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 2005-12-28

00:05:11 Quit actionshrimp (Read error: 110 (Connection timed out))
00:05:51 Quit ender` (" Never say "Oooops" ... always say "Ahhh, interesting..."")
00:15:11 Join DreamTactix291 [0] (
00:20:35 Join akaidiota [0] (
00:25:08 Part LinusN
00:32:13 Quit akaidiot (Read error: 110 (Connection timed out))
00:32:58Papricasomeone know where is the file that include the menu that open when hold "NAVI"?
00:35:24Papricagrr =\
00:35:38Sanwell, i would be the last person to know anyway
00:35:55Sanwhat file has the "Shutting Down"?
00:35:59Sanin it
00:36:09Papricadont know
00:37:21Sanbrb, reeboot
00:37:26 Quit San ()
00:42:53amiconnSlasheri: there?
00:42:58amiconnmuesli_-: wazzup?
00:43:31Paprica[::] <Paprica> someone know where is the file that include the menu that open when hold "NAVI"?
00:43:41muesli_-just good luck.. to shut down my system was just in my mind ;-)
00:44:11amiconnPaprica: onplay.c
00:44:21muesli_-would be theoretically possible to stream music from an external hdd via usb2go?
00:44:37Papricaohh 10xx
00:44:42amiconnHow do you mean, stream music?
00:45:14muesli_-you plug an external hdd via usb2go and play music from it
00:46:09amiconnIf usbotg and mass storage support gets implemented, then rockbox can access external HDDs, card readers etc
00:46:27amiconnThe type of access shouldn't matter, so all applications should be possible it copying files, playing music etc
00:46:48muesli_-thats a yes?
00:47:13amiconnYes - if someone starts working on it
00:47:25amiconnWriting an USB stack is a huge task was just a general q
00:48:35 Join linuxstb [0] (
00:49:56amiconnSlasheri: I have some new tagcache test results for you
00:52:18muesli_-do you know the name of that particalur usb2go chip?
00:54:45 Join mirak [0] (
00:55:11muesli_-cheers Moos
00:55:31MoosHi btw :)
00:55:49muesli_-hi & bye ;-) will go 2 bed right now ;)
00:55:59Mooshehe; good night
00:58:30 Quit gromit` (Remote closed the connection)
01:00:23 Quit muesli_- (Read error: 104 (Connection reset by peer))
01:06:44 Quit linuxstb ("CGI:IRC (EOF)")
01:10:46 Join DrumRBoy320|Away [0] (
01:18:44 Quit _Beowulf_ ("KVIrc 'Marmalade'")
01:23:48 Quit akaidiota (Read error: 110 (Connection timed out))
01:25:21 Join Jungti1234 [0] (n=jungti12@
01:35:03 Join saa[b_r]ider [0] (n=saab_rid@
01:46:31***Saving seen data "./dancer.seen"
01:47:31Jungti1234How are you?
01:47:38 Quit DreamTactix291 (Read error: 110 (Connection timed out))
01:47:45Papricaim fine u?
01:47:53Jungti1234me too :)
01:48:17Papricawant to see my quick screen?
01:48:37Jungti1234Karaoke remote
01:49:53Papricamm what can i add to it?
01:50:03Jungti1234and this
01:50:31Jungti1234I saw this.
01:50:52Papricabut not
01:51:23Jungti1234Is no it a clock there?
01:51:42Papricano clock in the status bar
01:51:57Papricacuz it take alot of space
01:52:27Jungti1234maybe it's possible
01:53:18Jungti1234Does unicode support?
01:53:38Papricathereis a unicode support
01:53:46Jungti1234How does it become if font is big?
01:53:50Paprica2005-12-06: Unicode support!
01:53:58Papricamm dont know
01:54:12Jungti1234Try test.
01:54:51Jungti1234Test it by unifont.
01:55:49Papricawhich font is unifont
01:56:21Jungti1234It's unicode support font.
01:56:23saa[b_r]idermorning guys
01:56:31Jungti1234good morning
01:56:33saa[b_r]iderunifont is called unifont :)
01:57:09saa[b_r]iderif you go to display/browse fonts you'll find a font named "unifont"
01:57:09Jungti1234saa[b_r]ider: are you Chinese?
01:57:25saa[b_r]idernope, I'm from saudi arabia, but I live in china
01:57:48Jungti1234Come to the South Korea.
01:57:59saa[b_r]iderand I speak chinese, so I have no problem reading chinese
01:58:22saa[b_r]iderwe have so many south koreans here with us in the university
01:58:22Jungti1234It's difficult that memorize Chinese character.
01:58:36Jungti1234yes. :)
01:58:52Papricawith unifont its look very ugly
01:58:55saa[b_r]iderto be honest, I do have problems reading chinese :D but I "can"
01:59:08saa[b_r]iderPaprica: too large?
01:59:24saa[b_r]iderthought so...
01:59:50Papricai want a const font to the status bar
01:59:54Papricabut idont know how
02:01:30 Quit mirak (Remote closed the connection)
02:03:08saa[b_r]iderwhat do you mean jungti?
02:03:31Jungti1234What ?
02:03:48saa[b_r]idersorry (was replying to an old message "about the size of unifont")
02:04:28 Quit Moos ("Happy Christmas @ all")
02:06:15Jungti1234Already, passed Christmas.
02:07:56saa[b_r]iderjungti, do you have a holiday on christmas?
02:08:40saa[b_r]iderlucky you :) we don't have a break at all....
02:08:45 Quit Jungti1234 (Excess Flood)
02:09:03saa[b_r]iderpaprica, how did you change the quick screen?
02:09:22 Join Jungti1234 [0] (n=jungti12@
02:10:16Paprica[03:09:07] <saa[b_r]ider> paprica, how did you change the quick screen?
02:10:28saa[b_r]iderwas it originally with controls on right, left, and bottom?
02:10:49Papricaand i change it to
02:10:57Papricaup down navi left right
02:11:20 Join webguest25 [0] (
02:11:28saa[b_r]ideryou didn't like the original design?
02:11:40Papricareally not
02:12:09Papricait looks like "shok" `]
02:12:10 Quit webguest25 (Client Quit)
02:12:40Papricain hebrew ;]
02:12:52saa[b_r]ideris it because you want to have more than just 4 options in your quickscreen?
02:12:59Jungti1234I it smart look like.
02:13:10 Join BreakStuf [0] (
02:13:49BreakStufcan i get some help with WPSs?
02:13:51Papricabut in my opinion the orginal design is ugly
02:14:07Jungti1234Korean people want this Design.
02:14:39Papricasay them that they need to wait
02:14:48Jungti1234BreakStuf: what ?
02:15:01BreakStufi am trying to modify engenier2 WPS
02:15:08Jungti1234yes. It must wait very much.
02:15:15BreakStufi moved all the codecs down, 9PX, and they don't show up
02:15:30BreakStufif i move them back, they get partially hidden by an image
02:16:00BreakStufwhen i skip, the coded picture will flash, but disapera
02:16:09BreakStufi mean disapear really quickly
02:17:28Jungti1234Does no one have a person who know?
02:18:20saa[b_r]iderjungti: is that just a design for a new UI, or is it from an iRiver player?
02:19:05Jungti1234It's new UI that we want.
02:19:42Jungti1234It's just a design.
02:20:04saa[b_r]iderfor the H300?
02:20:47saa[b_r]iderare the koreans working on Rockbox as well? or do they have something else?
02:21:27Jungti1234Korean doesn't have a person who is working in rockbox.
