Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

#rockbox log for 2005-09-28

00:00:09*tucoz blushes
00:00:58 Quit arkascha (Remote closed the connection)
00:01:39linuxstbHas anyone tried to compile the mp3 encoder for the target? I get an error at line 1471.
00:01:43preglowdoesn't look like the mp3 encoder has much of a psychoaccoustic model to me
00:02:05preglowbut then again, it's probably hard pressed doing just what it's doing now
00:02:30tucozdidn't antonius have it running on target?
00:02:31rasherlinuxstb: worked for me
00:03:18linuxstbmp3-encoder.c:1471: error: can't find a register in class `ADDR_REGS' while reloading `asm'
00:03:32rasherweird
00:06:34 Join Chamois [0] (i=Chamois@AStrasbourg-252-1-11-180.w82-126.abo.wanadoo.fr)
00:06:49Chamoishi
00:07:00rasherhi
00:07:05Chamoisi have a problem with the cache patch from slasheri
00:07:10Chamoisi have a splasj message
00:07:23Chamoiswith " no .rockbox"
00:07:36Chamoisand installation incomplete
00:08:13Chamoisanyone has the same problem ?
00:08:26rasherwhee
00:08:31rasherDon't think so
00:08:35rasherhaven't heard it yet at least
00:08:56Chamoisif i installt the latest daily build all is good
00:09:11Chamoisif i install a version with the dir patch it's ok
00:09:19Chamoisbut when i enable teh dir caching and reboot
00:09:29Chamoisi have this Splash message
00:10:09Chamoishmm
00:10:13Chamoisslasheri it's for you
00:10:51tucozChamois, you have to enable the cache, wait for the splash to disappear, move back to the file browser, and then reboot.
00:10:52linuxstbrasher: I cleaned my build directory, re-ran configure and it compiled fine....
00:11:10Chamoisi have enable the cache
00:11:15tucozah, ok.
00:11:28Chamoisand the splash doesn't dissapear
00:11:30rasherlinuxstb: Ah, fun
00:12:05 Join webguest29 [0] (n=3ed6f2c7@labb.contactor.se)
00:12:16webguest29hi
00:12:32tucozhello
00:12:42webguest29any news regarding the > v2.4 (and 2.5) recorder V1 battery charging bug?
00:12:57rasherI think there is no bug
00:13:05rasherjust incorrect battery percentage during chargin
00:13:12 Join arkascha [0] (n=arkascha@xdsl-213-168-119-63.netcologne.de)
00:13:17webguest29it does not charge my batteries correctly (charges only about 1 hour or so)
00:13:29webguest29no
00:13:53webguest29it says "100%" and "trickle charging" which means battery should be charged
00:13:59rasherah
00:14:10webguest29bit after 1 hour of playing, rockbox shuts down and batteries died
00:14:11*rasher knows nothing of the archos
00:14:33amiconnwebguest29: You're willing to test something? (Not atm, as I'm busy I can't prepare a test build right now=?
00:14:42webguest29sure!
00:14:59webguest29send it via email?
00:15:03amiconnI think you're the one who posted in the forum?
00:15:05webguest29mario.schmidt@mediavillage.de
00:15:08webguest29yes
00:15:11webguest29vpn-user
00:15:51tucozgood night
00:15:54webguest29There really is a bug in rockbox, since the archos firmware charges correctly
00:15:58 Part tucoz
00:16:09***Saving seen data "./dancer.seen"
00:16:28amiconnOkay, so I'll prepare a test build which logs battery voltage. What to do with it: Install build, run down batteries, let it charge fully (as far as it gets), then post the data file it creates
00:16:43webguest29ok
00:16:59webguest29shall i start with really charged batteries?
00:17:17webguest29they have 2400mAh so that may take a while...
00:17:52amiconnThe test should start with empty batteries, as the test build will log voltage while charging
00:18:19webguest29okay
00:18:36amiconnHopefully this shows what makes the algorithm think the cells are full when they aren't
00:18:43rasheramiconn: could this perhaps be made part of the official source? (not built by default)
00:18:53webguest29I think using the voltage is just a wrong way to do things
00:19:06amiconnI *suspected* something like that, although it isn't that bad here
00:19:10webguest29because newer batteries try to keep their target votlage as long as possible
00:19:31amiconn(charges for some hours and runs for some hours, but not for the runtime i'd expect from 2500 mAh cells)
00:19:47webguest29exactly.
00:19:50amiconnWe only have the voltage
00:20:01webguest29yes.
00:20:12amiconn...and perhaps a crude estimate of the current
00:20:15webguest29maybe we need some calibration
00:20:22webguest29a calibration file
00:20:46amiconnI think it would help to measure the voltage with the charging current switched off (briefly)
00:20:49webguest29new batteries should be calibrated using rockbox
00:21:17amiconn...but we need curves from various batteries, stock type as well as high capacity
00:21:24webguest29that way rockbox would know how to handle this battery type
00:22:19webguest29No. What i mean: When first time using rockbox, the user should do a full charge cycle with rockbox. rockbox loggs the voltage during that process
00:22:30webguest29and calibrates itself according to the voltage change.
00:22:41amiconnThat won't help much
00:22:50webguest29The voltage fluctuates with my batteries, too. but very little
00:23:29amiconnThe charging algo doesn't shut off based on the voltage. Doing so would be rather bad, as end-of-charge voltage is temperature dependent
00:23:48webguest29ah
00:23:54webguest29so when does it stop?
00:24:09webguest29it stops too early for me
00:24:34amiconnIt uses a delta-U approach, but either it is too snsitive, or something else is wrong
00:25:10webguest29do you know how the archos firmware handles it?
00:25:22amiconnno
00:25:38webguest29i wonder why you guys don' t have the same problem
00:33:52arkaschaany ideas yet to read an attached keyboard?
00:34:14webguest29attached to what?
00:34:27arkaschato an iriver in my case
00:34:51webguest29afaik thats not possible because the usb-port can' t act as master
00:35:07rasherThe usb is connected to a pure slave chip
00:35:09rasherhardware
00:35:18preglowwell
00:35:20amiconnhrmpf :/
00:35:23arkaschanot an usb keyboard, sorry, I have that small saitek thing here, attaches toa palm as well
00:35:24ender`it might be possible on H3xx though
00:35:27preglowyou can do a ps2 -> spi bridge
00:35:32preglowand connect via remote
00:36:25arkaschaI mean ok, we need a special cable, but the pin layout suggests it is possible
00:36:52preglowthe only port you can connect a peripheral to is via remote in
00:37:04pregloweither that or do some fancy stuff with the optical/line in port
00:37:23amiconnOr line in, perhaps
00:37:32amiconnMeh, ignore me
00:37:38preglowheh
00:37:46pregloweither way it's possible, but don't expect to see it
00:37:56preglowi for one can't possibly imagine what i'd use it for
00:38:08arkascharenaming files
00:38:15preglowright, but hey
00:38:24preglowi can make you a ps2->spi bridge if you just pay me
00:38:34arkascha:-; taking notes
00:38:54 Join webguest81 [0] (n=d5ee4049@labb.contactor.se)
00:41:02 Quit muesli- (Read error: 110 (Connection timed out))
00:43:09webguest29amiconn, you got my email adress for the test build?
00:44:24webguest29cause i have to leave for now
00:44:34webguest29thanks for trying to help me
00:44:38amiconnyes (there also is an irc log)
00:44:44webguest29ok
00:44:51webguest29thanks guys
00:44:55webguest29good night
00:45:54 Quit webguest29 ("CGI:IRC (EOF)")
00:58:53 Nick TiMiD is now known as TiMiD[ZZZzzz] (n=TiMiD[FD@asgard.valombre.net)
01:00
01:01:54 Quit Chamois (" HydraIRC -> http://www.hydrairc.com <- Leading Edge IRC")
01:05:17 Nick Sucka is now known as actionshrimp (n=NNSCRIPT@host81-156-209-15.range81-156.btcentralplus.com)
01:05:46 Join bagawk [0] (i=1000@unaffiliated/bagawk)
01:11:05 Quit Seed (Nick collision from services.)
01:11:12 Join Seed [0] (i=ben@l192-117-115-168.broadband.actcom.net.il)
01:15:33 Quit ender` (Read error: 113 (No route to host))
01:23:49 Quit webguest47 ("CGI:IRC")
01:31:12 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
01:51:59 Quit preglow ("leaving")
02:00
02:16:13***Saving seen data "./dancer.seen"
02:42:50 Quit Paul_The_Nerd ("Chatzilla 0.9.68a [Firefox 1.0.7/20050915]")
02:44:34 Quit bagawk ("Leaving")
02:48:36 Join _arkascha [0] (n=arkascha@xdsl-213-168-116-141.netcologne.de)
02:53:30 Quit arkascha (Read error: 110 (Connection timed out))
03:00
03:02:32 Quit _arkascha (Remote closed the connection)
04:00
04:05:50 Join QT_ [0] (i=as@madwifi/users/area51)
04:15:28 Join Mirthoneist [0] (n=mirthypo@c-66-177-213-74.hsd1.fl.comcast.net)
04:16:17***Saving seen data "./dancer.seen"
04:16:24MirthoneistJust wanted to show my support for Rockbox on iRiver H3XX by showing up here in the IRC channel! :)
04:16:52MirthoneistGood luck to all of who are developing it!
04:16:56fuzziewhee :)
04:17:01MirthoneistHehe ^_^
04:17:03 Quit QT (Read error: 113 (No route to host))
04:35:52 Join ashridah [0] (i=ashridah@220-253-121-239.VIC.netspace.net.au)
04:48:03 Join grimreap [0] (i=grimreap@c-67-165-167-67.hsd1.il.comcast.net)
04:49:38grimreapcan someone explain to me what the bookmark feature actually does on the ondio player?
05:00
05:14:35 Quit dpassen1 ()
05:29:42grimreapare you supposed to be able to like go back to the position on the audio using the bookmark?
05:31:42grimreapI can't figure out how to play the bookmark
05:40:10 Join BirdFish [0] (n=bradbox8@64.108.5.134)
05:40:49BirdFishJust wondering, and really don't mean to bother or annoy anyone. But does Rockbox plan to support the iRiver H10 models?
05:41:00 Quit cYmen ("zZz")
05:43:28fuzziei don't believe anyone has done any work on them
05:44:02fuzziethey're completely different from h1x0/h3x0.
05:44:25BirdFishYes :(
05:44:29BirdFishI was foolish enough to buy one.
05:44:43MirthoneistIs it bad?
05:44:47BirdFishI didn't know what I was getting into when I went with iRiver until after I bought it.
05:44:57BirdFishIt's not (intuitive)
05:45:08BirdFishNo Ogg support, no FLAC, no anything
05:45:11MirthoneistOh
05:45:15MirthoneistAhh, I see
05:45:18BirdFishiRiver really messed up when they made this model
05:45:33BirdFishAnd I've contacted managers and they don't plan to incorporate
05:45:35Mirthoneist:(
05:45:41BirdFishFux iRiver for going mainstream
05:45:48Mirthoneist:\
05:45:51BirdFishWhile I'm here
05:46:07MirthoneistYeah, as soon as the H10 came out, I had to choose between H300 or H10
05:46:13Mirthoneistwent with the H340 in the end
05:46:19BirdFishHow is evrything going with the h3xx models? Because I might sell my H10 and try to find one of those.
05:47:25MirthoneistAs far as rockbox, it
05:47:29Mirthoneistit's coming along I suppose*
05:47:44BirdFishiRiver firmware is fine on it though?
05:47:53MirthoneistUS firmware, no
05:48:08BirdFishYeah, I would DEFINITELY go with the Intl version this time as well.
05:48:11BirdFishI'd already decided that
05:48:12MirthoneistBut you can update to Korean firmware the same way you update US firmware
05:48:18Mirthoneistyeah
05:48:19BirdFishBut I was checking out the new iaudio x5 also
05:48:23BirdFishHas FLAC built into it
05:48:25Mirthoneisthas USB OTG I think with Int'l
05:48:26BirdFishVERY nice
05:48:26BirdFish:D
05:48:50BirdFishDoesn't updating firmware on a not so compliant device slow it down any?
05:48:50MirthoneistI didn't really like how the iAudio X5 was setup
05:48:53Mirthoneistlike, the buttons
05:49:01BirdFishIt does with the H10 I know at least
05:49:27MirthoneistAs in like, updating a diff. languaged firmware?
05:49:52BirdFishI'm also pretty pissed that iRiver is letting it's 20 gig H10 users down. All the updates are for the 10 gig users, and all the forums discuss the 10 gig rather than the 20. Basically, iRiver really screwed up this one.
05:49:55BirdFishHeh, sorry about the rant :P
05:50:01BirdFishNo
05:50:05BirdFishJust updating hte firmware
05:50:16MirthoneistI hear ya :)
05:50:22BirdFishIf you go from MTP to UMS on the H10 model, the menus run slower
05:50:28MirthoneistOh
05:50:40MirthoneistThe menus run the same on different formwares for the H3XX
05:50:45BirdFishThat's cool
05:50:46Mirthoneistfirmwares*
05:50:50Mirthoneistyeah
05:50:57MirthoneistIt's a neat device
05:51:06MirthoneistOnly problem I have it with it playing music
05:51:09BirdFishFrom what I've read about it, I think so too
05:51:29BirdFishWhat about it playing music?
05:51:35MirthoneistLike...
05:51:44MirthoneistWell, no Queing, or On the Fly playlists first off
05:51:47Mirthoneistbut...
05:51:57BirdFishRockbox will solve :D Woot!
05:52:04Mirthoneistif you're playing a playlist and you turn off your player, you have to restart the playlist
05:52:10BirdFish:/
05:52:15MirthoneistYeah, this is why I can't wait for Rockbox
05:52:17BirdFishCan you scroll through the playlist?
05:52:31MirthoneistHrm...
05:52:39MirthoneistI actually don't even remember o_O
05:52:50BirdFishThat would seriously be a bitch if it just read the file from start to finish every time.
05:52:55BirdFishI'd go insane
05:53:02MirthoneistI know :D
05:53:04BirdFish:)
05:53:11BirdFishIf you simple browse for files...
05:53:24MirthoneistI think it does, and since I use random, I had to unrandom, skips a few songsm, then random again to get a different song
05:53:24BirdFishand open a folder for an artist/album
05:53:35BirdFishWill it play all the songs in the folder, or just the one that you select?
05:54:01MirthoneistWell, that's another thing that I haven't really figured out...
05:54:02MirthoneistLike:
05:54:13MirthoneistYou can go to Artist, and it will show the albums
05:54:21BirdFishSo wait
05:54:25Mirthoneistbut you can't select to play like "All of that artist
05:54:32BirdFishThere is a full fledged menu structure?
05:54:38Mirthoneistyou have to go play all the album
05:54:41BirdFishOr do you mean after you create your own?
05:54:52BirdFishBecause I was under the impression that the menus were as you laid them out yourself on the hd
05:55:11MirthoneistNot really ;\
05:55:14BirdFishAh
05:55:16MirthoneistLike, you can create folder and stuff
05:55:20Mirthoneistfolderrs*
05:55:32MirthoneistI haven't experimented really with doing that
05:55:38BirdFishI see
05:55:39MirthoneistI guess you can sort the music in folders
05:55:42Mirthoneistand play my folders
05:55:57MirthoneistThat's how I play all my songs (I put all my songs in 1 folder, instead of using a playlist)
05:56:02BirdFishDoes it create it's menu from the id3 tags?
