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 2005-05-27

00:00:22amiconnThat unconditionally reverts everything...
00:00:36Bagderno mercy
00:00:46amiconnI sometimes do similar
00:00:55amiconn...but save my changes first.
00:01:09amiconncvs diff -u >blah.patch
00:01:20amiconnpatch -R -p0 <blah.patch
00:01:36BagderI actually have "diff -u" in my .cvsrc, so I never use "diff -u"
00:02:23preglowthat too seems like sound advice
00:02:35Bagder"update -dP" is neat to have there too
00:06:12Bagdermy current debug log browser patch is now at
00:08:28 Quit F1^Aison ("( :: NoNameScript 3.72 :: )")
00:08:43amiconnHmm. I don't understand what's going wrong here :(
00:10:45rasher(sim's working by the way)
00:11:29amiconnTry building an X11 iriver sim and start that
00:11:45amiconn(will only work if you installed cygwin x11)
00:12:27rasherI didn't
00:12:46rasherI aborted the whole cygwin thing when I realised I could crosscompiled the sim
00:14:52preglowyou should have aborted it when you noticed you have a linux box :P
00:17:31rasherThis center-scrolling of markun's is nice
00:18:04preglowsounds like a good setting, if it's easily doable
00:22:10 Quit tvelocity ("Leaving")
00:31:18 Join ashridah [0] (
00:33:37amiconnI found at least 4 problems in the mpeg thread. One significantly affects disk spinups (and hence battery life) in recording mode...
00:37:07preglowhow you can find anything in a 2000+ line source code file in such a short time puzzles me ;)
00:37:34amiconnIt wasn't that short of a time
00:37:57amiconnReading this mess over and over actually helps after some time
00:38:09 Join Kawoa [0] (~sss@
00:38:12Kawoahey anyone around
00:38:38*amiconn suspects something fishy in file.c now as well
00:38:58Kawoaumm does anyone of you ever seen a iriver
00:39:19preglowi'm fondling a h120 as we speak
00:39:32amiconnThe end-of-data-saved-at-beginning-of-file happens _only_ if the mpeg thread does _not_ use fysnc() before closing the file
00:39:53Kawoai jsut bough a h10
00:40:02amiconn...i.e. either when timesplitting or when encountering disk full
00:40:06Kawoado you think the tweaking will work on the h10
00:40:11crashd_Kawoa: no
00:40:16crashd_h10 has hardware very similar to the ipod
00:40:20crashd_and the ipod mini
00:40:35crashd_that's not to say developments wont happen on the h10 in the future
00:40:47crashd_just that, as of now, the most you can do (safely) is to change fonts on the h10
00:40:50crashd_which is a bit pap.
00:40:51Kawoado you know if there is a tweak for h10 somewhere?
00:41:20preglowwhat kind of tweak?
00:41:25 Join lostlogic [0] (
00:41:34preglowcrashd_: so you're lurking around, are you, any more luck on the disassembling?
00:41:46crashd_preglow: im not lurking, well, i am, but i've had exams this week :|
00:41:47crashd_and not yet
00:41:57crashd_although ive managed to locate about 90% of the data within the flash rom
00:41:58preglowKawoa: i don't know, we don't exactly peddle in tweaks
00:42:00crashd_like, images/text data
00:42:02Kawoai just wanna take a look at h10 tweaks
00:42:03crashd_and 00'ed it out
00:42:05ashridahKawoa: what are you looking for in a tweak?
00:42:11crashd_so dissassembly will be easier when i get round to looking at it again
00:42:20ashridahie, performance? replacing images?
00:42:20Kawoacuz i though maybe my iriver could play movies
00:42:29preglowahh, no, don't expect that
00:42:39preglowthat doesn't qualify as a 'tweak'
00:42:44preglowthat would be more like a 'miracle'
00:42:46crashd_that's a full blown miracle
00:42:49preglowcrashd_: ;)
00:42:50crashd_: ]
00:43:01Kawoaok just wondering thanks
00:43:13crashd_Kawoa: keep your eyes on the h10 forum on
00:43:13 Quit Kawoa ()
00:43:18crashd_just fuck off
00:43:50preglowbut go on
00:43:53preglowh10 rockbox would, yes, rock
00:43:59crashd_yes, yes it would
00:44:14crashd_just a case of figuring this minor problem out first, really
00:44:24amiconnHmm. close() just calls fsync(). Weird.
00:44:46HClamiconn: do you have your iriver yet?
00:45:06rasherhe seems too focused to have recieved it
00:47:06 Quit Harpy (Connection timed out)
00:47:39crashd_i need to find another h10 users who's got the motivation and knowledge to help with this project
00:47:44crashd_cursed misticriver forum users!
00:47:46*crashd_ shakes fist
00:48:46*webguest74 shakes fist back
00:49:15rasherthere's this other forum..
00:49:20rasherwhat's its name again
00:49:32rasherthe one that declared ware on rockbox
00:49:57webguest74the german one ?
00:50:48rashercould be
00:55:15rasherthat'd be it
00:55:33preglowthe germans that were the very epitome of maturity
00:56:18rasherBut at least they seemed slightly more tech-savvy than misticriver
00:56:26HClthats not hard
00:56:26webguest74I read that saga on their sight, seems some ego's were seriously bruised over there
00:56:47preglowthey were touchy
00:57:09ashridahwhy on earth did iriverlounge declare war on rockbox?
00:57:11preglowi understand their concern, but hell, when you run a german only forum with specs on it, you're not helping yourself
00:57:27ashridahand when exactly was the ravening horde supposed to descend? because, you know, it hasn't yet... :)
00:57:29preglowashridah: because people took their h1x0 specs and didn't link to their site for every line of text
00:57:49ashridahaah, that's right
00:57:56ashridahthey actually declared 'war' over that?
00:58:00preglowended up with rockbox removing all their info
00:58:11webguest74its a shame really, I'm sure there talent would have been welcome
00:58:20preglowthey knew what they were about
00:58:23ashridaha simple "don't infringe on 'our' stuff, plskthx" probably would have worked
00:58:37preglowgerman only forum...
