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 2007-10-16

00:02:32 Quit Bagder_ (Read error: 110 (Connection timed out))
00:02:38WalterEgomarkun - You gotta love a greetings that means something like "did you eat well today?" ... :D
00:03:26markunWalterEgo: you are refering to a greeting common in beijing, right?
00:04:32 Quit midkay ("Leaving")
00:04:37WalterEgoWell, I know nothing of Mandarin but I've been told that's about what 'ni hao' meant.. With the 'ma?' being the more intimate version.. But I may be totally mistaken. :p
00:05:41markunWalterEgo: "ni chi le ma" is a way of greeting which literally means "have you eaten?"
00:05:56markunni hao ma = "you good?"
00:06:40 Quit davina (Remote closed the connection)
00:06:41WalterEgoOh well thanks for putting that straight. :)
00:07:07markunI just started learning the basics myself :)
00:07:12markunbut now: sleep
00:07:28WalterEgoget as much as you can, sleep is invaluable. :p
00:07:46 Quit kkurbjun ("leaving")
00:09:14 Quit _Bagder_ (Read error: 110 (Connection timed out))
00:11:24 Quit bertrik ("bye")
00:12:44 Quit bluebrother ("leaving")
00:14:22pixelmaargh, just the moment I wanted to ask bluebrother a question
00:19:30 Join ddalton [0] (
00:19:58 Part ddalton
00:20:40 Quit WalterEgo ()
00:28:17*preglow tries to think of how to make a decent limiter without too much lookahed :/
00:28:46Davide-NYCohh ohh what for you thinking about making a limiter?
00:28:54*Davide-NYC excited
00:30:37preglowDavide-NYC: mainly to be able to listen to very silent stuff, like .spc files
00:31:03Davide-NYCoh, not a live limiter for recording
00:31:05preglowlimiter is probably the wrong word, thinking more about automatic gain control
00:31:10preglowdunno why i said limiter :/
00:34:29LloreanIt'd be nice to figure out some way to level out SPC files to match replaygained music
00:35:31 Quit ]RowaN[ (Read error: 110 (Connection timed out))
00:35:38preglowthat'd require realtime replaygain processing...
00:35:44Davide-NYCI gotta run. amiconn, I'm going to test while on the train. If I discover anything I'll let you know.
00:35:48preglownot impossible in any way, btw, but don't know i care
00:35:53 Quit Toxicity999 (Read error: 104 (Connection reset by peer))
00:35:56Lloreanpreglow: I'm not talking about replaygain specifically
00:36:04 Quit Davide-NYC ("ChatZilla [Firefox]")
00:36:24LloreanJust like, for formats like SPC, where most of the songs are the same amount of "quiet", the ability to add a pre-amp by-format or something.
00:36:40preglowi'm thinking agc so i can just enable it and stop caring
00:37:08preglowtracks that have silent parts and louder parts will suffer, but i can't think of any better solutions
00:40:30preglowi'm thinking a simple thing that smoothly adjusts the volume to the desired level, while quickly limiting the gain factors if clipping is detected on the output
00:40:45preglowthat last might be tricky since clipping doesn't happen until late in the dsp chain, though
00:41:22preglowhmm, no, that should be doable
00:42:28*Llorean wonders if Foobar can write SPC metadata.
00:42:44preglowi really don't want to use any lookahead to smoothly vary volume if clipping will happen since that'll need a largish buffer which will need runtime allocation
00:42:55preglowthe spc format DOES have a gain factor in the metadata, but i don't tihnk too much uses it
00:43:06preglowand i don't want to require people to preprocess their files for use
00:43:21LloreanI was just thinking for my own use, there.
00:43:26jhMikeSpreglow: you're saying SPC rippers don't set it?
00:43:35LloreanYour feature would be useful for other formats too, anyway.
00:45:45*amiconn would like to see a better resampler :/
00:45:47 Quit obo ("bye")
00:46:09preglowamiconn: me too, but i can't be bothered
00:46:12*jhMikeS would prefer to hear a better resampler :p
00:46:22amiconnhehe, of course
00:46:27preglowmuch work, will need tons of optimizing and tuning
00:46:37preglowand: i will almost never need it
00:46:58preglowi'm keeping a steady eye on that speex resampler
00:46:59preglowit looks pretty good
00:47:21preglowjmspeex says it's pretty fast, but even that will be tons slower than the one we have now
00:48:11 Quit ender` (" There are two major products that come out of Berkeley: LSD and UNIX. We don't believe this to be a coincidence. -- Jeremy S)
00:48:42preglowender strikes again
00:49:56*amiconn considers this low level optimisation stuff a major part of the fun
00:51:40preglowi consider it fun to a certain degree, but it really depends on what i'm doing
00:52:29*jhMikeS just wants to get the scheduler updates committed and worry about some size or simplicity issues later. Some size increase on all to add robustness and responsiveness to everything and keep the API somewhat unified.
00:53:35preglowwell, go on :>
00:53:40jhMikeSthis stupid delta table make me too neurotic and there's probably better things to save bytes on anyway rather than compromizing integrity
00:54:50jhMikeShehe. it's one of those you need to be on standby about if there are problems but it's been run on all the ipod variants so issues might be minor at best. worst? eh, I think the point is understood.
00:56:35preglowdid kkurbjun check out the mini1g packed i2s stuff?
00:56:48jhMikeShe said he would later
00:57:35jhMikeSI'd like to reduce that stuff to #ifdef CPU_PP502x already :P
00:58:17preglowme too
00:58:46 Join antgel_ [0] (
01:00:42 Join midkay [0] (n=midkay@rockbox/developer/midkay)
01:00:47preglowdid you get the swp replacement stuff going yet, btw?
01:01:15jhMikeSpreglow: that's done and swp is used when it works otherwise sw corelock
01:01:43jhMikeSipod color played dual-core SPC files just fine (and H10 of course)
01:01:44 Quit scorche|w ("CGI:IRC (EOF)")
01:02:04 Join kugel [0] (i=kugel@unaffiliated/kugel)
01:04:59preglowsounds good
01:05:16preglowhow much slower would you expect the sw implementation to be? noticable at all?
01:07:03jhMikeSmeasureable in tests - not noticeable for playback. straight-up queue_sends run at about 75% the speed of the SWP version. ~46000 vs. ~61000 per second.
01:08:14jhMikeSthose weren't gotten from a pure test loop either...there was checking that data echod back from COP was correct.
01:08:48toffe82I try a svn update and the answer is Skipped '.' , what is wrong ?
01:10:05preglowi would expect it to be measurable in tests, yeah
01:10:19toffe82ignore what I said :)
01:10:27preglowjhMikeS: so, what's left? just optimization?
01:10:52 Quit antgel (Connection timed out)
01:11:13***Saving seen data "./dancer.seen"
01:12:30jhMikeSpreglow: Perhaps. The 45000 is faster than the initial implementation using SWP so not too bad there.
01:12:44 Join mark__ [0] (
01:13:26 Quit kugel ("ChatZilla [Firefox]")
01:14:07jhMikeSBuffering times are noticably shorter and even mpegplayer runs faster (on Gigabeat of all things)
01:14:09XavierGrdoes anyone know what "svn: Delta source ended unexpectedly" means when I update my repository?
01:14:25preglowjhMikeS: so no reason not to commit?
01:14:49preglowi see no reason to keep the half-assed stuff in svn in any longer than we need to
01:14:50 Join kugel [0] (i=kugel@unaffiliated/kugel)
01:14:51jhMikeSnot that I can think of other than my delta table neurosis:)
01:15:06preglowis the size increase that huge?
01:15:38jhMikeS~700 bytes on SH, ~1200 on CF, ~6000 on PP DC
01:15:41jhMikeSsomething like that
01:16:01preglowwell, the sh targets are the only ones where we should really care much
01:16:26 Quit mark__ (Remote closed the connection)
01:17:32jhMikeSall swcodec gets the fancy sync objects too. but I want certain things like thread waits and creating threads frozen to be unified or else things will get awkward with too many api variations. SH might even save bytes by using more advanced things.
01:19:38 Quit Arathis ("Bye, bye")
01:20:33 Nick fxb is now known as fxb__ (
01:21:14 Join Toxicity999 [0] (
01:21:48*jhMikeS should probably just get a stiff drink first :p
01:22:54*preglow would like a stiff drink :/
01:23:53*preglow goes to squeeze his ardbeg bottle
01:23:56jhMikeSone thing, if some other target has a problem with SWP (haven't found anything but PP5020), change one #define and it switches it.
01:24:07preglowjhMikeS: got a patch hanging around?
01:24:15jhMikeSyeah, one sec
01:24:45jhMikeSmy tree doesn't included the codec but I have one on my FTP with all that...hold
01:25:53 Quit spiorf (Remote closed the connection)">
01:26:03jhMikeShope it applies still
01:26:19jhMikeSwoops, use http://
01:27:14preglowi was starting to wonder if the patch really was one meg zipped...
01:29:58 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
01:30:34preglowi need to reboot before i can use usb, apparently
01:30:45jhMikeSIt applied ok?
01:31:16preglowhaven't tried, just tried ocpying the spcs
01:32:00preglowi think i need to put this off until tomorrow anyway
01:32:02preglowjust saw the time :>
01:32:24jhMikeSok, night
01:36:20 Join barrywardell [0] (
01:36:30 Quit barrywardell (Remote closed the connection)
01:36:48 Join barrywardell [0] (
01:36:48 Part toffe82
01:36:57 Part barrywardell
01:38:01 Join barrywardell [0] (
01:38:52 Part [IDC]Dragon
01:41:39 Join ramon8 [0] (
01:44:38 Join zajacattack [0] (i=4261723c@gateway/web/cgi-irc/
01:50:46 Quit Rob2222 (Read error: 104 (Connection reset by peer))
01:52:36 Join kkurbjun [0] (
01:53:22XavierGrwow 2.5 fps gained with amiconn's commit for the mpegplayer
01:56:26 Join Rob2222 [0] (
01:58:01zajacattackthat's pretty good, what player?
02:03:51 Quit linuxstb ("ChatZilla [Firefox]")
02:05:04 Join kkurbju1 [0] (
02:06:43 Quit hcs ("Leaving.")
02:07:11 Join Soap [0] (
02:11:43 Quit Thundercloud (Remote closed the connection)
02:12:52barrywardellamiconn: I'm testing the H10 ADC on low battery. It's still ~0x14 different depending on PLL
02:15:34 Join JdGordon [0] (
02:17:29 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
02:18:22 Join Hartwell [0] (
02:19:37Hartwellany zune support in the future?
02:20:06JdGordonthe future is a very long time
02:20:10zajacattackprobably not
02:20:35Nico_PJdGordon: hi
02:21:10Nico_PI hope to have fixed at least some of the bugs you, pondlife and GodEater mentioned
02:21:23Nico_Pand pondlife set up a wiki page for testing
02:21:35JdGordonyeah i saw that..
02:21:49JdGordonI think i should setup git for it so i dont have to keep downloading 16mb archives :p
02:22:00Nico_Pindeed you should :)
02:22:05Nico_PI can help if you want
02:22:06zajacattackHartwell: you can see the work on the zune port at
02:22:48Nico_PJdGordon: I'd be thankful if you could try reproducing your crash on shutdwon when buffering... I couldn't even whith the same version as yours
02:22:57Nico_Pbut hopefully it's fixed
02:23:13 Quit kkurbjun (Connection timed out)
02:23:27JdGordondoh! didnt see karl was here
02:23:37JdGordonNico_P: i only had it once.. but illgive it a go
02:24:11JdGordonok, how to i get git going?
02:24:20Nico_Pyou have it installed ?
02:24:26*jhMikeS is going to take the DC plunge tonight to get the API disruption over with and let it sit for awhile
02:24:46JdGordondc plune?
02:24:55JdGordonNico_P: i do now
02:25:00 Nick antgel_ is now known as antgel (
02:25:21 Quit Mouser_X (Read error: 110 (Connection timed out))
02:25:32Nico_PJdGordon: ok, so first you need do do "git clone git://". This will clone the repo in a Rockbox subdir
02:25:42jhMikeSNico_P: create_thread with have a flags param, no fallback param. struct event will become struct queue_event
02:26:13 Part zajacattack
02:27:59 Join toffe82 [0] (
02:28:24JdGordonindexing objects.... slowly...
02:28:38Nico_PJdGordon: it's quite a big repo
02:29:08Nico_PJdGordon: after that you can do the rest of the steps of
02:29:09JdGordonok, done
02:29:10Hartwellthanks zajacattack
02:29:20Nico_Pbut what's important is to checkout the mob branch
02:29:58Nico_Pgit checkout -b my-mob origin/mob
02:30:51JdGordonok then?
02:31:13Nico_Pafter checking out you have the mob code and you're on the branch
02:31:28Nico_Pnothing left to do fo now but compile and test ;)
02:31:49*JdGordon is good at tha
02:32:52 Quit Hartwell (Remote closed the connection)
02:33:24Nico_Pto update: "git pull origin +mob:my-mob" (while on the my-mob branch), then "git remote update" if it wasn't a fest-forward update (it tells you when it is one... if it's not it means I've rebased)
02:34:20Nico_Pone very cool thing is git bisect
02:34:46JdGordongoing up one in he playlist viewer doesn crash it now, but playback stops
02:35:59JdGordonand its not crashing on shutdown while buffering
02:36:20JdGordonill leave this build on my sansa and let you know if anything comes up
02:36:37Nico_Pok, cool
02:37:15Nico_Pplayback doesn't stop for me in the sim when going up in the playlist... I did notice it on my player last time I tried though... forgot about it
02:37:44JdGordonpressing skip back like a madman works fine :)
02:37:50Nico_PJdGordon: I'll be off soon. if you find a bug and have a bit of time, git bisect helps you find the commit that caused the bug
02:37:53JdGordonno gap between tracks
02:37:53 Quit bb (Nick collision from services.)
02:37:58 Join bb_ [0] (n=bb@unaffiliated/bb)
02:38:14Nico_PI hope that soon doing that won't even need to trigger a rebuffer
02:38:41JdGordonit iwll if you go before the bufered tracks..
02:38:46Nico_Pit already works quite well except when you let it just play the codec thread ends up starving
02:39:05Nico_Psure but currently it always rebuffers
02:39:30Nico_PI want to avoid discarding tracks so that skipping back to them is instantaneous
02:39:39JdGordonah ok
02:39:50JdGordonso this version sticks the tracks id3 info on the buffer?
02:40:12jhMikeSlet's see what changes soon with that stuff
02:40:21Nico_Pyes, except that it copies current and next track info into static buffers
02:50:27 Join Calcipher [0] (
02:51:00Nico_Pbed time
02:51:26*pixelma agrees
02:51:59*Nico_P hopes to see lots of bugreports when he wakes up :p
02:52:22JdGordonyou mean you want to see none, so its working
02:52:43Nico_PI want to see reports about bugs that were there but were fixed
02:52:49CalcipherI'm curious, has anyone here had success with the rockbox install for the Sansa E200R series using the e200rpatcher.exe under windows XP
02:52:58Nico_Pbut that's asking a bit too much at this point :)
02:53:09 Part pixelma
02:53:33JdGordonCalcipher: whats the problem?
02:53:43 Quit Nico_P (Remote closed the connection)
02:54:08CalcipherI tried the instructions, attempted to go into manufacture mode, was prompted for drivers
02:54:21Calcipherand the drivers were used from the links provided
02:54:36JdGordonand they didnt work?
02:54:37 Quit kkurbju1 (Read error: 110 (Connection timed out))
02:54:41Calcipherthen I launched the e200r tool
02:54:52Calcipherand it gave me a fail error
02:55:02JdGordonwhich gave you a fail error?
02:55:34Calciphernow my players wheel light has been lit since
02:55:57Calcipherthe e200rpatcher
02:56:26JdGordonwhich gave you a fail error?
02:56:54Calcipherthe patcher said it failed to write
02:57:12JdGordonunplug the sansa and power it off (hold power untill it turns off)
02:57:28JdGordonstart e200rpatcher and then plug it in in manufac mode
02:57:49JdGordonapparently some are finiky about being in m mode for too long before something happens
02:58:26Calcipherok let me first boot into the regular firmware
02:58:33Calcipherjust to be safe
02:59:01Calciphernot so good
02:59:36Calcipherfalse alarm, that was just the hold message
03:00:00CalcipherI just bought this off my friend
03:00:28Calcipherhe went for an Ipod touch, I wanted this since I have trouble with my vision
03:02:39Calcipherso its in manufacture mode
03:02:58Calcipherwhere in device manager is it listed?
03:03:01JdGordonso press ener in e200rpatcher
03:03:26JdGordondunno, under usb devices maybe
03:03:59Calcipherthat was the smoothest tech support I've ever encountered
03:04:12CalcipherIt worked
03:04:54CalcipherSo I'll continue with the regular intructions. This method will allow dual booting correct?
03:05:27Calcipherglad I didn't have to pull out my linux cds, I use a screen magnifier in windows
03:05:47CalcipherUbuntu has accesability features, but sometimes its iffy
03:06:24Calcipherbeautiful, thank you both for being patient, I bet you get this alot hmm
03:07:02barrywardellthe frameworks method on this page sounds like a good way to build rbutilqt on the mac in the future:
03:07:10barrywardellI just tried it and it works
03:07:54CalcipherAs I was reading through the forums I noticed mention of sdhc support, and something about the implementation causing issues on some of the boot loaders
03:08:45CalcipherSince there is so much info, what is the actuall status of the sdhc support, or even the usb support, and also is the sdhc support only when your in rockbox ?
03:09:20JdGordonsdhc works
03:09:25JdGordononly in rockbox
03:09:37JdGordonand atm you have to use the OF for usb
03:09:53Calcipherwow, you guys are definitely talented
03:09:56barrywardellthe issues with bootloaders and sdhc are fixed
03:10:33CalcipherI was on the phone with sandisk tech support before buying this
03:11:01Calcipherasking about any firmware updates they may make for the e200s
03:11:15***Saving seen data "./dancer.seen"
03:11:15Calcipherthe guy sounded scared to answer
03:12:03 Join kkurbju1 [0] (
03:12:12CalcipherOk I have another question before I proceed
03:12:19 Nick kkurbju1 is now known as kkurbjun (
03:12:45Calciphernow just to let you know, I can't see very small print
03:13:31Calcipherand after the e200r patcher successfully applied the patch
03:13:58 Join amanda99 [0] (
03:14:05CalcipherI have not unplugged the player, and I can tell there are a couple of lines of text
03:14:11Calcipheron the screen
03:14:15 Nick amanda99 is now known as Aspiniou (
03:14:46Aspinioucipa cipa chuj
03:14:50DBUGEnqueued KICK Aspiniou
03:14:51***Alert Mode level 1
03:14:51***Alert Mode level 2
03:14:57scorcheAspiniou: you can stop now
03:15:07Mode"#rockbox +o scorche " by ChanServ (ChanServ@services.)
03:15:08JdGordonCalcipher: can you count how many lines there are?
03:15:34 Part Aspiniou
03:15:48 Join aliask [0] (
03:15:53JdGordonhe's back!
03:15:54JdGordonban him
03:16:04 Join psycho_maniac [0] (
03:16:52JdGordonCalcipher: it didnt definatly work... can you read the text on the screen? or at least how many lines of textt there are?
03:17:53CalcipherActually I used a magnifier, and it is the rockbox installer confirmation
03:18:04Calcipherand says to continue with step 2
03:18:12JdGordonok then it worked
03:18:19JdGordonunlock hold and press a button
03:18:24JdGordonthen continue with the steps on the wiki
03:20:10Mode"#rockbox -o scorche " by ChanServ (ChanServ@services.)
03:20:23scorchei really need to set some aliases so i cna do that quicker...
03:22:38JdGordonisnt there an autoop thing so you can kick/ban without actually being oped?
03:23:06scorchethere is, but autoop is discouraged on freenode, and i agree with that
03:23:37CalcipherHmm, so after the e200r tool I still need to run the sansapatcher?
03:23:47scorchethe opping isnt the slow is manually typing "/mode #rockbox +b %nick!*@*"
03:24:24JdGordonCalcipher: follow the instructions... iirc you have to put a e200 firmware on first, then sansapatcher
03:24:52***Alert Mode OFF
03:25:12CalcipherAh ok, If I lose rhapsody support I wouldn't care
03:25:49CalcipherJust checking since I had to take the R detour on the instructions
03:25:55CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
03:25:55*jhMikeS guesses alot of fingers may point to r15134 for a bit :p
03:26:16 Join webguest37 [0] (i=46403811@gateway/web/cgi-irc/
03:26:53webguest37Yay! The record feature on the c200 has been implemented! Thankyou!
