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 2005-07-02

00:00:11MO-PantsuI tried your line on it's own. no difference
00:00:19HClamiconn: it should probably be possible to uncomment part of the unused tagdatabase support functions
00:00:33amiconnBagder: 3.3.1 -> 3.3.5 saves almost 1 KB ...
00:00:35 Join ze [0] (
00:00:39HCli'm planning to use them later in order to be able to add fileentries to the tagdatabase on the fly
00:00:43 Join p3druskus [0] (
00:00:59amiconnHCl: That might be helpful
00:01:07MO-Pantsudoes your next track scroll too with 'artist - title'?
00:01:10Coldtoastwith my line, it displays "unknown artist -" on teh first track til it gets past the frst 2MB barrier. Then it loads the correct info and fills it in. When the track changes, it changes to the new info
00:01:13Coldtoastand all scrolls
00:01:47HClgah. too many interesting programs on discovery channel
00:01:50Coldtoastyou know that lines like this:
00:01:53*HCl just turns it off before he starts to watch it
00:01:54amiconnI hope that [IDC]Dragon finishes bootbox testing some day. Then rombox shouldn't be a problem anymore on all archoses
00:01:56ColdtoastArtist. %s%?ia<%ia|%?d2<%d2|[\root]>> %?ig<(%ig)|%t0>
00:02:03Coldtoastscrolsl teh WHOLE line?
00:02:07HClwhats bootbox?
00:02:18Coldtoastyou don't get static "Artist" with scrolling artist name
00:02:30Coldtoastthe whole line including "Arttist" scrolls
00:02:48amiconnOn archos, we have a dual-boot bootloader also developed by [IDC]Dragon which allows to boot 2 different firmwares from flash
00:03:25amiconnAny of the 2 can be either compressed (then it needs to be uncompressed into ram before execution) or uncompressed and executed directly from rom
00:03:51Coldtoastbtw MO-Pantsu
00:04:00amiconnThe first image serves as a backup and is only started if you hold a button while booting
00:04:01Coldtoastyou can actually alter a few of your conditionals
00:04:18amiconnIn our current flash packages this first image is the archos firmware, compressed
00:04:26Coldtoastyou don't need things like ?%Ia<%a|>
00:04:34MO-PantsuI just need a line that works....sigh
00:04:38amiconnThe second image, which is started normally, is usually rockbox
00:04:57ColdtoastI mean ?Ia<%Ia|>
00:05:11Coldtoastjust use ?Ia<%Ia>
00:05:12amiconn'rombox' means uncompressed rockbox, executed directly from rom and hence leaving some more free ram for buffering etc
00:05:33Coldtoast"The else part is optional, so the "|" does not have to be
00:05:33Coldtoastspecified if no else part is desired.
00:05:39Coldtoastthat's from teh docs
00:05:40amiconnThe problem is that this isn't possible on all units due to space constraints
00:05:44MO-PantsuI have aq headache now
00:05:47amiconn(the rom is only 256KB)
00:05:59MO-Pantsucan't be bothered anymore
00:06:06MO-Pantsuif it's some weird bug it's evading me
00:06:54Coldtoastaccept the dcc
00:07:06MO-Pantsuit's not downloading for some reason hold on
00:07:08amiconnSo [IDC]Dragon works on replacing the alternate firmware image (the archos one) with a stripped-down rockbox which is only able to charge (for archoses which have software controlled charging), allow USB access, and boot from HD
00:07:09 Quit MO-Pantsu ()
00:07:11Coldtoastthis works on mine
00:07:16amiconn*This* is bootbox
00:07:22 Join Rori [0] (
00:07:34Roritry again
00:07:58Roriah just paste the damned thing
00:08:30Coldtoastthere you go
00:10:46amiconnHCl: Could you please disable the unused parts for now?
00:11:37 Join Stryke` [0] (
00:12:57amiconnThe unused functions are only for updating the tagdb on the target, or is there more?
00:13:27HClnope, thats it
00:13:36amiconntnx :)
00:13:48amiconnLet's see if that brings back my rombox....
00:13:57*HCl goes to shower least temporrily
00:17:39Coldtoastcan you add a parameter to .wps so you can display %ff in KHz as well?
00:18:24Bagderyes you can
00:18:59amiconnRombox isn't back :(
00:19:11amiconnOutput is 524 bytes larger than max (192496)
00:19:33Bagderthen hardeep's next-dir thing is probably those ~500 bytes
00:19:57amiconnMust be more
00:20:37HClruntime database is still a fair amount of extra code..
00:20:56amiconnI guess your gcc 3.3.1 will give ~1500 bytes too much, but rombox was possible before even with that gcc
00:21:20BagderOutput is 644 bytes larger than max (192496)
00:21:35amiconnUh? Interesting...
00:21:46amiconnWhy is that?
00:22:13amiconnI would expect a roughly proportional decrease
00:23:21amiconnI know that it would bring rombox back if I compile with -Os (and I know the fix to make it run with that -O option)
00:23:45amiconn...but I don't think that we want -Os
00:24:15ColdtoastBagder: did you say you CAN get %ff to display as KHz?
00:24:26Bagdersure, just write the code
00:24:50Coldtoastheh. if I were conversant in C, I'd love to
00:25:39 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
00:27:06Bagdergood idea
00:33:19 Quit spiralout (Read error: 131 (Connection reset by peer))
00:43:52 Join TCK [0] (
00:45:45 Join bug_ol_paul32432 [0] (
00:46:48bug_ol_paul32432what is the easiest way to reset ur iriver??
00:48:54 Quit bug_ol_paul32432 (Client Quit)
00:49:08RoriOK guys. Me and Coldtoast have been testing the next track thing and I have a weird problem
00:51:11Rorifor some reason I can't get next track artist to display when playing the first track. Ever. Not even if I wait until the end of the track. Yes the condition is in there as advised by the wps docs
00:51:38Roriyes there is an id3 tag and tried swapping id3v1 to v2 and back
00:52:06RoriONLY if I FF a little on the first track played does it then display the artist for the next track
00:52:33RoriIt always shows the title when it buffers part way through the first track but never the artist for the next track unless I FF
00:53:12Coldtoastyeah. on my player I get the info right near the end but when he does exactly the same thing I do, his doesn't do it
00:54:23Roriusing his wps btw
00:54:28Rorijust to make sure
00:54:33Roritried various tracks
00:54:59Rorithey have correct ID3 tag info. using clean latest build of rockbox. Odd eh?
00:55:30Rorithe fact it reads the tag for artist after I FF makes it even odder
00:55:53amiconnThe tag is read always
00:56:06Roriit's just not displaying if I let it play
00:56:17amiconnI think there is a bug in wps code that it doesn't detect the info changed and the display should be updated
00:56:18Rorionly when it gets to track 2 it then shows artist correctly for track 3
00:56:33Roriyep methinks so too
00:56:51Rorithat bug has been driving me crazy for 2 days
00:57:13Roriwondering why I could not get next track info to display correctly
00:57:29ColdtoastI better head to be. it's 9am
00:57:46 Quit Coldtoast ("Peace and Protection 4.22")
00:57:50Rorithanks for the help
00:58:04Roriat least I know it's not me being dumb
01:09:01 Join webguest72 [0] (
01:09:07 Join n0bby [0] (
01:17:15 Quit Rick (Read error: 104 (Connection reset by peer))
01:17:29 Quit Ancelot ()
01:18:11 Join Rick [0] (
01:18:49 Quit Harpy (Read error: 145 (Connection timed out))
01:22:40 Quit webguest72 ("CGI:IRC")
01:25:24 Join wlad [0] (wlad@
01:30:22Roriis the top line always off limits for wps?
