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 2005-10-05

00:00:22linuxstbNo. But Rockbox is in need of some widgets of that sort.
00:00:34*linuxstb prods TiMiD[away]
00:00:54preglowindeed, i more or less took for granted that you could just call the file browser and have it return the chosen file
00:01:13linuxstbNope. Hence my sarcastic reply :)
00:02:35linuxstbBut the user has to browse at some point anyway. It's quick to browse for a .ss file and then click on it than it is to start the Sudoku plugin, select file->open and then browse for a file.
00:03:25preglowsure
00:03:49preglowbut there are some obstacles of that being possible, obviously
00:03:58preglows/of/to/
00:04:10linuxstbBut I think that if Sudoku can be started as a plugin, it will then need a "load game" option.
00:05:17*amiconn brings back his old idea for plugins & viewers, slightly modified this time
00:05:58XavierGrwhich is?
00:05:59amiconnMy idea is that plugins and viewers reside in the same directory, but they have 2 different file extensions
00:06:17XavierGryes thats a good idea!
00:06:24amiconnOne of the is visible when browsing plugins (it's a supported extension), the other is not
00:06:39XavierGrand not much of a hassle.
00:07:04amiconnThis way we can distinguish between plugins which are usable standalone, and which are only viewers (or 'hidden plugins')
00:07:31amiconnviewers.config decides which plugin is a viewer, as it does now
00:07:47 Quit ender` (Read error: 110 (Connection timed out))
00:08:01amiconnA plugin that is both a viewer and usable standalone would get the visible extension, and be listed in viewers.config
00:08:16amiconnYou can derive the other 3 possibilities...
00:09:12preglowsounds decent
00:09:24linuxstbYes, sounds good. But I like the extension ".rock" :). Any suggestions for the two extensions?
00:10:18XavierGr.rock and .vrock?
00:10:24linuxstbAlso, how does the build system know which is which?
00:10:52amiconnHmm, good point. makezip.pl needs to know that
00:11:19XavierGrwhy makezip is written in perl?
00:11:19amiconnnah, it's needed even earlier
00:11:25linuxstbAnd probably the make scripts - if they have different extensions.
00:12:17linuxstbAnother solution would be to allow viewers.config to specify the full path (i.e. viewers/viewer.rock).
00:13:03amiconnPossible, needs some more RAM though
00:13:08DBUGEnqueued KICK chris|peacock
00:13:08chris|peacockOS: WinXP Professional 5.1 Service Pack 2 (Build #2600) CPU: AMD Athlon XP 2000+ (Thoroughbred) , 1.67 GHz Video: Sharp LL-T17A3-B on nVidia GeForce4 MX 440 (Microsoft Corporation) (1024x768x32bpp 60Hz) Sound: SoundMAX Digital Audio Memory: Used: 286/512MB Uptime: 7h 14s HD: [C:] 2.58/31.56 GB [D:] 9.23/9.31 GB [E:] 5.46/5.68 GB Connection: Cayman 3000 series USB Network Adapter @ 10.0 Mb
00:13:31XavierGr?
00:13:40CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
00:13:40*preglow hates irc scripts
00:13:53chris|peacockcrispy cock love it
00:13:58fuzzie(ick, 60hz)
00:14:15preglowhe'll seen be joining our blind users
00:14:34XavierGrhaha
00:14:57amiconn...and doesn't solve the problem how to tell the build system which kind of plugin it is
00:15:05amiconnfuzzie: Erm, tft?
00:15:10fuzzieyes, i gathered
00:15:20linuxstbamiconn: It's already defined to be 32 characters per viewer (I think). We can keep that limit, and wouldn't have to change anything else at all.
00:15:53fuzziejust, assuming based on the bad resolution, shouldn't :)
00:17:07*preglow strokes his sight-saving lcd panel
00:17:57 Quit ashridah ("Leaving")
00:20:36 Join RotAtoR [0] (n=e@12-208-71-148.client.insightBB.com)
00:20:53 Part chris|peacock
00:26:07 Join webguest09 [0] (n=acd5652c@labb.contactor.se)
00:26:14webguest09hi
00:26:40XavierGrhi
00:26:52webguest09since rockbox is being expanded to run on other mp3 players like iriver, wonder how long it will be before it will run on an iPod :-)
00:27:28linuxstbHopefully not too long.
00:27:32webguest09yea
00:27:58fuzziethis is the point at which you hand some bars of gold to linuxstb
00:28:13webguest09not that i would buy an ipod because it is so inferior to the iriver h1x0 series. I reckon the h1x0 series is going to be highly collectable since it is no longer made and it is the most powerful mp3 player ever made
00:28:18linuxstbOr to my boss to give me a couple of weeks off work.