00:59:20preglowif every bloody thread hadn't been in german, it would have been a bit easier cooperating
00:59:52preglowbut then again, rockbox removed the info they complained about, but they still acted anal retentive about it
01:00:16preglowbut yes, it is a pity, they seemed like people who knew what they were about
01:00:58rasherit doesn't appear to be german-only
01:01:37preglowthe threads i saw most certainly were
01:02:18preglowi remember people having to translate the threads as well
01:03:24preglowi don't get their reasoning
01:03:33preglowbut it's of no consequence
01:03:46webguest74I think they wanted to retain control, and thereby get all the credit, that vanished when rockbox came along
01:04:16MoosCamaroI don't remenber the topic page of clash betwen Zagor and Konrad
01:04:40preglowthere was this thread about copyright
01:04:48preglowkonrad simply _REFUSED_ to take the point
01:04:50webguest74I guess they can just dowmload it here now :)
01:05:42 Join TCK [0] (
01:06:01MoosCamaroiriverlounge think than rockbox steal iriver infos
01:06:12MoosCamarobut it's open source ;)
01:08:13rasherIt's info that anyone could find, jesus
01:09:36amiconnWell, it's quite some work to find all the info, but I don't get why someone would publish this info when he doesn't want it to be of use for others
01:10:43preglowtheir main point seemed to be that they wanted people to frequent their site
01:11:08MoosCamaroi find it
01:11:09preglowwhich i understand perfectly, but a lot of the info was in german only threads, and you don't have much choice but to translate and mirror then
01:11:14MoosCamarothe clash
01:12:14MoosCamaro Zagor is in true
01:14:06MoosCamarothere is a bit of pride ;)
01:14:38preglowa small bit, yes
01:14:41***Saving seen data "./dancer.seen"
01:16:07webguest74high maintenance personality, ego needed lots of stroking
01:16:16preglowi especially love that thread
01:17:30MoosCamarohihi i remenber this :) the pride too ;)
01:22:52rasherthat's spectacular
01:25:22MoosCamarothis is pity, that would have been a good marriage
01:25:47webguest74of code
01:26:42MoosCamaroiriverlounge team + rockbox team
01:27:09webguest74iriverlounge rocks box :)
01:29:19MoosCamaroTang tried to do this marriage
01:30:52MoosCamarobut in vain :(
01:32:00MoosCamaroHi Tang if you read the logs
01:33:18preglowthey wanted to do their own thing, not join rockbox
01:33:39MoosCamaroit's true
01:34:33MoosCamarobut the goal was the even, a better firmware ;)
01:36:50MoosCamaroif they wanted really an alternate firmware, they could have collaborated
01:37:44MoosCamarothe pride of a put new project to leaves
01:39:22MoosCamarobut fortunately team it current is well ;)
01:39:50*webguest74 yawns
01:40:08webguest74way past my bed time
01:40:28webguest74nice chatting to you guys
01:40:39MoosCamarogood night
01:41:04 Quit webguest74 ("CGI:IRC")
01:41:27MoosCamaroI will go to sleep also, good night all
01:42:18 Part MoosCamaro
01:43:45 Quit preglow ("leaving")
02:25:02 Quit Stryke` (Read error: 110 (Connection timed out))
03:14:44***Saving seen data "./dancer.seen"
03:18:03 Join amiconn_ [0] (
03:36:22 Quit amiconn (Read error: 110 (Connection timed out))
03:36:22 Nick amiconn_ is now known as amiconn (
03:54:19 Join Stryke` [0] (
04:05:55 Quit cYmen ("zZz")
04:06:00 Join QT_ [0] (as@area51.users.madwifi)
04:15:58 Quit QT (Read error: 113 (No route to host))
04:48:49 Quit Stryke` (Read error: 110 (Connection timed out))
05:13:44 Join daven [0] (
05:14:47***Saving seen data "./dancer.seen"
05:18:15 Quit ldone ()
05:21:23 Join textchimp [0] (
05:22:16 Quit daven ("CGI:IRC (EOF)")
05:22:29 Join DMJC [0] (
05:48:38HCl*rubs eyes and peers into the light*
05:55:48 Join courtc_ [0] (
06:00:47 Quit Aditya ("Download Gaim:")
06:04:56 Quit courtc (Read error: 110 (Connection timed out))
06:15:57 Join kergoth [0] (
06:16:51 Join Stryke` [0] (
06:27:01 Quit zol1 (Read error: 104 (Connection reset by peer))
06:30:49 Nick izzy is now known as Ismo (
06:37:33 Join Lost-ash [0] (
06:48:58 Quit ashridah (Read error: 110 (Connection timed out))
06:49:11 Part kergoth ("Leaving")
06:51:29 Join LinusN [0] (
07:03:14 Nick Lost-ash is now known as ashridah (
07:03:15 Quit textchimp (Read error: 104 (Connection reset by peer))
07:07:20 Quit Stryke` (Read error: 110 (Connection timed out))
07:14:50***Saving seen data "./dancer.seen"
07:22:00 Quit Bager (Read error: 104 (Connection reset by peer))
07:28:07 Join cheriff [0] (
07:33:01 Join textchimp [0] (
07:37:38amiconnGood morning
07:38:11RickGood evening
07:48:09 Join odd [0] (
07:49:02amiconnLinusN: Did you read the log?
07:49:18amiconnSomething really strange is going on when recording....
07:49:20LinusNabout the fsync() discovery?
07:49:39amiconnI'm still unsure at which point it happens
07:50:11amiconnFor longer recording, some data is saved at the beginning of a file which doesn't belong there
07:50:32amiconnThis doesn't happen always, and the exact length varies
07:51:04amiconnIt seems the file code is not to blame, but it is a buffering issue
07:51:27amiconnI found 4 other problems in the mpeg thread that I want to talk about
07:52:06amiconnFor finding which exact part of data that is, I'll try adding timecode to the mpeg frames. The MAS supports this
07:53:11LinusNyes, it has the same limit as the frame counter
07:53:35amiconnYes, at it actually is the frame counter value.
07:54:35amiconnSo the test recording needs to stay <~ 3.5 hours
07:55:54amiconnAnyway, about the 4 problems I found:
07:56:12amiconnFirst ones are about playback
07:56:59amiconn(1) Lines 1405..1414: Doesn't this actually cause the loading to stop at _2 times_ MPEG_HIGH_WATER ?
07:59:07amiconn(2) Lines 1444..1478: Imho "Te very rare case.." isn't handled correctly here (lines 1473, 1474). In case the id3v1 tag is split _within_ then word 'TAG', the mas is fed 'T' or 'TA'
07:59:48amiconnNext ones are about recording:
08:01:10LinusN1) yes, seems so. We should perhaps correct that, as it may be confusing for people reading the code.
08:01:31LinusNand we can save 2 bytes of buffer space :-)
08:01:58amiconn(3) 1768..1808: In case of disk full condition, the file isn't closed. Apart from that, this is 2 times almost identical code, which appears at a 3rd place also, and could be made a function
08:04:05amiconn(4) 1841..1905. This ever only saves up to the end of buffer, not the other part in the next cycle as intended. It sends itself the same message again, but doesn't save because the condition at lines 1843..1845 no longer holds
08:04:17LinusN2) yes, i think i remember that i (in my infinite laziness) ignored that risk when i wrote that code
08:04:30amiconnThe last one significantly affects battery runtime when recording.
08:05:19amiconnI just wanted to ask you about your opinion; will probably try to correct these things myself.
08:06:45LinusN3) oops, and yes
08:08:47amiconn(5) The 1854..1870: Same as (3). (file not closed on 'disk full')
08:09:32LinusN4) aha, it doesn't save the rest as it is below the watermark
08:09:59LinusNdecreasing the available buffer space
08:10:24amiconn...and saving ~twice per buffer round
08:10:46LinusNi think all these issues should be fixed
08:11:01amiconnThis really happens, I measured the time between 2 save-file-operations
08:11:08LinusNi'm most worried about the fsync() issue though
08:11:21amiconnI don't think it's fsync()
08:11:27LinusNor rather fsync() related
08:11:41LinusNi don't think it's fsync either
08:12:08LinusNso it seems like the file pointer suddenly points to the beginning of the file?
08:12:21amiconnAh, and another one: At time splits, the 'split at a frame boundary' doesn't work
08:12:40LinusNit doesn't?
08:12:44LinusNis sure used to
08:13:36LinusNi remember testing this quite a lot when i implemented it
08:13:43amiconnAll my test recordings (even the 2 where the overwrite problem didn't occur) have the last frame truncated, with the next file matching...
08:14:22amiconn(using 10 hours time split)
08:15:28 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
08:15:34LinusNdoes the save-at-beginning bug only happen in the last file?
08:16:10amiconnNo. With my last test I checked this. It only happened in the first file
08:16:17amiconn(of a total of 2)
08:16:47amiconnThe strange thing is doesn't match any of the 2 file ends
08:16:59amiconn*is that the snippet...
