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 2005-06-27

00:00:48CassandraHmm. Do you think we should have a "design an iRiver default WPS" contest?
00:01:04t0masah, that's not fair to do in vacation time...
00:01:10t0masand not without conditional images
00:01:11rasherHrm, I'd put it off a bit
00:01:28t0maswait a little more... we can't do an iriver release anyway
00:01:30rashera case-like construct would be nice to have as well
00:01:57t0masor atleast a simple if()
00:02:10rashercase shouldn't be much harder?
00:02:30 Quit Coldtoast ("Peace and Protection 4.22")
00:02:43rasheror something like that
00:03:02t0masthat, and conditional images
00:03:05kenshinrasher: i'm starting in on trying to do exactly that (long way to go, though)
00:03:07t0masand maybe dynamic images...
00:03:17CassandraI love the WPS language. All the readability of ADA with non of the functionality. *grins*
00:03:21rasherkenshin: excellent
00:03:42rasherI don't believe I've ever seen ada
00:03:52CassandraHave you seen line noise?
00:03:58CassandraIt looks quite a lot like that.
00:04:10amiconnSlasheri: Shouldn't dsp code provide peakmeter data?
00:04:12Bagderyou mean like sendmail config? ;-)
00:04:20preglowamiconn: 'course it should
00:04:27preglowamiconn: rome wasn't built in one day, etc
00:04:44kenshinBadger: like Apache's mod_rewrite voodoo.
00:05:11t0maskenshin: Apache's mod_rewrite is read- and usable...
00:05:11preglowwhat, regex is pretty
00:05:29kenshinyeah, once you know how it works
00:06:23preglowsomeone embed perl in rockbox!
00:06:24Cassandrabadger: I was just thinking that.
00:06:27t0masResponse: 550 5.7.1 Rejected: listed at
00:06:39rashertucoz: doesn't seem too bad
00:06:40t0masthat's sourceforge...
00:06:52tucozno, not hello world at least :)
00:07:11rasherwell, I meant the language constructs
00:07:19rashera bit verbose perhaps..
00:07:27amiconnLinusN: Were you able to test the modified MFDR(2) values (i2c clock)?
00:08:31tucozit definitely looks nicer than ml
00:08:55rasherML is cute.. in a twisted horrible sort of way
00:09:02rasherNow, prolog..
00:09:02 Quit TBoy ("CGI:IRC")
00:09:18 Quit TCK- (Read error: 60 (Operation timed out))
00:11:28kenshinoooh. embed elisp instead of perl? ;)
00:11:40tucozI had a course in those lauguages, and found prolog nicer than ML. But, they do have their different uses I guess.
00:12:04Cassandralisp'd actually be a good choice.
00:12:14CassandraLightweight and good for string handling.
00:12:15Bagdertucoz: and all of them only in weird academic contexts ;-)
00:12:16 Quit Sucka` ("a bird in the bush is worth two in your house")
00:12:19kenshini had to write a scheme interpreter in univ.
00:12:30preglowban lisp!!!!
00:12:45CassandraOr forth, maybe.
00:12:52kenshinvive lisp!!!
00:12:57CassandraBut not Fortran. Ever under any circumstances.
00:13:03tucozBagder: true, I will never use either again
00:13:15t0maswhy not BF?
00:13:18rasherI'm told ML makes for nice compilers
00:13:25kenshinooh. Fortran brings up bad memories...
00:13:32t0masI wrote an interpretter in C long ago
00:13:37tucozrasher: for type-checks?
00:14:03t0masstill... brainfuck is the funniest language I know
00:14:08rashertucoz: yes
00:14:25Cassandrafunnier than Intercal?
00:14:45rasherI'd say whitespace has them both beat
00:14:46preglowi wrote a bf interpreter once
00:15:00tucozBut, the type checking algorithm that ML uses is probably used in other compilers
00:15:41rasherSure, but you don't have to worry about it yourself
00:16:14preglow <- language for heros
00:16:22tucozguess nto
00:16:27t0maspreglow: it's easy to write 1
00:16:34t0masI even wrote a debugging tool for BF programs
00:16:38rasherAnyway, I'm not going to touch ml if I can help it
00:16:54t0masthat one is just to funny
00:16:57CassandraML is only for nerdy compscis. :)
00:17:08t0masdebugging brainfuck with an ncurses tool :P
00:17:22kenshinCassandra: thanks for that vote of confidence!
00:17:32CassandraYou're welcome.
00:18:54tucozis the iriver build shaky at the moment?
00:19:16BagderI would say so, yes
00:19:18crwlhm, i seem to have a bunch of mp3's that crash the current build
00:19:26crwlabout a day or two ago they just didn't play
00:19:55rashernot crashing here..
00:20:05preglowi had an mp3 hanging me on play
00:20:16preglowand 48khz oggs will hang your player given time
00:20:25tucozBagder: in fuopl mode. i.e. frequent use of paperclip
00:20:33preglowit's just seen a big audio engine update, it's not surprisinmg
00:20:51Bagderfuopc here indeed
00:21:04*Bagder fuopcs tonight ;-)
00:21:08Bagdergood word
00:21:14Bagderhard to pronounciate
00:21:22preglowi can pronounce it just fine
00:22:44crwlhm, these files that crash my iriver seem to be normal VBR MP3's
00:22:53crwlthey're encoded with some lame 4.0 alpha though, it seems
00:22:54Bagderyes, mine are too
00:23:04Bagderno lame alpha though
00:23:17crwlbut others i tested work fine
00:23:19Bagderplain 3.9X series
00:23:25rasherAnything wrong with ?
00:23:58Bagderrasher: sounds like a good thing to me
00:24:11crwlw0w, now i got WPS data before it crashed
00:24:30Bagdermy crashes don't repeat on the same file when played again
00:24:41*Cassandra isn't getting crashes often.
00:24:50rasherBagder: same here.. patch looks correct too
00:25:06rasherand the fact that the id3v1 code does this already just adds to this
00:25:16*rasher hasn't had a crash in 30 minutes
00:25:17crwli haven't been getting crashes either (and these don't give illinstr, it just freezes), but these files didn't work earlier either
00:25:28crwleven though they seem to work fine in other players, including iriver original firmware
00:25:44preglowi don't get illinstr
00:25:46preglowi get i00
00:25:50BagderI get I00 too
00:25:50preglowunless that _is_ illinstr
00:25:58tucozMine crashed today when I was out running. Good thing I remembered a paperclip
00:26:12CassandraIn fact, I'm feeling as though I'm getting better bass from the new code. That can't be right can it?
00:26:20BagderI00 is _not_ illinstr
00:26:26preglowa rockbox user soon learns to be crafty when it comes to reset his h1x0
00:26:33preglowBagder: what is it?
00:26:38crwlwell, it just freezes here
00:26:46preglowso far i've used a piece of wood and a small brancjh
00:26:47rasherI ended up taking apart a keyring one day
00:26:47Bagderpreglow: I don't know, it has no name in the code for a reason I guess
00:27:13rasherbut now I have a paperclip in my keyring
00:27:25Bagderwise move
00:27:28kenshini slid a paperclip into the belt loop on my 120's case. the loop's purpose finally revealed itself to me! :)
00:27:36tucozI just put a paperclip in the case
00:27:39Bagderyou don't wanna get caught stranded without a paperclip
00:28:13tucozThat was after it crashed on the bus.
00:28:22preglowunless you happen to be a crafty man of the woods, like me!
00:28:36Bagdera buscrash!
00:28:41preglowoh, i've also reset it with a piece of chocolate wrapper
00:28:52rasherI used some tinfoil
00:28:53preglowand a shop receipt
00:29:01kenshinpencil tip
00:29:02preglowthe memories
00:29:06tucozyep, and no paperclip in sight. Once I uced a piece of a match
00:29:15CassandraThe best thing I have is a pushpin.
00:29:22CassandraI should put one in my purse.
00:29:52rasherI've been using one of those nails-with-plastic-attachment to fasten cables with
00:29:56rasherit has just the right size
00:30:22tucozpreglow, how the hell did you manage to get a paper for that use?
00:30:46crwleven mpg123 plays those crashing files fine without complaining anything
00:31:35Bagderthey've had years to mature
00:31:46tucozI tried on my bustrip. Didn't get it to work. Lecture notes, different types of papers. No use.
00:31:48preglowtucoz: like i said, i'm a crafty man of the woods, i'm used to surviving on gravel and pine needles, shaping paper into a resetting device posed no problem ;)
00:31:55crwlbadger, but mpg123 complains often if there's something wrong :)
00:32:06crwlit plays, though, but complains
00:32:15Bagdercrwl: right, but I'm quite sure these are problems in Rockbox and not with the files
00:32:23Cassandracrwl: The current code has bugs. This is the meaning of "alpha"
00:32:40*t0mas has found 4 iriver reset tools in different sizes :)
00:32:55*rasher tires out the "removing traling spaces" patch
00:32:59Bagdercrwl: do you get an error repeatably on certain files?
00:33:02tucoza pine needle might work. If you're gentle.
00:33:11crwlBagder, yes, i'm just interested why these particular files don't play, but others that seem quite alike, do
00:33:32t0masok, time to leave :)
00:33:36t0masc ya all in a week
00:33:37BagderI'm interested in getting a copy of a troublesome file, for my collection
00:33:37crwli've already played hundreds of mp3's with rockbox on iriver, but these don't seem to work :)
00:33:46amiconnrasher: What does that patch do?
00:33:52 Quit t0mas ("bye!")
00:33:54 Quit thegeek (Read error: 104 (Connection reset by peer))
00:33:56rasheramiconn: remove trailing spaces from id3v2 tags
00:33:59 Join thegeek [0] (
00:34:04amiconnAh, that one
00:34:07crwlBagder, yep, all songs in this album (which i 'acquired' from the 'net) seem to cause an instant crash
00:34:17rashervery small patch
00:34:27rasher(fortunately - had to apply it by hand)
00:34:43amiconnI noticed that some settings string have trailing spaces, supposedly to make sure that no remnant of another string are left behind
00:35:06amiconnHowever, that's not necessary, since lcd_puts() clears the rest of the line
00:35:46amiconn*settings strings; remnants
00:36:55Moosrasher: hehe and what's about the patch of this morning (about editing id3 tag)?
00:36:57amiconnOne day I'll clean that up, if I'm bored...
00:37:33rasheroh and checking for new rockbox in the usb-disconnect handler - please!
00:37:52rasherI more often than not find myself forgetting to rolo because I'm in some subdir
00:38:01amiconnI looked into this - it's not as simple as I though
00:38:28 Quit ghostiger2 (Remote closed the connection)
00:38:28 Quit ghostiger (Remote closed the connection)
00:38:55 Quit Harpy (Read error: 110 (Connection timed out))
00:39:30preglowLinusN: that's where i found it ;)=
00:39:46rasherOkay, patch worked (unsurprisingly).. want me to commit it?
00:40:53amiconnLinusN ?
00:40:55LinusNamiconn: no, i haven't had time to measure the i2c
00:41:13LinusNgoing away again tomorrow
00:41:31amiconnI know, that's why I asked
00:41:52LinusNcame home from the island last night at 2am
00:42:15LinusNand the elevator was broken (my wife is in a wheelchair)
00:43:04LinusNbut they sent a repairman promptly, took 20 minutes, phew
00:43:09Bagder <= now with crwl's problem as well
00:43:46LinusNBagder: a description of the problem with each file would be nice
00:44:18Bagderyes, I'll start to name them after who sent them and what the problem is
00:45:54Moosrasher: do you still think about this?
00:46:34LinusNMoos: i have an updated version of that patch
00:47:07 Join TCK [0] (
00:47:17LinusNMoos: want it?
00:48:18Moosunfurtunatly i can't do it :( , just rasher speaked of that this morning maybe can interested hiM
00:48:24rasherMoos: not really, since it was nowhere near compiling.. would take much more than what I could do
00:48:36Moosme too :(
00:48:36rasherLinusN: how much closer is it to working?
00:48:52Bagderand.... I00!
00:49:06Bagderfuopc it is
00:50:42LinusNLinusN: takes some work to apply, but you can try it
00:50:47MoosLinus: thanks nonetheless, i know you are too busy
00:50:53LinusNi haven't tried it myself
00:51:17amiconnI thought a bit about remote wps...
00:51:18BagderLinusN: maybe you should just submit your version to the tracker entry
00:51:30Moosexcellent idea
00:51:39LinusNBagder: sure
00:51:50amiconnit might be necessary to introduce a second gui thread for the remote
00:52:04amiconn..but that leaves the question how to distribute button events
00:52:10Bagderamiconn: I've been thinking about that too
00:52:12 Quit cYmen ("zZz")
00:52:19Bagdera lot ot tricky things pop up
00:52:59preglowi need to get back to writing before i cry myself to sleep
00:53:00CassandraI think this is another argument for rewriting the UI thread with virtual keyboard and screen handlers.
00:53:05preglowLinusN: have a nice trip
00:53:09preglowlater, all
00:53:17 Quit preglow ("bah")
00:53:26amiconnCassandra: Too much virtualisation won't do any good. I'm afraid
00:53:46CassandraVirtualisation solves many problems.
00:54:06amiconn...and introduces many other
00:54:28 Join elinenbe_ [0] (
00:54:42CassandraTrue. However, in my opinion, Rockbox is now so big and complex that virtualisation would actually significantly simplify the code.
00:54:53amiconnI'm not so sure
00:55:15 Join courtc [0] (
00:55:19CassandraAnd allow us to do stuff that just isn't possible under the current model eg timed recording.
00:56:01 Quit ac ("CGI:IRC (EOF)")
00:56:08amiconnVirtualisation will increase memory usage, code size and complexity, imho
00:56:18CassandraThe thing that would worry me most about virtualisation would be how it affects the executable footprint.
00:56:38amiconnMy idea is to introduce a second gui thread for units with a LCD remote
00:57:01amiconn...which handles all remote stuff, that is buttons, menu, wps etc
00:57:16amiconnThat requires a second button queue
00:57:22LinusNamiconn: sounds like a good approach
00:57:31amiconnInteraction would be done via some additional messages
00:57:38CassandraOnce you have two button queues, you might as well virtualise it, surely?
00:57:54elinenbe_any help here? I am setting up cygwin for the iriver enviroment −− previously it was set up for the archos enviroment, but when I checked out the wiki here: −− it says it will not work? is there any workaround?
00:58:10CassandraYou're half way there anyway, and your interthread communication is going to be incredibly complex., 'requesting' the remote lcd from the main thread, for use in plugins
00:58:12LinusNbuild your own compiler
00:58:36amiconnCassandra: I don't see why this should get overly complex
00:58:46rasherSpeaking of building compilers, I failed to build 3.3.4 sh-elf earlier today
00:58:48amiconnIt's not much more than what is currently done
00:59:09elinenbe_yeah −− I'll just do that... following the directions here should work −− right?
00:59:14amiconnCassandra: Right now, playback happens in a different thread that UI
00:59:20 Join ac [0] (
00:59:30amiconnA remote ui thread would provide just a second interface
00:59:44CassandraHmm. The problem is that pressing a button on the remote affects the main display and vice versa.
00:59:52LinusNrasher: what failed?
00:59:59amiconnCassandra: Not directly
01:00:28rasherLinusN: make.. let me finde the error
01:00:51amiconnImplementing my idea would allow to do crazy things... like browsing 2 different parts of the filesystem at once...
01:01:02LinusNamiconn: me likes it
01:01:20rasherLinusN: Yes, I saw.. not sure I did it correctly though
01:01:22Cassandra*mutters* Usability nightmare.
