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 2007-07-03

00:01:01gregjamiconn: I guess so
00:01:14gregjamiconn: I remember I bought mine just 2-3 months before video one came out
00:02:15scorcheamiconn: the 1g, yes"> <== new patch handling CLOCK_0xA0 during transition
00:03:30*amiconn wonders whether voodoo will help here
00:03:55amiconnThe patch is H10 enabled this time
00:03:56digin4in nano is there a way to set a timer to make the screen turn off?
00:06:05 Join piroko [0] (
00:06:32gregjamiconn: so you're expecting it to be 0x00 this time ? not 0xc0...
00:06:40 Quit Lear ("ChatZilla [Firefox]")
00:06:48amiconnIt will visibly always be 0xc000000
00:07:02 Quit AdmiralJ (Read error: 110 (Connection timed out))
00:07:04 Quit relaxed ("")
00:07:06amiconnHowever, it will be switched to 0 during clock transition
00:07:10 Join Dwyloc [0] (i=53688643@gateway/web/cgi-irc/
00:07:18gregjwhich won't be visible
00:07:23amiconnThis is what the OF does - just not on the minis
00:07:46amiconnOn the video it does, but since it's always 0 there it does effectively nothing
00:07:52amiconnThat's why I left it out first
00:08:43 Quit Entasis (Read error: 104 (Connection reset by peer))
00:08:49 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
00:09:22 Join miepchen^schlaf [0] (
00:09:26preglowlinuxstb: faster than libmad was initially, then :)
00:10:06gregjamiconn: lets see
00:10:19gregjamiconn: this time I am running it on lead, just to make sure it's not the power issue
00:10:42Dwylocdigin4: no its not currently posible to switch the screen off on the nano with rockbox while using the player.
00:11:34gregjbut patches are welcomed
00:11:37 Quit bdgraue (Remote closed the connection)
00:11:46 Quit piroko ("BitchX: it's wax ecstatic")
00:13:30gregjamiconn: it is 0xc000000
00:13:32gregjamiconn: all the time
00:13:38gregjjust like it was
00:13:40amiconnyes, sure
00:13:59amiconnYou won't see the 0, as this is for microseconds
00:14:02gregjanytihing else you need, it works for 2 mins now :)
00:14:14amiconnIt should work longer than that....
00:14:14gregjrock solid ;)
00:15:54 Join josh04 [0] (
00:16:15josh04Can anyone explain the wma plugin to me :P
00:16:31gregjin the making
00:16:53josh04I get "Beginning new iteration over decode_audio" constantly on my gigabeat f20
00:19:23 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
00:19:53DwylocAny reason why has not yet been committed as it makes the battery life percentage and run time estimates a lot more accurate on the nano and video?
00:21:22DwylocI know it may not be perfect but its a lot better than whats in SVN at the moment on the ipod nano and video.
00:21:29amiconngregj: Any news?
00:22:17 Join relaxed [0] (n=relaxed@unaffiliated/relaxed)
00:23:11gregjamiconn: still running
00:23:44DesiatoI'm having a nagging iPod Mini 2G problem - I want to power it while listening to it. I hold down the Menu butting when connecting the jack, but about 9 out of 10 times the thing acts like I'm not holding Menu. Any ideas?
00:23:45*petur feels a commit coming :)
00:24:10amiconnDesiato: YOu need to hold it fairly long
00:24:20scorchepetur: by the pricking of your trigger finger?
00:24:32DesiatoPerhaps - I'll try longer
00:24:51*amiconn is testing a lot these days with mini g1 and mini g2
00:24:54gregjI have the same problem
00:25:03gregjit immiedealy switches over to usb , and reboots
00:25:11gregjI have to press menu before plugging in usb
00:25:12gregjand hold it
00:25:15gregjand it works this way
00:25:21amiconnHold menu while plugging, and continue holding it for at least 10 seconds after plugging
00:25:22gregjit really should have 1-2s delay there
00:25:38gregjor should even better - ask
00:25:42amiconnThe 10 seconds are unwanted
00:26:00amiconnBut the requirement to hold menu while plugging is on purpose
00:26:25peturamiconn: where does the 10 sec come from?
00:26:43amiconnDunno... some timeout of the improved usb detection probably
00:26:47gregjamiconn: I am listeing to 'allegro con fuoco' - 'from new world'
00:27:07amiconnIt does this since it can distinguish real usb from an usb charger
00:27:18gregjamiconn: and besides the fact, that it just restarted itself (because of batery, even tho it is connected all the time!) - it seems to be rock solid. I can sign off for commit on my side :)
00:27:23DesiatoDo I need to be on the Now Playing screen when pressing Menu? Does it matter?
00:27:32gregjshouldn't matter
00:27:41amiconnNo it doesn't matter
00:27:51DwylocDesiato: I tend to have the same problem with my nano so I purchased a wall charger. This patch helps as well as it allows you to swaps the USB default insert mode
00:27:53DesiatoOK - thanks
00:28:17DesiatoCool - I have a wall charger too. I will try that.
00:28:22amiconnpetur: You said you know a H10 owner? Would be nice to test this on H10 as well
00:28:32*preglow off sleep
00:28:33amiconn(and I am interested in the CLOCK_0xA0 value)
00:28:58peturare you going to commit or shall I make a build?
00:29:11*amiconn guesses the H10 bootloader sets it up with 0xc0000000
00:29:24gregjI am worried, because it did not charge, with usb-disc mode off
00:29:38peturI'll first have to ask if he's still even using rockbox...
00:29:57amiconnI don't really understand what these bits are for, but from the OF it's obvious these bits should be 0 during clock change
00:31:25gregjdo you just use objdump/gdb, or have some other fancy methods
00:31:42amiconnobjdump and ida pro in parallel
00:31:56 Join SirFunk [0] (
00:31:56gregjida pro is nice
00:31:58amiconn(and grep, and several editors....)
00:32:21amiconnOne of my editors is taking 150MB ram right now...
00:32:33oboCan any wiki experts tell me why %YES% and %NO% aren't working on the LastFMLog page I just edited?
00:32:44gregjI hate ipods for that, they get into that stage where you cannot use it as external disc - it has to charge up a bit first
00:32:46 Quit ender` (" If God had intended us to go around naked, He would have made us that way.")
00:33:12gregjperhaps it is time to go back to home now
00:33:28gregjit is 23:32 here, which makes it 13h in office
00:33:31 Quit Thundercloud (Remote closed the connection)
00:33:40*amiconn is 1 hour ahead
00:34:34peturobo: you need to add something else to the page (meta-I-don't-know-what)
00:34:53gregjamiconn: cos you are on the continent
00:34:55*gregj is in uk
00:34:58scorchepetur: no he doesnt
00:35:24gregjamiconn: see, I don't care about wmv, but i wish someone added h.264 and mov - since there are good opensource implementations already
00:35:28gregjit shouldn;t be that hard
00:35:33amiconn%YES% and %NO% aren't working out of the box
00:35:38amiconn%Y% and %N% are
00:36:03amiconnFor the former, you need to add some kind of link. I looked that up once, but forgot what it was...
00:36:11scorcheno.... %x%
00:36:20bluebrotherobo: check TWiki/TWikiVariables
00:36:31amiconngregj: Forget h.264 (except _maybe_ on gigabeat)
00:37:12markuncpu power
00:37:47 Quit Desiato ("CGI:IRC (Ping timeout)")
00:38:08amiconnAnd mov is a just a container
00:38:32bluebrotherobo: afaiu you need to #set YES = <somelinktothegraphic> at the top of the page
00:39:21gregjamiconn: well, apple did that. besides - mpeg is veryyyysllloooow here
00:39:25obobluebrother: I'm looking at the BuyersGuide page, I can't see anything like that...
00:39:37gregjmencoder -of mpeg -oac lavc -lavcopts acodec=mp2:abitrate=192 -af resample=44100:0:0 -ovc lavc -lavcopts vcodec=mpeg2video:vbitrate=212000 -vf scale,harddup -ofps 25 -vf scale=224:176 -o satriani.mpeg
00:39:38amiconngregj: Yes, apple did - on the ipod video _only_.
00:39:41markunpetur: and you think h.264 will be faster??
00:39:42gregjmy onversion line
00:39:48amiconnAnd that has a hardware decoder for it
00:40:07markunoops, that was for gregj
00:40:10amiconnBut, as usual, this chip has no public docs
00:40:16*petur deflects highlight with mirror
00:40:18bluebrotherhmm. Right. Somewhat strange.
00:40:21linuxstbsaratoga (if you read the logs), a new version of the wma patch (which now works on big-endian CPUs - I had to fix some things in wmadeci,c) is here -
00:40:35bluebrotherbut using %Y% and %N% is shorter anyway ;-)
00:40:45obotrue :)
00:40:46 Nick amiconn is now known as {[ (n=jens@rockbox/developer/amiconn)
00:40:58 Nick {[ is now known as amiconn (
00:41:09scorcheobo: i changed it at least temporarily
00:41:23obothanks scorche
00:41:40gregjamiconn: how many mhz this cpu has (g4/photo)
00:41:58amiconnDual ARM7TDMI, 80MHz
00:41:59gregjthis mpeg player is not skipping frames well enough
00:42:06gregjdualcore ?
00:42:10amiconnWe're running it at max. 75MHz (78 with my patch)
00:42:16markunit skips them very well on the Gigabeat
00:42:24gregjnot on ipod
00:42:30josh04I watched all of primer on my way to london on saturday
00:42:34markungregj: talk to jhMikeS about it, he spent a lot of time on the skipping syncing
00:42:34josh04It's pretty impressive
00:42:56markunjosh04: yes, really not bad
00:43:09gregjsound is perfect
00:43:13gregjlemme try with 80kbps
00:43:26gregjit says 212k in manual, but that's dreaming
00:43:34amiconnlinuxstb: Do you think I should keep the 78MHz (if everything is stable)?
00:43:49linuxstbI can't think why not.
00:44:00scorchejosh04: that movie is complex enough without worrying about mpegplayer...
00:44:21josh04scorche: it was the second time round, naturally
00:44:25josh04I still don't get it
00:44:32markungregj: the bitrate in the kbit, right?
00:44:40scorcheread the wikipedia page and then watch it again
00:44:47gregj -lavcopts vcodec=mpeg2video:vbitrate=212000
00:44:50markungregj: I use vbitrate=400 for my gigabeat
00:44:59gregj400 ?
00:45:01markunquite a difference from 212000
00:45:14josh04I saw it, but the failsafe boxes still confuse me
00:45:36gregjso is that setting in kbits ? so I should just put 212 there ?
00:45:41*gregj scratches head
00:46:06gregjlets see
00:46:21markungregj: wait, they just have a strange syntax
00:46:40markungregj: 4-16000 (in kbit) 16001-24000000 (in bit)
00:47:47markunso you could use either 212 or 212000
00:47:50amiconnjhMikeS: around?
00:48:04gregj212000 was sluggish
00:48:23gregjcompletly not acceptable
00:48:29josh04linuxstb, i spot your name in the wiki. Can you give me a pointer for wma on the gigabeat ;P
00:48:35markungregj: how's the sample video?
00:48:44linuxstbgregj: mpegplayer can only skip decoding of B frames, so you may need to tweak mpegplayer to use more of them.
00:48:44markunElefants Dream
00:48:52linuxstb^I mean tweak mencoder
00:49:42linuxstbjosh04: Apply the patch, compile, install, listen...
00:49:49linuxstbIf it doesn't work, fix it...
00:50:44gregjvmax_b_frames ?
00:50:45josh04Do you mean the patch in flyspray? the one that comes as a folder called wma2wav?
00:51:17linuxstbNo, the patches I've been linking to in IRC today.
00:51:35linuxstbThis is the latest -
00:51:45josh04Ah, okay. Thanks a lot
00:51:45linuxstbThe ones in flyspray are not working codecs.
00:52:50gregjamiconn: encoding with more bframes makes it very slow (encoding)
00:52:53josh04that'll explain it. time to patch!
00:52:55gregjamiconn: will see the results tho
00:53:01jhMikeSamiconn: yeah
00:53:09gregjamiconn: but I guess, someone has to fix the docs on wiki than!
00:53:39amiconnjhMikeS: Do you have an unencrypted sansa firmware dump?
00:56:49*amiconn wonders whether gregj confused him with s/o else
00:57:08*amiconn also wonders whether gregj's ipod still didn't freeze
00:57:54 Quit Guile`` ("KVIrc 3.2.0 'Realia'")
00:58:04gregjamiconn: still not freeze :P, but I stil lhave problems with mpeg playback
00:58:19gregjamiconn: I am on 80kbp, plenty of b-frames, and stil the same problem!
00:58:59*amiconn still wonders whether gregj confuses him with s/o else
00:59:22linuxstbgregj: Encode with a lower fps.
00:59:27*linuxstb sleeps
00:59:30bluebrothergregj: is it a greyscale Ipod?
00:59:31gregjmanual says 212k!
00:59:34gregjnope, color
00:59:48gregjand I use 80k, still too much
01:00:32bluebrotherhmm. Thought you mentioned a 4G
01:00:47amiconn4G's come in 2 flavours
01:01:41bluebrotheroh −− thought 4G is always greyscale.
01:01:43gregjit is the - color/u2/photo thingie
01:01:48josh04Woohoo. Working first time. Thanks, linuxstb
01:01:49gregjbluebrother: wrong
01:02:01gregjbluebrother: take a look on linux on ipod site
01:02:05bluebrothergregj: already notices that ;-)
01:02:22 Quit josh04 ("Lost terminal")
01:03:46***Saving seen data "./dancer.seen"
01:04:11gregjamiconn: you're creating those patches, aren't you
01:04:19 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
01:04:25bluebrotherbtw, we have a nice manual which is different from the wiki. The manual doesn't mention anything about bitrates for mpegplayer ;-)
01:04:43amiconngregj: For video? Surely not
01:04:46gregjthan fix the woki
01:04:48 Join miepchen^schlaf [0] (
01:04:51gregjamiconn: I know, but cpu one
01:05:53 Quit jgarvey ("Leaving")
01:06:43amiconnyes, but what does that have to do with mpegplayer?
01:07:22gregjnothing :0
01:07:33gregjbut you are someone who does something here, and I can see it :)
01:07:58bluebrothergregj: the idea behind the wiki is that users can adjust the contents if they find something is wrong
01:08:00gregjnow mpeg player behaves better (on 80k), but every few s it stucks for 1s, and than tryies to catch up
01:08:10gregjbluebrother: true
01:08:32bluebrotherand I won't "correct" values I can't verify. Other users can ...
01:09:21gregjthat's why I am trying different settings for mencoder
01:09:47gregjonce I am done, and satisfied - I will give them in your capable hands (I don't have account for wiki, nor wish to setup one now - it is late)
01:10:10gregjipod doesn't charge the battery - if I do - menu+cable thing
01:10:13gregjjust doesn't...
01:10:28bluebrotherit doesn't charge or it doesn't display it's charging? ;-)
01:10:36gregjit doesn't charge
01:10:49gregjeven with lead connected it shuts down after a while
01:11:01gregjand reboots to b/w screen with battery-loading icon
01:11:03bluebrotherdo you have it playing?
01:11:23bluebrotherwell, it might not draw enought power from usb to actually charge.
01:11:24gregjmaybe it gets less power from the lead, than is required to sustain operations
01:11:43 Quit obo ("bye")
01:12:07bluebrotherI sometimes charged my Ipod that way, and it charged. Unless I was playing music, in which case the charging current wasn't high enough to actually charge.
01:12:16gregjwhich is ridiculous
01:12:18gregjbut anyway
01:12:35gregjwhat's the point of doing that, if you don't listein to anything meanwhile ?
01:12:47bluebrotherno. It's just another unknown / not enough known part of the hardware
01:12:55 Join Nico_P [0] (
01:13:03amiconnbluebrother: mini g2 charges, _very_ slowly. When playing music it depends on the boost ratio
01:13:05gregjmaybe you have to tell it somehow - charge
01:13:26amiconnThe fullsize ipods discharge overall (but slower than without the wire)
01:13:33bluebrotheramiconn: my 2G charges indeed. Maybe my music boosted too much ;-)
01:16:49bluebrotherwell, time for sleep
01:16:58 Quit bluebrother ("zzz ...")
01:18:02 Join chuckwag0n [0] (
01:18:45 Quit spiorf (Remote closed the connection)
01:19:26chuckwag0njust wondering if rockbox has the same feature as the apple software when it comes to scrolling (I have a million directories and it takes forever to get to the bottom
01:19:42chuckwag0nseems like when I change the scroll speed it makes no difference
01:21:15gregjit is a known issue (so says wiki)
01:21:45chuckwag0nso I probably need to make folders with letter?
01:23:23 Quit Rick ("I don't need to be here.")
01:24:41 Join Rick [0] (
01:24:47gregjchuckwag0n: probably
01:25:12chuckwag0ndoes my fm transmitter work?
01:25:34gregjmencoder -of mpeg -oac lavc -lavcopts acodec=mp2:abitrate=48:vmax_b_frames=16:vb_strategy=2 -af resample=44100:0:0 -ovc lavc -lavcopts vcodec=mpeg2video:vbitrate=90 -vf scale,harddup -ofps 25 -vf scale=224:176 -o satriani.mpeg
01:25:38gregjperfect setting here
01:25:47gregjcan someone stick those params on wiki for G4 ?
01:26:04 Join saratoga [0] (i=9803c6dd@gateway/web/cgi-irc/
01:26:13 Quit anathema (Read error: 113 (No route to host))
01:26:53peturregistering takes no time and write access can be granted the moment you tell us your wikiname
01:27:35petur(wikiname must be FirstnameLastname)
01:28:21saratogawhat is the proper check to see if a target is ARM?
01:28:57gregjpetur: grzegorzjaskiewicz
01:29:17peturall lowercase?
01:29:27gregjthis way
01:29:59gregjguess I have to check me email first anyway
01:29:59 Quit entheh (Remote closed the connection)
01:30:27amiconnsaratoga: #ifdef CPU_ARM
01:30:32peturwhere's the lastname? Normally it's FirstLast
01:30:40amiconn(this includes all arm variants)
01:30:42gregjpetur: it is FirstLast
01:30:42petur(note the capitals)
01:30:56gregjpetur: GrzegorzJaskiewicz
01:30:57peturoh sorry, I see
01:31:00gregjsimple, a!
01:31:28*petur blames 15in lcd on 1920x1200
01:33:37peturgregj: write permission granted, I guess you're waiting for that mail...
01:34:44 Quit chuckwag0n ()
01:35:23 Quit Dwyloc ("CGI:IRC (EOF)")
01:36:20peturthere you are
01:37:48saratogasomehow linuxstb's patches always break my SVN checkout and i have to delete it and start over
01:38:54gregjpetur: cheers mate
01:39:16*petur holds up glass
01:40:04gregjit is a way to say thanks very much in english
01:40:32gregjand I mean england, not us
01:40:42peturI know ;)
01:44:12 Quit stripwax (Read error: 104 (Connection reset by peer))
01:44:56 Quit relaxed (Read error: 110 (Connection timed out))
01:48:48 Part bushblows ("The day before 2 days after yesterday.")
01:52:25*petur gives up on fat32
01:53:15 Quit petur ("Zzzzz")
01:55:21 Join Thundercloud [0] (
01:59:07saratogaamiconn: I'm trying to get the arm asm from a block of code, but I can't seem to do it
01:59:19saratogaarm-elf-gcc complains that there is no crt0.o
01:59:27saratogais there some way to do this that I'm missing?
01:59:38 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
02:00:03 Join miepchen^schlaf [0] (
02:07:18 Join webguest81 [0] (i=44fbfdec@gateway/web/cgi-irc/
02:07:37webguest81how do you install rockbox on a 5g ipod video?
02:07:53saratogawebguest81: read the manual
02:09:42 Quit webguest81 (Client Quit)
02:09:47 Join webguest81 [0] (i=44fbfdec@gateway/web/cgi-irc/
02:10:33 Join webguest78 [0] (i=44fbfdec@gateway/web/cgi-irc/
02:10:42webguest78so where do i get the manual?
02:10:42 Quit webguest81 (Client Quit)
02:11:35DerPapst left bar.
02:11:42DerPapstthere is a link called manual
02:12:01webguest78ok thanks
02:12:41 Quit Thundercloud (Remote closed the connection)
02:15:33 Join Soap [0] (n=Soap@rockbox/staff/soap)
02:16:41gregjamiconn: it crashed!
02:16:45gregjamiconn: on my way back home
02:17:07gregjsays: undefined instruction at 0x00057fac
02:17:08 Quit webguest78 ("CGI:IRC (Ping timeout)")
02:17:14 Part pixelma
02:17:20amiconnThat's something completely different
02:17:28gregjperhaps :)
02:17:33gregjmaybe that's battery issue
02:17:44gregjare you aware of the fact, that apple remote doesn't work with ipod ?
02:18:04gregjI have this little thing that you plug into ipod - that uses extra 4 connectors next to jack
02:18:10 Quit Xerion (" ")
02:18:10gregjheadphones jack
02:18:17gregjthey don't work with rock box
02:18:17 Join JdGordon [0] (
02:18:34JdGordonmorning all
02:18:39DerPapstsure they don't
02:18:50DerPapsthey JdGordon :)
02:18:51krazykitgood evening, JdGordon
02:18:54gregji need/want them :)
02:19:05DerPapstgregj: write a driver ;)
02:19:21gregjso those are not working the same way other keys do ?
