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 2008-09-22

00:00:14 Join XstrangeX [0] (
00:01:09 Quit webguest49 ("CGI:IRC (Ping timeout)")
00:01:57 Quit ompaul (Client Quit)
00:03:17 Join webguest46 [0] (n=4770e3f1@gateway/web/cgi-irc/
00:03:53 Join bughunter2 [0] (n=Jelle@
00:03:54 Quit BlakeJohnson86 (Read error: 110 (Connection timed out))
00:04:18mmadiais there a commit log available?
00:04:55Unhelpful is one easy way
00:05:03Unhelpfulor use svn log
00:05:18XstrangeXmanuals down for anyone else?
00:05:28mmadiathanks Unhelpful
00:06:02webguest46i have them on my ipod like i can see them, but when i open them nothing happens/changes
00:08:24 Quit webguest46 ("CGI:IRC")
00:09:12 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
00:11:09 Quit bertrik ("Leaving")
00:17:47 Join Yondering [0] (
00:26:22 Quit robin0800 (Remote closed the connection)
00:29:48 Quit tvelocity[a] (Read error: 104 (Connection reset by peer))
00:39:44 Join barrywardell [0] (n=barrywar@
00:40:33 Quit Yondering ("Leaving")
00:40:40 Join BlakeJohnson86 [0] (
00:40:59 Quit rphlx ("Leaving")
00:41:17Nico_Pamiconn: you got a gigabeat S?
00:41:44Nico_Pnice! :)
00:41:51Nico_Pgot projects with it?
00:44:26 Join eddieFig [0] (
00:44:47 Part eddieFig
00:48:08 Quit ender` (" All power corrupts, but we need electricity.")
01:06:17 Quit ajonat ()
01:07:50 Join goffa [0] (n=goffa@
01:16:06 Join reacocar` [0] (n=reacocar@
01:17:05 Quit reacocar` (Client Quit)
01:17:05 Quit reacocard ("Coyote finally caught me")
01:20:06 Join massiveH [0] (
01:22:02***Saving seen data "./dancer.seen"
01:22:41 Quit mf0102 ("Ex-Chat")
01:23:47 Join reacocard [0] (n=reacocar@noslor.ST.HMC.Edu)
01:25:36 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
01:26:46 Quit Nico_P (Remote closed the connection)
01:31:17 Quit barrywardell ()
01:43:35 Join miepchen^schlaf [0] (
01:44:10 Quit stripwax (Read error: 104 (Connection reset by peer))
01:44:17 Quit miepchen^schlaf (Connection reset by peer)
01:44:29 Join miepchen^schlaf [0] (
01:46:00 Quit jhMikeS (Nick collision from services.)
01:46:06 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
01:48:35 Quit perrikwp (" ajax IRC Client")
02:00:00 Nick fxb is now known as fxb__ (
02:09:33 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
02:12:48 Quit MethoS- (Read error: 104 (Connection reset by peer))
02:13:52 Join MethoS- [0] (
02:15:46 Quit MethoS- (Remote closed the connection)
02:22:14 Quit oofus (Remote closed the connection)
02:24:54 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
02:43:25 Join HBK- [0] (
02:44:03soapdoes anyone know why on JdGordon's commit earlier today my build server barfed a segfault on the Mrobe 100 sim, built everything else asked of it fine, and then after amiconn's completely unrelated commit I am able to build the same sim w/o problems?
02:44:09 Quit HBK (Nick collision from services.)
02:44:54 Nick JdGordon|zzz is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
02:44:55soapDid the SVN update happed to barf on a file for that failed build? Need I test my PC's memory? Sunspots? ;) Config problem I missed on my end?
02:45:04JdGordonsoap: good old gcc bugs probably :p
02:45:44 Quit culture (Read error: 110 (Connection timed out))
02:45:56soapwhen I logged into rbclient after the failed build, I was able to reproduce the failure, it wasn't a one time thing.
02:46:31 Join star_jasmine [0] (n=arwyneve@
02:46:40soapand amiconn's commit (the following commit) didn't touch a file gcc should ever see in said build, yet it works fine now.
02:47:43star_jasminehi... I"m having a problem creating a voice in cygwin for rockbox. it tells me that there are too many open files, and that it cannot open c:\program files\viavoice tts\tts.log. how do I prevent this from happening, or close logs?
02:49:32star_jasmineactually, it says too many open files, as the error, siting the name of the log file it wishes to open
02:49:49star_jasmineany suggestions would be much appreciated
02:56:18 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/ajax/
02:58:26 Quit Rob2222 ()
02:58:43cool_walking_star_jasmine: I would try closing as many other programs as you can, and trying again.
02:59:29 Join Inverse [0] (n=InverseC@
02:59:39 Quit Inverse (Client Quit)
03:03:07 Join Rob2222 [0] (
03:06:33 Quit jhulst (Read error: 110 (Connection timed out))
03:09:15 Join AndyI [0] (i=AndyI@
03:12:00 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
03:17:56star_jasmineok thanks. so nothing else would cause this? I restarted the computer after the first time
03:19:31 Quit moos ("Rockbox rules the DAP world")
03:20:02 Quit jhMikeS (Read error: 110 (Connection timed out))
03:21:06star_jasmineand it still occured a second time
03:21:17 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
03:21:28star_jasminedoes cygwin keep files open after an exit and restart?
03:21:53 Quit AndyIL (Read error: 110 (Connection timed out))
03:22:06***Saving seen data "./dancer.seen"
03:23:44 Quit star_jasmine ()
03:24:31cool_walking_I have no idea. Just a guess.
03:25:32 Quit dabujo (Read error: 104 (Connection reset by peer))
03:26:13 Join perrikwp [0] (i=d1a8d351@gateway/web/ajax/
03:26:49 Quit cool_walking_ (" ajax IRC Client")
03:31:42 Join dataangel [0] (n=prophet@
03:32:52dataangelIf I have my plugin open a textfile like so: rb->open(ROCKBOX_DIR "/myfile.txt", "r");, then rockbox (at least in the simulator) ERASES the file's contents @_@ That is, it replaces it with an empty file. Don't modify the file at all, just open it with only read permission
03:33:51dataangelwhether or not I close file doesn't matter
03:33:57dataangelfile is in archos/.rockbox
03:34:21dataangelany ideas? :P I'm totally lost
03:36:07 Join setkeh [0] (
03:38:07 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
03:38:13saratogahopefully mp3 seeking issues are now solved
03:38:37saratogaseems seeking while the COP is still processing a previous frame caused bad things
03:40:43 Quit saratoga (Client Quit)
03:41:32linuxstbdataangel: The "open" command doesn't take a string as the last parameter - you want to use O_RDONLY.
03:41:51linuxstbfunction even...
03:42:17 Quit jeffdameth (Read error: 113 (No route to host))
03:42:18 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
03:42:19 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/ajax/
03:43:12dataangelhrm, it worked once
03:43:34linuxstbOnly by chance.
03:43:47linuxstbDidn't the compiler give a warning?
03:43:57dataangelno :P
03:44:28dataangelbut I dismissed it for giving a similar warning earlier that meant nothing :P
03:44:35 Join jeffdameth [0] (
03:44:39dataangelthat works, thanks :P
03:44:45*dataangel returns to corner with dunce cap
03:51:46 Quit DerDome (Read error: 110 (Connection timed out))
03:52:15 Join DerDome [0] (
03:58:39 Join Darksair [0] (n=user@
04:02:02dataangelWhat is the 3rd parameter to rb->strtok_r for?
04:03:46JdGordonso it can be reenterant
04:04:20dataangelbut what do I give it?
04:08:57 Quit Seed ("cu, Andre")
04:08:58 Join miepchen^schlaf_ [0] (
04:20:11*jhMikeS reads back and is astonished that amiconn has just purchased an overpowered device and figures this must be related to worldwide economic collapse somehow
04:23:15*jhMikeS keeps going back to make sure he didn't misread that (amiconn+gigabeast) ??
04:24:06 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
04:24:55 Join homielowe [0] (
04:35:25 Quit dataangel ("Leaving")
04:38:50mmadiacould someone take a look at this wps clipping? I'm having some issues rewriting margin tags as viewports+scrolling tags.
04:39:33 Join mikkas [0] (n=795b670c@gateway/web/cgi-irc/
04:39:56mikkashey everyone :)
04:40:12*mmadia nods
04:40:58mikkasi just had a quick question oh wise ones, I have a iriver h340.. it's running rockbox (woo) but it's version 060622 or something like that... I was hoping to update it..
04:42:13*scorche looks for a question
04:42:16mmadiai seek wise ones as well.
04:42:47mikkasheh. So anyway, do i have to uninstall rockbox (the current version on there)... or can i just copy over and it'll update?
04:42:57 Quit cool_walking_ (" ajax IRC Client")
04:43:18scorchebeing that it is so old, you will likely have to update your bootloader as well
04:43:35scorchejust "install" as said in the manual
04:43:59mikkasah excellent. cheers. :)
04:46:34 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/ajax/
04:47:57mikkasso to update my bootloader.. ill need the firmware again.. whatever the newested is..
04:48:35mikkasoh looks like there's an idiots guide hehe (utility)
04:51:57 Quit jhulst (Read error: 110 (Connection timed out))
04:54:34 Join ajonat [0] (n=ajonat@
04:54:49 Join andrewsund [0] (
04:54:54mmadiaaah, Ilumini needs kugel's build. oops :)
04:58:04andrewsundhey everyone. i bought a sansa today and will be retiring my ipod nano 1st gen. it's a little beat up, so resale value is going to be low. i figured i'd give it to a dev. anyone interested?
05:00:03 Join NinJew [0] (i=Bensawso@unaffiliated/bensawsome)
05:00:28scorcheandrewsund: thanks for the offer, but not many devs are typically active during these hours..perhaps you might get better results posting to the dev list?
05:02:44andrewsundyeah, i didn't want to join up to post a single question, but i'll give it a shot.
05:06:15mikkasis there going to be a rockbox version to run on the Zune?
05:07:07scorchemikkas: i left my crystal ball at work today, but for the forseeable future, one is currently working on it as far as i am aware
05:07:37mikkas^_^ you left it at work? lol
05:07:57NinJewlol nice
05:14:10 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
05:17:30 Part andrewsund
05:18:19mikkasarrgh i cant get the bootloader installed on my h40 :( when it asks for my hex file i give it the file.. but it says can't detect file..
