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 2005-05-31

00:09:47 Join austriancoder [0] (~austrianc@
00:10:02austriancoderhi all
00:16:44 Quit _aLF ("Leaving")
00:20:10rasherstrange.. running the windows sim under wine, I can't hit "."
00:34:04austriancoderi cant build the x11 sim :(
00:37:02rasherI just did, what's wrong?
00:38:34austriancoderah.. now it workd
00:39:23 Quit rasher ("CGI:IRC 0.5.4 (2004/01/29)")
00:39:33austriancoderhmm.. ok got one
00:39:41 Join rashums [0] (
00:40:43 Nick rashums is now known as rasher (
00:41:11 Join TCK [0] (
00:41:18 Join Dave99 [0] (
00:42:52austriancoderOT: does anybody knows a site with romatic sayings
00:43:03 Join webguest00 [0] (
00:43:31 Quit webguest00 (Client Quit)
00:49:31 Quit matsl ("Leaving")
00:50:23 Quit ehntoo ("Leaving")
00:52:42 Part Dave99
00:59:31austriancoderis enum part of c or is it part of c++?
01:04:06 Join ehntoo [0] (
01:16:49***Saving seen data "./dancer.seen"
01:24:03austriancodertree or double linked list for a gui?
01:48:08 Join ashridah [0] (
02:10:35 Quit preglow ("leaving")
02:14:25 Quit ehntoo (Remote closed the connection)
02:30:34 Join StrathAFK [0] (
02:37:05 Quit Strath (Read error: 110 (Connection timed out))
03:02:54 Join RED_M_CHIU [0] (
03:03:08 Quit RED_M_CHIU (Client Quit)
03:04:14 Join RED_M_CHIU [0] (
03:16:50***Saving seen data "./dancer.seen"
03:23:23 Join Adity1 [0] (
03:23:24 Quit Aditya (Read error: 131 (Connection reset by peer))
03:26:55 Join amiconn_ [0] (
03:35:27 Quit amiconn (Read error: 60 (Operation timed out))
03:35:28 Nick amiconn_ is now known as amiconn (
03:51:24 Join DMJC [0] (
03:56:09 Quit cYmen ("zZz")
04:05:28 Join QT_ [0] (as@area51.users.madwifi)
04:18:43 Quit thegeek (Read error: 110 (Connection timed out))
04:22:41 Quit QT (Read error: 113 (No route to host))
04:42:21 Join bipak [0] (
04:52:12 Nick StrathAFK is now known as Strath (
04:57:43 Quit bipak_ (Read error: 110 (Connection timed out))
05:10:16 Quit RED_M_CHIU ("( :: NoNameScript 3.81 :: )")
05:16:53***Saving seen data "./dancer.seen"
05:27:36 Quit Adity1 ("Download Gaim:")
05:32:52 Join thegeek [0] (
05:50:22 Join Lost-ash [0] (
05:50:57 Quit ashridah (Read error: 110 (Connection timed out))
06:07:47 Nick Lost-ash is now known as ashridah (
06:27:05 Quit courtc (Remote closed the connection)
06:28:35 Join courtc [0] (
07:10:40 Join Shagnar [0] (
07:11:00Shagnarhi guys
07:11:15ShagnarSlasheri did you release a patch with playback?
07:16:56***Saving seen data "./dancer.seen"
07:22:17ashridahno, he didn't.
07:22:20ashridahnothing to see here
07:22:26*ashridah nods
07:22:46ashridahin reality, anything he's been posting has been incredibly experimental
07:25:39ashridahunless you feel like cracking open your player, making a bdm or porting gdb to it so you can use that over serial, it won't do you much good
07:29:58ashridahjudging by the logs, he's still got some glitching to work out
07:30:29 Join LinusN [0] (
07:31:39BTKDaImMaikatamorning all
07:31:43 Nick BTKDaImMaikata is now known as Bager (~Bager@
07:44:20Shagnarashridah ah okay :)
07:44:38Shagnarwell, school now... till later ;)
07:47:18 Quit Shagnar ("( :: NoNameScript 3.81 :: )")
08:07:27 Join Chamois [0] (
08:09:49*HCl scratches his head
08:11:07 Quit ashridah (Read error: 110 (Connection timed out))
08:28:14 Join ashridah [0] (
08:30:40amiconnLinusN: It seems that I found the cause for the overwritten beginning of a recording in some cases.
08:30:53LinusNgreat, what was it
08:31:15amiconnSometimes the very first write() to the recfile takes so long that the recording buffer wraps meanwhile
08:31:48amiconnI'm not sure yet why that happens, the write() took 30 sec in one case...
08:31:57LinusNand that happens if the first write is small, but it has to spin up?
08:32:17LinusN30 secs????
08:32:29amiconnNo. I only tested without prerecording, so the first write is rather large, ~1.3 MB
08:32:41amiconn(buffer size minus watermark)
08:32:55LinusNah, the *very first* write
08:33:26amiconnI suspect this to be the search for a free cluster, but I'm puzzled why
08:34:00LinusNor it could be the open()
08:34:14amiconn(1) Iirc, this would happen at mount if fsinfo is undefined, judging from the boot delay some users observed on iriver
08:34:26amiconnNo, it's not the open(), it's the write()
08:34:40LinusNok, so the file is already opened
08:35:14amiconn(2) Scanning the whole disk from the debug menu takes 38 seconds here, so I wonder what causes the search to take almost the same time
08:35:32amiconnMy HD is filled ~35%
08:35:52 Join Nibbler [0] (
08:35:57amiconnI guess it's time to do some logf() debugging...
08:36:11LinusNamiconn: you could try to invalidate fsinfo and see if it happens
08:36:26amiconnCurrently I do some debug output to the bottom line in the recording screen
08:36:37LinusNamiconn: you don't have gdb?
08:36:59amiconnNo serial mod on my recorder
08:37:41 Quit edx (Read error: 145 (Connection timed out))
08:38:02LinusNi would start with patching the fat driver to not read the fsinfo, forcing a full free cluster search
08:40:29amiconnI have reworked the mpeg thread's recording part. It now correctly flushes when the watermark is reached. It also flushes in chunks on 8 MB units and Ondio, same chunk sizes as for loading
08:41:18amiconnThis allows flushing a bit more than doing it in one piece as the write pointer advances while the flush is in progress
08:42:08amiconnIt also saves > 200 bytes of code, only saving in one place
08:42:28LinusNbut the first-write problem is still there
08:42:54amiconnI was wrong about one problem. The file is closed when a disk-full condition occurs, only it is non-obvious
08:43:04LinusNchunks of 8MB?
08:43:12amiconnhaha, no
08:43:25amiconn'on 8 MB units'
08:43:45amiconnThe chunks are 1 MB for those, and 256 KB on Ondio
08:44:40amiconnI did not yet dig for the frame split problem
08:45:30amiconnBtw, the 'data' parameter in the queues is sometimes useful :)
08:45:46amiconns/queues/queue events/
08:47:10 Join B4gder [0] (
08:47:33LinusNmorning B4gder
08:59:06t0masmorning all
09:03:12 Quit B4gder ("CGI:IRC (Ping timeout)")
09:13:48Slasherihi :)
09:14:23Slasherigot glitches fixed from mp3 stream. That was pretty hard because the libmad buffer feeding structure is so strange
09:14:59*ashridah hands Slasheri the cookie of success
09:15:06t0maschinease? :P
09:15:22LinusNSlasheri: my eternal question - can i see the code? :-)
09:15:33SlasheriLinusN: very soon :)
09:16:58***Saving seen data "./dancer.seen"
09:17:48 Join bobTHC [0] (
09:18:51 Join textchimp [0] (~text@
09:19:10bobTHCmornin' folks !
09:19:50textchimpi was wondering, will it be possible when recording to start recording to a new file on the fly, without losing any data? or do buffer sizes prevent it or something?
09:21:26 Nick QT_ is now known as QT (as@area51.users.madwifi)
09:21:29 Join preglow [0] (
09:21:39 Join B4gder [0] (
09:22:09LinusNtextchimp: yes it is possible
09:22:52LinusNtextchimp: basically, everything that is in the rockbox manual today will work on the iriver as well
09:23:05LinusNprovided the hardware allows it
09:23:41QTbtw, how far away is MP3 playback on iRiver?
09:24:06SlasheriQT: it's working, not totally gapless at the moment
09:24:20textchimplinusN: that function is in the rockbox manual now?
09:24:30QTi guess no iriver user is used to listen to gapless playback ;)
09:24:40LinusNtextchimp: yes
09:24:57QTSlasheri: would say the port is ready for being installed on a n00b's player?
09:25:11t0masSlasheri's code is a hack...
09:25:27t0maswe still need an audio api sometime... and a lot of other things have to be changed...
09:25:29QTi see. then i rather wait some more days/months/whatever_it_takes
09:25:29Slasheriabsolutely not, the code is still experimental/testing only
09:25:43Slasherit0mas: there is simple api already
09:26:05t0masLinusN's or yours?
09:26:06QTi was reading that some folks from MisticRiver have installed the port already
09:26:19*LinusN can't understand why the word "API" has such a magical meaning
09:26:24t0masQT: lots of people have...
