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 2006-12-11

00:00:00peturdoes weird things here on my h340 with latest cvs
00:00:47aridesei haven't seen a lot of open source code, how does rockbox stack up?
00:00:56arideseit seems like there's almost no comments...
00:01:20aridesethen again, i comment every single line of my code
00:01:45 Quit Blork (Read error: 54 (Connection reset by peer))
00:19:39 Quit Sumguy21 ()
00:23:06 Quit petur ("Zzzzz")
00:29:26 Quit ender` (" Never say "Oooops" ... always say "Ahhh, interesting..."")
00:31:34 Quit Arathis ("[rl_bot quit]")
00:36:09 Quit noteventime (Read error: 104 (Connection reset by peer))
00:36:15 Join bawb2 [0] (n=bawb2@
00:37:44 Quit CriamosAndy ("( :: NoNameScript 4.03 :: )")
00:45:30JdGordonanyone know a decent gui flac->ogg/mp3 encoder? cbf with the cli
00:46:05 Join zivan56 [0] (
00:50:30pixelmamorning Jd ...
00:50:45JdGordonhehe morning again
00:50:59JdGordonI feel no better than I did 2 hours ago
00:51:47pixelmashould I wait with the report then?
00:51:54JdGordonna, shoot
00:53:06 Join matsl_ [0] (
00:54:18pixelmaok.. no change with the resume position... but an observation: in current CVS the MMC activity symbol appears for a short time after rebuffering (pure playback) - amiconn explained to me that this is a sign that the resume position is written...
00:54:34pixelmathat doesn't happen with your patch
00:54:59 Part matsl_ ("No reason")
00:55:19linuxstbJdGordon: You said you didn't want the command-line, but oggenc converts directly from FLAC to Ogg/Vorbis and maintains the tags.
00:56:00 Quit matsl ("Leaving")
00:56:26pixelmaJdGordon: and also very important - voice doesn't work anymore (that was also the case with yesterday's patch - buffer is big enough for the voice file to fit and the voice settings are displayed correctly
00:56:29 Quit MarcoPolo ("Bye !")
00:57:22JdGordonlinuxstb: i know, but im too tired to think to do the whle thig properly... how do i get it to convert every .flac under ~/flac/? (and even better, copy the .ogg intot he same folder structure except under ~/ogg/ ) ?
00:57:58JdGordonpixelma: ok ta, voice not working is a problem
00:58:23pixelmaand with the LCD mode "inverse" setting even weirder things are happening than with last week's patch
00:59:04*JdGordon found out that the h300 lcd can flip display on both axis' at the same time... makes for a very ineteresting display
00:59:37 Join pike2k [0] (
01:00:08pixelmawhile still in the settings screen it gets inversed correctly... go back one level - still inverted but line selector changes to pointer selector again - you can even set it to line selector again in the specific submenu...
01:00:31pixelma...going back to file browser/wps - the inversion disappears...
01:01:04pixelma...enter the settings again it shows that "inverse" is selected
01:01:22JdGordonsounds like fun :D
01:01:45 Quit pike2k (Client Quit)
01:01:51pixelmayup - that's what I thought :P
01:02:07 Join pike2k [0] (n=user@xbmc/staff/pike)
01:03:06***Saving seen data "./dancer.seen"
01:04:55pixelmathe font browser is ok now - also the statusbar drawing after loading a theme
01:05:45pixelmaand the new build picked up wps/font/lang/fmr from the config.cfg correctly
01:08:00 Join Arathis [0] (
01:10:54JdGordonpixelma: shuold that invert option be live? or only take effec when exiting the item?
01:11:47pixelmait is live in current cvs
01:12:19pixelma(and works)
01:12:37 Quit bawb2 (Remote closed the connection)
01:13:40 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
01:14:44 Join barrywardell [0] (
01:15:06 Part barrywardell
01:16:00 Join Thundercloud [0] (n=thunderc@
01:25:03JdGordongrr... LCD Mode setting doesnt work in the sim
01:25:17JdGordonand voice doesnt work on my h300 also...
01:26:22 Join Paul_the_Nerd [0] (
01:27:45pixelmaJdGordon: I'm not sure but I think voice was alright with the patch of December 2nd
01:28:22JdGordonI tihnk it was, because Im not gettin any slowness when the menu opens (which was happening when the voice was loading before)
01:28:35 Join barrywardell [0] (
01:29:12 Join l0de [0] (
01:29:29l0deHello sirs, has there been any progress towards rockbox on 80gb 5.5g ipods?
01:30:44Soapnone that I am aware of, check the New Ports forum.
01:31:56l0deI can't access forums from my work, unfortunately.
01:32:04l0deBut I'm sure you'd have heard if they had :\
01:32:26 Join bawb2 [0] (n=bawb2@
01:33:18Soapyea, my work is like that too.
01:33:41l0deAnything that includes the word "Forum" seems to get filtered out
01:33:44l0deBut I can irc ok
01:33:46*linuxstb works from home ;)
01:34:02l0deI want to put a proxy on my home computer so I can check stuff from work
01:34:15l0debut aren't sure how to do it on an xp system, and unwilling to learn and install linux
01:34:21 Join safetydan [0] (i=cbca159f@gateway/web/cgi-irc/
01:34:42safetydanamiconn: if you're looking for fixed point functions try
01:35:07safetydanThe code is GPL and has arm and m68k versions.
01:35:34safetydanThe header is also important
01:39:07 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
01:43:17barrywardelldan_a: any progress with crt0?
01:43:33JdGordonapart from resume info, which settings shhuold be saved every time the disk spins? (as apposed to only being saved when explicitly told to save)
01:44:36Arathisbarrywardell: appending "/usr/local/lib" didn't help my problem btw. (compiling arm-efl-gcc with got still the same error
01:45:57linuxstbbarrywardell: Do you have xcode set up to compile Mac OS "fat binaries" (or whatever the dual-architecture binaries are called) ?
01:46:47barrywardelluniversal binaries? yes, I did last time I checked anyway...
01:47:38linuxstbIt could be useful if you could compile ipod_fw.c (and maybe diskdump, after I add 5.5g support to it). But no rush - no-one's complained about the lack of mac-on-intel support so far...
01:48:30barrywardelli'll look into it. it's been quite a while since i've used xcode
01:48:48linuxstbI've never used it...
01:51:38 Quit JoeBorn (Read error: 110 (Connection timed out))
01:52:14JdGordonpixelma: does the resume info (and other settings) get saved correclty on a clean shutdown?
01:52:53pixelmayes they do (as far as I could see)
01:52:59JdGordonok, good
01:54:57 Quit obo ("bye")
02:00:51 Quit zivan56 ()
02:06:25 Join JoeBorn [0] (n=rootmeis@
02:08:12 Quit Arathis ("[rl_bot quit]")
02:12:54barrywardelllinuxstb, dan_a: here are the changes I need to make to crt0-pp.S to get rockbox to boot with the new bootloader crt0 on my h10 -
02:14:54 Quit niskel (Read error: 104 (Connection reset by peer))
02:14:56barrywardellthe only real change necessary is invalidating the cache when the cop wakes up and making the cpu wait for the cop to be asleep
02:15:00Ploujso, would it be a bad idea for me to buy my (non-geeky/technical) brother an iriver and install rock box on it since it's not stable/finished/feature-complete yet?
02:15:33Paul_the_NerdDepends on which iRiver you're referring to
02:16:14Ploujwell, someone suggested that iriver h320 is similar to ipod in price and functionality (of rockbox), so that one
02:16:27Plouj(ipod nano)
02:16:40Paul_the_NerdThe functionality of Rockbox on the H320 is much more reliable than the iPods, overall
02:16:58Paul_the_NerdBetter battery life than the original firmware, and the music is less likely to skip while doing things
02:18:25Ploujwhat are the downsides of installing rockbox on h320 then?
02:19:04Paul_the_NerdRockbox doesn't play WMA files, Rockbox doesn't have video, and if it's a US H320, Rockbox removes the ability for it to play DRM'ed music (as well, switching to a firmware that supports video does that as well)
02:19:46Ploujand recording, I heard, right
02:19:59Paul_the_NerdWhat about recording?
02:19:59Ploujconsidering that those features are not what my bro needs, it's good
02:20:10Paul_the_NerdRockbox has far better recording functionality than the retail firmware...
02:20:24Ploujthat's not what I heard in this channel, and read on the wiki
02:20:39Paul_the_NerdWhere in the wiki?
02:20:44Paul_the_NerdAnd who told you that?
02:21:30Paul_the_NerdThat page is over 7 months old
02:21:46Paul_the_NerdActually, closer to 9ish
02:21:55Ploujyeah, I've been told that
02:21:56 Quit ptw419 ()
02:22:05Ploujwell, I'll take you word for it then
02:22:20Paul_the_NerdYou could just read the manual.
02:22:27pixelmaPaul_the_Nerd: if you read the forums and latest logs... there seem to be some problems with recording on swcodec targets atm (H3x0 and X5 seem to be affected at least)
02:22:32Paul_the_NerdIt covers what functionality Rockbox has on the H300, and you could decide if that's suitable to your brother.
02:22:53Ploujmanual, eh
02:22:55*jhMikeS is reading the earlier logs and his head is going to explode
02:22:56Paul_the_Nerdpixelma: Yes, but that's not relating to the recording features, just a small glitch that occurs in long recordings, right?
02:23:06Paul_the_NerdThe dropped samples things?
02:23:15Paul_the_NerdOr is this the "strange button combinations can cause it to freeze" one?
02:23:27Paul_the_NerdEither way, it's not a matter of what Rockbox supports so much as a recent bug that'll get ironed out. :)
02:23:50jhMikeSYou just get a spontaneous stop and 1 Dec doesn't stop?
02:24:34pixelmajhMikeS: to whom was the question directed at?
02:25:27jhMikeSpixelma: I saw some discussion in the logs. Anyone that can clarify really. x5 and H340 is the issue?
02:27:48jhMikeSamiconn, peter and bger were talking about it
02:28:02pixelmathat's what amiconn reported (he did some tests) - I was there when he tried on an X5 and it seemed to occur reliably just after a non-predictable time - that's all I can say as I don't have one of those myself
02:29:58jhMikeSI'm just wondering if the problem actually starts on 6 Dec with the slight change in the DMA stuff
02:29:59 Join Thundercloud_ [0] (n=thunderc@
02:30:19pixelmathat was yesterday and today before noon (GMT) don't lnow if that was tested after your last commits
02:30:25JdGordonI need to zero a file, possibly fairly frequently, is it better to delete it completly or just truncate it to 0 ?
02:33:10jhMikeSyeah, in the 23:00's in the logs...but I'm sans x5 for a bit so can't check it out :(
02:38:59 Join Daya [0] (
02:40:06jhMikeSif no peakmeters and no increase in the size/time then I suspect DMA is just stopping now (guess it was the 21:00's :)
02:43:21 Part safetydan
02:43:57JdGordonpixelma: ok, I think I've fixed the resume info problem
02:44:07JdGordonnot 100% sure it wont thrash the disk again tho
02:45:52pixelmaI first thought it sounded good but reading your second statement - I'm not so sure anymore...
02:46:53barrywardelldan_a: i think my crt0 and bootloader updates are good to go into CVS. do i still need to special case for 3g ipods? will I just use the existing code?
02:47:18 Quit Thundercloud (Read error: 110 (Connection timed out))
02:47:26jhMikeSWhy's it show starting around 21:00 on 10 Dec if it was around noon? Today's log shows only up to ~02:45 *confused*
02:47:50pixelmaJdGordon: of course I'll test a new patch but not now - need some sleep
02:50:27 Part pixelma
02:57:40 Quit l0de (Connection timed out)
02:57:49 Join Pav [0] (
02:59:24 Quit Daya ()
03:01:53 Quit Pav (Client Quit)
03:02:03*jhMikeS is just going to try a blind DMA adjustment as long as the H120 takes it ok and hope for the best
03:02:15 Join pavel_ [0] (
03:02:37 Join theone [0] (i=43a26910@gateway/web/cgi-irc/
03:03:08***Saving seen data "./dancer.seen"
03:06:09 Quit JoeBorn (Read error: 60 (Operation timed out))
03:16:43jhMikeShrm, just got the stop here on H120....rrrr
03:17:05 Quit theone ("CGI:IRC (Ping timeout)")
03:17:06 Join JoeBorn [0] (n=rootmeis@
03:20:39jbahey jhMikeS
03:20:46jbayou see my message in the logs?
03:21:55 Join Brins0 [0] (
03:22:15Brins0heya all
03:22:19jbabasically it was a naughty yield, that wasn't yielding in the while loop at the end of audio_stop
03:22:32Brins0just need to ask this before I continue...
03:23:03Brins0does anyone know where to grab the archos cinema/podcast plugins?
03:23:25Brins0I was told some of you experts may know
03:24:11linuxstbRockbox is a replacement firmware for various mp3 players, *not* including the current Archos range....
03:24:14jhMikeSjba: yeah. good to hear you found the problem
03:24:22Brins0ah, k
03:24:29Brins0was worth a try, thx anyway
03:24:40jbacool, just thought I'd FYI you
03:24:53 Quit Brins0 (Client Quit)
03:28:10 Quit Thundercloud_ (Remote closed the connection)
03:30:51 Quit funky ("leaving")
03:30:59 Part Paul_the_Nerd
03:32:06jhMikeShmmm...DMA stop happened to me once but at 2m 39s precisely when a disk flush starts
03:35:40jhMikeSand a I had a complete and valid recording indicating a stop due to an error: DSR1 & 0x70 != 0
03:40:08barrywardelldan_a: do you have your 3g handy for testing?
04:02:50 Quit nudelin ("Hammer of Dawn is offline.")
04:07:27 Join [sellout] [0] (
04:10:56 Join nudelyn [0] (
04:14:13 Join jaebird [0] (
04:22:42 Join BHSPitLappy [0] (
04:30:32 Join Paul_the_Nerd [0] (
04:35:45pavel_I just want to thank yall for considering trying to find firmware replacement for the zune!!
04:36:59 Quit pavel_ ("Chatzilla 0.9.77 [Firefox 2.0/2006101023]")
05:02:26 Join yohannmisquitta [0] (
05:03:09***Saving seen data "./dancer.seen"
05:04:50 Quit jaebird ("Leaving")
05:06:59 Part yohannmisquitta
05:20:57 Join jaebird [0] (
05:23:53 Join DraX [0] (n=alex@xmms2/developer/DraX)
05:24:07DraXis aac realtime on iAudio X5L ?