01:46:55 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
01:46:56 Quit n0bby (Read error: 104 (Connection reset by peer))
01:47:32***Saving seen data "./dancer.seen"
01:55:09 Quit p3druskus ("leaving")
01:55:46 Part Moos
01:58:19 Quit hicks (Remote closed the connection)
02:01:05amiconnArgh! The runtimedb handling is bad bad bad! :(
02:01:21amiconnIt causes a spinup for *every* track change!
02:17:56 Join PaulJ_ [0] (
02:19:03 Quit PaulJ_ (Remote closed the connection)
02:19:09 Join amiconn_ [0] (
02:22:01thegeekshame on you HCl
02:23:15 Quit amiconn (Nick collision from services.)
02:23:15 Nick amiconn_ is now known as amiconn (
02:26:35*amiconn is severely annoyed by the runtimedb code
02:26:52*amiconn goes to sleep
02:29:21 Join n0bby [0] (
02:34:24 Quit courtc (Read error: 110 (Connection timed out))
02:35:03 Quit PaulJ (Read error: 113 (No route to host))
02:36:07wladanyone here know where does the damm ATJ2085 MCU boots on? (rom address)
02:48:28n0bbyi wouldnt anywat
02:48:33n0bby-t +y
02:50:35wladthere is absolutely no info on that chip ...
02:50:50wladjust a very imcomplete datasheet .....
02:52:58 Join kenshin_ [0] (
02:54:59 Nick wlad is now known as wladston (wlad@
02:57:28 Quit wladston ()
03:04:24 Join wacky_ [0] (
03:04:39wacky_hey guys, do you know a way to have warnings and errors highlighted in a bash shell, while running gcc ?
03:05:00kenshin_nope. never seen a way to do that.
03:05:48kenshin_but usually gcc just bombs on errors (making them obvious) and warnings are usually okay to skip
03:05:50wacky_I've seen that on a friend's computer, it was all highlighted.. and he didn't do anything really weird, it was by default on a Mandriva distro...
03:08:58wacky_there must be a package that tweaked /etc/profile.d in some way ?!
03:09:47kenshin_i've only seen highlighting on the bash prompt and ls
03:12:18kenshin_Mandriva may have customized bash to recognize gcc errors and warnings
03:13:10 Quit skel_ (Remote closed the connection)
03:13:34kenshin_and i don't see any gcc options to do this
03:15:50wacky_oh :) just installed gcc-colorgcc :)
03:15:55wacky_it's a perl wrapper that colorizes the output :)
03:19:04 Quit n0bby (Read error: 110 (Connection timed out))
03:24:30 Part wacky_
03:37:49 Join tvelocity [0] (
03:47:34***Saving seen data "./dancer.seen"
04:00:18 Quit Rick ("I don't need to be here.")
04:00:31 Join Rick [0] (
04:01:11 Quit Rick (Client Quit)
04:01:22 Join Rick [0] (
04:06:08 Join QT_ [0] (as@area51.users.madwifi)
04:09:01 Quit ShockerEngr (Read error: 145 (Connection timed out))
04:17:34 Quit QT (Read error: 113 (No route to host))
04:17:38 Quit kenshin_ (Read error: 110 (Connection timed out))
04:17:43 Quit xen` (Read error: 110 (Connection timed out))
04:20:10Rorianyone fancy tackling that wps updating display bug? :)
04:59:07 Join Leperkawn [0] (
05:02:42 Join n0bby [0] (
05:15:59 Join kenshin_ [0] (
05:17:08LeperkawnSo, what are your opinions on me getting a h3x0 through those free dealies and sending it to the rockbox devs?
05:24:26kenshin_the devs who need the 3x0 (for writing the bootloader) already have them
05:26:39LeperkawnWell, then.
05:26:45LeperkawnI will just have to sit by and wait.
05:32:00kenshin_have no fear, development is in progress
05:34:29 Quit n0bby (Read error: 110 (Connection timed out))
05:47:35***Saving seen data "./dancer.seen"
05:56:29 Join bagawk [0] (1000@bagawk.user)
05:59:08 Quit Leperkawn ("Chatzilla [Firefox 1.0.4/20050511]")
06:13:50 Quit kenshin_ (Read error: 110 (Connection timed out))
06:29:45 Quit bagawk ("Leaving")
06:43:13 Join cYmen_ [0] (
06:44:05 Quit cYmen (Read error: 104 (Connection reset by peer))
06:53:03 Join ashridah [0] (
07:03:01 Join ehntoo [0] (
07:39:12 Join gshgs [0] (
07:47:38***Saving seen data "./dancer.seen"
07:54:58 Quit Seed (Read error: 104 (Connection reset by peer))
07:56:49 Join Harpy [0] (
07:59:14 Join Seed [0] (
08:02:09 Quit Seed (Read error: 104 (Connection reset by peer))
08:06:09 Join Seed [0] (
08:06:34 Join tiegs [0] (
08:07:10 Quit tiegs (Client Quit)
08:13:51 Join AEnertia [0] (~aenertia@
08:27:06 Join StrathAFK [0] (
08:34:50 Quit Seed (Read error: 104 (Connection reset by peer))
08:41:44 Join Seed [0] (
08:45:08 Quit Thasp_ (Read error: 101 (Network is unreachable))
08:51:24 Quit Strath (Read error: 110 (Connection timed out))
08:56:29 Quit Harpy (Read error: 145 (Connection timed out))
08:58:57 Quit gshgs (Read error: 110 (Connection timed out))
09:03:31 Nick QT_ is now known as QT (as@area51.users.madwifi)
09:09:40 Nick StrathAFK is now known as Strath (
09:26:54 Quit ehntoo ("Leaving")
09:32:39 Quit pabs (Read error: 60 (Operation timed out))
09:47:40***Saving seen data "./dancer.seen"
10:08:56 Join einhirn [0] (
10:13:55 Join Thasp [0] (Thasp@
10:21:40 Quit TCK (Read error: 104 (Connection reset by peer))
10:30:34 Join webguest57 [0] (
10:32:42webguest57i have a question or rahter want your opinion
10:34:47webguest57what bmp wps do you think is the best\
10:50:06HClamiconn: ofcourse it causes a spinup for every track, it needs a binary search to locate the runtime record for the track, and people didn't want the database in ram, also, it needs to save the runtime database whenever it changes, cause it has to happen before the unit will be powered off
10:51:16amiconnI know *why* it happens with the current design, however, you should try to find a way to do without it
10:51:33HClhow? O.o.
10:51:34amiconn...and *not* holding the whole db in RAM
10:51:53amiconnThe current behaviour is inacceptable imho
10:52:21amiconnRockbox tries to save as much battery as possible by prebuffering, and the runtime database destroys this concept
10:52:24HClwell it could probably cache runtime data in the same way it buffers the audio
10:52:43amiconn...and it is not even deactivatable if a valid tag database is found :(
10:53:22amiconnThat's why I didn't further research the integration on archos although it seems that should be easy
10:53:28HCli have no idea how we would go around to caching the runtime info though
10:53:37HClmaybe the runtime info can be added to the id3 info?
10:53:53amiconnHmm, that sounds like a possible way
10:54:07HCliirc, on archos, it will still require it to save every time the data changes, cause it can shut down too fast?
10:54:35amiconnThe playback code (playback.c on iriver, mpeg.c on archos) stores id3 info and associated data for a number of tracks
10:55:00HClwell, you'd just have to move the loading to that, i guess the runtimedb code could put the runtime info into that (or even better imho, the playback code itself could update the runtime info field)
10:55:27HClbut it would still need to write to the rundatabase every time so that data isn't lost before shutting dwon
10:56:01amiconn...and then call the function to update the runtime database (mutliple entries) when it is rebuffering
10:56:14amiconn...or stopping
10:56:35HClwasn't there something with archos having a hardware turnoff thing?
10:56:48HCland rockbox not having time to do things before shutting down?
