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 2005-09-06

00:01:30 Quit Wett_ (Read error: 110 (Connection timed out))
00:04:37***Saving seen data "./dancer.seen"
00:06:29 Quit Wett ("Visitez !")
00:09:41flackThose are pretty good shots - pity they're such low res
00:10:34 Quit [-AIR-] ("I-n-v-i-s-i-o-n 2.0 Build 3515")
00:19:00 Join _DangerousDan [0] (
00:22:10 Join koniu_ [0] (
00:22:27 Quit koniu_ (Client Quit)
00:27:55 Join Maxime` [0] (
00:28:06 Quit Mxm`Pas`Bien (Read error: 104 (Connection reset by peer))
00:28:39 Quit ashridah ("Leaving")
00:28:42 Quit _DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
00:31:27 Quit DangerousDan (Read error: 110 (Connection timed out))
00:34:11darkskiezi need to buy a new player, my ihp140 broke, what do folks recommend
00:34:21OctalENPHI got a really cheap iPod. <_<
00:34:27*OctalENPH ducks.
00:35:45 Quit HCl (Read error: 54 (Connection reset by peer))
00:36:34 Join HCl [0] (i=hcl@2001:610:1908:8000:290:27ff:feca:8029)
00:38:08 Quit Moos (" HydraIRC -> <- IRC has never been so good")
00:43:02 Join TCK [0] (
00:49:30amiconnflack: Seems the IGP-100 has 16MByte RAM, like the H-110/H-115
00:52:33flackstrange why it buffers so often
00:52:43flackoh well, not the first strange thing iriver did
00:58:43amiconnHmm, I could need some help from Linus... It seems the problem with the MAS beep disturbing the recording is caused by a race condition somehow
00:58:50 Quit DMJC (Read error: 110 (Connection timed out))
00:59:52amiconnThe mpeg thread does some i2c memory reads while the main thread calls audio_beep() which does i2c register writes in quick succession
01:00:04amiconnBut: i2c is protected by a mutex...
01:00:52preglowlinus is soundly asleep
01:01:09amiconnYes I know...........
01:01:29amiconnThat's what logs are good for.
01:03:33 Part darkskiez
01:06:37 Quit preglow ("leaving")
01:12:01flackWould be nice if those FCC people would publish higher res. photos
01:15:29amiconnIt would be nice if they would publish *working* PDFs
01:20:43 Quit ender` (Read error: 110 (Connection timed out))
01:22:24flackworked for me using gpdf
01:23:06amiconnWell, adobe reader 7 has severe problems showing the pictures
01:23:27 Join Rob- [0] (
01:23:36amiconnIt shows them initially after loading, but scrolling them out of sight and back erases them
01:24:12amiconnBtw, it seems I found the cause for the recording beep bug
01:24:41flackSounds like acrobat is on crack.
01:26:35flackSo is this one of the annoying cases where finding the cause is nowhere near fixing the bug?
01:26:42flackOr might it actually get fixed?
01:30:14amiconnI'm just about to commit the fix :)
01:36:35amiconnThe fix is one single line
01:46:25 Join Sucka [0] (
01:55:02flackwow, it really is hard to tell what's going on here:
01:55:09flackSomething changed from yes to no - but what?!
01:55:17flackeh, no to yes
01:56:15flackAh, Ogg decoding support in Iriver
01:57:49 Quit hicks (Remote closed the connection)
02:04:41***Saving seen data "./dancer.seen"
02:06:30 Quit cYmen ("zZz")
02:44:19 Quit noC|andY`fRa (Read error: 104 (Connection reset by peer))
02:47:30 Quit dpassen1 ()
04:04:42***Saving seen data "./dancer.seen"
04:05:57 Join QT_ [0] (i=as@madwifi/users/area51)
04:12:12 Quit paugh ("Leaving")
04:16:39 Quit QT (Read error: 110 (Connection timed out))
04:19:39 Join midk [0] (
04:27:58 Quit flack ("CGI:IRC")
05:25:33 Join paugh [0] (n=pete@2001:5c0:8fff:ffff:8000:0:3e03:6822)
05:43:10 Quit paugh ("Leaving")
06:00:59 Join Andrew179 [0] (n=Andrew@
06:01:56Andrew179hey guys...I need some advice about buying a new portable player, since my iriver h120 recently died
06:02:54Andrew179I like the looks of the iaudio x5, but I haven't noticed any activity in the wiki about it
06:03:45Andrew179rockbox on the iriver was very nice, but new irivers are selling for way too much for me to get another one
06:04:17Andrew179so I guess my question is, does anyone know about the status of rockbox on the x5?
06:04:45***Saving seen data "./dancer.seen"
07:04:07midknot i!!
07:04:22midkha!! 59m 51s after your last message.. i'm so good.
07:04:29midker... 49s.
07:34:21 Join webguest40 [0] (
07:44:20 Quit Andrew179 ("Leaving")
08:04:46***Saving seen data "./dancer.seen"
08:32:44 Quit webguest40 ("CGI:IRC")
08:32:49 Join webguest40 [0] (
08:34:09 Join LinusN [0] (
08:45:16 Join kurzhaarrocker [0] (
08:46:17 Quit webguest40 ("CGI:IRC (EOF)")
08:51:27kurzhaarrockerLinusN: Somehow I have a feeling that you already noticed I updated the auto source selection patch ...
08:51:42LinusNi noticed
08:53:29kurzhaarrockerMaybe you even noticed my desire to break the ui design rules?
08:56:04 Join einhirn [0] (
08:56:35 Join ender` [0] (
09:05:01 Join ashridah [0] (
09:10:39 Join B4gder [0] (
09:32:00 Join DangerousDan [0] (
09:55:00 Quit Maxime` (Read error: 104 (Connection reset by peer))
09:55:07 Join Maxime` [0] (
09:59:56 Join Toni1 [0] (
10:00:11Toni1Hi there
10:00:29Toni1Linus, you're around?
10:01:25Toni1Does anybody know the status of %Fc issue?
10:04:49***Saving seen data "./dancer.seen"
10:05:23LinusNToni1: YES
10:05:37LinusNoops, caps lock :-)
10:06:05Toni1Linus, do you know, if someone is working on %Fc?
10:07:24LinusNi am
10:07:33LinusNToni1: what exactly is your problem
10:08:44Toni1Solve the %F(c) issue on the ReleaseToDo list.
10:09:00Toni1If not yet done
10:09:23 Join Zagor [0] (i=foobar@pdpc/supporter/sustaining/Zagor)
10:10:24LinusNToni1: yes, but what *exactly* is the issue for you?
10:11:00LinusNis it that the %Fc gives the wrong codec?
10:11:56Toni1Ah no, the issue on archos..
10:13:10LinusNToni1: you mean that you want to handle the case when the %F info isn't available?
10:13:57Toni1I did not investigate yet, but yes.
10:15:46LinusNwhat are the symptoms of your problem?
10:17:07Toni1I mainly like to get some more greens in the ReleaseToDos. So I thought, this might be a good starting point.
10:19:33Toni1If there is nothing to be done on that item, how about the playlist issue? Do you know, wether it is reproducable?
10:20:35LinusNitäs not
10:21:14LinusNiirc, none of us developers have experienced the problem
10:21:29Toni1Anything else I could help?
10:23:18 Join Wett [0] (
10:24:13LinusNToni1: you could always do some massive recordings to see if it's stable
10:24:41Toni1You mention my latest patch?
10:25:15 Quit midk ("Leaving")
10:28:36LinusNToni1: that patch was for iriver, wasn't it?
10:29:14Toni1Yes, right.
10:30:48LinusNthe release is for archos, so we want to test the archos stuff
10:31:34*amiconn has turned one more item green in the ReleaseTodo last night
10:31:47LinusNsaw that, great
10:32:22B4gdergood work amiconn!
10:33:21amiconnLinusN: While we primarily want to test archos stuff, I don't think iriver bugfixing should be neglected
10:33:38LinusNabsolutely not
10:33:40amiconnThe freeze is a good opportunity to fix bugs on all platforms, see my comment here:
10:33:46LinusNsaw that
10:34:03kurzhaarrockerLinusN sees EVERYTHING
10:34:10kurzhaarrockerLinusN is watching us
10:34:39B4gderhe has his spies everywhere
10:35:09kurzhaarrockerHe's probably Lord Vetinaris big brother.
10:37:23 Quit mbr ("leaving")
10:42:21 Join cYmen [0] (
10:43:52 Quit crashd (Read error: 110 (Connection timed out))
10:45:39 Part Toni1
10:45:56 Join crashd [0] (
10:47:07 Quit crashd (Client Quit)
10:47:33 Quit Nibbler (Read error: 104 (Connection reset by peer))
10:48:26 Quit linuxstb ("Leaving")
10:51:06kurzhaarrockeris it necessary to build binutils _before_ gcc?
10:51:40 Join thegeek [0] (
10:52:45 Quit thegeek (Read error: 104 (Connection reset by peer))
10:52:46Zagorafaik the binutils is used when building gcc
10:53:19kurzhaarrockerThanx. :(
10:53:29 Join thegeek [0] (
10:54:14B4gderHCl: how come you treat a zero crc32 as no crc?
10:54:39 Join mbr [0] (
10:54:44B4gderin the runtimedb
10:54:57HClum, why not 0? o.o..