05:18:38mikkas"could not detect firmware type"
05:20:25mikkasoh wait.. never mind.. im trying to use unsupported firmware.. the newest korean is 1.31k or something... ill have to downgrade
05:22:09***Saving seen data "./dancer.seen"
05:24:04 Join andrewsund [0] (
05:24:12 Part andrewsund
05:24:29 Quit homielowe ()
05:24:42 Join homielowe [0] (
05:25:17 Quit setkeh ("Leaving")
05:33:47 Quit mmadia (Read error: 104 (Connection reset by peer))
05:37:53 Quit NinJew ("t3h 1337 h4X0R sT4t80t h42 QuiT")
05:43:15 Join HBK [0] (
05:50:09 Quit XavierGr (Read error: 60 (Operation timed out))
05:50:45 Quit Horscht ("I got raided by the FBI and all i got is this lousy quit message")
05:51:37 Join vitja [0] (n=vitja@
05:52:00 Quit HBK- (Read error: 60 (Operation timed out))
06:03:44mikkasso i finally got it installed... however im getting bad bad static in the 1 channel of my headphones... what's up with that?
06:04:22 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
06:08:28 Quit mikkas ("CGI:IRC (EOF)")
06:08:46 Join JdGordon|phone [0] (
06:28:24 Quit JdGordon|phone (Remote closed the connection)
06:36:43 Quit jhulst (Read error: 113 (No route to host))
06:54:36 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
07:08:17 Quit bmbl ("Woah!")
07:09:33 Part toffe82
07:17:34 Join webguest20 [0] (n=456dd8de@gateway/web/cgi-irc/
07:17:55webguest20does rockbox work on the v2 e200 yet?
07:18:32 Quit webguest20 (Client Quit)
07:21:12 Quit FugitivePrisoner (Read error: 104 (Connection reset by peer))
07:21:53 Join FugitivePrisoner [0] (
07:22:11***Saving seen data "./dancer.seen"
07:29:05 Join einhirn [0] (
07:33:47 Quit faemir (Remote closed the connection)
07:36:02 Part FugitivePrisoner
07:36:09 Join FugitivePrisoner [0] (
07:36:18 Part FugitivePrisoner
07:49:51 Quit XavierGr ()
08:02:27 Quit miepchen^schlaf_ ()
08:13:30 Quit GodEater (" ajax IRC Client")
08:13:41 Join GodEater [0] (i=c2cbc962@gateway/web/ajax/
08:22:26 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
08:25:45 Join coatman [0] (
08:27:08 Quit BigBambi (Read error: 113 (No route to host))
08:33:14 Quit Nibbler (Read error: 110 (Connection timed out))
08:33:57 Join Nibbler [0] (
08:34:19 Quit massiveH ("Leaving")
08:34:30 Join massiveH [0] (
08:35:02 Quit massiveH (Client Quit)
08:39:12 Join massiveH [0] (
08:41:19 Join ender` [0] (
08:41:21 Quit massiveH (Client Quit)
08:41:33 Join massiveH [0] (
08:42:28 Quit massiveH (Client Quit)
08:44:47 Quit bughunter2 ("bye")
08:50:20 Join bertrik [0] (
08:50:55 Quit BHSPitMonkey (Remote closed the connection)
08:51:33 Join massiveH [0] (
08:53:29 Join Rob2223 [0] (
08:55:01bertriknice, just got 16 hrs runtime on my sansa c240 in a battery bench with r18560 (which does mp3 decoding partially on the cop)
08:56:49jhMikeSbertrik: what was it before?
08:57:33bertrikI'll have to check, slightly less than 15 h IIRC
08:59:13bertriklast time with the same mp3 set was 14:41 with r18085
09:00:13 Quit massiveH ("Leaving")
09:00:27 Join massiveH [0] (
09:01:07 Join vitja_ [0] (n=vitja@
09:02:10 Quit vitja (Remote closed the connection)
09:02:10 Quit vitja_ (Remote closed the connection)
09:02:21 Join vitja [0] (n=vitja@
09:11:37 Quit Rob2222 (Read error: 110 (Connection timed out))
09:12:36 Quit massiveH ("Leaving")
09:20:52 Join lasser [0] (
09:22:14***Saving seen data "./dancer.seen"
09:22:39 Quit bertrik ("Leaving")
09:23:12 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
09:52:39 Quit Acksaw ("Erection reset by queer")
10:00:06vitjagevaerts: I found, that was my bug ;)
10:00:31 Join Nico_P [0] (i=937f60dd@gateway/web/ajax/
10:01:05 Quit linuxstb (Read error: 60 (Operation timed out))
10:01:18 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
10:02:11 Join JUSTWJX [0] (n=79e8e0bc@gateway/web/cgi-irc/
10:02:26JUSTWJX#join rockbox-community
10:03:36JUSTWJX#join rockbox-community
10:03:58Nico_Pit's /join
10:04:04scorchemove the # to before rockbox and add a / at the beginning of the line
10:04:20JUSTWJX #rockbox-community
10:04:35scorchestill need "/join"...
10:04:57JUSTWJXhow to join rockbox-community?
10:05:13scorchewe juts told you
10:05:22ajonat /join #rockbox-community
10:06:19 Quit JUSTWJX (Client Quit)
10:06:51 Join JUSTWJX [0] (n=79e8e0bc@gateway/web/cgi-irc/
10:07:21JUSTWJXso complex~
10:08:23advcomp2019ajonat was right
10:08:33scorchefor the third time, "/join #rockbox-community"
10:09:40 Quit Nico_P (" ajax IRC Client")
10:12:49 Quit JUSTWJX ("CGI:IRC (Ping timeout)")
10:14:00 Join Acksaw [0] (
10:14:13 Join JUSTWJX [0] (n=79e8e0bc@gateway/web/cgi-irc/
10:15:38ajonathmm.. r18557 doesn't have any effect on an ipod nano, right?
10:19:43 Quit JUSTWJX ("CGI:IRC (Ping timeout)")
10:22:17 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
10:31:41 Quit Nico_P (Read error: 104 (Connection reset by peer))
10:31:42 Quit reacocard (".")
10:31:55 Join reacocard [0] (n=reacocar@noslor.ST.HMC.Edu)
10:34:44 Quit cool_walking_ (" ajax IRC Client")
10:34:56 Join culture [0] (
10:38:17 Join Frostbyt3 [0] (
10:42:28Frostbyt3website dont work
10:43:11Frostbyt3whats goin on
10:45:09 Part Frostbyt3
10:52:53 Join dany_21a [0] (
10:58:32 Quit XavierGr ()
10:58:57 Join BigBambi [0] (i=86ceaf40@rockbox/staff/BigBambi)
11:02:48 Join n1s [0] (n=nils@rockbox/developer/n1s)
11:05:46amiconnajonat: It does
11:08:13ajonatamiconn, #if (CONFIG_CPU == PP5024 || CONFIG_CPU == PP5022 || CONFIG_CPU == PP5020 || CONFIG_CPU == PP5002) && !defined(MPEGPLAYER)
11:08:34ajonati though nano had a PP5021C CPU
11:10:10ajonatanyway, I'm trying it today :) have to go now, bye!
11:10:15 Quit ajonat ()
11:10:16scorcheajonat: it does, but it is defined as a PP5022
11:10:26*scorche sighs
11:12:19markunscorche: second time today :)
11:22:16***Saving seen data "./dancer.seen"
11:52:32 Part dany_21a
11:53:19BigBambiCould this be due in any way to saratoga's decoding on COP?
11:58:13linuxstbBigBambi: I would have expected others to complain if that was the case, but you never know - you could ask him to try a daily from before saratoga's first commit.
11:58:33*linuxstb missed the "20th sept works fine" comment...
11:58:37 Quit culture (Read error: 104 (Connection reset by peer))
11:58:54BigBambilinuxstb: It is the second report as well
11:59:03BigBambiI've just merged the other post into it
11:59:23BigBambiHere is the combined one:
12:00:06BigBambiI don't have my H140 with me so can't test
12:00:08linuxstbWe need to send saratoga a coldfire target...
12:02:48*n1s tests on his h300
12:03:08BigBambin1s: cheers
12:04:35 Join daurnimator [0] (i=daurn@unaffiliated/daurnimator)
12:11:20n1syep, r18556 is fine r18557 has a nasty noise in the right channel with mp3 (left is fine)
12:13:52n1sand r18563 is still affected
12:18:42 Quit linuxstb (Read error: 110 (Connection timed out))
12:23:35jhMikeSamiconn: Congrats, you got a beast? I must admit I'm a bit surprised.
12:31:30 Join moos [0] (
12:35:49 Quit reacocard (".")
12:39:44 Quit scorche (Read error: 104 (Connection reset by peer))
12:40:45 Join scorche [0] (i=Blah@rockbox/administrator/scorche)
12:47:12 Join tvelocity [0] (n=tony@
12:47:41markunjhMikeS: what are you up to these days?
12:47:58markunworking on anything rockbox related?
12:50:15jhMikeSmarkun: yes. currently trying to figure out how to shoehorn this S charging code into things and have it work properly.
12:53:09amiconnDon't shoehorn it. The powermgmt code needs a thorough rework.
12:54:14 Join funman [0] (
12:55:19funmanhum no kugel here
12:55:33funmanhe told yesterday that he would be here 'tomorrow'
12:56:08jhMikeSamiconn: indeed. it shouldn't need much but it would be nice to have charging in SVN and it should be liveable for now. adjusting it should be simple. it really should have an event queue or it can miss important transitions to act accordingly.
12:57:22jhMikeSamiconn: what overall restructuring did you have in mind, if any?
12:58:18amiconnThe basic powermgmt loop needs to be reversed. Right now there is a "short wait" and a "long wait" function.
12:59:03amiconnThe loop should rather be always active and should call the "seconds-task" and the "minutes-task" at configured intervals
13:00:01amiconnThis is the only way to guarantee prompt reaction to events. I need this for suspend on 1st/2nd gen.
13:00:02jhMikeSamiconn: the S patch has 99% in the short wait function (so it is as you put it "reversed")
13:00:20amiconnThere should be no short wait or long wait functions
13:03:31amiconnThere should be functions for the various tasks, which get called from the main loop regularly, and which return immediately after doing their part of the job
13:07:01jhMikeSyes, that would flatten it out nicely. what about the NiCd code−− it has dependence on counting
13:08:25amiconnIt needs to be adjusted of course
13:08:35 Join lachlan12345 [0] (n=7a681d06@gateway/web/cgi-irc/
13:09:09amiconnIt's target specific, hence should be moved into the target tree
13:09:39 Join dany_21a [0] (
13:09:53funmanhi dany_21a
13:10:48dany_21ahello funman :)
13:12:51dany_21afunman: regarding GIT-repo, i am not sure what directory/file-layout we sholud use... i think it would be the best to use different and meaningfull names for the test.S-files. And not different branches or other GIT-foo
13:14:38jhMikeSamiconn: Hmmm...when? I did put a few things in to help target certain things in the patch since it needed that.
