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 2002-12-16

00:00:54HadakaDisciple1: you could try implementing real paged memorymanagement as well, with all empty/useless pages used by the mpeg thread as buffer
00:01:11Disciple1Hadaka: ouch
00:01:17LinusNHadaka: dream on
00:01:21Hadakawhaaaat? ;)
00:01:45Disciple1Hadaka: and a real mmu management, with swap on the drive ?
00:01:59HadakaDisciple1: no mmu, no virtual addresses :)
00:02:55Hadakabut application driven swapping would be nice ;)
00:03:03 Quit breaker ()
00:03:09BoD[]what about a port of netbsd ?
00:03:24LinusNHadaka: yeah, we could swap out not yet played data to disk! :-)
00:03:59Hadakaseriously though, there are three different memory models that are needed
00:04:09Hadaka1) while playing - as much as possible for the buffer
00:04:32Hadaka2) while browsing, perhaps playing at the same time - everything must work but the buffer doesn't need to be that big
00:04:49Hadaka3) while running and exclusive application - facilitate it's work as much as possible
00:06:49LinusNHadaka: I won't give that much thought until we find a nice way for the mp3 thread to use fragmented buffers
00:07:21LinusNfirst of all, it will need to scatter the disk reads
00:07:28 Part oZ
00:07:47LinusNthen the seek functionality will be a little trickier, along with the track change detection
00:08:13HadakaLinusN: requires the familiar notion of a buffer-page and a page directory, yes
00:08:22LinusNHadaka: yup
00:09:04LinusNi also don't care how the buffers are managed, i only care about the mp3 data handling
00:10:24 Part LinusN
00:10:30 Join LinusN [0] (
00:10:52Hadakayeah - the solution would have to be something middle-way between a specialized solution where the applications cooperatively ask and release stuff and a generic memory manager
00:10:54LinusNdamn D-Link Wireless AP
00:13:06LinusNHadaka: all i worry about is how the mpeg dma handler copes with the fragmented data
00:13:52LinusNnever buy a dlink 614+
00:13:58***Saving seen data "./dancer.seen"
00:14:07HadakaLinusN: there isn't necessary a need to fragment the buffer that much
00:14:13LinusNit drops the connection all the time
00:14:22LinusNHadaka: why not?
00:15:29HadakaLinusN: because in a cooperative any bigger allocations can take time, so the parts can be dragged from suitable areas
00:17:02 Part LinusN
00:17:09 Join LinusN [0] (
00:17:16LinusNHadaka: explain that firther
00:18:20Hadakawell as an entirely trivial example, the mpeg buffer could be kept entirely continguous and just reduced in size every time there is already played data at the start of the buffer so it can be done
00:18:33Hadakathis would ofcourse cause a huge wait on memory allocations
00:18:54LinusNnot very nice
00:19:00Hadakabut the approach could be used in a smaller scale to make sure fragmentation stays minimal
00:19:42LinusNas long as there is fragmentation, it won't matter if it's minimal or not, the mpeg thread must be able to deal with it anyway
00:20:18Hadakaisn't there always "fragmentation" of one sorts - I mean the buffer is not infinite so there is a spot where it will have to jump from the end of the buffer to the start of the buffer
00:20:35LinusNyes, but that's a very simple case
00:21:38Disciple1hmmmm, but what is the main interest to do that ? :)
00:21:56Hadakayup, but it's a smaller step to go from that to something a bit more complicated
00:22:01HadakaDisciple1: to achieve 1 and 2
00:22:19LinusNDisciple1: on occasions, a large buffer may be useful for other things, like caching large directories
00:22:30LinusNor for a game
00:23:11Hadakawhen I'm playing a playlist, I do not wish for one byte of extra memory spent that could be spent on the mpeg buffer thus prolonging battery life - and when I'm browsing I do not wish to be limited in what I can do - I wish to be able to sort entries, find, do tagging operations - all sorts of stuff which requires memory
00:23:15LinusNHadaka: take seek as an example, it would have to traverse a list of buffers to find the right location
00:23:49HadakaLinusN: yes, there would have a trivial directory then
00:24:37LinusNthe track change algorithm uses pointer arithmetics to detect track changes today
00:25:04LinusNthat has to be changed
00:25:21Hadakayeah I saw that when reading the dma code :)
00:25:40LinusNso what do we have in mind? a directory of 64K blocks?