05:56:03MirthoneistSo yeah, you can customize it
05:56:15MirthoneistWell, you can turn "DB function" on
05:56:22BirdFishGotcha
05:56:25MirthoneistWhich let's you choose by Genre, artist, etc
05:56:32Mirthoneistso I guess that works with ID3 tags
05:56:48BirdFishSo many things to consider
05:57:16BirdFishI'm going to punch the bastard at Best Buy that told me that the H10 had ogg support that could be downloaded
05:57:24BirdFish:P
05:57:31Mirthoneistlol
05:57:43MirthoneistBest Buy is usually good with returns
05:57:50Mirthoneistbut then again, it's not like you can exchange it
05:57:50BirdFishIt's been far too long
05:57:53Mirthoneist:\
05:57:55Mirthoneistoh
05:57:55BirdFishyeah
05:58:13BirdFishI thought maybe since iRiver was slacking in the 20Gig compartment, that it just wasn't out yet :o
05:58:13MirthoneistHehe
05:58:15BirdFishBig mistake ;)
05:58:24Mirthoneist:(
05:58:34MirthoneistYeah, my friend is highly considering getting the H10
05:58:35BirdFishOh well, live and learn :D
05:58:42MirthoneistI told him I'd personally get the H3XX
05:58:45BirdFishTell him to highly consider not doing it
05:58:53BirdFishIt'd be the worst decision he could make
05:59:01MirthoneistBut then again, he just really wants to play music, and I thought maybe the H10 would be better for him
05:59:06Mirthoneisthehe
05:59:14MirthoneistHe just has Mp3s
05:59:19Mirthoneistso like, I don't know...
05:59:19BirdFishIf he likes mp3s and wma's it's cool I guess
05:59:24Mirthoneistyeah
05:59:33Mirthoneistthe H3XX is tricky to navigate too
05:59:36BirdFishAnd if he's not exactly computer literate, mpt firmware would be good for him
05:59:52MirthoneistHe's computer litterate
05:59:58Mirthoneisthe wants drag and drop
06:00
06:00:05BirdFishYou can do that
06:00:11BirdFishIn either MTP or UMS mode
06:00:22BirdFishUMS is really the way to go
06:00:23MirthoneistOh, that's cool :)
06:00:38MirthoneistMTP, don't you need Windows Media Player to drag and drop?
06:01:02BirdFishBut if he wants to fully convert the device to UMS, he absolutely needs to get the 5/6 gig version, because 20 gig version doesn't have a UMS firmware that can be loaded yet.
06:01:37BirdFishBut with 20 gig, I reset > connect to computer > then power on the device while holding the select key, and it boots UMS for that period of time
06:01:38Mirthoneistoh, I seeee
06:01:40BirdFishThen it goes back to mtp
06:01:48BirdFishWith UMS, he needs to get EasyH10 though
06:01:54BirdFishAnd that'll build his database for him
06:02:08fuzziethat all sounds very horrible
06:02:09BirdFishOtherwise, he'd better have a sweet browser structure
06:02:16BirdFishIt is very horrible fuzzie
06:02:22BirdFishThat's why I'm ranting
06:02:22BirdFish:P
06:02:32BirdFishI'm pissed with iRiver right now
06:02:54BirdFishAlmost to the point of getting an iPoo :/
06:03:04BirdFishAnd I hate those damn "in" devices
06:03:05Mirthoneist:o!
06:03:07Mirthoneistlol
06:03:08Mirthoneistyeah
06:04:19BirdFishLOL, can you tell Calculus homework was a bitch tonight?
06:04:25BirdFish:P
06:04:45Mirthoneistlolz
06:04:55BirdFish50+ problems a night is a little overkill ;)
06:05:11BirdFishI tire of doing the same thing repetatively
06:05:12Mirthoneistgheesh
06:05:19MirthoneistI hear ya
06:05:30MirthoneistAlthough, Im only in Precalc
06:05:35Mirthoneist:p
06:05:39BirdFishJunior in highschool?
06:05:48MirthoneistSophmore in College :o
06:05:50Mirthoneistlol
06:05:52BirdFish:o
06:05:59MirthoneistI procrastinated with taking it
06:06:00BirdFishSenior in highschool :*)
06:06:10Mirthoneist:o
06:06:22BirdFishI'm big into computers though
06:06:32MirthoneistI am, kinda
06:06:35BirdFishSo I guess it fits with my goals in life
06:06:39BirdFishYou might not need Calc
06:06:46BirdFishI wouldn't take it if I didn't need it
06:06:50BirdFishFor damn sure
06:06:55MirthoneistI hope I dont need it :p
06:06:58BirdFish:P
06:07:02BirdFishBusiness Major?
06:08:00MirthoneistThat's what Im under now
06:08:11MirthoneistCause my damn Accounting teacher put me as it
06:08:20MirthoneistBut, I want to do something with computers
06:08:21BirdFishSo you don't like it I take?
06:08:26BirdFishThat's cool
06:08:28MirthoneistCorrect :)
06:08:29Mirthoneistlol
06:08:35BirdFishHardware/Software?
06:08:44MirthoneistI wanted to be a programmer :(
06:08:47Mirthoneisthehe
06:08:47BirdFish:D
06:09:02BirdFishTaking Fundamental Class this term
06:09:03MirthoneistAfter I took Programming 1, they cancelled all Programming 2 cvlasses, so that kinda failed
06:09:05Mirthoneist;\
06:09:12BirdFishThat sucks
06:09:15Mirthoneistyeahz
06:09:19BirdFishChange colleges man
06:09:22BirdFishI definitely would
06:09:34MirthoneistI would if I could :)
06:09:39BirdFish:)
06:09:43MirthoneistI have a fully paid scholarship, with a catch
06:09:51BirdFishWhat's that?
06:09:59MirthoneistHave to do 2 years of community college there, then I can go anywhere in Florida
06:10:06BirdFishWow
06:10:10Mirthoneistfor my las 2 years
06:10:12Mirthoneistyeah
06:10:12BirdFishHow did you manage that?
06:10:21MirthoneistIm not sure o_O
06:10:29BirdFisho_0
06:10:30MirthoneistI guess I did good in high school
06:10:38MirthoneistDidn't get into trouble, 3.5GPA
06:10:42BirdFishI'm really hoping to get a full ride somewhere
06:10:51MirthoneistYeah it's great
06:10:57BirdFish4th in rank of 200+ kids
06:10:59Mirthoneisttheres a lot of opportunities
06:11:04Mirthoneistnice!
06:11:05BirdFish4.5 GPA (I think)
06:11:11Mirthoneistwow :o!
06:11:25BirdFishI was #1 for my first 2 1/2 years
06:11:34MirthoneistGheesh
06:11:34BirdFishThen I got into girls and shit went down
06:11:38BirdFish:'(
06:11:40BirdFishFux girls
06:11:43BirdFish:P
06:11:50BirdFishI need to get my head outta my @ss
06:11:52Mirthoneistlol!
06:11:54MirthoneistI hear ya
06:11:58MirthoneistHaha :D
06:12:33BirdFishSo do you do any web designing or anything?
06:13:11MirthoneistI used to a little bit
06:13:18MirthoneistI'm not graphically artistic
06:13:30Mirthoneistso although I liked doing that stuff, it just never looked good
06:13:32BirdFishIt's cool. I'm about half/half really
06:13:43MirthoneistYou do your own websites?
06:13:45BirdFishI like to keep my stuff simplistic
06:13:49BirdFishYep
06:13:49MirthoneistHehe
06:13:51MirthoneistThat's cool
06:14:00MirthoneistThe simplistic pages are always the best anyways
06:14:03BirdFishI don't have my current site hosted right now though :/
06:14:04MirthoneistYou use HTML?
06:14:08BirdFishxhtml
06:14:10Mirthoneistohz
06:14:12BirdFish1.0 strict
06:14:16Mirthoneistoh my :o
06:14:20BirdFishWish it was xml
06:14:20BirdFish:(
06:14:26BirdFishBut that's beyond me for now
06:14:31MirthoneistHehe
06:14:38MirthoneistI've never dealt with XHTML
06:14:47BirdFishI definitely don't have the time to research it, and I've been told by my programming instructor that it's not exactly easy to pick up.
06:14:56BirdFishIt's not much different from html
06:15:00MirthoneistOh
06:15:15BirdFishReally, just keep tags in lower case, and instead of doing a </br> it's not <br />
06:15:17BirdFishStuff like that
06:15:22BirdFishMinor changes
06:15:28MirthoneistAh,, I see
06:15:32MirthoneistYeah, I went from HTML, to Frontpage ( :( ) to Dreamweaver
06:15:40BirdFishMeh
06:15:41Mirthoneistmight get back to HTML again
06:15:45fuzzieBirdFish: assuming your html was valid in the first place..
06:16:13BirdFishfuzzie: yes, a lot is assumed. I was just pointing out basics...
06:16:20***Saving seen data "./dancer.seen"
06:16:30BirdFishMirthoneist: Frontpage and Dreamweaver both do HTML, just more visually
06:16:40MirthoneistYeah
06:16:43BirdFishYour best bet is to do everything by hand (with a text editor)
06:16:45BirdFish:D
06:16:47BirdFishTotally leet
06:16:51MirthoneistYeah, that's what I used to do
06:16:53Mirthoneistlol
06:17:12MirthoneistI thought I could be lazy and do it in a WYSIWYG program
06:17:19BirdFishI just happen to use one now that highlights syntax and auto indents lines for me so that I can keep things cleaner easier
06:17:21Mirthoneistbut, they just aren't the same
06:17:27BirdFishDefinitely not
06:17:29MirthoneistAh, that's cool
06:17:30BirdFishYou lose control
06:17:34MirthoneistYeah
06:18:18BirdFishAnd then you forget the tags and when you go to impliment CSS into ASP.NET, you can't (heh, my father)
06:18:28Mirthoneistlol
06:18:28BirdFish:D
06:19:06MirthoneistYour dad is a web designer too?
06:19:08BirdFishEventually, probably next summer I want to get back into PHP
06:19:12BirdFishHell NO!
06:19:16Mirthoneistoh, lol
06:19:22BirdFishHe's a self-taught programmer
06:19:22 Quit Maxime` (Read error: 104 (Connection reset by peer))
06:19:28BirdFishThat writes software out of the house
06:19:29Mirthoneistoh my :o
06:19:36MirthoneistVery neat :)
06:19:37 Join Maxime` [0] (n=flemmard@82.229.205.156)
06:19:37BirdFish(retired police captain)
06:19:59MirthoneistDamn, I wish my dad could do that
06:20:07MirthoneistHe can barely find "My Computer"
06:20:09Mirthoneist:\
06:20:25BirdFishBut he doesn't feel the need to learn html syntax or anything and thus he has a hellova time trying to impliment CSS into anything (since he's porting his program to the web)
06:20:41MirthoneistOh, I see
06:20:51BirdFishA lot of the database stuff he has a guy in florida do
06:21:06BirdFishSo, I guess he kind of learns as he goes with it
06:21:16MirthoneistOh, that's neat
06:21:17BirdFishHe's by no means a professional programming :P
06:21:26Mirthoneistlol ^_^
06:21:34BirdFishBut he is the new techie at one of the local law firms
06:21:35BirdFishwoot woot
06:21:36BirdFishLOL
06:21:48MirthoneistHaha, that's sweet
06:21:54BirdFishHehe, yeah
06:21:55MirthoneistThat's a job I think I'd enjoy
06:22:26BirdFishYeah, it's kind of cool. Pain in the butt sometimes (I work on the puters with him occassionally)
06:22:34BirdFishBut check it out before you definitely decide anything
06:22:43MirthoneistYeah
06:22:46BirdFishI don't think it pays as nicely as one might hope :(
06:22:50MirthoneistI'll have to try a lot of different things
06:22:59MirthoneistAh, I don't care much about money
06:23:00BirdFishGet into something that deals with unix/linux systems
06:23:04Mirthoneistas long as I get a job I enjoy
06:23:06BirdFishYeah
06:23:08BirdFish:)
06:23:14MirthoneistI've never dealt with linux :(
06:23:32BirdFishThat's where most of the server techie stuff comes from
06:23:41BirdFishIt's cool
06:23:43MirthoneistOh
06:23:54BirdFishIf you like working through a command line (like me) you'll love it
06:24:03BirdFishBut then I seem to like doing things the hard way I guess
06:24:04Mirthoneisthehe ^_^
06:24:09Mirthoneistlol
06:24:19BirdFishfor instance, sometimes I telnet to IRC :P
06:24:21BirdFishLOL
06:24:29Mirthoneistlol!
06:25:47MirthoneistWell, Im gonna head out to bed
06:25:55MirthoneistMaybe I'll see ya around BirdFish
06:25:57BirdFish:) I was just about to say the same thing
06:26:01BirdFishYou'll definitely see me around
06:26:02BirdFish:)
06:26:02Mirthoneistlol :D
06:26:05BirdFishLater bud
06:26:07MirthoneistHehe, okiez
06:26:09MirthoneistG'Night
06:26:12BirdFishnight
06:26:20 Quit Mirthoneist ()
06:54:15 Quit rasher (Remote closed the connection)
06:58:13 Join amiconn_ [0] (n=jens@p54BD6985.dip.t-dialin.net)
07:00
07:03:21 Quit elinenbe (Read error: 104 (Connection reset by peer))
07:03:47 Join elinenbe [0] (i=elinenbe@207-237-225-9.c3-0.nyr-ubr1.nyr.ny.cable.rcn.com)
07:07:24 Join LinusN [0] (n=linus@labb.contactor.se)
07:11:31 Join Lynx_ [0] (n=lynx@tina-10-4.genetik.uni-koeln.de)
07:14:57 Join Paul_The_Nerd [0] (n=paulthen@cpe-66-68-93-2.austin.res.rr.com)
07:16:22 Quit amiconn (Read error: 110 (Connection timed out))
07:16:23 Nick amiconn_ is now known as amiconn (n=jens@p54BD6985.dip.t-dialin.net)
07:28:07 Quit Lynx_awy (Read error: 110 (Connection timed out))
07:28:08 Nick Lynx_ is now known as Lynx_awy (n=lynx@tina-10-4.genetik.uni-koeln.de)
07:35:53amiconnmorning
07:40:29LinusNmorning
07:40:40LinusNi don't like the reports we're getting about the v1 charging
07:42:02amiconnMe neither
07:42:22amiconnThe charging is definitely better than what we had before, but isn't perfect
07:42:32amiconnI suspected that we might get such reports
07:42:44amiconnI have slight problems myself
07:43:04amiconnNot as bad as in the reports, but it seems my rec v1 doesn
07:43:17LinusNi haven't used my rec20 much since i got my iriver, but i do get the sense that it doesn't charge them as much as the old algorithm
07:43:24amiconn't achieve the runtime that should be possible with 2500 mAh cells
07:43:56amiconnThe cells certainly aren't worn out
07:44:33amiconnYour comments regarding Ondio & resume are slightly off.
07:45:05amiconnWe use BUTTON_MENU in the code, but the "official" button name used elsewhere is MODE
07:45:20SlasheriLinusN: hi :) what do you think about the dircache: http://ihme.org/~miipekk/rockbox/dircache.diff
07:45:21LinusNsaw that
07:46:53Slasheri(i know there are two possible small bugs at the moment i will fix before the commit, if the code in other way looks good)
07:48:03LinusNSlasheri: a dircache file???
07:48:39LinusNi thought we were discussing a dynamic cache?
07:51:02Slasheriof course it's dynamic, but that file is for fast bootup without rebuilding the cache (that should be ok because the file tree browser is the only place that is using the cache and even if cache would be out-of-date there should be no risk of data corruption/lost)
07:51:13Slasheriand that is an option
07:51:30LinusNwhat's wrong with having it purely dynamic?
07:51:43Slasheriif user know he is going to use usb bootloader mode, the he should not skip the rebuilding at boot
07:51:56LinusNi certainly don't want to answer thousand of "it won't find my files" questions in the forums
07:51:58SlasheriLinusN: and that file is also using to calculate the space needed for the cache
07:52:06SlasheriHmm..
07:53:18LinusNand now long does it take to scan the drive when leaving usb mode?
07:53:33LinusNs/now/how/
07:53:49Slasherithat depends how much there are files but that scan is transparent to user
07:54:07Slasheriso user doesn't have to wait
07:59:22 Join koniu_ [0] (n=koniu@host-ip194-145.crowley.pl)
07:59:55LinusNit doesn't seem to handle usb mode while building the cache
08:00
08:00:48Slasherithat's possible
08:01:05Slasherido you think it should be fixed?
08:01:12LinusNof course
08:01:15Slasheriok :)
08:01:36LinusNif it takes long time to rebuild, it can be frustrating when it doesn't enter usb mode again
08:01:54LinusNfrankly, i don't like this concept at all, but that's me
08:01:59Slasheriah, yes
08:02:20LinusNhow does it reserve cache space?