02:22:02Jungti1234There is no programmer.
02:22:38saa[b_r]iderhow do they plan to implement that design? or are they just wishing that someone can implement it :)
02:23:27Jungti1234They can design only.
02:24:08saa[b_r]idernice ideas really... but I think we're still so far away from having something like that.....
02:24:18Jungti1234They can't make it.
02:24:46saa[b_r]iderI really wish they implement album art support....
02:24:59hyarionHCl: ping?
02:25:06Jungti1234Just it's dream.
02:26:07Jungti1234I desire it too.
02:26:32saa[b_r]ideryeah I saw that earlier.... I don't think this would be impossible.
02:26:59saa[b_r]iderthe H10 supports album art, doesn't it?
02:27:52Jungti1234I don't know.
02:28:43saa[b_r]idermost the koreans I know buy flash players, many of them are iRiver, but i've never seen anyone with an H100/H300....
02:28:48Jungti1234It doesn't support in H10.
02:29:30saa[b_r]iderso how come you bought an H300?
02:29:40Jungti1234They don't know HDD player's good point.
02:30:32Jungti1234I don't understand. I don't well English.
02:31:08saa[b_r]iderno problem.... well I guess iRiver are lucky, cause the iPod is not so popular in korea :)
02:31:41Jungti1234Have much iPod user.
02:32:04 Join saab_rider [0] (n=saab_rid@
02:32:10saab_ridersorry, d/c
02:32:53saab_riderI know that koreans prefer buying local products, rather than imported ones.... right?
02:33:59Jungti1234They want 'Hurry fast'.
02:34:18saab_riderhurry fast?
02:35:18Jungti1234Hurry! Hurry!
02:35:31Jungti1234They are urgent.
02:35:40saab_rideryou mean for new stuff?
02:36:53Jungti1234If buy, wait that it comes fast.
02:38:28Jungti1234Don't you understand it? :)
02:39:18saab_ridertrying :) what exactly do they want to be fast?
02:39:52Jungti1234They buy thing in on-line.
02:40:44saab_riderahh... you mean it takes a long time to get imported stuff, and it's easier and faster to get stuff from korea?
02:41:34Jungti1234!!! :)
02:41:37saab_ridercool :)
02:42:00saab_riderwhat about the iPod Nano, do they like it? because it's flash, not HDD based
02:42:44Jungti1234I don't know why iPod is no popularity.
02:43:50Jungti1234But, it's plain that it doesn't fit in Korean.
02:44:34saab_rideryeah, you love different designs.... iPod only comes in one basic design, and basically in only two different colors
02:44:58Jungti1234yes.. :)
02:45:11Jungti1234They want diverse color.
02:45:54saab_riderhow much does the mobi-blu cube cost in korea? any idea?
02:46:13Jungti1234ah mobi-blu
02:46:23Jungti1234It's a little popular.
02:46:31Jungti1234My friend bought it.
02:46:34saab_riderI want one :)
02:46:51saab_riderI don't know if I want to get a cube, or the new "peble" have you seen it?
02:47:11Jungti1234I will find.
02:47:46Jungti1234peble? What is it?
02:47:47Papricagood night
02:47:52Jungti1234good night
02:48:13 Nick Paprica is now known as Paprica[sleep] (
02:48:13DBUGEnqueued KICK Paprica[sleep]
02:49:23saab_ridergood night :)
02:49:32saab_riderbtw, isn't it morning jungti ;)
02:49:52saab_rideroh, it was paprica going to sleep... I think i'm still sleepy :D
02:50:08saab_riderJungti: I'll give you a link, I'm looking for it
02:50:53Jungti1234The South Korea is 10:50.
02:51:29 Quit saa[b_r]ider (Read error: 110 (Connection timed out))
02:53:41 Nick saab_rider is now known as saa[b_r]ider (n=saab_rid@
02:53:42 Join mirak_ [0] (
02:56:00saa[b_r]iderit's pebble :)
02:57:34Jungti1234very cute
02:57:51saa[b_r]iderI'm sure you understand more from the web-site than me :D
02:58:27Jungti1234hehe yes
02:58:55saa[b_r]ideron the pebble, the controls are under the screen, so it should be easier to use, because on the cube, the controls are on the side
03:00:17Jungti1234It's so small. :)
03:00:48Jungti1234I am hard as use because hand is big. haha
03:01:08saa[b_r]iderand the mirror effect is cool.....
03:01:31Jungti1234후꾸루무꾸루 - 매들아~ 한국말 쓰자. 통 뭔소린지 모르겠다.
03:01:45saa[b_r]iderI already have an HDD player with 40GB, so if I get something small, it should be REALLY small :)
03:02:38Jungti1234Hukkurumukkuru - yeadela~ hangukmal ssuja. tong monsorinji moruegetda.
03:03:28Jungti1234-> Kids ~ Korean use let's. Do not understand it is what speech.
03:04:08saa[b_r]iderthey made it for kids?
03:04:38Jungti1234They are malcontent because you unuse Korean.
03:05:14saa[b_r]iderI guess so :) does it say if it's out already? did they release it?
03:06:33 Join muesli_- [0] (i=muesli_t@
03:09:08Jungti1234saa[b_r]ider: It was released already.
03:15:41 Join saab_rider [0] (n=saab_rid@
03:16:17 Quit saa[b_r]ider (Read error: 104 (Connection reset by peer))
03:18:15muesli_-a page of "?" for me ;)
03:18:50saab_riderabout $US 150
03:20:44 Join DreamTactix291 [0] (
03:26:14Jungti1234Sorell released mp3 player who add GPS function.
03:27:06saab_rideris that the red one?
03:28:31Jungti1234I don't understand. sorry :(
03:31:53saab_riderthe MP3 player with GPS, was it red color?
03:35:08 Quit muesli_- ("ich will Khe!!!")
03:41:32 Join mikearthur [0] (
03:46:33***Saving seen data "./dancer.seen"
03:49:02BreakStufi got a quick question about WPS creation
03:49:49BreakStuf"%al %fbk %pc / %pt" why does this line obstruct some small images behind it (the spaces are the ones doing to obstruction)
03:59:45saab_riderbreakstuff, it seems all the smart people are sleeping :)
04:00:01 Nick saab_rider is now known as saa[b_r]ider (n=saab_rid@
04:04:17 Join reryan [0] (
04:04:41reryanlo rockbox
04:05:00Jungti1234hey BreakStuf
04:06:08reryananyone familiar with eclipse?
04:06:19Jungti1234BreakStuf: You know
04:34:17Jungti1234oh -_-;
04:48:24 Quit BreakStuf ()
04:48:40saa[b_r]iderbtw, that X5 still works ;)
04:50:39reryanError launching builder (make ?f uisimulator/win32/Makefile.eclipse clean all )
04:50:39reryan(Exec error:Launching failed)
04:50:55reryanwill haunt me to my grave
04:57:56 Join eEye- [0] (
05:07:37reryanis it supposed to s**t itself when it gets to dumb?
05:15:05 Join Mongey [0] (
05:15:24Jungti1234what is s??t?
05:16:40 Join mymomthelush [0] (
05:18:10 Part mymomthelush
05:25:27Mongeyanyone here that can edit plugins?
05:31:29 Quit Mongey ()
05:42:10 Join Sando [0] (
05:42:24 Quit Sando (Remote closed the connection)
05:46:37***Saving seen data "./dancer.seen"
05:46:49 Quit mikearthur ("Konversation terminated!")
05:47:01 Join Ikarius [0] (
05:47:12Ikariusholas, anyone here?