09:26:37t0masLinusN: because it has been the point to wait for for a long time ;)
09:26:58preglowthis is the audio api
09:27:21LinusNt0mas: we have not been waiting for an API, we have been waiting for some code to be written
09:27:49LinusNi couldn't care less about the actual api
09:31:27preglowthreads don't save the accumulator state, no?
09:31:36preglowno, they don't
09:33:51LinusNthe MAC is not preserved
09:34:01preglowwe might want to save the macsr
09:34:10preglowthe accumulators are less important
09:34:13 Join Harpy [0] (
09:34:13LinusNwe do?
09:34:28preglowsince they should always be cleared after we're done with them
09:34:42LinusNwhy save macsr?
09:34:56preglowLinusN: well, yeah, either do it there, or demand all the code reverts macsr to the state is was in when they found it
09:35:16 Join Lost-ash [0] (
09:35:29LinusNpreglow: i don't get it
09:35:39 Quit ashridah (Nick collision from services.)
09:35:44preglowto more easily be able to use MUL(x, y)-like macros with the emac unit without worrying about some other thread having set the emac unit to some freak mode
09:35:45 Nick Lost-ash is now known as ashridah (
09:36:02LinusNpreglow: i see
09:36:28preglowLinusN: more or less all the codecs have these macros, and doing a macsr set in the macro would be pretty wasteful
09:36:41LinusNsaving it won't solve the problem, would it?
09:36:59preglowno, not really :>
09:37:37preglowi think we'd need to require a default emac mode or something
09:38:05LinusNor you have to set macsr before starting your calculations
09:38:18preglowyeah, but that'd require manually doing so in every routine
09:38:31LinusNin every routine? why?
09:39:04preglowwell, you never know where/when macsr will be reset
09:39:09preglowand these MUL macros are spread all over the place
09:39:10LinusNyes you do
09:39:16LinusNyou know very well
09:39:31preglowafter a code grep, yes :P
09:39:54LinusNas soon as you return from a yield, you have to reinit the mac
09:40:12preglowthis won't be that much of a problem, really, most code should use the fractional mode, only permanent exception i can think of is libflac
09:41:04preglowwell sure, but then the codec plugin will have to assume how its underlying library does its calculations
09:41:29preglowbut well, yeah, like i said, fractional mode should be what everyone wants
09:41:54preglowbut ok, i'll commit some code from david bryant now
09:43:06tvelocitygoodmorning people...
09:43:55preglowthink i'll wait a couple of hours before i'll start calling it good
09:44:51preglowi wish this "summer" thing i've been hearing about kicked in soon
09:49:23preglowi'm reading wavpack at 450% realtime...
09:49:35preglowthis has got to be wrong
09:49:48 Quit Nibbler ("blubber")
09:50:55 Join t0mas_ [0] (
09:52:48 Join Nibbler [0] (
09:53:08bobTHCpreglow > @120Mhz ?
09:58:52bobTHCso it will be possible to reach 100%@48
10:01:53preglowit already is well above 100% at 48mhz
10:02:06pregloweven with disk writing
10:02:43textchimpwill vorbis still eat more battery than mp3, as it did with iriver firmware?
10:03:16LinusNtextchimp: probably
10:03:46preglowoh yes
10:03:51preglownot much doubt about that
10:04:36preglowvorbis has no static codec model, so needs to access memory quite a lot
10:05:46SlasheriYou can test it if you really want to, i don't guarantee that it will work at all :)
10:05:49*preglow grabs
10:08:27 Quit textchimp ()
10:08:54LinusNSlasheri: i'll have a look at it in a while
10:09:22SlasheriLinusN: ok, great :) however, not much has been changed
10:09:35 Nick Lynx_awy is now known as Lynx_ (HydraIRC@
10:15:59preglowworks like a charm
10:16:09preglowdidn't you mention being able to navigate while playing?
10:16:44 Quit Nibbler ("blubber")
10:16:52SlasheriHmm, the navigation is not possible until wps has been correctly integrated. I don't know how to do that
10:17:02preglowi think i've bugged it, btw :P
10:17:04SlasheriBut you can browse the other menus still
10:17:06preglowit refuses to play any more sound
10:17:23SlasheriYes, if you stop playback, you have to restart rockbox..
10:17:32preglowthat explains it
10:17:32SlasheriI don't know why is that required
10:19:03 Quit ashridah (Read error: 113 (No route to host))
10:19:20 Join Lost-ash [0] (
10:23:16 Join Nibbler [0] (
10:23:18preglowwell, it most certainly seems to work like a charm
10:24:35Slasherithat's great :) but there are still some issues with the buffer handling you might encounter
10:25:01preglowsure, i'm not exactly expecting perfect functionality yet
10:28:21 Join edx [0] (
10:30:06preglowgapless playback works like a friggin charm
10:30:25tvelocityany binaries?:P
10:30:57preglowi don't think you want binaries of something that needs a restart for every file :P
10:31:11preglowgood things come to those who wait
10:31:22crashd_just a smack in the mouth ;]
10:31:31tvelocitysure i want
10:31:41tvelocityi'm exceptionaly curious:P
10:34:38tvelocitywell anyway, i'll compile it and see for my self:P
10:40:19*LinusN added some rockbox kernel docs in the wiki:
10:42:04Slasherioh, that's very useful. Especially new people should find it much easier to start programming with rockbox
10:42:23preglowi still don't know much of rockbox's internals :>
10:42:28amiconnLinusN: Any thoughts about the nested mutex() calls yet?
10:43:28LinusN"just do it"
10:44:04LinusNi think tracking mutex ownership would be a Good Thing
10:44:20LinusNand would be a nice addition in the OS debug menu
10:45:42 Quit tvelocity ("Leaving")
10:47:44amiconnHow do you think this should be handled in the debug menu? Currently there is no list of mutexes...
10:48:19LinusNdo like queue_init() does
10:49:33LinusNand create a mutex_owner() function or something
10:50:06 Nick Lost-ash is now known as ashridah (
10:58:00 Quit ashridah (Read error: 131 (Connection reset by peer))
11:02:34amiconnLinusN: Your example for in RockboxKernel is not true. The button driver always sets it to NULL
11:03:06LinusNsilly me
11:05:15 Join ashridah [0] (
11:06:11 Join jipi [0] (
11:06:13 Join jipi_ [0] (
11:08:48preglowbut now we're starting to get somewhere in the codec department
11:08:51preglowfour realtime codecs
11:09:17ashridahisn't vorbis only realtime at 120MHz?
11:09:25ashridahisn't that a tad high?
11:09:39preglowwe'll be running at 120mhz when decoding anyway :P
11:09:54preglowi'm not saying they'll be stuck at the state they're in
11:10:11preglowi'm saying you can use them to play music realtime
11:10:12ashridahbut doesn't decoding require starting up the disk?
11:10:16preglowfeel free to further optimise
11:10:40preglowwhy, yes, if the tracks aren't loaded, why does that matter?
11:10:45*preglow vanishes for a biy
11:10:53ashridahwell, that's the highest heat output then :)
11:11:16ashridahi suppose if it's only doing it to fill the buffer, that wouldn't necessarily be a disaster...
11:11:47ashridahsince it could potentially be idle with the disk off for much of the time.
11:11:56ashridah(well, idle+decoding)
11:16:14LinusNthe disk heat issue was only a problem at 140MHz
11:16:24LinusNi haven't seen it at 120
11:16:48LinusNbesides, the disk will only spin up occasionally
11:16:55bobTHCOTF : i'm sure someone already give the link but :
11:17:00***Saving seen data "./dancer.seen"
11:17:13bobTHCaudiofile is written in C and support : AIFF/AIFF-C, WAVE, NeXT/Sun .snd/.au, Berkeley/IRCAM/CARL Sound File, Audio Visual Research, Amiga IFF/8SVX, and NIST SPHERE. Supported compression formats are currently G.711 mu-law and A-law and IMA and MS ADPCM.
11:17:14ashridahi thought you went down to 96 to keep it cool.
11:17:18*ashridah considers himself corrected
11:20:20B4gderbobthc: float float float...
11:20:52bobTHC:/ :/ :/
11:21:17 Quit jipi_ ("Leaving")
11:21:17 Quit jipi ("Leaving")
11:21:36preglowwhat's the difference between that and libsndfile anyway?
11:22:14bobTHCThe same calls for opening a file, accessing and manipulating audio metadata (e.g. sample rate, sample format, textual information, MIDI parameters), and reading/writing sample data will work with any supported audio file format.
11:22:33LinusN"supported" being the key here
11:22:57B4gderaudiofile and libsndfile seem pretty similar
11:23:11LinusNstill, it's great to have some code to look at when we want to parse those formats
11:24:14preglowi don't think very many of them will actually be of use
11:24:31LinusNme neither
11:24:37preglowlike adpcm, ulaw, etc
11:24:50bobTHChehe, that's why it's interesting ;)
11:25:27preglowi try to keep all the audio i have in its original format, but in the case of those, i don't even have any original material in those formats
11:25:37preglowand audio in ulaw format i'd rather throw away than listen to anyway
11:26:07preglow8khz sampling frequency is "charming"
11:26:53 Join Bumi [0] (
11:26:56Bumihey everyone
11:27:03Bumianything new on the iriver rockbox project?