02:19:25gregjhow do I approach that
02:19:35gregjI mean, finding right place to look for the key events
02:20:01amiconnThe remote uses serial protocol and the serial interface of the pp chip
02:20:06DerPapstit is connected to the serial afaik
02:20:36gregjand the serial is not implemented yet
02:20:43DerPapstso a serail driver would be neat
02:20:44gregjthat's the SPI one ?
02:20:58DerPapstthat could be used for debugging with gdb too i guess
02:20:58gregjwhat chip it is, and what's it called like
02:21:06amiconnJust rs232 serial (probably at ttl level)
02:21:26amiconngregj: It is built into the PP5020
02:21:36amiconnAnd again the no docs problem
02:23:11gregjsounds like cool stuff...
02:23:15gregjI hate where there is no docs
02:23:20DerPapstgregj: it is ttl
02:23:43amiconnJust hating it doesn't help... reverse engineering does
02:26:46gregjso it works for them
02:26:56gregjso they have som eworking implementation, and stuff, right ?
02:27:21JdGordonNico_P: still up?
02:27:51Nico_PJdGordon: yup
02:27:57JdGordonhowd you go?
02:28:08Nico_Ptrying to fix my bugs, I have a few changes coming
02:28:29 Quit DerPapst ("So Long And Thanks For All The Fish!")
02:28:44 Join DerPapst [0] (
02:29:02DerPapstgregj: the serial ony works on 1G-3G iPods.
02:29:10DerPapst*in iPodLinux
02:31:03saratogadoes it even work for them?
02:31:05JdGordonNico_P: haha, i love the graph :)
02:31:31gregjDerPapst: so thie one they have there, is completly different in g4 ?
02:31:43Nico_PJdGordon: I was trying to debug and got a bit lost
02:31:51gregjsounds like fun..
02:32:13gregjamiconn: any ideas what registers it sits in ? does that PP arm have interrupts ?
02:32:23gregjamiconn: I have no idea how that arm works
02:32:40DerPapstgregj: dunno how different it is to the one that is used in pp5002 chips but it is at least different enough on all PP502X
02:33:09gregjI can write simple program that would output what if finds in registers, and I can start guessing game from that
02:33:28gregjbut at least some register/ i/o range that's usefull, would be great for starters
02:34:10amiconnPP registers are scattered all over the 32 bit address space, and there's a lot of aliases as well (undecoded address bits)
02:34:30Nico_PJdGordon: looks like my last bug with long files is now in the main() code
02:34:35Nico_PI'll commit so you can see
02:36:04Nico_Pon a long file it just loops, trying to buffer a handle that doesn't exist
02:37:31JdGordonI changed the buffer size to 5mb so i dont have to look for massive files :p
02:38:39JdGordonit probably wont make a difference, but free_buffer() should go in the bufferning part, not playback
02:38:42gregjthere's some vague information on linux-on-ipod site
02:38:50gregjamiconn: can you help me a bit here, please ?
02:39:14JdGordonNico_P: probably inside need_rebuffer()
02:39:21gregjamiconn: judging from that, there's ser0 and ser1 interrupt for serials
02:39:53gregjamiconn: how would you approach that (besides trying to find the right pot in OF)
02:40:11Nico_PJdGordon: I didn't really know where to put it and the main() isn't going to stay unchanged anyway
02:41:08JdGordonalso, I think there is a huge probalem with the do loop in playback.. ive just changed it so it will only read 5 times before going back to buffering to simulate threads
02:41:42*gregj builds the chain at home, and stuff
02:41:43saratogahaha all 10 pairs of ffmpeg trig look up tables used by their wma decoder actually contain the same info
02:41:56saratogaI LOVE working with their code
02:42:13 Join toffe82 [0] (n=chatzill@
02:42:28Nico_Psaratoga: are you submitting patches to them ?
02:42:39saratogaNico_P: not at the moment
02:42:57saratogai'm not sure they'd accept it anyway, since it would require writing an MDCT function thats specific to wma
02:43:14saratogastuff like this happens to them because they try to make every codec use the exact same transforms
02:43:40saratogai may try and merge it back in later though
02:44:57 Join relaxed [0] (n=relaxed@unaffiliated/relaxed)
02:46:08JdGordonNico_P: i'm close :) it is rebuffering the track now (unless oyu already had that)
02:46:40Nico_Pit was rebuffering but very close to the end it started going nuts
02:46:41 Join piroko [0] (
02:47:07pirokoGrr... 10 minutes of playing an ogg vorbis file on the newest 4g ipod rockbox and a hard lockup... :(
02:47:37amiconnpiroko: Wait for the upcoming build. It should fix the freezing
02:47:49pirokoNo way. I'm super-excited now
02:48:08gregjpiroko: yeah, amiconn fixed it - looks like so
02:48:27gregjI had the same issue
02:48:38gregjso, go for beer, or try yesterdays build meanwhile
02:48:40pirokoI noticed there was a fix made to the ipod cpu code, I assume this isn't "THE fix"?
02:49:12amiconnThis was the first part, but not everything necessary to fix the freezes on all ipods
02:50:14gregjpiroko: this fixed it for nanos, not for g4
02:52:48saratogaif I have a look up table that has cos() from -90 to +90 degrees, is it worth cutting the table in half and adding a branch to exploit the odd symmetry of the function?
02:52:57saratogaor is the branch not worth it
02:53:28gregjdo it on a side, test, and stick in trunk
02:53:43gregjamiconn: can you point me to that patch again please ? unless you already commited it
02:53:43JdGordonNico_P: I tihnk the move_handle is where the problem is
02:54:23JdGordonah no.... its because m in main() isnt rechecked when the handle is moved
02:54:29Nico_PJdGordon: you sure ? it seemed ok to me
02:55:12Nico_Pthe handle moving is the reason why I disabled your cache btw
02:55:27Nico_Pit'd be good to find a way to reenable it
02:56:06gregjamiconn: I guess you did
02:56:30 Quit DerPapst (Read error: 110 (Connection timed out))
02:57:12JdGordonwoot :D fixed
02:57:27Nico_Pyou rock :)
02:57:47JdGordonmd5 failed though
02:58:07Nico_Pyes I expected that from what I had seen in the files
02:58:21Nico_Pbut I think we're not too far from a good file
02:58:56JdGordonthe read file is 5mb smaller... which is the size of the bufffer...
02:59:21Nico_Pmaybe there's one cycle missing
03:00:51JdGordonmove_handle only affects the first handle doesnt it?
03:01:03Nico_Perr no
03:01:11Nico_Pit affects the handle you give it
03:02:00JdGordonok, but its only used on the first handle... so the cache could be safe to add back
03:02:07JdGordonwith minor adjustments
03:02:22JdGordonyou want me to commit the almost working version?
03:02:29Nico_Pyes plsease :)
03:03:07Nico_PI think we'd need to test the cache to make sure it doesn't cause problems even when moving arbitrary handles
03:03:27Nico_Pyou never know what we'll want to do in the future ;)
03:03:49***Saving seen data "./dancer.seen"
03:04:30*JdGordon has an idea
03:04:36*gregj has two
03:05:10 Join disembowelment [0] (
03:05:20 Join kkurbjun [0] (
03:05:59 Quit disembowelment (Client Quit)
03:06:06kkurbjunamiconn, great job on the ipod mini fixes, the first patch helped out, I havn't tried the second yet
03:06:42amiconnYes, on mini the first one is almost sufficient
03:06:45*JdGordon doesnt want to add more global variables :'(
03:07:10amiconnStill got occasional freezes though
03:07:17kkurbjunamiconn, you were saying that you were looking into power consumption also?
03:07:36pirokoThis last commit that just got entered isn't for the 4g ipod then?
03:07:39amiconnI though these regs might be related, but obviously they aren't
03:08:19amiconnI mean, controlling both power and clock
03:08:28kkurbjunahh, I see
03:08:48kkurbjuncan the pp chips control their voltage rails?
03:08:49amiconnNeed to do some runtime tests (or perhaps add some wires to my mini)
03:08:59amiconnThat's a good question
03:09:14amiconnpiroko: It is, but not only
03:10:22 Quit Rincewind ("Cya")
03:10:25kkurbjunamiconn, to answer your question on the gameboy, it does play sound samples on it's 3rd channel at 44.1 khz
03:10:54pirokoamiconn: Excellent. I'll try it out now :)
03:10:57gregjamiconn: any chance, we can chat about that serial port ?
03:11:05Nico_PJdGordon: after removing the 5-limit thing which was causing an insanely long output on my test file, I get it out right :)
03:11:07kkurbjunbut my next patch will play at 11khz on all devices that support it except for the gigabeat
03:11:29JdGordonNico_P: cool :)
03:11:38Nico_Pit's 69.6 MB and it gets out with the right md5 with a 32 MB buffer
03:12:16JdGordonseems my caching is fixed also
03:12:21JdGordoncommit so i can make sure
03:12:24*amiconn wouldn't have expected an 8bit game device to output cd quality samples
03:12:35gregjyeah, right :)
03:13:20gregjatari65XE had a decent (for 8bit cpu) sound card, with more than 8bits sampling resolution
03:13:22kkurbjunthe gameboy advance can output up to ~139 khz on one channel I believe
03:13:23gregjit was either 10 or 12
03:14:07JdGordonNico_P: woot, yeah caching is back
03:14:18kkurbjunsorry, 131khz
03:14:39Soapamiconn, do your recent commits affect the 5021 targets at all, or just the 5020 and 5022?
03:14:58Nico_PJdGordon: cool :)
03:17:02JdGordonNico_P: check the svn log :)
03:17:23amiconnSoap: There is no 5021, at least not as seen by the code
03:18:06Soapthat is what I was pretty sure of, but not 100%
03:19:32Nico_PJdGordon: perfect :)
03:20:32JdGordonthis whole thing is going to get messy when we try adding the MoB... :'(
03:21:34 Quit GodEater (Read error: 110 (Connection timed out))
03:21:40Nico_Pyou mean having the metadata stay in the buffer while rebuffering a big file ?
03:22:25Nico_Pit might not be too messy
03:22:34JdGordonoh it will be :D
03:22:56Nico_Phopefully the free_buffer() function is generic enough to be helpful in most cases
03:24:28JdGordonyou mean move_handle() dont you?
03:24:40Nico_PI mean both
03:24:56JdGordonfree_buffer just moves the first handle
03:25:46Nico_Pyes but it can be extended to any handle and to leave some margin and so on
03:25:58Nico_Pwhat I mean is that at least we know it works well :)
03:26:38 Join z35 [0] (
03:27:55JdGordonah yes :)
03:28:10*piroko is testing out the latest ipod fix :-D
03:28:56pirokoSo should I expect no freezes in playback or just sparse ones?
03:29:36 Quit z35 (Remote closed the connection)
03:30:06 Join z35 [0] (
03:34:50 Quit jhMikeS (Nick collision from services.)
03:34:56 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
03:37:28Nico_PJdGordon: I actually love the graphs. I think they help understanding the process a lot
03:38:07kkurbjunmake zip
03:39:06Nico_Pand don't know if it's worth mentioning, but the buf_idx field of the struct is very much the same as the last_read pointer you had in your first implementation
03:39:29Nico_Pexcept there is one per handle, which may prove useful
03:40:21 Quit XavierGr (Read error: 110 (Connection timed out))
03:40:23JdGordoni got that :)
03:41:17JdGordoncan the buf_widx be moved relativly safely if we know its moving to a empt area?
03:42:06Nico_Pyeah, I think buf_widx is pretty safe to move forward as long as it doesn't get past buf_ridx
03:43:17JdGordonso we just need to figure out a way to link 2 items with the same handle, and to jump past some unfreed handles and we are done :D
03:45:38Nico_Pwhat happens if you don't bufclose a file before buffering the large one ?
03:46:08Nico_Pif you don't read it even
03:46:38JdGordonbad stuff probably
03:46:48JdGordonor nothing...
03:47:36Nico_Pmore like bad stuff it seems...
03:47:45JdGordonnothing... it loops indefinatly while the buffer is full
03:48:34Nico_Pbut with more intelligent freeing it could be better
03:48:57JdGordonits going to need 2 buf_ridx's
03:49:09JdGordonone for the start of the frozen bit, and one for the end
03:49:15JdGordonor something similar
03:49:25Nico_Phmm yeah you're right
03:50:31Nico_Pwell that would be a bit like the ring-buffer-in-the-buffer idea, wouldn't it ?
03:51:17Nico_Panyway I think I need to go to bed now... can't think clearly anymore
03:51:25JdGordonok, gnite
03:51:42Nico_Pthanks :) bye
03:51:54 Quit Nico_P (Remote closed the connection)
03:53:24 Quit ansivirus ("Leaving")
03:54:22 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
03:55:37pirokoGod rockbox kicks so much ass
03:56:13krazykithell yeah it does.
03:59:10pirokoWhat is the most well-supported device with rockbox? My iPod's hard drive is dying and I don't know how much longer it has...
03:59:39krazykitdepends on what features you want
04:00:02krazykitsome targets have input and recording
04:00:17krazykitothers have ludicrous battery life
04:00:19pirokoGreat battery life and stability with just playing music are my two biggest concerns
04:00:27krazykitdefine "Great"
04:00:43krazykitif 13 hours is good enough, the gigabeat should do fine. you can get them new for $140 or so
04:01:07pirokoThat's not bad.
04:01:52toffe82krazykit: what was your comment yesterday : all the gigabeat flash stuff is windows apps
04:02:24krazykittoffe82, yeah, they all seem to be .exe, the bmp reader etc
04:02:33pirokoOh that could be a problem
04:02:45saratogaif 13 hours is great, then you might want to look at the Sansa
04:02:52krazykitpiroko, no, no, this is hacking stuff. nothing ot do with installation
04:02:53saratogaits not well supported, but it gets at least that much
04:03:02pirokoAh ok
04:03:10toffe82krazykit: you talk about the gigabeat S
04:03:17krazykittoffe82, nah, the gigabeat f
04:03:19pirokoAre the ipods considered "well supported"?
04:03:35krazykitpiroko, not in comparison with some other targets.
04:03:47toffe82once you have rockbox on the gigabeat , you don't need anything else
04:04:12krazykittoffe82, i know, but stuff like holding down the button for so many seconds to turn it on is annoying
04:05:03toffe82I think it is a good thing because on the S there is no delay and you switch it to easley when you have it in your pocket
04:05:15saratogayes the slow turn on time for the F is maddening
04:05:18krazykithold switch :P
04:05:25saratogaparticularly when developing when you have to switch it on and off constantly
04:06:18krazykiti was hoping some work on it was going to be done at devcon west
04:06:39saratogai don't think we have access to the button timing for poweron
04:07:04krazykitlemme find the post, saratoga. maybe i'm misunderstanding
04:08:18krazykititem 4
04:11:39 Join toffe82_ [0] (n=chatzill@
04:12:49saratogakrazykit: thanks, hadn't seen that work
04:15:02toffe82_what are 2 second when you are going to listen an hour or 2 of music ? ;)
04:15:35krazykitminor irritations!
04:16:46saratogatoffe82_: eh, I've rebooted the gigabeat about 100 times today
04:16:50saratogaits starting to add up
04:17:27toffe82_if you dev I understand but for a user it is not so important
04:17:49 Join Febs [0] (
04:18:04saratoga400% realtime WMA decoding
04:18:15saratogaand I got the codec table size down 4KB
04:18:28 Join webguest56 [0] (i=489a6eef@gateway/web/cgi-irc/
04:20:10saratogahmm I made one of my nonconstant tables ICONST_ATTR, yet it still worked
04:20:17saratogai thought ICONST_ATTR was only for constants?
04:24:00 Quit webguest56 ("CGI:IRC (Ping timeout)")
04:24:29 Quit toffe82 (Read error: 110 (Connection timed out))
04:25:07saratogahow much IRAM does the gigabeat have anyway?
04:25:26kkurbjunsaratoga, I think it's only 4K
04:26:02saratogakkubjun: pretty sure its a lot more then that
04:26:12saratogai think i've already gone over that
04:26:41kkurbjunthe iram qualifiers don't do anythign on the gigabeat
04:26:50kkurbjunlike ICONST and IDATA and all that
04:27:05saratogai wonder why my code got faster
04:27:20kkurbjunyou can check it out in the linker script
04:28:22kkurbjunI'm pretty sure it's only 4k and that's why it's not used
04:28:36kkurbjunthe gigabeat has code and data cache also
04:29:38kkurbjunif you want to use the iram I think it's at address 0x4000000
04:30:23saratogayou're right
04:30:42saratogathe changes I made to move things into IRAM must have given a speed up then
04:31:53 Quit piroko (Read error: 110 (Connection timed out))
04:34:30 Join piroko [0] (
04:35:32pirokoParents are fighting... :'(
04:35:35 Join Tazer_gnome [0] (
04:36:43 Quit sneakums ("There shouldn't be this reign of silence, but what are the options when someone great is gone?")
04:37:21Tazer_gnomeI need some support on my ipod video 5.5 with rockbox.
04:37:41saratogaok. decoded a 128k wma file in real time on the Sansa
04:38:14kkurbjunsaratoga, is the wma decoder a codec now?
04:38:25saratogakkurbjun: yes
04:38:35krazykitwe didn't have a wma decoder before?
04:38:49saratogawe still don't, outside of me and linuxstb I think
04:39:04kkurbjunthats great, congratulations
04:39:21kkurbjunalot of people will be excited when the work is out I'm sure
04:39:44scorchebut will they have a clue? =P
04:39:45krazykithow long've you been working on it?
04:39:45pirokoPeople still use wma?
04:39:50 Join sneakums [0] (
04:40:13kkurbjun:), I don't, but I've come across a number of people that do
04:40:29kkurbjunmainly from allowing windows media player to do their ripping
04:41:17 Quit Tazer_gnome ()
04:41:58saratogahow do I make a patch verses the current SVN version?
04:42:16saratogatyping svn diff . > patch only gives me a patch relative to linuxstb's patch
04:42:24kkurbjunsvn diff > mydiff.diff
04:42:37saratogayeah that doesn't do what i need
04:42:52saratogait just gives me a patch that works on the patched version linuxstb gave me i think
04:43:17saratogai'd like to make a patch that includes his patch, and all the improvements I made to get PP targets running
04:43:43saratogahow do I tell SVN that i want the patch relative to the curernt rockbox svn
04:43:47kkurbjunIt should give all the differences from main svn
04:43:50kkurbjunthat's what I use
04:44:35saratogait gives me all differences relative to my SVN the last time it was patched
04:44:42kkurbjunI don't think it includes new files though..
04:44:51saratogawhich is not what I need
04:49:37*piroko is in ipod-stability heaven
04:51:33saratogaso no one here knows how to make a diff that incorporates a second patch?
04:54:37krazykitunfortunately, i suck at patch
04:56:24saratogalinuxstb: heres my updated version of wmadeci.c
04:56:29saratogapreglow: you too
04:56:36bagawksaratoga: you can just append it
04:56:52saratogait decodes wma in real time on my sansa
04:56:56saratoganow i'm going to bed
04:57:02 Quit saratoga ("CGI:IRC")
05:03:53***Saving seen data "./dancer.seen"
05:15:51pirokoAnd the room goes silent...
05:16:45*scorche relaxes in the silence
05:18:36 Join lextune [0] (
05:18:45 Part lextune
05:20:35 Quit EnterUserName (Read error: 104 (Connection reset by peer))
05:21:13 Join rhet [0] (
05:22:00 Quit piroko ("telinit 0")
05:23:32 Join piro [0] (
05:25:22piroI am seriously amazed at the ipod playing so perfectly :)
05:25:31rhethi, im having some issues with installing themes on my ipod..i copied over the contents of the theme folders into the .rockbox folder and the themes show up, but they don't look like they should and im not really sure what i did wrong..
05:25:55piroamiconn: Thank you so much for your hard work!
05:25:57 Join mickleby [0] (i=456b3a5c@gateway/web/cgi-irc/
05:26:14scorcherhet: make sure you have the fonts pack installed and that the build doesnt require patches
05:26:21 Nick piro is now known as piroko (
05:26:26scorcheother than that, it could be a broken theme
05:26:48krazykitrockbox is constantly worked on and themes break. some authors don't keep them up to date
05:26:57krazykitthose slackers.
05:27:07micklebygreetings assembled hordes of Ghengis Khan
05:27:23micklebyps: my theme was broke just the other day
05:27:37scorchemickleby: we are only like that if you mention that malloc should be in the core
05:28:10krazykitcore malloc! core malloc!
05:28:11rhetalright, i'll look into that scorche thanks.
05:28:40scorcheGodEater_: shall we go get the pitchforks?
05:28:57krazykitwait, you have pitchforks? shit, i take it back
05:46:13 Quit Soap (Remote closed the connection)
05:52:56 Join bootloader [0] (i=474872e4@gateway/web/cgi-irc/
05:55:22bootloaderumm is there anyone ther who can help me? I installed rockbox but want to uninstall it and the ipod patcher cant find the ipod
05:57:25pirokoAre you using linux or windows?
05:58:11pirokoSo iTunes isn't picking it up either?
05:58:31bootloaderdo you want to know exaclty all i have done?
05:59:03bootloaderor can we try to enter disk mode? How do you enter disk mode?