05:48:39IkariusI'm wondering a bit about flashing firmware.... downloaded rockbox 2.5 for AJB v1, and despite the release notes saying that bootbox is now included, I dont see the firmware files in what I downloaded
05:49:14Ikariusdo I download the latest test bootbox from the bootbox wiki page?
05:49:56 Quit eEye- ("fractal2 mirc script (ver 1.0betar2)")
06:00:41 Join lostlogic [0] (
06:00:47 Quit lostlogic (Read error: 104 (Connection reset by peer))
06:03:24 Join Lynx0 [0] (
06:17:24 Quit Lynx_ (Read error: 110 (Connection timed out))
06:17:25 Nick Lynx0 is now known as Lynx_ (
06:19:15 Quit saa[b_r]ider (Read error: 104 (Connection reset by peer))
06:19:47Ikariusanyone here?
06:20:40Jungti1234I'm here.
06:36:05 Quit `3nergy (Read error: 104 (Connection reset by peer))
06:40:52 Quit DreamTactix291 (Read error: 110 (Connection timed out))
06:57:03Ikariussorry, Jungti, you still around/
06:59:12Ikariuslooking for anyone whom can help me with where to get ver 2.5 bootbox
07:00:44IkariusI downloaded bootbox from the bootbox page to get the firmware_rec.bin file, and flashed that, and then tried to run "rombox.ucl", and it says "incompatible version"
07:01:06Ikariusso where the heck am I sposed to get the 2.5 version of bootbox?
07:09:26Ikariusmmm, ok, got it figured out
07:09:41Ikariusjust had to "run" the rombox.ucl file
07:09:52Ikariusdifferent than how I used to do it
07:11:49Slasheriamiconn: hi, did you get it working? :)
07:13:11Ikariushi slasheri
07:14:06SlasheriIkarius :)
07:14:28Ikariusheheh, ok, how does one build the id3 tag database?
07:14:37Ikariusis it done on the player with a plugin?
07:14:43Ikariuscant find this in the documentation
07:15:04Ikariusonly the release notes mention this functionality, and they just say "with the provided tools"
07:15:25Slashericurrently you need an external program (either java or perl version to do that)
07:15:37Ikariusok, where do I find that external program?
07:15:51Slasheribut with the new version (should be ready soon), you can build it much more easier on the player itself
07:15:53IkariusI have perl and java on my windows boxes, as well as my linux boxes ;)
07:16:00Slasherihehe :)
07:16:14Ikariusman, been a long time since I used rockbox
07:16:22Slasheriat least you will find the perl version from tools/ if you have cvs version installed
07:16:39Ikariusmm, ok, I'll go hit up your cvs tree
07:17:12Ikariusif I generally want to select an artist then an album, will the ID3 database work well for that?
07:17:19Ikariusor does it not cascade categories?
07:17:28Ikariusor rather, the ID3 browse mode
07:18:27Slasherii don't know how well the current implementation works.. i couldn't get it working at all! :D
07:18:43Slasheribut i think you should be able to do that
07:19:31Slasheriat least with the new tagcache you can do that - with genres also
07:20:40Ikariusgood thing I
07:20:50Ikariushave been totally anal about how my MP3s are tagged
07:21:11Ikariusgenerally rip with EAC then redo the tags with tag&rename
07:22:19SlasheriHmm, interesting. I pull the tags directly from some internet db at the encoding phase
07:22:47Ikariuscddb is OK, but spotty on quality
07:22:49Slasheri(with kaudiocreator one can specify the format also)
07:23:02Slasheriah, ok :)
07:23:04Ikariustag&rename pulls tag info from amazon
07:23:14Ikariusand supports mass tag editing
07:23:25Ikariusand puts JPGs of album covers into the tags
07:23:41Slasherihehe, interesting. That is probably a better then
07:23:42Ikariusand can rename files based on the tags
07:23:55Ikariusit's vastly superior to any other tagging program I've seeen
07:24:12Ikariuseven if it is a windows app ;)
07:24:23Slasherieasytag for linux is quite nice too. It can rename by tag or fill tags by filename also :)
07:24:46IkariusI'll check that out sometime
07:25:15Ikariuslong ago I spent quite a bit of time looking through tagging programs
07:25:29Ikariusand every single one I checked out sucked except for tag&rename
07:25:42IkariusI had to get my tags straight for my audiotron ;)
07:26:18Ikariussome time, I'll think about getting a roku or something like that to replace the audiotron
07:26:55Ikariusbut for now, audiotron + archos do pretty much everything I want with my MP3s
07:27:20IkariusI cant wait to demonstrate speaking menus on my archos to my iPod wielding friends ;)
07:27:38 Join _FireFly_ [0] (
07:34:51 Quit Dist (Read error: 145 (Connection timed out))
07:35:09 Quit Slasheri (Read error: 145 (Connection timed out))
07:42:17 Quit RotAtoR ("zzzzzzzz")
07:46:30 Quit Sucka (Read error: 110 (Connection timed out))
07:46:40***Saving seen data "./dancer.seen"
07:48:26 Join Slasher [0] (
07:48:35 Quit Slasher (Client Quit)
08:10:46 Quit _FireFly_ (Remote closed the connection)
08:17:35 Join saa[b_r]ider [0] (n=saab_rid@
08:18:39Bgerhi, Jungti1234
08:18:51amiconnHmm, now Slasheri's gone :(
08:21:20 Join DreamTactix291 [0] (
08:50:56 Join ender` [0] (i=ychat@
09:04:54 Quit DreamTactix291 (Read error: 110 (Connection timed out))
09:07:45Ikariusoh man, that's way too cool for school. my archos is now voicing directories
09:08:20Ikariusnext step, make iPod toting friends totally jealous ;)
09:41:44 Join Vlad0man [0] (
09:42:47 Quit Vladoman (Read error: 110 (Connection timed out))
09:45:01 Quit hyarion (Remote closed the connection)
09:46:44***Saving seen data "./dancer.seen"
09:56:58 Join gromit` [0] (n=gromit`
10:02:47 Quit ghode (Read error: 104 (Connection reset by peer))
10:04:44 Join ghode|afk [0] (
10:37:17 Join DreamTactix291 [0] (
10:59:38 Join Dist [0] (
11:02:39 Join Slasheri [0] (
11:07:27 Join pilot000 [0] (
11:14:45amiconnSlasheri: here now?
11:15:46Slasheriamiconn: yes, hi. Just configuring my new internet connection :)
11:16:59amiconnI have some results for you. I got tagcache sort-of working on archos
11:17:40amiconnFirst, I tried your idea to just align the entries (in line 710 of tagcache.c)
11:17:55amiconnThis didn't work right, the lists were very incomplete
11:18:29amiconnThat's understandable, as you use the length to 'walk' the list, but the length then no longer matches the true distance
11:18:53amiconnThen I tried to adjust the length itself. I changed line 700 to read:
11:19:00amiconnint len = (strlen(data) + 2) & ~1;
11:19:23amiconnThis actually works, but there are still problems, but first some benchmark data:
11:19:27Slasheriah, interesting
11:19:58amiconnStarting without any tagcache_* file, I called "Force tag cache update"
11:20:56amiconnDisk activity for ~ 9 min 30 sec, _tmp.tcd with 524399 bytes was created (there are 3233 mp3 files on my recorder)
11:21:16Slasherioh, that sounds quite good
11:21:23amiconnRoLo'ed to start again
11:21:32amiconnBoot time was extended by ~30 secs
11:22:45amiconnFile sizes: _0.tcd: 13344, _1.tcd: 7754, _2.tcd: 212, _3.tcd: 67041, _4.tcd: 233942, _idx.tcd: 66172
11:23:35amiconnInteresting detail: the length alignment only changed the size of _0, _1 and _2; the other file sizes remained constant
11:23:57Slasheriyes, that looks normal because not all tags are uniqued
11:24:08Slasherionly artists, albums and genres are uniqued
11:24:23amiconnYes, but that also means the length values aren't used in _3 and _4
11:24:33amiconn...otherwise it would crash on archos
11:24:50amiconn=> you could leave them out, saving 2 bytes/entry
11:25:00Slasherino, length is used (the same struct) but it's written directly to disk without loading to ram
11:25:24amiconnI know that it's the same struct, but the value is never used again (?)