11:27:13LinusNBumi: it has officially been cancelled
11:27:21Buminow realy
11:27:31Bumihey linusN
11:27:34Bumigood job BTW
11:27:37Bumivery good job
11:27:54BumiHUGE thumbs up from me
11:28:26B4gdergoodie, I'll tell my wife that and I'll get a few more hours in front of the computer :-P
11:28:35Bumisay, do you still need electrical stuff?
11:28:43LinusNBumi: like...?
11:28:43Bumilogic analysers and such
11:29:01LinusNBumi: you have some to spare? :-)
11:29:17Bumii actualy have one handy here
11:29:32Bumibut if it whold help i can do tests using it and such...
11:29:40Bumii dont think its something i want to send
11:30:12Bumihas anyone here heard from ziv from the begining of the week?
11:30:14SlasheriI have also access to some HP logic analyzers, spectrum analyzer, 2 GHz RF scope etc. :)
11:30:26LinusNziv? the hebrew guy?
11:30:26Bumioh, ok
11:30:39LinusNno, haven't heard from him
11:31:02Bumihe told me at the begining of the week he will have an updated version by the end of the week
11:31:15Bumidoas rockbox play MP3 allready? i mean actualy play it?
11:31:20LinusNsort of
11:31:24LinusNnot officially
11:31:33Bumiwhat doas that mean?
11:31:40LinusNbut Slasheri is working on it
11:32:30Bumigood luck Slasheri
11:32:37Slasherithanks ;)
11:32:41Bumiwhat are all of you guyses perfesion?
11:32:49Bumiyou all enginers?
11:33:02Bumielecrical or software
11:33:02*B4gder is
11:33:17Bumiwhich kind?
11:33:24preglowelectrical engineer here
11:33:27*LinusN is an embedded software engineer
11:33:44Bumihow did you become an emdedded software engineer?
11:33:46Slasheristill student only, but work experience much from embedded systems, fpgas etc.
11:33:51bobTHCand all this guys are wizard too, that's help ;)
11:33:51Bumii mean is there such a degry?
11:34:02*B4gder has no degree
11:34:06LinusNBumi: i hacked for years on my C64 and then i applied for a job
11:34:22Bumiwhats that?
11:34:28LinusNcommodore 64
11:34:41Bumii pressed enter before you:D
11:34:53preglowahh, i was afraid you hadn't heard of it
11:34:53Bumihow old are all of you?
11:34:57preglowin which case i would have to kill you
11:35:03DBUGEnqueued KICK preglow
11:35:10B4gderI'm 22
11:35:47Slasheri20 :)
11:35:50LinusNBagder: you just advanced a step on the kick-o-meter
11:36:03B4gderonly one?
11:36:11Bagerpreglow: about killing bumi:
11:36:53CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
11:36:53*Bumi throws a keyboard at Bager
11:36:57LinusN"industry leader in innovative killing approaches"
11:37:07preglownow this is a web service that innovates
11:37:08*Bager hides
11:37:11Bumiwatch out, am in the army :P
11:37:21BagerLinusN the best one is on the bottom of the page ...
11:37:25LinusN"HITMAN is a cruelty-free organization. None of our services have been tested on animals."
11:38:28preglowi'm surprised the americans haven't taken the site down yet
11:38:40Bagerpreglow i think that this is just a joke
11:38:49Bagersee the contact form ...
11:39:05preglowBager: of course it is
11:39:18ashridahBager: except that death threats aren't free speech.
11:39:19preglowBager: but still, americans have no sense of humour
11:40:58Bumiwell anyway
11:41:06Bumiam now learning to program for the army
11:41:22Bumiso i hope soon i will come back and cotribute back to my new firmware
11:41:48Bagerhaha <META name="ROBOTS" content="INDEX, FOLLOW">
11:42:03Bagerthey definitely want to become famous
11:43:13Bagerthis tells to web robots to follow and index the whole site (
11:44:00Bumisay, did anyone check the battery life for the iriver with rockbox?
11:44:06Bumii see you like programing
11:44:13BumiPerl RULEZ
11:44:30Bagerthis is sed command also ...
11:45:03Bumisay, i want to build some kind of program that will control my router
11:45:08Bumialcatel speedtouch pro
11:45:11Bumivia telnet
11:45:22Bumisince its web interface is shit
11:45:41Bumican anyone point me somewhere to see how telnet realy works?
11:45:51Bumii tryed to download a telnet activeX
11:45:58Bumibut it acts strange
11:49:12bobTHCtelnet is just a remote shell
11:49:23Bumifor some reson alot of clients i found on the web dont like working with the router
11:49:29Bumionly the windows telnet
11:49:32Bumiand putty
11:49:42Bumiand hyperterminal
11:49:56Bagerthese are "plain" terminals
11:49:59ashridahso use putty
11:50:19Bumii want to use a program
11:50:30Bumito run something like telnet scripts
11:50:37LinusNwhat exactly is it you want to do?
11:50:40ashridahuse plink
11:50:46Bumiso i can map ports with a butten click
11:50:51Bumiand port ranges
11:50:52ashridahif you want to script stuff, you want expect and plink
11:51:11Bumibut what if i want to make a GUI?
11:51:33BagerBumi then see the links above ....
11:51:37ashridahthen you do the expect part yourself, and still use plink. :)
11:52:16BumiPlink (a command-line interface to the PuTTY back ends)
11:52:28bobTHCRTFM it's never a loose of time...
11:53:03Bumiis this what expect you talked about?
11:54:25ashridahit's more for scripting purposes tho
11:55:12ashridahi'm not sure how you'd go about getting it working usefully in a gui
11:58:14 Join Lost-ash [0] (
11:59:18Bumihmm i looked around plink
11:59:29Bumibut i dont understand whats the difrance between it
11:59:32Bumiand normal telnet
12:01:04Lost-ashit's basically something you can open and send input into/outof
12:01:09Lost-ashwhereas putty is a gui
12:01:23Lost-ashplink's a tool that spits out stuff on stdout, and takes stuff on stdin
12:01:47Bumiwell, cant the normal telnet client windows has do the same?
12:03:19 Join ashridah__ [0] (
12:03:38Lost-ashwhich is useful for scripts, or using it to connect to stuff via ssh and running stuff remotely, then getting the input back. (or, instead of ssh, telnet)
12:03:38 Quit Lost-ash (Remote closed the connection)
12:03:41ashridah__i am getting SERIOUSLY over this
12:03:43 Quit ashridah (Nick collision from services.)
12:03:55 Nick ashridah__ is now known as ashridah (
12:04:55Bumiwindows telnet can do the same cant it?
12:05:00Bumiwell not the SSH part
12:05:06Bumibut the rest...
12:05:25ashridahi don't believe windows' telnet will be completely silent
12:05:38ashridahi'm assuming, you're referring to a cmd.exe version, not a gui version
12:05:52Bumiwell anyway
12:06:18Bumianyone know some nice ActiveX or cpp module that can hadle all the protocol
12:06:38*ashridah shudders
12:06:38Bumiand i do something like while for "User : "
12:06:52Bumisend Bumi
12:06:58Bumiwhats wrong with activexes?
12:07:49ashridahthey make my teeth itch
12:09:28preglowyou can just wrap a telnet program you know works
12:09:38preglowhook its stdin and stdout
12:09:39ashridahyes. like say, plink
12:09:48preglowi wrote a windows program that did that once
12:10:15 Join webguest14 [0] (
12:10:57ashridahi'm personally of the firm opinion that stuff like he wants isn't really something you want to do in a program :)
12:11:38ashridaha tolerant system like a scripting environment is going to be much less painful.
12:12:04ashridahparticularly things like expect. but, if you want shiny happy buttons, you want shiny happy buttons.
12:12:12ashridahi'm just not sure this is the best place to be asking ;)
12:12:47 Quit webguest14 (Client Quit)
12:14:10Mr_Wikmmmmm, I have a question, I don't have found the answer in the manual or forums...
12:14:36Mr_WikIs there any option to clear the dynamic playlist without shutting down the iHP unit ?
12:15:00ashridahin rockbox, or on the stock firmware?
12:15:35Mr_Wikin rockbox, as on the stock firmware there's not dynamix playlist ;)
12:23:43 Join leftright [0] (
12:25:07Mr_Wikanyone ? :/ ^^
12:25:07 Quit Bumi (Read error: 104 (Connection reset by peer))
12:25:46 Join Bumi [0] (
12:27:31 Quit ashridah (Read error: 60 (Operation timed out))
12:27:45Bagerofft: does anyone know how to center html table with CSS ?
12:32:34 Quit B4gder ("CGI:IRC (EOF)")
12:37:01 Join MoosCamaro [0] (
12:37:10MoosCamaroHi all
12:39:43BagerLynx_ :) 10x!!! a lot!
12:42:26Lynx_Bager: that's the first google hit for center table css ;)
12:44:25BagerLynx_ :) the problem is that i looked into
12:44:31Bagerand there was no info about it
12:44:59Bagerbut ... for IE it's enough to have body { text-align: center; } ....
12:45:37Bagerit's stupid how different behaves it from other browsers ...