10:55:05B4gderbecause a crc32 is 32bits
10:55:18B4gder0 is thus a valid crc32 value
10:55:22HClits either that or adding an extra field to the tagdatabase whether a file has an valid crc or not
10:55:27HCli know
10:56:13B4gderI just think it is a way too weird thing to do, imagine the day someone actually get a file with crc zero
10:56:38HClthen that one file won't work in the runtime database
10:56:54HCli need a way to determine whether a file has a valid crc or not
10:57:09HCli can take out the code entirely but that would work rather icky
10:57:18HClcause all files with no crc would share the same runtime data
10:57:28HClwhich is hardly what you want
10:57:48B4gderif you disabled the CRC on purpose you surely don't care
10:58:22HClwell, the 0 value is used elsewhere as well, like when for unknown reasons an io error occured when reading the file for crc.
10:58:42HClin any case
10:58:55HClits only an improvement over the proposed "1" value
10:59:24B4gderthat "improvement" showed a problem in the implemtation that I was unaware of
11:00:52HClwelp, i can bump the database version and add an crcvalid field, if you'd prefer that?
11:02:00HCli personally don't think that assuming a crc of 0 is invalid is so terrible
11:02:07B4gderI would prefer documentation
11:02:12HClsince there's a chance of 1 in 4 billion that one would actually encounter such a file
11:02:19 Join linuxstb [0] (
11:02:27B4gderCRCs aren't random
11:02:40B4gderso yes, until found out the chance is in a 4 billion
11:02:41HCli know that
11:03:20B4gderif this would've been documented
11:03:27B4gderI would've reacted a long time ago
11:03:42LinusNB4gder: read the source code :-)
11:04:13HClshall i add a comment on the wiki saying that crc 0 means the crc is invalid?
11:04:18B4gderyes please
11:04:41B4gderI'll make sure never creates a zero crc
11:04:44SlasheriHCl: i think there should never be that change
11:05:05 Join Nibbler [0] (
11:06:33 Join preglow [0] (
11:06:45HClSlasheri: well, the only alternative is adding a field that says whether the crc is valid or not to the fileentry
11:07:15LinusNso we should do that
11:07:51B4gderthe alternative is to hack the crc32 functions to never return 0
11:07:55RickI don't think that's much of a problem
11:07:59Rick1 byte
11:08:03Rickoooooh... ahhhh...
11:08:47B4gderthe one byte is not the problem
11:08:54B4gderthe all-people-need-to-adjust is the problem
11:09:13LinusNi don't see the problem
11:09:43HClyea, what B4gder said
11:09:55HCli'm just glad that the tagdatabase is detached from the runtime database
11:09:56LinusNsure, people will need to regenerate their databases
11:10:02HClso it won't invalidate the runtime database
11:10:48LinusNwe can't strive for backward compatibility for the daily builds
11:11:06HClyea, it would require way too much new code added
11:19:30Rickat least version the files
11:19:32Rickso if its old
11:19:37Rickyou can whine and bitch at the user saying its old
11:21:00CoCoLUSrockbox needs some original easter eggs ;)
11:26:48HClRick: the database already has versioning support
11:28:32 Join Paul_The_Nerd [0] (
11:30:24HClthere, i think i fixed it, i'll try it later.
11:33:29 Quit Paul_The_Nerd ("Chatzilla 0.9.68a [Firefox 1.0.6/20050716]")
11:39:13 Part kurzhaarrocker
11:39:17*HCl hmmz and thinks an hash-type field might be useful when improving hash algorhythms
11:39:45 Join Moos [0] (
11:39:50HCloh, i know. i can rename the hashvalid field to hashtype and have an hashtype of 0 be invalid.
11:41:06 Join kurzhaarrocker [0] (
11:43:16HClthat would allow converting between hash types when better hashing algorhythms become available
11:43:16LinusNHCl: good idea
11:43:21kurzhaarrockeramiconn, in case you are interested: When I compile gcc for sh1 with knoppix it fails to compile embed-bb.c because it doesn't find stdio.h, sys/types.h, (and more) included from tsystem.h
11:43:47LinusNHCl: psst: it's "algorithm"
11:43:53HCldarnit! i always get it wrong
11:43:54HCl :)
11:44:05*HCl will print it down somewhere to remember
11:44:32LinusNkurzhaarrocker: then you need the newlib fix
11:44:37thegeek"print it down";)
11:45:11kurzhaarrockerI already have put newlib in the gcc directory
11:45:24LinusNthen it should work
11:45:59HCli hate altering the runtime database, but. i should add the hashtype field to the runtime database as well
11:46:12LinusNwhy do you hate altering it?
11:46:23HClbecause it completely invalidates all runtime info gathered by users
11:46:38HCland requires them to start from scratch
11:46:39LinusNah, the *runtime* database
11:46:49LinusNi should learn to read
11:47:12HCli will write a java converter to update an old runtime database to the improved format.
11:47:19HClso people won't lose their database
11:47:53LinusNwrite a small script to migrate it
11:49:10HCli guess perl should work fine too
11:49:17HClits a simple adjustment
11:52:12kurzhaarrockerlet me guess: there's a -i missing in the parameters that are passed to xgcc: <snip>-isystem /home/knoppix/gcc-3.3.6/newlib/libc/include </snip>
11:53:24LinusNHCl: or maybe a plugin
11:54:01LinusNkurzhaarrocker: it should work if you put newlib in the correct place
11:55:07 Quit Moos (Read error: 104 (Connection reset by peer))
11:55:20LinusNkurzhaarrocker: but you need to reconfigure and rebuild
11:55:55 Join Moos [0] (
11:56:07kurzhaarrockerI put newlib in the gcc directory before calling configure
11:58:48kurzhaarrocker/home/knoppix/build/gcc/gcc/xgcc -B/home/knoppix/build/gcc/gcc/ -nostdinc -B/home/knoppix/build/gcc/sh-elf/newlib/ -isystem /home/knoppix/build/gcc/sh-elf/newlib/targ-include -isystem /home/knoppix/gcc-3.3.6/newlib/libc/include -B/home/knoppix/sh1/sh-elf/bin/ -B/home/knoppix/sh1/sh-elf/lib/ -isystem /home/knoppix/sh1/sh-elf/include -O2 -DIN_GCC -DCROSS_COMPILE -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -i
11:58:48kurzhaarrockersystem ./include -g -DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED -Dinhibit_libc -I. -I. -I../../../gcc-3.3.6/gcc -I../../../gcc-3.3.6/gcc/. -I../../../gcc-3.3.6/gcc/config -I../../../gcc-3.3.6/gcc/../include -c embed-bb.c -o libgcc/./embed-bb.o
11:59:14 Quit Rick ("I… don't need to be here.")
11:59:26kurzhaarrockerIsn't there a -i missing just before /home/knoppix/gcc-3.3.6/newlib/libc/include ?
11:59:28 Join Rick [0] (
12:00:47LinusNkurzhaarrocker: sure looks like it
12:02:59LinusNi don't know much about xcc, but it looks like the cmdline switch is "-isystem" and the argument is the path
12:04:51***Saving seen data "./dancer.seen"
12:09:12kurzhaarrockerThen I don't understand why it doesn't find the included files. I verified that they are there.
12:11:38 Join sylviapylvia [0] (
12:11:42LinusNkurzhaarrocker: weird
12:12:24 Part sylviapylvia
12:23:14kurzhaarrockerbut you were right. The option really is -isystem. It adds a dir to the start of the system include path.
12:27:18 Join hicks [0] (
12:30:58*kurzhaarrocker blushes
12:31:19kurzhaarrockerThere's a nested newlib - which I should have used
12:33:58 Quit kurzhaarrocker ("Verlassend")
13:00:33*B4gder grrrs
13:10:14*preglow hrmphs
13:10:49 Join noC|andY`fRa [0] (
13:21:03amiconnHmm. It seems the voice entries in the .lang files need some attention.
13:21:33amiconn(or it's just me needing to create up-to-date voice files)
13:22:37preglowi wonder how rockbox would work with a preemptive threading system
13:23:19B4gderwe'd need to add lots of locks and mutex calls all over
13:23:20Slasheriat the moment it absolutely would not
13:23:37Slasheribut indeed, pre-emptive threading would be nice
13:24:30preglowit's just that yield() bugs are pretty obscure, but i'm not certain preemptive threads would make our lives easier
13:24:49amiconnI think they would do the opposite
13:25:03B4gderI agree
13:26:12preglowi tend to agree as well
13:28:53amiconnThe yield() bug with the archos was so obscure because it was not just a missing yield()
13:29:28amiconnThe audio_beep() *did* yield before, although it wasn't visible. The i2c code yields when waiting for the device
13:30:18amiconnThe real problem was that both main and mpeg thread tried to use i2c and didn't yield explicitly., So one could block the other as i2c access is mutexed
13:49:13LinusNimho, preemptive multitasking opens the door to way more obscure bugs than the yield() bugs
13:50:22preglowvery probably true
13:50:37preglowi've bumped into my share of obscure threading bugs
13:51:15preglowand anyway, cooperative threading makes it easier to properly use things like code cache
13:57:28amiconnI think preemptive multitasking would also increase code size
14:00:31preglowbut anywho, if i were to code a wavwriter, any thoughts on how it should be implemented? i was thinking something along a file context menu entry
14:02:19LinusNwould be cool
14:02:25LinusNa transcode plugin :-)
14:02:26preglowamiconn says he doesn't want one
14:02:47preglowLinusN: yeah, linuxstb says he wants to code it as a plugin, problem is, we then need a way for plugins to load codecs
14:02:58LinusNwell, it would primarily be a debugging tool, so kind of agree
14:03:17preglowsure, me too, but this is a debugging tool that also has uses for those thusly inclined
14:03:34amiconnLinusN: I noticed that VBRI headers aren't fully supported in mp3data.c. I wonder why, there's quite some documentation on the net...