05:32:01Paul_the_NerdSome AACs are, I'm not sure if all are.
05:34:26 Join scorche [0] (n=scorche@rockbox/administrator/scorche)
05:34:35DraXPaul_the_Nerd: some?
05:38:23Paul_the_NerdAs far as I know, very high bitrates may not play realtime, but 128kbps ones should.
05:39:28 Quit aridese ()
05:42:18DraXso are the fonts in the WPS screenshots as ugly as they look? :)
05:47:39Paul_the_NerdI can't speak to the artistic tastes of various WPS authors. I use as a primary a black and white player with a text-based WPS, so it's pretty clear I'm not that concerned about aesthetics. :-P
05:47:48Paul_the_NerdThey aren't antialiased, if that's what you mean
05:47:55DraXyeah basically
05:48:33Paul_the_NerdFonts are 1-bit, bitmapped, so antialiasing isn't really an option, nor is it considered worth bothering with.
06:02:12JdGordonbarrywardell: do you need the new crt0 tested on the nano? or already done?
06:04:24 Quit bawb2 ("seacrest out!")
06:04:48 Join bawb2 [0] (n=bawb2@
06:14:23 Quit _Veseliq_ (Remote closed the connection)
06:17:22 Quit BHSPitLappy (Remote closed the connection)
06:19:44 Join BHSPitLappy [0] (
06:20:49 Quit bawb2 ("seacrest out!")
06:23:33 Join Mahesh [0] (i=3a444a52@gateway/web/cgi-irc/
06:24:30MaheshHi every body, I have a problem with my 6GB IPOD mini 2nd generation
06:25:02Maheshwhen i attached it to PC with USB it continuosly displays Apple and batter symbol with "!" symbol
06:25:16Maheshcan any body help me, what does this mean?
06:27:01 Quit BHSPitLappy (Read error: 104 (Connection reset by peer))
06:29:49 Quit JdGordon ("Konversation terminated!")
06:38:52 Quit perpleXa ("Leaving")
06:41:15 Quit Mahesh ("CGI:IRC (EOF)")
06:56:57 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
06:59:35 Join dantheman [0] (n=danielma@
07:00:56 Join slimeball [0] (
07:03:12***Saving seen data "./dancer.seen"
07:09:51 Join lee-qid [0] (
07:10:04 Quit lee-qid (Remote closed the connection)
07:13:23 Join bawb2 [0] (n=bawb2@
07:16:24 Part Paul_the_Nerd
07:17:53ashescan i reflash my old rca lyra with linux?
07:20:01jbahey any i300 or i100 hackers in here?
07:20:14jbahave a question about the cyprus usb chip initialization
07:21:07 Join webguest60 [0] (i=42d75940@gateway/web/cgi-irc/
07:23:25 Quit dantheman (Read error: 104 (Connection reset by peer))
07:24:52webguest60ipodpatcher isn't finding the pod; ipod works and shows up as a drive in win2k - any ideas?
07:26:54 Quit webguest60 ("CGI:IRC (EOF)")
07:26:59 Quit Slasheri (Read error: 131 (Connection reset by peer))
07:29:21 Quit mathgl ("Quitte")
07:35:06 Join ender` [0] (n=ender@
07:43:32 Quit jba ("Leaving")
08:05:38 Join MarcoPolo [0] (n=MarcoPol@
08:06:31JdGordonanyone around who knows the voice system? I need to track down why voice is not working, the voice menus setting is set and there is an english.voice file in the proper place... where do I start looking?
08:12:20 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
08:22:02jhMikeSJdGordon: You mean it stopped working in CVS?
08:24:40*jhMikeS might be able to get an idea with some input :)
08:28:41amiconnjhMikeS: Found my recording reports in the logs?
08:29:03jhMikeSyes. I know why (DMA error) but I don't know __why__
08:29:24amiconnStrange thing is that it works on H1x0
08:29:44jhMikeSI can make it more frequency by lower the interrupt priority. It run a few seconds and comes up with a configuration error
08:29:49amiconnI did not yet try to test-build for H1x0 w/o s/pdif support
08:31:07jhMikeSSetting DMA1 at level3 prio 3 (not same as timers), I get DMA err: 41, DAR1: 310EC290 (pcm_buffer address), BCR1: FF000000
08:31:23jhMikeSwonder why not a bus error or something
08:31:49jhMikeSBut it's most likely due to setting it at level 6
08:32:46amiconnI did a test recording of ~100 minutes on H140 y'day, which worked
08:33:07amiconn(line in)
08:33:42jhMikeSI tested right after reading the logs and it happened to me on the H120, but just once. It's definitely a DMA error from the description.
08:34:00amiconnYou mean you got an UIE screen?
08:34:26jhMikeSDSR1 & 0x70 type
08:35:44jhMikeSThat type of problem will post PCMREC_STOP and stop DMA transfers
08:36:13jhMikeSI just found a way to make it happen reliably within a few seconds
08:38:31jhMikeSagain I ask: why would that result in configuration errors that are supposed to occurr for alignment and size problems esp. with DMA_AA?
08:38:51amiconn0x41 means bus error on destinations. That doesn't make sense....
08:39:02 Join egotrippen [0] (i=c7616228@gateway/web/cgi-irc/
08:39:43jhMikeSDatasheet say CE is bit 6 (0x40), no?
08:39:55amiconnAh, yes
08:41:02 Quit Ensgabe ("Leaving")
08:41:03jhMikeSDatasheet also says if BCR24BIT = 1 in MPARK then bits 24-31 are supposed to be zeros, not 1s
08:41:22jhMikeSBut I'm seeing 0xFF000000 in BCR1
08:41:24 Join mathgl [0] (n=mathgl@
08:44:00amiconnI can think of 2 possible causes for a CE even with AA. (1) Destination isn't 32 bit aligned. Iiuc this is necessary even with AA and dest size == line, because source size is 32 bit
08:44:19amiconn(2) BCR is zero when trying to start a transfer
08:44:38jhMikeSIt's passed a constant from recording of 8192 bytes
08:44:58jhMikeSThe pcm buffer is 16 byte aligned as well
08:46:09jhMikeSI did try starting a transfer for kicks with BCR = 0 and just got a freeze
08:46:41amiconnDid you check SAR and DAR when the error happens?
08:47:05amiconn(and DCR too, since it is now changed with |= and &=
08:47:05jhMikeSI just gave those numbers...see above
08:47:09egotrippenhey all
08:47:32egotrippenit's been a while since i've been here, is it alright if i ask some 'current events' stuff?
08:49:30egotrippenis a stable 3.0 release still a goal? or a new iriver h300 bootloader that recognizes the ac plug?
08:49:54amiconnI wonder whether DCR1 &= ~DMA_EEXT; in the ISR might be the problem. This isn't atomic, same as DCR1 |= DMA_EEXT; in pcm_record_more()
08:50:15egotrippenhow's the direction of the project, and stuff?
08:50:44jhMikeSMight try a or_l and see
08:50:53amiconnYou could try to replace those with and_l / or_l
08:52:41jhMikeSPlayback has always been using |= and &= ~ as well. What would get in there and bother it anyway?
08:55:52jhMikeSnope, same CE bit thing
08:57:17 Quit bawb2 (Read error: 131 (Connection reset by peer))
08:57:18 Quit dan_a ()
08:59:08jhMikeSI'm wondering if something, because of lowering the priority, isn't getting more opportunity to clobber something. If it's that, I can't find it yet.
09:00:28JdGordonjhMikeS: soz, no its woring in cvs... its not working in my patch
09:00:37JdGordondon really know where to start looking
09:02:37LinusNegotrippen: i'm planning to update the bootloaders, but i am too busy atm
09:03:15***Saving seen data "./dancer.seen"
09:03:51LinusNi'm planning to do the rtc mod, but with a more common RTC chip with a crystal
09:04:06LinusNi.e not internal
09:04:23JdGordonh300 doesnt need an rtc mod... ?
09:04:25jhMikeSJdGordon: Make sure it loads the voice file for real to start with. Check talk.c out, maybe some logf-ing
09:04:34JdGordonok thank
09:04:42LinusNJdGordon: who said anything about h300?
09:04:56JdGordon <egotrippen> is a stable 3.0 release still a goal? or a new iriver h300 bootloader that recognizes the ac plug?
09:04:59*jhMikeS thought there was no room in there
09:05:41JdGordonLinusN: oh hmm... i thought your 2 meesages were follow on... dw
09:05:47LinusNJdGordon: ah...
09:05:52*JdGordon backs away slowly
09:06:03JdGordonthat h100 rtc mod looks awesome tho
09:06:07egotrippenok, thanks for the info
09:06:10 Nick myzar is now known as myzar|away (n=myzar@
09:06:27jhMikeSwhere's a link to it?
09:06:35JdGordonle wiki
09:06:47 Join simple [0] (i=cba20392@gateway/web/cgi-irc/
09:08:01simplehelp me
09:08:22*Bagder sighs
09:08:25Bagderthen ASK!
09:08:53simplea.... very esye quastion
09:09:07simpleeasy quastion pa
09:09:32*Bagder tries his mind reading abilities but fail
09:10:16egotrippenyou guys haven't implemented the ESP function yet?
09:10:24 Join petur [0] (i=d4efd6a6@gateway/web/cgi-irc/
09:10:37 Nick myzar|away is now known as myzar (n=myzar@
09:11:40LinusNsimple: ASK YOUR QUESTION
09:12:06peturbah... yelling in the morning... yawn
09:12:21*LinusN needs more coffee
09:13:15simpleipop A/S...
09:13:45*muesli__ nudges LinusN..hand some over pls!
09:14:25 Quit simple ("CGI:IRC")
09:14:47*Bagder shakes head
09:14:54*jhMikeS guesses he's up schitz creek and should just set the DMA1 priority back since it works
09:16:28LinusNegotrippen: question is which would have been the most effective in this case, ESP or EMP
09:16:58egotrippeni'd settle for a barrista function
09:17:09amiconnjhMikeS: Would be interesting to see SAR1, DAR1 and DCR1 values when the error happens
09:17:19LinusNi wonder, was "ipop" maybe "ipod"?
09:17:29egotrippeni'd guess
09:17:43jhMikeSamiconn: I gave 2 out of 3 them to you but I'll get them all. Gimme a minute
09:17:43LinusNbut i don't know about the "A/S"
09:17:43egotrippenor maybe it was a knockoff brand, and he wondering why RB wouldn't work
09:17:57*petur has some Kókómjólk - or rather wished he were in a place where they had plenty of it
09:18:03LinusNwould have made sense if Apple was a danish company :-)
09:18:26 Quit mathgl (Remote closed the connection)
09:19:11 Join safetydan [0] (n=dan@rockbox/developer/safetydan)
09:21:21amiconnjhMikeS: 2? I only found 1: DAR1
09:21:38 Join _Veseliq_ [0] (n=veseliq@
09:21:55egotrippenso, i'm still curious... is there a desire for the 3.0 release? i kind of fell out of things when that never materialized, and when the forums changed format
09:22:03egotrippenwhat are the project goals, nowadays?
09:22:14Bagderegotrippen: there isn't much work on getting a 3.0 done, no
09:22:31Bagderthe goals are still the same
09:22:43Bagderjust no particular work done on getting a release out
09:23:23egotrippenare there goals for more 'official' releases on other platforms? i mean, obviously the current system is working well enough for most people
09:23:36egotrippeni've been using a H300 build pretty much since audio started working on it
09:23:45LinusNgoals, yes - har work to meet them, no
09:23:54Bagderthere's none interested in being "release manager" and I doubt we'll manage to release without that
09:23:58jhMikeSamiconn: hehe...I read again and realized that :)
09:24:21Bagder... seeing how we failed the last time we'd have to make it differently the next time
09:24:39petura bugfixing period would be good tho... I've recently started looking at some to fill up the little bits of spare time
09:24:50jhMikeSamiconn: DSR1: 41, SAR1: 80000074, DAR1: 310EC3A0, BCR1: FF000000, DCR1: B00E7FFF
09:26:03Bagderpetur: I agree an official "bug fixing period" would be a good thing
09:26:29egotrippeni'm.... i'm not exactly phrasing this right, but i'm kinda sorry to hear that there's not more of a concerted effort
09:26:29 Quit petur ("CGI:IRC (EOF)")
09:26:38 Join p3tur [0] (i=d4efd6a6@gateway/web/cgi-irc/
09:26:39 Join My_Sic [0] (
09:26:41egotrippenseems like the farther along it gets, the harder it'd be to organize
09:26:48egotrippenseems like the farther along it gets, the harder it'd be to organize
09:26:58BagderI'm not sure about that
09:26:59 Nick p3tur is now known as petur (i=d4efd6a6@gateway/web/cgi-irc/
09:27:10egotrippensorry for the double. lag
09:27:19BagderIn my view, it is just about the very high goals we've had on "releases"
09:27:29Bagdergoals that are very hard to reach and take very long time...
09:27:42egotrippenmakes sense, yeah
09:28:18Bagderbut also because us devs don't have a lot to win on making releases as we manage quite fine with the dailies
09:28:28egotrippeni just remember a main part of the push back in may was to get a bug free release, so new work wasn't built on buggy bases
09:28:33 Join terriblyinept [0] (n=energy_O@
09:29:03 Join Slasheri [0] (i=miipekk@rockbox/developer/Slasheri)
09:29:06egotrippeni'm not a coder, so i don't have real working knowledge of the way it goes, but
09:29:17egotrippenthat sounded pretty good, y'know? :)
09:29:22BagderI agree
09:29:48Bagderbut our "bug fixing periods" tend to equal "only a few devs doing anything and the rest just waiting for the period to pass"
09:30:27Bagder(I'm not saying *I* am any better, just doing observations here)
09:30:27terriblyineptMeaning most devs write fairly good code with some bugs, and a few work to work out the kinks?
09:30:40LinusNand another problem is that the users that are supposed to test it are not running the latest version
09:31:15 Join JoeyBorn [0] (n=rootmeis@
09:31:27LinusNbut instead "waits until the bugs are ironed out"
09:31:32egotrippeneh, i'm guilty of that. i run a couple little mods that i'm not sure still apply
09:32:01LinusNso even if we fix bugs, we don't get the feedback we need
09:32:01egotrippeni keep meaning to check that, i've got a break from school coming up
09:32:25BagderI'm guilty of mainly playing music non-stop from a huge playlist using a single codec...