11:25:59amiconnBrowsing the db sort-of works, but with 2 significant bugs
11:27:14amiconn(1) Entering 'Artists' or 'Songs' splashes "Dir buffer is full!", and then only a fraction of the entries is shown
11:27:32Slasheriamiconn: that is to be expected if the dir buffer is too small..
11:27:43Slasheriso not all entries can fit to the buffer
11:27:44amiconnThe old db had a chunked browsing mechanism to go above the standard dir buffer size
11:27:56amiconnThis needs to be retained for archos
11:28:05Slasherii will look about that then
11:28:09amiconn...otherwise the tagcache isn't really usable
11:28:50amiconn(2) Entering 'Albums' and then an arbitrary album shows no (!) tracks at all
11:29:50amiconnGoing through 'Artists' -> artist -> album works...
11:30:51Slasheriinteresting, that must be a bug
11:30:53 Quit DreamTactix291 (Read error: 110 (Connection timed out))
11:31:00Slasheribut i will investigate those problems
11:31:24Slasheriit still looks quite good that the cache almost works on archos
11:32:12amiconnAnother (minor) thing: empty tag values are shown as empty string
11:33:13amiconnI did not yet check whether playback is working properly during background update
11:34:16amiconnBtw, the file size increase caused by the word alignment in unique_insert() is so small that we should do this for all platforms
11:34:47amiconnIt's just half a byte per entry on average
11:43:52Slasherisounds good :)
11:46:07amiconnBtw, the statistics are interesting too. It seems that disk access is indeed the bottleneck, and not CPU speed
11:46:48***Saving seen data "./dancer.seen"
11:47:01amiconnI'll test on Ondio next; depending on what part of disk access (seek or read) is the bottleneck, it will be faster or slower
11:47:36amiconnOf course the absolute numbers will be significantly smaller, as it's not possible to fit 1000s of tracks on a flash player
11:48:25amiconnOne big advantage with tagcache on Ondio will be that it's possible to have the db contain tracks from both internal flash and MMC, which is impossible with the offline db
12:05:20amiconnOkay, test data from Ondio:
12:06:19amiconn'Force tagcache update' => ~40 sec MMC activity (314 mp3 files, 302 on MMC + 12 internal)
12:06:40amiconnRoLo: Boot time increased by ~10 sec
12:07:37amiconn_tmp.tcd before RoLo: 52262; _0.tcd: 1520, _1.tcd: 548, 2.tcd: 62, _3.tcd: 6485, _4.tcd: 24830, _idx.tcd: 6292
12:07:42Slasherihehe, quite good :)
12:07:51amiconnBrowsing works, with the same quirks as on recorder
12:08:18Slasheriok, nice
12:08:22amiconn(Buffer full when browsing songs, and 'Album' -> arbitrary album is empty)
12:14:18amiconnThe point with the dir buffer is that one usually structures the files in a way to avoid too large dirs, and the advice is to set 'max files in dir' as little as possible above the largest dir size
12:15:03amiconnMy largest dir on recorder (and iriver) contains a little above 400 files, so I have 'max files per dir' set to 500
12:16:16amiconnThe database is different in that it might show more entries in one list (the song view should show all music files, which are >3200 on my recorder), so Zagor implemented the chunked browsing
12:18:26amiconnIncreasing 'max files per dir' just for db browsing would be a waste, and impossible for a fully upgraded archos
12:18:59amiconnA 120GB disk can easily contain >10000 audio files
12:28:20 Join darkless [0] (
12:35:06 Join Moos [0] (
12:39:26 Quit saa[b_r]ider (Read error: 110 (Connection timed out))
13:00:25 Quit edx (Read error: 110 (Connection timed out))
13:03:13 Join `3nergy [0] (
13:08:11 Join DreamTactix291 [0] (
13:15:16 Join amiconn_ [0] (
13:19:15 Quit pilot000 ("CGI:IRC (EOF)")
13:21:37 Join pilot000 [0] (
13:30:09 Quit DJDD_ ("Trillian (")
13:32:41 Quit amiconn (Read error: 110 (Connection timed out))
13:32:41 Nick amiconn_ is now known as amiconn (
13:33:23 Join Lynx0 [0] (
13:46:50***Saving seen data "./dancer.seen"
13:47:04 Quit pilot000 ("CGI:IRC (EOF)")
13:48:23 Quit Lynx_ (Read error: 110 (Connection timed out))
13:48:23 Nick Lynx0 is now known as Lynx_ (
14:02:01 Quit DreamTactix291 (Read error: 110 (Connection timed out))
14:14:45 Join lamed [0] (
14:18:42 Join saa[b_r]ider [0] (n=saab_rid@
14:18:46lamedI see nothing is quite happening around here latelty
14:19:45Weazel_because of the holidays i guess
14:20:31lamedhaven't thought of that (jewish)
14:40:30 Join edx [0] (
14:42:19 Quit Jungti1234 (Read error: 110 (Connection timed out))
14:52:44 Quit Bger ("BitchX-1.1-final -- just do it.")
15:00:55 Join Jungti1234 [0] (n=jungti12@
15:01:23 Join eEye- [0] (
15:13:15 Quit eEye- (Read error: 104 (Connection reset by peer))
15:21:06 Join Mongey [0] (
15:22:22 Quit Jungti1234 ()
15:23:38 Quit reryan ("Client Exiting")
15:26:47 Join Hooligan [0] (
15:27:00 Join mongey1 [0] (
15:27:08 Quit Mongey (Read error: 104 (Connection reset by peer))
15:29:02HooliganShould have the Record button on the H300s take you straight to the playlist when you're on the WPS. Would be nice.
15:30:27mongey1Hooligan; why cant you do it yourself?
15:30:48HooliganI mean, as a permanent change.
15:31:19mongey1so did i
15:34:40 Nick mongey1 is now known as Mongey (
15:35:38HooliganI don't know enough C to make any changes without worrying about breaking something I don't notice :P
15:35:57HooliganWell, that's a worry either way. But I'd probably do it.
15:36:10Mongeylol, same with me, matter of fact i dont know c
15:37:12HooliganA recent study shows that programming in a language you don't know increases your chance of encountering coding errors by 73%
15:39:30 Join DreamTactix291 [0] (
15:39:36 Quit Hooligan ()
15:44:54 Quit Nilisco ("changing servers")
15:46:41 Quit Mongey ()
15:46:54***Saving seen data "./dancer.seen"
15:59:12 Quit lamed ("CGI:IRC (EOF)")
16:13:22 Join Mongey [0] (
16:42:09 Join _FireFly_ [0] (
16:50:17 Quit Mongey ()
16:57:12 Quit Ikarius (Read error: 110 (Connection timed out))
16:58:59 Quit DreamTactix291 (Read error: 110 (Connection timed out))
17:03:13 Quit _FireFly_ (Remote closed the connection)
17:11:40Slasheriamiconn: i am still thinking how the implement sorting with the chunked browsing..
