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 2004-09-22

00:16:59 Quit ze (Read error: 104 (Connection reset by peer))
00:20:42 Join ze [0] (
00:28:50 Join bagawk [0] (lee@bagawk.user)
00:44:01bagawkhey midk
00:51:33***Saving seen data "./dancer.seen"
01:18:43 Quit bagawk (Nick collision from services.)
01:19:04 Join bagawk [0] (lee@bagawk.user)
01:21:43 Part midk ("Leaving")
01:25:26 Join midk [0] (
01:39:32 Part amiconn
01:39:37bagawkhi midk
01:54:50 Quit bagawk ("Leaving")
02:01:01 Join midk_ [0] (
02:01:11 Quit midk ("Leaving")
02:01:39 Nick midk_ is now known as midk (
02:19:35 Quit AciD (Remote closed the connection)
02:51:36***Saving seen data "./dancer.seen"
03:02:29 Join ze__ [0] (
03:02:47 Quit ze (Nick collision from services.)
03:02:53 Nick ze__ is now known as ze (
03:11:39 Quit midk ("Leaving")
03:11:56 Join midk [0] (
03:19:16 Join frsyj [0] (
03:19:30 Quit frsyj (Client Quit)
03:26:07 Quit midk (Remote closed the connection)
03:26:25 Join midk_ [0] (
03:26:31 Nick midk_ is now known as midk (
03:49:02 Join frsyj [0] (
03:53:45 Join bagawk [0] (Lee@bagawk.user)
03:54:17*frsyj is away - Automatically set away. - messages will be saved.
04:27:05 Quit plok ("Leaving")
04:37:15midkback in a bit
04:51:39***Saving seen data "./dancer.seen"
05:01:23 Nick frsyj is now known as plok (
05:28:04 Quit bagawk ("umount /dev/brain")
05:44:01 Part scott666_
06:51:40***Saving seen data "./dancer.seen"
06:55:12 Quit midk (Remote closed the connection)
06:58:09 Join LinusN [0] (
07:03:56 Join midk [0] (
07:05:17 Quit MrMoo (Read error: 110 (Connection timed out))
07:07:21plokLinus, I've pulled apart my AJB6k to see if the battery connections need resoldering (battery always drains very fast,unit powers off intermittently). Should I be able to see if the contacts need resoldering or would I not be able to tell just by looki
07:11:17 Quit midk (Read error: 104 (Connection reset by peer))
07:14:43LinusNplok: bend the small pcb outwards to see if it's firmly soldered to the chassis
07:18:40 Join midk [0] (
07:21:24plokAah, I'm at work at the moment, but from memory, when I put the batteries in the small PCB sections were visibly pushed outwards. I will check tonight to see if the connection has been compromised. Thx!
07:22:44LinusNyou're welcome
07:57:33dwihno\o/ wednesday \o/
08:01:02dwihnoWednesday, orange flavored chocolate. Stuff can't be better :)
08:02:07 Join ashridah [0] (
08:18:03 Quit Nibbler (Read error: 110 (Connection timed out))
08:21:05 Join amiconn [0] (
08:27:33LinusNfunny, since i committed my new adc code, my recorder sometimes senses OFF button keypresses when the drive spins up
08:27:50LinusNmust be some kind of voltage drop when the drive motor starts
08:35:34LinusNlet's hope some button filtering can cure that
08:39:31 Quit plok ("Leaving")
08:49:37 Join plok [0] (
08:50:00 Join [IDC]Dragon [0] (
08:51:43***Saving seen data "./dancer.seen"
08:52:02[IDC]DragonLinusN: my button filter code needs to get brushed up a bit, it's more alpha than I thought, and for recorder keypad only
08:52:18[IDC]Dragonso I didn't commit it yesterday
08:53:51 Join Nibbler [0] (
08:55:23*plok is away - Automatically set away. - messages will be saved.
08:58:38LinusN[IDC]Dragon: i see
08:59:07plokHas anyone experienced a very "sticky" player button on the archos. Stop and Forward (down and right) are very hard to press/won't spring back into position?
09:04:46 Quit ashridah (Read error: 110 (Connection timed out))
09:06:42 Join oxygen77 [0] (
09:06:54[IDC]DragonLinusN: but the main idea is to check if the last N measurements are stable within a certain corridor, and only than do the range check
09:07:42[IDC]DragonI can't explain your OFF button problem, because that's a digital input, lots of margin
09:08:06[IDC]Dragon(I only filter the channels with multiple buttons)
09:09:28 Join Zagor [242] (
09:10:13[IDC]Dragonhey Björn!
09:13:36LinusN[IDC]Dragon: yeah, the OFF button is unrelated to the ADC, so i can't explain the false detection
09:13:49LinusNunrelated on the v1, that is
09:13:58LinusNhi Zagor
09:14:34ZagorLinusN: how frequently does the spurious off occur?
09:15:01LinusNit seems to only happen when the disk spins up, and it's only occasionally
09:15:25LinusNhappened twice on my 35-minute drive to work
09:15:26Zagoris it any different if charger is connected or not?
09:15:39LinusNi haven't tried with the charger
09:16:05LinusNpeople have reported similar problems with the old adc driver as well
09:16:22LinusNbut i haven't seen it on my jukebox until now
09:16:44ZagorI have only seen ADC problem reports, not GPIO buttons
09:16:45LinusNi wonder if the adc conversion affects the current draw of the cpu
09:17:04LinusNspurious OFF events have been reported afaik
09:19:42mbrJust saw you are talking about the OFF button problem
09:19:52mbrMy recorder also suffers of this
09:20:00LinusNv1 or fm?