09:32:36terriblyineptWould there be any means, if a user kept updated, to know what sort of things should be played with to do said testing? On the user end exclusively, of course?
09:32:42jhMikeSamiconn: It's interesting too that DAR1 is still at the start of the pcm chunk. Something's perhaps goofing up the size. It should be past the end of the first chunk when it BCR1 runs down.
09:32:43LinusNand most of the unofficial patches are kludges that will never reach cvs
09:33:03LinusNmaybe not most, but many
09:33:16egotrippenone of the bigger mods i use is the 'insert albums randomly' patch
09:33:26scorchethat was committed
09:33:39peturmaybe we should mark such patches in the tracker as hacks that will not get in that way
09:33:40scorche...i think
09:33:59egotrippeni knew the one from the menu, continue to random album or something was committed
09:34:13egotrippenbut i like mine on the insert/insert next/etc. menu
09:34:36Bagderpetur: I've been toying with creating a new separate tracker to move patches that just won't get applied/committed, but I haven't yet sorted out what to call it nor how to use it...
09:34:50Bagdertoying with the idea that is
09:35:32peturanother idea I had was to have a textbox on the front page to enter a FS number in for quick access to the tracker
09:35:39egotrippenBagder: maybe 'mods' versus 'patches'?
09:35:46egotrippenor 'potentials'?
09:35:57egotrippenthat sounds like a good thing to mark
09:36:16Bagderyeah, but there needs to be a good definitions so that we don't just create another patch tracker and then have to deal with two instead of one
09:36:40terriblyineptpetur: I like that idea as a form of contribution, but as Bagder says, it seems troublesome to implement. Would it be worthwhile?
09:37:21egotrippencould it just be added as a marker on the current tracker? you'd need someone official to decide item to item, i guess
09:37:52 Join webguest20 [0] (i=c27f0812@gateway/web/cgi-irc/
09:38:29LinusNthe thing is that we don't have an official standpoint on what type of patches are "cvs material" or not
09:38:52LinusNthe idea might be very good but the implementation may suck
09:39:02webguest20I have a problem under cygwin. When I execute ../tools/configure I get the following error message: './tools/configure: line 22: syntax error near unexpected token `{
09:39:05 Join _FireFly_ [0] (
09:39:13egotrippensounds like you could use someone on the organizational end
09:39:15 Quit JoeBorn (Connection timed out)
09:39:20webguest20Any ideas why?
09:39:28egotrippeni dunno how you'd appoint someone to that position, but it might be worth looking in to
09:39:56webguest20The line 22 in configure contains "input() {". Nothing special...
09:40:27LinusNwebguest20: what shell are you using? is cygwin configured for DOS or UNIX line endings?
09:42:16webguest20LinusN: I think it's bash. Everything worked until the last CVS update. I've built RB before without problems.
09:43:06webguest20LinusN: I also tried sh < ../tools/configure −−> same error message
09:43:10LinusNi don't recall any changes in tools/configure lately
09:43:43Bagdernot since Nov 27
09:44:04webguest20LinusN: yes, that disappointed me as well. How can I check which line endings are in effect?
09:44:06BagderI blame cygwin somehow
09:44:28preglowone should always blame cygwin
09:44:32LinusNcygwin did mess something up with the shell lately
09:45:05LinusNbut the configure script still works fine for me
09:45:17webguest20Bagder: you may be right. I updated cygwin few days ago since I saw the message "update to make 3.81 to speed up builds" (or something like that) But IIRC I've built RB since then
09:46:05webguest20Would you recomment to update it back to an earlier version? If yes then which?
09:46:24preglowamiconn: :)
09:46:31Bagderline 22 also happens to *NEVER* have changed...
09:46:41Bagderit is from the original 1.1 configure :-)
09:47:29LinusNi recommend finding out what the problem is
09:47:31preglowamiconn: wouldn't it be best if the core and plugins shared the same code somehow? i don't like duplicated code...
09:48:31JdGordonegotrippen: does that insert folders randomly patch still apply cleanly? I'd love to have it commited but the otheres here dont see the need, or something...
09:48:36jhMikeSpreglow: agreed. more to maintain and debug imho
09:49:44egotrippeni don't know, i'll be able to check it soon
09:49:49egotrippenafter finals week =-)
09:49:58 Part terriblyinept
09:50:00egotrippeni love having the functionality, i hope you can get it through
09:50:23 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
09:51:44 Part webguest20 ("Will try to solve the problem and also report back if I'll succeed")
09:52:18JdGordondoes anyone like the idea of quickly going through the patches on FS and deciding as a group which features would never be commited, and which might be if done better, and which could go in whenever a dev has time?
09:52:27egotrippenit seems pretty self evident to me, RB being playlist based, that if the functionality should be included at all it should be in the context of playlist creation
09:52:30JdGordoncut the open patches count down
09:52:44peturI'd rather cut the open bugreports down
09:52:51peturmuch more important
09:53:08JdGordonbut not as easy to close them
09:53:14JdGordon_work_ needs to be done :p
09:53:33*JdGordon doesnt even know where to start on most of the bugs
09:53:35preglowJdGordon: the patch tracker isn't only there as a middle-stage before commiting
09:54:05preglowi think it should contain all kinds of patches people might find useful, but won't be commited anyway
09:54:12preglowperhaps we could make a 'wont-be-commited' flag
09:54:45LinusNi'd like to see the patch tracker as a to-be-committed only tracker
09:55:24Bagderperhaps one patches-for-inclusion and one patches-for-fun
09:55:25LinusNbut possibly with a "useless" status (in bluebrother lingo)
09:55:38preglowi think we definitely should have a patches-for-fun place
09:55:47LinusNpreglow: i agree
09:55:54JdGordonBagder: I tinhk you'll find, most people would submit their patches in the -for-inclusion section...
09:56:10Bagderyes, I would also say that they should submit them there
09:56:19Bagderthen we could move them over when we decide to
09:56:21egotrippenlike i was saying, you'd kind of need someone to determine which is which
09:56:49LinusNthere are basically two reasons why a patch wouldn't be committed:
09:56:54JdGordoneither way.. can a category be added for "not going to be commited" untill something better is worked out?
09:57:01LinusN1) We don't want the feature
09:57:07LinusN2) The patch is badly written
09:57:22egotrippenit's the first that gets tricky though, yes?
09:57:33LinusNthe 2) case can be remedies, 1) can not
09:57:34daurnimatorhey JdGordon
09:57:38JdGordonhey daurnimator
09:57:44LinusNremedied, even
09:57:49 Join bluebrother [0] (i=RfOoBz2j@rockbox/staff/bluebrother)
09:57:58JdGordon1) just needs some stong convincing
09:58:11JdGordondaurnimator: not much.. might be getting a neuros soon :)
09:58:45daurnimatoran OSD?
09:59:30daurnimatorwhen? :P
09:59:35 Join webguest62 [0] (i=c1713011@gateway/web/cgi-irc/
09:59:45JdGordonbday is next month.. so around then probably
10:00:03 Quit webguest62 (Client Quit)
10:01:39egotrippenthe build i'm using has this bug, i dunno if you guys would know if it's fixed yet....
10:02:06egotrippenimmediately after turning the player on, if i hit 'stop' before the resumed song is fully loaded, the player locks up
10:02:13egotrippenon the H300
10:02:18JdGordonI get that as weel
10:02:40egotrippenit's not a huge deal, especially because i keep a pen around that can hit the button
10:02:50egotrippenbut i was wondering
10:03:44peturthe bug is in the tracker, there's a fix for it too
10:07:02jhMikeSpetur: my patch is hardly a real fix. that problem needs mucho work to really resolve. :)
10:07:18JdGordonthat answers my questio then :p
10:08:36jhMikeSyou'll get codec failures afterward.
10:10:37 Join spiorf [0] (
10:11:42 Quit spiorf (Remote closed the connection)
10:11:50 Join spiorf [0] (
10:13:31*jhMikeS would like to make up a proper queue_send function which would eliminate that private queue in playback.c but is racking his brain on what could be fudging up recording DMA transfers when using a lower priority :)
10:14:07peturyou mean like postmessage and sendmessage in the windows api?
10:16:08jhMikeSpetur: yes, just like that so that each message is replied to with a value in turn instead
10:19:06 Quit petur ("worrrk")
10:19:14jhMikeSright now when waiting, threads don't know who's particular message a value was set for and it could change back to a wait value during the yields even if the message posted was responded to
10:19:52 Quit miepchen^schlaf ("Verlassend")
10:21:23 Quit egotrippen ("CGI:IRC (EOF)")
10:23:34bluebrotherJdGordon: I just realized that when the settings are saved in a file resetting them doesn't work like before.
10:24:05bluebrotherhow is this supposed to work then? Should still be possible
10:24:20JdGordonwith cvs?
10:24:28JdGordonsorry, hmm
10:24:36JdGordonreset shhuold be working
10:25:15JdGordonI just reset on my h300 and it worked as expected
10:25:25daurnimatorJdGordon: so... got anything cool i can play with?
10:25:30daurnimatoreg, new player or something
10:25:32bluebrotherno, with your patch. Haven't tried it but I realized that clearing the config block should work different
10:25:45bluebrotheror am I wrong? Do you still check some cookie in the config block?
10:25:47JdGordondaurnimator: nup, you can play with fs 6399 tho :)
10:25:56daurnimatorfs 6399??
10:26:10JdGordonbluebrother: hang on.. rese from the menu or on startup?
10:26:19LinusNJdGordon: does it work with both methods of reset, i.e the Rec-on-startup and the Reset in the settings menu?
10:26:22bluebrotherI'm thinking of startup, not from a menu
10:26:32JdGordonboth shhuold be wring
10:26:46daurnimatorJdGordon: you should know i have nothign to test with
10:26:54bluebrotherafaik the rec-on-startup simply makes the config block invalid, causing Rockbox to go back to defaults
10:27:10bluebrotherbut if the settings are in a file this won't work anymore, does it?
10:27:16JdGordonboth work correctly
10:27:23bluebrotherhmm, ok.
10:27:25JdGordonno, rec-on-startup causes it to not load from disk
10:27:47bluebrotherbut that's handled in Rockbox then, not the bootloader anymore?
10:27:59LinusNit has never been handled in the loader
10:28:08JdGordonrec in the bootloader causes OF to load
10:28:09bluebrotherit hasn't?
10:28:22bluebrotherah, ok. You're right.
10:28:24*JdGordon has to go, back in a few horus
10:28:42bluebrothermaybe I should get a coffee first before thinking of such stuff
10:28:54*bluebrother wrongly believed it was the bootloader
10:29:29preglowbarrywardell: yo, what was up with the patch that made it not work on ipod?
10:29:30bluebrothersorry for the confusion
10:36:04 Join miepchen^schlaf [0] (
10:36:30 Quit miepchen^schlaf (Client Quit)
10:40:38 Quit safetydan ("Ex-Chat")
10:40:49 Quit linuxstb (Read error: 110 (Connection timed out))
10:41:05amiconnpreglow: Didn't check the tracker and hence didn't see that patch
10:41:09 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
10:41:50amiconnAnyway, the fsincos() in there is mostly a plain copy & paste from eq.c (with the addition that passing NULL for the cosine pointer is allowed if you don't need the cosine)
10:47:16 Quit _FireFly_ (Read error: 104 (Connection reset by peer))
11:00:54jhMikeSfound the problem :) yay!
11:01:12jhMikeSor did I? crap
11:01:55amiconnmarkun: around?
11:03:16***Saving seen data "./dancer.seen"
11:05:06markunamiconn: yes
11:05:23amiconnmarkun: The gigabeat does have an rtc, doesn't it?
11:05:38amiconnThe sim doesn't see the #define for it...
11:06:08markunI know, but it's fixed in our source tree and will be ported back soon
11:06:59 Join sando [0] (n=lolsteam@
11:16:56 Join Genre9mp3 [0] (n=yngwiejo@rockbox/contributor/Genre9mp3)
11:20:13jhMikeSthe critical point is whether or not DMA interrupt priority is >= timer interrupt priority (4) anyway
11:22:40 Join webguest01 [0] (i=4a6829f8@gateway/web/cgi-irc/
11:22:50 Quit webguest01 (Client Quit)
11:23:00 Join webguest01 [0] (i=4a6829f8@gateway/web/cgi-irc/
11:23:01 Quit webguest01 (Client Quit)
11:24:13jhMikeSor maybe if just greater than tick interrupts ... :\
11:26:02amiconnThat's the case in cvs, isn't it?
11:26:44jhMikeSamiconn: yes
11:27:54amiconnCvs still causes those problems on H300 and X5
11:28:24LinusNhow often is the pcf50606 irq triggered, and which level is it?
11:28:50jhMikeSit's on a tick task right?
11:29:40jhMikeSthe only functional change to the DMA1 was putting it at 6
11:29:43amiconnThe pcf interrupt triggers when the first button is pressed (adc buttons only), and again when the last button is released
11:30:12amiconn(enabling and disabling the adc scan for main buttons)
11:30:56jhMikeSOf course H120 doesn't have the pcf, so why shouldn't DMA be able to run at level 3 prio 3 and timers at level 3 prio 0?
11:31:33jhMikeSticks rather
11:32:09LinusNwhy should it?
11:32:31LinusNdma should have higher prio than timers imho
11:32:32jhMikeSLinusN: just in principle. I haven't seen anything about it that should be a problem
11:33:13jhMikeSAnd especially the face that it turns up a config error and not a bus error or something
11:33:28LinusNif you set both to 3, the dma irq can't interrupt the timer irq
11:33:39amiconnBCR1 is (effectively) zero when the error happens
11:34:19amiconnThat looks like something is trying to start the dma when BCR1 is not yet set correctly
11:34:21jhMikeSIt should have advance DAR1 by that time but hasn't, it's at the initial address
11:36:17jhMikeSrecording always sets BCR1 to 8192 while
11:39:06amiconnWhat code calls pcm_record_more() apart from pcm_rec_dma_start() ?
11:39:12amiconnThat one does look fishy...
11:39:58jhMikeSamiconn: pcm_rec_have_more in pcm_record.c. removing that call from pcm_rec_dma_start doesn't affect anything.