00:26:13Hadakawell that's one solution
00:26:35LinusNand a second layer that manages those blocks, much like malloc()
00:26:55LinusNi mean to allocate smaller portions from a block
00:27:19Hadakayup, a slab allocator on top of those blocks then
00:28:26Hadakathe second layer can be almost duplicated from any allocator
00:28:37LinusNwe have one already
00:29:08Hadakaon the first layer it's probably beneficial to embed the mpeg thread there to make it perform nicely
00:30:20Hadakahmh, APR could be ported on rockbox ;)
00:31:12Hadakaapache portable runtime - the apache portability library
00:32:14Disciple1perhaps I can do a GUI for recorder ? :)
00:32:32Hadakarosh! rockbox shell!
00:33:41Hadakahmh, I wonder why anyone hasn't mentioned anything about the CVSROOT/modules patch
00:34:01LinusNi haven't had the time
00:34:18LinusNbut it looks useful
00:35:30Hadakait strikes me as very odd that no-one else has had anything to say about the matter
00:35:41LinusNme too
00:36:00LinusNbut zagor and bagder have been busy the last days
00:36:05Disciple1about what ?
00:36:26Hadakathe first thing I had to do when getting rockbox CVS was to hand edit a CVS directory and entries file since I didn't want to be update all the directories separately
00:37:18HadakaDisciple1: the rockbox CVS doesn't have anything in CVSROOT/modules right now
00:37:53LinusNHadaka: here's how it looks today:
00:38:00LinusNrockbox &apps &firmware &docs &tools
00:38:00LinusNrockbox-devel &apps &firmware &docs &tools &uisimulator &gdb
00:38:00LinusNrockbox-all &apps &firmware &docs &tools &uisimulator &gdb &www
00:38:00DBUGEnqueued KICK LinusN
00:38:00LinusNwebsite &www &docs
00:38:14LinusNzagor changed it 3 days ago
00:38:47Hadakaoh.. I don't receive commit mails yet
00:39:04LinusNHadaka: we haven't set up commit mails for all modules
00:39:37Hadakawell then everything should be fine - I just would've expected a reply on the list as well :)
00:39:58Hadakanaked@oro:~/src/rockbox$ cvs co rockbox
00:39:58Hadakacvs server: Updating apps
00:39:58HadakaU apps/FILES
00:40:09Hadakacvs server: Updating firmware
00:40:09HadakaU firmware/COPYING
00:40:39Disciple1Anonymous CVS doesn't work ! and I still havent sourceforge account :)
00:40:49HadakaDisciple1: works for me?
00:41:10Disciple1Hadaka: works for you ? Damn !
00:44:08Disciple1yeah, it works \o/
00:49:24adi|homedamn you linus.. you beat me to rejecting that mp3pro one
00:51:49LinusNadi|home: that was my goal
00:53:40adi|homeis there a specific reason we don't close requests like that?
00:53:52adi|homebecause we seem to have alot of them
00:54:30LinusNadi|home: because if we close them, they won
00:54:40LinusNt appear in the list on the site
00:54:54LinusNand then we will get new ones
00:55:08adi|homehehe dude.. if you havne't notice.. the one's where we generally answer 'ITS IN THE FAQ' aren't reading docs anyway ;)
00:55:11LinusNwe want the rejected ones to show, so people see that
00:55:12adi|homethey blindly submit
00:55:26LinusNof course
00:55:40adi|homeoh. the mic one you just commited.. do you want to reject that one too?