00:28:49webguest09Linus is it possible to record from the h1x0's FM tuner?
00:28:59XavierGrwhat's your job linuxstb?
00:29:05XavierGrYes
00:29:13preglowwebguest09: yes it is
00:29:20linuxstbXavierGr: I wish I knew. But basically programming.
00:29:25webguest09without running a cable from the lineout to the linein?
00:29:31preglowwebguest09: with some very possible interference from the harddisk
00:29:31XavierGrAhh I envy you.
00:29:48XavierGrnot so noticable on good signal though
00:29:56webguest09but in order to record you need to loop a cable from the line out to the line in?
00:30:02preglowwebguest09: no
00:30:03XavierGrno
00:30:08preglowwebguest09: it's recordable as is
00:30:13preglowwebguest09: just disabled in the iriver firmware
00:30:20webguest09oh ok
00:30:29linuxstbRecording from FM works now in Rockbox.
00:30:35preglowwebguest09: you can already record radio in rockbox if you jump through some hoops
00:30:43webguest09I tried it but didn't get any sound in the .wav file
00:30:57preglowlinuxstb: btw, have you compiled rockbox for arm yet?
00:30:58linuxstbDId you select line-in when recording?
00:31:05linuxstbpreglow: Most of it.
00:31:08webguest09yea i did. Ill have another look at it
00:31:20webguest09im missing the SRS WOW feature from iriver f/w
00:31:26preglowlinuxstb: what's left to compile then?
00:31:41preglowportalplayer uses an arm7 core, yes?
00:31:53linuxstbYes - arm7tdmi
00:31:58preglowarrh
00:32:15preglowif i get some cash coming in some day soon, i'll prolly get some ipod variant myself
00:32:34linuxstbI'm still working on the bootloader, so I haven't fixed all of the compile errors I am getting in apps/ due to missing button definitions etc.
00:32:35webguest09the ipod nano is rubbish
00:32:40preglowi wonder how the nano's flash will take to development
00:32:47preglowhow many write cycles can flash take these days?
00:33:05linuxstbProbably more than the USB socket.
00:33:35XavierGrOne day I would really like to know what is the magic behind writing the bootloader and all the proccess for it..
00:34:22linuxstbThere is no magic. But the bootloader is basically a mini-rockbox. It contains the LCD, ATA, FAT etc code from Rockbox. So that needs to be written for the bootloader to work.
00:34:28 Join Paul_The_Nerd [0] (n=paulthen@cpe-66-68-93-2.austin.res.rr.com)
00:34:49 Quit webguest09 ("CGI:IRC")
00:35:20XavierGrYes but you have to open your player see whats inside read the datasheets then decide how to code them individually (and I don't have a clue how to load code on a machine)
00:35:42XavierGrhi Paul_The_Nerd.
00:35:48Paul_The_NerdHello
00:35:52linuxstbI've tried to describe how my ipod bootloader will work on the IpodPort wiki page. The ipodlinux people did the hard work - I'm basically using the same method as them, but adapted to Rockbox.
00:36:23XavierGroh so you didn't start from scratch as Linus with the iriver?
00:36:39linuxstbNo. That's impossible iwith the ipod because there are no public docs.
00:36:59XavierGrand how the guys with the linux made it?
00:37:10Paul_The_NerdQuestion concerning the viewers discussion earlier: what happens if in viewers.cfg you type ../plugins/whatever.rock instead of just whatever.rock?
00:37:11linuxstbreverse-engineering of the Apple firmware.
00:37:45linuxstbI don't think anyone involved has dissected an ipod in the same way as Linus and others have done for Rockbox.
00:39:00XavierGrSo you say his attempt is unique? How about the archos models?
00:39:33XavierGror the same was done for archos too?
00:39:45linuxstbI'm talking about all Rockbox ports - I think they have all come into being as a result of dissection and documentation of the hardware, and only minimal (if any) r-e of the original firmware.
00:41:04linuxstbPaul_The_Nerd: That would probably work. But it doesn't seem elegant somehow.
00:41:16Paul_The_NerdIt's definitely inelegant.
00:41:28Paul_The_NerdBut it's a simple solution (assuming it works) that also doesn't break anything.
00:41:32linuxstbCunning though.
00:45:40XavierGrwell starnge that the iPod uses only a database. I had forgotten how an iPod sucks...
00:46:05XavierGrImprisoned to use only the database blah!!