08:02:40Slasherii understand that, but i think some people may like it
08:03:15SlasheriLinusN: when it does the first non-transparent cache build, it allocates the space used for cache + 64 KiB for reserve space
08:03:30LinusNand if it runs out of space?
08:03:47Slasherithen the cache is disabled
08:03:53LinusNuntil when?
08:03:58Slasheriand it will be rebuild on next boot
08:04:39LinusNok, so it's all or nothing
08:04:54Slasheribasically yes
08:05:22dwihno\o/
08:05:34Slasherihmm. in fact it might be possible to use the cache only partially but that adds more complexity to the caching code
08:06:37 Join ender` [0] (i=ychat@84.52.165.220)
08:07:32LinusNok, so how big will the cache buffer be initially, when there is no cache file?
08:08:30Slasherithat depends entirely about the number of files and filename length. On my player (with thousands of files, pretty full player) it's around 400 KiB
08:09:11LinusNno, i mean when calling dircache_build() for the very first time
08:10:07SlasheriHmm? It will build the cache and the buffer size really depends how much there are files :)
08:10:22LinusNso it can'
08:10:23Slasherithe first build is non-transparent
08:10:30LinusNok, so no resume
08:10:32Slasheriso it doesn't have to guess any buffer size
08:11:37LinusNand it will also disable it if the cache grows > DIRCACHE_RESERVE?
08:11:48Slasheriyes of course
08:12:57Slasheriit cannot exceed the reserved space or the system would crash or something else like that
08:14:49BirdFishVery cool idea
08:16:21***Saving seen data "./dancer.seen"
08:19:55LinusNSlasheri: i see a lot of situations where dircache_new_entry(9 would fail
08:20:40LinusNbut you don't handle it in for example dircache_mkdir()
08:20:57SlasheriHmm, that is handled by the dircache_new_entry() itself
08:21:08Slasheriit will disable the cache: dircache_initialized = false
08:21:13LinusNok
08:22:41LinusNwhat happens if i create a dir when dircache_do_rebuild() is running?
08:23:46Slasherithat's not a good idea. There is a chance that it will be missing from the cache when the rebuild is complete
08:23:59SlasheriHmm.. But i think it's possible to even fix that
08:24:02LinusNso the user will have to wait until the cache is rebuilt
08:24:12Slashericurrently yes
08:24:16LinusNhow does he know that it is rebuilt?
08:24:35LinusNi guess the same goes for files too?
08:24:45Slasheriwhen the disk activity stops.. but that's a good point, i will try to fix it
08:24:48Slasheriyes
08:25:29LinusNalso, will the cache file itself end up in the cache on the first rebuild?
08:25:52Slasheriit's just very rare that first thing user would do is to start creating files after booting up the unit. But indeed that is possible
08:26:03SlasheriLinusN: no
08:26:16LinusNrare or not, it is a source of error
08:26:21Slasheriyes
08:26:31LinusNtake recording for example
08:26:38LinusNleave usb and start recording
08:27:07LinusNit will create both files and directories
08:27:18LinusNand suck cpu power from the recording thread
08:27:44Slasheritrue.. i will fix that soon
08:28:18LinusNmy point is, we will probably find more issues like this
08:28:28Slasheri(it should be even quite easy, just using semaphores to lock and check for duplicate entries)
08:28:32Slasherihmm..
08:28:57Slasheriit's possible
08:29:14LinusNi don't have a good solution though
08:29:31LinusNif you want the entire tree in ram, this is where you end up
08:29:46LinusNor take the easy way out like iriver did
08:29:57LinusNannoying the users in the process
08:30:04Slasheribut eventually all the problems will be solved
08:30:05*LinusN gets some coffee
08:30:19LinusNnot all
08:30:22Slasheriheh, but my entire point was i don't want to annoy users
08:30:22Slasheri:)
08:30:33amiconnAnother point: What happens if I delete the cache file at runtime?
08:30:36LinusNi mean that iriver annoys users
08:30:45LinusNamiconn: nothing i guess
08:31:07Slasheriyep, that cache file is not needed at runtime
08:31:17LinusNthere is one problem that won't be solved, and that's the booot loader usb mode
08:32:27Slasheritrue.. then we just have to give some control to the user if he/she really wants to skip the cache rebuilding at boot
08:33:18Slasheriand i think that is acceptable if the user know that he is not going to use the bootloader usb mode
08:33:28amiconnRelated: What happens if writing the cache file fails because of disk full?
08:33:30LinusNthis is where the forums will get swamped with problem reports
08:33:50Slasherior another possible way would be to delete the cache file after leaving bootloader usb mode (but that requires modification to the bootloader)
08:33:57 Quit Paul_The_Nerd ("Chatzilla 0.9.68a [Firefox 1.0.7/20050915]")
08:35:23LinusNamiconn: i think the only thing that will happen is that the cache will have to be rebuilt the next boot
08:35:33LinusNthe cache file is only saved at shutdown
08:36:01ashridahwhat're you caching? the directory map?
08:36:19Slasheriashridah: yes, the directory tree
08:36:34amiconnI wonder why there is a file at all, if the cache rebuild can be done transparently
08:36:55LinusNamiconn: it isn't transparent at boot
08:37:17amiconnYes, but it can be made transparent, I think
08:37:20LinusNand it makes the initial caching faster
08:37:22Slasheri(it's if the cache file exist and user has selected not to skip the rebuild)
08:37:34Slasherithen the file header is only used for allocating the required space
08:37:38amiconn...and the needed RAM size could be stored in the settings structure
08:38:00Slasheriamiconn: no, that is saved on the cache file header
08:38:25amiconn(without connecting it to .cfg files - just the config sector entry the same way as the battery runtime data is stored)
08:38:46amiconnSlasheri: Yes I know, I say it *could* be stored there, for a fully dynamic cache
08:38:55amiconn..that doesn't use a file
08:38:58Slasheriamiconn: ah, yes
08:39:08Slasheri:)
08:39:15LinusNi think always rebuilding it is a waste of battery
08:39:58Slasheriit is.. but currently that is the only possible way if user uses the usb bootloader mode often
08:40:00amiconn...but not doing it creates inconsistencies, so we have no choice I'd think
08:40:35LinusNmy disk contents are mostly unaltered
08:40:45amiconnThere's one small advantage of caching I can see for myself, and that's recording from the microphone
08:40:45LinusNwhen i use it to play music
08:41:09amiconnLinusN: Yes, but rockbox can never know what happened between 2 rockbox runs
08:41:12LinusNso scanning in the background every time i boot is a waste of battery, then the cache file is much better
08:41:20amiconn(bootloader USB mode, iriver firmware...
08:41:25LinusNamiconn: i know
08:42:11amiconnWe do have delayed file open for recording, so recording from mic is undisturbed by disk activity until the buffer is filled the first time.
08:42:13LinusNthe cache file is basically unnecessary, but is a bettery saver
08:42:57amiconnHowever, that currently doesn't help on iriver because the recording file name is created at start of recording, and for non-RTC units it accesses the directory
08:43:07amiconn...to find the next free numbered filename
08:43:34amiconnThat's a non-issue on Ondio because disk access creates no noise, but it will be an issue on H1x0
08:44:23amiconnLinusN: I know, but it's also a source of inconsistency. If we could trust the FAT32 modification time stamp... but we can't
08:45:16LinusNi think the cache file inconsistency will be a huge source of error
08:45:22LinusNand confusion
08:45:44LinusNi mean, people tend to ignore simple instructions, and they don't read the manual
08:45:46amiconnI agree, as it will also happen with USB mode
08:45:49 Part koniu_
08:46:04LinusNamiconn: the cache is always rebuilt after Rockbox usb mode
08:46:12LinusNbut not after bootloader usb
08:47:26amiconnYes, and that willl lead to problems
08:48:31amiconnI thought about a (perhaps) simpler, but less resource-hungry (and less complete) caching system
08:48:54amiconnIt would never cache beforehand or across reboots, and only work for the browser
08:49:38amiconnIt would simply cache as many dirs in the browser's dir buffer as fit in there.
08:50:12amiconnIf a dir is opened that doesn't fit, it would start throwing away old dirs (lru)
08:50:22LinusNfragmentation?
08:50:30amiconnNone.
08:50:46amiconnIt would move all entries together after removing enough
08:50:53amiconn(memmove)
08:52:00amiconnThat also greatly simplifies lru - the first entry is always the oldest
08:52:44LinusNok, so increasing the dir buffer size will automatically increase the cache size
08:53:38amiconnyup
08:53:57amiconnMany small dirs would benefit from one large dir the user has :)
08:53:59LinusNthat method will force more reloads than slasheris approach
08:54:30amiconnYes, but it still saves reloads compared to no caching at all
08:54:42amiconn...and without additional cache RAM
08:55:43amiconn...and it *might* even save battery as it only loads the dirs the user is actually browsing
08:56:07LinusNnot necessarily
08:56:21Slasheriamiconn: but how can it know before hand what dirs user is going to browse and without annoying the user? :)
08:56:23amiconnThe current code loads the same dir over and over, and Slasheri's full cache system requires rebuilds
08:56:44LinusNamiconn: you mean a dir is only cached when visited?
08:56:52amiconnyes
08:57:07LinusNi don't think that is much of a benefit
08:57:17amiconnSlasheri: We can't for a not-yet-visited dir
08:57:28amiconnLinusN: why?
08:57:36LinusNwhen i look for a file to play, i rarely re-visit the dirs
08:57:45Slasheriamiconn: Hmm, yes.. And that was the entire idea of the caching
08:58:23amiconnLinusN: Don't you enter a dir, realise that it isn't what you want, then back out sometimes?
08:58:43LinusNyes, we would save the backout reload
08:58:51 Join einhirn [0] (i=Miranda@bsod.rz.tu-clausthal.de)
08:58:53amiconnSlasheri: My suggestion is somewhat in-between no caching at all and a full cache
08:58:54LinusNbut it will always load the next dir i enter
08:59:02amiconnyes
08:59:08LinusNso there will be a lot of spinups
08:59:25LinusNeven more, as the disk will spin down as i enter the cached parent dir
08:59:42LinusNso the browsing experience might even be more painful
09:00
09:00:38 Quit Vladoman (Read error: 110 (Connection timed out))
09:00:46amiconnDepends on how long a timeout you have set
09:00:58amiconn...and how fast you are browsing
09:01:14 Join Vladoman [0] (n=Vladoman@p54A7FC2D.dip.t-dialin.net)
09:01:27LinusNamiconn: of course
09:01:47amiconnFor me, it sometimes even happens that the disk spins down before I back out, and has to spin up again for just reading the dir I already visited
09:01:55LinusNbut in general, i don't think a "lazy" cache will add much to the browsing experience
09:02:01amiconnPerhaps I should set a higher spindown timeout...
09:02:48amiconnLinusN: Maybe. I already said it's somewhat in-between, but it doesn't require additional RAM
09:03:04amiconn...so it could work on archos
09:03:05LinusNit's certainly more KISS
09:03:36LinusNbut not nearly as effective as slasheris solution
09:04:16amiconn...or if I don't want to enable full caching because I don't want to set aside the additional RAM
09:04:25 Join B4gder [0] (n=daniel@static-213-115-255-230.sme.bredbandsbolaget.se)
09:04:38amiconnSlasheri: Is there a way to see how much RAM the cache takes? (debug menu?)
09:05:32 Quit Rori ()
09:07:14Slasheriamiconn: currently it puts that information through logf only but i think a debug menu entry would be a good
09:07:23amiconnWith such a lazy cache, it should be possible to back out up to the root from a deeply nested dir without a reload, provided the dir buffer is large enough
09:07:30Slasheriit could then show also the reserved buffer used
09:07:36amiconnyes
09:14:31amiconnHmm.
09:14:40 Nick Lynx_awy is now known as Lynx_ (n=lynx@tina-10-4.genetik.uni-koeln.de)
09:17:51amiconnSlasheri: On FAT32 you could perhaps check fsinfo on boot. If either freecount or nextfree have changed (I know windows is lazy), the disk was changed and a rebuild is needed
09:22:20Slasheriamiconn: oh, that sounds good
09:23:16Slasheriis it possible that windows doesn't update those entries even if i add some files?
09:24:40B4gderI'd guess so
09:31:26LinusNnot only possible, it's standard procedure
09:31:48LinusNwindows is known to not update fsinfo
09:36:57 Join XavierGr [0] (n=XavierGr@ppp45-adsl-28.ath.forthnet.gr)
10:00
10:13:32 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
10:13:53HCl#^#$5~#$
10:14:01*HCl gahs at internet explorer
10:14:19HClit had the nerve to save my creditcard number in its autocomplete >.<
10:14:44LinusNand how was it supposed to know it was a credit card number?
10:16:25***Saving seen data "./dancer.seen"
10:18:19XavierGrwhat autocomplete your credit number?
10:18:41XavierGrclear the cache
10:18:51amiconnLinusN: windows doesn't even update fsinfo.nextfree?
10:19:12XavierGrOh Linus is here.
10:19:34XavierGrLinus I am just programming the radio for multiple preset file support
10:20:54XavierGrI have a question though. How rockbox knows which is the latest WPS that the user has selected (same for font I guess)?
10:21:25XavierGrBecause I want to do the same with the fmr files. I want to remember the latest opened.
10:22:28Zagoramiconn: smetimes, sometimes not.
10:23:40amiconnZagor: Depending on the moon phase or somesuch? ;)
10:23:53VladomanNo, XP does not update FSINFO any more
10:24:00Vladoman98 and ME did
10:24:06Vladoman2000 i'm not sure
10:24:23Vladomanoh, hi guys
10:25:05B4gder:-)
10:26:17Zagoramiconn: yeah, probably. however nextfree is defined as hint where to start searching. it doesn't actually need to point to the next free block. freecount, however, is a blatant spec violation.
10:26:25LinusNXavierGr: check settings.c
10:26:37Vladomanwell, they made it, they break it :-)
10:26:47Zagoryup
10:28:39XavierGrRockbox code is so beautifully written. Functions like kbd_input and create_numbered_filename are sweet!!!
10:29:21XavierGrHmm so is it alright to use another one space for a fmr file?
10:29:54LinusNi guess so, if there's room for it
10:31:33XavierGrplenty!
10:31:48amiconnLinusN: Speaking about it: Is there any reason why the config block (on disk uses only half of the sector?
10:32:27amiconnA sector is 512 bytes, however, config block is only defined up to 0xFF
10:32:31XavierGrhmm not so plenty my entry can fit. but thats it.
10:32:40amiconnXavierGr: It can, easily
10:32:56XavierGryou can make it larger?
10:33:36XavierGrif someone wants to put another one then he should have a problem/
10:34:47XavierGrI got confused.
10:35:12XavierGrup to know it is used up to f4.
10:35:38XavierGrbut f4 + 20(dec) = 108(hex)
10:36:07XavierGrso it cant fit right?
10:36:58amiconnXavierGr: There are two options
10:37:21amiconn(1) You can extend it at the end should be possible, read my remark at 10:32:37
10:37:55amiconn(2) You can put it before the WPS file name, at 0xB8 - 20 = 0xA4
10:38:25XavierGrwhat do you prefer?
10:38:38amiconnThe current HD bit table only uses 31 bytes atm (worst case, so the end address of that is 0x2C + 31 = 0x4B
10:38:47amiconnPlenty of room before 0xA4...
10:39:55XavierGrok then I should do the 2nd. Thanks/
10:40:49amiconnZagor: I wonder how windows derives the free disk space without using freecount. I'd guess they scan the whole FAT.
10:41:15VladomanYes, they scan it at each connection, that's why it is so slow wiht usb 1.1
10:41:31Vladomanfor the new drive to come up
10:42:14amiconnSounds like a windows-like method ;/
10:42:47VladomanYes, if you see the actual USB MSC transactions you wonder who wrote the driver
10:43:01VladomanThey ask for disk capacity about 5 times
10:43:09Vladomanjsut to make sure, I guess :-)
10:43:26amiconnPerhaps someone connects a dynamical disk? ;)
10:43:47LinusNamiconn: in the days of yore, the entire path to the resume dir/playlist was stored in the config sector
10:43:52VladomanYeah, or the physical connection is not yet burned in ...
