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 2007-09-08

00:02:43Zagorany higher and you have to jump through a lot more hoops to get validated
00:03:20Zagornot that we validate today anyway
00:03:26 Quit jgarvey ("Leaving")
00:04:20amiconnStill weird... I'd either use html 4.0 transitional and do most of the formatting old-style for baaaaackwards compatibility, or use xhtml strict + css
00:04:49amiconnTo avoid switching the browsers to quirksmode
00:04:58Bagderaren't browsers mostly disconnecting the HTML version from the CSS anyway?
00:05:42Bagderbrowsers that can do 4.0 transitional aren't that many, and afaik the deal with css at least fairly well
00:05:55amiconnMost browsers, switch to mode where they try to do "best of both worlds" (with greatly varying success) if they see a transitional type
00:05:59Lloreanxhtml strict is (imho) a bit silly, since they're already rethinking some of the things they mandated about it, I thought. I tend to lean toward transitional.
00:06:16BagderI always just use transitional
00:06:25Bagderbeyond that gives more pain than gain
00:06:37amiconnI also did most of the time, but then used next to no css
00:07:08amiconnNot that I did much html lately...
00:07:12ZagorI doubt the two are as connected as you think
00:07:28 Join darkless [0] (
00:07:40 Quit freqmod (Read error: 104 (Connection reset by peer))
00:07:40BagderI've never even heard about that being a problem before
00:08:46 Join freqmod [0] (
00:09:20pixelmaZagor: call me old-fashioned but I can imagine it would be better to have the table widths at a fixed percentage and at worst have blank space inside the table but "jumping" blank space outside is worse especially when going that "boxed" look
00:09:55pixelmawithout trying out though
00:10:13Zagorpixelma: jumping? it simply doesn't expand
00:10:36Zagorsince there is no wider content that needs room
00:12:18pixelmathey'll look wider or more narrow depending on the longest line inside them - and so e.g. the centered table at the top will at one time really look good and at another time it wouldn't, depending on how wide the tables below are
00:12:56pixelmaand all depending on the size of your browser window
00:13:25pixelmathat was a lot dependencies
00:13:51 Join akaias [0] (
00:14:03ZagorI still don't understand what you're saying. the top box does not depend on the others
00:14:28Zagormaybe I'm unusally ... tired :-)
00:14:54pixelmano, it doesn't that's why - maybe I should do a drawin, can't explain that in words atm
00:15:20Zagorthe top table will always be this wide, no matter the others. the smaller boxes below will vary slightly depening on their content. that is a GOOD thing in my opinion.
00:16:01pixelmaIMO that's not good
00:16:14Zagorso they should rather wrap?
00:16:24 Quit Genre9mp3 ()
00:16:27Zagorand leave whitespace on the right?
00:17:38pixelmayou mean whitespace in the browser window or in the cell?
00:18:23pixelmaif you mean the former - they do ATM fo me in my browser window
00:18:36 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
00:18:38ZagorI don't know. what are you saying? :-) I'm saying if the content fits the screen, don't wrap. it the screen is wider than the content, don't pad.
00:19:47pixelmaI'm arguing the "fixed" width they have which is defined by the longest line
00:19:47 Quit Soap (Read error: 104 (Connection reset by peer))
00:19:56 Join Soap_ [0] (n=Soap@rockbox/staff/soap)
00:20:26Zagorpixelma: you mean that the subversion box is stretched because of the news box?
00:21:46pixelmano, that's not the point (which I seem to be unable to make for you at this point of time)
00:22:10Zagoramiconn: so quirks mode is entered when running transitional. no matter if I use css or not.
00:22:52amiconnBut when using css, it would be better to not trigger quirks mode
00:22:57Zagorexactly. so why not use css in html4
00:23:20Zagoryeah, for some borderline box doo-da. but xml is more bother than gain.
00:23:30Zagorxhtml I mean
00:24:15amiconnYes, html 4.0 strict would do as well
00:24:16 Quit spiorf (Remote closed the connection)
00:24:32Zagorunless you run msie...
00:24:33amiconn(but for some reason I do like some of the constructs in xhtml better)
00:24:54pixelmaZagor: trying to explain differently - why not make the tables (where 2 are side by side) like 50% of the screen width each (or 49%) or...
00:25:06 Nick Soap_ is now known as Soap (n=Soap@rockbox/staff/soap)
00:25:27Zagorpixelma: because then they become super-wide and full of space on a fullscreen browser. both ugly and hard to read.
00:26:53pixelmabut if you want to go the "boxed", it would adapt better to different browser window widths
00:27:23Zagoramiconn: I didn't know about the font style inheritence issue though. that could convince me to abandon transitional...
00:27:41pixelmaand looks differently every time content changes
00:27:51Zagorpixelma: stretching is not a good way to "adapt" in my opinion
00:28:00Zagorstretching without content, I mean
00:28:01 Quit desowin ("use linux")
00:28:31 Quit petur ("Zzzzz")
00:29:20pixelmadid you look at the current proposal in a bigger window? IMO everything falls apart right now, you don't feel a structure
00:29:21Zagoryes, the page has slightly different content at different times. different content == slightly different look. plain text has the same "problem" :-)
00:30:03Zagorpixelma: sure, it's ugly. don't browse in full screen :-)
00:30:21[aB]DaRkNeSSonce i dl a rockbox theme
00:30:22Zagorseriously, stretching is not the answer. that will just make it ugly in a different way
00:30:24[aB]DaRkNeSShow do i install it
00:30:46amiconnHmm, I always forget: was it MAX_INT or INT_MAX?
00:31:00pixelmaZagor: and my screen is quite average
00:31:12 Quit davina (Remote closed the connection)
00:31:14Zagorpixelma: look at the bottom mail table on the old/current page. do you think that is pretty? I don't.
00:31:37Zagorand above all, it's difficult to read
00:31:45pixelmaI didn't want to say that it would look good but IMO better than the other
00:32:25pixelmaZagor: no, the mail table doesn't look nicer to me
00:32:49pixelmajust saw the "new" one
00:33:48pixelmait would look better if it had the same width as the svn table, but currently it looks better than when it was as long as the longest entry - in my eyes
00:34:29pixelmaor if the svn table was the widest as before
00:35:03 Quit iamben (Read error: 104 (Connection reset by peer))
00:35:17pixelma(meaning both would have the same width)
00:37:37 Quit elinenbe (Read error: 110 (Connection timed out))
00:38:30 Quit freqmod (Remote closed the connection)
00:40:19pixelma[aB]DaRkNeSS: depends on the theme - if it comes with the correct file structure, it's enough to unzip it to the root of your player and then you can chose it through "browse themes"
00:40:29[aB]DaRkNeSSi figured it out
00:40:32[aB]DaRkNeSSbut i dont see album art
00:40:41[aB]DaRkNeSSand it looks diff then the pics
00:40:48[aB]DaRkNeSSif i manually add music
00:40:52[aB]DaRkNeSSdo i put it in .rockbox
00:41:17pixelmaalbum art is not possible if you use an official build
00:41:33[aB]DaRkNeSSoh rly
00:41:51[aB]DaRkNeSShow do i use an unofficial one
00:41:54[aB]DaRkNeSSto see it
00:41:58[aB]DaRkNeSSthe pics all show album art
00:42:41pixelmayou would either need an "unsupported build" from the right place (but it's called unsupported for a reason) or make your own builfd
00:43:17pixelmathere are some in the forums
00:43:48SoapBagder: I like the new layout for the front page. Now I can simply use as my first-stop
00:44:09BagderI agree
00:45:35[aB]DaRkNeSSwanna link me
00:45:38[aB]DaRkNeSSto one you might know of :p
00:46:07 Join iamben [0] (
00:46:28pixelmaI know nothing about unsupported builds, because a) I'm not interested and b) there's none for my player :P
00:47:03pixelma[aB]DaRkNeSS: but you can search around here
00:47:04Zagorfront page now replaced now
00:47:31*pixelma keeps old one in the browser...
00:48:05Zagoryou can still get it at
00:48:17[aB]DaRkNeSSi cant watch mp4 movies for osme reason
00:48:18[aB]DaRkNeSSon rockbox
00:48:21[aB]DaRkNeSSon my 80 gig vid ipod
00:49:49pixelma[aB]DaRkNeSS: please search a bit for info... there is a wiki, there is the forum (with a nice "search" feature) and this question is answered a lot of times...
00:49:51 Quit merbanan ("Leaving")
00:50:16[aB]DaRkNeSSi did search
00:50:17pixelmaor read the manual, especially the "Mpegplayer" part in the "Plugins" sections
00:50:21[aB]DaRkNeSSbut chillax homie
00:50:27[aB]DaRkNeSSit said if i click it in my database
00:50:30[aB]DaRkNeSSbut it didnt show up
00:50:49n1sHmm this crash.on trackskip in the 64 bit sim is pretty weird... whatever is placed after "next_track:" in either flac.c or vorbis.c crashes, doesn't seem to matter what is there really...
00:51:20*amiconn will soon see the effect of his libmad work
00:51:24Llorean[aB]DaRkNeSS: I assure you that it did not say to click MP4 videos in the database (or any sort of video in the database)
00:51:46Zagorn1s: does it really run up to that point, or could it be jumping there from a wild pointer?
00:51:51LloreanDatabase is a music-only feature.
00:52:14n1salso it only happends on a skip to a file played by the same codec...
00:52:26Zagorn1s: or what exactly do you mean by "crashes"?
00:52:37n1sZagor: Segfault
00:52:40***Saving seen data "./dancer.seen"
00:52:58Zagorif you put asm("nop") there will it segfault too? or a++;a++;a++ etc
00:53:08*amiconn is listening to mp3 in a 64 bit sim :-P
00:53:15n1samiconn, wow
00:53:33n1sZagor: I will try
00:53:38amiconnNow I need to check whether it doesn't break 32 bit sims, or targets
00:54:55amiconnThe fix really wasn't difficult, most of it was the bit handling stuff
00:55:36[aB]DaRkNeSSit didnt show in the filelist either.
00:56:00[aB]DaRkNeSSmaybe its only mpeg1 and 2
00:56:02[aB]DaRkNeSSlike it says
00:57:01Llorean[aB]DaRkNeSS: Did you think it was lying to you when it said that?
00:57:09pixelmathe sentence about the supported formats wasn't just arbitratrily put there ;)
00:57:41Zagorpixelma: if you feel like doing a revised suggestion of the front page, I am interested despite how I may sound.
00:58:12n1sZagor: neither the "nop" or a++ things crashed it went past them fine and hit the first statement that was there before (in flac.c samplesdone=ci->id3->offset;)
00:58:56n1sI tried reordering stuff but putting codec_init() first made it segfault in that function instead etc.. this is really over my head :-/
00:59:36[aB]DaRkNeSSwowee you nerds are pissy
00:59:48[aB]DaRkNeSSdont bust rhetorial questions on me
00:59:49[aB]DaRkNeSScutie pie
00:59:53pixelmaZagor: I'll get a real good idea, ATM it was just a suggestion how it could look slightly better (as I said I didn't try, it was just my feeling)
01:00:05Zagor[aB]DaRkNeSS: yeah, and we love helping people who insult us
01:00:21*pixelma gives up correcting typos...
01:00:31[aB]DaRkNeSSi didnt start until you started being a female dog
01:00:36[aB]DaRkNeSSi just installed it today
01:00:39[aB]DaRkNeSSdont be gay about it
01:00:40Mode"#rockbox +o Bagder " by ChanServ (ChanServ@services.)