08:17:16Rickadd a unique header to every write? :p
08:17:46LinusNhmmm, it could be related to 4)
08:18:10 Quit Rick (Read error: 54 (Connection reset by peer))
08:18:20amiconnThe exact length of this snippet varies, I got ~16 sec (q=5), 13 sec (q=4), 4 sec (q=3) and 2 sec (q=2)
08:18:54amiconnLinusN: Yes, that's what I think
08:18:58 Join Rick [0] (
08:19:19amiconn...or at least, want to correct, then try again
08:19:53amiconnApart from these problems, I didn't get any corrupt frames in >60 hours of recording!
08:24:21amiconnOkay, maybe there is the occasional frame that has a crc error (I have no tool for testing that), but no sync errors other than the ends of the overwritten snippets
08:26:07amiconn...recording from s/pdif, with independent frames set to on, and even at q=7 (where no overwriting happened)
08:30:28LinusNmaybe the overwriting bug can also overwrite random parts in the middle of a file?
08:30:52amiconnI don't know yet. Within my test sequence, it didn't
08:31:04 Join edx__ [0] (
08:31:40amiconnBtw, I also had an idea how to get the frame count when prerecording and timesplitting, as long as the total frame count did not yet saturate
08:31:45Slasherihi :)
08:32:25SlasheriLinusN: Do you think if the mpeg.c is the right place to implement codec plugin memory loading and initialization?
08:33:24LinusNmy plan is to abandon mpeg.c for sw codecs
08:33:33Slasheriokay, hmm. Should i do it in totally different file if i try something?
08:33:45LinusNand move all playback/recording to apps/
08:33:57LinusNSlasheri: may i ask what you are doing?
08:35:25Slasheriyes, i am trying to implement simple mp3 playback integration to rockbox =) yesterday i did an audio buffer in pcm_playback.c and now true basic mp3 playback is possible from viewer plugins
08:36:13LinusNmay i see the code?
08:37:13Slasherijust a moment, i find the url. However, currently it requires more coding before it could be inserted in cvs
08:37:34LinusNi have no doubts about that
08:38:10SlasheriHere, the patch has some strange and unnecessary code
08:38:22 Join Stryke` [0] (
08:40:55LinusNlooks nice and simple, but i'm not that fond of the memcpy()
08:42:38SlasheriAh, great :) Hmm, what would be a better way to do it? Some direct data insertion to memory without memcpy?
08:42:43LinusNit will probably add lots and lots af small pcm chunks when the audio buffer fills up
08:43:27LinusNit will yield() until the next pcm chunk is played, and add another chunk of the same size
08:43:43LinusNthe chunks will eventually get smaller and smaller
08:43:55LinusNjust a theory from reading your code
08:44:28 Join wacky_ [0] (
08:44:55wacky_Hey boys, I have a fix for a tiny bug in iriverify, a bug found by someone on the mailing lists...
08:45:01wacky_is anyone here ?
08:45:06*wacky_ checks de current.txt
08:45:11LinusNwacky_: shoot
08:45:44SlasheriHmm, that might be true.. Maybe there should be some minimum chunk size that will be added?
08:46:19LinusNmaybe, or simply a different approach to the buffer-full handling
08:46:23wacky_what's the best way for you to get a patch ?
08:46:38LinusNthe patch tracker
08:46:49wacky_would you like me to submit it over there ?
08:47:09 Quit edx (Read error: 110 (Connection timed out))
08:47:31 Nick edx__ is now known as edx (
08:47:35LinusNnot this time
08:47:40LinusNwacky_: real name?
08:47:46wacky_the patch fixes rockbox hanging when iriverifying a file twice
08:47:48wacky_Alexandre Bourget
08:48:52wacky_well I did the iriverify thing in the first place, so it's the minimum I must do, to maintain it a little :)
08:52:07wacky_alright :) thanks for everything..
08:52:12wacky_LinusN - you got the file ?
08:53:48LinusNyes, is this the issue when it hangs if you are converting an already converted file?
08:54:21 Quit cheriff (Remote closed the connection)
08:54:22LinusNcommitted, thanks
08:54:29wacky_thanks to you :)
08:54:53 Quit wacky_ ("good night")
08:57:50LinusNSlasheri: where were we? :-)
08:58:28LinusNah, yes
08:59:04LinusNi don't think we should worry that much about codec plugin loading and stuff
08:59:09 Join einhirn [0] (
08:59:10LinusNnot at this point
08:59:40LinusNi have lots and lots of uncommitted code for mp3 playback
08:59:55SlasheriHmm, okay. I just think that I could do some code if that would be useful
09:00:29LinusNi like your simple pcm buffer concept
09:01:20SlasheriI think it should be pretty easy to add file/codec buffer and to do some loader for that
09:01:32rasherit is indeed quite simple - even I understand it :)
09:02:00LinusNSlasheri: adding a buffer and load the codec isn't the hard part
09:02:38SlasheriHmm, what is the more difficult part?
09:02:39LinusNbut in this stage it isn't very practical to do it
09:02:56LinusNwe know too little about how we would want the codec api to look like
09:03:26LinusNand having the codecs as plugins makes it much more difficult to debug
09:03:37Slasheriah, i understand. I though that some simple working api could be a good point to start :)
09:03:42rasherjust use logf!
09:04:00LinusNwe can develop the api anyway
09:04:06LinusNjust not load it dynamically
09:04:21LinusNit is a pain to handle plugins in the debugger
09:05:05LinusNand imho it isn't the dynamic loading that is the most important at this point
09:05:55LinusNyour ring buffer yield() idea isn't really that good for the audio thread
09:06:08LinusNit blocks the thread while the audio is playing
09:06:42LinusNuntil the next chunk is played
09:07:49Slasheriyes, that was only for some testing. It is rather easy to make it non-blocking also if that is necessary
09:08:40LinusNi have some basic requirements from a pcm buffer:
09:09:05LinusN- it must not block the codec thread
09:09:43LinusN- it shall report to the audio thread when the last chunk of a track has been played
09:10:06LinusN(meaning that the audio thread must tell the pcm buffer which chunk it is)
09:10:50LinusN- it must let the codec thread decode data in fairly large chunks, so we can make effective use of the cpu_boost() function
09:11:11LinusNmaybe it should be in its own thread...
09:11:30*LinusN goes to the lab, bbs
09:11:44SlasheriHmm, interesting. At the beginning I had own thread for the buffer, but later i removed it
09:13:45SlasheriI have to go too.. See you later. I will think these ideas and could do some more coding at weekend :)
09:13:50rasherHow much does mkboot check? Injecting the bootloader into a ihp-100.hex went painlessly
09:14:16rasher(didn't do anything more)
09:14:53***Saving seen data "./dancer.seen"
09:15:39 Nick courtc_ is now known as courtc (
09:15:40LinusNit doesn't chech much at all
09:22:05 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
09:24:04rasherWas sortof tempted to flash it when patching went fine
09:24:57rasherI'll control myself
09:26:19 Join Aison [0] (
09:32:02 Join Harpy [0] (
09:56:52 Join chimpfrenzy [0] (
09:56:52 Quit textchimp (Read error: 104 (Connection reset by peer))
10:07:04 Join Bager [0] (~Bager@
10:28:42 Join bobTHC [0] (
10:29:05bobTHChi folks !
