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 2005-08-10

00:05:00 Join cpenner [0] (
00:06:17 Quit cpenner (Client Quit)
00:07:54 Join Stryke` [0] (
00:11:07 Join amiconn_ [0] (
00:11:49Stryke`anyone know any *nix mp3 player that can read/write replaygain id3v2 tags?
00:12:20crwli still suppose amarok can ;)
00:12:41Stryke`i looked at amaroks website
00:13:06Stryke`not yet, i'm afraid
00:13:09webguest33What does foobar's replaygain actually do?
00:13:20webguest33Is it comparable to vorbisgain, or does it alter the mpeg frames?'
00:13:38Stryke`very comparable to vorbisgain
00:13:47Stryke`writes tags to apev2 or id3v2, depending on setting
00:14:38Moosin id3v2 if you want to use it with Rockbox :)
00:14:45crwlStryke`, hm, i just changed a title tag in a mp3 file to "blahblahblahblahblahblahblahblahblahblahblahblahblahblahblahblahblahblahblahblahblah" in amarok and it shows ok at least in commandline madplay
00:14:51crwlcan't be id3v1
00:15:04Stryke`oh, i know it supports id3v2
00:15:08Stryke`its the replaygain i care about
00:15:21crwloops, oh :D
00:15:42Stryke`the only thing keeping me on windows at this point is foobar's replaygaining
00:15:49crwl(hm, there's something odd in those id3v2 tags amarok writes - the commandline id3v2 tool doesn't recognize them)
00:16:12crwlbut i think there's no tool that adds replaygain stuff to id3v2 tags
00:16:25crwlsomeone should patch the commandline version of mp3gain :P
00:16:34Stryke`mp3gain is a different animal
00:16:44Mooswhy don't use foobar?
00:16:53crwlit does the same calculations, doesn't it?
00:16:59Stryke`it changes the headers, a player doesnt have to be replaygain compliant to play it
00:17:01webguest33Well, it shouldn't be too hard to make it just write a tag instead of alter each frame
00:17:07Stryke`it doesnt store its info in tags
00:17:12Stryke`Moos: through wine?
00:17:12crwlStryke`, and that's why i said someone should patch mp3gain so it wouldn't do that
00:17:12Moosi use foobar and it work like a charm
00:17:35*webguest33 ponders filing a RFE with mp3gain
00:17:57Moosi believe mp3gain don't use id3v2 tags
00:18:20webguest33Which is the enhancement I'd be requesting
00:18:23crwlit stores its undo information somewhere...
00:18:25webguest33the ability to do that instead of altering the headers
00:18:39crwlah, in ape tags, it seems
00:18:41Mooscrwl: APE tag
00:19:10Stryke`APE tags are nice
00:19:26webguest33But why..
00:19:29Moosbut don't implemented in Rockbox :)
00:19:36 Quit amiconn (Read error: 60 (Operation timed out))
00:19:37 Nick amiconn_ is now known as amiconn (
00:19:46webguest33What's the point of APE tags?
00:19:46Stryke`yeah, they're so fast because they're at the end of the file like id3v1
00:20:13webguest33How's that faster?
00:20:26crwlthe cleanest way might be to patch mp3gain to not change the frames and just add the ape tag and then patch rockbox to understand APE replaygain tags :P
00:20:39Stryke`the whole file sometimes has to be re-written if the id3v2 tags were written without sufficient padding
00:20:39MoosLear it was the man who implemented Replaygain, maybe he will can the current patch of Frederic
00:21:17webguest33Stryke`: oh that, well that only goes when tagging, which is a one-time offence
00:21:44webguest33And having the tag at the beginning makes a whole lot more sense with streaming
00:21:46Stryke`well on the programming side, most software doesnt 'fully' support id3v2, which should allow unicode tags
00:22:02crwlall these mp3 tagging systems are pain in the ass
00:22:04webguest33most software doesn't support ape tags at all
00:22:34Stryke`foobar is still the best example of gapless/replaygain/ape or id3v2
00:22:55Moosfoobar is excellent
00:23:17Stryke`yes, i wish they're was a true *nix equivalent
00:23:48crwl*nix equivalent would be to screw all those zillion nonstandard ways of doing the same thing and use ogg vorbis ;P
00:25:06webguest33does foobar run on a multiuser windows system?
00:25:14webguest33(not running as admin)
00:25:19Stryke`yes, with per-user settings, i believe
00:28:55Mooshardeep: Hi, are you here?
00:37:38Moosprobably not
00:37:55Moosgood night (or day) all
00:38:05 Quit Moos (" HydraIRC -> <- The future of IRC")
00:39:18 Quit Luthion ("( :: NoNameScript 3.81 :: )")
00:44:18 Join TCK [0] (
00:50:02***Saving seen data "./dancer.seen"
01:05:46 Quit hicks (Read error: 110 (Connection timed out))
01:12:15 Part LinusN
01:23:19webguest33Ah. Let's hope for a smarter convbdf.
01:26:30 Quit t0mas ("CGI:IRC")
01:39:06stripwax_webguest33 - is that the problem?
01:43:48webguest33See the latest commit
01:47:15 Join EBounding [0] (
01:47:23webguest33Hm, or maybe the bitmap_t is in rockbox. Either way, it would be nice to be able to have >16 pixel wide fonts
01:53:32 Join Febs [0] (
02:01:23 Join sockerteze [0] (
02:09:27 Quit sockerteze ("Chatzilla [Firefox 1.0.4/20050511]")
02:26:44 Quit bluebrother^ (Read error: 110 (Connection timed out))
02:26:46 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
02:35:17 Quit EBounding (Read error: 60 (Operation timed out))
02:50:03***Saving seen data "./dancer.seen"
02:50:36 Quit stripwax_ (Read error: 110 (Connection timed out))
02:54:30 Quit webguest33 ("CGI:IRC (Ping timeout)")
03:01:32 Quit thegeek (Read error: 54 (Connection reset by peer))
03:06:02 Quit Asku (
03:06:02 Quit mbr (
03:06:02 Quit dwihno (
03:06:02 Quit crashd (
03:07:35 Join crashd [0] (
03:08:25 Join mbr [0] (
03:08:42 Join dwihno [0] (n=dw@
03:11:37 Join Asku [0] (
03:18:09 Quit edx (Read error: 104 (Connection reset by peer))
03:18:50 Join thegeek [0] (
03:22:19 Quit thegeek (
03:22:47NJointhegeek [0] (
03:27:07 Part courtc ("Leaving")
03:27:20 Join edx [0] (
03:31:26 Nick Sucka is now known as Sucka`zZzZz (
03:38:38 Quit thegeek (Read error: 54 (Connection reset by peer))
03:39:29 Join thegeek [0] (
03:41:05 Quit thegeek (
03:41:16NJointhegeek [0] (
03:45:23 Quit thegeek (
04:05:52 Join QT [0] (i=as@madwifi/users/area51)
04:16:01 Join EBounding [0] (
04:16:04 Quit tvelocity ("Leaving")
04:18:03 Quit QT_ (Read error: 110 (Connection timed out))
04:22:47 Quit EBounding ()
04:48:28 Quit edx (Read error: 110 (Connection timed out))
04:50:05***Saving seen data "./dancer.seen"
04:52:16 Quit hardeep ("Read error: 69 (Excessive tongue)")
04:59:01 Join solex [0] (
05:11:35 Quit solex_ (Read error: 110 (Connection timed out))
05:25:14 Join Speedforneed [0] (
05:29:06 Quit Strath ("Client closed")
05:36:03 Join Strath [0] (
05:37:10 Quit Strath (Client Quit)
06:16:20 Quit Febs (" HydraIRC -> <- State of the art IRC")
06:22:56 Join edx [0] (
06:38:06 Join DarkkOne [0] (
06:44:01 Quit alxcm ("leaving")
06:50:09***Saving seen data "./dancer.seen"
07:03:09ReKleSSanyone know why the rockbox simulator would be dying with a floating point exception?