13:14:47funmanyes it's simpler with different file names
13:17:02lachlan12345hello all. i'm a rockbox (ipod 5.5.G) user but i like to boot into the stock firmware for games and suchlike. this leaves me with an accessing music from both firmwares conundrum. i know i can build a database in RB off my itunes files, but i prefer to browse by folder view. is it in anyway possible to build and itunes database off of my music collection which i have stored on my ipod in normal files format
13:17:55funmandany_21a: what's the private repository url ?
13:18:09dany_21afunman: you mean for pushing?
13:18:24 Quit tvelocity (Read error: 110 (Connection timed out))
13:18:25funmanhttp ??
13:18:28dany_21a(vistit that site.,.. not that urtl)
13:18:51*jhMikeS thinks exactly what these tasks are needs to be well defined
13:18:51 Join tvelocity [0] (n=tony@
13:18:57funmanit only lists public git & http urls
13:19:39lachlan12345black bomb a myspace
13:19:41funmanlet's see if i login ..
13:19:53dany_21ahm... i see here as third point the push url...
13:20:01dany_21atry - maybe it works
13:20:28dany_21aand you have to upload your ssh-pub key
13:20:42Bagderlachlan12345: that's an issue for your OF and apple fanboys, we discuss how rockbox works here!
13:20:51funmandany_21a: ok I pushed my first commit, thanks you ;)
13:21:05funmanI deleted my message on the forum btw
13:21:17funmanAuthor: daniel <daniel@kubuntu.(none)>
13:21:24funmanyou might want to setup your global git config
13:21:57funmangit config −−global
13:22:01funmangit config −−global daniel@..
13:22:06lachlan12345Badger: 'fanboy' is a bit harsh- RB is still my primary firmware. all i want is harmony between the two systems!
13:22:10funmangit config −−global You
13:22:19***Saving seen data "./dancer.seen"
13:23:06 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
13:24:29dany_21athx for the hint funman, i thought by pushing it automatically rewrites it according to my useraccount on that site
13:24:58funmanno, git commits are 100% legal emails and you can change your name/email in each commit
13:25:38funmanyou can also commit a patch from someone else and you would only be listed as a Committer
13:25:46funmanbut you can fake the Committer line as well ;)
13:26:12dany_21aokay - still need to get used to GIT :)
13:26:17funmanI can give you hints on git if you want
13:26:32funmanI have learned when vlc transitioned from svn to git 6 months ago
13:26:46dany_21athx - will come back onto that, if i need some help
13:26:57n1sBagder: about bootloaders for the release, do you have some nice script to build "release" version? (with proper version numbers etc) and preferably build them all in one go? The test bulds i put up seem fine except for sansa c200 so maybe that should be excluded from release?
13:28:10funmanI will add buttons/other infos in the git repo this afternoon
13:29:36 Quit lachlan12345 ("CGI:IRC (EOF)")
13:33:14 Quit JdGordon (Remote closed the connection)
13:34:29 Join JdGordon [0] (
13:38:00funmandany_21a: we should also put a README
13:38:46 Join MethoS- [0] (
13:40:00 Quit funman ("leaving")
13:42:31BigBambin1s: I'll open a bug report on it.
13:54:26 Part LinusN
13:54:53 Join dabujo [0] (
13:55:40 Quit Zambezi (Read error: 110 (Connection timed out))
13:57:47 Quit vitja (Remote closed the connection)
14:05:00 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
14:05:37 Quit jeffdameth (Read error: 113 (No route to host))
14:13:04 Nick m0f0x_ is now known as m0f0x (
14:18:42 Join Seed [0] (
14:23:12 Join Schmogel [0] (
14:38:31 Join Zambezi [0] (
14:52:45 Join kugel [0] (n=chatzill@unaffiliated/kugel)
14:53:11kugeldany_21a: that's why I asked ffor support :)
14:53:49kugelalthough the e200 has more in common with the fuze than with the clip, e.g. size of the firmware file
14:57:16 Quit pabs (Remote closed the connection)
14:57:20 Join pabs [0] (
15:01:43 Join J-23 [0] (
15:07:32 Quit Seed ("cu, Andre")
15:07:50 Join LambdaCalculus37 [0] (i=44a04303@gateway/web/ajax/
15:11:48*gevaerts hands LambdaCalculus37 a mug of hot chocolate
15:12:40*GodEater is going to the dark side
15:12:46*LambdaCalculus37 thanks gevaerts and takes the mug, and gets out some biscuits
15:13:02LambdaCalculus37Oops... wrong channel. :P
15:22:00 Join mib_74uswniw [0] (i=cdcdc8ee@gateway/web/ajax/
15:22:04 Nick mib_74uswniw is now known as MarcGuay (i=cdcdc8ee@gateway/web/ajax/
15:22:22***Saving seen data "./dancer.seen"
15:28:33 Join webguest75 [0] (n=3eeb9ee8@gateway/web/cgi-irc/
15:29:58LambdaCalculus37Sixth what?
15:30:12webguest75Yeh this thing ate my whole line
15:30:17webguest75webclient kinda suc
15:30:20webguest75Lemme get on irc
15:30:25 Quit webguest75 (Client Quit)
15:31:06 Part J-23
15:32:55 Join MethoS-- [0] (
15:39:52 Quit tvelocity (Remote closed the connection)
15:40:05 Quit Zarggg ()
15:40:13 Join lalala [0] (n=3eeb9ee8@gateway/web/cgi-irc/
15:40:34lalalawhat server is this?
15:40:49lalalaSince it doesn't seem to be
15:40:59 Join mf0102 [0] (
15:41:02lalalaCuz there are 2 ppl on #rockbox over there
15:41:11lalalaIncluding me ...
15:41:25LambdaCalculus37There are 128 users on #rockbox right now.
15:41:48lalalaWell not on #rockbox on as stated on the site
15:41:53lalalaOr it has split
15:42:37 Join dark [0] (
15:42:40darkAh that helps
15:43:01 Quit lalala (Client Quit)
15:43:03darkShould script irssi to follow network hopping
15:43:04 Quit MarcGuay (" ajax IRC Client")
15:43:14darkAnyway, so here goes again
15:43:38darkMy gf would like me to put rockbox on her ipod. But I just want to make sure what generation it is, I don't have an ipod myself.
15:43:56darkHers is an 80GB I think, with color screen and capable of playing video. Ipod Classic
15:44:15LambdaCalculus37And iPod Classic will not work with Rockbox.
15:44:17 Join mib_cl72ftku [0] (i=cdcdc8ee@gateway/web/ajax/
15:44:22 Nick mib_cl72ftku is now known as MarcGuay (i=cdcdc8ee@gateway/web/ajax/
15:44:40darkAh, really? Mmm, sux
15:45:02darkTalking to a bot aren't I =P
15:46:50 Quit MethoS- (Read error: 113 (No route to host))
15:47:21LambdaCalculus37We only have one bot, and that's not me.
15:47:35darkAh, hehe, k
15:47:44LambdaCalculus37We have a list of supported players at
15:47:56gevaertsdark: the description you gave could also be for anipod video 5.5th gen
15:48:05darkYeh I saw, it listed ipod video but not classic, so I wanted to make sure
15:48:30darkShort reason why it won't work on classic?
15:48:47BigBambiencrypted firmware, new undocumented hardware
15:49:16darkThose are solid reasons
16:07:31 Quit Sargun (Read error: 104 (Connection reset by peer))
16:11:55 Quit Siku ()
16:19:15 Quit mf0102 (Remote closed the connection)
16:20:39 Join joetoo [0] (
16:22:01 Join desowin [0] (n=desowin@atheme/member/desowin)
16:22:16joetooWeird. I just looked in to find out where #rockbox was, and it said there wasn't one.
16:26:55joetooI notice that rockbox has basic video decoding support, but I'm interested in xvid. I see that the V.2 sansa e200's have divx, which makes me wonder if my V.1 has enough cycles to decode xvid. Looks like my OEM firmware only supports mjpeg in .mov, which obviously uses a lot of space.
16:27:34 Quit lasser (Remote closed the connection)
16:27:49markunjoetoo: interested as in "I want to work on it"?
16:27:54joetooAnyone know if there is 1) any way to play xvid or similar, 2) any plans to add such support, or 3) that this just ain't gonna happen?
16:28:29n1s1) if someone codes it 2) not afaik 3) see 1 & 2 :)
16:28:44joetoowell, I have some experience in embedded programming, but I've never tried to do it portably, nor have I ever worked with codecs.
16:29:35BigBambigood time to start then
16:29:49n1sit is a good opportunity to increase your experience then! but expect it to be a _lot_ of work
16:30:29n1sjoetoo: any idea about how much processing power divx uses in relation to mpeg2?
16:30:36BigBambiI think most people feel that given Rockbox has very good support for MPEG 1/2 that the advantages of using xvid are very very small compared to the amount of work
16:31:07joetooI'm sure it's more, but we are also talking about very small screens and possibly low framerates.
16:31:12markunBigBambi: but with players getting faster, I see how it would be easier to just copy an existing file over without transcoding.
16:31:45BigBambimarkun: Sure - I'm not saying that is completely pointless, but hithertoo nobody has felt the advantage is big enough to bother to do the work
16:32:03BigBambijoetoo: very small screens sure, but also very slow processors and low memory
16:32:08markunjoetoo: maybe you could try to do some benchmarks on a PC, it might give you an idea how much more complex mpeg4 (xvid/divx) decoding is.
16:32:24n1smarkun, BigBambi: and on the beast we have tv out so there's more use for larger (as in res) files :)
16:32:33BigBambivery true :)
16:32:44joetooin my case, I have a 2GB player, with microSD (not HC). Though I gather using rockbox enables micro SDHC. So storage space has been a concern for me.
16:33:01n1s(well, the ipod videos have tv out too but afaiu it's tied to the broadcom chip)
16:33:30joetoomarkun: that might make sense. I could transcode something to the right size, then see what the cpu load is with different codecs.
16:33:41BigBambijoetoo: Enough of a concern to write an xvid codec for Rockbox?