01:01:34rasheramiconn: And a new setting "link remote and main display" !
01:02:04amiconnWell, I think we should start with a partial interface
01:02:20rasherLinusN: the last error I get is "make: *** [configure-target-newlib] error 1" a second wps, which is replaced with our generic rockbox logo when nothing is playing
01:02:28LinusNlinking them could be as easy as replicating the events to both ui threads
01:02:40CassandraIt could be cool, I agree.
01:03:02*amiconn imagines browsing id3db on one screen, and filesystem on the other
01:03:08LinusNrasher: are you sure you put the soft link in the right place?
01:03:13 Quit markun ()
01:03:20rasherLinusN: no.. where does it go?
01:03:42amiconnLinusN: Is there a remote without lcd for iriver?
01:03:48CassandraWhy don't we have the id3db as a virtual directory in the root of the filesystem anyway?
01:04:04rasherI have gcc-3.3.4/newlib -> ../newlib-1.11.0/newlib
01:04:08amiconnCassandra: Hmm, good question
01:04:14LinusNrasher: sounds right
01:04:14elinenbe_amiconn: could you do something like a split-screen tree browser/WPS?
01:04:15rasheramiconn: I think there is for h300
01:04:19amiconnCould be done the same way as multivolume on Ondio
01:04:51rasherThat sounds like a lovely idea
01:04:54amiconn...only showing the <ID3DB> virtual volume if a valid database is found
01:04:56rasherswitching back and forth is a pain
01:05:05CassandraAnd then we wouldn't need the seperate ID3db Show Files setting.
01:05:13rasherit needs a better name than id3db, too
01:05:36rasheractually, tagdb is used in a lot of places already
01:05:37amiconnrasher: RockDB
01:06:11amiconnYes, tagdatabase is better as well
01:06:43amiconnThere is not only id3 support, even on archos one day
01:06:55amiconn(afaik mp3 can also have apev2 tags)
01:07:11elinenbe_we need "on target" database creation...
01:07:12CassandraTag is good.
01:07:19CassandraIt's what iRiver used.
01:07:49CassandraSomething snappy like "Tagbank" or "Tagbase" only less naff.
01:07:54LinusNamiconn: i believe there is an lcd-less remote for iriver
01:08:14amiconnLinusN: Hmm, that could be tricky then
01:08:44amiconnWe need to re-route remote button events depending on whether an lcd or an lcd-less remote is connected
01:08:58amiconnHope that's detectable at least...
01:08:59CassandraYou know what'd solve that problem?
01:09:09amiconnNo it wouldn't
01:09:28amiconnYou'd need to re-route anyway
01:09:48amiconn...and need to probe what's connected, virtualised or not
01:10:13CassandraIf you can't probe, you can always have it as a config option.
01:10:52amiconnLinusN: You don't happen to have such a remote, or know someone who has?
01:12:08LinusNi think it comes with the h300
01:12:21amiconnDidn't you acquire one?
01:12:31CassandraHow about we just call the database the "Tag folder"?
01:12:54LinusNamiconn: yes, a used one, without the remote
01:13:07amiconnAh, pity
01:13:18CassandraI thought they didn't actually come with a remote.
01:13:39amiconnLinusN: The H1x0 and H3x0 remotes are exchangeable, right?
01:14:09amiconn(I guess that because there are some offers of remotes for H1x0 and H3x0 on ebay)
01:14:26DBUGEnqueued KICK MO-Pantsu
01:14:26*MO-Pantsu »» np · 19 - Baltar Panics - BATTLESTAR GALACTICA, SEASON 1 - Bear McCreary (2005) · 00:28/01:46 [¦¦¦¦¦¦¦¦¦¦¦¦¦¦] · kbps (vbr) ««
01:15:14acamiconn: yep... i think i have read that you can use h1xx remote on h3xx remote
01:15:18LinusNamiconn: they have different buttons
01:15:35acah.. iAudio X5 has a ColdFire SCF5249
01:15:43LinusNac: yes
01:15:57HClMO-Pantsu: could you please turn that script off?
01:16:09MO-Pantsuit is off I manually did that
01:16:30rasherI suppose the h300 port will eventually benefit from a "treat remote as h100 remote" option
01:16:56HClwhats all this talk about the database? thats my territory :p
01:17:24CassandraHCl: I don't care who implements it. :)
01:18:15CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
01:18:15*LinusN has an H300 LCD remote
01:18:26HClpff, so much for credit to the hard working coders
01:18:27acits quite hard to get some scans of the hardware of the iaudio x5
01:18:27HCl :P
01:19:04LinusNac: is it hard to open?
01:19:41*rasher updates the ReleaseNotes
01:20:00acLinusN: i am doing some online research, before i buy it
01:20:09amiconnLinusN: Did you try the H3x0 remote on H1x0?
01:20:17amiconn(with iriver firmware)
01:20:28rasherhah, the editid3tags patch is in fact the oldest standing one
01:20:30Moosac: what's about your iriver?
01:20:47rasherLinusN: can I see the updated patch for that?
01:21:03acmoos: maybe i get it running again, but i am not very optimistic
01:21:30Mooswhat is the problem?
01:22:14acdont know it now... tomorrow i get a multimeter
01:23:16Moosi really hope for you it relive
01:24:00Moosthe resurection of the phoenix
01:24:45LinusNrasher: sf patch updated
01:25:18thegeekdo you guys use an ide when programming rockbox?
01:26:11CassandraYes. It's called "bash".
01:26:25 Part tucoz
01:26:59elinenbe_LinusN: that jzoss patch was a great piece of work... it was never committed though...
01:27:20rasherelinenbe_: a real shame
01:27:22elinenbe_rasher: good luck with that one!
01:27:25LinusNwell, afaik, it wasn't too kind on the existing tags
01:27:33elinenbe_it really is very useful...
01:27:47rasherLinusN: the latter version claimed to be
01:28:39LinusNi see that now
01:29:20rasherwow, lots of failed hunks
01:29:30rasherfun fun fun
01:29:48rasherHunk #1 succeeded at 402 with fuzz 2 (offset 351 lines).
01:30:42rasherI'm sure I won't
01:31:12rasherguess I should get any other changes out of the way first
01:34:19CassandraThis is the problem with patches. They go off faster than carrots.
01:36:28rasherI say we allow anonymous changes to cvs
01:36:34rasherworks for the wiki
01:36:45rasher(no, not really)
01:41:00LinusNthe problem with patches is that they seldom are ready for direct inclusion in cvs
01:41:15LinusNso we need the author to do some changes before we can commit it
01:41:37LinusNthen the author gets tired/annoyed and disappears
01:42:23*rasher remembers to run make in /tools/
01:43:12amiconnDifferent topic - the codec api should support version control the same way as the plugin api
01:43:40rasherDifferent topic still - it seems that building a simulator without logf support breaks if a plugin does logf
01:44:11amiconnrasher: Yes, that's because logf() is included conditionally in the api logf and non-logf api are incompatible
01:44:54***Saving seen data "./dancer.seen"
01:45:10rasherBut it works for the main code
01:45:29amiconnHmm, it should behave the same way for main
01:45:50rasherHm, I'm not sure
01:46:27CassandraWhat is logf anyway?
01:46:28amiconnWe would need a logf() stub for non-logf builds, but I don't like that either...
01:47:53amiconn...and when I was asked to include logf while I sorted the api, I even tried that
01:48:17amiconnFor some reason I didn't get it to compile, that's why I didn't add logf() back then
01:48:32amiconnNext day someone else did, with that conditional...
01:53:30LinusNtime to sleep
01:53:36LinusNnite all, cu in 3 days
01:53:44 Quit ac ("CGI:IRC")
01:53:57 Part LinusN
01:53:57Moosgood night Linus
01:54:08CassandraHave a good time.
02:04:18 Join courtc_ [0] (
02:04:40 Quit courtc (Read error: 110 (Connection timed out))
02:07:28rasherthink I have all the changes applied now.. now lets see what breaks
02:08:56rasherdon't get too excited just yet
02:09:02rasherit rearraged lots of stuff
02:09:06rasherI expect things to break horribly
02:09:45MoosIt's a good start
02:10:53 Quit shank (Remote closed the connection)
02:11:10rasheramiconn: you around?
02:13:13elinenbe_rasher: is this for the jzoss id3 edit patch?
02:17:58elinenbe_yes... good luck!
02:18:38rasherI'll need it..
02:26:12Moosooohhh David Bryant commit for wavepack :)
02:27:21Moosgood night all
02:27:47Moosrasher: good luck for the patch ;)
02:28:00 Part Moos
02:29:39 Join amiconn_ [0] (
02:45:20rasheroh excellent..
02:46:19 Quit amiconn (Read error: 110 (Connection timed out))
02:46:19 Nick amiconn_ is now known as amiconn (
02:49:29rasherthis plugin is such a mess
02:50:20rasherhalf of it is not using variable button-definitions
02:51:06rasherwhat a frikkin' mess
02:58:08CassandraNow you see why no-one got around to committing it.
03:00:30rasherI'm sure this is just the tip of it
03:00:33rasherit actually compiles now
03:00:41rashernow to see why it doesn't run
03:03:06rasherhrm, it does do *something*
03:08:32*kenshin wishes t0mas were still around to guide me on my WPS changes...
03:10:17CassandraNight all.
03:10:59 Quit Cassandra ("If I were actually witty, this quitline would be funny.")
03:11:30elinenbe_rasher: like i said... good luck!
03:11:36elinenbe_rasher: it'll be worth it!
03:12:37*rasher sprays logf calls all over edit_id3.c
03:21:26 Part kenshin
03:21:31 Join kenshin [0] (
03:25:03rashersomething happened
03:26:18rasherIt messed up the tags slightly, but it's actually displaying the tags of this file
03:27:06kenshinrasher: a good sign.
03:27:33rasherit's flaky though
03:27:39rasheropening the same file now broke
03:27:53kenshini finally got image load and display split into two components
03:28:02kenshinso images can be conditional, now
03:29:06kenshinstill needs some work i image. only my highly emptied WPS works so far! ;)
03:30:12kenshinwith t0mas out for a week, who can i ask WPS questions to?
03:30:29kenshinmy changes are heavily dependent on t0mas' code
03:30:44rasherI think Bagder knows about it
03:31:02rasheraw, dammit
03:31:08rasherit's crashing outside the plugin
03:32:09kenshinooooh. :-o
03:35:15kenshinrasher: think it's better to pre-load all images or load them only after they're requested for display?
03:35:40rasherI'd say preload
03:36:21kenshincool. that's what i coded it to do.
03:36:57rasheror you could end up with pauses in the middle of it all
03:37:27kenshinyep. that was my thinking.
03:38:34kenshinso i actually know something worth knowing! ;)
03:44:58***Saving seen data "./dancer.seen"
03:52:12rasherit still plays music though!
03:55:23*rasher -> sleep(3600*7)
03:56:15Stryke`takes arguments in seconds?
03:56:29rasherI hope!
04:00:54 Quit rasher ("leaving")
04:29:10 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
04:38:12 Nick courtc_ is now known as courtc (
04:45:11 Quit kenshin (Read error: 110 (Connection timed out))
05:30:42 Quit Febs_away ("Chatzilla [Firefox 1.0.4/20050511]")
05:45:02***Saving seen data "./dancer.seen"
06:19:41 Join Strath [0] (
06:33:23 Join Coldtoast [0] (
06:39:03 Join ashridah [0] (
06:43:53 Join Bgr [0] (~Bager@
06:43:53 Quit Bger (Read error: 104 (Connection reset by peer))
07:11:08 Quit hardeep ("[BX] Back wit anutha one of doz BitchX-rockin' beats!")
07:27:46ashridahheh. watching the audio thread view is more entertaining than watching the WPS :)
07:31:11Bgrmorning ;)
07:34:20 Quit Strath ("Client closed")
07:37:33 Nick Bgr is now known as Bger (~Bager@
07:39:14 Join B4gder [0] (
07:40:45amiconnmorning :)
07:41:05B4gderhi ho
07:44:02 Quit B4gder (Read error: 104 (Connection reset by peer))
07:44:23 Join B4gder [0] (
07:44:44Bgerveni, vidi, vici ... gone
07:45:03***Saving seen data "./dancer.seen"
08:10:44B4gdernew lang format suggestion posted
08:11:30Bgerwiki ?
08:11:35B4gdermailing list
08:12:09B4gderI'll make a wiki page later when the biggest flaws have been corrected :-)
08:12:41Bgerregarding translations: i can help translating bulgarian
08:18:20B4gderyou know any other bulgarian users?
08:19:28Bgerin fact - no
08:19:47Bgerbut someone must be using it ... at least the one who translated it...
08:20:06B4gderyes, at least someone once did
08:21:09Bgerok, then i'll skip it for now
08:27:18amiconnB4gder: Imho it might be a good idea to introduce the SOURCE mechanism to the plugin library build
08:28:11B4gderit is an oversight that it isn't already
08:32:32amiconnThat would save quite some compilation time, at least for cygwin users
08:32:55amiconnThe grayscale lib would then only be compiled on target, for the archos bitmap lcd
08:33:07amiconnThe player gfx lib for charcell display only
08:34:02B4gderhm, isn't SOURCES already used there?
08:34:49amiconnHmm, you're right...
08:35:03amiconnJust the #ifdef for the grayscale lib is wrong
08:36:29amiconn-> #if CONFIG_LCD == LCD_SSD1815
08:36:34amiconn(for now)
08:36:38B4gderseems right
08:38:07amiconnHmm, this #dfeine is partially defined for simulators as well
08:38:15amiconn(currently only for H1x0)
08:38:46B4gderah, right
08:40:23amiconnI think you changed that, because else the recorder lcd driver was used for H1x0 simulators, which caused graphical glitches
08:40:59amiconn...because of having too small fill bitmaps
08:41:02B4gderI think Linus fixed that
08:41:10B4gderbut yes, it was due to those rasons
08:41:23amiconnThat's unnecessary atm; they no longer use fill bitmaps
08:41:41amiconn...but it will become necessary again when 4-grey mode is added
08:42:14B4gderimho, if you want it to not be build for the sim, you should check SIMULATOR
08:43:08amiconnYes, as well as the lcd type
08:43:34B4gderthat way, we can have the LCD type or not in the sim build without it affecting the build
08:44:51amiconnI'm thinking about adding buffering to grayscale, perhaps as a secondary library
08:47:36amiconn...but to utilise it fully on archos (fullscreen + many shades), we would need to increase plugin RAM on archos
08:52:15SlasheriHi. I found that i should also attach a paperclip to my keyring before i get off to work ;)
08:52:40B4gderI didn't even take my iriver with me today
08:52:51B4gderlast night's experience was too annoying
08:58:09Slasheriah, i found one problem that might cause a crash.. fixing it
08:58:18Slasheri(if crossfading is enabled)
09:22:00 Join webguest84 [0] (
09:37:54 Join kurzhaarrocker [0] (
09:38:23*kurzhaarrocker detects proposals for modern technologies in rockbox
09:38:28kurzhaarrockerI'm shocked
09:39:53kurzhaarrockerHow could you possibly suggest a xml format for the languages, Bagder? Isn't that a rude violation of KISS? :)
09:40:03B4gderits not XML
09:40:11B4gderit is similar to it though
09:41:11B4gderand I'm always in favour of the best format for each particular job
09:41:30B4gder"best" of course being very subjective
09:45:06***No seen item changed, no save performed.