05:59:28pirokoTry this. Reboot your ipod into disk mode by holding down select and menu for about five seconds, then immediately hold down select and play until the disk mode starts up
06:00:02pirokoThen connect your ipod to your computer and see if iTunes picks it up then
06:00:09bootloader...ok one sec
06:00:15pirokoAlright :)
06:00:36bootloaderhow do i know if i started disk mode?/
06:01:00pirokoIt should pop up with the big circle with the slash through it
06:01:12pirokoOr a checkmark
06:01:17piroko(If it's not plugged into the computer)
06:01:22bootloaderok good one sec
06:01:42 Join aliask [0] (
06:01:46rhetis it possible to install multiple themes and then switch on the player?
06:01:58aliaskrhet: Yes
06:02:45pirokobootloader: Congrats! I'd just go with a restore from iTunes to restore yourself to iPod happiness
06:03:00bootloaderok so just restore ipod and all is well? then can i ask you how to install WPS properlly, because that is where all this went wrong
06:03:00 Quit kkurbjun (Read error: 104 (Connection reset by peer))
06:03:00rhethow would one go about doing that aliask? the only way i've seen to install themes is to replace the folders directly on the drive
06:03:30krazykitrhet, just extract the zip to the root of the player
06:03:39krazykitshould work just fine
06:03:45pirokobootloader: All should be well with the restore
06:03:58rhethm ok
06:04:27pirokobootloader: You'll have to re-install rockbox though if you still want it on your ipod
06:04:59 Part toffe82_
06:05:13krazykitdamn rockbox has spoiled me.
06:05:24krazykiti won't buy a player that's unsupported now
06:05:34pirokokrazykit: Nor will I!
06:05:53bootloaderyes i know, what happened was i installed rockbox it is awsome, i tried to install the Evil build of WPS then i tried to install a theme to the .rockbox/wps and that was bad i guess
06:06:12krazykitbootloader, the theme goes in the root of the drive.
06:06:45krazykitthe zip should be laid out like zip/.rockbox/{wps,backdrops,etc}
06:06:45bootloaderyes I figured that out when nothing was working :) i knew i needed to remove everything but i didnt know how
06:07:12krazykityou could've removed .rockbox then just extracted a new one in its place
06:08:08bootloaderso do i just install everything to the root then select my theme in the settings option on rockbox?
06:08:21pirokoShould do the trick ;)
06:08:47bootloaderok.... same for the fonts package? I dont know if i need it hough
06:09:02krazykityou want it
06:09:24krazykitmany themes expect it to be there
06:09:40bootloaderok so just download and extract to root?
06:09:56pirokoYep :)
06:10:18bootloaderok thanks for all the help i think i can do it now Thanks again
06:10:34pirokoNo problem. Ask if you get any further difficulties
06:10:47krazykittime for bed
06:11:05pirokoG'Night krazykit!
06:16:51pirokoIt's time for bed for me as well. Work in the morning :P
06:17:02pirokoG'Night all
06:17:32 Quit Nibbier (Read error: 110 (Connection timed out))
06:17:39 Quit piroko ("Zzzz")
06:18:12 Join Nibbier [0] (
06:19:54bootloaderwhen i try to install the Evil build WPS it asks if i want to overwrite some files do i say yes to all or no or what?
06:21:18bootloaderis anyone there anymore?
06:21:56scorcheyes to all
06:22:12 Join chrisjs169|bbl [0] (
06:22:33bootloaderok thank you very much
06:23:17 Part chrisjs169|bbl ("Konversation terminated!")
06:23:59 Join Bossier [0] (
06:25:52Bossierquick question: I just noticed some new commits dealing with cpu clocks by Jens Arnold, would I need to compile a bootloader to fully appreciate these improvements? that did look like something that would affect the bootloader to me, could be wrong..
06:27:12 Nick Bossier is now known as tdtooke (
06:27:48 Join jhulst_ [0] (
06:27:55 Quit jhulst (Read error: 104 (Connection reset by peer))
06:28:42tdtookeok, thanks, i'm really eager to try the latest out to see if the battery life might be improved
06:30:18 Nick tdtooke is now known as Bossier (
06:30:18 Quit Bossier ("( :: NoNameScript 4.02 :: )")
06:31:22bootloaderafter installing WPS the select button is really hard to get to respond...
06:31:56scorchea WPs has no affect on the function of rockbox..
06:32:46bootloaderwell that is not cool then
06:33:03 Join ptw419 [0] (
06:33:29 Quit rotator ()
06:33:42bootloaderbut i cant find my them i have downloaded
06:34:41bootloaderdo i extract themes to the root?
06:35:07bootloaderwhere do i select them from
06:35:27scorchewhat do you mean?
06:36:04bootloaderafter i extract the theme and i start rockbox back up the theme is not running
06:37:22bootloaderhow do i start the theme?
06:42:13 Quit jhulst_ (Read error: 104 (Connection reset by peer))
06:42:25 Join jhulst_ [0] (
06:43:58 Quit bootloader ("CGI:IRC (EOF)")
06:44:02 Join ansivirus [0] (
06:57:51 Join Pajama-B [0] (
07:03:37 Quit ptw419 ()
07:03:58***Saving seen data "./dancer.seen"
07:11:40alienbiker99does anybody here know how i to pop out a 5gen ipod screen?
07:21:11 Quit Pajama-B ()
07:23:16lazkasomething like this?
07:24:53 Quit jhMikeS (Nick collision from services.)
07:24:59 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
07:25:30 Join printfXh4 [0] (
07:29:03alienbiker99thanks, i just found a guide to do
07:30:18 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
07:36:07 Join chrisjs169 [0] (n=jack@unaffiliated/chrisjs169)
07:36:35chrisjs169jhMikeS: Another report of issues with the MicroSD patch on ABi
07:38:40 Quit rhet ("Leaving")
07:40:03 Quit billytwowilly (Remote closed the connection)
07:40:18 Join billytwowilly [0] (
07:41:40 Quit printfXh4 ("Leaving")
07:45:13 Quit Rob222241 ()
07:45:30 Join Rob2222 [0] (
07:47:48 Join My_Sic [0] (
07:50:35 Quit My_Sic (Client Quit)
07:52:45chrisjs169jhMikeS: I'm heading off for the night - will be on tomorrow
07:52:56chrisjs169err...later I guess
07:54:30 Quit scorche (Excess Flood)
07:54:38 Join scorche [0] (i=HydraIRC@rockbox/administrator/scorche)
07:56:36 Quit scorche (Excess Flood)
07:56:46 Join scorche [0] (i=HydraIRC@rockbox/administrator/scorche)
08:10:04 Join GodEater [0] (
08:13:51 Join bdgraue [0] (
08:16:35 Join Rob222241 [0] (
08:27:02 Join ender` [0] (
08:29:56 Join Siltaar [0] (
08:32:01 Join Bspec [0] (n=botxj@unaffiliated/bspec)
08:34:10 Quit Rob2222 (Read error: 110 (Connection timed out))
08:35:34Mode"#rockbox +o scorche " by ChanServ (ChanServ@services.)
08:36:21Mode"#rockbox +J 90,3 " by scorche (i=HydraIRC@rockbox/administrator/scorche)
08:36:34scorchebotnet attack going on...want to be safe
08:36:52*Bspec kisses scorche for using his settings :)
08:37:39*scorche shrugs at Bspec
08:39:04 Join ender1 [0] (
08:41:23GodEaterwhat the heck is +J ?
08:42:01GodEaterooh - nifty
08:42:19GodEaterhyperion is a fancy bit of kit and no mistake
08:42:29scorcheit has its issues
08:43:23GodEaterwell, all ircds to some extent
08:43:27GodEaterthere isn't a perfect one
08:46:17Bspec"fancy bit of kit" <−− so effin' cute :D
08:46:25Bspeci'm save that to my list of charming phrases to say
08:47:05GodEater"cute" ?
08:47:31 Quit ender` (Read error: 110 (Connection timed out))
08:47:49scorcheGodEater: you missed my call for pitchforks earlier
08:48:01GodEaterscorche: sorry :(
08:48:07GodEaterwhat were they in aid of ?
08:48:47GodEateris the guilty party still present ?
08:48:53*GodEater sharpens his scythe
08:50:40 Join kretender [0] (
08:59:21 Quit DataGhost ("NTOSKRNL.EXE caused a buffer overflow in System Idle Proce")
09:03:32 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
09:04:00***Saving seen data "./dancer.seen"
09:05:24 Join DataGhost [0] (
09:05:44 Join pondlife [0] (
09:06:24 Join Wiwie [0] (
09:06:31 Quit Wiwie (Read error: 104 (Connection reset by peer))
09:09:48JdGordonhey pondlife
09:10:26pondlifehi JdGordon
09:10:36 Join petur [0] (n=petur@rockbox/developer/petur)
09:11:32pondlifeJdGordon: r27 already??
09:11:52Mode"#rockbox -J " by scorche (i=HydraIRC@rockbox/administrator/scorche)
09:12:01Mode"#rockbox -o scorche " by ChanServ (ChanServ@services.)
09:12:06JdGordonyeah we kick ass :)
09:12:14Bspecya know, scorche you could use higher settings for +J
09:12:15JdGordonlong files are working
09:12:18pondlifeAnd big files now work ok?
09:12:28*pondlife isn't psychic
09:12:44Bspecfreenode recommends you tweak +J to suit your channel's normal join rate
09:12:58scorcheBspec: i know, but we dont need it
09:12:59Bspecand keep +J
09:13:06Bspecblarg, k
09:13:34pondlifeJdGordon: On another topic, I'm after a Gigabeat... from where I'm standing they seem particularly available in Australia, is this true?
09:13:49JdGordoni have seen them in shops
09:13:54pondlifeX60 ideally
09:14:32pondlife(Assuming that gives me Rockbox + ZIF HD connector and room for an upgrade.)
09:15:09JdGordonseems the X is on clearance stock in the major shops...
09:15:17JdGordonX30 for 280 AUD
09:15:29 Quit mickleby ("CGI:IRC (Ping timeout)")
09:15:43pondlife120 + P+P
09:16:54pondlifeAnyone know if the X30 is more restricted in physical HD thickness? I want to go to 80GB or more at some point.
09:18:09JdGordonis the X the one that was for AU and japan?
09:18:46GodEaterI think the X is easier to find in those locations certainly
09:18:52GodEaterI'm not sure it was exclusively for them
09:19:05pondlifeIt *is* capable of running Rockbox, right?
09:19:22petur30GB will be single platter I think
09:19:24JdGordonlooks like ebay is the way to go for one
09:19:32 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
09:19:38pondlifeJdGordon: Yep, that's where I've been looking mainly.
09:19:50 Join miepchen^schlaf [0] (
09:19:51peturso for 80GB+ you need double-platter
09:20:02pondlifeX60 is double-platter
09:20:12peturmust be
09:20:20pondlifeBut I'm always asleep when AUS ebay auctions end :(
09:22:52pondlifeWooh, WMA decoding got realtime on Sansa??
09:23:17JdGordonapparently :)
09:23:27JdGordondoes anyone actually have any wma's though ?p
09:23:47pondlifeI have a few, sadly
09:24:03pondlifeNon-DRMed, happily
09:27:07GodEaterpondlife: I thought people with a clue didn't use WMA ? ;)
09:27:32pondlifeTrue, but I also thought that people with a clue didn't transcode lossy to lossy
09:27:46GodEaterwho did that ?
09:27:58pondlifePeople without a clue, I guess :)
09:28:17pondlifeMy WMAs are downloaded radio programs.
09:28:32pondlifeUnavailable in MP3 or OGG
09:28:44GodEaterah - what a pain
09:28:55JdGordonI hate to say it, but real audio would be better than wma imho... my lectures are in .ra not .wma :p
09:29:28GodEaterget coding then ;)
09:29:43pondlifeIsn't RA an AAC variant?
09:29:50*scorche shakes his head at JdGordon
09:29:55*JdGordon doesnt know
09:30:23*pondlife was put off RA for a lifetime some time in 2002
09:30:34scorchepondlife: 9 and 10 are i believe
09:31:02pondlifeDontcha just love formats with versions...
09:32:53scorchelook at that collection...
09:33:38pondlifeHmm, what the world needs now is more container formats :)
09:33:48peturnow I know why the RA player was so bloated...
09:34:36 Join inversions [0] (
09:34:38scorchepetur: well, not to mention that they have a full web browser in it as well..
09:38:31 Quit aliask ("ChatZilla [Firefox]")
09:42:14 Join spiorf [0] (
09:43:35 Quit ender1 (Read error: 110 (Connection timed out))
09:44:58pondlifeJdGordon: you're using a linked list in the buffering code?
09:45:17JdGordonnico's idea, works nicely
09:45:28pondlifeaustriancoder also wants to use one for USB
09:45:38pondlifeMight be good to share stuff.
09:46:16pondlifeHe may well already have one by now, and we don't want 2 of them implemented in the core.
09:47:03JdGordonwell, there isnt really that much to share..
09:47:30JdGordonapart from the search and remove functions.. which wouldnt share well anyway if the structs are different
09:48:26pondlifeAh, ok... just thinking a generic LL code might be sharable.
09:50:36 Quit Weiss (Read error: 104 (Connection reset by peer))
09:53:12 Join Weiss [0] (
09:54:23peturhe already has one
09:55:24peturit should be in here: (some patches don't apply cleanly)
09:55:49petursearch for firmware/include/linkedlist.h
09:56:59JdGordonhe's gone doubly linked
09:58:53 Quit billytwowilly (Remote closed the connection)
09:59:09 Join billytwowilly [0] (
10:00:29 Join Jon-Kha [0] (
10:02:12markunwith Git it wouldn't matter that the a tree someone's working on isn't up to date if you want to try it out yourself, right?
10:05:01 Join linuxstb_ [0] (i=5343d4aa@rockbox/developer/linuxstb)
10:06:46 Join joshin_ [0] (n=joshin@VDSL-130-13-252-46.PHNX.QWEST.NET)
10:06:49linuxstb_Morning all. I want to try and commit the WMA codec soon, but there are still 3 warnings in wmadeci.c that I can't resolve (wrong pointer types). Can anyone help? (latest patch is on FS #4984)
10:07:18markunlinuxstb_: Don't know if I will be of much help but I will take a look
10:07:19 Quit joshin (Read error: 104 (Connection reset by peer))
10:09:20linuxstb_Also, anyone know if saratoga has SVN access? I don't think he does, but have lost track...
10:10:09markunlinuxstb_: hm, I got patch: **** malformed patch at line 3869
10:10:18markunPatching file wmadeci.c using Plan A...
10:11:18linuxstb_Hmm, the patch applies cleanly for me...
10:11:40LinusNfor me too
10:11:43markunmaybe the line is too long for FreeBSD's patch
10:11:43peturlinuxstb: I take it you saw 04.56.25 # <saratoga>
10:12:07linuxstb_petur: Yes, my patch includes that version (with a few more cosmetic tweaks)
10:12:33LinusNlots of commented code with c++ comments
10:12:48linuxstb_Indeed - it needs a cleanup.
10:12:49markunwow, that line is 38941 bytes long
10:13:02linuxstb_And it's commented out?
10:13:11peturheh, a bit over the 80 char limit :)
10:13:19markunlinuxstb_: yes
10:13:26linuxstb_If you're going to break a rule, do it in style...
10:14:11markunlinuxstb_: ok, applied now (with that line removed)
10:15:27*linuxstb_ is surprised - only 3KB binary size increase on ARM for the WMA patch...
10:16:04peturerrr a codec is a separate binary, no?
10:16:10markunthat's the metadata parser?
10:16:23peturah of course
10:16:52linuxstb_markun: Yes.
10:16:54JdGordonis there any way we can pull the metadata parsers out of the core?
10:17:09LinusNwhat's with the bswap.h file? don't we have that in system.h?
10:17:12linuxstb_I'm not sure the complication is worthwhile.
10:17:18linuxstb_LinusN: Yes, that's on my to-do list.
10:17:28linuxstb_There's still stuff there from ffmpeg that needs removing.
10:18:35JdGordonlinuxstb_: well, if we can get MoB working properly, the id3 struct would be in the audio buffer, so maybe the parsers could be moved to the codecs?
10:18:57***Server message 505: 'logbot_ :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
10:19:19***Server message 505: 'logbot_ :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
10:19:33linuxstb_JdGordon: The metdata parsing has to happen independently to decoding (at the time the file is buffered), so moving the metdata parsers out of the core would mean metadata parser plugins, with their own API and buffer.
10:19:36 Join obo [0] (
10:20:01linuxstb_Although there has been talk of making codec-specific loaders, so that would be combined with the metdata parsing.
10:20:08peturwas there a way to ask logbot when a nick was last seen?
10:20:27JdGordonwhy? they arnt read untill the track starts playing anyway arnt they?
10:20:38scorche/msg logbot_ seen nick
10:20:45pondlifeNext track info?
10:20:54JdGordonand that...
10:21:03linuxstb_Also, the metdata parsers need access to the entire file (tags can be at the end).
10:21:08***Server message 505: 'logbot_ :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
10:21:20scorchepetur: above
10:21:20linuxstb_So they use the file descriptor and read from disk, rather than the codec buffer.
10:21:42peturscorche: I don't get anything back
10:21:54GodEatermarkun: re your git question, the answer is "yes" - you can get and use their branch even if it's not up to date with your own
10:21:59linuxstb_So that obviously needs to happen when the disk is spinning - i.e. at the time the file is buffered.
10:22:04pondlifepetur: It never works for me either ;)
10:22:05JdGordonah ok
10:22:13scorchepetur: because logbot isnt identified, so you need to enable yourself to be able to recieve messages from unidentified people
10:22:25scorcheit might be quicker to just ask me who you want info on
10:22:47JdGordonlogbot_: seen austriancoder
10:22:53scorche-logbot_- austriancoder (n=austrian@ left the channel 3 days and 16 hours ago ("Kopete 0.12.4 :")
10:23:00JdGordonbah, its not smart :p
10:24:19 Quit linuxstb_ ("CGI:IRC")
10:26:53scorcheJdGordon: smart in what way?
10:27:28JdGordonit doesnt respond to channel
10:27:36amiconnpetur: /msg logbot_ seen <nick> does definitely work
10:27:43 Quit perplexity (Read error: 113 (No route to host))
10:27:47peturyes, I just found out
10:27:48JdGordonseen austriancoder
10:27:58amiconnBut since logbot_ isn't registered, you need to allow messages from unregistered users
10:28:01peturI had to /msg nickserv set unfiltered on
10:28:13*amiconn has that always set
10:28:15peturworks now
10:28:16scorcheamiconn: as i just said..
10:29:05Zagormeh, we need to register logbot on nickserv
10:29:22scorcheZagor: aye...i was goign to suggest that to bagder when he woke
10:29:39pondlifelol - "Notice from logbot_ (Freenode): You need glasses badly"
10:29:49amiconnZagor: Yeah, and that cute 'explain' feature still doesn't werk... ;)
10:30:40peturpondlife: how?
10:30:48***Server message 505: 'logbot_ :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
10:31:12pondlifeI asked when pondlife was last seen...
10:31:18amiconnAsking for 'seen' when the user is in the channel
10:31:42pondlifeHow did logbot_ *know* about my terrible eyesight :)
10:32:01peturhaha <logbot_ whispers>Yeah, I see you. Boy, you're ugly ;)
10:33:31 Join desowin [0] (
10:34:46 Quit inversions (Read error: 113 (No route to host))
10:36:37JdGordonis it ok to do a bandaid fix for FS #7387? just check if a card is inserted if trying to get into the <micro..> folder?
10:36:47 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
10:37:04LinusNlinuxstb: those 3 warnings - i assume they used to be malloc() calls?
10:37:34linuxstbYes, saratoga changed them.
10:37:54LinusNyou could cheat and use a (void *) cast
10:38:02linuxstbWMA is now one of the good guys and is malloc-free...
10:38:36JdGordon... objections to ?
10:39:08pixelmaJdGordon: wow... never noticed but it happens with the MMC too (usually I have it plugged in all the time or don't change to the menu)
10:39:29JdGordonok, so ill do it for all hotswap
10:39:57linuxstbLinusN: I'll do that if no-one works out a proper fix.
10:40:07LinusNhang on
10:41:27LinusNVLC_TYPE vlcbuf1[12288][2];
10:41:27LinusNVLC_TYPE vlcbuf2[7168][2];
10:41:27LinusNVLC_TYPE vlcbuf3[1536*2][2] IBSS_ATTR; //small so lets try iram
10:41:41LinusNthere you go
10:42:06linuxstbAh, I was trying the other way around - [2][...]
10:42:12pondlifeHmm, I'm getting errors displayed when using Flyspray - "Notice: Undefined index: 0 in /usr/share/flyspray/htdocs/includes/class.flyspray.php on line 344"... above a fully-working, but slightly misrendered activity.
10:42:34linuxstbVLC_TYPE is 16-bit, so I can divide the sizes by two.
10:42:58LinusNi just did
10:43:32LinusNthe sizes are correct in my pasted code
10:43:35 Join webguest31 [0] (i=c0231115@gateway/web/cgi-irc/
10:44:00LinusNah, no
10:44:08LinusNsilly me
10:44:12LinusNyou are correct
10:44:15 Join Guile`` [0] (
10:44:59JdGordonok... someone give me the OK to commit 7222 please....
