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 2002-08-15

00:02:21 Quit Zagor ("Client Exiting")
00:13:36***Saving seen data "./dancer.seen"
00:28:06RipnetUKgot the entire tree view scrolling now :)
00:28:18RipnetUKon hardware, nicely done with #ifdefs to turn it off
00:29:18dwihnoI just tested the ffw
00:29:20dwihnoI love it
00:29:22RipnetUKdoes seem to be a little slow on the real hardware now... will have to optimize the plotting of the window I think, was OK when only 2/3 of screen, so not too far away... will try and tidy up tomorow
00:29:41BagderRipnetUK: you still do the copying bit by bit?
00:29:57Bagderok, then you'll have a lot to gain still ;-)
00:30:00RipnetUKgonna optimize that a bit soon and do it 8 bits at a time
00:30:16RipnetUKyeah, plenty of & and | going on there, as well as shifting, and multiplying needlessly
00:30:29BagderI can imagine
00:30:53Bagderwhen you have that sorted out, I'll line up for a test patch ;-)
00:31:59RipnetUKi have stuck a compiled recorder firmware at if neone would like a look, but any comments can u email me, as I need sleep :)
00:32:15RipnetUKcant be arsed with diff tonight
00:32:19Bagderok, I might have a go, thanks
00:32:46Bagdernight RipnetUK
00:33:03BagderI have this terrible packet loss atm
00:33:19RipnetUKoh dear... u using parcelforce ;)
00:33:37Bagderno, plain fine adsl
00:33:48Bagderdon't know what's going on
00:34:12RipnetUKwhen that happens to me, usually 'other end'... anyway, gotta go, catch ya later
00:34:24Bagdergo sleep!
01:31:27 Quit datazone-work ("I like core dumps")
01:40:52 Join BoD[] [0] (
01:40:56 Part Bagder
01:41:00BoD[]hello !
01:41:45BoD[]hey I have a little question .. what does the mp3/m3u filter do exactly ?
01:42:08PsycoXulshows only mp3/m3u files if on, or all files if off
01:42:21BoD[]ah ok thanks :)
01:44:29BoD[]oh I have another question... is there an easy way to change the boot screen logo ?
01:46:26adi|homelast i checked you should be able to add a file to the archos root directory by the same name as the logo
01:46:33adi|homeand if that is there, the firmware should load it
01:46:44adi|homeif not, it pulls from source, but Im no 100% positive
01:47:14adi|homefile name rockbox112.bmp
01:47:40BoD[]well i'll try that
02:13:39***Saving seen data "./dancer.seen"
02:29:18 Join MeRWiN [0] (~merwin@
02:29:40MeRWiNadi: Wow, omnipresence.... home and work :)
02:29:49 Quit datazone (Remote closed the connection)
02:33:32adi|homei have lotsa talent :)
02:33:59MeRWiNi can see!\
02:34:12MeRWiNyou tried the fast forward/rewind yet?
02:46:54 Join datazone [0] ([85DgBAvim@
02:49:28dwihnoI tried it
02:49:30dwihnoI like it
02:56:20MeRWiNseem to work nicely
02:57:34BoD[]I noticed that the loading of a song is slower than with archos' firmware
02:57:46BoD[]but of course the loading of a playlist is 1000x faster
02:57:58BoD[]is this normal ?
02:58:02MeRWiNWe're working on speeding it up... it's a tough task with so little ra
02:59:09BoD[]but is it related to the loading of the playlists ? or not at all
02:59:49BoD[]like ... archos' firmware is slow because they do work to speed up the loading of songs after
03:02:21MeRWiNIt's not related to the playlists, no... we just do playlists better than them
03:02:30MeRWiNand at present, our file loading is a bit slower than theirs
03:04:13BoD[]i was just wondering :)
03:04:16BoD[]thanks !
03:04:46BoD[]i have to go to bed ... 3:02 :)
03:04:53BoD[]bye !
03:05:41MeRWiNbye :)
03:06:13 Quit BoD[] ("dodo")
03:39:09 Quit mistax ()
03:41:32adi|homei love the fact that it seems like each day we get more users :)
03:54:29MeRWiNI think the userbase is exploding
03:54:32MeRWiNespecially since 1.2
03:55:13PsycoXul1,2 cleaned up the mess that was 1.1
03:55:30MeRWiNPsycoXul: exactly
04:00:48PsycoXulsomebody needs to find and mp3 all the cut material from the hitchhikers guide to the galaxy radio show, the stuff that didn't make it to broadcast
04:01:28MeRWiNI'm sure somebody has it
04:02:00adi|homei just want the full mp3's period..
04:02:04adi|homei haven't heard it a,.ll
04:02:10adi|homei read the books.. but not the radio show
04:02:34MeRWiNi've never heard the radio show either
04:05:24PsycoXuli've got it
04:05:34PsycoXulits like 22khz 56kbps
04:05:51PsycoXulbut its mostly voice anyways, and is 6 hours long...
04:05:53MeRWiNthat's not bad for radio
04:06:02PsycoXul140mb total
04:06:10PsycoXulif somebody has a place i'll up it...
04:06:16PsycoXulthough i found it through google i think
04:12:02MeRWiNmaybe i'll have you up it tomorrow night when I'm at home. Apparently my home computer shut off :P
04:12:11PsycoXulheh ok
04:13:43***Saving seen data "./dancer.seen"
04:29:45datazonehey adi
04:30:32datazonehows the directory icons coming?
04:31:45MeRWiNspeaking of icons, are there going to be player icons for dirs, or is this what you're talking about?
04:39:38PsycoXulwe could have
04:39:52PsycoXulput them in the column where the selection arrow is
04:40:07PsycoXulexcept that it'd be covered up by the arrow... heh
04:41:11PsycoXulwe could put the icon for the entry being pointed at in the blank space but that might be confusing for people
04:43:45MeRWiNmove the words over to the right one more
04:44:04PsycoXulyeah but 11 chars is too short already
04:44:10PsycoXuland it's at 10 now with the pointer
04:44:18PsycoXulwhich i'd already like to try to find an alternative to
04:44:45datazoneadi was working on some icons
04:45:07datazonethe only thing that upsets me is that names shorter than the length of the screen always ends up in all uppercase
04:45:10datazoneso distracting
04:46:06PsycoXulif the player gets a compiled-in font and depending on the user-definable patterns functions/api we could invert the first char of the selection entry as a pointer and use all 11 chars...
04:46:19PsycoXulor we could do that and keep it with 10 chars and put the icons in the first char
04:49:53adi|homedatazone: their not.. i stopped working on it.. im planning on hit it more tonight...
04:50:34MeRWiNdidja see my id3 info displayer in the player? :-)
04:58:04 Quit Low[a] (Read error: 110 (Connection timed out))
04:58:07 Join Low[a] [0] (
05:21:30MeRWiNg'nite all
05:22:15 Nick MeRWiN is now known as MeRWiN|Sleep (~merwin@
05:22:46MeRWiN|SleepGot a flight at 6:30am EST (it's 11:30pm right now). Nice 4 hour nap ahead of me. heh.
06:13:45***Saving seen data "./dancer.seen"
06:18:33 Join aakil [0] (
06:43:54 Join Zagor [0] (
07:01:57 Quit aakil (Read error: 110 (Connection timed out))
07:06:19adi|homezagor you see the mail about the mailing list getting 'unfriendly;?
07:06:37Zagorhaven't got to it yet
07:06:59adi|homei respond to the best of my restrained ability
07:07:08adi|homebut i only spoke for myself so..
07:11:13elinenbe|storenick elinenbe
07:11:25elinenbe|storethat bug yesterday was just my build
07:11:38elinenbe|storeI got a clean CVS and it is great now.
07:11:44 Nick elinenbe|store is now known as elinenbe (
07:11:56elinenbebut there still seems to be a bug or two
07:12:17elinenbeif you fast forward through a song until the end, it will then start playijng the next song without updating all information on the display
07:12:28elinenbestatus bar knows as soon as the plug is plugged in, but it takes maybe 30 seconds to start the battery charging graphic
07:12:59Zagoryes, the fwd-to-end-of-song was noted in my mail response to hardeep
07:16:12 Join LinusN [0] (
07:16:18LinusNShow me the code!
07:17:26LinusNBeen up all night?
07:17:47LinusNso why are you up so early, it's not like you
07:18:05Zagorhehe. it gets too hot to sleep in the morning.
07:18:24Zagori can't wait for this darn summer to be over :-)
07:18:37LinusNhave to get up early to snap off some discussions in the mailing list
07:19:04elinenbeBoth LinusN and Zagor were on less then 8 hours ago. You guys need more sleep :-D
07:19:08Zagoryou left a warning in mpeg.c, btw
07:19:19Zagorelinenbe: hehe
07:19:23LinusNand you felt a nasty bug in wps.c
07:19:36Zagorwhat is that?
07:19:56LinusNwhen you play a file from the browser, the next file is played instead
07:19:58adi|homeLinusN: my reply wans't out of line on the list was it?
07:20:12LinusNon the contrary
07:21:14LinusNthe wps steps to the next file on key-up events, even those that were pressed in the dir prowser to play a file and when exiting the menu
07:21:23elinenbeLinusN: did you get my new bug reports?
07:21:28elinenbeif you fast forward through a song until the end, it will then start playijng the next song without updating all information on the display
07:21:39elinenbestatus bar knows as soon as the plug is plugged in, but it takes maybe 30 seconds to start the battery charging graphic
07:22:48LinusNelinenbe: the charging is supposed to start late
07:22:58elinenbeLinusN: why is that?
07:23:04LinusNit must gather some statistics about the battery status before charging
07:23:51LinusNthe fast forward has a lot off issues
07:23:56elinenbeLinusN: It can't just dump into charging mode and then refine itself as it gathers the data?
07:23:58LinusNbut it is a step forward
07:24:07LinusNelinenbe: why?
07:24:25elinenbeLinusN: also, the default value for the contrast is still quite low.
07:24:35LinusNfor the recorder?
07:25:02elinenbeLinusN: I just think that the status bar should display the charging battery −− for the users sake. they may think something is wrong as I did.
07:25:11elinenbeyeah, for the recorder.
07:25:17elinenbeit is quite dim here on my end.
07:25:29elinenbeI run it at 44.
07:25:30LinusNit's perfect for me
07:25:39LinusNat 44 it's totally dark for me
07:25:56elinenbeI can hardly see anything at the default value.
07:26:00LinusNi guess that's why the contrast is adjustable
07:26:17Zagorelinenbe: most people get confused with the battery status during charging and pull out the plug too early. that's why we decided to remove it.
07:27:10elinenbesuggestions: instead of a skip size for the ff/rw −− just let it ramp up the longer you hold down the button.
07:27:39Zagorelinenbe: yes. i'm working on repeat accelleration
07:27:51elinenbesuggestion: ff/rr into next and previous song respectively...
07:28:32elinenbesuggestion: keep the step size, but add the ramping feature into the options menu. some people will like it different ways.
07:28:38elinenbesuggestion: I need to go to bed.
07:28:46 Nick elinenbe is now known as elinenbe|sleep (
07:28:48LinusNsuggestion: go to bed
07:28:58LinusNif(tired) goto bed;
07:31:06elinenbe|sleepone last thing... there is a story on about mp3 players and the archos and rockbox get quite a few shout outs...
07:36:59Zagoryeah, I saw it too.
07:41:24LinusNZagor: are you working on the FF key-up "feature"?
07:41:41Zagornot right now, no. fix it if you have time.
07:43:14LinusNthe balance is quite lame on the Player at the moment, but it works
07:43:22LinusNthe implementation is lame
07:43:58Zagorok. i haven't looked at it. just noticed it leaves a warning, will you fix that too?
07:46:49LinusNi was thinking of a RAM usage setting
07:47:10LinusNwhere you can sacrifice buffer RAM for dir cache and stuff
07:47:37Zagorbah. why?