14:03:50preglowarbitrary seek support, perhaps?
14:03:51LinusNwhen i wrote it, there were no docs
14:03:59preglowwe'll need a dynamic buffer
14:04:28LinusNplus, the seek table is bigger than the xing
14:04:43LinusNand seeking involves more calculation than xing, iirc
14:04:43preglowi think it can be as big as you want it to
14:04:52***Saving seen data "./dancer.seen"
14:04:56 Nick Sucka is now known as Sucka`away (
14:05:32amiconnLinusN: The seek table is variable size with VBRI, however, I think that shouldn't stop us from supporting it
14:05:33preglowanyway, i don't think it's too widespread
14:05:38LinusNthe vbri reference code uses floats
14:06:19LinusNamiconn: the vbri seek table can be pretty big, and it uses longs instead of bytes
14:06:30LinusNso it will be considerably bigger
14:06:39LinusNbut that should be a big problem
14:06:44preglowwhat, xing seek table uses _bytes_ ?
14:06:46LinusNshould *not*
14:06:55preglowno wonder it has such low granularity
14:07:01LinusNbah, me silly
14:07:15LinusNpreglow: of course it doesn't
14:07:29LinusNi'm just babbling
14:08:05*preglow plays with his new iram
14:08:05LinusNanyway, the reason we don't support vbri is that the docs weren't there when i wrote the vbri parser, and i just haven't cared since
14:08:14 Join muesli- [0] (
14:09:54amiconnThe Xing seek table is a fixed 100-byte table
14:10:20preglowwoop, brb
14:10:43amiconnLinusN: You were right about the Xing toc using bytes
14:12:35amiconnHere are some docs about both vbr headers:
14:13:48amiconnThe only float value in VBRI is the delay
14:16:09 Join k0oba [0] (
14:20:27LinusNamiconn: lots of floats in that code
14:27:11amiconnWe aren't obliged to use this code, correct?
14:27:30amiconnI'm not even sure whether it is gpl compatible
14:30:03 Quit k0oba (
14:30:03 Quit cYmen (
14:30:03 Quit Rob- (
14:30:03 Quit edx (
14:30:03 Quit bluebrother^ (
14:30:03 Quit Asku (
14:30:03 Quit igor47 (
14:30:03 Quit coob (
14:30:03 Quit godzirra (
14:30:03 Quit t0mas (
14:30:03 Quit dwihno (
14:30:03 Quit ze (
14:30:03 Quit muesli- (
14:30:03 Quit DangerousDan (
14:30:03 Quit QT_ (
14:30:03 Quit crwl (
14:30:03 Quit odd (
14:30:03 Quit Strath (
14:30:03 Quit hicks (
14:30:03 Quit Nibbler (
14:30:03 Quit linuxstb (
14:30:03 Quit Zagor (
14:30:03 Quit Maxime` (
14:30:03 Quit pike (
14:30:03 Quit Bagder (
14:30:38CoCoLUSall hail, split
14:30:55LinusNamiconn: no, we shouldn't use that code
14:31:13NJoinBagder [0] (
14:31:13NJoink0oba [0] (
14:31:13NJoinmuesli- [0] (
14:31:13NJoinhicks [0] (
14:31:13NJoinNibbler [0] (
14:31:13NJoinlinuxstb [0] (
14:31:13NJoincYmen [0] (
14:31:13NJoinZagor [0] (i=foobar@pdpc/supporter/sustaining/Zagor)
14:31:13NJoinMaxime` [0] (
14:31:13NJoinDangerousDan [0] (
14:31:13NJoinQT_ [0] (i=as@madwifi/users/area51)
14:31:13NJoinRob- [0] (
14:31:13NJoinedx [0] (
14:31:13NJoinpike [0] (
14:31:13NJoint0mas [0] (n=Tomas@unaffiliated/t0mas)
14:31:13NJoinbluebrother^ [0] (
14:31:13NJoinStrath [0] (
14:31:13NJoincoob [0] (
14:31:14NJoinAsku [0] (
14:31:14NJoinigor47 [0] (
14:31:14NJoingodzirra [0] (
14:31:14NJoindwihno [0] (n=dw@
14:31:14NJoincrwl [0] (
14:31:14NJoinze [0] (
14:31:14NJoinodd [0] (
14:31:14Mode"#RockBox +o Bagder " by
14:31:17 Quit t0mas (Read error: 104 (Connection reset by peer))
14:31:18amiconnApart from that it's C++
14:31:27 Join t0mas [0] (n=Tomas@unaffiliated/t0mas)
14:46:58 Quit k0oba (Read error: 110 (Connection timed out))
14:48:44 Quit ashridah ("Leaving")
14:55:36LinusNi want replaygain support for archos
14:55:44amiconnLinusN: Lazy floating point routines...
14:55:58*amiconn doesn't care about replaygain
14:56:25LinusNamiconn: yeah, those floats don't seem necessary
14:56:44LinusNi don't care about replaygain either, but i know it's a feature people want
14:57:47 Join k0oba [0] (
14:59:05k0oba1st visit here i just installed rockbox on my old 6000 sounds something very cool since 2 year i am on the original firmware really limited and now i see a Light with the rockbox firmware thx to the team and if anyone is ok for givin me few tips....
15:00:22 Join muesli__ [0] (
15:01:10LinusNk0oba: welcome to a better world :-)
15:01:17k0obathx m8
15:01:54k0obait sounds so cool (like better sounds settings...
15:02:08k0obai just noticed it was ok also to play the mix where u stopped
15:02:13k0obaand not from the beginin
15:02:27k0obaerf how cud i live without
15:06:27amiconnLinusN: I have a slight problem with my intended playtime calculation changes. I have to prevent overflows, which are more likely with a formula like filetime = framecount * framesamples / frequency than with the current filetime = framecount * frametime / 1000
15:07:19LinusNlong long
15:07:46amiconnI'm hesitant to use long long just for this. It would cause the multiplication and division routines to be linked to the binary
15:07:52amiconn(muldi3 and divdi3)
15:08:03amiconndivdi3 is > 1 KB for SH1...
15:12:16preglowdoesn't replaygain for mp3 files act like mp3gain, in that it modifies the gain in each frame?
15:12:45amiconnpreglow: I don't think so
15:12:56preglowhmm, ok
15:13:13preglowwhat gain granularity does the mas have?
15:18:36 Quit Febs (Read error: 110 (Connection timed out))
15:19:09 Join rasher [0] (
15:19:35rasherDoes anyone know about the Hebrew fonts? They appear not to be in iso8859-8, which seems .. wrong.
15:21:13amiconnMAS3507 (or rather, DAC3550) has 1.5 dB, MAS3587 and MAS3539 has 1 dB granularity
15:21:18amiconn(main volume)
15:22:25 Quit muesli- (Read error: 110 (Connection timed out))
15:22:28preglowso, replaygain support wont yield anything more than mp3gain support
15:22:48amiconnThe MAS3507 prescaler has 1 dB granularity, the MAS3587/3539 prescaler has a linear scale
15:25:00LinusNrasher: no, i don't know much about the hebrew fonts
15:25:30rasherThey don't work with the current translation, that's for sure
15:26:33rasherYou get all sorts of ùöí type characters
15:26:54rasherUsing an iso8859-8 font from"> looks right
15:27:18rasherNot that I know hebrew, but I'm pretty sure it's supposed to be written using only those letters
15:28:37rasherI'll try and convert one of the current fonts to 8859-8
15:28:52rasherIf I can convince fontforge that it's a good idea
15:30:04muesli__gotta go 4 coffee :D
15:36:27 Quit DangerousDan (Read error: 104 (Connection reset by peer))
15:50:31 Quit muesli__ (Read error: 113 (No route to host))
16:04:53***Saving seen data "./dancer.seen"
16:06:55 Quit hicks (Read error: 110 (Connection timed out))
16:09:28amiconnpreglow: Afaik mp3gain doesn't need to be explicitly supported at all
16:09:51rasherIt doesn't.
16:10:37rasherLooks like fontforge has decided to let me do this afterall
16:11:03rasherAny objections to adjusting the Hebrew-* fonts to be iso8859-8?
16:11:05LinusNi don't see the connection at all between mp3gain and replaygain
16:11:50preglowamiconn: no, it doesn't
16:11:51amiconnIiuc both do the same thing, only in a different waY
16:12:17preglowthey both do the same, mp3gain additionally alters each frame to make all players support it directly, however, the intrinsic mp3 gain granularity is 1.5 db
16:12:48LinusNpreglow: but does mp3gain add replaygain tags?
16:12:53preglowLinusN: probably not
16:12:59preglowbut i have no idea
16:13:01amiconnWhile replaygain adds tags that store a level-correction value and works for various formats, mp3gain modifies all frames directly, and works for mp3 only
16:13:05LinusNso there is no connection then
16:13:11preglowi believe mp3gain adds tags as well
16:13:21preglowwhat kind, i know not
16:13:34LinusNit adds "undo" information
16:13:39MoosAPE maybe
16:13:58rasherYes, undo information in APE tags.. It's a horror show
16:14:12 Nick Mark_`away is now known as Mark_ (
16:14:25rasherArgh! Fontforge is reporting iso8859-8 wrong
16:14:45rasherapparently, 0xB? characters are some sort of weird dots
16:14:58rashereh, 0xC?