00:46:15XavierGrWhat if I have some mp3s untagged?
00:47:30 Join webguest51 [0] (n=c9816fa2@labb.contactor.se)
00:47:37 Quit webguest51 (Client Quit)
00:47:48Paul_The_NerdBy the way, that ../rocks/whatever.rock seems to work.
00:48:00Paul_The_NerdI tried it by moving the jpeg viewer and it went fine.
00:49:08linuxstbDid you delete the existing jpeg.rock from the viewers directory on your player?
00:49:42Paul_The_NerdYes
00:50:06Paul_The_NerdCtrl+X and Ctrl+v
00:50:22Paul_The_NerdSo, there's a quick 'n dirty solution at least
00:51:51 Join linuxstb_ [0] (n=linuxstb@i-83-67-212-170.freedom2surf.net)
00:51:52XavierGrthat can be made by default too. So there will be no need to do it again and again (after an update)
00:55:26Paul_The_NerdAlso, I'm having problems getting Cygwin set up properly, specifically the extra packages the page tells me to download
00:56:16XavierGrwhy not use the devkit?
00:57:07Paul_The_NerdI tried that too.
00:57:18Paul_The_NerdGave me all kinds of errors relating to cygwin1.dll
00:57:59Paul_The_NerdI'd gotten the other method to work before, but I've since replaced the HD in my laptop, and it didn't work this time.
00:58:29linuxstb_Are you trying to compile the m68k-elf-gcc and binutils yourself, or are you trying to install binaries of them?
00:59:23Paul_The_NerdI'm trying to install binaries via cygwin's setup.exe
00:59:53linuxstb_I'ld recommend getting the source. All you need from cygwin is the native (i.e. x86) gcc, make and Perl.
01:00
01:00:06Paul_The_NerdWhere would I get the source?
01:00:13Paul_The_NerdAnd how would I go about that?
01:00:23Paul_The_Nerdhttp://www.rockbox.org/twiki/bin/view/Main/CygwinDevelopment <−−- This is what I've been following.
01:00:25linuxstb_Follow the Wiki instructions - CrossCompile I think.
01:00:40Paul_The_NerdOkay
01:01:13linuxstb_http://www.rockbox.org/twiki/bin/view/Main/CrossCompiler
01:01:56linuxstb_The important thing is that you need binutils-2.16 (or later) and gcc-3.4.x (any "x" - higher the better) for iriver.
01:02:22 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
01:03:01Paul_The_NerdShould I stick with 2.16 with binutils, or go ahead and use 2.16.1?
01:04:15linuxstb_I would get the newest version. I'm using a CVS version of 2.16 (2.16.91 20050813). But standard 2.16 should be fine as well.
01:06:27 Quit linuxstb (Read error: 113 (No route to host))
01:07:30Paul_The_NerdAnd gcc 3.4.4?
01:10:48linuxstb_Yes. Perfect.
01:11:11 Nick linuxstb_ is now known as linuxstb (n=linuxstb@i-83-67-212-170.freedom2surf.net)
01:11:38Paul_The_NerdThank you very much
01:11:44 Join amiconn_ [0] (n=jens@p54BD4CFE.dip.t-dialin.net)
01:12:32 Quit amiconn (Nick collision from services.)
01:12:33 Nick amiconn_ is now known as amiconn (n=jens@p54BD4CFE.dip.t-dialin.net)
01:12:37amiconnBleh.
01:16:01***Saving seen data "./dancer.seen"
01:16:54 Quit matsl (Remote closed the connection)
01:33:15preglowbtw, any reason plasma isn't in cvs?
01:36:55linuxstbI don't think so. I haven't looked at it myself, but if you're happy with it, I would say to go ahead and commit.
01:37:23preglowhaven't actually looked at the source :>
01:37:34preglowworks very nice, though
01:37:41preglownice display of what the grayscale lib is capable of
01:37:58linuxstbI've just committed a "create new game" option in Sudoku. It's a menu option, and if you start sudoku.rock as a plugin (by browsing the viewers directory) it goes straight into this new mode.
01:38:50linuxstbIs it on the patch tracker?
01:38:58 Quit paugh (Read error: 54 (Connection reset by peer))
01:39:03preglowyes
01:39:17preglowhttp://sourceforge.net/tracker/index.php?func=detail&atid=439120&group_id=44306&aid=1280726
01:40:50linuxstbI'm compiling it now.
01:41:57linuxstbHow long is the Please wait... there for?
01:42:23preglowhmm?