10:51:31 Nick Lynx_ is now known as Lynx_awy (HydraIRC@
11:14:55***Saving seen data "./dancer.seen"
11:18:07 Join preglow [0] (
11:18:22Bagerhi, preglow
11:23:34 Nick QT_ is now known as QT (as@area51.users.madwifi)
11:28:43preglowi think i see a red build coming
11:29:15preglownope :V
11:29:50rasheroh my, vorbis optimizations
11:33:16 Quit odd_ (Ping timeout: 14400 seconds)
11:39:56rasherhrm, the audio-out patch gives IllInstr on vorbis2wav now
11:42:26rasherno! I'm playing oggs!
11:44:06rasherit almost manages q10 realtime
11:45:18preglowtry lower bitrates
11:45:20preglowaround q1
11:45:22preglowi think it'll choke then
11:45:32rasherq5 sounds fine
11:45:41preglowyes, but q5 still uses a sanely sized window
11:46:29rasherq1 made it explode
11:46:39rasherit just says "recoverable"
11:46:41preglowi don't know what i think of putting all those windows in iram, it would be better to have them in const memory, and copy to iram only those two you need
11:47:58rasherq3 gives illinstr
11:48:22preglowyes, q4 should work, though
11:48:42rasherwhat the.. now q5 isn't either
11:48:53rasherI just played that!
11:48:58preglowmemory init issue
11:49:02preglowi'll wager
11:49:15preglowsure you didn't play an mp3? :P
11:49:49rasherby the way, setting the remote size to 128px confuses the remote - more is needed indeed.
11:50:13rasherI am sure, yes.
11:50:17rashernow it did it again
11:57:02rashercan't make it work now
11:57:55preglowdoes it work fine without the playback code?
11:58:08rasherlet me check
11:59:57 Quit chimpfrenzy (Read error: 110 (Connection timed out))
12:02:42rasherseems so
12:02:58rasherbut then, it worked with the playback code a while ago
12:05:26rasherLinusN: mind if I try flashing the ihp-100, or would you rather have it in untouched condition?
12:06:34rasher(I'm so very curious)
12:07:52rasherlooks like it was the playback code that wrecked vorbis
12:09:11rasherworks with q1
12:09:16rasherand q3
12:11:47ashridahrofl. for a second there, i thought you lot were talking about quake
12:11:58ashridahi'm going "wtf, q5? "
12:12:03*ashridah <−− not quite with it
12:14:23rasherI doubt quake would be possible on iriver
12:14:29rasherdoom on the other hand...
12:15:02ashridahheh. anyone notice that a bunch of enterprising people have crammed doom1 onto the terminals in doom3?
12:15:40rasherhaven't seen doom3
12:16:19ashridahwell, it has fairly decent interactive programmable panels.
12:16:40ashridahbetter than an actual pushable switch, can make pin-number code panels and other animated gear to go with it.
12:16:47ashridahso someone recreated doom1 on it :)
12:20:00LinusNrasher: i'd like to have it untouched... :-(
12:21:53rasherLinusN: sure
12:22:52 Join B4gder [0] (
12:23:12rasherIt really is untouched. I keep being amazed at how nice condition it is in.
12:23:30LinusNrasher: he probably didn't use it very much
12:23:50*LinusN looks at Pedros latest commit and feels all warm inside
12:23:54rasherWell, he even bought it used himself!
12:24:17rasherBut yeah, he said he was selling it because he never used it
12:24:26rasherIt even has the stickers on the back
12:24:41B4gdera bit commit it is
12:24:45*LinusN gets some coffee
12:25:01rasherIt's most excellent
12:25:40rasherq10 oggs are not quite realtime
12:26:14rasherBut it's a nice step forward
12:26:59 Join webguest16 [0] (
12:28:10preglowcompiling libmad with O2 gives a very nice boost
12:28:56preglowwell, yes, but not really surprising ;)
12:29:02rasher"free" boosts
12:29:12rasherhow much?
12:29:48preglowfor 112 kbps it jumped from 314% peak to 335%
12:29:55preglow322% nominal
12:30:26preglowbrb, i'm wanted next door
12:30:53Seedpreglow: tell the police you didn't do it
12:31:17*B4gder grins
12:31:29rasherThat's indeed a nice boost
12:32:00 Quit webguest16 (
12:32:00 Quit Harpy (
12:32:00 Quit einhirn (
12:32:00 Quit ashridah (
12:32:00 Quit Bagder (
12:32:00 Quit AEnertia (
12:32:00 Quit ferenczy (
12:33:15 Join Harpy [0] (
12:33:22 Join MoosCamaro [0] (
12:33:30MoosCamaroHi guys
12:34:08MoosCamarohi Linus que tal? :)
12:35:00LinusNbien merci :-)
12:35:04B4gderaha, hablamos espanol ahora ;-)
12:35:43MoosCamarohaha espagnol+français :)
12:36:37NJoinashridah [0] (
12:37:54 Quit Harpy (Killed by (Nick collision))
12:37:57NJoinwebguest16 [0] (
12:37:57NJoinHarpy [0] (
12:37:57NJoineinhirn [0] (
12:37:57NJoinferenczy [0] (
12:37:57NJoinBagder [0] (
12:38:17 Join Harpy_ [0] (
12:39:13preglowwhat, i leave you for one minute and everyone goes spanish?
12:39:27 Quit Harpy (Connection reset by peer)
12:39:46Bageralmost everyone ;)
12:39:47 Quit ferenczy (Excess Flood)
12:39:53 Join ferenczy [0] (
12:40:46B4gderque pregunta!
12:41:04MoosCamarohi preglow: hablas un poco este ideoma?
12:41:40MoosCamarohaha para saber
12:42:33preglowyou've long since exceeded my spanish skills
12:42:37LinusNrasher: where is the cursor-in-the-middle patch?
12:45:40pregloware we btw aiming to eliminate all mallocs from the codecs?
12:46:35rasher(markun did it)
12:46:39B4gderpreglow: yes, removing them or doing them with some silly fake-malloc
12:46:57preglowB4gder: well, yeah, doesn't what we use now qualify as a fake-malloc? :>
12:47:06B4gderyes it does
12:47:24preglowminimizing the memory used for malloc would be nice
12:47:35preglowtremor has malloc all over the bloody place
12:47:40preglowflac as well
12:48:00 Quit Bagder (
12:48:00 Quit einhirn (
12:48:00 Quit webguest16 (
12:48:07B4gderisn't there some kind of penalty for that?
12:48:13preglowi really wonder if declaring variables 'register' does any good with current compilers
12:48:25preglowB4gder: why, yes, you get beaten with a cane if you ever set foot in norway
12:48:26B4gderI doubt that
12:48:38B4gder(the register part that is, not the cane)
12:49:11*LinusN whispers to rasher: "patch tracker...paaatch tracker..."
12:50:26NJoinwebguest16 [0] (
12:50:26NJoineinhirn [0] (
12:50:26NJoinBagder [0] (
12:50:29rasherblame markun
12:51:54preglowwill anyone slap me if i stuff libmad -O2 in cvs?
12:52:11LinusNpreglow: try it and see :-)
12:52:40preglowi'll just check if it produces correct output first
12:52:55LinusNgood idea :-)
12:53:08Bagerpreglow try various .mp3
12:53:28preglowBager: i think a layer3 and a layer2 will do
12:54:09*preglow catches the digital flow
12:54:21*B4gder tries to flow digitally
12:54:24preglowoh, how i am BURNING to code a replacement to the affront to mankind that is audacity
12:54:28rasherI have a n mp2 version of that also
12:54:50preglowan abomination unto god, it is
12:56:02preglowi wonder why mpeg layer2 is so slow compared to layer3
12:56:16ashridahwow, that is a major improvement to vorbis2wav
12:56:46rasherIt is indeed
12:56:48B4gderhero of the day, no doubt
12:57:44BagerLinusN did you see the update in
12:57:57LinusNyes i did
12:58:09Bagerany news on this front ?