11:40:18jhMikeSIt's specifically there to be called by the interrupt handler
11:41:24jhMikeSI should say "*replacing* that call" with direct register writes to start it
11:44:54jhMikeSpcm_record_more should probably be ICODE come to think of it :P
11:46:03amiconnThat wouldn't change much. There's no loop in it, and it's only called once every 8KB iiuc
11:51:55*Genre9mp3 points to:
11:55:36Bagderafaik, we don't have any "check if this is the same already in the buffer and just skip back and re-use if so"
11:57:01jhMikeSamiconn: It's for use by recording/effects plugins as well so the can start the next transfer then process samples
11:57:02Genre9mp3so this was the case always you say?
11:57:06BagderGenre9mp3: indeed
11:57:21BagderI believe there's a bug report for it too
11:57:24Bagdermany years old
11:57:40Genre9mp3aha...ok... good to know
11:58:04linuxstbThat's my understanding as well. But if that's the case, why are people told to use a playlist larger than the audio buffer when testing battery runtime? Surely it doesn't matter?
11:58:08jhMikeSEven making the recording start sequence identical to the previous code still gives the same.
11:58:36jhMikeSamiconn: Did you try resetting DMA1 level back to 7 to see if the problem stops on the x5?
11:59:33jhMikeSnot able to check myself atm
12:00:10*amiconn neither
12:01:56Genre9mp3linuxstb: good point. I thought that if the playlist wasn't larger than the audio buffer, the hd would never spin
12:04:29jhMikeSsuppose I'll remove tick tasks systematically and see if there's a culprit
12:05:34linuxstbSlasheri: Can you confirm how the buffering currently works? i.e. is there any difference in buffering between playing (on repeat) a playlist smaller than the buffer size and a playlist larger than the buffer size?
12:07:49Slasherilinuxstb: hmm, afaik there shouldn't be any difference
12:08:49Slasheribuffering always tries to buffer as many files as possible
12:14:28JdGordonhow does the resume info saving happen currently? is there a function other than settings_save() which causes a flush of the config block to disk?
12:17:31Slasherithere is playlist_update_resume_info()
12:18:19JdGordonthat just calls settings_save(),which is what I thought...
12:18:43JdGordonso on a crash it is not 100% that the resume info will be current
12:19:07JdGordonany idea how often hat function is called?
12:19:15JdGordontime wise during playback i mean
12:19:44Slasherihmm, always when buffering starts
12:19:47Slasheri(= disk spins up)
12:19:56Slasheriand when stopping playback
12:20:10JdGordonok ta
12:20:24Slasherilinuxstb: is the ipod_fw.c yet working? :)
12:20:43jhMikeSlet's see: usb.c, misc.c, backlight.c, button.c, lcd-h100-remote.c are all disabled and it's been going a couple minutes. did I miss anything?
12:21:47Slasherihmm, tick_tasks are interfering with recording?
12:23:01jhMikeSSlasheri: it seems so
12:24:03 Quit muesli__ ("ich will Khe!!!")
12:24:38jhMikeSWith all those disabled it ran awhile with DMA prio low but just pooped out.
12:24:52jhMikeSBefore it would only run about 5s
12:26:01jhMikeSAre there any others on the H120? What other interrupts do we have there?
12:26:14linuxstbSlasheri: No, not yet... I'm thinking of writing my own replacement tool that edits the firmware partition directly, rather than all this messing around with extracting a partition image, then extracting firmware files, then rebuilding a partition image etc etc.
12:26:58Slasheriok :) hmm, that sounds interesting
12:28:19linuxstbThe problem is that I now understand what needs to be done, but ipod_fw.c is an order of magnitude more complex than it needs to be for Rockbox.
12:32:54jhMikeSOops, missed adc_tick
12:39:03 Join _FireFly_ [0] (
12:41:57 Join muesli__ [0] (n=muesli_t@
12:43:12 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
12:46:11pixelmaJdGordon: reading your comment on flyspray... it's not true... there is a difference between cvs and your patch - config changes and resume position are not saved during rebuffering and you cannot ignore that because on all Archos which run on NiMH/Alkaline batteries rockbox doesn't do a clean shutdown but runs until the batteries are flat...
12:46:54pixelma..(which is wanted)...
12:47:14barrywardellJdGordon: I'd be glad of a nano test of that patch if you have the time CVS resume position gets written on every rebuffer (i.e. every minute in average) - in your patch it isn't - so in worst case I have to listen to my 40 songs or something that I heard before the batteries went dry - in cvs only the last minute or something
12:49:19barrywardellpreglow: the main problem with the patch was "getting the high part of the execute adress" - it was ok on the h10/Sansa which runs from DRAM start, but on the iPod it runs from 5MB into DRAM
12:50:33barrywardellthere were other problems too. eg. on the ipod the code in crt0 has to be below 0x100 to leave room for the boot table to be written there
12:51:13amiconnjhMikeS: Other interrupts shouldn't disturb recording
12:51:37amiconnAfaiu, recording is now on level 6, which is higher than most (all?) other interrupts
12:52:23amiconnBut even if another interrupt delays the dma isr, all that might happen should be a loss of a few samples
12:52:56Slasheriin that case the recording fifo could also overflow
12:53:05Slasheriand case an another interrupt and crash
12:53:24amiconnWhy would the recording fifo cause an interrupt?
12:54:07Slasheriah, so that kind of interrupts are disabled
12:54:21amiconnAll I would expect from an overflowing fifo is the loss of a few samples
12:55:59pixelmaJdGordon: again testing CVS and patched version of the same source
12:56:22jhMikeSamiconn: well, I did disable all ticks and it's been on for 1/2 hour
12:56:38 Quit Seed (Nick collision from services.)
12:56:45 Join Seed [0] (
13:00:27 Join petur [0] (i=d4efd6a6@gateway/web/cgi-irc/
13:03:20***Saving seen data "./dancer.seen"
13:06:49jhMikeSI also left the tick interrupt going btw
13:07:42jhMikeSseems related to actually calling the tasks
13:08:11 Join austriancoder [0] (n=austrian@
13:12:09austriancoderHi all... i am looking for a volunteer for reverse engineering a mpeg2-decoder card firmware.
13:21:06 Join tvelocity [0] (
13:21:11linuxstb:) Good luck...
13:22:10BagderI'm assuming the amount of people idling here just waiting for a new project to reverse engineer is rather small...
13:25:39 Join obo [0] (
13:30:45Bagderbtw, it seems austrianmicrosystems doesn't like me
13:30:47jhMikeSICR1 = 0x8c; /* Interrupt on level 3.0 */ <== !!! huh? try level 7
13:30:57Bagderno data sheet for me
13:33:30jhMikeSwait, never mind...i'm on the phone :P
13:34:46 Join funky [0] (n=repulse@unaffiliated/funky)
13:35:08*jhMikeS obviously can't discuss jewelry and interrupts simultaneously
13:35:46linuxstbBagder: Did they reply or just ignore you?
13:35:46Genre9mp3interrupted by phone
13:35:53Bagderlinuxstb: they just ignore me
13:36:10BagderI used their online form to request it
13:36:29Bagderour friend got it within a few days using the same thing
13:36:47peturI recently mailed regarding that mp3 player but they also ignored me :/
13:38:44amiconnjhMikeS: Hmm, playback is still non-maskable (level 7)?
13:39:04*amiconn wonders what's so special about level 7
13:39:17JdGordonpixelma: oh, ok, I didnt realise that... umm.. code wise there shuoldnt be a difference, so back to he drawing board i guess
13:39:33JdGordonbarrywardell: ok, ill try to test it tommorrow
13:39:57jhMikeSamiconn: yep
13:40:40jhMikeSHaving the tick interrupt going, but not calling tasks, and DMA at level 3 (prio 3) as well is going steady for about an hour though
13:41:16pixelmaJdGordon: thanks!
13:41:28JdGordonpixelma: the fix I mentioned this morning (last night your time) would mean it would save every 30s or so, is tha ideal?
13:42:43pixelmadonn't think so - it should only take advantage of "disk spinning" during rebuffer as it does now
13:42:49 Quit barrywardell ()
13:43:15amiconnjhMikeS: I had it working stable with level 6 and no disabled tick tasks on H140 for ~100 minutes
13:43:34JdGordonpixelma: do you know what logf option there are on the ondio? or is it sim only?
13:43:40jhMikeSamiconn: I only had the problem at level 6 one time on the H120
13:48:11 Join lee-qid [0] (
13:48:51pixelmaJdGordon: that is tested on the real target. I don't know about logf options (amiconn?). But pondlife reported that nothing is written during rebuffer on H300 too - you can test yourself: play something, change some settings, let it rebuffer and then reset
13:50:35JdGordonok, ive got a logf build on my h300 now, hopefully something fishy comes up
13:51:06JdGordonok, umm... I may have just spotted the problem
13:51:21JdGordonif (ata_disk_is_active()) is called before doin any writing
13:51:29JdGordonwhich wold explain it
13:52:53pixelmacool, let me know when there is another patch to test :)
13:53:37JdGordonI'm not 100% sure how best to fix this... but at least im 99% sure the problem has ben found
13:53:39 Join Nico_P [0] (
13:55:47 Join yewei [0] (n=yewei@
13:56:09yeweihi, I create a new lang file for rockbox. How can I commit it into cvs?
13:56:17yeweior how can I submit it
13:56:21jhMikeSwonder if I should check stack pointers to make sure they're actually in bounds on context swiches
13:56:21JdGordonput it on fs
13:56:39JdGordonyewei: flyspray, our patch tracker
13:57:32*JdGordon hopes this fix doesnt invoke endless config flushing
13:57:49jhMikeSthat really wouldn't catch an interrupt going over would it? :P maybe just increase the size of the fullest one and see
13:57:58yeweiJdGordon, put it on fs?
13:58:28JdGordonwhat language is it for?
13:58:38JdGordonor is it an update?
13:59:48yeweiIt for simplified chinese
14:01:25JdGordonBagder: does anything special need to be done when adding new languages? or do we just dump the .lang in langs?
14:01:39Bagderjust add it
14:01:42amiconnjhMikeS: The stack check in switch_thread should catch any overflow
14:01:43JdGordonpixelma: so far so good. no infinite looping
14:01:45Bagderthe file that is
14:01:58JdGordonyewei: dcc me the file and ill commit ti
14:02:04amiconn(except if that overflow only writes 0xdeadbeef, which is highly unlikely)
14:02:07JdGordonunless there are objects?
14:02:34jhMikeSamiconn: what if something allocates a large buffer and doesn't use it all?
14:03:08amiconnThen all is well...
14:03:36amiconnI doubt any function allocates a buffer which it never uses
14:04:08amiconn...and as soon as it uses it once, the overflow check will be triggered
14:04:12jhMikeSEven if the buffer starts say 16 bytes before the beginning of that stack buffer but only uses say 10 bytes?
14:05:19amiconnHmm, then it won't be caught
14:05:47jhMikeSyeah, *could* be missed but I'm all on far-flung notions atm
14:05:49Kasperlewow. rockbox now uses canary words for stack protection? :D
14:09:37yeweihi JdGordon, I've uploaded the chinese-simp.lang as task 6425
14:10:16 Quit _FireFly_ ("Leaving")
14:10:30JdGordonok, i hope my computer doesnt mangle it
14:11:04yeweiok. I'm a newbie. correct me, if I'm wrong. thx
14:11:10jhMikeSall I know is without calling the functions, DMA1 is happily sharing priority with the tick interrupt and the recording screen is running
14:13:14JdGordonyewei: done, thanks
14:13:43*Bagder tries a flyspray goto-box on the front page
14:13:55Nico_Pcould any of you point me to a nice IDE to use on linux ?
14:14:06yeweithanks Jd
14:14:08Genre9mp3Go to FS in frontpage box! woa!
14:14:35Nico_PBagder: i find it rather strangely placed....
14:14:43Bagderwhere would you like it?
14:14:45Nico_Pbut it's a nice idea
14:14:50JdGordonabove the search
14:14:57Nico_Pi would have seen it near the search box
14:14:59Bagderin the menu?
14:15:05bluebrotherI'd rather say below the search ;-)
14:15:18JdGordonBagder: can you make the search edit box for both but with a FS# button?
14:15:35Nico_Pjust above the search box seems good
14:16:03amiconnjhMikeS: The tick timer itself doesn't disable interrupts, does it? Some tick tasks might do that....
14:16:33Bagderits just that our web pages are not made very nicely
14:16:39BagderI blame zagor for that ;-)
14:16:47bluebrotherBagder: could the cvs table also auto-link FS#<number> entries?
14:16:50Nico_Pso, no ideas on IDEs ?
14:16:50JdGordonwhat ever happened to the guy who was goin to redesign the whole thing? :D
14:17:03Bagderbluebrother: that's a good idea too
14:17:07BagderI'll see what I can do
14:17:24Bagderbut first I need coffee
14:17:27bluebrotherI'd volunteer for doing some work on the website if I had the time
14:17:34JdGordonBagder: and last thing for the cvs table... make it actually use the newlines from the commit message?
14:17:44bluebrotherbut unfortunately I'm lacking the time :(
14:18:06Genre9mp3JdGordon: he just made some logo proposals
14:18:17Genre9mp3and vanished
14:18:27*bluebrother likes the old logo best ;-)
14:18:33amiconnJdGordon: How do you add newlines in a commit msg?
14:18:34jhMikeSamiconn: not sure, will check each one
14:18:57Genre9mp3bluebrother: Old logo??? you mean the current one ;)
14:19:02JdGordonamiconn: in the text editor that opens when you commit?
14:19:06JdGordonor are they stripped?
14:19:08 Quit firenx (Remote closed the connection)
14:19:10bluebrotherexactly ;-)
14:19:18 Join Thundercloud [0] (n=thunderc@
14:19:41*JdGordon assumes \n's were not stripped
14:19:50jhMikeSamiconn: backlight.c on x5 does
14:20:08amiconnJdGordon: I always use -m
14:20:09jhMikeSreading buttons from the pcf50606 does
14:20:17bluebrotherIIRC correctly I looked in the code once and found \n not stripped (but HTML is different on that ...)