01:00:48Kick(#rockbox [aB]DaRkNeSS :Bagder) by Bagder!n=daniel@rockbox/developer/bagder
01:00:49 Join [aB]DaRkNeSS [0] (
01:01:00Bagdernow behave or begone
01:01:19Zagorn1s: is ci corrupted then? can you set a breakpoint and inspect it?
01:02:13Llorean[aB]DaRkNeSS: The rules are pretty simple around here. There's guidelines posted, they include reading the manual before asking questions. If something doesn't work how the manual says it does, then feel free to ask about it.
01:02:19n1sZagor: I can try but I'm not really sure what I'm doing :)
01:02:44 Join sansaman [0] (i=485fd0ae@gateway/web/cgi-irc/
01:02:56sansamanany1 use it on sansa
01:03:00[aB]DaRkNeSSi did read the manual
01:03:22pixelma[aB]DaRkNeSS: if you have to answer the question a hundred times again and again - and put a lot effort in writing a manual and gather information elsewhere, you'll get disappointed...
01:03:22Bagderthere's no any1 here
01:03:45sansamanit wont boot on mine i am pissed and need help
01:04:00Bagderhe's not here either
01:04:12Lloreansansaman: Try using real words, most of us speak English.
01:04:17amiconnShort for Postleitzahl (area code)
01:04:18Bagdersansaman: see topic
01:04:31*Bagder goes to bed
01:04:54Bagderno I'm not god
01:04:59Bagdernot yet
01:05:16sansamanholy hell are you like 5?
01:05:20Lloreansansaman: Think of it as basic courtesy. You're asking someone here to spend time to help you, can you spend time to type the way they prefer to read so that they're comfortable while doing so?
01:05:39Bagdersansaman: get real, speak english
01:05:44sansamanwow at least that makes sense
01:05:57Bagderask your question
01:06:23sansamanMy Sansa e200 series mp3 player wont boot the rockbox current build
01:06:51 Quit ender` (" I have the heart of a small boy. I keep it in a jar on my desk.")
01:07:19Lloreansansaman: What happens, and did previous builds work?
01:07:23amiconnn1s: Btw, libmad now faces the same problem as all other codecs on 64 bit - it crashes on track change
01:07:32sansamanthis is my first build
01:07:47LloreanWhat exactly happens then?
01:08:04sansamanit just starts the regular sansa boot
01:08:18*amiconn tries running it under gdb control
01:08:25linuxstb_sansaman: Did you run the sansapatcher program?
01:08:32amiconnSomehow gdb makes playback distorted??
01:08:52n1samiconn: yep, I've been taking a look at it tonight, only thing I've found out really is that it doesn't happend if the next track needs a different codec...
01:08:56linuxstb_sansaman: There you go then... It's described in the manual.
01:09:08n1svery good job on getting mp3 working :-)
01:09:08sansamannot really
01:09:20 Join Fraser [0] (
01:09:21sansamanwhere is the patch
01:09:26Lloreansansaman: Yeah, actually it's very clearly stated in the manual. See the "Installing the Bootloader" section
01:09:36sansamanthank you
01:09:36*amiconn is puzzled
01:09:48n1samiconn: playback in gdb works for me but I can't do much useful stuff with it
01:09:52amiconnWhy does running under gdb make playback distorted?
01:10:05sansamannothing else would help
01:10:14amiconnIt's as if it skips chunks of samples
01:10:46amiconnn1s: A backtrace should already tell quite a bit
01:11:25n1samiconn: that's the main problem I don't really understand what it's telling me :-/
01:11:46sansamanthey hid that part until you guys said to do it
01:11:52sansamani swear
01:12:13n1samiconn: Zagor had a theory about ci being corrupted
01:12:51amiconnWrong cast somewhere? But that should trigger a warning...
01:14:37sansamanbig confusing download list for sansapatcher what do i do
01:15:36sansamannever mind
01:16:10sansamanYou all deserve a hug
01:16:35 Part sansaman
01:16:41 Quit Frazz (Read error: 110 (Connection timed out))
01:18:38XavierGrweis there a reason why the Rockbox logo and supported targets aren't center aligned in the front page?
01:18:51XavierGr-we at start
01:22:46 Join tchan [0] (n=tchan@lunar-linux/developer/tchan)
01:24:06 Quit ompaul ("night all")
01:26:10 Quit XavierGr (Remote closed the connection)
01:26:26 Join XavierGr [0] (
01:27:08ZagorXavierGr: bcause it looks strange full-screen
01:27:37Zagori'm not very happy with the look myself
01:28:59pixelmaZagor: but not centered doesn't look any better full-screen too ... :P
01:29:40LloreanWhat we need is one more table, to go to the right of the targets, all squishy
01:29:52 Quit miepchen^schlaf ("Verlassend")
01:30:14ZagorLlorean: ?
01:30:40LloreanWell, on my 16:9 screen at 1280x800 there's uncomfortable blue space in the top right area.
01:30:45LloreanAnother table would fill it
01:31:24Zagoryeah. this is a problem with widely varying browser widths. many people (myself included) don't browse full-screen
01:31:45LloreanMaybe no elegant solution
01:31:47*Llorean must go
01:31:47 Part Llorean
01:32:29 Quit newbyx86 ()
01:32:56 Join newbyx86 [0] (
01:36:41 Join frimpy [0] (
01:37:35 Part n1s
01:37:58 Quit BigMac (Remote closed the connection)
01:38:01 Join elinenbe [0] (
01:38:26 Quit obo ("bye")
01:38:40 Join BigMac [0] (
01:43:34 Nick fxb__ is now known as fxb (
01:44:11 Join kkurbjun [0] (
01:46:38 Quit Febs (Read error: 110 (Connection timed out))
01:52:50 Join RudMan [0] (
01:53:09*amiconn summons limbus
01:55:51 Join kkurbju1 [0] (
01:56:32 Quit kkurbju1 ("Leaving.")
01:57:27amiconnZagor: Could you disable limbus' server temporarily? "No space left on device" :\
01:57:37 Quit kkurbjun (Nick collision from services.)
01:58:36 Join kkurbjun [0] (
01:59:40Zagorunfortunately I can't. I haven't looked at the build system..
02:03:17 Part pixelma
02:03:47amiconnThe subversion list on the frontpage doesn't update
02:12:38 Quit iamben (Read error: 104 (Connection reset by peer))
02:14:41amiconnZagor: Your theory seems to be correct... the ci->request_next_track() callback seems to trash ci...
02:14:48amiconn(on 64 bit, that is)
02:14:57Zagorwow. that was just a wild guess.
02:15:07amiconnI'm about to step through that callback...
02:17:42 Quit Fraser ("Leaving")
02:24:33amiconnThe queue_send() call (apps/playback.c:1901) seems to trash ci ???
02:25:08 Quit elinenbe (Read error: 110 (Connection timed out))
02:25:31 Join chrisjs169|away [0] (
02:26:53 Quit jhMikeS (Nick collision from services.)
02:26:59 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
02:28:29 Join Febs [0] (
02:32:31 Quit linuxstb_ ("CGI:IRC")
02:35:11amiconnsdl-sim queue_send() calls switch_thread() (uisimulator/sdl/kernel.c:244) which obviously maps to yield(). ci is trashed in yield() directly after SDL_mutexV() is called (thread-sdl.c:39)
02:35:34amiconnNow someone understand that, please
02:35:45 Join iamben [0] (
02:36:50 Quit Zagor ("Client exiting")
02:40:39 Join annulus_ [0] (
02:42:29 Join advcomp2019_ [0] (n=advcomp2@
02:42:41 Quit advcomp2019 (Nick collision from services.)
02:42:43 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@
02:46:01 Quit iamben (Read error: 104 (Connection reset by peer))
02:51:12 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
02:52:44***Saving seen data "./dancer.seen"
02:53:41 Join iamben [0] (
02:58:47 Quit Nico_P (Remote closed the connection)
03:00:17 Join kugel_ [0] (
03:00:21 Quit kugel_ (Client Quit)
03:03:18 Join Thundercloud [0] (
03:18:43 Quit kugel (Read error: 110 (Connection timed out))
03:27:27jhMikeSamiconn: what problem are you trying to diagnose?
03:31:48 Quit Thundercloud (Remote closed the connection)
03:40:30 Join firelord [0] (
03:41:15 Part firelord
03:44:16 Join Jangari [0] (
03:45:01 Quit XavierGr ("One firmware to rule them all!")
03:57:26 Quit kclaf (Read error: 110 (Connection timed out))
04:14:39 Quit Jangari ("There's no wrong way to eat a Hoss")
04:22:23 Quit Pyromancer ("Leaving")
04:22:29 Join BigBambi [0] (
04:23:07 Join saratoga [0] (i=98039a9c@gateway/web/cgi-irc/
04:30:43 Quit BigBambi ("Leaving")
04:31:00 Join BigBambi [0] (
04:35:42 Join Pyromancer [0] (
04:43:14 Quit eigma (Read error: 104 (Connection reset by peer))
04:45:37 Quit Entasis (Read error: 104 (Connection reset by peer))
04:52:47***Saving seen data "./dancer.seen"
04:56:42 Quit kkurbjun ("Leaving.")
04:59:52 Quit jhMikeS (Read error: 104 (Connection reset by peer))
05:05:04 Join jhulst [0] (
05:05:37 Join bb [0] (n=bb@unaffiliated/bb)
05:16:01 Join bb__ [0] (
05:18:25 Quit bb_ (Read error: 110 (Connection timed out))
05:20:59 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
05:21:10 Join Entasis [0] (
05:27:30 Quit Entasis (Read error: 104 (Connection reset by peer))
05:27:45 Quit bb (Read error: 110 (Connection timed out))
05:28:53 Quit donutman25 ("ChatZilla [Firefox]")
05:35:17 Join Entasis [0] (
05:37:14 Quit n17ikh|Lappy (Read error: 104 (Connection reset by peer))
05:38:09 Join n17ikh|Lappy [0] (
05:43:51 Join kkurbjun [0] (
05:54:12 Join alienbiker99 [0] (
05:57:28 Join Llorean [0] (
06:09:17 Join Jangari [0] (
06:24:01 Quit idnar (Nick collision from services.)
06:24:03 Join idnar_ [0] (i=mithrand@unaffiliated/idnar)
06:27:52 Nick idnar_ is now known as idnar (i=mithrand@unaffiliated/idnar)
06:29:07 Join maxkelley [0] (
06:33:52 Join XavierGr [0] (
06:52:51***Saving seen data "./dancer.seen"
06:53:05 Join My_Sic [0] (
07:01:48 Quit saratoga ("CGI:IRC")
07:03:44 Part Llorean
07:16:04 Quit frimpy ("Leaving")
07:19:13 Join daurnimator [0] (n=fake@unaffiliated/daurnimator)
07:36:05 Quit gtkspert (Remote closed the connection)
07:41:00 Quit grndslm (Remote closed the connection)
07:42:01 Join gtkspert [0] (
07:50:07 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
08:04:31 Join Ebert [0] (
08:05:45 Join jack_ [0] (
08:06:02 Join webguest90 [0] (i=4362036a@gateway/web/cgi-irc/
08:06:31midkaywoo. neat new homepage, though i think it definitely needs a bit of work, i like the idea. :D
08:07:33 Quit chrisjs169|away (Read error: 110 (Connection timed out))
08:07:43 Join The-Compiler [0] (
08:08:06scorcheooo...i hadnt noticed
08:08:50scorchealthough i would like the commit log to be as it was before...
08:09:11scorchebut /recent is for that, i suppose
08:09:23midkaythat's my main complaint..
08:09:23 Join Zagor [0] (n=bjst@rockbox/developer/Zagor)
08:09:42midkayconsidering probably >95% of visitors are end users, the very most important thing to see on the front page as far as SVN is *what* was changed.