09:47:11kurzhaarrockerIs there a reason why you didn't make it _real_ xml?
09:57:04B4gdermostly because I don't want to write/use a complete XML parser
09:57:24B4gderand I don't like full XML when it is intended to be read and written by humans
10:03:27kurzhaarrockerI don't think it would be necessary to write e complete XML parser even if the file xml.
10:03:56 Join Musicmad [0] (
10:04:14kurzhaarrockerBut if the file is xml you'd have the benefits of syntax highlighting editors, format checkers, etc
10:04:35*kurzhaarrocker bows to Musicmad
10:05:22Musicmadif anybody sees rasher tell him there was a small mistake in the danish translation :)
10:05:36Musicmadneeds a "t" and a blank space.
10:05:42Musicmad"Indsæt undermapper?"
10:18:44 Part Musicmad
10:40:33 Join Godeater [0] (
10:41:22Godeatermsg nickserv register f566mgy
10:41:29Godeateroops :)
10:41:34B4gderooops :-)
10:41:44Godeaterlucky I don't use that password much eh ? :)
10:41:56Godeaterand will be using it even less now
10:53:44Slasherioh, i didn't have to use the paperclip a single time :) that looks promising
10:53:50 Join Harpy [0] (
10:54:10*B4gder announces the paperclip free day!
10:54:18 Join Chamois [0] (
10:54:24 Join Aison [0] (
10:55:36 Join cYmen [0] (
10:57:44 Nick Lynx_awy is now known as Lynx_ (
10:59:30BgerGodeater: soon your pw will be searchable in google :P
11:01:14GodeaterBger: I'm sure it's probably there already :)
11:01:48GodeaterI can only think of one internet accessible account that I still use that one for - good luck guessing where it is :)
11:02:40BgerGodeater: it's too early to find it yet
11:02:54B4gderI've heard people can _change_ passwords
11:03:56Godeateryeah, I've just done that actually, it just seems a little sad, I've had the same password there for over ten years now!
11:04:26Godeaterstill - better to change it than be forced to post as Anonymouse Coward from now on I suppose
11:05:16Godeateror Anonymous even.
11:08:25pabsslashdot hasn't had accounts for 10 years
11:09:08Chamoisthe new PLAY menu on iriver is uncompatible with the PLAY+joystick in the browser
11:09:57Chamoissomeone should fix thart
11:10:12 Quit kurzhaarrocker (Remote closed the connection)
11:10:54Godeaterpabs: huh? I got my account there in like 1995
11:11:37pabsthat's nice, except slashdot didn't have accounts then
11:12:20Godeaterdidn't it? Well it'sa bloody old account. Sub 10K user id.
11:13:21pabsi'm relatively sure the accounts were added in 98, possibly 97.
11:13:29pabsbut definitely not any earlier than that
11:13:57 Join ep0ch [0] (
11:14:15Godeaterah - I'm just getting old - don't remember anymore :)
11:16:00ep0chep0ch would just like to say that his 44.1 khz stereo Vorbis file (80 kbps) is 'almost' realtime without CPU boost (just 4% boost)!!
11:16:18ep0chwhich is damn impressive to me
11:20:38ep0chand for MP3 playback it looks like bittrate has little effect on CPU usage...
11:22:15ep0chi.e. a 128 kbps CBR MP3 uses about 48% boost, whereas a 235 kbps VBR uses just 15% more boost
11:22:29ep0chwierd no
11:22:59 Join ac [0] (
11:23:04achi all
11:23:15*ac adds now a new wiki site
11:25:30B4gderadd a link to it from the NonArchos page as well
11:25:37ep0choh nice, i would buy one if hit had a bigger disk to replace my ihp120
11:26:02acB4gder: i will add the link
11:26:47Bgerac: what's the exact model of the CPU
11:26:55Bgeralso, no images
11:28:03acone moment.. im am fightening with the wiki
11:28:34acBger: SCF5249
11:28:45Bgerac ?????= ?
11:30:14ac - it hates me... the link, which should be a link is none *grrrr*
11:30:31Bgerlet me see
11:31:13Bgerhm, u're editing it
11:31:26B4gderit is because you've used a non-wiki name I guess
11:31:48actry now
11:32:14B4gderyou'll need to make an explicit link to it
11:32:37B4gderor rename it
11:35:01acBger: could you edit the site?
11:43:44Coldtoastrockbox crashed on me. doh
11:45:08***Saving seen data "./dancer.seen"
11:45:21Coldtoastwhat's "ATA Error -11"?
11:48:59Bgerac ok
11:49:15Bgerfirst m68k toolchain under gentoo works ;)
11:49:40Bgermade with crossdev
11:50:09 Join Musicmad [0] (
11:51:15acif have made my toolchain on gentoo by hand
11:51:16Godeaterthere are ebuilds for the m68k toolchain ?
11:51:26BgerGodeater :
11:51:34Bgerfirst emerge crossdev
11:51:37Bgerafter that
11:52:17Bgercrossdev −−binutils 2.16 −−gcc 3.4.4 -s1 -t m68k-elf
11:52:20Bgerafter that
11:52:29Bgerbefore that
11:52:34Bgeredit your /etc/make.conf
11:52:45*Godeater cut and pastes furiously
11:52:52Bgerno, w8
11:53:12Bgeryou must put something like following in /etc/make.conf
11:53:46Bgerand mkdir PORTDIR_OVERLAY
11:53:58Godeaterthis is before running the crossdev command line ?
11:54:01Bgerafter that run crossdev
11:54:10Bgeryes, but crossdev will complain about it
11:54:47Godeatercan't you just run PORTDIR_OVERLAY="/blahblahblah" crossdev blah ?
11:54:48Bgerand all that follow is binutils-config -l => get the profile you need -> binutils <profilenumber> or <profilename>
11:54:50Godeaterwithout editing the make.conf ?
11:55:30Bgerand gcc-config -l => get the profile => gcc-config <profilenumber> or <profilename>
11:55:45Bgerand don't forget to comment out PORTDIR_OVERLAY
11:56:35Bgerbtw, u must run gcc-config & binutils-config as root (maybe it will work with a "portage" user, but i didn't test this)
11:56:39GodeaterI think I get it
11:57:03GodeaterI think you're right- gcc-config and bon-utils config both need root
11:57:10Bgerdon't forget to rerun gcc-config & binutils to switch to default profile
11:57:14Godeaterwow - rubbish typing today
11:57:36 Quit webguest84 ("CGI:IRC (EOF)")
11:57:55Bgerif you want to compile something other of course
11:58:51 Join webguest84 [0] (
12:01:32Bgerabout PORTDIR_OVERLAY="..." crossdev i guess you can :)
12:01:44Bgerbut you must make the dir on your own
12:02:19 Join Moos [0] (
12:02:23MoosHi all
12:02:33 Quit webguest84 (Client Quit)
12:03:04acBger: for what are you using PORTDIR_OVERLAY? It is used to use ebuilds, wich are not in portage
12:03:23Bgerac: the crossdev script complains
12:03:51acBger: ah ok
12:04:32Bgerdon't forget that it's usually used for building cross-platform gentoo :)
12:07:37 Join Musicmad_ [0] (
12:10:10Bgerac: about the wiki
12:10:13Bgerwhat do you want to change ?
12:11:40acBger: For X5 see iAudioX5HardwareComponents for more details. -> iAudioX5HardwareComponents should be a valid link
12:12:16Bgerthe wiki keywords must begin with uppercase letter (iirc)
12:13:58 Join rasher [0] (
12:14:21acBger: ahhh.. thanks for changes
12:14:47Bgeri'm not jonas
12:14:52Bgerrasher is jonas iirc
12:14:54*rasher <-
12:15:24rasherYup.. so.. I have an updated edit-id3 patch if anyone's interested.. it still not works, but it's pretty close
12:15:27Bgersorry for asking such FAQ, but about the problem convbdf: command not found ? :)
12:15:38rasherBger: make in the tools dir
12:16:40acrahser: merci
12:17:17 Quit DMJC ("Leaving")
12:18:36Bgerhm Godeater
12:18:41Bgermaybe i'm wrong
12:18:53Godeaterwhat about ?
12:19:14GodeaterI've not tried it yet - my gentoo is still in the middle of an 'emerge -uD world'
12:19:16Bgerabout gcc-config & binutils-config
12:19:25Godeaterwell I'll give it a try later
12:19:30Bgerhm, i've not run X yet :)
12:19:40Godeaterfor the time being I'll stick with the handbuilt toolchain I made
12:20:27Bgeryes, if it works, just use it
12:21:25Godeaterit appears to work certainly - although I've not actually loaded a built firmware onto my iRiver yet
12:21:38GodeaterI need to update my source tree again!
12:22:24BgerGodeater in fact, you don't need to run gcc-config & binutils-config
12:22:31Godeaterno ?
12:22:45Godeaterhow does it know to use a different gcc / binutils set then ?
12:22:54Godeateroh - the makefiles
12:22:58Bgeryep :)
12:22:59Godeaterduh - stupid question
12:23:02Bgerrockbox's makefiles
12:23:19*Godeater wears the pointy hat with the big D on it for a while
12:23:32*Bger emerges zip
12:24:24GodeaterI wish the firewall here let me use CVS properly :(
12:25:09 Quit Musicmad (Read error: 110 (Connection timed out))
12:25:26*Bger goes to read CVS docs
12:29:33Coldtoastthat's one crazy radio controlled toy!
12:32:24 Join |ep0ch [0] (
12:37:15aci have here the firmeware update for iaudio x5. it is a setup, which copies the fw to the player. is there a way to get the firmware out of the exe file?
12:39:50BgerB4gder why there is an *empty* "archos" dir in build dir for iriver h1x0 ?
12:40:19Bgerin THE build dir...
12:42:17B4gderthat is created when you setup a sim
12:42:35B4gderused as root dir
12:43:07B4gderwhen you make install, it unzips the zip file there
12:43:11Musicmad_Coldtoast: nice.
12:43:21 Join webguest77 [0] (
12:43:28Coldtoastcool huh?
12:43:34webguest77ac: just read the logs. Check this out.
12:44:07webguest77there is a .bin-file inside
12:44:43Bgerand it's unencrypted afaik
12:44:56acwebguest77: wow.. very good news
12:45:14webguest77also, looks unscrambled
12:45:22webguest77just like with the m3
12:45:31Musicmad_Coldtoast: unbelievable. I want one.
12:45:57Coldtoastyeah. me too
12:46:12acwebguest77: very good...
12:46:36webguest77good luck, got to go
12:46:39 Quit ep0ch (Read error: 110 (Connection timed out))
12:46:39 Nick |ep0ch is now known as ep0ch (
12:46:40 Part webguest77
12:57:19 Quit ep0ch (" Try HydraIRC -> <-")
12:59:32 Quit ashridah (Read error: 110 (Connection timed out))
13:02:24Coldtoastany of you use those kensington locks for your notebook?
13:02:39 Join ashridah [0] (
13:10:59rasherBagder: can't see any problems with your language-file proposal.. other than the danger that people may forget to update all translations, but no format-change will solve that
13:11:12 Quit Musicmad_ ("Trillian (")
13:12:12B4gderI'm currently thinking on how we can throw in translated plugins too into the concept
13:13:07 Quit edx (Read error: 110 (Connection timed out))
13:14:03rasherIt would be nice if the plugins provided the English text themselves somehow, such that they'd still only be one source file
13:14:18amiconnB4gder: I already made a suggestion for localised plugins
13:14:47B4gderamiconn: URL?
13:15:00amiconnNo URL; discussed it herein with Linus
13:15:11amiconn(about a month ago iirc)
13:15:20B4gderwe've discussed this many times
13:15:37amiconnThe idea was based on our current localisation system
13:16:13amiconn(1) A second lang directory in apps/plugins/, with translations split by language
13:16:41amiconn(it's more likely that one translator will translate all plugins than a plugin author knows all languages)
13:17:44amiconn(2) All strings for languages that are available for a plugin would be compiled into the plugin, as several 'string chains' the way the core works now
13:18:23amiconn(3) The plugin would then read the global UI language from the core, and switch to the same internally if that's available
13:18:41amiconnIf that is not available, english would be default
13:19:09amiconn(4) The system would be available for plugin authors as part of the lib
13:19:23B4gderI don't get (2). Would all strings for all languages be built in each plugin?
13:19:41amiconn(all valid ones for that plugin)
13:19:47B4gderwon't that become a problem?
13:19:52amiconnThere aren't too many
13:19:59B4gder24 languages
13:20:15amiconnyeps, and just a handful of strings perlanguage
13:20:18B4gderjust seems... like a potential waste
13:20:35amiconnAll other systems I can think of have just disadvantages
13:21:25B4gderwhat's the major disadvantage of loading the strings into a plugin buffer like the main lang system works?
13:21:47 Quit ac ("CGI:IRC (Ping timeout)")
13:22:44amiconnThe problem is splitting the files
13:22:56Coldtoastoh wow! that's awesome!
13:23:14Coldtoasthadn't seen the quick menu you get when you hold down ON in the file browser
13:23:25rasherIt's new as of yesterday
13:23:36B4gderamiconn: I would imagine system where a single file contains strings for multiple plugins
13:23:38amiconnSplitting per plugin would be basically the same as having the strings builtin, only that there is one more file per plugin -> unnecessary, imho
13:23:38Coldtoastsaw the other one
13:24:04amiconnSplitting perlanguage makes even less sense, because it destroys plugin independency
13:24:08B4gderits just that we have all strings for all languages
13:24:16B4gderso 10 strings in english, makes 10 strings in all languages
13:24:21B4gderso 24 langs, makes 240 strings
13:24:26amiconnSplitting per language *and* per plugin is just too many files
13:24:27B4gderwhen 10 is used
13:24:38amiconnYes, and?
13:24:44B4gderamiconn: that's why I think of combined files
13:24:53amiconncombined files?
13:24:57B4gderamiconn: it wastes memory
13:25:12B4gderamiconn: having strings for multiple plugins in a single file
13:25:27amiconnThat's the worst variant, imho
13:25:34B4gderbut only load the part for the plugin you use
13:25:47amiconnIt destroys plugin independency
13:25:55B4gderno one uses that today
13:26:10B4gderand we could support indenpendent as well
13:26:13amiconnIf the translation for one plugin is updated, you need to update the whole file
13:26:22B4gderthe script does that
13:26:27amiconnWhat's so bad about building the strings into the binary
13:26:31B4gderno one generates those things manually anyway
13:26:42amiconn10 strings * 24 languages * 20 bytes = 4800 bytes
13:27:07B4gderand if you write a plugin that uses 50 strings?
13:27:15B4gderit isn't impossible
13:27:21amiconnThen there are some more bytes needed
13:27:27B4gderI don't think it scales very well
13:27:35Coldtoastjust got "I04: IllInstr at 10000C84"
13:28:05 Quit rasher ("CGI:IRC")
13:28:15B4gderand I'm a bit amazed that you, amiconn, of all people is in favour of wasting thousands of bytes! :-)
13:28:48amiconnB4gder: Having strings for all plugins in one file creates another serious problem - ID number consistency
13:29:06B4gderit would only load the chunk for the particular plugin
13:29:36B4gderhaving an index in the start of the file
13:29:38amiconnYes. But what happens if the author of the first plugin adds 2 new strings?
13:29:49B4gderwe generate a new file
13:29:53amiconnAll subsequent IDs would shift
13:30:02B4gderthe IDs are per plugin
13:30:21amiconnThat means we need a plugin id as well
13:30:32 Join Cassandra [0] (
13:30:39B4gderthere's a plugin name used when loading/starting
13:30:39 Join preglow [0] (
13:31:00preglowi gotta say i agree with bagder
13:31:07amiconnB4gder: Hmm. I'm still not convinced...
