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 2004-08-23

00:03:52 Join LinusN [200] (
00:15:09 Quit _aLF ("Leaving")
01:00:26 Part LinusN
01:08:37 Join AciD` [0] (
01:22:05 Quit AciD (Read error: 110 (Connection timed out))
01:24:13 Quit Nibbler (Read error: 60 (Operation timed out))
01:26:31***Saving seen data "./dancer.seen"
02:23:00 Join gromit`` [0] (
02:34:12 Join Madman [0] (
03:01:23Madmananyone here?
03:04:21Madmanhey, do you know anything about pong on the jukebox recorder?
03:08:04 Quit Madman ()
03:26:35***Saving seen data "./dancer.seen"
03:42:11 Nick gromit`` is now known as gromit`TiRED (
03:48:23 Join hume22 [0] (~fgjkrj@
03:48:47hume22hello i have an archos jukebox 10 and i have the error dir buffer full
03:49:08hume22i never used rockbox before but i installed 2.2 and got the daily build of today
03:49:18hume22any help will be great
04:10:42 Join Dstar5 [0] (
04:35:10 Nick gromit`TiRED is now known as gromit`aw (
04:35:38 Join Madman [0] (
04:41:07 Quit AciD` (Remote closed the connection)
04:45:40 Join AciD [0] (
04:50:38Madmananyone know anything about the dir buffer full! error?
04:52:15Madmani just flashed my jukebox, it worked fine, but after i tried to install the latest daily build, it says, "dir buffer is full!"
04:58:39Madmananyone know?
05:02:14 Quit AciD (Read error: 60 (Operation timed out))
05:05:59Madmanit also wont allow me to install a new firmware, including all of the daily builds and the original archos firmware (1.7)
05:07:41Madmanit has v. 2.2 of rockbox in right now, because thats what came with the flash
05:13:18Madmaneither nobody is here or nobody knows
05:15:01Madmanits like v2.2 keeps on re-writing itself after i delete it...
05:23:10 Join hume23 [0] (~fgjkrj@
05:24:04hume22madman go into the options general, system then limits
05:24:23hume22and change the file limits and power off and on your player
05:24:55Madmandidnt work...
05:25:58Madmani think something else is wrong. i tried to delete the .ajz file and the .rockbox dir files, but they just rewrite themselves.
05:26:05Madmani did that from the computer
05:26:25Madmani cant even see any dir's in the jukebox
05:26:37hume22i had the same problem today
05:26:38***Saving seen data "./dancer.seen"
05:26:42hume22and that how i fixed it
05:27:01hume22hey do you know how to fix the laggint between changing tracks
05:27:11hume22it takes like 4 secs
05:27:18hume22before it was instant
05:27:47Madmando you have a daily build?
05:28:26hume22takes like 5 secs
05:28:29Madmanand its flashed? it shouldnt do that
05:28:42hume22whats flashed
05:28:51Madmanyour jukebox
05:28:52hume22first day i start using this firmware
05:29:09Madmanoh, well it should only take like 2 secs
05:29:31hume22the light starts blinking
05:29:41hume22like its searching hard
05:30:03hume22it didnt do that before i installed this firmware
05:30:35Madmanthat is weird, whenever i change the limit, not only does it reset when i turned it off, but i saw it jump from 0 to 50.
05:30:49Madmanthat means that the default is 0 when it boots up.
05:31:33hume22you have jukebox 10
05:32:00Madmanrecorder 20
05:32:51hume22oh i have jukebox 10
05:33:01hume22well didnt you delete some files_
05:33:06Madmanit should work about the same though
05:33:12hume22try reinstalling it again
05:33:16hume22and change the limit
05:33:34Madmanwhenever i delete something, it just rewrites itself
05:33:45Madmanim starting to think that its possesed
05:34:14hume22u think this rockbox is better than the original_
05:34:23Madmanmuch better
05:35:12hume22what does flashing do to your player
05:35:12Madmanor it was until it stopped working
05:36:22hume22u think if i flash it it will work
05:36:55Madmanit makes your jukebox run from ram so it goes a lot faster and boots faster
05:37:36Madmanif i were you, i wouldnt flash it right now.
05:37:38hume22hmm how long does it take_}
05:38:06Madmanwait untill i get mine working because i might have found a glitch or something
05:38:20Madmanits fast.
05:38:33hume22ok sounds good
05:38:50 Quit Ka_ (
05:38:50hume22could it damage it
05:39:14NJoinKa_ [0] (
05:39:46Madmanit could, but its highly unlikely
05:40:01Madmanthey just like to put out a disclaimer so that you dont sue them
05:40:10hume22ok cool
05:40:25hume22i wonder why they dont put rockbox on the damn thing from the start
05:40:53Madmandownload that video and see what flashing you rockbox will do
05:41:06 Quit Ka_ (
05:41:06 Quit dwihno (
05:41:06 Quit ze (
05:41:06 Quit mbr (
05:41:22NJoinKa_ [0] (
05:41:22NJoindwihno [0] (~dw@
05:41:22NJoinze [20] (
05:41:22NJoinmbr [0] (
05:41:24Madmanand the recorder in the video is the type that i have
05:42:22hume22it does load a lot faster
05:42:47Madmani dont know if your jukebox is compatible with it though
05:42:54hume22how long does it take to flash
05:43:05Madmana few seconds
05:44:24hume22it says it is compatible but i have to check with a sort of utility
05:47:00Madmanoh no...
05:47:37Madmansince i flashed it, the firmware is installed in the ram, so i cant delete it.
05:47:46Madmanor i dont know how to
05:50:22Madmanwow, it just started to work for no apparent reason
05:50:55hume22The first is a Recorder, FM Recorder or Recorder V2 model. Be sure you're using the correct package, they are different! The technology works for the Player models, too. Players can also be flashed, but Rockbox does not run cold-started on those, yet.