08:10:10midkaythose users don't care who did it, or what files were changed - most don't even know who these people are or what the files are for, they just want to see what's been modified.
08:10:17 Join witz [0] (n=witz@unaffiliated/witz)
08:10:40 Quit webguest90 (Client Quit)
08:10:46scorchea svn log without the message is just weird to see
08:10:48Zagorjust a quick paste of interesting links before I have to run: and <−−- disassembled
08:11:02scorcheZagor: aye...they have been linked here before
08:11:20Zagorah, ok. missed them in the log then. good, gotta go.
08:11:27 Quit Zagor (Client Quit)
08:17:04 Join Rob222241 [0] (
08:23:34 Quit witz ("good night!")
08:24:13 Quit perplexity (Read error: 113 (No route to host))
08:30:44[aB]DaRkNeSSanyone here use evilg fusion
08:31:06scorchethat is an unsupported build
08:32:06amiconnjhMikeS: [03:27:27] <jhMikeS> amiconn: what problem are you trying to diagnose? <== The sim crashes on track change or skip on linux/amd64, if the next track uses the same codec as the current one
08:33:37[aB]DaRkNeSSdoes anyone use it
08:33:57[aB]DaRkNeSSi didnt ask if it was supported
08:34:01 Join ivan` [0] (n=ivan`
08:34:17scorche[aB]DaRkNeSS: hence, it is offtopic here
08:35:16 Quit Rob2222 (Read error: 110 (Connection timed out))
08:36:19[aB]DaRkNeSSanyone use rockbox?
08:36:46scorcheplenty there something specific you would like to ask?
08:38:12[aB]DaRkNeSSwhy i cant get AA to show
08:38:32scorcheAA is not in the official build
08:39:03[aB]DaRkNeSSim trying with rockbox
08:39:06[aB]DaRkNeSSusing evilg
08:40:35 Quit ivan` (Remote closed the connection)
08:40:36jhMikeSamiconn: I saw the sequence you posted. Odd that ci should fall anywhere near that.
08:41:04amiconnI was single stepping through the code in gdb once I knew where it fails
08:41:28[aB]DaRkNeSSanyone wanna be cool
08:41:46 Quit n17ikh|Lappy ()
08:42:06scorche[aB]DaRkNeSS: this isnt about being cool...please see the link in the topic for the channel guidelines and read them
08:42:06amiconnAfaiu that single stepping only happens in one thread, so that yield() lets another thread run which trashes ci
08:42:42 Quit darkless (Read error: 104 (Connection reset by peer))
08:42:43jhMikeSIf the lock is unlocked by a thread other than the owner, it will crash the sim which seems what's plausible atm.
08:43:00amiconnThe trashing happening directly after the sdl mutex unlock points to that that
08:43:08amiconnThe lock is unlocked by the owner
08:43:38amiconnI mean that there is a problem with how the sim handles codecs, which only shows up on amd64 so far
08:43:41jhMikeScould've happened in the past some time
08:44:16amiconnMaybe it reloads the codec, even though it doesn't need to, and hence shouldn't
08:44:54jhMikeSI wonder if would do that with the dual core patch which makes sim threading act like rockbox threads.
08:45:10[aB]DaRkNeSSdont be gay scorche
08:45:14 Join ivan` [0] (n=ivan`
08:45:19amiconnI remember having read that there is a recent method to make exploiting buffer overflows harder, which will load an object to different addresses in memory
08:45:23scorche[aB]DaRkNeSS: last warning...keep it ontopic, or leave
08:45:42[aB]DaRkNeSSyou gonna kick me?
08:45:49jhMikeSamiconn: built into hardware?
08:45:55amiconnIf the linux kernel does that on amd64 but not on x86, it would explain a lot...
08:46:00[aB]DaRkNeSSim just asking questions
08:46:02scorche[aB]DaRkNeSS: possibly
08:46:05[aB]DaRkNeSSyoure being ghetto
08:46:06amiconnNo, not hardware, kernel
08:46:12[aB]DaRkNeSSgo ahead
08:46:16[aB]DaRkNeSSmr non @
08:46:19[aB]DaRkNeSSget some ops
08:46:26Mode"#rockbox +o scorche " by ChanServ (ChanServ@services.)
08:46:29jhMikeSbut pointers have to remain valid :\
08:47:11amiconnYes, they do as long as a lib is loaded
08:47:22[aB]DaRkNeSSgo ahead now scorche
08:47:28[aB]DaRkNeSSkick me for having interest
08:47:31[aB]DaRkNeSSin rockbox
08:47:37[aB]DaRkNeSSwhen clearly linux is so much better
08:47:51amiconnI think what's going on here is that the playback thread reloads the codec (hence it gets a new address), but it actually tries to continue using the old copy
08:48:18Mode"#rockbox +b %[aB]DaRkNeSS!*@* " by scorche (i=Blah@rockbox/administrator/scorche)
08:48:40amiconnI need to set some breakpoints in the playback thread as well, it seems
08:48:49scorche[aB]DaRkNeSS: you are muted for 10 min...if you wish to follow the guidelines after that time, feel free to do so
08:49:03 Join darkless [0] (
08:49:31jhMikeSamiconn: we load codecs as os modules in the sim and not just raw code?
08:49:59amiconnCodecs and plugins are loaded as shared objects (*nix) resp. DLLs (windows)
08:50:33amiconnI can't see how loading them as raw code would work
08:50:48jhMikeShmmm...which would make the data invalid if two copies are made. would also explain why voice would fail when playing.
08:51:55amiconnVoice loads another copy of the so/dll
08:51:57jhMikeScan't see why it wouldn't. it perfectly acceptable to have self-modifying code.
08:52:46amiconnWe cannot know which architecture the sim will run on (could be x86, amd64, ppc ...). Loading raw code would require it to be linked to a fixed address
08:52:54***Saving seen data "./dancer.seen"
08:54:56amiconnHow would you do that without knowing the architecture?
08:55:02jhMikeSWindows relocates DLLs. I can't believe these thing can't be made relocateable.
08:55:40amiconnThen it's not raw code anymore, and we can just use so/dll like we do
08:56:09 Join Llorean [0] (
08:56:14jhMikeSI was thinking to keep it as sim says, not the os.
08:56:21amiconnOtherwise we would have to do the reloacation "by hand", which would require even more knowledge about the platform
08:58:22Mode"#rockbox -b %[aB]DaRkNeSS!*@* " by scorche (i=Blah@rockbox/administrator/scorche)
08:58:28 Join spiorf [0] (
08:59:31jhMikeSas long as the reload can start with a fresh one. hmmm. by addresses you mean ci doesn't get updated to the new one?
09:00:27amiconnci does get a different address.
09:00:44jhMikeSwhat old address would it use? the old module handle?
09:00:51amiconnMaybe my theory is wrong, but I can't tell without extended debugging
09:01:50 Quit Ebert ()
09:02:09 Quit RudMan (Remote closed the connection)
09:02:46Mode"#rockbox -o scorche " by ChanServ (ChanServ@services.)
09:03:27jhMikeSthe currnet sim implementation of yield is just awful xP
09:04:33amiconnMaybe, but I don't think that's what causes the problem
09:04:48amiconnAnd it's a KISS solution that works...
09:06:10 Join n1s [0] (
09:08:21jhMikeSthere's a simpler one that works far better
09:09:18jhMikeSnot to mention running IO on a separate thread that isn't bound by the rockbox threads. that way rockbox threads can yield to each other while IO happens. huge performance improvement.
09:11:12 Quit The-Compiler (Read error: 104 (Connection reset by peer))
09:14:03 Quit Jangari (Read error: 110 (Connection timed out))
09:23:28 Join n17ikh|Lappy [0] (
09:24:10jhMikeScygwin sim uses the windows SDL I take it? there's no emulation there I take it.
09:24:59amiconnYes, as the "cygwin sim" is actually a windows sim
09:25:08amiconnIt's compiled with -mno-cygwin
09:28:23 Join Genre9mp3 [0] (
09:29:04 Join davina [0] (
09:30:51jhMikeSI should probably just do up the RB-accurate threading and IO separately from other scheduler updates. Not much to it except that kernel objects must be the same form in sim and target. I should check windows out for that but since I do know the windows thread APIs quite well, it should behave itself. (looked at SDL source)
09:31:18 Quit davina (Read error: 104 (Connection reset by peer))
09:32:06 Join davina [0] (
09:37:36 Join DiceRy [0] (
09:43:31 Join sasukee [0] (n=SK@unaffiliated/sasukee)
09:43:38 Join eigma [0] (
09:45:26 Join petur [0] (n=petur@rockbox/developer/petur)
09:48:44amiconnjhMikeS: Why should disk i/o run outside the rockbox threading in the sim, btw?
09:49:18amiconnIt doesn't do so on target, and the sim should behave as close to the target as possible
09:50:41 Quit tchan (Read error: 113 (No route to host))
09:51:56 Join ompaul [0] (n=ompaul@freenode/staff/gnewsense.ompaul)
09:51:57jhMikeSit behaves much more like a real device like that. think of it as a simulted DMA transfer and yielding happens just like the ATA driver.
09:54:27n1samiconn: foun dout one more thing about the trackskip issue, ci->filesize is always 6993184 for the first track played but is reset to the correct value when skipping... dunno if that matters
09:55:03 Join sasukee_ [0] (n=SK@unaffiliated/sasukee)
09:55:04amiconnjhMikeS: I don't know of any target that does disk dma except the ondios
09:56:09amiconnn1s: Hmm, no idea yet. Did you try the libmad fix already? :)
09:57:06n1samiconn: yep works very nicely :-)
09:57:43 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
09:59:29 Join sasukeee [0] (n=SK@unaffiliated/sasukee)
09:59:32jhMikeSamiconn: doesn't matter the simulation is more accurate and acutally runs. gigabeat also does it btw.
10:00:01 Quit sasukee_ (Read error: 104 (Connection reset by peer))
10:00:23jhMikeSthe dodgey execution when buffering disappears which makes it more target like
10:02:15 Join sasukee_ [0] (n=SK@unaffiliated/sasukee)
10:02:35 Join linuxstb_ [0] (i=5343d4aa@rockbox/developer/linuxstb)
10:03:54amiconnjhMikeS: I disagree. Several targets also operate dodgy when buffering (most prominent: ipod video)
10:05:19jhMikeSwell, none that I have do. perhaps that should be fixed on ipod video. so you think the worst should make the determination and make the sim run awfully?
10:05:51 Join t0mas [0] (n=Tomas@rockbox/developer/t0mas)
10:06:38amiconnIt's just a matter of cpu load. And I have no idea how to fix that on ipod video, unless someone figures out how to do ata dma on pp
10:06:41jhMikeSI'd suggest we implement emulators then if absolute 100% accuracy must be observed on all otherwise you'll never have.
10:07:02jhMikeSOn sansa I did periodic yields to keeping going.
10:08:11 Quit sasukee_ (Read error: 104 (Connection reset by peer))
10:08:24 Join sasukee_ [0] (n=SK@unaffiliated/sasukee)
10:08:32amiconnjhMikeS: The problem isn't yield. Buffering together with lcd updates draws so much cpu power that the priority stuff kicks in and slows down ui
10:08:49*scorche waves at t0mas
10:09:21 Join desowin [0] (
10:10:37 Quit sasukee (Read error: 110 (Connection timed out))
10:11:09jhMikeShmmm...seems rather unusual and not the standard I intend to work by but I can surely make it dodgey for an iPod video sim just to make sure it runs poorly there too. :)
10:12:32amiconnI'm not against making the sim scheduler more similar to the target, but I wouldn't put i/o outside the rockbox threading, unless the majority of our targets use disk dma
10:13:49amiconn...which they don't. Gigabeat does disk dma though, but only for aligned data
10:17:02 Quit eigma (Read error: 110 (Connection timed out))
10:17:03jhMikeSI think I'd do it just to make it not run like crap and one without disk DMA would still run more like it should. The stuff with iPod video I do know what's going on there.