12:58:11preglowi wonder why the hell registry.c was commited, there are no changes...
12:59:02B4gderwhitespace changes there were
12:59:34preglowthere are a lot of whitespace changes, it seems
12:59:55B4gderyes, I bet his editor does these nasty things or similar
13:00:20preglowyeah, every line he's touched is commited, even though it appears unaltered
13:00:45ashridahpreglow: he's probably using an editor like textpad that's stripping whitespace from the end of lines
13:00:48preglowprobably some tab abomination going on as well
13:01:17ashridahmy partner for an assignment used to do that. he'd keep fucking things up, because i'd have to keep merging the differences between code that differs only by age and whitespace
13:01:17rashergood thing he's not coding in whitespace
13:01:46preglowso THAT'S where monty put all the whitespace in tremor
13:02:10preglowhe sure as hell is stingy on it otherwise
13:14:58***Saving seen data "./dancer.seen"
13:17:00 Quit Bagder (
13:17:00 Quit einhirn (
13:17:00 Quit webguest16 (
13:19:16NJoinwebguest16 [0] (
13:19:16NJoineinhirn [0] (
13:19:16NJoinBagder [0] (
13:40:00rasherLinusN: did you try writing CMS Peripherals?
13:40:21LinusNno i had to run home
13:40:38 Quit lostlogic ("Going to the moon")
13:45:46 Join cYmen [0] (
13:47:41webguest16hi, what files do I need to look at when trying to add a codec to apps/codecs. I've changed apps/FILES and apps/codecs/Makefile
13:48:27LinusNwebguest16: look in tools/configure
13:48:49LinusNwhich codec is it?
13:48:53B4gderSOURCES is what controls the compilation
13:49:11webguest16ok, its libsidplay
13:49:30webguest16rasher, don't expect anything at all
13:49:47rasherwell just having it in there would be nice
13:49:49webguest16just trying to get rockbox to build it
13:50:27rasherI'm sure some nostalgic soul (nuge nudge) will optimize it eventually
13:50:54webguest16it's written i c though
13:50:57 Join kurzhaarrocker [0] (
13:51:07webguest16c plus plus
13:51:21rasherapparently that's not a problem per se
13:51:32webguest16ok, goodie
13:52:00preglowyou need to wrap it, but apart from that it shouldn't be a problem
13:52:06preglowjust make sure you disable exceptions and the like
13:52:22webguest16preglow, to make it build?
13:52:34preglowwebguest16: no, to make it not to horrible things to rockbox if it crashes
13:52:38preglowto = do
13:52:50preglowyou might also need to do it to make it link, though
13:53:05webguest16for now, I think I just try to add it in the way dumb is added to the source tree
13:53:29rasherthat's certainly a nice start
13:53:29preglowplease strip all unneeded files, though, that was not done with dumb
13:54:02preglowi wonder how well resid works with libsidplay
13:55:00webguest16no idea
13:55:10webguest16is that hardware?
13:55:21webguest16i.e resid
13:55:34preglowresid is just a very accurate sid chip emulation library
13:56:02webguest16preglow, ok.
13:56:17rasherand presumably slower?
13:56:22preglowof course
13:56:31rasheris this needed on a portable device?
13:56:46preglowwhat, if accurate emulation is needed on a portable?
13:56:48rasherI mean, if we can get both fast and accurate, I'm all for it
13:57:06preglowlibsidplay is fast, but not really accurate
13:57:16preglowi'm used to a real sid chip, so i notice it pretty easily
13:57:19rasherbut I'd rather sacrifice saccuracy bthan speed
13:57:29preglowi wouldn't
13:58:28webguest16preglow, the drawback with resid is that they say it needs a lot of processing power
13:58:34rasherI mean in the case where resid wouldn't actually run realtime
13:58:50rasherif it doesm then I'm all for it
13:59:04preglowrasher: that's not an option, of course
13:59:11preglowwebguest16: i don't believe resid needs THAT much processing power if you don't oversample
13:59:31preglowwebguest16: it still think it's worth a shot some day
13:59:35rasherwe'll see 
13:59:52rasheror you'll see.. this is waaaay beyond me
14:00:07webguest16preglow, if not to render accurate wav's at least on target
14:00:56preglowi see libsidplay uses operator new[] everwhere, i don't know how we can handle that
14:01:42preglowi really have no idea how that's handled internally
14:01:52preglowwebguest16: i don't really see the point in rendering wavs on target
14:02:00webguest16me neither
14:02:18HClgeeze its hot
14:02:24HCl25-36 degrees
14:02:29preglowraining like there's no tomorrow here
14:02:33LinusNc++ in rockbox? the answer is no
14:02:56HClwhy not anyways?
14:02:58preglowthen there'll be no sid playback for a while
14:03:03webguest16;) I just thought that one would be able to listen to accurate versions of the sids.
14:03:05preglowsid people seem to like c++ :V
14:03:24B4gdersid people are crazy, so it doesn't surprise me ;-)
14:03:43webguest16dumb is c right?
14:03:52webguest16argh, c plus plus
14:03:58LinusNis it?
14:04:00HClits not.
14:04:12preglowit's c
14:04:33preglowi thought that was what he asked ;)
14:04:53Bagerpreglow what's the problem with the new operator ?
14:04:59webguest16the webclient don't like pluses
14:05:05B4gderthe web client can't do + ;-)
14:05:29B4gderBager: we have no malloc so no heap so now new
14:05:30Bagerwe can always overload it as call to malloc
14:05:31preglowBager: well, where do you suppose the compiler gets the memory it allocated?
14:06:05preglowbut no
14:06:10preglowget rid of new[] anyway
14:06:18preglowit's like malloc, we don't want it
14:06:21BagerB4gder i know
14:06:21Bagerpreglow in the same place where is the fake malloc buffer for codecs ?
14:06:27rasherhow about playsid for amiga.. would that be any use?
14:06:31Bagerthis is something different
14:06:37preglowrasher: perhaps
14:07:08Bagerso there is a plan to remove malloc's from all codecs ?
14:07:19LinusNBagder: oh yes
14:07:38preglowit's all asm, rasher
14:07:52preglowi'd say forget that
14:07:53Bagerok, but that will change sources significantly ...
14:08:04webguest16so, libsidplay is out of the question?
14:08:06LinusNBagder: yes
14:08:06preglowi'm not about to rewrite 120kb of 68k asm to coldfire asm
14:08:24Bageriirc zagor (or maybe someone else) was against this
14:08:26LinusNpreglow: come on.... :-)
14:08:31preglowand in the case of libflac, very annoying
14:08:38Bagerpreglow: we are with you
14:08:54LinusNBager: i am against big modifications as well
14:08:58B4gderme too
14:09:01Bagerso ?
14:09:03LinusNbut removing malloc is essential
14:09:04B4gderbut I'm even more against malloc and new
14:09:06preglowi'm fearing for my sanity enough as it is
14:09:32LinusNpreglow: we are too :-)
14:11:03preglowgood, someone has to call the hospital when it goes wrong some day
14:11:21LinusNpreglow: look out your window...
14:11:23webguest16doesn't 120kb of asm sound nice?
14:11:28*preglow hides
14:11:33webguest16...and tempting
14:11:51LinusNpreglow: don't try any sweaters with long arms
14:12:15*Bager gives 2 preglow a valerian
14:13:21 Part webguest16
14:13:27rasherpreglow: when you say 300%, is that without hdd output?