07:47:48LinusNto be less hostile? :-)
07:48:42Zagorthose kinds of people can never be pleased. they are just pissed it can't do the same things as their 256MB unix workstation
07:49:09LinusNi assume it's a 256MB WinXP station
07:49:32Zagoruntil someone comes up with a *reason* for supporting huge dirs, there will be no such support
07:49:44Zagorand "because it's a limitation" is not a valid reason
07:49:49LinusNZagor: now you're not fair
07:50:03ZagorI'm not?
07:50:09LinusNthe reason is of course that they have more than 200 files in a dir
07:50:42Zagorand they're too lazy to organise their collection, so they scream at us instead. yeah, I know.
07:50:52Zagoris that a good enough reason for you?
07:50:53LinusNno, they are not lazy
07:51:03LinusNthe prefer to organize their files that way
07:51:21Zagornot so
07:51:22LinusNyour way isn't necessarily the *right* way
07:51:37Zagorthe single guy reporting >200 files said it was his download dir
07:52:33LinusNmeybe so
07:53:14Zagoranyway, our dir cache fix will probably bump that limit a bit for many people
07:53:31LinusNi have made some attempts to fix that
07:53:42LinusNit's not that simple as it may sound
07:54:16LinusNthe current cache is an array of structs, with a 260 byte name and an attribute
07:54:37LinusNand the cache itself is an array of pointers to those
07:54:47LinusNthat pointer array is sorted
07:55:36LinusNwe have the option to replace the 260-byte array with a char pointer to a packed buffer
07:55:58LinusNthen we'll have one array of pointers, one array of structs and one char buffer
07:56:28LinusNi think that's an array too mahy
07:57:46Zagorwe can skip the pointer array if we don't want the performance boost it gives to sorting
07:57:58 Join Bagder [0] (
07:58:08LinusNyeah, we can always sort the structs themselves
07:58:16CtcpIgnored 2 channel CTCP requests in 1 hour and 5 minutes at the last flood
07:58:16*adi|home slaps the happiness outta Bagder
07:58:17Zagorthe boost is not as large when we use a pointer instead of a 260-byte array
07:58:27Zagorhi bro
07:58:33LinusNBy the way, i have the dir-messup that elinenbe spoke of yesterday
07:58:56LinusNyou know, dirs displaying the wrong contents
07:59:00adi|homethe dir structure.. where is that held?
07:59:22LinusNadi|home: on the disk, silly! :-)
07:59:32adi|homeno no ono...
07:59:44adi|homei mean.. where is it defined ;)
08:00:07Zagorin dir.h
08:00:54Bagderthere's a guy in Congo that needs my help too! B-]
08:01:14adi|homegod i want that chick
08:01:25adi|homeAvril Levigne on MTV
08:01:28*adi|home drools
08:02:55Zagoryeesh, what is she? 15?
08:03:07adi|homedon't mean i don't want to nail here
08:03:13adi|hometheres just something about her
08:03:30adi|homesame thing gwen stefanni had when she first broke out
08:05:37Bagderfast-forward is coool
08:06:42Zagoryeah. and buggy :-)
08:07:13Bagdera minor detail I've noticed
08:07:21Bagderwhen I use F1 to get to the menus
08:07:27Bagderrun around there for a while
08:07:37Bagderthen press LEFT to go back to the WPS
08:07:45Bagderit very often counts as a LEFT in the WPS too
08:07:52adi|homein the directory structure.. how often do we reach the limit of 'MAXPATH;?
08:07:55Zagorthat's odd
08:08:03Zagoradi|home: very rarely
08:08:04adi|homei mean.. is it common to hit that limit?
08:08:13Bagderadi|home: it hardly ever happens
08:08:33adi|homehmmmm... okay... mind if i play devils advocate and brain storm for a sec?
08:08:50adi|homeelse ill find something else to think about
08:09:17adi|homeokay.. one guy bitchs about the 200 limit
08:09:23adi|homewe agree we don't nec. like the idea
08:09:25LinusNZagor: i told you that the key-up events from previous screens are sucked up by wps
08:09:33adi|homebut that right now its the best solution we have right?
08:09:55ZagorLinusN: uhh, right. i just haven't switched on my brain yet :)
08:10:09Zagoradi|home: not really, we have a better one
08:10:20adi|homewhats that?
08:10:30Zagora fixed-size memory buffer
08:10:32LinusNZagor: i am on the case
08:10:44adi|homeright, and you just allocate from it as nec yes?
08:10:51Zagorinstead of 200 256-bytes arrays, we store the names right after each other
08:10:58*adi|home grins
08:11:05adi|homeoh sure.. here i am thinking i had a unique solution
08:11:10*adi|home goes back to picking his nose
08:11:36adi|homehow would we deal with bubbling then?
08:11:53Zagorwe don't. the buffer is cleared every time we read a new dir.
08:12:11adi|homewell.. lets say we have the single length array...
08:12:24adi|homewith names stored sequentially, and the structure having a ptr to the location int hat arry
08:12:36adi|homewell.. are we still going to say that there is a limit of the # of files in a dir?
08:12:49adi|homeor would we just 'page' that memory when you reach the end of the list?
08:12:51Zagorno. we tell it like it is
08:13:07Zagorwe pop up a message "dir size limit reached"
08:13:12*adi|home nods
08:13:20LinusN"dir buffer full"
08:13:22adi|homeso... whats the downside of doing this?
08:13:24Zagorpaging kills sorting, and we'll just catch more flak for that if we should implement it
08:13:48***Saving seen data "./dancer.seen"
08:14:07Zagoradi|home: nothing, really. we've had the idea a week or so, just haven't fixed it yet
08:14:23adi|homeanyone taken a shot at it yet?
08:14:34LinusNi have, twice
08:14:39adi|homeand no luck?
08:14:53adi|homeor just nothing your happy with?
08:14:54LinusNbut it kind of bit me, as i couldn't decide how to do it in the best way
08:15:14adi|homemind if i ask the 2 ways you did it?
08:15:14LinusNi think i do now, after discussing it with zagor
08:15:30ZagorLinusN: new thought: we should have the structs in the buffer too
08:15:36Zagorumm, no
08:15:39adi|homezagor why?
08:15:45adi|homethat don't make sense
08:15:47Zagordamn. we'll get a fixed limit on the pointer array size anyhow.
08:15:48LinusNbut then we have to align them on 32-bit boundaries
08:16:03Zagoror struct array
08:16:09LinusNZagor: of course
08:16:18Zagoryeah, just my mind catching up :-)
08:16:22LinusNbut that limit should be quite high
08:16:28adi|homecan i assume that with the name buffer, we would need to add an int for name length to the dir struct?
08:16:39LinusNthat is also why i don't want a large struct array
08:16:44Zagorno, we null terminate the names
08:17:07adi|homeand for assignment to that array, we store a ptr so we know where to write the next name?
08:17:16adi|homeor just keep track of the full used length?
08:17:26Zagorwell, it's the same really
08:17:37*adi|home prefers ptrs *shrugs*
08:17:47adi|homeokay.. i think this is my last dead horse question
08:18:13adi|homethe actual dir structs.. umm.. would we have a fixed number of them?
08:18:33adi|homeie: a max of say 500 dir entries or something
08:19:17adi|homeso the solution is to not illiminate the dir max, its just to make it less likely to hit while conserving ram?
08:20:08Zagorwell I'd like the new buffer to be smaller than today's. so what we gain in improved use of memory is given to the mp3 buffer
08:20:27adi|homeso let me see if i fully understand this
08:20:49adi|homewe trash the d_name member of dirent
08:20:59LinusNno no no
08:21:08adi|homelet me finish :)
08:21:19Zagorwe don't touch dir.h, only tree.c
08:21:45adi|homehow does that fix our memory issue?
08:21:59LinusNthe cache is not an array of dirent's
08:22:08Zagorbecause that's where the cache is :-)
08:22:10 Quit RipnetUK (
08:22:10 Quit Hadaka (
08:22:10 Quit matp (
08:22:10 Quit adi|work (
08:22:10 Quit mbr (
08:22:10 Quit TummGBG (
08:22:10 Quit Low[a] (
08:22:10 Quit datazone (
08:22:10 Quit adi|home (
08:22:10 Quit ironi (
08:22:10 Quit bobTHC (
08:22:10 Quit fragglet (
08:22:11 Quit dwihno (
08:22:18Bagdersplit heaven
08:22:32NJoinLow[a] [0] (
08:22:32NJoindatazone [0] ([85DgBAvim@
08:22:32NJoinadi|home [0] (~adiamas@
08:22:32NJoinironi [0] (~ironi@
08:22:32NJoindwihno [0] (dwihno@Bald067.Baldakinen.Umea.SE)
08:22:32NJoinmbr [0] (
08:22:32NJoinTummGBG [0] (
08:22:32NJoinfragglet [0] (~fraggle@
08:22:32NJoinadi|work [0] (
08:22:32NJoinmatp [0] (
08:22:32NJoinHadaka [0] (
08:22:32NJoinRipnetUK [0] (
08:22:32NJoinbobTHC [0] (
08:22:38adi|homewelcome back :0
08:22:40LinusNwelcome back
08:22:50adi|homestruct entry
08:22:55adi|homemy fault
08:23:03adi|homeso "name" becomes a char *ptr
08:23:09 Quit RipnetUK (
08:23:09 Quit matp (
08:23:09 Quit adi|work (
08:23:09 Quit mbr (
08:23:09 Quit Hadaka (
08:23:09 Quit TummGBG (
08:23:09 Quit bobTHC (
08:23:09 Quit dwihno (
08:23:09 Quit ironi (
08:23:09 Quit Low[a] (
08:23:09 Quit datazone (
08:23:09 Quit adi|home (
08:23:09 Quit fragglet (
08:23:15 Join aakil [0] (
08:23:23NJoinLow[a] [0] (
08:23:23NJoindatazone [0] ([85DgBAvim@
08:23:23NJoinadi|home [0] (~adiamas@
08:23:23NJoinironi [0] (~ironi@
08:23:23NJoinbobTHC [0] (
08:23:23NJoinRipnetUK [0] (
08:23:23NJoinHadaka [0] (
08:23:23NJoinmatp [0] (
08:23:23NJoinadi|work [0] (
08:23:23NJoinfragglet [0] (~fraggle@
08:23:23NJoinTummGBG [0] (
08:23:23NJoinmbr [0] (
08:23:23NJoindwihno [0] (dwihno@Bald067.Baldakinen.Umea.SE)
08:23:31LinusNwhat a RIDE!
08:23:35adi|hometo another array say, 200*265 chars long?
08:23:48adi|homeokay.. let me try that again now that your back :)
08:23:53adi|homegot ya..
08:24:13adi|homeLinusN: what were the problems you ran into with your attempts to impliment it?
08:24:20LinusNthe sorting
08:24:45LinusNi have it figured out now
08:24:49Zagorhave you noticed playlist loading is faster now?
08:24:56Zagorand now the cpu is the limit, not disk?
08:25:01adi|homehow are you going to do it?
08:25:11*adi|home thinks he knows 2 ways he would do it, but not sure what is best
08:25:16Zagormy fat_read() fix surely sped things up
08:25:20*Zagor grins
08:25:29LinusNadi|home: i will have an array of structs, and move them around in the array
08:25:35*adi|home nods
08:25:41adi|homethats what i was thinking
08:25:51LinusNthe sorting will be slower, bu i don't think it matters
08:26:01adi|homeits either that or a second array of ptrs to the array of structs.. but thats just worse ;(
08:26:06 Nick Zagor is now known as Zagor|away (
08:26:12adi|homeitll be slower.. but it shouldn't be noticable
08:26:15adi|homethats what i did with my mud
08:26:27adi|homeand figure i have 10000 items or so
08:26:29LinusNit is actually *way* slower
08:26:34adi|homeand you didn't notice it...
08:26:49LinusNin terms of clock cycles
08:26:57adi|homehmmm.. why?