16:15:11B4gderhey, hebrew is weird all over not only in 0xb? ;-)
16:16:10rasherWell, up to 0x7? is fine
16:16:31rasherThing is, fontforge has a "preview" thing that shows what the character's supposed to be according to the encoding
16:16:36rasherDoesn't show these dot-things
16:18:17rasherI'm confused.
16:18:35preglowdidn't there exist a unicode patch for rockbox at some stage?
16:18:46rasherStill does
16:18:53preglowdoes it still work?
16:19:00rasherIt's not that old
16:19:09rasherBut not complete either
16:19:30preglowuses utf8?
16:19:32rasherIt got assigned to me :-O
16:19:50rasherThink so
16:20:06preglowi think i'll try the wav patch out
16:20:10preglowor is anyone else working on that?
16:20:30 Join zezayer [0] (n=zezayer@
16:21:15LinusNpreglow: please do
16:21:20Moospreglow: I'm wondering if Lear don't working on it already, but not sure
16:21:33LinusNwe need to investigate all these ogg playback problems people are reporting
16:21:49LinusNi want lear to add replaygain support for archos
16:21:58Mooshehe :)
16:22:01rasherEhm.. What's the point of the Hebrew-small and Hebrew-medium fonts?
16:22:12rasherThey're practically identical to Hebrew-large
16:22:16LinusNoh :-)
16:22:18preglowgive me a file and i'll check that out as well
16:22:32LinusNpreglow: misticriver
16:23:03preglownow, there's a site i haven't been to in a while
16:23:08rasherThere was one related to chained vorbis streams
16:23:13rasherapparently that'll crash
16:23:21preglowlear said he knew how to fix those easily
16:24:07LinusNgotta go
16:24:15Moosciao :)
16:24:26Moosc you later
16:24:33pregloware there actually people adding _id3_ tags to vorbis files???
16:24:38LinusNwill hold a speech about reverse engineering and rockbox, with daniel and björn
16:24:52rasherpreglow: Worse, there are programs that default to doing that
16:24:52LinusNcu guys
16:24:56 Part LinusN
16:25:16preglowrasher: sweet god
16:25:29rasherI know!
16:25:45preglowlinusn: convert them to our cause!
16:25:51rasherRockbox still shouldn't crash :-\
16:26:08preglowdepends on the tag version
16:26:11 Quit B4gder ("Lämnar")
16:26:26preglowyou can't bloody prepend arbitrary data to something and expect it to work
16:26:36preglowgod, how i hate incompetent people, id3 with ogg
16:26:37preglowthat's rich
16:26:43rasherOf course not
16:27:07rasherActually, I was talking of panic-type crashes
16:27:14rasher"codec failure" is acceptable
16:29:01rasherThe codec may crash - Rockbox shouldn't.. And I don't think it does, actually
16:29:42preglowbut ofcourse, i agree
16:29:54preglowit should never crash, however badly you abuse it
16:29:58rasherAh, the hebrew part of the Hebrew-small/medium fonts differ
16:30:06rasherJust not the latin1 part
16:30:13preglowwhy is that even included, then?
16:30:27preglowthe non-hebrew parts
16:30:41*amiconn wonders what Linus wants to discuss with Björn and Daniel
16:30:43preglowthe non-hebrew parts of the small and medium fonts are the same size?
16:30:58preglowamiconn: sounds more like they're talking to other people
16:31:02rasherThey still need to define the glyphs, or there'll be blanks
16:31:30preglowrasher: ahh, i thought the format allowed non-existing glyphs
16:31:40rasherAh, no.
16:31:57preglowif they're not even the same size as the hebrew portion, they sound kind of useless
16:32:06amiconnpreglow: Ah, seems I slightly misread the statement
16:32:12rasherSpeaking of - there should be some "missing glyph" bitmap
16:32:32amiconnThe ascii part needs to be included in all fonts
16:32:40amiconn...for several reasons
16:32:55rasherSome Russian person complained about his screen going blank after selecting russian language
16:33:04rasherturns out he had chicago loaded...
16:33:27rasher"his screen" being the menus
16:33:34HClits hotttt x.x.
16:34:54preglowthe patch didn't apply cleanly
16:35:36Mooswav patch?
16:36:59 Join ripnetuk [0] (
16:37:06 Quit ripnetuk (Remote closed the connection)
16:44:04 Nick Mark_ is now known as Mark_`afk (
16:44:09preglowamiconn: how long does it take to generate a full set of voice files for a language?
16:44:35amiconnHow do you mean, a full set?
16:44:47preglowwell, just one voice file, then
16:44:58rasherWhy I think a "no glyph" bitmap is a good idea:
16:45:07amiconnYou will always ever need only one voice file per language. You can choose from several voices
16:45:22amiconnOne voice takes ~5 minutes on my laptop
16:45:29amiconnErgh, wrong
16:45:44amiconn*All* voices together take ~5 minutes
16:45:52preglowall voice files?
16:45:54preglownot bad
16:46:06amiconn(a total of six voice files; 5 english and 1 german)
16:46:19rasherAny objections to committing the 8859-8 versions of the Hebrew-* fonts? They lack some of the characters from the current versions, but I'm not sure how they'd ever be used.
16:46:23amiconn~50 seconds/voice
16:46:58preglowstill some problems with the new macsr setup
16:47:05rasherAlso, the current versions aren't compatible with the current hebrew.lang
16:47:20amiconnThe upload takes much longer; that's why I asked for a way to automate it when localisation v2 will be added
16:48:08amiconnThis will cause the number of voice files to multiply. If I wanted to support all 6 voices, there would be 36 files
16:48:16preglowdoes the voice ui work with wavpack playing for you guys?
16:48:17amiconn(6 voices * 6 platforms)
16:49:19amiconnPlatforms: Player, recorder v1, recorder fm/v2 (they're identical), ondio sp, ondio fm, iriver H1x0
16:49:38preglowit doesn't here, and i suspect that's a result of the CODEC_DSP_ENABLE thing
16:49:54rashermore than likely I'd say
16:50:05amiconnThe first 5 would be in bitswapped mp3 format, the last one in non-swapped mp3 or whatever voice code we will decide upon
16:51:12amiconnAutomated upload should be easy with curl, only precondition is that I'd need an account on the server
16:52:39 Join kurzhaarrocker [0] (
16:52:57rasherShould be possible to do automated upload to the wiki as well
16:53:20preglowflac is so foolishly slow
16:53:29rasherAssuming a utility that can "POST" files
16:53:32preglowamiconn: well, i don't think that should be much of a problem
16:53:53kurzhaarrockerDoes anybody know how to fill a jukebox more quickly than by copying data via usb1.1?
16:54:05preglowdoing it with usb2? :]
16:54:23rasherTake out the harddisk, connect it using a 2.5->3.5 adaptor?
16:54:54preglowthe CONFIG_DSP_ENABLE flag has to go
16:55:14kurzhaarrockera good idea - but it's not worh the effort, rasher
16:55:48*kurzhaarrocker copies iso cd images just to fill the hd.
16:55:52amiconnkurzhaarrocker: How do you mean, fill? Just take up some space to test short-on-space condition?
16:56:01kurzhaarrockeryes, amiconn
16:56:23rasherLooks like curl can do HTTP POST with binary data
16:56:28amiconnWrite a simple plugin that generates (2GB-1)-sized files
16:57:04amiconnThe ata driver should be able to pump ~4MB/s, almost 4 times faster than USB1.1
16:58:07 Join DangerousDan [0] (
16:58:13preglowis that when boosted?
16:58:15kurzhaarrockerThat idea is even cooler :)
16:58:25amiconnpreglow: Talking archos here...
16:59:15preglowi'll just fix a macsr save in dsp.c now, since that runs in codec context, and i'll commit the new macsr saving stuff
16:59:24preglowthen i'll start optimising some codecs
16:59:24amiconnNice :)
16:59:35preglowprovided that something doesn't pop up to busy me again today
17:01:46amiconnkurzhaarrocker: Is it your recorder that can only do USB1.1 or your computer?
17:01:58kurzhaarrockerboth :(
17:02:14amiconnoh :(
17:02:34 Quit zezayer (Remote closed the connection)
17:03:44kurzhaarrockerIt takes 12 minutes to copy a knoppix cd image to my jukebox.
17:06:21 Join bagawk [0] (
17:14:39rasheramiconn: HTTP POST with curl is easy: curl -F ""
17:15:28rasher"userfile" being the name of the file input field
17:16:05rasherOf course, you also need -u JensArnold:password to be logged into twiki
17:16:28rasherStill, automation should be possible
17:16:53amiconnYes, that could work too
17:17:15amiconnHowever, uploading that many voice files to the wiki might be undesirable
17:17:37amiconnThe wiki does automatic versioning, so this would eat lots of space over time
17:18:29rasherare the old versions of files actually saved
17:20:23rasherAh, it is
17:20:28 Join TCK- [0] (
17:20:32rasherThey are.
17:20:47 Nick Mark_`afk is now known as Mark_ (
17:25:22rasherHCl: Did you see my proposal for databox/searchengine?