07:10:45 Quit DarkkOne (Read error: 104 (Connection reset by peer))
07:34:31 Join Cka3ka [0] (
07:34:51Cka3kahey hey
07:35:04Cka3kaits so quiet :-\
07:40:16Cka3kaanyone still awake?
07:41:15 Quit Cka3ka ()
07:45:24 Join Coldtoast [0] (
07:45:28Coldtoasthey all
07:45:53Coldtoastwhich build doesn't have the volume bug? I need to download the last one before the bug
07:46:50ColdtoastI haven't been able to use Rockbox since the volume bug and don't really want to have to keep trying dailys til I find the last one that doesn't have it
07:51:27Coldtoastis thre a reason it's been unfixed since it was introduced? it's been about a week since the prob has been there and it makes Rockbox pretty much unusable. have ppl been unable to find what the problem is?
08:15:17 Join Nibbler [0] (
08:21:35 Join NibbIer [0] (
08:29:01 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
08:31:29 Join xNibbler [0] (
08:32:12 Quit Nibbler (Read error: 104 (Connection reset by peer))
08:36:12 Part amiconn
08:41:54 Join Nibbler [0] (
08:42:00 Quit NibbIer (Read error: 104 (Connection reset by peer))
08:43:06 Join amiconn [0] (
08:50:12***Saving seen data "./dancer.seen"
08:51:33 Join NibbIer [0] (
08:52:03 Quit xNibbler (Read error: 104 (Connection reset by peer))
08:53:36 Part NibbIer ("Leaving")
08:53:43 Join LinusN [0] (
09:00:27 Join einhirn [0] (
09:02:17 Quit Nibbler (Read error: 104 (Connection reset by peer))
09:06:06 Join bluebrother^ [0] (
09:18:30bluebrother^Coldtoast: still here?
09:18:52bluebrother^ths bug has been discussed yesterday (or was it the day before?), just read the logs.
09:22:01 Join webguest30 [0] (
09:22:56Coldtoastah ok
09:23:30ColdtoastI found the last build that doesn't have it: 20050730
09:24:01Coldtoastat least now I can take my h140 to the gym and use it. heh
09:24:07bluebrother^to me that was 050726 ...
09:24:28Coldtoastinteresting there'd be a difference
09:24:57LinusNwhich bug?
09:25:25bluebrother^the volume bug.
09:25:28Coldtoastwhen you change the volume, when it gets to around 80% (83 for me) it slows to a crawl and the volumd doesn't actually change
09:26:00Coldtoastthe WPS shows voluem changes about once a sec but the volume doesn't change and the player tends to become unresponsive
09:26:22Coldtoastif you then change all the way down to 0, for example, the WPS shows the volume change but the volume doesn't actually change
09:26:28Coldtoastbut the track keeps playing fine
09:26:58Coldtoastto turn it off, you have to hold in the button and it eventually responds
09:27:18LinusNi sat all night trying to fix that bug
09:27:31LinusNit's really hard to reproduce for me
09:27:48Coldtoastit does it 100% of the time for me
09:28:09Coldtoastif I change volume at full speed
09:28:19LinusNdoes it matter which song you're playing?
09:28:35bluebrother^I sometimes had the impression this bug happens less when compiled with debug mode on.
09:28:36ColdtoastI've only tried MP3 tho
09:28:36LinusNwhicih build?
09:28:45Coldtoastfrom every build after 20050730
09:29:06ColdtoastI'll try the latest now tho if you want
09:29:11LinusNplease do
09:29:27bluebrother^for me this occured after amiconn's timer changes on 050727
09:31:17Coldtoastit doesn't do it with the latest!
09:31:39LinusNdid it reset the settings?
09:32:05Coldtoastand I had 20050730 on there
09:32:21ColdtoastI've tried builds between 20050730 and this latest bleeding edge and some have tho
09:32:41Coldtoastyes it does
09:32:51LinusNsee? it comes and goes
09:32:51Coldtoastbut it does it much closer to 100%
09:33:16LinusNas far as i can see, it's a timing problem, and can happen anytims
09:33:51Coldtoasthow frustratng for you guys
09:34:17bluebrother^ok, got it.
09:34:22LinusNas soon as i'm closing in on it, it goes away
09:34:52bluebrother^just installed the latest daily. Reproducing took *way* longer than before but I could trigger the bug :(
09:35:08CoCoLUScouldn't we just revert amiconn's timer code changes?
09:35:24LinusNi'm not sure thay have anything to do with it
09:36:21CoCoLUSyou just said it's most likely a timing problem
09:36:28Coldtoastwill reverting break anything done since?
09:37:08LinusNColdtoast: reverting it will be a pain
09:37:30LinusNand i want to know the cause of the bug before i do anything to fix it
09:37:31Coldtoastyeah. I'd imagine so. it's been 8 days
09:40:24LinusNwhat happens is that the uda1380 locks up and stops responding on the i2c bus
09:40:42LinusNthe lagging comes from the timeouts in the i2c driver
09:41:13LinusNi've been trying to analyze the i2c bus, but it's difficult when you can't reproduce the bug on demand
09:41:39*LinusN gets some coffee
09:41:43Coldtoastwish I could send you my player
09:43:42 Join UncleBill [0] (
09:43:51UncleBillhello friends
09:45:33UncleBillI'm interested in playing with the source and building for the iRiver, but nervous I'll crush it. how dangerous is just messing around without owning one of these P&E Wigglers?
09:45:43UncleBillheya LinusN, Coldtoast
09:46:17CoCoLUSuncle, as long as you don't mess with the bootloader, you're quite safe
09:46:41UncleBillis there any functionality in the bootloader beyond booting the image?
09:46:47CoCoLUSusb mode
09:46:50CoCoLUSfor recovery
09:46:54LinusNfor safety
09:47:22UncleBillwhere can I read on building a development environment and building from source?
09:47:24LinusNnobody has killed a player yet by installing rockbox
09:47:44Coldtoastwonder if anybody's managed to revive their player
09:47:57LinusNrevive from what?
09:48:08Coldtoastfrom being bricked
09:48:34Coldtoastby the stupid Iriver fw
09:48:34UncleBillwhy there it is. don't know how I missed it. thx
09:49:13Coldtoastif there had been a bootloader for the h300, my friend wouldn't have had to RMA his player. heh
09:49:16LinusNUncleBill: the "documentation" section on has all you need
09:49:32Coldtoasthis got bricked the second day he had it
09:49:41UncleBillthis is a good thing you guys have going
09:50:10Coldtoasthe thinks it's cos he created a corrupted DB and enabled DB mode
09:50:30CoCoLUSuncle, rockbox is great yes, as long as you don't try to change the volume... ;)
09:54:25CoCoLUSit's interesting how one accepts bad products when there's no alternative
09:54:52CoCoLUSeverytime i use the original firmware now, i get seizures having to endure that font size
09:54:54Coldtoastbad products?
09:55:01Coldtoastoh. heh
09:55:25Coldtoastit's funny tho
09:55:38ColdtoastI've always liked my h140
09:55:53Coldtoasteven before rockbox,
09:56:14ColdtoastI never realised how rough the iriver fw sounded tho. For things liek changing tracks
09:56:41Coldtoastwith teh iriver fw, I get a little glitch of sound just before it switches tracks. I'd never noticed it before
09:57:25CoCoLUSyou never noticed that -huge- gap between tracks?
09:57:34Coldtoastof course I noticed that
09:58:22Coldtoastthat doesn't bother me as much as the shitty noise you hear when changng tracks tho
09:58:40Coldtoastrockbox used to do it but Slasher fixed it
10:17:40 Join ashridah [0] (
10:18:21 Quit merbanan (Read error: 104 (Connection reset by peer))
10:18:23 Join banan_ [0] (
10:24:21 Quit banan_ (Read error: 104 (Connection reset by peer))
10:24:22 Join banan__ [0] (
10:50:15***Saving seen data "./dancer.seen"
11:24:42*amiconn spots LinusN, aka Mr. mpeg.c ...