08:27:18LinusNbecause it takes only a single move to move a 32-bit pointer
08:27:37 Quit mbr (
08:27:37 Quit TummGBG (
08:27:37 Quit matp (
08:27:37 Quit Hadaka (
08:27:37 Quit RipnetUK (
08:27:37 Quit adi|work (
08:27:37 Quit dwihno (
08:27:37 Quit bobTHC (
08:27:37 Quit ironi (
08:27:37 Quit datazone (
08:27:37 Quit Low[a] (
08:27:37 Quit adi|home (
08:27:37 Quit fragglet (
08:27:37LinusNbut the compiler will probably use memcpy to move the structs
08:27:44Bagderyou missed him ;-)
08:28:01NJoinLow[a] [0] (
08:28:01NJoindatazone [0] ([85DgBAvim@
08:28:01NJoinadi|home [0] (~adiamas@
08:28:01NJoinironi [0] (~ironi@
08:28:01NJoinbobTHC [0] (
08:28:01NJoinRipnetUK [0] (
08:28:01NJoinHadaka [0] (
08:28:01NJoinmatp [0] (
08:28:01NJoinadi|work [0] (
08:28:01NJoinfragglet [0] (~fraggle@
08:28:01NJoinTummGBG [0] (
08:28:01NJoinmbr [0] (
08:28:01NJoindwihno [0] (dwihno@Bald067.Baldakinen.Umea.SE)
08:28:02LinusNbut the compiler will probably use memcpy to move the structs
08:28:10adi|homewould you guys make up your mind ;)
08:28:11 Quit mbr (
08:28:11 Quit TummGBG (
08:28:11 Quit matp (
08:28:11 Quit Hadaka (
08:28:11 Quit RipnetUK (
08:28:11 Quit adi|work (
08:28:11 Quit dwihno (
08:28:11 Quit bobTHC (
08:28:11 Quit ironi (
08:28:11 Quit datazone (
08:28:11 Quit Low[a] (
08:28:11 Quit adi|home (
08:28:11 Quit fragglet (
08:28:29NJoinLow[a] [0] (
08:28:29NJoindatazone [0] ([85DgBAvim@
08:28:29NJoinadi|home [0] (~adiamas@
08:28:29NJoinironi [0] (~ironi@
08:28:29NJoinbobTHC [0] (
08:28:29NJoinRipnetUK [0] (
08:28:29NJoinHadaka [0] (
08:28:29NJoinmatp [0] (
08:28:29NJoinadi|work [0] (
08:28:29NJoinfragglet [0] (~fraggle@
08:28:29NJoinTummGBG [0] (
08:28:29NJoinmbr [0] (
08:28:29NJoindwihno [0] (dwihno@Bald067.Baldakinen.Umea.SE)
08:28:29DEBUGParse error in OnTopic(from = "", line = "#rockbox :rockbox - home of the revolving door")
08:28:31LinusNi'm getting dizzy!
08:29:03Mode"#rockbox +o adi|home " by ChanServ (ChanServ@services.)
08:29:07Mode"#rockbox +t " by adi|home (~adiamas@
08:29:45adi|homehmmm.. when you do a direct struct copy, how does the compiler do it?
08:29:49adi|homeuse memcpy or what?
08:29:56LinusNi repeat:
08:29:58LinusNbut the compiler will probably use memcpy to move the structs
08:30:03*adi|home nods
08:30:06adi|homemissed it
08:30:32adi|homehmm... well
08:30:48adi|homewould having a second array of char *'s be all that bad... for sorting?
08:31:01adi|homei mean.. you lose another 200+bytes
08:31:13adi|homebut you gain on expanding dir entries limit, and speed.
08:31:14LinusNit will use up 4*MAX_FILES_IN_DIR bytes
08:31:33adi|homeokay... still < 1k no?
08:31:49BagderI don't think its necessary
08:31:55Bagderthe sorting won't take that long anyway
08:32:01Bagdereven if it becomes "much" slower
08:32:09adi|homewould it be noticable?
08:32:12LinusNi agree
08:32:16adi|homei mean.. im all for impliment it and find out
08:32:23adi|homeppl complain.. deal with it then
08:32:23LinusNi'm on the case
08:32:24Bagderlet's at least try it first with the struct copy approach
08:33:06Mode"#rockbox +ooo LinusN Bagder Zagor|away " by adi|home (~adiamas@
08:33:13Mode"#rockbox +o adi|work " by adi|home (~adiamas@
08:33:14*Bagder bows
08:33:26*LinusN is honored
08:33:42adi|homeis there anyway to find out how much v1.2 has been dl'ed?
08:34:00Bagderjust by scanning the logs
08:34:03*adi|home sways back and forth to Van Morrison
08:34:11BagderI don't think Zagor runs any analyzers on a regular basis
08:34:17*adi|home is just curious how popular we've gotten in that sense..
08:34:26Bagderlemme try a quick grep
08:36:01Bagdersince August 6th...
08:36:16Bagder826 daily target builds were downloaded
08:37:44Bagderand amazingly...
08:38:06Bagder1.2 mod/ajz files were downlaoded 1059 times
08:38:14LinusNare you joking?
08:38:28HesThat's target builds only?
08:38:33Hesand simulators?
08:38:37Bagderegrep 'rockbox-1\.2.*\.(mod|ajz)'
08:38:48Bagdershould be release target builds only
08:38:52HesNot bad at all.
08:39:39Bagdersimulators: 469 downloads
08:40:01Heswhat if you add awk '{print $1;}'|sort|uniq to get unique hosts?
08:40:07Bagdergood idea
08:40:57Bagderit makes 635 unique hosts that downloaded the 1.2 release
08:41:09Bagdercut "-d " -f1 | sort -u | wc -l
08:43:02HesHm... why was the config block revision increased again?
08:44:43Hesit should be upgraded only and only if data moves to another location in the rtc block, or rather, when space is reused for a new purpose
08:45:29HesIf we change a definition of a single byte, we can pick a new unused byte for the new value and forget about the old one without upping the revision
08:45:36adi|homebtw.. when does the t-shirt contest close?
08:45:47Hesso that users don't loose their old config
08:45:53Bagderadi|home: the submission period runs through August
08:47:29LinusNHes: i upped the revision again because Zagor reverted my last revision bumb
08:47:41HesLinusN: i think he reverted it for a reason
08:47:49LinusNno, he was ignorant
08:47:58HesWhy was it bumped?
08:48:12LinusNbecause the contrast default was 0 on older configs
08:48:35HesIsn't a contrast value of 0 unusable?
08:48:57adi|homeanyone wanna give me a ballpark on the price of an average pc for home use?
08:49:00LinusNthat is why i had to bump it
08:49:32LinusNand no, a contrast value of 0 isn't necessarily unusable
08:49:49Hes if (rtc_config_block[0xa] != 0xFF) {
08:49:49Hes global_settings.contrast = rtc_config_block[0xa];
08:49:49Hes if ( global_settings.contrast < MIN_CONTRAST_SETTING )
08:49:49Hes global_settings.contrast = DEFAULT_CONTRAST_SETTING;
08:49:49Hes }
08:49:54Bagderadi|home: starting at something like 800 USD for a basic, new, one
08:49:58HesThat's what settings.c reads now, to get over the value of 0
08:50:16HesZagor fixed it that way
08:50:18LinusNHes: why forbid 0?
08:50:38HesWhy not, why would it be unusable?
08:50:43 Join matsl [0] (~matsl@
08:50:51LinusNand why that kludge instead of bumping the revision?
08:50:52Hes... why would it be usable?
08:51:11HesBecause bumping the revision is unnecessary and makes the thing forget the old config.
08:51:42HesRead the original spec I wrote for the config block in the beginning of settings.c?
08:51:49LinusNstill, that forbids a pereftly valid contrast setting
08:52:05Hesbut perfectly useless? 8-)
08:52:42HesOnce we have settings savings, I think we should make the effort to avoid forgetting settings between versions
08:53:12HesConfig memory is reset to 0xff and initialized with 'factory defaults' if
08:53:12Hesa valid header & checksum is not found. Config version number is only
08:53:12Hesincreased when information is _relocated_ or space is _reused_ so that old
08:53:12***Alert Mode level 1
08:53:12Hesversions can read and modify configuration changed by new versions. New
08:53:12***Alert Mode level 2
08:53:12Hesversions should check for the value of '0xff' in each config memory
08:53:15Heslocation used, and reset the setting in question with a factory default if
08:53:17Hesneeded. Memory locations not used by a given version should not be
08:53:20Hesmodified unless the header & checksum test fails.
08:53:22HesSorry for the large paste 8-)
08:53:25LinusNand i also bumped it because the balance default was wrong
08:54:42adi|homeanyone have a shot of the 20 gig studio (or any studio) running rockbox?
08:54:43LinusNi could have let the revision stay at 0, but then all people would have the sound on the right ear after doanloading the latest build
08:55:05HesThat can be fixed by putting the balance into a new location in the RTC which is unused currently
08:55:17Hesand setting the current location to 0xff
08:55:33LinusNHes: and fragmenting the setting space? i think not.
08:55:48HesThat byte won't be unused for a long time...
08:56:28LinusNHes: why go through all that trouble just to keep the settings between unsupported daily builds?
08:57:04LinusNbumping the revision is far more easy
08:57:24HesBecause if we don't keep them between unsupported daily buildings, will we keep them between major releases? Or will we write a converter loader for every major release?
08:57:35HesBumping the revision is far to easy to do
08:58:08LinusNi don't think people will complain if they lose the settings when upgrading from 1.x to 1.y
08:58:48HesBut they'll be delighted if they don't 8-)
09:00:10LinusNi don't want to have half-dead code in the settings just to convert settings between upgrades
09:00:23LinusNlike the contrast "fix"
09:00:45LinusNthe code should be clean and easy to figure out
09:01:07*matsl is away: I'm busy
09:01:21HesIt is pretty simple as it is. Those things can be removed at the next large release (or so) when the version is bumped
09:01:34LinusNHes: but will it?
09:01:40HesWhy not.
09:01:55LinusNwill someone go through the settings code and remove unnecessary checks?
09:02:04LinusNi don't think so
09:02:05HesIt's not a very large piece of code. I could do that.
09:02:37LinusNthen i'll let you be in charge of the settings code, and let you add new settings to it
09:03:13***Alert Mode OFF
09:03:34HesUh. I wasn't asking for that 8-)
09:04:24LinusNmy point is that the code will be a lot simpler and easier to maintain if we just bump the revision whenever it becomes incompatible
09:04:48LinusNinstead of adding conversions, like in the contrast case
09:05:01HesYou're right about that. And my point is that it doesn't come too bad if we make a little effort at keeping the old settings.
09:05:03LinusNor fragment it, like you suggested
09:05:36LinusNi see your point, and i'll keep that in mind the next time i change a setting
09:06:13adi|homeid have to agree with LinusN
09:06:21adi|homeppl have the tendance to forget stuff
09:06:23LinusNthe problem was actually only because we have saved unimplemented settings (contrast and balance)
09:06:29adi|homemay as well keep a clean house as we go.
09:06:55Hessaving unimplemented settings was my bad, yeah
09:07:16LinusNwe will probably have the same issue when we implement the poweroff setting
09:07:26LinusNand the resume
09:09:40LinusNHes: speaking of poweroff, have you looked into that?
09:10:15HesNope... only been looking at the charging lately
09:10:34Hesbut I'm thinking of leaving it untouched for a while to gather usage experience on it
09:10:43Hesit seems to be working somewhat properly now.
09:11:08HesUwe has some interesting ideas for a next-generation charger but they seem a bit harder to implement properly.
09:13:03HesMental note: poweroff timer must be disabled while charger is plugged in
09:14:40*Bagder sneaked in a Rockbox reference on slashdot (again)
09:16:21HesBtw, personally i'd rather tune the poweroff/resume into a good setting from an invalid default instead of going through the other settings again... contrast is of course a special case, having a bad default is really confusing 8-)
09:16:56HesOkay, enough of that, time to drive to work.
09:17:08LinusNhappy driving
09:19:01Mode"#rockbox -oo adi|work Zagor|away " by ChanServ (ChanServ@services.)
09:26:22*Bagder ehums
09:26:50Bagderwhen trying to search on it says:
09:26:56Bagder"Sorry, our traffic is currently busy. Please try again."
09:45:29HesA hot sunny day, a 750cc engine between the legs, rockbox in the pocket of the leather jacket and a pair of EX70's under the helmet - what more can a man ask for commuting?
09:45:50BagderA/C ?