00:55:41LinusNthat's why it's so fun to answer them like that
00:56:00LinusNit's certainly possible, with a hardware mod
00:56:11LinusNcall it sarcasm if you like :-)
00:56:26adi|homehehehe k
00:56:40LinusN"can i drive the car, mom?" -"Sure, when you're 18"
00:58:28LinusNtalk about feature request spamming
01:00:59adi|homebe nice if on the listing page, we showed the status
01:01:00 Join kSADerfos [0] (
01:01:07adi|homeso we could see easily what is / isn't rejected
01:01:35 Join e-potis [0] (
01:01:39 Quit e-potis (Client Quit)
01:02:09LinusNadi|home: we don't?
01:02:37adi|homesorry. i mean the resolution
01:02:52LinusNadi|home: we show the resolution
01:03:11adi|homenot on the straight listing
01:03:25LinusNadi|home: you mean on sourceforge?
01:03:37LinusNok, we have no control over that
01:03:44LinusNlook at the rockbox site
01:03:46adi|homelike i said.. i wish we could
01:04:25adi|homelike i said.. once again we are hoping they are going to read it ;)
01:04:37adi|homewe can't even prevent dupes of stuff thats there already
01:04:40adi|homesilly users ;)
01:05:40Disciple1good night all ! see ya'
01:05:47 Quit Disciple1 ("time2slepp")
01:06:33 Part LinusN
01:07:08 Join LinusN [0] (
01:07:42LinusNlike the silly request i rejected earlier today, when a guy wanted us to implement a "Hold feature"
01:08:14adi|homei liked that response
01:08:43LinusNwho do people call a keylock "Hold"?
01:09:58kSADerfoscreative users :)
01:10:10LinusNwell, Archos calls it Hold as well
01:10:30LinusNis it a bad translation from french?
01:13:02kSADerfoswho knows
01:13:29kSADerfoseverybody call it hold
01:13:47kSADerfoseven my philips discman has a "hold feature"
01:14:48 Nick kSADerfos is now known as e-potis (
01:23:59 Join breaker [0] (
01:37:02 Quit breaker ()
01:43:42 Part LinusN
01:45:08 Quit e-potis ()
01:59:46 Join Sachem [0] (
02:00:04Sachemhey , just found out about this project
02:00:15 Quit probonic ()
02:00:17Sachemit's excellent, I hated the archos firmware
02:00:27hardeepas do we all
02:00:51adi|homehehe welcome to a brand new world
02:01:00Sachemindeed,I was suprised to see anyone working on it,
02:01:40SachemI do wish I had the recorder though
02:02:29SachemI purchased this mp3 player almost a year ago, it's pretty good - just wish it wasnt so heavy and big
02:02:30adi|homeheheh its worth the purchase...
02:02:36adi|homesell your unit.. buy the recorder :)
02:03:07SachemIf I had a choice though, I would pick the ipod
02:03:19CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
02:03:19*Sachem thinks everyone thinks hes a trader!
02:04:44adi|homei banish thee!
02:05:17Sachemhey do you guys use music match, it works with the archos, in transfering
02:06:06SachemI've been using it for a month or so now,it's pretty good. Havent found anything like it, do you guys know of any other programs that can interact with the archos in anything else?
02:06:08hardeepi prefer to use EAC/lame for ripping CDs and copying files manually using explorer
02:06:47Sachemthat's what I used to do, but the transfer times were outrageous sometimes, you would have to charge it 100% before you transfer or it somehow dies,or maybe cause Im not using usb 2.0
02:07:07hardeepi don't see how music match would improve that performance though
02:08:49Sachemseems to do something
02:08:49SachemI wish winamp had that
02:08:49Sachemmusic match is bulky
02:08:50DBUGEnqueued KICK Sachem
02:08:50Sachemwhat headphones do you guys use? My originals just happened to break today
02:09:04Hadakakoss porta-pro
02:09:18Hadakaand I don't use a software to copy mp3's either
02:14:01***Saving seen data "./dancer.seen"
02:18:52adi|homedrag and drop is your friend..
02:26:34 Quit BoD[] ("fuck *.*.*.*.*.*")
02:40:01Sachemhows that archos fm radio mp3 player, did it come out yet?
02:40:47Bryantdo any of you guys know how i can connect to two servers using one irc window?