11:25:16amiconnLinusN: are you around?
11:27:38ReKleSSanyone know if alexspy comes on here?
11:28:05ReKleSSor, has anybody else done anything about the remote wps/file tree?
11:30:40 Join arfo_____ [0] (
11:31:34 Nick arfo_____ is now known as arfo (
11:31:34 Quit arfo (Client Quit)
11:32:31 Join arfo [0] (
11:35:34LinusNamiconn: i'm here now
11:35:51LinusNReKleSS: no, we haven't worked on it
11:36:42amiconnLinusN: I found a bug in mpeg.c, immediately before my trip to Rome
11:37:05amiconnmpeg.c, line 691 caused a NULL pointer access
11:37:13ReKleSSthat's the main thing stopping me from using rockbox instead of the original firmware, I'll start on that when I get a chance... probably this weekend sometime
11:37:23amiconnObviously it doesn't happen always
11:37:49LinusNif (num_tracks_in_memory() > 1)
11:39:00LinusNhow can that cause a NULL pointer access?
11:39:14LinusNwhich version of mpeg.c?
11:39:25 Join XavierGr [0] (
11:39:33ReKleSSwell that was good timing...
11:40:16ReKleSSXavierGr: you're alexspy, right?
11:40:25amiconnLinusN: Ah, that was my modified version (the modifications don't influence this part; only recording)
11:41:24amiconnIt's line 696 in the cvs version
11:41:42amiconn(the track->mempos access)
11:42:03XavierGryeah and hi
11:42:21ReKleSSerr... hi :p
11:42:31ReKleSShow much progress have you made on the remote lcd stuff?
11:43:02LinusNok, so track_offset is off limits
11:43:51amiconnMy suspicion is that it happens if lines 664..672 detect a trackchange, and then line 691 checks num_tracks_in_memory() again, the count has not yet been updated
11:43:53XavierGrWell it displays correctly the tree and the wps. Also there are many options to change and save. But there's still a bad bug that I can overcome by enabling an option to render the main unit OR the remote. See link
11:45:38amiconnLinusN: I'm considering to permanently enable the memory guard on archos, "catch zero area" mode. Should help uncovering some bugs, but it's less convenient for the user...
11:45:50XavierGrUnfortunately this is not a smart way to implement the remote. So it is a temporary way to use the remote until someone takes the task to do more proffesional work. Maybe after a lot of work I could make that happen we will see, no promises though
11:47:15XavierGramiconn: I have a new snake2.c file with 2 bugfixes of the hiscore. Are you interested?
11:49:27XavierGrAlso have you noticed that the simulator crashes (when changing char pointers in some manner) when the real player runs them correctly?
11:50:16XavierGr file = "/pic.jpg";
11:50:16XavierGr filepath = "/Image";
11:50:16DBUGEnqueued KICK XavierGr
11:50:16XavierGr filename = rb->strcat(filepath,file);
11:50:16XavierGr rb->splash(HZ, true, filename);
11:50:41XavierGrthis will crash the simulator but my player splashes a box with the whole filename.
11:51:26SlasherXavierGr: eh, you can't do that
11:51:39 Nick Sucka`zZzZz is now known as Sucka (
11:52:10XavierGrmy player runs fine with it...
11:52:11SlasherXavierGr: you are trying to copy a longer string to the filepath that it has allocated memory
11:52:16Slasherso it will overwrite something
11:52:47XavierGrany better way to copy the whole string into 1 variable?
11:52:48Slasherand generally you should never modify a constant string
11:53:46Slasherchar buf[40]; strncpy(buf, file, sizeof(buf)-1);
11:53:53Slashersomething like that should work
11:54:40Slashernow i have to go ->
11:56:08XavierGrbye Slasher.
12:00:43 Part arfo
12:03:45LinusNamiconn: it should be safe to change it so it checks the return value of get_trackdata() instead of num_tracks_in_memory()
12:08:56amiconnThe problem is that I don't fully understand the 'magic' done in transfer_end()
12:10:23amiconnWhat do you think about enabling 'memory guard' to aid debugging?
12:10:33LinusNfine with me
12:11:04LinusNtransfer_end() is not among the easiest parts of mpeg.c :-)
12:12:52LinusNhow do you go about to reproduce the bug?
12:14:33amiconnI just enabled the memory guard and started playing music. After a couple of tracks I got a UserBrk interrupt.
12:15:23amiconnI spotted the exact point with the help of the .map file and a disassembler listing
12:16:18amiconn(fyi, I was using my Studio when this happened, but it should be model independent)
12:20:29 Join Moos [0] (
12:20:48MoosHello guys ! :)
12:23:03LinusNamiconn: i'll see if i can check it out this evening
12:23:43 Join Febs [0] (
12:23:52XavierGrhi Moos and Febs.
12:24:03FebsGood morning.
12:24:28Moosmorning guys
12:29:50 Join psychocydd [0] (
12:30:14 Part psychocydd
12:34:32Coldtoastanybody want to go to work for me?
12:36:42LinusNtough question, where do you work?
12:38:25Coldtoastnightclub. I'm a bouncer
12:41:03XavierGrwhat's a bouncer?
12:41:53Coldtoastnightclub security
12:42:30XavierGrI don't think I can do that.
12:42:55Coldtoastsure you can. it's great
12:43:14Coldtoasteach say is better than the next
12:43:15LinusNColdtoast: like "no membership card? get lost!"
12:43:16XavierGrI am not so strong to be a security guy.
12:43:50Coldtoastnah. like "Excuse me. We're non-smoking here. Just need you to take that to the smoking area"
12:43:59XavierGrTough job I would say and dangerous.
12:44:14Coldtoastor "If you do that one more time I'm going to have to ask you to leave"
12:44:22XavierGrI hope that you get paid well.
12:44:29Coldtoastit's ok
12:44:52Coldtoastonly have to work ~25hrs a week and it's easily enough to live on
12:48:09Moosis there any battle with hot guys who drunked much?
12:48:34ColdtoastI'm doing well tho. so far, in the 2+ years I've been doing it, I've not been charged with assault yet nor have I had to visit hospital
12:49:07Moosapear you work in a cool club
12:49:24ColdtoastI thought one time I might have had a broken nose but it wasn't. was just a bit crunchy. heh
12:50:02Coldtoastthe only things that have really happaned are I've damaged my hands once each
12:50:17***Saving seen data "./dancer.seen"
12:50:54ColdtoastI can tell you one thing too. NEVER EVER hurt your thumbs. they're teh WORST things to hirt. worse than anything else I've had happen cos they take months to heal. MONTHS
12:51:54ColdtoastI'd rather have a broken nose than damaged thumbs
12:51:58ReKleSScoldtoast: what did you do to damage them?
12:52:05Moosis it your regular work all the year?
12:52:14Coldtoastrestraining a guy for too long
12:52:47Coldtoastmy you know the "heel" of your thumb? the big meaty part your thumb comes out of?
12:53:24ReKleSSI can kind of see what damaging that would do...
12:53:30Coldtoastthat part went pretty much black and the bruising went all teh way around to the top of my hand near teh thumb and was yellow
12:53:51Coldtoasttook probably 7 months to heal properly
12:54:11Coldtoastanyway. time to go get ready. later
12:54:25Moosgood day
12:56:22XavierGrhave a nice day (night) in your work
12:57:12 Join leftright [0] (
12:58:04leftrightLinusN, you around /
12:59:34leftrightI managed to see RB crash twice on H140, with error IO3 adderr, 31022E3E
13:00:29leftrightit occured when using FIle Tree playback and it was transitioning from lat song in the album to first song in next album
13:01:41amiconnLinusN: I'll see whether I can remember the album where this happened. Perhaps it needs a special constellation of buffer wrap & track change
13:02:04leftrightmp3 file, no crossfade, RG on, 'Move to next folder set'
13:11:33Febsleftright, I've noticed the same thing:
13:12:14FebsExcept that I had RG off.