12:49:45 Quit Hadaka (Read error: 145 (Connection timed out))
12:50:42 Join asdsd____ [0] (
12:50:43SlasheriNow cpu frequency control works nice, there is long interval between cpu_boost calls
12:51:03LinusNSlasheri: nice
12:51:54MoosCamaroHi Slasheri, i read logs, very good progress, we hope in you :) for this ;)
12:53:46MoosCamaroit's wonderfull
12:58:40DBUGEnqueued KICK asdsd____
12:58:40asdsd____,00hey guys, any one working on the h3x0 yet?
12:59:26LinusNnot really
13:00:37preglowin a way we all are!
13:01:18preglowanyone here have any experience with gdb bugging like a turd on x86-64?
13:01:36 Join Naked [0] (
13:01:48CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
13:01:48*LinusN has to reboot
13:01:54 Nick Naked is now known as Hadaka (
13:02:05 Part LinusN
13:05:30 Quit Bumi ("( :: NoNameScript 3.81 :: )")
13:07:19 Part asdsd____
13:16:05 Join LinusN [0] (
13:17:02***Saving seen data "./dancer.seen"
13:18:59LinusNSlasheri: u there?
13:19:27SlasheriLinusN: yes :)
13:19:42LinusNi have some suggestions
13:19:56LinusN1) you should handle USB connections in your threads
13:21:14Slasherithat should be easy to fix i think
13:21:22LinusN2) use the new logf() function for debugging
13:21:44LinusNre usb handling, look at mpeg.c:1531
13:21:56Slasheriok :)
13:22:41LinusNwhy do you store fd in the track structures?
13:22:50LinusNonly one file should be opened
13:23:00Slasheriah, yes that's unnecessary
13:23:09Slasherii will change it to a global
13:23:37LinusNi see that you move the track data around with memmove instead of using pointers
13:24:14LinusNthat will not work with the mp3entry structs as they work today
13:24:33LinusNsince they use pointers into their respective buffers
13:24:38 Quit austriancoder ("using sirc version 2.211+KSIRC/1.3.12")
13:24:54LinusNconsider doing it like mpeg.c does it today
13:25:07Slasherihmm, that was because codec loader can't yet handle ring buffer format
13:25:21Slasherihowever, i can change the buffer to ring buffer if you like that more
13:25:56Slasheribut anyway, memmove should move no more than the watermark amount of data
13:25:58LinusNit's not a matter if liking
13:26:14Slasherihmm, ok
13:26:22LinusNthe mp3entry structs depend on not being moved
13:26:34LinusNsince they use pointers into themselves
13:27:15 Join oxygen77 [0] (
13:27:48LinusNwe will have to rework the mp3entry stuff anyway, but if you want it to work now you'll have to use a ring buffer instead
13:28:26*preglow kicks gdb, sighs, then fetches fprintf
13:28:27Slasheriokay. do you think i should change the buffer to a ring buffer?
13:28:51Slasherii think it's not very hard to do
13:28:52LinusNyou'll have to do it if you want wps to work without modification
13:29:06Slasheriok, then i will do that :)
13:30:47LinusNbug: in codec_request_buffer_callback(), you don't unlock the mutex if *realsize == 0
13:34:30SlasheriI should not :)
13:34:41SlasheriIt will be unlocked by codec_release_buffer_callback
13:35:19Slasheriyes of course it's bug.. :D
13:35:24LinusNit sure looks weird to set buflocked to false without actually unlocking
13:36:25LinusNin fact, it also looks weird to have both a mutex and a buflocked flag
13:36:29 Join ashridah [0] (
13:37:17SlasheriI do that because i didn't know if it's good idea to call mutex_unlock many time without mutex_lock etc.
13:38:56SlasheriAnd i think that codec_advance_buffer requires some information if the mutex is locked or not
13:40:49SlasheriI decide to make the buffer locking that way that codecs may use the function calls in almost any order without accidentally locking the buffering code
13:41:16LinusNif you really need that info, you could use bufmutex.locked
13:41:45Slasherioh, i will look that
13:41:48 Quit lostlogic ("Going to the moon")
13:43:05preglowdoes anyone know how i make gcc spit out source code line numbers in the assembly it makes? i can't seem to find that option anywhere in the manual
13:44:31Bagder-fverbose-asm ?
13:44:32amiconnLinusN: bool mutex.locked will probably go away when recursive mutexes are implemeted
13:45:04amiconnWe should provide a mutex function to read locking status
13:45:10LinusNamiconn: true
13:46:33preglowBagder: thanks
13:46:48 Join jipi [0] (
13:46:48LinusNBagder: you are a loser and a lame hacker
13:47:00Bagderdarn, you caught me
13:47:02LinusNthe bleeding edge builds have the wrong plugins
13:47:15Bagderthey do?
13:47:20*Bagder investigates
13:47:37LinusNdailies work, but not bleeding edge
13:47:44Bagderright, same problem I guess
13:48:25*Bagder jumps in
13:49:11Bagderrm -rf * is much nicer than make clean ;-)
13:49:33LinusNrm -rf / is a nice thing to do as root
13:49:45Bagderbeen a while since ;-)
13:51:16Bagdernow commit something!
13:51:21SlasheriLinusN: Hmm, there is one problem if i use mutex.locked flag in advance_buffer: it's not possible to know if the buffer was locked by the buffering code or the codec. And we don't want to unlock the mutex unless it's locked by codec
13:51:35amiconnLinusN: dd if=/dev/zero of=/dev/sda ?
13:52:29 Join Shagnar [0] (
13:52:32LinusNSlasheri: if you have such dependencies, maybe you should rething the entire locking?
13:53:01SlasheriLinusN: but maybe that is not a big problem, because if codec want's to use advance_buffer_loc, the buffer should be locked already by codec
13:53:59 Quit jipi ("Leaving")
13:53:59preglowreading x86-64 asm is a nightmare :///
13:54:00SlasheriHmm, yes. Anyway, it should not cause problems if there are not that kind of buffer cheks at all and codecs are written right
13:54:14LinusNHCl: hi
13:54:31HCli don't suppose anyone has seen markun?
13:54:46preglowHCl: he was on yesterday
13:54:47LinusNi have never seen him, i don't know how he looks like
13:54:52Bagdermarkun ( signed off 20 hours and 54 minutes ago ()
13:55:04Bagder "/msg logbot seen markun"
13:55:42*HCl just found out he hasn't uploaded the latest version of their compiler..
13:56:49 Join ted2005 [0] (
13:57:04HClwhere is he :/
13:57:08HCland whats up?
13:58:10 Join markun [0] (
13:58:18markunHCl: I was on ICQ..
13:59:29SlasheriLinusN: the extra buffer check was quite needless. I removed that flag completely without any problems
13:59:50HCli wasnt
14:00:00HClits great that you're on anyways
14:00:41 Join muesli- [0] (
14:01:26 Quit markun (Client Quit)
14:01:39SlasheriLinusN: btw, do you think i should limit that cpu_boost(state) will be called only when state really has changed from last value or could there be some problems if it's called several times with the same value?
14:02:20LinusNit has a counter, so you need to "unboost" it as many times as you have boosted it
14:02:48Slasheriok, so maybe i should not worry about it
14:04:18rasheras long as you unboost the same number of time (and the counter doesn't overflow)
14:04:33rasherwhich is what LinusN said
14:05:42 Join Lost-ash [0] (
14:06:04 Quit ashridah (Nick collision from services.)
14:06:13 Nick Lost-ash is now known as ashridah (
14:09:08Slasherirasher: yes. When the pcm buffer goes empty, the counter might get boosted several times (normally something 1-3) and when buffer is full, it will be unboosted quite fast to zero
14:12:25Slasherihmm, could there be any troubles if i unlock a mutex twice? Will it get locked normally at next time?
14:13:23LinusNin the current implementation yes, but it will not work in the future
14:14:24Slasheriok. this would be nice to do everytime a codec quits so we can be sure the mutex has been unlocked. Oh! i forgot i could read the mutex state from the flag, so no problem :)
14:14:33 Join webguest40 [0] (
14:16:09webguest40.oO(I am lost!)
14:16:25*Bagder points out the direction ====> there you go
14:16:32webguest40hi bobTHC
14:16:37bobTHC=========> []
14:16:50LinusNSlasheri: no, you should keep track of the mutex instead
14:16:50bobTHCwhat's up ?
14:16:56Slasheriand now i realized other thing too: The multitaskin is not pre-emptive, so i have a few extra mutexes there :)
14:18:06webguest40does someone know a URL where I can join different channels? I need to join some other ones? thx fellows!
14:18:22webguest40I am looked behind a firewall....
14:20:43*webguest40 encounter that /join is block by this web irc gateway
14:21:02Bagderwell of course
14:21:13 Join austriancoder [0] (
14:21:15Bagderwhy would our server server the whole world to do irc on freenode?
14:23:09 Quit oxygen77 ("CGI:IRC")
14:23:23webguest40Bagdar: I would release some ressources if I had an alternative :-D
14:23:52Bagderimagine the abuse
14:24:12Bagderwe'd be banned instantly
14:25:19webguest40? For letting people join many different channel through your web irc gateway or what do you mean?
14:25:37preglowpeople would start spamming sooner or later
14:25:43preglowand then the rockbox server would be banned
14:26:30webguest40okay, so there is little chance to find a web gateway which is configured for open access...