17:26:11rasher(seeing as "making search plugins more userfriendly" is on the ReleaseTodo)
17:34:52 Quit TCK (Read error: 110 (Connection timed out))
17:39:48kurzhaarrockerdoes fat_size(IF_MV2(0, ) &disksize, &diskfree) return values in kilobytes?
17:39:49HClrasher: not yet
17:39:58HCli'm gonna shower and get dinner, then i'll look at it
17:41:58*amiconn finally found a way to build the build date into the fat driver as the starting date :)
17:42:11amiconn...for non-rtc platforms, taht is
17:42:26amiconnNo more 2003-08-01 :)
17:43:23kurzhaarrockerWhy does the build date have to be in the fat driver?
17:44:43amiconnWell, the non-rtc units need some faked date to use when writing files
17:45:15amiconnThe starting date used to be fixed at 2003-08-01
17:45:31amiconnIt's certainly better to let it start at the build date
17:45:57kurzhaarrockerSo you have to build rockbox faster than you write the file....
17:46:33kurzhaarrocker.. in order to have the correct time stamp for the new file
17:46:44kurzhaarrocker(it was supposed to be a joke)
17:50:47HClrasher: what does the save search option do?
17:50:59rashersave an .rsp
17:51:08rasherI think.. it's been a while since I wrote that up
17:51:14rasherI'll have a look again
17:51:46HCli'm a bit confused over that
17:51:59rasherYeah, it saves an .rsp
17:52:09HClbut didn't you just open databox?
17:52:14HClexactly what would you be saving
17:52:37rasherWait... hang on
17:53:02rasheredit coming up
17:53:10HCli'm gonna look at dinner in the meanwhile
17:55:09 Nick Mark_ is now known as Mark_`afk (
17:56:36*HCl punches f5
17:56:52HClah yes, that makes much more sense
17:57:10rasher<END> was of course being interpreted as a html tag
17:57:35HClsure, i agree with that
17:57:49HClthe only problem is that i looked at directly adding to the playlist a while ago and couldn't figure it out
17:58:05rasherI guess that can be left out
17:58:20rasherWell, not permanently
17:58:21HClwe need that feature
17:58:27rasherBut for starters
17:58:33HClone thing though
17:58:42HClwhat about running searches that we previously saved?
17:59:00rasher"If the plugin is opened using "Open with" on an .rsp file, the file should come up in "edit mode", the way databox does right now if you enter an existing filename."
17:59:12HClso then you have to browse it before you can run it?
17:59:26rasherHm.. is that a problem?
17:59:31HCli dunno.
17:59:43HCli kind of wanted to be able to directly run a search
18:00:01HClwithout further input required
18:00:07amiconnYOu can
18:00:48amiconn.rsp would be associated with searchengine, so that is what gets run if you simply play the file
18:00:54 Quit bagawk ("Leaving")
18:00:54 Join massoud [0] (
18:00:59rasherThat'd remove the possibility of being able to quickly *edit* a search
18:01:00amiconn(or "Open with.."->searchengine)
18:01:24rasherIn fact, with my proposal that's *the way* to edit a search
18:01:30amiconnYou would then use "Open with..."->databox to edit
18:01:42rasherOh, my idea is to unify the two
18:01:48rasherSo there's one plugin
18:02:00HClyea, they'd be merged into one..
18:02:06 Join dpassen1 [0] (n=dpassen1@resnet-233-61.resnet.UMBC.EDU)
18:02:20rasherI think splitting them is not very intuitive
18:02:45HClthe idea was that databox was for editing and creating, and searchengine for running
18:03:22rasherI know, I just think having it be a two-step process is not very userfriendly
18:03:27amiconnrasher: So how would you execute a query then, without going into edit mode first?
18:03:32rasherI wouldn't
18:03:43 Join Newbyone [0] (
18:03:49HCli know, the original idea was that we were going to have an interface where you can do an exec() like operation for plugins
18:03:51rasherBut it'd be a question of running the file, pressing select (or whichever button), and it's run
18:04:01HClso you'd call the searchengine plugin from databox
18:04:20rasher(because <END> would be selected when a file is run)
18:04:27rasher(I think I forgot to write that)
18:04:32HClmore than that...
18:04:37HClone click to confirm end
18:04:42HClone click to confirm run
18:04:47HCland then another click to confirm play
18:04:58***Saving seen data "./dancer.seen"
18:05:00NewbyoneHCl: hello, is it possible to have words replaced querry for us non dev., pure nowbie?
18:05:15HCli don't understand your question
18:05:57NewbyoneIf I want to use your plugin, is it harder to understand how is work
18:06:03rasherHCl: I understand that this would be a bit slower, but I think the benefits and increased userfriendlyness is worth it
18:06:19Newbyonesyntax is complicated for lambda people
18:06:33HClrasher: thats my problem.. it would decrease userfriendlyness to actually run a search..
18:06:47rasherI don't think so
18:06:54Newbyoneis it possible to have in databox words not sign
18:06:54amiconnrasher: Independently whether the plugins get merged or not, we still need a way to make a plugin both a viewer and a standard plugin
18:07:09HClNewbyone: i don't understand what you mean with words not sign
18:07:25rasherYou get to see the query, and have the possibility of editing if you want to
18:07:34rasheramiconn: That's true
18:07:38Newbyone< >... are word signifiant, no? :)
18:07:56HCl< > tags mean tags that belong to the song
18:08:09Newbyonefor exemple
18:08:18HClrasher: we could add an item "run and play" on top of the main menu to reduce it to 2 clicks
18:08:42HCland open the menu by default when opening an existing query
18:08:47rasherYeah, the second menu could be removed
18:08:50NewbyoneHCl: but for user lambda is not very intuitive, can it be more readable?
18:09:06HCllambda? last time i checked thats not in databox
18:09:08Newbyonefor you guys is easy
18:09:41Newbyonescuse lamda people=x user
18:09:44HCli highly suggest that users use the search song / artist / album options in the browse tagdatabase for basic searches
18:10:15Newbyoneyes but in this way we can't use your plugins
18:10:26HClpersonally, i don't really see how databox could be made more intuitive, without resorting to graphical widgets
18:10:37Newbyoneare they just for you dev? or something?
18:10:37HCllet me put it this way
18:10:51HClno, but i highly suggest that people that don't understand it don't use it o.o;
18:11:06Newbyonepitty for us
18:11:07HClif you get databox to allow you to save your query, then its correct
18:11:21Newbyoneis not the Rockbox politic :)
18:11:27HClor you could try to explain to me whats hard about it :/
18:11:34HClbut i seem to lack in understanding that sometimes :/
18:11:48 Join amiconn_ [0] (
18:11:49 Join Lear [0] (
18:11:52HCli specifically designed databox in such a way so that its idiot-proof, it simply refuses to save invalid queries
18:12:26HClin theory, as long as you mess around with it enough and try to understand it
18:12:35HClas long as it allows you to save, your search query should be correct.
18:12:44Newbyonejust for exemple before you did this plugins, you wanted something for "create me 1980's songs..., this is very understandable for newbiees
18:12:58HClin databox it would be something like
18:13:06HClyear >= 1980 and year <1990
18:13:13Newbyonebut < > = ... not the same, you understand me?
18:13:28HCli'm sorry, i'm afraid that i can't offer you a better interface o.o;
18:13:33rasherNewbyone: So you want to be able to write natural language and have Rockbox understand what you mean?
18:13:38HClmy plan is to add a series of pre-defined search queries to rockbox at a certain point.
18:13:44 Join webguest80 [0] (
18:13:53NewbyoneHCl: don't worries
18:13:56HCllike all songs of the 1980
18:14:04Newbyonejust newbee point of view
18:14:21HClthats one of the reasons why i want to be able to run searches directly
18:14:27HClso that newbies can use predefined searches
18:14:33rasherNewbyone: That's simply not going to happen. You have to formulate the query in a precise manner - which is the databox syntax
18:14:37HClwithout having to worry about them editing and screwing them up or them having to understand how it works
18:14:50Newbyonemake sentences more easy no, it's why said you "words"
18:15:17HCli really wouldn't be able to change much aside from ">" to "greater than"
18:15:20HClwhich is less readable imho
18:15:24Newbyonebut not dramatical fortunatly :)
18:15:27rasherHCl: Yes, I agree. Let me make proposal
18:15:35rasherI think I have a way
18:15:43Newbyoneyes guys
18:16:43 Quit TCK- (Read error: 110 (Connection timed out))
18:16:47Newbyoneoh need to go, thanks for heared a newbee point of view about databox
18:16:52Newbyonebye all
18:16:54HClnp, byes :)
18:16:58 Quit Newbyone ("CGI:IRC")
18:17:10HClmy first priority with the searchengine is possibilities, not userfriendlyness
18:17:29*kurzhaarrocker killed a read ReleaseTodo
18:17:33*kurzhaarrocker goes home
18:18:05rasherHCl: I have a solution to the predefined searches thing as well
18:18:28HClgood :)
18:18:38 Quit kurzhaarrocker ("Trillian (")
18:19:10rasherHow does a .rockbox/searches/ dir, and a menu entry that shows .rsp files there sound?
18:19:34rasherCould be populated with some examples by default
18:19:46rasherAnd people can then save whatever they want into that dir
18:20:07rasherwell, I'm writing..
18:20:21 Quit OctalENPH ("No windows for this server")
18:20:39HClsounds good :)
18:25:48 Join TCK [0] (
18:26:05HClafk, food
18:26:46rasherK, suggestion done now.