13:34:50 Join Liehbeth [0] (
14:08:24XavierGris there a fixed space memory for plugins?
14:09:10XavierGrBecause when I try to make some big char arrays into the jpeg viewer the pictures aren't displayed correctly
14:24:32 Join tvelocity [0] (
14:28:25 Quit Febs (Read error: 110 (Connection timed out))
14:30:39LinusNXavierGr: check the map file
14:41:26leftrightanyone looked into the feature of ReplayGain going to track mode whenever shufle is enabled ?
14:43:51 Part Liehbeth
14:45:36CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
14:45:36*leftright hugs ReplayGain
14:50:18***Saving seen data "./dancer.seen"
14:54:54XavierGrLinusN: I wish I knew what is the map file.... :p
14:55:04 Join hicks [0] (
14:59:51LinusNXavierGr: each .rock file has a .map file
15:00:05LinusNin the build dir
15:00:15XavierGrok thanks
15:06:07XavierGrLinusN: When I make changes to a specific plug-in. Do I have to replace the .rockbox folder again or can I just throw the plugin there for replacement?
15:13:29LinusNjust replace the changed files
15:14:39amiconnconvbdf is a mess... :-/
15:18:56LinusNamiconn: tell me about it
15:19:21LinusNtook me some time to find the 16-pixel bug
15:22:14amiconnYes, and the strange thing is that it shouldn't happen
15:22:32amiconnAfaiu convbdf should work for arbitrary widths
15:23:29amiconnI also don't understand why it uses this bitmap_t type. Using unsigned char for bitmap data would seem much more straightforward to me
15:23:29 Join Maxime [0] (
15:23:44amiconnNo endian hassle, ...
15:23:54LinusNyes, it could be a lot simpler
15:24:51LinusNstill, just using a 32-bit type would probably suffice
15:25:18amiconnI'll probably throw out all old (non-ROTATEd stuff), then do some rewriting to get rid of the 16 pixel limit
15:25:37LinusNhave fun :-)
15:25:44amiconnHaha :-/
15:26:07LinusNi'll see if i can add memguard to the iriver as well
15:26:24LinusNat least for the flash
15:26:27amiconnI didn't find a comparable feature in the datasheet. Did you?
15:26:53LinusNthere is a hardware breakpoint unit
15:27:05amiconnDoes that allow to catch ranges
15:27:14 Join muesli- [0] (
15:27:30amiconnThe user break controller in the SH1 is really useful for these purposes
15:28:08amiconnDo you think renaming the UserBrk interrupt would be helpful?
15:32:21amiconnAt the moment, rockbox displays "UserBrk" when memguard is enabled and an illegal access happens
15:32:57amiconnWe don't use the user break controller for anything else, so it might be useful to change the display name something giving a hint that an illegal memory access happened
15:34:08amiconnBtw, on coldfire, there already is a 'hint' if you try to access certain areas (those that contain nothing)
15:34:23amiconnDoing this causes the bus to hang.
15:34:38amiconnI had to learn this when implementing the grayscale lib
15:37:08amiconnI loaded a (byte) index from memory, but didn't obey the fact that the coldfire doesn't do any longword extension (signed or unsigned), unlike the SH1
15:37:42amiconnI had to add a clr.l before the load operation...
15:38:21 Quit ashridah ("Leaving")
15:45:37 Nick banan__ is now known as merbanan (
15:48:18LinusNok, i gotta go now, cu later
15:48:22 Part LinusN
15:55:24 Quit webguest30 ("CGI:IRC (EOF)")
16:07:04XavierGramiconn:do you know how can I extract a global_setting.value from a plugin?
16:07:33XavierGrExample: I want to know the value of MAX_FILES_IN_DIR the user has from a plugin
16:07:46XavierGrWhich file I must include?
16:08:50 Quit muesli- (Read error: 110 (Connection timed out))
16:13:59amiconnXavierGr: From a plugin, you should never include other .h files than plugin.h or files from apps/plugins/lib
16:14:13amiconnThe global settings are available via the plugin api
16:14:37XavierGrso to call the value I type rb->global_settings.value?
16:15:46amiconnglobal_settings is also a pointer in the api, so you need to use
16:16:05XavierGrok thanks!
16:16:27amiconnSome plugins already use that, e.g. video.c
16:16:50XavierGrOne day I must understand this saperation of plug-ins and main threads
16:16:54amiconn(not yet functional on iriv3er)
16:17:18Suckahooray! just got an email saying my iriver was repaired under warranty and theyre sending it back
16:17:20*Sucka dances
16:17:30XavierGrThat's awesome
16:17:42XavierGrwell we all know iriver and it's replacement policy
16:17:53Suckai sent it back to the retailer actually
16:18:14Suckacheaper postage :D
16:18:32XavierGrWhatever, I have heard many stories were the user was to blame but iriver replaced his player.
16:22:05XavierGrwhat was your case?
16:22:18XavierGrWhat did you do to your player?
16:40:27 Join ghostiger [0] (n=ghostige@tor/session/x-31680a8dfbecfa05)
16:45:05 Join thegeek [0] (
16:47:33 Quit tvelocity ("Leaving")
16:50:21***Saving seen data "./dancer.seen"
17:17:39XavierGrwhich is the best way to reset a char array or a char *pointer?
17:25:23 Join muesli- [0] (
17:26:51 Quit UncleBill ("Client exited.")
17:39:16 Join Lear [0] (
17:39:29 Join UncleBill [0] (
17:39:36UncleBillhello friends
17:39:54muesli-g'day mate+
17:40:02UncleBillheya muesli-
17:42:22LearXavierGr: null pointers are bad. :)
17:42:37XavierGrI am lost...
17:42:52XavierGrI use too many char pointers and then I cant controll them
17:43:26LearI was thinking about the one in the code you sent to the mailing list...
17:43:47XavierGrforget that I must send an e-mail to warn not to look at it
17:44:08XavierGrit is all wrong. I found something more quick and easy.
17:44:19XavierGrThough the char pointer hell still exists
17:44:42XavierGrI am trying to make single char arrays but then I am running out of plug-in memory
17:45:06Learsounds like you're creating huge arrays...
17:46:10XavierGrwell not sooo huge. I have cuted them for my test needs and still...
17:46:35XavierGrthe bigest char array I've got is 2 dimension [10][50]
17:46:38XavierGris that big?
17:46:58XavierGrThough if I am carefull I will not need more that [3][MAX_PATH]
17:48:25Learno, that's not very big, but you must do something to run out of plugin memory (you do mean as in link error?)
17:48:49XavierGrAnother problem is that I am not used in using pointers. Ahhh old days in Visual Basic were more easy with strings...
17:49:23XavierGrwell I dont get an error, it is just that when I maximize the size of the arrays the jpeg viewer will not render the files correctly.
17:49:28XavierGrStrange isn't it?
17:52:24XavierGrYou said that you were thinking something about the mail I sent in the list. Do you have a nice advise for me Lear? I could use some help.
17:53:44LearThe code in the mail calls on a function to load a directory, but the "context" pointer is never initialized, and thus null.
17:54:05Learas for the rendering problem, as I don't know much about the inner workings of the jpeg loader, I can't really help you there...
17:55:19XavierGrWell I solved that problem.
17:55:21Learas for pointers, it really isn't that hard, once you get used to/learn the concept, but it is something you really need for C programming. :)
17:56:18XavierGrNow I use something completely more light and fast to see the contetnts of a directory. (though still there is a gray area which I dont understand)
17:56:26XavierGrI will PM it to you.