10:19:13*amiconn wonders how multiprocessor PCs ensure cache coherency
10:19:29 Quit sasukeee (Read error: 113 (No route to host))
10:19:30amiconnI mean multiprocessor, not just multicore
10:19:33jhMikeSThe attempt to kick priority in can actually make the situation worse because the priority implementation isn't complete.
10:19:43jbitamiconn: bus snooping
10:20:00*jhMikeS wishes PP had bus snooping. :\
10:20:15amiconnHow? The external bus is much slower than the L1 caches, isn't it?
10:20:40jbitamiconn: yup, but there are algorithms based on cache line size/etc
10:20:53jbitto determin if something has been changed or not
10:20:54jhMikeSI'd imagine some arbiter would be needed or else you'd get the line-overlap problem making all processor's data invalid.
10:21:20jbitcache coherency is over rated anyway :)
10:21:29jbitespecially on embedded systems
10:22:02jhMikeSwho needs it ... Prefetch aborts at 0xC0EDBABE are so much better ;)
10:22:37jbitwell on embedded machines you should always know what's in your cache and waht's acutally in ram anyway (not talking out my ass, done this for many systems)
10:22:41 Quit sasukee_ ("Leaving")
10:25:29jhMikeSit seems there's the need to watch alignment of buffers that should never get line-read by another core or else a later flush can blast another processor's data even though they never read or write the same addresses.
10:27:30amiconnjhMikeS: Hmm, perhaps have a second set of sections for those?
10:27:49amiconncop_data, cop_bss etc...
10:28:30 Join cendres [0] (n=ashes@2001:5c0:8fff:ffff:0:0:0:13)
10:29:12amiconnIn fact that should be the only two
10:31:11cendresi used to be subscribed to the rockbox mailing list, and unsubscribed because i didn't seem to have any conversation to contribute. but now i have a couple ideas
10:31:22 Quit Siku ()
10:31:36cendresi have an iriver
10:32:41cendresim curious why linux isn't used for this. ipaq's have linux based systems. i'm not fluent in C, and rockbox's source code doesn't make sense to me because im more familiar with code that interacts with a desktop, and opie, seem to be close to ideal for an iriver and rockbox system
10:34:01cendres'familiar linux', and even 'intimate linux', should work well on an iriver, and would provide the ability to play movies too
10:34:02jhMikeSamiconn: hmmm could work. it doesn't seem to need alignment > 16 bytes but things like the allocator implementation in mpegplayer would still need some explicit coding I think.
10:35:18cendresa linux based system, or even netbsd bases system, would help with things like network time. if the mp3 player is running an ntp client, when it becomes connected to usbnetwork the time and date can be updated (irivers don't keep perfect time, they lose several minutes after a couple months)
10:35:24Lloreancendres: Linux wouldn't magically make the devices faster.
10:35:49cendresi understand there is no linux distribution that would allow anyone to use an iriver
10:35:58cendresbecause the drivers don't exist, yet
10:36:02LloreanBesides, this is #rockbox, Rockbox is not Linux based.
10:36:02 Join tchan [0] (n=tchan@lunar-linux/developer/tchan)
10:36:26cendresim curious what advantages rockbox has over linux
10:36:37LloreanA lot less complicated.
10:36:42cendreswhy wasn't linux used as the base for rockbox?
10:36:51LloreanBecause there was no reason to.
10:36:59jbitlinux isn't as suitable for embedded devices as some people think
10:37:21sbehcendres: because embedded linux is still to fat
10:37:31*jhMikeS prefers writing his own kernel code without answering to Torvalds
10:37:35scorchelinux contains a lot of cruft that we dont need, and we have full control over everything and anything
10:37:48linuxstb_jhMikeS: Yes, you just have us lot instead... ;)
10:37:59cendresrockbox looks like it was written from scratch, which is great and very specific for our needs. but it would be less effort to port linux to iriver, because many of the linux applications which would run on the iriver would be maintained by another development group
10:38:17cendresand so, the rockbox developers could spend time on better things
10:38:19*jhMikeS mis-parsed linuxstb_'s statement: "us lot" ?
10:38:40cendresim talking about uclibc, or even dietlibx
10:38:55 Join Siku [0] (
10:39:29Lloreancendres: Rockbox isn't designed around apps
10:39:46Lloreancendres: Rockbox is designed to be a good music player. Porting linux and adding a bunch of overhead goes exactly contrary to this purpose.
10:39:48cendresi would agree that a redhat port to iriver would be serious overkill. im talking about ipaq-like systems. the familiar-linux pda linux is about 4MB
10:40:16cendresand includes a lot of applications
10:40:24n1scendres: how would it be less effort to port linux when rockbox is already working on the devices?
10:40:24cendresit might be cut down to 2MB
10:40:28Lloreancendres: Rockbox is NOT about Applications.
10:40:58Lloreancendres: Feel free to start your own project with the purpose of turning MP3 players into PDAs though, or talk to iPodLinux about becoming cross-platform, but you seem to have misplaced the intent of the software.
10:41:06cendresn1s: im not specifically talking about today. im curious why when rockbox was born that the developers didn't use linux for the kernel
10:41:24amiconncendres: There are rockbox targets which have a *total* RAM size of 1MB
10:41:28jbitcendres: the extra-cruft people talk about isn't just all the applications in "redhat" etc, it's all teh random stuff in the kernel that really isn't needed for an embedded device and just bloats stuff up (or even causes more cpu/ram usage, which decreases baterry life)
10:41:41pixelmacendres: do you know on what devices Rockbox was born?
10:41:42*jhMikeS is pleased to say he hasn't used a non-dual core rockbox kernel in weeks and is using a dual-core spc codec that has tight interaction between processors using kernel object. not a that leaves pesky PP5020 to conquer.
10:41:57LloreanjhMikeS: Well snap, snap.
10:42:00cendresi think i understand that rockbox is mp3 player, just like the software for my television. it'd designed to do what it needs to, and nothing else
10:42:11amiconnjhMikeS: Did you use voice?
10:42:11jhMikeShopefully only a couple days :)
10:42:30jhMikeSI can use voice just fine. mpegplayer problems don't change of course.
10:42:51cendresbut it would be very nice to have network time protocol on my mp3 player
10:43:57cendresso once a week i can usb connect it, and the mp3 player's clock will be accurate within a minute during the week
10:44:03Lloreancendres: Yes, it sounds like a wonderful idea. Sacrifice Rockbox's ability to play music efficiently to keep the clock in sync...
10:44:06jhMikeSthe spc codec as I've implemented here is a more authoritative test than mpegplayer with the shared semaphore and event use
10:44:13*Llorean wears a watch.
10:44:23amiconncendres: I am very curios how you think that could work...
10:44:45 Quit linuxstb_ ("CGI:IRC")
10:44:48jbitcendres: why do you need ntp for that? when the new usb stack for rockbox gets working, you could easily write a pc side program and a rockbox side plugin to do that
10:44:54cendreshotplug ntpd, so it only runs when the usb network is online. this way the cpu and ram usage is as minimal as possible
10:45:04n1scendres: you couldn't just hook it up to a computer and start communicating with the net, you'd need to add it's own usb nic probably...
10:45:08amiconnThere is no usb network, and can't be on irivers
10:45:25Lloreanamiconn: On the h300's "media" port?
10:45:25amiconnIriver usb is done in hardware, via a usb-ata bridge
10:46:04cendresi didn't realize that... the iriver doesn't have a network interface. that's why it shows up as a disc drive
10:46:15*jhMikeS just hammered on voice alot to check problems
10:47:18amiconnLlorean: Yes perhaps, but we don't have a driver for that yet
10:47:52amiconnAnd linux wouldn't magically solve any driver problem. Drivers need to be written no matter whether the kernel is linux or rockbox
10:47:57LloreanStill, I think the most KISS solution is "check your watch once a week, and fix it"
10:50:37*jhMikeS checks his atomic watch and tends to his radiation burns
10:50:53jbiti'm going to hazard to say that writing drivers for rockbox is easier than for linux too :)
10:51:46*pixelma 's time supplied by the RTC in the M5 is more accurate than her watch :)
10:52:32cendresi own a fiori pocket watch (made in canada). i set it to ntp time when i first got if. after 1 year it still hasn't lost a single second
10:52:32*jhMikeS agrees the RTC in Cowons keeps pretty accurate time.
10:52:57***Saving seen data "./dancer.seen"
10:53:21n1sso anyone familiar with the codec api stuff?
10:53:34cendrestime is important to me. i need to be able to trust my watch/clock
10:54:07*jhMikeS has never dealt with the code api ;)
10:54:17jhMikeScodec even
10:54:55n1sif I put a breakpoint at codec_start and print api->filesize it is correct, the first line in that function for sims is ci=api; and after that both have an incorrect value
10:55:04*amiconn just got a data abort on his 2nd gen right after resuming
10:55:41amiconnSomething seems to be wrong again with voice :(
10:56:17jhMikeSn1s: just the assignment trashed it?
10:56:39n1sjhMikeS: it seems so to me but it's really strange
10:57:35jhMikeSdoes api change value?
11:00:50n1sjhMikeS: now I'm not sure...
11:01:04n1scould be the codec_main function too, hmm
11:01:28jhMikeSI thought it hadn't gotten that far yet
11:02:07n1sjhMikeS: not at first it seemed like it changed right before the assignment
11:02:12 Quit Genre9mp3 (Nick collision from services.)
11:02:14 Join Genre9mp4 [0] (n=yngwiejo@rockbox/contributor/Genre9mp3)
11:03:25 Nick Genre9mp4 is now known as Genre9mp3 (n=yngwiejo@rockbox/contributor/Genre9mp3)
11:03:47n1sjhMikeS: this is the output i get from gdb
11:04:08n1sminus the typos
11:04:37jhMikeSsounds like something that might happen with stacks and lack of thread sync...but that's just a wild guess atm.
11:06:55n1sThe strange thing is that the filesize is always 6993184 for the first file played, regardless of actual filesize btw 3218946 is the correct size for the file
11:07:38 Join Jdgordon [0] (
11:08:05amiconnn1s: Even more weird:
11:08:14jhMikeSCannot access memory at address 0x0 means ci is NULL, correct?
11:08:21amiconnIf I try to print api->filesize, I get "Cannot access memory at..."
11:08:33amiconnThat's _before_ assigning to ci
11:09:18n1samiconn: yep, I got something similar when trying to use DEBUGF it would segfault if I tried to print api before the assignment
11:09:42amiconnThis is plain svn code with no DEBUGF
11:09:59amiconnJust a breakpoint at codec_api.c:36
11:12:30amiconnHmm, one step later it works... maybe breakpoints at the opening brace of a function do something I don't expect?
11:13:08JdgordonBagder: please go back to the old frontpage (at least for the svn part)...
11:14:28scorcheJdgordon: aye...i do miss the old svn chart (although it is in /recent)
11:14:40amiconnn1s: Eh? I also get this api->filesize change...
11:14:54amiconnBefore the assignment it's corrent, after it it's wrong
11:17:22 Join hannesd [0] (
11:18:33amiconnn1s: filesize is the only element that changes (try 'print *api in gdb) - and it's the very first
11:19:38 Join seablue [0] (
11:19:39amiconnHehe, and api->filesize == api here
11:21:02jhMikeSwhat's RTLD_NOW mean?