14:20:32JdGordonamiconn: I got out of that habbit after bad commits :p
14:20:44jhMikeSKeys would go dead with the remote on x5 if it didn't
14:21:38jhMikeSI thought IRQs were disabled at a level <= the irq level not <
14:22:01amiconnIt is <=, except on level 7
14:22:09LinusNjhMikeS: that is correct, but level 7 is special
14:22:48amiconnJdGordon: What does -m have to do with bad commits?
14:23:10JdGordoncommiting wrong files, when you just do cvs commit -m "blaa"
14:23:16JdGordonand missing files and such
14:23:35*JdGordon thinks ata_disk_is_active() in ata_mmc.c is wrong
14:23:42amiconnI always do cvs commit -m "blah" <file1> ... <fileN>
14:23:47 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
14:23:50jhMikeSok, I see on p 3-6...
14:23:51JdGordonreturning mmc_mutex.locked; is almost backwards to ata.c
14:24:29*petur always right-clicks on the file/tree and uses the tortoise cvs context menu :D
14:25:03 Part austriancoder ("Kopete 0.12.3 :")
14:25:06jhMikeSthe peakmeter code disables interrupts shortly but it seems for naught on level 7. but it's still running with that.
14:25:13amiconnpetur: It's weird... In general I prefer a GUI over the commandline, but for CVS it's different
14:25:27amiconnTortoise is too heavily integrated for my taste
14:25:37daurnimatori like torteise
14:26:02 Quit linuxstb (Read error: 110 (Connection timed out))
14:26:20peturI only do diff and patch on commandline, the rest is via tortoise
14:26:39daurnimatormost things i prefer cmd line
14:26:51daurnimatorbut.... version control needs an interface
14:27:22Genre9mp3Bagder: The Recent CVS activity table is a bit narrow now(?)
14:27:31jhMikeSI do need some sort of coherency there for the peakmeters...what to do? mask it temporarily?
14:28:19*linuxstb_ always previews commits with cvs diff
14:28:29 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
14:28:39*petur too
14:28:53amiconnlinuxstb: I do that as well, with cvs diff >diff.txt
14:29:31JdGordondoes anyone dissagree that ata_disk_is_active() in ata_mmc.c shhuold always return 1 ?
14:29:38linuxstbOr just "cvs diff | less"
14:30:16peturOr just right-click and select cvs diff
14:30:17jhMikeSif something should always return the same value why return anything?
14:30:27*petur hugs tortoise cvs
14:30:32pixelmaJdGordon: if I only knew what that means
14:30:38 Join perpleXa [0] (n=perpleXa@unaffiliated/perplexa)
14:30:38JdGordonbecause its mostly a dummy function in targets which use that file
14:30:40daurnimatorjhMikeS: to keep a standard for 1 on working
14:30:51amiconnJdGordon: Nope
14:30:53linuxstbpetur: At the command-line I do "cvs diff", and then edit that command to turn it into cvs commit - no chance of mistakes...
14:31:11amiconnata_disk_is_active in ata_mmc is correct the way it is written
14:31:39jhMikeSguess you need real returns in other impl. :\
14:31:39JdGordonamiconn: the problem is that unless the rest of the code is using that funtion wrongly, it means that the disk is spinnig, and not nessacerily being written to
14:31:55amiconnYes, and?
14:32:06amiconnThe MMC is active _only_ when it's accessefd
14:32:10JdGordon.. so settings will only ever be written on the ondio if something else is writing at the same time as we try to save them
14:32:14amiconnSo the implementation is correct
14:32:35amiconnWhy does settings_save check for ata_disk_is_active at all?
14:32:51JdGordonto not waste a spinup incase it was called too slowly
14:32:53amiconnThat doesn't make sense, as settings_save should only register an ata callback
14:33:00JdGordonall the ata callbacks should do that
14:33:45amiconnIn fact none of the callback registering should, that's the whole purpose of the callback handling
14:34:36amiconnThe only exception would be an asynchronous callback, but then it's not the registering which needs to do the check, but the function that's invoked by the asynchronous callback
14:34:55JdGordonsettign saving is async
14:35:17amiconnThat doesn't make any sense to me..
14:35:25JdGordonsetting saving has to be done in the main thread because it can stkov the ata thread
14:35:39JdGordonso a sys event is sent to all queues, handled in default handler
14:35:48JdGordonwhich then calls the actually dumping file
14:36:05amiconnThat's more than ugly imo
14:36:11JdGordonso is_active is called, which on the ondio gives the wrong return value
14:36:20JdGordonI know its ugly, but it works
14:36:26amiconnNo, the value isn't wrong
14:36:40JdGordonI can just skip the call on the ondio
14:36:55amiconnThe proper fix is to check why settings_save uses so much stack, and fix that
14:37:13amiconnEverything else is considered an ugly hack
14:37:56JdGordonapart from the "fluff" which gets saved (acceptable settin values" the saving file cant be optimized very much
14:38:26amiconnDid you check how much stack it consumes?
14:38:49amiconnThe 'fluff', as you call it, shouldn't have any significant effect on stack usage
14:39:23 Join funky_ [0] (
14:39:27 Quit funky_ (Client Quit)
14:39:29JdGordonI havnt checked, but it stkov'ed on bootup when I first had it calling properly
14:39:48JdGordon... unless we use this as an excuse to increase the ata stack size anyway?
14:40:48*jhMikeS notices a train of lengthy commit descriptions of late
14:41:15JdGordonmakes for goot morning readin :p
14:42:26Genre9mp3They seem lengthier now, too
14:42:38Genre9mp3the table wasn't that narrow before
14:43:47amiconnFor some reason the right table margin is now just left from the sponsors box...
14:44:40 Join n1s [0] (
14:45:08pixelmamissing some </div> or </p> somewhere?
14:46:56linuxstbSeems to be a </table> missing.
14:48:11pixelmathere are no </p>s anymore
14:48:42JdGordonpixelma: ok, new patch on fs... lemme know how it goes, im goin to bed
14:48:48 Quit JdGordon ("Konversation terminated!")
14:50:00 Part yewei
14:52:59 Join miepchen^schlaf [0] (
14:55:40Nico_Pso... what do you guys use to code rockbox ?
14:55:49Nico_Pi've been using kate up till now
14:56:25*petur mentions M$ DevStudio and ducks
14:57:05peturnotepad2 is nice too (on windows)
14:57:13Nico_Pdo these support things like jumping to a definition ?
14:57:22*n1s nags people to say what they think of
14:57:31Nico_Pon windows i like notepad++
14:57:40Nico_Pbut i'm looking for soething on linux
14:58:20Bagderhow's the flyspray box now then?
14:58:46Nico_PBagder: perfect :)
14:59:00LinusNn1s: looks ok. needs testing on all targets though
15:00:16n1sLinusN: ok, anybody up for testing?
15:00:29bluebrotherBagder: maybe add a "search" header like the other sections?
15:00:39n1shas been tested for several days on h300 at least
15:03:23***Saving seen data "./dancer.seen"
15:05:01 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
15:05:11 Quit lee-qid (Read error: 110 (Connection timed out))
15:07:31pixelmaBagder: the wiki page doesn't look right (the left side menu is too wide) propbably there are now some <p /> on it?
15:08:04pixelmawiki main page that is
15:08:39Bagderyeps, fixed now
15:10:37 Join Rondom [0] (
15:10:46Bagderhm, the wiki front page and the front page do show the menus very different
15:11:43Bagderoh well, I'll ignore that for now
15:13:14Genre9mp3hehe... evey left-side menu now seems it has it's own personality :)
15:13:28Bagderyeah but I ignore that too for now
15:13:55pixelmaBagder: on the front page there are -no- </p>s anymore and on the wiki main page I could still find <p /> - checking in a different browser now (maybe a caching problem)
15:13:58Bagderits mostly due to crappy building of the html files
15:14:18Bagderdue to Zagor's fondness of doing makefiles using wildcards
15:14:18jhMikeSok...having no tick tasks called and just burning CPU cycles in the tick interrupt makes the problem occur just the same ...
15:18:25preglowwhich indicates? :>
15:21:21*jhMikeS is indicated to be in the Twilight Zone
15:24:12oboNico_P: kscope?
15:24:29 Join MadDog011 [0] (
15:24:33 Part MadDog011
15:24:46Nico_Pobo: never heard of it... i'll look it up
15:26:02 Join _FireFly_ [0] (
15:33:09 Quit perpleXa (Read error: 110 (Connection timed out))
15:34:20amiconnjhMikeS: Hmm, so it seems we must not mask the audio dma interrupts, because latency is obvoiusly important
15:35:00amiconnI wonder whether using a source size of 'line' would allow for some more latency
15:35:20jhMikeSI suppose that is the case, yet they keep going with one another in contention (tested in a cheap reverb plugin)
15:35:37amiconnIt should, because the audio i/f also allows movem.l for reading
15:36:06jhMikeSamiconn: I tried it and it was paperclip time
15:36:30amiconnManual chapter 17.5 says that dma is enabled when PDIR is full, so there should be 6 samples available
15:38:08 Join neOliv [0] (i=52e3e414@gateway/web/cgi-irc/
15:38:25jhMikeS17.5 first paragraph?
15:39:37jhMikeSAnything but longword source froze it
15:39:55neOlivHi there.
15:40:09neOlivposted a theme to the forum
15:41:08neOlivand as bascule stated I may need more permissions to upload it the the wiki gallery
15:41:29neOlivis he right to think I may find someone able to help here?
15:42:07peturneOliv: wiki name?
15:43:33peturneOliv: done... happy wiki editing ;)
15:43:52 Join JoeBorn [0] (n=rootmeis@
15:44:08neOlivthx! THAT is fast service!
15:50:20 Join x1jmp [0] (
15:51:30Nico_Pobo: thanks for poiting me toward kscope, it's the best i've found so far
15:53:38x1jmpIs it possible to use some MiB of RAM in a plugin?
15:54:00Bagderx1jmp: if you "eat" the music buffer
15:54:31x1jmpBagder: that's just around 3 MiB on my iriver
15:54:46Bagderit's more like 30
15:54:52x1jmpRAM, but the h120 semms to have 32
15:56:08 Join My_Sic [0] (
15:58:06x1jmpah, i misread the number of the audio buffer ;)
15:58:43x1jmpcan it be used safely some way (when playing music)?
15:59:08n1sno playback has to be stopped
15:59:53Bagderwith music, you can only use the plugin buffer
16:00:02Bagderwhich is the rest of the memory where the plugin itself is loaded into
16:00:43 Join Moos [0] (i=d9805c3e@gateway/web/cgi-irc/
16:00:58 Quit JoeyBorn (Read error: 110 (Connection timed out))
16:02:39MoosGrr...! no JdGordon around :(
16:03:18x1jmpcould music still work, if i only use space at the end of the audio buffer?
16:03:57 Part LinusN
16:04:25Moosif you read the logs JdGordon: you seems to have broken the "save settings" before shutdwon since days and days now :( no possibility to use dir/tagcache anyway cause the settings doesn't save (your ata changes)
16:04:51x1jmpok, thank you
16:05:10peturMoos: don't they save when you rolo?
16:05:30Mooshi petur, didn't try
16:05:38Mooslet's see ...
16:07:45 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
16:08:44Kasperlethat's interesting. i've been using the same rockbox build for one to two weeks now, and yesterday it started losing its settings
16:09:00neOlivpetur: thx for the wiki edit permissions. first submission done.
16:09:13neOlivhope it will prove useful ;)
16:10:00 Quit Nico_P (Remote closed the connection)
16:10:07 Join My_Sic [0] (
16:17:32 Join freqmod [0] (
16:17:36jhMikeSamiconn: is movem atomic?
16:19:10 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
16:20:32 Join barrywardell [0] (
16:21:41 Join Criamos [0] (
16:25:03amiconnMoos: This should be fixed as of 06-Dec
16:25:47amiconnjhMikeS: Atomic in what way?
16:26:24 Join terriblyinept [0] (n=energy_O@
16:27:35jhMikeSamiconn: as in movem.l %d0-%d3,(%a0) will occur without interruption for example?
16:28:05 Quit neOliv ("CGI:IRC (EOF)")
16:28:09Moosamiconn: that what I thought too, but it seems that the bug is still here, I use one yesterday night build
16:28:59terriblyineptOdd question, and maybe I'm a bit too concerned with how things function, but is anyone aware of how a dualboot for the H10 would work?
16:29:06Moosamiconn: no possibility to use dircache or database correctly since few days now, frustration here :(
16:29:26amiconndatabase works fine here
16:30:50Mooshere, if I don't activate db, it show me "-1" in the db debug screen for the progress, then, it canot do commit steps
16:31:40MoosI discovered this annoying bug unfortunatly, in turning off the dircache option, and since canot use the chaches anyway :(
16:31:40x1jmpIs there any flash based device which runs rockbox?
16:31:41amiconnjhMikeS: I don't know
16:32:04Bagderx1jmp: archos ondio, ipod nano, sandisk sansa possibly in the future
16:32:16amiconn...and Iriver iFP 7xx
16:32:23amiconn(also not yet playing music)
16:32:46Moosamiconn: can you enable the dircache thing for see the bug?
16:33:01Mooslike I know you don't use it :)
16:33:10x1jmphow fast is the flash storage in comparison to ram?
16:33:20amiconnYou cannot compare them
16:33:41amiconnMass storage flash isn't directly addressable like ram or flash eprom
16:33:50peturand it has limited write cycles
16:34:16amiconnAccess speed depends on the interface, mostly
16:35:00x1jmpadn how long does it take until the first bytes of data arrives after a request (HDD players need some seconds to spin the disk up)?
16:35:47jhMikeSamiconn: can't find any info either but it seems like a burst couldn't be interrupted or it would screw up
16:36:46Bagderx1jmp: what request?
16:37:47 Join lee-qid [0] (
16:37:55 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
16:37:58x1jmpBagder: When I want to read a file on a HDD which is not spinning it needs a few seconds until I get the data, how fast are flash devices in that case?
16:38:28Bagderor possibly less
16:39:48 Quit x1jmp (Remote closed the connection)
16:41:27 Join My_Sic [0] (
16:47:03 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
16:53:18 Quit Moos ("CGI:IRC")
16:56:17barrywardellterriblyinept: dualboot on the H10 will give the option to load either Rockbox or the Original Firmware on startup by holding a button as it boots
16:57:54terriblyineptbarrywardell: *nods* I understand the concept. I understand you're the developer? I was actually wondering, if you had the time, about what the strategy for getting dualboot running is? A patched version of the H10 firmware with the Rockbox bootloader attached, or something else?