05:51:00hume22what those this mean
05:51:15hume22i mean does
05:52:30Madmancold started means by just pressing on... i thing
05:53:05hume22so it might not work
05:54:30Madmanit should
05:55:23Madmanits working for me now, im going to try to install one of the daily builds
05:55:27hume22you think i should flash it
05:56:38Madmanback up all of your stuff first
05:58:03hume22to my hard disk
05:58:12hume22you mean to my pc
05:58:18hume22or to the same player
05:58:40Madmanto the pc
05:58:50Madmanim not saying anything is going to happen
05:58:53Madmanbut just incase
05:59:01hume22ok lets see
05:59:23Madmanby the way, i got mine to work.
05:59:40Madmani held f1 and on to boot it to the original os
05:59:52Madmanthen i turned it off and booted to rockbox
06:00:19hume22im going for it
06:00:21Madmanthen i had to execute the ucl file to make the dailybuild load up when it boots
06:00:24hume22hope it works
06:00:41Madmanthat was for anyone who is going to read the log who had my problem
06:02:20Madmanmy jukebox is going like 10 times faster now
06:02:58Madmanif you get it to work, tell me
06:06:21Madmanhave you installed it yet?
06:10:18Madmanill be back
06:10:21 Quit Madman ()
07:26:39***Saving seen data "./dancer.seen"
07:37:35 Join oxygen77 [0] (
07:55:34 Quit hume23 (Read error: 110 (Connection timed out))
07:55:58 Quit hume22 (Read error: 110 (Connection timed out))
08:20:25 Join midk [0] (
08:21:29 Join Bagder [241] (
08:21:53midkhey bagder
08:25:38*Bagder hates coming to work only to find that someone took your "target" so now I can't test my code
08:26:30Bagderbtw, I bought a new rec20 from a UK guy last week
08:26:43midkoh, great
08:26:44Bagdernot new then, used of course
08:26:47midkwhen do you get it?
08:26:58midki doubt you could easily find a new one, they're so old
08:27:00BagderI hope I'll get it in a few days
08:27:38BagderI miss my rockbox
08:27:38midkgreat great
08:40:03 Join Madman [0] (
08:40:33 Quit Madman (Client Quit)
08:53:51 Join amiconn [0] (
09:13:22 Join Zagor [242] (
09:26:43***Saving seen data "./dancer.seen"
09:29:41 Join kurzhaarrocker [0] (
10:33:21 Part kurzhaarrocker
11:01:13 Nick midk is now known as midk|sleep (
11:01:15midk|sleepnite everyone
11:01:25Bagdersweet dreams
11:19:53 Join lImbus [0] (
11:26:47***Saving seen data "./dancer.seen"
11:39:48 Join mbrisou [0] (~mbrisou@
11:48:24*oxygen77 is away: chui pas l
11:48:57 Part mbrisou ("Leaving")
11:51:57 Join kurzhaarrocker [0] (
11:52:22kurzhaarrockerBagder do you at least have a backup of your music collection?
11:54:19 Quit Hes ("Changing server")
11:54:19 Join hessu [0] (
11:58:38 Join mroizo [0] (
12:00:25 Join Strath [0] (
12:00:59Strathhey kurzhaarrocker
12:01:31kurzhaarrockerHi :D
12:23:52 Quit kurzhaarrocker (
12:23:52 Quit Dstar5 (
12:23:52 Quit webmind (
12:24:16NJoinkurzhaarrocker [0] (
12:24:16NJoinDstar5 [0] (
12:24:16NJoinwebmind [0] (
12:24:31Bagderkurzhaarrocker: I do have my music backed up, so no worries there at least
12:24:47 Quit kurzhaarrocker (Read error: 104 (Connection reset by peer))
12:25:01 Join kurzhaarrocker [0] (
12:35:01 Join Nibbler [0] (
12:59:36 Join ze__ [20] (
12:59:36 Quit Nibbler (Read error: 54 (Connection reset by peer))
12:59:41 Join Bagder_ [241] (
13:00:32 Quit Bagder (
13:00:32 Quit ze (
13:00:32 Quit dwihno (
13:00:32 Quit Zagor (
13:00:32 Quit mbr (
13:00:32 Nick ze__ is now known as ze (
13:00:41 Nick Bagder_ is now known as Bagder (
13:01:01NJoinZagor [242] (
13:01:01NJoinmbr [0] (
13:01:54 Nick kurzhaarrocker is now known as kurzhaarrocker|m (
13:02:08 Part kurzhaarrocker|m
13:05:04NJoindwihno [0] (~dw@
13:09:36 Join AciD [0] (
13:26:48***Saving seen data "./dancer.seen"
13:27:26*oxygen77 is back (gone 01:39:01)
13:34:57 Quit oxygen77 ("Cho")
13:51:40 Quit mroizo ("Chatzilla 0.9.64a [Mozilla rv:1.7/20040614]")
14:17:49 Join kurzhaarrocker [0] (
14:18:04kurzhaarrockerIs it really desirable to have french text in the wiki?
14:18:50webmindnext to english or french only ?
14:18:52Strathdo you realize how many people who speak french there are working on archos stuff?
14:19:16kurzhaarrockerIt's a page in french only
14:19:26webmindthen it could use translation :)
14:19:26Strath(besides archos employees ;) )
14:19:52kurzhaarrockerTo me it seems as it might be a translation of the english battery faq
14:21:33kurzhaarrockerStrath: But should we mix the languages within one wiki?
14:22:11Strathup to you... i don't speak french, so it doesn't matter to me ;)
14:23:08kurzhaarrockerThe grades I received for my french lessons prove that I don't speak french either.