11:23:13 Join lazka [0] (
11:26:20amiconnn1s: Maybe it's a symbol confusion only hitting on amd64 atm?
11:26:36amiconnThe codec api pointer is called ci both in the core and the codecs
11:26:44amiconn...which are shared objects
11:27:41 Nick Jdgordon is now known as JdGordon (i=jonno@rockbox/developer/JdGordon)
11:29:08jhMikeSthe symbol should be local to the module
11:29:50amiconnWell, _should_
11:30:03amiconnIt's not static at least
11:32:03desowinI don't quite like new website look
11:32:05jhMikeSeach dll has it's own data so I really don't see a problem there.
11:32:26 Quit hannesd ("Client suicide")
11:34:51 Join ender` [0] (
11:36:14amiconnjhMikeS: Both the codec and the core export 'ci'
11:36:20amiconn(on linux)
11:36:39*amiconn just learned about 'nm'
11:38:51jhMikeSummm...why should they export it? eek.
11:39:29amiconnI think symbols are exported by default
11:39:38 Quit t0mas ("Bye bye!")
11:39:49amiconnOn windows it's specified which symbols are exported from plugins and codecs
11:39:56amiconn(iirc - about to check)
11:40:07jhMikeSso, on linux there's a namespace for resolving symbols rather than an explicit GetProcAddress or runtime link?
11:40:12 Join t0mas [0] (n=Tomas@rockbox/developer/t0mas)
11:41:14 Join stripwax [0] (
11:41:40amiconnI'm no linux expert....
11:43:42 Nick bb__ is now known as bb (
11:51:50amiconnCould someone perhaps try something on 32 bit linux?
11:53:15*linuxstb could
11:53:46amiconnWould be nice... Step 1: compile an swcodec sim (preferably H300 but not required)
11:53:54 Join rvvs89 [0] (n=rvvs89@unaffiliated/rvvs89)
11:54:16amiconnStep 2: load it with gdb −−args ./rockboxui −−background
11:54:20linuxstbWhat audio files (if any) will I need for this?
11:54:23amiconnDon't run yet.
11:54:40amiconnAny will do - the strangeness already happens in codec_crt0.c
11:54:51linuxstbI'm just compiling a h300 sim now...
11:54:57amiconnSet a breakpoint at codec_crt0.c:37
11:55:28amiconngdb will ask whether this symbol will appear in a dynamically loaded object. Answer yes
11:55:38amiconnThen run, and try playing an audio file
11:55:52amiconnIt should immediately hit the breakpoint
11:55:59amiconnThen comes the interesting part
11:58:11linuxstb...still compiling...
12:02:22 Quit JdGordon ("Konversation terminated!")
12:02:45jhMikeSamiconn: this is a cygwin sim? the only 32-bit sim I'd have is an actual linux one on vmware.
12:04:43jhMikeScygwin's not an emulator so it uses windows functions underneath, right?
12:05:41linuxstbamiconn: OK, breakpoiint hit...
12:05:53amiconnjhMikeS: yes
12:06:14amiconnlinuxstb: Now print values: print/x api should show an address
12:06:29amiconnThe interesting thing is what print/x ci shows
12:06:43amiconnDoes it also show an address (0x0), or the actual structure?
12:06:46linuxstbci is 0, api is some address.
12:07:08amiconnWith just 'print/x ci' ?
12:07:31amiconnBecause, if I do 'print/x ci' on amd64, it shows the whole struct contents
12:07:43linuxstbYes, "print/x ci" displays "$3 = 0x0"
12:07:49amiconn...and 'print &ci' shows an address, which is not 0x0
12:08:13amiconnAnd the breakpoint is before 'ci = api'
12:08:44amiconnOkay, two more checks: print api->filesize
12:08:55amiconnIt should show the filesize of the file you selected for playback
12:09:11amiconnThen 'step' over 'ci = api', and print/x ci again
12:09:29linuxstbYes, filesize is correct.
12:09:57linuxstbci is now set to api
12:10:14amiconnAnd print ci->filesize also shows the correct size I presume?
12:10:56linuxstbYes. Here's my gdb session -
12:11:54amiconnOkay. Thanks for testing
12:12:18*amiconn wonders why the loader seems to confuse symbols on linux/amd64
12:18:08 Join JdGordon [0] (
12:22:39 Join SiM99 [0] (
12:25:24SiM99have i understood the manual correctly here - the ipod firmware (on the flash rom) is not touched, everything is done via the copying of files?
12:25:26JdGordonany e200r's here?
12:28:14 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
12:34:50linuxstbSiM99: Yes. A Rockbox install only modifies the contents of the disk, not the flash ROM.
12:34:54 Join FOAD_ [0] (
12:37:15 Quit JdGordon ("Konversation terminated!")
12:48:28 Quit FOAD (Read error: 110 (Connection timed out))
12:48:29 Nick FOAD_ is now known as FOAD (
12:50:07 Join agm3nt [0] (
12:53:00***Saving seen data "./dancer.seen"
12:53:25 Quit darkless (Read error: 113 (No route to host))
12:56:01 Join Frazz [0] (
13:02:04 Join datasleep [0] (
13:02:11 Quit datachild (Nick collision from services.)
13:03:54 Join pepie34 [0] (
13:11:59 Join Lear [0] (i=chatzill@rockbox/developer/lear)
13:15:18 Join FittizioAmbiguo [0] (
13:15:20 Join obo [0] (n=obo@rockbox/developer/obo)
13:15:22 Part FittizioAmbiguo ("Sto andando via")
13:30:38 Join JdGordon [0] (
13:30:52SiM99thanks linuxstb
13:31:09SiM99lol, took a while to respond there, sorry... was reading up on things
13:31:24SiM99got sidetracked into looking at ipod cases n things too :\
13:31:59 Quit agm3nt (Read error: 113 (No route to host))
13:35:58 Quit jhMikeS (Nick collision from services.)
13:36:04 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
13:36:14 Join Agm3nt [0] (
13:36:15 Join Thundercloud [0] (
13:36:57 Quit Febs (Read error: 110 (Connection timed out))
13:37:24 Join perplexity [0] (
13:39:01JdGordonanyone with an e200 here (preferably an R, but either is ok)
13:45:55JdGordoncan I get you to stick the e200r .rom on it and run some code in manufacturing mode?
13:46:08JdGordonIt keeps failing to copy on mine
13:46:40 Quit stripwax ("Miranda IM! Smaller, Faster, Easier.")
13:47:43 Quit thegeek (Read error: 113 (No route to host))
13:47:44desowinumm, there is risk isn't there?
13:48:05 Join kclaf [0] (i=kclaf@
13:48:42JdGordonumm... not really
13:48:55JdGordonworst happens you recover the e200 rom and its fine again
13:51:12JdGordondesowin: actually... unless youve started the recovering stuff it
13:51:17JdGordoncant be bothered with it now
13:54:29JdGordondesowin: you started? or can i leave for a while?
13:56:33desowinno, I didn't start
13:58:05 Quit JdGordon ("Konversation terminated!")
14:01:32 Quit tchan (Read error: 110 (Connection timed out))
14:04:48 Join decayedcell_ [0] (i=792cfd7c@gateway/web/cgi-irc/
14:15:32 Quit hcs (Read error: 110 (Connection timed out))
14:17:24 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
14:18:36 Quit pepie34 ("Ex-Chat")
14:22:10 Quit Nico_P (Remote closed the connection)
14:27:28 Quit BigBambi (Read error: 110 (Connection timed out))
14:31:47 Part pixelma
14:31:47*petur buys an h320 as (backup) recorder
14:34:56 Quit datasleep ("[Water is] the only drink for a wise man.")
14:36:09 Join JdGordon [0] (
14:38:09 Join darkless [0] (
14:41:36 Join Febs [0] (
14:41:52peturmessage for datasleep: don't drink water, fish fuck in it
14:42:45JdGordondesowin: im back if your still game ot have a play?
14:50:24amiconnjhMikeS: With your sim threading update, the sim acts *very* strange on linux/amd64
14:51:43amiconnIf you start playing a track, it switches to the wps. Peakmeters are moving, but no sound output, elapsed time stays at 0. More than 10 seconds later, it actually starts playing...
14:53:01***Saving seen data "./dancer.seen"
14:53:14amiconnAnd the sim draws ~90% cpu power - on an Athlon64 3800+ (!)
14:53:45peturso much like the old win sim
14:53:56amiconnClosing the sim doesn't work
14:55:22jhMikeSthere's a little dealy with the use of the sim threads anyway that should be addressed
14:55:30amiconnOn windows it also draws all available cpu power...
14:55:31jhMikeSactually the button routine
14:56:15amiconn...when playing music, and also in mpegplayer
14:56:24amiconnThe latter didn't happen with the old threading
14:57:01jhMikeShappened like crazy here on 32 bit...totally nasty
14:57:41amiconnYes, it's rather nasty, especially on a laptop
14:58:21amiconnmpegplayer needed <50% cpu on my laptop before
14:59:29 Join funky_ [0] (
14:59:40 Nick funky_ is now known as kubrick (
14:59:44 Nick kubrick is now known as funky (
14:59:47oboI don't get any audio delay, but otherwise acts the same
14:59:49jhMikeSodd, it runs so nice and smooth here all the time now. I hope nothing is trying to access sim_read/write from a non-rockbox thread somewhere.
15:01:08jhMikeSfull rebuild/reinstall?
15:01:26amiconnjust make && make install
15:01:37 Join miepchen^schlaf [0] (
15:01:57jhMikeSdoes it run normally after buffering is complete?
15:02:47amiconnThe delay is even longer than I thought first - around 30 seconds
15:02:59amiconnAnd I can't say it runs normally
15:03:16amiconnThe sim doesn't react to close attempts
15:03:19amiconnHave to kill it
15:03:24amiconn(on linux)
15:03:28decayedcell_um is it just me or is the 5G iPod still freezing when you plug in USB
15:03:35jhMikeSthat issue I might expect for closing
15:04:18 Quit Febs (Read error: 110 (Connection timed out))
15:04:23jhMikeSwhat happens after buffering is done though?
15:05:06amiconnIt plays music, until the next track change/skip. Then it crashes like before
15:05:18amiconnBut this time, the app doesn't exit on crash(!)
15:06:00Soapdecayedcell_: when are you plugging in USB? During boot or after? For if you are plugging it in after Rockbox is loaded, I'm not sure what to say as on my 60GB 5th gen I get perfect detection and reboots into Apple's disk mode.
15:06:02 Join bluey- [0] (
15:06:06amiconnIt just closes the gui window, prints 'Xlib: unexpected async reply (sequence 0x73)' to the console, then hangs in te background
15:07:14amiconnSoap, decayedcell_: It's a known problem affection all ipods. It depends on the individual build for some reason. One tiny change in totally usb-unrelated code can make it work or break it
15:08:22oboin a 64bit sim, cpu load drops for me when not in the default WPS - also the peak meter activity seems totally unrelated to the audio
15:08:23jhMikeSfor one thing, gui_message_loop was posting to queues and that thread is not a rockbox thread but this in effect has to be treated like an ISR.
15:08:33amiconnjhMikeS: 20 seconds this time. And indeed it only starts playing after buffering is complete (which takes way longer than before)
15:08:53jhMikeSaudio sync was never very tight with the sim
15:09:05obono audio delay, audio is playing as it buffers (tried ogg, mp3 and flac)
15:09:25amiconnOops, and a simple screendump makes the sim freeze now
15:09:35jhMikeSI know about that and why that happens
15:10:58decayedcell_Soap after it has loaded and sitting in root menu I plug in USB and it has freezed twice now
15:11:24SoapThank you for the information, amiconn. I so very rarely use the featuere, it must be coincidence/good luck I haven't had a "hang" on USB insertion for months and months.