10:44:56amiconnI wonder what happens if someone constructs a 'mass storage device' that deliberately sends different answers for each request...
10:45:28 Join tucoz [0] (n=martin@hornved.ii.uib.no)
10:45:47LinusNamiconn: and Jörg mistakenly adjusted the comment so that it seemed that the max is 0xff
10:46:11tucozLinusN, hi, I read somewhere that you might be interested in a change in the fm-preset format.
10:46:15LinusNwhen he revamped the bit allocation scheme for the settings
10:46:33Vladomanamiconn: I could try that on a JBM020 :-)
10:46:55LinusNtucoz: yes and no
10:47:36tucozOne question though. If the format were like 95.6:descr, would we have to handle the old format as well?
10:48:00LinusNprobably not
10:48:54amiconnWhen fiddling with FM anyway, it might make sense to allow 0.05 MHz steps
10:48:59tucozyes
10:49:00amiconnSome people asked for that
10:49:22tucozI think so too, I saw the stepsize is set to 0.1 Mhz now
10:49:44amiconn..although in Germany there are no xx.x5 frequencies used by terrestrial stations
10:49:49tucozBut I do not know the sensitivity of the fm-chips in the different players.
10:50:13amiconn...only broadband cable uses that
10:50:23XavierGrwhy the code utilizes the "lets say 9100000" format for the station 91.0?
10:50:29tucozHz
10:51:02tucoz91.0 Mhz = 91x1000000 = 91000000
10:51:08tucozHz
10:51:19XavierGryes but is this necessary?
10:51:33XavierGrI dont think that the chip has hz detail.
10:51:38amiconnIt's the simplest solution
10:51:46XavierGroh then ok
10:52:03amiconnkHz would be more unintuitive than Hz, and MHz requires handling decimals
10:52:08tucozMy radio uses a 91.0:descr format
10:52:14tucoz:)
10:52:38amiconnAn example of rockbox philosophy - KISS
10:52:39tucozCrappy coded though. I just wanted to try to change it, and it works nice.
10:52:49XavierGrso tucoz you suggest using a 0.05 step?
10:53:12 Quit ashridah (Read error: 110 (Connection timed out))
10:53:18tucozIf that is what is wanted I think that should be done.
10:53:30amiconnThe .fmr files aren't primarily made to be edited by hand
10:53:45amiconn...although it's possible, since they're ascii
10:54:21tucozamiconn: maybe if loading of .fmr files is implemented
10:54:32XavierGrwell in my code (for fmr loading and saving) I don't think that would be a problem. Although it is quite handy to be able to write one by hand.
10:55:38 Join ashridah [0] (i=ashridah@220-253-122-8.VIC.netspace.net.au)
10:55:45XavierGrsilly me "my code"? well modded code more like me. excuse me.
10:56:05tucozWell, the format is ok as it is now I think. It is fairly easy to read, and keeps the code simple
10:57:00tucozBut I think there are some things that could be changed in the fm-radio. 1) number of presets 2) load preset file 3) change freq range
10:57:31tucoz3) because the chip let's you do so, and some people like in japan need that.
10:58:38XavierGr1 and 2 will be fixed with a patch that I prepare.
10:58:51amiconn3) may need attention concerning the old Samsung tuner chip
10:59:07tucozXavierGr, great
11:00
11:01:40tucozyou have AM-radio as well on that chip?
11:01:51LinusNno
11:02:11amiconnThe chip can do AM, but the components needed for AM aren't present afaik
11:02:20LinusNexactly
11:02:21tucozah, I see
11:02:24amiconnI don't have an FMR
11:02:26XavierGrso many people ask for AM. Is that so great, I cant tune a single station in my home radio in AM
11:03:04amiconnLinusN: Does the Samsung tuner allow FM frequencies outside the european range?
11:03:04tucozI think the americans use am radio more than people in europe
11:03:25LinusNamiconn: i think so
11:03:26amiconnIirc the chip itself does, but perhaps there is a filter blocking them
11:03:50SlasheriHmm, what is the frequency range the radio chip can do on iriver?
11:04:04Slasherior is there any limitations at all?
11:04:09amiconnyes
11:04:14amiconn76-108 MHz
11:04:17Slasheriah :/
11:04:52Slasheriit would be cool if it could go over 140 MHz amateur radio bands
11:05:09tucozWith the bad reception of the iriver?
11:05:43amiconnThe old crappy samsung tuner allows 10-160 MHz fpr FM
11:06:29tucozhehe
11:06:51tucozdo you have something against the samsung tuner?
11:06:58amiconnYes
11:07:00tucozhaha
11:07:31amiconnIt is very sensitive to electrical interference, and is not very sensitive
11:07:38amiconn...to the actual signal
11:07:46tucozsounds like a bad deal
11:08:25tucozhave to eat, see you
11:08:31 Part tucoz ("Leaving")
11:13:50XavierGrSlasheri I tried your cache patch and I can say I am very pleased by the results!
11:15:48XavierGrAt start I thought that it would work for the fisrt folder to give time for the disk to spin, but no. I browse through various folders and spinning at all.
11:16:02XavierGrHow it works? Does it caches all the folders?
11:17:44LinusNyes
11:18:41*solexx_ can't wait for it to be committed
11:18:46XavierGrand what if there are too many folders?
11:19:35LinusNXavierGr: then the cache is disabled and you have to reboot to enable it again
11:19:46LinusNafaik
11:19:48B4gderugh
11:19:59XavierGrhmm that's why I saw a reboot warning once in a while?
11:20:14LinusNyup
11:21:49XavierGrLinus: About the fm preset scenario. Lets say that someone loads his own preset. Then he adds one (or removes) do we overwrite the changes automatically or do we wait the user to select save preset list (and then overwrite or make a new one)?
11:22:02solexx_XavierGr: how many folders do you have?
11:22:16XavierGr800 maybe
11:22:24XavierGr3300 files
11:22:36LinusNXavierGr: today it saves it immediately
11:22:51LinusNi think it should keep doing that
11:22:53XavierGryes I know it saves it in a default fmr file.
11:23:16LinusNit should save in the currently selected preset file
11:23:26XavierGrok we can keep that routine (that saves the default fmr) but what about the other presets.
11:23:39LinusNit should save in the currently selected preset file
11:23:53B4gderI agree
11:23:58XavierGrautosave changes or manually save them (manual option will be standartd one way or the other)
11:24:48XavierGrso it will autosave any changes, right?
11:25:02LinusNit should do it exactly as today, with the only difference that it saves to the currently selected file
11:25:12XavierGrok got it.
11:25:25LinusNthat's my opinion, others may disagree
11:26:02B4gderI say go with the "least surprise" design rule
11:26:17B4gderand then I think it should save in the newly selected one
11:26:24LinusNfull ack
11:27:48XavierGrso in order to retrieve the latest filename I call gloabal_settings.wps_file (in my case fmr_file after I have done the changes)
11:29:47LinusNyes
11:38:19XavierGrpff again I can't get used to the idea that in C I can't type "char a = char b" to copy a string!!!
11:39:13LinusNhighlevel loser :-)
11:39:52XavierGryou got me! Visual Basic was one of my favorites...
11:40:16XavierGrand it is for Windows Programing
11:40:35SlasheriXavierGr: just think logically and you will see that it is not possible *a and *b are pointers so doing a = b would change a's pointer to b's pointer and then both a and b would always point to the same string
11:42:02LinusNSlasheri: clear as mud to a vb programmer :-)
11:42:19Slasheri:D
11:42:21Slasherihehe
11:42:26XavierGrLOL
11:42:34 Join Lynx0 [0] (n=lynx@tina-10-4.genetik.uni-koeln.de)
11:42:59XavierGrSlasheri is it right with your patch to ask every once in a while rebooting?
11:43:30XavierGrI mean okay I like the cache but if every time I delete a folder or file it needs reboot...
11:43:41SlasheriXavierGr: no that is a bug
11:43:48XavierGrphew!!
11:43:51linuxstbXavierGr: You may already have implemented the presets this way, but I don't think we need a "save presets" option. The act of adding or deleting a preset should imply a save.
11:43:58Slasheriunfortunately it doesn't happen on my unit but it will be fixed before the real commit
11:44:03XavierGrcan't wait for commitment.
11:44:06Slasheri:)
11:44:23XavierGrLinusN: what do you think of this new addition?
11:45:17LinusNwhich addition? the cache or the presets?
11:46:46XavierGrcache!
11:46:59XavierGrpresets are still pending...
11:47:03XavierGrbut on the way
11:47:31XavierGramiconn was not so happy about it.:p
11:47:35ZagorI don't like the cache, but you knew that already :-)
11:47:48LinusNi like the cache, but i don't like the problems that go with it
11:48:02LinusNXavierGr: see the irc log for my opinions
11:48:04XavierGrBrowsing was so fast the other day....
11:48:08XavierGrok
11:48:35B4gderI don't know enough facts about the cache to have an opinion yet
11:48:43linuxstbI'm undecided about the cache - it's something users definitely want, but developers don't :).
11:48:50B4gderI'm worried about the "too big to fit"-problems
11:49:12LinusN"08.01.54 # <LinusN> frankly, i don't like this concept at all, but that's me"
11:49:38B4gderI'm likely to join up on LinusN's side on this
11:50:15LinusNthe major problem is how to handle the situation when the hard drive is updated using the boot loader usb mode
11:50:49linuxstbOr the iRiver firmware USB mode?
11:50:58amiconnFully dynamic cache without a cache file, I'd say
11:51:23amiconnEven if it costs a little more battery
11:53:52 Join Febs [0] (n=Febs@207-172-122-81.c3-0.rdl-ubr4.trpr-rdl.pa.cable.rcn.com)
11:55:01 Quit Lynx_ (Read error: 113 (No route to host))
11:55:01 Nick Lynx0 is now known as Lynx_ (n=lynx@tina-10-4.genetik.uni-koeln.de)
11:55:18LinusNlinuxstb: yes, that too
11:56:13SlasheriLinusN: i think the most reliable solution for that problem would be to add a unlink(DIRCACHE) to the bootloader when exiting the usb mode. This should not compromise the bootloader reliability to handle emergency situations and maybe it could be added to next bootloader that is released?
11:56:46LinusNi'd say that the bootloader shouldn't write to the file system at all
11:56:48B4gderand if you only update rockbox then?
11:56:59Slasheriah, probably you are right
11:58:23Slasheribut is it possible in theory that deleting a single while could made a corrupt filesystem worse?
11:58:51Slasheriif the filesystem is very corrupted, even the unlink function should not succeed?
12:00
12:00:59linuxstbSlasheri: The update could have happened using the iRiver's firmware. So the bootloader won't know about it.
12:01:06LinusNit's more a design decision
12:01:15LinusNthe boot loader should do as little as possible
12:01:46LinusNand definitely not delete files, imho
12:02:15Slasherilinuxstb: yes, but the possibility for that is very small
12:02:37SlasheriLinusN: true..
12:02:53LinusNSlasheri: small, yes, but when it happens, we will get annoying questions in the forums
12:03:15Slasherihmm, what about the eeprom? =) or is there any unused space? or maybe writing a value to some ram address rockbox could read on next boot up?
12:04:02LinusNi just thought of this scenario:
12:04:38LinusN- enter bootloader usb mode to add some music before i go to work
12:05:01LinusN- i remove the usb cable and want to run to the bus
12:05:21LinusN- then i can't shut off the player until the dir scan is complete, can i?
12:06:04LinusNif i do, i will have to reboot twice to restore the cache file, right?
12:06:05Slasheriwith current implementation you can't (or yes you can, but that will be a forced shutdown after the 8 seconds). But instead you can start playing music
12:06:29Slasherione boot should be enough
12:06:35LinusNok
12:12:09XavierGrok question for the presets now.
12:12:32XavierGrAs you know I just made a new dir_filter for fmr files.
12:12:57XavierGrso if a user has some fmr files he can see them and open them.
12:13:14XavierGrcurrently it will pop up the radio and load the presets
12:13:19XavierGris that acceptable?
12:13:29LinusNseems awkward to me
12:13:35XavierGrwhy?
12:13:51LinusNi think a preset file selector is better
12:14:10XavierGrexplain please?
12:14:12LinusNlike the language selection
12:14:28XavierGror the font or the wps
12:14:31XavierGryes I did the same
12:14:32LinusNyes
12:14:54LinusNok now i understand what you mean
12:15:16LinusNmaybe that browser should be in the radio screen itself
12:15:29XavierGrso is it acceptable when someone opens the file that the radio screen pops up and loads the presets?
12:15:41XavierGrso make it saperate.
12:15:45LinusNi'd rather have the preset selector in the fm radio menu
12:16:03linuxstbXavierGr: Are there any limitations on where the fmr files can be stored?
12:16:04XavierGryes there is that already
12:16:14XavierGrwell standart ones
12:16:23XavierGron the rockbox dir it will be rememvered
12:16:28***Saving seen data "./dancer.seen"
12:16:35LinusN.rockbox/fmpresets/ ?
12:16:52XavierGrwell I set it .rockbox but I can change it.
12:17:09LinusNi think it's best in a separate dir
12:17:17LinusNlike the langs and plugins
12:17:31linuxstbSo I'm guessing you are not allowed subdirectories? - e.g. country/city.fmr The Wiki list of presets is getting very big.
12:17:45XavierGrokay fine but the I will have to make the entry for the configuration more than 20 characters right?
12:17:55LinusNlinuxstb: but are you likely to have all those preset files on your player?
12:18:10XavierGras Linus said
12:18:18linuxstbYes - I would like to just download them once, and forget about them. Rather than having to remember to get them before I travel somewhere.
12:18:29linuxstbBut it wouldn't be a huge problem to have them all in one directory.
12:18:37XavierGrwell that a little stretched
12:19:19LinusNi agree that it would be cool to have a giant preset database on the player
12:19:39XavierGrso Linus you suggest to disable browsing support of the files?
12:19:49LinusNXavierGr: nah
12:20:06LinusNbut maybe it shouldn't start the fm...
12:20:15LinusNi'm a little undecided
12:20:26XavierGryes I don't know either.
12:20:32LinusNi'd say disable it
12:20:47LinusNor start the viewer
12:21:08XavierGrdisable what the initialization of the FM?
12:21:44LinusNmaybe it would be best to send the preset file to the text viewer instead
12:22:54 Quit goa (Read error: 110 (Connection timed out))
12:26:24solexx_something completely different: is there a reason why "push-to-right" doesn't start playing the currently selected track in the playlist viewer?
12:26:37LinusNi guess not
12:26:56solexx_it's different from the file browser although it looks almost the same
12:27:02LinusNspeaking of the pl viewer, i think it should leave it after selecting a file
12:27:28solexx_wouldn'thurt either
12:27:50*solexx_ curses his space bar
12:29:01 Nick solexx_ is now known as solexx (n=jrschulz@d164064.adsl.hansenet.de)
12:30:11XavierGrso if someone opens it. then the viewer should pop-up and load the file as the last opened preset file then, right?
12:32:33XavierGrhow do I open a file with the viewer?
12:33:17LinusNXavierGr: imho, it shouldn't do anything else than send it to the viewer
12:33:40XavierGrnot set it as the current preset?
12:34:42XavierGrbut then how could one load presets. Because even inside the radio the user loads preset after a rockbox_browse is called to select the file.
12:34:59LinusNtrue
12:35:08XavierGrso...
12:35:35XavierGrmaybe poping the FM is still good.
12:35:48XavierGrWhy someone wants to open a preset file?
12:35:49 Join preglow [0] (n=thomjoha@hekta.edt.aft.hist.no)
12:36:25XavierGrnot to mention that the preset files are difficult to find unless saved outside the current rockbox folder
12:36:27LinusNit should probably load the preset file, but not start the fm
12:36:51XavierGrok that is the easiest for me to code.
12:36:58LinusNgood
12:37:03linuxstbI think the fmr files should behave the same as the wps files when you select them in the normal file browser. But I forget what that behaviour is.
12:37:22XavierGranyway we will ask for feedback before commiting.
12:37:25LinusNlinuxstb: it selects the file (and displays it for a second)
12:37:43XavierGryes
12:37:46linuxstbAnd it sets it to be the current WPS?