03:27:27JdGordonjhMikeS: cool, so does the cop do anything yet?
03:27:51webguest37What do you mean?
03:27:56jhMikeSwe'll just keep things low for now. it should be able to be used like anything else now though.
03:28:49 Quit webguest37 (Client Quit) says i dont have permissions to view that task
03:29:52 Join TypeleveN [0] (n=Typeleve@
03:30:56*jhMikeS wonders what the score will be here
03:31:02CalcipherI tried the rbutil install method
03:31:47CalcipherI'm guessing the player isn't in MSC mode, so autodetect didn't work, let me see if thats the case
03:32:13JdGordondid you put the vanialla pp5022.mi4 on it yet?
03:32:42Calcipherno, so far I have only ran the R patcher
03:32:51Calcipherand just tried rbutil
03:33:05 Quit TypeleveN (Client Quit)
03:33:16lostlogicnice work jhMikeS
03:33:50jhMikeSlostlogic: thanks...I hope the build server didn't stick on me :P
03:34:01lostlogichehe, I'm reading the diff now
03:34:02 Join TypeleveN [0] (n=Typeleve@
03:34:09JdGordonna, a big commit like that makes sense to take a while longer
03:34:25lostlogicthe builds sent to my server have already completed
03:35:17jhMikeSnot bad on the red considering
03:35:32JdGordonyour lucky amiconn is alseep :D some more commits so he doesn't see that...please :)
03:36:04JdGordonouch on the delta
03:36:35jhMikeSI'll get to trimming things where possible of couse.
03:37:08JdGordonoh, red is only sim.. not so bad
03:37:10jhMikeSThe archos deltas are higher than what I measured
03:37:49jhMikeSmy player builds only went up about 750 bytes ... hmmm
03:38:38lostlogicwell it boots :-P
03:38:41lostlogicand plays
03:38:53 Join meoblast001 [0] (
03:39:28meoblast001does anyone know a good place where i can find rockbox themes for Sansa c200
03:39:37jhMikeSlostlogic: normally a good thing indeed :)
03:39:42meoblast001i know its newly supported but i know some site has to have a few
03:40:04meoblast001i could only find 1 on the whole internet
03:40:08lostlogicspecially since my build is a little bit special (no tagcache, no dircache, codec on COP and codec has no yields)
03:41:50Soapso, jhMikeS - what are the implications of your commit?
03:42:06jhMikeShopefully not a mob with torches
03:42:20meoblast001i guess not... ok ill bbl
03:43:16SoapI've been reading a lot of your conversations prior to this commit - and I'll admit, as a layman, I'm not 100% sure what differences a user should expect to see...
03:43:27lostlogicusers shouldn't see any yet :-P
03:43:30jhMikeSYou can use COP threads just like threads anywhere else using queues, mutexes, and whatever safely...and a general speedup for all things. The playback engine and other things can use it freely.
03:43:58JdGordoncan threads swap between cores?
03:44:08Soapso you've only opened the door?
03:44:40jhMikeSat the moment yes. to include a playback rewrite to enable it now would be going way too far.
03:45:09psycho_maniaccould somebody please explain the new big comment in svn? something about mulitcore support.
03:45:15Soapok, cool, that's what I was ASSuming looking at the diffs, but I wasn't sure.
03:45:48jhMikeSpsycho_maniac: the kernel is multiprocessor safe in a general way now
03:46:01 Quit jhulst ("Konversation terminated!")
03:46:25psycho_maniacdoes this effect how rockbox runs on the 502x players?
03:46:52 Join kkurbju1 [0] (
03:47:05jhMikeSit will. I just laid the groundwork to be able to use the other cores which will make things run much better.
03:47:43psycho_maniacand this is for the 80gig ipods? i connected my ipod before i could check the info on it.
03:48:06 Quit kugel (Read error: 110 (Connection timed out))
03:48:13lostlogicpsycho_maniac: his changes impact all multicore targets, which includes all currently supported ipods
03:48:29jhMikeSpsycho_maniac: all PP502x devices. PP5002 is on the TODO list.
03:48:46 Quit kkurbjun (Read error: 110 (Connection timed out))
03:49:06psycho_maniacso that means NOT my player?
03:49:19jhMikeS80gig is a PP502x target
03:49:53jhMikeSaye, "no newline at end of file" ?? how lame :P
03:50:03psycho_maniacwhat is a pp5002 ?
03:50:39jhMikeSan older portal player chip found in iPod 1g/2g/3g
03:51:04psycho_maniacoh ok i see.
03:51:06 Join webguest34 [0] (i=4cb431de@gateway/web/cgi-irc/
03:51:14lostlogicjhMikeS: do we have no multicore support at all on 5002?
03:52:30jhMikeSnope, it's built as single core only
03:52:33webguest34Will Rockboy be optimized to take advantage of the multicore support?
03:52:56jhMikeSif someone decides to do it, sure
03:53:25*jhMikeS wonders why his compilter didn't complain about the newline
03:54:33*jhMikeS should probably update the gcc for x86
03:54:34webguest34so far on my H10 20gb with the new build, the only major difference I've noticed is Brickmania having a smooth framerate while music is also playing
03:54:36lostlogicwow man, this is a huge body of work you've done.
03:54:50webguest34before, Brickmania was jerky while music was playing
03:55:17lostlogicIIRC nothing is actually running on core2 at this point in the official builds.
03:55:33jhMikeSaye...been working on it a few months. using Peterson's algorthm on PP5020 chips where SWP is broken finally got it going.
03:55:45 Quit qweru ("moo")
03:56:30webguest34well thank you fo your work :)
03:57:00jhMikeSyou're welcome. you might notice some benifits right off, yes.
03:57:24lostlogicjhMikeS: unless I'm hallucinating, I got a 1% speedup on ogg decoding 224% vs 223% realtime
03:57:33lostlogictheoretically not noticeable to a human though
03:58:23jhMikeSswitch_thread is smaller. lags that would happen because the scheduler didn't detect the thread wakeup should be gone.
03:58:25meoblast001im back
03:59:16meoblast001does anyone know where Sansa c200 themes can be found? i only found 1 on the whole net
03:59:24jhMikeStimeout list removal is very lazy and timeouts are only checked when the earliest thread is due to wake up
03:59:25lostlogicmeoblast001: there really just might nto be any posted online yet
03:59:37 Join Mouser_X [0] (
03:59:48meoblast001lostlogic: any good sites you know of that generally have a lot?
04:00:00lostlogicmeoblast001: rockbox WPSs and themes are pretty easy to make, I suggest making your own that suits your needs
04:00:08lostlogicmeoblast001: if it's not on our wiki WPSs page then no.
04:00:33jhMikeSnow that thread.c is cozy 2595 lines, I'm sure I'll get to hear later endless lectures on complexity :)
04:00:47meoblast001lostlogic: whats the difference between a theme and a wps?
04:00:50lostlogicjhMikeS: my statement huge body of work was code for a lecture :-P
04:01:17lostlogicmeoblast001: a theme changes font and color and other system attributes, a wps is just a layout for the while playing screen
04:01:18jhMikeSI _try_ to do as little as I can while meeting my goals.
04:02:01meoblast001lostlogic: is there a way to get the layout (like the battery) to appear everywhere on the system... i dont like the default
04:02:07jhMikeSBut really I'd rather 1% of the code support simplifying the other 99%
04:02:27meoblast001lostlogic: was that no to me?
04:02:32lostlogicjhMikeS: yeah, better to have the complexity in the kernel where few have to work on it
04:02:36lostlogicmeoblast001: yes
04:02:51lostlogicjhMikeS: it's definitely harder to read thread.c now :(
04:02:59psycho_maniaci have a question. now i have 3 folders in my networking thing. this happend after my last svn up, i now have filesystem, and homes, what is this?
04:03:01meoblast001lostlogic: should i request that feature?
04:03:13jhMikeSI was hoping the comments helped somewhat? no? Maybe it should be split?
04:04:13lostlogicmeoblast001: doesn't make sense to me to have a customizeable status bar outside of the WPS, but request anything you want :-P
04:04:40lostlogicjhMikeS: yeah, it's understandable, I just always have some difficulty mentally ifdefing ;)
04:04:43webguest34Question - in the database uner "A to Z" - why isn't there a subfolder for "Track"? - It would be much easier to search for songs by title if they were sorted in folders marked A to Z
04:05:19Mouser_Xlostlogic: Well, it would be nice to see a few things. Time, date, perhaps battery, outside the WPS screen (I thought there was a theme that did that though?)
04:05:36lostlogicMouser_X: the status bar shows those things... ...
04:05:43Mouser_XAh, well then.
04:05:49lostlogicI thought meoblast001's question was about customizing the status bar
04:05:54Mouser_XI guess I either missed them, or my theme removed them.
04:06:32lostlogicjhMikeS: I'm definitely stil lreading though, specially since I'm such a bystander these days ;)
04:06:39meoblast001lostlogic: my question was about being able to customize features like the battery icon outside of the now playing window
04:06:47meoblast001if thats what your wondering
04:06:52lostlogicmeoblast001: ok, cool.
04:07:10meoblast001lostlogic: so thats not currently possible?
04:07:20lostlogicmeoblast001: correct, afaik.
04:07:52webguest34as far as i know
04:09:36jhMikeSlostlogic: should give some reading enjoyment for awhile :)
04:10:50 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
04:11:53meoblast001lostlogic: are you sure because i dont want to sound stupid
04:12:00JdGordonsomeone was just casing my house with me home!
04:12:12lostlogicjhMikeS: did you find out that SWP was broken by test?
04:12:14Mouser_Xmeoblast001: You could alway try, and see what happens.
04:12:57Mouser_XJdGordon: My brother's place was broken into while he was home (it's small place. I don't think they were too bright. He shot at them..)
04:13:00jhMikeSlostlogic: difficult to test directly actually but it's removal or addition would make or break it for H10
04:13:10CalcipherAlright, rockbox is running, very nice
04:13:14 Quit webguest34 ("CGI:IRC (Ping timeout)")
04:13:36lostlogicjhMikeS: gotcha
04:14:29CalcipherThanks again, I had originally thought that after the R bootloader patch process I had to continue with the original install, thats where my confusion came from, but I looked it over and saw I had to follow a different set of instructions, sorry for the trouble.
04:14:42jhMikeSany dual core test plugin would simply lock up instantly using it
04:15:26jhMikeSon pp5022/24 it's quite fine and preferred since it's faster and smaller
04:15:41*Soap can tell a diff between 224 and 223
04:17:32lostlogicsoap: ie it feels faster with jhMikeS's commit?
04:17:55Soapit's 1 faster, isn it?
04:18:07Soap /sarcasm
04:18:25Soapyou don't have to get mean
04:18:37jhMikeSit goes to 11 now :P
04:18:38lostlogicahahaha :)
04:19:49Soapnow that you have this offloaded, a week of rest and squashing, and you should be ready to receive my Nano ;)
04:20:04SoapI'll never ask for status updates, I promise.
04:20:25jhMikeSah...why not. gotta be something that can be done. I just have no clue what atm.
04:20:48lostlogicSoap: what's he need your nano for?
04:21:31jhMikeSit's the type that freezes when warm
04:21:36Soaplostlogic - my Nano is one of the minority which suffers freezing and data corruption with the recent clock reworking Amiconn did based upon his RE of the Apple OF.
04:21:45 Join shnee [0] (
04:21:55lostlogicahh, sweet!
04:22:05lostlogicjust keep some lN2 in your pocket with it.
04:22:07Soapand yes, the firmware misbehavior can be mitigated by cooling.
04:22:22 Quit troxor ("leaving")
04:22:39 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
04:22:44Soapthe strap-on peltier kills the battery-life.
04:22:49 Join miepchen^schlaf [0] (
04:23:01lostlogicproblematic indeed.
04:26:53*jhMikeS looks at the diffs more closely since the colored ones always show the debugging crud accidentally left laying around
04:28:47jhMikeSugh, left a whole bunch of NOCACHEBSS stuff that's really not nescessary. :\
04:29:50meoblast001lostlogic: if themes only do colors and stuff... how is mine able to do linear colors.... thats generally not a color related thing
04:30:05meoblast001although it is colors
04:30:18SoapI don't think lostlogic said "only"
04:31:31meoblast001so can it change the generic black battery icon?
04:31:40Soapthemes can address all the eyecandy of rockbox - from icons, to backdrops, to pretty gradient selector bars, to colour-by-file-extension, to WPS screens, to colours in general.
04:31:52SoapWPSs are the While Playing Screens.
04:32:12meoblast001so is there anyway to change the bar at the top of my screen with the battery and the time and the volume
04:32:53SoapDo you mean the default status bar at the top of the screen? IIRC the status bar itself is NOT current themeable, except by a tracker-hosted patch.
04:33:24meoblast001whats a tracker hosted patch?
04:33:41scorchea patch hosted in the tracker :)
04:34:04meoblast001do you know where one might be that allows me to do this?
04:34:24scorcheas the name he tracker
04:35:22SoapWhat I ment by "a tracker hosted patch" was a patch, that while unofficial - and currently deemed unworthy of inclusion into "mainstream" Rockbox, nonetheless is hosted in the official patch tracker (flyspray) and not floating around in the ether on somerandombody's website.
04:36:49kkurbju1jhMikeS: Ipod Mini G1 packed I2S works :)
04:37:35CalcipherI really am amazed at how well all these things work together, the rockbox dev community is definitely good at what it does
04:37:57kkurbju1I changed the define to CPU_PP502x, do you want me to commit it?
04:38:11SoapMy suggestion would be to search flyspray with the keywords "status bar", and expand out from there if you do not find what you are looking for. That being said - I am not aware of one of the "Unsupported Builds" from the forum grouping of the same name which are using this patch...leading me to believe that it is hopelessly out-of-date. tdtooke has been doing a very impressive job of...
04:38:11CalcipherRight now I'm using the rbutil on my E280r with speech enabled, and some new larger fonts
04:38:12Soap...keeping all the rejected eyecandy patches current, and if he hasn't done so with this one - it either still works, or it might never work again. Calling this one a macro, scorche?
04:38:21scorchehell yes
04:39:58Calcipherthank you all yet again, you made my damn good mp3 player, into a serious little feature filled bastard
04:40:44jhMikeSkkurbjun: with LE16_2? You can pull the selection in i2s-pp.c entirely then. So yeah.
04:41:59jhMikeSkkurbjun: dump the 2nd fifo format (halfword as well). Not needed now either.
04:42:02 Quit Lambuntu (Remote closed the connection)
04:42:20jhMikeSIn pcm-pp.c
04:42:35kkurbju1eh, I'll let you simplify it
04:42:49jhMikeSok, np. 5 min to that.
04:43:28 Quit TypeleveN (Read error: 110 (Connection timed out))
04:43:53*jhMikeS will start trying to use DMA there
04:45:06 Join Lambuntu [0] (
04:54:00 Quit animeloe ("Leaving")
04:55:37 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
04:58:40 Quit aliask ("ChatZilla [Firefox]")
05:00:26 Join webguest42 [0] (i=424b3cb2@gateway/web/cgi-irc/
05:01:05webguest42hey where can i download the actual tool for putting the themes and fonts and games on my ipod at?
05:01:52Mouser_XTry the Wiki. RockboxUtilityqt?
05:01:56Mouser_X(Is that right?)
05:03:09webguest42ya, thanks ive been trying to find it it got earased from my pc and i forgot where to download it from.
05:04:25webguest42hey i keep aplying the base wad file to doom and it loads but it always freezes as soon as i get to the games play
05:05:10Mouser_XSorry, I've only used DOOM like, twice. I can't help
05:06:21webguest42is there anyone else in here who can help me
05:08:38krazykithave you also searched the forum?
05:10:23webguest42it wasnt very helpful
05:10:29webguest42it wasnt cleer
05:11:13webguest42i just want to know what files i need to get in order to run doom on my ipod video
05:11:18***Saving seen data "./dancer.seen"
05:11:27krazykitthe wiki tells you what files you need
05:11:45krazykiti followed the wiki and doom worked no problem.
05:13:01webguest42ya ive downloaded it but do i just extract it to the ipod and the file will find its way or where do i put it
05:13:35krazykitthe wiki tells you.
05:13:58krazykitit specifically tells you where to put each file
05:14:12 Quit Mouser_X (Read error: 104 (Connection reset by peer))
05:15:38 Join bb [0] (n=bb@unaffiliated/bb)
05:16:23krazykitwebguest42, in fact, i don't know what you're extracting, as the wiki tells you to put two files into the right place, neither of which is zipped.
05:17:33 Join eigma [0] (n=cat@
05:18:44webguest42ok i got the three files it said to put on there and added them, is that it i dont understand what the addons are?
05:18:53webguest42for doom
05:19:02krazykitdoes the standard game work?
05:19:34krazykitaddons go where the wiki says: .rockbox/doom/addons
05:20:51 Join webguest54 [0] (i=dce9d1f6@gateway/web/cgi-irc/
05:21:16 Join Mouser_X [0] (
05:21:38webguest42do i need prboom?
05:22:32 Quit webguest54 (Client Quit)
05:22:37 Join webguest54 [0] (i=dce9d1f6@gateway/web/cgi-irc/
05:24:09krazykitwebguest42, no, you don't need prboom. all you need is specified in the wiki, which, at minimum, is 2 files.
05:25:26webguest42ever since i installed rockbox on my ipod it seems to consume more battery life than with the normal apple firmware why is that?
05:26:39krazykitbecause, as is discussed several times in the forums, it's not known how to put some of the hardware to sleep and other power-saving modes.
05:26:47psycho_maniacNo work has yet been done to optimise power consumption on other models.
05:27:11Lloreanpsycho_maniac: that's a lie.
05:27:23psycho_maniacwell i got that from the wiki
05:27:37LloreanWhere in the wiki does it say *no* power optimization has been done?
05:27:42 Quit bb_ (Read error: 110 (Connection timed out))
05:27:48LloreanAnd, more importantly, when is this dated?
05:27:52psycho_maniaci should of changed the words aroudn a bit. this is only for the ipod.
05:28:04LloreanMy question still stands
05:28:08psycho_maniacin the ipodport wiki page.
05:28:12webguest54can someone pls tell me if there are any other alternatives to rockbox? I want to play flac files on my 80g video ipod. I gather that rockbox will not run on it from discussions i have read on the forum....
05:28:46Lloreanpsycho_maniac: It's certainly not on that page
05:28:56psycho_maniacalright then
05:29:02Lloreanpsycho_maniac: And the IpodStatus page says "more work could be done" which very clearly requires that some work has been done.
05:29:10LloreanI'd very much like to know which page you got it from so that I can correct it.
05:29:22webguest42ok doom still doesnt work it loads to the first level then it just freezes as soon as the level starts the controls dont work. i have installed all three files that the forum told to install and it still doesnt work. is there something that im missing?
05:29:46krazykitwebguest54, which 80gb ipod? if it's a 5.5g ipod, rockbox will run on it. if it's the ipod classic, you're out of luck, as no 3rd party software runs on the device.
05:29:53Lloreanwebguest54: Rockbox runs on All HD based iPods except the "iPod Classic" recently released
05:30:33webguest42its the 5.5 released in late 2006
05:30:57alienbiker99then rockbox runs on it
05:30:59Lloreanpsycho_maniac: That's actually in reference to a very specific set of fixes made to 1st through 3rd generation, whoever wrote it managed to phrase it very poorly
05:31:05webguest54i am not sure about the latest classic but I bought mine in HongKong 4 months ago
05:31:17 Quit eigma ()
05:31:40psycho_maniacthen why is it under 4th gen and higher?
05:31:57LloreanBecause that fix can't be made to 4th gen and higher yet
05:32:26 Quit meoblast001 ("Leaving")
05:32:46webguest54hi krazykit.....wht is a 5.5g ipod (excuse my poor understanding)
05:32:59webguest54what is a 5.5g ipod?
05:33:14webguest42its the 80 gig video released in late 200p
05:33:32webguest54oh ok...then i may be in luck
05:34:22CalcipherExcuse me folks, I have a quick question for anyone using the sansa e200 series players and rockbox
05:34:35psycho_maniacask it
05:35:17CalcipherOh and also on the windows platform, I just wanted to know what the most compatible non annoying program for transfering and maintaining music on the player is
05:35:42CalcipherI just got one today and got it rockboxed, I never even installed the cds
05:35:51DogBoyany file manager you choose to use
05:36:42CalcipherI was going to use Winamp 5, but I noticed it reads the player as USB mass storage L: and K:
05:36:44krazykitCalcipher, i do it by hand, personally, but i hear good things about mediamonkey's sync options. winamp and foobar probably have plugins for this too
05:37:31CalcipherOh and winamp was just dropping folders on the root
05:37:38Calcipherso thats not very organized
05:37:59 Quit webguest42 ("CGI:IRC (EOF)")
05:38:45psycho_maniaccorrect me if im wrong, but you can play FLAC files in rockbox?
