Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

#rockbox log for 2005-01-16

00:00:50jypbtw... Is there a "watchdog" on the already supported rockbox targets ?
00:01:03jyp(any of them)
00:01:22Zagornot that i know of
00:01:37amiconnWhat do you mean? The SH1 cpu features a watchdog timer, but it's not used by rockbox.
00:02:13jypMaybe I can diable it :P
00:02:23amiconnHowever, it is used by the archos firmware to avoid a corner-case possibility for lockup
00:02:49jypIt didn't occur to me that it was possible to disable... I'm checking
00:03:41amiconnOn the SH, it is disabled by default
00:04:12jypAs it is I run my code via a hook in the archos firmware
00:04:20jypso I inherit all their settings
00:04:52jypmh; datasheet says watchdog is enabled by default..
00:05:01jypSo I guess it can be disabled
00:05:13jyp... Another related question ...
00:05:43jypNow I have to run with DC plugged in; or else the device shuts itself off
00:06:07jypDo you think this can be related with the watchdog ?
00:06:19jypSubquestion, in what way ?
00:06:40amiconnMaybe there is a port pin that you have to set for holding the power.
00:07:23amiconnThis is needed on some of the SH-based archos units as well (Recorder fm/v2, Ondio)
00:07:43jypActually I want to keep this behaviour until I have something more stable...
00:08:06jypIt allows me to reset easily :P
00:08:15***Saving seen data "./dancer.seen"
00:08:34jypI just wanted to make sure disable watchdog would keep it this way
00:08:35amiconnDoes the Gmini have hardware poweroff?
00:08:44jypI think not
00:08:56jypOr maybe ...
00:09:15amiconnThe mentioned SH-based boxes have that.
00:09:57jypIs the hw. poweroff accessible by a lambda user ?
00:10:24amiconnIf the software doesn't release the port pin to switch off power when the Off button is pressed (e.g. crashed), you can still poweroff by holding the Off button long enough (>10 sec)
00:10:46jypOk, thanks for the info
00:12:47jypStrangely enough, the watchdog is disabled by archos firmware
00:12:58jyp(if I trust the emulator traces)
00:17:28amiconnBagder, Zagor: One of the video.rock bugs I mentioned is cosmetic, and it's related to the scrollbar() function:
00:18:01 Quit silencer_ (Read error: 110 (Connection timed out))
00:18:51amiconnIf scrollbar() is called with very large parameters, there is an internal overflow which causes the displayed bar to be totally wrong
00:19:14amiconnI think this should be fixed within scrollbar() itself. What do you think?
00:19:28Zagorhow big are the values?
00:20:19amiconnIt is called with max_shown= <file position> and items=<file size>, in bytes
00:20:46amiconnThe width varies, but with my font it is around 80 pixels
00:21:23amiconnSo if the sizes go above (2^31-1)/80 bytes (around 26 MB), the overflow occurs,
00:21:50amiconnbecause of (width-2) * (max-min)
00:22:53Zagori agree this should be fixed in screollbar()
00:23:11amiconnI thought of something like:
00:23:30amiconnwhile (item > MAX_INT/width) {
00:24:05amiconnand likewise for the vertical variant order not to loose precision for low values
00:25:22Zagorhmm, can't we solve it without a loop?
00:26:09amiconnYes, but this would require more than 1 division, which would probably be more costly (3 divisions in total)
00:26:43amiconn..even 4:
00:28:46amiconnint temp = MAX_INT/width;
00:28:51jyp(watchdog successfully disabled)
00:28:58amiconnif (item > temp) {
00:29:21amiconntemp = item / temp + 1;
00:29:35amiconnitem /= temp;
00:29:40amiconnmin /= temp;
00:29:46amiconnmax /= temp;
00:30:54amiconnThe shifts within the loop are certainly faster than all those divisions.
00:35:28amiconnThere will be a maximum of 7 loop iterations (equalling /= 128) on the archos units, since the width can't be higher than that. It may be 8 on the iriver.
00:37:47 Join dethbrg [0] (
00:41:09dethbrgcrap, I /j the channel thinking someone could help me locate music like "ap collective" or "this function is void" I'll give it a few minutes before I /part
00:41:38dethbrgWith respect ^^^^^ to all working on the Archos Jukebox 6000, Studio 20 and Recorder MP3 players
00:50:11amiconnZagor: I just made scrollbar() ~26 % shorter (before fixing the problem) :)
00:54:46 Join midk [0] (
00:57:09 Join webguest25 [0] (
00:57:47 Quit webguest25 (Client Quit)
01:00:11 Join lenny [0] (
01:01:45 Part dethbrg ("~poof~")
01:02:04amiconnZagor: Does the code police allow the use of limits.h ?
01:02:41Zagorwhat are you thinking of doing?