14:26:40Bagdervery slim
14:26:52preglowwhy not just use an ordinary irc client?
14:26:54preglowit's even preferrable
14:27:18HCl<.< >.>
14:27:32*HCl whistles innocently and shuts something down on his webserver
14:27:32BagderI usually run a irc bouncer on another machine to deal with irc through proxies
14:27:34preglowi hate firewalls
14:28:27 Join asdsd____ [0] (
14:28:34 Part asdsd____
14:28:54webguest40Badgar, name some of the usefull programs you use for thatt. I have a unix server running at home. So maybe I can look into this in the future...
14:29:28Bagderis the name of the one I've used
14:29:31SlasheriHmm, can i trust that memcpy etc. will never cause a task switch and remove almost all mutexes from code?
14:29:42Slasherigreat :)
14:29:44 Join b0bTHC [0] (
14:30:14webguest40muh? Sounds strange
14:30:29 Join webguest44 [0] (
14:30:39 Quit bobTHC (Nick collision from services.)
14:30:57 Nick b0bTHC is now known as bobTHC (
14:31:07 Quit muesli- ("ich will Kühe!!!")
14:31:14 Join muesli- [0] (
14:31:22amiconnSlasheri: Some other functions may cause thread switches (e.g. read() )
14:31:40webguest40Bagdar, thx.
14:31:43Slasheriamiconn: ok, i will protect them
14:32:58 Quit webguest40 ("#rockbox thanks for your patience")
14:37:05 Quit webguest44 ("CGI:IRC")
14:53:50 Quit Nibbler (Read error: 104 (Connection reset by peer))
15:02:28 Quit muesli- (Read error: 145 (Connection timed out))
15:02:28 Quit thegeek (Read error: 54 (Connection reset by peer))
15:10:44 Quit ashridah (Read error: 148 (No route to host))
15:11:50 Part leftright
15:14:14 Join Aditya [0] (
15:17:04***Saving seen data "./dancer.seen"
15:19:00 Join oxygen77 [0] (
15:28:02 Join Nibbler [0] (
15:33:04 Join ashridah [0] (
15:35:31austriancoderi am allowed to change the sound settings ranges in sound.c for iRiver?
15:35:45LinusNof course
15:36:40austriancoderfine :)
15:39:37LinusNgotta go now
15:39:39LinusNcu around
15:39:50 Part LinusN
15:41:39 Join thegeek [0] (
15:41:39 Quit oxygen77 ("CGI:IRC (EOF)")
15:50:06 Join muesli- [0] (
15:50:40muesli-does anybody know if the 1xx series charger fits to the 3xx series!?
15:51:04rashersounds like it's the same, but I've no idea
15:51:21muesli-me too :|
15:51:38rasherthere are a couple of people here who own both
15:53:00muesli-have to ask in mc probably
15:55:16muesli-err mr
15:57:38*rasher ponders porting matrix from ipodlinux to rockbox
15:58:44rasherjust displays matrix code
15:58:52rasherscreensaver style plugin
15:59:04rasherthink it's a port of cmatrix
15:59:07rasherfor linux
15:59:23muesli-are there errors within in the matrix?
16:00:21rasherUses a lot of malloc though.. lord knows why
16:00:52Bagerrasher to be more interesting for porting to rockbox maybe ? ;)
16:00:54bobTHCblue or red pill rasher ?
16:00:58SlasheriHmm, logf doesn't support %s in format?
16:01:15Bagderit should
16:01:24SlasheriTrying to use logf("File: %s", "something"); for example prints only "File: "
16:01:33rasherBager: I think that's a likely explanation
16:01:45Bagerrasher ;)
16:02:34*Bager wonders when he'll have enough time to dig into rockbox ...
16:06:05BagderSlasheri: it worked fine for me before
16:06:29SlasheriHmm.. doing some more tests
16:06:51Bagderand still do in my sim
16:11:12SlasheriHmm, maybe the problem is that file names are too long so they are not displayed at all
16:12:03Bagderthe output should be chopped at 16 letters then
16:12:24Slasheriah :)
16:13:13*Bagder checks
16:14:23Bagder logf("test %s", "thisisaverylongtexttoadd"); works in my sim
16:15:07austriancoderi cant build my sim on x11 :(
16:15:13 Quit t0mas_ (Remote closed the connection)
16:15:25Bagderwhat happens?
16:17:31Bagderis this a non-patched source tree?
16:17:47austriancoderyep.. it should
16:18:07Bagderit doesn't look like that to me
16:18:36austriancoderi got the error yesterday night too,,
16:19:00BagderI have no pcm.c for example
16:19:40austriancoderah.. i have the sound playing patch applied
16:20:05austriancoderbut i think i will recompile libc and hope the best
16:21:05austriancodermust go now
16:21:09austriancodersee you later
16:21:12 Quit austriancoder ("using sirc version 2.211+KSIRC/1.3.12")
16:24:42rasherI don't think hte sound-patch is entirely simulator-proof
16:26:12 Quit muesli- (Read error: 110 (Connection timed out))
16:29:53 Quit Hadaka (Read error: 145 (Connection timed out))
16:41:53 Quit thegeek (Read error: 60 (Operation timed out))
16:42:50 Join oxygen77 [0] (
16:52:23 Quit ted2005 ()
16:57:32 Quit ashridah ("Leaving")
17:02:05*rasher expects the latest commit to result in wacky garbage on the remote because of the init problems
17:03:27 Join webguest33 [0] (
17:04:33rasherhm, nope
17:06:16 Quit oxygen77 ("CGI:IRC")
17:07:30 Quit webguest33 (Client Quit)
17:08:18amiconnrasher: ?
17:09:25rasherwell the remote lcd is initialized wrongly, shifted 4 pixels to the left
17:09:56rasherand the rightmost 4 pixels are then left blank.. when I tried it they were randomly filled
17:13:43amiconnIt looks like Linus fixed this
17:13:44 Join webguest61 [0] (
17:14:28amiconn(from looking at the code; lcd-h100-remote.c, line 393)
17:14:51webguest61just want to see if anybody is here
17:15:35rasherstill doesn't look quite right
17:17:04webguest61omg my english is bad but can you tell me how to create an dictionaryfile without linux?
17:17:06***Saving seen data "./dancer.seen"
17:17:48webguest61you know for the dict-plugin
17:18:26t0masI can help you
17:18:31t0masI think
17:18:38webguest61ou thank you:)
17:18:46t0masprivmsg me...
17:18:59rasherwait.. I think I'm forgetting to compile or update or something
17:19:07rasherprobably nothing wrong with the remote
17:26:13rasherBagder: do I need any patch to use logf?
17:28:02 Quit webguest61 ("CGI:IRC")
17:28:04 Join webguest61 [0] (
17:31:25 Quit pabs (Remote closed the connection)
17:31:51 Join kergoth [0] (
17:37:26 Join Naked [0] (
17:37:30 Join pabs [0] (
17:37:45 Nick Naked is now known as Hadaka (
17:41:49rasherBagder: /home/rasher/rockbox/build-env/src/rockbox/build/iriver/apps/debug_menu.o:(.rodata+0x7c): undefined reference to `logfdisplay'
17:42:09 Part kergoth ("Leaving")
17:42:36amiconnrasher: Iiuc you just need to configure with logf support, then do a _full_ rebuild
17:43:10rasherthought I did that
17:43:12rasherguess not
17:45:42rasheryeah, that worked
17:45:53rasherexcept now I need to figure out how to use logf from a plugin
17:46:05 Join tvelocity [0] (
17:46:30bobTHClogf recall me too much the natural logarithmic function and confused me ;)
17:46:59 Quit webguest61 ("CGI:IRC")
17:48:06rashermatrix.c:79: warning: implicit declaration of function `logf'
17:48:53t0massounds bad
17:48:59t0mashave you included the header?
17:49:02amiconnIf you want to use it from a plugin, you need to add it to the plugin api of course, then use rb->logf()
17:49:10t0masor isn't there a logf header?
17:49:32amiconnt0mas: You should never include any core header other than plugin.h from a plugin
17:49:43t0masyes, I didn't see it was a plugin
17:49:49rasheramiconn: ah, want me to add that and commit?
17:50:28amiconnIt can't be added unconditionally, and we still need a solution to 'fold it away' within a plugin
17:50:34t0masrasher: if you do... don't forget to raise the plugin ver thing
17:50:50amiconnOtherwise plugin sources would get rather ugly
17:51:32rashernevermind, I have the sound patch applied.. won't commit anything like this
17:51:41amiconnIn the core, it's as simple as #define logf() , when compiling without logf, but that won't work in a plugin
17:54:00t0masamiconn: shouldn't we just add it to the plugin api and make it an empty function there if it's not compiled in? smart compilers won't take the trouble of jumping to it and back right?
17:55:20 Quit Hadaka (Read error: 145 (Connection timed out))
17:56:12amiconnt0mas: A call via a function pointer cannot be optimised away for the same reason as such functions cannot be inlined
17:57:10preglowSlasheri: why does volume max out at 75 ?
17:57:41Slasheripreglow: scaling has been done incorrectly
17:58:06preglow85, i mean
17:58:20preglowlemme guess, you multiply by 3? :P
17:58:25t0masamiconn: then maybe our best bet is to add a define logf() to plugin.h ?