02:41:17 Join _seb_ [0] (
02:42:07Sachemwhich irc client
02:47:35Sachemjust do , uhm /server -m
02:47:47Sachemadd nick and port
02:58:05hardeepSachem: the fm radio mp3 player is out, however rockbox software doesn't work with it currently
03:08:22adi|homeno currently about it...
03:08:32adi|homeRockbox doesn't have the intention of supporting it
03:12:15 Quit Bryant ()
03:13:10 Join Bryant [0] (
03:13:32 Quit Bryant (Client Quit)
03:16:10 Join Bryant [0] (
03:22:51 Quit Bryant (Read error: 104 (Connection reset by peer))
03:34:25 Join Bryant [0] (
04:14:02***Saving seen data "./dancer.seen"
04:37:51 Quit Sachem ("Just cause he wear a kufi dont make him right, Just cause you don't understand him don't make him nice.")
05:04:10 Nick _seb_ is now known as seb-sleep (
05:32:11 Join elinenbe [0] (]I[
05:36:17 Nick elinenbe is now known as elinenbe|sleep (]I[
06:14:04***Saving seen data "./dancer.seen"
08:01:45 Nick dw|gone is now known as dwihno (dwihno@Bald067.Baldakinen.Umea.SE)
08:14:06***No seen item changed, no save performed.
08:55:56 Join Zagor_ [242] (
08:56:01 Nick Zagor_ is now known as Zagor (
09:03:13hardeepheya Zagor
09:03:46hardeepi think i've found a bug in the file writing code...
09:03:51hardeepyou can see it in:
09:04:25hardeephowever, i don't see how it's happening...
09:05:17Zagori've ran file contents checks to death...
09:05:33hardeepif you begin writing at the end of a sector but the size spans the sector, the end of the sector is being zeroed
09:05:50hardeepi've debugged all the way to ata_read_sectors and everything looks right
09:05:56hardeeper ata_write_Sectors
09:52:18hardeepmorning :)
09:53:46hardeepZagor: a couple more items which may help...
09:54:27hardeep1. I printed out the data before and after the headbytes write call (file.c:361) and it was correct which implies that the data isn't being modified
09:55:08hardeep2. I saved the file right after that write (ie. before anything else was done to it) and it still showed the zero bytes
09:55:44hardeepso, it seems like something funky is happening in the headbytes write... i just couldn't track down what
09:57:38Zagoris this easily repeatable for you?
09:57:50hardeepyeah, 100% with my queue code
10:00:33hardeepsearch for queue_add in to see how i'm doing the writing
10:02:30Zagorlooks ok to me
10:04:38Zagori'm rather swamped with work today so I can't look more closely, i'm afraid. :-(
10:05:23hardeepyeah, I know that feeling. :( I'll keep on poking around
10:05:55hardeepthis is the last issue i know of before i submit queuing for review
10:13:14 Join TotMacher [0] (
10:14:07***No seen item changed, no save performed.
10:19:20hardeepsleep time
10:19:22 Quit hardeep ("BitchX: all the things phone and hop won't include")
10:20:20 Quit TotMacher ()
10:33:19 Join dwihno_ [0] (dwihno@Bald067.Baldakinen.Umea.SE)
10:33:19 Quit dwihno (Read error: 104 (Connection reset by peer))
10:46:33 Nick dwihno_ is now known as dwihno (dwihno@Bald067.Baldakinen.Umea.SE)
11:16:45 Join Bagder [0] (
11:16:54Bagderhey ho
11:21:17CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
11:21:17*dwihno gives Bagder and Zagor some pepparkaksfil
11:29:18BagderZagor: you preparing a "donation" web page?
11:29:28Zagornot right now, no
11:29:35Zagorbusy @work :(
11:30:50 Join LinusN [0] (
11:31:31Bagderhi LinusN
11:45:55dwihnoLinus! :D Yay! The power of three will set you free!
11:46:14dwihnoThree different heroes with three magic abilities :-)
11:46:39LinusNdwihno: are you high on starka såsen?