12:37:49LinusNyes
12:38:04linuxstbSounds logical for the FMR files as well then.
12:38:10LinusNso doing the same for fmr files seems logical
12:38:29solexxIMO you could skip the display part
12:38:31linuxstbBut do we need to display it? Would a splash() be better?
12:38:35XavierGrdisplay them for a second and load them
12:38:52solexxI never understood the use of displaying the wps
12:39:17solexxit's too short to read or recognize it anyway
12:39:19LinusNsolexx: me neither, it's pretty useless
12:39:32linuxstbBut the user needs some feedback to say what has just happened.
12:39:35LinusNi think it's a debug feature that remained
12:39:49XavierGrexactly I set it to load the presets only
12:39:49solexxdisplaying the wps doesn't tell the user "wps loaded" either
12:39:50LinusNa "selected" splash is way better
12:39:59solexxyep
12:40:05XavierGrand when you press right nothing happens but the preset loads
12:40:19XavierGrLinusN is right.
12:40:23linuxstbSo "WPS selected" and "Presets selected"?
12:40:50XavierGrMaybe though still I think that poping the FM is still good behaviour
12:41:08XavierGrwhat if I have a preset in the root?
12:41:13XavierGrIt will be useless.
12:41:19LinusNyes it will be
12:41:29linuxstbIt will be loaded temporarily - like WPS (I think).
12:41:50LinusNXavierGr: i don
12:41:53XavierGrbut if the FM will not pop up then it will be lost again
12:41:56LinusNt think i mind that much
12:42:08linuxstbIt will be lost after the next boot.
12:42:11XavierGrto be useless or to pop the radio
12:42:23LinusNthe question is where it would save the presets if you change them
12:42:39linuxstbBut selecting a WPS file doesn't bring up the WPS
12:42:55XavierGryes because there is no need to.
12:43:00linuxstbLinusN: Back to the same file. It's just that that file would not become the default after the next boot.
12:43:12XavierGrbut if someone wants to open presets there is a big chance that he wants to listen to radio
12:43:42XavierGrthat's why I am talking about poping the radio
12:43:45 Join [IDC]Dragon [0] (n=d90a3255@labb.contactor.se)
12:43:49LinusNlinuxstb: the way it will work, it only looks for preset files in the .rockbox/presets/ dir
12:43:53solexxI say: do it
12:44:00XavierGror at least that's the way I think of it
12:44:30solexxthe wps files are "settings", the fm stations are more like "application specific documents"
12:44:32LinusNXavierGr: do as you wish, i don't mind if it starts the fm radio
12:44:33linuxstbLinusN: All I'm saying is that whatever we do, I think WPS and FMR should be consistent.
12:45:06*[IDC]Dragon saw a cache debate
12:45:21XavierGrlinuxstb: I don't think that we must relate those two.
12:45:31XavierGr2 diffent uses for both.
12:45:37LinusNlinuxstb: but they are, as solexx pointed out, not really the same thing
12:45:46XavierGrWPS is needed only in playback while presets are needed for radio
12:46:03[IDC]Dragonlong in the past, I was kicking an idea around: have a "smart" disk cache and also handle playback with it
12:47:11linuxstbIn terms of simplifying how the Rockbox UI works, I think it's useful to think of them as the same thing. That's all I'm saying.
12:47:49XavierGralso Linus I would like to ask you about a preset addtion I thought. We can utilize a button to toggle between preset scaning and frequency scanning. I think it is a very good thing to add.
12:48:00solexxlinuxstb: jpeg files are opened with the appropriate plugin, too
12:48:11solexxok, fm is not exactly a plugin, but...
12:48:25XavierGrI see solexxes argument
12:48:30LinusNXavierGr: i'm not sure i like that, but what the heck...
12:48:42XavierGrCool!
12:48:56XavierGrI 'll try and see.
12:49:08LinusNXavierGr: remember that we have multiple platforms
12:49:38XavierGrYes before everything I will try to see if the keymaps allow such thing.
12:49:41LinusNXavierGr: are you using cvs now?
12:49:47XavierGrno
12:49:51LinusNplease do
12:49:59XavierGrhow?
12:50:05XavierGrI need access
12:50:09LinusNsyncing your patches will be hell if you don't
12:50:17LinusNyou don't need write access
12:50:38XavierGroh you mean to update the cvs
12:50:44LinusNyes
12:51:04XavierGrwell I am using a cvs update before everything
12:51:20XavierGrthen I fresh copy that folder to make my hacks
12:51:33LinusNcopy?
12:51:36LinusNwhy?
12:51:50XavierGryes I have a rockbox-devel folder that I update once in a while
12:51:58XavierGrthen I copy the whole folder to make my changes
12:52:05B4gderskip the copy
12:52:09B4gderand you'll be happier
12:52:13XavierGrjust to avoid downloading the whole source afain
12:52:17LinusNok, and do cvs update in that folder?
12:52:17XavierGr^again
12:52:20B4gderjust edit, then update
12:52:25B4gderin the same tree
12:53:03XavierGrIt is just that I am afraid to spoil the rockbox-devel folder with my changes so I copy it to be sure that if anything goes wrong I will have the backup
12:53:18LinusNthis is what cvs is for
12:53:18B4gderif you need a fresh tree, then check out another copy there
12:53:23B4gderyou can checkout as many as you need
12:53:56LinusNin fact, cvs should work fine in the copy as well...
12:53:58XavierGrso I can type "cvs update" after the cahnges I have done?
12:54:07B4gderXavierGr: yes!
12:54:08LinusNyes, that's the whole point
12:54:16B4gderthat's how all of us work
12:54:17XavierGrand no duplicates or wrong code?
12:54:29B4gderif so, those are yours ;-)
12:54:41LinusNit will merge the current cvs with your changes
12:54:42XavierGrhmm its funny that I do the update before my changes
12:55:07preglowdoesn't seem like the mailing list is too happy with me
12:55:20LinusNpreglow: oh?
12:55:47preglowbah
12:55:57preglowforget it, i signed up with a mail alias instead of the one i use to send
12:56:01preglowno wonder it wont relay my mails
12:57:38XavierGrso if I make some cahnges in radio.c and apply my changes and update to cvs, what if a dev make changes to the radio.c?
12:57:49XavierGrafter waht I have done?
12:58:02XavierGrthat the case I am afraid
12:58:27preglowno way for me to change the list address without resubscribing, no?
12:58:39preglowXavierGr: then you get a collision if you're unlucky
12:58:40 Join B4gd3r [0] (n=daniel@static-213-115-255-230.sme.bredbandsbolaget.se)
12:59:01B4gd3rshaky connection
12:59:01preglowXavierGr: cvs will mark up for you where yours and the other dev's changes collide so you can merge them yourself
12:59:21XavierGrfine I will do this from now on.
12:59:32B4gd3rbut such collides are rare
12:59:34preglowXavierGr: if you changed different places cvs will just merge the file and you'll never notice anything
13:00
13:00:14XavierGrLinusN: Did you got any time to check the filescrolling jpeg viewer? Not that I am in a hurry, just asking.
13:00:21LinusNnot yet
13:06:48preglowgreat, THERE it accepts my mail twice
13:08:23XavierGrAlso Linus one final comment for the FM on iriver. If someone exits the FM (and continue playback) and then reenters you can hear a very loud pop when the radio starts?
13:08:28XavierGrhave you noticed it?
13:08:45LinusNhaven't tried :-)
13:09:13XavierGrtry it, it is quite loud!
13:09:27LinusNit always initiates the playback when entering the fm screen, so i'm not surprised
13:09:36B4gd3rpop music! ;-)
13:09:41XavierGrLOL
13:09:52amiconnAt least the volume change is fixed :)
13:11:28*preglow vanishes
13:14:44 Quit B4gder (Read error: 110 (Connection timed out))
13:15:26 Nick B4gd3r is now known as B4gder (n=daniel@static-213-115-255-230.sme.bredbandsbolaget.se)
13:15:59[IDC]DragonB4gder: back to l33t level 1?
13:16:24B4gdery34h! ;-P
13:41:51 Nick QT_ is now known as QT (i=as@madwifi/users/area51)
13:53:51amiconn[IDC]Dragon: There are 2 problems with gcc >3.3.x and archos
13:54:57amiconn(1) Code size increases, and we don't want that
13:55:16amiconn(2) gcc 4.0.x doesn't work at all because of a silly decision of the gcc team
14:00
14:01:07 Join Chamois [0] (i=Chamois@AStrasbourg-252-1-10-114.w82-126.abo.wanadoo.fr)
14:01:47Chamoishi
14:02:22Chamoisanyone did experience the splash problem i spoke about with the dir caching patch from slasheri ?
14:06:34[IDC]Dragonamiconn: ok, then rather not 4.0.x :-(
14:06:53SlasheriChamois: i will investigate that problem, it is probably a bug
14:07:28Chamoisok
14:07:41Chamoisbut it appears only on my iriver
14:07:43Chamoisstrange
14:08:32[IDC]Dragonis gcc 4.0 beneficial for Iriver?
14:11:51amiconn[IDC]Dragon: The problem with gcc 4.0.x and archos is that gcc 4.0+ does only allow (weak) symbol aliases to symbols defined in the same source file, but SH1 firmware/system.c defines aliases to symbols defined in an asm() block, which the compiler can't see
14:11:53SlasheriChamois: i have some ideas what might cause that, i will try those soon
14:11:55 Quit Febs (" HydraIRC -> http://www.hydrairc.com <- State of the art IRC")
14:12:51SlasheriChamois: do you have very deep directory structure (over 8 nestled directories)? However it shouldn't cause that problem
14:13:39Chamoishmm
14:13:45Chamoisyes i have
14:14:46amiconn[IDC]Dragon: The problem is http://gcc.gnu.org/gcc-4.0/changes.html , under "New languages and language specific improvements" / "C family"
14:15:22amiconn"This is because it's meaningless...." is nonsense as we can clearly see
14:16:31***Saving seen data "./dancer.seen"
14:25:48 Quit Chamois (" HydraIRC -> http://www.hydrairc.com <- 100,000+ downloads can't be wrong")
14:41:26preglowamiconn: well, have you spoken to the developers?
14:41:38[IDC]Dragonamiconn: hmm, do we have any contact to gcc maintainers?
14:42:09[IDC]Dragonseems this would make gcc unuseable for embedded
14:42:36[IDC]Dragonif we can't call into assembler
14:43:09preglowjust file a bug and see what they say
14:43:24preglowmailing list would probably be more sensible
14:44:47[IDC]Dragoniriver has no asm files?
14:46:34amiconn[IDC]Dragon: It has, but it doesn't use (weak) aliases to symbols defined in asm blocks
14:47:00preglowthere has to be a way around that on archos as well
14:47:08[IDC]Dragonnaively asking, why do we have to in Archos?
14:47:08preglowi'll admit calling into asm blocks like that isn't very pretty
14:47:41amiconnIn fact we have to, if we don't want to make the exception handling either bloated, or hard to understand
14:48:27amiconnpreglow: This is the exception handling. We need one entry point per exception number because it's the only way to find out this number
14:48:43LinusNpreglow: it all comes down to the fact that the SH1 stack frame doesn't contain the vector number
14:48:53preglowi don't think sh1 is alone in this
14:49:14preglowi suggest someone drop a mail on their list, though
14:50:29amiconnsee firmware/system.c lines 569..572, 592..838, 905..1015
14:54:31preglowi don't see how having that feature can hurt anyway
15:00
15:06:10linuxstbA feature idea: Every plugin has a menu assigned (if possible) to a standard button or button combination. This menu (unless the plugin disables playback) would have some default playback-related options (e.g. "volume", "pause/resume playback" and "skip current track"), plus any menu items required by the plugin itself.
15:07:09LinusNlinuxstb: a good idea
15:07:28linuxstbAre the pause/resume/skip options easy to implement?
15:07:55preglowas long as the plugin itself controls this, it sounds like a nice idea
15:08:07preglowhowever, this menu would of course be best assigned to a common button :>
15:10:09linuxstbI haven't thought about how to implement it, but it would seem sensible for the plugin API to handle the common menu items in some way.
15:13:28LinusNlinuxstb: it would seem sensible, yes, i think we could do it with a default handler in the plugin library
15:15:04linuxstbDo you mean a button handler that would intercept the menu button press?
15:15:14HCleh..
15:15:30HClwhy don't we just have two threads and allow people to control the music with the remote...?
15:15:46linuxstbBecause not everyone uses the remote...
15:15:53*linuxstb has lost his remote...
15:15:55HClmk, well.
15:16:04HClas long as the two ideas won't exclude eachother
15:16:06HClits fine with me
15:16:10linuxstbAnd not all targets have remotes
15:16:17preglowbut still
15:16:30preglowas long as we're not hijacking a button from all plugins now
15:16:39linuxstbpreglow: Yes, I think we would have to.
15:16:59preglowwell
15:17:02preglowsome plugins use all buttons
15:17:03linuxstbBut the benefit is that the plugin could replace key combinations with menu items.
15:17:13linuxstbMuch more user-friendly.
15:17:15preglowthat would do
15:17:16preglowyes
15:17:18preglowi agree then
15:17:40HClincidentally
15:17:48HClsomeone has to fix patience on iriver
15:17:50HClto use grayscale
15:17:50linuxstbFor some reason, plugins (apart from Rockboy) have not used menus until I added the API for Sudoku.
15:18:00HClso its easier to identify the color of the cards
15:18:11linuxstbI think I saw a patch for that.
15:18:20HClhmm, i'll look at that soon then
15:18:31HCli haven't really been motivated to code lately
15:19:49LinusNif we use the default handler concept, it's up to the plugin to call the default handler
15:22:24linuxstbSo we would add the menu to default_event_handler(), along with functions to allow the plugin to add items to that menu?
15:24:11LinusNthe adding of items didn't occur to me
15:25:29LinusNi wouldn't add it to the default handler itself
15:25:31preglowwell, you'd need that if you rob the plugins of buttons to utilise
15:26:30linuxstbI was also thinking that in general some plugins (especially the ones with lots of button combinations) could benefit from their own menus anyway.
15:26:41LinusNyes
15:26:44amiconnI think all plugins that have a menu should be converted to use the core routines, then such features can be added to the plugin's menu
15:27:06amiconn...providing the necessary functions as part of the plugin library
15:27:22preglowagreed, i think it's a marvelous idea
15:27:32amiconnPlugins that don't use a menu atm but would benefit from such should get one
15:27:40amiconn(like the text viewer)
15:27:52linuxstbNow the big question - which button?
15:28:03preglowa/b on iriver?
15:30:40amiconnThe menu should alwas use the standard menu button of the target if at all possible
15:31:13amiconnAB on iriver, F1 on archos recorders, MENU on archos player, long MODE on Ondio
15:32:41linuxstbSo I can take the definition of TREE_MENU from tree.h?
15:33:41amiconnYes, provided it is available without explicitly including tree.h
15:34:09amiconnPlugins should never include anything else than plugin.h or .h files from the plugin library
15:35:07linuxstbSo there is no central set of #defines for button aliases?
15:35:14linuxstb(at the moment)
15:38:43linuxstbI now have a far more controversial suggestion - making the main menu the "root window" of Rockbox, instead of the file briowser. The first item in the main menu would then become "Browse Files"
15:42:21B4gderwow
15:42:53B4gderwhy is that better?
15:44:19linuxstbI think it's a nicer "home" to come back to.
15:45:04linuxstbTwo new items on the main menu would be "browse files" and "WPS" (or "Now Playing" or similar)
15:45:23[IDC]Dragonthe Ondio stock firmware has such
15:45:32linuxstbLeaving the WPS or the file browser or the FM radio would take you consistently back to the main menu.
15:45:42[IDC]Dragona quick menu with an entry for each button, basically
15:46:22amiconnI wouldn't like that.
15:46:24[IDC]Dragontop entries are browsing, resume, record, FM or such
15:46:26linuxstbI've started implementing it, and will put a test build up for people to try. The code changes seem trivial, but it's a completely different way of thinking when using Rockbox.
15:46:43amiconnIt means more keypresses for selecting the music to play
15:46:43linuxstbBut personally I think it's more natural.