17:58:29t0masand use it without rb->
17:58:29SlasheriBagder: logf("Filling file buffer..."); prints "Filling file bu"
17:58:36Slasheripreglow: yep, that's correct :)
17:58:41t0masSlasheri: max 16 chars
17:59:06SlasheriAh, yes.. :D
18:00:11SlasheriHmm, why is the limit 16 chars? There should be room for about 24 characters per line..
18:00:44amiconn21 chars with the system default font
18:00:59amiconn...on the iriver remote lcd
18:01:05SlasheriInteresting.. Then i must have some other font
18:01:32amiconnlogf() doesn't set the font?
18:01:47preglowsetvol((vol*167116) >> 16) instead
18:02:12Slasheripreglow: thanks, i will change that
18:02:42amiconnpreglow: looks like precision overkill to me
18:02:54preglow167117, i mean
18:03:07preglowamiconn: yeah, it is, but all our platforms have 32 bit ints anyway :-)
18:03:43preglowapart from gmini, of course..
18:04:05preglowanywho, i just tossed it out, i don't expect it's code that'll end up in the final implementation
18:04:37preglowSlasheri: will your current code just loop all files in a directory?
18:04:57Slasheripreglow: it depends playlist configuration i guess
18:04:59amiconnsetvol(vol*255/100); would at least look more readable
18:06:22preglowit's slower :P
18:07:07amiconnNot necessarily. Loading large constants like 167117 may take longer
18:07:36preglownot longer than a div, i assure you
18:07:43preglowbut it's a non-issue anyway
18:08:19rasherthought I'd never see the day when people were optimizing turning the volume up
18:08:31preglowthen you don't really know me, now do you
18:08:38Slasheripreglow: the scaling works well :)
18:09:03preglowyeah, it should
18:09:12preglowfound out why you have to restart after stopping, btw?
18:09:40Slasherinot yet
18:10:07preglowthink i'll take my rockbox for a little walk
18:10:18preglowthis time it'll entertain me with music
18:11:15SlasheriI already had it yesterday playing oggs for me :)
18:11:30rasherit is most excellent
18:13:13SlasheriHmm, there is still some bug with cpu_boost.. If i change it manually from debug menu, system will immediately crash
18:28:06 Join thegeek [0] (
18:29:57 Join asdsd____ [0] (
18:30:04 Part asdsd____
18:39:48HClis it committed yet?
18:40:59*HCl and markun wrote an .net compiler
18:41:02SlasheriHCl: Will be committed soon, when LinusN comes back :)
18:41:07HClnice :)
18:41:24HClthen its time to fix the restart bug, and to hack in the wps
18:45:14SlasheriYes. And the cpu_boost bug should be fixed
18:45:31SlasheriIt will eventually crash the rockbox
18:46:41rasherBy the way, I'm nhaving trouble going into usb mode
18:46:54Slasherirasher: It's fixed now
18:47:01SlasheriWill be included in the cvs..
18:47:02rasheror my iriver is.
18:47:08rasherah, great
18:47:15rasherjust thought I'd mention it
18:47:21Slasherithanks anyway :)
18:48:48HClwhats the cpu boost bug
18:49:39SlasheriI don't completely understand what causes it but it has something to do with pcm playback and dma transfer
18:50:17SlasheriHmm, i think that with logf i was able to locate the stop/start bug :)
18:51:41HClstop start bug?
18:53:14SlasheriCurrently you have to restart rockbox if you want start stopped playback again
18:53:21CoCoLUS * fn~HCl and markun wrote an .net compiler
18:53:23CoCoLUSfor which language?
18:53:34HCloh. you call that the start stop bug :)
18:53:39HClCoCoLUS: "bla"
18:53:43HClour own language
18:54:19ChamoisLinus we need you to commit
18:54:43*HCl goes to catch some sleep..
18:54:47HClafk :/
18:57:20 Quit bobTHC ("Smoke Weed Every Day !")
19:05:40preglowSlasheri: even with this level of audio support, i can't say i see the need to start the iriver firmware again
19:07:29Chamoisonly to flash with a new bootloder
19:10:05t0masSlasheri: after your commit I can look at wps...
19:10:09t0masor is that working already?
19:12:49Slasherit0mas: oh, that would be great. The wps is not working yet
19:13:06Slasheripreglow :D
19:13:06HCland after that i can look at the runtime database :p
19:13:09t0masok, I've read a lot of wps code
19:13:15t0maswhen chaning it to graphical here
19:13:30Slashericool :)
19:13:45SlasheriHmm, i think i know how to fix cpu_boost
19:13:50SlasheriI will add a mutex there :D
19:14:07preglowwhy should that matter?
19:14:29SlasheriBecause that is a serious problem and will crash rockbox at random intervals
19:14:38preglowyeah, but why?
19:14:46t0masyes, but why would a mutex fix it?
19:14:55SlasheriI think that the crash will happen when there is a dma transfer going
19:15:09preglowi can see no reason why it should crash for a dma transfer
19:15:12SlasheriSo i will try preventing the cpu_boost happening at same time
19:15:17preglowbut i guess linus knows this
19:15:30SlasheriHmm, okay. But i will test if it makes any difference
19:17:10***Saving seen data "./dancer.seen"
19:22:38 Join Gibbed [0] (
19:23:01 Quit rasher (
19:23:01 Quit Nibbler (
19:23:01 Quit edx (
19:23:01 Quit courtc (
19:23:01 Quit DMJC (
19:23:01 Quit Mr_Wik (
19:23:01 Quit gromit` (
19:23:01 Quit CoCoLUS (
19:23:01 Quit odd (
19:23:01 Quit Seed (
19:23:01 Quit silencer_ (
19:23:01 Quit mbr (
19:23:01 Quit rob- (
19:23:01 Quit Strath (
19:23:01 Quit Shagnar (
19:23:01 Quit bipak (
19:23:01 Quit TCK (
19:23:01 Quit _DangerousDan (
19:23:01 Quit ze (
19:23:01 Quit preglow (
19:25:23 Join Harpy_ [0] (
19:26:01NJoinpreglow [0] (
19:26:05 Join thegeek_ [0] (
19:26:09 Quit thegeek (
19:26:09 Quit Harpy (
19:26:09 Quit Chamois (
19:26:09 Quit QT (
19:26:09 Quit Rick (
19:26:09 Quit einhirn (
19:26:09 Quit t0mas (
19:26:09 Quit Bagder (
19:26:09 Quit HCl (
19:26:09 Quit TiMiD[FD] (
19:26:09 Quit Plugh_ (
19:26:09 Quit crashd_ (
19:26:19 Join Stryke` [0] (
19:26:28 Nick Gibbed is now known as Rick (
19:26:41 Join crashd [0] (
19:26:41NJoinHCl [0] (
19:26:41NJoinNibbler [0] (
19:26:41NJoinShagnar [0] (
19:26:41NJoinedx [0] (
19:26:41NJoincourtc [0] (
19:26:41NJoinbipak [0] (
19:26:41NJoinDMJC [0] (
19:26:41NJoinStrath [0] (
19:26:41NJoinTCK [0] (
19:26:41NJoinrasher [0] (
19:26:41NJoinMr_Wik [0] (~Mr_Wik@
19:26:41NJoingromit` [0] (~gromit`
19:26:41NJoinCoCoLUS [0] (
19:26:41NJoin_DangerousDan [0] (
19:26:41NJoinodd [0] (
19:26:41NJoinSeed [0] (
19:26:41NJoinmbr [0] (
19:26:41NJoinsilencer_ [0] (
19:26:41NJoinrob- [0] (
19:26:41NJoinze [0] (
19:27:18NJoinPlugh_ [0] (
19:27:20NJoinChamois [0] (
19:28:24CtcpVersion from freenode-connect!freenode@connect.utility.freenode
19:28:52NJoinQT [0] (as@area51.users.madwifi)
19:29:02 Join webguest01 [0] (
19:32:44 Join ted2005 [0] (
19:33:32CtcpTime from ted2005!
19:33:35NJoint0mas [0] (
19:36:09 Join _aLEX [0] (
19:36:17NJoinBagder [0] (
19:36:17 Join TiMiD[FD] [0] (~TiMiD[FD]@
19:41:33 Join webguest71 [0] (
19:43:08t0masSlasheri: does it make difference?
19:45:45 Quit webguest71 (
19:45:45 Quit TiMiD[FD] (
19:45:45 Quit Bagder (
19:46:22Slasherit0mas: i was unable to use mutexes there :/
19:46:29Slasheritrying to think some other way
19:46:32 Quit webguest01 ("CGI:IRC")
19:50:42 Quit thegeek_ (Read error: 104 (Connection reset by peer))
19:52:27 Join muesli- [0] (
19:53:47 Quit ted2005 (
19:53:47 Quit Stryke` (
19:53:47 Quit Rick (
19:53:47 Quit _aLEX (
19:53:47 Quit tvelocity (
19:53:47 Quit pabs (
19:53:47 Quit amiconn (
19:53:47 Quit bill2or3 (
19:53:47 Quit elinenbe (
19:53:47 Join LinusN [0] (
19:53:49NJoin_aLEX [0] (
19:53:49NJointed2005 [0] (
19:53:49NJoinStryke` [0] (
19:53:49NJoinRick [0] (
19:53:49NJointvelocity [0] (
19:53:49NJoinpabs [0] (
19:53:49NJoinamiconn [0] (
19:53:49NJoinbill2or3 [0] (
19:53:49NJoinelinenbe [0] (~elinenbe_@
19:53:50amiconnbah :(
19:53:50NJointhegeek [0] (
19:54:19preglowSlasheri: here's your chance! bagder linus about the cpu_boost issue!