16:34:06joetoowell, writing one is a bigger deal than porting one to a different libc
16:34:16markunjoetoo: someone got started on it a while ago, but I don't know if you could reuse any of his work:
16:34:31BigBambiThat is all my point is - if the difference matters enough to you then you will do the work. So far the difference hasn't been important enough to anyone
16:34:47JdGordonjoetoo: ignore them... do it and the worst that will happen is that you dont get a usable framerate, in which case its likely someone will come and help optimize it...
16:35:06markunjoetoo: ignore JdGordon :)
16:35:08BigBambijoetoo: OK, port. But given things like no floating point, no malloc, it could (I say could, I don't know) almost be writing
16:35:44BigBambijoetoo: I'm not trying to discourage you incidently, I'd like to see it personally :)
16:35:44*JdGordon goes back to wondering how the hell to do a recent bookmark list :(
16:36:49joetooThe framerate issue is another concern. It is imaginable that at a low framerate, the deltas between frames would be huge no matter what you do. Some codecs may not save much space or may have very poor quality at low frame rates.
16:37:04 Quit Schmogel (Read error: 54 (Connection reset by peer))
16:37:13 Join toffe82 [0] (
16:37:25markunjoetoo: you could start with the ffmpeg or xvid source code, don't know which would be better.
16:37:27 Quit amiconn (Nick collision from services.)
16:37:29joetoosounds like further research is in order.
16:37:34 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
16:39:00 Join saratoga [0] (n=9803c264@gateway/web/cgi-irc/
16:40:04markunsaratoga: any plans for further codec speedups?
16:40:45BigBambisaratoga: I've just seen your reply to the forum - it seems a bit of an odd bug then!
16:40:51 Join Schmogel [0] (
16:41:22saratogaBigBambi: not to mention amiconn benchmarked on his Iriver yesterday
16:41:37BigBambiWhich iriver?
16:41:47saratogaone of the Coldfire ones
16:42:14saratogahe didn't mention problems, though i suppose he may not have noticed either
16:42:33saratogamarkun: yes many, but little time right now
16:42:57JdGordonBigBambi: for a recent bookmark list.. is the bookmark filename enough?
16:43:02 Join lukstr_ [0] (n=lukstr@
16:43:04markunsaratoga: anything as big as the shared imdct?
16:43:07JdGordonor it has to be as it is now?
16:43:11saratogadoes the problem happen for all tracks on all CF or only some?
16:43:26saratogamarkun: for AAC?
16:43:29 Part lukstr_
16:43:37markunsaratoga: for any codec
16:44:17saratogaeventually i want to put AAC on COP and look into getting HE working on more targets
16:44:28LambdaCalculus37saratoga: I updated my Gigabeat F's build to r18563 this morning. It's working fine; no MP3 glitches here.
16:44:42saratogai want to better optimize our IMDCT and figure out why it sucks on the Gigabeat
16:44:44n1ssaratoga: i tested with 2 mp3 files i had on my h300 (both from the same album and both had it, don't have any more mp3's than that album on there
16:45:33 Join jeffdameth [0] (
16:45:35markunsaratoga: lack of IRAM maybe?
16:47:17saratogan1s: is it only the one channel thats screwed up?
16:47:29n1ssaratoga: yes, only the right channel
16:47:59LambdaCalculus37Hmmm... I just got an ATA error -32 error on my Gigabeat F when coming out of Bootloader USB mode.
16:48:36LambdaCalculus37It turns on and boots Rockbox just fine now, though... wonder what that was...
16:49:13saratogais "&*frame->sbsample_prev[ch]; " the same as &(*frame->sbsample_prev)[ch]; ?
16:49:20saratogaif not i bet thats the problem
16:50:09desowinover 19 hours on sansa e280 with r18560!
16:50:23LambdaCalculus37desowin: Sweet! :)
16:50:38saratogadesowin: got a before benchmark? or one against the retail firmware?
16:50:59desowinsaratoga: yes, I benchmarked it twice, I'll upload the new one in a moment
16:51:03saratogais someone around who can compile and test on CF?
16:51:05JdGordonsaratoga: extra brackets are always good :)
16:51:08desowin(on rockbox, didn't on OF)
16:51:25desowinbut before I made tests with ogg, this one is with mp3
16:51:34 Nick delt_zZzZ is now known as delt (i=1000@
16:51:59LambdaCalculus37desowin: How was the test with Vorbis?
16:52:22saratogathe ogg test won't be quite comparable, but is still interesting since it gives some idea how much COP costs verses being fast but single core like Ogg
16:52:36desowinLambdaCalculus37: 18:22
16:53:13LambdaCalculus37desowin: Wow! Still very good runtime!
16:53:29 Quit pixelma2 (Read error: 110 (Connection timed out))
16:53:34 Nick delt is now known as delt_zZzZ (i=1000@
16:53:37*LambdaCalculus37 updates the build on his iPod video and gets it ready to run a benchmark
16:53:48 Join pixelma2 [0] (n=marianne@rockbox/staff/pixelma)
16:57:49desowinbattery benchmark is up in the wiki
17:02:00LambdaCalculus3719 hours, 19 minutes, 9 seconds... lots of 9's there, desowin. ;)
17:02:50 Quit einhirn (Read error: 110 (Connection timed out))
17:05:27 Join mf0102 [0] (
17:05:58amiconnsaratoga: I only benched speed, I didn't actually listen
17:06:34 Join J-23 [0] (
17:08:15n1ssaratoga: i can build and test on h300
17:15:17 Join {phoenix} [0] (
17:20:23 Join massiveH [0] (
17:22:26***Saving seen data "./dancer.seen"
17:24:49joetooI see that w.r.o says that gcc-4.1.0 fails to build Rockbox properly 'on some inline assembly and with the Tremor codec'. Do we know if this continues to be true for all gcc >= 4.1 ? I managed to build for sansa e200 without errors using 4.1.2, (although I have not tested the binary). Is this a platform specific problem due to the inline asm? Or is it that my binary likely won't run (or will brick my sansa) ?
17:26:56joetooHmm. Perhaps that was a dumb question. The webpage seems to have a recent modification date.
17:29:21n1sjoetoo: it can be platform dependant, I have tried (and built working rockbox) with gcc 4.0 4.1 and 4.3 for coldfire for example although they required some small tweaks to the code and compiler options to work
17:33:06 Nick JdGordon is now known as JdGordon|zzz (n=jonno@rockbox/developer/JdGordon)
17:33:14joetooOK, so it may be that using the recommended compiler is only absolutely necessary when testing code before submitting patches.
17:36:27BigBambiJdGordon|zzz: Sorry, I got called away for work
17:36:37BigBambiJdGordon|zzz: For me, yes as I only use the filetree
17:37:20BigBambiI guess though if you are thinking of extending bookmarks into database that people there will need more than the filename as the filename on e.g. ipod is often meaningless
17:37:21 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
17:41:41 Join mmadia [0] (
17:43:32 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
17:44:02n1sjoetoo: different gcc versions may work but it is like with unsupported builds, we haven't tested them and don't support them
17:44:12 Join bughunter2 [0] (n=Jelle@
17:44:23joetoon1s: gotcha
17:47:08 Quit BigBambi (" ajax IRC Client")
17:49:48jhMikeSsaratoga: here's an mpa patch that makes it atomic without the synth_running variable needed:
17:50:19n1ssaratoga: mp3 working nicely now, with r18564 :)
17:50:48 Quit massiveH ("Leaving")
17:56:06 Join jgarvey [0] (
17:57:07 Quit desowin ("KVIrc 4.0.0 Insomnia")
18:00:05jhMikeSsaratoga: actually, I think I can simplify that
18:00:34 Quit MarcGuay (" ajax IRC Client")
18:02:13 Quit saratoga ("CGI:IRC")
18:15:02 Join BlakeJohnson861 [0] (
18:15:02 Quit BlakeJohnson86 (Read error: 104 (Connection reset by peer))
18:15:03jhMikeSsaratoga: (for the logs I guess) :
18:17:00jhMikeSn1s: if you feel so inclined, test that last refinement por favor :)
18:17:39 Join faemir [0] (
18:18:31 Join star_jasmine [0] (n=arwyneve@
18:20:13 Quit m0f0x ("a")
18:21:16 Join culture [0] (
18:22:39bughunter2jhMikeS: i'm just curious but, on line 28, "if(die){", don't you need to release the semaphore in there too?
18:23:04 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
18:24:10jhMikeSline 28?
18:24:23n1sjhMikeS: sounds just great :)
18:24:26bughunter2jhMikeS: yeah, the last paste
18:24:36jhMikeSoh, in the patch
18:25:03bughunter2just curious about why there is no semaphore release there?
18:25:32jhMikeSshouldn't have to because iff die is set to 1, then we signal the synth and wait for the thread
18:26:05bughunter2jhMikeS: yeah, so why isn't the semaphore released? i don't know anything about the piece of code there, but i was curious so i checked out the pastebin link :P
18:26:40jhMikeSbughunter2: you probably should look at context. if die=1 then the sem isn't waited on again.
18:27:01bughunter2jhMikeS: that's true, but i see it as malloc/free more, like, if you alloc, you should free
18:27:04bughunter2isn't that the same with semaphore?
18:27:17bughunter2perhaps a bit of a newbie question, i haven't worked with semaphores really
18:27:21jhMikeSwait/release don't allocate memory, they just change the count
18:27:30bughunter2yeah i know that i think
18:27:38bughunter2about the not allocating part that is..
18:27:43jhMikeSthey get reset later
18:27:51bughunter2i just know a bit of why they're used but not what happens internally in them
18:28:01 Quit kugel (Read error: 104 (Connection reset by peer))
18:28:01jhMikeSn1s: sounds great to test...or you checked it?
18:28:02 Join Seed [0] (
18:28:42jhMikeSbughunter2: kernel.c/thread.c should say more than I can in IRC really if you wan't to dig deep
18:29:01bughunter2jhMikeS: k :)
18:29:32n1sjhMikeS: ah, a bit ambiguous, tested, on h300 sounds ok :)
18:29:51jhMikeSbughunter2: I mean, I did write the code for them but detailing that in IRC just a PITA :)
18:30:22bughunter2i'm checking them out @ svn
18:30:25jhMikeSn1s: it was broken on H300 before?
18:30:29bughunter2online i mean
18:30:54jhMikeSn1s: I thought that the problem was seeking on COP or something...guess there was more to it.
18:31:27n1sjhMikeS: before r18564 it was broken on h300, r18564 fixed that and your patch didn't break it
18:31:34jhMikeSn1s: but, thanks. I am appreciative.