15:11:29amiconnjhMikeS: Funny enough, the screendump itself seems to work
15:11:41Soapfeature even.
15:11:57 Quit spiorf (Read error: 104 (Connection reset by peer))
15:12:20amiconnSoap: Eh? How do you update builds, upload music etc?
15:12:35jhMikeSamiconn: it's because a non-rockbox thread is forbidden to use the sim_xxx stuff now and the button "driver" calls screen_dump
15:12:40 Join spiorf [0] (n=spiorf@
15:12:42Soapamiconn: shutdown and reboot into original firmware.
15:12:54amiconnUh, why that?
15:13:02amiconnTakes much too long for my taste...
15:13:15Soapdo most my updating in the mornings before work, so the iPod has been off all night. I just toggle the hold switch and plug it into the computer.
15:13:19jhMikeSbecause it uses kernel.c mutexes as well.
15:13:23amiconn(that is, if the detection works as intended)
15:14:07jhMikeSthings are quite literally like on target. you wouldn't call screen_dump from an interrupt handler, eh?
15:14:35decayedcell_although amiconn IIRC it was working fine before the last change to the USB code, I can't pinpoint it because the subversion history has changed on the main page
15:14:56amiconnThere is no version to pinpoit
15:14:57decayedcell_anyway going to sleep now
15:15:22amiconnIt happens ever since usb detection learned the difference between usb port and usb charger
15:15:22 Quit decayedcell_ ("CGI:IRC (EOF)")
15:15:23jhMikeSI might have to settle down the activity a wee bit so the audio thread gets cpu time.
15:15:34amiconnBut not all builds show that problem
15:16:27 Join Zom [0] (
15:16:48amiconnjhMikeS: That's why the old threading called SDL_delay(1); every few yields...
15:17:56jhMikeSsince it's IO related, it's better to place it in the place where disk activity is occuring instead of all the time
15:18:00amiconnWould it be possible to implement the equivalent of cpu sleep in the sim?
15:18:40jhMikeSI could probably implement all the core features actually
15:19:37amiconnI mean core_sleep()
15:19:48jhMikeSthough, the rockbox threads really do block now except during the I/O
15:20:15amiconnThen why does the sim draw all available cpu power?
15:20:19jhMikeSIt should effectively exist now
15:20:23jhMikeSamiconn: all the time?
15:21:02amiconnHmm, no
15:21:05 Nick jack_ is now known as chrisjs169 (
15:21:16amiconnBut it does in a peakmetered wps as well as in mpegplayer
15:21:38amiconnA peakmetered wps also caused high load with th eold threading, but mpegplayer didn't
15:22:23jhMikeSbecause mpegplayer furiously spins when waiting for the frame presentation. I get what's going on totally now.
15:23:06amiconnHmm, I would think it should sleep when waiting...
15:23:50amiconnI wonders what's up with the peakmeters - they are throttled in the code (20fps iirc)
15:24:26 Join tchan [0] (n=tchan@lunar-linux/developer/tchan)
15:24:37amiconnHmm, and without peakmeters, the linux/amd64 sim doesn't need 20+ seconds for buffering...
15:25:04amiconnSo it's actually a problem of the rockbox threads drawing all available cpu power, leaving nothing for the i/o therad
15:25:54amiconnSo there are actually 2 things to solve imho. (1) make the rockbox threads never draw _all_ available power (by adding an occasional SDL_delay(1) as before
15:26:23amiconn(2) fix mpegplayer (and peakmeters?) to wait properly instead of spinning. Might also solve peakmeter problems on target...
15:26:39jhMikeSobo: is that a dual core processor?
15:27:21amiconn(2) doesn't have to do with your threading changes, but became evident because of these
15:27:23jhMikeSI wonder if the native scheduler actually queues threads when they're running on different CPUs
15:27:54amiconnMine are all single cores (laptop is a 1st generation Pentium M, linux box a single core Athlon 64)
15:29:03jhMikeSgiven that only one thread may run at a time, all the rest are blocked, so not running. but one thread could run furiously.
15:29:47 Join My_Sic [0] (
15:31:58jhMikeSthe IO thread will likely be blocked at some point while performing IO internally leaving an active rockbox thread just running and running. same for other yield spins.
15:32:49jhMikeSmpegplayer on target shouldn't delay. 1ms resolution isn't so good for smooth video on target.
15:33:24jhMikeSbut we can't even be that precise there...hmmm.
15:33:46 Part bluey- ("Leaving")
15:34:10jhMikeSdo peakmeters spin? I haven't looked.
15:35:01amiconnI don't think that a 10ms granularity will be visible. The lcd panel itself smoothes things out
15:35:20amiconnI don't remember, but it looks like they do
15:35:25amiconngtg now
15:36:26jhMikeSyep, they do and sleep(0) isn't end of tick based either in the sim
15:44:02 Nick fxb is now known as fxb__ (
15:54:23 Join |desowin| [0] (
15:54:33 Quit |desowin| (Remote closed the connection)
15:55:00 Quit desowin (Read error: 113 (No route to host))
15:58:02 Join hannesd [0] (
16:04:39 Quit Lear ("ChatZilla [Firefox]")
16:11:03 Quit JdGordon ("Konversation terminated!")
16:22:54 Quit iamben (Read error: 104 (Connection reset by peer))
16:23:29 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
16:23:56 Join BigBambi [0] (
16:25:15 Join ddalton [0] (
16:25:36 Join barrywardell [0] (
16:28:39 Join iamben [0] (
16:29:39 Quit ddalton (Client Quit)
16:34:11 Join desowin [0] (
16:37:46 Quit linuxstb (Read error: 110 (Connection timed out))
16:42:02 Join Buschel [0] (
16:52:35 Quit petur ("switching")
16:52:45 Join petur [0] (n=petur@rockbox/developer/petur)
16:53:05***Saving seen data "./dancer.seen"
16:58:32 Quit BigBambi (Remote closed the connection)
17:03:19 Quit barrywardell ()
17:03:27 Join Kingstone [0] (n=king@
17:10:17 Join datachild [0] (
17:15:26 Join barrywardell [0] (
17:16:25 Quit tchan (SendQ exceeded)
17:17:55 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
17:18:03 Join tchan [0] (n=tchan@lunar-linux/developer/tchan)
17:18:58 Join advcomp2019_ [0] (n=advcomp2@
17:19:08 Quit advcomp2019 (Nick collision from services.)
17:19:18 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@
17:21:16 Join frimpy [0] (
17:23:29 Quit Thundercloud (Remote closed the connection)
17:24:13 Join Thundercloud [0] (
17:27:17 Join aliask [0] (
17:34:08Buschelquestion regarding scrollwhell-acceleration: in list.c (gui_synclist_do_button(...)) there are two methods for scroll-acceleration. 1st is done when HAVE_SCROLLWHEEL is defined, the other is used if this is not defined. For iPODs this define is not used, the other acceleration method does not work for iPODs.
17:34:29Buschelwhy is HAVE_SCROLLWHELL not defined for iPODs?
17:35:17oboBuschel: because button-clickwheel needs to be updated to use the HAVE_SCROLLWHEEL api
17:35:28SoapAs I understand is because the iPod's scrollwheel driver is not as advanced as the Sansa's.
17:35:54 Join linkmaste [0] (
17:36:00Buschelhmm, what is button-clickwheel? some software component?
17:36:10linkmastedoes rockbox have any firmware for the Creative Zen MicroPhoto?
17:36:25oboit's the rockbox driver for the wheel on ipod 4gs and up
17:36:41Soaplinkmaste: No, Rockbox does not have a current port for the Creative Zen series of players.
17:36:42obolinkmaste: no - only devices listed on the frontpage of
17:36:49 Quit linkmaste (Remote closed the connection)
17:36:52 Quit Genre9mp3 ()
17:38:40 Join My_Sic [0] (
17:38:53 Quit My_Sic (Read error: 104 (Connection reset by peer))
17:39:06 Join My_Sic [0] (
17:39:35 Join Presence [0] (
17:39:37Buschelobo: do you know what the function button_apply_acceleration() does? can this be somehow connected to the iPODs get_ipod_scroll_wheel_clicks_per_second()-function?
17:40:41 Quit hcs ("Leaving.")
17:43:54 Join BigBambi [0] (
17:44:43oboBuschel: have you read the commit message for r13959?
17:48:48Buscheljust read it. does it mean I just need to set an appropriate value for WHEEL_ACCELERATION_FACTOR?
17:50:26oboit means the ipod wheel driver needs to be updated :)
17:51:08*jhMikeS apologizes for all the trouble he causes ... and yes, appropriate things need to be defined.
17:52:10obojhMikeS: the wheel on the sansa: from what I understand it only reports direction?
17:52:25 Part frimpy ("Leaving")
17:52:44jhMikeSobo: basically. no touch indication.
17:53:07obohow many clicks does it do per rotation?
17:53:56jhMikeSQuite a number. I've never counted them. It might be around 64
17:55:01*obo will have another go at getting his head around this :)
17:55:16 Join Shane_e250r [0] (i=d08a1da3@gateway/web/cgi-irc/
17:56:11jhMikeSI counted 26 list moves for moving once around with my thumb but it slips so I doublt that's it. since it's slow speed, 52 click would have happened.
17:56:51*jhMikeS gets more precise using liquid paper
17:56:52Shane_e250rI would like to edit the information about the sansa e200r on the Wiki could someone please grand me access.
17:59:28barrywardellwhat's your username?
18:00:10SoapTeenage Suicide...
18:01:07barrywardellShane_e250r: access granted
18:01:33Shane_e250rThank you
18:01:40SoapHaven't you seen the movie 'Heathers'?
18:01:42jhMikeSobo: what's other info do you need about HAVE_SCROLLWHEEL?
18:04:35 Quit aliask ("ChatZilla [Firefox]")
18:06:03obojhMikeS: I don't know yet, need to concentrate on it - will shout later though if needed
18:07:15*jhMikeS would have thought there'd have been a big clamor for scrollwheel settings but noone seems to be asking for them :\
18:09:08jhMikeSdid I miss a flyspray post or something?
18:12:20 Quit gromit` (Remote closed the connection)
18:12:40 Join gromit` [0] (
18:14:02 Join SliMM [0] (n=chatzill@
18:14:23SliMMhey, i have a veri big problem with rbutil qt
18:14:33SliMMvery* lol
18:14:51n1sSliMM: tell us about it
18:15:01SliMMit says 'no ipod found' when trying to install the bootloader
18:15:10n1sSliMM: which OS?
18:15:17SliMMwin xp sp2
18:15:20 Quit Shane_e250r ("CGI:IRC (Ping timeout)")
18:15:30n1sDo you have admin privileges
18:15:34SliMMthe autodetection works great
18:15:39SliMMyeah, i do
18:15:56n1sdoes regular ipodpatcher work?
18:16:00iambendo you have an ipod?
18:16:37SliMMi didn't try ipodpatcher, but it detects my ipod well in rbutil
18:16:52SliMMiamben: haha >:P
18:17:12n1swell, could you try with ipodpatcher?
18:17:51SliMMyeah, right away
18:17:55SoapSliMM: the current build of RButilQT does the same to me on W2K SP4 with admin priv.
18:18:01Soapipodpatcher has no problem.
18:18:31SliMMoh, that's great news, 'cause i have updated the apple firmware
18:18:47SliMMand it apparently erased rolo
18:19:21SliMMi like the new home page layout
18:20:25SliMMwhat's with the usb stack and other things going on lately?