15:47:00B4gderI think I'll have to test it before I give a judgement
15:47:12linuxstbamiconn: I accept that.
15:47:13amiconn[IDC]Dragon: Ondio FM has 4 entries, Ondio SP has only 2
15:47:21[IDC]DragonI found it quite nice
15:48:01linuxstbIt's also that you can do a lot more in Rockbox now than just play files - but it's biased towards that aspect.
15:48:29[IDC]Dragonresume is one button, like before
15:48:32amiconnYes, however, playing music is the primary purpose of a DAP
15:48:57[IDC]Dragonother people record often, or use the radio
15:49:34linuxstbThe default option in the main menu will be "browse files" - so it's only one extra keypress.
15:55:26amiconnStill one keypress too much ;/
15:56:21amiconnThe archos ondio firmware uses something like our quick menu, left/right/up/down for 4 functions (only up/down on ondio SP)
15:56:23LinusNStart in browser = yes
15:57:12 Quit ashridah ("Leaving")
15:57:15B4gdernew option!
15:57:17B4gder:-)
15:57:38linuxstbAll I'm asking is to test my demo build with an open mind.
15:57:59*B4gder will
15:58:17XavierGrme too sounds great!
15:59:01XavierGrNow that I think of it I haven't heard any suggestion that I disliked sofar in Rockbox...
16:00
16:00:39amiconnLinusN: Hmm. Perhaps a Start in.. option allowing to choose between main menu, browser, recording screen (...)
16:01:05amiconnThat wouldn't add an extra option, as we already have "Start in recording screen"
16:02:04amiconnThe question is how the menu will be handled if we already are in the browser. It might become possible to loop...
16:03:03B4gderthat could be... interesting
16:03:13preglowlinuxstb: i'd love to try this change
16:03:20linuxstbEither the browser calls the menu (and the menu exists), or the menu calls the browser, and the browser exits.
16:03:40linuxstbSo at startup, one or the other would be called.
16:04:17amiconnI'd rather do it a bit different to make the behaviour more consistent
16:04:55amiconnIt would require this "root menu" to be independent, not callable from anywhere except at startup
16:05:42amiconnThis root menu would then check the "Start in.." setting and execute the matching menu option without a button press
16:06:19amiconn(of course unless the setting is "root menu", then it would display itself and wait for a button)
16:06:26LinusNi like it
16:06:52amiconnThe question is whether it should be possible to go back to this root menu if it's not your preferred starting point
16:07:21LinusNhmmm
16:07:23[IDC]Dragonhow about a configurable root menu (Jehova!)
16:07:35[IDC]Dragon</irony>
16:07:37LinusN /kick [IDC]Dragon
16:07:46amiconnI think it has to, or otherwise functions need to be callable from multiple points
16:11:41 Nick paugh is now known as AliasCoffee (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
16:12:29 Join Chamois [0] (i=Chamois@AStrasbourg-252-1-21-160.w82-126.abo.wanadoo.fr)
16:13:39XavierGrnow that is what I call strange or magic!!!!
16:14:04XavierGrI planted some splashes onto my code to see output(I know I should have used logf)
16:14:20XavierGrthen I removed them all with replace.
16:14:21 Join rasher [0] (n=jonas@62.79.64.148.adsl.hs.tiscali.dk)
16:14:52XavierGrbut one was splashing even if I had removed all of them.
16:15:17XavierGrI commented a line compiled and then uncommented and the splash never apeared again????
16:15:22solexx*PANIC*
16:15:36XavierGrdo not PANIC
16:15:39solexxupdating size on empty dir entry 19
16:15:52solexxi don't - my iriver does!
16:16:08XavierGrthat's the error you got?
16:16:10solexxdidn't touch it for quite a while, it was just playing
16:16:13preglowsourceforge cvs is hell
16:16:15solexxXavierGr: yep
16:16:34***Saving seen data "./dancer.seen"
16:16:37solexxoh, battery empty...
16:17:10B4gdersourceforge in general should be avoided for sanity
16:17:48*solexx curses rockbox for not shutting down when battery is *very* low
16:19:47 Join muesli- [0] (i=muesli_t@Bbc87.b.pppool.de)
16:20:18 Part LinusN
16:20:34XavierGryou should always charge your player at around 40% of the remaining battery.
16:20:56XavierGrLi-ion and li-polimer tends to hold more if you charge it around that.
16:21:09B4gdersays who?
16:21:25preglowi've never heard of that
16:21:31XavierGrBattery University.com
16:22:07XavierGrOn the other hand using li batteries to the end and over usage kills them more quick
16:23:09solexxafk
16:24:33preglowyes, that's known
16:24:56preglowif you let a li-* battery run to too low a voltage, its capacity will be permanently lowered
16:25:08XavierGrexactly.
16:25:21XavierGrthe golden line for charging is at 30-40%
16:25:53muesli-hum...
16:26:05muesli-i always let my battery dry out :-|
16:26:20 Join wacky [0] (n=wacky@modemcable006.177-201-24.mc.videotron.ca)
16:26:21crwli hear you should charge when the battery's still about 60-70% full
16:26:24B4gderspeaking percentages in this aspect seems so non-techie
16:26:43crwlheh
16:27:19muesli-btw is misticriver down?
16:27:33B4gderno
16:28:09XavierGrhow long was your last visit there muesli?
16:28:13muesli-cant reach it
16:28:32wackyhey guys, do you think a port for the iAudio is something that will come after the H3xx ?!
16:28:32muesli-http://www.misticriver.net/
16:28:36muesli-no connect
16:28:56B4gderwacky: if there are people who wants it and works on it, then sure
16:28:59XavierGrI have no problems to log in.
16:28:59preglowwacky: different people are working on those ports
16:29:08preglowwacky: either might come first
16:29:15B4gderbut the iAudio port has stalled for months
16:29:18muesli-XavierGr who knows..will try l8er..
16:29:23wackyotherwise, I think I'd return that thing. I can't stand bad firmwares anymore, now that I've sighted your software
16:29:26preglowbut can't exactly claim i've seen austriancoder too much lately, though
16:29:43wackyyeah.. that's it! haven't seen him either
16:29:55B4gderI mailed him a week or so ago
16:30:01B4gderand he said he'd be back on track soon
16:30:09wackygoooodd :)
16:30:12muesli-XavierGr so you recommend charging at 30% already?
16:30:25 Quit tvelocity (Success)
16:30:42wackyguys, you who know the hardware a little bit more than me.. do you think porting pitchtune.sf.net to the Rockbox would make something horribly slow ?
16:30:47wackywhat's the speed of the display ?
16:30:59preglowyou can't just straight port that
16:31:08wackyI know.. I mean the concept
16:31:13wackyheh :P
16:31:20preglowunless you want to code a gtk port for us
16:31:26wackyhahah :)
16:32:03preglowsomething like it is not impossible, though
16:32:13wackyjust wondering.. if the LCD and hardware stuff are fast enough to draw the samples that come through the mic. input...
16:32:36preglowwell, we already do it on the remote
16:32:42wackyhuH??
16:32:44preglowas for the main screen i have no idea
16:32:55preglowwacky: when you record, you can have the waveform be viewed on the remote lcd
16:33:07preglowrealtime
16:33:28wackyreally ?!? hey... I've leaved my beloved iRiver alone too much
16:33:58wackythe main unit's LCD is broken.. so I was just waiting for remote support.. but that doesn't mean I can control recording from the remote huh?
16:35:35preglownot entirely, no
16:35:51wackyparts ?!
16:39:01preglowyeah, but can't remember what
16:40:58wackyI'm going to try that out :)
16:49:37[IDC]Dragonmaybe LinusN has a display in that bag of chips
16:51:49XavierGrmuesli: yes 30% is okay.
16:54:19wackyan LCD ? yeah.. but he's a bit far away :)
16:54:45XavierGrhey muesli check this out another remote support by someone called firefly!
16:54:57 Quit muesli- (Read error: 110 (Connection timed out))
16:55:11XavierGrhttp://forums.rockbox.org/index.php?topic=1549.0
16:56:15wackya good one this time ?
16:57:52XavierGr:( well it seems better than mine. It is based on latest builds too.
16:58:03XavierGrBut it lacks remoteconfiguration :p
16:58:20XavierGrEasy thing to add I will email him the details
16:58:56XavierGrThough I don't think that it will make cvs...
16:59:12wackywooooo!
17:00
17:02:23B4gderI would like an approach that doesn't pollute the entire source code with #ifdefe HAVE_REMOTE_LCD
17:02:34wackyhey that's neat!
17:02:54wackyeverything works! so I've got my player back on track! :)
17:03:46wackywhere's the recording menu ?
17:03:55preglowin the debug screen
17:04:08preglowi don't think that'll take kindly to this patch
17:05:05 Join linuxstb_ [0] (n=linuxstb@i-83-67-212-170.freedom2surf.net)
17:08:25wackyhmm.. the recording screen isn't on the remote :P So I don't know what I'm doing :)
17:10:14wackydo you know the keymapping for the recording screen ?!
17:10:23 Quit AliasCoffee (Read error: 60 (Operation timed out))
17:19:50 Quit Chamois (" HydraIRC -> http://www.hydrairc.com <- IRC has never been so good")
17:20:32 Quit linuxstb (Read error: 110 (Connection timed out))
17:22:44 Join goa [0] (i=hd@gate-hannes-tdsl.imos.net)
17:24:02*wacky 's tired of that BitchX client
17:24:05 Quit wacky ("laksdfj")
17:25:06 Join wacky [0] (n=wacky@modemcable006.177-201-24.mc.videotron.ca)
17:25:24wackyouch.. that 'epic' client has absolutely no colors ?!
17:26:43 Quit wacky (Client Quit)
17:26:51 Nick linuxstb_ is now known as linuxstb (n=linuxstb@i-83-67-212-170.freedom2surf.net)
17:27:40 Join cYmen [0] (n=cymen@nat-ph3-wh.rz.uni-karlsruhe.de)
17:28:43rasherWacky should've loaded a script.
17:29:00rasherBut really, he should just use irssi.
17:29:07fuzzie:)
17:29:08XavierGrwhy rockbox_browse always return false?
17:29:29XavierGrit should return true on file selection and false on cancel!
17:30:04rasherhrm, there are only "return false" and "return" statements
17:30:18rasherWhat does "return;" mean in a bool function?
17:30:50fuzzie"you're violating the spec!", i assume
17:31:03XavierGrsorry didn't catch what you said, rasher
17:31:36rasherXavierGr: the rockbox_browse function either does "return false;" or "return;" - there are no return true; statements in the function
17:32:17XavierGrreturn? where do you see the blant return?
17:32:23XavierGrI can see only a return false.
17:32:48rasherline 1150
17:33:04rasherand 1504
17:34:05XavierGrbut these are not on the rockbox_browse function
17:34:22rasherhaha
17:34:27rashersilly me
17:34:33XavierGrlol
17:34:42rasherWasn't reading properly
17:34:59fuzzieHeh.
17:35:08XavierGranyway what a "return" means on a boolean function? I know that you need to return something
17:35:15rasherhm, dirbrowse() returns a bool
17:35:27XavierGrreturn true or false
17:35:34rasherXavierGr: it's not there, it'd probably give either a warning or a compiler error
17:35:42rasherI was imagining things.
17:36:13XavierGrit is just that I call rockbox_browse to select a file.
17:36:26XavierGrif the user cancels the code executes anyway.
17:36:45rasherrockbox_browse should possibly return the return value of the dirbrowse() call
17:36:49XavierGrbut in the examples of lets say font selection it works. Strange
17:36:49rasherI'm thinking
17:37:21rasherThat'd make sense, methinks
17:37:31XavierGrI think that dirbrowse never returns something
17:37:37rasherit doesn't make sense that rockbox_browse always returns false, at least
17:37:41rasherdirbrowse returns a bool
17:37:50rashereither false in the case of a few errors, or true
17:37:50rasherbut
17:37:56rasherthat's not what you want either, I guess
17:38:54XavierGrhmm dirbrowse return true in the end.
17:39:17XavierGrit is strange because I can see the same thing for something else working right.
17:39:53XavierGrAnd it is not big deal for what I am doing,it will not cause an error, it is just that it will spin the disk needlesly
17:40:21XavierGrrockbox_browse will return false either if I push right or left.
17:40:57XavierGronly diference is that if it is a known filetype and you press left the code will not run for the file.
17:42:11XavierGranyway i Will find something different.
17:42:14 Join muesli- [0] (i=muesli_t@Bbca6.b.pppool.de)
17:47:42XavierGrhow many presets do you suggest to be the limit?
17:47:46XavierGr40 or 50?
17:48:12 Quit Nilisco (Read error: 110 (Connection timed out))
17:49:18XavierGranyone? a simple suggestion?
17:49:52rasherI think someone was saying paris had 55 stations
17:50:10ghode|afk64?
17:50:43 Join Nilisco [0] (i=nilisco@wrath.shellfx.net)
17:51:06XavierGrok then I will double it. 64 to be sure
17:51:51XavierGrIt is just a define so if the devs think that it is a waste we should change it. (Though I think the code stops claiming memory if there are less than 64 entries)
17:52:35rasherI saw a h340 in a shop today
17:52:55B4gderXavierGr: nope
17:53:04B4gderwe have no dynamic memory
17:54:05XavierGrso do you think that 64 is too much?
17:55:32B4gderI guess not
18:00
18:02:54 Quit B4gder ("time to say moo")
18:07:40 Quit muesli- (Read error: 110 (Connection timed out))
18:16:36***Saving seen data "./dancer.seen"
18:26:26 Join Lear [0] (n=chatzill@h36n10c1o285.bredband.skanova.com)
18:37:08 Quit einhirn ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
18:41:01 Join ashridah [0] (i=ashridah@220-253-122-177.VIC.netspace.net.au)
18:41:25 Nick Lynx_ is now known as Lynx_awy (n=lynx@tina-10-4.genetik.uni-koeln.de)
18:49:00 Quit XavierGr ()
18:51:10 Quit rasher ("Ex-Chat")
18:51:18 Join rasher_ [0] (n=jonas@62.79.64.148.adsl.hs.tiscali.dk)
18:51:21 Nick rasher_ is now known as rasher (n=jonas@62.79.64.148.adsl.hs.tiscali.dk)
18:54:40 Quit [IDC]Dragon ("CGI:IRC")
18:54:55LearAnyone played with Tremor on Windows?
19:00
19:00:12 Join arkascha [0] (n=arkascha@xdsl-213-168-116-124.netcologne.de)
19:01:35 Join _aLF [0] (n=Alexandr@mut38-2-82-67-66-128.fbx.proxad.net)
19:02:28amiconnlinuxstb: Regarding your ipod porting effort: I don't think it would be mandatory to support the itunes database
19:03:59amiconnI'd think the main reason for rockbox on ipod would be to get rid of the now-mandatory pair ipod-itunes, and allow plain filesystem playback
19:05:08rasherIt's probably a good idea to do it. Some people are too attached to iTunes by now.
19:05:54amiconnI think these people are better off if they stay with the apple firmware
19:06:24amiconnItunesdb support might be a nice complement, but I won't give it top priority
19:08:16*ender` tried itunes - too much eye candy
19:08:25BirdFishWhy would someone change from Apple firmware if they wanted to use iTunes?
19:08:30amiconnApart from full usb msd compatibility I wouldn't see the point in porting rockbox. From several statements it seems the apple firmware is user friendly, unlike iriver or archos firmware
19:08:42*fuzzie mumbles something about vorbis.
19:08:53linuxstbGapless and other audio codecs
19:08:59BirdFishfuzzie: but iTunes doesn't allow vorbis
19:09:06fuzzieit does
19:09:20fuzzieadmittedly no-one has a qt7 codec which can read tags, yet
19:09:26BirdFishI thought that iTunes converted everything to Apple format first
19:09:26amiconnThe user unfriendliness of the archos firmware was one of the reasons why the rockbox project was started
19:09:30fuzzieno
19:09:30linuxstbBut it refuses to sync them to the ipod - because the ipod can't play them.
19:09:30 Join tvelocity [0] (n=tony@ipa147.3.tellas.gr)
19:10:10amiconnlinuxstb: ipod doesn't do gapless???