18:32:04n1sthere was dual core breakage fixed in r18563, np
18:34:02 Quit BigBambi (Remote closed the connection)
18:35:24star_jasminehi everyone... I am having a cygwin problem. I took someone 's suggestion yesterday of closing as many programs running, but to no avail. I still got the message too many files open cannot open c:\program files\viavoice tts\tts.log when trying to create a sapi 4 voice under cygwin. I am reinstalling it at the moment.
18:35:30 Join bertrik [0] (
18:35:34star_jasminedoes anyone have any suggestions?
18:41:16 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
18:45:45 Quit MethoS-- (Remote closed the connection)
18:49:00 Join miepchen^schlaf [0] (
18:49:59 Quit Schmogel ("Miranda IM! Smaller, Faster, Easier.")
18:54:30*gevaerts looks at his ipod mini battery_bench results
18:56:29joetoohmm. unzipping onto my sansa gave many messages of "chmod failed". Should I delete .rockbox, remount with different permissions, and retry? Or does it not matter?
18:57:13ameyerthat's to be expected
18:57:26ameyerfat32 doesn't support unix permissions
18:57:33ameyerhence, chmod failed
18:57:39gevaertsI get 15:33 now (r18561), and I got 12:45 two weeks ago (cut-off at the same battery voltage). Something like a 20% improvement
18:57:50gevaerts320kbps mp3, default settings
18:58:05 Join dabujo_ [0] (
19:03:27 Quit star_jasmine ()
19:04:40 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
19:05:26 Quit Nibbler (Read error: 113 (No route to host))
19:06:00 Join lasser [0] (
19:10:23gevaertsI've been wondering. Wouldn't it be useful if the battery_bench output includes the svn revision?
19:11:39jhMikeSgevaerts: Probably, in order to keep info together. I guess it would be a simple addition. Go for it!
19:12:52*gevaerts gets to work
19:14:04 Join FOAD_ [0] (
19:14:46 Nick fxb__ is now known as fxb (
19:14:48 Quit dabujo (Read error: 110 (Connection timed out))
19:15:54bertrikmaybe log the target name too :)
19:16:39gevaertsGood idea
19:16:55 Join Plouj [0] (
19:16:56Ploujhi guys
19:17:01Ploujhave you heard of: ?
19:17:33gevaertsAny reason why the first sentence is split over two lines, even if the following lines are just as long as the first sentence would be?
19:17:47gevaertsPlouj: yes. Have you seen the price?
19:18:19Ploujis it more than a personal mp3 license? :p
19:19:16 Quit linuxstb (Read error: 110 (Connection timed out))
19:19:53*gevaerts finds it interesting that the battery_bench plugin prints usage instructions to its output file
19:21:25bertrikyeah to read them, you have to abort the bench ... :/
19:22:31***Saving seen data "./dancer.seen"
19:23:23 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
19:24:45 Quit {phoenix} ("Konversation terminated!")
19:25:50PloujI guess a gp2x would be cheaper
19:26:15 Join perrikwp|school [0] (i=9821449a@gateway/web/ajax/
19:28:18 Quit perrikwp|school (Client Quit)
19:31:08 Quit FOAD (Read error: 110 (Connection timed out))
19:31:08 Nick FOAD_ is now known as FOAD (
19:32:56 Quit jhulst (Read error: 113 (No route to host))
19:39:02 Quit Darksair ("ERC Version 5.3 (IRC client for Emacs)")
19:45:55 Join Siku [0] (
19:45:56 Quit XstrangeX (Read error: 113 (No route to host))
19:47:43 Part dany_21a
19:51:21 Join bluebrother [0] (n=dom@rockbox/staff/bluebrother)
19:51:30 Join {phoenix} [0] (
20:00:51 Quit ameyer (Remote closed the connection)
20:01:19 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
20:01:54*gevaerts starts to think that getting the svn revision in a plugin needs some work
20:02:52Nico_Pgevaerts: there's svnversion. that's what the splash screen uses
20:02:54 Join FugitivePrisoner [0] (
20:03:06Nico_Pit's not always correct though
20:03:07bluebrothersvn revision in a plugin?
20:03:39gevaertsNico_P: appsversion I think
20:07:33gevaertsCurrently plugin_api seems to consist only of function pointers. Is there a reason not to have data pointers there?
20:08:30jhMikeSgevaerts: there are some data pointer like the frambuffer
20:08:50gevaertsah, yes
20:09:08gevaertsSo no fundamental reason against also putting appsversion in there
20:09:38jhMikeSnone I can see if the core version is what's important and not the plugin version
20:10:14gevaertsI'm working on adding info to the battery_bench output, so we need the core version
20:12:24 Part lasser
20:12:38bertrikversion/credits is a plugin and displays the revision and date
20:14:36gevaertsbertrik: that one just calls show_logo()
20:15:40 Part Plouj ("oh well")
20:17:13 Join herrwaldo [0] (
20:19:08 Join Schmogel [0] (
20:20:09 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
20:21:00jhMikeSsaratoga: (even more for logs) - additionally, no mopping up COP thread between tracks:
20:23:56 Join Xerion [0] (
20:24:43 Join massiveH [0] (
20:25:13 Join dataangel [0] (n=prophet@
20:25:36 Join timostk [0] (
20:25:39dataangelIs there a way for plugins to exit without returning from the plugin_start function, kind of like exit() in standard C?
20:26:25 Part timostk ("Leaving")
20:26:36 Join Yondering [0] (
20:26:45 Join timostk [0] (
20:26:54 Quit timostk (Client Quit)
20:27:01gevaertsCan people have a look at FS #9415 and give their opinion?
20:27:48gevaertsdataangel: not as far as I know. Why?
20:28:11dataangelgevaerts: It's just sort of a pain to propogate return codes all the way back
20:29:07 Join shotofadds [0] (n=rob@rockbox/developer/shotofadds)
20:29:16*gevaerts doesn't talk about setjmp() and longjmp(), which hopefully aren't available for plugins anyway
20:29:36 Quit Schmogel (Read error: 54 (Connection reset by peer))
20:29:39*shotofadds thinks there's nothing wrong with the occasional "goto exit;"
20:29:40 Quit DaCapn ()
20:29:41jhMikeSdataangel: you can call exit_thread but then not much will happen after that :)
20:29:51 Join Schmogel [0] (
20:30:07*gevaerts agrees with shotofadds
20:30:13 Quit Schmogel (Read error: 104 (Connection reset by peer))
20:30:34 Join Schmogel [0] (
20:30:44gevaertsIt doesn't always fit the code structure though
20:30:48*jhMikeS wonders thinks implementing setjmp/longjmp is a nice idea (hehe)
20:30:59 Join GodEater_ [0] (n=ge@rockbox/staff/GodEater)
20:31:44gevaertsjhMikeS: I think we need them. Go for it!
20:32:03*gevaerts hides to make sure
20:32:10jhMikeSserious? ftw
20:32:16*bertrik never used setjmp longjmp
20:32:37 Quit Schmogel (Read error: 104 (Connection reset by peer))
20:32:50gevaertsjhMikeS: I don't see a serious need, no
20:32:53jhMikeSthe thread code for the sim does (things are too scrambled to exit them gracefully without it)
20:32:56 Join Schmogel [0] (
20:33:11 Quit Yondering ("Leaving")
20:33:51gevaertsbertrik: I once built a nice try/catch mechanism in C using setjmp/longjmp. It worked pretty well
20:35:33jhMikeSgevaerts: aha, we need exception handling
20:36:24gevaertsjhMikeS: there's this annoying "Write normal C code. Don't redefine the language." bit in docs/CONTRIBUTING
20:36:51 Join vitja [0] (n=vitja@
20:37:09vitjagevaerts: hi
20:37:11jhMikeSmeh, rules
20:37:49gevaertsvitja: hi. Found the bug?
20:38:25vitjathere was dead loop under interrupt handler )
20:38:50*jhMikeS shyly ask what the bug was
20:39:06vitjaalso I found a problem that SCSI_TEST_UNIT_READY fails
20:39:19gevaertsjhMikeS: tcc usb bulk transfers
20:39:20vitjasomewhy usb_exclusive_ata() returns false
20:39:48vitjathen I manually set it to true, but thinks doesn't go much better
20:40:06 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
20:40:12jhMikeSgevaerts: tcc uses the same controller or just the drivers?
20:40:14vitjabut I see some large packets about 4K going from player
20:40:40gevaertsjhMikeS: this is a new driver
20:40:49vitjagevaerts: can you take a look at updated driver it's cleaned up a little
20:41:18vitjawhat do you think about usb_exclusive_ata()? before doing tests I run ata_init() is that right?
20:41:44jhMikeSgevaerts: but the non-hardware-device code move over well?
20:41:50J-23good night!
20:41:59gevaertsjhMikeS: that's what we're trying to find out
20:42:02 Part J-23
20:42:51gevaertsvitja: ata_init() is fine. My guess is that there's something wrong with the SYS_USB_CONNECTED_ACKs
20:43:19vitjawhat's that?
20:43:24jhMikeSgevaerts: ok, I don't mean to poke too much :) I just had curiosity piqued.
20:44:15Nico_Pgevaerts: IIRC appsversion is set using the svnversion script
20:44:16vitjagevaerts: here it is: vitja/usb-tcc77x.c">
20:45:23gevaertsvitja: usbstorage calls usb_request_exclusive_ata(), which sends a USB_REQUEST_DISK event to usb_queue. The handler for that sends SYS_USB_CONNECTED to all queues, so all threads get a chance to exit if needed. They are expected to send a SYS_USB_CONNECTED_ACK back
20:45:57vitjaokay will take a look tomorow
20:46:01gevaertsWhen all of those have come in, exclusive_ata_access is set to true, and usb_exclusive_ata() returns true
20:46:40gevaertsvitja: any idea if this should work as-is on cowon D2 or Logik DAX?
20:46:41vitjanow I'm going to commit some changes in header that will allow to compile driver
20:47:05vitjagevaerts: some old piece of code worked the same way on D2 so I think yes
20:47:18gevaertsok. That means I can test as well
20:47:19vitjathe one proble in D2 is it doesn't have TMODE it has MODE
20:47:55gevaertsDo they behave the same? i.e. can I just replace TMODE by MODE?