19:54:43preglowwhat, now i'm starting to confuse badger/bagder the other way as well
19:54:47*LinusN hides
19:55:32NJoinTiMiD[FD] [0] (~TiMiD[FD]
19:56:46Slasherihi LinusN :)
19:57:13NJoinBagder [0] (
19:57:55Slasheriyep, the cpu_boost will occassionally crash rockbox and i think it might have something to do with dma transfers. If i change the cpu speed from debug menu while music is playing, the system will crash immediately
19:58:18LinusNhow nice
19:58:33preglowi cherish this behaviour like a fine wine
19:58:52preglowSlasheri: what kind of audio latency are we having now, btw? something close to a second?
19:59:59 Join FireEgl [0] (Ariel@Atlantica.US)
20:00:00Slasheripreglow: Hmm, when the audio buffer is full, it's something like few seconds
20:00:29Slasheribut that's a good thing?
20:01:03preglowon was thinking about things like an adjustable eq
20:01:16preglowit's going to take close to two seconds before people can hear the difference in the sound
20:01:21preglowbut it's not a big problem, i guess
20:01:59amiconnThe latency should be kept low imho
20:02:25preglowthat'll destroy our cpu_boost strategy :/
20:02:37amiconnHmm :/
20:02:55Slasheriah.. But it's easy to decrease the latency by decreasing audio buffer size
20:03:26preglowwe should have someone do a bunch of tests to see how much of a difference there is
20:03:33amiconnFor the voice ui the latency should be kept really low. Music can take a bit more, but I guess > 0.5 s will be irritating
20:03:37*preglow summons rasher
20:04:26rasherWhy, here I am.
20:04:36amiconnRemember my recent fade-out fix for archos? It was necessary because the MAS latency gets higher when the bitrate is lowered
20:05:21Slasherithe start latency is already pretty low when audio buffer is empty
20:06:06 Quit t0mas ("oh no! it's evil... rebooting linux :) brb")
20:06:07amiconnSlasheri: Not only start latency, stop also latency matters
20:06:24amiconnBah, word order :(
20:06:38amiconnstop latency also matters
20:06:47Slasherihmm, okay. The stop is immediate if it's done by stopping dma transfer
20:07:00amiconnFor the voice UI, we need to be able to almost immediately stop one clip and schedule another
20:07:13Slasherishould be not a problem
20:07:19amiconnThen it's okay
20:09:07LinusNSlasheri: so calling cpu_boost with music playing may hang it, right?
20:09:44SlasheriYes, it will hang it after audio has been played a few minutes
20:09:50SlasheriThe interval is random
20:10:03SlasheriAnd from debug_menu the hang is immediate
20:10:31SlasheriHmm, i have to test something
20:11:30SlasheriYou might be not able to reproduce the problem with that earlier patch where dynamic cpu_boost was disablef
20:12:38LinusNSlasheri: could it be disk accesses that cause it?
20:12:58SlasheriMaybe, i don't really know what causes that
20:14:30 Join t0mas [0] (
20:15:04SlasheriLinusN: I think you should be able to reproduce it
20:15:05 Quit Chamois (Read error: 104 (Connection reset by peer))
20:15:15SlasheriJust go to the debug menu and press joystick
20:15:22SlasheriThere should be instant crash
20:15:32LinusNwith music playing?
20:15:43 Join Chamois [0] (
20:22:23 Join muesli_ [0] (
20:22:36 Quit t0mas (Remote closed the connection)
20:24:43 Join t0mas [0] (
20:24:45 Quit t0mas (Read error: 54 (Connection reset by peer))
20:25:02*LinusN puts the kids to bed
20:27:42 Join t0mas [0] (
20:27:45 Quit t0mas (Read error: 54 (Connection reset by peer))
20:28:00 Join Mr_Wik_ [0] (
20:29:43amiconnLinusN: I didn't perform any further tests, but I might have found the cause for the extremely long first write() when recording
20:29:55 Quit Mr_Wik (Read error: 110 (Connection timed out))
20:30:34 Nick Mr_Wik_ is now known as Mr_Wik (
20:30:37Mr_Wikroh Mr_Wik :D
20:31:31amiconnDepending on the MAS signals, my new transfer routine might spend too much time in the loop
20:31:57amiconnThe loop runs within an interrupt... and that might slow down the free cluster search
20:32:28amiconnIt seems I forgot a condition when the loop should be left...
20:33:09 Join Tangleding [0] (
20:33:14amiconnI'll check my theory soon...
20:33:21TangledingHello :)
20:33:36TangledingHi Chamois, hi amiconn
20:34:00Tangledinghave you get your iHP140 or still not?
20:34:49amiconnI can pick it up at the customs office; I hope to be able to do that tomorrow :)
20:35:06TangledingHo at custom
20:35:19Tangledingyou get extra fees? :(
20:35:27amiconnLet's see...
20:35:41TangledingHope not :)
20:36:45TangledingThings have got accelerated it's cool
20:36:56Tangledingis Bjorn here?
20:37:44TangledingI wanted to know if i didn't encountered issue with paypal finaly
20:38:00Tangleding(seen nothing on my account report but maybe too soon)
20:40:09 Quit muesli- (Read error: 110 (Connection timed out))
20:50:25 Join t0mas [0] (
20:50:27 Quit t0mas (Read error: 54 (Connection reset by peer))
20:54:35 Join asdsd____ [0] (
20:54:39 Part asdsd____
20:55:57*amiconn digs out the old recording transfer logic analyser graphs
20:56:27MoosCamaroTang are u r ?
20:56:34MoosCamaroHi amiconn
20:59:35 Join TCK- [0] (
20:59:36 Quit TCK (Read error: 104 (Connection reset by peer))
21:07:58 Join t0mas [0] (
21:08:18 Quit FireEgl (Read error: 104 (Connection reset by peer))
21:10:54Tangledinghello HCl, hello t0mas
21:13:21HClSlasheri: anything new?
21:15:07Tangledingslasheri is here?
21:17:14***Saving seen data "./dancer.seen"
21:22:17 Quit Chamois (" HydraIRC -> <- The professional IRC Client")
21:23:22muesli_psss..dont disturb him ;)
21:24:56 Part tvelocity ("Leaving")
21:25:31 Quit muesli_ ("ich will Kühe!!!")
21:26:35SlasheriHCl: Some bugs have been fixed etc. Now investigating gapless mp3 playback
21:26:45HCl :)
21:27:35CoCoLUShow about a compiled image for the lazy people slasheri :)
21:28:09Tangledingcool way for investigate indeed
21:28:59SlasheriCoCoLUS: The whole thing will come soon to cvs so there is no need for that
21:29:05SlasheriTangleding: yes?
21:29:19CoCoLUScan't wait for soon :)
21:29:20Tangledingif you encounter an issue about the lame stuff for gapless
21:29:34Tangledingyou can contact Gabriel Bouvigne at hydrogen audio
21:29:43Tangledingsee this post
21:29:50Slasheriah, ok :) I will keep that in mind :)
21:30:10Tangledinghe's very cool guy :)
21:30:25 Join Chamois [0] (
21:30:46Tangledingchamois is back
21:31:09preglowdoesn't explain much other than the info exists
21:31:10 Join muesli- [0] (
21:31:11preglowwe know that
21:31:45Tangledingin fact it was just a tip
21:31:48 Join austriancoder [0] (~austrianc@
21:32:11Tangledingi imagine he can supply more info on this if needed
21:33:25preglowahh, yes
21:33:28preglowi imagine he can
21:33:36preglowa quick google doesn't yield accurate information
21:34:24Tangledingokay anyway it's just an option
21:34:43Tangledingif you get out without needing his help it's cool
21:34:59Tangledingif you dodn't understand sthing thought
21:35:16Tangledingthe guy is very helpful
21:35:22Tangledingand sympathetic
21:36:28Tangledingof course Peter Pavlovski can be a solution so since he developed foobar gapless feature with lame using the header
21:36:29preglowthis is his site, no?
21:36:39preglowi believe the info we want can be found there
21:36:40Tangledingit's his site
21:36:45preglowthat describes the info tag
21:38:01Slasheripreglow: Hmm, great site
21:38:25SlasheriI think that header info could be even read by codec plugin itself
21:38:54SlasheriShould be quite simple to get the important fields
21:38:57preglowi think that's appropriate for mp3
21:39:05preglowlibmad should never need to see it
21:40:19preglow32 bit integer filed containing the exact length in bytes of the mp3 file originally made by LAME excluded ID3 tag info at the end.
21:40:22preglowthere it is
21:40:38preglowor no
21:40:46SlasheriThe first byte it counts is the first byte of this LAME Tag and the last byte it counts is the last byte of the last mp3 frame containing music.