19:10:32*BirdFish sees where a Rockbox port with iTunes capabilites could be nice now. But points out that most iPod users are of the "social crowd" and have them merely for the sake of having a dap.
19:10:43ender`about user friendliness, truthfully, i preferred the original H120 interface to rockbox
19:10:54BirdFishOh?
19:11:10BirdFish/ignore ender`
19:11:17BirdFish:P
19:11:59LearMaybe it looks a bit better, but that doesn't make it user friendly, imho...
19:12:26 Join linuxstb_ [0] (n=5343d4aa@labb.contactor.se)
19:12:37linuxstb_amiconn: No, the ipod is far from gapless.
19:12:45amiconn*When* do dap manufacturers learn their lessons? :\
19:12:46ender`when eg. setting options, i like to see all available options at once, while with rockbox i usually get 2 lines of text and 75% of unused screen space
19:13:42rasheramiconn: I believe they already did
19:13:53LearYou do have a point there. Still, should be easy to fix...
19:13:54rasheramiconn: noone outside a few whiny geeks care about gapless
19:13:55 Join Mxm`Pas`Bien [0] (n=flemmard@fbx.flemmard.net)
19:14:09 Quit Maxime` (Read error: 104 (Connection reset by peer))
19:14:14BirdFishrasher: gapless playback could be a cool feature
19:14:46*ender` doesn't care about gapless, while ReplayGain is very important to him
19:14:51amiconnrasher: Archos could do gapless in 2000. Why today's dap manufacturers can't do that is beyound me
19:14:57BirdFishrasher: dap manufacturers are just rushing everything our right now in an attempt to beat another in the "looks pretty, must be cool" race
19:15:02rasheramiconn: They don't have to.
19:15:15linuxstb_The ipod hardware is also capable of high quality (96KHz 16-bit stereo) recording. But the Apple firmware doesn't support it.
19:15:19BirdFishReplayGain is a must
19:15:33BirdFishVorbisGain is even better!
19:15:50linuxstb_But I would say the main problem is the reliance on iTunes or a third-party clone.
19:15:53rasherBirdFish: same thing
19:16:00BirdFishrasher: no
19:16:09BirdFishVorbisGain is for OggVorbis files
19:16:15BirdFishReplayGain is for Mp3 files
19:16:16BirdFish...
19:16:24rasherVorbisgain is an implementation of ReplayGain for vorbis files
19:16:30amiconnender`: It's exactly the other way 'round for me, but that doesn't change the image of today's dap manufacturers at all, in fact
19:16:33LearNope, ReplayGain is for MusePack! :p
19:16:35rashermp3gain is an implementation of ReplayGain for mp3 files.
19:16:37ghode|afkit's kind of funny, ask many ipod owners about gapless and they would say that its nearly perfect
19:16:45LearFor Mp3, use Mp3Gain.
19:18:04LearHeh, either it is gapless or it isn't, so I don't see it could be "nearly perfect"...
19:18:07linuxstb_Also, the iPod is (I think) the only other DAP with a third-party open source firmware. So it's the easiest available new target for Rockbox.
19:18:27amiconnNah, I wouldn't say easiest
19:18:40BirdFishI'm not sure what the big deal about gapless playback is though
19:18:59BirdFishI'm not sure I always want my songs playing directly back to back
19:19:05BirdFishA nice pause helps when changing genres
19:19:12amiconnBirdFish: Encode a mix album, and play that
19:19:25amiconnThe track change must not be audible
19:19:34amiconnIt should sound exactly as the source CD
19:20:01BirdFishSource music is always nice
19:20:03BirdFish:D
19:20:43BirdFishAnd once I get rid of this h10 piece of shit and get a real dap, I'll start loading all my music in a much higher bitrate
19:21:13BirdFishthe iaudio x5 does FLAC :D
19:21:42*BirdFish wonders if Rockbox has ever thought of including Monkey's Audio in their firmware
19:22:00*rasher points BirdFish at the SoundCodecs wiki page
19:22:08*Lear doubts the APE code is GPL compabitle...
19:22:13rasherIt isn't.
19:22:14*BirdFish thanks rasher
19:23:38linuxstb_amiconn: I'm curious - which other players would be easier?
19:24:25amiconnlinuxstb: I'm not sure, but I'd think the iAudio players, and of course H3x0
19:24:37LearOdd.. I have one gcc3 build and one gcc4. If I try to compile a "partial" gcc3-build with gcc4 (or vice versa), I get really strange errors...
19:24:56amiconnThe bad thing with ipod is that it uses a closed-docs core chip
19:25:24amiconnEven though the ipodlinux guys reverse engineered a lot, you said yourself it's badly documented
19:25:38linuxstb_I wouldn;'t disagree with that.
19:26:27amiconn...and I'd expect some 'neat' surprises when working on a port
19:27:17amiconnHmm, then that may also happen with documented chips; I should know about that :/
19:27:23linuxstb_I'm expecting to be able to get a 95% working port working relatively easily. The other 5% will be due to lack of knowledge by ipodlinux.
19:27:31*amiconn just says mas bitshift
19:30:46amiconnAnother ipod drawback is the variety of generations
19:31:39amiconnIt's hard to define whether it's valid to say 'rockbox runs on ipod' if it runs on just one or two of them
19:31:58linuxstb_I wouldn't call it a drawback. The good thing is there has been a general stability between all the different models.
19:32:14linuxstb_Just compare them to the range of iRiver or Archos players...
19:32:43amiconnArchos isn't bad in that respect
19:32:55amiconn(well, the range rockbox is running on)
19:33:11rasherigp-100->ihp-100->ihp-300 wasn't too bad
19:33:26linuxstb_True, but there are probably more Archos targets for Rockbox than there are iPod variations.
19:33:47amiconnAll of them use the same CPU (SH1) at almost the same clock, all of them use a MAS as the audio dsp, all except the players use the same lcd...
19:34:48amiconnThere are six archos targets, with two of them being binary identical with only a slight difference in scrambling
19:34:57amiconnThat's easy to beat with ipod generations...
19:35:45amiconnSmall quiz: Which 2 of the 6 archos targets are the binary identical ones? ;)
19:35:52linuxstb_I'm just saying the situation is similar.
19:36:48 Join Slyck [0] (i=Acksaw@spc1-stok5-4-0-cust5.bagu.broadband.ntl.com)
19:37:32SlyckHi
19:39:06amiconnlinuxstb: If we want to support all the different generations, we would need to acquire at least one of each.
19:39:20amiconnDo you think it's possible to tell the generation of an ebay offer?
19:39:27linuxstb_Or acquire developers with them.
19:39:41amiconnHmm, of course
19:40:13Slyckanyone seen those Boomtubes?
19:40:20linuxstb_The photos and descriptions at http://ipodlinux.org/Generations will help if you want to go shopping
19:40:47rasherI think they're pretty easily disinguishable
19:40:49rashereh
19:50:15 Quit ashridah ("Leaving")
19:50:57 Join dpassen1 [0] (n=dpassen1@resnet-233-61.resnet.UMBC.EDU)
19:51:08 Part Slyck
19:52:24LearYay, the "Sandvall"-patch produces bit-identical output, it seems like. Good. :)
19:53:23LearAnd when running in Windows, it's about 30% faster...
19:53:38Lear(Than the stock lowmem version of Tremor, that is)
19:55:17LearBut that only makes it slightly faster than the stock tremor (again, in Windows)...
19:55:39rasherwhat's the point then?
19:56:30LearTremor_lowmem uses much less memory, which would solve a few problematic ogg files... And could perhaps be a little faster.
20:00
20:00:28 Quit arkascha (Read error: 110 (Connection timed out))
20:01:47amiconnLear: I thought rockbox already uses tremor-lowmem, am I wrong?
20:02:07LearOuch, the Tremor that was the base for the Rockbox port is really fast... about 40% or so, compared to current Tremor.
20:02:59Learamiconn: I'm pretty sure, yes. The code differs quite a bit (compared with the lowmem version I've checked at least), and Rockbox simulator memory behaviour is similar to that of the non-lowmem version.
20:09:43amiconnMaybe the lowmem version can be made faster on iriver because lower memory consumption allows to put more things in iram?
20:12:25LearDoubt that; much of the difference is from dynamically allocated stuff based on the codebooks in the file, and how the file is read into work buffers.
20:13:07 Join DangerousDan [0] (n=Miranda@newtpulsifer.campus.luth.se)
20:13:58 Join zeekoe [0] (n=zeekoe@wekkerradio.kabel.utwente.nl)
20:16:40***Saving seen data "./dancer.seen"
20:21:18rasherLear: have you tried with gcc4.0.2 ?
20:21:58LearNo, wasn't released when I downloaded 4.0.1...
20:22:10rasherAh
20:22:24rasherI'll try
20:23:55LearIs it released now?
20:24:26rasherThere's a 4.0.2 directory on the server at least
20:31:47 Join solexx_ [0] (n=jrschulz@d144209.adsl.hansenet.de)
20:46:03 Join XavierGr [0] (n=XavierGr@ppp45-adsl-28.ath.forthnet.gr)
20:48:45 Quit solexx (Read error: 113 (No route to host))
20:49:36 Quit Mxm`Pas`Bien ()
20:50:13 Nick TiMiD[ZZZzzz] is now known as TiMiD (n=TiMiD[FD@asgard.valombre.net)
20:50:19TiMiDhi
20:51:05XavierGrhi
20:51:08TiMiDabout remote : will firefly patch be accepted (in that case I stop my developpement) ?
20:51:24XavierGrI don't think so.
20:51:29TiMiDok
20:51:45XavierGrWhile it does the work (like my older remote build) it approaches the wrong way.
20:51:52CoCoLUSthere's -another- remote patch?
20:52:03rasherI think a combination of your work TiMiD and firefly's is the way to go
20:52:05 Join Maxime [0] (n=flemmard@fbx.flemmard.net)
20:52:05TiMiDyes, it's like the one you did before
20:52:22XavierGrThe remote must be programed in a saperate thread with diferent .c files that will contain the remote rendering routines
20:52:25rasherLear: danger
20:52:31TiMiDrasher: the problem is that it's a completely different approach
20:52:34rasherLear: /home/jonas/rockbox/apps/codecs/libmad/synth.c: In function ‘synth_full’:
20:52:37TiMiDI don't duplucate code
20:52:37rasher/home/jonas/rockbox/apps/codecs/libmad/synth.c:646: error: can't find a register in class ‘DATA_REGS’ while reloading ‘asm’
20:52:53rasherTiMiD: I know, but some places aren't easily generalized
20:53:06TiMiDso the existing code have to be more deeply modified
20:53:10rasherlike rolo ouotputting "Rolo loading" on both screens, etc
20:53:17XavierGrthe #ifdef remote isn't a good way to go
20:53:48Learrasher: 4.0.2 isn't released yet...
20:53:49XavierGrproblem with both patches is that it floods the main code with remote specific code.
20:53:51TiMiDWhat I intend to do is to avoid lcd_* calls in application files
20:53:59rasherLear: Ah :)
20:54:02Learbut it sounds like more cases of the same problem...
20:54:07XavierGrthat is not acceptable. we must program the remote saperately
20:54:34XavierGrthen you are on the right path TiMiD
20:54:41TiMiDby getting the job done with some gui functions
20:55:17TiMiDfor example I started a list "widget" which can be used in both menus and filetree
20:55:22XavierGrit would be best to saperate your code to remote local files (that will include some of the original code)
20:55:23LearRemote should not be different .c files, imho (if you mean like remote-tree.c, etc.). Makes maintenance harder...
20:55:29TiMiDbut the integration will be painful
20:55:46TiMiDnope
20:55:50TiMiDno separate files
20:55:52Leartimid: and playlist viewer.
20:55:56XavierGrOk then what you suggest in order to avoid floodinf the code?
20:56:53TiMiDplaylist viewer too
20:57:24TiMiDmaybe also a wps "widget" (you can't call this widget because it's not that flexible)
20:57:28LearA sort of graphics context, keeping the needed state for remote _or_ main. Widgets and gfx render to the context, which maps the calls to the right display. Might reqiure same kind of display in main unit and remote though...
20:58:17TiMiDhttp://timidzone.free.fr/pub/rockbox/remote/screen_access.h
20:58:29TiMiDI handle multiple screens like that
20:58:53TiMiD(not finished though)
21:00
21:00:11XavierGrhmm I got a little confused. What are you trying to do exactly?
21:00:39XavierGrBecause all main rendering functions for the remote are already present
21:00:46TiMiDto allow acces to both screens with the same interface
21:01:01XavierGrwhat is missing is to call those functions.
21:01:12TiMiDso code is same for remote and main display
21:01:49XavierGrhmm better discuss this with a dev
21:02:03LearHe, exactly what I had in mind... :)
21:02:03XavierGrI can't help you much
21:02:21TiMiDfor example lcd_set_drawmode(DRMODE_SOLID|DRMODE_INVERSEVID); or lcd_remote_set_drawmode(DRMODE_SOLID|DRMODE_INVERSEVID); becomes display->set_drawmode(DRMODE_SOLID|DRMODE_INVERSEVID);
21:02:52TiMiD(since the functions have exactly the same prototype)
21:03:01XavierGrinteresting now I begin to understand
21:03:06TiMiDI handle this with fn pointers
21:03:51XavierGrbut still sometimes you must call the function specifically with arguments that will fit current screen.
21:05:58 Join hd [0] (i=hd@gate-hannes-tdsl.imos.net)
21:06:21 Quit Maxime ()
21:07:03TiMiDXavierGr: no, LCD_WIDTH becomes display->width :)
21:07:40TiMiDwell I didn't tested "advanced functions"
21:08:06TiMiDbut until now, everything is working as expected
21:08:10 Join Maxime [0] (n=flemmard@fbx.flemmard.net)
21:08:14 Quit DangerousDan (Read error: 110 (Connection timed out))
21:10:47TiMiDThe only problem I see is the huge amount of code to understand and modify / reorganize
21:11:20rasherI believe you should start by creating the widgets
21:11:27rasherif possible
21:11:53TiMiDI started with the list (from scratch) and the status bar
21:11:54rasherand then start submitting patches for the rest of rockbox to use these
21:11:54 Quit Maxime (Read error: 104 (Connection reset by peer))
21:11:57rasherother people can do that
21:12:04rasheras well, once the widgets are there
21:12:18 Join Maxime [0] (n=flemmard@fbx.flemmard.net)
21:12:18TiMiDwhen status bar will be working, I will try an integration with the filetree to test
21:12:42TiMiDthe problem is that the code is in intermediate state
21:13:09TiMiDold pieces of code to allow thewhole thing to compile andnew code used in small parts
21:15:33 Join DangerousDan [0] (n=Miranda@newtpulsifer.campus.luth.se)
21:15:34 Quit Maxime (Read error: 104 (Connection reset by peer))
21:19:07 Join hd- [0] (n=light@gate-hannes-tdsl.imos.net)
21:20:42 Join Maxime [0] (n=flemmard@fbx.flemmard.net)
21:22:24 Quit goa (No route to host)
21:22:25 Nick hd- is now known as goa (n=light@gate-hannes-tdsl.imos.net)
21:28:17 Quit Maxime ()
21:30:42rasherLear: I suck. I hadn't cvs upped to get your changes
21:31:17Lear:)
21:31:29rasherso I guess the 4.0.2 might've worked as well
21:31:34rashergoing to stick with the 4.0.1 release
21:31:44CoCoLUSmaybe someone can help me: how can i rip multiple tracks from a cd as a single wav file?
21:32:09LearInteresting... Sandvalls patch includes lots of tables, but many of them doesn't seem to be used...?
21:32:22LearExact Audio Copy (Windows) can do that easily...
21:32:24 Join Maxime [0] (n=flemmard@fbx.flemmard.net)
21:32:32 Quit hd (No route to host)
21:33:01XavierGrSo it looks like TiMiD will get this done... I didn't have a clue about function pointers.
21:33:05fuzzieI think someone needs to write a coldfire compiler which isn't ICE-happy.
21:33:11fuzzieSomeone get to that.