05:38:53krazykityes, you can
05:39:16krazykitthere's a codec page on the wiki for reference
05:39:35psycho_maniacyes i looked at that.
05:39:35CalcipherNow I'm wondering if setting the player on MTP mode will make it more compatible with winamp 5, but will this cause problems with rockbox?
05:39:44psycho_maniacwebguest54: you can play FLAC on rockbox. so what is the problem?
05:40:05krazykitCalcipher, shouldn't cause problems, as long as the files get on there.
05:40:10Calcipherrockbox is fuckin amazing, so many formats supported
05:40:57Calcipherah ok, then I'll give it a try, sounds like I may have to switch back to msc mode if I want to add stuff to rockbox
05:41:07Mouser_XIt's why I have a portable player now (the formats I use are rather wonky to most players.)
05:41:53krazykitCalcipher, just note that in MTP mode, only Sansa-supported files can be transported, so not vorbis, flac, etc
05:42:09JdGordondont use mtp
05:42:39CalcipherSo what I have been doing since installing rockbox is turning the player off then connecting to the USB
05:43:04webguest54psycho_maniac....I was asking if rockbox runs on my i found out that my ipod is a 5th generation 80gig ipod with video that was made late, I guess it runs rockbox...and I am ready to study the manual etc
05:43:12Calcipherthen it loads the original firmware and quickly goes into docked mode with the usb transfer image displayed
05:43:44Calcipherso to get into the normal firmware I need to hold down the left button when booting correct?
05:44:04 Quit Lambuntu (Remote closed the connection)
05:44:08CalcipherOh ok, so scratch fucking everything up in MTP mode haha, forget I asked
05:44:30CalcipherI think I'll listen to JdGordon
05:44:35webguest54another question for you seasoned guys:
05:45:02webguest54once i start using rockbox is it goodbye to Itunes? or can i have both running on my ipod? thx
05:45:11 Join Lambuntu [0] (
05:45:25Mouser_XYou can, but Rockbox won't use iTunes database at all.
05:45:36Mouser_XNor will it be able to play the DRM stuff.
05:45:53DogBoywhat drm stuff
05:45:56krazykitwebguest54, if you choose to sync with itunes, you MUST use the rockbox database to play music
05:46:25krazykitDogBoy, the itunes music store purchases, except the more expensive drm-free ones
05:46:26Mouser_XOh yes. I forgot. Itunes *mangles* the filenames.
05:46:44webguest54thx krazykit
05:46:51DogBoykrazykit, I was kidding, I refuse to buy that stuff
05:51:38webguest54thx to all for your answers...really appreciate it
05:52:56CalcipherCan I ask what I might be avoiding by not using MTP mode, I noticed a large amount of forum posts all over mentioning the player working with winamp and WMP10 and 11 easier
05:54:09alienbiker99you cant transfer ogg and flac among others to use in rockbox
05:55:05Mouser_XI don't use Winamp to transfer files. I use Windows Explorer, and simply dump the files where I want them.
05:56:03Mouser_XYou could do the same (though, depending on how you have your music setup, there might be something that's more useful to you. Mine is a mess, so it's probably better organized on my Gigabeat, than it is on the PC).
05:57:54CalcipherI see, well I was just looking to use my normal music player to do all my music tasks, I don't have much music in other formats other than mp3 so I don't mind not having that ability in MTP mode
05:58:44CalcipherI get the feeling theres more to avoiding MTP mode than just the music file types limitation, when dealing with a rockbox set up
05:59:25alienbiker99its the limitation of files that can be transfered
06:00:04 Join kubiix [0] (
06:00:26CalcipherMan, so each time I reboot this thing its going to rebuild the db??
06:00:38Calciphereven if no changes were made to the music?
06:01:33JdGordonwe know how to stop it for 3 OF verions
06:01:45CalcipherI'm brand new to the player, so I don't know if theres a sleep or power save mode, instead of powering off to avoid the loading time
06:01:47JdGordonbut the later ones changed their system so it doesnt work anymore
06:01:58JdGordonno sleep
06:02:01JdGordonjust poweroff
06:02:47Calcipherdamn bastards, and I bet the other additions to the newer OFs are worth having?
06:03:03*JdGordon wouldnt know...
06:03:12JdGordoni only use the OF for usb
06:03:32Calcipheractually that doesn't happen in rockbox
06:03:45Calcipherso your right its no so bad if its only for USB transfers
06:04:03CalcipherHey why do you suggest I stay away from MTP mode?
06:04:13JdGordonbecause mtp sucks
06:04:28CalcipherSeems like it would make using the player with certain programs smoother
06:04:42alienbiker99then use it
06:04:57Calcipherdoes the SD card still show up when your in MSC mode?
06:05:21Calcipherso then I can add random non supported formats to the sd card probably
06:06:22Calcipherthanks for all the help jd
06:06:40 Quit ramon8 (Read error: 110 (Connection timed out))
06:07:56webguest54another question: will I be able to keep and update my itunes podcasts on the rockbox?
06:09:38DogBoypodcasts are a good reason _not_ to use itunes
06:10:03webguest54also, if there were 2-3 major reasons to use rockbox on the ipod, what would they be....(i just want to make sure i am doing the right thing)
06:10:19webguest54why dogboy?
06:11:08DogBoyfrom what I understand (not using itunes) the feed urls are difficult to sus out of itunes
06:11:15Calcipherbeautiful, now syncing with winamp is running perfectly
06:11:23LloreanDogBoy: Most modern podcasts are just a standard RSS feed
06:11:27scorchewebguest54: see the WhyRockbox page and choose your own "major reaons"
06:11:38DogBoyexactly Llorean
06:11:44DogBoybut try finding them in itunes
06:13:01CalcipherAh, plenty of Deep purple and Maiden...making me happy I bought this thing, and that Rockbox rocks so much!
06:15:21webguest54thx scorche...i will
06:16:38webguest54also, is there a specific mp3 player that rockbox runs better than on anything else?
06:17:48krazykitdepends on what features you want, really
06:20:37Mouser_XRockbox runs like a dream on the Gigabeat...
06:20:46Mouser_X(I have a Gigabeat F40)
06:21:59webguest54krazykit...i was used to minidisk players....i want crystal clear music...i may be crazy but this is what i like and i have not seen an mp3 player able to do this....flac files sound great on my pc...if i could do the same on the ipod that i have that would be nice....but then i am open to suggestions...gigabeat oh yes
06:22:37Mouser_XThe iPod is not the player you want then. I've heard from many people that its audio hardware is terrible.
06:23:09scorchewell, that also depends on your headphones...
06:23:36Mouser_XHowever, I've heard many people say that the Gigabeat sounds really good, and very clear (I'm not one to ask. I don't pay attention to those things. I'm just saying what I've heard.)
06:23:58Mouser_Xscorche: I have $15 earbuds.
06:24:14krazykiti've been happy with the sound quality on higher-end phones
06:24:35scorcheMouser_X: well, i was referring to the issue of sub 32 ohm headphones
06:24:40DogBoyI was saying that a while back but then...
06:25:01DogBoywhat issue is that scorche
06:25:06DogBoyI don't know about it
06:25:26scorcheSoap will tell you all about it ;)
06:25:52 Quit webguest54 ("CGI:IRC")
06:25:59 Join webguest54 [0] (i=dce9d1f6@gateway/web/cgi-irc/
06:26:05DogBoyok thanks
06:26:33scorcheDogBoy: there are threads in the forum about it if you are interested enough to search
06:26:51DogBoyseems like we've been here before
06:27:01 Quit lazka_ (Remote closed the connection)
06:28:20CalcipherI... I couldn't help it, I had to put Dio on there...
06:28:32CalcipherKing Diamond next!
06:28:40krazykitCalcipher, please keep it on topic
06:28:50Calcipherah, sorry
06:28:55DogBoydon't forget the slim whitman
06:50:40 Quit kkurbju1 (Read error: 110 (Connection timed out))
06:57:36 Quit Calcipher ("—I-n-v-i-s-i-o-n— 2.0 Build 3515 with A Pack Fix By")
07:00:54 Join EnterUse1Name [0] (
07:04:08 Part toffe82
07:05:15 Quit Toxicity999 ("Leaving")
07:08:11 Join Toxicity999 [0] (n=bryan@unaffiliated/Toxicity999)
07:11:20***Saving seen data "./dancer.seen"
07:12:24 Quit EnterUserName (Read error: 110 (Connection timed out))
07:22:31 Quit psycho_maniac (" HydraIRC -> <- IRC with a difference")
07:25:14 Nick fxb__ is now known as fxb (
07:29:52 Quit tchan (SendQ exceeded)
07:31:22 Join tchan [0] (n=tchan@lunar-linux/developer/tchan)
07:34:43 Join advcomp2019_ [0] (n=advcomp2@
07:35:11 Quit advcomp2019 (Nick collision from services.)
07:35:17 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@
07:36:15 Quit webguest54 ("CGI:IRC")
07:38:26 Nick fxb is now known as fxb__ (
07:41:58 Quit jhulst (Read error: 113 (No route to host))
07:46:05 Quit parafin|away (Read error: 101 (Network is unreachable))
07:54:01 Quit Mouser_X (Read error: 110 (Connection timed out))
08:12:10 Quit Seed ("cu, Andre")
08:12:41 Join Rob222241 [0] (
08:19:35 Join Mouser_X [0] (
08:19:42 Join Seed [0] (
08:23:17amiconnQuite a big red delta :(
08:23:42amiconnjhMikeS: The delta for SH1 is larger than what you got because it's not only binary size change
08:24:01amiconnIt's (binary_size_delta + ram_usage_delta) / 2
08:24:13amiconnHover over the cell to see the details
08:24:35jhMikeSamiconn: I figured that out after a closer look. I do think it's a good deal more than it should be as well.
08:24:51amiconnjhMikeS: One reason for the increase is probably the extra argument to create_thread() (the flags)
08:25:12amiconnNow it has 5 arguments on SH instead of 4, meaning it is now a stackparm function...
08:25:43amiconnWhat are those flags used for, btw? I've only ever seen zero when looking at the diffs...
08:25:47jhMikeShmmm....I was under the impression that SH used 8 regs for regparms
08:25:54amiconnNo, it uses 4
08:26:15amiconnIt has 8 scratch regs though, r0..r7
08:27:06jhMikeSa) to extend functionality without adding more parameters. b) right now there's a flag that keeps a new thread from actually running until you thaw it. This functionality is rather small itself.
08:28:36amiconnWhy introduce functionality that isn't used (unless you have plans to use it *soon*)?
08:28:39jhMikeSI had thought about just using a struct that can be statically initialized and just a pointer passed to create_thread. Probably would save size and be less awkward.
08:28:57jhMikeSI do have plans to use it. playback.c is already using it.
08:29:11 Join ramon8 [0] (
08:29:54amiconnEh, one single thread...
08:30:07amiconnDoes that make sense to have on hwcodec then?
08:30:38 Quit Rob2222 (Read error: 110 (Connection timed out))
08:30:55jhMikeSHow much thread variation and behavioral difference should be tolerated?
08:32:09jhMikeSFor plugin devs they'll constantly have to watch their back for it and code around it here and there. A wait to condense the function call nicely would save more than the 68 bytes that takes.
08:33:07 Join Thundercloud [0] (
08:33:34amiconnI don't think having a struct instead of several arguments will save code size. Then the struct needs to be filled with values instead of the stack
08:35:47amiconnHmm, if the struct is pre-filled (static const) it might. Maybe that should be tested...
08:36:19*amiconn started a compile cycle
08:36:19 Join henfon [0] (n=ircap8@
08:36:20jhMikeSexactly. I saved a ton doing that with recording and most (all) threads are create with static parms.
08:36:47 Part henfon
08:37:29jhMikeSaudio_set_recording_options was nearly always used with global_settings values only. so one function to fill in the struct and just modify the ones that are different. worked nicely.
08:37:31 Quit BigBambi (Read error: 110 (Connection timed out))
08:37:42amiconnDidn't you say you moved thread structs from iram to dram?
08:38:07jhMikeSnot the thread structs. switch_thread itself.
08:39:00jhMikeSdual core _must_ have them in IRAM of couse. switch_thread should be there as well because cache fills could throw off the time-critical cop/cpu mailbox code
08:40:31jhMikeSI realized this little thing later anyway.
08:40:45amiconnWell, on single core it doesn't make that much sense to have it in iram. Thread switching is used constantly, but the time spent in there is way less (or at least should be) than in the actual application code
08:41:19amiconnSo the precious iram should be saved and used for tight, time critical loop stuff
08:41:49jhMikeSfar, far, less, yes though data is more commonly accessed than any particular function
08:42:22amiconnOn coldfire it makes sense to keep the thread structs in iram as it has no data cache, but on SH I would probably even put the structs in dram
08:42:46*amiconn will need some spare iram soon(ish) on SH
08:43:26jhMikeSI did arrange them to be both packed and aligned by sorting member size so no gaps should exist within them.
08:44:48 Join chris_mt [0] (
08:45:18jhMikeSI just needed this big stuff taken care out some details later and refine it. Somehow it's easier to see what to do after getting a break.
08:46:07JdGordonhehe.. commit and abandon == a break ?
08:46:27amiconnI doubt that's what he meant
08:46:34*JdGordon was joking btw
08:47:06jhMikeShaha...when's the last time you saw me do that on any big project?
08:47:30 Join parafin|away [0] (
08:48:38jhMikeSJdGordon: if you were referring to switching cores as "the thread x switches cores all by itself problem". No it's not that. A thread can hop deliberately.
08:50:11amiconnNow we need dual core on PP5002...
08:51:49JdGordonjhMikeS: yeah, i did mean automatically switching, but thats probably not so good anyway
08:51:52Llorean And some code that takes advantage of it for playback. :-P
08:52:13JdGordonis there any limitation to what threads can go on the cop?
08:52:20GodEater_multi-core doom!
08:52:20amiconnIsn't that what this commit was mainly about?
08:52:30JdGordonI mean, is there any reason to not put scolling, backlight onto the cop to free up the cpu a bit?
08:53:11 Join bertrik [0] (
08:53:33jhMikeSJdGordon: no techincal limit
08:53:48amiconnI'd rather keep most threads on cpu, and only put a few selected, computing intensive ones on the cop
08:53:57jhMikeSJust extreme caution
08:54:56 Quit midkay ("Leaving")
08:55:49jhMikeSJdGordon: if you kill a cop thread from the CPU in Debug OS Stacks, the main thread hops to the COP to kill the thread and hops back. that's just a minor use though.
08:56:20JdGordonthat sounds reasonable
08:57:28amiconnHmm, codec thread is on cpu still...
08:57:50 Quit shnee ("Konversation terminated!")
08:57:59JdGordonand moving it across causes playback to crash
08:58:27 Join ender` [0] (
08:59:24JdGordonbacklight is fine on the cop
08:59:28JdGordonscrolling doesnt work though
08:59:34*JdGordon assumed that thread was self contained?
08:59:54amiconnScrolling has a tick task that sends messages
09:00:25amiconnAnd accessing the framebuffer from both cores is a bad idea
09:00:45JdGordonyeah, i tihnk ive seen some artifacts because of that presumably
09:01:19jhMikeSplayback need some real thread sync before you can plop codec on COP
09:03:07jhMikeSlcd drivers can be secured. really mutexes are fast now and should really be used. spinlock should become a dual core true spinlock which can be used to sync interrupts and threads on both cores at the same time.
09:04:01amiconnlcd would have to flush cache all the time...
09:04:15amiconnAnd we don't want the framebuffer to be uncached for sure
09:07:21jhMikeSwell, it can be accessed as both. even switched by changing the base on the fly if some plugin wants to use it that way.
09:07:49amiconnI cannot imagine why we would want that
09:08:01amiconnUncached access is slower by factors
09:08:45jhMikeSmpegplayer didn't seem to suffer for using the uncached address
09:08:58amiconnFor code I measued a factor of around 7 iirc
09:09:20amiconn(running a tight loop from sdram with cache disabled vs. cache enabled)
09:09:24jhMikeSI think MrH's experiments revealed code is far more imporant to cache than data.
09:09:44chris_mtby uncached do you mean writing directly to lcd_framebuffer?
09:10:03amiconnNo, writing to its uncached alias address
09:10:04jhMikeSlcd_framebuffer + 0x10000000 vs. just lcd_framebuffer
09:10:56chris_mtso which one is cached? i used lcd_framebuffer for a patch to the ines port but it would be great to speed it up further
09:11:24***Saving seen data "./dancer.seen"
09:12:45jhMikeSlcd_framebuffer is cached
09:12:59chris_mtI see.
09:13:51chris_mtI haven't looked at the driver code - does this then get moved to the real frame buffer in batches?
09:14:54jhMikeSin whatever rectagle is specified by lcd_update_rect. it's not batched.
09:14:55amiconnColdfire buffers so incredibly fast compared to PP...
09:15:13jhMikeSH10 is horribly slow at buffering
09:15:45amiconnjhMikeS: small h10 became a lot better due to the fast lcd driver, especially when using a wps with peakmeters
09:16:19amiconnI think the peakmeter fps needs to be adjusted to the target's lcd. Slow lcd updates needs less fps in peakmeter, or it will take forever
09:16:27jhMikeSbig H10 took forever since I've had the thing.
09:17:00amiconnBuffering on 2nd gen with default wps takes almost one minute with default wps, but around 20s with iCatcher
09:17:07jhMikeSwe must be leaving some clock running too slowly
09:17:17amiconnOn small H10 I now get 20-ish seconds even with peakmeter
09:18:19amiconnBut on both H180 and H340, buffering takes around 6 seconds for mp3...
09:18:46jhMikeSsounds reasonable. about 12-ish on big H10 though
09:18:58amiconnI think we will get close to that when we can put codec on the cop
09:19:16amiconnI think what we see here is just the difference in mp3 decoder efficiency
09:19:45 Join petur [0] (n=petur@rockbox/developer/petur)
09:19:48jhMikeSI think SWCODEC peakmeters should just wait on the queue and not sleep()
09:19:51 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
09:20:12amiconnThen they would suck even more power...
09:20:34amiconnOr maybe I misunderstand what you mean... ?
09:20:36jhMikeShardly, they'd spend 1/20 sec doing nothing. I could test this.
09:20:59amiconnHmm, but sleep() achieves the same, no?
09:21:32jhMikeSJust wait on the queue for the entire peakmeter frame duration instead of polling it and doing sleep()
09:21:42 Join ddalton [0] (
09:22:00 Part ddalton
09:24:55amiconnSWCODEC just needs to be highperf always false
09:26:08amiconnBoth the highperf business and polling every single tick (as a compromise) is only needed on hwcodec
09:26:19jhMikeSthen you'll still make 40x more trips through the threading code every second
09:26:39jhMikeSor no it's 1/20 second
09:26:58jhMikeSso 10x more
09:27:35jhMikeSthe get_action + the sleep call each loop
09:27:39 Quit bertrik ("bye")
09:28:48amiconnI'll try that anyway - just a switched #if and a changed condition...
09:29:55amiconnIt's in fact not 10x or 5x more
09:30:06amiconnThe big loop is just executed twice
09:30:39amiconn...because next_big_refresh is HZ/10 and next_refresh is HZ/PEAK_METER_FPS which is HZ/20
09:31:11jhMikeS button = get_action(CONTEXT_RECSCREEN, HZ/PEAK_METER_FPS);
09:34:22*jhMikeS wonders why /player and /recorder still exist :)
09:35:11 Join davina [0] (
09:45:05 Quit chris_mt ("ChatZilla [Firefox]")
09:45:24 Join CaptainSquid [0] (
09:47:40 Quit Mouser_X (Read error: 110 (Connection timed out))
09:51:12 Join qweru [0] (
09:52:26 Join ivoreus [0] (
09:53:12 Quit ivoreus (Client Quit)
09:54:32 Join safetydan [0] (n=safetyda@rockbox/developer/safetydan)
09:55:27 Quit qweru (Client Quit)
09:59:34 Quit atsea-34 (Read error: 104 (Connection reset by peer))
09:59:44amiconnGot an undefined instruction on 2nd gen :(
10:00:35amiconnNow a prefetch abort...
10:01:25amiconn...a stkov...
10:02:18jhMikeSstkov where?
10:02:23JdGordonyouve alsmot got the whole set!
10:02:48amiconnjhMikeS: It's strange - it just showed an address (iram area) instead of the thread name...
10:03:13jhMikeShmmm...idle stack. those should be big enough
10:03:13amiconnThe undefined instruction was in iram as well
10:03:44amiconnThe prefetch abort was at 0xc0000002 - no code should ever jump there
10:04:13jhMikeSof course not
10:04:49jhMikeSwhat was running to get that one?