14:25:10webmindkurzhaarrocker, well demanding english only would not be nice to the french.. and the otherway around for the non-french
14:26:13kurzhaarrockerIs there a wiki-police? :)
14:30:21Bagderyou applying for the job? ;-)
14:30:28 Join oxygen77 [0] (
14:30:39lImbusI do speak french, amongst other languages, and it looks to me like JocelynDelalande ist currently translating the BatteryFAQ.
14:30:48*kurzhaarrocker hides from the wiki-police job
14:40:23 Join Nibbler [0] (
14:50:03amiconnHi all
14:50:25kurzhaarrockerHi amiconn. Any news from the recording-bug front?
14:50:54amiconnBad news for all core developers: It seems like there is some function in rockbox which trashes innocent memory.
14:51:14Bagdercan you repeat the problem?
14:52:19amiconnUnfortunately, this bug is not triggered often, but yesterday I observed the following odd behaviour:
14:53:31amiconn(1) I was listening to some music (shuffled dynamic playlist, created via on+play on the directory and selecting playlist->add)
14:54:08amiconn(2) As I went away for a short time, I only stopped playback, but didn't shut down the box
14:54:48amiconn(3) When I came back, I hit ON to resume, but didn't get the resume question. Instead, it told me about a corrupt playlist control file.
14:55:28amiconn(unfortunately my brain wasn't fast enough to realize I should save that file for analysis. I didn't have my computer with me)
14:57:01amiconn(4) The file browser was still in the dir where my files are located (via follow playlist = yes). When I hit "Left", it should have gone one dir level up (which it did) - but the cursor was not positioned on the dir I just left, but on some other random file!
14:57:38amiconn(5) Switching the box off and back on returned everything to normal (of course I had to create a new dynamic playlist)
14:58:23Bagderthis may explain why others have reported broken playlist control files
14:59:10amiconnThat all combined is why I believe there is something trashing other thread's memory - in my case it may have corrupted the playlist control as wel as the browser's dir buffer
14:59:35BagderI think you're right
15:00:17amiconnThe "only" thing is - we have to find out what it is
15:02:38Bagderyeps, and it is likely hard to repeat using the sim
15:03:47amiconnYesterday I got an idea how to check if there is something wrong with rombox, i.e. trying to write to the rom - using the user break interrupt feature of the sh1
15:07:14amiconnI'll look into it this evening - perhaps it may help to get the rombox recording issue sorted
15:16:01amiconnkurzhaarrocker: That recording bug issue is really strange. I'll have to do further test recordings. Maybe the MAS firmware itself has a bug...
15:17:02kurzhaarrockerI still am confused because I never encountered any audiable recording bug problems. But to be honest I usually record with a patched firmware from January/04.
15:26:51***Saving seen data "./dancer.seen"
15:33:04 Join [IDC]Dragon [0] (
15:34:56amiconnhi Jrg
15:48:04 Quit [IDC]Dragon ("CGI:IRC (Ping timeout)")
16:36:25 Quit Strath ("Client closed")
16:46:00 Join Lee_ [0] (
16:46:13Lee_hey Dstar5 is still here lol
16:46:21Lee_i was using it on a cell phone
16:46:55kurzhaarrocker? A jukebox on a phone?
16:47:05 Quit Dstar5 (Nick collision from services.)
16:47:15Lee_no, i was using IRC on a phone
16:47:41 Nick Lee_ is now known as dstar5 (
16:47:45*kurzhaarrocker shakes his head
16:48:01dstar5i used cgiirc lol
16:49:08dstar5it didn't work well on the phone
16:49:14dstar5it was weird
16:49:27dstar5it had names of pages like fmain, fuserlist, and fmsg
16:49:39kurzhaarrockerThen use a wired phone! MUAHAHAH!
16:55:35 Join [IDC]Dragon [0] (
16:56:13 Quit [IDC]Dragon (Client Quit)
16:57:01 Join [IDC]Dragon [0] (
16:57:21dstar5hey [IDC]Dragon
16:57:36dstar5is that board working yet? :)
16:59:21[IDC]Dragonhi dstar5
16:59:33[IDC]DragonI still haven't tried
16:59:41[IDC]Dragonlots of other worries:
16:59:55[IDC]Dragona HD in my PC died
17:00:11[IDC]Dragona part of a tooth of mine broke off
17:00:35dstar5thats really bad
17:00:37[IDC]Dragonthe HD is replaced now, but not my tooth
17:01:02dstar5i lost everythig on my drive about 5 times within a week
17:01:14dstar5turned out to be a bad IDE cable
17:01:37[IDC]Dragonafter the first one, it wasn't much to lose, I guess :-(
17:02:18[IDC]DragonI was lucky, recovered most of the data
17:02:26 Part kurzhaarrocker
17:02:34 Part oxygen77 ("Cho")
17:02:38dstar5thats good :)
17:02:48dstar5what went wrong with it?
17:02:49[IDC]Dragonand bought 2 new disks, I have a RAID now
17:02:57dstar5nice :)
17:03:03[IDC]Dragonthe one just died of age, I guess
17:03:37[IDC]Dragonit still works, but has plenty of errors
17:03:38dstar5how much storage on the disks?