01:03:00amiconnUsing the INT_MAX #define
01:03:23amiconnThere seems to be no MAX_INT #define (!?)
01:03:28 Quit lenny (Client Quit)
01:04:02Zagorcan't see any harm in using that
01:07:28amiconnThe fix is working, and the fixed scrollbar() is still smaller than the original, unfixed & unoptimized one :))
01:08:31Zagorsounds good
01:26:55amiconnAll green builds :) ... time to sleep
01:27:49Zagorgood night
01:27:58 Part amiconn
01:37:31 Quit jyp ("Leaving")
01:56:37 Quit Zagor ("Client exiting")
02:08:20***Saving seen data "./dancer.seen"
02:39:06 Join silencer [0] (
02:46:17 Quit Stryke` (Read error: 110 (Connection timed out))
03:15:07 Join gromit` [0] (~gromit`
03:29:07 Quit midk ("Leaving")
03:30:22 Join MooMaunder [0] (
04:08:24***Saving seen data "./dancer.seen"
04:17:08 Join DMJC [0] (
04:25:41 Quit MooMaunder (Read error: 60 (Operation timed out))
04:35:32 Quit silencer (Nick collision from services.)
04:35:33 Join silencer [0] (
04:35:39 Quit silencer (Nick collision from services.)
04:37:15 Quit DMJC ("Leaving")
04:40:33 Join silencer [0] (
04:56:43 Join TuDo [0] (
04:58:05 Quit TuDo (Client Quit)
06:08:25***Saving seen data "./dancer.seen"
08:08:28***No seen item changed, no save performed.
08:23:10 Join methangas [0] (
09:24:19 Join webguest70 [0] (
09:24:51 Quit webguest70 (Client Quit)
09:45:13 Quit Nibbler (Connection timed out)
10:01:05 Join amiconn [0] (
10:08:31***Saving seen data "./dancer.seen"
10:44:42 Join amiconn_ [0] (
10:49:45 Quit amiconn (Nick collision from services.)
10:49:45 Nick amiconn_ is now known as amiconn (
11:31:36 Join Nibbler [0] (
11:37:10 Part amiconn
11:47:26 Join TexJoachim [0] (
12:03:10 Join Digital007 [0] (
12:05:02 Quit Digital007 (Client Quit)
12:08:33***Saving seen data "./dancer.seen"
12:13:16 Join ashridah [0] (
13:08:27 Quit Nibbler ()
13:14:46 Join MooMaunder [0] (
14:08:36***Saving seen data "./dancer.seen"
14:14:32 Join edx [0] (
14:23:04 Quit MooMaunder (Read error: 60 (Operation timed out))
15:23:22 Quit ashridah ("sleep")
16:01:52 Join R3nTiL [0] (~zorroz@
16:08:40***Saving seen data "./dancer.seen"
16:33:16 Quit R3nTiL ()
16:55:00 Join amiconn [0] (
17:41:30 Join Spida [0] (
17:50:45 Part TexJoachim ("Bye!")
17:53:16 Quit Spida_ (Read error: 110 (Connection timed out))
18:08:42***Saving seen data "./dancer.seen"
18:18:49 Quit Ka_ (Remote closed the connection)
18:20:21 Join Ka [0] (
18:23:03 Quit gromit` (Remote closed the connection)
18:31:13crash_we have a new device where rockbox will work?
18:41:57amiconnIncredible... platforms rushing in
18:42:25 Quit amiconn (" cu l8er")
18:46:44 Quit silencer (Nick collision from services.)
18:46:45 Join silencer [0] (
18:46:51 Quit silencer (Nick collision from services.)
18:51:46 Join silencer_ [0] (
18:55:45crash_ah so you also didnt know :)
19:05:14 Join R3nTiL [0] (~zorroz@
19:10:37 Join Vassago [0] (
19:19:08 Join midk [0] (
19:27:00 Join gromit` [0] (~gromit`@
19:39:42 Quit Vassago ()
19:46:36 Quit R3nTiL (Read error: 110 (Connection timed out))
19:58:07 Quit midk ("Leaving")
20:07:30 Join preglow [0] (
20:08:45***Saving seen data "./dancer.seen"
20:23:27 Join Tang [0] (
20:24:16TangHello :)
20:54:32 Quit edx ()
20:54:46 Join Stryke` [0] (
21:12:18 Join jyp [0] (
21:31:16TangIs there somebody?
21:31:26TangSems not...
21:32:26preglowsure, but doing other things
21:37:31TangReal life once again...
21:43:56 Quit Stryke` (Read error: 110 (Connection timed out))
22:01:29 Quit Tang ("Chatzilla 0.9.66 [Mozilla rv:1.7.5/20041108]")
22:08:47***Saving seen data "./dancer.seen"
22:20:09 Join Zagor [242] (
22:28:47Zagorthat rockbox on carbon pic is fake. compare with
22:29:34preglowdid anyone actually say it was real?