18:27:51 Quit amiconn (Read error: 110 (Connection timed out))
18:27:52 Nick amiconn_ is now known as amiconn (
18:29:49 Quit Wett ("Visitez !")
18:38:54 Join merbanan [0] (
18:50:54HCli'd rather have the list of searches in the main rockbox rather than inside a plugin, but okay
18:51:31preglowanyone thought anymore about widget sets? :P
18:53:02rasherHCl: you could make a "browse .rps files" like "browse .cfg files"
18:53:07rasherAnd do away with the list
18:53:19rasherand just make it jump straight to editmode
18:54:17HCli don't want newbies to get dumped in edit mode though
18:54:29HClmaybe we can add some extra parameter when called from browse .rps files?
18:54:46HCllike "-<filename>" rather than "<filename>"
18:54:59HCl- meaning it should be run and added to the playlist immediately
18:58:52pillty linus for the last WPS related commit
19:00:24 Quit noC|andY`fRa (Read error: 104 (Connection reset by peer))
19:00:56rasherYou wouldn't get dumped to editmode if you run a file
19:01:29rasherSo no need to do that
19:02:18rasherYou'd get dumped to the "Run and play / Run and save playlist / Edit" menu
19:04:13HClfirst i'm gonna take a nap, then check my database enhancements, then commit those, and then i'll see what i can do.
19:04:27rasherStill needs a way for the plugin to be in two places
19:04:32rasherdon't know how that'd work
19:10:48amiconnrasher: Imho, /.rockbox/searches should simply be shown in a sub-browser from the core, rather than as an extra choice in the plugin
19:11:04amiconnSame method as the .fnt / .wps etc... browsers
19:11:18rasher Any Hebrew people here/reading logs?
19:12:24amiconnWe would have a far easier job with these search thingies if we would have a way that a viewer plugin can distinguish whether it was invoked via "open with" or via running the file
19:12:38amiconnopening an -
19:12:59amiconn.rsp with "open with" would then open it for editing, while playing would execute the query
19:13:25amiconnThis wouldn't be a problem when both plugins stay separate though
19:14:07 Quit Maxime` ()
19:14:09rasherI don't think they'll stay seperate?
19:14:17amiconnWhy not?
19:14:20rasherAnd I agree, that'd be a nice way of handling it
19:14:49rasherWell, because having them split is a usability nightmare in my opinion
19:15:03amiconnNot if we have a way to chain them
19:16:16rasherIt'd still be very non-obvious
19:16:42rasherAnd require creating a way of chaining them
19:17:00 Join Maxime [0] (
19:17:01rasherAnyway, Run, select edit is quicker than opening context menu and then open with
19:17:21rasherSo I withdraw my support of seperating run from open with
19:17:57amiconnImho that'd be the best approach brought up so far...
19:19:08amiconnChaining would also use that method, only without requiring to be able to distinguish in the plugin
19:19:40rasherWhat's wrong with the "Run and play / Run and save playlist / Edit" menu?
19:21:18amiconnExecuting the query would take longer
19:21:18 Quit webguest80 ("CGI:IRC (EOF)")
19:21:23rasherYou'd still need to differentiate between "run and play" and "run and save playlist"
19:21:45rasherAnd "edit" would be hidden in non-obvious obscurity
19:22:08amiconnWhat would be the benefit of having these 2 modes?
19:22:34amiconnYou can always run a saved playlist with a simple click, and you can also save the current playlist
19:23:04rasherSo you want it to always play the query?
19:23:16amiconn...apart that the latter doesn't make much sense to me for a playlist resulting from a query. You can always re-run the query
19:23:30rasherBut it might give different results next time
19:23:35rasherdepending on the query, of course
19:24:14amiconnWith your proposal, reaching edit mode would faster, but I think queries will be way more often just run than edited
19:24:42rasherA lot of people would need to ask how to edit a query
19:24:49rasherIf it was hidden in "open with"
19:25:05rasherAnd running a query is just pressing right twice
19:25:09rasherinstead of once
19:25:23rasher(differing keylayouts apply)
19:26:40amiconnWe still need a solution for the plugin being bivalent
19:27:20amiconnWhile thinking about this, I'd suggest to make the viewers.config mechanism a bit more flexible
19:27:41amiconnCurrently we always show all viewers in "open with"
19:28:25amiconnWhile I do not want to restrict the list to just these plugins that support the extension (that would e.g. prevent viewing a .cfg with the text viewer)...
19:29:09amiconnI would like to have a mechanism whether a viewer is a catch-all, like the text viewer, or a specialised one
19:29:56 Join webguest97 [0] (
19:31:00rasherYeah, a flag field in viewers.conf would be helpful
19:32:05rasherThat could also say "this viewer is to be found in rocks/"
19:32:14rasherMaking way for duality pretty easily
19:32:17amiconnRegarding your suggestion for the save filename: Imho it should be something like searchNN.rsp, with NN being created by the numbered filename function like the configNN.cfg files
19:32:45rasherForgot about that
19:34:40 Quit webguest97 (Client Quit)
19:36:55 Quit TCK (Read error: 110 (Connection timed out))
19:39:35 Join _aLF [0] (
19:55:50Mark_`afkcould the %xl be expanded to include %xl1 %xl2 etc
19:56:10Mark_`afkso theres 36 instead of 26
19:56:29Mark_`afki used them all even without the battery, volume or rating stuff
19:56:48Mark_`afkeven more would be better, if theres a way
20:04:59***Saving seen data "./dancer.seen"
20:10:15 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
20:26:08 Join paugh [0] (n=pete@2001:5c0:8fff:ffff:8000:0:3e03:6822)
20:40:56 Nick Mark_`afk is now known as `afk (
20:44:50 Join Wett [0] (
20:50:43 Join matsl [0] (
20:52:01 Nick `afk is now known as iMark (
21:08:35 Join TCK [0] (
21:14:55 Join ]RowaN[ [0] (
21:15:17]RowaN[how (in?)accurate is the iriver battery meter in rockbox?
21:18:09rasherWhich one?
21:18:19rasherThe voltage? The percentage? The estimated time left?
21:18:23]RowaN[the wps one that estimes the time left
21:18:33rasherNo idea.
21:19:00]RowaN[k. says 2 hours left.. just wondering if that'll last me on my commute, of it its lying heh
21:19:09]RowaN[i'll find out 2moro wont i
21:20:12rasherWhy not just charge it?
21:21:00rasherConfused about this hebrew deal..
21:21:06]RowaN[wont it wear the battery out if i unneccessarily charge it all the time?
21:21:19iMarkits lithium
21:21:26rasherLetting it go completely down is way worse
21:21:27iMarkmore charges is actually good for it
21:21:36]RowaN[woo really, cool
21:21:37]RowaN[i love to learn!
21:21:53rasherThere's something on wikipedia about this
21:21:59]RowaN[in that case i'll get a charger at work too
21:22:02iMarkor that battery university site
21:22:32]RowaN[i trust you
21:22:55rasherIt's Lithium polymer, isn't it?
21:24:13]RowaN[guys i need a cable for my soundcard.. its got digital out, but its not optical.. Creative tell me i need a "mono-mini-jack to RCA" cable, but no where seems to know what that is
21:24:40]RowaN[trying to link it to my amp (which has optical and coaxial inputs)
21:25:48rasherWhat type is the coaxial input?
21:27:27]RowaN[hmm let me see
21:27:50]RowaN[it just says digital in
21:29:10rasherWhat type is the plug?
21:29:21rasherThe physical plug
21:30:11]RowaN[the plug socket u mean
21:30:49rasherThat should be fairly easy to find
21:31:03rasherAssuming it doesn't require something of the wire
21:31:04]RowaN[shame on me
21:31:39]RowaN[maplins say they dont have em
21:31:41rasherIt's not the most common type of cable
21:32:22]RowaN[aparently its quite standard for pc soundcards to have this kind out digital out, so im surprised it wasnt easier to find a cable to buy
21:33:39rasherI guess most places that have lots of cables have them for "real" audio equipment
21:34:26]RowaN[so where should i look do you think?
21:35:30 Join arkascha [0] (
21:38:42paughget that adaptor and use a standard cable with it
21:39:15amiconniMark: While it's true that deep discharge isn't good for lithium batteries, keeping them at full charge isn't the best treatment either
21:39:37]RowaN[is that mono "mini" though?
21:40:22paughyeah i think 3.5mm is also known as mini
21:40:23]RowaN[coz i think u can get 2.5mm as well as 3.5mm
21:40:37amiconnrasher: What's the problem with hebrew (apart from the bug in splash() - that's a known problem)
21:41:24*amiconn received his 5th rockbox today :)
21:41:32rasheramiconn: Font problems
21:41:39rasherThe hebrew fonts are fubar
21:41:44rasheror was
21:41:47rasherI thought I had fixed it
21:41:59rasherBut now someone is telling me the problem is still there
21:43:40rasherAlthough I suspect he may be doing something wrong, because screenshots show characters that are no longer in the font
21:47:14amiconnHmm, perhaps he didn't install everything, and still has the old fonts
21:47:23rasherThat's my thinking
21:47:33rasherwaiting for a reply
21:47:45rasherIt *looks* right to me
21:48:29iMarkami- is a 1/3 to 1/2 discharge an ok time to charge it?
21:48:47iMarki *hate* the battery running out when im out of the house
21:49:28 Join solex_ [0] (
21:50:56rasherHrm, why are the Rockbox source headers not specifying GPL version?