17:12:15Slasheriamiconn: maybe that means that we will need to sort the db at the same time while committing it
17:15:31 Join Nilisco [0] (
17:32:50amiconnSlasheri: Yes, chunked browsing needs a sorted database
17:35:11Slasheriok, so sorting while committing should solve that :)
17:35:54amiconnYes, if it's not too difficult to implement
17:37:37amiconnIt would even save sort runs, sorting once during commit instead of each time during display
17:37:39 Join markun [0] (
17:38:15amiconnI'm not sure whether this will work with on-the-fly commits when the tagcache is in RAM
17:40:07SlasheriHmm, on-the-fly committing is not even implemented yet but i will investigate that soon
17:42:16Slasherithat would also make tag editing on-the-fly harder, because tags would get out of sort until next boot
17:42:35 Quit Rob2222 (Read error: 104 (Connection reset by peer))
17:42:48amiconnHmm, what will happen when the buffer RAM gets all used up on commit?
17:43:28amiconnI saw that while this is a foreground operation, you don't use the whole audio buffer, but just 256KB
17:43:28Slasherithe commit will fail and tagcache files will get erased
17:43:33Slasheribut that is quite rare to happen
17:44:17 Join Rob2222 [0] (
17:44:19 Quit Rob2222 (Read error: 104 (Connection reset by peer))
17:44:24amiconnI don't understand why...
17:44:28Slasherihmm, i am not sure if we can do partial commits
17:44:59Slasheriin fact that might be possible. But the temporary file might needs to be erased after that
17:45:01 Join Rob2222 [0] (
17:45:24amiconnI mean allocate_tempbuf()
17:45:56Slasherithat is a bug
17:46:07Slasheriit should use the whole audiobuffer
17:46:09amiconnI don't understand what this is good for
17:46:35Slasheriyes, it was before it allocated only smaller buffer and did the committing on the background
17:46:54Slasheribut i think it's better to do on the foreground because it might need more memory
17:46:57***Saving seen data "./dancer.seen"
17:47:01Slasheriand with sorting it definately needs more
17:47:53amiconnWith your variable-length string structures the database gets smaller, but the old-style tagdb can easily exceed the archos RAM size
17:48:13amiconnThat's why there was discussion about using merge sort for sorting
17:53:34lostlogicxhow would I test something I'm trying to do to improve LCD update speed?
17:54:05 Nick lostlogicx is now known as lostlogic (
18:02:50amiconnlostlogic: My method would be to write a test plugin that times lcd_update()
18:03:18amiconnJust call it a certain number of times and check the number of ticks passed
18:08:04 Join webguest04 [0] (
18:08:20webguest04hello everybody
18:08:35webguest04i have a question concerning the recording monitor
18:09:22Nibblerhi webguest
18:09:26webguest04i'm a musician and play the bass... sometimes there is no amplifier around and i want to practice... so here's what i do:
18:10:07webguest04i connect the bass to a minidisc and hit record and pause... this way, i am able to hear myself
18:10:09amiconnSlasheri: Some more test results, iriver h140 this time:
18:10:12webguest04with earphones
18:10:21lostlogicamiconn: I wonder if I could profile it :-D yeah, I'll probably just do that.
18:11:00webguest04now the problem is, that the signal from the bass is only mono, i hear it on one side only...
18:11:07amiconnForce tag cache update: 4 min 10 sec (3157 audio files), _tmp.tcd 497145, Rolo: boot time +20 sec
18:11:52amiconn_0.tcd: 11248, _1.tcd: 7452, 2.tcd: 208, _3.tcd: 62985, _4.tcd: 222207, _idx.tcd: 63192
18:12:01amiconn...and the same quirks as on archos
18:12:28webguest04is it possible to write a plugin for rockbox that makes a mono signal from line in stereo and redirects it to line-out and headphone out?
18:12:34 Join akaidiot [0] (
18:12:50mirak_webguest04: I play bass too
18:13:31mirak_webguest04: I did that trick with a tape player
18:13:54webguest04it's good for playing on long train "rides"
18:13:55mirak_I used a dummy tape you put into cars players
18:14:09mirak_and puted it into the walkman
18:14:11amiconnSlasheri: I noticed another quirk, not sure if we could/should do something about it. On my iriver, I have the very same album stored 3 times in different formats (mp3, and for testing flac and wavpack).
18:14:33mirak_webguest04: well at this point I don't mind playing without amplification at all :D
18:14:45amiconnIn 'Artists'->(artist), the album is shown only once, but when entering the album dir, all tracks are shown 3 times
18:15:00webguest04yeah, but it's quite different without amplification
18:15:53webguest04do you have experience in rockbox programming?
18:15:54mirak_webguest04: well I sweet most with the big sound
18:16:14amiconnI also found the reason why my first try on archos gave an incomplete _tmp.tcd, as I reproduced the same problem on iriver by accident:
18:16:41amiconnThe background update thread seems to react on SYS_ messages by stopping the background scan
18:17:38mirak_webguest04: not much
18:17:38amiconnTry it yourself: Start a fresh scan with the charger unplugged, then plug the charger after a minute or so => the background scan will stop
18:17:48webguest04mirak: do you know if it's possible?
18:18:14webguest04mirak: i mean write a plugin that redirects its mono input to a stereo output
18:19:49webguest04should be trivial i think
18:20:06webguest04but i never had a look at the plugin architecture
18:22:43mirak_can't you record mono from line in ?
18:25:13 Join muesli_- [0] (i=muesli_t@
18:25:15 Quit webguest04 ("CGI:IRC (Ping timeout)")
18:26:16Slasheriamiconn: ah, interesting
18:26:40Slasheriyes, it will stop if the thread notices the command queue is not empty
18:27:46amiconnThe SYS_ events are broadcast to all threads
18:28:03 Join Sanitarium [0] (n=test@
18:28:04amiconnIt makes sense to stop the scan on SYS_SHUTDOWN
18:28:22amiconn...but not for other SYS_ events
18:29:19amiconnErm, I mean SYS_POWEROFF
18:32:00SlasheriHmm.. but then i would need to design the function differently to return occasionally and check what type of messages are in the queue
18:32:22Slashericurrently it can't detect the type of messages
18:34:38 Join dpassen1 [0] (
18:37:29 Join DreamTactix291 [0] (
18:39:43amiconnSlasheri: I think this will be necessary anyway. There is quite a number of SYS_ events which need different handling
18:39:47 Join RotAtoR [0] (
18:40:02amiconncheck firmware/export/kernel.h
18:40:13Slasheriamiconn: yes, true.. i will find some way to do that
18:43:07lostlogichow big of an IRAM LCD buffer do you think makes sense? 1kbyte? 1kpixel? more? less?
18:43:29Sanitariumhow do you declare something?
18:46:50 Quit Rob2222 ()
18:48:17SanitariumI am adding a menu and it says it is undeclared
18:48:23Sanitariumhow do I fix that
18:53:49preglowdo you even know c programming?
18:54:36Sanitariumbut I am learning
18:55:51 Quit dpassen1 (Read error: 110 (Connection timed out))
18:56:22 Nick Sanitarium is now known as San| (n=test@
18:57:57mirak_lostlogic: that's on kernel iram space ?