10:56:51HClor am i mistaken
10:56:51amiconnYes, but you should ignore that
10:57:24amiconnThere are 2 targets (archos player and recorder v1) which have a fast hardware shutdown (~ 1 sec)
10:57:45amiconnAll others have a much longer delay, so we do software shutdown as on the iriver
10:57:55amiconn(rec fm, rec v2, ondiops)
10:58:20amiconnThey still have a hardware shutdown, but that requires to hold the button > 10 seconds
10:58:24HClwell, my suggestion is then to add all the info the runtime database needs about a file to the id3 tag thing
10:58:35amiconnThink of it as the archos equivalent of the iriver reset button
10:58:55HCllike file entry info, current file entry offset, runtime info, current runtime entry offset..
10:59:27amiconnThe units with fast hardware shutdown also have a safe shutdown feature in rockbox, which takes care of flushing all info to disk
10:59:49SlasheriHmm, on certain situations id3 structs might get cleared before actual buffer refilling
11:00:03amiconnWe can't prevent the user from doing a hw shutdown, but if he does this, it is his own fault if some runtime info gets lost
11:00:51amiconnHCl: Also keep in mind that the hd spins up more often on archos due to the small RAM, so the potential loss isn't so big
11:00:58SlasheriHCl: for example with dynamic playlists a flush event causes that behaviour (buffering code clears everything but the current track_info struct) and disk spinup will follow later
11:01:23amiconn...and even on iriver there is a potential danger to loose some runtime data - when the battery gets too low
11:01:31*HCl is a bit confused, but isn't quite awake yet..
11:01:35amiconn...even with your current design
11:02:25HClthen we could probably just have an array of structures containing that info for each song next to the track info thing?
11:03:06amiconnSlasheri: The important point would be to call the runtimedb callback *before* any such event processing, be it a normal buffer refill or a flush due to skip, new track selected etc
11:03:21Slasheriamiconn: Hmm, that is possible
11:03:38amiconnOf course this menas the runtimedb code shouldn't take too long
11:03:40SlasheriBut i think the database should not flush to disk before the actual spin up is necessary
11:03:48HClit doesn't take long at the moment...
11:04:04HClthe binary search is pretty fast, much faster than i had expected
11:04:11amiconnThat's nice
11:04:46amiconnI tried to build a tagdb with your java songdb, however it seems this doesn't work with the runtimedb here (??)
11:05:02HClit should as long as the generated paths are okay
11:05:15HClthey should start with a / and follow with the absolute path to a file
11:05:16 Quit webguest57 ("CGI:IRC (EOF)")
11:05:28HClyou might need to grab the new songdb.jar from the wiki
11:05:29amiconnMay that be caused by the fact that I ran it under windows? The paths are displayed with backslashes....
11:05:40HClnah, i fixed that..
11:05:45HClyou probably have an oldish version
11:05:46amiconnI grabbed the latest one as of today ~1 am
11:06:04amiconnMaybe I used −−strip the wrong way?
11:06:08HClthe slashes should be / and i *know* i fixed that o.o.
11:06:18HCli dunno
11:06:20HCli run it with..
11:06:21amiconnDo I need to strip the drive letter?
11:06:38HClwell, it does that for you, but if you have −−strip, it should be added there
11:07:00amiconnOkay, perhaps I should try again...
11:07:12HCljava -jar SongDB.jar C:\windows\profiles\hcl\desktop\music −−strip c:\windows\profiles\hcl\desktop\music
11:07:15HClthat worked fine here
11:07:29amiconnI prefer to run it directly on the unit
11:07:35HClyea, same
11:07:35amiconnI used
11:07:38HClthat was just a test example
11:07:51amiconnjava -jar L:\ −−strip L:\
11:07:51HClfor my unit i use java -jar −−dirisalbum /
11:08:00HClyou can leave out the strip entirely
11:08:07HClit will strip the drive letter by default
11:08:23amiconnThe paths are printed with \ on the screen
11:08:33HClyea, thats okay
11:08:36HClas long as the final path
11:08:40HClin the end with longest filename
11:08:43HClis having / slashes
11:09:04HClits most probably the \ at the end of your strip
11:09:09HClremoving the first / slash
11:09:20HCli haven't added code to force paths starting with / yet
11:09:21amiconnI think it would be better to print the mangled paths instead of the source paths, at least for tracking down problems
11:09:48amiconn(taking into account −−strip, −−add and \ ==> / replacement
11:09:48HClwell.. the paths it shows are merely for showing, it doesn't process those since they are directories..
11:09:53HClbut okay
11:09:58amiconnAh, yes
11:10:39amiconnI think it would still be better to show them processed as well. I might not be the only one who gets confused...
11:10:53HClyea, i'll change that and also force a / at the beginning..
11:11:00amiconnI don't remember how the perl version does it
11:11:26 Join [IDC]Dragon [0] (
11:11:57amiconnIt *seemed* to me that the java version is slower, but I'm rather unsure about this
11:12:10amiconn(not the disk scanning step, but the sorting & output step)
11:12:17[IDC]Dragonhi there
11:12:36[IDC]Dragonbefore i gotta run off again:
11:12:52[IDC]Dragoni just made a BootBox wiki article
11:13:00[IDC]Dragonwith .ucl files
11:13:08amiconnFor testing?
11:13:15amiconnI.e. flash as secondary?
11:13:21[IDC]Dragonyes again
11:13:23amiconn(with rockbox_flash)
11:13:59amiconnAnything special to observe?
11:14:08[IDC]Dragonsee the text
11:14:54amiconnYou have tested on Ondio SP???
11:15:17[IDC]DragonFM, sorry
11:16:46HCluploading fixed version
11:16:59[IDC]Dragonis there a way to delete a wrong attachment? so far I've hidden it.
11:17:16HClamiconn: try the new songdb.jar..
11:18:37[IDC]Dragonbtw, what is this dynamic db?
11:18:48amiconn[IDC]Dragon: Should I do that (I guess it's th eplayer .ajz)?
11:19:04HClproof-of-concept quality, at the moment, but it will keep track of playcount and other things
11:19:18HClin order to be able to do some nice searches on music
11:19:30[IDC]Dragonamiconn: or educate me how to delete
11:19:38amiconnThere is no special delete action, to delete an attachment, you have to move it to the Trash web, topic TrashAttachment
11:20:18[IDC]Dragonhow can I move it?
11:20:44amiconnAttach -> (attachment to delete) -> action -> move attachment -> To: Trash topic: TrashAttachment
11:20:46[IDC]Dragonah, found it
11:22:21amiconnHCl: Thanks.
11:23:18amiconnHCl: Perhaps the runtime db should be switchable for now?
11:23:24[IDC]Dragonare all Rombox builds too big now (except Player)?
11:23:38amiconn(To avoid the spinup for those who want the tagdb, but not the runtimedb)
11:23:39HClamiconn: i would, but i'm not sure how to add a config option o.o
11:23:58amiconn[IDC]Dragon: Player and Ondio SP are still romboxable
11:23:59HClit shouldn't be hard to add, just a return -1 in the rundb_init depending on a variable
11:24:09amiconnHCl: I can do that
11:24:34HCli could also use quite some help on moving the runtime data so that its pre-cached, since i have 0 experience with that
11:24:46[IDC]DragonI've read about 2.5, is that prepared already?
11:24:56 Join hicks [0] (
11:25:41amiconnHCl: Unfortunately I can't do that now, as I'm still busy with graphics... That requires much more work than a config option
11:25:43[IDC]Dragonsorry for my noob kindof questions these days
11:25:47HClthats okay
11:26:08HClthe code isn't meant to be production quality yet, we needn't be in a hurry, the option to turn it off is a good idea though.