20:48:05 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
20:48:53vitjathink that should be done in one of the header
20:49:21*gevaerts looks at shotofadds
20:51:02vitjagevaerts: also it relays on charger_inserted() so it should be implemented for target
20:51:40gevaertsOK. If that's not there I can cheat a bit
20:52:55vitjasure. I've commited necessary changes to compile the code
20:53:30vitjaalso I defined USB_BASE for tcc780x
20:54:03 Quit moos ("Rockbox rules the DAP world")
20:54:40gevaertslinuxstb: what's the current DAX state? Since it's also tcc77x (I think), I'd like to test with that
20:55:36linuxstbgevaerts: The flash drive is extremely unreliable. Most of the time Rockbox doesn't mount the disk. I have some local changes I need to commit though (I don't think SVN builds at the moment)
20:55:46vitjagevaerts: all you need is lcd driver working
20:56:08linuxstbWhich it does on the DAX (and m200 - same driver)
20:56:42 Join petur [50] (n=petur@rockbox/developer/petur)
20:57:23 Quit gevaerts (Nick collision from services.)
20:57:33 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
20:58:49 Join p3tur [50] (n=petur@rockbox/developer/petur)
20:59:10 Quit petur (Nick collision from services.)
20:59:15 Nick p3tur is now known as petur (n=petur@rockbox/developer/petur)
21:01:23 Join p3tur [50] (n=petur@rockbox/developer/petur)
21:01:56*shotofadds wonders if gevaerts is still looking
21:02:10 Quit petur (Nick collision from services.)
21:02:11shotofaddsbtw. d2 has charger_inserted()
21:02:15 Nick p3tur is now known as petur (n=petur@rockbox/developer/petur)
21:02:44gevaertsshotofadds: any ideas about this TMODE/MODE thing?
21:02:51*gevaerts doesn't actually know what they are
21:03:38 Join tvelocity [0] (n=tony@
21:03:44 Quit jhulst (No route to host)
21:05:20 Quit Nico_P (Remote closed the connection)
21:06:40 Join funman [0] (
21:09:09shotofaddsgevaerts: looks like they're the same register. Maybe we should tidy up one of the tcc77x.h or tcc780x.h to make them consistent?
21:09:17gevaertsshotofadds: I think so, yes
21:09:30*shotofadds wonders what local changes linuxstb has ... anything juicy?
21:09:52gevaertsvitja: bootloader/telechips.c redefines usb_acknowledge() and usb_wait_for_disconnect(). That's the reason you don't get exclusive ata
21:10:20linuxstbshotofadds: No, just obvious things to make it build
21:10:38*linuxstb should commit before other people break his patch
21:12:07 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
21:14:12shotofaddsgevaerts: you'll also need USBD_IRQ_MASK, etc. do you want me to update tcc780x.h?
21:14:31gevaertsshotofadds: please do. I guess you can do that faster than I can
21:14:33shotofaddsand remember we don't use IRQs in the bootloader on 78x
21:14:56shotofaddsnothing like a bit of consistency ;-)
21:15:19funmanyou don't use IRQs for sd/nand data transfer ?
21:15:20vitjagevaerts: no that's doesn't links so I make the #ifndef HAVE_USBSTACK
21:15:31*linuxstb commits his pending changes
21:18:14LambdaCalculus37linuxstb: Excellent, I can start having some more fun with my m230 now.
21:18:31 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
21:19:06vitjashotofadds: I'm going to commit USBD_IRQ_MASK and DEV_USBD, ok?
21:21:13shotofaddsvitja: yeah sure. I was just about to do that (and also rename MODE to TMODE ..)
21:22:33***Saving seen data "./dancer.seen"
21:22:49vitjashotofadds: take a look at the patch first: vitja/tcc780x.diff">
21:23:02amiconnlinuxstb: How much ram does the logikdax have again?
21:23:29 Quit funman (Read error: 104 (Connection reset by peer))
21:23:30amiconngevaerts: Btw, I have no problems at all using rockbox usb on the beast via my "problematic" hub
21:23:49amiconnSo the stack is definitely ok, and the problem is some PP register setup
21:24:03shotofaddsvitja: that looks good to me. don't forget to fix system-tcc780x.c!
21:24:10gevaertsamiconn: I think I tried it at devcon, but I'm not sure now. Definitely PP indeed
21:24:54LambdaCalculus37So the new commits for m200 and Logik DAX should allow a normal build to compile without error, correct?
21:25:01gevaertsamiconn: have you tried a PP5020 device with it? I can't remember
21:25:43amiconnYes. Same problem, a bit less likely though
21:26:14amiconnThat might just have been due to the different (better?) cable of the H10 (vs. ipod cable of my mini)
21:26:29amiconnI should probably try it on the Photo as well (same cable as the mini)
21:27:15linuxstbamiconn: 2MB - built into the SoC
21:27:20 Quit Nico_P (Remote closed the connection)
21:27:26shotofaddsLambdaCalculus37: looks like it, but don't expect the bootloader to be able to read the rockbox binary with the current NAND driver
21:27:35vitjashotofadds: done
21:28:14 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
21:28:15amiconnlinuxstb: OK, so it's not too tight. How much codec ram and plugin ram do those targets have atm?
21:29:22*linuxstb notices he left some experimental sizes in the config files he committed :(
21:31:10amiconnUmm, those are *huge* buffer
21:31:22amiconnDo those even work?
21:31:26LambdaCalculus37shotofadds: It doesn't work on the m240 you have, does it?
21:31:35*LambdaCalculus37 wonders if it'll work on his m230
21:31:56amiconnThey're even larger than on largemem targets...
21:32:30gevaertsvitja: is usb expected to work if I use tcctool, or do I need to install for real?
21:33:04shotofaddsLambdaCalculus37: I wouldn't expect it to work. Reading works if every sector was written to the device sequentially.. but that's rather unlikely in the real world!
21:33:10vitjagevaerts: should work from bootloader call usb_test() after ata_init() in bootloader/telechips.c
21:33:38*shotofadds looks up at 20:14
21:34:05LambdaCalculus37shotofadds: Okay. I'll try it, but I won't expect anything.
21:34:19*linuxstb revisits that part of his commit
21:34:43amiconngevaerts: There are several data pointers in the plugin api, for a looong time...
21:34:56amiconnframebuffer, global_setings, ...
21:35:08gevaertsamiconn: I noticed since then. They're a bit hidden between the function pointers though
21:36:09*shotofadds wonders if there are any Swedes around
21:36:46gevaertsshotofadds: how do you tell the D2 to boot from USB again?
21:36:51shotofaddsthere's still a link to the feature requests tracker on the pages linked from the build table
21:37:13shotofaddsgevaerts: hold one of -/M/+ while inserting the cable
21:37:15amiconnjhMikeS: Do you know whether the i2s master/slave roles are dictated by the hardware, or are just a matter of setup?
21:37:18linuxstbamiconn: Look at the m200 config file, rather than the DAX - that's what I meant to commit....
21:38:00amiconnRight now the audiocodec is master in the ipods, while it is slave in the H10. I wonder whether this can be unified.
21:38:21 Quit pixelma2 ("-")
21:38:31 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
21:38:32amiconnlinuxstb: Why 64KB plugin buffer?
21:38:34vitjashotofadds: there is also usb interrupt handler missed
21:39:01linuxstbamiconn: It's just arbitrary for now - I chose the same values as the iFP as a starting point.
21:39:29vitjashodanX: that should be called USB_DEVICE()?
21:40:01vitjashodanX: sorry
21:40:05amiconnAh, hmm. My first idea was that those targets shoul duse 32KB plugin buffers, same as archos
21:40:26amiconnBut then I'm not 100% sure anymore, because of arm code size vs. SH1 code size
21:40:39amiconnPerhaps 48KB would be a good size
21:40:42shotofaddsvitja: you mean USB_DEVICE vs. USBD_IRQ? I really don't mind which we use .. just so long as they're the same!
21:40:44linuxstbamiconn: Plugin building is disabled for now anyway, so when someone gets them building, we can see how much RAM is needed.
21:41:08gevaertsamiconn: another difference is that these are flash targets, so they can probably get by using less buffer
21:41:28shotofaddsvitja: none of the existing 78x handler names have a _IRQ suffix, so I'm reluctant to change that.
21:41:38amiconnCodecs are different, of course. They need a bit more unless you want to do without many of the codecs...
21:41:45vitjashotofadds: should be nice
21:42:09amiconngevaerts: Even on the flash targets there are a few reasons to keep the main buffer large
21:42:35vitjagevaerts: you may also want this patch vitja/bootloader-allow-printf-wrap.diff">
21:42:39amiconnEverything that needs a whole file at once will profit. Voice, plugins like the jpeg viewer...
21:43:15jhMikeSamiconn: it depends on the bits you set if you mean the imx31
21:43:34amiconnjhMikeS: I mentioned ipods and H10...
21:43:39 Quit shotofadds (Read error: 104 (Connection reset by peer))
21:43:52linuxstbamiconn: How big is a typical swcodec voice file?
21:43:57 Join shotofadds [0] (n=rob@rockbox/developer/shotofadds)
21:44:03gevaertsvitja: I'm not far enough for that to be useful yet. Since the D2 bootloader doesn't do interrupts, I'm working on DAX, but that doesn't have charger detection yet
21:44:05jhMikeSamiconn: not in the sentence addressed to me, but I don't know about PP
21:44:25linuxstbgevaerts: There is no charger...
21:44:30vitjagevaerts: you may say it's allways inserted
21:44:37amiconnjhMikeS: No, but the one following it (explaining why I asked)
21:44:37shotofaddsgevaerts: do you really need to use the bootloader?
21:44:43gevaertslinuxstb: good point :)
21:44:49jhMikeSamiconn: nvm...I read "i2c". it is software, yes
21:45:01gevaertsshotofadds: actually, no. the main build should also work
21:45:26amiconnvitja: If there is no charger, it should be handled like a target without a charger, not faked
21:45:38jhMikeSamiconn: Anything AS3514 runs i2s as master, the Wolsons run as slave iirc
21:45:45shotofaddswhy do we need charger detection for USB anyway?
21:45:49gevaertsusb detect and charger detect need to be separated
21:46:05amiconnjhMikeS: Nope. In the ipods, the wolfsons run as master. That's why I was asking ... ...
21:46:08vitjaamiconn: I mean make usb detect return USB_INSERTED
21:46:32vitjashotofadds: because it usualy implements the same code
21:46:39jhMikeSamiconn: I was talking on from the SoC PoV
21:46:55shotofaddsvitja: not on the d2, it has both a wall-charger and USB-charger...