22:29:54 Join MooMaunder [0] (
22:30:14 Join [IDC]Dragon [0] (
22:30:15Zagornot that I know, no. but it's better to act than react :)
22:32:38[IDC]Dragonhi guys!
22:34:00[IDC]DragonPieter in the forum seems to have problems withe file system corruptions on Ondio
22:36:18[IDC]Dragonwhich I'm sad to hear
22:36:26Zagorsounds bad
22:36:43 Join bagawk [0] (
22:37:06jypZagor: I got ATA working; thanks for the support ;)
22:37:25 Join Nibbler [0] (
22:37:30[IDC]Dragona reason not to enable multivome too hastily for the HD models
22:37:39Zagorjyp: excellent!
22:37:58[IDC]DragonI'm a bit lost on how to debug this
22:38:10[IDC]Dragonjyp: congratulations!
22:38:14jypThere's pic of root dir listing somewhere in there
22:38:29jyp(I can't access the site right now)
22:39:00[IDC]Dragonme neither, it seems
22:39:15 Join Elektroluder [0] (
22:39:55 Quit bagawk (Client Quit)
22:40:03 Join bagawk [0] (
22:40:22 Part bagawk
22:40:50jypNext week I'll start submitting patches
22:41:05Elektroludercan someone answere me a question to rockbox?
22:41:16ZagorElektroluder: only if you ask it :)
22:41:50Elektroluderdoes rockbox even work for the archos gmini 220?
22:41:57Zagornot yet
22:41:58jypHahaaaaa ;)
22:42:07jypI'm working on it
22:44:00Zagorjyp: is the led for power or disk? hardware or software controlled?
22:44:18jypI don't manage the led
22:44:41jypIt remains as it is at the time I take control
22:44:46jypsame for backligt
22:46:15Zagorthe id3 database for my 8000+ songs is only 2.3 MB large. i thought it would be bigger.
22:46:46 Join webguest23 [0] (
22:47:16 Quit webguest23 (Client Quit)
22:50:05 Quit Elektroluder ("ChatZilla 0.9.61 [Mozilla rv:1.7.3/20040910]")
22:50:14Zagorhowever the artist "folder" is almost 900 entries long so i need to add support for display partial folders
22:53:26Zagor[IDC]Dragon: well obviously we need more details from him. if it's a small mmc, maybe he can simply make a complete image available
22:54:05[IDC]Dragonbut that'll be post-mortem, whatcan we learn from it?
22:54:40[IDC]Dragonat least, pretty tough
22:55:10Zagoryes it is
22:56:27[IDC]Dragonmaybe I'd run into such problems, too, if I just would use my Ondio more often :-/
23:04:14 Quit jyp (Read error: 104 (Connection reset by peer))
23:04:14 Join jyp_ [0] (
23:04:41 Join amiconn [0] (
23:04:54[IDC]Dragonhi Jens
23:05:19amiconnI read about the Ondio fs corruption. I doubt this has to do with multivolume.
23:05:43amiconnI also got this twice, and that was before multivolume existed.
23:06:01[IDC]Dragonthat doesn't comfort me
23:06:27[IDC]Dragonany idea what provokes it?
23:07:00amiconnOne occurence was accompanied by a (slight) fs corruption that was detectable with chkdsk.
23:07:25amiconnThe other occurence simply had 2 blocks swapped, with no corruption otherwise.
23:07:43amiconnA complete (not quick) format of the flash seems to have cured it.
23:08:04[IDC]Dragonswap, hmm
23:08:17[IDC]Dragonwhy no quickformat?
23:08:17amiconnI really don't know what caused it, but it might be related to the flash itself.
23:09:07amiconn(quickformat) Because after the first occurence, I did a quickformat (only).
23:10:14[IDC]Dragonwhich should give you a shiny new file system...
23:10:29amiconnYes, but a long format touches more blocks.
23:11:02amiconnPerhaps there were some faulty blocks, that the flash chip needed to be touched in order to detect and remap them.
23:11:42[IDC]Dragonsounds exotic
23:12:23[IDC]Dragonand like I read the datasheet, replacement should be transparent
23:13:12 Join Stryke` [0] (
23:13:44amiconnCan you imagine a firmware bug that may cause *data* blocks to be swapped?
23:14:45 Quit methangas (" Like's GUI? Then try HydraIRC -> <-")
23:14:47[IDC]Dragonbug causes are sometimes hard to imagine
23:14:54amiconnI checked both FATs after the second occurence (block swapping). They were identical in, cluster chains contiguou,s for these files.
23:15:06[IDC]Dragonblock=sector or cluster?
23:15:37amiconnHmm, I'm not 100% sure.