13:31:14CassandraAnyone have any idea where I can find a list of UTF-8/unicode equivalents for the old codepage 850 ASCII codes?
13:31:25amiconn...although this might have advantages when it comes to voicing plugins
13:31:49B4gderI'll do some further thinking and then write down my ideas/suggestions
13:32:06amiconnYour new language handling proposal has some disadvantages
13:32:16preglowwith a lot of strings, which some plugins will undoubtedly have, it WILL waste tons of space
13:32:20amiconn...or rather, areas of possible improvement
13:33:00preglowwhen we have plugins with the complexity of rockboy, you can't rule out much
13:33:04amiconnB4gder: It should be possible to assign an adapted translation to more than one target, to avoid unnecessarily repeating it
13:33:05B4gderamiconn: I'm all ears
13:33:21B4gderamiconn: right, I thought about that but couldn
13:33:23amiconnLike, a specific string for ondiosp *and* ondiofm
13:33:26B4gdert think of a nice way
13:33:45amiconnor even worse: for recorder v1, v2 and fm
13:33:59amiconnor for ondiofm and recorder fm
13:34:26B4gderany idea on how?
13:35:06amiconnFirst thought - comma separated target list left of the colon
13:36:45amiconnYour proposal creates a file size problem for voicing, but that can't be avoided I think
13:37:30amiconnToday I am building 6 voices, which total to ~9 MB
13:37:32B4gderhwo does it make any different size?
13:38:01amiconnWhen I am building per target, this will multiply by 7 (currently)
13:38:32B4gderyou mean the totaly size of all voices?
13:38:47B4gderthen yes, it will be more outputs
13:38:59B4gderbut each voice will be better
13:39:03amiconnUploading all them will get tedious
13:39:10B4gderor at least they can be better
13:39:16preglow9 megs isn't that much
13:39:16amiconnYes of course
13:39:19preglowdo you do it that often?
13:39:29amiconnNot that often
13:39:49amiconnI could automate the upload, but probably not when uploading to the wiki
13:40:13amiconn63 megs *will* take a while via ADSL with 128 kbps upstream
13:40:54preglowoh, indeed
13:40:56CassandraWould it be possible to have a "common" file and a suplemental platform file.
13:41:01amiconnI'll probably drop the MS voices soon
13:41:09B4gderI could offer a different upload way to allow automation
13:41:12amiconnThey don't sound that nice
13:41:21CassandraSince the majority of strings are common to all platforms.
13:41:29amiconnI'll possibly buy some UK english AT&T voices instead
13:41:40CassandraThat would be cool ami.
13:42:04CassandraI'll happily throw in a tenner towards them.
13:42:36preglowhow much do they cost?
13:42:42amiconnI think about complementing my female german with a male, and buying the 2 uk english voices as well
13:42:43preglowisn't this was the rockbox donation money is for anyway?
13:42:50amiconnpreglow: about 28 EUR
13:42:52B4gderamiconn: I like the idea with a comma-separated list on the left side of the colon
13:42:56amiconnper voice
13:43:44preglownot too bad, actually
13:45:12***Saving seen data "./dancer.seen"
13:45:35amiconnCassandra: I think splitting the file in two isn't nice from a user point of view, and it may create additonal problems
13:45:59Cassandra*nods* Probably better to have 1 file per platform, yes.
13:49:24amiconnB4gder: Would it be possible to automate the voice file upload? How about other contributors
13:51:05B4gdereveryone that wants to do automatic would need to get an account/details from ue
13:51:19amiconnHmm, ok
13:51:42amiconnOne would then link to an external dir from the wiki?
13:51:53B4gderyes I think so
13:52:01amiconnSounds good
13:55:26amiconnB4gder: Does curl support authentication?
13:55:41B4gderall sorts of them
13:55:53amiconnNice, so I could use that, directly from my build script...
13:55:56 Join bipak_ [0] (
13:56:07amiconn...given that there is a win32 version
13:56:11B4gderthere is
13:56:24B4gdercurl exists for just about every platform there is
13:56:38amiconnEven for Amiga?
14:07:23 Join oxygen77 [0] (
14:07:40 Quit oxygen77 (Client Quit)
14:09:30 Quit elinenbe_ (" HydraIRC -> <- Try something fresh")
14:12:09 Quit bipak (Read error: 110 (Connection timed out))
14:30:42 Quit Godeater ("CGI:IRC (Ping timeout)")
14:31:02Bgercleaning codecs/lib
14:31:22Bgermake[2]: *** No rule to make target `clean'. Stop.
14:33:33amiconnB4gder: Another thing concerning your localisation proposal
14:34:06amiconnSince the language files will become target dependent, they'll need to include a target id
14:34:16amiconnSame goes for the voice files
14:34:24B4gderyes, makes sense
14:34:52amiconn...and they should refuse to load (displaying a message like the plugins) if they don't match
14:35:56Coldtoastis there ftp access for the bleeding edge builds?
14:36:16B4gderwhy do want that?
14:36:37Coldtoastoh. I was going to set up SyncBack to check for updates
14:36:45Coldtoastbut it can only do it with FTP, no thttp
14:37:04B4gderso script curl to do it ;-)
14:37:25Coldtoastnever heard of curl
14:37:26B4gdercurl -z [previous download] URL -o newfile
14:37:35 Join RotAtoR [0] (
14:38:57B4gderits my baby
14:39:23B4gderalthough quite grown-up these days :-)
14:41:29preglowcurl supports telnet???
14:41:35preglowexactly how does one download from telnet? :P
14:41:44*preglow remembers zmodem...
14:42:21B4gderits hard to make it automatic, but telnet is a stream as well
14:42:46B4gderor can be at least
14:43:17B4gderthen telnet is not curl's strongest side
14:43:44preglowyeah, sure, i know it can be done, i just can't imagine a standard way
14:43:54 Join Godeater [0] (
14:43:55B4gderthere is no standard way, no
14:44:19preglowbut i guess that's more of a feature needing scripting
14:49:02 Join Boogder [0] (
14:49:23 Join edx [0] (
14:49:44Boogderall of a sudden my client refuses :-/
14:49:51Boogderresorting to web client
14:51:59amiconnHmm, logbot still suffers from amnesia
14:53:22 Quit Bger (""In the other world, in paradise, the beauty of women surpassed even the beauty of Bulgarian women" Adaloloddin Mohammed Balh)
14:53:30 Join einhirn [0] (
14:54:27amiconn /msg logbot explain * -> -logbot- *Shrug* I don't know
14:54:40amiconnHe used to be able to explain a number of things...
14:55:04Boogderzago reinstalled a long time ago
14:55:11Boogderand since then the explains are gone
14:55:37amiconnI know... but they could have been restored
15:02:17 Quit B4gder (Read error: 113 (No route to host))
15:24:38 Join [1]Chamois [0] (
15:33:11 Join Bger [0] (~Bager@
15:38:10 Join B4gder [0] (
15:38:17 Quit Boogder ("CGI:IRC")
15:38:35preglowgotta go
15:38:36 Part preglow
15:41:33B4gderthat latest commit looks weird to me
15:42:08 Quit Chamois (Read error: 110 (Connection timed out))
15:42:09 Nick [1]Chamois is now known as Chamois (
15:42:53B4gderSlasheri: shouldn't it be audiobuffer_pos + copy_n ?
15:44:57SlasheriHmm, nope :)
15:45:08SlasheriThe fillpos itself means audiobuffer_pos + audiobuffer_fillpos
15:45:12B4gderwell it looks weird
15:45:15***Saving seen data "./dancer.seen"
15:45:28B4gdersetting it to an absolute value when it can loop
15:45:41B4gderand then afterwords, you add 'length'
15:45:47Slasheriyep, but pcm_flush_fillpos will reset it to zero :)
15:45:51B4gderadding some comments in that code wouldn't be bad
15:45:56Slasheriso that should be ok
15:46:25Slasherisure, comments would be good..
15:46:28B4gderso why first set it and then reset it to zero if lengh >0 ?
15:47:05Slasheripcm_flush_fillpos looks from that variable how many bytes to add in a chunk, then it will reset it to zero
15:47:31B4gderso a if(length > 0) flush; else fillpos = copy_n would make sense
15:48:00B4gderjust trying to understand
15:48:01Slasherino, that fillpos has to be set before calling that flush. Just check the function
15:48:40B4gderI think it is dubious to make such a use of a global variable
15:48:50B4gderbut that's me
15:48:50 Join west-acre [0] (
15:49:31SlasheriHmm, yes.. I think the pcm buffering really is quite misleading at the moment :/
15:49:48 Join tvelocity [0] (
15:51:07godzirraIs there a place showing the changes between one day and the next?
15:51:10godzirrafor the daily builds?
15:51:36B4gderSlasheri: so why the while(length >0) if it never loops?
15:52:02B4gdergodzirra: yes, on the daily build page there's a link to it
15:52:10SlasheriB4gder: Hmm, it can loop one time (at least it should be able to do that)
15:52:12godzirraAhh, just found it. Thank you.
15:52:30SlasheriIt loops if buffer wraps
15:52:31B4gderSlasheri: no, if length is > 0 it gets set to zero you just said
15:52:43B4gderah no
15:52:44SlasheriNo, fillpos gets set to zero, not length :)
15:52:46B4gdersorry, me confused
15:52:50Coldtoastshould this work B4gder?
15:52:55Coldtoastcurl -z -o
15:53:09B4gderColdtoast: yes, I believe it will
15:53:20ColdtoastI get Illegal Dat eFormat
15:53:25Coldtoastrunnign it on WIN
15:53:38B4gderand there is an existing file?
15:53:48Coldtoastnot cos of format differences between *nix time and WIN?
15:54:21 Join Flemmard [0] (
15:54:31Coldtoasthmmm. hold on
15:54:32 Quit Maxime`Mrn (Read error: 131 (Connection reset by peer))
15:55:09Coldtoastah. doesn't matter. got it
15:55:37Coldtoastnice little app
15:57:23B4gderSlasheri: I really think you should make slightly more efforts on producing code that others can understand and follow
15:57:36B4gderit will reduce the load from you
15:57:44B4gderand allow others to help out more
15:58:41*B4gder imagines a low level driver to linux sounds dev instead of uda1380
15:58:51SlasheriB4gder: Hmm, maybe i could try to clarify the pcm buffering code, that is where most of the problems have happened.
15:59:09amiconnB4gder: ...and another driver to directsound?
15:59:15godzirraWhen I try and delete a directory, it gives me the normal prompt "Delete hit play, else hit anything" but when I hit play to delete the directory, it won't delete it.
15:59:27B4gderamiconn: I assume that is something on windows?
15:59:49B4gderit would be very useful for playback debugging
15:59:52godzirraShould I just report that in here for iriver? The website says don't report them on the site.
16:00:13amiconngodzirra: PLAY in that sense means SELECT on iriver
16:00:33amiconnThat's because we do not yet have platform specific strings
16:00:55amiconn...but hopefully will have them soon
16:02:16godzirraColdtoast: where did you have your wps again? For some reason when I copied the new iriver build onto my player, everything copied as 0 byte files.
16:02:17HClbleh, people suck.
16:02:51godzirrahowdy HCl. Sucky Monday to you too.
16:03:08HCli wasn't wishing people a sucky monday
16:03:17HCli was just saying that a lot of people are needlessly mean, immature and stupid.
16:03:20HClhow goes?
16:03:22godzirraI wasn't wishing you one... it was a statement. and yes, they are. :)
16:03:28HCl :/
16:03:33amiconnBagder: If someone implements that, we could also build a 'pseudo mas' using libmad and the pcm driver
16:05:31godzirraIS next song data available on irivers, out of curiousity?
16:10:20B4gderamiconn: that would also... be totally cool
16:10:22godzirraCrap. I got *PANIC* Updating size on empty dir entry 36
16:10:27godzirraand now I can't turn off my iriver.
16:11:05godzirraWhich part?
16:11:18HClUpdating size on empty dir entry 36
16:11:24godzirraThats what I got.
16:11:31godzirraon my iriver when trying to go into an mp3 directory.
16:11:32HClon rockbox or normal firmware? what does that even mean?
16:11:47godzirraI think it was because I didnt reset after deleting and updating my .rockbox directory maybe?
16:11:51godzirrathat was on today's daily build.
16:11:54godzirraon rockbox.
16:12:28godzirraafter resetting it worked fine.
16:12:34HClits never a good idea to not reset after an update.
16:12:46amiconnHmm, that's a fat driver panic
16:12:47godzirraYeah.. was early morning stupidity on my part.
16:13:26amiconnMaybe we should close all files by force before usb?
16:13:52amiconn(Same thing we do with hotswap if the card is removed) least all files opened for writing
16:15:09godzirraIt would keep it from killing stuff when people forget to reset, in that case, wouldnt it?
16:15:33HClsounds like a good idea to me, as long as you restore the open files afterwards.
16:15:39HClwhich may be impossible
16:16:00amiconnI think closing them is better even without reopen.
16:16:19godzirraWon't the code reopen them when it needs them anyways?
16:16:21amiconnWhen the fat driver doesn't find the entries it expects, it will panic (bad)
16:16:37HClif you don't reopen the files
16:16:41HCli know that at least the database
16:16:48HClwill crash horridly.
16:16:58godzirraoh. That'd be bad.
16:17:10amiconnHowever when it does find the expected entries, but these are the wrong ones, it may corrupt the filesystem
16:17:31amiconnHCl: The database is closed and reopened correctly
16:18:26HClif the database is the only thing
16:18:31amiconn..and the database is not a big problem (yet), cause it is read only
16:18:37HCli dunno whether we have other constant open files
16:18:46HCli have a patch laying around here that opens it in read write
16:18:50amiconnI'm thinking of .playlist_control
16:19:07amiconnThat is opened read/write
16:19:41godzirraWell, I got am emssage telling me .playlist_Control couldnt be read, and then it went away and started playing.
16:19:44godzirraafter I hit reset.
16:22:18godzirraAnother problem I'm having... with the daily build. If I hold down play to get my playmode menu, I get the new one (where I can hit left to change shuffle, right to change repeat mode and down to change show files) and when I change it to repeat mode all and shuffle mode on, it will only play files in the current directory... Is that normal?
16:22:39godzirranevermind.. apparently I need repeat mode shuffle.
16:27:30 Join n0bby [0] (
16:28:10n0bbyyet another feature request: wps codes for all the stuff thats normally in the status bar
16:28:23n0bbyfor new graphical wps's
16:35:08godzirraACtually I Was wrong... it still does only repeat in the current directory.
16:35:16godzirrano matter how you set the repeat mode...
16:35:26godzirraI cant figure out how to shuffle through all mp3s on the new iriver build
16:36:10B4gderit only plays the current playlist, yes
16:37:12godzirraThere's no way to shuffle through all mp3s on the device anymore?
16:37:17godzirraother than adding them all to a playlist?
16:37:23B4gderno, there never was
16:37:33godzirraOn the iriver firmware there was, which is why I ask.
16:37:58B4gderwe don't write that and never did
16:38:40B4gderand actually, I never understood how the iriver firmware did
16:38:46godzirraI realize that. I was just asking if there was a way to do that.
16:39:01godzirraOh well. I loved that function.
16:39:15B4gderI see, but with phrases like "anymore" you give the impression we have taken something away
16:39:21B4gderwhich we haven't
16:39:44godzirraSorry for my grammatical faux pas. I just changed to rockbox Friday so I didnt know it wasn't a feature.