17:03:49[IDC]Dragon160 GB each
17:04:01[IDC]Dragon(the broken one, and the 2 new)
17:04:12dstar5i have 1 160gb drive
17:04:28dstar5it was very cheap, got it for $90
17:04:35[IDC]Dragonthose are the most economical
17:04:49[IDC]Dragongot mine for %u20AC84
17:05:00[IDC]DragonEUR 84, that is
17:05:19[IDC]Dragonabout the same, I guess
17:05:27dstar5ahh, that would be about what i got mine for
17:05:38dstar5euro/dollar are quite close
17:06:00[IDC]DragonI wanted a larger one, but they are overproportionally expensive
17:06:07[IDC]Dragonlike 300 GB
17:06:26[IDC]Dragonthen I remembered the RAID feature of my mainboard
17:06:55[IDC]Dragonworks pretty well now, with that disk pair, blazingly fast
17:07:03dstar5 1 EUR = 1.22005 USD
17:07:23[IDC]Dragonand times two, for the stripe
17:07:49[IDC]DragonSamsung makes very quiet disks
17:07:51dstar5now what about 2 radtors ;) 10,000 rpm each
17:08:18dstar5my drive is western digital
17:10:05[IDC]Dragonsounds noisy?
17:10:47 Quit AciD (Connection timed out)
17:11:03dstar5no, very quite, except when the drive is powering on, it has a loud spinup
17:11:22dstar5much quiter than my dads 40gb drive in the same machine'
17:11:38 Join mattzz [0] (
17:12:19[IDC]Dragonhi mattzz!
17:12:31[IDC]Dragonhome again?
17:13:02mattzzhi joerg - yup
17:13:21mattzzwasp experience...
17:13:49[IDC]Dragonwasp? you got stinged?
17:14:36mattzzwe were on that company trip last weekend. Thousands of wasps. Unfortunately the only one who got stinged was my daughter. But she is fine again.
17:15:36dstar5i am going to go write some c0de
17:15:38dstar5see you
17:15:55mattzzproject camp is finished as well - so we are going back to normal - kind of...
17:16:04 Quit dstar5 ("umount /dev/brain")
17:16:37[IDC]Dragonmattzz: didn't know you've got family...
17:16:58mattzzthree fine women ;-)
17:17:18[IDC]Dragonthree! impressive.
17:17:34mattzzwell, two girls, one woman to be precise
17:17:44[IDC]DragonI suspected that.
17:18:26mattzzIs your phase of bad luck over?
17:19:35[IDC]Dragonno, I need to get this code to work.... ;-)
17:26:52***Saving seen data "./dancer.seen"
17:27:56 Join slazZ [0] (
17:29:30 Quit slazZ (Client Quit)
17:29:42 Part Bagder
17:41:30 Join edx [0] (
17:43:08 Part Zagor
17:44:39 Quit mattzz ("CGI:IRC")
17:47:23 Join mecraw [0] (~lmarlow@
17:54:59 Part lImbus
18:14:30[IDC]Dragonamiconn: r u there?
18:14:50amiconn[IDC]Dragon: Yep; just came home :)
18:17:46[IDC]Dragonsaw your idea to use the user break controller to trap bad writes, good idea!
18:18:03[IDC]Dragon(poor man's MMU)
18:19:36amiconnyup. Unfortunately the user break controller does only provide control over one area, which must be sized 2^n
18:20:59amiconnOtherwise it could be used to track down other bugs, e.g. the memory trashing I suspect to be the cause for the odd behaviour I expereinced yesterday
18:22:28[IDC]Dragonthe ISR can filter that out, but severly degrades performance
18:22:44amiconnI'm actually thinking of a way to do this with the user break controller (by capturing writes to the area that contains the dir browsers buffer, and then filtering in the isr
18:23:21amiconnAll that means I have to figure out how to implement such an ISR
18:23:39[IDC]DragonI once looked into the UBC, to allow gdb to step through ROM
18:23:45 Join _aLF [0] (
18:23:56[IDC]Dragonbut it turned out impractical
18:24:17amiconn[IDC]Dragon: Iirc the sh1 features a single step mode
18:24:43[IDC]DragonI'm not aware of that
18:28:00[IDC]Dragonwhile you do such anvanced stuff, I'll rotate the font, if you don't mind
18:29:46amiconnOk, I remembered wrong. But it should be possible to mimic a single step mode by utilizing the UBC
18:31:05[IDC]Dragoniirc, branches were an unsolvable problem
18:31:23[IDC]Dragonrendering it useless for a debugger
18:31:30[IDC]Dragonbut I don't remember why
18:34:47amiconnProbably because if a user break is triggered by a delay slot instruction, the break interrupt is delayed until after the instruction
18:35:06amiconn(Just reading the notes)
18:45:47amiconn[IDC]Dragon: For simply capturing writes to the ROM, it looks like I only have to initialize the UBC; the exception is already handled in system.c. Do I miss something?
18:45:48[IDC]Dragonamiconn: please remind me how we wanted to do font rotation
18:46:09[IDC]Dragon(UBC) sounds OK, yes
18:46:30[IDC]Dragon(font) the glyphs are 16 bit words
18:46:56[IDC]Dragonso using 8 bit (rotated or not) would break compatibility anyway?
18:48:11amiconnYes and no. The font glyph data is defined as being of bitmap_t type, which is 16bit.
18:48:55amiconnIn order to not break compatibility, the pointer itself has to stay that way, that means, each glyph must be short aligned.