17:57:20 Join bagawk [0] (n=lee@unaffiliated/bagawk)
17:59:27 Join Nibbler [0] (
18:04:03XavierGrI have to go. Later all!
18:04:05 Quit XavierGr ()
18:18:18muesli-cya mates
18:27:00 Quit Maxime ()
18:27:00 Quit muesli- (Read error: 104 (Connection reset by peer))
18:35:27UncleBillI can't seem to get a development environment setup. I'm using cygwin and building for iRiver (m68k-elf). binutils-2.16 builds and installs fine. then I add the binutils bin to the head of my path. when I make gcc-3.3.4, I get errors in cc9Tass2.s about fmovem.l. any ideas?
18:42:47bagawkUncleBill, it might just be easiest to use BC's dev kit
18:43:20UncleBillI'll check it out
18:43:30bagawkhumm gcc 3.3.4
18:43:42bagawkI think they expect you to use a newer gcc version
18:43:45UncleBillI'm just following the wiki directions
18:43:56bagawkI tink they use that for archos SH1 players only
18:44:00UncleBillit's very specific not to use newer gcc.
18:44:46bagawkThats it
18:45:02bagawkYou need gcc 3.4.X
18:45:32bagawkRead the download the source/coldfire section
18:46:23UncleBillyep. I was doing that as we talked. it compiled fine this time
18:46:59bagawkThat was very fast
18:47:17UncleBill(the wiki says this is a patch for a failure in binutils, not gcc, so I didn't think that comment was applicable)
18:47:35UncleBillI had been building for maybe 5-10 mins now
18:50:24***Saving seen data "./dancer.seen"
18:51:45 Join _aLF [0] (
18:57:27UncleBillbagawk, BC's dev kit has a list of "supported mp3 players", and they're all Archos's. are you sure it will work for iRiver (which has a different cpu)?
18:58:15bagawkbagawk, yes, but I just though of a even better way, since you already have downloaded and installed cygwin
18:58:31UncleBillspeak on, friend
18:58:34 Quit Seed (Read error: 110 (Connection timed out))
18:58:47bagawkRerun the cygwin setup.exe
18:59:06bagawk(it will keep everything you have now)
18:59:12UncleBillwow, that is cool
18:59:35UncleBillthe wiki could use a bit of organization in this area (imho)
18:59:51bagawkBut it is a wiki
18:59:56bagawkAnyone can change itr
19:00:17UncleBillwhen I have everything going, I believe I'll do that
19:00:43UncleBillok, I may go back to the cygwin packages you pointed out, but for now, I have everything installed from source. I'll play with that a bit
19:02:47 Join synd [0] (
19:05:31bagawkUncleBill, great
19:10:07UncleBillthanks for the help, bagawk.
19:10:12UncleBillgotta flee to work now. later, all
19:10:26 Quit UncleBill ("Client exited.")
19:23:15 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
19:38:28 Join zezayer [0] (n=jake@
19:47:28 Quit bagawk ("Leaving")
19:58:51 Part zezayer
20:01:01 Join cpenner [0] (
20:02:14cpennerhow do I go about filing a bug report for rockbox firmware for the iriver?
20:05:16 Quit cpenner (Client Quit)
20:10:59 Join einhirn [0] (
20:25:48 Join [1]Moos [0] (
20:25:49 Quit Moos (Read error: 104 (Connection reset by peer))
20:25:51 Nick [1]Moos is now known as Moos (
20:26:58Coldtoastoh wow!
20:27:10Coldtoast"podcast" is now in te Oxford dictionary!
20:27:20Coldtoastthat didn't take long!
20:39:35 Join preglow [0] (
20:41:28 Quit Speedforneed (Read error: 104 (Connection reset by peer))
20:42:12preglowcoding playback speed support for iriver rockbox should be a pretty quick hack
20:44:00Coldtoasthey preglow
20:47:28 Join Seed [0] (
20:47:39 Join webguest81 [0] (
20:48:50preglowyo, seed
20:48:55preglowany news on musepack? :P
20:50:04Seedno, sadly
20:50:11Seedno good, no bad
20:50:25***Saving seen data "./dancer.seen"
20:51:49preglowrockbox support is still what i left it at, it seems
20:53:08SeedI'm sad to not have been the person who created it
20:53:18Seedor things would have gone very very differently
20:53:52Seedone day it'll all be revelead
20:53:57Seed<evil laughter>
20:54:02 Join stevenm [0] (
20:54:37preglowhi, stevenm
20:55:08preglowmusepack port will require a ton of work, and i dont know if well ever get seeking working
20:55:17stevenmpreglow, hello
20:55:18preglowbut hell, i dont seek all that often anyway
20:56:32Learwhy would seeking be so hard (not being very familiar with the musepack port...)?
20:56:54preglowlear: its hard to seek in musepack streams without walking the entire stream¨
20:57:19stevenmit is like that with many tracker formats also
20:57:31preglowlear: the musepack people themselves recommended backing up the decoder state at certain points to facilitate seeking, which i believe speaks volumes about how nice a format it is for seeking
20:58:09 Quit ghostiger (Remote closed the connection)
20:58:09SeedLear: I'm sure people will accept it even without seeking at first.. but that's not a way to add a codec
20:58:14Learyou mean walking as in decoding pretty much the entire stream (up to the seek point)?
20:58:16Seedsorry Lear :)
20:59:08preglowlear: as in huffman decoding it until you reach the place you wanted to get to
20:59:09 Join ghostiger [0] (
20:59:18preglowlear: actually decoding the data shouldnt be necessary
20:59:32preglowas in synthesizing the audio
20:59:50leftrightLear: dont you think it would be nice if RG went automatically to 'Track' if shuflle is selected ? :)
21:00:01Learokay, could be worse, I guess... But what about the new stream format that has been discussed?
21:00:23Learleftright: as an option, perhaps. I don't want it, but I see the point in doing so.
21:00:26preglowyou got an url?
21:00:44preglowi havent seen it, but at least they know about the problem, and there are several nice fixes for it
21:01:10Learyou mean the stream format? I has been mentioned on Frank Klemms web pages...
21:02:27Coldtoastdo any of you "reliably" get the volume bug?
21:02:34preglowi do
21:02:59Learcoldtoast: as in sluggish response when rapidly changing it?
21:03:00ColdtoastLinus said he randomly does. he''ll start to zero in on it and then it goes away
21:03:11Coldtoastit's more than that lear
21:03:26Coldtoastwhen it gets sluggish, try dropping the volume to 0
21:03:45Learyes, I know it stops working altogether...
21:03:58Coldtoasterr.. yeah
21:04:18ColdtoastI tried the latest bleedign edge and I have to put the volume up much higher before I get it
21:04:50ColdtoastI was getting it around 80% but I tried the latest bleeding edge and it doesn't do it til about 95%
21:05:10Learpreglow: a very old link I had doesn't work, and googling didn't show anything likely in the first few pages...
21:05:15 Part synd
21:06:03preglowlear: didnt find it myself
21:06:07Learpreglow: at least partial mirror is available here:
21:06:29preglowi hope they get it going soonish
21:06:42preglowi like the simplicity of the format, but not its current flaws
21:08:19 Quit stevenm (Read error: 104 (Connection reset by peer))
21:09:27Learcoldtoast: based on when the bug seems to start appearing, I suspect the timer changes from July 26. I'd like to hear what amiconn has to say about that. :)
21:09:44ColdtoastCoCoLUS thought that too
21:10:16Coldtoastsaid when the changes were implemented is when he started getting it. I don't get it til the build fromt eh 30th
21:10:39Coldtoastactually, 31st. 30th is ok
21:11:18 Join LinusN [0] (
21:11:42LearI downloaded a number of builds, and first found it in the July 27 build (and I tried hard in the 25 build).
21:12:54preglowhi, linus
21:14:56 Join dpassen1 [0] (
21:21:21Learhm.. should obsoleted strings be removed (as in emptied) immediately?