16:40:32B4gderautomatically going from folder to folder is not easily done, without sacrifising some of our basic beliefs
16:40:59B4gderand it is so easily done by simply making a playlist and using that
16:41:18godzirraThats fine. I'm still getting used to how everything works. I havent made a playlist yet.
16:42:56ColdtoastI used to use dir skipping myself
16:43:08Coldtoastreally liked it
16:44:22godzirraHmm. About how long should it normally take when you hit "next" before it plays the next song?
16:45:57amiconnIt *should* be instantaneous if the track is buffered, and take disk spinup time plus a few 1/10 seconds if not. It currently takes much longer on iriver
16:46:26B4gderI figure crossfade makes it longer as well
16:46:34B4gderat least feel longer
16:46:48ColdtoastI tested withand without and there was no diff I don't think B4gder
16:46:59B4gdersure, but that's a problem
16:47:02B4gdernot how it should be
16:47:04amiconnB4gder: Imho it shouldn't crossfade when skipping tracks
16:47:09Coldtoastthought it might even be linked to the anti-skip buffer length
16:47:12Coldtoastbut no
16:47:18amiconn...but then that's me, I think it should *never* crossfade
16:47:28B4gderI kind of like crossfade on next
16:47:39B4gderat least sometimes
16:47:42ColdtoastI agree with amiconn. no crossfading when you skip tracks
16:47:59Coldtoastbut it doesn't work once you get to the end of the buffered tracks B4gder
16:48:02B4gderbut it kind of goes with the concept that crossfaded next is slower
16:48:16B4gderwhat doesn't work?
16:48:18Coldtoastso you get FADE... FADE... FADE...abdupt skip... abrupt skip... abrupt skip
16:48:30B4gdersure, if there's no data it can't fade
16:48:38B4gderthat's not much to do about it
16:48:47Coldtoastthen disble it all together on skip cos it sounds a bit didgy
16:49:02west-acrei cant seem to get hold of the grayscale patch 1162865 for iriver h1xx
16:49:05west-acreany suggestions
16:49:07B4gderso you rather have "didgy" all the time?
16:49:11west-acrethe sourceforge page dont worx
16:49:26Coldtoastmy idea was to disable crossfade for, like, the first 5 secs or something
16:49:31godzirraWhen I hit "next" I get about 4 seconds before the next song starts.
16:49:46Coldtoastso if you skip quickly along, no fading but if you leave it longer, it fades
16:50:15B4gderI think the current concept holds fine
16:50:27Coldtoastok. I disagree tho
16:50:28B4gderif it can, it fades, it if can't it doesn't
16:50:42west-acrei cant seem to get hold of the grayscale patch 1162865 for iriver h1xx
16:50:47B4gderbut I could agree with an option that disables fading on next
16:50:59B4gderwest-acre: so wait until the sf page is up again
16:51:08Coldtoastthat'd be better I think
16:51:10B4gderwest-acre: I trust you know the patch is a source code change
16:51:11godzirraDisabling fading on next would be nice, I agree.
16:51:27B4gdersf == sourceforge
16:52:02west-acreoh yeh
16:52:17amiconnwest-acre: I guess the grayscale patch won't apply cleanly anymore
16:52:28Coldtoastcool. got Windows crappy task scheduler to run curl every 20mins or so
16:52:35B4gderI'm quite sure it doesn't apply cleanly
16:52:52amiconnToo many gfx changes... with yet more to come
16:53:22B4gderI believe amiconn's path is the better grayscale road
16:53:48ColdtoastB4gder: if you disable crossfade on next, would you also do it on Prev?
16:53:58Coldtoastit tries to fade Prev too doesn't it?
16:54:02godzirraColdtoast: where is your wps file you use?
16:54:03HClmarkun made it as a quick hack to get it to work anyways...
16:54:10B4gderof course, it would be on skip either way imho
16:54:18Coldtoastyeah. cool
16:55:16 Quit n0bby (Read error: 60 (Operation timed out))
16:55:17godzirraAlthough maybe something else is wrong... when I hit my select button and hit the left arrow to go up a directory or right arrow to go down a directory, it takes about 3 seconds before it displays the folders.
16:55:31godzirraAnd I dont have 200 mp3s there or anything, they're ordered by Artist\CD\Mp3s
16:55:51B4gderthat's when it needs to spin up the disk I guess
16:55:54Flemmardgodzirra: you have defragmented your hard drive?
16:56:10godzirraFlemmard: ;p
16:56:19Flemmardit helps
16:56:24godzirraAre you serious?
16:56:35godzirraHow do I defragement my hard drive from linux?
16:56:39godzirraColdtoast: sweet.. thanks :)
16:56:44Flemmardwith the original firmware it speeds the startun
16:56:49Coldtoastjust unzip it into the .rockbox dir
16:56:52Flemmardgodzirra: erm good question
16:57:37godzirrawith no answers? :)
16:57:51godzirraI need knoppix for windows. lol
16:58:06Flemmardgodzirra: maybe there's a tool with knoppix
16:58:13B4gderI've never defraged my mp3 players
16:58:25godzirraB4gder: do yours take 3+ seconds to switch directories?
16:58:48B4gderif the disk isn't already spinning, I guess it is about that time
16:59:02FlemmardB4gder: wit the original firmware, it taked a long time to start, i've defragmented, i've won half of the time
16:59:06B4gderbut after the first switch it is, and then it is swift
16:59:10godzirraB4gder: you're right.
16:59:19godzirraB4gder: I just checked, the first one takes a few and the next ones are fast.
16:59:29B4gderthat's how rockbox works
16:59:48godzirraI think thats just how harddrives work isnt it? :)
16:59:55B4gderyes, and rockbox
17:00:11B4gdersincei it doesn't cache dirs
17:00:30godzirraSo is there anything a non-C programmer can do to help the project?
17:00:43godzirraI know just enough C to be dangerous, but I do lots of perl and sys admin stuff.
17:00:57B4gdersure, writing docs/wiki, testing, responding to posts/forums qs
17:01:23B4gderbuying beers for developers
17:01:25CassandraI wonder if all these crossfade fixes explain why I wasn't having stability problems with recent builds. I don't use crossfade. It annoys me.
17:01:45B4gderCassandra: sounds plausible
17:02:06B4gderI have it enabled and suffered a lot
17:02:32godzirraooh. how do you turn off crossfade?
17:02:47godzirraB4gder: I don't buy beer for anyone sorry...
17:02:54godzirraB4gder: however I do homebrew kickass stout.
17:03:17B4gderok, that'll do :-)
17:04:03godzirraGod I hate tftp already. Isn't there a way to do it without setting mac addresses manually?
17:04:08godzirraugh. wrong window.
17:04:09Slasheribtw, the crossfade still crashes. In fact i am just running more tests to figure out if the problem is really related to crossfade or not
17:04:21 Quit Godeater ("CGI:IRC")
17:04:34godzirracrossfade seems to work fine for me. I just don't like it :)
17:05:07Slasherioh.. in fact i forget to upload the newest fix to my iriver.. :D
17:05:13Slasherino wonder why it crashed ;)
17:05:43*Cassandra pokes sourceforge with a stick.
17:09:56godzirraIs there an easy way to create a playlist in linux for everything on my hard drive?
17:10:06godzirra(For my iriver, obviously.. not my linux box)
17:10:48CassandraErm, why not do it in Rockbox?
17:11:15godzirraIs there an easy way to say "add everything to my playlist"?
17:11:43CassandraRoot directory - menu - playlist options - create playlist should do it, I believe.
17:11:48B4gdergodzirra: for linux "find . -type f -printf "/%p\n" >full.m3u
17:12:25godzirraCassandra: sweet, thanks!
17:12:39CassandraMust remember to add a note about that to the manual.
17:12:40godzirrahrm... 10,000 files is the limit in playlists right?
17:12:46godzirraThat would be good. :)
17:12:49godzirranevermind. I hit 2.5k and stopped
17:12:50CassandraIt's configurable.
17:14:23godzirraahh ok. I just remember reading something in the faq.
17:14:29godzirrathe answers were quite humorous.
17:14:32 Quit tvelocity ()
17:14:36Coldtoastthe FAQ for iriver maybe
17:14:52B4gderthat limit is the same for all models
17:15:29godzirraNo, in the normal faq.
17:16:19 Quit DaKi][er ()
17:18:02amiconngodzirra: The limit is configurable; absolute maximum is 20000
17:18:30B4gderwhich must seem like endless to iriver users ;-)
17:19:52godzirraB4gder: are iriver harddrives smaller than normal or something?
17:20:05godzirraB4gder: mind if I add your how to make playlists on linux to the wiki?
17:20:15B4gdersmaller than Archos' ones, yes (1.8" compared to 2.5")
17:20:28godzirraOH, I knew that.. I meant smaller in diskspace.
17:20:42B4gderyes, the 1.8" are smaller diskspace-wise
17:20:55B4gder2.5" exists up 120GB
17:21:02B4gderup to
17:21:51godzirranevermind.. its already there. bleh.
17:21:56godzirrawow.. nice.
17:22:28godzirraI wonder how big the 1.8's get.
17:22:47B4gderthey've talked about a 80gb to be released this year
17:22:56ColdtoastI might buy a 2000mAH battery
17:22:56godzirraI'm trying to figure out where to buy one.
17:22:57amiconnB4gder: I just read about 160 GB 2.5" disks today
17:23:30amiconn...using perpendicular recording (new technology)
17:24:05godzirraWow. $150+ for 30 gig 1.8" drives.
17:24:09godzirraeasier to just buy an archos.
17:24:34godzirraand thats on pricewatch.
17:25:46amiconnThey're 5400 rpm
17:26:23 Join n0bby_ [0] (
17:29:55 Quit Moos ()
17:30:56 Join Moos [0] (
17:31:34 Quit Moos (Client Quit)
17:37:35 Join Moos [0] (
17:40:08 Join preglow [0] (
17:40:25preglowslasheri: these latest fixes work only for crossfading?
17:40:38preglowi believe there was a problem with I00 occuring yesterday
17:43:29Coldtoastwhat's I04?
17:44:35Slasheripreglow: yes
17:45:01Slasherii don't know if they are any unstability without crossfading
17:45:19***Saving seen data "./dancer.seen"
17:45:43ColdtoastI got near the end of a track earler and hit Next and rockbox crashed with an I04 error
17:46:04Slasheridid you have crossfade enabled?
17:46:22Slasheriok, might be related to that
17:46:34amiconnpreglow: I wonder what I00 actually is
17:46:42amiconnIt seems to have no name
17:46:52ColdtoastI thought it migh tbe, what with the bugs found over the past day or so
17:47:19ColdtoastI got an ATA Error -11 earlier too. Had that twice so far
17:47:38Coldtoastplugged my h140 into USB, powered on
17:47:52Slasheriyep, crossfade still crashes
17:47:52Coldtoastdisconnected USB and got that when Rockbox tried to boot
17:48:16SlasheriColdtoast: i have also got that a few times. Maybe some hard disk initialization problems
17:49:24godzirrahow odd. My crossfade doesnt work for the daily build from today.
17:49:29godzirraerr does work rather
17:49:48 Quit Coldtoast ("Peace and Protection 4.22")
17:53:55 Join TCK- [0] (
17:55:19 Join Coldtoast [0] (
17:56:28 Quit ashridah ("Leaving")
17:57:56bill20r3how can I power my H340 off usb while it's playing?
18:02:28 Quit TCK (Read error: 60 (Operation timed out))
18:03:18preglowwhat makes you think we know? rockbox isn't even ported to h3x0 yet
18:03:47bill20r3just hoping that someone may know.
18:04:07godzirraCassandra: Any chance of getting more info on how the gameboy emulator works as well, in the docs? I'm not sure how to push "start" or "select" :)
18:05:14 Quit west-acre ("—I-n-v-i-s-i-o-n— 2.0 Build 3515")
18:05:31Coldtoastbill20r3: have a look in the manual
18:05:59Coldtoastif you set it to Charge mode, does that do it?
18:08:17*bill20r3 looks
18:09:35 Quit n0bby_ (Read error: 110 (Connection timed out))
18:11:08bill20r3looks like you can disable usb-charging, but that's not really what I'm after
18:11:37Coldtoastdisable? I thought you were asking how to ENABLE it?
18:11:48Coldtoastyou want to power it and not charge it?
18:13:14bill20r3I want to power and/or charge it, but while it's playing
18:13:25bill20r3but instead as soon as I plug in usb it goes to the 'data connected' screen
18:13:41Coldtoastas I said. change the USB mode to Charge and see if you can play tracks
18:13:42bill20r3so it's powering it like I want, but it stops it from actually playing.
18:13:59Coldtoastyou need to change the mode from Data to Charge
18:14:16bill20r3I think it's on charge, but I'll doublecheck
18:14:38bill20r3yeah, it is.
18:15:18Coldtoastok. and it still says "Data Connected" or something?
18:17:18Coldtoastwell isn't that just STUPID then?
18:17:24bill20r3yeah, it is.
18:17:40bill20r3maybe I can get a cable with the data lines cut, and just the power connected.
18:17:51bill20r3I'm just trying to avoid having to bring my charger into my office.
18:18:08Coldtoastmake one
18:18:18bill20r3yeah, shouldn't hard.
18:18:26Coldtoastnot at all
18:18:34Coldtoast\you can buy em
18:19:39Coldtoasttho.. I don't like the quality of their cables at all. I bought a USB/Charge cable for my h140 and the third time I went to charge, the cable was dodgy and I had to get it in JUST the right position
18:19:51Coldtoastuntil eventually it stopped working all together
18:20:07Coldtoastin fact, I'd go so far as to say the cable I bought off of them was shit
18:20:54bill20r3I've probablly got a spare usb cable lying about I can modify and use
18:20:59 Quit B4gder ("go go go")
18:27:50HClSlasheri: does your proposed patch to me still apply or do you have something better by now?
18:28:00HCli was wanting to start developing again tomorrow
18:30:35 Join Sucka [0] (
18:50:23SlasheriHCl: it does apply as before :)
18:58:10amiconnBagder: ar64nd?
18:58:17amiconnErm, around?
18:58:22amiconn(stoopid numlock)
19:04:12 Join webguest53 [0] (
19:04:37webguest53i don't know if this is a known prob or not but
19:04:53webguest53holding play button was used for page scrolling
19:05:13webguest53but now, at the same time, it pops up that mini menu...
19:05:20amiconnAH, yes
19:05:45*amiconn points to Cassandra
19:14:55 Quit webguest53 ("CGI:IRC")
19:16:00 Join matsl [0] (
19:18:06 Join Stryke` [0] (
19:24:47HClSlasheri: k
19:24:52*HCl will work on it tomorrow
19:25:58Stryke`last night, i had what i think is a nice idea: the ability to treat a playlist as a virtual folder, which you can browse, etc.
19:26:27amiconnYou can browse a playlist
19:26:38Stryke`only after playing it, right?
19:28:12amiconnxyz.mp3 -(context menu) -> playlist -> view
19:28:47Stryke`it seems all my best ideas are already implemented ;-)
19:28:51 Nick Lynx_ is now known as Lynx_awy (
19:28:53godzirraAlways a good sign :)
19:31:29amiconnOf course I meant m3u, not mp3 ;)
19:32:15Stryke`yeah, i gathered
19:33:41Coldtoastmy h140 is SO odd
19:33:57Cassandraami: Any suggestions for how to fix that?