14:14:12preglowrasher: without hdd output and dithering
14:14:29rasherAh, more calm now
14:14:36rasherI was getting only 130%
14:14:57LinusNthe dithering takes a lot of power
14:15:13preglowunnecessarily, imho
14:15:17preglowshould be able to make that quite fast
14:15:28preglowbut then, i haven't tried
14:19:25preglowrasher: 130% at 120mhz, btw?
14:19:48preglowrasher: this is without the playback patch, i assume? :P
14:19:56preglowrasher: what bitrate?
14:20:08rasherand yes, without playback patch
14:20:13preglowmy, dithering really does gobble cpi
14:20:41 Quit B4gder ("Lämnar")
14:21:01HCllets fix that :D
14:21:16 Part kurzhaarrocker
14:21:39preglowgo ahead
14:21:43HClno :P
14:21:44preglowsome asm will do you good
14:21:55HClbusy enough with our compiler
14:22:46preglowbut not nearly amused enough!
14:22:59*preglow pushes HCl towards libmad
14:23:12*HCl stumbles over libmad and falls onto the couch
14:24:07HClnah, i'm no good with asm stuff.
14:24:30LinusNyeah, he is! :-)
14:24:30rashernot you, the codec
14:24:33preglownor am i, but it didn't stop me from trying!
14:24:40 Join tucoz [0] (
14:24:50preglow -falign-loops -falign-labels
14:24:50preglow -fcrossjumping
14:24:50HClmaybe when i finished our compiler assignment
14:24:56preglowyes, thanks for the paste
14:25:12preglowguess what i've been reading
14:25:22LinusNa makefile!
14:25:24HClthe hitch hikers guide to the galaxy
14:25:30preglowwrong! gcc docs
14:25:44rasherHCl was pretty close
14:25:59preglowi wonder what the reason behind assigned the side mouse button to paste was
14:26:14*preglow sighs and corrects himself: assigning
14:26:56LinusNit's on the middle button for me
14:27:07preglowyeah, me as well, but the side button has been assigned to paste as well, obviously
14:27:13preglowi usually never touch it
14:27:38*rasher notes
14:27:49LinusNi hate having a zillion buttons on the mouse, especially on the sides
14:28:08preglowyes, me as well
14:28:14preglowi have just one
14:28:19LinusNrasher: yes, i wanted to apply that patch, but i had second thoughts
14:28:22*HCl patpats his trackball with a scrollwheel
14:28:32LinusNand then i suddenly had better things to do
14:28:33preglowi hate trackballs :P
14:28:42HClpreglow: they're great for moving around in 3d spaces
14:28:50HClit works awesomely with 3dmax and maya
14:28:56preglowi seldom do that
14:29:33*HCl forces his plextor to eat what it just spit out
14:30:14*preglow sighs in relief as he remembers what 'plextor' is
14:30:31HCl :P
14:30:37preglowa moment i thought you were talking about your cat
14:30:48HCli would never do that to my kitty
14:30:59HClmy cat is named moby o.o
14:31:11preglowgood, plextor is not a very fine name
14:31:17HCl :P
14:32:56preglowi hate x86, it's impossible to get a straight cycle count
14:33:32rasherLinusN: man, that "patch" is broken
14:36:07*HCl shakes his head at dutch radio, grinning.
14:36:19HClthey're handing out "really wrong" music awards to artists
14:36:23HCland because they're all english
14:36:26HClthey think its a good thing ;x
14:40:55rasherthe tetris looks nice
14:41:17rasherI like the menu
14:41:34LinusNyeah, i'm having second thoughts about the tilted archos version
14:42:17rasherhow so?
14:42:36LinusNi think it's perfectly playable without rotating it
14:42:49LinusNso we can get rid of the rotation option
14:43:21rasherah.. maybe people like having it larger
14:43:32rasherbut then, I can't really comment on it at all
14:45:06LinusNin his patch, it's not that much larger anyway
14:46:33*LinusN has to go
14:46:38LinusNcu guys
14:46:47 Part LinusN
15:02:08 Join Aditya [0] (
15:04:58 Join elinenbe [0] (~elinenbe_@
15:09:51 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
15:15:00***Saving seen data "./dancer.seen"
15:30:55 Quit Bager (Read error: 110 (Connection timed out))
15:44:26 Join Bager [0] (~Bager@
15:47:52 Join Adity1 [0] (
15:48:34 Quit Aditya (Read error: 131 (Connection reset by peer))
15:51:06 Quit ashridah ("Leaving")
16:19:06preglowSeed: what, libmusepack has dies?
16:20:25preglowwell, i'm not fixing that for now
16:27:45tucozpreglow, is libsidplay out of the question for rockbox?
16:29:42 Join ashridah [0] (
16:29:48tucozand in that case, why is stuff written in c plus plus not wanted?
16:30:16preglowdon't ask me
16:30:23preglowi think c++ is ok as long as it's fairly fancy c++
16:30:31preglowit's NOT fairly fancy
16:31:04preglowno stuff like new, exceptions, rtti
16:31:17tucozoh, your'e right
16:33:58preglowbut that shouldn't be a problem
16:34:08preglowbut i've never tried to call c++ code from c, so i don't know
16:34:56tucozme neither, but I sure want sid playback on my rockbox
16:36:22tucozjust a question, when compiling c for rockbox, is the m68k-elf-gcc able to do this?
16:36:27tucozc plus plus
16:37:16preglowyes, i'd most certainly think so
16:37:21tucozok, great
16:37:22preglowagain, as long as it doesn't too many fancy things
16:37:27preglowlike require a ton of standard libs
16:37:42preglowi'd love to try, but i don't have time right now
16:38:14tucozyes. It's very hard to figure out how to adapt the Makefiles in libsidplay to rockbox
16:38:49tucozas I am far from a Makefile/linker expert
16:39:15 Quit ashridah ("Leaving")
16:40:23*preglow summons pbvas
16:55:40preglowlibsidplay is sidplay2?