18:58:56 Quit DreamTactix291 (Read error: 110 (Connection timed out))
19:03:12San|amiconn: is this good
19:07:40 Quit San| (Read error: 104 (Connection reset by peer))
19:07:45 Join Rob2222 [0] (
19:20:53amiconnlostlogic: What should an IRAM LCD buffer be good for?
19:21:05amiconnWe can't fit the whole framebuffer in IRAM
19:21:17lostlogicmirak_: yeah
19:21:25lostlogicamiconn: see post on the dev list
19:22:36mirak_I need a little resume. for a plugin, a variable declared outside a function is part of the binary code of the plugin.
19:22:42amiconnI don't see how using IRAM will speed up LCD updates if the framebuffer itself is in SDRAM
19:22:58lostlogicam testing
19:23:02mirak_the variables inside functions are put on the stack
19:23:17amiconnWhat will help update speed is using burst mode for reading from SDRAM, but we don't need IRAM for that
19:23:22mirak_the stack is in iram right ?
19:23:23amiconnJust 4 registers
19:23:36lostlogicyep, doesn't help, oh well was worth testing.
19:23:49amiconnmirak_: Depends on the thread. Not all thread stacks are in IRAM
19:24:12mirak_amiconn: for a plugin you can have more than one thread ?
19:24:45amiconnlostlogic: I hope that more H300s will appear on ebay the next days. Then I'll acquire one
19:25:33mirak_amiconn: why don't you wait the successort ?
19:26:16amiconnWhich successor?
19:26:33amiconnThere is no direct successor afaik.
19:26:36mirak_I wouldn't buy a H300 actually
19:26:39mirak_it's a bit to old
19:26:53lostlogicsigh, was hoping that that idea on the devml would speedup the LCD update, I'm going to try copying bytes to registers as well, but I doubt that will help either
19:27:01mirak_no but we are expecting one amiconn
19:27:29amiconnThe H300 is actually the iriver with the best button concept for me
19:27:42 Quit Rob2222 ()
19:27:58amiconnH10 introduces that strange slider, which is similar crap as the ipod touchwheel
19:28:25mirak_maybe it's best but I think it's a bit crappy
19:28:31mirak_I have one though
19:28:40mirak_the multidirectional cross is a pain to use
19:29:09mirak_well not that horrible, but it could be better
19:30:07mirak_but well if you want rockbox support
19:30:19mirak_I would think of the ipod video because of video
19:30:58amiconnI won't buy an iPod, for several reasons
19:31:10amiconnVideo should be possible on H300 as well
19:32:18mirak_amiconn: well if you an help to optimise xvid :D
19:32:21amiconnI don't really care about video; it's a nice gimmick, nothing more
19:32:52mirak_that's what I though, but I used it sometimes in the train
19:32:59amiconnIf I would care, I would buy a PMP like one of the archos AV series
19:33:05mirak_if you don't want to read or sleep it's nice
19:34:16mirak_I am not sure it's worth working on video. I am not sure we can beat what iriver did with his firmaware amiconn
19:34:52mirak_I don't know rockbox enough to be sure of that, but it seems hard
19:34:56mirak_at least for mpeg4
19:35:24mirak_maybe with a less compressed format we could have a better frame rate
19:36:13amiconnI'd think mpeg2 should be within reach for usable frame rates
19:36:17mirak_if it's possible to achieve 25 frame per seconds I don't mind having file of 1 giga instead of 200megs for a 220*176 video
19:36:17 Join Rob2222 [0] (
19:36:23 Join Ikarius [0] (n=ikarius@
19:37:02mirak_we need anyway to reencode videos
19:37:52amiconnOn archos we have realtime video ;-P
19:38:44mirak_you have a gemini ?
19:38:49amiconn750MB for a 3-hour film isn't too bad (including audio)
19:39:05amiconnmirak_: No, I mean rockbox on archos
19:39:21mirak_in 220*176 750mb ?
19:39:28amiconnOf course not
19:39:45mirak_in grayscale ?
19:40:11amiconnActually it's 67fps black & white, but it looks like grayscale
19:40:27mirak_is there some compression on video ?
19:41:16mirak_I am motivated to try to do a mpeg2 plugin like I did for xvid
19:41:34mirak_and see if that's better
19:41:47mirak_we would lose on battery probably however
19:42:39amiconnI don't think the difference will be that big
19:43:03amiconnIn fact we might even save battery power with mpeg2, depending on how fast we'll get the decoder
19:43:07mirak_on battery or decode speed ?
19:43:38amiconnOn archos, I did an interesting test with video playback:
19:43:48 Quit `3nergy (Read error: 104 (Connection reset by peer))
19:44:45 Join ataxy [0] (
19:46:13amiconnStandard cells (1500 mAh) allow ~10 hours of music playback (160 kbps mp3). 160 kbps means rebuffering every ~80 seconds
19:46:58***Saving seen data "./dancer.seen"
19:47:14amiconnVideo playback requires rebuffering every ~20 seconds (given the same audio bitrate, plus the constant video bitrate of 480kbps)
19:47:55amiconnVideo playback runs for >4 hours with the same stock batteries...
19:50:39ataxyhi is markun online
19:51:06ataxyoups yeah just saw im in the list
19:51:37ataxywant to know if you have received the source code for the gigabeat
19:52:24markunI'm here
19:52:34markunno, didn't receive it yet
19:53:28markunmirak_: did you decide which mpeg2 decoder you will use?
19:53:52markunataxy: I just tried to run some code off the Gigabeat, but it failed..
19:53:57 Quit Rob2222 (Read error: 104 (Connection reset by peer))
19:54:33ataxyok i am waiting for it to and want to know what they are sending exactly aside from the kernel code
19:54:53ataxyoh you did
19:56:06ataxyi really hope to see something come out cuz its a great little machine but the job they did os wise, pardon me the word, sux
19:57:17 Quit akaidiot ("( :: NoNameScript 3.81 :: )")
19:58:31preglowmarkun: so, whats going on in the gigabeat camp?
20:00:59markunpreglow: Just replaced a binary file called from the firmware update with a small script that should have displayer "I_Love_Rockbox" but it didn't work.
20:02:11markunpreglow: take a look at this:
20:02:24markunIt's what I got from 'strings ledcont'
20:03:03preglowmarkun: received the source code yet?
20:03:30markunLooking at "usage : ledcont ledNo Cont Bright" I guess it's only used to change some led (backlight?) brightness
20:04:32markunbut objdump shows that many interesting sounding functions are also included in the file
20:04:37amiconnpreglow: Afaiu, no special option is needed to make gcc generate arm thumb code. I googled for 'configure gcc arm thumb' and found this:
20:05:52 Join San| [0] (n=test@
20:05:54preglowno, not generate, but we dont want just thumb code, we want both 32 bit and thumb code
20:06:15preglowand libgcc thats built by default isnt compatible with that
20:06:50 Join Rob2222 [0] (
20:07:04amiconnWhy both?
20:08:16preglowthumb code is usually used for ordinary code, while 32 bit code is used for hand-optimised parts
20:08:25preglowand the most cpu-intensive parts
20:09:10preglowso we'll probably need to use the -mthumb-interwork flag for gcc
20:09:14preglowand that breaks currently
20:13:37amiconnI what way does it break?
20:15:09markunpreglow: can't we just generate all thumb code and swith to 32 bit mode in the asm blocks ourselves?
20:15:48mirak_markun: no I didn't, so if you have a suggestion maybe
20:16:05 Quit Rob2222 (Read error: 104 (Connection reset by peer))
20:16:08amiconnpreglow: Maybe −−enable-interwork while building gcc will do the trick?