21:22:28LearI'd like to commit an update for the id3 info screen, where among other things, only LANG_ID3_NO_INFO is used when information is missing...
21:24:21LinusNLear: any and all improvements to the id3 info screen are welcome
21:25:05LearI suspected that. Thought it looked a bit silly with a ten row screen (in my case) and then only show text on two...
21:25:46LearBut should I obsolete the no longer needed strings as a part of the commit? I think I've seen that, but...
21:26:44LinusNyes do that
21:28:29 Quit dpassen1 ("Leaving")
21:29:28preglowlinusn: so, rockbox has already got a playback speed config screen for platforms that support it?
21:29:44LinusNLear: regarding the resampler, will it choke if the codec changed the resampling factor on the fly?
21:29:57LinusNnot playback speed, pitch
21:30:13preglowwell, theyre directly related
21:30:14LinusNas in sample rate
21:30:38preglowthe resampler should handle a changing delta just fine
21:30:53LearDon't think so. There's a "last_sample" that will be zeroed, causing a bad sample or two, but no more than that. :)
21:31:07preglowah, but dont reset that
21:31:11preglowjust change the delta
21:31:15preglowthats all thats needed
21:32:04preglowsomeone should make the resampler do a stereo stream in one pass ;)
21:32:06LinusNi want the mp3 codec to switch sample rate depending on the decoded frames rather than trusting the info from the id3 parser
21:32:07LearPerhaps the set_frequency command should be changed (or a new one added), so that last_sample isn't touched, and perhaps not the phase either...
21:32:12preglowanother week, and i might be coding again
21:32:36preglowlinusn: yes, definitely
21:32:50LinusNi'll test that
21:32:51preglowlinusn: there are plenty of files with wrong sfreq in the id3 tag
21:32:59Learlinusn: yes, I gathered that, though one might want to make the id3 metadata parser a bit more robust (though I'm not sure how feasible that would be).
21:33:01LinusNthe freq is not in the tag
21:33:32LinusNthe id3 parser is quite robust if you ask me
21:33:52preglowlinusn: but yeah, playback speed/pitch control for iriver could be quickly hacked together by using the resampler
21:34:01LinusNit's just that it does a lot more than parse id3 tags
21:38:02LinusNi see two options for the frequency change
21:38:13LinusN1) add a SWITCH_FREQUENCY config
21:38:26LinusN2) add a RESET_RESAMPLER config
21:39:42preglowhow about just setting the resampler after the sfreq in the first frame? it should never change
21:40:13preglowbut of course, this is The Correct Way
21:40:15LinusNi will only set it if it changes
21:40:25Learperhaps should have said, "mp3 metadata", there's lots of stuff parsed there...
21:41:36LinusNi'll set it after decoding the first frame, then i know if it's a valid frame
21:42:03Learthe dsp has a "full reset" and a "set sample frequency" command, but full reset isn't always executed.
21:42:19Learmaybe "set new sample frequency" should be added...
21:42:26Learvery simple to do...
21:42:40LinusNand full reset doesn't reset the sample_data
21:43:50LinusNhaving a SWITCH_FREQUENCY without resetting sample_data would allow us to set it on each frame regardless
21:45:27LinusNthe mpa codec sets stereo mode on the fly
21:45:42LinusNso it could just as well set the delta as well
21:46:51LinusNso, should i add DSP_SWITCH_FREQUENCY?
21:48:28 Join hardeep [0] (i=hardeeps@freeshell.ORG)
21:49:43LinusNwhy are the DSP_xxx symbols defined in playback.h?
21:51:16Leartrue, it doesn't reset the sample data, but the default is no resampling anyway...
21:51:47 Quit Coldtoast ("Peace and Protection 4.22")
21:51:59Learsome configurable items refer to the playback, and it's an enum...
21:52:30LinusNi still don't get it
21:53:10Learyes, DSP_SWITCH_FREQUENCY, or something like that... (DSP_CHANGE...)
21:53:31LinusNi though all the DSP_xxx items belonged to dsp.c
21:54:13Learthe codec api has a configure function. some of the settings involve the buffering in playback.c. True, all DSP_* items belong to dsp.c, and playback.c passes any unknown settings to dsp.c.
21:54:45LinusNok i get it
21:54:55Learwell, not all belong to dsp.c; the enable setting is in playback.c.
21:55:04LearSo as to not call dsp_process unless needed.
21:55:18LinusNbut that is a CODEC_xxx symbol
21:56:10 Join incognito [0] (
21:56:42Learah, yes...
21:58:31incognitodoes anyone know if anybody is optimizing mpc playback on the H1x series? Its pretty cool that they play, although it cuts out every 5-10 seconds...
21:59:18Learnot at the moment, afaik...
22:01:34incognitodoes anyone want to speculate on a september 5th or earlier release with realtime playback of mpc?
22:01:53Learwhy that date?
22:02:03incognitojust threw it out there
22:03:28LinusNrealtime mpc will come when someone bothers to work on it
22:03:57preglowincognito: we more or less need to rewrite the mpc code to make it work fast enough
22:05:48incognitoso it can be much more optimized i gather?
22:05:56preglowoh yes, tons
22:06:01preglowbut the current code doesnt make it easy
22:06:07incognitoi see
22:06:13incognitothanks for the response
22:06:13preglowi just did the easiest optimisations and left it at that
22:06:42incognitoi was pleasantly surprised to hear my mpc files play at all on a portable :)
22:07:59preglowbut yeah, tons and tons of potential, you just need to find someone skilled in fixed point math and willing to do some work
22:08:16incognitoyeah, but then again, ogg vorbis is getting pretty damn close to mpc transparency rates
22:09:34preglowits really bitter, because musepack is such a simple codec compared to for example vorbis
22:10:14incognitovery fast decoding also, at least on pc architecture
22:10:40preglowyes, it _should_ be fast
22:11:29LinusNon-the-fly frequency switch worked nicely
22:11:47preglowwell, it should
22:12:24LinusNonly problem will be the time estimate
22:12:45LinusNwhich is based on the frequency
22:13:02LinusNfound by the metadata parser
22:14:47preglowwhat does that base its findings on?
22:14:54preglowthe sfreq, that is
22:15:17LinusNthe metadata parser attempts to find the first valid mp3 frame header after the id3v2 tag
22:15:44LinusNthe initial sample frequency is taken from that
22:17:26LinusNbtw, samples_done isn't updated correctly when (re)winding
22:17:48LinusNso the resume info is wrong when stopping the playback
22:19:00preglowi dont think its ever correctly updated when seeking
22:19:15preglowgapless playback is pretty broken at the moment as well
22:19:22preglowbut not much i can do about it right now
22:19:34preglowyeah, at least in the build i use now
22:19:54preglowand its pretty recent
22:20:39LearI haven't noticed anything... I'm not using the most recent build, true, but I have there been any changes regarding that?
22:20:41preglowill contact gabriel bouvigne and ask him exactly how to implement gapless mp3, so i dont have to guess anymore
22:21:02LearIt looks pretty much the same as foobar...
22:21:04preglowlear: its pretty marginal, but id guess at least a thousand samples of noise/silence is inserted between tracs
22:21:10preglowlear: thats where i took it from
22:21:28LinusNsounds gapless to me
22:21:35LinusNjust tried
22:21:39preglowok, ill get me a new build
22:24:35preglowi can most definitely hear a gap
22:25:09LinusNhand me the two tracks
22:25:51preglowgimme a sec, ill check more
22:26:13Learhe, that's a new one... a woman with acrophobia drove up on a high bridge by accident, and didn't dare continue. The police had to rescue her.
22:26:30Learand that's with a lame encoded track (vbs)?
22:28:26preglowive i also encoded a sine wave as two tracks
22:28:44preglowthat case will always glitch, thanks to mp3 weaknesses, but it shouldnt glitch this bad
22:29:28 Join LaMeD [0] (
22:30:53LaMeDHey. I would like to search the repository for patches that have been accepted. I'm looking for one specific patch that I can't recall it's name :/
22:31:16Learso search! :)
22:31:35LaMeDyeah.. where?