19:33:59Coldtoastone build the LCD inversion has stripes missing the next, it's solid
19:34:33Stryke`Coldtoast: i know i've seen verticle stripes on my 120
19:34:46Stryke`i forget which build, haven't tried it in a while
19:35:34Coldtoastweird stuff
19:35:51Stryke`ill check now, latest daily build
19:36:04 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
19:36:28Stryke`lots of stripes
19:37:13Stryke`no one else has seen these?
19:37:30godzirranot I
19:37:33godzirraI'm on a 120 as well
19:37:45godzirraI almost just said "take a screen shot and show me"
19:37:59Stryke`ill do a screen dump
19:38:08godzirraooh neat. didnt know you could dothat.
19:38:46amiconnStryke`: I guess the stripes won't show up in a screen dump
19:38:47Coldtoastdoes it go: thin black slightly thicker black, thin black, thick black, thick, thick, thin, very thick, medium, thick, thin, thin, thin,medium, thick, medium, medium?
19:38:58Coldtoastthey don't amiconn. I tried
19:39:02ColdtoastI can take a pic tho
19:39:10Coldtoastgot my camera not too far away
19:40:10amiconnMaybe the lcd init is incomplete somehow
19:40:17Stryke`yeah, it just takes a normal shot (non-inverted)
19:40:37amiconn...and sometimes an internal value of the lcd controller is wrong because of that
19:41:05amiconnI'll look into the init anyway when switching to greyscale
19:41:07Stryke`at least now, its more than an isolated problem
19:41:59Coldtoastit's not any sort of a hassle but interesting that one build it's fine, the next, stripey
19:42:10 Join Zoom2 [0] (
19:42:18Stryke`is it possible we have early runs? when did you buy?
19:42:38ColdtoastNov last year but it's branded as an iHP
19:42:50Stryke`oh, mines approaching 2 years
19:43:04 Quit preglow ("CGI:IRC")
19:43:06amiconnColdtoast: I'd rather think the behaviour is random
19:43:28amiconnIf you boot the same build again and again, it'll sometimes happen, sometimes not
19:43:59amiconnThe probability seems to depend on the individual device, e.g. I never observed it
19:44:13ColdtoastI'll boot a bunch of times and see what happens :)
19:45:24***Saving seen data "./dancer.seen"
19:48:27Coldtoastyeah. it is a random thing
19:48:40godzirraMines approaching a year I think. And I don't have it.
19:48:48godzirraabout november is the same time I bought mine.
19:48:50godzirralast year.
19:49:28 Join preglow [0] (
19:49:37preglowSlasheri: you read the mailing list?
19:50:17MO-PantsuGlastonbury 2005 :D
19:50:18Coldtoastbooting doesn't change anything but powering off and on a few times, the stripe pattern changed
19:51:09amiconnColdtoast: Yes, I meant cold boot, not reboot
19:51:24amiconnA reboot doesn't switch off the lcd
19:51:39Slasheripreglow: yes, the buffering problems?
19:51:59preglownegative codec buffer value
19:52:01preglowthat's strange
19:52:07Slasheriyep, that's really bad
19:52:14preglowand never above 10% pcm buffer?
19:52:16preglowthat's insane
19:52:17Slasherii have even myself noticed it sometimes
19:52:34preglowespecially for a 192 kbps mp2
19:52:52Coldtoastamiconn: non-inverted is perfect tho. I have the line cursor set to Bar and that's perfect too
19:52:56Slasherihmm.. i will need to find some mp2 to test the performance..
19:53:05SlasheriMaybe i have some DAB recordings
19:53:17preglowi've got some mp2 files
19:53:21preglow192 kbps, as it happens
19:53:24preglowthey work perfectly
19:53:32preglow44.1khz, tho
19:54:10preglowit does, however, not work very fast :/
19:54:47preglowit is conceivable resampling might choke it
19:54:59preglowas the current resampler is dog slow
19:55:39preglowouch indeed
19:55:44preglowit BARELY runs when buffering
19:55:53preglowit's constantly at the 10% mark
19:56:00preglowresampling would have choked this easily
19:56:22preglowi wonder why the flaming hell mp2 is so slow
19:56:35preglowit should be faster than mp3, really
19:57:12Slasheriwhile buffering, the system tries to keep the audio buffer around 10% by slowing down reading from hard disk
19:57:26preglowwill it always succeed?
19:57:32Slasheriit should
19:57:43preglowto the point where it doesn't buffer at all if it needs to?
19:57:47Slasheriit will give all performance to codecs if that's necessary
19:58:03preglowif it notices it can't buffer without having pcm empty itself, it should give a warning in some way
19:58:10preglowthat might save us a lot of problems
19:58:14Slasherihmm, a good idea
19:58:27preglowpeople WILL experience this, once we get heavier dsp going
19:58:37preglowand having an explicit warning will save us much debugging
19:59:03preglowwell, ok
19:59:10preglowresampling will kill 192kbps mp2 easily, as far as i can see
19:59:19preglowi'll reply to him
20:01:34preglowSlasheri: i think we should disable dithering for now
20:01:36 Quit TCK- (Read error: 104 (Connection reset by peer))
20:02:11Slasheripreglow: ah, yes. Would you like to do that? (just changing the one true to false should be enough)
20:02:46preglowwill do
20:03:00 Join TCK [0] (
20:04:31 Join hubbel [0] (
20:05:24MO-PantsuUS Supreme court ruled 9 to 0 against Grokster peeps. Just in case you missed it.
20:05:46bill20r3the streets are finally safe for megacorporations
20:05:49preglowthat dithering is _SLOW_
20:07:40preglowhow can it be so slow
20:07:41preglowit's a miracle
20:09:40 Join Strath [0] (
20:12:03godzirra9 to 0 against grokster peeps for what?
20:16:41 Join n0bby [0] (
20:20:01 Join bagawk [0] (
20:20:49 Join hardeep [0] (
20:21:38godzirraOh. the p2p one?
20:21:52godzirraI'm going to find someone who got shot and go sue the gun company now.
20:22:58preglowSlasheri: that static int channel; isn't very elegant
20:23:46Slasheripreglow: yes it's not. feel free to change that :)
20:23:57preglowmight when i have time
20:24:02preglowi need to make a stereo resampler
20:24:08preglowthat does both channels in one pass
20:24:57preglowall the resampling calculations are duplicated now, which is kind of a waste
20:24:57bagawkhey hardeep
20:25:09preglowi also need to make downsample work in-place, so we save a memory copy
20:25:31preglowwell, that might not matter too much anyway
20:25:39preglowsince the arrays probably never are in iram
20:26:05hardeephi bagawk
20:26:43preglowslasheri: or might they be?
20:26:46 Quit bagawk ("Leaving")
20:27:51SlasheriHmm, i tested to put them in iram but i didn't notice any performance improvement
20:28:07SlasheriYep, maybe i should try that again..
20:28:20preglowi have always, without exception, gotten gains
20:29:22preglowbut yeah
20:29:26preglowi'll fix it later some time
20:29:36Slasherigood :)
20:30:22preglowbtw, what is the rockbox policy on using long longs in non-critical parts of the code?
20:30:46preglowthe resampler delta calculation should really either be using long longs or floats
20:30:59preglowas it is, it will overflow badly for sampling rates above 65khz
20:31:00 Quit Coldtoast (Read error: 54 (Connection reset by peer))
20:32:31Plugh_someone tell bagawk to give me more than 30 seconds to respond to a PM
20:34:42 Quit Sucka ("a bird in the bush is worth two in your house")
20:43:42 Join Sucka [0] (
20:48:27 Quit n0bby (Read error: 110 (Connection timed out))
20:52:55 Quit Cassandra ("If I were actually witty, this quitline would be funny.")
20:55:56Slasheripreglow: Hmm, what do think about changing all size_t in pcm_playback and playback.* to long?
20:56:00 Quit Stryke` (Read error: 110 (Connection timed out))
20:56:02SlasheriThat might solve some problems
20:57:50Slasherioh, it might even solve the crashes.. I think i do it
20:59:20preglowhow can they even be used interchangably?
20:59:23preglowsize_t is unsigned
20:59:47Slasheriyes.. and there might be some comparision problems..
20:59:57Slasherii think if everything is signed, the code works better
20:59:57preglowand anyway
21:00:00preglowwhat should it change?
21:00:03preglowthey're both 32 bits
21:00:16Slasherisome MIN and MAX comparisions at least
21:00:21 Join Coldtoast [0] (
21:00:32Slasheriat least i got different values with logf now :)
21:00:36SlasheriIt "looks" better
21:00:39 Quit TCK (Read error: 104 (Connection reset by peer))
21:00:42SlasheriI will do more testing
21:00:50 Join Lear [0] (
21:01:24preglowif you got different values, it's undoubtedly something that needs checking anyway
21:02:23LearWhen making a diff with cvs, can you get it to include new files as well? Easier with only one file to upload to the patch tracker...
21:03:59amiconnOh, Coldtoast is gone
21:04:23amiconnAnybody else here who had that intermittent striped display problem with inverse mode?
21:06:41amiconnHmm, it seems I should look more thoroughly
21:10:15preglowamiconn: any thoughts on my 'long long' question?
21:10:48 Quit Sucka ("Leaving")
21:10:51amiconnWe have to keep an eye on code size, but on iriver, this should be less critical
21:11:03preglowit's only an issue for software codec platforms
21:11:12preglowwhich shouldn't have large space constraints
21:11:13preglowand besides
21:11:15amiconnlong long is definitely less evil than floating point
21:11:16preglowit's just one place
21:11:52amiconnYes, but that one place lets gcc include some library functions, depending on what operations you will use
21:12:05preglowthat's right
21:12:08preglowi forgot it doesn't inline code
21:12:23amiconnIt may depend
21:12:30preglowfor 68k-gcc it doesn't
21:12:34preglowi checked it some time ago
21:12:41Mooshi guys
21:12:43preglowit does like with floating point, and adds some libgcc functions
21:12:48amiconnAddition is likely to be inlined, but multiplication probably not
21:13:09Moosamiconn: i've got this "problem"
21:13:13amiconnWhat about sacrificing that last bit instead going for long long?
21:13:24preglowi don't want to sacrifice precision when i don't have to
21:13:29preglowbut anywho
21:13:34preglowit's not much of an issue
21:13:48preglowusing the linear interpolator for anything above 65khz will sound very, very bad
21:13:58preglownow that i think about it
21:14:05amiconnMoos: I'm just preparing a test build with fixed up display init. Hold on
21:14:29amiconnpreglow: Didn't you talk about sophisticated filtering in the iriver firmware?
21:14:34preglowamiconn: indeed
21:14:46Learpreglow: I'm about to submit a patch with int64_t... :)
21:14:53preglowLear: doing what?
21:15:13Learpreglow: better metadata support for ogg vorbis (comments and stuff).
21:15:15preglowamiconn: i guess them being able to do such things really demonstrates how much faster their codecs are than ours :/
21:16:39preglowlear: do you happen to know how hairy it will be to support multiple ogg streams in one file?
21:16:59preglowi'd love for us to support that
21:17:29Learpreglow: well, pretty much, yes. At least the reading the information part. Not sure how it should be shown to the user though.
21:17:54amiconnMoos, Coldtoast: Here is a test build (current cvs with modified lcd init):">
21:17:59preglowlear: just wonder how seeking, for example, would handle that, there's not advanced warning that there are multiple ogg streams in one file, no?
21:18:30preglowunless i remember incorrectly, you can pretty much just concatenate a couple of files into one file and it'd be perfectly legal
21:18:46Learpreglow: as it is now, my patch detects if there are several streams and calculates total play time (well, sort of), but doesn't support reading tags for the other streams (the codec could be used for that as well, so it might not be needed).
21:19:00preglowlear: nice
21:19:12 Join Sucka [0] (
21:19:18Learpreglow: it's easy to detect chaining, and seeking isn't much harder because of it; just need to keep track if which stream to seek in.
21:22:58Moosamiconn: it's work very fine now
21:23:16amiconnTried switching off/on multiple times?
21:23:25Moosi trying
21:24:23Slasheripreglow: i fixed the negative buffer indicator, unfortunately it wasn't a bug i think
21:24:37Moosamiconn: never work for me before this
21:24:49amiconnNice :)
21:25:08Moosi tested stop an reboot few times and always work, good work
21:25:18amiconnBtw, I just reordered the init sequence according to the datasheet, and inserted the required pause of 100 ms
21:25:26amiconnThen I'll commit this
21:25:44 Quit Zoom2 ("CGI:IRC (EOF)")
21:27:20preglowslasheri: how?
21:28:22Slasheripreglow: codecbuflen was increased before cur_ti->available was increased. So the buffer length went temporarily to negative side. Of course that was a bug but i doubt it caused any harm
21:30:44Slasherioh, that was vice versa. codecbuflen was increased too late..
21:30:49preglowseems people are starting to ask for ipod rockbox
21:31:03 Join Stryke` [0] (
21:33:29preglowso it was just a bug in the display
21:35:27Slasheripreglow: oh!
21:35:38Slasherithe mp3 decoding is now ultra fast with dithering disabled =)
21:36:13Slasherionly about 20% boost ratio now and with dithering ~60%
21:36:51Slasheriit's even faster than vorbis i think
21:37:43 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
21:37:53 Join Stryke` [0] (
21:38:19preglowSlasheri: yes, i know
21:38:26preglowSlasheri: like i said, the dithering is rEMARKABLY slow
21:38:27Slasheriboost ratio ~15% for 128k mp3 file
21:38:39Slasherithis is cool
21:39:34Learslasheri: I browsed the UDA1380 spec today. AFAICT, it supports 24-bit audio. No need for slow dithering then. :)
21:39:53hubbellear: but the CPU only supports 16 bits when using DMA
21:40:13SlasheriHmm, 24 bit audio would be great but that dma support is a problem
21:40:15Learhubbel: argh! so there was a reason it wasn't used.
21:40:24preglowhubbel: do you know what the impact will be on using interrupt driven audio?
21:40:46preglowslasheri: it is faster than vorbis, it's always been
21:40:50Learand I guess the hardware implementation prevents sample rate changes; or is it just that we don't know how to do it?
21:41:09preglowlear: it allows them, but only 44.1khz, 22.05khz, etc
21:41:26hubbelpreglow: not really but we can only write 5-6 samples at the time, so one interrupt for every 5-6 samples
21:41:36preglowlear: and then, we have the optical out to worry about too, that can support a minimum of 32khz
21:41:47preglowlear: so we will probably opt for never leaving 44.1khz
21:41:53preglowthen again, we might make it an option
21:41:55preglowi can't see why not
21:42:07preglowwith all the decision making removed to dsp.c, this should be fairly easy
21:42:16preglowhubbel: ouch
21:42:17hubbelpreglow: definitely
21:42:48 Part hubbel
21:43:45Coldtoastamiconn: you fixed the LCD issue
21:44:09preglowalways using 44.1 should be a nice default
21:44:20amiconnBoost ratio ~25% for 448 kbps mp1. Was >80% before
21:44:38preglowyes, i once again feel obliged to comment on how slow that dithering is
21:44:42Stryke`amiconn: your fixed LCD build, its non-dithering as well?
21:44:55Stryke`thanks so much
21:45:03amiconnLatest bleeding edge already has the fix
21:45:26***Saving seen data "./dancer.seen"
21:45:57ColdtoastI'll never actually USE inversion myself but ther was an issue. good to know it's sorted
21:46:42Learhe, trying a 128 kbit file, CPU at 48 Mhz, boost < 8% before the large buffer fill started.