01:42:46preglowwhen running? if so, couldn't say, i'm just using rasher's precompiled version
01:42:46linuxstbI started the plugin, a "please wait..." splash came up, and it is still there...
01:43:03preglowi never see any splash
01:43:18linuxstbI'll reboot my iriver. Looks like something went wrong.
01:44:57linuxstbMmm. Same problem. Pressing STOP gets rid of the splash, but then the plugin exits back to the browser.
01:45:31Paul_The_NerdHey, at least it behaves consistently.
01:45:51preglowhmm, i've no idea
01:45:55preglowyou'd have to ask rasher how he did it
01:46:00preglowbut i'm nearly falling asleep here
01:46:04preglowneed to go, bye all
01:46:09linuxstbSame here. Goodnight.
01:46:15 Quit preglow ("foop")
01:47:57 Quit dpassen1 ()
01:50:59Paul_The_NerdMaking GCC fails. =/
01:55:11*Paul_The_Nerd tries the patch
02:00
02:34:13 Join actionshrimp [0] (i=dave@dhcp-163-1-214-173.seh.ox.ac.uk)
03:00
03:07:09 Quit actionshrimp ("a bird in the bush is worth two in your house")
03:07:28 Quit RotAtoR ()
03:16:03***Saving seen data "./dancer.seen"
04:00
04:06:01 Join QT_ [0] (i=as@madwifi/users/area51)
04:14:47 Quit goa (Read error: 110 (Connection timed out))
04:16:25 Quit QT (Read error: 110 (Connection timed out))
04:26:37 Join Aramil [0] (n=tony@ipa151.0.tellas.gr)
04:30:03 Quit tvelocity (Nick collision from services.)
04:30:13 Quit Paul_The_Nerd ("Chatzilla 0.9.68a [Firefox 1.0.7/20050915]")
04:30:16 Nick Aramil is now known as tvelocity (n=tony@ipa151.0.tellas.gr)
04:51:05 Quit XavierGr (Read error: 110 (Connection timed out))
05:00
05:16:05***Saving seen data "./dancer.seen"
05:52:03 Join goa [0] (i=hd@212.87.148.33)
06:00
06:33:17 Quit webguest95 ("CGI:IRC (EOF)")
07:00
07:16:08***Saving seen data "./dancer.seen"
07:29:13 Join B4gder [0] (n=daniel@static-213-115-255-230.sme.bredbandsbolaget.se)
07:34:57Bgermorning :))
07:35:29B4gdermorning!
07:35:58Bgervery quiet night here :) unbelievable
07:37:50BgerB4gder: any news on logs ? (except that 04 oct is there) ?
07:38:04B4gdernope
07:39:46Bgerif I continue to ask these questions, i can very easily become "Mr. Where are the logs" :)
07:42:31B4gderand then we might very well have you teach you a lesson! B-]
07:42:36*B4gder grins
07:42:42B4gderhave to
07:42:50B4gdergee, can't type now either
07:43:04Bger:)
08:00
08:04:04 Join LinusN [0] (n=linus@labb.contactor.se)
08:08:37LinusNi have restored the missing logs between 0930 and 1003
08:08:49B4gdergood work LinusN
08:10:29 Join ender` [0] (i=ychat@84.52.165.220)
08:14:27SlasheriLinusN: hi, do you have had time yet to try the dircache? Do you think if there is anything bigger changes necessary to the architecture before it can be committed?
08:19:44LinusNbeen busy
08:20:39Slasheriah, ok :)
08:22:30amiconnLinusN: Thanks for restoring the logs :)
08:22:41LinusN:-)
08:23:01 Quit Zagor ("Client exiting")
08:24:11amiconnNow it's possible to refer to my idea I had on Friday. I'd like to hear some opinions, please check http://www.rockbox.org/irc/rockbox-20050930.txt 21:05:14...21:15:05
08:29:07LinusNwhere would that counter be?
08:30:15LinusNshould the filesystem code keep a counter for every dir it updates?