09:20:09mbrOnly if battery is below 50%
09:20:17mbrand only if disk is spinning
09:21:12LinusNi can imagine that it is a problem with inadequate decoupling or something like that
09:22:34LinusNi wonder if it happens if "disk poweroff" is off?
09:23:19LinusNhmm, it was already off on my jukebox
09:23:58mbrmine too
09:25:29mbrfor now i "fix: it with turning keylock on :(
09:26:05mbrDo you think it could be a battery problem?
09:26:09mbrbad batteries?
09:26:15LinusNi don't think so
09:26:56LinusNi still wonder why the faster ADC conversion triggers the problem for me
09:27:06mbrI wasn't sure because on disk access my backlight also gets a litte bit darker
09:27:20mbrand i can't remember how it was before
09:30:37Zagorwe need to measure the current draw with different AD use
09:38:27[IDC]Dragonmabe we shouldn't treat the binary port C inputs as analogue
09:39:48 Join amiconn_ [0] (
09:42:49Zagorwe don't use port C, we use AN[0-8]
09:44:48Zagor[IDC]Dragon: how do you mean?
09:46:47Zagorbtw, for those who don't read the misticriver forum: the ihp100 only has 16MB ram, 120/140 32MB
09:48:15Zagorthere's also a comment saying 120+ being able to drive a mic in the line-in port, while <120 can't
09:51:14LinusNZagor: the analogue inputs can be read "digitally" in Port C
09:53:14LinusNwould work with the ON/OFF buttons on the FMR, and we did that in earlier versions of the Player button code
09:54:05LinusNhowever, it is a bad idea to mix the two ways, since the port c bits are invalid during the conversion
09:55:27[IDC]DragonLinusN: all of them, or those configured as analogue?
09:55:50Zagorbut since this also happens on v1, which has off on a gpio port, I don't see the relevance.
09:56:22[IDC]Dragongood point
09:56:32 Quit amiconn (Read error: 110 (Connection timed out))
09:56:32 Nick amiconn_ is now known as amiconn (
09:57:39LinusN[IDC]Dragon: only those that are being converted
09:58:07[IDC]Dragonwhich is the whole batch of 4?
09:58:33[IDC]Dragoneven if one inbetween is configured as digital in?
09:58:51LinusNthe docs aren't clear on this, but i wouldn't trust any of the 4 during the scan
09:59:01*[IDC]Dragon could dig it up in the datasheet, excuse laziness
09:59:23LinusNonly the bits that are included in the scan
09:59:31LinusNare affected
10:00:00LinusNbtw, there is no digital/analog configuration
10:00:19LinusNthe portc bits are always active
10:00:45LinusNbut the respective bits are invalid during the conversion
10:25:05Bagdernot yet "configure upgrade" compatible
10:25:23Bagderbut a lot simpler makefiles, if I make say so
10:31:57Bagderthe goal here is to move platform-specific knowledge to configure or keep in the sources
10:32:18LinusNBagder: the current binutils CVS includes the EMAC patch for the coldfire
10:32:42LinusNi don't know how well it plays with gcc though
10:32:54Bagderit'll be exiting
10:33:11LinusNit seems like gcc 3.4 has better coldfire support
10:33:42[IDC]DragonZagor: I found a buglet in fat.c, panicf("Writing before data\n") doesn't work as intended
10:34:10 Join MrMoo [0] (~me@
10:34:41[IDC]Dragonbecause start is already offsetted by the partition start, versus fat_bpb.firstdatasector is not
10:35:23Zagorright, good catch.
10:35:26[IDC]DragonI suggest to do the offset within transfer(), not by the caller
10:35:34[IDC]Dragonsaves code, too
10:35:44Bagder"Code generation for the ColdFire processors family has been enhanced and extended" (gcc 3.4 changelog)
10:36:10oxygen77hello all, I have a question on wav playback (and I know it's on nodo and that it has been discussed a lot)
10:36:27Zagor:) ask away
10:36:51oxygen77what I want to know is what is known about code that can be downloaded on the MAS
10:36:55oxygen77(for linav)
10:37:23LinusNoxygen77: it is possible to d/l code to the mas
10:37:37LinusNand we have code for WAV playback for the 3507D
10:37:58oxygen77I have also the code for the 3587F
10:37:59LinusNbut that code requires a different h/w configuration than the jukebox has
10:38:12LinusNoxygen77: nice, can i have it?
10:38:20LinusNdo you have 3587f dox as well?
10:38:32oxygen77oh ok
10:38:46oxygen77the one from micronas website
10:38:52Bagderh4ck3r lingo linus ;-)
10:38:59LinusNthe standard data sheets?
10:39:56Zagorhow is the mas connected to the cpu on the avos?
10:40:36[IDC]Dragonwe'd need a microcode that can take wav input at the serial port
10:40:48[IDC]Dragonnot just parallel
10:41:18oxygen772 connections
10:41:19[IDC]Dragonbecause out parallel is input only :-(
10:41:37LinusNthe wav code for the mas3507d is serial
10:41:47Zagor[IDC]Dragon: yes and the serial port is too slow for 44.1 kHz wav
10:42:00LinusNbut it is not using the demand protocol
10:42:00oxygen77parallel and 22c
10:42:02[IDC]DragonI'm not sure
10:42:23LinusNwav playback over i2c? i doubt it
10:42:43oxygen77no but PIO
10:42:53oxygen77i2c is used for control
10:43:09[IDC]DragonZagor: we may use 3 MBit
10:44:37oxygen77LinusN, did you disasm the wav code?