21:47:53vitjashotofadds: ok but usb_detect is kind of get_vbus_status()
21:47:58jhMikeSamiconn: The SoC interface setup is master for AS3514 and slave for Wolfsons
21:48:16amiconnjhMikeS: Then it's not so on the H10
21:48:27vitjabecause usb cabel is allways charger
21:48:46amiconnvitja: USB power != charger on many targets
21:49:04jhMikeSamiconn: The i2s on the Soc for H10 was setup as slave the last time I worked on it
21:49:24vitjaamiconn: ok, so usb_detect() should be implemented for each target
21:49:45 Join MethoS- [0] (
21:50:00linuxstbAh yes, the DAX can be powered by USB (but not charge)
21:50:03amiconnusb_detect() is not for detecting usb power, but for detecting an actual usb connection
21:50:04shotofaddsvitja: there is a stub for usb_detect in usb-cowond2.c. Currently it always returns true...
21:50:18shotofaddsI'm not sure why
21:52:23 Join Hillshum [0] (
21:54:23 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
21:58:01*gevaerts decides to commit FS #9415. If anyone complains it's easy to revert
21:58:33*shotofadds thinks the logic in wmcodec-telechips.c is kinda funny :)
21:59:19 Quit LambdaCalculus37 (" ajax IRC Client")
22:02:21vitjashotofadds: really so why not reduce that all to 0x34
22:06:18shotofaddsvitja: maybe that should be changed to "#if defined(HAVE_WM8985) || defined(HAVE_WM8731)"...?
22:07:02vitjaand WM8751
22:07:43shotofaddswhich player uses WM8751?
22:07:58 Quit {phoenix} (Remote closed the connection)
22:08:16vitjain last commit I see it in audiohw_init()
22:08:18linuxstbGigabeat F, m:robes and meizus
22:08:42 Join Yondering [0] (
22:09:02shotofaddslinuxstb: so probably shouldn't be in wmcodec-telechips.c then...
22:09:10vitjaI think that list of all targets is better
22:10:10gevaertsvitja: did you do something about UNCACHED_ADDR in usb_storage.c?
22:10:31shotofaddsvitja: well, that's what we have right now (in a strange fashion). so maybe just leave it ;)
22:10:41*shotofadds goes for a drink
22:10:46linuxstbshotofadds: The D2 has an external WM codec chip?
22:11:02vitjagevaerts: use I added defined(CPU_TCC77X) || defined(CPU_TCC780X) for the first case
22:11:33vitjashotofadds: yeah things could change and addresses too
22:11:50 Quit mmadia (" , #haiku")
22:13:00vitjagevaerts: and and them too usb.c to get USB_FULL_INIT defined
22:16:04vitjaI'm going to sleep now, so good luck
22:16:40*gevaerts needs the USB_DEVICE irq handler. Off to the iaudio7 files for reference...
22:17:01vitjaok I'll give you full patch for iaudio7
22:17:44vitjabut tcc78x handles interrupts in another way, you can simply rename USBD_IRQ to USB_DEVICE
22:18:04gevaertsah, ok
22:20:17 Quit bluebrother (Nick collision from services.)
22:20:20 Join bluebroth3r [0] (n=dom@rockbox/staff/bluebrother)
22:20:50 Nick bluebroth3r is now known as bluebrother (n=dom@rockbox/staff/bluebrother)
22:23:02Hillshumso what do i do to patch the sansa v2 OF?
22:23:04Hillshumso what do i do to patch the sansa v2 OF?
22:30:24 Join JiriTechet [0] (n=59b01f46@gateway/web/cgi-irc/
22:35:37 Quit JiriTechet ("CGI:IRC (EOF)")
22:35:47amiconnjhMikeS, saratoga (logs): Shouldn't the check in mpa.c be for (NUM_CORES > 1) instead of CPU_PP?
22:36:21amiconnThis way it would (1) be possible to compile a single core PP build for testing purposes and (2) it would already be prepared for other dual core targets
22:36:53 Join lasser [0] (
22:38:31 Join JiriTechet [0] (n=59b01f46@gateway/web/cgi-irc/
22:40:44amiconngevaerts: Do you have an idea how difficult it would be to write masstorage support for firewire?
22:41:33gevaertsamiconn: not really, no. The actual SCSI part should be the same as for USB, but I don't know anything about firewire
22:41:35 Quit JiriTechet (Client Quit)
22:41:48 Join beta2k [0] (
22:42:48 Quit HBK (Read error: 104 (Connection reset by peer))
22:42:50 Join HBK- [0] (
22:43:37 Join webguest43 [0] (n=59b01f46@gateway/web/cgi-irc/
22:44:40 Join saratoga [0] (n=9803c264@gateway/web/cgi-irc/
22:44:59saratogajhMikeS: I can't test right now, but I think that patch you posted would probably deadlock from time to time seeking
22:45:22saratogathough maybe I misunderstand
22:45:28 Quit webguest43 (Client Quit)
22:45:59 Join webguest44 [0] (n=59b01f46@gateway/web/cgi-irc/
22:46:09saratogaactually maybe not since I see you've moved around the semaphores
22:46:36saratogaregarding the low mem targets, perhaps now is a good time to explore Nico_P's low mem buffering code?
22:47:12amiconnI don't think that's necessary
22:47:16saratogai'd be very interested to see a port start using that, so we can work on supporting very low mem targets better
22:47:43Nico_Psaratoga: the patch is already usable
22:47:52 Join reacocard [0] (n=reacocar@WL-156.CINE.HMC.Edu)
22:48:17Nico_Pthough it hits on runtime quite heavily
22:48:37saratogaNico_P: why the hit?
22:49:04saratogai don't understand the codec selections in the latest patch either, IIRC Vorbis uses more memory then WMA
22:49:09Nico_PI'm not certain, but my guess is that hitting flash more frequently eats more power
22:49:45amiconnIt most probably does. The flash won't go to sleep immediately after being accessed
22:50:01 Join fml [0] (n=4fd3ed80@gateway/web/cgi-irc/
22:51:06amiconnsaratoga: Do you think a similar split could be done for the ac3 codec?
22:51:27 Quit fml (Client Quit)
22:51:41amiconnThat one isn't even realtime on PP yet... (well, when testing with the common 48kHz ac3 files, i.e. resampler also running)
22:51:53saratogaamiconn: I've never looked at our AC3 codec, but IIRC its an IMDCT codec very much like WMA
22:52:15saratogaso you could probably split the IMDCT + windowing code
22:52:46saratogaits probably against the standard, but it might be possible to hack it into using the IMDCT lib
22:52:55saratogathat'd probably gain you a lot of speed if you need it
22:53:55shotofaddslinuxstb: yes, the wm8985 is external to the SoC - as seen here:
22:54:04 Quit HellDragon (Read error: 104 (Connection reset by peer))
22:54:25 Quit Yondering (Remote closed the connection)
22:54:49 Join ameyer [0] (
22:55:10 Join MethoS-- [0] (
22:55:13saratogahmm the AC3 IMDCT does look very much like the other transform codecs, so maybe we could use the imdct lib
22:55:18 Join funman [0] (
22:55:40saratogai think i remember reading that the spec actually specified the exact fixed point IMDCT to be used, but we can probably get away with using our own
22:56:04bertriksaratoga, with mp3 on COP I got 16 hours runtime on my sansa c240 (up from 14:41 h)
22:56:06 Quit beta2k_ (Read error: 110 (Connection timed out))
22:56:37 Quit mcuelenaere ()
22:57:18*gevaerts can beat that with his ipod mini
22:57:21webguest44Hi everyone, I'd like to upload a theme to the Rockbox wiki and I need write permissions to do this. Could anyone give me these permissions please? The theme is basically an update of my previous Plain Text theme for e200, but uses viewports to display different parts in different color and WORKS with current builds.
22:58:02domonokywebguest44: tell us your wikiname, and promise not to spam the wiki :-)
22:58:50 Join HellDragon [0] (
22:59:38saratogaamiconn: the ac3 codec's windowing code could probably be replaecd with the ASMed one used by WMA/AAC as well
23:00:04webguest44Aha, my nickname is incorrect (I used the web-based IRC client). So my nickname is JiriTechet and I only hope my theme won't be regarded as a spam :-)
23:00:17saratogabertik: any chance you tested the OF?
23:00:26saratogai'm curious if we beat any retail PP firmwares yet
23:00:37domonokywebguest44: done
23:00:55amiconnsaratoga: On ipod 1st/2nd Gen we already did
23:01:04webguest44Wow, that was fast, thanks!
23:01:12saratogaamiconn: thats right I'd forgotten
23:01:13gevaertssaratoga: my ipod mini 2g with cf mod went from about 12:45 to 15:30
23:01:25gevaerts(320kbit mp3)
23:01:27amiconn(against runtimes stated by manufacturer, not measured)
23:01:57bertriksaratoga, no, I haven't tried the OF yet
23:02:20saratogai should test mine, but i just don't want to deal with unmodding the bootloader and such
23:02:37saratogait was annoying enough to get the bootloader off and the database refresh gone the first time
23:02:38amiconnWhy would you need to do that?
23:02:54saratogapresumably not having the Sandisk bootloader could change things
23:02:58desowin_saratoga: I think we already did for sansa e200, I might do test with OF now, butthe problem is there's no battery_bench plugin in OF ;)
23:03:10saratogasince the hardware init is different
23:03:26saratogadesowin_: play music and record on the line in at 8khz and 8 bit
23:03:32amiconndesowin_: Hook it to your pc soundcard, and start recording. 8kHz mono is sufficient
23:04:17saratogaactually i need to test if removing the OF bootloader changes things, that could give us some clues about hardware initialization on the sansa
23:04:20bertrikI can't find a claim about runtime for the c200's on the sandisk website
23:04:43 Part domonoky
23:04:45*amiconn thinks this comparison would probably be easiest on H10
23:04:46saratogabertrik: you can probably get a good idea just by scaling the 20 hours for the e200 by the change in battery size
23:05:20 Quit webguest44 ("CGI:IRC (EOF)")
23:05:32amiconnNo database refresh, and UMS for easy swap between rockbox bootloader and OF
23:05:47bertrikthat would be 20h * 530 mAh / 750 mAh = 14:08
23:06:09amiconnAnd H10 (small) runtime isn't great in general, so no excessive waiting ;)
23:07:34bertrikBTW, the AS3514 inside the PP5024 allows undervolting from 1.2 down to 1.05V, I'm sure runtime can be extended even more that way
23:07:58saratogafor what its worth, the ABI review says Sandisk says 15 hours and the reviewer claimed 13 hours (c200)
23:08:03funmanPP5024 is the Sansa c200 SoC ?