10:05:29amiconnI just booted. It happened at the end of the dircache scan
10:06:10amiconnAnother crash ("just" freeze this time)
10:06:17jhMikeShmmm...could be one of those hidden overflows?
10:06:34jhMikeSor is swp broken on that model too?
10:06:38amiconnStrange - it worked at the first boot with latest svn, but now it crashes every time :(
10:06:45 Quit rasher (Read error: 145 (Connection timed out))
10:07:11jhMikeSafter running dircache?
10:07:52amiconnAs long as dircache is scanning, I can scroll around the browser. As soon as dircache is done, it freezes or crashes
10:09:37amiconnNow I forced a foreground scan by removing nvram bin
10:09:53amiconnAFter the foreground scan I got an undefined instruction :( another one
10:10:48jhMikeSDoes not scaling affect anything?
10:10:51amiconnIt's unusable :(
10:11:11jhMikeSwhich processor is that?
10:12:28jhMikeSI'll double check things for a stupid mistake.
10:12:34 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
10:13:53 Join spiorf [0] (
10:14:18jhMikeSwhat if dircache is shut down?
10:16:15amiconnThen it crashes right away on boot
10:16:24Zagormeh, usb docs are soo vague.
10:16:30amiconnThe addresses are different each time
10:16:31 Join obo [0] (n=obo@rockbox/developer/obo)
10:16:57Zagorwould you say this means in/out can or can not have the same endpoint number? "A stream pipe to a device is bound to a single device endpoint number in the appropriate direction (i.e., corresponding to an IN or OUT token as defined by the protocol layer). The device endpoint number for the opposite direction can be used for some other stream pipe to the device."
10:17:24amiconnI'd say it can
10:17:42LloreanSounds like a "can" to me, because it sounds like it's saying you can choose not to, suggesting you don't have to choose not to.
10:17:43amiconnAfaik, the IN endpoints are numbered independently of the OUT endpoints
10:17:54 Join bluebrother [0] (i=5SXq1MdZ@rockbox/staff/bluebrother)
10:18:10peturbut I've never seen them with the same number...
10:18:42petur(windows & windows CE)
10:19:18ZagorLlorean: yeah that's what I'm reading too. and in another place it says endpoints are uniquely identified by their number _and_ their direction.
10:19:37Zagorbut as you say, no devices seem to do that. so I'd better go with the reality instead of over-interpreting the map...
10:19:55*amiconn recompiles the core with sw corelock...
10:20:17LloreanYeah, de facto tends to trump de jure.
10:20:50amiconnjhMikeS: Hmm, could a broken swp influence PP5002 at all?
10:20:58amiconnWe don't run dual core yet...
10:21:36*amiconn found that his change for 5002 in config.h was useless...
10:21:44jhMikeSamiconn: it I made an error
10:22:20amiconnDidn't you want to get that firewire card? ;)
10:23:37jhMikeSyeah, it's about time. I think I can fix this now though.
10:25:20 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
10:29:16pondlifeLlorean: Do you think it'd be a good idea to lock one of the two "Nano not working" threads? They may have started out as different symptoms (audio glitch vs crash?) but that seems to have been lost now.
10:31:45amiconnDisabling highperf in the swcodec peakmeter cuts buffering time in half on G5 with peakmeters enabled
10:31:51 Quit spiorf (Remote closed the connection)
10:31:55 Join ddalton [0] (
10:32:14 Part ddalton
10:32:33 Join GregoryHouse [0] (
10:33:53jhMikeSamiconn: See if that does it?
10:33:55GodEater_pondlife: I was thinking the same thing this morning
10:34:25pondlifeWell, go for it.. I'd do it myself but I've not been involved in the discussion so it seemed a bit heavy-handed.
10:34:42GodEater_no - I think I'll let Llorean do it - he's been *more* involved
10:34:55GodEater_he can decide which one can go
10:35:37jhMikeSjust made a commit. SVN up and see if that clears it. no swp unless specified.
10:38:19amiconnJust 18 seconds buffering time on G5 with default wps now
10:42:03 Join random_desu_is_s [0] (
10:42:19pondlifebluebrother: Don't suppose you could provide an updated RbUtilQt for Windows that I could test?
10:42:38bluebrotherpondlife: sure, gimme some minutes
10:42:40 Quit random_desu_is_s (Client Quit)
10:42:42pondlifeNo rush
10:44:38 Join random_desu_is_s [0] (
10:44:49 Quit random_desu_is_s (Client Quit)
10:45:10amiconnjhMikeS: Looks like it's working now...
10:45:48amiconnHowever, in one try I managed to start playback and it did only output silence. WPS was moving etc. Stopping and restarting playback fixed it
10:46:22jhMikeSguess that answers that for swp on PP5002
10:46:48*jhMikeS goes to the FW card store
10:48:32Zagorthey're rather niched
10:49:39amiconnjhMikeS: It's broken?
10:51:30 Quit Soap (Read error: 110 (Connection timed out))
10:51:31jhMikeSamiconn: if it's working now, it sounds like it. That's the only change there.
10:52:17amiconnOkay, so PP5004 fixed the cache bug of the 5002. 5020 added new stuff. 5022 fixed swp
10:52:55amiconnI wonder what bugs there were in PP5000 (which presumably existed) then...
10:52:57 Quit CaptainSquid ("Miranda IM!")
10:53:10*jhMikeS wonders if ARM told them they're not going to do any more half-a$$ed core hardening work for PP
10:53:40 Join lee-qid [0] (
10:54:23jhMikeSthere was a PP5000? it's probably just a bug all over.
10:57:20*amiconn likes it when a bug fix means a green delta...
10:58:08 Join aliask [0] (
10:59:23*jhMikeS actually prefers those as well
11:01:01bluebrotherpondlife: at the usual location ...
11:04:40 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
11:05:03pondlifebluebrother: Thanks
11:05:25 Quit pondlife ("disconnected has pondlife")
11:11:27***Saving seen data "./dancer.seen"
11:13:27markunanother speedup?
11:13:30*amiconn just found that the jpeg viewer now has dithering
11:14:15amiconnmarkun: No, not yet, but there will be a couple soon, hopefully
11:14:21*jhMikeS put that in there oh so long ago :p
11:15:15amiconnjhMikeS: It's probably because of that menu which always annoys me so that I just click 'Quit' immediately without reading further
11:15:41markunjhMikeS: did I read something in the logs about buffering being faster on the Gigabeat? Or is that not committed?
11:15:50amiconnWe really should make the exit button actually exit plugins, and put the menu on a separate button, preferably the usual menu button
11:16:11jhMikeSmarkun: it's just faster because of the scheduler.
11:16:46 Join ddalton_ [0] (
11:16:47jhMikeSnot hugely but there's a small improvement
11:16:59ddalton_I keep forgetting but how do I log ddalton off?
11:17:24linuxstbddalton_: /ns ghost ddalton password
11:17:58 Nick ddalton_ is now known as ddalton (
11:18:13ddaltonok thanks
11:19:10jhMikeSamiconn: r12987 added the gray and dither stuff to jpeg.c
11:21:15 Quit idnar (Nick collision from services.)
11:21:17 Join idnar_ [0] (i=mithrand@unaffiliated/idnar)
11:22:47amiconnThe test_fps results for 1st gen and 2nd gen are interesting. First one would expect exactly the same results, as they have the same CPU and LCD controller
11:23:32 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
11:23:35amiconnBut the 1st gen is ~0.5% slower. These are the 0.5% sacrificed to the 'Wheel Power God' on 1st gen...
11:24:17jhMikeSWheel Power God?
11:24:36amiconnThe mechanical wheel on 1st gen draws quite some power when enabled (12 mA)
11:25:00amiconnIt is switchable (via gpio).
11:25:35amiconnSo we just enable it for a short time every tick in order to check whether it moved, and only keep it enabled when it moved
11:26:01amiconnWhen hold is enabled, the whell is kept disabled of course
11:26:24jhMikeSno int is available to activate it when needed?
11:26:40amiconnHow would the wheel fire an interrupt when it's disabled?
11:27:28amiconnThe button+wheel driver uses interrupts for the actual readout
11:27:34 Quit bluebrother ("reboot")
11:27:38jhMikeSwouldn't nescessarily come from the wheel controller itself
11:28:01markunjhMikeS: was the scheduler improvement committed?
11:28:33jhMikeSmarkun: you missed that enormous things on the from page? :)
11:28:49amiconnThe wheel has no controller.
11:29:23jhMikeSguess it disappeared from there already
11:29:39markunjhMikeS: the multicore commit?
11:30:23amiconnIt's probably just a LED, 2 photo transistors and a sectored wheel
11:30:42amiconn...working like a opto-mechanical mouse
11:31:02jhMikeS12mA sounds like a LED draw
11:31:35amiconnThe wheel needs a minimum time to get stable readout after enabling it
11:31:51markunjhMikeS: didn't miss that one :)
11:32:09amiconnMy tests got stable readings after 20..30 us. I settled for 50 us to be on the safe side
11:32:19amiconnThat's the 0.5% - 50 us per tick
11:32:25barrywardellso I think I can either just include the shift and shift all values in powermgmt-h10.c correspondingly
11:32:26jhMikeSyou'd think some kind of hall-effect sensor would have been used
11:32:31barrywardellor forget about the shift altogether
11:33:34jhMikeSamiconn: does OF just poll it?
11:33:49amiconnI didn't check, but I doubt it
11:34:08amiconnFor getting the actual movement, polling would always be too slow
11:34:22amiconnWe use interrupts for this
11:34:34barrywardellthe h10 scrollpad sounds similar, but we know when to enable it because a gpio bit gets set
11:35:18amiconnPolling misses steps when the wheel is moved fast, and then it loses direction
11:35:43amiconnWe had that before I switched to using interrupts... which required to stabilize clock scaling (sic)
11:37:34amiconnbarrywardell: That shift is a strange thing... it means that zero isn't true zero
11:37:54 Join rasher [0] (n=rasher@rockbox/developer/rasher)
11:38:01jhMikeSe200 had that problem before interrupts. It could have even moved the list in the wrong direction if turned too quickly.
11:38:06amiconnI wonder what state delivers the correct zero value - pll disabled or pll enabled
11:38:22amiconnjhMikeS: yup, exactly that thing...
11:38:26barrywardellpll disabled gives steadier readings
11:38:58amiconnI don't mean stability, but what situation gives a zero reading when the input voltage is zero
11:40:40barrywardellyeah, I know. but if it's more unstable with pll disabled, then maybe it gives a wrong zero reading too
11:41:08amiconnI think we should apply it if it is a true offset
11:42:22barrywardellit's also possilbe that I just couldn't see the change since the range of voltages is quite small compared to the total
11:42:29barrywardelland the readings are still jumpy
11:43:06 Join CaptainSquid [0] (
11:43:13jhMikeSamiconn: I think the reversal is called "aliasing". Nyquist sampling applies there as well.
11:43:35amiconnah, yes
11:43:36 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
11:44:01barrywardellany objections to changing a couple of the defines in pp5020.h, like so:
11:44:51amiconnLooks ok to me
11:45:50 Join Redbreva|work [0] (i=c1713011@gateway/web/cgi-irc/
11:46:25 Join jba [0] (
11:48:40 Quit jba (Client Quit)
11:48:43 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
11:48:55jhMikeSyeah, always update that as meanings become clearer
11:50:41pondlifeAnyone object If we raised the max playlist size from 20000 to 30000? No impact on existing users, is there?
11:50:50 Quit Redbreva|work ("CGI:IRC (Ping timeout)")
11:51:29amiconnWho would use such insanely long playlists??
11:51:34jhMikeShow bout 32768 so it's a nice round number?
11:51:39*GodEater_ was thinking that
11:51:59rasheramiconn: shufflers
11:52:00amiconnThere is no impact on existing users as the actual limit, and hence the reserved ram, is a user setting
11:52:25pondlifeamiconn: I'm insane right now, then :)
11:52:30amiconnrasher: How? Putting the whole contents of the player multiple times into the same list?
11:52:43pondlifeNo, only once
11:52:48pondlifeDatabase > Tracks
11:53:09amiconnI can't fit more that ~3000 tracks on a 20GB target
11:53:36rasher1) A 80gb target exists 2) not all songs are multiple megabytes
11:53:40amiconn(with usable quality, that is)
11:54:00pondlifeLots of short 128k MP3s here
11:54:04amiconnOkay, maybe with sid or so... but then you'll run into dircache trouble ;)
11:54:28amiconnDo you use dircache?
11:54:29pondlife21591 tracks here, what's the dircache limit?
11:54:37amiconnThe dircache limit is 6MB
11:55:00jhMikeSyeah, I want my HVSC in one playlist
11:55:04amiconnHow many files fit there depends on average dirname/filename length
11:55:25amiconnCheck dircache status in the debug menu to see how much is taken
11:55:34pondlifeSo dircache quietly fails?
11:55:42jhMikeSuse 8.3 filenames internally?
11:55:51amiconnIt fails, but then tries to scan every boot
11:56:20amiconnjhMikeS: No, that wouldn't work reliably
11:56:21jhMikeSoh were talking dircache, heh. not database.
11:56:36 Join miepchen^schlaf [0] (
11:56:57*jhMikeS is a bit exausted
11:57:02pondlifeHmm, indeed, it's not working here ion sim... but no sign of a scan taking place either
11:57:05pondlifeon sim
11:57:54amiconnI don't know what dircache is supposed to do in the sim
11:58:07pondlifeSame as the target, I'd have thought
11:58:16jhMikeSdoes dircache hold paths in a tree by component?
11:58:37amiconnjhMikeS: I think it does, but I'm not sure. It's Slasheri's work
11:59:17amiconnI have ~6800 files on my H180, and dircache takes ~430KB
11:59:57jhMikeSI wonder about other data structures that might effectively compress it nearly no overhead
12:00:12jhMikeS*with nearly
12:02:45 Nick fxb__ is now known as fxb (
12:05:08jhMikeSamiconn: I seem to recall you saying you didn't use dircache.
12:09:24 Join spiorf [0] (
12:09:32 Quit spiorf (Read error: 104 (Connection reset by peer))
12:10:27 Join spiorf [0] (
12:17:12*preglow does the big commit dance
12:17:56amiconnjhMikeS: Meanwhile I do, but only on targets with >= 32MB ram and slow spinup
12:18:20amiconn(the latter means that I don't use it on ipod mini, as its microdrive needs only 500ms spinup time)
12:18:52jhMikeSpreglow: you missed the party
12:18:59preglowjhMikeS: any fun?
12:19:13amiconnDircache doesn't help much when using directory talk clips though
12:19:42jhMikeSamiconn: I definitely wouldn't bother using it on sansa
12:20:13jhMikeSpreglow: lots of fun...except I didn't score so well :(
12:20:55JdGordonthe only reason to enable dircache on sansa is for the auto update for database
12:21:01pondlifeamiconn: You still get crashes using dircache
12:21:03JdGordonZagor: success?
12:21:12Zagorbulk transfer is alive
12:21:13pondlifeZagor: gogogogogogo!
12:21:13amiconnpondlife: Didn't get any recently...
12:21:29amiconnZagor: What was the problem?
12:21:33Zagorturns out the controller was picky about when the registers were inited. I couldn't do them all in the beginning.
12:21:44pondlifeamiconn: Just wondered, I never had that problem.
12:21:47amiconnAh, init order matters...
12:21:53JdGordoncongrats Zagor :)
12:22:01amiconnReminds me of the H300's otg controller...
12:22:20Zagornot only order, but I have to wait for the connection interrupt before configuring bulk endpoints
12:22:22jhMikeSa common afliction datasheets neglect to ever talk about
12:22:35pondlifeHmm, maybe we'll be revisiting the H300 OTG at some point.... ;)
12:23:02preglowZagor: much work needed to make this stuff work on ipods too?
12:23:12jhMikeSGigabeat supports OTG too.
12:23:21amiconnjhMikeS: The data sheet in fact mentions those limitations, but it wasn't very clear. In case of the OTG controller it is that not all status transitions are valid
12:23:29preglowjhMikeS: thunderbird labeled your commit mail as spam :P
12:23:33markunjhMikeS: well, USB host at least
12:23:35Zagorpreglow: I don't know what/if the differences are. those with the same controller should need very little modification
12:23:41amiconn...and if you try one that isn't allowed, it's simply not executed...
12:23:45jhMikeSpreglow:'s probably right :P
12:23:46markunnot sure if we can do the OTG auto negotiation
12:23:51preglowZagor: i thought most of the usb stuff was on the pp chip
12:24:01Zagoryeah I mean those with the same pp as the sansas
12:24:20Zagoror perhaps all PP work the same regarding usb. I just don't know.
12:24:26preglowi don't expect they've changed that too much
12:24:30amiconnpreglow: On pp502x it should be all the same. Maybe the 5020 has a bug that's fixed in 5022 though
12:24:31jhMikeSmarkun: for gigabeat?
12:24:42*amiconn really doesn't trust pp anymore in that respect
12:24:43preglowamiconn: wouldn't surprise me...
12:25:09preglowat least we know what register accesses to look for in disassemblies
12:25:26amiconnpreglow: That would mean it should be easily ported to mini G2, Nano and Video. G4, Color and Mini G1 could be problematic, but hopefully just work too
12:25:28markunjhMikeS: yes
12:25:43jhMikeSmarkun: why not? (not that I know atm what that is)
12:25:54preglowthis will be so cool
12:26:00amiconnG3 is a whole different thing, as (1) the PP5002's internal USB controller is different and even (2) it's not used, because it would be USB1.1 only
12:26:15amiconnThe G3 has an external USB controller which is not (yet) known
12:27:43*barrywardell congratulates Zagor
12:27:52markunjhMikeS: I'm not too familiar with OTG myself, I'll read up on it a bit
12:28:13jhMikeSThe tetrahub seem to handle that stuff
12:28:30 Join bagawk_ [0] (n=lee@unaffiliated/bagawk)
12:28:31markunbut I'm more interested in normal USB host and device support than supporting OTG
12:29:11amiconnHmm, doesn't the gigabeat already have device support, as that is a hardware bridge?
12:29:25jhMikeSI think a good deal is just handled by the tetrahub from an i2c ROM and we just handle packet on target.
12:30:45*amiconn is undecided what to try next :/
12:31:01 Quit Thundercloud (Remote closed the connection)
12:31:02 Join japc [0] (n=japc@
12:31:35markunamiconn: sure, but we could use the device also for other things besides UMS (audio playback, serial port for debugging)
12:32:12amiconnAudio playback?
12:33:39 Quit bagawk (Read error: 110 (Connection timed out))
12:33:41markunwell, maybe not so useful
12:34:25markuna friend was working at an office where the PC's didn't have a sound card
12:34:43amiconnAH, using it as a sound card
12:34:48markunperhaps a Gigabeat could be used as a USB sound card
12:34:51markunyes, sorry
12:35:09amiconnFor some reason I was thinking about the opposite - connecting a pair of USB speakers to a rockbox target
12:35:22markunwhy not :)
12:35:39amiconnUsing it as a sound card would also allow using it like streamripper
12:36:56 Nick idnar_ is now known as idnar (i=mithrand@unaffiliated/idnar)
12:37:15markunare there USB clocks, so we can sync the time? :)
12:37:34 Join bluebrother [0] (i=CCblBtlP@rockbox/staff/bluebrother)
12:37:48jhMikeSgigabeat could record using an external adc
12:38:25markunif only we could find some foxconn connectors :(
12:38:45amiconnWe could also implement MTP ;)
12:39:01barrywardelland the ipod protocol :P
12:39:01amiconn^^ _BAD_ joke
12:39:16markunamiconn: you know some people will ask for it ;)
12:39:22*bluebrother still wants a coffee plugin *g*
12:39:25LloreanI guarantee we'll get feature requests for MTP at some point.
12:39:25jhMikeSadd that to the 3.0 realease TODO list
12:39:27maxkelleyYay drm!
12:39:40LloreanEasy syncing with WMP and other such things, and a host aware of the formats the DAP can play, etc.
12:39:54markunjhMikeS: maybe we should skip 3.0 and just release 4.0
12:40:04LloreanI think were MTP not so closed, it might not be such a bad idea.
12:40:18jhMikeSsure - 3.0 was a NULL release :)
12:40:34bluebrotherwenn do we release 3.0.1?
12:40:42jhMikeSafter 4.0?
12:40:49bluebrothergood idea.
12:40:53pixelmaLlorean: but then you also need to implement building the itunes database... ;)
12:41:17Lloreanpixelma: Building the iTunes database, no thanks. Transparently building ours as files are synced to the device? Maybe.