10:44:51LinusNi have the source
10:44:59Zagor[IDC]Dragon: how?
10:45:16LinusNit depends on SDI and SDO having the same clock
10:45:26oxygen77and you konw how it works?
10:45:28LinusNwhich it hasn't on the jukebox
10:46:09ZagorBagder: i don't think "configure update" is terribly important
10:46:14Bagderme neither
10:46:26Bagderits _reaaally_ hard to maintain
10:46:35LinusNso we might be able to write a wav playback module that can use the demand protocol like the mpeg decoder does, but there is no docs on how to do that
10:46:41BagderZagor: I just tried this instead:
10:47:04Bagderadd "2\n\2\nN\n\n" in a file, then run "../tools/configure < file"
10:47:28Bagderthat is a lot easier
10:51:40LinusNi worked a little on wav playback some time ago, got some funny noises, but i couldn't figure out how to implement the demand protocol and the buffering
10:51:40LinusNthe wav module from micronas is simply a pass-through device
10:51:46***Saving seen data "./dancer.seen"
10:52:34BagderZagor: and do try my 'make clean' ;-)
10:52:49Bagderleaves nothing but the Makefile
10:52:58Zagorthat's nice
10:53:11Zagorcommit this, it's better in every way
10:54:26LinusNBagder: "clean" is using wildcards?
10:54:38Bagderjust fixed to clean better
10:55:00LinusNBagder is a makefile ninja
10:55:59LinusNZagor: i want the broken iriver NOW!
10:56:17BagderZagor: will you take care of running configure on the automatic builds?
10:57:21Bagderwe could alter the scripts to use the "configure < input" approach isntead of "configure update"
10:57:31Bagdersince the update one isn't working anymore
10:57:58amiconn[IDC]Dragon: I now have a strong suspicion what causes the serial mmc communication problems, as I now understand some dirty trick that the archos xfer routines do. Will have to prepare some special test pattern and hook up the scope to verify this
10:58:12Bagdercommit coming up
10:59:00LinusNecho -e "2\n2\nN\n" | configure
10:59:39Bagderbut one return nire, isn't it?
10:59:40Bagderfor language
11:00:46Bagdergotta go, a little girl is waking up...
11:01:25Zagorehe, can't use '2' for all builds ;)
11:07:30oxygen77LinusN, do you want the code (bin version) of the wav decoder used on av3xx?
11:08:05 Join AciD [0] (
11:08:22LinusNoxygen77: nah, i won't have the time to analyze it, and to be honest, wav playback isn't top priority for me
11:10:41[IDC]Dragonhey, let's not lose this
11:10:57oxygen77I want to have it work on the av to try ogg playback
11:11:25oxygen77could you send me the source you have?
11:11:36LinusNoxygen77: hang on
11:11:38[IDC]Dragonoxygen77: how about putting it intoa twiki page?
11:12:18oxygen77[IDC]Dragon, what do you mean?
11:13:55[IDC]Dragonwe should keep that code
11:14:26oxygen77it's only a bin code directly from archos firmware
11:15:51Zagordon't put archos code in the wiki. it's not legal.
11:16:14oxygen77yup, that's why I'm asking ;)
11:18:38 Quit AciD (Read error: 104 (Connection reset by peer))
11:19:29 Join AciD [0] (
11:48:35[IDC]Dragonoxygen77: if nobody wants it, email it to me
11:54:20[IDC]DragonZagor: how strict should I be with my #ifdef HAVE_FAT16SUPPORT ?
11:54:39[IDC]Dragonright now, I have bracketed every tiny extension
11:55:21LinusNyou mean that the code could coexist without the #ifdef?
11:55:51[IDC]Dragonwe could relax it a bit, leaving afew unused members and statements in case it's not defined (FAT32 only)
11:56:16[IDC]DragonLinusN: coexist?
11:56:55LinusNcoexist == both fat16 and fat32 in the driver, without conditionals
11:57:16LinusNi see no problems supporting both types
11:57:22[IDC]Dragonif I define HAVE_FAT16SUPPORT, the code can doboth, yes
11:57:33LinusNthe only reason not to would be code size
11:57:38[IDC]Dragonif I don't define it, it's FAT32 only
11:58:01[IDC]Dragonyes, in order not to carry ballast for the HD models
11:58:55LinusNi think you can be the judge of that, whichever you feel is best
11:59:08LinusNthe only problem would be cluttered code
11:59:37LinusNif there are lots of ifdefs, that is
11:59:40[IDC]Dragonthe I'd commit the strict version first, I can still relax it
12:01:16[IDC]Dragonthedefine is currently checked at 17 places
12:07:54LinusNi agree, commit what you have, and we can strip the ifdefs later if we want
12:31:58 Quit oxygen77 (Read error: 110 (Connection timed out))
12:32:25 Join oxygen77 [0] (
12:47:30Zagorwhoa, $50 donation just came in
12:51:50***Saving seen data "./dancer.seen"
12:54:28[IDC]Dragonsomebody we know?
12:54:50[IDC]Dragonwhat's the donation balance, if I may ask?
12:54:57Zagori don't know him anyawy :)
12:59:19Zagorfund balance is about a thousand dollars
13:01:05[IDC]Dragonnot bad for getting some boxes
13:03:05[IDC]DragonZagor: is there some script for the fat test?
13:03:21[IDC]Dragonlike a regression test or so?
13:03:46[IDC]Dragonoh, I overlooked that
13:05:39[IDC]Dragonhmm, looks like I can't easily tun that on Windows
13:06:50Zagorcygwin should be able to run it
13:07:03Zagoror even a standalone bash.exe
13:07:16[IDC]Dragoneven building the disk image, mounting it, etc. ?