09:45:52*Bagder grins
09:46:00HesHm... a 1100cc engine would be nice.
09:46:18HesA/C: open the jacket
09:49:53HadakaI have an mpeg running in the lower right corner of my screen
09:50:02Hadakaand it takes 0% cpu power
09:50:16HesCPU's aren't what they used to be
09:56:33 Nick matsl is now known as matsl|away (~matsl@
10:03:55 Quit aakil (Read error: 110 (Connection timed out))
10:13:52***Saving seen data "./dancer.seen"
10:18:14 Join notch [0] (
10:18:43adi|homeokay.. opinion time
10:18:46Bagdermorning notch
10:19:42Bagderadi|home: I like the general idea, but you need to work on the finish
10:19:53adi|homethats what i thought
10:20:00Bagdernot that I know how ;-)
10:20:04adi|homei couldnt' think of the right way to phrase it :)
10:20:13notchthe old maling list is getting heated these days! :-)
10:20:21Bagderhehe, yeah
10:20:40notchI sense we are on the precipis of a flame war :-))
10:20:48BagderI feel *a bit* responsible for that "Hostility" thread
10:21:04adi|homeyou should
10:21:07adi|homebut hell
10:21:12adi|homethey deserve it
10:21:17adi|homei was _so_ tired of that thread
10:21:22adi|homebut but but but
10:21:32adi|home"why can't i have it MY way!!!"
10:21:35BagderI'll try to be friendlier in my next mail ;-)
10:23:09notchMakes for a more interesting read I think :-)
10:24:47LinusNi was about to answer to that thread, but i could really find the words
10:25:06LinusNit was too silly to respond to
10:25:21HesIt'll settle down after a while.
10:25:44LinusNjust saying 'are immediately saying "show me the code" or similar, whenever anyone posts a suggestion' is pretty childish
10:26:04notchOhh I'm soo tempted to call robert tweed a bad name :-)
10:26:14notchbut i'll behave...
10:26:23BagderI was tempted to reply, but I thought I better not ;-)
10:26:44LinusNi think the 200-file discussion went out-of-hand
10:26:53Bagderit certainly did
10:27:00Bagderit turned just plain stupid
10:27:04dwihnoInteresting thread...
10:27:04HesYeah, the other Robert was generally right, but the 200-file thread was crap indeed
10:27:50LinusNwe can't please everyone
10:27:59*dwihno is pleased
10:28:03dwihnoJust thought you should know :)
10:28:22Bagdernah, we need to take a stand too, we can't take everything
10:29:09LinusNwe have made a choice not to sacrifice RAM, and if people want to do things differently they'll have to do it themselves
10:29:50Bagderif people were around on IRC more, they'd have a better understanding of us poor misunderstood developer dudes ;-]
10:30:05LinusNwe are of course willing to discuss it, like we did on the 200-file limit
10:30:13notchHe would just have to move some files into another directoy anyway... big deal...
10:30:45LinusNstill, i think the 200-file limit could seem silly for the uninitiated
10:31:02Bagderindeed, the initial question wasn't out of place
10:31:39HesHey, does tuning stack sizes make sense? power thread is using like 11% and I know it won't allocate much more while it runs
10:32:19Bagderand the full size is 2K or what is it?
10:32:24LinusNHes: i was thinking the same thing
10:32:33LinusNwe can trim the stacks a little
10:32:35 Join notching [0] (
10:32:39Hescould reduce the stacks of many threads to save a little memory.
10:32:48HesJust a little, but it's memory anyway.
10:32:56notchingokay who kicked me?
10:33:19*Bagder senses hostility B-]
10:33:41Topic"rockbox - your favourite doorstop" by Bagder (
10:34:02LinusNnotch: nobody kicked you
10:34:20Bagdernotch has a split personality
10:34:36LinusNyeah, he's a psycho. let's kick him
10:34:46*Bagder chuckles
10:34:48notchingI know! what's going on! mIRC has a lot to answe for!
10:35:09LinusNmIRC? You are a psycho! :-)
10:35:14notchingyou can kick notch if you like but not notching :-)
10:35:24*Bagder runs X-chat on win32
10:35:30Kick(#rockbox notch :LinusN) by LinusN!
10:36:04*LinusN runs X-chat on linux, with an X11 server on his NT box
10:36:25dwihnoBecause each entry takes up about 260 bytes of RAM, and we think this memory
10:36:25dwihnois better used as an MP3 data buffer.
10:36:25dwihno> couldn't someone change this in the cvs?
10:36:25DBUGEnqueued KICK dwihno
10:36:25dwihnoYes, we can change it, but we don't want to.
10:36:33dwihnoMy favorite quote from you Linus
10:36:41dwihnoTalking about the 200 dir limitation etc.
10:36:45LinusNThat was hostile!
10:36:56LinusNI am a very mean person
10:37:06CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
10:37:06*Bagder slaps the "hostile person" sticker on LinusN
10:37:16Bagdernow shame on you!
10:37:21LinusNi'd better kick myself
10:37:26 Quit datazone (Read error: 110 (Connection timed out))
10:37:31Kick(#rockbox LinusN :LinusN) by LinusN!
10:37:36*notching is hiding and doesn't want to be kicked again
10:37:40 Join LinusN [0] (
10:37:55Mode"#rockbox +o LinusN " by Bagder (
10:38:07*LinusN repents
10:38:24Bagdernow will you behave?
10:38:36LinusNi mean YES!
10:38:38*Bagder stops playing and tries to do some work instead
10:38:40Kick(#rockbox LinusN :LinusN) by LinusN!
10:38:55 Join LinusN [0] (
10:39:06dwihnoLinusN: so what happens when there's more than 200 files in a dir? :)
10:39:18LinusNit says "200 file limit reached"
10:39:24dwihno"Do the 10 meg RAM hw mod" ? ;)
10:39:25Bagderit does now
10:39:26LinusNand shows only the firsty 200 files
10:40:02Bagder"error: we recommand a brain surgery since you seem to be crazy judging from the amount of files in a single directory"
10:40:26LinusN"ERROR 21: Silly User"
10:41:25Bagderwith the mod, we will mostly likely be able to read more than 200 files though, using the same amount of mem
10:43:25Hesooh, you have a planned something?
10:44:37BagderHes: yes, changing the dir cache system to better use all the available memory
10:44:52Bagderall the available dir-cache mem that is
10:45:41dwihnoHas anyone got a clue how much memory the archos firmware uses for the mp3 buffering?
10:46:10HesAh, ok, nice.
10:46:57Bagderdwihno: it would be possible to dig up from dissassembling, but I don't know if anyone has bothered
10:47:38LinusNit can probably be determined by measuring hard disk activity
10:47:43Hesor roughly by timing the disk read intervals for a known-bitrate stream
10:47:49Bagderyes, that too
10:47:55Bagderthat's probably easier ;-)
10:48:46dwihnoI have a feature request... I'd love to play the games while in USB mode ;D
10:49:11Bagderit would be possible
10:49:12LinusNitäs easly done
10:49:19LinusNit's easily done
10:49:30Bagderit isn't such a bad idea really
10:49:31dwihnoYeah, but it's not really that useful :)
10:49:46Hesnot bad, not very important, but cool 8-) I have another one...
10:50:02Bagderbtw, is the charging running now while in USB mode?
10:50:16BagderI think it should
10:50:29Hesthe fast forward's jump length (amount of stream time jumped at a time) could go up while going forward for a longer time
10:50:35Hesup to a certain jumping speed
10:50:37dwihnoNobody has bothered to add my USB connected icon thingy ;)
10:50:41BagderHes: I agree with that one
10:50:47LinusNdwihno: i forgot
10:50:54Hesit takes _long_ to fast forward some 45 minutes (i have loads of 1-hour mp3's)
10:51:13BagderHes: I would prefer that instead of having the step-time as a config option
10:51:22Hesi run an mp3 archive of my favourite radio station which broadcasts really good stuff at times when i'm not at home, or i'm sleeping 8-)
10:51:58Hesand the speed needed for going 45 mins forward is really too fast for the average user going forward a minute or so
10:52:58dwihnoBagder: the initial stepping time should be configurable, and the "acceleration" in the ffw/rew should be logarithmic, thus making it go ffffffw after 5-6 secs of 'regular ffw' (my opinion, as always)
10:53:46BagderI think the config part can go away if it is made correctly
10:53:57Bagderat least I think it's worth an attempt
10:54:32dwihnoLinusN: show me the code! :D *act hostile* :-)
10:55:22BagderI think the id3 reader bugs
10:56:09Bagderor perhaps the 300-byte thing hit me
10:56:14 Nick LinusN is now known as Linus|lunch (
11:05:09Hesexponential growth for step time up to a hard limit
11:05:51adi|homei really need a better way to finish that t-shirt.. i just don't know how
11:05:53*adi|home pouts
11:05:54Hesstarting from a configured low limit
11:08:03PsycoXulrock-hard firmware
11:11:12BagderI miss a 'tags' target
11:11:18Bagderin the makefile
11:11:30*Bagder will add one later
11:12:24Bagderemacs with etags rocks
11:15:42dwihnoetype rocks
11:18:37 Nick Low[a] is now known as Lowfiler (
11:18:42Lowfilergood morning :)
11:25:28dwihnoI have a 110% serious question
11:25:35dwihnoDo you guys know what målgången i årsta is?
11:25:54Bagderan address?
11:34:49Bagderetags target added
11:35:38Bagderuse "../tools/configure update"
11:35:52adi|homedwihno: no.. what is it?
11:37:54 Join edx [0] (
11:38:07Bagderhey Felix
11:38:13edxgood morning :)
11:38:35adi|homeopinions wanted
11:38:50adi|homesexy naked ladies under 24 and older then 18 accepted
11:39:16Linus|lunchadi|home: haha! But you have misspelled "boundaries"
11:39:27adi|homeheheh k
11:39:37edxthat one is cool
11:39:47Linus|lunchmaybe the pictures and text could be larger
11:40:10edxmake that rockbox logo bigger :D
11:40:29*edx wants a list of names on the back :)
11:40:38edxbut we wont have two-sided shirts, will we?
11:41:45Bagderwe made the contest one-side since we want to have a Haxx logo on the other side
11:44:24edxahh :) - the back that Robert Hak designed - its has names and haxx logo :D
11:44:51edxplus I like it very much
11:47:29*adi|home likes all the submissions so far
11:48:13adi|homeokay.. typo fixed.. reload
11:48:19adi|homeim getting a second version no
11:57:22MeRWiN|Sleepgrrr... overslept
11:57:33 Nick MeRWiN|Sleep is now known as MeRWiN (~merwin@
11:59:01adi|homeokay.. version 2 up
11:59:09adi|homeopions please
11:59:42MeRWiNhah... nice adi :)
11:59:56adi|homemerwin check out the first one ..
12:00:04adi|hometell me which you like better
12:00:06adi|homeand if you have suggestions
12:00:29MeRWiN2nd is better
12:02:26 Join Jonmann [0] (jonmann@
12:03:05Topic"RockBox: Hostile answers upon demand" by adi|home (~adiamas@
12:03:07JonmannDoes the Jukebox take running? I mean, it's HDD-based and all.
12:03:21adi|homeummm.. i don't understand the question
12:03:31Linus|lunchadi|home: place the rockbox URL somewhere
12:03:36edxadi: put the desktop pc to the right - makes it look more dynamic ;)
12:03:39adi|homeon the shirt?
12:03:46Jonmannadi, the hard drives I know don't like shaking.
12:03:50Linus|lunchJonmann: don't run with the Jukebox
12:04:05JonmannLinus, okay. Then it's not for me. Thanks.
12:05:34 Part Jonmann ("gone")
12:09:06adi|homeLinus thats not ture
12:09:16adi|homeyou can do it.. just depends on _how_ long your going to
12:11:37adi|homeokay.. version 3 up
12:13:04MeRWiNi think it's priceless2
12:13:21MeRWiNthey're the same pic :)
12:13:48adi|homei know.. i had to fix it.. hold on :) - Everywhere you want to be
12:13:57***Saving seen data "./dancer.seen"
12:14:12adi|homereload :)
12:14:23adi|homethat sound better?