16:58:21barrywardellI have a working version running myself
16:58:52barrywardellat the moment it just boots a decrypted version of the H10 firmware that's in a file on disk
17:00:07barrywardellSo, the Rockbox bootloader replaces the H10 firmware (stored in a mi4 file) and then loads either rockbox.h10 or OF.bin from disk
17:01:37terriblyineptHow was the decryption managed? Did you do that on your own?
17:02:24 Join webguest09 [0] (i=c2cbc95c@gateway/web/cgi-irc/
17:03:03Bagderterriblyinept: H10 uses the mi4 file format which was cracked by our friend using the name MrH
17:03:25***Saving seen data "./dancer.seen"
17:03:50*Bagder runs off
17:04:39 Join miepchen^schlaf [0] (
17:05:09terriblyineptterriblyinept: Good, good. Do you foresee any minor problems, past the point of implementing the dualboot, with using a decrypted version of the H10 firmware?
17:05:28 Quit tvelocity ("Αποχώρησε")
17:05:40 Nick DreamThief|off is now known as DreamThief (
17:06:22terriblyineptbarrywardwell: Pardon me for all the questions.
17:12:22barrywardellnot really
17:12:29 Join My_Sic [0] (
17:12:57barrywardellit should be possible to even use a normal H10 firmware without having to decrypt it first
17:13:16barrywardelland just get the Rockbox bootloader to decrypt it
17:13:32amiconnArgh, LinusN left...
17:13:51barrywardellterriblyinept: sorry for the delayed response
17:14:28terriblyineptbarrywardell - I wanted to thank you for your work for porting Rockbox to the H10. If there's anything I can do on the user end, or learn about to contribute as a developer, leave me a message. I'm headed out for now. I'm getting an H10 today, I believe the 5G model, in the mail.
17:16:00terriblyineptbarrywardell - (I've worked with Rockbox on the H10 already; not yet on the 5G, though.)
17:16:26barrywardellthe best thing you can do is testing and reporting problems. there's loads of information about starting to contribute as a developer in the wiki
17:16:51terriblyineptbarrywardell - Thanks! Have a great day.
17:17:38*amiconn found something very suspicious regarding CACR
17:17:45 Part terriblyinept
17:24:42amiconnCheck firmware/target/coldfire/crt0.S line 174 and firmware/target/coldfire/system-target.h line 137, and the MCF5249UM.pdf page 84
17:24:54amiconnThe invalidate bit is bit 28, not bit 24
17:25:43amiconnBit 9 also doesn't match the comment
17:27:18preglowdo we have any trouble that might be caused by this?
17:29:25markunamiconn: does it make sense to have CPU_DEFAULT = 100MHz? I don't see an easy way to clock it any lower.
17:30:04preglowthe crystal freq is 100mhz???
17:31:44markunNo, 16.9MHz
17:32:08 Quit webguest09 ("CGI:IRC")
17:32:26markunbut the timing for the LCD depends on HCLK which depends on FCLK (CPU clock)
17:32:49markunI will read the datasheets a bit more, maybe I missed something
17:33:14 Quit miepchen^schlaf (Read error: 60 (Operation timed out))
17:37:33 Quit muesli__ (Read error: 110 (Connection timed out))
17:37:57Slasheriamiconn: i see there is CINV bit but it's not told what bit that is, weird
17:38:59Slasherithe bit 28 (CPDI) does not look the correct one either
17:39:18amiconnLooks like I confused the missing CINV and CPDI
17:41:12amiconnThe SCF5250 manual shows bit 24 is indeed CINV
17:42:23 Quit daurnimator (Read error: 54 (Connection reset by peer))
17:42:39 Join daurnimator [0] (i=daurn@unaffiliated/daurnimator)
17:50:58 Join miepchen^schlaf [0] (
17:53:07barrywardelllinuxstb: I think i have a universal binary ipod_fw for you. I can only test on my ppc mac though
17:53:22 Join daurn [0] (i=daurn@unaffiliated/daurnimator)
17:53:23 Quit daurnimator (Read error: 104 (Connection reset by peer))
17:55:14linuxstbbarrywardell: Nice. Give it to Bagder so he can put it on (I can't test on Intel either - we'll soon know if there are problems).">
17:57:21barrywardellI think Bagder is away at the moment, so I'll point him to that file next time I see him
17:57:34linuxstbWhat was involved in compiling it?
17:59:27 Quit petur ("*plop*")
17:59:34barrywardellmake an xcode standard command line project, in project settings make sure to use the universal target SDK (in general settings), and set architectures to "ppc i386" in build settings.
18:01:06 Quit lee-qid ("aufwiederbyebientotsayonara")
18:02:01 Quit Rondom (Remote closed the connection)
18:02:50 Join Rondom [0] (
18:03:13linuxstbOK. I should probably upgrade to a recent xcode and try it.
18:04:29barrywardellyes, you need a fairly recent version - i have 2.3
18:05:28 Quit Rondom (Client Quit)
18:06:07 Quit _FireFly_ ("Leaving")
18:09:04 Join Rondom [0] (
18:09:52 Join Ekron [0] (
18:10:47 Join Gnelik [0] (
18:15:49 Quit miepchen^schlaf (Read error: 60 (Operation timed out))
18:15:56 Quit barrywardell ()
18:16:07 Join miepchen^schlaf [0] (
18:16:13 Quit daurn (Read error: 104 (Connection reset by peer))
18:16:15 Join barrywardell [0] (
18:18:39 Join terriblyinept [0] (n=energy_O@
18:19:28dakydamn my mp3player ( Maxfield - Ivy 1 gb ) sucks so the firmware are shit and and
18:19:47dakyi think i buy the SanDisk Sansa e 250 DAP :D
18:20:35 Join Ribs [0] (n=ribs@
18:30:14 Join Arathis [0] (
18:30:19 Join Nico_P [0] (
18:36:49 Join muesli__ [0] (n=muesli_t@
18:37:21 Part terriblyinept
18:38:05 Join PaulPosition [0] (
18:38:12 Quit My_Sic (Read error: 131 (Connection reset by peer))
18:39:21 Join homecable [0] (
18:39:58 Quit blue_lizard (Read error: 110 (Connection timed out))
18:40:02 Join blue_lizard [0] (
18:41:39 Join petur [0] (n=petur@rockbox/developer/petur)
18:44:06 Quit barrywardell (Read error: 145 (Connection timed out))
18:45:44 Join webguest23 [0] (i=d92caed0@gateway/web/cgi-irc/
18:49:09 Quit webguest23 (Client Quit)
18:55:57 Part Gnelik
18:56:05 Join Gnelik [0] (
18:56:22 Join piquadrat [0] (
18:56:33 Join My_Sic [0] (
18:58:18 Quit Arathis ("[rl_bot quit]")
19:03:26***Saving seen data "./dancer.seen"
19:04:36 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
19:05:00 Join miepchen^schlaf [0] (
19:07:01 Quit idnar (Nick collision from services.)
19:07:04 Join idnar_ [0] (i=mithrand@unaffiliated/idnar)
19:15:07 Join SoulDeaD [0] (i=logrus@
19:15:22SoulDeaDhi folks
19:26:49PaulPositionhow come I can't check-out cvs anonymously anymore?
19:27:21PaulPositionTells me empty password ain't good enough. Didn't have that problem last time I tried, but then again I just re-installed vmware...
19:28:01PaulPositionduh.. Nevermind, a typo on my part :blush:
19:28:04 Nick Everybody|away is now known as Everybody (
19:29:54 Join terriblyinept [0] (n=energy_O@
19:31:21 Quit Rudy4Pez (Read error: 60 (Operation timed out))
19:33:21 Quit muesli__ (Read error: 145 (Connection timed out))
19:33:45 Join muesli__ [0] (n=muesli_t@
19:34:13 Join Thundercloud_ [0] (n=thunderc@
19:39:02 Quit PaulPosition (Read error: 131 (Connection reset by peer))
19:39:53 Quit Rondom (Read error: 60 (Operation timed out))
19:40:17 Quit Gnelik (Read error: 110 (Connection timed out))
19:41:07 Part terriblyinept
19:42:57 Join Lear [0] (i=chatzill@rockbox/developer/lear)
19:45:26 Nick idnar_ is now known as idnar (i=mithrand@unaffiliated/idnar)
19:45:47 Join jborn__ [0] (
19:46:32 Join matsl [0] (
19:48:23 Join ender1 [0] (n=ender@
19:49:01 Quit hcs (
19:49:02 Quit daky (
19:49:02 Quit corecode (
19:49:02 Quit crwl (
19:49:02 Quit jrockway (
19:49:02 Quit jborn_ (
19:49:02 Quit Bg3r (
19:49:02 Quit FOAD (
19:49:02 Quit crashd (
19:49:02 Quit spug (
19:49:02 Quit luckz (
19:49:02 Quit netmasta10bt (
19:49:02 Quit Kohlrabi (
19:49:02 Quit BHSPitMonkey (
19:49:02 Quit SUSaiyan (
19:49:02 Quit vr (
19:49:02 Quit ScoTTie (
19:49:02 Quit Bagder (
19:49:41NJoinluckz [0] (
19:49:45NJoinvr [0] (
19:49:50 Quit MarcoPolo ("Bye !")
19:49:55NJoinjrockway [0] (
19:50:04NJoinspug [0] (
19:50:04NJoincrashd [0] (
19:50:07NJoinKohlrabi [0] (
19:50:30NJoinFOAD [0] (
19:50:39NJoincorecode [0] (
19:51:59NJoinhcs [0] (n=agashlin@rockbox/contributor/hcs)
19:52:25NJoincrwl [0] (
19:52:53 Join ScoTTie [0] (
19:52:54***Server message 505: 'logbot :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
19:52:59NJoindaky [0] (
19:53:17NJoinSUSaiyan [0] (
19:53:26NJoinBagder [0] (
19:53:37 Quit ender` (Read error: 145 (Connection timed out))
19:53:50NJoinBg3r [0] (n=bager@
19:54:20 Quit SoulDeaD ()
19:58:32 Quit Thundercloud (Connection timed out)
19:59:17piquadrathi everybody. I'm trying to apply the latest scrobbler.patch, but there are several rejects. Someone got a build for iAudio X5 with scrobbler?
19:59:38peturscrobbler is in the cvs version
19:59:50peturjust enable it in the settings
20:00:16peturGeneral -> playback ->
20:00:20 Nick Everybody is now known as Everybody|away (
20:00:24piquadratah, cool, didn't know that
20:00:46*pixelma points towards
20:01:43piquadratawesome! Thank you!
20:02:13NJoinBHSPitMonkey [0] (
20:02:40 Join MarcoPolo [0] (
20:04:11 Join vertic23 [0] (
20:04:35vertic23hey, I got some problems with my player (iAUDiO x5v)
20:04:48vertic23it stucks when I try to play mp3 ...nothing happens - frozen
20:04:54vertic23yesterday everything was fine
20:06:10peturdid you update rockbox inbetween?
20:06:28vertic23I still got a CVS daily build from 9-22-2006
20:06:44vertic23should I try doing a new one?
20:06:58peturor reset your settings do I do that
20:07:39vertic23doing the reset button on the player?
20:08:11peturhold rec while rockbox starts I think
20:08:39vertic23ah okay
20:08:55peturor menu -> manage settings
20:08:58vertic23I'm doing scandisk atm ...after that I'll try
20:08:59 Nick ender1 is now known as ender` (n=ender@
20:19:28 Join Paul_the_Nerd [0] (
20:21:43vertic23won't work... at all
20:21:56vertic23how to proceed
20:22:05vertic23maybe I'll try a new daily build?
20:22:37vertic23fuck ...hopefully my player isn't fucked up
20:25:14 Join Mongey [0] (i=Mongeyc@
20:25:22MongeyAny french people here?
20:26:28Nico_PMongey: yes
20:26:28*Mongey waits
20:26:37Mongeywanna help me out?
20:26:50Mongeyim doing a project on french Tv stations
20:27:18Mongeyso i was wondering what the main tv station is
20:27:32Mongeyand what type of shows are on it
20:27:58Nico_PI'd say the main station is TF1
20:28:12Nico_Peven though it's not the best ^^
20:28:17vertic23how do you update the daily build?
20:28:26vertic23do I delete .rockbox?
20:28:38vertic23and then put the latest on it again?
20:28:45vertic23or how do I update rockbox
20:29:09n1svertic23: copy a new build on top of your old
20:29:16Nico_Pvertic23: you just unzip your new version in the root of your DAP, overwriting everything
20:29:28MongeyNico_P: thanks
20:30:21vertic23and now I am fixed hopefully :P
20:31:16Nico_PMongey: you should also have a look at "France Télévisions"... it's the public stations
20:31:22Nico_Pvery important too
20:31:35Mongeywhat about TV5 ?
20:31:37*petur returns
20:31:43vertic23wb petur
20:32:29Nico_PTV5 is only available on digital TV, so it's not very popular
20:32:32vertic23aight... new daily build fixed it!
20:32:38vertic23I am so happy
20:32:44vertic23which theme are you using
20:32:56Mongeyty so much
20:33:02 Quit piquadrat (Remote closed the connection)
20:33:02 Part Mongey
20:36:21vertic23petur is this one available for my X5V?
20:36:43peturit should be, it comes with the default install
20:37:12 Nick myzar is now known as myzar|away (n=myzar@
20:37:14Genre9mp3s/should be/is
20:37:35vertic23yeah got it..
20:37:40*petur points to Genre9mp3 and whispers 'creator'
20:37:43vertic23it's Browsing .wps files, right?
20:37:54*Genre9mp3 hides
20:38:59vertic23 <−−-
20:39:06vertic23I wish it would look like this here
20:39:12vertic23but the text is all over the right side
20:39:24peturvertic23: browse themes
20:39:39vertic23well... I just "activated" this theme
20:39:50vertic23and.. the "text"-part is way too much right
20:39:58vertic23so I can't read what I am listening etc.
20:40:22peturhow long before you'll make your own custom wps? :)
20:41:08vertic23hehe, well.. it should work, right?
20:41:16vertic23because it's for my player...
20:41:26vertic23hmm, maybe it has been updated.. checking...