08:30:40amiconnNo, just one global revision counter that is updated if something is changed
08:31:41LinusNaha
08:32:06LinusNthat's a good idea
08:32:13B4gderI like it too
08:32:43amiconnThis still means some unnecessary re-reads, but it shouldn't need much resources, and it avoids passing around these falgs everywhere
08:32:51LinusNi like colddy's quit msg: "PG = Boy gets girl, R = Villain gets girl, X = Everybody gets girl"
08:34:35amiconnThe fs code knows best when something changes, e.g. opening a file with creat() doesn't necessarily mean the dir is changed - the file might have existed before
08:36:23LinusNbool filesystem_changed(unsigned long *counter)
08:36:44LinusNthe function compares and updates the counter you pass it
08:37:25amiconnYes, sounds good
08:40:54amiconnOnce this mechanism is in place, all the separate tracking (like the have_recorded variable in recording.c) can go away, and the plugin return value can be used for other things
08:41:45amiconnMenu return values can also be used for other things
08:46:21 Join Vlad0man [0] (n=Vladoman@p54A7DCAD.dip.t-dialin.net)
08:47:08 Join _FireFly_ [0] (n=icechat5@pd95b7c08.dip0.t-ipconnect.de)
08:47:24_FireFly_moin
08:56:52Bgerinsert coin
08:56:54Bger:)
08:57:30_FireFly_??
08:58:22*B4gder inserts
08:58:27B4gdernow dance dammit!
08:59:39 Join einhirn [0] (i=Miranda@bsod.rz.tu-clausthal.de)
09:00
09:01:34*Bger dances
09:01:53B4gderaaah, these toys are the best
09:04:54Bger;)
09:05:06 Quit Vladoman (Read error: 110 (Connection timed out))
09:09:41 Join linuxstb_ [0] (n=5343d4aa@labb.contactor.se)
09:10:22linuxstb_Morning all.
09:10:37_FireFly_moin :)
09:11:11linuxstb_Does anyone have any more thoughts about the viewer/plugin dilemma? Sudoku is now both, but the user has to browse to the .rockbox/viewers/ directory to start it as a plugin.
09:12:31linuxstb_There seem to be two workable solutions: 1) amiconn's idea of putting them both in the same directory, but with different extensions. Only the plugin extension would be a supported filetype.
09:12:50linuxstb_The problem is changing the build system to give plugins and viewers different extensions.
09:13:56linuxstb_or 2) Allowing viewers.config to point to .rock files in either the viewers/ or plugins/ directory. This would need a (small?) change to buildzip.pl to put the viewers in the right destination directory.
09:13:56LinusNnot a problem
09:14:22LinusNthere are other plugins that would benefit from a dual role
09:15:13linuxstb_Am I right in saying that the filename limit for viewer names is 512 bytes in total - for filenames and icon data?
09:16:12***Saving seen data "./dancer.seen"
09:16:16B4gderwe should remove the "*towav.rock" entries from there
09:16:36linuxstb_My suggestion is to change viewers.config to include the directory - e.g. plugins/sudoku or viewers/jpeg
09:16:36B4gderas we only have one of them left
09:16:39amiconnTelling the build system about the plugin type shouldn't be difficult.
09:17:19amiconnmakezip.pl processes viewers.config anyway, and nobody says the source viewers.config must have the same format as the target's viewers.config
09:18:18linuxstb_amiconn: The only drawback with your approach is that I have "show files" to all, and when I browse to .rockbox/plugins, I will see both viewers and plugins together.
09:18:33amiconnYes
09:18:58B4gderI think I favour approach 2 really
09:19:14amiconnUsing different dirs is perhaps better, but we still need to tell the build system about the type
09:19:27B4gderyes
09:19:28linuxstb_That's already done in viewers.config
09:19:50B4gderor the code would first check the viewers dir and then the plugins dir
09:20:24amiconnlinuxstb: You mean by the path? That should work...
09:20:30linuxstb_amiconn: Yes.
09:20:39amiconn...but only for plain plugin, plugin+viewer, and plain viewer
09:20:43linuxstb_B4dger: But then we need to give the build system extra information.
09:21:02B4gderyes, about where the plugin should end up
09:21:05amiconnFor a possible fourth type, hidden plugin, we would still need a separate mechansim
09:23:13B4gdermy suggestion how to proceed: 1) make the SOURCE viewers.config include the dest path, 2) make the TARGET viewers.config look like today 3) make the code search for viewers in first the viewers dir and then in the rocks dir
09:24:33amiconnHmm, keeping the (relative) path in the target viewers.config as well means less directory lookup
09:24:54B4gderyes, but uses slightly more space
09:25:07amiconnYes, 6 or 8 bytes per viewer
09:25:11B4gdernot saying that's an argument
09:25:24B4gderkeeping the path is clean
09:25:34B4gderand is more future-proof for other changes
09:25:57amiconnWe could instead shave off the .rock extension then
09:26:03B4gderindeed
09:26:33amiconn(less elegant though)
09:28:25linuxstb_Sounds good to me. But is it a big problem to keep the directory name in viewers.config? It's only 8 chars for each of the (currently) 24 viewers. How much extra code would it be to check both directories?