12:41:20jhMikeSRhapsody support too
12:41:57bluebrotherand ITM support :)
12:41:59*amiconn still fails to understand how MTP / itunes could be "easier" than plain UMS access
12:42:07jhMikeSIf it exists, we're _obligated_ to support it :)
12:42:10preglowdepends what you're used to
12:42:26*preglow likes the stack usage debug screen :>
12:42:34bluebrotheroh, and I'm _desparately_ missing proprietary sound effects.
12:42:37Lloreanamiconn: With the proper host application, MTP "just works". You put your CDs in, say "Do your thing" and don't need to think. Perfect for people who wish they didn't own a computer, but want to listen to music anyway. :)
12:42:54preglowjhMikeS: wouldn't "View OS Stacks" be better renamed to "Thread list" or something now?
12:42:56jhMikeSpreglow: don't go killing threads willy nilly now
12:43:25amiconnYeah, thread list sounds better
12:43:30preglowwhat's this idle business?
12:43:33jhMikeSpreglow: I agree, but now there's two items that aren't threads
12:44:01preglowthat's what i'm asking about, i guess
12:44:36preglowand how do i kill a thread? :P
12:44:50 Quit linuxstb ("Leaving")
12:44:51jhMikeSpreglow: 1) when COP threads exit plugins and the COP idles, it continues to idle on the stack of the last thread that ran. problem is, the plugin in not technically loaded anymore.
12:45:13jhMikeSyou need to enable it in debug_menu.c
12:45:28preglowthink i'll just not do that, then
12:45:30jhMikeSpreglow: they're also the stacks used when switching cores
12:45:43preglowok, so the percantage is still just plain old stack usage?
12:46:06 Join petur2 [0] (n=petur@rockbox/developer/petur)
12:46:09 Quit petur (Nick collision from services.)
12:46:09preglowpretty easy to interpret that as cpu idle percenage...
12:46:11 Nick petur2 is now known as petur (n=petur@rockbox/developer/petur)
12:46:23jhMikeSif you don't use idle stacks when switching cores, both cores will be using the same stack at the same time for a bit...crash
12:46:46jhMikeSrename it "Idle Stack"?
12:46:51preglowsounds good
12:47:23jhMikeSI thought it would be understood by the context though but that's np
12:47:46preglowwell, context currently is that the screen is named "stack view", but mostly seems to just display thread info
12:47:53preglownot too clear
12:48:49preglowwhat does T status mean?
12:49:29jhMikeSBLOCKED_W_TMO more accurately
12:50:36preglowis rockbox capable of blocking on read(), or does it pretty much only block on queue reads?
12:50:58jhMikeSblocking on read?
12:51:25preglowlike other oses do when reading from a network socket, for example, i don't really know if we use read() for any fancy stuff or do stuff explicitely
12:51:29jhMikeSyou mean background I/O?
12:51:38Nico_PjhMikeS: are you intending to change the SVN playback code to make it use the new sync features ? It would be a waste of time IMHO
12:52:06jhMikeSNico_P: I only did things to be compatible there
12:52:44jhMikeSmutexes cannot be used in the manner they were for codec swapping. semaphores have to be used and are more proper for that.
12:53:32Nico_PjhMikeS: and now you recommend using mutexes and semaphore to sync things as much as possible ?
12:54:46jhMikeSas much as nescessary. no more of course.
12:55:14 Join Thundercloud [0] (
12:55:31preglowthey pretty much should be used where they should be used, i guess, when commicating between threads :>
12:55:33jhMikeSfor threading to be advantagous, you really need to sync as little as possible
12:55:40preglowhello preemptive threading! :D
12:56:07*jhMikeS hugs preemtive threading
12:56:52 Join homielowe [0] (
12:58:44Nico_PjhMikeS: typically, in the playback code, what would need sync ?
12:58:59preglowjhMikeS: lemme guess, you don't use an editor with a 80 char wide window? :>
13:02:55jhMikeSpreglow: um nooo
13:03:32jhMikeSNico_P: not sure specifically right now. sort of depends on the MoB layout and stuff.
13:04:11Nico_Pyeah of course
13:04:33jhMikeSI would think the handle allocation would
13:04:58Nico_Phow ?
13:05:00jhMikeSIsn't that called directly by many threads?
13:05:17Nico_Pyes, that's true. two threads could call bufopen at the same time
13:06:03Nico_Pcurrently the only user is the audio thread though
13:06:37jhMikeShow does the WPS get the metadata?
13:07:08Nico_Pin most cases it accesses the static struct for the current or next track
13:07:17Nico_Pin some cases it might use bufgetdata
13:08:55amiconnjhMikeS: I have a question regarding your comment here:
13:08:59amiconn(all the way down)
13:09:34amiconnI think mac and msac are equally fast, and there is just a mistake in the timing table in the mcf52xx user manuals
13:10:20preglowjhMikeS says he's done tests
13:10:23jhMikeSI actually tried msac in the SPC codec and the boost change was quite high
13:10:23preglowi thought it was just a mistake too
13:10:48amiconneh? Really odd then
13:11:08preglowit's really weird if that is so, it's pretty unheard of in dsp circles for those two to have different timinig
13:11:12amiconnmac.w, mac.l, msac.w are all listed as single cycle
13:11:13jhMikeSjust negating one parameter and using mac was faster. I could verify again by changin two instructions.
13:11:28***Saving seen data "./dancer.seen"
13:11:42amiconnand mac.w, mac.l msac.w and msac.l with parallel load are all listed 2 cycle
13:11:56amiconnOnly msac.l (without parallel load) is listed as 3 cycles...
13:12:54jhMikeSI'll check again and make sure nothing else factored in. I have a hard time accepting it myself.
13:13:19amiconnOtoh, the mcf5249 and scf5250 manuals agree on this... weirdo
13:14:03jhMikeSI'll make sure I try a full 8-voice SPC so that it gets a nice workout
13:14:17preglowwow, that new mpegplayer menu was annoying
13:14:17amiconnjhMikeS: Btw, the yuv conversion for cf also uses emac saturation for clamping
13:14:46amiconn(for quite a while now, but doing two-line zig-zag is even better :) )
13:14:59Lloreanpreglow: I agree. It could very much use banishing.
13:15:57 Quit advcomp2019 ("Leaving")
13:16:02jhMikeSamiconn: it makes nice quick work of that indeed
13:16:39preglowjhMikeS: seems to me mpegplayer is more stable after your commit
13:16:42jhMikeSpreglow: I don't know why it shouldn't just play something like a normal viewer
13:17:04amiconnWell, if there's something to resume, it should ask
13:17:17amiconnBut _only_ then
13:17:46LloreanYeah, if there's a resume point, some button that's not "Select" or "Right" should cancel, and any other button resume.
13:17:51jhMikeSit should ask about resume when exiting if anything unless it plays to the end
13:18:27jhMikeSpreglow: are you saying it's been unstable still?
13:18:38preglowjhMikeS: i've had some weirdness with hanging
13:18:45preglowah, no
13:18:47preglownot unstable now
13:18:50preglowworks just fine
13:19:06preglowi don't get why i can only seek at the start menu...
13:19:12jhMikeSodd...could be the fact the kernel object are actually dual-core safe
13:19:30 Join jac0b [0] (
13:19:36LloreanI never did ask why the seeking is in 30 second increments.
13:19:44markundid any of the mpegplayer seeking implementors even visit irc?
13:19:45jhMikeSwhy? :)
13:20:14jhMikeSwhy should they...they're so 1337
13:20:19safetydanwow, mpegplayer is almost smooth on the h120 now
13:20:26pondlifeWhat FPS?
13:20:52safetydanpondlife, can't see it. I think graylib interferes with printing the fps
13:20:54markunpreglow: I also don't understand why they implemented it in this way
13:21:22jac0bI have a question about posting to the wiki.
13:21:23markunand duplicating code to generate the thumbnail
13:21:28preglowsome discussion on the matter would be sweet, they seem to have done good work
13:21:40markunjac0b: thanks for all your battery stuff btw!
13:21:47markunwhat's your question?
13:21:48 Quit jhMikeS (Read error: 104 (Connection reset by peer))
13:22:11markunjac0b: btw, whas your last 26 hour benchmark without replacing the PCB?
13:22:23 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
13:22:56*preglow ponders making jpeg.c decode images in the background with another thread
13:22:59jac0bI was asked in the forum to post a how-to on the wiki, is that okay or should I make a post in the forum with the how-to
13:23:17LloreanThe wiki is for things like how-tos.
13:23:21markunjac0b: static info should always be in the wiki imo
13:23:31LloreanThe forum is, where possible, best kept for discussion and question/answer type things.
13:23:32jac0bmarkun: yes I didn't replace the PCB just the wires
13:23:44markunjac0b: I'll order one of those batteries as well then
13:23:59Isolinearpreglow: If you did that, how would it manifest itself when using the jpeg viewer?
13:24:13preglowIsolinear: basically not making you wait when zooming in some cases
13:24:17Nico_Ppreglow: excellent idea
13:24:25amiconnsafetydan: The graylib doesn't allow printing the fps the normal way while it's active, due to how it works
13:24:34Isolinearpreglow: That was my guess.. I say go for it. :)
13:24:41jac0bllorean: so posting the how-to in the wiki is better than posting in the forum?
13:24:44*preglow needs to learn to start finishing stuff
13:24:45amiconnThe fps would need to be printed using graylib functions
13:24:54Lloreanjac0b: Yes.
13:25:10Nico_Ppreglow: btw, maybe it would be possible to use the buffering API to store pictures in memory in advance
13:25:15safetydanamiconn, I was just going to print it to the lcd remote but that sounds better
13:25:15jac0bllorean: thanks
13:25:18amiconnpreglow: That will only work if there's enough RAM
13:25:25IsolinearWhile on the topic, why is it that my jpegs have a limited zoom when I view them with music playing?
13:25:25preglowamiconn: sure, it will obey the usual limits
13:25:35LloreanNico_P: For a slideshow like feature?
13:25:38amiconnsafetydan: The unbuffered graylib functions have no text output
13:25:44jhMikeSpreglow: on COP? :P
13:25:51Nico_PLlorean: yeah for example
13:25:55amiconnBuffered has them, but would be slower for mpegplayer
13:25:56jac0btoday I will be posting the how-to with the info on what battery I used
13:26:10preglowjhMikeS: well, why the hell not. if you've done your stuff correctly, there shouldn't be too much hardship involved :P
13:26:15jac0band how to put it together
13:26:16pondlifeHmm, mpegplayer is far from smooth on H300 sim.
13:26:46amiconnThe current graylib isn't that suited for moving content anyway, because (1) all the bit shuffling is slow and (2) the content change interferes with the greylevel generation mechanism, so it becomes grainy
13:26:52preglowonly cache stuff, lord oh lord, how i wish the cache snooped the bus
13:26:56jhMikeSpreglow: haven't had a problem myself with using it
13:27:03amiconn(not noticeable on H100 because of the dog-slow panel, but on others)
13:27:34amiconnI have plans for a way faster implementation with more greylevels, faster updates, and no graininess when changing content
13:27:39jac0bdoes anyone know if the gigabeat has a black front cover that is metal?
13:27:45jhMikeSand yeah, I hate that cache stuff. it's the worst part of this.
13:27:57preglowh10 and ipod video people needed for testing:
13:28:15markunjac0b: my friend's F10 has a black metal fron cover
13:28:28amiconnpreglow: Ignore that patch, we'll go asm soon
13:28:34safetydanamiconn, I'm curious how you'd do that as I thought the current routines were as fast as they could get on the h120 lcd controller?
13:28:46preglowamiconn: how soon is soon?
13:28:57preglowamiconn: i don't see any harm in commiting it unless we're going asm tomorrow
13:29:09amiconnsafetydan: The controller output has nothing to do with that. I'l switch from using planar data to 2 arrays of byte-packed data
13:29:28amiconnOne holds the actual image, and the other holds the current phase for each pixel
13:29:29jhMikeSif more instructions were squeezed out, then e200 and gigabeat should get that too.
13:29:30jac0bmarkun: well people with the plastic black covers might not be able to use the 4G iPod batteries in their player
13:30:05amiconnpreglow: Depending on how I proceed with the cf asm, it could be today for H10...
13:30:28safetydanpondlife, looks to be about 8 fps
13:30:59jac0bmarkun: my g/f has a black plastic front cover and I bent it when I put it in last nite but this morning the touchpad was non responsive again.
13:31:18markunjac0b: which player does she have?
13:31:24amiconnsafetydan: Practically, my new idea will do what halftone.c does when generating .RVFs, just live, and of course without using floating point :)
13:31:25preglowamiconn: sure, and what about the other targets in that patch?
13:31:28jac0bthe F10
13:31:56markunI could have sworn my fried's was metal, I can check when I visit him on thursday
13:32:04markunso the battery is a bit big?
13:32:36jac0byeah a bit
13:32:44jhMikeSpreglow: you must mean small H10?
13:32:54jac0byou know the black rubber feet on the OEM battery
13:33:06markunsame company, right?
13:33:08safetydanamiconn, I'm always amazed at what people can squeeze out of these things. that will be interesting to see
13:33:14preglowjhMikeS: and video and nano...
13:33:25*jhMikeS has big H10
13:34:09markunjac0b: no, don't really remember the rubber feet
13:34:10jhMikeSI ought to throw those optimizations in lcd-as-memframe.S and try it
13:34:27pixelmapreglow: amiconn said yesterday that the c200 version won't work or at least won't have a noticeable effect because the bottleneck is somewhere else
13:34:31jac0bmarkun: same battery company but that battery is only a 1000mah and not a 1200mah like the iPod battery
13:34:41 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
13:34:44markunjac0b: hence the small size difference
13:35:04IsolinearAny downside in upgrading my iPod battery to a high capacity one?
13:35:18markunIsolinear: longer charge times? :)
13:35:55IsolinearAnd conversely, longer runtimes especially with Rockbox, yes? :)
13:35:59jac0bmarkun: well if you look at the OEM battery from the side with the rubber feet on it the iPod battery is about the same thicjness
13:36:02 Join TheRock [0] (
13:36:10markunIsolinear: also, non official batteries have 'exploded' from time to time
13:36:38IsolinearAny brands or sources I should avoid? lol
13:36:54preglowamiconn: are you going to asm optimize the other targets that patch addresses too?
13:37:05 Quit ddalton ("leaving")
13:37:32 Quit spiorf (Read error: 110 (Connection timed out))
13:37:45jac0bmarkun: these are the batteries I am suggesting people get
13:37:48markunIsolinear: jac0b should know more about sources to avoid (because they lie about the mAh value)
13:37:59markunjac0b: yes, saw that
13:38:11markunquite cheap
13:38:30 Join spiorf [0] (
13:38:40jac0bisolinear: is yours a 4G ipod?
13:38:41 Quit safetydan ("Leaving")
13:39:07 Quit GregoryHouse (Connection timed out)
13:39:31IsolinearAlthough I just got my girlfriend a 4G, so it's all applicable... :)
13:39:45amiconnmarkun: Yeah, manufacturers of cheap replacement LiIon batteries often lie about their capacity
13:39:52IsolinearWhat are the stock capacities for 4G and 5G anyway?
13:40:14amiconnFor my digicam, I have 2 batteries, a branded 800mAh one and an unbranded 1000mAh one. Guess which runs longer?
13:40:40markunamiconn: so far I was lucky with my no brand batteries
13:41:22markunat least the ones for my canon and h120 they last longer than the original batteries
13:41:56IsolinearCan one use a multimeter to test the "actual" capacity of a LiIon battery?
13:42:22preglownot in any matter that would be more efficient
13:42:41IsolinearStill have to run the charge down?
13:42:54preglowafaik, yes
13:42:59IsolinearThought so.
13:45:04linuxstbpreglow: I'm pretty confident Zagor's code will just work on ipods - I've tested austriancoder's code occasionally, and it's always done the same on my ipod as it did on his Sansa.
13:45:46*Llorean is excited with the thought of USB on the horizon.
13:46:13*preglow too
13:46:14*linuxstb is excited with the thought of hiding ipods from itunes
13:46:17preglowamiconn: ping...
13:46:22IsolinearLlorean: USB for what?
13:46:44markunIsolinear: usb stack for rockbox
13:47:07LloreanIsolinear: It means "My nano may not be bog slow, and I can finally nuke the OF"
13:47:29markunlinuxstb: I hope anyone can show the gigabeat S guys what they need to change to get it working with their players
13:47:44IsolinearAhhh, we're still using Apple disk mode, aren't we?
13:48:01preglowIsolinear: deedey
13:48:15peturmarkun: CameronSino seems to be a good brand of replacements...
13:48:42jac0bisolinear: I would just make sure it is a cameron sino
13:48:43IsolinearGotcha... Which probably also means that my iPod won't need to reboot every time I go from USB to playing music and back? :)
13:49:14Lloreanmarkun: I'll second the CameronSino thing, at least my H120 one had a runtime along what it should with its capacity, and the battery hasn't exploded yet. :)
13:49:14preglowmy, that will be sweet
13:49:18Isolinearjac0b: That a brand of battery?
13:49:41IsolinearGuess so.. lol
13:49:43jac0bisolinear: yep
13:49:52jac0bsorry I am at work
13:50:03IsolinearYeah, I gathered it was from what Llorean said before you could respond.. ;)
13:50:38IsolinearSo how far off is USB for iPods then?
13:51:10Llorean"When it's done."
13:51:20IsolinearOf course... :P
13:51:30IsolinearAllow me to rephrase...
13:51:38IsolinearIs someone working on it? ;)
13:51:42LloreanAt least one.
13:51:48markunLlorean: but does Zagor intend to have it working soon?
13:52:00linuxstbmarkun: Why wouldn't he?
13:52:04Lloreanmarkun: Sounds like it.
13:52:16Isolinearlinuxstb: Just to mess with our heads.. :)
13:52:59markunlinuxstb: perhaps it was more of a viewport/fontcache priority for him :)
13:53:29 Quit iamben (Read error: 104 (Connection reset by peer))
13:53:31barrywardellmarkun: it sounds like henearly has it working already
13:55:20ZagorI'm working as much as time permits. writing usb-storage code as we speak.
13:55:46markunZagor: great!
13:55:48Zagordo you want it live? ;)
13:55:58Zagorinq_data.DeviceType = DIRECT_ACCESS_DEVICE;
13:56:55 Quit jac0b ("ChatZilla [Firefox]")
13:58:12 Quit Nico_P (Remote closed the connection)
14:01:33 Join Arathis [0] (
14:02:36linuxstbZagor: No webcam? ;)
14:03:46pondlifeUSB webcam, will be active when coding complete ;)
14:03:54linuxstbAh that vicious circle..
14:03:59pondlifeIncluding host support
14:04:24linuxstbAnd of course USB->Ethernet and a webserver plugin...
14:06:45jhMikeSpreglow: was there some upate to the mpegplayer seeking code that I missed?
14:07:44peturlinuxstb: no, support for wireless usb stick and an ftp server to sync your music to
14:08:48pixelmamarkun: the sansa OF's stupid "database refresh" might give some more motivation to work on it
14:10:26jhMikeSpreglow: I think you're right. It seeks movies that would just hang before.
14:10:51markunpixelma: yes, stupid behaviour of an OF can be very motivating :) A long list of motivations to port rockbox to the Gigabeat
14:14:08markunJdGordon: after your touchpad and mouse for sim commits a USB mouse could actually work :)
14:16:26 Join agm3nt [0] (
14:26:40 Quit barrywardell ()
14:27:15Aveinteresting commentary on the fs7510 again.. so maybe we do get a "rollback" of akind
14:31:14markunpreglow: another ambisonics guy
14:33:51preglowmarkun: the gigabeats would be able to decode ambisonics for sure
14:34:01preglowmaybe even coldfire if the codecs are easy
14:36:08preglowdepends how much filtering one would want to do, though
14:36:41 Join barrywardell [0] (
14:36:46preglowonly thing i know for sure is i can both encode and decode a huge number of sources to a huge number of loudspeakers on this pc with a very nice cpu load
14:37:03 Join Morey [0] (
14:38:12 Join bluey- [0] (
14:39:21Moreypreglow: I'm one of the authors of the mpegplayer seek functionality. We have a plan to get seeking from within the player as well. Also the startup menu needs work, I just started a new patch that almost eliminates the delay in startup.
14:39:48preglowMorey: only thing i feel strongly about is that i don't want a menu to appear when i would expect the thing to just start playing, everything else looks pretty good to me
14:41:00Moreypreglow: After the seek is moved to the play menu we should be able to just ask for resume (if a bookmark exists) on startup.
14:41:34MoreyThis is not an easy task and is taking me a little time to do, but it's in the works.