23:08:14Bagderfunman: no, the e200
23:08:39saratogabertrik: didn't we already play around with the AS voltages?
23:08:43*bertrik was confused
23:09:19 Quit Bagder (Read error: 104 (Connection reset by peer))
23:09:32bertriksaratoga, yes I think so, but they're still at their defaults as far as I know
23:10:09 Quit petur (Read error: 104 (Connection reset by peer))
23:10:13 Join Bagder [241] (n=daniel@rockbox/developer/bagder)
23:10:59 Join petur [50] (n=petur@rockbox/developer/petur)
23:11:22 Quit MethoS- (No route to host)
23:11:52 Quit petur (Read error: 104 (Connection reset by peer))
23:12:49 Join petur [50] (n=petur@rockbox/developer/petur)
23:13:28 Quit massiveH (Read error: 104 (Connection reset by peer))
23:14:17 Join dabujo [0] (
23:14:54 Join massiveH [0] (
23:16:38saratogabertrik: when you get a chance, post a patch and the next time I do power measurements I'll explore the impact of the AS voltages
23:17:45saratogalinuxstb: does the DAX really only have 128kb allocated to codecs? or is that 128k and some other buffer?
23:18:06 Join Nibbler [0] (
23:19:03linuxstbsaratoga: I'm sure we'll increase it, I just chose a relatively small starting point. There's still that nasty 512KB malloc buffer as well.
23:19:06 Quit petur (Read error: 104 (Connection reset by peer))
23:19:21 Join matsl [0] (
23:20:55saratogalinuxstb: I'd kill the malloc buffer and leave the codec buffer alone
23:21:10saratogaor kill malloc and shrink the codec buffer
23:21:19 Join BlakeJohnson86 [0] (
23:21:23saratogamight "encourage" someone to gut the malloc stuff from vorbis
23:21:36amiconnsaratoga: Back to 444% on cf with your fix. So no speed change at all - good
23:22:05saratogaand then eventually we could reclaim some or all of that 512KB on the normal disk targets
23:22:08 Join petur [50] (n=petur@rockbox/developer/petur)
23:22:34***Saving seen data "./dancer.seen"
23:22:59 Quit petur (Client Quit)
23:23:33 Nick synergist is now known as christopher (
23:23:57 Nick christopher is now known as synergist (
23:24:48 Quit BlakeJohnson861 (Read error: 110 (Connection timed out))
23:25:34 Part lasser
23:26:24 Quit shotofadds (Read error: 110 (Connection timed out))
23:26:36 Quit ompaul (Client Quit)
23:28:00 Join Yondering [0] (
23:28:00 Quit reacocard (".")
23:28:35 Quit Nibbler (Read error: 104 (Connection reset by peer))
23:28:54 Quit jgarvey ("Leaving")
23:29:21 Join Llorean [0] (
23:30:27LloreanBagder: Regarding the upcoming release, I'm so out of touch that I have no idea what's been happening but I wanted to drop a reminder regarding default settings that may need changed (larger limits for files in folder, playlist length, and default enabling of dircache)
23:30:34Nico_Psaratoga: in the commit message for MP3 synth filter on COP, you mention a diff between test_codec and real world perf increase. why is there one?
23:31:11 Quit dabujo_ (Read error: 110 (Connection timed out))
23:31:20LloreanNico_P: I thought even when it was first introduced it wasn't expected to mirror real world performance
23:31:44saratogaNico_P: test_codec doesn't use the buffering, DSP or playback code, so the split between the processors isn't realistic
23:32:16saratogain reality the CPU is still much more worked then the COP, so the near perfect scaling with cores in test codec is unrealistic
23:32:24Nico_PI see
23:32:47Nico_PI wonder if we could switch e.g. buffering to the COP
23:33:09saratogaNico_P: I'd very much like to make the buffering code thread safe, but its beyond my abilities i think
23:33:27saratogaparticularly the insert pcm function
23:33:56saratogai guess thats playback more then buffering
23:34:02Nico_Pyes, it is
23:34:29 Join MarcGuay [0] (
23:34:36saratogasince buffering has to be synchronized with decoding (so the input buffer doesn't get overwritten while its in use), putting buffering on the COP would not be my first choice
23:34:37amiconnLlorean: Imho the files per folder limit is already too high
23:34:45 Quit funman ("leaving")
23:34:49saratogahowever, if its easy to do, I would certainly like to see it done
23:34:51*amiconn lowers it immediately on any new target
23:35:02Lloreanamiconn: For the average user, though, it's not.
23:35:05saratogathen maybe we could put all of the codecs on COP
23:35:17LloreanIt shouldn't be raised for archos, but many people seem to lump all their files in one folder, or have many, many, many artist folders
23:35:19saratogaand leave playback on GUI on the CPU
23:35:31LloreanBut on swcodec we have the RAM to spare to set it somewhat higher, in most cases.
23:35:47LloreanA very large portion of us agreed this would make sense for the release branch, at least.
23:35:55LloreanWhen it was discussed some time ago
23:35:56amiconnIt is already a lot higher than is sane to have actual files per folder
23:36:02saratogaputting it higher on the release would be nice, mostly just so that people don't complain about problems with it
23:36:02LloreanPeople aren't sane, though.
23:36:20saratogamost sensible people will be using SVN builds anyway I think
23:36:23amiconnIt's set to 1000 by default. I usually lower it to 300
23:37:56saratogaNico_P: basically, data flows from the buffering code, to the codec, then to playback/DSP, so ideally we'd pick one point along that chain and move things to the COP
23:38:29saratogaright now mp3 moves to the COP and then back to the CPU so theres two transitions which is a bit ackward (but not a serious limitation for MP3 at least)
23:38:43Lloreanamiconn: Maybe just for the 80gig iPod then, but we definitely have users who from time to time report bugs about the error you get when it's set too low.
23:39:04LloreanWe can't expect users to be sane, so we need to set insane defaults when we can spare some RAM for it, and document that "lowering it could increase battery life"
23:39:11gevaertsWhat error message do you get?
23:39:28amiconnsaratoga: Didn't you say that mp3 performance on gigabeat was rather bad? Was that F/X or S?
23:39:47*amiconn just checked on the beast, and got 800% realtime with 192kbps mp3
23:39:48Lloreangevaerts: Something like "Directory buffer full" maybe? I can't remember exactly
23:40:01amiconnThat's ~33MHz for realtime
23:40:08saratogaamiconn: I looked at the F
23:40:14amiconn(beast running at half speed in svn)
23:40:18*gevaerts thinks that rephrasing the error could help, at the risk of angering our beloved translation maintainer
23:40:26saratogai don't remember what MP3 was, but WMA/AAC/Ogg were quite poor
23:40:32saratoganeeding nearly 2x the clock speed as PP
23:41:02amiconnWell, the beast needs less for mp3. No WMA or AAC to test with. OGG I can check
23:41:13saratogai'd expect a huge gain on the beast
23:41:19saratogait has the fast multiplier
23:41:25saratogaand much bigger caches
23:41:50rashergevaerts: As long as it's not done in the 3.0 branch, I'm down with it
23:42:33saratogaa fixed point mul on ARM9 is 7 cycles (of which I think only 3 are pipelined) while its just 3 on the beast (all pipelined)
23:42:43Lloreangevaerts: Possibly for future releases
23:43:23gevaertsrasher, Llorean: it would be nice for 3.0, but 24 hours is not really enough for translations to get updated I guess
23:43:27LloreanBut I feel that in the release version, we should have defaults lined up with the shotgun approach: Solve potential problems, with the option to disable things to increase battery life (rather than having things optimized to increase battery life but require referencing the manual just to be able to access your music period if you're a non-expected setup)
23:43:48gevaertsLlorean: I think I agree
23:44:05*Nico_P agrees too
23:44:28LloreanAnd since this may be the last hour or two I have on the 'net before the release, I figured I'd try to drop by and re-mention this.
23:44:58saratogaARM9 is a bit weird since its still all terrible like ARM7, except with slightly better load/stores and much higher clock speeds (which actually eat up the load store performance if you don't have IRAM)
23:44:58Nico_Pdoes the manual mention the release (as per )?
23:44:59 Join Nibbler [0] (
23:45:38gevaertsLlorean: any chance of sending a mail with all those issues, so we can remember them?
23:45:41saratogai expect once we get slower ARM9 cores like the AMS chips, it'll become necessary to pay attention to ARM9 performance
23:45:46 Quit matsl (Read error: 60 (Operation timed out))
23:46:31 Join mmadia [0] (
23:46:51Bagdermy approach to the release is the harsher: I want to release whatever we have when the time is up, as I have very little actual rockbox fiddling time I only herd the release action not so much the contents of it
23:47:07amiconnsaratoga: Results for the same ogg track (~175kbps): ipoc color: 205% realtime, 39.1MHz. Beast: 760% realtime, 34.7MHz
23:47:27Lloreangevaerts: Sent
23:47:27*amiconn curses the beast's button map
23:47:46LloreanBagder: Changing the defaults should at least be trivial, right?
23:48:16amiconnThe release will only have beta quality anyway....
23:48:52LloreanStill, it will be the highest quality release for SWCODEC
23:48:56LloreanAnd we'll move up from there.
23:49:02LloreanThe release ball needs to start rolling.
23:50:02*gevaerts doesn't think that "we release at 23:23" and "we still want to do some things" are incompatible
23:50:03 Quit bmbl ("Woah!")
23:50:22 Part toffe82
23:54:25amiconnWhat I'm missing most in the wps is a _stop_ button mapping
23:54:59amiconnThe mappins which exist are 100% logical, but there are lots of dupes, and otoh missing functions
23:55:10Lloreanamiconn: The power button doesn't stop?
23:55:17LloreanTop button on the right side?
23:55:28*Llorean hasn't seen his S in a week, but thought it was the same as the F.
23:55:37amiconnShort press also brings up the menu, and long press shuts down
23:55:54Nico_Pamiconn: short press on POWER should stop
23:55:56MarcGuayI thought it stopped - testing.
23:56:27MarcGuayWorks for me.
23:57:01 Quit n1s ()
23:57:02amiconnI'm running plain svn as far as button mappings are concerned
23:57:20Nico_Pworks for me too... plain svn also AFAIK
23:57:20saratogaamiconn: those ogg results look a bit slow
23:57:37saratogai think my sansa was doing 192k at ~250%
23:58:49 Join homielowe_ [0] (

Previous day | Next day