13:07:33Zagorah, no I forgot that
13:07:52Zagora great reason to switch to linux ;)
13:08:56[IDC]DragonI once tried, I can say to my defense
13:09:16ZagorI'm just joking
13:09:20[IDC]Dragonbut knoppix failed to properly install lilo
13:10:17[IDC]Dragonjust a few weeks ago I gave up the 20 GB hole in my HD partitioning, which I reserved
13:11:22ZagorI'm thinking about displaying last 5 donations on the website. do you think people would mind having their names listed?
13:12:41[IDC]Dragonas a donator, I would welcome some "official" appreciation
13:13:19[IDC]Dragonpeople donate money during TV galas just to see their name scroll through
13:13:28amiconn[IDC]Dragon: As we discussed the way of browsing both internal & external flash at the same time and I said that I'd prefer the unix way, there is another reason why this way should be preferred: This way, the path /.rockbox would be always valid
13:13:46[IDC]DragonHi Jens!
13:14:03[IDC]DragonI'd prefer that way, too
13:19:29[IDC]Dragonhow's th MMC in general?
13:26:10amiconn[10:58:00] <amiconn> [IDC]Dragon: I now have a strong suspicion what causes the serial mmc communication problems, as I now understand some dirty trick that the archos xfer routines do. Will have to prepare some special test pattern and hook up the scope to verify this
13:26:53[IDC]Dragonwhy is the sync mode so tricky?
13:31:13amiconnWhich the SH1 SCI, it is rather tricky to receive well-defined amounts of bytes, because (1) the SH tries to recieve (and clocks the SCI) as long as the receiver is enabled, and then stops on overrun error. (2) the SH SCI uses double buffering
13:33:25amiconnThe archos routines use a rather clever trick to circumvent this problem
13:41:40[IDC]Dragongood that you peeked in...
13:41:56[IDC]Dragonsorry, I thought this would be easier
14:09:13 Quit MrMoo (Read error: 54 (Connection reset by peer))
14:22:33[IDC]DragonZagor: the testsuite fails a test
14:22:41[IDC]Dragonon my FAT16
14:23:10 Join pike|| [0] (
14:23:17[IDC]Dragonit's strange, it panics with a non-empty dir entry
14:23:28Zagorthat's what it's for :)
14:23:52[IDC]DragonI haven't touched the dir entry handling
14:24:03[IDC]Dragonotherwise, it went quite far
14:24:13Zagortry it with a fat32 disk and see if the same thing happens
14:24:35[IDC]DragonI have no disk image at hand
14:24:48Zagorhow are you testing fat16 then?
14:25:02[IDC]Dragonwith a FAT16 disk image ;-)
14:25:15[IDC]DragonI meant, no FAT32 image at hand
14:25:31[IDC]Dragonthe one I have is the dumped Ondio
14:25:40ZagorI won't mention linux again, promise :)
14:26:11 Quit pike (Read error: 104 (Connection reset by peer))
14:26:11 Nick pike|| is now known as pike (
14:27:16[IDC]Dragonthis will have to wait for later
14:27:47Zagorif you send me the code I can run some tests
14:28:25[IDC]Dragonthanks, we'll do that later
14:29:00[IDC]DragonI have to mergethe code first, unfortunately the one I modified is 2 revisions old
14:29:39[IDC]Dragonthat was my working PC simulation, a bit aged
14:30:06[IDC]Dragonnot too bad, I will patch it
14:30:48 Quit Nibbler (Read error: 60 (Operation timed out))
14:43:57 Join MrMoo [0] (~me@
14:51:49[IDC]DragonZagor: I've emailed you the merged fat.c now
14:51:53***Saving seen data "./dancer.seen"
14:52:11 Join Poddan [0] (
14:52:17Poddantjena alla
14:52:43[IDC]Dragontjena Poddan
14:52:46Poddanvad står MDB för i Sound Settings?
14:53:03Poddanär det extra bas, eller något helt annat?
14:53:09ZagorMicronas Dynamic Bass (och prata engelska, det är en intl kanal)
14:53:12[IDC]Dragonno, I don't speak svenska
14:53:24Poddanoh sorry xD
14:54:40kaboofaich liebe apfels!
14:54:57[IDC]Dragonsorry: Äpfel
14:55:02*Bagder ducks to avoid the language throwing ;-)
14:55:16kaboofaare you guys going to stop developing rockbox for archos and just move to iRiver?
14:55:29Bagderwe support Archos now
14:55:34kaboofayeah, i know this.
14:55:36BagderiRiver is a possibility
14:55:38Zagorkaboofa: no
14:55:38Bagderin the future
14:55:52*kaboofa does happy fat-polish-kid-dance
14:56:05[IDC]DragonZagor: you'd need to define HAVE_FAT16SUPPORT somewhere suitable
14:56:17*kaboofa goes back to programming :)
14:56:20kaboofasee you all later.
14:57:13Poddanwhat do u think the optimal settings on MDB are for Metal, Rock etc?
14:58:26Zagorit all depends on your headphones/speakers. I use flat for everything.
14:58:58PoddanI have Sony headphones, the ones that are REALLY comfortable, the ones that look the Koss Plug
14:59:05Poddanso the bass is really good in those
14:59:24ZagorEX-70, yes. I use those too.
14:59:42Poddannice :D
14:59:57Zagorno need for bass amplification with those, imo
15:00:06Poddanthat's why I'm trying to configure the sound settings for that perfect sound....but that's just impossible
15:01:05Zagormaybe because "perfect sound" is subjective?