20:16:10mirak_ffmpeg is still as horrible to extract
20:16:49mirak_extract a plugin from
20:18:52markunmirak_: maybe this
20:19:09markundon't know if it's good
20:19:29preglowamiconn: makes sense...
20:20:29preglowmarkun: the problem is that gcc has to use special branch instructions to jump to and from thumb code, incidentally, this instruction also doesnt support conditional execution, so therell be a performance hit
20:21:37preglowhmm, no, thats not right. i dont know why theres a performance hit, then
20:22:07preglowmarkun: but did you get the source code in the mail yet?
20:22:48markunpreglow: no, not yet
20:23:32 Join Rob2222 [0] (
20:29:21preglowthats an unreasonably long time for shipping a cd
20:29:44preglowwhy wouldnt they let you download it instead?
20:29:46markunThey told me to wait 4 weeks
20:35:28mirak_markun: source code for what ?
20:36:45ataxythey told me a month to
20:37:04markunmirak_: linux source for the toshiba gigabeat
20:38:11 Join DreamTactix291 [0] (
20:40:08 Join _FireFly_ [0] (
20:41:43 Join Lear [0] (
20:57:33 Quit saa[b_r]ider (Read error: 110 (Connection timed out))
20:58:36lostlogichmm, 5% performance improvement in h300 LCD by using 7 registers to hold values from sdram and then writing them from there out to the device.
20:59:53amiconnI'd expect significantly more...
21:00:37lostlogic3-5% is what I'm getting... I'm only doing it for 7 words, I'll do 14words as soon as I remember movehi syntax
21:01:14amiconn...and 4 registers should be sufficient. 4 registers allow using line burst reading
21:01:37lostlogicoh yeah, I didn't use movem. that's why I"m not getting better.
21:01:40lostlogicwill fix.
21:02:12amiconnmovem.l (source),%d0-%d3, then a sequence like this:
21:02:53amiconnswap %d0, move.w %d0,(lcd), swap %d0, move.w,(lcd)
21:02:56_FireFly_oh asm-"gurus" among themselves ;)
21:02:59amiconn...etc for d1..d3
21:03:12amiconnOf course the source has to be line aligned
21:05:05 Join actionshrimp [0] (
21:05:06lostlogic*nod* how does postinc addressing work with movem? increment to the next byte after the last byte read?
21:05:26LearNot at all on coldfire, afaik.
21:05:30amiconnmovem always reads longwords, and there is no postincrement or predecrement
21:05:43lostlogicoh, so I see (just got to that page in the PRM)
21:05:54lostlogicthat simplifies in a way.
21:06:07amiconn...but that's no problem. lea.l (16,%a0),%a0 will do the trick
21:06:33lostlogicI'll leave the increment out of the asm block.
21:07:17amiconnI would do the whole lcd_write_data() in asm
21:09:12lostlogicthe burst read from memory, write to LCD should be ASM, but the loop can be done outside with no penalty, and the writing of the trailing few words can be done outside of assembly...
21:09:16lostlogicat least that's how it seems to me
21:10:47lostlogiclcd-h300.s:78: Error: invalid instruction for this architecture; needs 68000 or higher −− statement `movem (%a1),%d1-%d3' ignored
21:11:19lostlogicfigured it out as you said that.
21:11:41lostlogicthink the framebuffer is line aligned?
21:12:29amiconnYes it is. However, you'll need to handle leading non-aligned words as well
21:12:51lostlogichow ... irritating
21:13:32amiconnlcd_write_data often will often be called for a part of the framebuffer only
21:13:36lostlogicuse %16 on the start address?
21:14:19amiconnYou could do that, however, it would increase the amount of data to be transferred
21:16:29amiconnWell, if lcd_update_rect() e.g. requests to write words 4..11, you would transfer words 0..11 if you just line-align the start address
21:17:32lostlogicoh, no no I just meant I would mod 16 it to determine how much to pretransfer.
21:17:47amiconnAh, yes
21:18:27amiconnI would do that loop-less in asm
21:20:09amiconnThere's another possibility for optimisation
21:20:48amiconnThe LCD controller allows to write the GRAM backwards. That'll save one swap instruction per register
21:21:01lostlogicwas wondering about that
21:21:46lostlogicbah, broke the LCD driver atm
21:22:32lostlogicseems like I'm writing to a mode register weird or something, but move.w to f0000002 should only ever write data inside lcd_write_data
21:23:32amiconnpreglow: I'm currently test-building binutils + arm-elf-gcc with −−enable-interwork −−enable-multilib
21:24:18amiconnSeems these options are commonly used for arm cross compilers
21:27:10Slasheriamiconn: interesting. Looks like this describes the one playback bug you recently had:
21:28:08amiconnYes, looks like the same problem at least
21:28:10LearBtw, anyone who knows why the expression "(1 << 32) - 1) yields the result 0xFFFFFFFF on target while 0 in the simulator (x86-cygwin)?
21:28:15markunamiconn: do you have a ARM player?
21:28:21amiconnThe weird thing is that it doesn't happen always
21:28:25amiconnmarkun: no
21:29:53amiconnLear: I have no idea. 0xFFFFFFFF would be correct for all 32 and 64 bit platforms
21:29:54preglowamiconn: cool, lemme know how a test build of rockbox turns out with -mthumb and the interwork flag enabled
21:29:58amiconnThe wonders of x86...
21:30:18preglowamiconn: im very curious as to how big the binary ends up, its huge with 32 bit instructions
21:30:36preglowamiconn: and yeah, you probably need to add .asm32 (or something) in the inline asm parts
21:30:39amiconnI know, I've built ipod nano rockbox once for a test
21:30:41Learamiconn: maybe a gcc bug... :)
21:31:01amiconnpreglow: .code 32 or .arm
21:31:10preglowright it is
21:31:12amiconnThe other option is .code 16 or .thumb
21:37:44markun.arm and .thumb look nicer :)
21:37:54lostlogicwhole screen updates are nearly 2x as fast using movem.l than not. (21s/1000 vs 40s/1000)
21:42:54lostlogic(4 line flood) why does this break?
21:42:54lostlogic int precount = (int)p_bytes%16/2;
21:42:54lostlogic while(precount−−)
21:42:54DBUGEnqueued KICK lostlogic
21:42:54lostlogic *(volatile unsigned short *)0xf0000002 = *p_bytes++;
21:42:54lostlogic count-=precount;
21:43:20Learlostlogic: part of your tremor patch applied.
21:43:25 Quit Moos ("Happy Christmas @ all")
21:43:29lostlogicLear: render point?
21:43:46LearYes, that and iram tweaks.
21:43:57lostlogicawesome, thanks
21:44:22Learmakes quite a difference.
21:44:30markunLear: how much faster is it?
21:44:50markunah, read it in the commit log :)
21:47:01***Saving seen data "./dancer.seen"
21:48:03 Join tvelocity [0] (
21:58:48 Join stripwax [0] (
21:59:27preglowhi, stripwax
21:59:55 Quit DreamTactix291 (Read error: 110 (Connection timed out))
22:00:12stripwaxwas wondering if anyone has any ideas about my forum posts re Auto Change Directory only working if you're viewing the wps or main menu
22:00:34stripwaxbut i'm just getting latest cvs to see if it's since been fixed
22:10:55Slasheriah, i think i finally found the crossfade and track skipping glitches :)
22:11:10Slasheriand they were indeed connected together
22:11:18lostlogiclcd_update speedup patch submitted to tracker.