21:54:37 Quit Sucka`away ("a bird in the bush is worth two in your house")
21:55:11rasherHebrew confirmed fixed. Excellent.
21:59:47rasherwould be a shame to release with a full Hebrew translation, and no working fonts
22:02:11 Join LinusN [0] (
22:02:40amiconnHi again LinusN
22:03:19rasherSplitedit fix in the patch tracker
22:03:25LinusNrasher: saw that
22:03:34LinusNwill check it out in a minute
22:04:39amiconnI've received my Ondio FM today :)
22:04:40rasherI'm still confused what the Hebrew fonts were up to before my fix
22:05:01***Saving seen data "./dancer.seen"
22:05:02amiconnI'm just removing a confusing option for it :)
22:05:08 Join _DangerousDan [0] (
22:05:09 Quit solex (Read error: 110 (Connection timed out))
22:06:01LinusNamiconn: nice
22:06:27amiconnOndio has no s/pdif, but rockbox allows to select s/pdif as source...
22:06:31]RowaN[hmm americans have it good
22:06:44amiconns/source/recording source/
22:07:29amiconnI've got the "all good" Ondio FM type (new MMC circuitry, philips tuner)
22:07:48LinusNamiconn: ebay?
22:08:07amiconnNo, new, from a german webshop via amazon marketplace
22:08:19amiconn't was shrink-wrapped...
22:08:43amiconnAlready fully flashed ;)
22:09:05LinusNit's not every day you buy shrink-wrapped antiques :-)
22:09:32amiconnAll my archoses are already bootboxed
22:09:43amiconnRunning without any problems
22:10:08LinusNamiconn: that reminds me, how does a-b repeat work on the ondio?
22:10:15amiconnIt doesn't
22:10:37amiconnThe code isn't enabled, and we'll have a hard time to find a button to map it to
22:10:43LinusNok then
22:10:50amiconnPerhaps my quickmenu idea will be helpful...
22:10:54LinusNquadruple-click? :-)
22:11:05*Lear wonders why make clean is so dog slow on cygwin...
22:11:18LinusNLear: how about adding replaygain support for the archos?
22:11:21amiconnI already added double-click in one place for Ondio...
22:11:31rasherLear: The answer to that question lies in the end of your sentence.
22:11:35 Quit arkascha (Remote closed the connection)
22:12:13LearHm... Don't know, is it tricky to know exactly when the track change occurs? Because that's when the volume needs to change.
22:12:41LearBut only partial support is possible anyway (clip prevention isn't possible)...
22:12:53LinusNof course
22:13:25LinusNLear: audio_set_track_changed_event()
22:13:55LinusNhmmm, maybe not
22:13:57amiconnLinusN: Not quite
22:14:08LearHaven't really looked at it, but maybe I should. I do have a player lying around, so I can't blame it on not being able to test it. :)
22:14:13amiconnCurrently it does only handle one registration
22:14:19Lear(player as in studio...)
22:14:20LinusNsaw that
22:14:23amiconn...which is used for the runtime db
22:15:20LinusNall these silly callbacks are annoying me
22:15:22LearStill, a good start for where/how to do it.
22:15:39 Join webguest09 [0] (
22:15:50amiconnLinusN: I don't see why callbacks are silly
22:16:07LinusNamiconn: how about calling system_reboot() instead of rolo_load() in panic.c?
22:16:11webguest09wwhich is technically superior, flac or wavpack ?
22:16:28rasherI'd say wavpack.
22:16:43LearHeh, is there even room for replaygain support on the archos? :) If the volume function takes a db value, that would save some code...
22:16:45LinusNthe callbacks are silly because all they do is add a level of complexity
22:16:52amiconnIn the case of runtime db on archos, directly calling the runtimedb functions from the playback code would mean direct calls from firmware layer to application - certainly not any better
22:18:00LinusNyes, that's why the playback code should be moved
22:18:24LinusNthe callbacks are there only to solve that problem
22:18:50amiconnIn fact they were there before I added runtimedb for archos
22:18:53LinusNif the playback code had been in apps, the callbacks would not be needed
22:19:05LinusNyes, they were added by the a-b repeat
22:19:12amiconnOtherwise I would had to add them
22:19:24amiconnNo, they were added way before that Slasheri, for better runtimedb support
22:19:59LinusNah ok
22:20:10rasherHrm.. Hebrew fonts are small... any objections to adding a larger one from my fonts-page?
22:20:11amiconnThe first runtimedb code did a disk spinup after every track...
22:20:23rasher(suggested by the person who was testing my fix)
22:20:29 Quit DangerousDan (Read error: 110 (Connection timed out))
22:20:29LinusNrasher: go ahead, if we are allowed to distribute it
22:20:46rasherIt's from the Debian/Ubuntu package of
22:20:47rashercan't imagine we're not
22:20:54rasherI'll find the precise terms
22:21:03amiconnLinusN: I agree that system_reboot() would be way better than rolo_load() both in panicf() and uie()
22:21:14amiconnrolo_load() often doesn't work
22:22:56amiconnI wonder whether there is sufficient time left before release to fix the splash() issue with hebrew/arabic
22:23:08LinusNah, bidi?
22:23:10rasherCripes, /usr/share/doc/xfonts-75dpi/copyright is 75kb
22:23:50amiconnLinusN: Yes. The problem is that splash() lcd_putsxy()s each word in a line separately, so the word order is reversed for rtl languages
22:24:03amiconn...because the rtl handling is done in the lcd driver
22:24:18rasherAh, the font is one of those "Public domain font. Share and enjoy" ones.
22:24:28LinusNrasher: good
22:24:31amiconnsplash() would need to lcd_putsxy() each line as a whole after line-wrapping
22:25:09amiconn...and lshachar probably needs to adjust the hebrew translation back to normal, as it is now adjusted to this reversal
22:25:27amiconnThis won't work well with varying font sizes anyway
22:25:49rasherI'm just going to go with the one he suggested:">
22:27:10LinusNamiconn: speaking of font sizes, i think the xtal font has a too big line spacing
22:27:38rasherThe nimbus font "fixes" this
22:28:17rasherHoly crap.. The tab completion of bash just scared me.
22:28:37rasher"cvs add <tab>" only cycles through files which are not known by cvs
22:30:07rasherBy the way, I noticed this earlier.. why does the Rockbox source-header not specify a GPL version?
22:30:49LinusNi dunno
22:31:40LinusNwe do refer to docs/COPYING
22:31:43Bagderlet's blame Zagor ;-)
22:31:45LinusNwhich is v2
22:31:55rasherAh, true
22:32:13LinusNbut we could of course be more explicit
22:32:15Bagderit should mention that in the boilerplate anyway
22:32:40rasherThe license says that if you don't specify a version, you can use any version
22:32:48rasherIncluding v1
22:32:55rasherI'm wondering what was wrong with that
22:33:56Bagderno idea
22:35:11 Quit Zagor (Remote closed the connection)
22:35:15amiconnBagder: I hope I've added the config magic for FAT date the right way...
22:35:26Bagderit looked fine
22:35:44Bagderit'll show rather quickly otherwise ;-)
22:36:41amiconnHmm, 'make zip' still adds an empty codecs dir for archos... :(
22:36:57Bagderthen something has changed
22:38:28preglowhow'd you speech go?
22:38:30LinusNamiconn: no codecs/ in my zip
22:38:34 Nick iMark is now known as iMark`away (
22:38:58LinusNamiconn: oops, now i see it
22:39:06*LinusN cleans his glasses
22:39:13BagderI see it too
22:39:30Bagderhow silly
22:39:36BagderI'll fix
22:39:53amiconnYou could sell it as preparation for wav support ;)
22:41:03LinusNok, so now we have 4 fonts that looks like chicago?
22:41:23 Join hicks [0] (
22:41:25webguest09will development of archos firmware end after release of v2.5 ?
22:41:44rasherLinusN: yup
22:41:55amiconnwebguest09: Surely not
22:42:10LinusNchicago-12, nimbus-12, nimbus-14, xtal-14
22:42:23LinusNchicago-12 is lame, misses so many chars
22:42:36rasherI wanted to extend chicago to include 8859-15, but fontforge can't read it
22:42:58Bagderamiconn: I think I fixed it now
22:43:40LinusNit would be nice if someone could clean up in the rockbox font mess
22:44:08rasherSome of the fonts baffle me..
22:45:16rasherCompletely weird encodings
22:46:18amiconnAfaik all fonts should use ISO8859 encodings, except cyrillic which should use Win1251
22:47:05rasherIt would be nice to have a bdf editor
22:47:25rasherFontforge is so much more, and bdf seems like a feature hidden underneath it all
22:47:45LinusNfontforge is not good for pure bdf editing
22:49:03 Nick iMark`away is now known as iMark (
22:49:14amiconnLinusN: the system_reboot() for uie() is in the wrong place.
22:49:32amiconnThe way it is now, it wouldn't work on the player
22:51:11webguest09reasher: do you have a link to your fonts site for h140 please ?
22:52:03rasherIt's not for h140
22:52:12rasherIt's for all Rockbox platforms (except player)
22:52:12 Quit webguest09 ("CGI:IRC (EOF)")
22:52:15amiconnLinusN: There's a bug in the radio screen.
22:52:20LearI've added lots of chars to nimbus-12, but I also changed some... (And a few need fixes. :))
22:52:42rasherEw.. I remember why I stopped using xmbdfed.