10:45:20JdGordonor 6358
10:45:24LinusNVLC_TYPE vlcbuf1[6144][2];
10:45:25LinusNVLC_TYPE vlcbuf2[3584][2];
10:45:25LinusNVLC_TYPE vlcbuf3[1536][2] IBSS_ATTR; //small so lets try iram
10:46:54LinusNJdGordon: 7222 looks ok to me
10:47:11linuxstbLinusN: Thanks.
10:48:14 Quit midkay ("Leaving")
10:48:19peturLinusN: c++ style comment? :p
10:48:23JdGordonLinusN: I actually think 6358 looks better, cept it uses strchr instead..
10:48:35LinusNJdGordon: hang on
10:48:50linuxstbpetur: Blame saratoga ;)
10:50:25pondlifeJdGordon: Which one results in smallest binary size?
10:50:27 Join ompaul [0] (n=ompaul@freenode/staff/gnewsense.ompaul)
10:50:45JdGordonpondlife: dunno, i guess ill check :p
10:51:11LinusNJdGordon: i wonder why 7222 has 48 illegal characters and 6358 only 9
10:52:31JdGordonthe first bit in 6358 is only for folders which possibly has less illegal chars?
10:52:55 Join midkay [0] (n=midkay@rockbox/developer/midkay)
10:52:56JdGordonil merge the two ... I dont like the hex cases
10:53:03pondlifeI don't think the rules are different for files/folders..
10:55:06amiconnThere are actually very few illegal chars
10:55:13amiconnCheck the fat specs
10:55:20pondlife48 sounds too many, surely
10:56:41pondlife\/:*?"<>| makes 9
10:57:34LinusNThe following characters are not legal in any bytes of DIR_Name:
10:57:34LinusN? Values less than 0x20 except for the special case of 0x05 in DIR_Name[0] described above.
10:57:34LinusN? 0x22, 0x2A, 0x2B, 0x2C, 0x2E, 0x2F, 0x3A, 0x3B, 0x3C, 0x3D, 0x3E, 0x3F, 0x5B, 0x5C, 0x5D,
10:57:34DBUGEnqueued KICK LinusN
10:57:34LinusNand 0x7C
10:57:36pondlife? and * might actually be valid under FAT, but not Windows - I'm not certain of that
10:59:02LinusNsince the FAT spec uses hex codes, i think we should too
10:59:35ZagorI'd prefer a comment with the char next to the hex in that case
10:59:47LinusNa good idea
11:00:05 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
11:00:32peturZagor: did you have a look at that patch?
11:00:32 Join miepchen^schlaf [0] (
11:00:46LinusNthe current code disallows 0x20 in the wrong way
11:01:06Zagorpetur: only very quickly. I hope to get some more time today.
11:01:09pondlifeHmm, 0x20 is disabled??
11:01:12LinusN0x20 is only forbidden in the first position
11:01:26LinusNbut the current code forbids it in all positions
11:01:43pondlifeWindows 2000 allows a leading space on FAT32
11:01:53amiconn0x20 is forbidden as leading or trailing char DOS short names
11:02:08peturZagor: since I think the ./.. entries are now set correctly and chkdsk still complains, maybe it checks some other things like creation time?
11:02:09pondlife03/07/2007 10:04 8 test
11:02:09pondlife20/05/2005 20:16 <DIR> Temp
11:02:33LinusNlong names can contain 0x20 anywhere
11:02:46peturiirc, our fat code always generates a longfilename
11:03:14 Join scorche` [0] (n=scorche@rockbox/administrator/scorche)
11:04:02***Saving seen data "./dancer.seen"
11:04:11 Quit scorche (Read error: 104 (Connection reset by peer))
11:04:21 Nick scorche` is now known as scorche (n=scorche@rockbox/administrator/scorche)
11:07:48 Join Xerion [0] (
11:07:54JdGordon is down :'(
11:08:00JdGordonis 0x20 . ?
11:08:28LinusNpetur: from the fat spec: "The DIR_FileSize field on both entries is set to 0, and all
11:08:28LinusNof the date and time fields in both of these entries are set to the same values as they were in the
11:08:28LinusNdirectory entry for the directory that you just created."
11:09:09 Join kfazz [0] (
11:13:52JdGordon ?
11:13:52linuxstbDoes anyone object to the wma codec being committed in more or less its current state? It needs lots of cleaning, but it at least compiles without working, and I think it will be better for saratoga to work in SVN than with patches.
11:14:10LinusNlinuxstb: go! go! go!
11:14:12linuxstb^that should be "compiles without warnings"... ;)
11:14:12pondlifeNo, I say get it in there.
11:14:26pondlifecompiles without working would be ok with me :)
11:14:34JdGordoncompiles without working :) even i can do thaat :p
11:14:46linuxstbIt compiles without working on Coldfire...
11:14:47CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
11:14:47*pondlife resists the urge...
11:15:20LinusNJdGordon: i think the /* space */ comment is misleading
11:15:37JdGordonwoops, so it is.. got carried away
11:16:02LinusNit looks fine apart from that
11:16:13JdGordoncontrol chars?
11:16:51linuxstbOK, I'm just compiling for a few more targets, then I'll commit it.
11:17:11LinusNJdGordon: "control chars" is probably good
11:18:07JdGordonshould len > 255 be MAX_PATH?
11:18:29JdGordontook that from lear's version
11:19:30LinusNJdGordon: no, that is correct
11:20:42JdGordonbbs, dinner
11:21:03peturLinusN: thanks, I'll add that to my fix tonight and see if tha solves everything...
11:21:39*petur gets way too much interrupts from payed work
11:23:32LinusNJdGordon: Long names are limited to 255 characters, not including the trailing NUL. The total path length of a
11:23:33LinusNlong name cannot exceed 260 characters, including the trailing NUL.
11:24:18LinusNbtw, long names don't have the same character limits as short names do
11:25:11LinusNthe + , ; = [ ] characters are allowed in long names but not in short names
11:25:44*linuxstb does the new codec in SVN dance
11:25:56*LinusN joins the happy dance
11:26:55*LinusN also does the "WMA playback stutters" bug report dance
11:27:08 Quit desowin (Read error: 113 (No route to host))
11:27:39petur20% realtime on coldfire :/
11:28:30 Join maffe [0] (
11:28:43 Part maffe
11:29:00pondlifeHow come Coldfire is so much worse than PP??
11:29:15 Join maffe [0] (
11:29:46 Join Entasis [0] (
11:29:48amiconnSpeaking about demanding codecs - I was able to listen to 128kbps aac on H1x0 without skipping, but boost ratio stays at 100% :(
11:30:31 Join austriancoder [0] (n=austrian@
11:30:48peturoh look who's there...
11:31:15peturaustriancoder: how's coding?
11:32:48austriancoderI am trying to get file storage gadget driver compiling...
11:35:07peturI had troubles getting your patches applied and build
11:35:44*amiconn calls for PP5020 testers...
11:36:19*austriancoder has a sansa
11:39:40 Quit Febs (Read error: 110 (Connection timed out))
11:40:36 Join crop [0] (i=c27f0812@gateway/web/cgi-irc/
11:41:23cropJdGordon: ping
11:43:42croplinuxstb: ping too
11:46:01cropJdGordon: hey. Shouldn't the #include hotswap.h be also protected by HAVE_HOTSWAP? Just as a matter of purity?
11:46:53JdGordonmeh, does it really matter?
11:47:10croplinuxstb: am I right in the assumption that the custom WPS page should be updated since we now have one more codec? Or does the wps parser have to be updated first?
11:47:44cropJdGordon: not really, but I always eat with clean hands :-)
11:47:55JdGordoneach to their own I guess....
11:48:27preglowamiconn: that extra stuff is gleaned right out of retailos?
11:48:31preglowlooks weird...
11:49:40JdGordonLinusN: so its good to go?
11:49:47cropJdGordon: :-)
11:49:50LinusNi think so
11:50:06LinusNi assume you have tested it...?
11:50:26JdGordoncourse I have..... im not thaat bad...
11:50:50JdGordonhmm... int..ernet is um... stuffing up.... commit will have to wait a few min...
11:51:32 Quit Siltaar (Remote closed the connection)
11:52:04*JdGordon loves his custom sansa bootloader, boots from the mSD to speed up testing :D
11:52:28linuxstbcrop: Yes, the CustomWPS page needs updating, and so do the WPSs in SVN...
11:53:23*linuxstb curses those 64-bit sims and goes to fix some DEBUGF warnings
11:55:31amiconnpreglow: Yes it is.
11:56:05 Join DerPapst [0] (
11:56:07 Join desowin [0] (
11:56:16croplinuxstb: done (only the former :-)
11:56:40*DerPapst serces wmas but can't find any...
11:56:43pixelmalinuxstb: I'd put updating the SVN wps's that use bitmaps for codecs on ice... - because almost all of them (with the exception of iAmp IIRC) can't take more bitmaps and are already missing some codec bitmaps...
11:56:44JdGordonpixelma: you dont know the # for the rec path patch do you?
11:57:00pixelmaJdGordon: 7201
11:57:01DerPapstah.. windows has some default ones :)
12:00:37 Part maffe
12:00:43 Join maffe [0] (
12:00:44DerPapstpixelma: that would fix the problems with these WPSes FS #7345
12:01:43pixelmaI know that there are patches out there
12:03:19austriancoderdoes anybody know where I can find the latest sansa emulator? seems to be not up 2 date
12:04:19linuxstbpixelma: Yes, I agree about the WPSs.
12:05:05 Join Nico_P [0] (
12:05:53DerPapstpixelma: don't wanted to offend you...
12:06:26Nico_Plinuxstb : congrats to you and saratoga on the WMA codec :)
12:07:13preglowlinuxstb: the ffmpeg code itself is still not synced to their svn, no?
12:08:21 Join lee-qid [0] (
12:08:54*DerPapst is gelling lots of compiling errors :-/
12:09:44linuxstbpreglow: IIUC, saratoga just took marsdaddy's code and continued hacking.
12:09:52linuxstb(so no, it's not synced to ffmpeg svn)
12:10:44pixelmaDerPapst: didn't feel offended.... by what?
12:10:54JdGordonamiconn: the disk cant spinup when the rec screen is entered to check the dir because it will stuff up the recording right?
12:11:11JdGordonshould I then make it a seperate setting to check the rec dir is OK?
12:11:59JdGordonor pixelma...
12:12:04JdGordonor really anyone... :p
12:15:18pixelmajust from a user perspective: I don't want a separate setting - don't know the internals
12:16:16Nico_PJdGordon: just finished reading your comments in buffering.h
12:16:44pixelmaI thought there was an idea how to solve this problem but can't remember atm
12:16:52JdGordonpixelma: without the setting it would have to wait for the disk to stop spining before rec starts...
12:17:53JdGordonNico_P: that was my post-shower brain dump... didnt touch that code all day and it might be useless... but what you think?
12:18:16 Quit ompaul ("Leaving")
12:19:10Nico_Psounds fine to me... there's just the last paragraph I'm not sure about: we probably can get all the info we need about the metadata before reading the audio, can't we ?
12:19:54Nico_Pbitmap and cuesheet will be easy to look for before evening opening the audio file... we just need the name
12:20:04JdGordonyeah, thats what I thought
12:20:28Nico_P"image and buffer will not be known, which is where it might get tricky." -> it's this part I don't understand
12:21:15JdGordonumm... like I said.. braindump :p umm...
12:21:21JdGordoni guess ignore that
12:21:33Nico_Pok :)
12:24:33linuxstbamiconn: Do you need a pp5020 tester?
12:24:43JdGordoncan anyone with a hard disk player check if the rec screen causes a spin up when its entered with svn?
12:25:46pixelmaJdGordon: it does - at least with swcodec...
12:26:06JdGordonok, so I guess I dont need to worry about it
12:26:47JdGordon:'( h300 battery just died... pixelma you happy with how 7201 works?
12:30:37pixelmahaven't tried the latest because the patch didn't apply cleanly (one of the 2 failed hunks didn't look like easy to fix at least not at first glance)
12:30:58JdGordonok, ill upload a new one
12:30:59pixelmawhen I wanted to try on saturday
12:32:40 Quit crop ("CGI:IRC")
12:32:52JdGordonnew one is up
12:33:49 Quit kretender (Read error: 104 (Connection reset by peer))
12:36:06 Quit pondlife ("disconnected has pondlife")
12:49:58 Join Lear [0] (i=chatzill@rockbox/developer/lear)
12:51:02 Part maffe
12:56:10LearJdGordon: Btw, the new filename check is more restrictive than necessary (because short name char limits are applied to long names). Also, the char2dos function doesn't really do anything as it is...
13:00:18 Join printfXh4 [0] (
13:00:43preglowthere doesn't seem to be a page on sansa runtimes, anyone got an estimate on how long it'll run rockbox on a charge?
13:00:48preglowi've seen a graph that says something around 10h
13:04:03***Saving seen data "./dancer.seen"
13:14:06preglowright, right
13:14:09preglowthanks too
13:14:31peturTak Tak?
13:19:05preglowhmm, and am i right if i say that the newer european models of the sansas do have fm radio?
13:24:19 Quit DerPapst ("So Long And Thanks For All The Fish!")
13:27:37GodEaterdoesn't "Tak" = "Roof". I thought it was "Tack" for "Thanks" ?
13:27:55 Quit Zagor ("Client exiting")
13:28:07peturI was thinking Icelandic
13:28:27 Quit miepchen^schlaf (Read error: 113 (No route to host))
13:28:37 Join Thundercloud [0] (
13:29:05GodEateroh ok :)
13:29:13peturhmmm sorry, it's Takk
13:29:27petur(means thanks in Icelandic)
13:29:42gregjello chaps
13:29:44peturDak = roof in Dutch
13:29:56gregjGodEater: Tak means Yes in Polish
13:30:09pixelmaGodEater: that's swedish... "takk" in norwegian IIRC
13:30:56GodEaterI just assumed petur was talking swedish, given the predominance of them in #rockbox ;)
13:31:15preglowpetur: takk in norwegian, too
13:33:34 Join midkay_ [0] (
13:34:30 Quit Thundercloud (Remote closed the connection)
13:34:39linuxstbpreglow: Thinking of buying a Sansa/
13:34:50 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
13:34:55preglowlinuxstb: no, but a mate is
13:35:59 Join Genre9mp3 [0] (n=yngwiejo@rockbox/contributor/Genre9mp3)
13:36:20 Join Thundercloud [0] (
13:36:52 Quit spiorf (Read error: 110 (Connection timed out))
13:37:12 Join spiorf [0] (
13:42:13 Join kretende1 [0] (
13:42:43 Join freqmod [0] (
13:43:39*linuxstb looks at the binary sizes table and wonders why the WMA codec added about 4KB on some ARM targets, but only 3KB on others.
13:46:50 Join Bob7k [0] (
13:49:25 Quit midkay (Read error: 110 (Connection timed out))
13:49:51preglowhi, bob
13:50:28Bob7ki checked back, last time i was at rockbox was december i think, the AV500 project was down to crashing the player
13:50:49Bob7kbut now they got rockbox kinda working on it, i think
13:51:08LinusNgood for them, whoever they are
13:52:02markunBob7k: do they have a website?
13:52:31Bob7kyou didnt know it? lol, here it is
13:52:38markunWhat's the URL?
13:52:44Bob7kit wont let me paste it
13:53:17Bob7ki guess not too many people bought the AV500
13:53:18markunLast time I checked those guys didn't have anything working on the AV500
13:53:35Bob7kthey got a video to crash it, aparently thats a step
13:54:00Bob7know they got MediOS on it, but i dont know how to install, or use it
13:54:05markunMaybe I just missed something
13:54:20markunMediOS is from
13:54:29Bob7kive rockboxed my iPod 5th gen, way better than what apple loaded onto it (all my stuff was in WMP11, not iTunes)
13:54:38Bob7kforgot about that
13:54:56Bob7ki got them confused
13:55:16Bob7ki forgot, rockbox has nothing for the AV500. sorry
13:55:23 Join jac0b [0] (
13:56:04*LinusN giggles
13:56:06Bob7kbut its opensource, so im sure theres something in the code to get rockbox running on the AV500
13:56:09 Join Febs [0] (
13:56:42markunsure, there could be some useful stuff in there
13:57:10jac0bI have just put rockbox on a gigabeat but I can't save my settings and I get a error when trying to put a song on the playlist
13:57:18Bob7ki just liked the way rockbox ran on my iPod, and wanted it on my archos, or anythingn similar
13:57:19markunBob7k: the biggest problem is the lack of open documentation for the CPU (which we call DM320 here)
13:57:38 Quit GodEater (Read error: 110 (Connection timed out))
13:57:40Bob7kyeah, wish archos turned over the source code for all thier old players
13:57:59LinusNjac0b: sounds like you are missing the .rockbox directory
13:58:13markunLinusN: I was hoping for an email from a girl.. only to find out it was you :)
13:58:29Bob7kbut i doubt that they will do that. would be nice if they released the source code on all discontinued players and stuff
13:58:33LinusNmarkun: sorry to disappoint you :-)
13:58:33jac0bno its there rb starts up
13:58:44 Join GodEater [0] (
13:58:52jac0bI am listening to it right now
13:58:54amiconnJdGordon: Disk spinup on entering the rec screen is okay, but it must not spin up when starting the recording
13:59:02markunjac0b: maybe do a scandisk?
13:59:14 Quit lazka ("I'm off now")
13:59:24 Join lazka [0] (
13:59:43jac0bmarkun: do a scandisk when I get it connected to my computer
14:00:00markunpreglow: TTSBox has finally ordered a DAP, a Gigabeat F20
14:00:09Bob7kanyone know how many mhz the AV500 has?
14:00:56markunBob7k: maybe the guys in #archopen know
14:01:19 Join perplexity [0] (
14:01:22Bob7kmarkun: i didnt think they had an IRC server
14:01:31GodEaterthey don't - they use this one
14:01:54Bob7kthat explains my confusion over MediOS
14:02:06markunBob7k: this is all I found
14:02:40markunBob7k: this server, but not this channel :)
14:03:26Bob7kim still learning the IRC server thing stuff
14:03:56markunBob7k: just type "/joing #archopen"
14:04:06markuneh, /join
14:04:15Bob7kin this window, or the one in the back? i think its the one in the back
14:04:22amiconnlinuxstb: I want the new clock setup code to be stress tested as much as possible on PP5020
14:04:31markunBob7k: you can use any window I think
14:04:43amiconnI am especially interested in H10 behaviour, as that was not tested before
14:04:44Bob7ki did it on the back one, it worked
14:05:06 Quit webguest31 ("CGI:IRC")
14:05:07peturamiconn: I've asked a friend to load current build on his H10 and report back to me
14:05:15*amiconn would also like jhMikeS to update his core locking patch
14:05:22 Join webguest31 [0] (i=c0231115@gateway/web/cgi-irc/
14:08:17 Quit billytwowilly (Remote closed the connection)
14:08:33 Join billytwowilly [0] (
14:08:39lazkaI have a 600kb WMA file that crashes my player.. if someone is interested
14:09:00Bob7kis it for the AV500?
14:09:10Bob7kwhat player does it crash?
14:09:23GodEaterBob7k: one of the rockbox ones at a guess
14:09:28GodEaterwhich means not the AV500 ;)
14:10:12Bob7kGODEATER: yeah, id totally love rockbox on my AV500, my sister loved it on my iPod, and everyone i showed it to liked it
14:10:43GodEaterlinuxstb thought about porting it to the AV series once upon a time. Not sure where he got to with it.
14:11:13*GodEater finishes dropping linuxstb in it :)
14:12:02Bob7kive got an AV500 30 gig, they stopped making the 100Gb ones
14:12:27Bob7kanyone know how to install medios?
14:12:45amiconnlinuxstb: The 4K increase is on arm targets without recording, and 3K on arm targets with recording.
14:12:52*amiconn wonders about the reason
14:13:12markunamiconn: yes, me too
14:13:25 Join maffe [0] (
14:13:59markunamiconn: if you disable recording the size increases or could it be a coincidence?
14:14:59amiconnI guess something got enabled for all targets on accident which was only enabled for recording capable targets before
14:17:57amiconnOr the wma metadata parser needs some function that was already linked before, but only on recording targets
14:22:45Bob7kanyone know how to downgrade an AV500 to 1.5.12 from 1.5.13? PM me
14:24:45GodEaterBob7k: did you try #archopen yet ?
14:24:56Bob7kthats why i need to downgrade
14:25:22GodEaterso that you can speak to the people in #archopen ?
14:25:25*GodEater is confused
14:25:40Bob7ki tried to use the link to archos' site, and my AV500 came up with an error, something about the firmware being too old for my hardware
14:26:04GodEaterBob7k: so go ask the guys in #archopen - they should be able to help.
14:26:06Bob7k#archopen is completly inactive
14:26:16Bob7knobody is saying anything there
14:26:22markunbastards :)
14:26:33GodEaterBob7k: well this isn't the right place to ask I'm afraid.
14:26:39jac0bI was looking over the install instructions for the gigabeat and saw that I did not eject/unmount it after I installed rb. Could that be why I am having problems?
14:26:49markunjac0b: could be
14:29:47preglowhrm, dircache seems to do a full foreground scan each time i turn on my h120 now
14:31:20dionoeathe new wma codec is wma1 and 2 only right ?