11:26:10amiconn[IDC]Dragon: The 2.5 snapshot is already taken (a week ago or two)
11:26:30HCli'm probably gonna look at getting the rating in the wps menu (?) that i haven't even discovered yet
11:26:32 Quit gromit` (Read error: 104 (Connection reset by peer))
11:26:36[IDC]Dragonso I should check out that label for BootBox
11:26:37HClafter i wake up
11:26:44HCland eat breakfast
11:27:19[IDC]Dragonmust have happend secretly
11:27:35 Join Chamois [0] (
11:27:39 Join gromit` [0] (~gromit`
11:27:46[IDC]Dragonso your latest gfx stuff is not in
11:29:29 Join PaulJ [0] (
11:29:39[IDC]Dragonok, I'm off and stop asking questions
11:29:54[IDC]Dragon...will stop
11:30:35 Quit [IDC]Dragon ("CGI:IRC")
11:31:06amiconnHCl: In 'General settings'->'Playback': 'Gather runtime data' yes/no
11:31:11amiconnWould that be acceptable
11:32:01HClsounds fine
11:32:02HCl :)
11:32:25HClmight want to add (experimental) to it for now
11:33:12amiconnOkay, will do that
11:33:27amiconnThis is only the language string which is easily changed later
11:33:41*HCl goes to buy breakfast and stuff..
11:47:43***Saving seen data "./dancer.seen"
11:52:01 Join preglow [0] (
12:00:26 Quit PaulJ (Read error: 145 (Connection timed out))
12:17:39amiconnHCl: When I'm thinking about it, caching the playtime info shouldn't need changes to the runtime db code, only to the playback code
12:17:56amiconnWell, almost
12:18:16 Join ep0ch [0] (~ep0ch@
12:18:33amiconnThe playback code would the simply call the callback not after each track, but only when it is rebufferung, and then it would call it for all 'passed' tracks
12:18:56amiconnThe id3 structure would then contain the actual playtime, which is a good thing imho
12:19:06amiconnNo need to keep track of it in the database code
12:19:18 Join [-AIR-] [0] (
12:19:23 Nick [-AIR-] is now known as west-acre (
12:19:25HClmhm, you just have to move the data it requires to the id3 structure
12:19:42amiconnThe data should already be there
12:19:47amiconn(I think)
12:20:00SlasheriHmm, that is a good idea
12:20:01amiconnThere should be an additional parameter which would tell the runtimedb whether it should fsync() or not
12:20:13amiconn(To avoid fsyncing everytime)
12:20:18HClwhat do you mean the data should already be there?
12:20:21SlasheriBefore every flush that data from played tracks can be passed to the database
12:20:26amiconnThis would then be true for the last call in each batch
12:20:28HClthe runtime database requires quite a bit of knowledge
12:20:58amiconnThe id3 structure contains path, length, and already played time
12:21:02HClit needs the offset of the fileentry of the file, the data that was in the file entry
12:21:11HClthe offset of the runtime entry of the file
12:21:21 Join Moos [0] (
12:21:23HClthe rating, playcount, and all the other runtime data
12:21:34HCland ofcourse the hashes of the runtime entry and the file entry
12:21:42amiconnHmm, but this is handled by the runtimedb itself???
12:22:00HClat the moment, yes, but there's only a single location where it stores this info
12:22:09HClso it would have to be moved to the id3 structure
12:22:15amiconnNo, why?
12:22:49amiconnThe way the runtimedb works won't change, the only thing that changes is the point in time when it is called
12:22:52HClbecause it would still require a harddisk spin up otherwise?
12:23:05HCli'm confused.
12:23:12HCl :/
12:23:16amiconnCurrently it is called after each track that finished playing
12:23:45amiconnWith the new concept it would be called by the playback code *when the playback code needs to rebuffer*
12:23:49HClyou're wanting to keep track of the playcount, and look the file up when you're planning to write to it?
12:23:52amiconn...which needs a spinup anyway
12:24:05Slasheriin future it can be called n times for each track info in buffer while we have to rebuffer / erase the track structure
12:24:15amiconn...and then it would be called *for all passed tracks* *in succession*
12:24:31HClyou can't do that, since it loads runtime db info
12:24:35HClwhen needed
12:24:40amiconnYes, and?
12:24:52HClyou can't call it in succession cause the data wouldn't change
12:25:06*HCl is confused x.x
12:25:29HClthe current code insists on getting called when a new song gets loaded
12:25:33HClyou can't call it at other locations
12:25:47HClprior to a song getting loaded it needs to load the runtime info of that song
12:25:52amiconnThat would not happen
12:26:08HCland at the end of it it needs to write the runtime info back after it has been altered
12:26:16amiconnHmm, why would it need the runtime info prior to loading?
12:26:28HClbecause it contains stuff like rating, previous playcount, volume adjustment
12:26:37amiconnImho it should be sufficient to do it all at the end (or even delayed further)
12:26:54HClno, we want playcount to be increased after 50% of the file has been played or so
12:26:59HCland volume adjustment to work, obviously
12:27:25amiconnWhy would you require to increase playcount at the extact time 50% have passed?
12:27:37HClnot exact, but 50% or 75%
12:27:41HClbecause at the moment its horrible
12:27:43HCleven when you skip tracks
12:27:44amiconnIt should be sufficient to bump playcount when we know that 50% have been played
12:27:46HClit still counts them as played
12:27:55HClyes matter when that info is passed to the db
12:28:05HClbut still. you NEED the info from the runtime database
12:28:18amiconnOkay, then this needs to be split in 2
12:28:19HClin order to be able to show playcount, change the rating properly and show the rating of the file
12:28:27HCland volume adjustment as well
12:28:33amiconnOne callback would update all data for the passed tracks
12:28:48amiconn...which would be called before rebuffering
12:28:59amiconn...and another callback for upcoming tracks
12:29:10amiconn...which would be called when buffering these tracks
12:29:23HClit would still be tricky, since you'd have to know which track is actually the current one
12:29:30HCland you'd have to store all the info in an array
12:29:42amiconn...and the relevant playback info (which would be volume adjustment only) would then be stored in the id3 struct
12:29:54HCli prefer rating and playcount as well
12:29:56amiconn...which already is in an array
12:29:58HClso you can actually show that on wps
12:30:05amiconnYes, okay
12:30:25HClbut yea, that would work
12:30:31HClas long as i can load the data into the id3 struct
12:30:37HCland fetch the data from the id3 struct to write it back
12:31:22HClpreferably by just calling a event_fetch_runtime_info(blah *id3)
12:31:23amiconnImagine a rebuffer event: 3 tracks have passed, and 2 new one get buffered
12:31:32amiconnThe following would happen:
12:31:34HCland an event_write_runtime_info(blah *id3)
12:31:50HCland just call that many times
12:32:09amiconn(spinup) -> call to runtimedb for passed track (3x) -> rebuffer -> call to runtimedb for upcoming track (2x) -> (spindown)
12:32:46HClas long as you use the two events i just described and i'm able to store the runtime data in the id3 tag
12:33:04amiconnyes, exactly
12:33:16HClpreferably, i'd like to store the fileentry and rundbentry in the id3 as well
12:33:26amiconnI think this would work pretty well; the wps already accesses the info in the id3 struct
12:33:33HClit would prevent doing 2 binary searches when 1 will do
12:33:41amiconnHow large are these? Just pointers?
12:33:45HCljust ints
12:33:52HCloffsets into the databases
12:34:04amiconnChange that to explicit longs if possible
12:34:12HClyea, i think they are..
12:34:37HClhmm, or not *changes*
12:36:00amiconnWe don't want to limit the databases to 64 KB on gmini, if that port will be finished some time, i think? ;)
12:36:18HClhm? are they limited to 64kb?
12:37:12amiconnIf you use ints, then yes
12:37:16HClah :P
12:37:18amiconnGmini has a 16 bit cpu
12:37:23HCl :P
12:37:35HClwell, changed it, i'll go over the code later to see if i missed some
12:37:53amiconnHmm. I do have that option prepared here, currently testing...