20:41:33petursome themes require unofficial builds or patched builds
20:41:46vertic23ah.. wait.. it needs another font
20:42:05vertic23is it on it already (in general)
20:42:16peturdid you install the fonts
20:42:37vertic23got it ...should work ...
20:42:42vertic23just checked that
20:42:51vertic23but still the "aspect-ratio" is kinda wrong
20:42:56peturthen make sure you use the theme to install, not the wps
20:43:13peturtheme = wps + other settings like font
20:46:16vertic23hmm, and where can I do it... or _how_?
20:46:25 Quit Paul_the_Nerd (Read error: 110 (Connection timed out))
20:46:25vertic23I'm not sure if I know what you are talking about
20:46:46peturjust a sec
20:47:53 Quit freqmod (Read error: 145 (Connection timed out))
20:48:05peturstrange, font is included and the theme has theme cfg file and wps file
20:48:51peturwhat's the problem really?
20:48:59vertic23do you want me to make a photo -g-
20:49:17vertic23all the id3 stuff - and battery display is way too much right
20:49:20vertic23so I can't read it
20:49:34vertic23it's like this "Battery: 7"
20:49:44vertic23Battery: 70% would be the whole info
20:50:05 Join _FireFly_ [0] (
20:50:08vertic23..but after the 7 the LCD display is ending -g-
20:50:31peturcan you manually select the font he mentions?
20:50:39vertic23how do you mean?
20:50:46vertic23I know I got this font
20:51:34peturand did you unzip the whole theme onto your player (theme and wps dir)?
20:51:54peturand then selected it by using 'browse themes'?
20:52:25vertic23actually not.. :P
20:52:29vertic23I did browse .wps :P
20:52:35vertic23where is browse themes
20:52:52peturgo to menu
20:52:55peturthere it is
20:53:27vertic23hmm but still this theme sucks -g-
20:53:36vertic23because he doesn't scroll the titles..
20:53:50vertic23so I only can read the first x letters of a title.. bah
20:54:09peturah, that's because standard rockbox doesn't do scrolling and margins together
20:54:29peturcome back next year ;)
20:54:34vertic23I see :P
20:54:43vertic23with my old silk theme it worked anyway
20:55:42vertic23DancePuffDuo rocks =)
20:55:45 Join MikeW [0] (
20:55:49vertic23but now he took the gay orange...
20:56:25petursome themes don't reset everything
20:56:48vertic23sucks :P
20:58:19 Join PaulPosition [0] (
20:58:25 Nick Everybody|away is now known as Everybody|ingame (
20:58:26vertic23I love my mp3 player again =)
20:58:41vertic23today was the first day without it
20:58:48vertic23it was horrible
20:59:28*petur wonders where the cfg files are in cvs
21:00:57 Quit spiorf (Remote closed the connection)
21:01:51linuxstbpetur: I think they're generated from wps/WPSLIST
21:01:51pixelmaIIUC they are generated from the wps-files and WPSLIST
21:02:02pixelmahaha :)
21:02:14peturlinuxstb wins
21:02:57n1sdoes databox, work with the current database?
21:03:25n1swe build it but it includes database.h from the old database...
21:03:30***Saving seen data "./dancer.seen"
21:04:40linuxstbNo, I don't think it works. I think it's left in case someone wants to adapt it for the database formally known as tagcache.
21:04:53 Join tucoz [0] (n=martin@rockbox/staff/tucoz)
21:05:10n1swill anybody cry if I disable building it (not deleteing)
21:05:20tucozI talked with HCl about it some time ago
21:05:32linuxstbI won't. I wouldn't even cry if you deleted it (CVS remembers...)
21:05:52preglowlinuxstb: time to delete libFLAC? :-)
21:05:53tucozI think we agreed (sort of) that it is safe to delete it
21:06:22tucozI can not remember when this discussion was though.
21:06:35n1sok, I'll do it
21:06:42tucozIs dumb still built?
21:06:44 Quit My_Sic (Read error: 145 (Connection timed out))
21:07:09 Join barrywardell [0] (
21:07:55linuxstbpreglow: I don't think we'll ever want the decoder, but what about encoding?
21:09:32preglowlinuxstb: been a while since i encoded a flac, but i think i remember concluding it's not entirely realistic
21:09:50preglowbut if we ever do do encoding, let's please use flake or something instead of libflac, i hate libflac :>
21:10:39linuxstbflake is floating point...
21:10:51linuxstbOr at least, uses some floating point.
21:11:04 Join My_Sic [0] (
21:11:04preglowso does libflac
21:11:11preglowthe entire lpc deal is floating point based, afaik
21:11:26preglowhmm, or no, i might be thinking about something else
21:11:41linuxstbI thought the libFLAC issue with encoding were the wide use of 64-bit integers.
21:11:49preglowyeah, that's what i remembered too
21:12:36preglowbut can't just whoever does encoding recommit it? i don't know what i think about keeping code around in cvs "in case it's needed"
21:12:47preglowi'd also like to remove dumb for the same reason
21:12:54amiconnHehe, I had to use some trickery in bounce.c to avoid using 64 bit ints
21:13:13linuxstbpreglow: Well, I can't bring myself to do it, but wouldn't shout if you did.
21:13:25*tucoz notices a new search box on Nice :)
21:13:44amiconnThe cordic fsincos has nice precision, but the angle argument spanning the full 32 bit range isn't very handy
21:13:54preglowamiconn: why not? i find that extremely handy
21:14:20 Join t0mas [0] (n=Tomas@rockbox/developer/t0mas)
21:14:24preglowit's very precise, and you get free phase wrapping if you do calculations on the argument
21:14:28 Join lowlight [0] (i=c730190b@gateway/web/cgi-irc/
21:15:15barrywardellPaulPosition: i've got an updated bootloader patch. crossfeed and pitch adjustments now work. and it works on ipod color. were there any other problems that you noticed?
21:15:21amiconnYou cannot do a full circle in N steps without roundoff problems if you don't either go 64 bit for intermediates or take special care and process the remainder (as I did)
21:16:02preglowamiconn: what format would you prefer it to be in?
21:17:28amiconnHmm, and the free phase wrapping isn't exact iiuc
21:17:47amiconnThe comment says 0..0xffffffff equals 0..2*pi
21:17:56preglownot exactly true 0x100000000 is slightly more than 2*pi
21:18:07preglow0xffffffff is _almost_ 2pi, if i remember
21:18:16amiconnIt says so in the comments...
21:18:26preglowthe comments weren't changed for the new implementation, obviously
21:18:47preglowbut anyway, i don't know, the new sincos is done by safetydan, and i don't know how it works
21:19:10 Quit My_Sic (Read error: 104 (Connection reset by peer))
21:19:17amiconn"phase has range from 0 to 0xffffffff, representing 0 and 2*pi respectively."
21:19:34preglownot even true for the old implementation
21:19:42preglowso that's not completely accurate, no
21:20:03preglowunless my recollection is far off, 0xffffffff will be 2*pi - 1/0xffffffff
21:20:36preglowbut anyway, that sincos is intended more for dsp than anything else
21:20:51preglowin what form would you prefer the argument to be?
21:21:29amiconnNot 100% sure, but certainly something that leaves room for multiplication without overflowing 32 bits
21:21:39preglowthat would be way too imprecise for dsp
21:21:47preglowtwo different implementations would probably be a good idea
21:22:00preglowthe other sin/cos implementation i am thinking of putting in also uses the very same argument format
21:22:21preglow <- this one
21:22:35preglowit's more precise, and might even be faster than the current one
21:24:04amiconnForget the m68k asm in there
21:24:09peturregarding : would be a solution that is too simple? Hard menu recursion protection
21:24:15amiconnIt's not coldfire compatible
21:25:56 Join Rob222241 [0] (
21:25:57preglowamiconn: i'm not talking about that
21:26:02preglowamiconn: i'm talking about the generic c
21:26:22preglowamiconn: the m68k asm is worthless for us anyway, it assumes a 16 bit multiplier, and we've got a better one
21:26:47tucozwould it be ok if we stopped building dumb?
21:26:51preglowtucoz: hell yes
21:26:52amiconnWhy would we need 2 sin/cos implementations?
21:26:55preglowtucoz: i'd even like to delete it
21:27:02preglowamiconn: this one would replace the one we've got
21:27:04tucozhmm. that could be arranged
21:27:16lowlighttucoz: here...
21:27:17preglowamiconn: like i said, it's more accurate, perhaps faster, and probable smaller in code size
21:27:25tucozlowlight, hehe
21:27:26amiconnIt's more accurate?
21:27:29preglowamiconn: yup
21:27:31lowlightbeen sitting on that
21:27:43preglowamiconn: i'm almost 100% sure, i wrote a small test program
21:28:18tucozlowlight, i think my local copy is equal as your patch :)
21:28:33LearAnyone ever heard of Cygwin being much slower on some computers? And if so, what it might be caused by?
21:28:38lowlightpreglow: you might as well remove libflac. if someone wants to make an encoder then then they should start with the newer 1.1.3
21:28:42amiconnpreglow: Smaller and faster is always good, and if it's more precise as well...
21:28:51preglowamiconn: btw, i tested the sin/cos routines separately, not the sincos combined one
21:28:59amiconnBut somehow I doubt it's smalkler, the current one is very compact already
21:29:13preglowamiconn: yeah, that's just a guess i have, i've yet to see if it's true
21:29:29amiconnEven on SH it's just ~160 bytes code and 128 bytes table data
21:29:37amiconnArm and coldfire are smaller
21:30:31*linuxstb remembers you can't delete directories in CVS...
21:30:32amiconnSIze doesn't matter that much here
21:31:09amiconnlinuxstb: But they should be pruned if they're empty and you cvs up -dP
21:31:22*preglow mutters "svn"
21:33:12tucozamiconn, i think that is correct
21:33:41 Quit Rob2222 (Read error: 60 (Operation timed out))
21:34:06tucozat least the empty manual directories gets pruned by -dP
21:34:47 Join donutman [0] (i=502968b5@gateway/web/cgi-irc/
21:35:18amiconnIt's the -P which does this
21:35:36*linuxstb commits the death of libFLAC
21:35:39donutmanhello, sorry to be a pain in the a** for asking, but ive not kept up with rockbox progress for a while and want to find something out quick- do you have video with sound and/or radio on the h10 yet?
21:36:44*preglow spits on libFLAC's remains
21:37:06linuxstbI feel so much lighter now...
21:37:27preglowlinuxstb: do dumb in the same go?
21:37:44linuxstbNo, just FLAC. I know nothing about dumb.
21:38:07lowlightdumbs not used
21:38:07 Quit donutman (Client Quit)
21:38:13preglowstill compiled, isn't it?
21:38:16tucozonly built
21:38:17tucozyes, it is
21:38:24linuxstbpreglow: Good point...
21:38:46preglowso removing that will actually have some practical benefit as well, heh
21:38:50*linuxstb leaves the pleasure of dumb to someone else
21:39:14preglowah, that was the nicest cvs update i've seen in a while
21:39:33lowlightand then remove the directory
21:39:48*tucoz can do that, in case no one else wants to
21:39:58lowlightgo ahead
21:40:04preglowgo ahead
21:40:06*preglow goes brb
21:40:37 Quit Rob222241 (Read error: 104 (Connection reset by peer))
21:40:54 Join Rob2222 [0] (
21:40:57amiconn <== huh?
21:41:15linuxstbamiconn: I was just looking at that...
21:42:59 Join netmasta10bt [0] (
21:43:45 Join barrywardell_ [0] (
21:43:45 Quit barrywardell (Read error: 104 (Connection reset by peer))
21:44:05 Join lee-qid [0] (
21:44:30PaulPositionBarrywardell - Hi... Say, is that last e200-dualboot-v2 patch specific to e200? (ie, did you erase the dualboot code for H10s?)
21:44:40*Genre9mp3 wonders why someone would leave his telephone number in a mailing list...
21:44:56barrywardell_no, it should apply to the h10 too
21:45:08 Quit lowlight ("CGI:IRC (Ping timeout)")
21:45:24 Nick barrywardell_ is now known as barrywardell (
21:45:32 Join perpleXa [0] (n=perpleXa@unaffiliated/perplexa)
21:45:43tucozhmm. to remove dumb, i'll have to first delete all the files locally, and then cvs remove them right?
21:45:49PaulPositionbarrywardell - Gives me a rockbox error -1, but I might have f*** something.. Rockbox loads fine though.
21:46:12barrywardellit expects the file to be called OF.bin now, rather than Original.mi4
21:46:15*n1s likes when everybody rates the severity of 'their' bugs as 'high' or 'critical' :-/
21:46:28barrywardelltucoz: yes
21:46:38PaulPositionbarrywardell - Ah, should have looked at the code I guess. Thanks, I'll try that. :)
21:46:49 Join Thundercloud__ [0] (n=thunderc@
21:46:50 Join mathgl [0] (
21:47:19PaulPositionbarry - Still decrypted and beheaded?
21:47:36tucozbarrywardell, do you know if it is possible to do this recursively and with wild cards?
21:47:53 Join midgey [0] (
21:47:58barrywardellPaulPosition: yup. thee only change is renaming the file
21:48:28amiconnbarrywardell: Why not use <insertfavouritename>.sansa the same way as the rockbox binaries?
21:48:28PaulPositionThanks. :) And how is that Sansa port going? A friend of mine just ordered three of 'em for the whole family.. :p
21:48:36barrywardelltucoz: not sure. i think cvs does accept wildcards
21:48:46amiconnThe header could indicate that this is the original firmware and needs special treatment
21:48:57amiconnThe same file could then be used to rolo the of
21:49:03tucozbarrywardell, thanks. i'll see what google says.
21:49:50barrywardellgood point amiconn. i'd left the header out for simplicity up to now
21:50:06barrywardellbut it would definitely be good to keep it eventually
21:51:01PaulPositionbarrywardell - Works great, albeit frighteningly fast to load the OF, much faster than, well, the OF on itself. :o
21:51:28n1sheh :-)
21:51:40PaulPositionthanks :D The crossfeed and pitch don't freeze anymore either, very well done.
21:52:15barrywardellwow, great! possibly faster because it doesn't need to be decrypted by the player
21:52:35linuxstbAnyone here with win2k and an ipod?
21:53:31amiconnHmm, .mi4 files are loaded from disk and decrypted every time? Or are they loaded once and then flashed?