17:08:46preglowoooh, i think that's very cpu hungry
17:09:42tucozok, I think that it has to modes, where one is resid
17:10:04preglowit comes more from the fact that the 6510 emulator is cycle accurate
17:10:07preglowbut yeah, i know little about it
17:10:18tucozdon't really know, but I had a look at sidplay aswell, with the old libsidplay
17:10:36preglowbut having libsidplay2 would be wonderful, of course
17:10:59tucozI think it supports more sid-types than sidplay
17:13:07tucozMight have another look at the old libsidplay. the makefiles are driving me crazy. all dependencies to here and there
17:13:22preglowit also has way better playback than sidplay
17:13:53tucozyes, I haven't tried sidplay though
17:15:03***Saving seen data "./dancer.seen"
17:17:42Seedpreglow: not "died", "rebranded" ;)
17:21:06preglowcvs renaming is such a bother :/
17:21:21bobTHCscript it ;)
17:21:24preglowbet i'll just go on calling it libmusepack
17:24:18Seeda 144 MHz Palm Zire 71 plays .mpc files at 170%. iRivers will need a lot of hand optimization before 100% is achieved
17:25:29preglowSeed: this i know
17:25:46preglowi even know what to optimize
17:25:49preglowjust need people to do it
17:30:54 Quit bobTHC (Read error: 60 (Operation timed out))
17:33:25SlasheriHmm, i think that the rockbox integrated multi-codec system for playback is possible to implement. At next week you might have something to test :)
17:34:00tucozhappy happy, joy joy
17:34:47preglowyes, just be aware that linus too is working on this, so if you end up trying to make a more complete system, you should talk with him about it
17:35:52Slasheriokay, i will do as simple as possible.. some code to start with
17:37:44SlasheriI just need something to code and this rockbox looks interesting =)
17:48:32preglowhell yes, please continue what you're doing
17:51:47Slasheribtw, is there any way to get "what is currently playing" display to show on iriver? That would help testing
17:58:52preglowno idea
17:59:15Slasheriokay, i will need to find out that later
18:22:12 Join einhirn [0] (
18:22:32 Join cYmen_ [0] (
18:23:41 Quit cYmen (Read error: 104 (Connection reset by peer))
18:35:50 Nick DMJC is now known as DMJC0skeeo (
18:45:14 Join jipi [0] (
18:47:36 Quit jipi (Client Quit)
19:09:22 Join Stryke` [0] (
19:15:07***Saving seen data "./dancer.seen"
19:23:12 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
19:44:02 Join StrathAFK [0] (
20:02:37 Quit Strath (Read error: 110 (Connection timed out))
20:20:57 Join tvelocity [0] (
20:21:33 Nick StrathAFK is now known as Strath (
20:23:15SlasheriUh, now the rockbox plays directly when i select a file :)
20:23:23SlasheriStill much to do but great beginning
20:24:14SlasheriThere is a dynamic ram loader for codec plugins (that LinusN probably hates at the moment) and filebuffer. But _much_ more to do now :)
20:25:39preglowpatch :P
20:25:46Slasherinot yet ;)
20:28:40 Join Stryke` [0] (
20:29:41 Join elinenbe_ [0] (~elinenbe_@
20:30:00 Quit elinenbe (Read error: 131 (Connection reset by peer))
20:30:00 Nick elinenbe_ is now known as elinenbe (~elinenbe_@
20:35:30rasherSlasheri: do you get the wps?
20:35:58Slasherirasher: No, i don't know how it should be installed/configured/programmed.. :/
20:36:03Slasherionly blank screen now
20:41:48MoosCamaroBagder work on it
20:42:14preglowblank screen is just fine
20:42:22 Join markun [0] (
20:42:22SlasheriBut you will get a patch at least on next week. Then i hope it is able to do true gapless playback, follow playlists and _maybe_ even have the wps
20:42:25preglowit works just for mp3 still?
20:42:31preglownext week????
20:42:35preglowi want it now :/
20:42:40Stryke`true gapless :-)
20:42:56Slasheriyes, but i will extend it to work on all files because it supports the multi codec architecture
20:43:10Slasheriit needs more coding and testing to be stable & useful
20:43:28markunWhat is a good place for the center-scrolling option in the settings menu? File view?
20:44:26rasherI'd say under displisplay->scrolling
20:44:42rashercall it File scroll mode or something
20:45:24markunA bit strange maybe because scrolling is about the scrolling of a line of text, but ok.
20:45:47rasherWell.. it just says "scrolling" doesn't it?
20:46:04rasherbut.. I'm not really sure either
20:46:44markunThe center scrolling also works for menu items, so I don't know if file scroll mode is clear enough.
20:46:49rasherI'd probably put it under scrolling, but I can see the arguments for putting it under file view
20:47:00rasherdidn't know it did
20:47:16rasheror perhaps it didn't in the first patch?
20:47:17markunI wanted to fix it but it worked out of the box :)
20:47:24markunYes, it does
20:47:36markunWell, at least plugin browsing.
20:48:00rasherI think that's just a tree
20:48:09rasherit doesn't work in the "open with" menu
20:48:25*rasher finds a huge font
20:49:03*markun is doing the same :)
20:49:05rashernot large enough, it seems
20:49:11amiconnI didn't have a look at it yet, but I think we should do a poll on the ml
20:49:33amiconnIf it looks good, we could just make it the default behaviour, and go without an option
20:49:47amiconn...of course except on the player
20:49:49rasherto preserve space?
20:50:01amiconnCould you say 'center' within 2 lines?
20:50:11rasherwell it should work on the player as well I guess
20:50:16 Join muesli- [0] (
20:50:25rasherbut be totally uncalled for :)
20:50:25amiconnrasher: To save space for one, but also to reduce option clutter
20:50:32rasherI agree
20:50:38markunrasher: you are right, doesn't work in the menu yet..
20:51:14rasherI'd expect it to work the same both places
20:51:21SlasheriHmm, do you know how many bits is size_t?
20:51:25amiconnCenter scrolling isn't really hard to code; that's what the keyboard input line does as well (horizontally in that case)
20:52:39markunI was very happy today when all the ogg files (q3) I tried decoded at about 150%
20:52:42rasherI like how even the largest font in rockbox holds more text than the iRiver firmware can
20:52:45amiconnrasher: How do you think could center scrolling work on the player? With just 2 lines...
20:53:06rasheramiconn: I'd expect it to behave just like it does now
20:53:11rasherbut for different reasons than now
20:53:17amiconnAh, yes
20:53:43rasherperhaps the cursor would be at the top of the screen instead
20:53:43amiconnmarkun: Btw, how does your center scrolling behave when there is an even number of lines?
20:54:23rasherI was going to say.. it should pick the center+1, I think it picks center-1
20:54:26markunDidn't find any problems. Did you take a look at the code?
20:54:42amiconnNo, I don't mean problems, but how does it behave
20:54:44rasherof course, I'm nitpicking
20:55:14rashermarkun: I have a problem here
20:55:23rasherout of 7 lines, it's picking line 3
20:55:27rasherwhich is clearly wrong
20:55:47amiconnWith an even number of lines (e.g. 4), it should imho stay at center + 1/2 (3) when scrolling down, and on center -1/2 (2) when scrolling up as long as possible
20:56:20rasheractually I disagree.. shouldn't it allow you to see most of what's coming?
20:56:48amiconnImho this would be more intuitive, showing you the direction your are moving into.
20:57:35amiconnDoing it the other way round would make the cursor line jump 1 position _in the opposite direction_ of where you are going to move the cursor
20:57:49rasheralright, I surrender
20:58:11amiconn...and just fixing it at one of the 2 possible 'center' positions would look lame, imho
20:59:08markunIt's now more or less fixed. If you move up from the very bottom of the list the cursor stays at center + 1/2. From the top it stops at center - 1/2.
20:59:43markunBut if you move from the top and it's at center - 1/2 and you go back up it stays at - 1/2. Confusing?
20:59:56markunIt was not intetional, but that how it behaves right now.
21:01:19markunSo if you go are first going down and you go back up you want the cursor to move?
21:03:02rasherwell, it's consistently staying at Center-1 now
21:03:13rasherwith an uneven number of lines
21:03:17markunNot if you go up from the bottom of the list.
21:03:30markunAh, that's not good.
21:03:48rasherworks if I go from the bottom-up
21:04:01rasheryes, something's going wrong
21:04:14rasherI'm using the iso-8859-15 font
21:04:18rasherwhich gives 7 lines
21:04:37markunI'm using one with 9 lines.
21:05:06markunWhe I go up from the bottom it is at the center. When I go down from the top it's center -1.
21:05:34rashersame thing happens for me with 7 lines
21:05:54markunI'll change the code a bit
21:06:13amiconn7 lines only on the iriver lcd? Huge font that is!