12:14:27adi|homeEverywhere you want to be?
12:14:56MeRWiNthe mastercard slogan :)
12:15:08adi|homeheheh bastard ;)
12:17:14adi|homeokay.. reload :)
12:17:39MeRWiNI don't know which I like better though
12:18:10MeRWiNI can't believe I overslept for my flight home :)
12:18:18adi|homeYOU WHAT?!
12:18:26adi|homedude i was going to offer to wake you up too ;)
12:18:37adi|homehow much did you miss it by?
12:18:49MeRWiNwell, I woke up an hour before it was supposed to leave. There'
12:18:57adi|homeheheh sorry
12:19:01MeRWiNThere's another flight leaving 3 hours from now, so i'm going to catch that one
12:19:11adi|homesee if you can get a hold of the airline .. switch tickets
12:19:23MeRWiNI technically didn't miss it since I changed it before I actually missed it
12:20:02MeRWiNIt's a business trip, so we get use of the travel agency. Already switched. First class all the way (gotta love last minute tickets on Northwest Airlines... half the time it's automatic upgrades)
12:20:31adi|homewhere are you now?
12:20:46MeRWiNHarrisburg, PA... going home to Seattle.
12:20:51adi|homeno shit?
12:21:08adi|homewas just talking to my ambulance crew about that place yesterday :)
12:21:15adi|homethere is _nothing_ there :)
12:21:21adi|homewhat were you doing?
12:21:22MeRWiNnope, not really anything there
12:21:33MeRWiNTesting high speed internet (GPRS).
12:21:49adi|homeyour company hireing at all ;)
12:21:52MeRWiNheard of T-Mobile (you guys have it in LA now)
12:21:54*adi|home needs another job
12:22:07adi|homeand who is 'you guys'?
12:22:20MeRWiNadi: aren't you in LA?
12:22:24adi|homehell no
12:22:27adi|homeNY state
12:22:36MeRWiNoh. heh... well, then VoiceStream
12:22:42adi|homeim about 20 mins west of white plains ny
12:22:47adi|homeahhh.. ok
12:23:20MeRWiNVoiceStream/T-Mobile is always hiring..., click on Jobs at the bottom of the screen
12:23:21MeRWiNtons of jobs
12:25:45MeRWiNI'm User Acceptance Testing (UAT). Get to travel around the states making phone calls. NOT like the "can you hear me now" guy though. heh. I've actually got a brain.
12:26:30MeRWiNSo, I just found out that harrisburg is the capitol of PA. I never knew that
12:27:14HesMeRWiN: I've been working with the T-Mobile/one2one guys in UK...
12:27:14adi|hometo far for me to get to or i would have met you for a drink
12:27:55Hestheir SMSC interface (XML over https) sucks badly 8-)
12:27:55adi|homeMeRWiN: how often are you out to the east coast?
12:28:38MeRWiNoccasionally. I was going to go to Wayne, NJ next week, but it looks like i'm not
12:28:43MeRWiNor at least, probably not
12:30:00adi|homeno shit?
12:30:05adi|homeif you do, give me a heads up
12:30:10adi|homethats only like 30 mins away
12:30:17adi|homewe can do dinner or drinks one night or something
12:30:23adi|homegive ya somthing to do ;)
12:30:31MeRWiNadi|home: nice :) sounds good
12:30:43MeRWiNHes: XML over https? haven't heard of it.
12:30:52MeRWiNhes: what is it?
12:31:25HesWe transport SMS'es to/from their customers, encoded in XML files which are transported over https...
12:31:47MeRWiNhes: ahhh... That's an interesting way to do it :P
12:35:13 Join Blaster_Master [0] (
12:38:35 Nick matsl|away is now known as matsl (~matsl@
12:38:39*matsl is back (gone 03:37:31)
12:40:55Topic"RockBox: Don't forget! The great T-Shirt contest is underway!" by adi|home (~adiamas@
12:41:50 Join Zagor [242] (
12:42:26Zagorhey all. logging in from the office.
12:43:27 Nick notching is now known as notch|lunch (
12:46:41MeRWiNhey zagor. were you the one doing hardeep's patch for ff/rew?
12:47:06MeRWiNZagor: read my latest post... his new patch to fix the EOF problem doesn't quite work right
12:47:23MeRWiNjust a heads up
12:47:50Zagoryup. going through the mail now.
12:47:55Zagordescribe the problem
12:48:28Zagorit simply does not always work?
12:48:42MeRWiNIf you FF from like 3:15 to an EOF that is 3:45, the track change will work properly (since it's within the buffer still). If you FF from like 1:00 to 3:45 then it won't work
12:50:21 Nick MeRWiN is now known as MeRWiN|Shower (~merwin@
12:51:42 Nick Linus|lunch is now known as LinusN (
12:54:02 Join mistax [0] (
12:54:20adi|homei did a quick check
12:54:27adi|homeno less then 33 messages about the 200 file limit
12:54:45adi|homeZagor do you manage the tshirt submissions?
12:54:58Zagorno that's bagder's job
12:55:04Zagorhe's the tshirt guy :)
12:55:14adi|homehehehe k
12:55:21adi|homecause i followed the rules.. sent 3 emails
12:55:30adi|homebut they are all variations on the same idea
12:55:34adi|homeso should be listed togeather :)
12:55:45Zagoryeah, I saw that. i'm on the list too.
12:56:28adi|homei don't know which one i like better.. so i figured hell.. let the users decide
12:57:36Zagorman this screen sucks :(
12:58:14LinusNAt last I found the dreaded "wring contents in directory" bug!!!!
12:58:20ZagorI'm having a crisp 20" TFT at home. CRTs should be outlawed!
12:58:32ZagorLinusN: nice! what was it?
12:58:59LinusNthe path component match used strncasecmp()
12:59:33LinusNthat meant that "Ace Of Base - Happy Nation" matched if it looked for "Ace Of Base"
12:59:38 Nick matsl is now known as matsl|away (~matsl@
12:59:43*matsl|away is away: I'm busy
13:00:24 Nick LinusN is now known as LinusN|meeting (
13:00:27adi|home"Barenaked" on mtv.. jennifer love-hewitt climbs out of bed.. looks in the mirror.. her hair is a mess but her makeup is perfect
13:00:35adi|homei can just hear my gf saying 'cha−−- right'
13:01:52 Join Moufle [0] (~moof@
13:03:17adi|homeZagor.. someone had a suggestion i think i like
13:03:24Zagorwhat was that?
13:03:25MoufleI can't manage to boot my freshly compiled ajbrec.ajz. This happen since I enabled Loadable fonts -> I get a white screen after logo, and no HD activity, and the blind nav/play is impoossible.
13:03:28adi|homeany chance we could get the FAQ sent to someone when they join the mailing list?
13:03:49MoufleAre there some settings to do for fonts?
13:03:49adi|homeor better yet
13:03:52ZagorMoufle: did you copy the system.ajf file to your archos?
13:03:52adi|homea remind to read it
13:04:08Zagoradi|home: i responded to your mail
13:04:12MoufleZagor: THis must be missing, is it in the doc?
13:04:26ZagorMoufle: loadable fonts is work in progress. no docs for that :-)
13:05:02Zagoror: we suffer a shortage of doc writers. do you want to contribute? ;-)
13:05:17MoufleZagor: huhu ok, so this was not a FAQ ;) I am a power user : i compile but not hack the code
13:05:57MoufleZagor: I will soon have exams, but it could be interesting, BTW since I'm frech i could create a little howto
13:06:05Zagorok. yes, the system.ajf is the loadable font, so it is required. without it, we can't even display the error message that the font wasn't found...
13:06:14Moufleof course
13:06:24MoufleI'm gonna copy it
13:06:58Zagoryou can experiment with other fonts too, but we only support loading monospaced fonts yet (no propotional)
13:07:04Bagderwe should have a small internal one
13:07:22ZagorBagder: yes. we actually link them, but don't have code to use it
13:08:19Bagderbtw, I'm all for a '.rockbox' directory on target for stuff like that
13:08:22MoufleAnd i guess I must create the file but converting the font I choose w/ an utility ?
13:09:09adi|homeZagor is the system.ajf in the cvs?
13:09:38Bagdersystem.ajf is not in cvs, it is generated from a .bdf font
13:09:47*adi|home nods
13:09:52Moufleby /tools/bdf2ajf
13:09:58adi|homeheheh k...
13:10:28adi|homeMoufle: if you want to write up a doc, ill toss it in the CVS :)
13:10:49MoufleYes, I'm gonna write a quick HOWTO
13:13:01adi|homeokay.. VERY BASIC
13:13:07adi|homedesc of loadable fonts in the FAQ
13:14:21adi|homeokay.. do we have any valid .bdf font files around?
13:16:49Zagorin firmware/fonts
13:16:55adi|homeyeah.. got it :)
13:17:38*Zagor is switching to another computer. brb
13:17:39 Quit Zagor ("Client Exiting")
13:19:59MeRWiN|ShowerOkey all... ttyl. Gotta catch a flight (2nd attempt, hopefully better than the first)
13:20:15MoufleWhat's the best font to use for loadable fonts
13:20:31adi|homedunno.. just playing with it now :)
13:20:47 Quit MeRWiN|Shower ()
13:20:47BagderAlex provided two in his patch package
13:22:10 Join Zagor [242] (
13:23:49MoufleMy little howto is written
13:23:57MoufleWhere do I put it?
13:24:01Zagormail it to me
13:24:05adi|homeumm.. mail it up :)
13:26:02MoufleZagor: it's posted, tell me when you get it
13:26:19Zagorgot it
13:27:37Mouflecorrect export PATH=***SLASH****home
13:27:53Zagoractually, it's not quite that complex. the makefile already converts a font for you, so you always get a system.ajf. the only special thing is that you have to copy a second file to the archos
13:28:14adi|homewhat file does the makefile convert for you?
13:28:17MoufleIt didn't generated it for me ;)
13:28:28Zagorare you sure? did you get an error message?
13:28:54 Quit Blaster_Master ()
13:29:07MoufleIt should come in the build dir?
13:29:36Zagorjust tested, works here
13:29:50adi|homei didn't get one either.
13:30:34Mouflethe Makefile the configure script writes contains the instructions?
13:30:47adi|homenope... no system.ajf file
13:31:25adi|homei do ../tools/configure in the build
13:31:27adi|homeand then make
13:31:34Moufleme too
13:31:49adi|homeand get no system.ajf
13:31:53Zagoryou're right, it doesn't work.
13:32:06Zagorhow come it worked for me before? /me is confused
13:32:08*adi|home sticks his tongue out at Zagor
13:32:11adi|homeha ha.. told you so.
13:32:21MoufleZagor: maybe you use the main makefile?
13:34:49adi|homehmmm.. i cant get any of the new fonts to work
13:34:55adi|homedo we know if they work at all on the recorder?
13:34:59Moufleneitjer do I
13:35:08Zagoruse alt6x10
13:35:13Moufle(I got an AJBR20)
13:35:15adi|homei just tried that on
13:35:25adi|homeit has to be translated to system.ajf right?
13:36:33adi|homehmmm then that one don't work for me
13:36:48Zagorfixed Makefile now
13:37:22adi|home../tools/bdf2ajf -f ../firmware/fonts/alt6x19.bdf -o system.ajf
13:37:31Zagoruse the new makefile instead
13:37:41Bagder../tools/configure update
13:37:52Zagorno, it's a new in firmware/
13:38:01*Bagder hides
13:38:38MoufleIt works at compilation
13:39:32Mouflebut not on my target :/
13:39:44Moufleno logo screen, no display
13:39:47adi|homeim testing too
13:40:16ZagorMoufle: how long since you tested? maybe your contrast setting is low
13:40:37Zagortry pulling a battery for a minute to reset the settings
13:40:47Zagorthe font patch doesn't affect the logo display
13:40:55adi|homehmm.. thats a prob too..