18:49:41amiconnThis doesn't prevent us to use byte alignment within the glyph, because the actual drawing is handled by lcd_bitmap()
18:50:12[IDC]Dragonand all chars are 6 byte wide, which happens to be even
18:50:46[IDC]Dragongot it
18:51:36amiconnSo basically, convbdf has to "learn" an additional output format ("rockbox raw"), with the glyphs already rotated and byte packed
18:52:20amiconnThen the makefile has to use that to build font.[ch], and the rotate_font_bits() call taken out from the font init
18:52:38[IDC]Dragonor just change the C output format, we won't use both
18:52:44amiconnOf course the function itself has to stay, because it is needed for loaded fonts
18:52:52 Join AciD [0] (
18:54:13amiconn[IDC]Dragon: Further ideas for font handling:
18:54:36amiconn(1) change the bitmap offsets into an array of short. This would require 2 different font structs, as discussed a while ago
18:56:06amiconn(2) Do the font rotating (and perhaps changing to short offsets) on the fly for loaded fonts too. This would either allow larger fonts with the same font buffer size, or allow us to decrease the fonts buffer size
18:58:17amiconnFont buffer size is currently 9000 bytes
18:59:06 Join dstar5 [0] (
18:59:27dstar5yay checkers has a new life in it :)
18:59:33dstar5i just figured out how to finsih it
19:03:27 Join zeekoe_ [0] (
19:09:22 Nick zeekoe_ is now known as zeekoe (
19:14:04amiconn[IDC]Dragon: Capturing ROM writes via UBC works - I can't flash an .ucl with an UserBreak-enabled ajbrec.ajz. It complains about the plugin wanting to write to the ROM
19:17:31[IDC]Dragonwhow, that was quick!
19:17:54amiconnIt's only 3 lines in system.c :)
19:18:00[IDC]Dragonwhat do you do in your ISR?
19:18:26amiconnI didn't write a special ISR yet - only capturing the write via the default ISR
19:18:50amiconnBAR = 0x02000000; /* start address of ROM */
19:18:50amiconnBAMR = 0x003FFFFF; /* match any ROM address */
19:18:50amiconnBBR = 0x00F8; /* Break on any write access, CPU & DMA, IF & data, any size */
19:19:47[IDC]Dragonso we won't know who wrote where
19:20:07amiconnIf I want to use this for tracking other wrong writes, of course I have to write my own ISR
19:20:10[IDC]Dragonbut maybe we don't have to, if everybody behaves
19:20:45[IDC]Dragonbtw, you used an old flash plugin? The current one shoudn't run if ROM-based
19:20:50amiconnWe will know who writes there, because the default ISR outputs the source address
19:21:04 Join etelcontar [0] (
19:21:06[IDC]Dragonah, OK
19:21:31[IDC]Dragonso, go and record (or whatever)
19:21:35amiconnNo, I did run the flash plugin from a rolo'ed ajbrec.ajz (which had the rom-write detection enabled)
19:22:01etelcontarhey guys, sorry to interrupt, but I wonder if you can help me out...
19:22:15[IDC]Dragona flash protection ;-)
19:22:48amiconn[IDC]Dragon: This was just a first test, now I'll flash an .ucl with that, and try as much functions as I can think of
19:23:05amiconnetelcontar: ?
19:23:39 Quit etelcontar (Client Quit)
19:23:52 Join etelcontar [0] (
19:23:53[IDC]Dragonyou have to ask, else I'm afraid we can't help
19:24:20etelcontarhey.. my archos just locked up .. I upgraded to 2.2 the other day and now it is stuck on "error".. any thoughts?
19:24:43[IDC]Dragonwhat error? when?
19:24:59zeekoeetelcontar: you "upgraded" to 2.2? :-P
19:25:06[IDC]Dragonflashed or regulear disk loading?
19:25:07etelcontarI went to play a tune.. now the LCD diplay just says "Error"
19:25:20etelcontarregular disk loading
19:25:53[IDC]Dragondoes it happen again, after power cycling?
19:26:02etelcontarhow do you power cycle?
19:26:14[IDC]Dragonshut it off, then on again
19:26:36etelcontarThere does not seem to be an "off" switch on this thing
19:26:51etelcontarStudio 20
19:26:56***Saving seen data "./dancer.seen"
19:27:19zeekoelol... how do you normally turn it off?
19:27:41etelcontarI set the idle timeout to 1 minute and unplug the power
19:28:39zeekoehey.. you're right... i can also see just on & menu on studio 20
19:28:47[IDC]Dragonhold stop for a while
19:29:02etelcontarhow long? I did that already
19:29:43etelcontarok.. that worked! (took about 30 seconds)
19:30:11[IDC]Dragonif it's not crashed, this should be shorter
19:30:25[IDC]Dragon(it's the normal power off, afaik)
19:31:05etelcontarany song I try to play gives me the "Error"
19:32:21 Quit AciD (Read error: 104 (Connection reset by peer))
19:33:47dstar5could it be there is no .rockbox?
19:33:57dstar5i am not sure how that errors look on the player
19:36:19[IDC]Dragonetelcontar: you may want to try the Windows installation on the daily build page
19:36:41zeekoeetelcontar: what version did you use before
19:36:44etelcontardo I need a build environment for that?
19:37:05etelcontarI had 1.4 before (I think)
19:37:18zeekoeok :)
19:37:25zeekoeyou'll be amazed :)
19:37:26[IDC]Dragonit's an install exe that doas everything for you, the most foolproof we have to offer
19:37:51etelcontarok, I will have to try that wehn I get home .. I have no USB cable here at the office
19:38:13 Join AciD [0] (
19:38:20[IDC]Dragonactually, where is that installer?
19:38:32[IDC]DragonI thought we have it in the daily build
19:38:54dstar5we do
19:39:00[IDC]Dragonah, rightmost column
19:39:48 Join de [0] (
19:39:57dstar5i think that should just be made bold
19:40:00[IDC]Dragonetelcontar: download the upper right link from
19:40:09dstar5make it standout for people that are new
19:40:26[IDC]Dragon(most recent Windows installer)
19:41:02 Quit de (Client Quit)
19:41:05zeekoemaybe put it on the download page
19:41:05dstar5or just make it seperate from the table, and make it above it
19:44:18amiconn[IDC]Dragon: Good news: rockbox doesn't seem to try writing the rom (in all cases I could test so far)
19:44:48 Quit dstar5 ("umount /dev/brain")
19:45:09zeekoeamiconn: is that good news?