11:47:38dwihnoLinus: Nah, I lack sleep :)
11:47:54LinusNme too, always
11:50:31dwihnoYes, but you have the Smilla siren - "Guaranteed to keep you awake"
11:51:31Jet8810god damnit, I cant fall asleep ;)
11:51:38Jet8810woke up at 5 and been tossing and turning for an hour
11:51:45Jet8810so screw it, ill just get some work done lol
12:11:20 Join TotMacher [0] (
12:14:11***Saving seen data "./dancer.seen"
12:21:38 Part LinusN
12:55:04 Join pyvasene [0] (~pyvasene@
12:59:28 Quit Bagder ("Client Exiting")
13:01:07 Join matsl [0] (
13:01:47*matsl is away: I'm busy
14:07:04 Quit Jet8810 (Read error: 104 (Connection reset by peer))
14:14:14***Saving seen data "./dancer.seen"
14:28:06 Join LinusN [0] (
14:32:00 Quit ken0_ (Read error: 104 (Connection reset by peer))
14:32:25 Join ken0_ [0] (marklar2@
14:34:59 Quit ken0_ (Read error: 104 (Connection reset by peer))
14:35:37 Join ken0_ [0] (marklar2@
14:56:13 Part LinusN
15:56:46 Quit TotMacher ()
16:06:39 Quit matsl (Remote closed the connection)
16:14:18***Saving seen data "./dancer.seen"
16:15:50 Nick elinenbe|sleep is now known as elinenbe (]I[
16:31:45 Nick dwihno is now known as dw|gone (dwihno@Bald067.Baldakinen.Umea.SE)
17:01:33 Part Zagor
17:35:02 Quit pyvasene (
17:35:03 Quit elinenbe (
17:35:03 Quit noy0m (
17:35:03 Quit MrSnazz (
17:35:23NJoinpyvasene [0] (~pyvasene@
17:35:23NJoinelinenbe [0] (]I[
17:35:23NJoinnoy0m [0] (
17:35:23NJoinMrSnazz [0] (
17:47:29 Join montykid [0] (
17:50:42 Quit pyvasene ()
17:50:51 Part montykid
17:54:49 Join probonic [0] (
17:59:37 Quit ken0_ (Read error: 104 (Connection reset by peer))
18:00:08 Join ken0_ [0] (marklar2@
18:09:15 Quit probonic ()
18:14:21***Saving seen data "./dancer.seen"
18:23:53 Quit ken0_ (
18:23:53 Quit elinenbe (
18:23:53 Quit noy0m (
18:23:53 Quit MrSnazz (
18:24:33NJoinken0_ [0] (marklar2@
18:24:33NJoinelinenbe [0] (]I[
18:24:33NJoinnoy0m [0] (
18:24:33NJoinMrSnazz [0] (
18:26:28 Join hardeep [0] (hardeeps@
18:28:46 Quit elinenbe ()
19:05:14 Join webmind_ [0] (
19:08:22 Join TotMacher [0] (
19:08:25 Quit webmind (Read error: 54 (Connection reset by peer))
20:14:24***Saving seen data "./dancer.seen"
20:22:39 Quit hardeep ("[BX] Been around the world and found that only stupid people are breeding.")
21:07:46 Join hardeep [0] (hardeeps@
21:24:26 Nick seb-sleep is now known as _seb_ (
21:30:21 Join pd [0] (
21:30:43pdhow do i know what ROM firmware i have?
21:30:52pdon a jb 6000
21:36:34 Quit pd ()
22:14:28***Saving seen data "./dancer.seen"
23:05:35 Nick _seb_ is now known as seb-away (
23:07:38 Join JimmyDurante3 [0] (
23:13:08 Join totmach3r [0] (
23:13:44 Quit totmach3r (Client Quit)
23:21:39 Quit TotMacher (Read error: 60 (Operation timed out))
23:24:38 Join Jet8810 [0] (
23:26:01 Join NiarkY [0] (
23:31:05JimmyDurante3anyone heard if anything greater than 40 gigs is working w/ the archos?
23:40:19 Join BoD[] [0] (
23:40:26BoD[]hello !!

Previous day | Next day