14:31:27peturpreglow: join the club
14:33:05amiconnpreglow: Do you shut down directly from wps?
14:33:54amiconnThat's what I observe here. Shutdown from wps -> foreground scan on next boot. Shutdown after stopping explicitly -> no fg scan
14:36:50preglowamiconn: hmm, i'll check it out when i'm done with it now
14:39:34linuxstblazka: Yes, I'll have your WMA file...
14:40:09 Quit desowin ("use linux")
14:40:42linuxstbamiconn: So I should just play a boosting album on repeat with current SVN on my Color?
14:41:32jac0bWMAs now work?
14:41:47linuxstbGodEater, Bob7k: My AV (potential) target gathering dust is an AV300, not AV500.
14:42:22linuxstbjac0b: Initial support is in the current Rockbox builds, but there's still work to do.
14:42:36*dionoea feels like his question was ignored :)
14:43:04lazkalinuxstb: it seems it has some errors (some sound-dropouts) but here it is:
14:43:07linuxstbBob7k: Godeater suggested I was interested in porting Rockbox to the AV300, I'm not.
14:43:23preglowdionoea: i think it could potentially do all the wma files out there, as long as they're not lossless
14:43:25linuxstb^I meant AV500... (I am interested in the AV300)
14:43:32preglowbut i'm not sure at all
14:44:17*GodEater was confused and hangs his head in shame.
14:44:25dionoeawmva 3 specs aren't public are they ?
14:44:33dionoeawhich mackes it wma 3
14:44:37linuxstbNo wma specs are public afaik
14:45:00linuxstbThe ASF spec is published, but under strict (and not open source friendly) conditions.
14:45:17dionoeaok, let me rephrase that: no open source wma3 decoder exists right?
14:45:27preglowi don't even know what versions of wma exist
14:45:38dionoea1 2 and 3 :) must have info about that
14:45:47linuxstbThe ffmpeg codec handles codec_id 0x160 and 0x161 - wmav1 and wmav2
14:47:56linuxstblazka: Which device are you running Rockbox on?
14:49:31 Quit midkay_ ("Leaving")
14:49:44 Join midkay [0] (n=midkay@rockbox/developer/midkay)
14:49:58lazkalinuxstb: sansa
14:51:23linuxstbBy "sound dropouts", do you mean errors in the decoding, or just pauses because it's decoding too slowly?
14:55:58Bob7kmy AV500 isnt playing the crash boat clip, anyone know whare i can find it somewhare else?
14:57:19 Quit jac0b ("ChatZilla [Firefox]")
14:57:27peturBob7k: again, this isn't the place for AV500 support
14:57:40 Quit hcs (Read error: 110 (Connection timed out))
14:57:42 Part Bob7k
14:59:21 Part LinusN
15:03:17*amiconn was right
15:03:36amiconnlinuxstb: The difference is the 64 bit division (_udivdi3)
15:04:07***Saving seen data "./dancer.seen"
15:06:31 Join XavierGr [0] (
15:06:44linuxstbHmm, I'm not sure if I can avoid that - ASF files contain the file duration in usecs...
15:08:36preglowi think we'll survive that
15:08:54linuxstb1KB of code for that seems a bit excessive though...
15:09:10 Join kaaloo [0] (
15:09:16 Join desowin [0] (
15:09:37amiconnThe coldfire version is larger...
15:09:42 Part kaaloo
15:10:05amiconn(at least the one for cf variants without 'div' insn)
15:10:36 Join pondlife [0] (
15:11:35preglowdoubt we'll ever need that
15:11:49preglowall of our coldfires have div
15:15:01amiconnDunno how large udivdi3 for those is
15:18:07 Join saratoga [0] (i=9803c6dd@gateway/web/cgi-irc/
15:18:16 Quit SirFunk (Remote closed the connection)
15:21:53pondlifeThe viewvc diffs for the 64-bit WMA commit are mainly blank... e.g
15:22:02pondlifeIs this correct?
15:22:15pondlifeAh keywords only...?
15:22:44saratoga"wow, that line is 38941 bytes long"
15:23:07saratogai guess thats the complete set of MDCT coefficients I put in there code for debugging
15:23:24*petur wonders if saratoga checked his mails...
15:26:42lazkalinuxstb: by "sound dropouts" i mean decoding errors with VLC and gstreamer. my sansa crashes before it starts playing (loading... for ever)
15:27:06 Quit perplexity (Read error: 113 (No route to host))
15:27:09 Quit webguest31 ("CGI:IRC")
15:28:26saratogapreglow: regarding the sansa runtime, someone in the forums reported 12-14 hours in rockbox
15:28:40saratogabut I believe the battery bench plugin does not work for it, so its just an estimate
15:28:48saratogabattery life does seem quite good though on mine
15:29:07saratogadefinately better then my old Ipod (running the Apple firmware)
15:32:50saratogacan I see that wma file that didn't work?
15:33:05saratogaalso, we should really make it clear that the current codec cannot decode files under 32kbps
15:42:09 Part maffe
15:45:16saratogalazka: I noticed your download link and checked out that file
15:45:36saratogaits <32kbps so my decoder cannot handle it
15:46:07preglowdo you think it'll be hard to support?
15:46:13saratogayes I do
15:46:28saratogathe older fp version of ffmpeg I have handy can't decode it either
15:46:42preglowmaybe it's been fixed now
15:46:51saratogaand the "line spectral pair" code in SVN is horribly broken is several ways
15:46:57preglowany opinion on how much work an ffmpeg svn merge will be?
15:47:09saratogai do intend to fix it though
15:47:15saratogaI'm not sure
15:47:23saratogaare they intested in an int decoder?
15:47:36preglowthey should be asked, at least
15:47:47preglowi think they're only interested if people are willing to maintain it too
15:47:54preglowbut i'm by no means certain
15:47:59saratogai was thinking of cleaning it up and releasing it as libwmafixed or something
15:48:20saratogathe problem I see is that it will be slower then the fp version on almost any system that could run ffmpeg
15:48:27preglowin that case it'd be better for everyone if you tried to merge your stuff with ffmpeg
15:48:43preglowwell, sure, but ffmpeg is also used on systems with no fpu
15:49:32saratogaok, then I should probably post on their mailing list
15:50:18saratogai was planning to sync all their fixes anyway
15:51:19saratogaregarding cleaning up the source code, the logs mentioned not liking // comments
15:51:25saratogawhat else should I be doing?
15:51:47preglowwell, not much
15:51:53preglowuse tabs
15:52:18preglowbut anyway, don't replace // comments that are there in the ffmpeg tree
15:52:42preglowif a change has no good reason apart from changing the syntax, don't do it, since it'll just make a later merge harder
15:53:14saratogai assume that this will be added as a second ffmpeg codec, right?
15:53:29saratogasecond wma codec I mean
15:56:05saratogaI assumed this would go in as wmadeci and they'd keep wmadec as a seperate codec
15:56:25preglowwouldn't that duplicate quite a lot of code?
15:56:33saratogayes it would
15:56:36preglowthe current codecs they have share code between floating point and fixed point code
15:56:51saratogathey just #define the fixed functions then?
15:56:53preglowmacros like MUL and ADD are used to abstract between versions
15:57:11preglowthis is also how it's done in all of our current codecs
15:57:18preglowin the ones also supporting floating point operation, that is
15:57:38saratogaprobably easier to just check out their current source and start adding the fixed stuff line by line then
15:59:51saratogaI'll take a look at merging the recent ffmpeg improvements today
16:00:36 Quit pondlife ("disconnected has pondlife")
16:02:40 Join pondlife [0] (
16:05:17preglowsaratoga: do you have a list of high-prio stuff you want to do?
16:05:59 Quit Zagor ("Client exiting")
16:08:09saratogapreglow: I guess the first thing is finding out if anyone has problem samples in higher bitrate files
16:08:30saratogai'm not really comfortable with some of the code in the current version, though so far I haven't been able to make a file that breaks it
16:08:37saratogaother then very low bitrate ones of course
16:08:41 Quit pondlife ("disconnected has pondlife")
16:08:57 Join pondlife [0] (
16:09:06saratogathen I want to work on performance, its barely realtime on the PP players running at 100% boost
16:09:23saratogathen I want to merge in the ffmpeg changes
16:09:35saratogafinally, I want to see if it can be merged back into ffmpeg
16:09:44saratogawith that last one being a pretty distant fourth
16:11:26linuxstbsaratoga: I think it would be nice to spend some time cleaning the code - continue moving things back to their separate .c/.h files, removing unused debugging code (or at least making it work with DEBUGF and appropropriate defines) etc.
16:11:51linuxstbAlso, if the codec will crash on all < 32kbps files, I should probably reject them in the metadata parser.
16:12:17saratogalinuxstb: yes that would be nice
16:12:34saratogaalso, I have some SVN questions
16:12:49 Quit Nico_P (Remote closed the connection)
16:12:50linuxstbWas that in reply to my first sentence, or second (or both)?
16:13:02saratogahahha, second
16:13:07 Quit pondlife (Client Quit)
16:13:10saratogai'll begin on the first now
16:13:17saratogaanyway, I never did figure out how to generate a patch that combined my changes with your first patch
16:13:21saratogahow is that done with SVN?
16:13:27linuxstbDid you do "svn add" on the new files?
16:13:28 Join pondlife [0] (
16:13:41saratogayou mean the ones in the patch?
16:13:42linuxstbIf not, then "svn diff" wouldn't have included them.
16:13:49saratogaah, no
16:14:17 Join miepchen^schlaf [0] (
16:14:19saratogai need to see if that book google sent me explains SVN
16:14:25preglowyeah, the ffmpeg merge is pretty low-prio, it doesn't really concern rockbox directly
16:14:33linuxstbThere's also the "UsingSVN" wiki page.
16:14:36 Join thegeek [0] (
16:14:44saratogayes thats how I can manage to run diff in SVN ;)
16:15:24 Join webguest35 [0] (i=c4dbb919@gateway/web/cgi-irc/
16:15:26linuxstbMy advice regarding svn would be to always do "svn diff" before an "svn commit", so you can preview the changes you're about to commit.
16:16:03linuxstbAnd the preference is for small, single-change commits, and for white-space/formatting commits to be done separately.
16:16:34*amiconn would like to see wma realtime on coldfire
16:17:26linuxstbsaratoga: So < 32kbps files are definitely unsupported at the moment?
16:17:45saratogalinuxstb: yes absolutely
16:17:52saratogaeveryone I've found crashes the decoder
16:18:24linuxstblazka's sample works fine in a recent mplayer (which uses ffmpeg), so I assume the problem is fixed in SVN ffmpeg.
16:18:35saratogagreat, that gives me a place to start
16:19:45webguest35hi guys
16:19:51 Join SirFunk [0] (
16:20:31saratogalinuxstb: should I change the LGPL license to GPL now that its in rockbox?
16:20:39saratogaor do we just leave the ffmpeg headers alone?
16:20:43linuxstbYou can't just change the license...
16:21:13saratogalinuxstb: I thought changing LGPL to GPL was legal?
16:21:24 Join Crash91 [0] (n=crash@
16:21:57 Quit webguest35 ("CGI:IRC (EOF)")
16:22:28Crash91does anyone know the status of USB connection for the Sansa?
16:22:38LearYou can pick any license you like, if you're the copyright holder...
16:22:41GodEaterCrash91: it's still being worked on.
16:22:48 Quit atsea- (Remote closed the connection)
16:23:01Crash91@GodEater Thanks
16:23:07 Join perplexity [0] (
16:23:09saratogaCrash91: Austriancoder posted about his progress earlier today in the logs
16:23:24linuxstbsaratoga: I've committed the < 32kbps check.
16:23:52 Quit pondlife ("disconnected has pondlife")
16:23:56austriancoderCrash91: at the moment: code clean up and gadget driver for sortage
16:24:35Crash91Looks like it could be up soon :)
16:26:19 Join pondlife [0] (
16:26:40GodEaterlinuxstb: did you ever pass comment on my suggestion for meeting with Febs *at* his hotel and then wandering off to find somewhere to drink from there ?
16:26:41Crash91do i have to be root to use Sansapatcher (Fedora 7)?
16:27:02GodEaterCrash91: unless Fedora has got some odd device permissions, then yes
16:27:14saratogalinuxstb: would it be poor form to put all the fixed point code in one .h file?
16:27:38saratogai dont really like putting code in those, but it seems annoying to put the #defined operations seperately from the function calls
16:27:40linuxstbGodEater: No... I guess we could do that, but I think a pub would be nicer if we can think of one...
16:28:08GodEaterI don't know many near Aldwych at all - so I'm open to suggestions
16:28:08 Quit joa ("( :: NoNameScript 4.02 :: )")
16:28:10preglowfinding a pub in london
16:28:18preglowwhat hardships you face
16:28:23linuxstbsaratoga: Normal functions should go in a .c file, "static inline" or macros belong in a .h
16:28:29GodEaterpreglow: linuxstb has very strict requirements ;)
16:28:37 Nick joshin_ is now known as joshin (n=joshin@unaffiliated/joshin)
16:28:43GodEaterand we're trying to avoid tourists too
16:28:46Crash91lol, like what
16:29:00preglowi prefer my pubs with no televisions and loud music and with a decent selection of beers
16:29:03preglowthat would be it
16:29:17GodEaterthat would be a good one to start it
16:29:23preglowthough the selection of beers is of course what matters
16:29:23GodEaterlinuxstb requires Guinness too only 15 lol
16:29:36FebsGodEater, Mrs. Febs and I will try not to act too tourist-y.
16:29:45linuxstbYou need to be unlucky to find a pub without Guinness in London
16:29:58pregloweven most norwegian "pubs" have guinness these days
16:30:05linuxstbYou'll leave your baseball caps and cameras in the hotel then?
16:30:06 Join Arathis [0] (
16:31:27GodEaterand the "I love London" T-shirts too
16:31:32Febslinuxstb: fortunately, I am not a big fan of wearing baseball caps, except when actually playing baseball.
16:31:48Crash91Ive been having some wierd problems with the WPS lately
16:32:19GodEaterlinuxstb: right - I shall go exploring near Aldwych tonight then
16:32:27preglowbe sure to sample the beer
16:32:44Crash91Text is off centre, etc. read full account here:
16:33:00GodEaterpreglow: I want to be able to remember which pubs I went to though :)
16:33:25GodEaterCrash91: I presume you've got all the right fonts ?
16:33:42Crash91im using HelvR12
16:33:55GodEaterIs that what's specified by the WPS though ?
16:34:24saratogalinuxstb: right now I've got the CORDIC trig function copied into the codec
16:34:42saratogais that the correct thing to do, or can I use the one already in rockbox (i forget where...)
16:36:05preglowi think it's quite ok to keep it in the codec
16:36:15preglowat least for now
16:36:34Crash91the tag is simply %ac%?ia<%ia|%fn>, and this happens with other fonts
16:37:03GodEaterCrash91: did you download and install the from the "Extras" link under downloads on the rockbox home page ?
16:37:19GodEaterbecause to me - it sounds like you didn't
16:37:31Crash91Yes, i did
16:37:47JdGordonamiconn: ok cool, so should I commit the recording path patch?
16:37:49Crash91this is only tested on sim for one theme
16:38:08Crash91but ive had the same problem with another on the player
16:38:15GodEaterCrash91: well I'd post about this in the rockbox WPS forum if I were you
16:38:21GodEaterI'm far from a WPS expoert
16:38:22Crash91I tried updating the sim etc.
16:38:27Crash91kk will do
16:38:57 Join [IDC]Dragon [0] (i=c19e01ec@gateway/web/cgi-irc/
16:39:05austriancoderhow got the sansa emulator ( working?
16:39:48Crash91GodEater: I recieved this as a reply to my thread on Lodist forums
16:39:49Crash91it may just be too complex. i read somewhere that if its too complex it will only load the first part of it.
16:40:03GodEaterCrash91: yes I read that
16:40:06GodEaterthat's not quite right
16:40:37GodEaterI believe it's possible to use the sim to tell you about WPS errors you may haev
16:40:43saratogaaustriancoder: I believe Toni was the only one really working on that, and he seems to be doing Sansa stuff directly now
16:40:47JdGordonaustriancoder: Toni did.. he is almost never here
16:40:47GodEaterbut again I'm not sure on the details - I don't do WPS design very much.
16:41:21Crash91Ok, posting topic.....
16:41:24[IDC]Dragonhello Rockbox world
16:41:34austriancodersaratoga: okay.. i tried oit to run.. but.. failed
16:41:58[IDC]DragonI'm unsure about 2 commits from today
16:42:02*amiconn welcomes RareGuest aka [IDC]Dragon ;)
16:42:21amiconn[IDC]Dragon: I have a question regarding the rvf video converter tools
16:42:27amiconn(the commandline stuff)
16:42:41[IDC]Dragonthis ilegal char checking, would it break the multivolume names in <> ?
16:42:51amiconnWould you mind putting them in svn? And what license are they under (the headers state nothing)
16:43:23amiconn[IDC]Dragon: No it won't, as these dirs are just virtual
16:43:28[IDC]Dragonand the FS #7387 fix, shoudn't that test for multivolume instead of hotswap?
16:44:00saratogaaustriancoder: getting it to run typically involved a lot of time with the VS6 debugger and asking Toni for hints
16:44:08JdGordon[IDC]Dragon: no, you cant unmount multivolume.. only hotswap
16:44:21pixelmaArathis: do I remember correctly that you have an H10?
16:44:37[IDC]Dragonok, please excuse my unqualified remarks
16:44:49Arathispixelma: yes. an H10 with 20gb
16:44:55JdGordonpixelma: amiconn: commit 7201?
16:45:19pixelmaArathis: have you tried the latest svn on it?
16:45:34[IDC]Dragonamiconn: I have no problem with gpl'ing the command line tools
16:46:16[IDC]Dragonthe DirectShow filter has to stay closed source, sorry
16:47:32Arathispixelma: no. I've got an svn build from yesterday. any h10 specific changes?
16:48:05amiconn[IDC]Dragon: The ds stuff is (1) too unflexible and (2) not cross platform anyway
16:48:06[IDC]DragonOT: is sourceforge a good starting point for a fresh open source project, or what othe service might be better?
16:48:16*amiconn always creates rvfs via commandline
16:48:29pixelmaArathis: yes (not only H10 but affects it) - see the recent commits on the front page
16:48:31GodEater[IDC]Dragon: depends on what you're looking for
16:48:54[IDC]Dragonrevision control basically
16:49:12pixelmaArathis: the one of 3 Jul 00:42
16:49:16GodEaterdo you have a preference for a tool to do that ?
16:49:25[IDC]Dragona project home page or a forum would be nice, too
16:49:42GodEater[IDC]Dragon: ah - then your options narrow considerably at that point :)
16:49:58[IDC]Dragonsvn would do, cvs if I have to
16:49:59Lear[IDC]Dragon: Paths with '<' or '>' would probably cause problems, yes...
16:50:36Crash91rockbox is linux-based right?
16:50:36 Join amiconn_ [0] (n=jens@rockbox/developer/amiconn)
16:51:01GodEaterCrash91: not in the least
16:51:06JdGordonLear: char2dos doesnt do anything?
16:51:15Crash91From scratch then?
16:51:39[IDC]DragonGodEater: forum or webspace are available elsewhere, too, so the main thing is code revision control
16:51:44JdGordonwith bits from the leaked win2k sources
16:51:54pondlifeLear: Have you seen ?
16:52:16 Quit amiconn (Nick collision from services.)
16:52:17 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
16:52:17Crash91GodEater: How do you explain this in my rockbox-info.txt
16:52:18LearJdGordon: Not any more, since all paths with such chars are rejected before it is called.
16:52:19linuxstb[IDC]Dragon: I've no idea what it's like, but google have a project hosting service for open source projects. No forums (I think), but SVN and a wiki.
16:52:19*JdGordon was joking if there are any MS lawyers lurking :p
16:52:21GodEater[IDC]Dragon: I've become a git convert - so I'd recommend a place that uses that ;)
16:52:22Crash91gcc: arm-elf-gcc (GCC) 4.0.3
16:52:22Crash91ld: GNU ld version 2.16.1
16:52:22Crash91Host gcc: gcc (GCC) 4.1.2 20061115 (prerelease) (Debian 4.1.1-21)
16:52:22DBUGEnqueued KICK Crash91
16:52:22Crash91Host system: Linux
16:52:26 Join bluebrother [0] (i=s3axidIe@rockbox/staff/bluebrother)
16:52:35 Quit XavierGr (Nick collision from services.)
16:52:37GodEaterCrash91: that means it was BUILT on linux
16:52:38 Join XavierGr [0] (
16:52:45linuxstb[IDC]Dragon: (very minimal page...)
16:52:58Crash91Ok, sorry for being a noob :P
16:53:21[IDC]Dragonyes, I saw Google "competes" Sourceforge
16:53:34Learpondlife: No, but just read it. Sounds like the problem that was hopefully fixed.
16:53:35 Join PaulJam [0] (
16:53:41[IDC]Dragonbut I don't trust these guys very well...
16:54:02pondlifeLear: I was wondering if that would be related... I'll ask them to try a new build.
16:54:18Crash91has anyone seen this revolutionary development by google?
16:54:24JdGordonLear: ah yes, your right
16:55:12 Join Esp_ [0] (n=Esp_154@
16:55:20LearJdGordon: And fat_checkname should be relaxed a bit...