19:45:23zeekoeamiconn: now you still know the answer to the rec problems...
19:46:02etelcontarwell, thanks for the info.. I will try the Windows installer tonight. In the mean time.. is there any way I can blow away the archos.mod file and return to the Archos until then?
19:46:23etelcontarwithout a USB cable that is
19:47:22amiconnzeekoe: No, I still don't know what causes it, but I now know what does not cause it..
19:48:15amiconnI think I'll have to rely on Linus doing some logic analyzing in order to understand what's going on there
19:49:02amiconnI tried changing the hard mas reset into a soft reset. Unfortunately this didn't help anything
19:50:07 Quit etelcontar ("CGI:IRC")
19:51:39amiconnI just triggered the corrupt playlist control file bug again
19:52:14 Quit hessu (Read error: 104 (Connection reset by peer))
19:53:39 Join uski [0] (~uski@2001:7a8:3bb9:0:defa:ced:dead:beef)
19:55:43 Nick midk|sleep is now known as midk (
19:58:21amiconnIt's actually quite easy to trigger the bug:
19:58:52amiconn(1) Set the cursor over a directory containing music. Do not enter it.
19:59:15amiconn(2) Press On+Play and select playlist->add. The music will start playing
19:59:38amiconn(3) Hit OFF to stop, the hit ON again -> corrupt playlist control file
20:03:08amiconnSo it seems that the playlist control file handling is wrong if the playlist is generated that way. It doesn't have to do with trashed memory
20:03:11amiconn[IDC]Dragon: ?
20:04:57 Join maikeul [0] (
20:10:46 Join pfavr [0] (
20:14:01 Join Strath [0] (
20:14:51Straththe new main page design for just went live
20:17:18midkstop advertising. ;)
20:17:48Strathi'm not
20:18:27Strathi'm just letting people know so they can marvel at the pretty new main page
20:18:53midkresize those logos in photoshop.
20:19:14Strathwhats wrong with em?
20:19:55midkat least in mozilla.
20:20:47Strathwhat size would you like em?
20:21:29midkthe same size...
20:21:36midkdah, they're just pixelated
20:22:04Strathi supose i could blur em a bit
20:22:17midkwhy are they pixelated?
20:22:32midktake the large rockbox logo and downsize it, not the small one and increase its size.
20:22:32Strathno idea...
20:22:51Strathrockbox400.png, resized to 200
20:23:02 Quit gromit`aw (Read error: 110 (Connection timed out))
20:23:07Straththats what it was/is
20:23:29Strath(redid again, looks the same to me)
20:23:40midklooks fine?
20:23:57Strathi didn't say that ;)
20:24:15midkhow are you resizing it?
20:29:24Strathi think i know....
20:29:33Strathi had left it as indexed colour
20:30:23 Join lImbus [0] (
20:31:13[IDC]Dragonamiconn: sorry for not paying attention
20:31:42[IDC]DragonI think my rotated font works, that distracted me
20:31:47midkstath: ho.. :(
20:32:33midksee the rockbox/avos logos at the left
20:34:54Strathwow that looks bad
20:35:04Strathbut thats just your libs
20:35:34Strathlooks fine in Firefox/win32
20:35:47midkfirefox looks the same
20:36:11Strathyou want me to upload a screenshot?
20:36:45 Join jakesir [0] (
20:37:07[IDC]Dragonhi Jake
20:38:51jakesirwhat's the latest to my problem?
20:40:28Strathmidk, uploading to temp, i'm going to sleep
20:41:00 Nick Strath is now known as StrathAFK (
20:42:10[IDC]Dragonjakesir: you have read my emails?
20:42:33[IDC]Dragonsince then, nothing changed, I had to give up
20:42:59[IDC]Dragonplease tell me what to flash on it before returning it
20:43:09jakesirummm.... 2.2
20:43:18jakesirand i'm happy to live with that
20:43:42jakesirdoes it play alright under 2.2?
20:44:01[IDC]Dragonyou shoudn't change that, since your box has this disk read/write problem
20:44:29jakesiri'll stick to 2.2 forever
20:44:45[IDC]DragonI think it plays, 1 bad bit every 8KB is probably not very audible
20:45:00[IDC]Dragonalthough I wonder
20:45:07jakesirwell, i mainly use it for audio books
20:45:13jakesirso, it won't matter too much
20:45:17[IDC]Dragonhave to test, I think I never played mp3
20:45:53jakesirdid you see my HD content of audiobooks?
20:46:06[IDC]Dragonand it has this severe starting problem, at least for me
20:46:44[IDC]Dragonno, I'm too disceet to look at your disk
20:46:44jakesiri have that same starting problem when the box is openned. it goes away when the casing is closed
20:47:40[IDC]Dragonseriously, I only put Rockbox on and off
20:48:09midki trust idc
20:48:41jakesiryou can delete the whole HD, i don't care
20:48:52[IDC]Dragonno, why should I?
20:49:29jakesiri meant that i have back-ups??
20:50:51jakesirisn't 2.3 due soon?? it's been 5 month...
20:51:19[IDC]DragonI could as well put a daily build on it
20:51:29jakesirthat's fine
20:51:48 Quit pfavr ("ChatZilla 0.9.61 [Mozilla rv:1.7.2/20040810]")
20:53:33 Join jakesirs [0] (
21:00:06midkallchan say brb
21:00:15 Quit midk (Remote closed the connection)
21:06:26amiconn[IDC]Dragon: The playlist control file bug isn't reproducable on the sim :(
21:06:46[IDC]DragonI would have been surprised
21:07:13[IDC]Dragondid I say the rotated font is working?
