be faster
23:21:03senabfaad must now be connected to nero now then, hence the license change
23:23:36 Quit thegeek ("( :: NoNameScript 4.03 :: )")
23:24:13senabi currently make a experimental patched rockbox build and release it on my website, under the gpl am i required to release the source? anyone committing as we speak?
23:24:19 Quit barrywardell ()
23:24:33Bagdersenab: yes you are
23:24:38linuxstbPaul_The_Nerd: I'm not 100% sure. I think the first implement is only floating point, and then a fixed-point version may happen afterwards.
23:24:51Paul_The_Nerdlinuxstb: Gotcha.
23:25:01 Join barrywardell [0] (
23:25:35Paul_The_Nerdsenab: A lot of people don't realize it, but you're actually required to host the whole of the source code you're providing binaries for, not just a patch or diff against source code at another location, if you want to be fully compliant.
23:25:40linuxstbsenab: IIUC, anyone that downloads a binary from you is entitied to request the source you used to build that exact binary. The easiest way to fulfil that obligation is simply to make the source available every time you release a binary.
23:26:08linuxstb(just make it part of your script that generates the binaries)
23:26:22*lamed crossing fingures he haven't broke anything ...
23:27:02senabi thought so, i am hosting the source anyway it's just i was wondering whether it's a requirement
23:27:08*lamed is crossing fingers he haven't broke anything ...
23:27:38amiconnlamed: I don't like that solution...
23:29:44lamedamiconn: I tried to discuss that around the mailing list, no one replied. I discussed that with daniel and he thought it's a good idea. anyhow, I'd love to know your thoughts, we can always revert
23:30:24markunsenab: I think you don't have to host it unless someone asks for it
23:31:08linuxstbmarkun: No, it's not a requirement to host the source, but the alternative is to provide it on request...
23:31:20*linuxstb know what he would prefer to do
23:31:24amiconnInstead of working around a deficiency, we should remove the deficiency.
23:31:56amiconnImho we should port all screens (except debug stuff) to use the ui font.
23:32:11amiconnThen we could even restrict the sysfont to pure ascii
23:33:01senabmarkun, linuxstb: i get requested for it all the time anyway so its easier to just host it
23:33:27*Paul_The_Nerd agrees with amiconn's idea on that one.
23:33:46linuxstblamed: A little bit of red on the build table...
23:34:06Paul_The_Nerdsenab: Was that post to imply that you were changing the name from "Experimental" to "Senab's" or was that directed at Jon, btw?
23:35:02lamedlinuxstb: tt amiconn: brb, fixing reds
23:35:30senabpaul: that i was changing the name from experimental to senab*, it wasn't meant to jon
23:35:54lamedhow do I submit a quiet one? (without it being on the cvs table?)
23:35:56amiconnPorting some screens to use the UI font is a little difficult right now (e.g. the quickscreen, if we want to keep it). Viewports will make those things easier
23:35:57Paul_The_Nerdsenab: Okay, that's what I thought. The "Just kidding" seemed aimed more at my comments anyway. ;)
23:36:28Bagderlamed: submit a quiet what?
23:36:48Bagdernothing escapes the cvs builds and logs
23:37:16lamedbagder: without it being on the main page
23:37:16senabJust made a new WPS for the 5G if anyone wants to have a look. It's called frostBox
23:37:24Bagderlamed: all commits will appear there
23:38:45linuxstblamed: The traditional commit message is "oops"...
23:38:55Paul_The_NerdYou shouldn't mess with tradition.
23:39:33amiconnlamed: Your solution has several disadvantages. One is inconsistency in the UI for non-latin scripts, another is binary size increase
23:40:11amiconnIt forces duplicating quite a number of strings
23:40:51lamedlinuxstb: I've noticed. Bagder: how do you commit whitespace changes and stuff?
23:41:06lamedamiconn: I agree it's much better to translate all menus
23:41:09Bagderyou commit everything the same way
23:41:13 Quit Arathis ()
23:41:28lameddo we have an overall consent over that, so one can start working on it?
23:41:51 Join Arathis [0] (
23:43:07 Quit Arathis (Client Quit)
23:44:03lamedI can imagine a pitch screen using the user font, what about the recording screen on the recorder small display?
23:44:44lamedI'm not sure what you've ment with "inconsistency in the UI for non-latin scripts"
23:44:50amiconnThe recording screen can be simplified a bit, imho.
23:45:15Paul_The_NerdI wonder...
23:45:26Paul_The_NerdYou could also have a simplified multi-font for that?
23:45:35Paul_The_NerdWhere the font's set, and swapped out while you're on the recording screen?
23:45:48Paul_The_NerdSo you don't need to add space to the buffer, but it takes a spin-up to get to the record screen?
23:46:11amiconnI mean that some screens show the native script (cyrillic, hebrew, hangul, whatever), and others show latin text
23:47:01amiconnPaul_The_Nerd: Multi-font support is another story. It's been discussed for years, and I'm sure it will come
23:47:32amiconnSteps taken now (like porting everything to work with variable font sizes) will make it easier...
23:47:51 Join rotator [0] (n=e@rockbox/developer/rotator)
23:48:26 Quit netmasta10bt ("BitchX-1.0c17 -- just do it.")
23:48:46lamedamiconn: I'm still not sure if you understand the patch purpose. you don't see those hebrew characters on the screens that uses the internal font, so the meantime solution would be using latin letters.
23:48:56 Join netmasta10bt [0] (
23:49:05senabis on screen font anti-aliasing on the cards in the future too?
23:49:09amiconnlamed: I do understand it, and I don't like it
23:49:43linuxstbsenab: General view is that it's too CPU intensive...
23:49:48lamed I mean that some screens show the native script (cyrillic, hebrew, hangul, whatever), and others show latin text - > they all can only show latin text.
23:49:54amiconnsenab: Nope, unless someone comes up with a super-render-magic. Text rendering is slow as it is
23:50:03Paul_The_Nerdsenab: That requires some sort of translucency support too.
23:50:15amiconnlamed: Yes, and that's inconsistent. It defeats the purpose of a translation to some degree
23:50:31 Join Arathis [0] (
23:50:49 Quit Arathis (Read error: 104 (Connection reset by peer))
23:51:03amiconnThe screens which can only show latin today are that limited because they use the sysfont - and that is what should be changed
23:51:28 Join Arathis [0] (
23:51:38senabpaul: thats true you'd need some sort of alpha layer
23:52:05amiconnYour solution adds 48 strings, which wouldn't be necessary when the screens are adapted, which take up space, and even take up space for no reason when used with latin scripts
23:52:29amiconnIt's also a rather major change to drop language strings again
23:52:38 Quit Rob2222_ ()
23:52:48amiconn...because it will render all earlier .lng and .voice files incompatible
23:53:14 Join Rob2222 [0] (
23:53:37 Quit Arathis (Client Quit)
23:54:23 Join Arathis [0] (
23:55:47lamedah. ok. that I can agree with you. well then, as you can see there are only 48 new strings for now, and for that price people could use their own native languages (as without this they only get squares around some screens, prefering not to use the translation at all). reverting is as easy as SEDing LANG_SYSFONT_ back to LANG_ strings and removing (future to be) unnecessary LANG_SYSFONT_ strings from the .lang files.
23:57:41lamedmmmm. I wish you have took the time to talk to me about that beforehand. it was a few hours work. you may scrollback if it sounds best for you :)
23:58:53senabPaul: Are you Llorean then?
23:58:55amiconnI talked about the problem here, a few days ago

Previous day | Next day