16:55:29JdGordonwas about to say that
16:55:56JdGordonLear: the while() in fat_checkname can be removed I tihnk....
16:56:21Esp_oh dear.. my 1 gen ipod nano 2gb just reported bad CRC with ipodpatcher.exe reading the partitiontable. Is this a known problem
16:56:38LearBut then you're basically back to what it was before... :)
16:57:38JdGordonbah, then leave it as is...
16:57:58CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
16:57:58*JdGordon tihnks its better to fail than create a file/folder with unexpected names
16:58:08Learfat_checkname checks the long name, which allows a few more chars (and as IDCDragon mentioned, with some special multivolume handling).
16:58:28pixelmaJdGordon: I just tested the recpatch and got a "Panic: recfile -116" again - just took a dir on the mmc called "recordings" set as recording dir, started recording and stopped again... (didn't take it out or something. But have to check because that MMC is rather full - but even if it has run out of space it should give me the "disk full" splash...
16:59:47 Join jgarvey [0] (
16:59:56pixelmait didn't write a new file... and there are 7MB available (would do for a few minutes mp3)
16:59:57Crash91does anyone here own a sansa?
17:00:38 Quit Entasis (Read error: 104 (Connection reset by peer))
17:00:43JdGordonCrash91: a few of us..
17:01:13Crash91well, here you go, you might like them
17:01:30Esp_strangely iTunes seems to be able to recover the nano, but it's still faulty after restarting it. It even shows the serial and sw-version but unable to get the flash-size
17:02:45GodEaterthat would be my choice - but that's just me ;)
17:03:11linuxstbEsp_: It sounds like your flash disk is failing. Is it still under warranty?
17:03:15pixelmaJdGordon: yes, happens when it tries to write the data
17:03:37saratogalinuxstb: what does it mean when I'm suddenly getting hundreds of "pointer targets in initialization differ in signedness" messages from every file in rockbox?
17:03:55JdGordonsaratoga: bad gcc .... i have the same problem
17:03:59Esp_linuxstb: probably not. it's over 1 year old
17:04:09***Saving seen data "./dancer.seen"
17:04:16saratogaJdGordon: can I safely ignore it?
17:04:32linuxstbsaratoga: Is that building the sim?
17:04:34JdGordonyeah, but be careful before commiting anything incase there are other warinings
17:04:59pixelmaJdGordon: wait... seems I copied from the old build... *slaps forehead*
17:05:08JdGordon*phew* ")
17:05:11saratogalinuxstb: no, gigabeat
17:05:15JdGordonprob gonna get the same problem anyway
17:05:20JdGordonsaratoga: you on fiesty?
17:05:24linuxstbsaratoga: Which version of arm-elf-gcc do you have?
17:05:29Learpixelma: what do paths on the mmc look like?
17:05:36saratogaCentos, 4.0.3
17:05:42Lear(In Rockbox)
17:05:44saratogait didn't do this before
17:05:46JdGordonLear: /<MMC1>/...
17:06:20 Join forehead [0] (i=54bd747f@gateway/web/cgi-irc/
17:06:45saratogamy build worked anyway, so I guess its nothing
17:06:47JdGordonLear: the point of the fat checking was so we can set the recording path to that and ave it fail if the card isnt inserted
17:07:20amiconnJdGordon: The fat code should only ever check to-be-created entries, never the complete path
17:07:36JdGordonit does
17:07:43 Join richcoosa19 [0] (
17:08:05amiconnThen <MMC1> in the path should work...
17:08:24Crash91Why is everyone switching to Ubuntu these days??
17:08:25LearIndeed. Didn't think about that...
17:08:42[IDC]DragonGodEater: interesting site, thanks
17:08:46JdGordonamiconn: no because it tries to create it if its not there
17:08:48richcoosa19Is anyone else having problems creating the database on a Sansa e200 using SVN?
17:09:02amiconnJdGordon: But it _is_ there when the card is inserted...
17:09:11JdGordonyes, so it works
17:09:16JdGordonif its not inserted it fails
17:09:21preglowCrash91: because it's nice?
17:09:36Crash91preglow: Fedora is nicER
17:09:40amiconnJdGordon: Yes, and that is what it should do
17:09:43preglowmatter of opinion, i don't think it is
17:09:49GodEater[IDC]Dragon: <−− read that too incase you've missed our conversations on it recently
17:10:04scorcheCrash91: please try to stay on-topic..
17:10:05amiconnIt's just that the recording code shouldn't panic but rather error out
17:10:09 Join Soap [0] (n=Soap@rockbox/staff/soap)
17:10:09JdGordonamiconn: ok cool, so then we are all happy
17:10:18[IDC]DragonI've missed a lot lately
17:10:24JdGordonyeah, I dunnow hy the ondio is stuffed, works on the sansa and h300
17:10:29JdGordoncould be a hwcodec thing :(
17:11:02saratogaI've cleaned up wmadeci.c a bit, anyone want to commit the patch for me?
17:11:16saratogajust moves constants and fixed point functions into proper files
17:11:20linuxstbYou can't commit it yourself yet?
17:11:20 Quit forehead ("CGI:IRC (Ping timeout)")
17:11:30saratogastill waiting for confirmmation from linus
17:11:35[IDC]DragonI haven't even followed Rockbox switching to svn, now you're unhappy again and want to switch to git!
17:11:55scorche[IDC]Dragon: not everyone...
17:11:57 Join Nico_P [0] (
17:12:00saratogaalso, I think this version should be faster on Coldfire
17:12:05*amiconn doesn't want to switch to git
17:12:14saratogaseems I left some debug code in that was slowing that version down a bit
17:12:15Nico_Pamiconn: why ?
17:12:28amiconnI can't see an advantage that would be worth the hassle
17:12:29[IDC]Dragonis there a windows client for git?
17:12:33*Nico_P goes read the logsd
17:12:36 Join maffe [0] (
17:12:41amiconnTo the contrary, I think it's better to have one central repo
17:12:47scorche[IDC]Dragon: not that i know of
17:12:48[IDC]Dragon"TortoiseGit" perhaps? ;-)
17:12:52bluebrotherafaik there is only a cygwin version of git.
17:12:52LearCygwin client at least.
17:13:04bluebrotherwhich is dog slow, as far as I heard.
17:13:16scorchewell, it is cygwin =P
17:13:49Esp_linuxstb: ah.. found the diagnostic menu.. that should be handy...
17:14:02*bluebrother also doesn't see a reason to switch version control system
17:14:28*pondlife also votes for sticking with svn
17:14:31bluebrotherif we have a git mirror everyone who wants to can use git.
17:14:44bluebrotherand all other can stick to svn :)
17:14:45[IDC]DragonI don't want to raise such obstacles to my project
17:14:51GodEateryes, the windows filesystem does not lend itself to a fast git implementation apparently
17:15:06scorche[IDC]Dragon: what project is this?
17:15:25[IDC]Dragonmy house control stuff
17:15:37GodEater[IDC]Dragon: 'twas just a suggestion - sourceforge is probably still better for your needs
17:15:47[IDC]Dragoncode for Atmel AVR microcontrollers, mostly
17:16:06GodEateris that X10 based stuff ?
17:16:06lazkarichcoosa19: you are right. building database doesn't work with latest SVN.
17:16:23[IDC]Dragonno, it's RS485 based
17:16:36LearJdGordon: I think I'll adjust my fat filename check and apply that then (different char check in fat_checkname, smaller code).
17:17:16JdGordonas long as it fails trying to create anything with < im happy...
17:17:18mtoupsoh, git...
17:17:19 Quit mtoups ("leaving")
17:17:54 Join mtoups [0] (n=mtoups@HENSON.ISR.CS.CMU.EDU)
17:17:56[IDC]Dragoncan that google version control be used by tools, or only via web frontend?
17:18:20[IDC]Dragonwhat's under the hood?
17:18:21LearJdGordon: Yep, no change there.
17:19:03pixelmadon't want to end up with 2 <MMC1> folders again ;)
17:19:25GodEater[IDC]Dragon: I've no idea
17:19:30amiconnDeleting an entry with < > created by rockbox can be tricky on a pc....
17:19:48*amiconn did it with a disk editor
17:19:56 Join atsea- [0] (i=ariel@gateway/tor/x-9615000c180555e5)
17:20:00GodEater[IDC]Dragon: looks like svn
17:20:18 Join The-Compiler [0] (
17:20:50JdGordonany objections to so the ata led icon turns on when the disk starts spining up instead of once the spinlock is aquired? (fix 7393)
17:21:03JdGordon... and someone with a hd target test please
17:21:52pondlifeJdGordon: I can't test right now, but I'd prefer that behaviour
17:25:07 Join joa [0] (n=joe@
17:25:58JdGordonpixelma: so did the new version work?
17:27:45linuxstbsaratoga: Your patch will break big-endian targets. You need to make sure you #include <config.h> (which is included by most, if not all Rockbox .h files, like codecs.h or codeclib.h) in any files where you test ROCKBOX_BIG_ENDIAN.
17:28:20pixelmajust put it on my Ondio... just started testing
17:28:40pixelmabut that shouldn't take too long
17:29:55amiconnlinuxstb: Shouldn't codecs see config.h automagically (because it's included from codecs.h)?
17:30:12linuxstbamiconn: Yes, but all .c files in a codec lib need codecs.h
17:30:31linuxstbBut yes, the solution is to add codecs.h to any files that do need it.
17:30:33amiconnYes, *but* ??
17:31:03linuxstbI'm not sure what you're saying...
17:31:13pixelmano... still the same "Panic: recfile -116" when it tries the write the data. That's with a patch applied to rev13774 and more available space on the MMC
17:31:26JdGordon:'( ok bugger
17:31:40linuxstbamiconn: Sorry, I meant to say "not all .c files" need it...
17:32:16Crash91People, i am making a rockbox backdrops site, does anyone want to help me??
17:32:24JdGordonpixelma: this is with the mmc inserted right?
17:32:43JdGordonjust making sure... you cleared the setting first then set it to the mmc path?
17:33:17pixelmahmm... didn't clear the settings
17:36:36pixelmanow I cleared the recording directory first, then set a new one (which is on the MMC) - no change
17:36:38saratogalinuxstb: heres the patch with that fixed
17:36:49 Join tchan1 [0] (
17:37:10saratogai've begun merging ffmpeg changes, so I can't test it, but all I did was add 1 line to the diff and fixed the line count, so it should work
17:37:18saratogaat least i hope
17:37:44 Quit tchan (Connection timed out)
17:37:48linuxstbYou can have multiple local copies of SVN...
17:38:33pixelmaJdGordon: did you try that patch with a dir on a microSD in your Sansa - shouldn't it behave similarly? (Only if hwcodec or swcodec recording doesn't matter here though)
17:38:48JdGordonpixelma: what about a path not on the mmc?
17:38:53JdGordonand yes, works fine on the sansa
17:39:00JdGordonseems to be a firmware hwcodec issue
17:40:49pixelmathe check if the directory exists is now executed when entering the recording screen - there could be differences because on swcodec it has to load the codecs (guessing though)
17:41:07pixelmafirst part should have been a question
17:42:33JdGordonyes, the check happens before any recording code gets started
17:44:23pixelmawhat do you mean with a path not on the MMC?
17:44:51JdGordoninstead of /<MMC1>
17:45:16Esp_linuxstb: The ipod is working again. after a few resets in the diagose menu and a iTunes recovery. The flash scan wouldnt budge from 0% at first but after a few resets it finally started counting up
17:48:00pixelmaJdGordon: will try but have to free a few MB on the internal memory first
17:48:39 Quit XavierGr (Nick collision from services.)
17:48:42 Join XavierGr [0] (
17:49:27 Join TrueJournals [0] (
17:50:11 Join bonbonthejon [0] (
17:50:52pixelmaJdGordon: same thing...
17:50:59TrueJournalsWhat version of gcc do I need in order to run the cross compiler? I keep getting "configure: error: installation or configuration problem: C compiler cannot create executables."
17:51:01bonbonthejonRockbox is complaining about not being able to access the playlist file
17:51:16TrueJournals(I have gcc 4.1.2)
17:51:29gregjello folks
17:51:34alienbiker99oh wow wma was commited
17:51:48gregjTrueJournals: I use 4.1.2 too, and it runs okay here
17:51:52pixelmaah... the other H10 owner here I know of
17:52:02gregjalienbiker99: initial support, I wouldn't rely on it
17:52:28TrueJournalsgregj: Hmmm... then am I doing something wrong or what?
17:52:30*JdGordon wishes he had a ondio :'(
17:53:06gregjTrueJournals: try fsck.vfat the system
17:53:18gregjI don't think it is compiler related
17:53:24alienbiker99yeah, i dont use wma's, but its a good commit
17:53:35gregjarm-elf-gcc (GCC) 4.1.2
17:53:44gregjwma is shaite in terms of quality
17:53:48gregjbelow even mp3
17:55:06JdGordonpixelma: well, I have no idea why its doing that... and unless you want to spend hours trying to figure out where it stops with some gui_syncsplash()'s i guess the patch will sit again
17:55:39gregjbtw, before I start any hacking
17:55:49gregjhow do I get verified ? any devels here ?
17:57:08JdGordonpixelma: can you apply and see what filename it says its trying to open please?
17:57:35peturgregj: your code goes in the tracker, devs review and commit
17:57:53pixelmaJdGordon: additionally, right?
17:58:03gregjnope, I am not talking about it
17:58:15gregjI know someone is going to review it first that's the way it works :)
17:58:18peturoh, irc
17:58:30petur /msg nickserv help
17:58:31gregj(I am KDE devel, and few other projects)
17:58:34gregjpetur: nope :)
17:58:45dionoeagnome ftw!
17:58:46*petur gives up
17:58:51gregjpetur: someone said yesterday, that even before commiting anything - I have to get "real life veryfied"
17:58:54gregjdionoea: die!
17:59:06*gregj is on gnome and mac os x now :)
17:59:07bonbonthejonyay KDE
17:59:08*Crash91 agrees with dionoea
17:59:09JdGordongregj: we dont accept anonymous contributions...
17:59:31JdGordonso as long as your patches have _a_ real name on it thats all the verification it needs
17:59:33 Quit maffe (Remote closed the connection)
17:59:49gregjJdGordon: ie, if I say - I am GJ - by god blessing I give thy patch - pass it on
18:00:08 Quit [IDC]Dragon ("CGI:IRC (EOF)")
18:00:10 Join maffe [0] (
18:00:22gregjI was affraid I have to fly to norway, and show up my passport to someone, or something :)
18:00:36peturnope, it's Sweden :p
18:00:37JdGordon... rockbox is "based" in sweden....
18:01:02gregj(which I wouldn't mind, but not necessarily now)
18:01:07gregjJdGordon: ok, sorry
18:01:24peturgregj: next years devcon then?
18:01:42gregjI met a nice Sweedish girl once, but she didn't left me any other contact than mobile number - tht was dead after one month :(
18:01:50gregjshe was cool :)
18:01:58gregjpetur: if I lst
18:02:11gregjie, won't give up
18:02:54 Part Crash91
18:03:07JdGordonpixelma: falling asleep here, if your keen, can you do a logf build and attach the debug log to the fs page? assuming of course it is written to disk before the panic
18:03:08 Quit bonbonthejon ("ChatZilla [Firefox]")
18:03:28gregjI will give up if someone will not give me a hand with that serial port thing on 5020
18:03:31TrueJournalsNo dice :'(
18:03:49Arathispixelma: on my h10 rockbox runs fine with the latest svn
18:04:04 Quit XavierGr (Nick collision from services.)
18:04:07 Join XavierGr [0] (
18:04:14*petur hands gregj logf and splashes....
18:04:23 Join webguest80 [0] (i=18cfadbe@gateway/web/cgi-irc/
18:04:29pixelmaJdGordon: yes, will report back on fs (don't know about the logf though)
18:04:35gregjpetur: what's that
18:04:51 Quit webguest80 (Client Quit)
18:04:51peturlogging and splashscreens
18:04:57peturfor debugging
18:04:58 Quit jhMikeS (Nick collision from services.)
18:05:03peturand the sim of sourse
18:05:04 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
18:05:09gregjoh, so there is no logging stuff there ? badness
18:05:17gregjcurse :)
18:05:21gregjcurse of no logs :)
18:05:58peturyes you can log to mem and then write to disk, build a logf build
18:06:03pixelmaArathis: good news :)
18:06:21Arathisand wma support works :D
18:06:45gregjpetur: oh, I thoguht you say - 'for starters, you can write logging module'
18:07:02pixelmaArathis: frequency scaling should give you a bit more battery life too
18:07:29gregjwhen it works
18:07:48JdGordonpixelma: one more before bed.. what happens if you try recording to the mmc when it it not inserted?
18:08:02Arathispixelma: is that an option or a code fix?
18:08:26pixelmacode fix
18:10:08Arathiswill make a bat-benchmark :)
18:10:32pixelmaJdGordon: gives a splash "can't write to recording directory - off = abort"
18:10:46Arathisbut wma playback slows down the player.
18:11:35pixelmaJdGordon: that's when I try to enter the recording screen, of course
18:12:47JdGordonok bloody hell... the path is getting mangled...
18:12:51JdGordondoes the ondio have an rtc?
18:12:58peturArathis: the codec is very new and needs optimizing, it needs a lot of cpu right now
18:13:44ArathisI know. Just wanted to mention it ;)
18:13:51gregjI just updated rockbox on my computer
18:14:00gregjit never goes beyond splash screen
18:14:43JdGordonpixelma: please
18:15:36JdGordonbah, no missed one...
18:15:40gregjwow, it is just terribly slow
18:16:22pixelmaJdGordon: should I apply that one on top of the other or replacing it
18:16:47JdGordonrevert firmare/mpeg.c first I think
18:16:50JdGordonwont really make a diff
18:17:12 Join My_Sic [0] (
18:17:28 Quit petur ("work->home")
18:17:31*gregj feels like talking to /dev/null
18:17:52TrueJournalsYay! I installed glibc and that fixed it :-D
18:18:40 Quit Nibbier ("bei uns hat noch niemals nicht koana koan durscht ned ghabt")
18:19:20 Join Siltaar [0] (
18:20:25Nico_Pgregj: what's your problem ?
18:20:33Nico_Pit's slow ?
18:21:00 Quit My_Sic (Client Quit)
18:22:06 Quit Thundercloud (Remote closed the connection)
18:23:18pixelmaJdGordon: I can't patch the pastebin'd one... or I'm doing something wrong
18:24:21 Join n1s [0] (
18:25:13LearJdGordon: Why reject a name that starts or ends with a space?
18:25:44JdGordonthats the standard apparently
18:28:34pixelmaI get "Index: firmware/mpeg.c
18:28:35pixelma−−- firmware/mpeg.c (revision 13770)
18:28:35DBUGEnqueued KICK pixelma
18:28:35pixelma+++ firmware/mpeg.c (working copy)
18:28:35pixelma@@ -2263,11 +2263,13 @@
18:28:36***Alert Mode level 1
18:28:36pixelma void audio_record(const char *filename)
18:28:38pixelma {
18:28:40pixelma mpeg_errno = 0;
18:28:42pixelma+ if (filename)
18:28:44pixelma+ {
18:28:51pixelmaclip board mishap :(
18:28:59Nico_Ppondlife: here ?
18:29:25 Quit freqmod (Remote closed the connection)
18:29:52pondlifeNico_P: kind of
18:30:05Nico_Ppondlife: seen the buffering api code ?
18:30:09pixelmaJdGordon: what I wanted to paste is that I get "patch unexpectedly ends in middle of line" for both hunks in the pastebin code
18:30:26pondlifeNico_P: Not the latest stuff, but I hear it's working
18:30:36LearJdGordon: From what I can see, leading and trailing spaces are "ignored", whatever that means...
18:30:46Arathispixelma: with the current build rockbox won't build a tag-cache on my player :(
18:30:59Nico_Ppondlife: yes it is :)
18:31:11Nico_Pthere still is work left to be done of course
18:31:25Nico_PI mainly wanted to know how you liked it
18:31:33pondlifeWell, I like it working!
18:31:45Nico_Pbut the threading will have to be done in rockbox
18:32:31CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
18:32:31*gregj will repeat,
18:32:42gregjcurrent build on G4, is sluggish - slow
18:32:46gregjany ideas ?
18:32:47Nico_PI'd like to do as much as possible in the standalone app
18:33:24Nico_Pgregj: that might be related to amicon's latest changes... have you tried a slightly older build ?
18:33:42pixelmaArathis: hmm weird... I heard that it also doesn't work on sansas
18:33:43Nico_PJdGordon: any idea what I could try implementing next in the app ?
18:34:03gregjNico_P: I will try to revert that change, hold on
18:34:13Nico_Pmaybe the connection between an audio file and its metadata
18:34:22Nico_Pgregj: just download a build from 2 days ago
18:34:36gregjNico_P: tht must have been something else, I recompiled with clean build dir
18:34:36JdGordonNico_P: yeah that.. or umm... i dunno...
18:34:43pondlifeNico_P: Best you see if you can get some of this into Rockbox I think.... if you wait for me then you may be waiting a while.