09:28:52B4gderwe only have 16 viewers now ;-)
09:29:11*B4gder removed the ones we don't build anymore
09:29:23linuxstb_Sorry - I forgot the multiple entries.
09:29:42linuxstb_So it's 128 extra bytes to store the full path (relative to .rockbox).
09:29:50B4gderI say we add the dir names
09:29:58B4gderand cut off the extensions
09:32:08B4gderanyone against?
09:32:54linuxstb_Not me.
09:32:57LinusNi'm for it
09:34:37linuxstb_bbl.
09:34:41 Quit linuxstb_ ("CGI:IRC")
09:34:41B4gderok, I'm on the case
09:44:07HClthe only good part about being sick
09:44:14HClis the feeling you get when you finally get better
09:44:25HCland you can stay at home watching anime because you're still not completely healthy
09:44:54 Quit _FireFly_ (Read error: 110 (Connection timed out))
09:53:54B4gderseems to work now
09:56:07B4gderthe viewers.config file is 45 bytes bigger
10:00
10:02:14B4gdercommitted
10:13:36*Bger fsck-ed up the linux's crypto/aes.c :)
10:32:07 Join ashridah [0] (i=ashridah@220-253-122-129.VIC.netspace.net.au)
10:33:02Bgerhttp://www.cits.rub.de/MD5Collisions/
10:34:04Bger2 VERY different .ps files with same md5sum
10:39:03Bgerat least it looks this way
10:47:29ashridahomg! hashes have collisions? never!
10:49:15ashridahshould be using sha1 anyway
10:49:26Bgersha256 better
10:50:03Bgerbtw, the 2 files differ verry little (in the header)
10:50:31Bgerbut the difference in the result is enormous :)
10:56:05 Join pilot000 [0] (n=51dd1b21@labb.contactor.se)
11:00
11:01:04 Join Paul_The_Nerd [0] (n=paulthen@cpe-66-68-93-2.austin.res.rr.com)
11:01:43 Quit Bger ("damned bx")
11:02:28 Join Bger [0] (n=Bager@83.222.160.88)
11:04:59 Join Paul_The_Nerd_ [0] (n=paulthen@cpe-66-68-93-2.austin.res.rr.com)
11:09:15 Join webguest52 [0] (n=446a3a97@labb.contactor.se)
11:09:42 Nick webguest52 is now known as phxson (n=446a3a97@labb.contactor.se)
11:11:25phxsonhmm kickass web irc client
11:11:55ashridahBger: of course, one could vastly decrease the incidents of collisions by using multiple hashes
11:11:58B4gderbut you can't type + ;-)
11:12:17B4gder(in the irc client that is)
11:12:24phxson
11:12:27phxsondang
11:12:27Bgerashridah yep
11:12:31B4gderhehe
11:13:03Bgerbut it's more easy to use longer hash
11:14:22ashridahthat also helps, but it's not perfect.
11:14:51Bgerashridah it helps for sure...
11:15:24Bgerand other (better) algorithm
11:16:13***Saving seen data "./dancer.seen"
11:20:07 Quit Paul_The_Nerd (Read error: 110 (Connection timed out))
11:20:22 Nick Paul_The_Nerd_ is now known as Paul_The_Nerd (n=paulthen@cpe-66-68-93-2.austin.res.rr.com)
11:25:23 Join Febs [0] (n=Febs@207-172-122-81.c3-0.rdl-ubr4.trpr-rdl.pa.cable.rcn.com)
11:29:45ashridahthat site is actually well timed. the students i tutor at uni have just finished doing an assignment involving md5 as a poor-mans tripwire
11:29:57ashridahand they were supposed to look into issues with md5sum
11:33:40Bgerhehe :)
11:34:33Bgertaken from en.wikipedia.org/wiki/MD5
11:34:38HClhmm
11:36:13 Quit pilot000 ("CGI:IRC (EOF)")
11:41:46 Quit linuxstb (Remote closed the connection)
11:48:25gromit`http://slashdot.org/article.pl?sid=05/10/04/147205&tid=154
11:48:56gromit`seems rockbox hasn't been the only victim...
11:55:08 Quit phxson ("too drunk for the tech-speak")
11:57:21 Join preglow [0] (n=thomjoha@hekta.edt.aft.hist.no)
11:57:22ashridahoh twikicide?