22:53:12*Lear used a text editor and a ruby script...
22:53:32amiconnIf the status bar is on, and I enter the radio menu, then return, the status bar vanishes. It reappears after a button press that causes a status bar change
22:53:34 Quit edx (Read error: 110 (Connection timed out))
22:54:00amiconn(e.g. changing volume, or starting a recording)
22:54:00LinusNamiconn: cool
22:55:49amiconnYour uie() change even caused the player build to turn red...
22:59:56amiconnFound the cause for the vanishing status bar. Look at lines 571..574 in radio.c ...
23:00:15LinusNalready fixed it
23:04:00rasherGeez. Why must every program capable of editing bdfs suck so.
23:04:58rasherMust be a curse. Remember never to add a bdf-editor plugin to Rockbox.
23:06:10 Quit Lear ("Chatzilla [Firefox 1.0+/undefined]")
23:08:04rasherAh, this one isn't so bad. Once you can get past it looking like 80s kitchen furniture (and the fact that it's tcl/tk):
23:10:08rasherIt assumes you're trying to make a 8859-1 font, but that's acceptable.
23:15:06amiconnThe blue-violet part of the Ondio case changes colour over time, probably caused by sunlight
23:15:25amiconnIt's quite blue at first, but turns more and more violet over time
23:16:07 Join ghostiger [0] (i=debian-t@tor/session/x-1bb2c90cc50933ce)
23:16:21 Join ghostiger2 [0] (i=debian-t@tor/session/x-3c36e3197dd19fa2)
23:17:32 Join iamlost [0] (
23:18:22 Join mirak [0] (
23:19:05iamlostmirak: hi
23:19:51 Quit ghostiger (Client Quit)
23:19:53 Join chenco [0] (n=ghostige@tor/session/x-971535161745f20e)
23:23:02 Part iamlost
23:23:10amiconnLinusN: Regarding my quickmenu idea: This would practically replace (or complement) two-button combos by two-button sequences, which has 2 advantages
23:23:43amiconn(1) There are no restrictions as with the combos (buttons that can't be detected together at the same time)
23:23:55amiconn(2) It allows for one-handed operation
23:24:29rasherBy the way, there is a pretty serious problem with non-latin1 translations
23:24:42amiconnIt will probably require a button beep with voice ui, to provide feedback for the first keypress
23:24:57rasherUsing the default font gives gibberish for those translations
23:25:09rasher(quickscreens, etc)
23:25:10LinusNamiconn: i'd like to try it in action, do you have any test code?
23:25:31amiconnLinusN: Not yet...
23:25:44LinusNi gotta go to sleep
23:25:51amiconnrasher: Yes, unfortunately not easily solvable
23:26:08amiconnThe same gibberish will appear if you try such a translation on the player
23:26:13 Quit chenco ("Leaving")
23:26:13 Quit ghostiger2 ("Leaving")
23:26:19LinusNnite all
23:26:20rasherI'm guessing unicode is the way to fix this
23:26:25 Part LinusN
23:26:26amiconnLinusN: nite
23:26:31rasherWith a "complete" default font
23:26:31amiconnrasher: Yes
23:26:43 Join chenco [0] (n=ghostige@tor/session/x-9c013ebec1de0953)
23:26:57 Join ghostiger [0] (n=ghostige@tor/session/x-22573bfeab5033d1)
23:27:40rasherPlayer will be left in the dark
23:27:46rasherNot much to do about that
23:28:09 Quit _DangerousDan (Read error: 104 (Connection reset by peer))
23:28:16amiconnThe player should also get unicode support, of course with a limited character set
23:28:25preglowso, what bugs count as release stoppers?
23:28:52amiconnIt may even be possible to support some non-8859-1 languages
23:29:22amiconnImho cyrillic and greek are within reach
23:31:14rasherI thought the player didn't allow redifining the charset?
23:31:31amiconnYes, but it allows defining up to 8 characters
23:31:39amiconn(only 4 with the old lcd)
23:32:08rasherpreglow: .. all of the ones marked "investigating" or "not fixed" for now, I think
23:32:12amiconnIt helps here that the player lcd does only show 22 characters at once
23:32:31rasherTrue, but that's still 14 short - how would that work?
23:32:35 Join DangerousDan [0] (
23:32:57preglowthat's quite a lot
23:33:04amiconnSo if we have a script that needs only a certain percentage of redefined characters, this will hopefully be sufficient
23:33:36amiconnE.g. cyrillic uses many letters that look equivalent to latin letters
23:33:37 Join Mxm`Pas`Bien [0] (
23:33:41rasherAh, like that
23:33:58amiconnThe player code already does dynamic reuse of the 8 definable characters
23:34:17rasherI'm thinking some of those bugs are less critical (at least the "sometimes shuffle may play the same song more than once")
23:34:52amiconnThat means, as long as the current display content needs at max. 8 redefinable characters, it will look perfect
23:34:54preglowis it easily reproducible?
23:35:03rasherpreglow: Don't think so
23:35:29amiconnWith more alien characters, the code would resort to using replacement characters for some of them
23:35:51rasherHrm, cyrillic has "A E K M H O P C T Y B"
23:36:13rasherand, interestingly, 4
23:36:21rasherdepending on how the player shows that
23:36:28amiconnThe replacements will be sorted by "ugliness", i.e. the most ugly replacements will be used last (hopefully practically never)
23:36:54amiconnI alraedy worked out a complete replacement table with a russian user
23:37:19rasherAh, excellent.
23:37:57rasherAll these hidden pages
23:38:29amiconnIt's linked from the DocsIndex page...
23:38:50amiconnWhat more direct link did you expect?
23:39:15amiconn(It's under Rockbox Design Docs)
23:39:39rasherNothing specific, it just reminded me
23:39:57rasherthat the rockbox website has a lot of semi-hidden pages
23:40:00rasherand duplicated pages
23:40:11amiconnGrr, I can't find the default wps in the code...
23:41:07rasherline 1560 ish
23:41:26amiconnDidn't expect this to be within a function...
23:41:51rasherSomewhat obscure place
23:41:51 Quit TCK (Read error: 110 (Connection timed out))
23:43:23rasher gotta love the optimism
23:44:02*amiconn did his first wps change in more than 18 months...
23:46:19rasherHm, would be nice if "codec failure" contained a error return value
23:46:35preglowand a message, heh
23:47:13preglowi daresay we've got enough code space left for that kind of user friendlyness on iriver
23:47:44preglowthe codec plugins are littered with errormessages that have been commented out
23:48:13rasherThey should just use the standard rockbox way of passing errors
23:48:23preglowwhich is?
23:48:47rashera negative int, and if the error comes from another function, add that times 10
23:49:14rasherres = somefunction(); if (res < 10) return res * 10 - 3;
23:49:16preglowyes, but that's a far cry from anything the user might understand
23:49:31rasher< 0, of coures
23:50:05amiconnCommon errors should be reported to the user in cleartext, localised
23:50:06preglowit would be nice to avoid posts like 'what does codec failure: -50 mean?#?#??#' when the message could have been 'codec failure: chained streams not yet supported'
23:50:24amiconnThis is done in a number of places on archos too
23:50:47preglowamiconn: sure, but passing temporary messages as english shouldn't be too nasty
23:51:02preglowsane people use english lang file anyway :]
23:51:03 Quit Maxime (Read error: 110 (Connection timed out))
23:51:04amiconnObscure errors that shouldn't happen during normal operations can be reported as short english text along with the error value
23:51:32amiconnpreglow: The passing should be done the standard rockbox way
23:51:50amiconnThe ui thread is responsible for displaying the cleartext messages
23:51:54preglowamiconn: would be nice, yes, but then it'd be pretty much up to the callee to decide what it means, yes?
23:52:07preglowand that's a pretty nasty place to put codec specific messages
23:52:44amiconnWhat codec specific messages do you want to throw at users???
23:52:47preglowbut anywho, permanent error messages should of course be localised anyhow
23:52:53preglowsomehow, i means
23:52:59preglowamiconn: like the example above
23:53:06preglowwith the chained ogg streams
23:53:39preglowamiconn: or perhaps 'he aac profile not supported'
23:53:57amiconnEither a codec doesn't manage to decode some stream, the a general error message should be sufficient, perhaps accompanied with the error code to give the devs a clue what's going on
23:54:12amiconnEverything else would be too techy for the average user
23:54:43amiconnE.g. the jpeg viewer just displays "unsupported -4" if it encounters a progressive file
23:54:43iMarki recon the released builds should have some kind of logging support
23:55:00preglowi dont
23:55:07amiconn..and that string isn't localised just because it's not yet possible to localise plugins
23:55:11iMarkwhy not?
23:55:40preglowbecause error messages should tell the user all he needs to knwo
23:55:56iMarkyeah, but when the developer needs info on a random crashbug
23:55:56preglowand log buffer takes space that should be used for music
23:56:12preglowiMark: if it's crashed, you can't view the log buffer anyway
23:56:39preglowiMark: and if you're suggesting we spin up the disk to write every log line, you're clearly had too much vokdka today :P
23:56:52amiconnpreglow: Yeah, an average user probably won't have a bdm wiggler... ;)
23:57:25preglowamiconn: so they say
23:57:26iMarkthey should put the bdm wiggler in the board next to the firewire800 port and wifi chip :P
23:57:56preglowiMark: of course, there's a very lucrative modding business just waiting to get started here

Previous day | Next day