21:07:34 Join midk [0] (
21:07:35amiconnSo it looks like there might be a bug in either the file handling code (bad) or the fat driver (even worse) :((
21:07:52[IDC]Dragonon the first shot, I was not prepared for that, suspected some old code to be running
21:08:11amiconn[IDC]Dragon: (font rotation) nice :)
21:08:50[IDC]DragonI changed the algorithm to something wrong, to find out if it has an effect, but it did
21:08:59 Quit midk (Read error: 104 (Connection reset by peer))
21:09:05 Join midk [0] (
21:11:15 Quit jakesir (Read error: 110 (Connection timed out))
21:11:31[IDC]Dragonsorry for uncomfotable playlist control file debugging
21:11:49[IDC]Dragondo you really think it's the file system?
21:11:59[IDC]Dragonnot some astray pointer?
21:13:28amiconnWhat I did understand so far: (1) If a playlist is generated the way I described (via On+Play) on the dir), it is much longer than otherwise - every single file is listed on a separate line
21:14:15amiconn(2) The corrupt files just end prematurely, but not always at the same position (and the position is not sector aligned)
21:17:05[IDC]DragonI have to leave now.
21:17:23[IDC]Dragoncu guys
21:17:30 Part [IDC]Dragon
21:22:38 Join zeekoe_ [0] (
21:27:00***Saving seen data "./dancer.seen"
21:33:54 Join zeekoe2 [0] (
21:38:01 Nick midk is now known as midk|shower (
21:40:40 Quit zeekoe (Read error: 110 (Connection timed out))
21:48:04 Join dstar5 [0] (
21:51:56 Quit zeekoe_ (Read error: 110 (Connection timed out))
21:56:44 Nick midk|shower is now known as midk (
22:03:48 Quit midk (Read error: 104 (Connection reset by peer))
22:03:57 Join midkay [0] (
22:09:03 Join Bagder [0] (
22:09:25amiconnhi Bagder
22:09:48 Quit midkay (Read error: 104 (Connection reset by peer))
22:10:06amiconnBagder: di you read about my findings about the corrupt playlist control file bug?
22:10:40*Bagder checks
22:12:15zeekoe2oh... if you're talking about bugs anyway, when you press play, then on, a bit later off, the resume point is set to the point where you pressed on
22:12:23zeekoe2is that supposed to happen?
22:12:44dstar5i found a bug in flash
22:12:50Bagderthe resume point is set all the time
22:13:00zeekoe2Bagder: it is???
22:13:02dstar5if you are charging using archos firmware, and plug in usb, you get ata error -31
22:13:11 Quit uski ("brb")
22:13:12dstar5same if you press on when archos charging
22:13:14zeekoe2Bagder: but now when for example browsing files
22:13:17amiconndstar5: Yes, we know
22:13:46amiconndstar5: I found that last week, reported it to [IDC]Dragon, and he said Linus pointed it to him 2 weeks ago...
22:14:19amiconndstar5: It has to do with the cold start detection not being perfect...
22:14:58dstar5time for a new nick
22:15:02 Nick dstar5 is now known as bagawk (
22:15:44 Join midk [0] (
22:15:46Bagderamiconn: maybe you should talk to hardeep about that playlist control bug, as he's the author of that
22:16:46amiconnhardeep has been offline for almost 3 weeks now...
22:21:27 Quit jakesirs ("I-n-v-i-s-i-o-n 2.0 Build 3515")
22:26:44 Quit midk ("just STOP it arspy")
22:27:58 Join midk [0] (
22:30:28 Quit midk (Client Quit)
22:30:36amiconnBagder: It found out some more: If I create a playlist that way, DON'T stop the playback but plug in USB and look into the .playlist_control, it looks alright
22:31:14amiconnAfter unplugging USB and re-plugging it, it is chopped the same way as it is when I hit OFF to stop
22:31:21 Join midk [0] (
22:35:14 Join LinusN [200] (
22:36:49amiconnhi LinusN
22:38:01 Quit midk ("just STOP it arspy")
22:39:23 Join midk [0] (
22:42:44amiconnLinusN: Read about the .playlist_control woes?
22:43:51bagawkhey LinusN
22:44:04bagawkthis is my new nick (i was dstar5)
22:46:09LinusNamiconn: yes
22:46:52LinusNbagawk: hi
22:47:51LinusNi have suspected file system problems for a long time
22:49:34amiconnLinusN: It seems rockbox likes to annoy me - while I try to hunt down one bug (or just want to use it), it presents me with another one :(
22:49:59LinusNwelcome to the club...
22:50:44LinusNi have also suspected that the recording bug is a file system problem as well
22:52:07LinusNso, how shuld we implement file/dir move in the UI?
22:53:57Bagderprobably cut and paste style
22:54:01Bagderwith on+play
22:55:00LinusNmy thought as well
22:55:29 Join Zane [0] (
22:55:58Zanewho here knows where to get a good chat server?
22:56:38amiconnLinusN: If you decide to do some logic analyzing - could you try to capture the recording init while running from ROM?
22:56:59amiconnLinusN: Changing the hard reset into a soft reset didn't help...
22:57:27amiconnThe reset itself worked, but trying to record still hangs at 30 bytes
22:58:17 Quit bagawk (Read error: 104 (Connection reset by peer))
22:58:35 Quit Zane (Client Quit)
22:59:37LinusNamiconn: so the recording bug is still worse in ROM?
22:59:58amiconnIt depends about which bug you are talking...
23:00:02 Quit midk (Remote closed the connection)
23:00:14LinusNah, so the 30-byte bug is only in rom?
23:00:45amiconnThe bug which trashes either individual frames or the whole remainder of the file seems to be independent from where rockbox is running
23:01:29amiconnHowever, the bug where recording hangs after 30 bytes is very prominent when running from rom, and occurs only rarely when running from ram
23:01:33LinusNis it any different if you have prerecording enabled?