14:42:57MoreyAs for now I think you will find my new patch less annoying. I worked out some of the delays and did some coding cleanup.
14:43:20 Quit barrywardell ()
14:43:28markunMorey: good to see you on IRC
14:43:37markuneasier to communicate
14:44:40Moreymarkun: Can't stay on for long, but wanted to let ppl know I'm aware of the problems, complaints, and compliments.
14:46:13MoreyTask #7971 is the new patch.
14:48:10markunok, great
14:48:39markunpersonally I don't we need such a special preview screen for seeking at all
14:48:55markunbut I haven't looked at the mpegplayer since your patch got committed
14:49:03MoreyjhMikeS: The update to the seeking code fixed an issue where an mpeg got chopped off and the stream didn't start at time 0. Now it takes into account the stream start time as 0.
14:49:28preglowmarkun: do gigabeats do digital out?
14:49:28MoreyPlus offsets the ending of course.
14:49:45linuxstbpreglow: No.
14:50:07linuxstbpreglow: Although I have a feeling there is I2S in the dock connector..
14:50:32jhMikeSMorey: what about where looking ahead into the guard buffer when parsing can access garbage?
14:51:07 Quit ramon8 (Read error: 104 (Connection reset by peer))
14:51:09markunlinuxstb: there is
14:51:46markunpreglow: is it possible to make a simple i2s -> s/pdif convertor?
14:52:41MoreyjhMikeS: There should be checks to handle that.
14:52:43jhMikeShmmm...that sounds like a nice little kit project
14:52:53 Nick parafin|away is now known as parafin (
14:53:04preglowmarkun: if you can get the proper chips it shouldn't be too hard
14:53:17preglowmight even get away with just one chip
14:54:30 Join roolku [0] (
14:56:37jhMikeSMorey: but what about further increment + lookaheads? I only see one.
14:56:51 Join kugel [0] (i=kugel@unaffiliated/kugel)
14:58:39preglowmarkun: looks like it'll do the trick
15:00:39preglowmarkun: surface mount, though, prepare for soldering fun
15:00:49jhMikeSso we can mod the gigabeat cradle to output s/pdif?
15:01:00MoreyjhMikeS: Not sure, i'll have to ask Gwynne. He worked on that code.
15:01:45markunjhMikeS: either mod the cradle or make our own
15:02:25jhMikeSMorey: I only see one check. Everything else looks the same as before. Every increment could put a lookahead past disk_buf_end. Just my .02
15:03:13jhMikeSmarkun: well, just make a new pc board and nicely package it in there :)
15:04:22 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
15:09:22jhMikeSMorey: One change I'd like to make in there is to use the normal message queues instead of the hack that was there before the full COP support. Are you relying on the current code alot for anything? Will changing it be terrbly disruptive to anything?
15:09:29 Part hcs
15:11:05MoreyjhMikeS: it would be nice to get #7971 in svn first. then you would not disrupt anything.
15:11:33***Saving seen data "./dancer.seen"
15:13:45jhMikeSok, I'm waiting a bit anyway before integrating tons of stuff
15:15:40MoreyjhMikeS: The start menu takes about 7 seconds to load in which it's determining the length of the mpeg. That new patch got that time delay down to about 3 seconds on my e200.
15:16:33 Quit CaptainSquid ("Miranda IM!")
15:19:01 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
15:19:24jhMikeSMorey: there's code in mplayer that does some figuring based on first/last timestamps and a few other heuristics. not much looking around is done.
15:21:48MoreyjhMikeS: One issue that I fixed was in alloc.c where mallocs were zero'ing out the data. This is not needed and was really slowing things down.
15:21:49markundoes the statement between brackets meen that wav playback in not gapless?
15:22:09jhMikeSMorey: the calloc function?
15:22:59MoreyjhMikeS: No, calloc should zero out the data. Malloc should not, and it was. This may have been a mistake on our part, but I removed it in the new patch.
15:23:52*jhMikeS saw a small writeup clamining rockbox is gapless only because it inserts a small fade between tracks and wanted to correct this error but comment registration was closed. :\
15:24:59 Join tictoc [0] (i=tabac@gateway/gpg-tor/key-0xB9002659)
15:25:17markunjhMikeS: where did you read it?
15:26:03jhMikeSon about rockbox and the m5
15:26:15markunI might have an account there
15:27:21 Quit Morey ("Ninja IRC v1.5.8.1(#1) exiting after 50m37s of use")
15:27:22LloreanIt's a different login than the forums, or something
15:27:30LloreanI'm logged into their forums, but it tells me I must be logged in.
15:27:46*preglow thinks it's about time the ipod scrollwheel accel patch went in
15:27:51preglowanyone have any objections?
15:27:51jhMikeSyeah, it works about as well as cowon support
15:28:20linuxstbpreglow: What settings does it add?
15:28:23jhMikeSnot really. sansa settings range needs tweaks but that's about it.
15:28:35preglowoh, there's that, we might not want the settings
15:29:02linuxstbpreglow: I think it's OK to add them initially, to give people chance to play with them, but later we may want to remove them.
15:29:13markunjhMikeS: I only have an account for the forums :(
15:29:21preglowlinuxstb: remove them after people have had the chance to tweak them? there will be complaints...
15:29:26Lloreanpreglow: I very much would prefer if the wheel settings where hard-coded. :)
15:29:38linuxstbi.e. so we can work out what they should be for all the different targets (assuming they will need to be different).
15:29:42pregloweither we remove them now, or with great difficulty later
15:29:58LloreanOr... hide them in debug?
15:30:02linuxstbpreglow: We're dictators - just put a comment in the commit message saying the settings are temporary... Or just remove them now...
15:30:25preglowideally we would have a dev tweak each target, then just use that as a hardcoded value in the commit
15:30:27linuxstbDebug menu might be a nice idea.
15:30:44preglowi don't really think so, i think we should at least _try_ to go hard-coded
15:30:45LloreanPut them in debug, and have the commit comment say "Settings can be adjusted in the debug menu, with the intent to choose a final value"
15:31:20LloreanI like the Nano values of the last version of the patch I tested, which is I believe the second to last entry in the tracker.
15:31:24linuxstbpreglow, Llorean: For example, can the two of you agree on the values for the Nano?
15:31:24JdGordonmarkun: usb mouse would be.. umm... interesting :p
15:31:42Lloreanlinuxstb: The default ones in the patch are good for the Nano, I think.
15:31:42preglowi liked the last patch i tried, which i believe is v12, haven't tried the latest
15:32:03LloreanYeah, v12 for Nano was fine by me
15:32:06preglowi will however do so in ten minutes
15:32:12 Join barrywardell [0] (
15:32:17*linuxstb will try and test the patch this evening on his Color and Video
15:33:24markunjhMikeS, Llorean: you could try to PM the author:
15:33:31Lloreanmarkun: Already did that. :)
15:33:54LloreanExplained that Rockbox had true gapless since before M5 or even X5 support, and that crossfade is actually an entirely separate feature.
15:34:05LloreanI was even polite.
15:34:30markunLlorean: wow, congratulations on that last bit :)
15:34:34Zagorhrmf. /me spots typedefs in the hotswap code.
15:35:17*JdGordon doesnt remember adding any
15:35:43linuxstb"svn blame" never forgets...
15:35:52Zagorand tSDCardInfo
15:36:17JdGordonsvn blame shows who commited it.. not who coded it :)
15:36:48markunwith git it would show who codec it..
15:37:07bluebrotherhow that?
15:37:09jhMikeSmarkun: thanks. if Llorean doesn't get to it, I'll try later.
15:37:16markunjhMikeS: he already did
15:37:48markunbluebrother: because it makes a difference between the person who wrote it and the person who commit it
15:38:02linuxstbmarkun: How?
15:38:06markunfrom what I understood during a discussion at the GSoC summit
15:38:17JdGordonnot if the person adding it to git was doing it for someone else....
15:38:44linuxstbsvn blame will give you a revision number, and the commit message should indicate if the committer wasn't the author.
15:38:45markunwell, assuming that it was pulled from the git repository of the person who wrote it
15:38:55markuninstead of from the patch tracker
15:40:04bluebrotherbut the guy who wrote it still needs to check it in −− even if it was a different repository ;-)
15:40:19Zagoris cid and csd the only textual information available about sd cards?
15:40:37bluebrotheriirc, yes.
15:40:40 Nick Tanuva|Zzz is now known as Tanuva (n=tanuva@
15:41:00 Join pondlif1 [0] (
15:41:00 Quit pondlife (Read error: 104 (Connection reset by peer))
15:41:08ZagorI finished the ata/scsi layer before I realized my target doesn't have ata :-)
15:41:19JdGordonwell done :)
15:41:30jhMikeSmarkun: ah, woops. didn't read enough.
15:41:44linuxstbZagor: I thank you on behalf of ipod and H10 users ;0
15:42:03Zagorlinuxstb: yeah, to bad you guys will have to debug it ;-)
15:42:11markunbluebrother: no, it can be pulled, or what are you trying to say?
15:42:48bluebrothermarkun: if the origin is a patch (e.g. from the tracker) you can't figure out who wrote it as that person didn't check it in to any repository
15:43:07linuxstbZagor: I thought the Sansas implemented the same ata_* API though/
15:43:27Zagorlinuxstb: yeah, almost. only the basic identification and capacity reporting differs.
15:43:55markunbluebrother: isn't that was I said? "well, assuming ... instead of from the patch tracker"
15:43:59Zagorso hopefully your debugging duty will be light
15:44:17bluebrothermarkun: seems I missed the last part ;-)
15:44:35linuxstbZagor: Sounds like you're close to mounting...
15:44:48Zagoryeah, I'm an optimist :)
15:44:55*markun should play around with git a bit
15:45:31barrywardellZagor: what information do you need?
15:45:34*bluebrother is curious when he'll be able having the Ipod doing some usb stuff
15:45:54markunbluebrother: when it's done ;P
15:46:14bluebrotherdamn. I feared that answer. Bastards! :P
15:46:14markunshould have put quotes around it since it's llorean's line
15:46:15Zagorbarrywardell: nothing really important. there is a vendor and prouct text field in the scsi protocol that would be nice to enter something in.
15:46:34|Rain|Zagor: you should probably just use whatever the OF for that platform uses
15:46:37Zagorcsd and cid appear to be the candidates to use, but what are they? and in what bit order
15:46:54Zagor|Rain|: no this is information that is on the card/disk
15:47:17Zagors/bit order/byte order/
15:47:35|Rain|Zagor: you're talking about the vendor/model/revision info? or something else?
15:47:50 Join jgarvey [0] (
15:47:52barrywardellZagor: they're fully detailed in the SD specs
15:47:57Zagor|Rain|: yes, in the scsi INQUIRY struct
15:49:41|Rain|Inquiry command
15:49:44|Rain|Vendor: SanDisk
15:49:44|Rain|Product: Sansa e280
15:49:44|Rain|Revision level:
15:50:40Zagor|Rain|: I'm still interested in reading it from the card. this is supposed to be a multi-platform driver.
15:50:58preglowi think there are some makefile dep errors around with regard to lang file updates
15:51:07preglowi needed to make clean before the patch built
15:52:37preglowZagor: any guesstimates on when you'll be copying files to your sansa via your own usb code? :>
15:52:44 Join |Rain|_ [0] (
15:53:20Zagorpreglow: possibly tonight
15:53:34*preglow crosses fingers
15:53:42|Rain|_Zagor: that's why I suggested just reporting what the OF does, although I doubt anything matches on that instead of the USB ID. I know it's supposed to be multi-platform, but you probably WILL want to match the OF USB ID, so you've probably already got some hardcoded platform-specific stuff
15:54:34|Rain|_Zagor: but I'm not against the idea of providing meaningful information, either (although −− what do you do for hotswap SD cards? the INQUIRY will probably only happen at device connection, unless I'm mistaken...)
15:55:04Zagorwe support hotswap already. and so does usb-storage.
15:55:05barrywardellZagor: the CID has product name string, manufacturer id, oem id, ... I think that's the best to use
15:55:28|Rain|_Zagor: yes, but the INQUIRY happens whether there's a device inserted or not, and I don't believe it happens again when media is inserted
15:55:29Zagorbarrywardell: do you have a link to a good spec? I only find partial ones.
15:55:34markunnice, rockbox shirt in China:
15:55:47 Quit lee-qid ("aufwiederbyebientotsayonara")
15:55:52barrywardellpage 86
15:55:59|Rain|_s/a device/media/
15:56:10Zagor|Rain|: I'm pretty sure we read it when new media is connected, but if not that is easy to fix
15:56:21preglowLlorean: i'd say the latest patch is pretty good on nano
15:56:32|Rain|_Zagor: I'm talking about the host OS
15:56:44Lloreanpreglow: Pretty much the same as v12?
15:56:53preglowdidn't try v12, only v11, it seems
15:56:55 Quit |Rain| (Read error: 104 (Connection reset by peer))
15:57:00Zagor|Rain|_: the host OS has to support multiple inquiry. otherwise hotswap harddisks wouldn't work.
15:57:19Zagoranyway, let's deal with that when/if it happens
15:57:25 Nick |Rain|_ is now known as |Rain| (
15:57:48|Rain|Zagor: for real hotswap SCSI drives, I always have to force a rescan −− but okay
15:57:59Zagorbarrywardell: thanks
15:58:12Zagor|Rain|: maybe you're right. we'll see.
15:58:38|Rain|tbh I'm just anxious to see the code. the vendor could be "turkey" and the model could be "giblets," and I'd still be happy :P
15:58:46preglowouch, i think i've found a bug in dir browser limits here
15:59:52preglowif i set the max entry limit low, try to open a dir with too many, increase the max entry setting, then look at the dir again, all of the new files can't be used, plus the "open with" menu is clobbered
16:01:28preglowwtf, it seems like these settings need you to restart, but don't really tell you
16:03:02Arathisamiconn: what does the common H10 user (like me) have from the last commit?
16:04:10Lloreanpreglow: I thought increasing max files used to tell you.
16:04:12 Quit |Rain| ("unborking")
16:04:29preglowdoesn't anymore
16:04:47preglowthe tree code really shouldn't use the new value, but i think someone has blundered here
16:04:52 Join scorche|w [0] (n=8dc5049d@rockbox/administrator/scorche)
16:05:15 Join |Rain| [0] (i=rain@2001:440:eeee:fffb:42:0:0:2)
16:06:26preglowthe people who redesign menus aren't always so careful to retain old functionality either
16:07:39barrywardellArathis: which commit?
16:08:31preglowi truly wish people would start using the code style in the file they're editing
16:08:38preglowso preciously few even try to do this, it seems
16:09:33Arathisbarrywardell: "Improved H10 ADC driver."
16:09:51Zagorgotta go. see you later.
16:09:52 Quit Zagor ("Client exiting")
16:10:10barrywardellArathis: it will hopefully make the readings from the ADC less jumpy (battery, remote, scrollpad)
16:10:24 Join midkay [0] (n=midkay@rockbox/developer/midkay)
16:11:06Arathisoh, cool :D
16:12:24pondlif1preglow: Code style..? I only changed 2 characters...
16:12:32preglowpondlif1: not talkinga bout you
16:14:32 Quit blithe (Read error: 104 (Connection reset by peer))
16:14:52 Part pondlif1 ("disconnected has pondlife")
16:15:00 Join pondlif1 [0] (
16:15:02 Join blithe [0] (
16:15:28pondlif1Hmm, my IRC client's confused...sorry
16:15:30 Part pondlif1 ("disconnected has pondlife")
16:16:05 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
16:17:38lostlogicproblematic indeed.
16:17:40preglowseems quite a few places use the new value with no realloc of buffer
16:17:42lostlogicgah, sorry
16:19:16 Join Hammer89 [0] (
16:22:42Hammer89does the e200 bootloader need to be updated on my player? (haven't installed a new bootloader in months)
16:26:08markunHammer89: probably not. Are you having problems?
16:26:55Hammer89no.... but I haven't updated rockbox in months
16:27:11Hammer89so I was just checking before I did, and got a nasty surprise ;)
16:27:29 Quit animeloe ("This computer has gone to sleep")
16:27:48markunI don't know anything about the sansa players, maybe someone else can answer
16:30:06scorche|wif it is working for you, then there isnt really a need to change it, but it shouldn't hurt you if you do
16:30:16Hammer89okay... thanks :)
16:30:51markunscorche|w: the gigabeat really needed a bootloader update a few times
16:32:55scorche|wmarkun: aye...i /think/ this one is just to fix issues with OF loading
16:33:35 Quit JdGordon ("Konversation terminated!")
16:35:09roolkuwhat is this warning in the x5 sim "oggmalloc.c:8: warning: dereferencing type-punned pointer will break strict-aliasing rules"?
16:35:17preglowjhMikeS: new flyspray entry concerning you
16:35:18 Quit zicho (Read error: 104 (Connection reset by peer))
16:35:23roolkuI didn't even touch that file?
16:35:48roolkum5 sim
16:37:06 Join zicho [0] (
16:38:47 Join toffe82 [0] (
16:40:06pixelmait's already there in previous commits - sims that were built on 64bit machines if I see correctly. First occurance in jhMikeS' commit but since there were other warnings as well and the usual 2 in sim builds, it's hard to spot
16:40:54pixelmafirst I see is the H120 sim in the big commit
16:47:43Hammer89one other thing... does rockbox play m4a's?
16:48:13amiconnjhMikeS: Did you write the arm idct for mpegplayer, or is that something libmpeg2 already had, and we just didn't use it before?
16:48:33amiconnI wonder why it's asm-in-c instead of a proper .S file...
16:48:59preglowamiconn: do you plan to write asm versions of nano and video yuv blit too?
16:49:18preglowok, i will ignore that flyspray entry, then
16:49:25amiconnIn fact nano could use the same as H10 will, if we split out the driver
16:49:50amiconn(more efficient - doing the double-line zig-zag as already used for gigabeat, e200, and c200)
16:50:16amiconn...and also for X5 and H300, although that's of course different asm ;)
16:50:58amiconnColor could for the type 1 lcd, but not for type 0, unless someone finds out what controller that actually is and we find a datasheet
16:51:15preglowamiconn: btw, do you know anything about the limits menu? did those entries use to tell you that you have to reboot?
16:51:32preglowthen they are bugged
16:51:42preglowcurrently i get corrupted entries when making it bigger
16:51:43amiconnOr, rather they didn't tell
16:52:02amiconn...but the new value used to only apply after reboot
16:52:10amiconnThen someone messed them up
16:52:24preglowboth zagor an linus have put in uses of the .max_files_in_dir member
16:52:33preglowafaik that value should only be used on boot, yes?
16:52:54preglowtree.c does back it up to a separate variable, but other files don't see that
16:53:04amiconnUmm, I don't know the exact details of the implementation
16:53:26amiconnBut you're probably right, and the settings member must only be used on boot
16:53:38preglowanyway, right now the behaviour is messed up, if i increase the size, the "open with" menu gets seriously garbled, and new file entries don't work at all
16:54:05*amiconn is trying to get his head around idct :/
16:55:25preglowdo the yuv blit asm first :>
16:55:53amiconnWhich target(s) have worse video performance, coldfire or PP?
16:56:00preglowcoldfire, afaik
16:56:43preglowi never tried playing a video on coldfire
16:56:51preglowand only do so on pp for testing
16:58:53preglowi'm quite confused by this stuff, some of the .max_files_in_dir use is pretty old
16:59:31pregloweither the code has been rewritten in funny ways, or this bug has been in rockbox for a good while
17:00:52pondlifeProbably a bit of both
17:01:37preglowmostly zagor have touched the files, it seems
17:01:40preglowthink i'll ask him later
17:04:34*pondlife wonders where the MajorChanges link on the front page went...
17:05:43markunsomeone talked about RTC problems on the Gigabeat a while back. I also got some now (but don't know how to reproduce)
17:06:01 Quit TheRock ()
17:06:02 Join mf0102 [0] (n=michi@
17:06:11krazykitmarkun, what, that the clock unsets itself sometimes?
17:07:32LloreanHm, Slashdot says DRM-free iTunes music now costs the same as the encumbered stuff.
17:08:47 Part agm3nt
17:08:47roolkumarkun/ krazykit: yes, happens to me occasionally usually after midnight?
17:10:09krazykitroolku, i've turned off my player only to turn it on 20 minutes later to have the time unset. i can't figure out the common factor in all of this.
17:11:36roolkukrazykit: I am not saying it is cause/effect. I just notice it most often when I want to listen to an audiobook before I go to sleep
17:11:38***Saving seen data "./dancer.seen"
17:14:44 Quit aliask ("ChatZilla [Firefox]")
17:18:04 Join ToHellWithGA [0] (
17:18:35markunLlorean: what's the use of the encumbered stuff then?