13:40:59MoufleIt has always been ok, and blind nav (pressing 3x play to get a song) doensn't works
13:41:05adi|homehow do you get _to_ the contrast setting.. if you can't see it :)
13:41:18Zagoradi|home: simplest is to reset the settings
13:41:32Zagorwe have to work something out about that
13:41:46Bagderperhaps pressing down a single key for N seconds
13:41:47Bagderin the main menu
13:41:58MoufleZagor: did you had it working on atarget?
13:42:03ZagorMoufle: yes
13:42:12adi|homewell.. we could just do on boot... have a function key act for contrast
13:42:16Zagorif your logo doesn't show, it's the contrast problem. not the font patch.
13:42:36adi|homezagor on your target..
13:42:44adi|homegeneral settings
13:42:54adi|homewhat is the location of contrast?
13:42:57adi|homehow many times down?
13:43:15Mouflegonna put out a battery
13:43:50Zagormaybe some obscure key chord, like ON+MENU+UP/DOWN
13:44:18adi|homezagor.. to get to contrast from boot
13:44:20BagderZagor: hm, f1+f2+f3 could work
13:44:32adi|homemenu => down 4 => right => then up or down.. right?
13:44:43ZagorBagder: yeah, maybe. what for recorder?
13:45:11ZagorON+MENU+PLAY or something
13:45:35adi|homewe would just do...
13:45:45Zagorsets it to default contrast
13:45:45adi|homeand ON+LEFT
13:45:58adi|homebe a fast way of setting contrast
13:46:06Moufle: now it says : Read acces error, Can't read from HD
13:46:12Zagoradi|home: those are too valuable key chords. we'll want to use them for something more important
13:46:19BagderZagor: it should probably be an all-settings reset
13:46:26ZagorMoufle: that sounds bad
13:46:29ZagorBagder: yes
13:47:17MoufleZagor: and it looks like it's the Orig firmware that says this
13:47:55ZagorMoufle: how are your batteries?
13:48:13Moufleok i think
13:48:32Zagorharddisk errors are often a sign of low batteries
13:48:47Mouflei will charge it for some hours
13:49:12Moufleanyway thx for the fixes, I hope you'll be able to modify/post my little HOWTO
13:48:55ZagorI think maybe a printed line in ../tools/configure is enough. "Don't forget to copy system.ajf to your archos!"
13:49:10Zagorwhen you select loadable fonts
13:49:17Zagornow that I fixed the makefile
13:51:30Zagorthanks anyway
13:52:35 Part LinusN|meeting
13:53:36 Join LinusN [0] (
13:54:39 Part LinusN
13:56:09adi|homeits definatly a bug.. not contrast
13:56:17adi|homei wrote down the steps to get to the contrats
13:56:20adi|homei load the fonts
13:56:23adi|homei follow the steps
13:56:23Lowfilerheya :)
13:56:26adi|homechange the contrast
13:56:30Lowfilertried the daily build from today
13:56:33adi|homereload w/o contrast
13:56:39Lowfilernice fastfwd :)
13:56:42adi|homeand the contrast setting hasn't changed
13:56:50Zagoradi|home: tried the bleeding build?
13:56:57ZagorLowfiler: yeah, isn't it?
13:56:58adi|homei updated my cvs
13:57:02adi|homeand compiled locally
13:57:18adi|homeunless you made changes in the last 10 mins
13:57:26 Join LinusN [0] (
13:57:53Lowfilerits great :)
13:57:57Zagoronly in firmware/Makefile
13:58:16adi|homeyup.. i updated that
13:58:40Zagori'll do a fresh checkout and try here. hold on.
13:58:53adi|homebesides.. my contrast is at 47
13:58:57LinusNok guys, how long is the average .mp3 file name?
13:59:02adi|homei should see something...
13:59:15adi|homefigure "band - album - name"
13:59:23ZagorLinusN: hehe. 25-40 chars, I'd hazard a quick guess
13:59:24BagderLinusN: you mean without path?
13:59:24adi|homefigure 45 chars per field?
13:59:31adi|home130 chars maybe?
13:59:33Zagoradi average?
13:59:41BagderLinusN: I'll give you my average in a sec
13:59:43adi|homehmm.. well
13:59:48adi|homeim not average ;)
13:59:56adi|homei do dirs by band name
14:00:01 Part LinusN
14:00:03adi|homeand only the song name for mp3
14:03:50Bagderfind . -name "*mp3" | awk 'FS="/"; { print $NF}' | egrep -v '^\./' | wc | awk '{ print $3/$1 }';
14:04:18 Join LinusN [0] (
14:04:26BagderLinusN: 34.1117
14:04:31LinusNf*cking tunnel
14:04:40Bagderadi|home: I win ;-)
14:04:41LinusNok, so 30-40 bytes then
14:04:49adi|homebut you have to figure...
14:04:54LinusNi missed you discussion
14:04:59adi|homemine is generally artist - song.mp3
14:05:09Bagderadi|home: mine are too
14:05:21adi|homelotsa ppl do artist - album - song
14:05:33adi|homeso id figure may 45 to be safe
14:05:39Moufle"you know that you have done too much computer" ... When you get to know the mean size of your Mp3 collection
14:05:42adi|homeie: ours + (our/2)
14:05:47LinusNmine are often artist - album - tracknum - name.mp3
14:05:59BagderLinusN: find . -name "*mp3" | awk 'FS="/"; { print $NF}' | egrep -v '^\./' | wc | awk '{ print $3/$1 }';
14:06:06adi|homei do that for broadway showtumes
14:06:30Topic"Who needs friends. We have => find . -name "*mp3" | awk 'FS="/"; { print $NF}' | egrep -v '^\./' | wc | awk '{ print $3/$1 }';" by adi|home (~adiamas@
14:07:19Bagderthe egrep in there is a bit lame, it should be possible to do it without that
14:07:33Bagderbut I leave that for improvements in the next release ;-)
14:08:45LinusNif we assume that the average filename length is 45, we can have 500 files in an 22Kb buffer
14:09:28Zagori say 16kb is enough, then
14:09:34Zagorthat's 354 files
14:09:37*adi|home nods
14:10:01Bagder"364 files should be enough for everybody" ;-)
14:10:07LinusNhow many files did the 200-file-limit guy have?
14:10:09adi|homethank you gill bates
14:10:16 Nick Zagor is now known as BjornGates (
14:10:23adi|homei don't think he said
14:10:24HesHa, 655 files for me
14:10:33HesHe probably has half a thousand or so
14:10:41Hesmore than the limit, anyway 8-)
14:10:44adi|homeprob doesn't have an key to rep the internet on his keyboard
14:11:09 Nick BjornGates is now known as Zagor (
14:11:25adi|homeokay.. i need a nap.. be back later
14:11:29Bagdernight adi
14:11:55Bagderwe use 200*260 today, don't we?
14:12:08Zagornice saving
14:12:12LinusNok, i set the average file size to 40 bytes and the dir cache size to 400 files
14:12:16Bagdernice shaving ;-)
14:12:35ZagorI noticed the playlist loader also uses 50k. cut that down to 16kb too, I think, now that the disk code is much faster
14:13:17LinusNit should be the same size as the dir cache
14:13:33Bagderyes why?
14:13:34Zagorit's the text file loader buffer
14:13:37LinusNbecause dir play must work
14:13:59***Saving seen data "./dancer.seen"
14:14:36LinusNi should put those #defines globally somewhere
14:14:46Zagornow I understand why it was so big. I thought I remembered it being much smaller before
14:14:58LinusNit was 500 bytes or so
14:15:15ZagorLinusN: apps/limits.h
14:15:21Zagorwhy not expose them in the open
14:15:38LinusNdo we have any more limits?
14:15:45LinusN10000 files playlist?
14:15:54Zagorthe playlist size is the only one I can think or right now
14:15:59Bagderwe probably have some more too
14:16:36Zagorlimits.h is a bad name, though.
14:16:53Zagorlimits.h defines type sizes
14:17:18Zagormight aswell avoid the collision
14:20:43BagderAdiamas' latest ones added
14:21:48Zagorthose mastercard spoofs are popular these days
14:22:25Zagorthe "priceless" gag is from a mastercard commercial in the us
14:22:41*Bagder learned something
14:31:07LinusNWe just saved 50Kb of memory!
14:31:07LinusNand increased the max-file limit to 400
14:31:19Bagderreally cool
14:31:24Bagderdoes it work too? B)
14:31:28Zagorthe power of clever code
14:31:29LinusNof course not
14:31:43LinusNthe power of short file names
14:32:28Hadakathose "priceless" gags have been around for decades
14:35:38ZagorBagder: what must be in CVS to work? I'm trying to make a fake CVS dir so I can run 'cvs update' from the project root
14:35:51Zagorit works at home, but I fail here
14:36:11BagderZagor: don't fake it, just checkout "." in a suitable directory
14:37:06Zagor..and if I don't want the www module? :-)
14:37:16LinusNthen delete it again
14:37:25Zagorno. next update puts it pack
14:37:30LinusNafter waiting for the huge devcon images
14:37:30Bagderremove it from CVS/Entries perhaps
14:38:36BagderI always get . and just ignore that I get the www too
14:38:55Zagorthe simple solution to my problem was to set CVSROOT in an env variable
14:38:56Hesi do checkout/update apps firmware uisimulator tools
14:42:46*Zagor is checking in ata_delayed_write()
14:45:14LinusNZagor: is there a way to spin up the disk without accessing the disk?
14:45:26LinusNlike a SPIN_UP command?
14:45:52LinusNjust curious
14:45:56Bagdertime to get an ice cream
14:49:05LinusNZagor: does it still say "Saving Settings"?
14:49:21Zagoryes, but so fast nobody sees it :-)
14:49:28Zagorso that can be removed
14:49:36LinusNit must be removed
14:49:40LinusNit is a lie
14:50:08LinusNZagor: is there a way to tell if the settings aren't saved yet?
14:50:18Zagornever was
14:50:27LinusNof course it never was
14:50:31LinusNit is a new feature
14:50:51Zagortry it before you diss it, ok?
14:51:05LinusNwhy would i diss it?
14:51:13Zagorsounded like it last time :-)
14:51:23elinenbe|sleepI have a serious bug.
14:51:28 Nick elinenbe|sleep is now known as elinenbe (
14:51:37elinenbeI just grabbed the latest CVS
14:51:50LinusNno, i meant if there is a way to tell, and we could display it somewhere, like the 'info' screen
14:52:18elinenbethen I loaded onto an ajbr6k the fimware and the font
14:52:30elinenbe(this bug only occurs when I use the loadabel fonts)
14:52:33LinusNand *BOOOM*
14:52:39ZagorLinusN: it's a local bool in ata.c: delayed_write. true if a write is pending
14:52:53elinenbeI get a few pixels on the left side of the screen.
14:52:55Zagorelinenbe: copied the font?
14:53:09LinusNyes he did
14:53:10elinenbeyeah −− the default one that is made with the new makefile
14:54:53elinenbea pattern repeates itself down the lefthand side (i take picture and upload)
14:56:48elinenbeis disk writing working properly now?
14:57:09Zagoryes, but the file system support isn't ready. only sector writing.
14:57:17elinenbethat is a good start
14:57:51elinenbeno more RTC stuff?
14:58:01LinusNwhat makefile creates the system.ajf font?
15:00:59LinusNmy recorder is completely blank with the default loadable font
15:00:59 Quit mistax (Read error: 110 (Connection timed out))
15:01:23elinenbeLinusN: same thing here.
15:01:31elinenbejust check out that picture
15:01:51Zagorvery strange
15:02:08LinusNelinenbe: you are lucky!
15:02:13LinusNyou get pixels on the screen
15:02:16ZagorI can't test it myself now, the usb port on this computer has gone completely bonkers :-(
15:02:17elinenbeare settings now saved to disk (to the sector)?
15:02:25Zagorelinenbe: only on players
15:02:29LinusNelinenbe: only on player
15:03:00elinenbewill that be coming for the recorder too?