21:46:47Learmp3 that is...
21:48:36 Join ep0ch [0] (~ep0ch@
21:48:48preglowi think it's very possible for libmad to be realtime at 48mhz for some files
21:48:49ep0chwhat have you just done to mp3 playback?
21:48:56preglowep0ch: raped it in the ass
21:49:08ep0chshall i just read the log?
21:49:16ep0chnice job
21:49:18preglowi disabled dithering
21:49:28Learso I tried an ogg (Q5). Boost rate about 30%...
21:49:30preglowdon't tell me you can tell the difference
21:49:30ep0chahh like i said the other day
21:49:50amiconnBoost ratio ~20% for 384 kbps mp2
21:49:55ep0chi can see a huge difference in the debug screen :)
21:49:59preglowoh yes
21:50:02preglowit's heaps faster
21:50:05ep0chhell yeah
21:50:14ep0chnow for wavpack :p
21:50:21preglowwavpack has always been fast
21:50:28preglowand lossless codecs dont need dithering anyway
21:50:39preglowwavpack doesn't even need to boost the cpu here
21:50:48ep0choops i meant mpc
21:50:58amiconnpreglow: They might if the lossless source is >16 bits
21:51:06preglowamiconn: ah, sure
21:51:15preglowamiconn: and i'm thinking we'll want to dither if we do any processing
21:51:19amiconn(not common, I think)
21:51:21preglowep0ch: you do it, i've had it with the stupid codec
21:51:40preglowi just got i00 on a wavpack file
21:51:42preglowfirst file i played
21:52:17preglowit's consistent
21:52:18preglowwhat the ehll
21:52:49ep0chwhy does the pcm buffer need to be so big?
21:53:05preglowfor increased efficiency and gapless, i guess
21:53:05ep0chhelps with crossfading?
21:53:09preglowi'd like to see it become smaller
21:53:40ep0chi'd rather see the "when to top up the buffer" value lowered
21:53:43preglowalso, the current cpu boosting strategy works best with larger buffers
21:54:05ep0chso the cpu aint switching to boost all the time
21:54:29ep0chbut am sure there's a reason for it being what it is...
21:55:23preglowwhy does my lossless wavpack file crash
21:55:46amiconnHmm, with resampling the boost ratio is a bit higher. Still, resampling seems to eat less cpu than dithering
21:55:59preglowamiconn: and that's really saying something, if you ask me
21:56:00amiconn~42% with 320 kbps 48 kHz mp3
21:56:33preglowas it is, i should be able to increase resampling efficieny easily
21:57:10ep0chusing asm?
21:57:19preglowusing a clevrer stereo strategy
21:57:24preglowasm can come later
21:59:03preglowall the deltas and positions only need to be calculated once
21:59:10preglownow it happens twice
21:59:21preglowall the looping also happens twice
22:00:11amiconnThe saved performance should go into some more sophisticated resampling, imho
22:00:29amiconnUpsampling doesn't sound nice for higher ratios...
22:00:47preglowoh no it doesn't
22:00:52preglowwe should use fir filters for upsampling
22:01:05preglowand again, lucky us, the emac unit is ideally suited for fir filtering
22:01:39amiconnI'm curious whether and how iriver used the emac
22:01:47preglowmost definitely
22:02:04amiconnPerhaps they use a compiler with emac support?
22:02:05preglowyou'd be foolish not to use it for for example resampling
22:02:12preglowi don't know if there is such a beast
22:02:20preglowi'm under the impression iriver uses assembler a lot
22:02:23Learha, I'm pretty sure I have dithering enabled now, yet CPU boost is only about 17-20%
22:02:36Learthat's for 128 kBit MP3...
22:02:42preglowlear: what did you change?
22:02:48preglowlear: i doubt that
22:02:53ep0ch128 kbps was about 50%
22:03:03ep0chthis morning anyway
22:03:33amiconnFound a bug :(
22:03:35Learwell, first I removed that nedless & 0xffffffff in prng (though I doubt that made much difference), then I moved the dithering buffers to IRAM. :)
22:03:38preglowin what?
22:03:51amiconnI just played a 22 kHz mp3 track
22:04:09amiconnPlayback hung at end of track
22:04:17amiconn...instead of playing the next
22:04:25 Quit RotAtoR ()
22:04:40amiconnrockbox is completely frozen
22:04:54preglowamiconn: no secret that playback currently has bugs
22:04:57*amiconn takes paperclip
22:05:05preglowi too have managed to make rockbox just hang
22:05:08preglowno exception or anything
22:05:54Coldtoastshould release a Rockbox h1x0 case that comes with a paperclip
22:06:12Coldtoastheh. actually. I got the first crash in days today
22:06:52Leardecode is slower, but boost hovers at 21-22 percent now
22:08:25amiconnHmm, the resampler seems to produce some crackling noise in the 22.05->44.1 case
22:08:52preglowamiconn: oh?
22:09:06preglowwhat amplitude are we talking here?
22:09:31 Quit Flemmard (Read error: 104 (Connection reset by peer))
22:09:53amiconnIt's noticeable, but not as loud as the music itself
22:10:54preglowi haven't got a 22.05khz file
22:10:57amiconnGot 2 hangs again
22:11:07preglowas a matter of fact, i've never even tried upsampling
22:11:09amiconnIt's also noticeable with 24 kHz, but not wih 32
22:12:15 Join Maxime`Mrn [0] (
22:12:32preglowit might be several things
22:12:32amiconn16, 12, 11 and 8 kHz have it too.
22:12:46preglowit's a miracle 8khz even works ;)
22:12:55preglowi don't believe the resampling buffer is large enough to contain it, haha
22:13:16amiconnApart from that noise it works, like the other sample rates
22:13:25preglowany way of giving me a test file?
22:13:50preglownow what the hell
22:13:59preglowsomeone with sense in this country?
22:15:24Learpreglow: the idea isn't new. Wasn't it Venezuela or something that said something similar a couple of years ago?
22:16:23preglowi can remember someone having said similar things
22:16:24Learpreglow: to promote access to government documentation for everyone, and also the local IT industry
22:16:39preglowi just didn't expect it from anyone in this particular government :P
22:18:54LearNah, it was Peru, according to a ./ comment...
22:19:13Learoops, s/.///./ :)
22:21:25amiconnpreglow: pls check pm...
22:23:42 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
22:23:47 Quit preglow ("CGI:IRC")
22:23:50 Join Stryke` [0] (
22:24:00 Join preglow [0] (
22:24:03preglowamiconn: let rip
22:24:44Slasheripreglow: Hmm,
22:24:52SlasheriDo you think we have now a true gapless mp3 playback?
22:25:17Slasheriindeed i fixed one bug from the codec mpa..
22:25:19preglowno, i don't
22:25:36Slasheriok, but anyway please test if you can :)
22:25:45preglowoh, you did?
22:25:48preglowwhat was the bug?
22:25:53Slasherijust a moment, i find it
22:26:00preglowplease do
22:26:42Slasherithe code inserted sample to audiobuffer which was Synth.pcm.length in size. I changed that size to Synth.pcm.length - start_skip
22:27:15Slasheriso before that change, random bytes got inserted to pcm buffer if start_skip was set
22:27:26pregloware you completely certain?
22:27:26Slasheriyes :)
22:27:32Slasherii think so
22:27:32preglowi have never ever heard random bytes
22:27:37preglowand i know what they sound like ;)
22:27:39preglowbut anywho
22:27:43preglowi'll check the gapless
22:27:50Slasherithey might have been zeroes also
22:27:56ep0charghhh page scrolling is almost useless now thanks to the quick menu :(
22:28:02preglowyou have any idea why things crash so much now, btw?
22:28:15Slasheripreglow: i don't.. :/ i am trying to figure it out
22:28:33ep0chmove page scrolling assigned key to the record button?
22:28:47Slasherii will leave the unit playing without crossfade when i go to sleep and see tomorrow if it has crashed
22:28:59Learpage scrolling? :)
22:29:01preglowit crashes without crossfade, oh yes
22:29:06preglowall the time
22:29:14Coldtoastep0ch: how many pages?
22:29:15ep0chyeah holding down play while scrolling the filetree
22:29:21amiconnSlasheri: I have crossfade disabled since it exists, yet I get frequent hangs with my testset
22:30:10Slasheriamiconn: ah, ok. i will try to find some problems but that's not easy
22:30:54Slasheriat least i hope it gives no more any I00 errors
22:30:59Learep0ch: ah, so that's how to reach the quickmenu. Makes it easier to toggle shuffle (I've never needed page scrolling anyway).
22:31:17Coldtoastme either
22:32:05ColdtoastI don't have any more than 2 pages so I can just move the joystick UP while at the first litem and is wraps around to the last.
22:32:34ep0chhmm i have just over 6
22:32:51Coldtoastman. really?
22:32:53ep0chlots of artists
22:32:57Coldtoastheh. organise your player better :)
22:33:00ep0chand default font
22:33:14ColdtoastI have the alphabet split up
22:33:17 Quit Chamois (Read error: 104 (Connection reset by peer))
22:33:20Stryke`you wouldn't believe how many pages i have using artist/album/track
22:33:26amiconnPaged scrolling can be nice. My testset folder has 210 files == 14 pages
22:33:57Coldtoast@ for stuff I download from websites, A-E, F-J, K-O, P-T, U-Z
22:33:58ep0chpage scrolling could be enhanced by scrolling by alphabet and not page
22:34:12ep0chand not calling it page scrolling :p
22:34:14 Join Chamois [0] (
22:35:42Slasheripreglow: yes, i think it's 100% gapless now :)
22:35:48preglowSlasheri: do you have any idea how large buffers might be passed to the resampler?
22:35:55 Join Flemmard [0] (
22:35:59preglowthe delta precision isn't good enough
22:36:13preglowso i either need more frac bits, or i need to split the delta into a integer part and a frac part
22:36:19Slasheripreglow: with mp3s around 1 kB, oggs 4 KiB
22:36:50preglowsure? and mp3 frame is 1152 samples
22:36:54preglow*4 is 4kb
22:36:56Slasheribut dsp limits that size also
22:36:59 Quit matsl (Remote closed the connection)
22:37:05SlasheriNot very many bytes..
22:37:20preglowi think i'll split the delta into frac and integer parts
22:37:23preglowthat way i can take all buffer sizes
22:37:27preglowand get ultimate precision
22:37:37Slasheriok, sounds good :)
22:37:46preglowtesting gapless
22:38:12Slasheripreglow: MAX_CHUNK_SIZE in dsp.c is the maximum amount of data passed to resampler
22:38:15preglowno, it's not completely gapless i'm afraid
22:38:28preglowi still get a small glitch
22:38:28Slasheriand the *4 should be removed from samplebuf
22:38:31preglowin a silent part
22:38:39SlasheriHmm :/
22:38:49Slasherii didn't hear any glitches
22:39:04ep0chthis for a lame encoded mp3?
22:39:25preglowSlasheri: no, it's a very problematic transition, i don't even think foobar gets it 100% right
22:39:41preglowso for all intents and purposes, mp3 is gapless for lame _VBR_ (this can easily be fixed) files at the moment
22:39:52preglowBUT, i'm still not sure how to handle files that do not have lame headers
22:40:16ep0chfoobar doesnt do that, and foobar sets the standard :)
22:40:29preglowfoobar doesn't do what?
22:40:31Stryke`foobar does, however, have a DSP Gapkiller
22:40:34ep0chmaybe some dsp could iron out the glitch
22:41:20preglowStryke`: foobar doesn't need one
22:41:28preglownot for my files, at least
22:41:31Stryke`nor mine
22:41:36Stryke`but i know it exists
22:41:39preglowep0ch: the glitch cant be heard for normal files
22:41:46preglowi've just found myself a particularily nasty transition
22:41:53preglowand i heard more of a gap in the left channel, for some reason
22:41:56preglownot gap, really, but a gltich
22:48:07 Join ]RowaN[ [0] (
22:49:56 Quit Maxime`Mrn (Read error: 110 (Connection timed out))
22:50:02]RowaN[i think my iriver is fried... rockbox crashed and i couldnt reset it with a pin until about 2 hours later (was travelling from work) .. the hd led was on perminantly, and now it cant seem to read the hd
22:50:50Slasheriyou should try a paperclip instead..
22:51:00Slasheripins are not good and may damage the reset switch
22:51:04]RowaN[dont carry them with me either i must say
22:51:27Learcould low batter explain the hd problem?
22:51:43]RowaN[nah ive got it plugged into the mains now and same prob
22:51:45preglowlow batt does give weird things
22:51:48preglowlet it charge
22:51:51preglowTHEN try again
22:52:00 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
22:52:07 Join Stryke` [0] (
22:52:17]RowaN[its been charging for about 20mins
22:52:43Slasheri]RowaN[: did you let the battery go out or has it been constantly hanged until you plugged in the charger?
22:52:58SlasheriIf that's the case, you haven't managed to push the reset button correctly
22:53:24]RowaN[battery didnt die.. hd was still spinning when i got home (loveryly 2 hour commute i face daily in each direction)
22:53:35Slasheriok, then try to find a paperclip :)
22:53:42]RowaN[i did reset it when i got in
22:53:42Slasherithat will solve the problem
22:53:48 Join wintermute [0] (
22:53:48]RowaN[only way i could turn it off
22:54:03]RowaN[it was very warm =/
22:54:16Slasheriand now it doesn't start?
22:54:27wintermuteAm I correct in assuming that the current tag database only works on iRiver builds?
22:54:45]RowaN[it starts... stuggles to bring up the initial list of dirs, then when i try to go into a dir the hd access led is very agressive but it doesnt get anywhere
22:54:52Bagderwintermute: no
22:55:02Slasheriah, really weird
22:55:16Slasherimaybe the hdd has failed if your unit is old..
22:55:21]RowaN[anyway, i dropped my iriver last month for those who remember, this is just the final nail in its coffin
22:55:39wintermuteBagder: I just rebuilt the database with the on the TagDatabase page, and Rockbox throws a 'Unsupported database version: 3' error on reboot.
22:55:46Slasheriyou should also try check the disk for bad sectors with computer
22:55:55Bagderwintermute: then you need to update your rockbox
22:56:08]RowaN[scandisk wouldnt know a bad sector if one bit it in the ass
22:56:14wintermuteBagder: I just installed the current daily build.
22:56:35wintermuteBagder: Except I forgot to flash it. Damn. Hold on.
22:56:53*Bagder holds
22:57:03*amiconn spots Bagder
22:57:06preglowSlasheri: man, don't you zero out the resampler structs anywhere?
22:57:22Slasheripreglow: Hmm.. :D
22:57:28amiconnBagder: Some more ideas concerning extended localisation handling...
22:57:30Slasherino, i don't do that ;)
22:57:32preglowit's got a pos variable, you know, heh
22:57:42preglowcould be reading the miles of bounds
22:57:54preglowout of bounds
22:58:09Slasheriok, it should be zeroed then
22:58:23preglowoh yes indeed
22:58:32wintermuteBagder: Okay. Reflashed, and it reads the database, but it gives me the same problem it's always given me, namely:
22:58:35SlasheriHmm, maybe DSP_SET_FREQUENCY is a good place to do that
22:58:46preglowyeah, i'd say so
22:59:03Slasherigood. would you like to add a memset or something there? :)
22:59:21amiconnBagder: *If* we're going for the solution to have the plugin language files per language for all plugins where each plugins just loads its section, it would be natural to extend that concept back to the core
22:59:37wintermuteBagder: When browsing the database, I get quite a few entries in, for example, Artists that are nothing more than a single letter. Same for albums and songs. And I can't seem to figure out why certain songs' tags aren't read correctly.