17:18:38 Join random_desu_is_s [0] (
17:19:03pondlifeIs the Iriver H300 capable of a wake-up alarm?
17:20:01Lloreanmarkun: As of this new pricing change? No clue at all
17:20:09Lloreaniiuc the unencumbered stuff is even at a higher bitrate.
17:20:41markunmaybe some people feel better using DRM-ed stuff :)
17:20:55Llorean"Hah, now nobody can steal MY music!"
17:20:57pondlife"Files you can trust"
17:21:01pixelmapondlife: I believe so but needs it enabled in the bootloader too IIRC, I think XavierGr played around with it - probably he knows more
17:21:07 Part bluey- ("Leaving")
17:21:23pondlifeAccording to the wiki, the SVN bootloader supports it.
17:22:03ArathisGreat. I was just so suprised about the new very smooth midi playback of my big H10 and than it crashes at playback. It's one specific file and a specific location in playback. :/
17:24:09 Join sin613 [0] (
17:24:33pixelmapondlife: could remember about a patch but that's all I know...
17:25:19pondlifepixelma: Thanks
17:25:41Arathisbarrywardell: ever experienced such a midi crash?
17:36:54 Join AmbiquitY [0] (i=51f1b1f0@gateway/web/cgi-irc/
17:38:00AmbiquitYI've got a question.. where could i find an older build for ipod nano?
17:38:00pixelma..lo :)
17:38:17GodEater_you couldn't - you'd have to make one yourself.
17:38:25GodEater_we don't archive them for very long at all
17:38:48 Quit Thundercloud (Remote closed the connection)
17:38:54bluebrotherthere are archived builds −− about the last 4 weeks.
17:39:03AmbiquitYwhere can i find them?
17:39:10bluebrotheron the website.
17:39:39AmbiquitYyes but where? i can only find the current builds
17:39:51GodEater_click the "old" link under the Nano
17:40:00*bluebrother wonders if it's that hard to find the "Daily builds" link on the "current build" page.
17:40:16GodEater_depends how lazy you are
17:40:32bluebrotheryeah, reading can be quite difficult ;-)
17:40:53AmbiquitYhuh i dont see a daily build link under current build..=s
17:41:33AmbiquitYaah now i see.. sry about that =)
17:41:57AmbiquitYdidnt know daily builds where older builds then the current..
17:42:46AmbiquitYso uhm thx guys
17:43:09 Quit AmbiquitY ("CGI:IRC (EOF)")
17:44:38 Part Hammer89
17:45:03*GodEater_ wonders when he'll be back to say "it didn't help"
17:45:33 Join n1s [0] (
17:48:42barrywardellArathis: I've never played a midi file on my H10
17:49:35pixelmabluebrother: in jewels.tex the complete button table is within an opt for each keypad, thus multiplying the tables. I prefer one table with the "opts" in each row and already rewrote it this way, however I had a small problem to work around (but commented it) - care to take a look?
17:50:02pixelmaor maybe you have a different opinion...
17:51:05 Quit Rob222241 (Read error: 104 (Connection reset by peer))
17:53:11Arathisbarrywardell: okay ^^
17:53:18ArathisI'll try other files then
17:53:25Arathisother midi files
17:53:26bluebrotherpixelma: sure.
17:54:52pixelmawould be the new jewels.tex
17:56:10pixelmaArathis: if it's only one file it could be the midi itself, e.g. I also have one midi that completely freezes my M5
17:56:31n1sArathis: do you have a link to that file or could you send it to me?
17:57:05n1sand also does it crash with a message or just lock up?
17:57:37Arathispixelma: it works on my pc though and I haven't tested less than three files till I updated rockbox today (after about two month)
17:58:26n1spixelma: I'd be interested in that too, might be able to work something out
17:58:39pixelmaArathis: it's the same here - the file plays on my computer and my cellphone too
17:58:57Arathisn1s: I could send it to you, but would like to further test some other files first. and it crashes while my screen is already black so I'd need to switch backlight to "on" and test again
17:59:01n1spixelma: plays on your computer as in the rockbox sim?
17:59:22Arathisfor me "plays on pc" means plays with timidity
17:59:46pixelmahaven't tried that yet - plays in errmm... mediaplayer
18:00:06pixelmawindows mediaplayer
18:00:18ArathisI'm a bit busy right now so it could take some time to test and respond
18:02:24 Quit jhMikeS (Nick collision from services.)
18:02:30 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
18:03:36pixelman1s: just discovered that it plays on my M5 but it struggles at the begining (but not throwing buffer misses), wasn't patient enough... but it doesn't play right. Will send it to you, I think that's easier than to try explaining it here...
18:03:38 Join inomiad [0] (
18:05:43inomiadanyone have idea if there's anyone working on 2nd gen nano port? i've been checking the site every now and then for a year, but no progress i guess?
18:06:19scorche|wno one is working on it currently
18:07:38inomiadsniff :I
18:08:57*markun wonders how many new DAPs Bageder and LinusN will bring back home from China to port rockbox to..
18:09:50 Join Thundercloud [0] (
18:09:54inomiadhopefully 2nd gen nano ;)
18:10:26pondlifeOnly if you can break into Apple HQ and snaffle the encryption details..
18:10:30 Join Crash91 [0] (n=evil91@
18:10:45scorche|wand a few datasheets too if you have the time
18:10:53scorche|w(assuming they even have them)
18:11:36inomiadit's really that different hardware, huh?
18:12:01markuninomiad: yes, very different
18:12:06pondlifeXavierGr: around?
18:12:11linuxstbmarkun: They need something to do on the flight home..
18:12:26pondlifeone each, maybe
18:12:43markunlinuxstb: let's hope they didn't forget to compile a blackfin toolchain :)
18:13:07linuxstbTelechips is popular in China...
18:13:37markunyes, also
18:13:42markunand rockchip?
18:16:36 Join Lear [0] (i=chatzill@rockbox/developer/lear)
18:17:11inomiadokay, thanks for the info anyway, guess i'll have to live with default ipod firmware ;)
18:17:31 Quit inomiad ()
18:21:13 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
18:22:44Arathisn1s: okay, I'm getting Buffer miss! for that file. But I had a Data abourt with another file followed by Buffer miss! although the file should be at it's end where that happens.
18:23:37 Quit linuxstb ("Client Exiting")
18:23:39n1sArathis: Buffer miss! is a performance issue only and known about, however a "Data abort" is bad
18:24:24n1sdoes the Data Abort happen at the same position in the same file every time?
18:25:10ArathisI'll test
18:26:16pondlifeArathis: Might be worth trying in the sim too, if you can repro the crash.
18:26:37 Quit obo ("bye")
18:28:13 Join Rob2222 [0] (
18:29:09Arathispondlife: later perhaps. I'd need to run vbox and build it first and as said am a bit busy
18:29:32pondlifeOr pass the file onto one of us to try?
18:29:47pondlifeFlyspray would be a good place
18:29:48Arathis.. mised the buffer miss :/ but it was at the same location in the same file and I haven't found it in another file yet
18:30:27Arathiss/'buffer miss'/'data abort'
18:31:56 Quit Crash91 ("Bye Bye!")
18:32:16 Part pondlife ("disconnected has pondlife")
18:32:18n1sArathis: could you give me a link to that file or send it to me?
18:32:36Arathisin a second
18:33:41 Quit Lear ("ChatZilla [Firefox]")
18:34:38*ender` yawns
18:35:00Arathisn1s: port seems to be closed ..
18:35:15sin613which models have fm tuners?
18:38:07amiconnpreglow: around?
18:39:55sin613n/m, found the list
18:40:56Arathisn1s: I'm getting the data abort after the last notes/final chord and it says "at 400C16C (0)"
18:41:06 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
18:41:16n1sArathis: thanks, I'll look into it
18:42:01rasherI've noticed this from many people, but isn't it generally better to just ask whatever question you have, rather than saying "ping" "around?" etc. That way, the person has a chance to answer when he gets back (and you might not be there anymore) (this is not directed specificly at you, amiconn)
18:44:24|Rain|some things demand interaction, though... something sufficiently complex is frustrating to both parties if the message is lost in scrollback or not completely understood by the recipient (when the sender is now away instead)
18:44:27 Join midgey [0] (
18:44:28 Join bertrik [0] (
18:45:54 Quit midgey (Client Quit)
18:46:13 Join midgey [0] (
18:47:26rasher|Rain|: then something along the lines of "I want to talk to you about <foo>"
18:47:48 Join Domonoky [0] (
18:48:05scorche|wrasher: but then you just scare them off ;)
18:48:23rasherSeeing a "Are you there?" in your scrollback is rather frustrating when that person is no longer there
18:49:03|Rain|it can be even more frustrating if you know what they want and WANT to talk to them about it :P
18:50:17 Quit GodEater (Remote closed the connection)
18:52:57 Join lazka [0] (
18:53:04 Join Lear [0] (i=chatzill@rockbox/developer/lear)
18:54:14 Quit gromit` (Read error: 104 (Connection reset by peer))
18:55:54 Join Blindbricks [0] (i=519c1f9d@gateway/web/cgi-irc/
18:56:00BlindbricksHi guys
18:56:05BlindbricksA really quick question
18:56:25BlindbricksHow can I open the original Ipod firmware with rockbox installed?
18:56:29Blindbricksis it possible?
18:56:30 Join gromit` [0] (
18:56:45Domonokyhold menü while booting, i think..
18:56:49krazykityes, and the manual explains how
18:56:55BlindbricksIt does?
18:56:58BlindbricksAh good thanks
18:57:10 Join GodEater [0] (n=bryan@rockbox/staff/GodEater)
18:57:26BlindbricksI just installed today rather good
18:57:33BlindbricksAlthough I need to look into mpeg games
18:57:36 Quit Blindbricks (Client Quit)
19:11:39***Saving seen data "./dancer.seen"
19:11:55 Nick EnterUse1Name is now known as EnterUserName (
19:13:26 Quit random_desu_is_s ("ChatZilla [Firefox]")
19:18:29preglowamiconn: am
19:18:51amiconnNm, found the place where I went wrong meanwhile
19:19:10amiconnI thought the coldfire idct asm would be different from the C code that should resemble it
19:19:34amiconn(other than the different intermediate storage method), but it isn't
19:20:11preglowi thought it was
19:20:38amiconnIn mirak's patch there is new C code that resembles his asm
19:21:00amiconnBtw, I already spotted 4 instructions which can be collapsed into 2
19:21:44 Quit sin613 (Read error: 113 (No route to host))
19:21:59amiconn(in one of the matrix muls, two parts of each stroke are identical, so with trivial reordering, I can use a move.l %accX, %accY)
19:22:16amiconn...replacing two mac.w instructions each
19:24:37 Join desowin [0] (
19:29:35 Join SirFunk [0] (
19:36:59 Join linuxstb [0] (
19:37:55 Join sin614 [0] (
19:45:01 Quit spiorf (Read error: 104 (Connection reset by peer))
19:45:12amiconnSome further reordering allows filling almost all emac latency, and saving a couple of instructions
19:45:53 Join spiorf [0] (
19:48:18 Join styleism [0] (
19:49:23amiconnThe row loop is just 68 instructions now :)
19:50:11 Part styleism
19:50:28 Join kkurbjun [0] (n=kkurbjun@alamode.Mines.EDU)
19:55:17 Join mirak [0] (
19:57:20 Join Lars_G [0] (n=Lars@unaffiliated/lars-g/x-000001)
19:57:23XavierGrpondlife: I am here now
19:57:24Lars_GHi all.
19:57:36XavierGrah crap he is not here
19:57:44Lars_GThese here who listen to podcasts. What do you use to fecth and sync them?
19:59:20sin614amiconn: idct optimizations?
19:59:23LearJuice to fetch, manual sync (only listen to a few...)
19:59:54Lars_Gok, thanks lear
19:59:54sin614i stick to google reader, only download things with interesting descriptions
20:02:58 Join ilgufo [0] (
20:05:05 Join atsea-34 [0] (i=atsea-@gateway/tor/x-66aec4112b41a718)
20:06:47 Quit ilgufo (Client Quit)
20:07:58 Quit petur ("connection reset by beer")
20:09:10 Part Lars_G
20:12:25 Join ilgufo [0] (
20:12:26 Quit japc (Read error: 110 (Connection timed out))
20:16:48 Quit mirak (Remote closed the connection)
20:18:54 Join hannesd [0] (
20:23:43 Join ompaul [0] (n=ompaul@freenode/staff/gnewsense.ompaul)
20:23:55bertrikI'm trying to fix a bug, but I get a confusing error message during compile
20:24:13bertrikI added lang.h to apps/plugins/lib/playback_control.c
20:24:52bertrikThis is what I get
20:25:57bertrik* I mean I added #include "lang.h" to apps/plugins/lib/playback_control.c
20:26:39nicktasticIt is a problem with the build environment, not the code
20:26:53 Join WalterEgo [0] (
20:27:16nicktasticBeyond that, *shrug*
20:28:50bertrikok maybe I'll try it under Linux, I have little hope though
20:29:48rashernicktastic: why do you think so?
20:31:14nicktasticrasher, Because it is a make error, not a compiler error
20:32:18nicktasticAdding a header or otherwise changing some code shouldn't throw make errors, excepting some weird auto-header-dependency-resolution madness
20:32:24bertrikI'm concerned about the *** No rule to make target line, not the modification time in the future
20:33:12LearWell, lang.h isn't your typical include file. It is generated during build.
20:33:36nicktasticAgain I should just keep my mouth shut ;)
20:35:44linuxstbbertrik: You can't include any .h files in plugins apart from plugin.h. If you want access to what's in lang.h, you will need to export it via the plugin API.
20:35:46bertrikI realise I haven't tried a clean build yet, I'll try that first
20:36:07bertriklinuxstb: thanks
20:36:31 Quit XavierGr (Nick collision from services.)
20:36:34 Join XavierGr [0] (
20:36:44bertriki wondered already how the main firmware and plugins interacted
20:37:32 Join Zagor [0] (
20:44:32bertrikIs there any support for internationalisation at all for plugins?
20:46:26 Part sin614
20:46:27markunbertrik: don't think there is right now
20:48:01Zagorbertrik: I'm still confused about this. Is every endpoint bidirectional? i.e. both IN and OUT endpoints can receive requests from the host?
20:48:43bertrikAFAIK, endpoints are always unidirectional, except for EP0
20:49:06bertrikI find the USB spec a bit vague on whether endpoint numbers or endpoint addresses need to be unique
20:49:20Zagoryeah that's what I have assumed all along, only both usbmon (linux kernel usb snoop module) and the imx31 docs conflict with this
20:49:21bertrik(endpoint address = endpoint number | direction bit)
20:49:53Zagor"if endpoint 3 (transmit direction) is configured as a bulk pipe, then we can expect the host will send IN requests to that endpoint."
20:50:21markunZagor: do you know where the pp specific code in the USB driver is and what needs to be written for the i.mx31 to get it working with the Gigabeat S?
20:52:43bertrikZagor: USB is basically a polled system, every transfer (from host to device or vice versa) is always initiated by the host
20:52:46 Quit desowin ("use linux")
20:53:00Zagormarkun: I think only the interrupt vector is pp specific
20:53:33Zagorbertrik: yeah I know. but I expected all request to come in at the same OUT endpoint. but they don't...
20:53:58markunZagor: and maybe some defines for the addresses of the USB hardware?
20:54:04Zagormarkun: ah, of course
20:54:18*bertrik is confused
20:54:32Zagorbertrik: guess what I am...
20:54:47markunbut if it's only that, then the Gigabeat S should have USB shortly after you got it working
20:55:22bertrikZagor: what kind of request do you mean?
20:56:09bertrikfirst packet of CBW?
20:56:27Avebtw, this diff here ( , the stuff that makes people happy, see the difference between the speeds
20:56:43Aveis CPUFREQ_MAX used anywhere but as a label?
20:56:55Avebc the two changes arent same valued
20:57:13Avepll setup says 75 MHz whereas MAX 78
20:57:39amiconnThat patch contains at least 2 bugs
20:57:51amiconnThe one that you mentioned, and the actual pll setup
20:58:29Aveyeah that one is hairy so I cant even comment on that
20:58:44amiconnThe pp5022 pll must be run at >= 96MHz, and the post divider needs to be used for lower frequencies
20:58:52bertrikZagor: I'm not familiar with the freescale controller so I don't really know when you get an interrupt
20:58:56amiconnIt's actually not really dfficult
20:58:56Zagorbertrik: hmm, perhaps my problem is the reversed. I get the request on the OUT ep and writes my response to the IN ep, and that fails.
20:59:32bertrikOUT is from host-to-device, IN is from device-to-host
20:59:39amiconnAnd it's not a bug fix, it's just a workaround that sacrifices speed
20:59:50Zagorno, that's not it
20:59:57Avebut, that aside, it works for me as well, if only I also set MAX define to 75
21:00:27bertrikwhat kind of endpoint setup does the original firmware use?
21:00:34Aveso the pll control bit is just a divisor for the base clock running at 96 megs or such
21:00:49Zagorbertrik: same type
21:01:39 Nick bagawk_ is now known as bagawk (n=lee@unaffiliated/bagawk)
21:02:53amiconnAve: No, the base clock is 24MHz
21:03:05 Quit SirFunk (Remote closed the connection)
21:03:30amiconnThe pll can be programmed to produce an (almost) arbitrary frequency, using a multiplication and a division factor.
21:03:40amiconnThis frequency must be >= 96MHz on PP5022
21:03:58amiconnThen a post divider can be applied, dividing that clock by 1, 2, 4 or 8
21:03:59Aveok so the control word has a divisor and multiplier bits
21:03:59bertrikdo you have a link to the lsusb output of a sansa (sorry, no libusb installed on windows)?
21:04:31preglowZagor: yo, you have any idea how the limits stuff in the menu is supposed to work? you seem to have touched the code, and right now, the code uses the current value here and there without checking to see if there's actually enough room buffer_alloced
21:04:46Aveok that makes sense now
21:05:13amiconnYes, plus some more
21:05:20*bertrik is installing libusb-win32
21:05:54amiconnbits 0..7 are the pll divider, bits 8..15 the pll multilpier
21:06:08amiconnbit 20..21 choose the post-divider
21:06:26amiconnbit 31 enables the pll
21:07:05amiconnFor 75MHz cpu clock, you need to run the pll at 150 MHz, and post-divide by 2
21:07:40Zagorbertrik: I'll throw it up
21:08:02Zagorpreglow: as best I can remember they were to be used to the initial buffer allocation at boot.
21:08:09bertrikAh, the sansa DOES use two endpoint numbers, 0x01 and 0x82
21:08:31Zagoroh yes
21:09:41Zagormy sandisk usb drive does not, however. it uses 0x01 and 0x81
21:10:42amiconnPerhaps the PP usb has a limitation (read: bug) that using the same endpoint number for in & out doesn't work?
21:11:37preglowZagor: yeah, exactly, and the limits at boot where supposed to be used until next boot, right, not the current values?
21:11:42***Saving seen data "./dancer.seen"
21:11:47Zagorpreglow: yes
21:11:56preglowZagor: there's a bug here currently, memory is corrupted, and i suspect it is because the current numbers are used here and there
21:12:06Zagoraha, badness
21:12:24preglowonly tree.c saves the the current number, plenty of other spots use the current number, but i don't know if any of them are able to do so safely
21:13:53LearHmm, tree.c:348 looks bad...
21:14:55 Part sheppard
21:15:20preglowLear: i tried changing that to max_files, didn't work out
21:15:30preglowi think the bug is in some other place
21:15:41preglowLear: at least i thINK i did so, i sometimes mess up
21:16:48 Join merbanan [0] (n=banan@
21:16:52 Quit ompaul (Client Quit)
21:17:15LearAnd you don't have any usage scenario to clue in to what it could be?
21:17:18preglowanyway, to reproduce, just start out with a lower limit, raise it, and see some dir
21:17:34preglowall the new files will not function, and the "open with" menu will be busted
21:18:35amiconnThat only reproduces the bug if you have dirs with more files than the previous limit, right?
21:18:36Aveamiconn: er ok so.. mul 25, div 4 and post 2 would work
21:18:41 Quit barrywardell (Read error: 110 (Connection timed out))
21:18:51Learpreglow: Using database in that case?
21:18:53preglowLear: no
21:19:05preglowamiconn: well, yeah, i don't expect the buffer will overflow if not
21:19:18preglowLear: i've got database files, but not loaded to ram, and using only the file browser
21:19:24LearWell, tagtree.c:1189 looks bad, regardless.
21:19:27*amiconn didn't manage to exceed the limits for ages...
21:19:43amiconnMy largest dir has a little less than 300 files
21:19:43preglowi set it low
21:19:46preglowto 200
21:20:18LearOught to be filetree.c:223 in this case...