15:03:03Moufleit looks like display was reversed by 90°
15:03:26Moufleyou can see n° artist song in the wrong way
15:03:26LinusNelinenbe: someday, yes
15:03:31Zagorelinenbe: yes. but currently recorder only uses rtc
15:05:25elinenbeZagor & LinusN: it would be nice for the recorder to use the disk too. That way my settings would be presistent. Currently every time I replace my batteries I lose everything :(
15:05:37Zagoryes, it will
15:05:47LinusNelinenbe: you are a minority
15:05:56LinusNmost people keep their batteries in the unit
15:06:05 Quit Moufle ("Client Exiting")
15:06:18Zagorstill, with the delayed writing there is no real reason not to write to disk on recorder too
15:06:30Zagoralso, this enables true resume Real Soon Now
15:06:44Bagderwe should add it on the recorder too, indeed
15:06:47LinusNisn't the recorder able to be partitionless?
15:06:48elinenbesome of us have a 1 hour charger and use the jukebox 20 hours a day :)
15:06:57PsycoXuloh btw are directories supposed to repeat? i don't see any repeat option...
15:06:58ZagorLinusN: both player and recorder are
15:07:07ZagorPsycoXul: all playlists repeat
15:07:13Zagordirs are playlists too
15:07:14LinusNPsycoXul: the repeat option is always ON
15:07:18PsycoXuli see
15:07:38Zagorone of the we-really-should-fix items :-)
15:07:42PsycoXulcan we get that in the settings?
15:07:57PsycoXuland how about parse/filename fallback for no-id3 tracks
15:08:00LinusNnow we come back to the repeat problem...
15:08:26LinusNyes, what is repeat?
15:08:37LinusNfor playlists it is simple
15:08:46LinusNfor dir play it isn't that simple
15:09:07Zagorisn't repeat a fairly entrenched concept? cd players have had it since 1985 or something
15:09:11LinusNmaybe two settings, loop and repeat
15:09:22Zagorrepeat one and repeat all
15:09:43Zagornobody will understand the difference btwn loop and repeat (neither do I)
15:09:57LinusNok, so without repeat, the dir should play to the end of the dir?
15:10:11LinusNand with repeat it should repeat and loop
15:10:29Zagorand repeat-one repeats the same song
15:10:42LinusNthe playlist code can't handle non-repeating lists today
15:11:34Zagori know. fixing...
15:11:52LinusNthere is a pile of interesting problems with that
15:12:03Zagoris there? explain
15:12:31LinusNit has to do with the negative index parameter being less than -1
15:12:50LinusNand that the user toggles the mode while playing
15:12:59ZagorI don't see the problem
15:13:58LinusNthe code sets the index to 0 when the negative index goes past the start of the list
15:14:33Zagorwe already fixed that. and without repeat, that is not allowed. so I still don't see a problem
15:14:45*Zagor is thick-headed
15:15:04LinusNif the user disables loop when wrap is already made you can't press PREV
15:15:44Zagorah, right. that old thing :-)
15:15:58LinusNFarbror Frej haunts us again
15:16:25LinusNjust try to imagine all special cases, and test with farbror Frej
15:17:03LinusNby the way, the setting name loop_playlist isn't very good
15:17:03Zagorbuh, I can't test without working usb. I'll hold this work 'til tonight
15:17:14ZagorI know. it should be repeat_all
15:17:26 Join Moufle [0] (~moof@
15:17:35ZagorMoufle: working better?
15:17:37LinusNZagor: do we really need repeat_one?
15:17:53Zagoryes, I think so. for completeness, if nothing else.
15:17:59MoufleZagor: not at all, may I send you an error log via a /msg?
15:18:08ZagorMoufle: sure
15:18:45adi|homefrabror Frej?
15:18:47adi|homewtf is that?
15:18:59Lowfilerheya uhm
15:19:04Zagor20 second comedy tracks
15:19:13LinusNin 32kbit/s
15:19:16Zagorexcellent for testing many concurrent files in the buffer
15:19:30Lowfileris it possible that the battery-info is not really correct in v1.2
15:19:56LinusNeverything is possible
15:20:25ZagorLowfiler: 1.2 battery code is very early. it's improved a bit since then and is growing better day by day.
15:20:57Lowfilerok thats good. btw: its a nice idea to show it in %
15:21:55Zagoryeah. we like everything in % :-)
15:22:09Lowfiler i c :)
15:22:24Zagorjust remember, it's % voltage. not % battery time
15:23:05Lowfileruh ok
15:23:53Lowfilerdoes somebody know if its possible (ya i know - everything is possible :P) to play ogg-vorbis with the jbr
15:24:01Lowfilerin the future
15:24:25Zagornot unless a miracle happens
15:24:27Lowfilerthen u have to flash the decoder
15:24:42LinusNsomething like that
15:24:57LinusNthe RAM is probably to small
15:25:04LinusNfor both code and data
15:25:12Lowfilerhmm yes thats right
15:25:15LinusNin the decoder chip
15:25:23Hadakaoh, damn, I forgot the answer alread - does, or does not, gcc-3.2 have sh1 support?
15:25:33ZagorHadaka: dunno
15:25:37LinusNi'm not sure
15:25:46BagderI don't think anyone found out yet
15:26:02LinusNHadaka: go find out!
15:26:10Hadakahaha :)
15:26:59Lowfilerhmm the bug in archos-os 1.27d is nice - i use 5GB hd-space but it shows me that >18gig are free :)
15:27:33ZagorLowfiler: haha
15:27:46LinusNi use 5Gb and i have 35Gb free
15:27:49*Zagor can laugh, since we don't even try to calculate free space
15:28:10PsycoXulhey yeah when's rockbox gonna show disk space usage info? heh
15:28:15ZagorLinusN: really? I have about the opposite..
15:28:24LinusNPsycoXul: as soon as you have sent us the patch :-)
15:28:28ZagorPsycoXul: when you submit that patch ;)
15:28:41LinusNZagor: i exaggerated a little...
15:28:44*Bagder says "show us the code" and runs to hide
15:28:47Lowfiler[15:29] <LinusN> i use 5Gb and i have 35Gb free <<?? 40gb
15:28:55Zagor40 gig rocks
15:29:16Lowfilernice :)
15:29:21LinusNBagder: :-)
15:29:35Lowfilersome1 here who has jbmm
15:30:03Lowfileri read that it is a really buggy thing
15:30:20Zagortoo bad the 60gig draws so much power, or I'd buy one just for the bragging rights
15:30:39ZagorLowfiler: I wouldn't be surprised.
15:30:41Hadakawell, sh1 has not been obsoleted - so that that would mean that it is still in-tree atleast
15:30:55LinusNHadaka: good
15:31:07LinusNi think they lack a maintainer for the sh1
15:31:27Hadakayeah, it might be that it doesn't work, but atleast it isn't removed
15:31:51HadakaI wonder when toolchain is going to be updated to 3.2
15:32:12Lowfilerbtw: can somebody of you build a option to leafe the backlight allways on?
15:32:57Lowfilerwould be cool
15:33:20ZagorLowfiler: just set it to 0
15:33:31LinusNno, then it will be always off
15:33:39Zagorreally? try it! :-)
15:33:43*Zagor just did
15:33:48Lowfilerno its always on now
15:35:13LinusNok, i remembered it wrong
15:35:33LowfilerThat will use a lot of power but i dont care
15:35:38LinusNwhat did we say in the last backlight discussion?
15:35:55LinusNoff, on, 1,2,3,4,5,6,7,8,9,10?
15:36:08Hesoff, on, 1,2,3,4,5,6,7,8,9,10...30 at least
15:36:12Zagoryes, something like that. only don't stop at 10
15:36:20LinusNok, 30 then?
15:36:46Zagorwhy a low upper limit? give'em 250 it that's what anyone wants
15:37:01*Hes is gone, bye
15:37:09 Quit matsl|away ("Client Exiting")
15:37:26LinusNZagor: because I planned on using the set_enum() function
15:38:04Zagorwell how about 1-10 in seconds, then 15,20,25,30,60,90,120,180 or something
15:38:14LinusNgood idea
15:38:16Lowfilerthat's good :)
15:39:04Lowfilerwill the next version include file-options? (delete, new etc.)
15:39:13Bagderpeople don't stop doing rockbox talk in there
15:40:11LinusNsilly sods
15:40:15Zagornot really
15:40:32Zagorit's not bad having user talk in there while we have more dev-talk on our list
15:40:48Zagorafter all, rockbox is now part of their world. so they discuss it.
15:41:42Zagorjust like the discuss cassette adapters, chargers, carrying cases and anything else
15:44:30Lowfileri think if you will add all functions (that are not integrated in rockbox) from the original firmware, no-one will use the original anymore
15:44:59Bagderwe will
15:45:39Lowfilermaybe archos will contact u then :)
15:46:50adi|homedidn't someone have a cli for building a playlist easily?
15:47:08Lowfilerdirectly on the player?
15:47:08Bagderfind . -name "*mp3" > allthemdarnedfiles.m3u
15:55:24BagderZagor: did you ever write down those quick "create me a play list" commands anywhere?
15:56:50Zagorno. a good candidate for the manual
15:57:13LinusNbacklight setting committed
15:57:44adi|homeBagder: what are you talking about?
15:58:01adi|homegod damnit..
15:58:04Bagderadi|home: we should write down those "one-liners" that creates playlists
15:58:04adi|homewtf is going on
15:58:09adi|homemy archos is mounted rw
15:58:14adi|homebut after like 5 commands
15:58:15Bagderfor linux and windows
15:58:16*Zagor just set up an rdate cron job for the web server
15:58:22adi|homei get 'ro file system; errors
15:58:28LinusNgotta go now, bye!
15:58:39Bagderadi|home: it means your fs is screwed up
15:58:42Bagderbye LinusN
15:58:55Lowfilercu linusN
15:58:57 Part LinusN
15:58:58Zagoradi|home: that's just what I got here too
15:59:01Bagderadi|home: that's that the fat fs does when it finds weirdnesses
15:59:17Bagder(in linux that is)
15:59:23Zagori think it's the usb layer that's broke down
15:59:24adi|homeBagder: you feel like cating me your fstab?
15:59:25Moufleor your linux kernel
15:59:43MoufleI had similar stuff on an ancian kernel
16:00:13Bagdercheck the kernel/messages log for details
16:00:25Bagderadi|home: for the archos mount you mean?
16:00:45Bagder/dev/sda1 /mnt/archos vfat noauto,umask=0 0 0
16:00:57adi|homeyeah.. hmm not good
16:01:05adi|homeAug 15 10:00:36 truman kernel: Filesystem panic (dev 08:01).
16:01:05adi|homeAug 15 10:00:36 truman kernel: FAT error
16:01:05adi|homeAug 15 10:00:36 truman kernel: File system has been set read-only
16:01:05DBUGEnqueued KICK adi|home
16:01:05adi|homeAug 15 10:00:36 truman kernel: Directory 11374: bad FAT
16:01:19BagderI got that too once
16:01:23Hadakaadi|home: have you done fsck.vfat on that?
16:01:24Bagderfsck time
16:01:29Hadaka/dev/scsi/host0/bus0/target0/lun0/part1 /mnt/jukebox vfat defaults,user,noauto,p
16:01:32Hadakaosix 0 0
16:01:46Bagderposix, wazzat?
16:02:06adi|homebummmm.. if i fsck it.. won't that clear the contents?
16:02:10Hadaka posix Allow two files with names that only differ in case.
16:02:14Hadakaadi|home: no
16:02:22Hadakaadi|home: mkfs clears contents ;)
16:02:43BagderI had to mkfs when that happened to me
16:02:49adi|homeokay.. should i assume
16:02:55adi|homefsck -t vfat?
16:02:55Hadakadosfsck has worked like a dream on my archos
16:03:25Hadakadosfsck -r /dev/sda1
16:03:31Hadakaand don't have your filesystem mounted
16:03:48adi|homeshould i assume, no action?
16:04:09adi|homeoky... first error..
16:04:12Hadaka -r Interactively repair the file system.
16:04:15adi|homeThere are differences between boot sector and its backup.
16:04:30Hadakawell for me, I just let it fix all the errors
16:04:50Hadakaand - it doesn't make the changes right away
16:04:58Hadakait makes them at the end and asks for confirmation
16:05:26Hadakaso just say yes to everything - and decide at the end whether to say yes or no to the final write
16:07:04adi|homewhile its running.. if there are no errors..