23:17:32amiconnNo, must have been cluster (iirc the playing time of the swapped blocks)
23:17:55amiconnBut then it's really hard to believe it was caused by the flash...
23:18:30[IDC]Dragoncould have been the fat cache
23:18:53[IDC]Dragonwhich is hopefully bulletproofed now
23:19:10amiconnYes, maybe this has done the trick.
23:20:13amiconnMy sister is rather good at finding bugs. She found the video.rock bug I fixed...
23:20:45[IDC]Dragonemberassing, how long this was in
23:20:52amiconnMay I replace your Ondio FM photos with scans, btw?
23:21:09[IDC]Dragonyes, sure
23:21:57[IDC]Dragon(video bug) I meant the missing close()
23:22:27amiconnShe found the problem because she wondered about the resume point not resetting.
23:22:45[IDC]Dragonthe slider overflow may have been difficult on Ondio
23:23:18amiconnFor a change, this time I caused the bug (missing close) when I did the plugin mass-adaption for Ondio.
23:23:40[IDC]Dragonoh, I already blamed myself
23:24:13amiconnExpect a big Win32 simulator commit soon :) Featuring different pics (also as preparation for clickable buttons), different LCD sizes & colours, and some keyboard adjustments.
23:24:48[IDC]Dragonwill it compile with MSVC?
23:25:03preglowwhy is that an issue? :V
23:25:04amiconnI dunno. Probably it needs adjustments.
23:25:16preglowi thought you guys used gcc
23:25:20amiconnI don't have MSVC (costs money, unlike cygwin)
23:25:42*preglow hugs mingw
23:25:46amiconnpreglow: The Win32 simulator is compilable both with GCC/cygwin & MSVC
23:26:30amiconnThe MSVC makefile will need adjustment for sure.
23:27:00[IDC]Dragonit has always been the ugly little brother
23:35:43amiconnI'm a bit undecided about the aspect problem (recorders, ondios): Would it be better to have an 1:1 aspect (cleaner view), or the target's aspect (looking more natural)?
23:36:38[IDC]DragonI think we can only afford target aspect ratio when greatly magnified
23:37:10[IDC]Dragonelse we need 1:1 to get a meaningful display
23:37:35amiconnI tried target aspect. I think it doesn't look too bad.">
23:42:19[IDC]DragonZagor: we need to "long police" Rockbox for the Gmini
23:43:03[IDC]DragonI'm wondering if we should leave that to them, or we start with the modules we're faminliar with
23:43:04*jyp_ thinks 'Not be confused with "long island police"'
23:43:33BagderI would say we can start
23:44:12jyp_Alright ... I'm going to submit a patch with what I've done already perhaps ?
23:44:22Bagderplease do
23:44:35jyp_fat.c is the main offending file
23:44:47[IDC]Dragonyeah, I figured
23:44:54jyp_... that i've worked on
23:45:10amiconnZagor: What's your opinion on the aspect issue (Win32 sim)?
23:45:54Zagorfrankly i think the surrounding player picture can be ditched. that solves the aspect problem and decreases the software size :)
23:46:10Bagderexactly what I said the other day
23:46:32amiconnZagor: I left the picture there on purpose, for adding clickable buttons.
23:47:00amiconnAnd it wouldn't solve the aspect problem either.
23:47:31Zagorwell no, but i don't think it needs solving. it's just a UI simulator.
23:47:34Bagderwhat aspect problem are we talking about?
23:47:47Bagderthat the pixels aren't shaped the same way on the screen as on the target?
23:48:20amiconnIt's not a big problem though.
23:48:33BagderI've never given it any thoughts
23:49:06Bagderbut I wouldn't mind a resizable sim-window
23:49:46amiconnThe Win32 sim is resizable since a long time
23:50:14Zagoris it scaling the player image then?
23:50:41jyp_Do you have a script (or whatever) to indent the sources as required by your standards ?
23:50:52Zagorjyp_: emacs! :-)
23:50:59amiconnZagor: yes.
23:51:06jyp_I'm using xemacs...
23:51:19Bagderjyp_: check tools/
23:51:19jyp_I guess I can set it up accordingly
23:51:51Zagorit's in the tools directory
23:52:09Zagoroh, you already said that :)
23:52:52[IDC]DragonI'm off, bye
23:53:07 Quit [IDC]Dragon (" HydraIRC -> <- Get hot chicks here!")
23:53:13BagderZagor: you heard anything about the carbon port besides the pic in wiki?
23:53:34Zagorit's not a port, it's just a photoshopped image. i have removed it.
23:53:42BagderI wondered if it was
23:53:51Bagderfooled me
23:53:56Bagderbut I didn't check closely
23:54:10Zagorit was the iriver boot photo
23:58:06Bagdermaybe he is in a time zone that has reached april 1st already? ;-)
23:58:19Zagorperhaps :)

Previous day | Next day