22:31:45LaMeDhow do i search the cvs?
22:31:53Learpatch tracker on sourceforge.
22:38:20LinusNam i right in that sample_count is only used to detect out-of-bounds seeking?
22:38:25LinusN(in mpa.c)
22:40:07Learsamplescount or sample_count?
22:40:07 Quit incognito ("CGI:IRC (EOF)")
22:40:21LinusNsorry, samplecount
22:41:32LinusNand it seems as if it is supposed to contain the number of samples remaining
22:41:43Learseems like it, yes. I'd have expected it to be used to detect eof...
22:41:44LinusNfor some reason
22:42:00 Join dcranford [0] (
22:42:10amiconnLear: I don't think my shared timer implementation has to do (directly) with the slow response bug
22:42:26dcranfordHi there, im trying to install bootloader 5 using fwpatcher, and it doesnt recognize the firmware
22:42:27amiconnIt only changed code which has absolutely nothing to do with i2c
22:42:41dcranfordBTW, this is on iriver h120
22:42:51amiconnThe only core feature affected is the backlight fading
22:43:02LinusNamiconn: i don't think it has anything to do with it either
22:43:23LinusNdcranford: which firmware are you patching?
22:43:34dcranford1.65 EN-US
22:43:53preglowi dont think samplecount has much of a function, out of bounds seeking can be detected in other ways
22:43:55Learwhen I tested it, the first build with the problem was the one that only contained the timer changes, so...
22:44:43Learand the slow response bug seems to be timing related somehow (make volume changes slower, and there's no problem).
22:44:49LinusNdcranford: so you browse to the newly downloaded ihp_120.hex and fwpatcher says it doesn't recognize it?
22:45:05 Join dpassen1 [0] (
22:45:19LinusNpreglow: i'll change it
22:45:43Learpreglow: sine waves to cause a glitch, yes, but in real music, I can't detect any gaps, and that's what counts, imho. :)
22:46:07preglowlear: they will always glitch, but i think theres a bigger gap than there should be
22:46:29LearI didn't hear much of a gap though.
22:46:50preglowlinusn: i think i orignally introduced it to know how many samples to play before discarding the samples that contain padding
22:47:37amiconnpreglow: I don't know exactly why that is, but Linus' fix for the id3v1 trimming fixed gapless playback for me
22:47:50dcranfordLinusN : that is my way of doing it....downloaded fwpatcher (July 22nd) from rockbox site....then download h120-bootloader.bin to desktop, then launch fwpatcher, browse to desktop, and fwpatcher doesnt see it
22:48:05amiconn...with the album that caused "negative gaps" before
22:48:15dcranfordLinusN : if I type in the name of the bootloader software I get that it is unrecognized firmware
22:48:15LinusNfwpatcher contains bootloader.bin
22:48:29LinusNdcranford: you need to download the original iriver firmware
22:49:04dcranfordoh I see, you download the iriver firmware and then the fwpatcher merges the 2 together?
22:49:10LinusNfwpatcher patches the original firmware with the boot loader
22:49:11amiconnLinusN: I hope to have correctly remembered the album that caused the illegal mem access a week ago; playing it now...
22:49:33amiconn...and yes, it already crashed
22:49:55dcranfordI see I thought it was like a bios update utility, where you download the utility then download the bin file and upload it to the player using that utility
22:50:03LinusNpreglow: ah, yes
22:50:14amiconn...within the first track (!)
22:50:28***Saving seen data "./dancer.seen"
22:51:09amiconnTrying again, watching it to find the exact spot...
22:52:11LinusNthe stop_skip is never handled
22:52:44dcranfordLinusN : ok the fwpatcher says it has patched successfully, do I now copy the file to the H120 and go through the firmware update process
22:53:18Leartrue, and wouldn't you need samplecount to detect when to do that?
22:53:25LinusNyes i would
22:53:51LinusNor i could detect that it is the last frame
22:54:08LinusNand only play a fragment of it
22:54:25preglowhow do you detect the last frame reliably?
22:54:28LinusNbut the decoder doesn't know
22:54:41amiconnLinusN: It happens at the very end of the first track
22:54:51LinusNamiconn: interesting
22:55:08amiconnI'll try whether this happens with other albums as well
22:55:42amiconnIt seems the first track has a length and bitrate that needs almost exactly 3 buffer refills to play completely
22:56:09dcranfordnow I suppose I need to get the rockbox software, where can I get the latest pre 2.5 software? Also I was wondering there hasnt been any updates to the testing of the H120 software since 7/20
22:56:24LinusNsummer time, you know
22:56:31LinusNgo to the daily builds page
22:57:00dcranfordLinusN : ok got this "pretty" stable ?
22:57:06amiconnLinusN: 192 kbps mp3 (cbr), length 3:31, with a personal rombox build on my Studio
22:57:07LinusNas stable as it gets
22:57:29LinusNamiconn: does it crash if you rolo a daily?
22:57:50amiconnI can try
22:58:03 Quit lostlogic (
22:58:12LinusN"rolo a daily" ... talk about rockbox lingo... :-)
22:58:34NJoinlostlogic [0] (
22:59:02 Quit Seed (Read error: 104 (Connection reset by peer))
22:59:18amiconnUsually I always run personal builds, especially on the Studio
22:59:25 Join Seed [0] (
22:59:34amiconn(You know, the HD poweroff issue...)
22:59:57amiconnMy Studio is working absolutely fine with it, but obviously not all players do
23:00:05LinusNah, yes
23:00:12LaMeDLinus, other VIP: It's nice that you'r here, because I just wanted to make a small comment and don't know where to place it exactly: about two month ago a patch has been accepted that makes the marker stay at the middle of the screen, only getting to the ends on start/end of lists. I wanted to put out that this is not implemented on playlists and menus. (!)
23:00:37BagderLaMeD: feel free to fix
23:01:00LaMeDI wish... ever heard of the disengajment in israel?
23:01:36leftrightwoman are difficult to get along with
23:02:13LaMeDsorry... disengagement. well i'm a soldier there. and trying to learn C out of .txt manuals i've put on my player
23:02:43dcranfordsweet, ogg is working gapless on this
23:03:35preglow.txt manuals and lots of practice will do just fine
23:03:53LaMeDI promiss I'll fix everything right after knowing C. now could any of you please just let this be known? I donno, put this on the right place on the requests?
23:04:18BagderLaMeD: I'm sure most of us know this
23:04:41Bagderbut adding a bug tracker entry about it could help reminding us
23:04:47LaMeDpreglow: could you please write me a C compiler for the H runtime? so I could practice when I'm away
23:05:34LaMeDBagder: yeah. I have no idea how to phrase that in as a bug tracker
23:05:51LaMeDso would you?
23:06:21leftrightanything else?
23:06:31LaMeDcoffee. thanks.
23:06:31LaMeDI'll b back when i'll know C. swear.
23:07:20dcranfordLinusN : are we going to get some equalizer settings?
23:07:26amiconnLinusN: It happens with a rolo'ed daily as well, only the address is different (obviously)
23:07:42LinusNdcranford: i guess so, eventually
23:07:48LinusNamiconn: good
23:07:58amiconnSo it has most likely nothing to do with a buffer wrap vs. track change
23:08:06LinusNhand me the files, and i'll fire up my gdb
23:08:13amiconnThen it should happen with other tracks as well, will try...
23:08:16preglowim going to work on an equalizer soonish
23:08:42dcranfordLinusN : just wondering, the bass settings / treble settings are ok I would just like more presets ......also the volume seems to be a bit lower than stock firmware? Is it just me?
23:09:00preglowdcranford: it is if you have volume at 100 and use the eq
23:09:47dcranfordpreglow : ah so the bass/treble settings bring down the volume?