21:33:29*rasher suggests punching the gcc developers
21:33:51fuzziei've spent quite enough time punching the m68k gcc people for other reasons :<
21:34:20 Join hd [0] (i=hd@gate-hannes-tdsl.imos.net)
21:36:25 Join Aedaren_ [0] (n=Aedaren@71-35-210-44.slkc.qwest.net)
21:36:40Aedaren_hey is recording working on the h1x0 yet?
21:37:01Aedaren_and if so, where is it in the menus?
21:37:41 Quit Maxime ()
21:37:50rasherAedaren_: a recording-test function is available through the debug menu "info > debug (keep out!) > pcm recording"
21:38:14rasherIt's not close to how recording will work eventually
21:38:38Aedaren_right, thanks.. will it kill my player if I try it out?
21:38:41Aedaren_:)
21:39:19rasherNot at all (but you'll have to restart once you're done)
21:39:28rasherAnd it may crash
21:39:40rasherThere's a reason why it's in the debug menu after all
21:39:45rasherBut it's not "dangerous" as such
21:39:53rasherJust not very good
21:39:59Aedaren_thanks!
21:40:38 Join Maxime [0] (n=flemmard@fbx.flemmard.net)
21:43:22 Part Aedaren_
21:43:49 Quit goa (No route to host)
21:43:49 Nick hd is now known as goa (i=hd@gate-hannes-tdsl.imos.net)
21:43:52 Quit Maxime (Read error: 104 (Connection reset by peer))
21:46:43 Quit linuxstb_ ("CGI:IRC (Ping timeout)")
21:55:23XavierGrwhy do I get an error regarding alac codec when compiling?
21:55:38XavierGrI get past it with make -k but it is annoying
21:56:25LearI've needed to run tools/configure, but I don't know why...
21:56:57Bagderit's because it doesn't build alac otherwise, but still tells the linker to link with it
21:57:01XavierGrdo you get the same error?
21:57:31BagderXavierGr: always try to rerun tools/configure first when you get problems like that
21:58:04LearWould be nice if configure could remember autoconf settings... :)
21:58:31Bagderit once did, but it turned out to be rather painful to support properly
21:58:42Bagdersince it changed quite a bit at times
22:00
22:00:21BagderI have a patch that adds a big notice on the screen if the configure script is newer than what was used to generate the root makefile
22:00:22XavierGrI type an alias to tools/configure
22:00:23XavierGrrbconf
22:00:45Bagderugha
22:03:37 Quit goa (No route to host)
22:06:05 Join ]Maggie[ [0] (i=fupl@t2-17.ath.cx)
22:12:25 Quit BirdFish (Read error: 110 (Connection timed out))
22:16:04 Join Maxime [0] (n=flemmard@fbx.flemmard.net)
22:16:43***Saving seen data "./dancer.seen"
22:25:23 Join matsl [0] (n=matsl@1-1-4-2a.mal.sth.bostream.se)
22:29:56 Quit Maxime (Read error: 104 (Connection reset by peer))
22:30:39 Join Maxime [0] (n=flemmard@fbx.flemmard.net)
22:37:20 Join mirak [0] (n=mirak@AAubervilliers-152-1-43-229.w83-114.abo.wanadoo.fr)
22:40:28LearHm... What happens with the result when using the emac on 16-bit integers in fractional mode?
22:40:45*rasher summons preglow
22:43:10XavierGrI am always suprised with rockbox. Due to a bug a filename takes 2 slashes "/.rockbox//test.fmr". Not only it will not crash but it will load the file fine :)
22:43:41rashersounds logical
22:44:42mirakhi
22:44:42XavierGrwhy it should crash on the second slash
22:44:47XavierGrhi
22:44:51mirakhow is it going for the H300 ?
22:44:58XavierGridle...
22:45:15Bagdermirak: it sits waiting
22:46:06mirakwhy ?
22:46:14mirakthere is a problem or lack of time ?
22:46:26XavierGrlack of time I suppose.
22:46:36Bagdera big lack of time even
22:46:38XavierGrLinusN is a very busy person
22:46:50XavierGrwho else is on H300?
22:47:13Bagderwell, many people could write h300 code if they wanted
22:47:22Bagderbut I guess everyone waits for the bootloader
22:47:31XavierGrbut LinusN is in charge ofthe bootloader right?
22:47:45Bagderyes, he's the only one with both a h300 and a bdm wiggler
22:47:59XavierGrthe preciouss bootloader, my preciousssss
22:51:34preglowLear: so, rockbox works just fine with 4.0.x now?
22:51:45preglowLear: tried to do any codec performance measurements? ;)
22:52:12XavierGrLets say that I have a char pointer *name. Is it right to allocate memory for it "memset(name, 0, 40)" and then strpy a value there?
22:52:27XavierGr^strcpy
22:52:43Bagdermemset() doesn't "allocate" memory
22:52:51Bagderit fills it with a set byte value
22:53:06XavierGrso that is wrong suppose.
22:53:17Bagderyes, at least rather pointless
22:53:19fuzzieand you should use strncpy to not overflow a static buffer :)
22:53:52Bagderor even memcpy() since strncpy() is generally an evil function ;-)
22:54:32fuzziewell, memcpy is more work because you'd have to grab the string length and work out how much to copy yourself, but shrug
22:54:42Bagderyes
22:54:48Bagderbut strncpy() zero-pads
22:54:54Bagderand it may not zero-terminate
22:55:06XavierGroverflowing it, is impossible.
22:55:17XavierGrto my case
22:55:26fuzziewell, if you're memsetting beforehand then you can just provide it with bufferlen-1 for n.
22:55:32fuzziebut i wasn't aware it zero-padded
22:55:46fuzziethat is indeed evil :)
22:55:59Bagderit is very annoying
22:57:11*HCl yawns
22:57:12HClhello.
22:57:54XavierGrhi HCl!
22:58:11HClhey :3
22:58:13HClhow goes?
23:00
23:00:00XavierGrinteresting...
23:00:17XavierGrSoon I will finish the load/save preset patch!
23:00:28XavierGryou?
23:02:24HClsleepy
23:02:27HCli'd lay on my bed
23:02:32HClbut i put my kitty behind my blankets
23:02:39HCland he's being very happy laying there
23:02:43ghode|afkdoes anyone know how to reover a hdd mp3 player that has a bad file system without opening it up?
23:02:43HCland i don't want to disturb him
23:02:54HCldepends on the mp3 player
23:02:58HClwith rockbox, usb bootloader
23:03:03HClotherwise, open it up, heh
23:03:29ghode|afkyeah i wish this has rockbox on it ><
23:03:52HClalternatively
23:03:56HClyou might be able to claim warranty
23:04:15dpassen1what mp3 player (curiosity)
23:04:17ghode|afkhttp://www.actronelectronics.com/detail.asp?product_id=38 - this one, bought it for my bro's birthday but now it won't boot after i tried transfering some files ;/
23:04:27HCl :(
23:04:28ghode|afkyeah going to try them tomorrow morning
23:05:15ghode|afki have bad luck with mp3 players apart from iriver, bought a phillips hdd070 and that sucks
23:05:31 Join Moos [0] (i=DrMoos@m50.net81-66-159.noos.fr)
23:05:35HCli only have an iriver
23:05:37XavierGrHCl: you have a kitty? (yeah I remember from your site) you are lucky, I love cats, though I don't have one. :(
23:05:38ghode|afkthink i'll go ipod next time :/
23:05:43HClyesh ^^
23:05:48HClhe's so cute :)
23:06:37XavierGrprrrrr :p
23:06:49Vladomanmiaow!
23:06:55HCl :p
23:07:11XavierGrhow old is he?
23:07:34HClabout two, i think
23:07:39HClthough i'm not completely sure
23:14:03 Quit mirak ("Ex-Chat")
23:14:07 Join einhirn [0] (i=Miranda@szgt-d9b8e95e.pool.mediaWays.net)
23:16:54Learstrlcpy is better than strncpy...
23:18:11Learpreglow: Yes, I run a 4.0 build now. I haven't made any exact measurements, no. Looking on the audio thread screen a couple of times...
23:18:12XavierGrstrcpy,strncpy. strlcpy, memcpy, memncpy I say whatever...
23:19:01XavierGrand I don't think that a function is better than other, it is just some uses of that function are better than others.
23:19:33Bagderwell, some functions makes it easier to do mistakes than others
23:20:06Bagdera million buffer overflows have taught us that
23:20:22XavierGrsure but if there is no possibility for error it is quite troublesome.
23:20:51rasherThere's always possibility for error.
23:20:56rasherError is always an option.
23:21:07Bagderthere will always be errors
23:22:58XavierGrso you mean that if I call a strcpy there always a chance that an overflow will occur?
23:23:08HClyes
23:23:16LearRegarding errors, I read on wikipedia about a small chunk of code in the Xbox. 512 bytes only, doing the initial bootstrap. 3 bugs made the it wide open for hackers (in the positive sense)...
23:23:19Bagdernot if you check the length properly first
23:23:29HClyea
23:23:34HClxbox is fun :p
23:23:38*HCl has one
23:23:51Bagderand in the xbox, they even left two of those three bugs when they updated it (iirc)
23:24:09HClyou can still hack in to it with software only :)
23:24:17HClthough you have to get the software on the harddisk first
23:24:49Bagderthey need a kernel patch due to bad pci
23:27:29XavierGrI wish rockbox could support dynamic memory...
23:27:37XavierGrwhy it can't anyway?
23:27:39Bagderwhy?
23:27:39Zagorno you don't
23:27:43Bagderit can
23:27:45Bagderwe just won't
23:28:33XavierGrevery time I want to enter a new filename variable I think of all the wasted bytes cause of the char array[MAX_PATH]
23:28:51Zagorthey are only waste if you declare it 'static'
23:28:51Bagderthink of it as always counting for the worst case
23:29:36XavierGrthen okay. Though in the filescrolling jpeg viewer it has to be static so...
23:29:55Bagderin the plugins the space isn't wasted anyway
23:30:04Bagdersince it just thrown away when the plugin exits
23:30:14Zagorthe reason we don't have dynamic memory is that it doesn't solve any problem, it just makes bugs easier
23:30:46Zagorstatic allocation means you always make room for everything you want to do −− you can never run out of memory.
23:30:50XavierGrwell you can clearly get a little more memory...
23:31:04Zagorno, that is an illusion
23:31:14Bagderif you can, you do it wrong now
23:31:57XavierGrand what of all this dynamic memory allocation in other projects?
23:32:08Zagordynamic allocation means you're never really certain where the limit is.
23:32:12XavierGrAnd I can see some uses of it.
23:32:26XavierGrin big n-dimensional arrays for example.
23:32:28Zagorbecaust most other projects live on PCs with virtual memory and disk swapping
23:32:32BagderXavierGr: do you much embedded programming with limited ram?
23:33:06Bagderyou have gigabytes of ram and blazing disk speeds on any regular PC
23:33:14XavierGrI am hardly a programmer what do you expect. I just express my thoughts and experiances.
23:33:41preglowBagder: we'll need a more proper malloc for codecs, i can't see how we'll be totally rid of it
23:33:44Bagderbesides, most programmers are sloppy and MANY programs are written badly
23:34:00Bagderpreglow: I realize that
23:34:20Zagorwell, we only need it because we don't want to modify the codecs to not need it :-)
23:34:38BagderLinusN has tested our malloc code in a "real" project so I'm quite sure we can get a functional malloc added once we feel like it
23:34:59preglowBagder: a better one should be a quick deal
23:35:08preglowBagder: the one we have now doesn't even support free
23:35:16preglowBagder: rockbox has a proper malloc??
23:35:27Bagderno, I meant the one that is actually a true malloc
23:35:47preglowokies
23:35:53Bagderhttp://daniel.haxx.se/projects/dbestfit/
23:36:10preglowbut yeah, nothing great, just enough that we don't have enough buffer for a long consecutive streak of one codec that does some mallocs and free
23:36:13preglowlike flac
23:36:49amiconnHmm, I'd prefer to not introduce malloc, but rather change the codecs if at all possible
23:37:54LearVorbis has memory problems too, but I don't think malloc would help there...
23:38:20amiconnImho malloc without mmu is good for headaches
23:38:42HCli agree with amiconn..
23:38:46fuzziefaad does malloc all over the place :<
23:38:58fuzziehave yet to get it to run far enough to hit any, though
23:39:13Zagorit's pretty strange that such performance-intensive code is so sloppy with its' memory handling
23:39:23Zagorbut quite common
23:40:02amiconnAlthough I'm not a codec expert, but I would think such type of code wouldn't need dynamic memory at all
23:40:25fuzziewell, faad2 at least was obviously designed without thinking of such things
23:40:28amiconnI mean it's not like codecs suddenly need heaploads of ram
23:40:34fuzzieso while it's not necessary, they're used it all over the place because it makes things easier
23:40:41amiconnThey work on frames and spit out samples
23:41:02amiconnIt should be simple to allocate the buffers to do that statically
23:41:37LearOr at least alloc some stuff initially, and not anything more after that...
23:42:26amiconnAllocating and deallocating all the time just causes overhead
23:43:39 Quit Lear ("Chatzilla 0.9.68.5.1 [Firefox 1.4/undefined]")
23:44:20amiconnI wonder about the rule to keep the changes to the codecs to a minimum. Imho there are already so many changes that this rule won't help anything anymore
23:44:29BagderI agree
23:44:45Bagderwe came up with that before we realized that the codecs are all messed up
23:44:55amiconn(emac routines, variables in iram, some mallocs already removed...)
23:44:57 Quit ansivirus ("Leaving")
23:45:06Bagderwe thought they were better than what they actually are
23:46:28linuxstbI don't see a real problem with changing the codec libraries significantly. faad seems to be the only one that is still actively developed.
23:46:35linuxstbEverything else is pretty static.
23:46:53Bagderyes, it seems we mostly agree on this
23:46:58fuzziethe faad changes are irrelevant anyway
23:47:01fuzziegiven the license
23:47:42linuxstbThe real problem is that it is a tedious job to strip them out of the codecs.
23:47:43preglowi think tremor has memory leaks as it is
23:47:48preglowvery tedious
23:47:56preglowespecially since there some times isn't a clear upper limit to the malloc
23:49:16linuxstbYes - some codecs seem to support theoretically large limits. But in practice, only a small range of values are used. That kind of thing isn't obvious.
23:49:22preglowtremor is like that
23:49:34preglowvorbis is designed to allow vast changes in the encoder
23:49:37*Bagder goes to sleep
23:49:42preglowwhich might suddenly quire different memory requirements
23:49:46preglowBagder: nitey
23:51:56preglowamiconn: yes, merging a new codec version is most definitely going to be manual work as it is
23:52:16linuxstbThere are some things like seek tables which are not that easy to define a maximum size for though. What do we say is the maximum duration of file Rockbox will support?
23:52:23preglowand metadata
23:52:46preglowi wouldn't like a defined maximum duration unless it is very large
23:52:58preglowin which case we might as well do it dynamically anyway
23:53:03linuxstbI guess FAT32 gives us some limits.
23:53:05amiconnlinuxstb: Theoretically rockbox can support files with up to 2^31-1 milliseconds playtime
23:53:16amiconn(before the 32 bit playtime calculation overflows)
23:53:28amiconn...provided the file is no larger than 2GB
23:53:47preglowbut getting rid of all mallocs is just not going to happen
23:53:50preglowat least i'm not going to do it :>
23:54:04linuxstbIt can be an aim though.
23:54:25preglowsure
23:54:33preglowit already is
23:55:03amiconnThat'd be 24 days 20 h 31 min 23 s ...
23:55:46linuxstbWhat bitrate is that for a 2GB file?
23:55:47amiconnNot very likely to fit that into 2GB ;)
23:56:39preglowSeed: awake?
23:57:41amiconnlinuxstb: In fact it would be possible: 8 kbps
23:57:46preglowthe only good thing i can say about musepack currently is that they don't malloc :/
23:58:00preglowand they make up for that by having the static decoder struct being a behemoth
23:58:16preglowamiconn: so we WILL have problems when we get speex files ;)
23:58:32preglowsupport, even
23:58:51CoCoLUS7,8125023574109783406900077905157, to be exact.
23:58:54amiconnpreglow: You think someone could speak non-stop for more than 24 days? ;)

Previous day | Next day