12:38:14amiconnIs there a method to display the runtimedb contents?
12:38:30HCla crude one... testdbv2.c in tools/
12:38:41HClit requires you to set all the defines in the source to the values it spits out
12:38:46HCland recompile
12:38:49HClbefore it will work
12:38:51amiconnHmm, my rockbox.rundb is still 8 bytes :(
12:39:05HClwell, not much contents to display, then
12:39:07HCl :x
12:39:16amiconnI did play a number of tracks...
12:39:20HClRRD1,0 is what it should contain
12:39:26HCldid you turn the option thing on?
12:39:47HClthe remote should print out some rundb info
12:40:03HClat the moment it ignores files with a hash of 0
12:40:22amiconnYes I turned it on
12:40:30HCland it ignores files on which it fails to find a fileentry in the tagdatabase
12:40:52amiconnI have a database generated with your latest java songdb, which contains all files on my iriver
12:41:13HClwell, i guess you can toss me your database and i can check it briefly with testdbv2
12:41:27HClpretty much the thing you need to look for is whether the fileentries seem to have the correct path
12:41:43HClyou could also add a logf to print out the results of the binary search
12:41:48amiconnI'll check whether the option is working
12:41:58amiconn(that's possible even without getting entries)
12:42:02amiconnThen commit it
12:42:46amiconnIf I delete the rundb, the reboot, the file should be recreated when the option is on
12:42:53amiconn...otherwise not
12:44:10amiconnI also added proper handling for the usb case (like with the tagdb)
12:45:59amiconnThe volume adjustment has to be done by the playback code anyway, so the id3 struct should be the best place to keep it
12:47:25preglowabout time to start renaming the id3 struct, perhaps? :)
12:47:42amiconnIt's called mp3entry now
12:47:49preglownot much better
12:48:10preglowdoesn't matter much, but is misleading now that we support a bunch of codecs
12:49:26amiconnHCl: Committed.
12:49:27 Join Coldtoast [0] (
12:53:01HClk :)
12:53:12*HCl needs more sleep :/
12:53:32HClshould i work on implementing those two event things when i'm awake?
12:53:44HCli just need someone to make the event interface much like the old one
12:54:13amiconnI think this would be cool
12:55:17amiconnI mean, that it would not need to spinup after each track anymore
12:55:25amiconnI just have to find out why my database refuses to work, then I could add the event handling for archos
12:55:33HClk :)
12:55:47*HCl prods Slasheri a bi
12:55:51amiconnI already tried, but I thought something was wrong because my runtimedb didn't get populated
12:56:19HClas long as the option is enabled and the tagdatabase is loaded, there are two options left to why it could go wrong
12:56:21amiconnCould be that just my tagdb was wrong...
12:56:24HCl1) the hash of the file is 0
12:56:28HCl2) the binary search failed
12:56:47HClthe binary search fails if the tagdatabase has wrong filename entries
12:56:47 Join Christi-S [0] (
12:57:06amiconnhi Christi
12:57:16*HCl goes to nap a bit :/
12:57:21SlasheriHCl: Hmm :)
12:57:30amiconnHCl: The has shouldn't be 0 with the java version, right?
12:57:31 Quit ep0ch ("goto")
12:57:41HClSlasheri: do you think you could implement the two events that were proposed?
12:57:51HClamiconn: nope, java version should be hashing
12:57:56HClyou can try the testdbv2 tool
12:58:03HClit allows raw dumps of the database
12:58:09SlasheriHCl: that doesn't sound a good thing. If you mean the start-of-track and end-of-track events?
12:58:19amiconnPerhaps I will... I have to decide between 4-grey coding and database adaption...
12:58:28HClmmm, ask amiconn for details :/
12:58:51amiconnSlasheri: No start-of-track and end-of-track when they actually happen
12:58:55HCli'm not 100% sure what i should call the events, but pretty much, there's one that gets called repeatedly for every id3 entry that needs to be saved
12:59:08HCland one that gets called repeatedly for every id3 entry that needs runtime info added to it
12:59:24*HCl goes for a nap now :/
12:59:31Slasheriamiconn: Hmm, only start-of-track and end-of-track when codec starts / ends?
12:59:33amiconn...but 'upcoming track' for each track that gets buffered, and 'passed track' for each track that has passed, *at the time the rebuffering takes place*
13:00 completey avoid additional spinups
13:00:14Slasherithat might work
13:00:32amiconnI didn't look at how playback.c handles it, but I know a bit how mpeg.c handles it
13:00:40amiconnIt should be quite simple
13:01:00Slasheriso when we are buffering new tracks, generate an event when we start buffering a new track?
13:01:07amiconnmpeg.c has an array (16 entries) to keep info about all loaded tracks
13:01:32amiconnIt has a read pointer and a write pointer and is used as a ring buffer
13:01:43Slashericurrently playback.c generates the event at exact time (with audio buffer latency taken in account) when track change
13:02:07amiconnThe read pointer advances at each track change, and the write pointer advances advances for each newly buffered track
13:02:19Slasheriyep, playback.c has that ring structure also
13:02:30amiconnFor my proposal to work, we would need a third pointer, 'trailing read'
13:03:14amiconnThe buffering code would call the 'new track' callback each time it advances the write pointer, which happens while buffering
13:03:44Slasherithat sounds possible
13:03:53amiconnImmediately before buffering new tracks, it would count the 'trailing read' pointer up to the read pointer, and call the 'passed track' callback for each of these
13:04:34amiconnSo n times callback for passed tracks, then rebuffer and call new track callback m times
13:04:36SlasheriHmm, is there a need for trailing read? I guess no because we know the buffered track count
13:05:07amiconnI don't know how playback.c does that, but the read pointer in mpeg.c advances at the time of a track change keep track of current info for wps
13:05:38amiconn...but we need to tell the runtimedb about all tracks that have passed
13:06:11Slasheriyep, that is possible with the current implementation without any new pointers
13:06:24amiconnOkay, so you have 3 pointers?
13:06:57amiconnOr you keep track about the write position with th buffered tracks count?
13:06:58Slasheriread_index, write_index and track_count. These are enough to find the passed tracks
13:07:08amiconnAh, yes
13:07:44 Join Hansmaulwurf [0] (
13:08:24Slasherii can implement that event :)
13:08:40amiconnmpeg.c does that a little different afair, but that shouldn't be a problem
13:09:02amiconnmpeg.c doesn't keep the track_count separately iirc
13:11:05 Quit ashridah (Read error: 110 (Connection timed out))
13:11:21amiconnI think these 2 new event could be added while the current one should stay as well
13:11:35amiconnThen HCl could hook the runtimedb to these 2 new callbacks
13:11:40Slasheriof course, i think it that way too :)
13:11:53amiconn...and finally the current trackchange callback can go away
13:12:26amiconnThe 'new track' callback should have an additional parameter
13:12:40amiconn...telling the runtimedb whether this is the last call in the batch
13:12:54 Join ashridah [0] (
13:12:56Slasheriah, ok
13:13:01amiconnThis way it only needs to fsync() at the last call
13:13:40amiconn...and the ata_sleep() in the playback code should go after these calls
13:13:42 Quit edx (Read error: 110 (Connection timed out))
13:15:15HClsounds good :)
13:15:56 Join PaulJ [0] (
13:18:14 Quit Christi-S ("If I were actually witty, this quitline would be funny.")
13:21:27amiconnSlasheri: Just a detail: It is perfectly possible that there are no callbacks at all when rebuffering, in case we're playing a very long track that was already running at the last rebuffer event, and will still play at the next rebuffer event
13:22:01amiconnMuch less likely on iriver than on archos, but still possible even with mp3
13:22:11Slasheriamiconn: yes, of course :)
13:22:19amiconnI had some 100 MB mp3 track....