11:57:25ashridahs/oh/of/
12:00
12:09:28 Quit ze (Read error: 104 (Connection reset by peer))
12:10:09 Join ze [0] (i=ze@ca-dstreet-cuda1-c6a-130.snbrca.adelphia.net)
12:10:28*Bger needs a little c help :(
12:10:56Bgeranyone ?
12:12:11gromit`yep ashridah
12:12:56gromit`what kind of help Bger ?
12:13:26Bgeri don't understand something...
12:13:33Bgerit's like
12:15:29Bgeri for(j=0;j<9999;j++) { ... /* no j here */ } ; /* more source ..., no j used */ for (j=0;j<16;j++) { ... } ... in the second for cycle j is 9999?
12:16:13 Quit tvelocity (Remote closed the connection)
12:17:37Paul_The_Nerdj=0 is an assignment though, j should start at 0...
12:18:21Bgermore exactly www.cr0.net:8040/code/crypto/aes/aes.c line 824
12:18:43Bgeri think there's an overflow somewhere...
12:20:03Paul_The_NerdIn the second loop J starts at a value of 9999 for some reason?
12:20:19Paul_The_NerdIs that what you meant?
12:20:32BgerPaul_The_Nerd did you see it? yes, at least gdb says so
12:20:53Paul_The_NerdLooking at it, I don't see why it would...
12:21:08*Bger too
12:21:11Bgern=0
12:21:23Bgerm=0
12:21:39ashridahBger: wait, has gdb actually executed the initial condition?
12:21:44Paul_The_NerdI don't know how GDB works at all though, so I don't kno..
12:21:52Paul_The_NerdYeah, I was about to ask that in a roundabout way.
12:23:01Bgerashridah gcc -gdb aes.c -o aes.o ; gdb aes.o
12:23:05Bgerbreak main
12:23:07Bgerrun
12:23:17Bgeruntil <linenumber> ...
12:23:50ashridahBger: okay, is linenumber after the for() line, or on it?
12:24:05Bgerafter the for
12:24:32ashridahthat is weird
12:24:55Bgerhm, forge
12:24:56Bgert
12:25:00Bgeri'm stupid
12:36:05 Quit Paul_The_Nerd ("Chatzilla 0.9.68a [Firefox 1.0.7/20050915]")
12:37:01 Quit Febs (Read error: 110 (Connection timed out))
12:46:18Bgeram i right that 1) x86 is little endian; 2) in little endian le32_to_cpu() and cpu_to_le32() is equal to its parameter?
12:46:56B4gderit is little endian, yes, so such functions won't need to convert anything
12:47:19Bgerthen why the f... it doesn't work :(
12:48:24Bgerit doesn't pass test vectors and that's equal to it doesn't work...
12:52:48 Join XavierGr [0] (n=XavierGr@ppp15-adsl-30.ath.forthnet.gr)
13:00
13:07:17 Join tvelocity [0] (n=tony@ipa151.0.tellas.gr)
13:16:15***Saving seen data "./dancer.seen"
13:20:22BgerB4gder do you have little time ?
13:20:27B4gdersure
13:21:01Bgercan i give u something to look at ?
13:21:07Bgermaybe via mail
13:21:21B4gdercertainly, daniel at haxx.se
13:21:29Bgerok, 10x
13:38:34BgerB4gder i just sent it
13:39:22B4gdergot it
13:45:22B4gderits quite a job to figure out the flaw in that
14:00
14:04:28Bgeri suppose so :(
14:04:32Bgerbut
14:04:38Bgeru have the diff...
14:04:56Bgerbetween the kernel version (i suppose it's working) and mine...
14:06:07B4gderyes, but I couldn't spot any obvious problem using that
14:06:22Bgerme too :( that's the problem ....
14:07:58Bgeri also checked whether u can use aes_encrypt(ctx,buff,the_same_buff), seems ok
14:10:05Bgerhm, i didn't give you aes.h
14:11:44Bgeri mean byteorder.h
14:13:26Bgersent
14:14:13Bgerif you don't have time or don't want to see it, just don't do it
14:14:19Bgerthank you very much anyway
14:36:02 Quit B4gder ("time to say moo")
14:41:54XavierGrmoooo! :D
14:42:02Bger:))))
14:47:34Bgerthis "moo" doesn't have any meaning in swedish, i suppose ?
14:48:22 Join Paul_The_Nerd [0] (n=paulthen@cpe-66-68-93-2.austin.res.rr.com)
14:49:24Paul_The_NerdMornin' y'all.