15:01:39Bagdersubjetive to others, objetive to me! ;-P
15:01:55Poddanyeah, would be a nice function if u could set the settings for each song :D
15:02:03Bagderis it called "objective" ? I doubt that
15:02:41Poddanyeah, objective :D haha, no idea actually
15:02:52ZagorBagder: it is
15:03:15 Join maikeul [0] (
15:03:28Zagor"a. Uninfluenced by emotions or personal prejudices: an objective critic. b. Based on observable phenomena; presented factually: an objective appraisal."
15:05:07[IDC]Dragonnot open, boo
15:05:23BagderI like wikipedia
15:05:29Poddanso happy that I have my Jukebox back...the HD failed on me for about 4 months ago...but now it's upgraded to a 40 GB =)
15:06:15Zagoropen? as in editable?
15:07:03[IDC]DragonI thought w're so open here, it'll match
15:08:29 Join elinenbe [0] (~elinenbe_@
15:09:04 Part elinenbe
15:09:14 Join elinenbe [0] (~elinenbe_@
15:09:19Zagoryeah, but I don't know of any such open dictionary. wikipedia is more of an encyclopedia.
15:10:50elinenbewikipedia has a dictionary as part of the site
15:11:42 Quit gromit` (Read error: 110 (Connection timed out))
15:11:48Zagornot a good one. try looking up "objective" and get a distinct definition and some synonyms
15:17:32 Quit Poddan ("CGI:IRC")
15:21:34 Join R3nTiL1 [0] (
15:23:22[IDC]DragonZagor: running fat.c already?
15:23:49*LinusN is checking out binutils from CVS, to fix an assembler bug
15:24:01*[IDC]Dragon whistles, looks up in the sky
15:24:26[IDC]DragonLinusN: what bug?
15:24:41[IDC]Dragonwhat CPU?
15:25:02LinusNcoldfire, internal error
15:25:04Zagor[IDC]Dragon: sorry, have to do some other stuff first
15:33:25*oxygen77 is away: chui pas là
15:36:31 Join srn [0] (
15:37:46srnHi Linus. Why did you change the UDA1380TT Stereo audio coder/decoder info back in the Wiki. Is it not correct that it can NOT decode mp3?
15:39:51LinusNsrn: because the Philips official name is like that. check the philips web page
15:40:05LinusNnone of us believe that it can decode mp3
15:40:17srnyeah, I know but it can confuse.
15:40:42LinusNnever mind, you can change it back if you wish
15:40:43Zagorwell it's a misleading name. just because we know what it is doesn't mean other people will be confused by a chip claiming to be a codec
15:40:55 Join ze__ [0] (
15:40:56Zagor*will not*
15:41:07LinusNi agree, my bad
15:42:24LinusNi have reverted my change
15:42:41Zagoroy, the bleeding build table is confused by all my test builds... :)
15:42:44srncool, Linux. I am looking forward to hear about the BDM interface - I hope it works. c u :)
15:44:26 Quit srn ("CGI:IRC (EOF)")
15:51:28 Quit ze (Read error: 110 (Connection timed out))
15:51:28 Nick ze__ is now known as ze (
15:51:54 Nick R3nTiL1 is now known as R3nTiL (
15:55:54*LinusN just fixed his first gas bug :-)
15:56:29LinusNEMAC register handling bug
15:57:15*LinusN is fixing emac bugs with emacs :-b
15:59:07LinusNgotta go
15:59:09 Part LinusN
16:08:02 Quit R3nTiL ()
16:09:13[IDC]DragonZagor: I have a suspicion about the FAT16 bug, because the root dir changes the cluster when it happens :(
16:17:09[IDC]Dragongrmph, the compiler should warn when I check an unsigned for < 0
16:18:37Zagorfound the bug?
16:19:44[IDC]Dragonbetter now
16:20:48[IDC]Dragonchanged the parameter of get_next_cluster() to signed
16:21:15[IDC]Dragonit wasn't detecting the root dir case
16:21:32*[IDC]Dragon re-runs the suite
16:23:08Bagderthe compiler does warn for that if you pass on pickier flags
16:23:21Bagdergcc that is
16:23:34amiconn[IDC]Dragon: Does changing the cluster numbers to signed not break FAT32 compatibility? It would if cluster numbers > (2^31 - 1) are allowed with FAT32...
16:24:51Zagoryes but 2^31 is 65 TB with 64 sectors/cluster...
16:24:53[IDC]Dragonno, clusters are 28 bit
16:25:08Zagorah, correct. bad memory
16:25:20amiconnAh ok.
16:25:29[IDC]Dragonthey are just internally signed, this doesn't go to disk
16:25:42amiconnShould be called FAT28 then...
16:26:08[IDC]Dragonstill a lot of data
16:26:17Zagorluckily, they are not stored with 28 bits each...
16:26:34BagderZagor: you know what debian package that has mkfs.vfat ?
16:26:37[IDC]Dragonupper 4 are reserved to M$
16:28:05ZagorBagder: dosfstools contains mkdosfs, which is what mkfs.vat uses. can't login to my home machine right now
16:28:21Zagor(to check if it's changed in unstable vs. testing)
16:28:25amiconn8 TB with 32 KB/ cluster. 2 TB is the limit for 32 bit sector numbers anyway...
16:28:41*Bagder just upgraded his harddisk
16:28:53[IDC]Dragonthe suite completes now
16:29:09[IDC]Dragonno filesys checks in cygwin, though
16:29:20[IDC]Dragonit's just not bailing out
16:29:33 Join methangas [0] (
16:30:00[IDC]DragonZagor: can you remove the unsigned,or shall I send the file again?