13:22:46Slashericallbacks will be called only when track really begins/ends
13:23:04amiconnAnd also, the 'new track' callback must be called at the initial buffer filling, and the 'passed track' callbacks at stop
13:23:27amiconn...but I think you got the point :)
13:23:32Slasheriyep :)
13:23:49Slasherii will try to test that these special cases works also
13:26:08amiconnHow large is your track info array?
13:26:51amiconnOn archos we use 16 even with the small RAM. Some users managed to exceed that, playing tiny snippets from a language learning CD...
13:27:29SlasheriHmm, i haven't measured it but it might be several thousands bytes long because we cache 10 id3 and mp3data structures. However, it should be enough to pass the mp3entry structure only to the database
13:27:32 Quit tvelocity ("Leaving")
13:27:52amiconnYes, I mean the number of entries
13:27:55amiconn...only 10 :/
13:28:00SlasheriAh, yes 10
13:28:09SlasheriThat can be changed with the MAX_TRACK define
13:28:32Slasherihowever, i have never had a situation that even 10 tracks in buffer at a time ;)
13:28:54amiconnI can imagine that happen with some albums I have
13:29:22Slasheriok, that number should be increased if necessary
13:31:23amiconnI just checked. D.Trance Gold CD1, encoded with lame @192 kbps
13:31:31amiconnThe first 12 tracks are ~27 MB
13:33:55amiconnI noticed some strange effect with the backlight fading. It still flickers if there's a lot of bitmap drawing, but I don't understand why.
13:34:09amiconnTry snow.rock and logo.rock, you'll see what I mean...
13:35:14Slasheriyes, i know that.. I think some interrupts might override the timer2 interrupt
13:36:36Slasherior maybe the interrupts are disabled while updating display
13:36:38amiconnHmm, now I can't make it happen
13:36:57Slasherijust try to write bunch of text to the remote lcd with logf
13:37:06Slasherithat will cause that undesired effect always
13:37:35amiconnAh, now I understand...
13:37:59amiconnIt does happen with snow.rock and logo.rock, because these draw on the remote as well
13:38:07amiconn...but only if the remote is plugged!
13:38:22Slasheri:D so the problem is with the remote display code
13:38:35amiconnRather, with the remote lcd data transfer
13:39:40amiconnI know why
13:39:48Slasherithat's great :)
13:40:08amiconnThe port handling of the remote lcd transfer interferes with the port handling in the backlight dimming interrupt
13:40:23amiconnThat's tricky...
13:41:12amiconnThat's because the port set/reset is non-atomic...
13:41:39*amiconn checks the coldfire manual
13:42:05amiconnWe might need an equivalent to the SH1 read-modify-write AND and OR
13:43:36amiconn..and it seems it is possible :)
13:44:16 Join Febs [0] (
13:45:06amiconnThat would be an equivalent to the and_b / or_b / xor_b macros for SH1
13:47:09amiconnThe problem is that statements like GPIO1_OUT &= ~0x00000004; are non-atomic
13:47:46***Saving seen data "./dancer.seen"
13:48:13SlasheriHmm, yes. So it should be possible to make it atomic with some assembly code?
13:48:28amiconnso on archos, we have and_b(~0x04, &PADRL); as an equivalent (obviously for byte witdh only)
13:48:40amiconnThis is actually a small assembly macro
13:48:50amiconn...and that should be possible on coldfire as well
13:48:56Slasherigood :)
13:49:32amiconn...with something like and.l %dn,(%am)
13:50:31amiconnThis has to be used for all port operations on ports that could be potentially changed in an interrupt as well
13:50:43amiconn...which are practically all on iriver :/
13:50:59amiconnNot too bad, but quite a number of changes
13:51:37SlasheriThat's true.. but if it's possible to do it with some nice macro it shouldn't be so bad
13:52:37amiconnYeps. Look at firmware/export/system.h lines 82 ff how this is done for SH1
13:53:20amiconnWe would then have and_l(), or_l() and xor_l()
13:53:22Slasheriah, that seems good :)
13:53:46amiconn..even a bit more flexible since both the address and the data can be variables on coldfire
13:56:21 Join bipak [0] (
14:04:48amiconnw00t, that actually works :))
14:05:04Slashericool =)
14:05:39amiconn..and it decreases code size a bit
14:05:49Slasherinice :D
14:13:23amiconnI converted the whole remote lcd driver and backlight.c for now. All other places doing port manipulation should probably be changed as well
14:14:24 Quit bipak_ (Read error: 110 (Connection timed out))
14:19:02amiconnThe flickering fade is no more :)
14:45:29HClany progress on the events thing or should i just go and look at implementing rating?
14:45:34HClhow do i access the wps context menu?
14:47:24SlasheriHCl: I will implement it soon (now working with mono playback)
14:47:34HClk :)
14:52:13 Quit Chamois ("Leaving")
14:55:02 Join edx [0] (
14:58:49HClwho can tell me about the wps menu?
15:00:29*HCl thinks he got it
15:04:13 Quit ze (Read error: 104 (Connection reset by peer))
15:04:17 Join ze [0] (
15:14:41HCli'm bored :/
15:15:00preglowand i'm out
15:15:08preglowsee you guys some day in the not too distant future, i hope
15:15:14 Join tvelocity [0] (
15:15:21preglowif i can stand 28k8 it might even be pretty soon!
15:15:26 Quit preglow ("leaving")
15:17:30*HCl prods amiconn
15:17:33HClany luck?
15:34:29amiconnSorry, no. Didn't investigate further for now
15:34:34amiconnGotta go, bbl
15:34:39 Part amiconn
15:36:22 Join xen` [0] (
15:36:40xen`Is there anyone on the .mod playing ?
15:36:54 Quit tvelocity ("Leaving")
15:36:55xen`I'll probably take a look about if it's possible
15:37:16xen`(if nobody on it)
15:47:49***Saving seen data "./dancer.seen"
15:58:59 Join tvelocity [0] (
15:59:10tvelocitya ETA i antigrafi!
15:59:17tvelocitymia wra*
15:59:19tvelocityoops wrong channel
15:59:30tvelocitydamn xchat focus
15:59:33 Quit AEnertia (Client Quit)
16:07:49 Quit xen` (Read error: 145 (Connection timed out))
16:14:29 Join kenshin [0] (
16:19:10kenshinanyone know when t0mas will be back?
17:22:13 Join Godeater [0] (
17:23:33 Quit Godeater (Client Quit)
17:29:15 Quit tvelocity ("Leaving")
17:37:27 Join DangerousDan [0] (~Miranda@
17:39:42 Quit kenshin (Read error: 110 (Connection timed out))
17:47:53***Saving seen data "./dancer.seen"
17:48:06*HCl prods around
17:48:14HClSlasheri / amiconn?
17:49:10SlasheriHCl: you will get it soon :)
17:50:49 Join bagawk [0] (1000@bagawk.user)
17:51:35 Join pabs [0] (
17:52:21HClk :)
17:52:26*HCl is falling asleep here
18:04:36 Join muesli- [0] (
18:09:56 Join courtc [0] (
18:13:19 Quit edx ()
18:18:21 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
18:19:12 Join kenshin [0] (
18:19:43 Join Lost-ash [0] (
18:22:42 Quit ashridah (Read error: 110 (Connection timed out))
18:23:44 Quit hicks (Remote closed the connection)
18:24:07 Quit muesli- ("ich will Khe!!!")
18:26:25 Join hicks [0] (
18:34:19SlasheriHCl: Done, now testing
18:36:20 Nick Lost-ash is now known as ashridah (
18:36:34 Quit ashridah ("Leaving")
18:39:21 Join webguest91 [0] (
18:41:41 Join XavierGr [0] (
18:49:48SlasheriHCl: it should work now, i will commit
18:51:45XavierGrdoes anyone know what HAVE_LCD_CHARCELLS define means?