23:10:47LinusNyou mean "rock" "jazz" etc?
23:10:54dcranfordLinusN : eq presets for "rock, jazz
23:11:10dcranfordLinusN : you got it
23:11:28 Join Strath [0] (
23:11:30preglowdcranford: the overall volume is reduced if there is a possibility for distortion, which there is with volume at 100 and additional eq gains
23:11:31dcranfordwhy do treble/bass settings bring down the volume?
23:11:39amiconndcranford: Boosting treble/bass brings down the volume if the boost can't be fulfilled otherwise
23:11:46dcranfordah I see
23:12:24 Part leftright
23:12:39amiconn...without causing really ugly distortions
23:12:54amiconnThe iriver firmware is cheating here; try the following
23:13:19 Quit LaMeD ("CGI:IRC (EOF)")
23:13:22amiconnIn the iriver firmware, set bass & treble to flat, the max out the volume
23:13:39amiconn(be sure to use the real max)
23:13:47amiconnThen try increasing bass or treble.
23:13:57amiconnYou won't hear the faintest change...
23:14:03dcranfordyeah it gets fugly fast
23:14:10amiconnThe iriver firmware caps treble/bass instead
23:14:52dcranfordwell, im using the crappy iriver stock phones for this good phones are at work
23:15:25amiconnLinusN: I just tested; it happens with all tracks, on all units
23:15:28preglowbut please, if anyone wants a new software based eq, id love to hear how people would like it to be
23:15:35preglowi dont use them myself, so i dont know how i want it
23:15:50amiconnJust start an arbitrary archos with a fairly recent build, enable memguard, then start playing an album
23:16:01amiconnIt will crash at the first track change...
23:16:03dcranfordpreglow : like I said an adjustable eq with at least 8 channels
23:16:11LinusNamiconn: ok
23:16:21dcranfordpreglow : then presets for rock,jazz,pop,classical
23:16:23preglow8 bands, you mean?
23:16:34dcranfordpreglow : yes bands (sorry)
23:16:37preglowand what do you mean by adjustable?
23:16:52preglowjust gain adjustable?
23:16:56preglowor q/center freq as well?
23:17:16dcranfordpreglow : well you have the presets right...then you have the ability to manually set the seq as well
23:17:27amiconnLinusN: Of course, use the zero area protection...
23:17:32LinusNhow silly, codec_set_elapsed_callback only updates the elapsed time if it is later than the current one
23:17:49preglownot necessarily, heh, i could store the coefs in a table to avoid having the coef calculation code in fixed point
23:18:06dcranfordsay the rock preset is a bit too bassy, you can just go into the manual eq and set it how you want it exactly
23:19:29 Quit dcranford ()
23:19:42Learlinusn: should always update, afaict...
23:20:08preglowthe gain of each band can be adjusted without recalculating the filter coefs, change in band placement needs recalc
23:20:22LinusNi don't understand the if()'s in that function
23:20:32LinusNwhy is it conditional?
23:21:14LinusNi understand the latency, but not the other condition, in the "else"
23:21:24Learsorry, didn't check the playback.c code... :)
23:22:39LinusNit only updates if the new elapsed time is either larger than the current, or at least 2ms smaller
23:23:27Learyeah, so the value doesn't change unless it is a "large" difference. don't see the need for that, I must say...
23:23:59*preglow looks forward to starting audio effect box plugin
23:24:08preglowand coding rockbox in general
23:26:17 Join thomjoha [0] (
23:26:25 Nick thomjoha is now known as prethom (
23:26:32prethompreglow: begone!
23:26:38 Quit preglow ("CGI:IRC")
23:26:54 Quit Seed (Nick collision from services.)
23:27:00 Join Seed [0] (
23:27:08prethomi wonder how using this shell as a dev box would work...
23:27:23prethomnot having a linux server of my own any more sucks madly
23:32:13 Quit webguest81 ("CGI:IRC (EOF)")
23:34:02amiconnLinusN: I'm just reading the part about the breakpoint hardware of the coldfire. Seems it is working much the same way as the ubc in the SH1...
23:37:00amiconnProgramming it is a bit different though, as it uses a special coldfire instruction instead of just a couple of memory-mapped registers
23:44:25amiconnIt shouldn't be difficult to replicate the archos memory guard functionality on iriver with that
23:44:56prethomdoes it eat much resources?
23:45:05amiconnThe only thing I need to know is at which address the ROM is mapped
23:45:17amiconn(RAM and IRAM are known)
23:46:28amiconnprethom: It doesn't eat any resources (apart from a bit of code space) until it triggers
23:47:42amiconnWell, perhaps the cpu draws a tiny bit more power
23:48:00LinusNamiconn: rom is at 0
23:48:06amiconnI don't know whether the breakpoint logic is switched off in case it is unused
23:49:19amiconnLinusN: Hmm, then it makes no sense to distinguish between illegal zero(-page) accesses and illegal ROM writes
23:49:54amiconnIt makes it impossible to declare zero(-area) reads illegal...
23:51:01LinusNwe don't use the rom today
23:51:03amiconnROM code will be read from ROM, and I strongly hope rockbox will run from ROM one day on iriver
23:51:25LinusNcode from rom? why?
23:52:23amiconnWhen rockbox will run from ROM, it will obviously be read from ROM, and will also read constant data from ROM
23:52:35LinusNget real. you want to sacrifice cpu performance to save a few hundred kbytes?
23:52:53LinusNi see absolutely no point in running from rom
23:52:57amiconnI do.
23:53:01amiconnBoot time.
23:53:25LinusNso you want to waste cpu performance for a faster boot time?
23:53:26prethomboot time is great as it is
23:53:27amiconnPerhaps we can copy from ROM, but that also requires reading
23:53:41LinusNyes, but then the memguard is off
23:53:42prethomwell, then enable the checking after copying
23:53:53amiconnprethom: Boot time of iriver rockbox takes ages compared to archos rockbox boot from ROM
23:54:09prethomyeah, id expect, but its still great
23:54:14amiconnLinusN: On archos, memguard survives rolo
23:54:21prethomwhats the best archos unit, btw?
23:54:26amiconnWe would need to switch off then
23:54:45amiconnprethom: I can't say there is a 'best' unit; it depends what you want
23:54:52Bagderprethom: a subject of preference, if you like the AA nimh or custom LiIo batteries basically
23:55:09amiconn...and what capacity, whether you want a radio etc
23:55:22prethomwell, radio as a feature is always good
23:55:28prethomcant see why i wouldnt want it
23:55:37amiconnThe only units I would avoid (if not for development) are the Studio and the recorder V2
23:55:51LinusNin my not-so-humble opinion, rombox on iriver is a waste of developer time and a waste of cpu performance
23:55:55prethomthink im getting my hands on an fm recorder, well see
23:56:01Bagderthere's only one archos model with FM and harddrive, the FM Recorder
23:56:15Bagderone model that rockbox runs on that is
23:56:16amiconn...but the FM recorder's radio is rather bad
23:56:35 Quit _aLF (Read error: 104 (Connection reset by peer))
23:56:38prethomi dont much care about it anyway
23:57:03 Join _aLF [0] (
23:57:09amiconnRecorder V1: No radio, but spdif in/out, runs from 4 AA NiMH cells
23:57:30prethomim not going to go out of my way to get an archos anyway, but if this girl i know wants to donate her fm recorder, ill just take it
23:57:37amiconnFM recorder: Bad radio, but no spdif in, runs from a custom LiIon battery, a bit lighter
23:58:16amiconnOndio FM: Good radio (newer models only), no spdif at all, flash memory, much lighter, runs from 3 AAA cells
23:58:25Learre radio, how much work is it in getting it working on the iriver? don't we have the "chip driver" already?
23:58:35amiconn(rechargeable or alkaline)
23:58:44amiconnLear: Yes we do

Previous day | Next day