16:31:37Zagorsend it again, but I found the machine i use for testing doesn't have vfat in the kernel(!) so I can't test properly until tonight :(
16:33:08[IDC]Dragonno hurry, while no MMC driver
16:33:54[IDC]Dragonit'll probably take a while to run, anyway
16:35:42BagderVendor: FUJITSU Model: MHT2080AT
16:35:48Bagderand it even works! ;-)
16:36:23BagderI noticed it was a Hitachi I took out
16:38:19[IDC]Dragonthat's a nice one
16:38:28[IDC]Dragonthe new one, I mean
16:39:58[IDC]Dragon100GB are outrageously expensive
16:41:04dwihnoThere are 2.5" 100 GB disks now?
16:41:40Bagderabout 2 - 2.5 times the price of the 80GB
16:41:59dwihnoInsane :)
16:42:06dwihnoInsane pricing for insane people :)
16:43:21[IDC]DragonI payed $1000 for 2GB, 10 years ago
16:44:06*[IDC]Dragon stops telling old war stories now
16:46:44 Join Lynx_ [0] (lynx@
16:51:55***Saving seen data "./dancer.seen"
17:13:36*oxygen77 is back (gone 01:40:11)
17:20:46 Join webguest44 [0] (
17:21:01 Quit webguest44 (Client Quit)
17:21:28 Part Zagor
17:24:58 Join mecraw [0] (~lmarlow@
17:43:41 Join teej [0] (
18:00:18 Quit pike (Read error: 54 (Connection reset by peer))
18:20:11 Join pike [0] (
18:25:35 Quit [IDC]Dragon ("CGI:IRC")
18:44:34 Join ze__ [0] (
18:51:56***Saving seen data "./dancer.seen"
18:59:36 Quit pike ("2.0 Build 3515")
18:59:36 Quit elinenbe (Read error: 104 (Connection reset by peer))
19:00:00 Quit ze (Read error: 110 (Connection timed out))
19:00:00 Nick ze__ is now known as ze (
19:02:50 Join kaouete [0] (
19:03:05 Part Lynx_
19:04:52 Join pike [0] (
19:04:54kaouetei have a strange behavior with my archos, i think it's linked to the battery, but its' strange. When i am listening music, when i reach about 70% of battery, the music stops for 1/2 sec, play 1/2 sec and stop completely :/
19:05:07kaoueteis it linked to rockbox or maybe should i change my battery ?
19:06:06 Quit pike (Client Quit)
19:08:08 Quit teej (Read error: 60 (Operation timed out))
19:08:33kaouete(and i cant even start it now without re-charging it)
19:42:41Bagdertry a new set of batteries
19:56:33kaoueteok, are 2200mAh supported now ?
20:36:50 Join elinenbe [0] (~elinenbe_@
20:51:57***Saving seen data "./dancer.seen"
21:29:15 Join HenrikB [0] (
21:35:22Bagderkaouete: 2200 mah has always been supported
21:37:28 Quit HenrikB (
21:37:28 Quit mecraw (
21:37:28 Quit MrMoo (
21:37:28 Quit oxygen77 (
21:37:28 Quit Sebulba02 (
21:38:57NJoinHenrikB [0] (
21:38:57NJoinmecraw [0] (~lmarlow@
21:38:57NJoinMrMoo [0] (~me@
21:38:57NJoinoxygen77 [0] (
21:38:57NJoinSebulba02 [0] (
21:39:23 Join mecraw_ [0] (~lmarlow@
21:39:53 Join HenrikB_ [0] (
21:41:43 Join [IDC]Dragon [0] (
21:42:12[IDC]DragonHello Henrik(s)
21:44:48elinenbehello [IDC]Dragon
21:44:59[IDC]DragonHi Eric
21:51:21 Quit oxygen77 (
21:51:21 Quit MrMoo (
21:51:21 Quit mecraw (
21:51:21 Quit HenrikB (
21:51:21 Quit Sebulba02 (
21:51:52 Join scott666_ [0] (
21:52:32NJoinHenrikB [0] (
21:52:32NJoinmecraw [0] (~lmarlow@
21:52:32NJoinMrMoo [0] (~me@
21:52:32NJoinoxygen77 [0] (
21:52:32NJoinSebulba02 [0] (
21:53:24 Quit oxygen77 (
21:53:24 Quit MrMoo (
21:53:24 Quit mecraw (
21:53:24 Quit HenrikB (
21:53:24 Quit Sebulba02 (
21:54:04 Join Zagor [0] (
21:54:07NJoinHenrikB [0] (
21:54:07NJoinmecraw [0] (~lmarlow@
21:54:07NJoinMrMoo [0] (~me@
21:54:07NJoinoxygen77 [0] (
21:54:07NJoinSebulba02 [0] (
21:54:40 Quit mecraw (Connection timed out)
21:55:46elinenbe[IDC]Dragon: how is the Ondio progress coming?
21:59:04 Quit HenrikB (Read error: 110 (Connection timed out))
22:01:45[IDC]Dragonelinenbe: pretty good, I've implemented FAT16 now, needs testing
22:03:06[IDC]DragonJens is working on the MMC driver, these 2 things are the key components
22:03:22Zagor[IDC]Dragon: fat.c line 294 is missing 'int'. it gives a warning in gcc.