18:53:36 Quit bagawk ("Leaving")
18:55:06 Join amiconn [0] (
18:57:24XavierGrhi amiconn
18:58:48XavierGramiconn: I see that while there are scroll configuration variables in lcd-h100-remote.c there are not present to configure in settings.c and settings-menu.c
18:59:30XavierGrI tried to add them but I got lost...
18:59:31amiconnNobody did this so far because the remote isn't really used with scrolling yet
19:00:34XavierGrdoes anyone know what HAVE_LCD_CHARCELLS define means?
19:00:37amiconnThere is some oddity too which I don't understand why Rick did it
19:00:58amiconnThe remote scrolling thread isn't created for the simulator. Imho it should
19:01:31amiconnYes, that's simple, and should be obvious. It means the target has a character cell based display, as opposed to bitmap based
19:01:31XavierGryeah all my attempts to test code is making a normal build to test in my player it is very tiresome
19:01:58XavierGrlike seven segment displays?
19:02:16amiconnThe only device that has this is the archos player/studio
19:02:23XavierGroh I got it
19:02:33amiconnIt has a 2 line x 11 character, 5x7 point dot matrix display
19:02:55amiconnHowever, you can't freely access the single dots
19:03:32amiconnThe display ram contains character values, and the 5x7 dot matrices are taken from a character generator, which is mostly in ROM
19:03:42XavierGrso player/studio is very limited in text output I guess...
19:04:08amiconnIt's still possible to support many rockbox features on that limited display
19:04:38amiconnDepending on the exact lcd type (there are old and new players), 4 or 8 characters are user definable
19:05:25amiconnThis is what the playergfx library exploits - it allows (tiny) bitmap grahpics on this display!
19:07:26XavierGrso what about the remote thread you proposed. Do you think that it will have files like remote-wps.c and remote-tree.c e.t.c
19:07:38XavierGror it will be implemented in current code?
19:11:01XavierGrany plans for the remote uisimulator on win32 yet?
19:11:55 Quit DangerousDan (Read error: 131 (Connection reset by peer))
19:12:41 Join Ancelot [0] (
19:16:09 Join Stryke` [0] (
19:16:45 Quit Ancelot (Client Quit)
19:27:41 Quit amiconn (Read error: 110 (Connection timed out))
19:32:17 Join amiconn [0] (
19:35:20 Join road_runner [0] (
19:36:37road_runnerQuestion: I found something very useful at sourceforge and there's no apperent reason why RockBox wouldn't include that patch on their build. where's the place to suggest to use it?
19:37:21Nibblerwell, there is some patch-area at i think
19:38:36road_runnerthanks i'll check it out.
19:40:48 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
19:40:55 Join Stryke` [0] (
19:47:39 Quit courtc (Read error: 110 (Connection timed out))
19:47:57***Saving seen data "./dancer.seen"
19:48:27 Join courtc [0] (
19:48:59 Part amiconn
19:56:02 Quit Stryke` (Read error: 145 (Connection timed out))
20:08:35 Quit XavierGr ()
20:13:49 Join Bger_cgiirc [0] (
20:14:24Bger_cgiircBagder / B4gder : about Li-Ion batteries... it seems you
20:14:29Bger_cgiircare right
20:15:47Bger_cgiircsome quotes
20:15:50Bger_cgiirc"Another easy way to destroy an Li-Ion battery is by discharging it too far."
20:16:06Bger_cgiirc"The Li-Ion cell should never be allowed to drop below about 2.4V, or an internal chemical reaction will occur where one of the battery electrodes can oxidize (corrode) through a process which can not be reversed by recharging."
20:16:43Bger_cgiirc"The Li-Ion typically is discharged to 3V/cell. The lowest low-voltage power cutoff is 2.5V/cell."
20:17:05Bger_cgiirc"During prolonged storage, however, a discharge below this voltage level is possible. Manufacturers recommend a trickle charge to raise such a battery gradually back up into the acceptable voltage window. Not all chargers are designed to apply a charge once a Li-Ion battery has dipped below 2.5V/cell."
20:17:11 Quit Hansmaulwurf ("( :: NoNameScript 3.81 :: )")
20:18:11Bger_cgiircsorry for the flood...
20:26:09 Join CheeseBurgerMan [0] (~Dan@
20:38:01 Join Yokaloshi [0] (
20:38:51 Join PaulJ_ [0] (
20:39:01Yokaloshiguys, since #linux + #debian are totally useless i have come seeking someones advice cos u ppl are in the know.....
20:39:15Yokaloshiwhich version of debian should be used with an amd-k6?: [alpha] [arm] [hppa] [i386] [ia64] [m68k] [mips] [mipsel] [powerpc] [sparc]
20:40:17Yokaloshior am i wrong and none of you have the slightest idea?
20:41:59 Quit CheeseBurgerMan (Client Quit)
20:47:45 Join CheeseBurgerMan [0] (~Dan@
20:47:58Bger_cgiircyokaloshi: i386
20:49:55 Quit hicks (Remote closed the connection)
20:51:26Yokaloshiwhat kernal?
20:51:43Slasherijust use the latest unstable or stable version
20:51:49Yokaloshioki doki
20:51:51SlasheriAnd always newest kernel, 2.6.xx should be ok
20:52:14Yokaloshi2.6 on the debian disk didnt work
20:52:19Yokaloshii'll try something older
20:52:22 Join hicks [0] (
20:52:35 Quit west-acre ("I-n-v-i-s-i-o-n 2.0 Build 3515")
20:54:11 Quit Yokaloshi ()
20:55:17 Quit PaulJ (Read error: 113 (No route to host))
20:59:41SlasheriHmm, the voice ui seems quite easy to port
21:01:57MoosHi Slasheri, not needed a new codec like speex?
21:13:44SlasheriMoos: We can implement the new codec later. If i understood correctly we should modify the codec loader to support two codecs loaded at a time
21:14:05SlasheriSo we could have the voice codec always loaded & ready
21:16:09Moosa ok, goody :)
21:17:19Bger_cgiirchm, what do you think about the idea to put rockbox.iriver in .rockbox folder ?
21:17:33Bger_cgiircin *the* rockbox folder...
21:18:02Bger_cgiirci know that it's a must for the archos devices to be in the root folder
21:19:07Bger_cgiircbut this comes from being compatible with original archos firmware
21:20:02Bger_cgiirchowever, iriver's firmware never had posibility to load the firmware from the disk
21:22:23Bger_cgiircyes, i know that for making such change it's needed to change the bootloader
21:31:16 Quit Thasp ()
21:48:00***Saving seen data "./dancer.seen"
21:48:46road_runnerImho, would clean up some mess, not having files running around at your Root folder. specially when it's out in the public.
21:50:30Bger_cgiircyes, that's my reason to offer this
21:55:33ghode|afkum its only one file...
21:57:06 Quit kenshin (Read error: 145 (Connection timed out))
21:57:42Bger_cgiircbut don't you think that all rockbox specific stuff is good to be in the .rockbox directory ?
21:59:13hicksEither that or modify the rockbox file list to hide rockbox.iriver when normally viewing
22:01:15Bger_cgiirchicks, when you hide it, it's still there
22:02:36hicksyes, but is there a major problem with it been there if you can't see it? kinda like the .rockbox folder is there but not visible?
22:04:13road_runnerThat's just a part of the problem. the other is you don't want no one to exidently delete the file. if it's inside a directory it's less likely to happend.