18:35:08Nico_Phmm got to go for a bit bbs
18:35:11gregjNico_P: builds are for sissies
18:35:31 Join Nibbier [0] (
18:36:10gregjreal man compiles from sources
18:36:57gregjthe wmv's here won't play :/
18:37:07gregjbut those are vids, so I gues that must be the problem
18:37:19Arathispixelma: and I get an error when loading a file regarding playlist controll file (-116) although playback starts normaly. in addition I can't save my config. rockbox sets is as default, but I can't save it in a seperate file
18:38:37***Alert Mode OFF
18:39:29 Nick tchan1 is now known as tchan (
18:39:31Arathisand now it seems it's freezed ...
18:40:46 Quit digin4 ()
18:42:14 Quit TrueJournals ("Leaving.")
18:42:57pixelmaJdGordon: additional info in the "Panic" splash.. "recfile: /<MMC>/recordings/rec_0001.mp3", looks like the correct path though...)
18:43:15JdGordonshouldnt it be /<MMC1>/ ?
18:43:38pixelmayes... it is. It was a typo
18:44:27JdGordon:( ok i give up for tonight
18:45:14 Quit JdGordon ("Konversation terminated!")
18:46:39 Quit pondlife ("disconnected has pondlife")
18:47:10 Join low_light [0] (i=c730190b@gateway/web/cgi-irc/
18:47:48gregjhow can I contact the mpeg guy ?
18:48:11gregjI have little mpeg file, that either drops the sound after one minute, or produces all sorts of garbage sounds
18:48:15 Quit bdgraue (Read error: 104 (Connection reset by peer))
18:48:21 Join bdgraue [0] (
18:48:31Learpixelma: I see why that name causes problems: it contains a '.'. Yes, the dot. :) I'll fix soon.
18:48:38 Join lukaswayne9 [0] (
18:49:31pixelmaLear: you mean the ".mp3" - shouldn't that happen to JdGordon too?
18:49:54LearSure should...
18:49:56gregjdoes he show up on irc from time to time ?
18:50:24pixelmaLear: maybe he didn't test it after his own commit though...
18:50:40gregjand the mpeg player stopped responding to key events
18:52:13n1sgregj: mpegplayerguy = linuxstb
18:52:35gregjI sometimes feeling here like no one listeins
18:53:43 Quit lukaswayne9 ("Ex-Chat")
18:54:01pixelmaArathis: are you around for testing a bit - amiconn couldn't (let) test the changes on H10 before because there was no H10 owner around
18:54:16gregjlinuxstb: pik
18:54:44alienbiker99what are you testing on the H10?
18:54:47gregjlinuxstb: are you interested in about first 5MB of that file to check it please ?
18:54:55Arathispixelma: was that a question? :)
18:55:15pixelmaactually yes :)
18:55:31richcoosa19does anyone else get "Error accessing playlist control file (-116)" when playing an mp3 on a Sansa e200?
18:56:20pixelmaalienbiker99: I don't test but I mean the changes this night - do you also have an H10
18:56:21LearMy guess is that it isn't Sansa-specific...
18:56:42alienbiker99yes i have one, i just dont know what i would be looking for, thats all
18:56:43richcoosa19I tried it both on SVN and on the latest build and I get the error on both.
18:58:00pixelmaalienbiker99: the question was if it still runs stable - the changes enabled frequency scaling which made it crash regularly before (which is why it was disabled)
18:58:04richcoosa19Also the Database is not building
18:59:16pixelmaLear: then that has to do with the '.' in ".rockbox/" too, I guess
18:59:55Learpixelma: Rather, the "." in ".playlist_control".
18:59:59alienbiker99ok let me get it and put the build on it, and ill try it
19:00:47Arathisrichcoosa19: same things on my h10. db not building, playlist error. plus at least one freeze
19:00:59pixelmaLear: I just thought about ".rockbox" because of the people who can't build the database...
19:01:38 Quit billytwowilly (Remote closed the connection)
19:01:52LearDatabase files also contain a '.' in the name...
19:01:57 Join billytwowilly [0] (
19:02:17pixelmaaye... almost all files do
19:03:15alienbiker99booting up now
19:03:46LearFix checked in...
19:03:51pixelmaArathis: so database not building and the playlist control error would be unrelated to the change I meant
19:04:12***Saving seen data "./dancer.seen"
19:04:19alienbiker99hmmm db not building for me...
19:05:11pixelmaalienbiker99: yes, that was a different problem - the one that Lear just committed a fix for
19:05:22alienbiker99oh ok, gotta update then
19:05:34 Quit SirFunk (Read error: 101 (Network is unreachable))
19:07:30alienbiker99ill be back when its done building
19:07:40 Join ender` [0] (
19:09:19 Join thegeek_ [0] (
19:09:26 Quit thegeek (Read error: 104 (Connection reset by peer))
19:10:05richcoosa19Just resynced with SVN/Compiled and it's fixed! Thanks a million!
19:12:50LearGot a bit worried when I stopped playback here. Had the "bookmark created" splash remain on screen. Causing the file list to refresh fixed it though. :)
19:14:23 Quit kretende1 (Read error: 104 (Connection reset by peer))
19:14:37 Join ender1 [0] (
19:14:59 Quit obo ("KVIrc 3.2.6 Anomalies")
19:16:31 Join My_Sic [0] (
19:23:04 Join petur [0] (n=petur@rockbox/developer/petur)
19:25:17PaulJamhi, my h300 with the current svn build gives me a splash with "filetype array full" during startup. what does that mean?
19:25:46peturtoo many types in viewer config I think
19:25:56peturWMA probably
19:26:30peturthere's a define somewhere that sets the size of the filetype array
19:27:36peturPaulJam: is that SVN or unofficial? do you have extra types defined?
19:28:13PaulJami just downloaded it from
19:28:20bluebrotherwasn't that limitation changed a while ago?
19:29:03PaulJamand i haven't modified the viewers.config
19:29:23peturOK, Iḿ building here now and I'll check
19:29:37 Quit ender` (Read error: 110 (Connection timed out))
19:29:51*petur cursus ubuntu keyboard driver
19:31:26 Quit My_Sic (Read error: 104 (Connection reset by peer))
19:31:34peturyep, here too.... will increase the buffer
19:32:37 Join My_Sic [0] (
19:33:09*preglow curses size of wma decoder struct
19:38:03XavierGrpreglow: are you trying to optimize it?
19:38:14pixelmaLear: nice finding... that was really the problem that made the patch I was testing fail
19:38:17*preglow then sees the size of the output buffer and faints
19:38:25preglowXavierGr: trying a quick couple of hacks, yes
19:38:30 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
19:40:45XavierGrdoes one have to add something to add a plugin to the open with dialog?
19:42:50XavierGrok thanks I will make it a viewer then
19:42:58 Join ifhar [0] (i=57f29bd0@gateway/web/cgi-irc/
19:43:05peturXavierGr: new filetype?
19:43:25XavierGrno I just want to test the wma codec on test_codec plugin
19:43:40ifharhello. i need help with formatting my iaudio x5. can anyone help?
19:43:52peturoh ok... I was just increasing the array size of filetypes ;)
19:43:55n1sXavierGr: just add it to SOURCES in plugins
19:44:05preglowlinuxstb_: there?
19:44:08XavierGrn1s yes I already did that to compile it
19:44:14n1sit's already in wievers
19:45:00XavierGrhmm I just want the test_codec rock to appear on the open with context menu
19:45:35n1sXavierGr: well, it really should if you use the viewers.config file that's in svn...
19:47:52 Join bdgraue_ [0] (
19:48:34 Quit bdgraue (Success)
19:49:03richcoosa19ne1 heard anything on the problem with the LCD on the newer e200's? two pixel horizontal line
19:49:53 Quit linuxstb (Read error: 113 (No route to host))
19:50:10peturhey, our build table still hasn't got that fancy countdown... :) *pokes Bagder*
19:52:21peturPaulJam: should be fixed now...
19:53:24 Join bootloader [0] (i=474872e4@gateway/web/cgi-irc/
19:53:33*preglow hears his first wma in rockbox
19:54:12preglowalso his only wma
19:54:14saratogapreglow: what target?
19:54:17preglowsaratoga: h120
19:54:21preglowhaven't got my nano here right now
19:54:28saratogadid you add my last patch?
19:54:39preglowcan't say i did, it's straigt svn
19:54:42preglowanything important?
19:54:55saratogai accidently left a float in there that only compiles on Coldfire
19:55:06saratogano idea if it got optimized out or not
19:55:12preglowreally, i'm timing this bastard now
19:55:20preglowdoesn't look too fast, heh
19:55:30preglowi only have a huge wma file to test with
19:55:34PaulJambtw, is the problem with the disk loosing its label still there?
19:56:05bootloaderhow do i install the EvilG-Fusion WPS patch right? Just extract to root of the ipod 30Gb Video? And then how do i make a Theme I downloaded work?
19:56:18saratogapreglow: check the "fixmul32" function in wmadeci.c
19:56:32saratogatheres a float mul in there for debug reasons
19:56:43saratogahopefulyl optimized out, but worth deleteing to be sure
19:56:54preglowsaratoga: oh, i've replaced that
19:57:05 Quit ifhar ("CGI:IRC (Ping timeout)")
19:57:17saratogahmm, my patch moved the fixed point stuff into seperate files
19:57:24saratogai'm guessing we're now out of sync
19:57:50preglowdoesn't matter, it's just out of curiosity
19:57:57peturPaulJam: I haven't seen that lately, but as there hasen't been a bugfix, it must still be there
19:59:50markunbootloader: look here
20:00:33markunbootloader: and for info in general about compiling rockbox:
20:01:08bootloaderok thank you i seem to be really messing up when i try to do WPS i can install rockbox but anything after i just mess up lol
20:01:16preglowsaratoga: did you and linuxstb use the same test file? got it around somewhere?
20:02:57bootloaderone more question why does rockbox not respond well to the click wheel? It goes slow and hardly ever recognizes when i press the select button
20:03:47 Quit Genre9mp3 ("I don't suffer from Rockbox psychosis. I enjoy every minute of it.")
20:06:15 Join Thundercloud [0] (
20:07:34bootloaderhow do i install the fonts package to an ipod just extract to root after I have installed rockbox using ipodpatcher?
20:08:06 Quit My_Sic (Read error: 110 (Connection timed out))
20:08:44preglowsaratoga: currently i've asm optimized both fixmul32 macros, speed testing a small 64 kbps file now
20:08:55preglowsaratoga: CMUL is a very obvious candidate for asm too
20:10:00preglowsaratoga: looks like the entire need for fixmul32b vanishes if we optimise CMUL
20:10:39bootloaderdo the fonts need to go into the .rockbox folder or just into root?
20:11:34PaulJambootloader: they go to .rockbox/fonts/
20:13:31bootloaderoh well ok then where does the WPS go into rockbox/wps? then themes to rockbox/themes?
20:13:43preglowsaratoga: hmm, with these two optimisations, my file seems to be 90% realtime
20:18:15 Join webguest42 [0] (i=57f29bd0@gateway/web/cgi-irc/
20:18:22 Quit webguest42 (Client Quit)
20:18:28 Join ifhar [0] (i=57f29bd0@gateway/web/cgi-irc/
20:19:48bootloaderthere is no rockbox/fonts where do they go?
20:20:39n1sbootloader: if there isn't a .rockbox/fonts folder, just create one
20:21:12linuxstb_preglow: I'm around now.
20:21:15 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
20:22:04preglowtest file, please
20:22:06preglowthe one you used
20:22:44linuxstbpreglow: No, we didn't use the same test track. I found a few on Usenet I've been using, as well as some webcasts I captured, and some of the mplayer collection here -
20:23:04linuxstbThe main one I was using was "Bangles - Walk Like An Egyptian.wma" from mplayer ;0
20:23:28bootloaderdo i have to run the ipodpatcher after i extract all of these?
20:23:31preglowthat's the 80 kbps one?
20:23:32linuxstbThat's the one I mentioned speed test results for.
20:23:41preglowk, i'll try that out, then
20:24:15n1sbootloader: try reading a bit in the manual
20:24:32linuxstbgregj: I used to be the "mpegplayer guy" (I did most of the early work on it). Then jhMikeS worked on it so I don't recognise it anymore. I'm not sure if he's still interested in doing more work on it though.
20:24:54preglowokiedoke, testing it now
20:25:20preglowwith a bit of luck, it's a wee bit closer to realtime now
20:25:50 Quit bootloader ("CGI:IRC (EOF)")
20:27:40 Join Rondom [0] (
20:27:53 Join bdgraue__ [0] (
20:28:29preglow95% realtime
20:28:38linuxstb\o/ not bad for 5 minutes hacking...
20:28:55 Quit bdgraue_ (Read error: 104 (Connection reset by peer))
20:28:57preglowthe usual deal
20:29:13preglowi just substituted with asm two places
20:29:52preglowmy only wma file plays realtime now!
20:29:54linuxstbUsing any IRAM yet?
20:30:04preglowabsolutely none
20:30:06linuxstbThen we're laughing...
20:30:20 Join luebecktricker_ [0] (
20:30:30linuxstbI'm wondering if any of those many buffers in the wmadecodecontext struct can be combined.
20:31:01preglowwho knows, i did want i wanted to do right now. just couldn't sit by not trying something so obvious, heh
20:31:28preglowi'll just brush it up a bit and commit it
20:32:27linuxstbCan saratoga's "fixed32" format be sent directly to the DSP layer via pcmbuf_insert?
20:33:11preglowis it converted right now?
20:33:23linuxstbYes, the codec returns 16-bit interleaved samples.
20:33:27preglowyeah, so i see
20:33:32preglowis it deinterleaved internally?
20:33:34linuxstbThat's high on my list to change.
20:33:43preglowdeinterleaved 32 bit == optimal
20:34:00linuxstbIt's decoded into the "frame_out" array in wmadecodecontext
20:34:17linuxstbSo yes, deinterleaved 32-bit internally.
20:35:05preglowthat's just fine and dandy, then
20:35:27preglowi didn't even need to write any code for this
20:35:34preglowripped it straight out of libmusepack and tremor
20:35:39preglowcopy&paste living
20:35:48linuxstbI'm sure the memory requirements can be reduced drastically - there seem to be lots of copying from one buffer to another, where things could be done in-place.
20:36:03preglowanywho, it's time for some grocery shopping, apparently
20:36:04 Quit luebecktricker_ ("Verlassend")
20:39:07 Join SirFunk [0] (
20:39:55 Join merbanan [0] (n=banan@
20:40:47linuxstbsaratoga: Around?
20:41:43 Join Themes [0] (i=474872e4@gateway/web/cgi-irc/
20:41:48linuxstbpetur: Regarding your filetypes commit, it was probably me adding wmv and asf - so two new extensions. wma itself was already there.
20:42:17 Join entheh [0] (
20:42:45Themesi have installed the "flow" theme into rockbox/themes but it does not show up in the browse themes menu what do i do?
20:43:35peturlinuxstb: I saw ;)
20:45:42ifharcan someone help??? i'm trying to format my iaudio x5 with rockbox
20:46:11 Quit thegeek_ (Read error: 104 (Connection reset by peer))
20:46:30linuxstbWhat's the problem?
20:47:13 Join thegeek [0] (
20:47:21 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
20:47:31Themescan someone help me with the themes?
20:47:47 Quit thegeek (Read error: 104 (Connection reset by peer))
20:47:49 Quit bdgraue__ (Read error: 104 (Connection reset by peer))
20:47:52 Join miepchen^schlaf [0] (
20:47:56 Join bdgraue__ [0] (
20:47:56 Join thegeek_ [0] (
20:48:12ifharlinuxstb: don't know how to do it.....
20:49:14linuxstbifhar: Have you found the installation chapter in the manual?
20:49:28PaulJamThemes: what do you mean by "installed"? most themes have already the correct folderstructure in the .zip file, that you just have to extract the .zip to the root of your player
20:50:47ifharcan't find the actual manual. thought all this would be online but can't seem to find it
20:51:53FebsClick the link that says "Manual." It's on every page of the site.
20:53:02ifharfound it but it doesn't say how to reformat the drive, which i wanna do coz playback is faulty
20:53:31FebsIt doesn't say that because reformatting the drive is not necessary for installing Rockbox.
20:54:17ifharno no, i already have rockbox, but i wanna format the device because it doesn't play files right
20:54:51 Join MrKeuner [0] (n=kudo@unaffiliated/mrkeuner)
20:55:19ifhari think it might be coz i dropped it once so i thought if i format the drive it might help
20:55:21 Join ompaul [0] (n=ompaul@freenode/staff/gnewsense.ompaul)
20:55:47MrKeunerhi, may rockbox have some kind of settings which might lead to practically lesser amount of battery time?
20:56:08Febsifhar: most issues can be resolved with remedies short of reformatting.
20:56:22MrKeunerFor example backlight turns of but the lcd does not go off when playing oggs in ipod nona
20:56:50 Quit Rob222241 ()
20:56:57FebsIn your case, I would try running utility that scans the disk and corrects errors.
20:57:09ifharFebs: sounds good, where do i find that?
20:57:20FebsWhat OS?
20:57:44ifharmac os x
20:58:38ThemesPaulJam: I meant i extracted the theme to the rockbox/theme folder is that wrong?
20:59:10peturMrKeuner: I think switching off the LCD will not solve that much battery power. iirc, we keep the lcd on for targets where you can read the display without backlight
20:59:11FebsThemes: extract the theme to the root of your player.
20:59:17linuxstbAnyone know the maximum number of bytes a utf8 character can be?
20:59:25PaulJamThemes: does the .zip contain a .rockbox folder?
20:59:40 Join pearldiver [0] (
20:59:50ThemesFebs: so should i first delete the one i extracted from the rockbox/theme folder?
21:00:01MrKeunerpetur: it is really convenient but, may be if it could be changed it would have been even better
21:00:05ThemesPaulJam: I will check
21:00:26pearldivercongrats with WMA, whoever was involved!
21:00:50FebsThemes: if the theme is configured properly (i.e., as PaulJam says), then you can just extract it to the root, and it will automatically put the files and folders where they need to be.
21:00:54MrKeunerpetur: I do not have data but I feel like I am having drastically lower battery time
21:01:15peturMrKeuner: compared to apple os?
21:01:22MrKeunerpetur: yes
21:01:24ThemesPaulJam: yes ot is all contained in a .rockbox folder, should i delete it and extract to root?
21:01:38peturMrKeuner: we know, and it's not the LCD
21:01:42FebsThemes: yes.
21:01:43Learlinuxstb: Four, I think. Could depend on actual UCS version...
21:01:44PaulJamThemes: yes
21:01:44ThemesFebs: yes PaulJam is right i just extracted wrong
21:02:02MrKeunerpetur: what is it if I may ask, as if I would understand
21:02:21peturMrKeuner: if we knew, it would have been fixed
21:02:24pearldiverhows the wma performance across the targets?
21:02:25linuxstbLear: I'm parsing ASF metadata, which is stored in utf16, and need to convert to utf8, and check I have enough buffer space for the result...
21:02:32Themesok im restarting my player now thanks for the help
21:03:36Learlinuxstb: According to Wikipedia, 4 is maximum, but it is rarely used. 3 is a more common upper limit.
21:03:44MrKeunerpetur: can I still boot into apple os I would like to compare the battery times
21:04:14peturMrKeuner: check the manual, I think you have to flip HOLD while booting
21:04:17***Saving seen data "./dancer.seen"
21:04:19linuxstbLear: I'll use four anyway, it's not going to make much difference.
21:04:24linuxstbmarkun: Are you around?
21:04:25Febsifhar: I don't know much about Macs, but Google found this for me:
21:04:30ThemesPaulJam: thank you it is working
21:04:39ThemesFebs: thnak you it is working
21:05:20MrKeunerpetur: that did it thanks
21:05:39markunlinuxstb: I am
21:06:01linuxstbmarkun: Are you responsible for unicode.c ? ;)
21:07:00preglowlinuxstb: six, i think
21:07:07preglowfor utf8
21:07:09linuxstbI need to use the utf16LEdecode function for the ASF metadata parser, and it doesn't return the number of bytes consumed in the utf16 stream.
21:07:23 Quit billytwowilly (Remote closed the connection)
21:07:37markunlinuxstb: doesn't it return a pointer?
21:07:40 Join billytwowilly [0] (
21:07:44markunlinuxstb: I'll check
21:07:58linuxstbmarkun: It returns a utf8 pointer - so you know the output size, but not how much of the input has been consumed.
21:08:19markunah, that's not nice
21:08:26linuxstbThe only place I can find that function being used is the id3v2 parser, which seems to assume 2 bytes.
21:08:46Learlinuxstb: But it always consumes count utf16 chars, does it not?
21:09:05linuxstbYes, but IIUC, a utf16 char can be 32 bits.
21:09:45linuxstbA so-called "surrogate pair" according to the comments in the function I'm reading.
21:09:46LearYes, but the code checks for that. (Assuming that's what the surrogate pair check thingy is. :) )
21:09:49markunLear: they added some hack to allow for more than 2^16 chars
21:10:03 Join obo [0] (
21:10:28linuxstbLear: The decoding function does, but that fact isn't returned to the calling code.
21:10:31 Quit ansivirus (Remote closed the connection)
21:10:42Themeswhen you convert the itunesDB to rockboxDB does it have to actually convert each filer or just reorginize?
21:11:05LearYes, it could actually consume count + 1 chars...