22:11:50 Nick Paprica[sleep] is now known as Paprica (
22:11:50DBUGEnqueued KICK Paprica
22:12:12preglowSlasheri: hooray!
22:12:35Slashericommitting now
22:12:42 Join solexx_ [0] (
22:13:08preglowstripwax: weird issue
22:13:19preglowseems i need to install a new build
22:13:56Slasherihehe, committed :)
22:15:49Slasheripreglow: the problem was i buffer overflow that even simulator couldn't catch
22:19:50lostlogic*gives in and signs up for the CVS ml*
22:20:59 Quit tvelocity ("Leaving")
22:25:01 Quit solexx (Read error: 110 (Connection timed out))
22:27:45lostlogicLear: prefer the prettier-in-C, uglier-in-ASM style of assembly endlines. I'll remember that if I have a next patch.
22:30:31 Quit _FireFly_ ("Leaving")
22:31:58preglowSlasheri: could this crash rockbox?
22:32:19Slasheripreglow: i doubt that, but it will corrupt the audio frames
22:32:34Slasheripreglow: it can't overflow so much that it would crash it
22:32:47lostlogicman it's good to have my Shure E3s back. Can listen to music at -60dB indoors.
22:33:31 Quit darkless (Read error: 104 (Connection reset by peer))
22:33:55 Join darkless [0] (
22:37:20 Join saa[b_r]ider [0] (n=saab_rid@
22:39:39Learlostlogic: how is the bass with those? I have E2c:s, and they are a bit weak in that department on some tracks...
22:40:11 Join akaidiot [0] (
22:40:12lostlogicLear: also a bit weak, crystal clear, but lack boom. I think you have to go up to E4 to get real bass response from shure.
22:40:34Learpretty expensive though.
22:45:42amiconnlostlogic: The precount computation you've posted is wrong
22:46:55lostlogicon patch tracker, or in here?
22:47:45lostlogicnot %16/2? why?
22:47:51amiconnYour formula calculates the number of words between the start address and the next-lower aligned address
22:48:10lostlogic16-%16/2 then?
22:48:38amiconnYes, but with the exception that 16 becomes 0
22:49:30amiconnI'd use the following: ((-(int)p_bytes) & 0xF) / 2
22:51:23amiconnIt'll probably end up as the same asm code as ((-(int)p_bytes) % 16) / 2
22:51:32amiconnAh, no
22:51:45 Quit akaidiot ("( :: NoNameScript 3.81 :: )")
22:51:53amiconn'%' obeys the sign, and the implementation is compiler dependent iirc
22:52:20lostlogic0xF is probably more clear in this situ eh?
22:55:58 Join leftright [0] (
22:56:56leftrightSlasheri: trying your fix now, and I dont hear any artifacts when skiiping, purrrfect
22:57:08Slasherileftright: sounds good :)
22:57:30leftrightthats with xfade off, sounds good
23:01:33 Quit BHSPitMonkey (Read error: 104 (Connection reset by peer))
23:01:35 Join akaidiot [0] (
23:05:30 Quit stripwax ("Trillian (")
23:06:59 Part leftright
23:07:00preglowpretty build table
23:07:30 Quit ataxy ("CGI:IRC (EOF)")
23:07:34 Join ataxy [0] (
23:09:12 Join YouCeyE [0] (n=YouCeyE@unaffiliated/youceye)
23:10:28 Quit ataxy (Client Quit)
23:14:34 Join BHSPitMonkey [0] (
23:16:50 Join LinusN [0] (
23:18:39LinusNSlasheri: u there?
23:22:47San|he was 20mins ago
23:31:45DrumRBoy320|Away[16:57] Slasheri's last activity
23:33:02 Quit markun (
23:33:02 Quit thegeek_ (
23:33:02 Quit Rob- (
23:33:05muesli_-its 23:33 ;)
23:33:07 Join markun [0] (
23:33:33NJoinRob- [0] (
23:35:42 Join dreadhead [0] (
23:35:51 Join thegeek [0] (
23:37:11 Quit Rick (Read error: 104 (Connection reset by peer))
23:37:16 Join Gibbed [0] (
23:38:07San|no, its 22.38
23:38:17DrumRBoy320|Awayw/e :P
23:38:27DrumRBoy320|Awaycan someone give me an example for a trigger script?
23:38:35San|[21:57] <Slasheri> leftright: sounds good :
23:38:35San|[22:18] <LinusN> Slasheri: u there?
23:38:48DrumRBoy320|AwayWHATEVER MAN!
23:38:57DrumRBoy320|Awaycan someone give me an example for a trigger script?... please?
23:39:08San|haha, wrong window
23:39:08lostlogicnone of you have the time right... it's just 16:39 now :-P
23:39:13 Join DreamTactix291 [0] (
23:39:50 Join mikearthur [0] (
23:39:52DrumRBoy320|Awaycan someone give me an example for a trigger script??
23:40:00lostlogicstop repeating yourself!
23:40:18DrumRBoy320|Awaysorry, im repeating it to people who are just joining... lol
23:40:28DrumRBoy320|Awayincase they know
23:41:16muesli_-Paprica ?
23:42:00San|pappy! seem not to be registered
23:42:14muesli_-thus you cant query..
23:42:20Paprica[::] -NickServ- You have already identified
23:42:56muesli_-i queried you but you didnt answer
23:43:11Papricai dont see
23:44:06muesli_-did you update your brick mania.rock on our website?
23:45:06 Join linuxstb [0] (
23:45:22Papricai have no complete what i do noe
23:45:42Papricaafter i will complete
23:45:50Papricaim update it
23:45:54muesli_-oki..why dont you commit your .rock at all?
23:45:57 Quit markun ("leaving")
23:45:59muesli_-commit to cvs
23:46:27Papricacuz its dont ready to be in the cvs
23:46:38Papricai want to complete it first
23:47:03***Saving seen data "./dancer.seen"
23:47:37muesli_-just out of curiosity
23:49:34amiconnLinusN: Did you see the tagcache statistics for archos in the logs?
23:50:10amiconnLooks promising... only some glitches to fix
23:50:12LinusNi had a hard time making slasheris patch work at all
23:50:38amiconnHis latest patch compiled fine for all targets I tried, with just one warning
23:50:59LinusNmine didn't work at all, the tagcache thread wasn't started
23:51:17LinusNand the forced update didn't work either
23:51:30amiconnYes, Slasheri forgot the init in some earlier diffs
23:51:43LinusNthis was yesterday
23:51:56LinusNurl to a newer one?
23:53:23 Join `3nergy [0] (
23:54:23amiconnThe dcc'd version is the one that worked for me, applied to a clean cvs checkout. The only change to make it work on archos is changing apps/tagcache.c line 700 to read:
23:54:32amiconnint len = (strlen(data) + 2) & ~1;
23:55:08 Part dreadhead
23:55:34LinusN10 minutes for 3300 files?
23:55:45amiconnCheck today's log starting at ~11:15 and ~18:10 for stats
23:56:11amiconn10 minutes plus the 30 second at next boot
23:56:45LinusNnot *that* shabby
23:56:56LinusNi thought it would be longer
23:57:06amiconnIt seems HD access is the limiting factor, not CPU speed
23:57:26amiconnWe have asm optimised ata transfer on archos :)
23:57:54LinusNwe could have dma on iriver ;-)
23:58:24amiconn...if there were enough DMA channels
23:58:47LinusNwell, we have two channels, we could use an allocating scheme
23:58:55LinusNfor the second one
23:59:00 Join ataxy [0] (

Previous day | Next day