22:04:13Bger_cgiircnot the same when you attach the unit to a PC
22:04:29hicksyeah thats true
22:04:56Bger_cgiircthis too
22:07:00 Quit road_runner ("CGI:IRC (EOF)")
22:23:58 Quit Bger_cgiirc ("CGI:IRC 0.5.4 (2004/01/29)")
23:02:12 Quit Febs ("Chatzilla [Firefox 1.0.4/20050511]")
23:02:34 Join stoffel [0] (
23:07:02 Join Musicmad [0] (
23:07:02 Join PaulJ [0] (
23:09:31RoriI found a workaround for that 'Next track' not displaying bug
23:10:03RoriSomeone suggested using t0; in front which forces a refresh to display after buffering
23:10:08 Join noel_sad_sogn [0] (
23:11:10noel_sad_sognHey.... anybody there?
23:11:11RoriI don't see why wps can't pull the ID3 tag info without having to buffer the next track first
23:11:24BagderRori: then write a patch
23:11:37RoriShow me how ;)
23:11:46Coldtoastthat's your fave response Bagder :_
23:11:47Bagderquestioning functions when you don't know how things work is odd
23:11:54RoriI'm not complaining just pointing it out heh
23:12:04BagderColdtoast: yes it is! ;-)
23:12:21BagderRori: of course we would've done that if it was that simple
23:12:42RoriI look from laymans point of view. If there is a technical reason not much else I can do or say as long as someone tells me that is the reason
23:12:58Bagderwell _everything_ more or less is possible
23:13:05Bagdergiven enough time and sweat
23:13:09Roriof course
23:13:21Coldtoastwhere's the ID3 info stored in an MP3?
23:13:23Roribut some things have limitations due to hardware
23:13:33BagderColdtoast: in the begining or the end of the file
23:13:51Coldtoastbeginning OR end? how odd
23:13:59Bagderv1 vs v2 id3
23:14:02Roriso you have to read the entire file to read the tags?
23:14:07Coldtoastbeginning or end depending on what version of ID3 it is I assume?
23:14:07 Join kenshin [0] (
23:14:11Bagderyou can seek to the end
23:14:17BagderColdtoast: yes
23:15:49RoriDoes it seem like every time I type something here it's a whinge? If so I apologise. I'm just speaking my mind without knowing much really.
23:16:15RoriTake it with a pinch of salt ;)
23:16:18Bagderyou say things "I don't see why..."
23:16:28Bagderwhich sounds as if you question things
23:16:59Bagderat least in my ears
23:17:00RoriWell I don't because I don't know much so I say it without actually asking directly. Maybe I should just ask instead of putting it in such a way as it seems like a whinge :)
23:17:36Coldtoastthink I'll go watch War of the Worlds when I wake up
23:17:45RoriNothing wrong with questioning things. But I guess it is how it comes across that counts.
23:17:50RoriI saw it
23:17:54Roriwell half of it
23:18:04RoriI was in the loo puking up a lot whilst at the cinema
23:18:13Coldtoastis it THAT bad?!?
23:18:17Rorisomething I ate did not agree with me
23:18:25Roriyeah it was THAT bad ;)
23:18:38RoriI saw Tom Cruise' acting and it made me vomit
23:18:46ColdtoastI actually like Tom now
23:18:56RoriIt was an OK movie from what I saw of it
23:19:17Coldtoastnever really used to but he's done some films I LOVE in recent years; Minority Report, Vanilla Sky, Last Samurai
23:19:26Roriwas lacking a little from a modern day spin on the story I thought. It stuck too close to the book which is a bit out of date now.
23:19:41RoriBut the fighting machines were awesome!
23:19:46Coldtoastthen I saw the full vid of the guy squirting him and how he reacted and I now really like him
23:20:09Roriand the sound they make will make you shat your pants
23:20:19Coldtoastnuff said
23:20:29Coldtoasthope so
23:20:43Coldtoastnot expecting it to be as enjoyable as Batman Begins but I hope I like it
23:20:52Roriit's enjoyable
23:20:55ColdtoastBatman Begins is absolutely AWESOME
23:21:06Roridon't try to look too close at the plotholes
23:21:28RoriBatman Begins had some plotholes but I looked past them
23:21:41RoriBecause the acting and ideas in it were great
23:22:17Coldtoasttiem to get my player and update it
23:22:22RoriI want to see wotw again to catch the bits I missed while in the toilet chucking my guts up
23:22:34Roridownloading a cam
23:22:43RoriI think I have that right since I paid to see it anyhow ;)
23:22:58Roriand it will get deleted afterwards since the quality sucks
23:23:18RoriI will just flick to the bits I missed
23:23:54Coldtoasteh... no comment
23:25:21 Quit PaulJ_ (Read error: 110 (Connection timed out))
23:28:47Coldtoastthere's still a prob with mono MP3s
23:29:12ColdtoastI get L+R fine, which is gret, but I also get a constant ticking thru them
23:29:43ColdtoastBagder: if you play that MP3 I found from the other day, you should hear it
23:30:40Coldtoasthold on
23:30:55Coldtoastit's the resampling I guess. the ones that tick are 22KHz
23:31:24Coldtoastthe 64kbit, 44.1KHz ones are clean
23:39:01 Quit Bagder ("Off to search for that connect-resetting peer guy!")
23:40:02 Join Bagder [0] (
23:42:52dionoeahmmm ... the simulators seem to mess up the gnome control panel ... thats weird :)
23:43:08Bagdernot for me
23:43:19dionoeait doesn't do it every time
23:43:26 Join ghostiger [0] (~ghostiger@fdf84575d351d3c3.session.tor)
23:43:26Bagderit never did it for me
23:43:34dionoeait just happened twice since i've used it... really weird
23:43:37Bagderand I've used the sim on x11 since day 1
23:43:47dionoeathe panel vibrates and cpu usage sky rockets
23:43:47Bagder(day 1 the sim existed that is)
23:43:55dionoeakillall gnome-panel fixes it ...
23:44:13BagderI say that is a bug in gnome then
23:44:19dionoeapossible :p
23:44:36Bagderthe sim uses only plain x11 functions
23:47:04 Quit noel_sad_sogn ("CGI:IRC (EOF)")
23:48:01***Saving seen data "./dancer.seen"
23:48:36*Bagder just made a little alsa app that says "glitch"
23:49:52Coldtoasthey Bagder: how do you disable a soundcard in Linux?
23:50:06ColdtoastI have 2 in this machine and need to disable one cos PCM gets sent to the wrong one
23:50:20zeunload its module?
23:50:20Bagderunload the driver for it
23:50:21ansivirusColdtoast: you can try rmmod
23:50:44ColdtoastI did try using the hotplug blacklist but it stopped working]
23:51:12 Quit PaulJ (Remote closed the connection)
23:51:18Coldtoastby "stopped working" I mean it stopped beign disabled. Something must have been calling the driver
23:51:22Bagderalso, the card is often identified by an alias in /etc/modules.conf that could be changed
23:51:59ColdtoastI have a good reason for hacing 2 cards, btw
23:52:00Bagderbut I'm not really skilled in linux sound
23:52:16Coldtoastone's my pro card which I need for ASIO. the other's an Audigy 2
23:52:23dionoeaColdtoast: you blacklisted it in /etc/hotplug/blacklist.d/<some filename> ? (works fine here)
23:52:32ColdtoastI use the pro card for recording guitar and the Audigy 2 for gaming
23:53:08Coldtoastis there another blacklist file in that die dionoea?
23:53:27dionoea"in that die" ?
23:53:37Coldtoasterr. die=dir
23:53:53dionoeawell ... you can put as many files as you wish
23:54:06dionoeathey all get parsed when hotplug is launched
23:54:17Coldtoastin /etc/hotplug/ is there a file called "blacklist"?
23:54:31dionoeathat was the old way of doing it i think
23:54:43dionoeabut its there all right
23:54:44Coldtoastaah. then I was doing it the old way
23:54:53Coldtoastyeah. that may be why it doesn;t work too well
23:54:55 Join muffti [0] (
23:54:55dionoeawhat distro are you using ?
23:56:06mufftii need some help installing rockbox
23:58:49mufftii've got a archos jukebox studio 10

Previous day | Next day