21:11:50preglowmarkun: it's not really a hack, it's always been part of utf-16. ucs-2 is the precursor and only allows 16 bit characters
21:13:06ifharthanks Febs
21:13:41 Join kaaloo [0] (
21:14:12linuxstbmarkun: My problem is that I know the size of the utf16 buffer in bytes, and I need to decode it into a limited size utf8 buffer (the id3v2/v1 buffers in the id3 struct). Maybe I should write a new function...
21:14:37 Join ansivirus [0] (
21:15:21markunlinuxstb: or replace it?
21:15:36markunpreglow: is vfat UCS-2
21:15:57linuxstbmarkun: I'm not sure what the id3v2 parser is doing...
21:16:04amiconn/me not preglow though ;)
21:16:35amiconnHmm, not sure whether surrogate pairs are allowed in recent implementations
21:17:20 Join Domonoky_ [0] (
21:17:23pixelmaArathis: tried again with a build after Lear's fix?
21:17:52 Part kaaloo
21:21:20Learlinuxstb: There are other unicode.c functions that could use a output buffer size argument, so that there won't be a buffer overflow.
21:21:20markunlinuxstb: or maybe we can give an extra argument with a pointer to the end of the buffer?
21:22:57amiconnIt could just return the changed input pointer, couldn't it?
21:23:04amiconn(would require a ** argument)
21:24:56 Quit The-Compiler (Remote closed the connection)
21:28:56 Part maffe
21:30:04 Join maffe [0] (
21:30:09*amiconn had an idea how we could save some power on PP targets in a CPUFREQ_MAX sleep() situation
21:30:20amiconnOnly if both cores are sleeping though
21:31:03 Quit MrKeuner ("Ex-Chat")
21:31:28 Quit ifhar ("CGI:IRC (EOF)")
21:33:00 Quit richcoosa19 ("[BX] The best part of waking up is BitchX in your cup")
21:34:53*n1s finds new binsize saving potential :-)
21:35:28 Join Wiwie [0] (
21:35:29markunn1s: many kbytes again?
21:35:31*amiconn finds it interesting that Lear's fix even saved more on binsize than JdGordon's wannabe-fix added
21:36:02 Join richcoosa19 [0] (n=richie@
21:36:09amiconn(on arm)
21:36:11n1smarkun: no, a couple of hundred bytes so far, but for both SH and coldfire, and probably arm too...
21:36:32preglowlinuxstb: mind testing the code?
21:36:37preglowon something coldfire, that is
21:37:37*pixelma would like to test her one wma file on M5 :)
21:37:42 Join amiconn_ [0] (n=jens@rockbox/developer/amiconn)
21:37:47preglowsure, please test using my latest commit
21:38:07 Quit amiconn (" HydraIRC -> <- Would you like to know more?")
21:38:23 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
21:38:40preglowwhat the hell, i was certain i fixed that bloody warning i'm about to get
21:39:17preglowoh, i did
21:40:26 Quit PaulJam (".")
21:40:52 Join TrueJournals [0] (
21:43:13ThemesWhen you convert the itunesDB into a rockboxDB does it convert the files or just reorginize them?
21:44:02Learpreglow: No iram at all yet?
21:44:10preglowabsolutely none
21:44:21preglowboth the usual suspects for iram use are bigger than iram
21:44:50 Quit andrew__ (Read error: 110 (Connection timed out))
21:45:55 Quit Themes ("CGI:IRC (EOF)")
21:47:15 Quit bdgraue__ (Remote closed the connection)
21:47:35 Join bdgraue__ [0] (
21:50:21 Join webguest86 [0] (i=97c98a9f@gateway/web/cgi-irc/
21:52:34 Quit miepchen^schlaf (Read error: 113 (No route to host))
21:52:55 Join miepchen^schlaf [0] (
21:53:50 Quit Rondom (No route to host)
21:57:54LearArgh, profile build causes an ICE...
21:59:39 Part webguest86
22:03:48 Quit pearldiver (Read error: 110 (Connection timed out))
22:04:55preglowLear: where?
22:04:57gregjlinuxstb: well, I don't have a week to dig into , but I can give you the first 5-6MB of that movie, so you can at least test it on your side, and tell me what you think, alright ?
22:05:04 Join bobj [0] (i=c66c9f3f@gateway/web/cgi-irc/
22:05:24LearTremor, bitwise.c
22:06:02 Quit bobj (Client Quit)
22:06:34n1sLear: maybe try lowering the O level, we build Tremor with O3 now...
22:06:59Learn1s: My thought exactly. Pretty sure that part worked before.
22:07:09 Join bdgraue [0] (
22:14:13 Quit bdgraue__ (No route to host)
22:14:46 Join amiconn_ [0] (n=jens@rockbox/developer/amiconn)
22:14:54 Quit amiconn ()
22:14:55pixelmapreglow: my 128kbps wma doesn't play realtime but it's not as bad as linuxstb's ape track, I tried (very personal first impression though)
22:15:00 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
22:16:51preglowpixelma: i wouldn't expect it to either, do you have test_codec handy?
22:17:59pixelmadoesn't have it handy - but like to compile it. Might be fun.
22:20:00 Join Quelsaruk [0] (
22:20:10pixelma*don't too
22:20:32 Quit TrueJournals ("Leaving.")
22:23:18 Quit low_light ("CGI:IRC")
22:24:12 Quit SirFunk (Remote closed the connection)
22:27:13 Join dubbae [0] (i=4cb431de@gateway/web/cgi-irc/
22:28:05 Join ComputinChuck [0] (
22:28:39 Join stripwax [0] (
22:29:26ComputinChuckhello, i just installed rockbox on my ipod video and it won't save any config files when i shut it down. also when i try to save the config in the "Manage Settings" menu, it says it "failed"
22:29:57 Join My_Sic [0] (
22:30:07dubbaeI'm using the current H10 5GB build - is the clock scaling feature used to overclock the device?
22:30:36scorchedubbae: no
22:30:38Domonoky_dubbae: not overclocking, underclocking :-)
22:30:41pixelmaComputinChuck: what build are you using (when did you download it)?
22:31:09ComputinChuckpixelma, i just downloaded it this morning
22:31:24amiconnMorning being when?
22:31:37amiconnThere is the concept of timezones, you know? ;)
22:31:52ComputinChuckhaha, yes
22:31:58amiconnOr simply tell the version...
22:32:08pixelmaComputinChuck: I guess you downloaded one that was broken...
22:32:10 Quit ompaul ("night")
22:32:13ComputinChuckmorning was about 3 hours ago
22:32:30dubbaeso the default clock speed settings for H10 5 GB would be Frequency: 78000000 and boost_counter 1, correct?
22:32:43 Join lukaswayne9 [0] (
22:32:54amiconnThe default is 30MHz, 78MHz is boosted
22:33:07dubbaeso isn't that overclocking?
22:33:17amiconnThe cpu is specced up to 80MHz
22:33:22pixelmaComputinChuck: that was about the time a fix was committed, I suggest to just try out the newest "current build"
22:34:16ComputinChucki just need to replace the .ipod file right? i don't need to install the boot loader and everything again
22:34:33scorcheComputinChuck: not the bootloader, but the whole folder you need to replace
22:34:44scorchejust extract over it and it should overwrite the files
22:34:45dubbaeI don't notice any real performance increase when the processor is boosted - is that normal?
22:34:47 Quit entheh ("^~")
22:35:50 Part n1s
22:36:33scorchedubbae: what kind of performance increase are you looking for?
22:36:47dubbaeRockboy runs the same
22:36:54dubbaeI figured maybe it would be faster
22:37:02 Join DogBoy [0] (
22:37:40scorcherockboy uses it fully boosted, i believe
22:38:08amiconndubbae: You are not supposed to change the boost counter by hand. This is debug stuff
22:38:27amiconnRockbox boosts/unboosts as necessary to keep up with its tasks
22:38:30dubbaeI thought it was some neat new overclock setting
22:38:35dubbaelike Palm Pilots used to have
22:38:49dubbaemy mistake
22:38:59amiconnAnd the goal is that you don't notice the changes too much
22:39:45amiconnThis is for saving power, not for overclocking. (you can overclock the cpu if you make your own build, but expect it to crash within a more or less short time)
22:40:02dubbaeand the battery would drain faster as well, I presume
22:40:31 Quit Arathis ("Bye, bye")
22:40:40dubbaeso basically - I should leave the debug setting at 30MHZ and not worry about it
22:41:00scorcheso basically, you shouldnt touch it at all and not worry about it, yes ;)
22:41:05ComputinChucki read somewhere that rockbox's usb charging ability was about 1/5 that of the Apple firmware's, is that true when the player is in hardware disk mode?
22:41:06dubbaelol, ok
22:41:16 Quit atsea- (Read error: 104 (Connection reset by peer))
22:41:26scorcheComputinChuck: disk mode is apple's
22:41:54dubbaeis any work being done on FM Radio for the H10?
22:42:05ComputinChuckok, so as long as it's in disk mode it'll charge just fine
22:42:17dubbaeI know someone mentioned the Rockbox firmware couldn't get power to the radio, or something to that effect
22:42:55pixelmapreglow: it's almost 92% realtime
22:44:11 Part maffe
22:44:25 Join maffe [0] (
22:45:21preglowthen it'll be zipping along once we optimise more and use iram
22:46:33preglowdamn, commited today and already it's almost realtime, i'll call that a success
22:46:46 Quit ComputinChuck ("Leaving")
22:46:54 Quit lukaswayne9 ("Ex-Chat")
22:47:04 Quit XavierGr (Nick collision from services.)
22:47:07 Join XavierGr [0] (
22:47:34alienbiker99that is impressive, but the work has been going on for a while no?
22:48:19preglowyeah, but it's only recently breakthroughs have been happening
22:48:30amiconnpreglow: Hopefully progress doesn't stop here :|
22:48:40*amiconn thinks about aac on cf
22:48:43preglowamiconn: i would be greatly surprised if that happened
22:49:01pixelmaooh... and it seems I missed an ape optimisation, test_codec gives 105% realtime even though "real" playback stutters now and then (wps without peakmeters too)
22:49:11saratogapreglow: I'm around now
22:49:24pixelmawith linuxstb's sample track
22:49:32saratogayou've left that trig table I put in IRAM there right?
22:49:47preglowoh, i didn't even notice it
22:49:59saratogathat made a pretty big difference
22:50:22saratogathe VLC stuff could probably use IRAM too, but I don't understand it well enough yet
22:50:31saratogaand some of it is really huge
22:51:05preglowmax block size is 2048 samples?
22:51:18DogBoydo I need to install podzilla first in order to get it dual booting with rockbox?
22:51:25saratogapreglow: yes
22:51:42saratoga(2048 or 4096 I guess depending how you look at it)
22:51:59preglowwell, per channel
22:52:36saratogawell its a 2048/4096 transform thanks to mdct weirdness
22:52:54preglowsure, just talking output buffer here
22:53:05preglowi want it in iram
22:53:08preglowand want to see if it's practical
22:53:25preglow2048 max size would mean 16k iram used for that, which is acceptable
22:53:59saratogalinuxstb mentioned this
22:54:07 Part dubbae
22:54:19saratogawhich output buffer?
22:54:28preglowthe final one you pass to dsp.c
22:54:37preglowit also seems data is converted to 16 bits, that is unnecessary
22:55:27saratogathat can contain more then one block though
22:55:35preglowyeah, wma_decode_frame converts to 16 bits, i see
22:56:08 Quit Domonoky_ (Read error: 110 (Connection timed out))
22:56:20saratogalinuxstb said each superframe could have up to 10 frames, each of which could be 2048 samples, at least if I understood him correctly
22:56:25preglowalso it interleaves, we don't want it to do that. with these changes, i bet the codec'll be usable realtime on both cf and arm
22:56:29saratogaI've only seen much smaller blocks
22:56:35saratogaerr frames
22:56:48preglowyeah, but it looks to me like not decoding a whole superframe at once will be trivial
22:57:26saratogayeah shouldn't be too bad
22:57:31saratogajust making sure you're aware of that
22:57:45preglowsure, i am
22:57:48saratogahave you profiled the codec yet?
22:58:09preglowi'm quite busy with other work i need to have done before i go on vacation
22:58:55saratogahow involved is it?
23:02:13saratogayes, the wiki page is rather vague
23:02:24preglownot very much at all, though i can't remember if it's easily doable for arm targets. you essentially just run the codec as normal with a profiling build, then post-process a file to have results
23:02:29amiconnpreglow: Btw, you can optimise away the need for a 2nd temp in fixmul32() for coldfire
23:02:42amiconn...without needing additional cycles
23:03:08amiconnOr rather, saving one cycle...
23:03:16preglowamiconn: very possible, i just took that code from the libmusepack source. afaik, it's yours
23:03:49amiconnCould be; don't really remember
23:04:19***Saving seen data "./dancer.seen"
23:04:33preglowbtw, how? by reusing y?
23:05:34amiconnthe asl.l can be split in 2 shifts with literal shift count (simple)
23:05:44preglowahh, yeah
23:05:47preglowbut would it be worth it?
23:06:09preglowduh, of course it would
23:06:10amiconnAnd the moveq.l / lsr.l / or.l sequence can be shortened
23:06:18amiconn..because it's a shift by 16
23:06:45amiconnswap %[x] / move.w %[x], %[t1]
23:07:44*amiconn will try that
23:07:57preglowahh, beh
23:08:12preglowi keep forgetting the partial reg tricks
23:08:31*amiconn loves this cf trick :)
23:08:42amiconnm68k cisc heritage....
23:08:54*preglow hugs cisc
23:08:59 Part maffe
23:09:28 Join maffe [0] (
23:09:30*amiconn goes hunting for a wma test track
23:09:52preglowlinuxstb pasted an url for some
23:10:38saratogathose are the ones I used
23:11:13*ender1 yawns
23:12:05preglowsaratoga: but yeah, this'll simplify midterm evaluation tons
23:12:39preglowseeing as evaluating something in working condition is pretty simple, heh
23:13:22saratogasounds good
23:14:19preglowbut yeah, i think one of your priorities should be merging latest ffmpeg svn
23:14:40preglowsince getting low-bitrate files working would be pretty important
23:14:41ender1is there something wrong with my build server - it hasn't been used in a while
23:14:48preglowit smells bad
23:15:29peturender1: talk to Bagder
23:15:37 Nick ender1 is now known as ender` (
23:15:59BagderI believe I disabled it a while ago when it acted funny... I'll re-enable it again
23:16:29ender`funny? what was wrong?
23:18:10 Join Guile` [0] (
23:19:49 Quit bdgraue (
23:19:49 Quit billytwowilly (
23:19:49 Quit obo (
23:19:49 Quit linuxstb (
23:19:49 Quit petur (
23:19:49 Quit Siltaar (
23:19:49 Quit jhMikeS (
23:19:49 Quit joa (
23:19:49 Quit Nico_P (
23:19:49 Quit desowin (
23:19:49 Quit GodEater (
23:19:49 Quit spiorf (
23:19:49 Quit austriancoder (
23:19:49 Quit kfazz (
23:19:49 Quit Guile`` (
23:19:49 Quit jhulst_ (
23:19:49 Quit chrisjs169 (
23:19:49 Quit RaRe (
23:19:49 Quit darkless (
23:19:49 Quit bospaadje (
23:19:49 Quit Lynx_ (
23:19:49 Quit DogBoy (
23:19:49 Quit goffa (
23:19:49 Quit alberink_ (
23:19:49 Quit Farp (
23:19:49 Quit blithe (
23:19:49 Quit GodEater_ (
23:19:49 Quit Bagder (
23:20:06NJoinDogBoy [0] (
23:20:06NJoinbdgraue [0] (
23:20:06NJoinobo [0] (
23:20:06NJoinbillytwowilly [0] (
23:20:06NJoinlinuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
23:20:06NJoinpetur [0] (n=petur@rockbox/developer/petur)
23:20:06NJoinSiltaar [0] (
23:20:06NJoinjhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
23:20:06NJoinjoa [0] (n=joe@
23:20:06NJoinNico_P [0] (
23:20:06 Join desowin [0] (n=desowin@unaffiliated/desowin)
23:20:06NJoinGodEater [0] (
23:20:06NJoinspiorf [0] (
23:20:06NJoinaustriancoder [0] (n=austrian@
23:20:06NJoinkfazz [0] (
23:20:06NJoinGuile`` [0] (
23:20:06NJoinchrisjs169 [0] (n=jack@unaffiliated/chrisjs169)
23:20:06NJoinjhulst_ [0] (
23:20:06NJoinRaRe [0] (
23:20:06NJoindarkless [0] (
23:20:06NJoinbospaadje [0] (
23:20:06NJoinLynx_ [0] (
23:20:06NJoinBagder [0] (n=daniel@rockbox/developer/bagder)
23:20:06NJoinFarp [0] (
23:20:06NJoinalberink_ [0] (
23:20:06NJoinGodEater_ [0] (
23:20:06NJoingoffa [0] (n=goffa@
23:20:06NJoinblithe [0] (n=blithe@
23:20:06***Server message 505: 'logbot_ :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
23:25:39 Join borisyeltsin [0] (
23:25:47 Quit billytwowilly (Read error: 104 (Connection reset by peer))
23:25:48 Join lazka_ [0] (
23:26:03 Join acb5649 [0] (i=4a841a2a@gateway/web/cgi-irc/
23:26:21amiconnCodecs aren't relinked if only the lib changes
23:27:05preglowthe codec build bug annoys the hell out of me
23:27:46LearAnd on sim builds they always relink...
23:28:17acb5649may i please have more permission in flyspray? close tasks and assign?
23:29:35amiconnpreglow: The opt saves 164 bytes for wma.codec on H1x0
23:29:38acb5649nd view reports?
23:29:41amiconnTesting speed now...
23:30:00preglowamiconn: and a couple of cycles too, but i'd be surprised if it's measurable
23:30:29 Quit Wiwie ("Miranda IM! Smaller, Faster, Easier.")
23:30:31amiconnwell, one cycle per fixmul32(), plus potentially some register shuffling
23:31:01 Quit richcoosa19 (Remote closed the connection)
23:31:37amiconnThe coldfire fixmul32() assumes PRECISION == 16
23:32:07 Quit jgarvey ("Leaving")
23:32:48 Quit joa (Read error: 104 (Connection reset by peer))
23:33:51acb5649i got it to work!!
23:34:55 Quit acb5649 ("CGI:IRC (EOF)")
23:35:02 Join BK-- [0] (
23:35:11amiconnhmm :/
23:35:33amiconnIt got slower....
23:35:47saratogaquestion: why I should be getting a profile.out in the players root when I do a profile build right?
23:35:56amiconnLess than a percent though
23:35:58saratogai'm pretty sure its profiling, the codec got slow as hell
23:36:02saratogabut no log file
23:36:27 Quit Guile`` (Connection timed out)
23:38:44preglowamiconn: sounds like a coincidence to me
23:39:30 Quit lazka (Read error: 113 (No route to host))
23:42:29 Part maffe
23:43:16 Join freqmod [0] (
23:44:50 Join maffe [0] (
23:47:07Learsaratoga: Added calls to start and stop profiling?
23:49:53 Join mario_ [0] (
23:50:11mario_hey, how do I find Rockboy on my ipod
23:50:34scorchemario_: it was explained to you in #ipodlinux...
23:50:51saratogaLear: yes I have
23:51:00mario_I can't find it on my ipod what is it under
23:51:05mario_what heading
23:51:27saratogaci->profile_thread(); correct?
23:51:32LearThen you should get a profile log, after the stop has been executed.
23:51:40saratogaand then ci->profstop();
23:51:43amiconnpreglow: As asl.l does exactly the same as lsl.l, another instruction can be shaved off :)
23:51:47LearDon't remember really. Was some time ago I did it last time...
23:52:02scorche[13:22:48] <rvvs89>
23:52:02scorche[13:25:06] <mario_> I searched through the plugin but I couldn't find rockboy
23:52:02scorche[13:25:24] <DerPapst> it's a viewer
23:52:02DBUGEnqueued KICK scorche
23:52:02scorche[13:25:37] <DerPapst> to select a rom from the filebrowser
23:52:03scorche[13:25:43] <DerPapst> s/to/you
23:52:04***Alert Mode level 1
23:52:04scorche[13:25:58] <rvvs89> "To start a game open a ROM file saved as .gb or .gbc in the file browser. "
23:52:21saratogaits odd because it becomes very slow, so i imagine its profiling, but then I don't get the log file written out
23:52:34 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
23:54:07Learsaratoga: But the functions sound right.
23:54:12mario_where is the file browser? is it a folder
23:54:41scorchethe file browser is what is used to browse your files, oddly enough
23:55:10amiconnpreglow: :)
23:55:12mario_ok I put them in a folder i made
23:55:21mario_but cant find them
23:55:29mario_will they be in my database
23:55:37scorchechange your viewmode to something besides music
23:56:32preglowamiconn: hahaha
23:56:44amiconnWHat's so funny about it??
23:56:44preglownow let's hope saratoga doesn't change the format
23:56:58preglowjust the size of it
23:57:43saratogayou mean the fixed format?
23:57:47preglowsaratoga: yeah
23:57:54preglowamiconn: did you check which of the arguments would be the best to clobber?
23:57:58saratogano way I'll change that
23:58:04 Quit davina (Remote closed the connection)
23:58:27amiconnpreglow: I tried clobbering y as well. Got bigger and slower (a tiny bit)
23:58:44amiconnThough that was for the first version
23:58:48preglowok, good

Previous day | Next day