22:03:50[IDC]DragonHi Zagor, I didn't notice you're here
22:04:04Zagori just came in
22:04:07[IDC]Dragonamiconn sounds no good
22:04:21[IDC]Dragonah, with that split
22:04:52[IDC]Dragonmissing int's go unnoticed in MSVC
22:05:03amiconn[IDC]Dragon: I got an MMC today, so I can test external access
22:05:47amiconnIt is indeed the case that when you plug in USB with an MMC inserted, the original firmware asks you to remove the card (for reset).
22:06:32[IDC]Dragonyeah, I'm afraid without a power switch it has to be that annoying
22:07:25amiconnYou can then insert and extract the card at will while in USB mode. When the card is not inserted, the PC presents you with the contents of the built-in flash
22:09:41Zagorahhh, the test case creates a too big root dir... :)
22:09:48 Part oxygen77 ("Cho")
22:10:15[IDC]Dragonoh, it runs that far?
22:10:29[IDC]Dragondoes it gracefully fail?
22:10:47Zagorit never runs the fat driver at all. it fails creating the test disk :-)
22:11:04amiconn[IDC]Dragon: While this does work with the USB mode of rockbox, rockbox doesn't notify the PC host that the card was changed
22:11:05[IDC]Dragonhow can?
22:11:36Zagorit sets up the test disk by creating lots of files on it, and many files with long names. too many to fit in a fat16 root dir. i need to make a subdir for this instead.
22:11:37[IDC]Dragonamiconn: the partition is re-mounted after USB anyway, iirc
22:12:22[IDC]DragonZagor: ah, that part, I had to skip that (no ext dir, no mounting)
22:12:24amiconnOf course. I mean that the archos fw disables and re-enables the usb bridge whenever a card is inserted or removed while in USB mode
22:17:07HenrikB_Any makefile ninjas here? No rocks are created for me after Bagder's makefile update
22:17:17[IDC]DragonZagor: it would be a more inteesting test case if some of the "seed files" get deleted, too, to create FAT holes
22:19:49Zagoryes. but this is enough to make it fail anyway
22:20:07Zagormkdir error: file exists
22:20:07ZagorFailed creating directory
22:22:23Zagorrather strange error
22:23:28BagderHenrikB_: plain recorder build?
22:24:02[IDC]Dragonno good
22:24:20[IDC]Dragonin the driver this time, I guess?
22:24:44HenrikB_plain recorder on cygwin
22:25:59BagderHenrikB_: any error message?
22:26:56amiconnBagder: Same here; tried FMR and player so far. No error message for FMR, the player build generates plenty of them :(
22:27:47HenrikB_Bagder, nope it just enters and leaves the plugin directory. The libplugin.a is built though.
22:27:51amiconnPlayer build doesn't even complete, exits with error on building sysfont.o
22:28:04Bagderamiconn: then rebuld convbdf
22:28:12HenrikB_Same for me
22:28:20Bagderrun make in tools
22:28:22amiconnBagder: Did you change it today?
22:28:58Bagderit now produces #ifdef have_lcd_bitmap
22:29:00Bagderin the code
22:29:14 Quit methangas (" Try HydraIRC -> <-")
22:29:24amiconn"make clean" errors out too...
22:30:04BagderI need details
22:30:18amiconn$ make clean
22:30:18amiconnmake[1]: Entering directory `/home/Administrator/rb-patched/firmware'
22:30:18amiconncleaning firmware
22:30:18DBUGEnqueued KICK amiconn
22:30:18amiconnmake[1]: Leaving directory `/home/Administrator/rb-patched/firmware'
22:30:18amiconnmake[1]: Entering directory `/home/Administrator/rb-patched/apps'
22:30:18***Alert Mode level 1
22:30:18amiconnmake[1]: *** [/home/Administrator/rb-patched/build/player/dep-apps] Error 1
22:30:20amiconnmake[1]: Leaving directory `/home/Administrator/rb-patched/apps'
22:30:22amiconnmake: *** [clean] Error 2
22:31:07Bagderdoes dep-apps get created and contain dependencies?
22:31:50amiconnBuilding just failed.. cleaned dir and try building from scratch..
22:32:09amiconnBuilding still doesn't work (player, that it):
22:32:18amiconn$ make >make.log
22:32:18amiconnmake[1]: *** [/home/Administrator/rb-patched/build/player/dep-apps] Error 1
22:32:18amiconnmake: *** [all] Error 2
22:32:49amiconnThe file "dep-apps" gets created.
22:32:51Bagderah, it fails for me too now
22:32:54Bagderfor player
22:33:55amiconnPossibly a hint: while most paths in that file are absolute, some are not?
22:35:57Bagderit is used within the respective dir, so it can use relative path names
22:36:05amiconn[IDC]Dragon: Really interesting finding: Initializing the card into SPI mode *works* *with my transfer routines* for external MMC, but not for internal flash :/
22:36:47amiconnMaybe I have to reset the flash beforehand...
22:38:11Bagder... but if I invoke make again it works
22:40:01amiconnHmm, same here. Strange
22:40:19***Alert Mode OFF
22:40:39amiconnAfter using "make clean" or building in a new dir, it fails. A second make call completes the build
22:40:55BagderI'm on it
22:41:27amiconnPlugins are missing for player too
22:51:59***Saving seen data "./dancer.seen"
23:03:28 Quit AciD (Connection reset by peer)
23:04:56[IDC]DragonZagor: I need a way of reproducing that error
23:05:24[IDC]Dragonhow small of an image can you generate?
23:05:48[IDC]Dragonmaybe we can transfer one, plus the next failing command
23:06:59Zagorhang on a bit, i'm debugging a little first
23:07:37[IDC]Dragonok, hang me ;-)
23:11:09Zagorheh, my 128MB disk image compressed to 305KB :) would you like it?