21:53:36PaulPositionbarrywardell - Makes a lot of sense yes. :thumbsup:
21:54:52barrywardellamiconn: decrypted every time on the H10. they're flashed on the Sansa, but I'm not sure if they're decrypted first.
21:57:21peturamiconn: any problem with preventing main menu recursion the hard way? patch at the bottom. Or isn't life this simple?
21:59:33amiconnThis is just the main menu itself, right?
21:59:57amiconnI can't imagine we want the same menu active 2 times
22:00:05amiconnI think that should work
22:01:44*amiconn hopes the main menu doesn't have more than one return path
22:02:11peturno it doesn't
22:05:39 Quit MikeW ()
22:07:32amiconnThe menu recursion hard-crashes on recorder
22:07:41amiconnI04: IllInstr
22:07:49 Quit homecable (Read error: 131 (Connection reset by peer))
22:08:56 Join joey- [0] (
22:09:47amiconnIf I go back after 4 recursions, it does a different thing. First it displays all files in plugins (as if it would be 'show all'), and after leaving everything, the red led starts flashing quickly, with the buttons frozen
22:10:22peturfor me it also showed all plugins, but with the extensions flashing
22:10:42amiconnYes, because the H300 doesn't have a LED it could flash ;)
22:11:10 Quit Thundercloud_ (Connection timed out)
22:12:15 Quit bluebrother ("leaving")
22:12:39 Join Cassandra [0] (
22:12:54peturhmmm... the buildserver ate my commit
22:13:01CassandraGods - CVS, how I hate thee. Let me count the ways ..
22:13:19CassandraNope, no time. Someone please help me create a new directory under the root.
22:13:25petursmall Gods?
22:13:33 Join bluebrother [0] (i=ujpHxcxr@rockbox/staff/bluebrother)
22:13:41*petur thinks tortoise again
22:13:52CassandraMore momegs, really.
22:13:52amiconnpetur: Ipod is less spectacular. Most of the time simply freezes (but also does the show all thing)
22:13:59CassandraAnd Eris. It's bound to be her fault.
22:14:32CassandraAnd even if it isn't, she'd be upset if she didn't get blamed.
22:15:04*bluebrother notices dumb got removed
22:15:18Nico_Pwhat is dumb ?
22:15:18*tucoz wonders where linuxstbs commit did go
22:15:27tucoza .mod .s3m etc library
22:15:41bluebrothertucoz, you're too fast for me ;-)
22:15:41peturah.. there it is... seems to be busy evening for the server ;)
22:16:01tucozbluebrother, did you want to do that?
22:16:02PaulPositionwtf? Really? I could listen to, like, demo-compo music from 15 years ago with that?? :o
22:16:04CassandraWhat do you mena "no such file or directory, you pissant piece of software.
22:16:20tucozPaulPosition, but dumb was not working
22:16:28PaulPositionAh. lol, okay.
22:16:35bluebrothertucoz, no, I was referring to the answer
22:16:35tucozPaulPosition, only the lib was committed. but no codec was written
22:16:39 Join _DreamThief [0] (
22:17:04*bluebrother is a bit tired.
22:17:15 Nick _DreamThief is now known as DreamThief|off (
22:17:40CassandraWhy is it complaining that I can't add to CVSROOT/Emptydir? I don't want to do that.
22:18:08peturcomputers tend to do what you ask them, not what you want them to do :p
22:19:23linuxstbCassandra: Are you trying to create a new top-level module in CVS?
22:20:53godzirraSo whats the status on the gigabeat F series?
22:21:09Cassandralinuxstb, yes.
22:21:12bluebrotherbtw, what's the state of that mod codec in the tracker? Mod support would be great
22:21:13markungodzirra: we're close to having audio playback
22:21:20CassandraPlease, tell me how to make it happen.
22:21:21 Join lowlight [0] (i=c730180b@gateway/web/cgi-irc/
22:21:25markungodzirra: better come to #gigabeat for more info
22:21:57tucozbluebrother, i have listened to mods on my h120. but that was before it was committed to cvs.
22:22:08tucozi would assume that it works
22:22:10CassandraBefore I gnaw off my own (or worse) someone else's arm.
22:22:24bluebrotherwas committed? dumb?
22:22:38tucozoh. maybe the mod patch was never commited
22:22:44godzirramarkun: awesome.
22:22:45tucozi thought it was
22:22:48bluebrotheror do we have a mod codec in cvs?
22:23:09bluebrotherI believe we don't, but I'm a bit out of sync currently
22:23:19Genre9mp3isn't there a mod patch somewhere?
22:23:25 Quit lowlight (Client Quit)
22:23:28bluebrotheryes, in the tracker
22:23:33linuxstbCassandra: I normally create an empty directory somewhere on my hard disk, "cd" into it, then use the "cvs import newdir vendor version" command. Where newdir is the name of the directory that will be created in CVS, and I've never understood what vendor and version should be - I normally use something like "cvs import newdir dave start"
22:23:34Cassandrabluebrother: maybe you should check yourself into CVS.
22:23:38tucozGenre9mp3, yes. But i thought that was committed. But I was probably wrong
22:23:58Cassandralinuxstb, I'll try that. ta.
22:23:58bluebrotherhehe ... but I fear someone else is doing changes to me
22:24:04Genre9mp3no.. only sid was commited a while ago
22:24:21linuxstbCassandra: You then do a "cvs co" for your newly created directory, and then add files in there as normal.
22:24:24 Join lowlight [0] (i=c730190b@gateway/web/cgi-irc/
22:24:24tucozright. i mixed those up.
22:24:32 Quit DreamThief (Read error: 60 (Operation timed out))
22:25:10Genre9mp3this mod patch is not based on dumb btw
22:25:40 Quit MarcoPolo ("Bye !")
22:26:14linuxstbCassandra: BTW, I'm working on a rewrite of ipodpatcher which should make its incorporation into your installer a breeze. The aim is that it will do what ipod_fw does as well, so a single command like "ipodpatcher 1 bootloader-nano.bin" will do the install...
22:28:37Cassandralinuxstb, cool. I've already designed how to fit it all into the installer in my head.
22:29:20CassandraI was holding off on that anyway while waitng for the toolchain to settle down again.
22:29:42CassandraErm, the command claimed it worked. Don't know where it thinks it imported it to though.
22:30:16CassandraOh no, it's just a new module. Wonder why there were no logs.
22:30:57linuxstbMaybe the emails only fire on certain modules - Bagder will need to add rbutil.
22:31:08linuxstbBut it's there:
22:32:17*linuxstb spots CRLF line terminators.....
22:33:17amiconnlinuxstb: Shouldn't it be possible to find ipods automatically instead of probing those numbers?
22:33:46linuxstbamiconn: I'm sure it is, but I don't trust myself...
22:34:01Cassandralinuxstb: bugger.
22:34:12linuxstbBut I agree that should definitely be added at some point.
22:35:03amiconnOnly problem would be if more than one ipod is connected
22:35:04CassandraI'll have to figure out a way of doing that for the installer anyway.
22:35:53bluebrotherthe ipl installer2 has some ipod autodetection
22:36:15linuxstbThis page seems to be a good approach, but I don't think anything implements it yet:
22:36:39linuxstbbluebrother: I thought that was broken now that Apple have removed the SysInfo file from the ipod_control folder...
22:37:12bluebrotherI had a short look into it, and from what I understood they parsed some information from the partition table
22:37:49bluebrotherbut as I don't have an ipod I couldn't test it
22:38:11linuxstbBut that's for later, first priority is to get it working with manual detection.
22:38:30 Quit _FireFly_ ("Leaving")
22:38:52CassandraStep back a moment. How does a human work out which PhysicalDevice is the Ipod?
22:38:54linuxstbI also don't develop on windows, so my first job is porting ipodpatcher to Unix...
22:39:09 Quit joey- (Read error: 104 (Connection reset by peer))
22:39:11CassandraIt's the removeable storage with a small boot partition and a data partition.
22:39:21CassandraWe ought to be able to codify that somehow.
22:39:47oboUSB product and vendor IDs may work...
22:40:04obofor make/model detection
22:40:24Cassandraobo: That might be as
22:40:32Cassandrasensible approach too, yes.
22:40:37linuxstbI just don't want to get it wrong and be responsible for trashing someone's hard drive... So I've just made it the user's final responsibility so far.
22:41:00linuxstbBut there are already a lot of checks in ipodpatcher to make sure it's writing to an ipod.
22:41:13linuxstb(and the new version will have even more)
22:41:37linuxstbi.e. it will modify the firmware partition in-place, so if there isn't one there, it won't do anything.
22:42:19*linuxstb waves goodbye to dumb
22:43:33 Quit Ribs (Remote closed the connection)
22:43:49linuxstbIs anyone else just receiving the dumb CVS emails, or is it just my email being slow?
22:44:12CassandraI had them not too long ago.
22:44:51 Join joey- [0] (
22:45:11lowlighttucoz: you forgot FILES too
22:45:58 Nick joey- is now known as homecable (
22:46:50 Join amiconn_ [0] (n=jens@rockbox/developer/amiconn)
22:47:15tucozand thanks for telling me
22:51:03peturamiconn, Slasheri: dircache setting saving is still not ok (as Moos said). After scanning is done the first time, the subsequent shutdown doesn't spin up the disk and as such doesn't save the dircache size.
22:52:31CassandraUnix line delimiter is lf, right?
22:53:37 Quit amiconn (Read error: 145 (Connection timed out))
22:53:37 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
22:55:00 Quit Lear ("Chatzilla 0.9.75 [Firefox 2.0/2006101023]")
22:55:10*n1s wonders why rbutil files are versioned
22:58:29Cassandran1s - me too. Strange are the ways of CVS. Best not to ask.
23:03:34***Saving seen data "./dancer.seen"
23:03:55 Quit PaulPosition ()
23:03:57CassandraArrr! Now why can't I delete a file?
23:04:10Cassandra"remove requires write access to the repository"
23:04:34CassandraErm, but I put the code in there in the first place. Doesn't that rather imply that I may have write access?
23:05:52n1s"Strange are the ways of CVS" :-P
23:07:23*petur gives format_display() in gwps-common.c a puzzled look
23:10:49 Quit barrywardell ()
23:11:17 Quit t0mas ("Doei")
23:11:39 Join Mmmm [0] (
23:13:12 Quit Everybody|ingame ("( :: NoNameScript 4.02 :: )")
23:15:10 Join Everybody [0] (
23:15:20 Join PaulPosition [0] (
23:15:38 Quit PaulPosition (Client Quit)
23:16:37 Quit Mmmm (Remote closed the connection)
23:19:36 Part lowlight
23:25:15 Join Esoterik [0] (
23:25:29CassandraIt's going to be one of those nights, isn't it?
23:26:23peturbetter quit early then
23:27:11EsoterikHey all , was just wondering if anyone could help for a sec? :P
23:28:38EsoterikIm trying to get my ipod 5.5 setup with a rockbox build but so far in trying to find the number assigned to my ipod i have no luck... can anyone tell me what im doing wrong if i give a run through of what ive done or havnt done so far?
23:28:38peturwhy not just ask the question
23:29:22peturbetter find an ipod expert in here :)
23:29:27Esoterikive got all files in the need location , i see my ipod in my computer , but i ran ipodpatcher 0-26 and nothing
23:29:42BagderEsoterik: 80GB?
23:30:05Bagderthen I figure it should work... but I'm not the ipod expert either
23:30:25 Join JoeyBorn [0] (n=rootmeis@
23:30:32bluebrotheriirc 5.5G 30GB still needs a patched ipodpatcher
23:30:47bluebrotherbut no ipod experience here.
23:31:39EsoterikAhh.. so thats what i forgot / missed... im gunna find one and give it another go :P
23:33:35CassandraWell, I've got WinCVS talking to the repository, and everything in in the manner it's supposed to be in. I think that counts as a win.
23:33:43CassandraI should probably stop there.
23:34:02CassandraOh, bagder, can you add the rbutil module to the notification system please?
23:34:15 Join Paul_the_Nerd [0] (
23:34:28tucozis that module possible to check out now?
23:34:39peturWinCvs? better than tortoise cvs?
23:35:00Cassandrapetur: No idea. I don't normally use CVS under windows.
23:35:06CassandraIt's kind of ... monolithic.
23:35:23CassandraAnd seems to gratuitously use services.
23:35:38CassandraAlso not user friendly. But this *is* CVS we're talking about.
23:35:50peturtortoise cvs intergates nicely into the explorer
23:36:05CassandraThat sounds better then.
23:36:46tucozi found out that it was in fact easy to "delete" a directory tree with cvs. Just cvs remove -f <dir>/* ; cvs commit
23:37:20tucozBut i found that out after i had done the recursion manually :/
23:37:35 Quit JoeBorn (Read error: 145 (Connection timed out))
23:38:19peturand if we _ever_ switch, there's :D
23:38:52 Nick Everybody is now known as Everybody|away (
23:38:57Cassandrapetur: Right after someone implements track seeking, I expect.
23:39:15tucozaudible track seeking?
23:39:35 Quit midkay (Read error: 110 (Connection timed out))
23:40:05tucozas far as i know, track seeking is implemented already
23:40:15*petur sends himself to bed
23:40:15tucozat least for the codecs i use
23:40:21 Quit petur ("Zzzzz")
23:41:04 Join My_Sic [0] (
23:41:22bluebrotherI had a look in wincvs quite some while ago ... and went straight to tortoisecvs. Still cervisia isn't as good :(
23:41:32linuxstbEsoterik: Which version of ipodpatcher are you using, and which version of Windows?
23:41:57 Quit lee-qid ("aufwiederbyebientotsayonara")
23:42:13 Part tucoz ("Good night")
23:42:39EsoterikLinuxstb: i got it to work :D... i had the wrond HDD not to mention the wrong patcher... but all is well so far thanks
23:43:09linuxstbDid you use ipodpatcher v0.4?
23:43:11bluebrotherhmm, I should also get some sleep
23:43:20 Quit bluebrother ("set mode sleep on")
23:43:47Esoterikumm i got it from Psiuyo's build or link i should say , im not sure what the build is...
23:44:17linuxstbSo that's the unofficial version... The latest official version should have worked as well.
23:45:19Esoterikoh well thats good to know... it was probably just the fact i had the wrong root folder to begin with but i changed both at the same time
23:48:53Cassandratucoz: with audio, I meant.
23:53:41 Quit Ekron ("Leaving")

Previous day | Next day