18:20:59iambenthe new homepage makes it harder to follow changes
18:21:14 Join e-ric [0] (
18:21:57*iamben changes his rb bookmark
18:22:01SoapSliMM: The USB stack is one of the GSoC projects. An attempt to bring a software USB stack to Rockbox (PP targets only?). It is not finished, and not in use even though it is in SVN.
18:22:52SliMMSoap: what's the use of the USB stack?
18:23:59*SliMM is enjoing Rockbox again, for ipodpatcher worked
18:24:25SliMMdid i miss an 'y'?
18:24:30peturusb connectivity from within rockbox?
18:24:41SoapSliMM: The use of the USB stack is that Rockbox can do its own USB. Ya know...file transfers for a start w/o rebooting into original firmware.
18:24:47 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
18:25:00 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
18:25:19SliMMah, great :) maybe usb otg later *grin*
18:25:43e-richi, i just upgraded from 2.5 to current build on archos recorder v1, and i wonder what happened to the "poweroff disk" option. it was quite useful to tweak either for battery time or smaller gaps between songs.
18:25:56barrywardellpetur: what's the current status of the USB stack?
18:27:20 Quit BigMac ("Konversation terminated!")
18:27:21linuxstbe-ric: I have a feeling it's now always enabled. It should never have affected gaps between songs though.
18:27:46 Join webguest74 [0] (i=44769e2b@gateway/web/cgi-irc/
18:27:58 Quit webguest74 (Client Quit)
18:28:02e-ricis there anything else that affects gaps?
18:28:15 Join Febs [0] (n=chatzill@rockbox/administrator/Febs)
18:28:16linuxstbOnly how the files were encoded.
18:28:19SliMMi have another question: don't you think that in 'repeat one' mode, the 'next' and 'prev' buttons should move to the next and prev song rather than to the current one?
18:28:52 Join miepchen^schlaf [0] (
18:29:04linuxstbe-ric: I assume you're talking about a normal transition - i.e. when tracks are playing as normal, rather than skipping manually?
18:29:10 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
18:29:41e-ricno i meant skipping normally
18:29:43 Join Ebert [0] (
18:30:59linuxstbAh, then yes, it's possible the power-off setting can mean the disk takes longer to spin up, and if the next track isn't already buffered, will cause a longer delay.
18:31:52SliMMthat wasn't a technical question, anyone could answer it
18:32:38 Quit midkay ("Leaving")
18:32:52linuxstbe-ric: It seems nls made that change - on 23 January this year. Maybe the IRC logs from around that can explain the decision.
18:34:17linuxstbHere's the commit -
18:36:04barrywardelllinuxstb, e-ric: I think most of the delay is in spinning up the disk, not powering on
18:36:37linuxstbbarrywardell: Yes, "most"....
18:36:53e-ricits still a noticeable difference
18:37:16barrywardelli meant most as in unnoticably different
18:38:18barrywardelle-ric: how noticable?
18:38:25 Quit iamben (Read error: 104 (Connection reset by peer))
18:38:32 Join iamben [0] (
18:41:23 Join The-Compiler [0] (
18:44:13 Join saratoga [0] (i=41becb3b@gateway/web/cgi-irc/
18:45:28e-ricabout a second
18:46:51e-ricits about 3 seconds from standby vs 4 seconds from poweroff
18:48:11 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
18:49:08 Part n1s
18:50:42e-ricanother question, database - gather runtime data and On-Left On-Right in the WPS doesnt work for me, are that limitations on the recorder v1?
18:52:19e-rici initialised database but enabling the "gather runtimedata" doesnt have an effect, nothing in history and cant add ratings.
18:53:08***Saving seen data "./dancer.seen"
18:54:09e-ricin the WPS i can skip to next (or random) dir when i skip on the last file, but not with ON-Right
18:55:17barrywardelle-ric: 1 second sounds like a lot. I don't think it should make that mch of a difference
18:55:30 Quit HellDragon (Read error: 104 (Connection reset by peer))
18:55:37 Join nichos [0] (
18:55:48 Join HellDragon [0] (n=Nocebo@unaffiliated/helldragon)
18:56:40nichoshi all, I'm trying to get rb installed on a ipod photo. I had the loader working, but the filesystem was HFS. So I made it FAT32, but now when I run ipodpatcher it says there's no ipod found. It's in disk mode, and i've unmounted it.
18:56:46nichosusing linux, any ideas? Thanks.
18:58:25krazykitthe easiest option is to find someone with windows and use itunes to fix it. however, there's a manual way to do it in linux, lemme find the link
18:59:02nichoskrazykit i've been following those instructions, but I re-did it on the mac
18:59:08nichosdisk util says it's a windows ipod
18:59:15nichoser, sorry, itunes says it's a windows ipod
19:00:05krazykitand itunes on the mac will happily transfer files?
19:00:55nichoslet me try, i didn't test
19:01:02nichosfiles files, or music files?
19:01:10krazykitmusic files
19:01:23krazykitjust to confirm that somethings not funny with the ipod itself
19:01:29 Quit linuxstb (Read error: 110 (Connection timed out))
19:02:36nichositunes says it is...let me disconnect the ipod and see if it's playable
19:03:06 Quit ompaul (Remote closed the connection) showed the folder w/the exclamation mark, and rebooted
19:04:13 Join linuxstb__ [0] (
19:04:55krazykitnichos, i'd try the stuff on again. make sure you've downloaded the correct mbr and such
19:06:20nichosok, i just did that and it's working. I'm screwing up somewhere converting to fat
19:06:23nichosthanks krazykit
19:07:15krazykitlike it says, mkfs.vfat -F 32. i think mkfs.vfat defaults to fat16 on most systems
19:07:40nichosi know i did that. the only thing i was 100% sure of was the /dev/disk#####
19:07:54nichosthat is, it shows up as /dev/disk2s2
19:08:13krazykitare you running a bsd?
19:08:20nichoskinda, mac;)
19:08:23 Join BigMac [0] (
19:08:25krazykitoh, right, right.
19:08:34krazykitthey do /dev/ differently :P
19:08:34nichosso when i run mkfs, do i disk2, or disk2s2
19:08:44nichosyeah, it didn't show up on linux like that:)
19:08:49krazykitwhichever is the partition, which i think is disk2s2
19:09:08nichosahhh, ok - that's probably where i lost it:)
19:10:13barrywardellit should be disk2s2
19:10:33 Join SiM_99 [0] (
19:11:04barrywardellthat corresponds to the data partition
19:11:18nichosohhhh, yeah, i didn't see that on linux
19:11:19barrywardelldisk2 is the whole disk
19:12:24 Quit linuxstb_ (Read error: 110 (Connection timed out))
19:13:32barrywardellso use /dev/disk2 in place of /dev/sdx, /dev/disk2s1 in place of /dev/sdx1 and /dev/disk2s2 in place of /dev/sdx2
19:15:19nichosthanks - i'll redo it with that
19:15:31nichosthat's pretty lame the mac version doesn't all to set up as fat
19:16:59SliMMis it difficult to add a menu option or a button combination to enable 'poket mode' or something like that?
19:17:13SliMMto disable most of the controls and the screen
19:18:51 Quit Nico_P (Remote closed the connection)
19:21:10 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
19:21:22e-ricSliMM: theres a button combination for hold (lock buttons), it doesnt disable the screen (at least on archos) tho
19:25:17SliMMe-ric: hold disables phisically (i think) all the controls, but a 'pocket mode' option should leve the pause, forward and rewind buttons enabled
19:25:54 Quit iamben (Read error: 104 (Connection reset by peer))
19:28:30 Quit SiM99 (Read error: 110 (Connection timed out))
19:30:41 Quit hannesd (Connection timed out)
19:31:34 Join iamben [0] (
19:32:49 Quit linuxstb__ (Read error: 113 (No route to host))
19:34:36 Join hannesd [0] (
19:39:51 Quit XavierGr (Nick collision from services.)
19:39:54 Join XavierGr [0] (
19:39:56barrywardelldoes anybody know what the status on the usb stack is?
19:40:26 Quit hannesd ("Client suicide")
19:42:34Soapbeyond "in progress"?
19:42:49SoapIt does the PC connection - but no file transfer IIRC.
19:44:33barrywardellis austriancoder still actively working on it?
19:45:47 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
19:49:24 Quit amiconn (Nick collision from services.)
19:49:31 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
19:50:59 Part hcs
19:51:06 Join ompaul [0] (n=ompaul@freenode/staff/gnewsense.ompaul)
19:55:11 Quit saratoga ("CGI:IRC")
19:58:33peturit already responds correctly to several inquiries but there is still work...
19:58:57 Quit perplexity (Read error: 113 (No route to host))
19:59:15peturI admit I haven't given it much attention recently, and this weekend isn't much better...
20:06:20 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
20:06:44barrywardellpetur: thanks. I think I'll just try it when I get my H10 or Sansa from work on Monday
20:07:00 Join RudMan [0] (
20:10:14e-ricany idea why skip dir with ON+Right or database -> gather runtime data does not work on my archos v1 rec?
20:13:59 Quit The-Compiler (Remote closed the connection)
20:23:46 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
20:24:14 Join miepchen^schlaf [0] (
20:24:47 Quit BigBambi (Read error: 110 (Connection timed out))
20:26:39 Quit barrywardell ()
20:26:49peturamiconn: my sim still crashes here when playing mp3, ubuntu x86. This is a backtrace:
20:27:05 Join eigma [0] (
20:27:53peturI am however very new to gdb & co
20:28:36peturit looks like it is a crash in SDL
20:33:25 Join SiM99 [0] (
20:38:52 Join webguest21 [0] (i=474fc6da@gateway/web/cgi-irc/
20:39:19 Quit webguest21 (Client Quit)
20:40:48 Quit Agm3nt ()
20:46:41 Quit iamben (Read error: 104 (Connection reset by peer))
20:49:29 Join linuxstb__ [0] (
20:50:53 Quit SiM_99 (Read error: 110 (Connection timed out))
20:50:55 Nick linuxstb__ is now known as linuxstb (
20:52:47 Quit SliMM (Read error: 104 (Connection reset by peer))
20:53:11***Saving seen data "./dancer.seen"
20:54:46 Nick fxb__ is now known as fxb (
20:58:59 Join barrywardell [0] (
21:01:55 Quit Ebert ()
21:03:46 Quit linuxstb_ (Read error: 110 (Connection timed out))
21:11:11 Join iamben [0] (
21:13:16 Join perplexity [0] (
21:18:35 Join webguest82 [0] (i=0cafe63b@gateway/web/cgi-irc/
21:18:46 Quit webguest82 (Client Quit)
21:21:13 Quit n17ikh|Lappy ()
21:22:11 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
21:28:32 Quit RudMan (Read error: 110 (Connection timed out))
21:30:34 Join thegeek [0] (
21:32:37 Nick fxb is now known as fxb__ (
21:33:52 Quit ender` (Read error: 104 (Connection reset by peer))
21:34:18 Join ender` [0] (
21:34:32 Quit billytwowilly (Remote closed the connection)
21:34:52 Join n17ikh|Lappy [0] (
21:35:51 Quit linuxstb (Read error: 110 (Connection timed out))
21:35:53 Join Ebert [0] (
21:38:59 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
21:42:24maxkelleyI'm having trouble building the ui simulator, because adc-target.h is missing. I know this happened previously with the e200 build, so how was it fixed? I'm using c200.
21:46:32linuxstbDid you download the source .tar.bz2 or via SVN?
21:46:35 Join billytwowilly [0] (
21:46:45maxkelleysource tar bz.
21:46:58maxkelleywell, bleeding.
21:47:07linuxstbIt's possible that file is missing then... That happens occasionally...
21:47:18maxkelleydo all builds need it?
21:47:27linuxstbEach build will have their own adc-target.h
21:47:56maxkelleyI don't see one for the e200.