22:59:49Bagderamiconn: yes, I thought about that too...
22:59:51amiconnThat is, there would be one language file per language, for both core and plugins
23:00:02amiconnSame goes for voice files
23:00:14preglowSlasheri: no, you do it, please, my dsp.c is hacked to pieces right now
23:00:24Slasheripreglow: ok :)
23:00:50Bagderamiconn: ... and I'm trying to come up with a system that supports both separate individual files for easy user replacements, and one big provided by the default build
23:00:59amiconnThe routines for sectioned loading would then be part of the core anyway, plugins could reuse them
23:01:26Bagderwintermute: sounds like some kind of bug
23:01:26ep0chhow come the pcm buffer refills itself at about < 20% with MP3, but for Vorbis it refills when the pcm buffer is about < 75% ??
23:01:35amiconnAnd when I think about it, we're not giving up independent plugins, only that independent plugins can't have localisation right now
23:02:27amiconnEach language and voice file would need a platform id, and each section would need a plugin id
23:02:36amiconn...or a special id for core
23:02:42wintermuteBagder: If I were to report this bug, who should I inform? The tracker on the site, the mailing list, or an individual dev?
23:02:57Bagderwintermute: the tracker is the best place
23:02:59Learep0ch: I tend to see it the other way around. Vorbis always filled at very low levels, whereas MP3 often - but not always - refills earlier.
23:03:08wintermuteThanks for your help, sir.
23:03:17ep0chno its the other way :)
23:03:21*Bagder bows
23:03:22Learep0ch: Not that I've tested Vorbis much with the audio track viewer...
23:03:40ep0chsurely its codec independant though, and the pcm buffer controls when to boost?
23:03:50amiconnBagder: To fully exploit the concept, IDs which have neither a string nor a voice for a certain platform should be left out on that platform
23:04:38Bagderno need to include empty strings
23:04:39Learep0ch: should be, but different codecs use different block sizes for filling the PCM buffer...
23:04:54amiconnThat's a task for genlang/binlang
23:04:57Bagder...unless they're marked as DEPRECATED
23:05:08ep0chbut about 512k difference?
23:05:19amiconnDeprecation should be marked in a special, clear way
23:05:22 Quit preglow (Read error: 104 (Connection reset by peer))
23:05:35amiconn(per platform)
23:05:37Bagderamiconn: yes, I'm thinking keywords without using quotes on the right side of the colon
23:05:46Bagderall actual strings have quotes
23:05:53 Join preglow [0] (
23:05:59preglowyes, that box was overdue for a crashing
23:06:08amiconnIt's necessary to deprecate per platform if we leave out empty IDs
23:06:11ep0chSlasheri: don't suppose you can fill my lack of knowledge regarding the pcm buffer?
23:06:15 Join muesli- [0] (
23:06:33Bagderamiconn: yes, we need to support that per-platform
23:07:23amiconnHaving per-platform voice files automatically solves the codec problem
23:07:31amiconnWe can just use what we want
23:07:32Slasheriep0ch: yes, the pcm buffering controls the boost
23:07:35Bagderah, right
23:07:39preglowman, what the hell
23:07:48ep0chso how come different codecs appear to boost at different pcm levels
23:08:03ep0ch[22:01] <ep0ch> how come the pcm buffer refills itself at about < 20% with MP3, but for Vorbis it refills when the pcm buffer is about < 75% ??
23:08:07preglowmy box just halts at building firmware file now
23:08:14amiconnBagder: I think voicing at least some plugins is very important
23:08:28amiconn(Thinking about our blind guys trying to flash...)
23:08:28SlasheriHmm, interesting.. Only crossfading should affect that
23:08:33Bagderyes, that'll be awesome
23:08:40ep0chswitched off atm
23:08:41preglowwhat the flaming hell?
23:08:46SlasheriIf crossfading is enabled, buffer is kept very filled all the time
23:08:51preglowscramble started and consumed all the bloody ram on my box
23:09:27ep0chi lie
23:09:32preglow-rwxr-xr-x 1 thomj users 822292036 Jun 27 23:06 apps/rockbox.bin
23:09:34ep0chi changed it halfway!!
23:09:37*preglow thinks he knows why
23:09:38ep0chthanks for explaining
23:09:40amiconnBagder: ...and having voice per-platform will indeed help to get better voice quality
23:09:55preglowLear: yo, did you make your long long change actually work? :P
23:10:04Bagderpreglow: looks like a lds problem
23:10:08preglowBagder: indeed
23:10:16preglowoh don't think gcc actually generated all that code :P
23:10:18amiconnpreglow: Check the .map
23:11:28Learpreglow: sure. I did need to edit though...
23:11:30preglow*default* 0x00000000 0xffffffff
23:11:33preglowis that right?
23:11:37preglowLear: explain...
23:12:04Learfrom my patch: + *(.eh_frame)
23:12:15preglowlength 0xffffffff sounds a bit excessive
23:12:17Learat about line 172 or so
23:12:22preglowLear: what's that do?
23:12:48Learbut that didn't give the problem you have; the linker complained about a section it didn't know to handle, so I had to add it.
23:13:19preglowso you've got it in the ro_data section?
23:13:43Learyes, but I don't know for sure if that is the right thing really.
23:13:50preglowwell, it sure didn't HANG now at least
23:13:54amiconnLear: There's only one sane way to handle .eh_frame, drop it
23:14:00preglowwhat IS eh_frame?
23:14:06Learand how do you do that?
23:14:47Bagder /DISCARD/ : {
23:14:47Bagder *(.eh_frame)
23:14:47Bagder }
23:14:59Bagdersee plugins/ does that too
23:15:39LearI guess whoever applies my patch does that then? :)
23:16:08preglowwhat is it?
23:16:11wintermuteAlright, the bug is on the Tracker under User Interface. Thanks for the help, all.
23:16:15preglowand why should it be discarded?
23:16:16amiconn.eh_frame contains debug info
23:17:25Bagderthat's my understanding too
23:18:26amiconnLear: You did introduce long long, right?
23:18:53preglowthat's what i've done, at least
23:19:10Learamiconn: in a patch on the tracker yes. But the Vorbis codec uses it as well...
23:19:52amiconnYes, and the codec linking uses, which drops .eh_frame
23:20:05Learchanged to discard, still builds and nothing in the map...
23:20:12amiconnI guess that .eh_frame section comes from a gcc lib routine
23:20:23amiconn...that deals with long long
23:20:31Learamiconn, yes, 64 bit divide, IIRC.
23:20:37 Quit wintermute ("CGI:IRC")
23:21:32preglowi dont do divide
23:21:43preglowi do...
23:22:15amiconn.eh_frame most likely means "exception handler (or handling) frame"
23:23:32 Quit muesli- (Read error: 54 (Connection reset by peer))
23:23:38Learbtw, a fully charged h140 should play for more than ~9 hours, shouldn't it?
23:23:53Bagder"Exception Frames"
23:23:58MO-Pantsuoptimization of battery life will come later no doubt
23:24:04Bagderamiconn: right on!
23:24:06MO-Pantsuright now more important things
23:24:25BagderLear: yes it should
23:24:49amiconnLear: probably even longer now that dithering is disabled
23:24:53Learbadger: so the run-time calculation needs work then :)
23:24:57MO-Pantsuany progress on the LCD remote front?
23:25:04BagderLear: yes it certainly does
23:25:16amiconnLear: Someone needs to do runtime measurements
23:25:18Learamiconn: or optimized? :)
23:26:06MO-Pantsuas in any progress towards a working gui for the lcd?
23:26:44Learamiconn: as in letting the battery_test plugin run for a while...?
23:26:58amiconnBattery_test won't tell us much
23:27:27amiconnThe test should be carried out while playing real mp3 data
23:27:50amiconn...and measuring how long it runs
23:28:04preglowcasting a long long to a long DOES give the lower 32 bits and nothing else, yes?
23:28:08Leardarn, I was hoping not having to redo the kind of measurement I did for the player, where I wrote down the voltage every half hour or so, when using 1800 mAh batteries...
23:28:23Bagderpreglow: it probably sign-extends it
23:28:33Bagderunless unsigned of course
23:28:34amiconn...using a well defined bitrate, sample frequency, and repeat all with total filesize >> 32 MB
23:28:37preglowunsigned, yes
23:29:04MO-Pantsuis anyone actually working on the remote lcd gui?
23:29:19amiconnMO-Pantsu: I don't think so
23:29:20BagderMO-Pantsu: thinking, talking, discussing yes
23:29:40amiconnI made a suggestion, but remote lcd gui is low-pri imho
23:29:54amiconnlow-low pri, I'd say
23:30:02MO-PantsuI use it a lot too
23:30:19 Join Cassandra [0] (
23:30:27MO-PantsuI wonder what percentage of users use the remote
23:30:30Bagderwhy, there's nothing on it? B-]
23:30:37Flemmard(i use it)
23:30:39amiconnAh, our paged-scrolling breaker ;)
23:30:40Bagderexcept for the debug log
23:30:41CassandraI use the remote in winter.
23:30:45MO-PantsuI used to use it a lot with iRiver FW that is
23:30:46preglowit's late
23:30:47CassandraIn the summer, it's too bulky.
23:30:48LearRegarding the dither again, I played a 320 kbps file, boost was 44% with optimizations.
23:30:51preglowi can't spend any more time on this
23:31:05preglowLear: how much iram do you use?
23:31:23CassandraCan someone tell me which iRiver buttons I can pick up keypresses from at the same time.
23:31:34preglowCassandra: ON + any one of the others
23:31:39CassandraI know that PLAY and STOP can be used at the same time as the arrow keys.
23:31:41preglowCassandra: same applies for the remote, afaik
23:32:05Learthe dither buffers aren't large, 32 bytes I think.
23:32:11CassandraI can't use MODE or REC at the same time as the arrow keys though, can I?
23:32:12amiconnPLAY (ON) and STOP(OFF) are independent on both main unit and remote
23:32:21 Quit Aison ("( :: NoNameScript 3.72 :: )")
23:32:21amiconnAll others are dependent
23:32:31amiconnCassandra: no
23:32:37preglowwhat, i though play was the only one that was wired on its own
23:32:37CassandraDamn. This is not good.
23:33:02CassandraI can't really use STOP+key for anything. Too much danger of accidentally turning the unit off.
23:33:11CassandraSo I have to choices.
23:33:11amiconnpreglow: STOP is as well, but obviously STOP can't be used as a shift
23:33:37Cassandra1) Move the quick menu to MODE and lose the quickest way of operating it.
23:33:50 Join RotAtoR [0] (
23:33:55Cassandra2) Disable page up/page down in the WPS.
23:33:59preglowamiconn: but i can't code any more on this now, you should try your bresenham idea
23:34:43Cassandrapage up/down is less of a necessity on the iRiver.
23:35:29CassandraWish they'd had more independent button controllers.
23:38:27amiconnIt's not the wps, it's the browser
23:38:40amiconn...but I'd prefer to keep paged scrolling
23:38:50 Quit Lear ("Chatzilla [Firefox 1.0.4/20050512]")
23:39:09CassandraAny other opinions?
23:42:10Coldtoastcan you have it so the quick menu is only in wps? in the browser it's scrolling?
23:42:49ColdtoastI'm thinking maybe there's no need for the quick menu in the browser
23:43:04 Join webguest89 [0] (
23:43:05CassandraProblem is it's useful in the broswer too (changing displayed files)
23:43:25Coldtoasthmmm. just for one option?
23:43:47Coldtoastcan you move that option somewhere else?
23:44:14Bagdermy god what an annoying site
23:44:20Bagderthat font showing url
23:44:38Coldtoasthmmm. what about if you put the option in the menu you get where you hold Play down?
23:45:02hardeepanyone know of a site that has some free (legal) ogg files? i wanted to test something
23:45:27***Saving seen data "./dancer.seen"
23:45:29BagderI have a problematic ogg from a user ;-)
23:45:39Coldtoastnot from wacky?
23:46:06Coldtoastwacky had a whole album of problematic oggs
23:46:15Bagderthat's one of wacky's, yes
23:46:26Bagderogginfo apparently explained the problem
23:46:28Coldtoastyeah. they don't work at all
23:46:35Coldtoastyeah? what was it?
23:46:39Bagderthere's something wrong with the encoding
23:46:54Bagderthat negative something ;-)
23:47:07preglowgranulepos :/
23:48:00preglowwhat the hell is that font comparison for?
23:48:07Bagderhis changes to the fonts
23:48:18preglowhe's just shifted some of them
23:48:25preglowanyway, i know what i need to do
23:48:32preglowi need to convert the cp865 dos font to rockbox format
23:48:40preglowand then i need to make me a frigging ansi based wps
23:48:49ColdtoastCassandra: the menu with playlist management is an important menu whereas the Quick menu isn't. My suggesstion is put the file display type at the bottom of that, forget the quick menu in the browser and have it as page scrolling, which it looks like is important to some
23:48:54Coldtoastmy 2cents
23:49:35Coldtoastthat way you have quick access to the file disply option in the browser still
23:50:23CassandraWell, the other alternative is to stick the Quick menu on long press on A-B and just live without chording
23:50:24Coldtoastwhat's on the Record button tho? that doesn't seem to have anythign assigned to it
23:51:11CassandraWell eventually it'll have recording assigned to it.
23:51:24Coldtoastin the browser menu?
23:51:29CassandraCould assign a quick menu to a long press on that, but still wouldn't have chording.
23:51:33amiconnColdtoast: Did you try the lcd fix?
23:51:33preglowif this ipod port ever takes off
23:51:38preglowi'll bloddy well have to buy one
23:51:43Coldtoastamiconn: yeah. perfect
23:51:43preglowbloody, even
23:51:54amiconnColdtoast: okay, nice
23:52:00Bagderpreglow: I could very well imagine one too if it takes off
23:52:06preglowi'd want a mini
23:52:27preglowbut yeah
23:52:31preglowthere's no reason it can't take off
23:52:44Bagderonly lack of efforts from people
23:52:46preglowthe ipodlinux guys seem like nice enough fellows
23:52:55preglowwe've got a lot of ground to cooperate on
23:53:03Bagdervery true
23:53:26preglowand my dream of doing arm coding comes true as well!
23:53:36preglowbut yeah
23:53:53preglowit'd rock seriously, both as a new port, and considering that the ipod is the most popular dap out there
23:54:30ColdtoastI'd definitely buy an ipod if rockbox were ported
23:54:37Bagderwe basically needs someone with an ipod, some brains and some will
23:55:57preglowthe bootloader work has even been done, which i dare say is the worst hurdle to getting a port started
23:56:23Bagderyes indeed
23:57:42amiconnHmm, I'm not sure. I don't like the ipod at all, I think it's ugly
23:57:54preglowit's not the prettiest thing around, no
23:57:55preglownot by far
23:58:01BagderI would do it for Rockbox, not for an ipod
23:58:03preglowbut i abso-bloody-lutely love the wheel thing
23:58:08Coldtoastbut I hear the wheel is amazingly good
23:58:15preglowit makes navigating far more precise
23:58:20ColdtoastI've never even held an ipod
23:58:38 Quit Flemmard (Read error: 110 (Connection timed out))
23:58:43preglowi don't exactly think my h120 is pretty anyway, that's not why i got it
23:58:43Bagderme neither
23:58:54ColdtoastI do :)
23:59:00ColdtoastI love the look of the h1x0

Previous day | Next day