23:11:40[IDC]Dragonif it check ok and fails the next command, yes
23:12:22Zagorwell it's not quite that polished. i'll see what I can do...
23:13:24[IDC]Dragoncertainly I don't mind if you find the problem :-))
23:14:12 Join AciD [0] (
23:14:24Zagoroops... it was actually a bug in the test script :)
23:14:52[IDC]Dragonbug found then :-)
23:16:35Zagoryeah, now the test runs further but fails a bit later on
23:17:06[IDC]DragonI'm running out of smileys
23:17:46Zagorhehe. well be happy, the test script makes debugging a lot nicer than having to rely on angry users who just crashed their disk :)
23:18:31[IDC]Dragonyes, the script is great
23:18:49[IDC]Dragonwith those checkdisk's inbetween every step
23:19:16[IDC]Dragonbut that's for Linux only
23:21:11 Quit scott666_ (Read error: 110 (Connection timed out))
23:21:42 Join scott666_ [0] (
23:21:52Zagorlooks like it's rename that fails
23:22:31HenrikB_Goodnight and happy (bug) hunting
23:22:38Zagorgood night!
23:22:41[IDC]Dragonmaybe the root is full?
23:22:49[IDC]Dragonwhat does it say?
23:23:20Zagorit's not clear what is happening, there is no error message
23:23:35Zagorso i have to debug the test code also :)
23:23:36 Quit HenrikB_ ("Lämnar")
23:24:04[IDC]DragonI'm not doing debug output on full root dir, perhaps I should
23:25:08Zagori think we've changed file.c since the test script was last used...
23:25:53[IDC]Dragonfor me, it complained when renaming, pathnames not starting with /
23:26:03Zagoryeah, that's it. rename() used to not take a dir in the second argument. now it must be present.
23:26:34CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
23:26:34*[IDC]Dragon is relieved, sees hope
23:26:34Zagorall tests completed for images with 4 and 32 sectors/cluster. failed on 1 sector/cluster.
23:27:19Zagorah, simply too many clusters. not a driver issue.
23:27:30[IDC]Dragonhow soon does it fail?
23:27:56[IDC]Dragontoo many clusters would give FAT32
23:28:15Zagorright, so mkdosfs complains because I told it to make fat16
23:28:38[IDC]Dragonyou need to stay below 64 MB then
23:29:04[IDC]Dragonno, 32 even
23:29:14Zagorsame thing happens with 128 sectors/cluster, although in the opposite direction
23:29:31[IDC]Dragonis that allowed?
23:29:33Zagorso everything works in the 4, 8 and 32 sectors/cluster tests
23:30:00ZagorI don't think so, no. :)
23:30:07[IDC]Dragonvery good!
23:30:45[IDC]Dragonlet's check the legal corner cases
23:31:14amiconnZagor: For fat16, up to 64 sectors/cluster are allowed and understood by all OS'es that support fat. However, WinNT allows 128 sectors/cluster
23:32:47[IDC]Dragon2 GB, that is
23:33:03Zagoryes, but the fat32 spec says never to use more than 32KB/cluster
23:33:07amiconnWith WinNT it's 4 GB
23:33:19amiconnWe're talking FAT16?
23:33:33Zagorsorry, the fat spec
23:34:26amiconnThis feature of WinNT is called "NT large clusters". I think this does not need to be supported by rockbox, but it should not mount such partitions then
23:34:31Zagor"Values that cause a cluster size greater than 32K bytes do not work properly; do not try to define one."
23:34:52[IDC]DragonI'm facing the same spec :-)
23:34:58Zagorwe support them, i just didn't make an image that had them correctly
23:35:14 Join iriver_srn [0] (
23:35:31amiconnAh ok. Perhaps I should fire up a NT4 VM and create one?
23:35:56elinenbeis the Ondio flashable?
23:35:57Zagorno need, you can do it with mkdosfs. I tested the fat32 code on it.
23:36:02[IDC]DragonI'm forced to goto bed
23:36:08Zagor[IDC]Dragon: ok, good night
23:36:18[IDC]Dragonvery sorry to leave this interesting moments
23:36:39[IDC]Dragonc u, will read the logs tomorrow
23:36:47amiconnnite Jörg
23:36:50 Quit [IDC]Dragon ()
23:37:20Bagdermake fixes coming up in a few minutes
23:40:52Zagorbtw [idc]dragon: the additions to support fat16 are so small I think we should not #ifdef them. and that will allow us to make the changes even smaller, by combining the 16- and 32-bit code in fat_recalc_free, find_free_cluster, update_fat_entry and fat_read_entry.
23:41:03Bagderamiconn: now try my fixed version, just committed
23:41:38Zagorworks better
23:41:58Bagdernow I tried the Ondio build too ;-)
23:41:58amiconnBagder: Do I have to redo configure?
23:54:40 Quit iriver_srn ("Miranda IM! Smaller, Faster, Easier.")
23:55:28amiconnBagder: Seems to work correctly now for all targets I tried. However, I noticed some things:
23:55:35Zagorweird slashdot story: "Iriver H320 almost hits the market". Haven't people had it for weeks, at least?
23:55:51amiconnBagder: (1) target builds do use dep-xxx files now, while the simulator builds still use the .deps directory
23:56:10Bagderyes, I've left the sim build as-is for now
23:56:56amiconn(2) Creating/removing the dependencies (make/ make clean) for the target is way slower than for the sim? This was also the case with the old build system though
23:57:54BagderI can't explain that though...

Previous day | Next day