16:07:10adi|homeis there any kind of output?
16:07:21Bagderuse -l or -v for that
16:07:26Bagder -l List path names of files being processed.
16:08:10Hadakajust let it run
16:08:15Hadakait's going to take a while
16:08:49 Quit Moufle ("Client Exiting")
16:09:23CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
16:09:23*Bagder regged a suitable name on yahoo to reply to Qs
16:09:33adi|homewhats that?
16:10:54adi|homei wrote a script in college to screw with the reg process on yahoo
16:11:14adi|homei think i had like ~2k names registerd when it stopped working :)
16:11:28BagderZagor: <input type=hidden name=as_sitesearch value="">
16:14:02***Saving seen data "./dancer.seen"
16:21:53adi|homeokay.. why you said this was going to take awhile..
16:21:57adi|homeare we talking.. 15-20 mins
16:22:03adi|homeor we talking 2-4 hrs?
16:24:11Bagderanother "do you use rockbox" poll
16:27:28adi|homeheheh didn't take you long to notice did it ;)
16:27:58adi|homei wanna get answers on how many of them are using it, and if not, why
16:29:54adi|homevote again.. i fixed the polls
16:33:38 Join Moufle [0] (~moof@
16:41:00 Join mistax [0] (lamamba_@
16:49:23notch|lunch100% rockbox :-)
16:51:47 Quit Moufle (Remote closed the connection)
17:00:21BagderI'm off
17:00:25 Part Bagder
17:09:18dwihno"In short, buy an Archos and load RockBox!!"
17:11:18matphi guys. quick question: do we use much battery power when paused ?
17:11:30adi|home4/11 tshirst are my desing
17:11:45adi|homei wouldn't expect so matp but i dunno
17:12:20matpI mean, it's only scrolling the screen. I wouldn't have tought so either. Thanks.
17:12:33fraggletthe guy wanted it for jogging though
17:12:48fraggletthe archos manual explicitely says you shouldnt use it while jogging iirc
17:16:33 Join LinusN [0] (
17:17:06LinusNi noticed that the bitrate and total time is wrong in VBR tracks
17:17:25Zagorooh, bad
17:17:34Zagormust be the ff patch
17:17:37LinusNthe bitrate is always 128kbit
17:17:48Zagor(was in meeting last hour)
17:18:48LinusNwell, time to eat. just wanted you guys to know. bye!
17:19:11 Part LinusN
17:26:49Zagortime to go home. see you soon.
17:26:50 Quit Zagor ("Client Exiting")
17:38:38*notch|lunch must get diet coke or will die
17:38:54 Quit notch|lunch ()
17:47:48 Join aakil [0] (
18:14:05***Saving seen data "./dancer.seen"
18:15:34 Nick Zagor|away is now known as Zagor (
18:34:41 Quit aakil (Read error: 110 (Connection timed out))
19:03:11 Nick Zagor is now known as Zagor|pub (
19:49:23 Part elinenbe
20:14:06***Saving seen data "./dancer.seen"
20:15:40ironihi ppl
20:22:24 Join freshmaker [0] (
20:24:31freshmakerid3 tags seem to be broken in recent cvs. can anyone confirm this?
20:36:39RipnetUKneone know a good time server (ntp)? mine (machester uni) seems to be down
20:38:30 Join rwood [0] (
20:39:31 Quit RipnetUK (Read error: 113 (No route to host))
20:39:47 Part rwood
20:57:28Zagor|pubfreshmaker: confirmed
21:03:17freshmakerZagor|pub: Rats. *My* favorite pub has no WLAN ;-)
21:03:58freshmakerAny idea when it will be fixed? It's that FF/BW patch, isn't it?
21:03:59 Nick Zagor|pub is now known as Zagor (
21:04:05Zagorhehe, i'm home again :-)
21:04:22Zagoryes, that was prolly what broke it. looking at it now.
21:04:36Zagor(then I'm doing resume before anyone else gets the chance! ;)
21:08:27Zagorfound it. i missed one line from the patch
21:25:34 Join elinenbe [0] (
21:52:35adi|workdidn't someone here say they were doing SMS bundling?
22:02:30 Join Electrocut [0] (
22:07:28adi|workZagor you around?
22:12:50HesGood evening.
22:14:07***Saving seen data "./dancer.seen"
22:22:56 Quit edx ("later")
22:23:27freshmakerHas anyone tried to compile with '-Os'? saves about 5k, but I cowardly didn't try it on my jukebox.
22:27:37Zagorhere noe
22:27:58Zagorfreshmaker: never tried it
22:29:03freshmakeranyway, it a bit too early for compile time fine tuning.
22:31:48adi|workZagor.. Smash...
22:31:54adi|workyour my buddy :)
22:32:11*adi|work started looking into sms today...
22:32:19adi|workany docs you could recommend?
22:33:19ZagorI've got plenty. what layer are you interested in?
22:33:35adi|workfor now..
22:33:35 Join rwood [0] (
22:33:40adi|workrecipient to gateway
22:33:49adi|workother then that i can't tell you much :)
22:33:57*adi|work is just feeling out the water as yet
22:34:10Zagoryou want to send or recieve?
22:34:19adi|worksorry... send
22:34:30adi|workwith confirmation (where possible)
22:34:33freshmakersome GNU make guru around? how do i use a conditional with two 'anded' expressions, like "if eq(a,b) and defined(foo)"?
22:34:55ZagorGSM messages always reach their destination, or timeout. I don't think many SMSCs deal with confirmation.
22:35:29Zagorfreshmaker: use multiple lines
22:35:38Zagorif eq(a,b)
22:35:44Zagorif defined(boo)
22:36:37freshmakerthanks, thought about that, but it's a bit ugly, though.
22:37:30Zagorafaik, it's the only way. make is but a simple timestamp comparer :)
22:39:49rwoodthis is randy wood, i sent in the ROLO code a week or so ago - i'm looking to get a recorder to finish the task if no one gets to it first. is USB 2.0 worth the price difference between the 10 and 20?
22:40:01Zagorrwood: oh, absolutely!
22:40:33Zagorwelcome, by the way :-)
22:40:46Zagorand thanks for the rolo patch
22:41:00rwoodZagor: you are a smooth talker - i'll order today and should have it early next week - you are welcome on both
22:42:03HesI have a recorder 20... usb2.0 really rocks
22:42:41Hesdon't have an USB 2.0 interface on my work machine or laptop, it takes ages to transfer a large amount of data
22:42:49rwoodmy computer has 1.1 any suggestions on a good PCI card
22:42:52 Join RipnetUK [0] (
22:43:00Heswhen compared to the home box which I bought the card for
22:43:36HesI got a cheap 5-port one which has the nec chip (first USB 2.0 interface chip that shipped)
22:43:52Heswhich is alright, although never reaches full USB capacity
22:44:10ZagorI have a nec chip too
22:44:32rwoodi'll look for that also
22:45:13rwoodby the way, has anyone tested rolo on a player?
22:45:30HesThe chip is probably on a huge amount of brand cards
22:45:40Heslinux/Documentation/usb/ehci.txt talks about it
22:49:35HesMy card is Datafab's PCI-USB2N
22:49:58 Quit Electrocut (Read error: 104 (Connection reset by peer))
22:50:11HesThe ROLO code is a Very Importand and Good enhancement a lot of people have been missing
22:50:31HesI'd just love to be able to load the archos firmware while rockbox doesn't record just yet
22:51:35Zagorit will enable a kind of task switching behaviour between different applications. imagine an address book application or calendar that you switch to, and then back.
22:52:12rwoodHes: my best guess is that i am not getting the hardware back to an initialized state before transferring to the new app
22:52:25rwoodi don
22:52:37Zagorrwood: you should talk to Linus. he tried it and had some thoughts.
22:53:00rwoodsorry - i don't think that it will take that much to get it going
22:53:02Zagorfor one, you don't need to relocate. just put the loader code at the top of the RAM using a linker segment.
22:54:37rwoodZagor: i didn't want to get into changing the design - i just worked within the constraints as i saw them - i'll check in later so see if linus is around
22:54:57Zagordo so
22:55:31rwoodi've got to go - i'll check back later
22:55:50 Quit rwood ()
23:02:06 Join Lear [0] (
23:03:48LearZagor: volume shortcut (menu+left/right) seems broken on the player in CVS. It just jumps to the next track.
23:06:20Leartried disabling "discard first release event", but that didn't help...
23:06:33LearAlso, a status_draw is missing too.
23:09:32Zagorthe button situation is not optimal. i'm thinking of alternative ways to handle it.
23:10:41Learyeah, the ff_rewing vs. next/prev isn't the that easy to understan... :)
23:15:52Zagoryou have wps 1.70 right?
23:17:18dwihnowps wps wps wps wps wps wps wps wps wps wps wps!
23:18:01freshmakerso, then. i have a new patch for the recorder that optionally leaves out screensavers and games. together with '-Os' the result is 95k -> 67k. is it worth it?
23:18:55Zagorjust a couple of #ifdefs ?
23:19:11freshmakeryup like so.
23:19:21Zagorsure, why not
23:19:25freshmakermostly in the makefiles
23:19:53Zagordid you make an option in tools/configure too?
23:20:27freshmakerthat's still todo. and then some tests.
23:20:43freshmakerexpect the patch in one hour or so.
23:21:31Zagori'll probably be in bed by then, but we'll see :)
23:24:18LearAh, the volume *does* change, the problem is that the release is considered to be a track change!
23:25:08LearLooked at the release handling in button.c. Looks a bit odd to me; should "BUTTON_REL | diff" really be posted on release?
23:26:17Zagoryes. diff says which buttons are no longer pressed (but was last time) and the REL bit says it/they were released and not pushed
23:26:37Zagorthe problem is most likely in wps
23:30:06LearI'm not entirely sure on that... In this case, left/right is released while menu still pressed; wouldn't that - with the "diff" post - generate e.g. a left | rel even, without the menu button? => track change...
23:31:00Zagorwell, it's simple elimination. the bug came alive today. wps.c was changed today. button.c was last changed the 7th. :)
23:31:35ZagorLear: yes, that's why there's a complex state handling in wps
23:31:41Zagorno next until MENU|REL
23:32:30Learotoh, one of the wps.c changes was changes in release mask (though that was yesterday), which could trigger a bug/shortcoming in button.c...
23:34:15LearI see no "if(menu_button_is_down)" checks in ff/rewind or prev/next...
23:35:48Zagoryou're right. that needs to be added.
23:35:54Zagorbutton.c still does the right thing
23:36:10Zagorthe event was a PREV/NEXT release
23:36:19Learcould be, yes. Just about to try that addition (should that be only for players?).
23:36:52Zagoronly players use MENU+PREV/NEXT for volume, so #ifdef that yes
23:37:01LearAnd that "if" should read "if (!menu...)"...
23:37:33Zagorif(!menu_button_is_down) mpeg_next()
23:38:41 Quit freshmaker ("ChatZilla 0.8.9 [Mozilla rv:1.1b/20020809]")
23:39:25Learthat'd better be "else if(!menu_button_is_down)" on the whole block, wouldn't it?
23:41:11ZagorI see now ff_rewind bypasses the keylock. not very nice :)
23:41:19LearSeems to work too.
23:41:35Zagorthat keys_locked should be the outermost if
23:42:14LearBut only "bypasses" in the release handling, right? ff_rewind won't be set if keylock is on...
23:43:10Zagortrue. but the logic is still better if keylock is checked first.
23:43:27Zagorless complex to follow
23:43:28Learclearer, at any rate, yes.
23:47:47LearOK, think I have something now; adds keylock check and menu button (for players)...
23:49:39Learargh; showing "Keylock ON" if locked on release when in menu isn't good. Another #ifdef...
23:52:35LearSeems to work now too. Generating patch...
23:57:51Zagorhi there
23:57:56 Join freshmaker [0] (
23:57:59elinenbeso much seems to have been done today... great job guys.
23:58:53elinenbeI do think however you should implement playing divx files in 2 bit color on the LCD, and until then I am going to bitch non-stop about other features I WANT, WANT, WANT!!!! :)

Previous day | Next day