23:03:09amiconn(see my description within the Wiki article)
23:03:17LinusNyou should try to increase the sleep(20) in init_recording
23:03:21amiconnHmm, I didn't try prerecording yet
23:03:29 Join bagawk [0] (
23:04:21 Join midk [0] (
23:04:33amiconnYes, increasing the sleep() time within mas_reset() lowers the probability of triggering the bug, but doesn't get rid of it
23:05:05LinusNhow long have you set it to?
23:06:11amiconn200 ms gave me 60% failure rate, 500 ms gave 20% (this is when running from ROM)
23:06:15LinusNamiconn: i'm not talking about the delay in mas_reset()?
23:06:38LinusNi'm talking about the sleep(20) in init_recording()
23:07:12bagawkamiconn: leaving the mas_reset() out was a bad thing?
23:07:20LinusNwhere it waits for a frame to be encoded
23:07:25bagawk(it has been working for me fine)
23:08:07LinusNthe mas reset should not be necessary
23:08:13amiconnLinusN: That one I didn't touch (yet)
23:08:19LinusNamiconn: try it
23:09:16amiconnIf we leave out the mas_reset() we have to make sure all parameters are reset to default. That's why I tried the soft reset (via the reset bits in the control register)
23:09:47amiconnI still wonder why this hang does (almost) only happen when running from ROM...
23:11:29LinusN1) timing
23:11:41LinusN2) write-to-rom
23:12:11amiconnLinusN: There is no write-to-rom, I checked it today (with the help of the user break controller)
23:12:34LinusNso it's probably a timing problem
23:13:24amiconnAnd the check itself worked perfectly - I stumbled upon it myself when trying to flash an .ucl via an ajbrec.ajz containing that "rom protection"
23:13:57amiconn-> Exception I0C:UserBrk at 0x091F8082
23:18:22*amiconn tries a recording init with sleep(50) - compiling...
23:18:27 Nick zeekoe2 is now known as zeekoe (
23:23:01*amiconn flashing...
23:23:49*amiconn testing...
23:24:47*LinusN is holding his breath
23:27:02***Saving seen data "./dancer.seen"
23:28:13amiconnError rate is almost the same - around 65% (20 tries)
23:28:23 Join [IDC]Dragon [0] (
23:28:50[IDC]DragonLinusN almost suffocated...
23:29:11[IDC]Dragonsorry for log-peeking
23:29:23LinusNthe 30-byte bug is most likely related to the first-frame "trick"
23:29:40LinusNi have seen it lots of times when developing the init code
23:30:01amiconnYeah, this took a bit longer because I stumbled once more about my rom-write catch (my flashed .ucl did still do that, so I had to boot with F1+On)
23:30:29 Quit bagawk ("umount /dev/brain")
23:31:14[IDC]DragonLinusN: the font file format we use is something "standard"?
23:31:55[IDC]DragonRockbox proprietary?
23:32:01amiconnLinusN: I wonder why that first frame trick is necessary in the first place...
23:32:13LinusNamiconn: what happens if you set the delay to something really long, like half a second?
23:32:19LinusNamiconn: read the comment
23:32:31amiconnsleep(50) _is_ half a second...
23:32:43[IDC]Dragonbecause I wonder if we should/could make it more efficient
23:32:43*LinusN is silly
23:32:52midkdhahahah what a joker.
23:34:16LinusNamiconn: the irq handler is edge triggered, on the falling edge of EOD
23:34:58[IDC]Dragonwell, I'm off for sleep
23:35:07amiconnnite [IDC]Dragon
23:35:08 Quit [IDC]Dragon ("CGI:IRC (EOF)")
23:35:08LinusNthe trick records some data, and the drain_dma_buffer() will clear the dma buffer and raise EOD
23:37:03 Nick midk is now known as midk|gone (
23:40:01amiconnLinusN: Yes, I read this. But I still don't get 2 things: (1) why the mas doesn't set eod high if it is started in monitoring mode directly, (2) why it shouldn't be sufficient to force it high by reading _without_ recording anything first
23:40:57LinusN1) This is why i call it "the silly MAS"
23:41:13amiconn(Otherwise I think I shouldn't hurt to do this, and obviously extending that initial recording period doesn't change anything)
23:41:33LinusN2) clearing an empty dma buffer has no effect
23:42:04 Quit AciD (Connection timed out)
23:42:32 Quit Bagder ("Leaving")
23:43:49amiconnOn a related topic: I tried a 10 hour recording with m/s stereo switched off, and did not get a corrupt file. I'll have to do more tests (with a well-defined source, i.e. CD) to find out if it is really related to the m/s stereo setting
23:44:39amiconnMaybe the mas firmware has another bug (why else would archos switch off m/s when it is on by default)
23:45:03LinusNcould be
23:45:19 Join AciD [0] (
23:47:16amiconnLinusN: Do you have logic analyzer captures of the archos firmware initializing the mas recording mode?
23:47:51LinusNah, no
23:48:45LinusNbtw, are we using B2 versions of the MAS?
23:50:05amiconnYeps. Most likely all recorder have that version, because if I understand the specs correctly, this is the only version capable of recording
23:50:27LinusNand the black magic is performed?
23:55:38amiconnThe following things would be interesting to see as logic analyzer diagrams, containing EOD, PR, /RTW and the i2c communication:
23:55:53amiconn(1) recording init done by archos firmware
23:56:02amiconn(2) recording init done by rockbox
23:56:12amiconn(3) recording init done by rockbox running from rom
23:57:33amiconn(if such complex analysis would be at all possible; and I forgot the mas reset pin)
23:58:55amiconnBtw: the i2c communication would be interesting by itself

Previous day | Next day