21:20:36preglowyeah, what does ft_load really do?
21:20:43preglowit says it's dircache related, but it always gets called
21:20:50LearLoad a directory, I think...
21:21:21LearThat dircache is not the firmware/dircache.c, I'm pretty sure, but rather the internal buffer, allocated in tree.c...
21:21:27preglowahh, right
21:21:31preglowthen i got confused
21:21:39preglowbut yeah, i can't test right now
21:21:51preglowwill do later
21:23:30LearLooks indeed like the problem. It handles the name cache properly, but not the array pointing into the name cache.
21:23:46Zagorbertrik: perhaps I'm just confused. here's some output from usbmon and kernel.log:
21:23:47preglowin that case, i think this has been a bug for a long time
21:24:30Zagorthe first two lines show the sucessful reception and ACK of the INQUIRY command
21:24:38LearIndeed. But I guess most users don't go about changing those limits very often... :)
21:24:44bertrikZagor: this is your code interacting with Linux usb-storage?
21:24:51Zagorbertrik: yes
21:25:26preglowLear: i don't really know what i think about those settings, but then i guess i don't really like wasting memory when it can be helped
21:25:34Zagorthe next four lines is my failed attempt at reponding with the INQ struct and csw.
21:25:56bertrikWhat helped me a lot, was to run the USBCV utility from
21:26:10bertrik(you need a high-speed hub for that)
21:26:32Zagor...and windows :)
21:26:40 Quit billytwowilly (Remote closed the connection)
21:27:25Zagorlooks good though
21:28:28bertrikyes, looks good, current problem is probably in the BOT protocol. low-level USB comms should be quite OK if you get this far
21:29:33 Join webguest97 [0] (i=50d81efc@gateway/web/cgi-irc/
21:32:40bertrikit seems bulk IN still doesn't work
21:33:47Zagoryeah. i'll try a separate number
21:34:31 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
21:35:37 Join SkinInd95 [0] (
21:36:38 Quit SkinInd95 (Client Quit)
21:36:56rasherZagor: you should hold the code ransom for a fundraiser period once you get it working
21:38:57 Quit midgey ()
21:40:37Zagorno improvement with different numbers
21:41:57bertrikmmm, hard to say what's wrong
21:43:03bertrikI'll have a go at it tomorrow
21:43:31bertrikI think I'll browse a bit more through the Linux freescale code
21:44:28bertrikZagor: that is also the base you used for your code, right?
21:45:01ZagorI looked at it, but nothing is copied
21:46:13linuxstbZagor: Google would have paid you $4000 if you were a student...
21:46:46 Join webguest02 [0] (i=4c407df0@gateway/web/cgi-irc/
21:46:56 Join ompaul [0] (n=ompaul@freenode/staff/gnewsense.ompaul)
21:47:51 Quit webguest02 (Client Quit)
21:49:00pixelmais that a new record?
21:49:04 Quit linuxstb (Read error: 104 (Connection reset by peer))
21:49:30rasherpixelma: not unlikely
21:49:38 Join linuxstb [0] (
21:50:05rasherThe cgi:irc login screen should have a short notice about having patience
21:50:09 Join webguest43 [0] (i=4c407df0@gateway/web/cgi-irc/
21:50:25pixelmaoh... welcome back
21:50:31-->"help" received from wlw (
21:50:40-->"join #panfs-client-push" received from wlw (
21:51:03 Quit mf0102 ("Verlassend")
21:51:44webguest43is anyone here?
21:51:53markunwel, I am
21:52:03scorche|wspeak for yourself!
21:52:03webguest43can i get some help with my ipod?
21:52:10webguest43im having trouble rebooting it
21:52:14markunwhat kind of help do you need?
21:52:28scorche|wwebguest43: well, is it frozen, or do you just want to shut it down?
21:52:45webguest43well, i had just put on some new icons to my ipod. and i put it in my charger
21:52:54webguest43then when i went to turn it back on, it wouldnt boot
21:53:05scorche|whold select and menu
21:53:15webguest43tried that
21:53:23webmindwebguest43, might the battery be empty ?
21:53:35webguest43if its plugged it, i get the battery icons with a lightning bolt
21:53:39webguest43but i cant do anything
21:53:41scorche|wturn hold on and off, then hold menu and select for up to a min
21:55:05webguest43i tried that, and it flashed the screen 4 times but still at the same screen
21:55:24webguest43i tried putting it into disk mode, but it went to the same screen
21:56:57 Join billytwowilly [0] (
21:58:19webguest43if I unplug the ipod, and go into the diaognostics menu, i can chose diskmode, but it just goes back to the same screen
21:58:35webguest43could it be a bootloader problem?
21:59:10scorche|wdefine "same screen"
21:59:30webguest43 screen that has a lightning bolt next to a battery icons
21:59:36webguest43but it doesnt seem to be charging
21:59:47webguest43its an ipod colour if that helps at all
22:00:56 Quit newbyx86 ()
22:01:19Zagorehe, if I stop being stupid bulk IN works when it gets a separate number
22:01:19 Quit webguest97 ("CGI:IRC (EOF)")
22:03:11 Quit Lear ("ChatZilla [Firefox]")
22:04:25 Quit homielowe (Read error: 110 (Connection timed out))
22:05:29webguest43any ideas what could have gone wrong? im suspecting the bootloader, because the rockbox firmware is default so the ipod would try that first, then if theres a rockbox error it just stops loading at the apple screen. however, if i have the hold switch on while i reboot, it just halfboots then tries that again, and again, etc
22:06:08rasherAre you doing this with the charger connected?
22:06:27webguest43what i just said?, no
22:06:39markunZagor: was just a silly mistake?
22:06:56rasherwebguest43: Have you tried this: "<scorche|w> turn hold on and off, then hold menu and select for up to a min" without the charger connected?
22:06:56scorche|wwebguest43: if you cant get into disc mode, then something else is wrong
22:06:59Zagorthe quick test hack was a bit too quick
22:07:33 Nick Tanuva is now known as Tanuva|Zzz (n=tanuva@
22:07:48Zagorso back to usb-storage fixing now
22:08:03webguest43ok, now it wont even do what i said
22:08:08webguest43if i unplug it
22:08:12webguest43and hold menu and select
22:08:36scorche|w...and turn the hold switch on and off?
22:08:40webguest43it shows an icon with a battery and an excelmation mark inside a trianglur box
22:08:41*preglow puts champagne in freezer
22:08:57scorche|wwebguest43: do it again then
22:09:44 Join major_works [0] (n=a08e9562@
22:10:09webguest43does the icon meen theres something wrong with my battery?
22:10:12webguest43or that its low power
22:10:45webguest43also, if i plug it into my computer, my pc wont recognise it, its not even shown under my computer
22:10:51preglowlinuxstb: eh, doesn't wma get the metadata in the usual way?
22:10:54rasherwebguest43: Have you tried this: "<scorche|w> turn hold on and off, then hold menu and select for up to a min" without the charger connected?
22:11:13rasherAnd you held menu and select for a long time?
22:11:23webguest43ill even try it agian
22:11:27webguest43plugged in or not?
22:11:37rasherHow long would your battery last on an average day, before this?
22:11:54linuxstbpreglow: What do you mean?
22:12:02preglowlinuxstb: i can't see any case for it in metadata.c
22:12:13preglowin get_metadata()
22:12:48preglowi don't see any entries for that either
22:12:51webguest434hours, in rockbox doing games and stuff. in apple OS, not exactly sure (long time though)
22:12:56preglowit's here...
22:13:07*preglow looks at vim with suspicion
22:13:28 Join qweru [0] (
22:13:37maxkelleyas you should.
22:13:46 Nick parafin is now known as parafin|away (
22:13:58preglowlinuxstb: doesn't look like filesize member is filled
22:14:40 Join _michael [0] (
22:14:41 Quit amiconn (Nick collision from services.)
22:14:43 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
22:15:05linuxstbpreglow: Hmm... Will you fix?
22:15:15 Quit ilgufo ("So Long, and Thanks For All the Fish -")
22:15:42preglowlinuxstb: sure, i'll just stuff the usual filesize = line somewhere in get_asf_metadata
22:18:35preglowi get a bloody error in settings_list.c when applying/unapplying the scrollwheel accel patch
22:18:46preglowthere is a dependency problem here somewhere
22:18:59preglowa make clean fixes it
22:19:05rasherYour dependency on beer?
22:19:07preglowseems there's some trouble with voice ids
22:19:27 Quit nicktastic ("Leaving")
22:19:36preglowhey, i can quit drinking anytime i want to!!!!11!
22:21:03markunbut you don't want to of course :)
22:21:03webguest43should i just leave my ipod connected to the charger and see what happens in a couple of hours?
22:21:07 Quit merbanan (Remote closed the connection)
22:21:19markunpreglow: did you see we had fun with belgian beers in NY?
22:21:36preglowmarkun: i did not...
22:21:56markunthe attached picture
22:22:17rasherwebguest43: try that
22:22:21preglowdaamn, i want one of those glasses :P
22:22:24preglowbeen looking for them for ages
22:22:35markunstella or chimay?
22:22:51 Join petur [0] (n=petur@rockbox/developer/petur)
22:23:07markunpetur: was is the beer trigger which brought you here? :)
22:23:30peturno, my car.. just got home
22:24:23preglowso your car triggers on beer talk on irc now? :P
22:26:19maxkelleyONTOPIC! :p
22:26:39webguest43 should i just leave my ipod connected to the charger and see what happens in a couple of hours?
22:26:44peturdevs always talk on tpic :p
22:26:57peturtopic even
22:27:09bluebrotherisn't beer a topic in here? ;)
22:27:15preglowdamn straight it is
22:27:30 Join lee-qid [0] (
22:30:01 Quit ompaul ("restarting client - new config")
22:30:36 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
22:31:20rasherIn PERL, can't you do this: my ($foo, $bar) = function(); sub function { return ("foo", "bar"); } #?
22:31:54rasherI'm seeing some strange behavior
22:32:07parafin|awayyou can, AFAIK
22:32:10Zagoryes you can
22:32:20rasheroh right... I think I know.. It's something to do with lists collapsing in strange and frustrating ways
22:32:40rasherThat's come to bite me in my buttocks
22:33:20 Join newbyx86 [0] (
22:34:21preglowreally? what you're doing there is just fine
22:34:34rasherWell my code is not quite as simple
22:35:09rasherI'm trying to return two hashes and two scalars (really LibXML::Element objects)
22:35:25rasherAnd perl is having none of it
22:36:09webguest43so guys, any idea what idea could be wrong with my ipod?
22:36:10preglowyou're returning them by reference?
22:36:17 Quit webguest43 ("CGI:IRC")
22:36:32rasherpreglow: I think not.. I'm on that lead.. how do I do that?
22:36:40preglowi don't know that you can return several hashes or lists without doing so by reference
22:37:01preglowknow what, i don't really remember, i haven't used perl for years, even though i wished i had opportunity to do so :/
22:37:10Zagor\%hash and \$element
22:37:19rasherWell that's deceptively simple
22:37:34rasherAnd how do I access them, then?
22:37:37preglowyeah, then do my ($hash1, $hash2) = function(); and %$hash{element1};
22:37:37preglowi think
22:37:52preglowof course
22:37:58preglowi love perl, but i just haven't had anything to use it for :/
22:38:25rasherLooks like I can do \($element, %hash)
22:38:27 Quit kubiix (Read error: 104 (Connection reset by peer))
22:38:30rasherIf I'm reading this right
22:38:40preglowwell, that just makes a reference to an anonymous list
22:38:44preglowand returns that
22:38:58preglowwhich is fine
22:43:20 Join BigBambi_ [0] (n=Alex@rockbox/staff/BigBambi)
22:43:49 Quit ompaul (Client Quit)
22:44:46n1spreglow: the problem with lang/voice ids appears when the enabled features for a target is changed. that dependency would require rebuilding almost all of apps/ when either features.txt or the config-*.h fiel changed for your target
22:45:12n1sBut as makefiles don't cooperate with me...
22:45:28bluebrotherpixelma: sorry, didn't got around looking at the latex paste earlier
22:45:46pixelmafeel free to improve :)
22:45:49bluebrotherbut I haven't found a better solution. I guess opt is doing some things here.
22:46:35bluebrotherI should find the time to read the latex book completely and understand the opt package −− but I don't think I'll find time to do this
22:47:05pixelmaI tried quite a few possibilities and most of the times the output of latex wasn't very helpful - the most out of it I got was something about the bottomrule
22:47:32 Quit spiorf (Read error: 110 (Connection timed out))
22:47:49 Quit major_works ("major_works has left the building...")
22:48:15rasherSo now I get: Can't call method "getElementsByTagName" on unblessed reference at ./genlangxml line 231.
22:48:25 Join spiorf [0] (
22:48:43rasherI know I need to use bless($input, <something>), but what that something is..
22:49:05rasher(this is for my xml for language files experiment - not just any off-topic chatter)
22:49:42ZagorI've never bothered with object-oriented perl, so I can't help you there
22:50:04rasherI'm beginning to understand that decision..
22:51:05preglown1s: right
22:52:57 Quit davina (Remote closed the connection)
22:55:02 Join davina [0] (
22:55:41 Quit _michael ("Kopete 0.12.4 :")
23:01:56 Quit BigBambi (Read error: 110 (Connection timed out))
23:03:22 Quit hcs (Read error: 113 (No route to host))
23:04:30 Join eigma [0] (n=cat@
23:04:36 Join contxt [0] (
23:05:00 Part contxt
23:05:19 Join jac0b [0] (
23:05:31amiconnphew, 575 lines asm file ...
23:05:49jac0bI am going to post my how-to for the gigabeat battery
23:05:58markunjac0b: looking forward to it
23:06:05jac0bshould I make a new page or add it to the already gigabeat page
23:06:09markunnew page
23:06:24markunGigabeatBatteryMod or something?
23:06:45jac0bdo I have rights to make a new page?
23:06:45markunmaybe Upgrade instead of Mod
23:06:51markundo you have write access?
23:07:01jac0bI have a wiki username
23:07:09markunwell, you need write access too
23:07:13markunwhat's your wiki name?
23:07:36 Quit eigma (Client Quit)
23:07:38markunok, I'll add you to the list
23:07:42 Quit ToHellWithGA (Nick collision from services.)
23:07:46 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
23:07:51 Join ToHellWithGA [0] (
23:08:16jac0bDoes anyone have a Gigabeat S series?
23:08:55markunjac0b: ah, you already are in the list..
23:09:01 Quit davina (Remote closed the connection)
23:09:03toffe82jac0b: yes
23:09:43*preglow _finally_ has low-latency audio in linux
23:09:54jac0bmarkun: how do I create a page
23:10:09 Join eigma [0] (n=cat@
23:10:24rasherjac0b: just enter the address you want to create
23:10:29markunjac0b: go to the URL and click "create"
23:10:38pixelmadoes someone actually have the powers to correct commit messages (Zagor)?
23:10:38jac0bI was wondering if I could get a person with a S series to try out this battery thing I ahve done with the F series
23:10:58markunjac0b: why don't you ask toffe82?
23:11:07toffe82jac0b: the battery on the S is thinner
23:11:09 Join advcomp2019 [0] (n=advcomp2@
23:11:10preglowpixelma: we all do, no? i think the svn admin command lets you
23:11:27amiconnIt has to be enabled on the server side
23:11:29*rasher wrestles some more with perl refs
23:11:34amiconn...but currently isn't
23:11:42pixelmaI thought there was something not set up right?
23:11:44***Saving seen data "./dancer.seen"
23:11:48jac0btoffe82: hmm well on ebay they were saying a battery for the gigabeat could fit a S series and a F series
23:12:38toffe82jac0b: I don't think that the battery of the F fit in the S, I will have a closer look tonight if I have time
23:13:17rasherpreglow: do you know, if I have a reference to a hash, how do I get a specific value? $$hash{'foo'} doesn't seem to work
23:13:27markunjac0b: he has a big collection of various gigabeats :)
23:13:28toffe82jac0b: if one is good for the S, it is good for the F
23:13:33jac0btoffe82: okay cool, also check the description on this
23:14:16jac0bit says the F and S series
23:15:10toffe82perhaps it is a the same domension as the one of the S, so it fit on the F without problem
23:15:22preglowrasher: it should
23:15:39jac0bmarkun: should I make a topic parent?
23:15:41preglowrasher: $hashref->{'foo'} will also work
23:15:49preglowrasher: but $$hashref{'foo'} should be just as fine
23:16:02markunjac0b: maybe GigabeatInfo ?
23:16:34preglowrasher: so i guess you haven't really got a hash reference after all
23:16:53rasherpreglow: not entirely unlikely, I must admit
23:17:17preglowtry printing the reference variable
23:17:17rasherAh, there we go.
23:17:25preglowit should say "HASH(XXXXX)" if it is a hashref
23:17:34rasherA series of stupid mistakes lined up
23:17:42preglowthe best kind of lineup
23:17:45rasher(such as not quoting a $ in a debug message)
23:18:28 Quit Domonoky ("Trillian (")
23:25:07 Quit petur ("Zzzzz")
23:25:12 Quit hcs ("Leaving.")
23:27:53rasherI take it the order of strings in the binary languagefiles doesn't matter?
23:28:18 Quit eigma ()
23:29:29amiconnIt does matter
23:29:38amiconnHmm, or not
23:29:50amiconnI'm probably confusing something here
23:30:25*amiconn is busy with cf asm
23:30:29rasherSeeing as their ids are written into the languagefile, it seems reasonable that oder doesn't matter, but I'm a bit afraid of digging into the language-loading code, so if anyone knew, that'd be helpful
23:30:48rasher(numeric ids, generated from the English language file)
23:30:57amiconnThe good news is that the asm function already decodes well... :D just move the dct block to iram now
23:31:29preglowamiconn: performance figures! :P
23:33:43rasherOkay, I'm pretty certain order doesn't matter. I'll not waste my time trying to retrieve the elements from the xml document in a specific order, then - something libxml seems unwilling to do, unless I'm missing something
23:34:07rasherlanguage_strings[id] = ptr .. That looks reasonable
23:35:07n1srasher: as long as the ids correspond to the ones in lang.c i guess you should be safe
23:35:39rashern1s: that should happen automagically, with any luck
23:36:36n1sthat caused a couple of fun bugs with the v2 conversion where the binary file got a gap in the ids so it got out of sync at a certain id
23:36:41rasherWe'll see once I attempt building with this system..
23:36:50amiconnHmm, less than expected :/
23:40:08 Join Diamondiceman200 [0] (i=c35d1584@gateway/web/cgi-irc/
23:40:47 Quit lee-qid ("aufwiederbyebientotsayonara")
23:40:53 Quit jgarvey ("Leaving")
23:41:31Diamondiceman200how do you use the recording function on the iPod 60g?
23:45:03markunjac0b: how's it going with the wiki?
23:45:42jac0bmarkun: just have to post some pic
23:46:02rasherHrr.. libxml is inexplicably converting my wonderful utf-8 into what I can only assume is latin-1
23:47:54linuxstbrasher: libxml stores all strings as utf-8 - at least, that's my experience with it. Are you using the Perl wrapper though/
23:48:03 Quit billytwowilly (Remote closed the connection)
23:48:24amiconnpreglow: With just the idct buffer in iram I'm above realtime on X5 now
23:48:40amiconn(for widescreen)
23:48:41rasherlinuxstb: I am, yes. Anything I need to look out for?
23:48:43 Quit bertrik ("off to read a book")
23:48:46amiconn32..33 fps
23:49:20amiconnI'll experiment with the whole decoder struct in iram, but I need to sacrifice some iram somewhere else then...
23:49:25rasherBoth $doc->encoding and $doc->actualencoding return UTF-8, but when I print a textnode it ends up as latin-1
23:49:31amiconn...otherwise it won't compile for MCF5249
23:49:50rasherAnd the actual file is indeed utf-8
23:50:11amiconnFullscreen is around realtime (24..25fps)
23:50:56amiconnOn H300, widescreen struggles... around 18fps
23:51:32markunamiconn: was the H300 OF 10 of 15fps?
23:51:48linuxstbrasher: I don't know, I've never used the Perl wrapper, just the native C.
23:51:49amiconnDidn't try fullscreen on H300 yet
23:53:16rasherlinuxstb: dang.. it's really very strange. It's an xml file I wrote using libxml itself (constructed from a non-xml langfile)
23:55:59markuntoffe82: do you have a good gigabeat F frontplate?
23:57:47keanuZagor, so I hear success?
23:58:43 Join freqmod [0] (
23:58:44 Join billytwowilly [0] (
23:58:53Zagorkeanu: yeah it's going ahead step by step.

Previous day | Next day