21:40:50SlasheriOh, that is really cool
21:41:19SlasheriHehe, we pasted the same thing twice ;)
21:41:39preglowbut seems like that is just a simple file size
21:41:40Tangledingeh eh seems you won't need to "bother" Gabriel himself
21:42:11SlasheriIt might be still worth to try out
21:42:48 Quit TCK- (Read error: 54 (Connection reset by peer))
21:43:48LinusNthat integer isn't that helpful is it?
21:44:22LinusNi mean, our mp3 parser finds that out anyway
21:45:05Stryke`for gapless, you'd want to look at Encoder delays
21:45:05LinusNin fact, rockbox on archos is already gapless on mp3 frame level
21:45:23LinusNStryke`: exactly
21:45:26Stryke`bytes $B1-$B3
21:47:19 Quit thegeek (Read error: 104 (Connection reset by peer))
21:47:35preglowof course
21:47:36 Join thegeek [0] (
21:47:41preglowi just skipped that entry
21:47:43preglowvery wise of me
21:48:06 Quit t0mas ()
21:48:08preglowthere you can see how many zeroes have been used to pad the last frame
21:48:49LinusNthat's the info we need for 100% gapless at the pcm level
21:49:08preglowand the info is provided straight for use, how convenient
21:49:12LinusNstripping the id3 tags is simple, and rockbox already has code for doing that
21:50:01preglowit handles id3v2 as well?
21:50:24LinusNyou don't think highly of rockbox, do you? :-)
21:50:32LinusNof course it does
21:50:35preglowyou never know, some people refuse to support id3v2
21:50:53Stryke`it strips APEv2?
21:51:01preglowi don't think there are many mp3s with ape tags
21:51:03LinusNStryke`: nope, not yet
21:51:57Stryke`shouldn't be difficult
21:52:32preglowapev2 is fairly simple
21:52:44LinusNand fairly unnecessary imho
21:52:58preglowwell, like i said, some people hate id3v2
21:53:05preglowwith a passion
21:53:10Stryke`to me, apev2 is the best of both words (id3v1 and v2)
21:53:16SlasheriHmm, sounds quite straightforward to do if rockbox already supports reading all necessary information
21:53:19LinusNi think id3v2 is an extremely silly "standard"
21:53:30preglowyes, but id3v1 is even worse
21:53:35preglowsimple and plain shitty
21:53:55LinusNSlasheri: i think our mp3 parser should extract the necessary info from the lame tag
21:54:13rasherWhat if there's no lame tag?
21:54:39LinusNthen we can't do pcm level gapless
21:54:42preglowthen we just use the entire last frame
21:54:47preglowthere's not much of a gap anyway
21:54:55LinusNSlasheri: mp3data.c:516
21:55:05preglowbut you'll just have to live with it, there's no simple way to detect if a frame has been padded
21:55:08LinusNfill in the blanks
21:55:45SlasheriLinusN: ok :)
21:56:32rasherI see.. I mostly care about ogg anyway. Just thought that it'd be possible to do gapless without having to encode in special ways
21:56:53preglowthe mp3 creators never thought about it
21:56:58*LinusN has to go away for a while
21:57:04rasherThe bastards.
21:57:14*preglow feels alone and forgotten
21:58:24 Join FireEgl [0] (Ariel@Atlantica.US)
21:59:01 Quit Shagnar ("( :: NoNameScript 3.81 :: )")
21:59:13SlasheriLinusN: Hmm, it would be great if get_mp3file_info could read all the information from a given buffer address instead of file handle
21:59:48SlasheriMaybe we could a little modification to support this?
22:01:40 Join ehntoo [0] (
22:01:52 Quit Stryke` (Read error: 131 (Connection reset by peer))
22:03:21 Quit Tangleding ("Chatzilla 0.9.68a [Firefox 1.0.4/20050511]")
22:04:34 Join t0mas [0] (
22:12:05HClanything i can do?
22:12:07amiconnLinusN: Is there a reason why some places use #pragma interrupt while others use __attribute__ ((interrupt_handler)) in a declaration?
22:16:06Rickamiconn: different coders? :p
22:18:35amiconnLinusN: Another question, perhaps even more important. Why is /IRQ3 set to edge sensitive?
22:25:36 Join webguest59 [0] (
22:26:32 Quit webguest59 (Client Quit)
22:30:52amiconnIgnore me; found out why... at least for playback
22:31:34 Join webguest62 [0] (
22:33:03amiconnIs there a way to force fsinfo 'hint' recalculation, i.e. without writing a file, and also without causing a full scan including free space calculation?
22:34:00webguest62hi, all
22:35:10webguest62Just a quick (I hope) question. Anybody here have an URL for archos hard drive upgrade howto?
22:35:29Slasheribtw, there is the "ticking" problem with remote lcd if screen is updated while audio is playing..
22:36:04muesli-so there is no ticking anymore?
22:36:07SlasheriSo this must be a hardware bug
22:36:24webguest62I've got a 40 gb travelstar I want to swap into a recorder10
22:36:26Slasherino, there is. but only when screen content is changed
22:37:14Bagderwebguest62: see the docs page on the rockbox site
22:37:21muesli-hmpf i have feared this
22:38:06muesli-folks have to send in there units to fix it. wouldnt make sense to assume a software issue
22:40:50Slasheribut that problem is easy to solve just by not using the headphone jack on remote control or purchasing earpuds that have integrated volume control. Then set the volume from the earpuds low enough and the ticking will be suspended
22:41:28webguest62Bagder: thanks, I knew it existed, but couldn't remember where.
22:43:42amiconnSlasheri: This sounds like that ticking is caused by crosstalk from the lcd data lines to the headphone output
22:44:25Slasherithat is my guess too
22:44:28amiconnPerhaps it can be avoided by driving the lcd slow enough
22:44:47Slasherihmm, maybe :)
22:50:26preglowyeah, that'd be worth a shot
22:52:37amiconnIt's uncertain whether this would help; on archos there was a similar issue:
22:53:10amiconnOn the recorders, the original firmware produces a faint background noise while playback
22:53:58amiconnRockbox doesn't do that, and it drives the lcd way faster than archos...
22:55:58austriancoderSlasheri: where can i get your current patchset for rockbox?
22:58:10Slasheriaustriancoder: If you want to test it, you can find the code here: Take the snapshot.diff, not the other. Then you have to copy the all separate files to right place and compile
22:58:58 Quit muesli- (Read error: 60 (Operation timed out))
22:59:25 Quit FireEgl (Read error: 54 (Connection reset by peer))
23:01:09austriancoderhope i can commit some new uda1380 and sound.c code tonight
23:01:17preglowanything fun? :V
23:02:28austriancodernot really ;)
23:04:04 Quit ehntoo ("Leaving")
23:09:14 Join FireEgl [0] (~FireEgl@Atlantica.IPv6Tunnel.Info)
23:14:41 Quit FireEgl (Remote closed the connection)
23:17:17***Saving seen data "./dancer.seen"
23:21:53 Quit Harpy_ (Read error: 60 (Operation timed out))
23:29:50 Quit Bagder ("Off to search for that connect-resetting peer guy!")
23:30:31 Join Bagder [0] (
23:33:32 Quit t0mas (Remote closed the connection)
23:44:22 Join XavierGr [0] (
23:44:57 Quit ze (Read error: 110 (Connection timed out))
23:46:51XavierGrHello all!
23:46:57Bagderhey ho
23:47:30XavierGrI just wanted to compile the win32 uisimulator but when I execute the make file an error pops-up in the bash
23:48:13XavierGrmake[1]: /home/Xavier/rockbox/tools/convbdf: Command not found
23:48:13XavierGrmake[1]: *** [/home/Xavier/rockbox/build/firmware/sysfont.o] Error 127
23:48:13XavierGrmake: *** [all] Error 2
23:48:27Bagderyou must run make in the tools dir first
23:48:30rashercd ../tools
23:48:44XavierGrlets see...
23:50:31XavierGrcollect2: ld returned 1 exit status
23:50:31XavierGrmake[1]: *** [/home/Xavier/rockbox/build/rockboxui.exe] Error 1
23:50:31DBUGEnqueued KICK XavierGr
23:50:31XavierGrmake: *** [all] Error 2
23:50:46XavierGrI got the source from the daily build
23:50:58thegeekgot the same error
23:51:04thegeekwith the cygwin devkit
23:51:04Bagderand no error before that?
23:51:16 Join FireEgl [0] (~FireEgl@Atlantica.US)
23:51:58XavierGrthe first time I executed the code it was fine(except when I was running a plug-in from the simulator) all was build
23:52:10XavierGrI will try again getting the code from cvs
23:53:37 Quit Chamois (Read error: 54 (Connection reset by peer))
23:54:03 Join Chamois [0] (
23:55:05Bagderso there was no other error?
23:55:26XavierGrthe first time or now?
23:55:44Bagderwhen you get this "ld returned 1 exit status"
23:56:06XavierGrno, no other error
23:56:25XavierGrbut I remember that when it compiled I had the source downloaded from the cvs
23:56:39Bagderit shouldn't matter
23:56:44XavierGrthen I moved the uisimulator dir from the daily build
23:57:00Bagderthat is totally wrong
23:57:57 Join ze [0] (

Previous day | Next day