21:06:25*amiconn also has 7 lines on his archoses
21:06:43preglowSlasheri: size_t is big enough to be able to index all arrays possible
21:06:53preglowSlasheri: it's usually the same size as the address space
21:07:04preglowSlasheri: hence 32 bits for coldfire
21:07:15rasheramiconn: it is
21:11:08rasherHeh, that font gives 3 lines on recorder
21:15:09***Saving seen data "./dancer.seen"
21:20:26Slasheripreglow: great, then it's fine. I had a feeling that perhaps it's less but it was bug elsewhere in code :)
21:29:20 Join t0mas [0] (
21:37:09markunHi t0mas!
21:37:59 Quit muesli- (Read error: 148 (No route to host))
21:38:05markunok, center scrolling now works as it should.
21:42:36t0mas5 days away... and 173 new emails... excluding mailinglists :X
21:42:55t0massome people don't understand the "I'm out of office" reply I guess
21:43:30SlasheriHmm, can i load plugins inside a thread? Currently a only made the rockbox to crash :/
21:43:30rasheroh are you one of those people :(
21:43:33Slasherihi t0mas :)
21:43:59rasherauto-replies are like 25% of the spam I get
21:44:16t0masit sends just 1 replay
21:44:19rasherauto-replies to mail I never sent in the first place, that is
21:44:25t0masand only on office email
21:44:41t0masso it won't respond in forged viruses and things like that
21:44:49Slasherit0mas: sorry, i haven't had time for your plugin.. maybe i could look it more at next week if i get this mp3 playback fully working :)
21:45:06t0masnp, I was away :)
21:45:09t0masand a little busy now
21:45:16Slasheriah, okay :)
21:45:18t0masand news I missed last week?
21:45:36rasherSlasheri made crude audio output
21:46:52Slasherit0mas: it was able to play mp3 files through viewer plugins
21:47:01Slasherinow i am integrating it into rockbox
21:47:15Slasheriand implementing simple multi codec architecture
21:48:06preglowwell, everything's a thread in rockbox, afaik
21:49:25 Join webguest56 [0] (
21:49:46Slasherihmm, yep. But is it possible to create a separate thread where i would initialize plugins?
21:52:37webguest56HD FM radio, would that be possible with the H140 ?
21:52:50SlasheriAh, it was probably bug in the code :)
21:52:54preglowwhat do you mean HD?
21:53:10webguest56High Definition, digital radio
21:53:52rasherWhat on earth is that. Last I heard fm was fm
21:54:02rasherfm radio, that is
21:54:21webguest56its starting to catch on fast in the US and is known as DAB in the UK
21:54:35webguest56Digital Audio Broadcast
21:54:48rasherNot a chance
21:54:58preglowthere's just an fm tuner
21:55:19webguest56yes that and software to decode the signal ?
21:55:52preglowDAB isn't even transmitted in the same frequency band
21:55:56preglowso there's no chance in hell
21:56:22amiconnThe iriver only has a one-chip fm receiver with analog outputs
21:56:43webguest56ah its the analog outputs which are the killer ?
21:56:54preglowamiconn: you can record fm
21:57:01amiconn...and yes, DAB uses different frequency bands than fm
21:57:04preglowamiconn: theoretically, iriver disabled it, they say
21:57:23preglowwebguest56: no, the chip being an FM chip is the killer
21:57:30amiconnpreglow: Yes, it's just the same as recording from an arbitrary analog source
21:57:32preglowwebguest56: there's no possible way for this to work
21:57:37preglowamiconn: aye
21:57:43webguest56ta, thank you
21:58:02preglowwould have rocked, though
21:58:09preglowwe've got more than enough juice to decode mpeg layer2
21:58:38amiconn..yet slower than layer3 ;)
21:59:46preglowyes, i don't quite get that yet
21:59:58preglowthey utilize different routines for decoding
22:00:17preglowand i haven't optimized the layer1/2 one the least, but hell, there's not that much fancy going on
22:00:24preglowi've already optimized the subband part, which they share
22:06:42preglowdab was developed in the 80's, layer 2 must have been cutting edge then
22:09:55SlasheriHmm.. Now i got the codec running on thread but it's really slow :/
22:26:48tucozI think libsidplay will not work for rockbox, as the m68k-elf-gcc don't support c
22:26:58tucozc plus plus
22:27:23amiconnI think you would need to build m68k-elf-g++ for that
22:27:39tucozamiconn, I think so too
22:27:42amiconnThe suggested build procedure sets −−enable-languages=c
22:28:57tucozyes, I've know.
22:29:32tucozI meant, I know
22:29:45preglowyou don't need g++ for that, all the g++ command does is add a few libs
22:29:48preglowlibs we don't want to use
22:29:57preglowgcc should handle it
22:30:26amiconnOkay, but then gcc needs to be built with c++ support, correct?
22:30:27tucozbut when I try to compile a .cpp file with m68k-elf-gcc it says cplusplus is not supported
22:31:19 Quit tucoz ("CGI:IRC (EOF)")
22:32:25 Join tucoz [0] (
22:33:37tucozBut, I do not think that the core devs would want to add c++ support for .sid's to work
22:41:14amiconnWell, I would not want to add c++ code to rockbox in general
22:42:08 Join belgarath [0] (
22:42:46belgarathhas anybody managed to get onto
22:43:27 Join godzirra [0] (
22:43:49godzirraHowdy guys. Just want to say I think your projects rule, but I was curious if any more progress has happened on the iRiver.
22:44:06godzirraI'm using an ihp120 and am eagerly awaiting a release for it ;)
22:45:06tucozbelgarath, what's wrong with misticriver? The webpage loads fine here
22:45:36belgarathok thanks I'll see what I can do about it
22:47:03belgarathweird it doesnt load in ie or firefox but does on aol software...
22:48:15rasherI think it's just having a hiccup
22:48:35t0masI'm away again :(
22:48:47t0mas5 days w/o rockbox... and now I'm away again untill sunday somewhere
22:49:16 Quit t0mas ("bye")
22:50:09 Part MoosCamaro
22:50:38 Quit belgarath ("CGI:IRC")
22:53:17 Join Yokalosh [0] (
22:53:17 Quit markun (Read error: 104 (Connection reset by peer))
22:54:13YokaloshGuys can i ask a q plz?
22:54:38webguest56godzirra, could I suggest that you read yesterdays and todays logs in their entirety :)
22:55:03Yokaloshi just flashed my jbr 10 and the speeds are amazing now
22:55:52Yokalosherm, with the old firmware when i was out and about my recorder occasionaly stooped playing
22:56:02Yokaloshand it was getting more and more frequent
22:56:17Yokaloshwill the rockbox stuff fix this problem or what?
22:57:07 Quit webguest56 ("CGI:IRC")
22:57:49godzirrawebguest56: Oops.. will do :)
22:58:48Yokaloshhello? can anyone help me?
22:59:56rasherYokalosh: stopped playing or turned off?
23:00:40godzirraSo seems like playback on the iriver still isn't running?
23:02:05 Part tucoz ("Leaving")
23:15:11***Saving seen data "./dancer.seen"
23:18:16 Quit Yokalosh (Read error: 110 (Connection timed out))
23:18:40preglowseems to be a strong anti c++ sentiment around here :>
23:23:57SlasheriHmm, anyone else encountered slow code execution inside a thread?
23:27:56preglowhow slow?
23:28:16Slasherivery slow, realtime mp3 playback not possible
23:28:22SlasheriI am not sure what is causing this..
23:28:50preglowthings not yielding correctly?
23:29:31Slasherinot sure.. in the "main thread" i have while(1) yield(); after create_thread
23:32:57 Join webguest37 [0] (
23:33:00 Part godzirra
23:33:13Slasherigood night.. more coding tomorrow :) ->
23:34:13 Quit webguest37 (Client Quit)

Previous day | Next day