21:48:56maxkelleyI think that mark arigo, who made the c200 build, added it accidentally in the dependency list.
21:49:05maxkelleyoh, or not.
21:49:59linuxstbI've just updated the FILES list that is used to create the .tar.bz2 - the next bleeding edge file should be fine.
21:50:21maxkelleyif I download it now, it should be fine?
21:50:23 Quit billytwowilly (Remote closed the connection)
21:50:44maxkelleyactually, to make it easier, what's the svn server?
21:50:53linuxstbYou need to wait until the build has finished here -
21:51:07linuxstbInstructions for SVN are at
21:53:38maxkelleynow it's working..
21:53:40 Join e-ric_ [0] (
21:53:44maxkelleythanks :)
21:53:49linuxstbSVN or the download?
21:54:28maxkelleyI used SVN.
21:54:41linuxstbOK. Always easier...
21:55:00maxkelleyI wanted to use it, but was too lazy to find out the SVN server.
21:55:24maxkelleyunsupported keypad, eh.
21:55:49linuxstbYes, low_light mentioned there is still some work to do on the sim...
21:56:15maxkelleyI emailed him, he told me that I should build the sim and tweak the plugins to work with the screen of the c200.
21:58:07maxkelleyoh, that's for the keypad for the plugins.
21:58:29maxkelleywhich are disabled by default, but to test them you need them enabled :P
21:58:55 Quit barrywardell ()
22:02:01Buschelobo, still there?
22:05:00Buschelobo: i've tried to merge the changes in sansa-bscrollwheel stuff to iPOD scrollwheel, and I#ve got problöems with the result :)
22:05:14Buschelahem, several typos
22:05:51 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
22:05:53Buschelobo: result is that the scrolling only happens at all, if I am rotating very fast
22:06:03 Join miepchen^schlaf [0] (
22:06:07 Join Genre9mp3 [0] (
22:06:12Buscheldo you know how to parametrize this stuff correctly?
22:06:32obotried adjusting the WHEEL_ACCELERATION_FACTOR?
22:07:01maxkelleyYAY, FIXED IT.
22:07:13Buschelnot yet, I am not sure if I also have to adjust some other constants in the scrollwhell stuff itself
22:07:37 Join billytwowilly [0] (
22:07:49 Join Zagor [0] (
22:09:28 Quit e-ric (Read error: 110 (Connection timed out))
22:09:46maxkelleylinuxstb: how do I commit a change? (I know the command, but how do I log in?)
22:10:19 Quit BigMac (Read error: 110 (Connection timed out))
22:11:06Buschelodo: now trying to enlarge WHEEL_ACCELERATION_FACTOR by a factor of 10
22:13:28linuxstbmaxkelley: You can't unless you've been offered SVN write access. Use "svn diff > mypatch.diff" and post a patch to the patch tracker.
22:14:09maxkelleyI figured that.. thanks.
22:18:52 Join BigMac [0] (
22:18:53 Quit ender` (Read error: 104 (Connection reset by peer))
22:19:14 Join ender` [0] (
22:20:34 Quit billytwowilly (Remote closed the connection)
22:20:45 Quit petur ("gone taping :)")
22:22:46maxkelleywhat's error 23 mean on flyspray? I registered, got the confirm code, etc.
22:23:38iambensounds like a bug, you should report that =)
22:23:41maxkelleyah, never mind.
22:23:43maxkelleyhehe :)
22:23:58linuxstbflyspray gives odd errors sometimes, but most of the time it's worked anyway...
22:26:52 Join qweru [0] (
22:28:15 Join billytwowilly [0] (
22:29:14maxkelleyYeah, it didn't process my confirmation.
22:30:04Zagorlinuxstb: had time to look at sansapatcher yet?
22:30:20ZagorI'm itching to install :)
22:31:03BagderZagor: not tried the e200tool way?
22:31:09 Quit jhulst ("Konversation terminated!")
22:31:26Mode"#rockbox -o Bagder " by Bagder (n=daniel@rockbox/developer/bagder)
22:31:27ZagorBagder: umm, no.
22:32:15maxkelleywhee, my first patch :)
22:32:17Buschelodo: no difference. seems like the non-accelerated scrolling is too slow −− trying another thing
22:32:22Bagderdespite its name, e200tool works fine on the c200 too
22:32:38linuxstbZagor: Sorry, not yet. I'm currently doing things with ipodpatcher....
22:33:19Bagderaha another c200 user
22:33:47maxkelleyBagder: on the topic of c200/e200tool, my c200 starts with a black screen, so I reboot into manufacturing mode and upload the pribootloader, then it reboots, and goes into the firmware.
22:34:04maxkelleyit gets stuck on the "connected" screen, however. and when I restart, it's back at the black screen.
22:34:12maxkelleyit's almost like it's forgetting its bootloader.
22:34:33maxkelleyhave you seen this happen?
22:34:44linuxstbIIUC, all e200tool does (in manu mode) is transfer an application to RAM and run it.
22:34:47Bagderno, but I don't have a c200
22:35:05linuxstbSo if you transfer the bootloader, all it's doing is running the bootloader.
22:35:08Bagderlinuxstb: at least when used the way we tell users for installing...
22:35:30maxkelleywell, I run e200tool recover priBootloader.rom
22:35:36ZagorI'm a sansa newbie. which installation instruction do I use for e200tool on c200?
22:35:57maxkelleyZagor: the c200 isn't really supported at all.. basically.
22:36:33BagderZagor: #1 boot in manufacture mode, #2 e200tool recover priBootloader.rom, #3 goes into recovery mode, put your rockbox bootloader mi4 there #4 disconnect
22:36:53Zagormaxkelley: yeah I know. I'm not a rockbox newbie (possibly the least newbie of all :-)
22:37:03 Quit Kingstone ()
22:37:04Bagderafaik, the bootloader cannot yet run the OF...
22:37:10maxkelleyah, ok, sorry :)
22:37:52 Join ender1 [0] (
22:45:43 Quit ompaul ("take it easy")
22:51:19 Quit desowin ("use linux")
22:51:25 Join low_light [0] (i=4cbdfa51@gateway/web/cgi-irc/
22:53:13low_lightZagor: make a rockbox c200 build, boot the OF & attach the usb, unzip to the c200
22:53:15***Saving seen data "./dancer.seen"
22:54:09low_lightZagor: then make a c200 bootloader. Enter recovery mode on the c200 (hold on + rec + insert usb). copy the firmware.mi4 file to the device. unmount and disconnect.
22:54:23 Join My_Sic [0] (
22:54:37low_lightRockbox should now boot
22:54:58Bagderah right, you can go straight into recovery mode
22:55:17 Join Rincewind [0] (
22:55:23 Quit ender` (Read error: 113 (No route to host))
22:55:34Zagorlow_light: neato, I'll get cracking
22:55:40low_lightYou will need to put an original firmware.mi4 file back (using recovery mode) if you need to transfer anything throught usb
22:55:48linuxstbSo the regular firmware upgrade process doesn't work on the c200? (putting files in the root)?
22:56:09low_lightUnfortunately I haven't been able to get dual boot working.
22:57:11low_lightlinuxstb: I don't think so. I think I tried, but don't really remember. Perhaps Zagor should try that.
22:57:46amiconnZagor: Somehow I think that the commit msg would be a more interesting thing to put in the frontpage table than the files changed
22:57:47linuxstbThere are some (brief) instructions in this forum thread -
22:57:47Zagoraha, there's no usb in c200 yet?
22:57:57low_lightI've always used recovery mode.
22:58:08BagderZagor: right, no usb in any PP target
22:58:20amiconnBagder: Did you see my suggestion for some additional info in build table tooltips in y'day's log?
22:58:48BagderI saw you wanting info, yes, but I didn't see what/how
22:59:08Bagdermaybe I was a bit too quick
22:59:10low_lightlinuxstb: I see. Perhaps that works. Someone can try.
22:59:42 Join barrywardell [0] (
22:59:47low_lightBut you'll still need to use recovery mode to put the OF back in order to use usb.
23:00:13amiconnFor tracking both "special" warnings and the oscillating build sizes, I'd want a tooltip to show which individual build server built that build, plus the gcc version used (host or cross-gcc)
23:00:39 Join agm3nt [0] (
23:02:26amiconnIdeally it would also show the host architecture
23:03:05amiconnRight now I don't know which build servers are 64 bit, and which are 32 bit
23:03:17 Join ptw419 [0] (
23:03:36Bagdernobody does
23:03:59amiconnThe build scripts could execute uname -a
23:04:00Bagderthe build system needs to figure that out and display it
23:04:07 Part low_light
23:04:29Bagderuname -a isn't good enough I believe
23:04:53Bagderin the 64bit kernel/32 bit userspace cases
23:07:29 Join frafra [0] (
23:07:34frafrahi all
23:07:46frafraI've a problem
23:09:06frafraI'm a linux user, and my ipod nano is totally broken. how to reset completely it?
23:09:37 Quit barrywardell ()
23:14:58 Join k-man [0] (n=jason@unaffiliated/k-man)
23:15:25maxkelleyAh, hell, I missed low_light.
23:16:14 Quit nichos (Read error: 104 (Connection reset by peer))
23:20:38 Quit Rincewind ("Cya")
23:21:15Bagdertry the build table hover how
23:21:28Bagdergcc version is there
23:21:56amiconnAh, nice
23:22:29amiconnI hope adding the server name is also possible?
23:22:39Bagderyes, working on that next
23:23:53amiconnWho would run 64 bit kernel + 32 bit userspace? Is there any advantage in doing that?
23:24:12amiconnSounds like unnecessary complication to me...
23:24:21Bagderyou get the advantage of 64bits in the kernel, but no 64 bit pains in userspace
23:24:27Bagderand its not a complication at all
23:24:55BagderI have 32bit userspace, I can just boot a 64bit kernel with it
23:25:13Bagderand I do occasionally
23:25:23amiconnHmm, I can't see the advantage
23:25:48Bagderwell, that's beside the point though
23:25:54maxkelleyare plugins allowed to use the volume keys, or will that conflict with music playing in the background?
23:26:34zei have a primarily 64-bit system but with 32bit userspace compat
23:26:55Bagderbuild server add too now
23:27:09zethe latter mainly just for binary-only junk and the increasingly rare 64bit-unsafe opensource thing
23:27:14amiconnI just run 64 bit because I can (64 bit all the way without 32 bit compatibility)
23:28:34 Quit frafra ("Ex-Chat")
23:29:09 Join jhulst [0] (
23:29:46Bagderok, I'll install this new showbuilds version for real now, seems to work
23:31:23amiconnhmm, interesting....
23:31:27amiconnSlasheri: around?
23:31:42Bagdermy point was only that uname -a shows the kernel 64/32 bit, not the userspace situation
23:33:00amiconnI didn't think that someone might run a 64/32 bit combination doesn't run the recommended arm-elf-gcc
23:38:08 Join ender` [0] (
23:38:56amiconn...and it looks like that server is the reason for the oscillating arm target builds least for the larger oscillations
23:40:03amiconnIt looks like we should also log the binutils version...
23:41:05amiconnSome oscillations can't be caused by gcc version. Either it's host arch, or binutils
23:41:21BagderI'll add binutils output now
23:53:23 Join rasher_ [0] (n=rasher@rockbox/developer/rasher)
23:54:03 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
23:55:27maxkelleyall files should be in SVN, right?
23:55:53maxkelleyor is there a files list for that as well, like the bleeding builds one?
23:56:10maxkelleybecause I'm missing some files for the flipit plugin.
23:56:22 Quit ender1 (Read error: 110 (Connection timed out))

Previous day | Next day