14:50:10Bgerafternoon :)
14:55:58Paul_The_NerdI have the distinct curse of being in Texas... it's quite morning still here. =/
14:57:49 Join linuxstb [0] (n=d57b9aa9@labb.contactor.se)
14:57:53Bgerhehe:)
14:58:17linuxstbpreglow: This could simplify the Rockbox FLAC codec: http://www1.mplayerhq.hu/cgi-bin/cvsweb.cgi/ffmpeg/libavcodec/flac.c?cvsroot=FFMpeg
14:58:29linuxstbIt's an independently implemented flac decoder.
14:58:40linuxstbi.e. it isn't libFLAC.
15:00
15:03:51Bgerlinuxstb is this integer decoder ?
15:04:05 Join _FireFly_ [0] (n=FireFly@p54A450A0.dip.t-dialin.net)
15:04:17_FireFly_hi
15:04:40Paul_The_NerdHello
15:05:09linuxstbBger. Yes. All lossless decoders are integer (as far as I know).
15:06:03Paul_The_NerdI'd imagine that's almost a necessity.
15:06:36linuxstbIt stills needs extra code to parse the file headers completely - but that is almost done by the code in metadata.c already.
15:07:15Paul_The_NerdOh, I had a question: Why does the tagDB need to be prepared by an external program rather than a plugin? (Other than speed, I mean, is there any other good reason?)
15:07:59linuxstbProbably because it was easier to write it in Perl/Java than C.
15:08:10Paul_The_NerdAh
15:08:23linuxstbI would like to see a C version of it though - based on the code that now exists in apps/metadata.c
15:08:40Paul_The_NerdI was just wondering, because it only handles certain filetypes, whereas you've already got the utilities to get all the supported metadata built in.
15:08:46linuxstbI would write it myself, but there's too many other things on my to-do list.
15:09:27linuxstbWhen the tagdb was first written, metadata.c didn't exist. But by definition, metadata.c has routines to parse all formats that Rockbox can play. So the hard work is already done.
15:09:35Paul_The_NerdI have a long and established history of picking the most difficult tasks as a means of familiarizing myself with a new program, environment, or whatnot. Converting that may just qualify. ;-)
15:11:36Paul_The_NerdAt the very least it'll fill a large portion of my 'spare' time until I inevitably go mad.
15:12:35linuxstbPlease do. I'll be happy to help with any metadata.c questions. But I don't know the tag database format at all.
15:16:17***Saving seen data "./dancer.seen"
15:21:50amiconnThere are two main reasons why the database isn't created on target
15:22:21amiconn(1) Creating the db on target would be sloo-oow, especially on archos but also on iriver
15:23:04amiconn(2) You'll need sophisticated techniques like creating the db in chunks, doing merge sort etc due to the limited ram
15:23:19Paul_The_NerdAh
15:23:20ashridahyay, external merge sort!
15:23:46amiconnIt should be possible, just not very practical
15:24:01amiconnThere is already some C code for db creation in cvs
15:25:08amiconnIt's another of those half-baked things...
15:25:13Paul_The_NerdAH
15:25:15Paul_The_NerdAh, even
15:25:28Paul_The_NerdWell, at the very least I'll look at it.
15:26:33Paul_The_NerdRight now the db creation utilities that are on the site don't create a working one for me. I'm not sure if it's because I have mixed tags (Though I'm fairly certain the Ape and the ID3 tags have the same data) or if it just hates me.
15:27:09amiconnThis code was intended for double use: for building an exectutable host tool and a target plugin
15:27:34*Bger spots probable error ...
15:27:36linuxstbAnyone know how big a typical tag database is? I've never tried to create one.
15:28:12Paul_The_NerdI didn't really pay attention.
15:28:29amiconnlinuxstb: It highly depends on your music files's structure, not only the number of them but also the structure
15:29:08amiconnLike, what is the maximum number of tracks with the same album tag, the maximum number of tracks from the same artist etc
15:29:18linuxstbI know it's hard to define "typical". How big are your databases for example? Are we talking 500KB, 5MB, 50MB ?
15:29:31amiconnMine is ~1.4 MB
15:29:36Paul_The_NerdI seem to remember it being about 2.something on mine.
15:29:41Paul_The_NerdLike, 2.2mb or so.
15:29:50amiconn(on archos recorder and H1x0 with all my tracks)
15:30:00linuxstbSo at least on the iRiver, the 32MB RAM should be enough to build it?
15:30:13amiconnYou can't rely on that at all
15:30:26amiconn..and don't forget the H100
15:30:34Paul_The_NerdBut mine was... weird. Like, it'd have more than one entry for the same song, and such.
15:31:20Paul_The_NerdIs the database updateable? Like, how hard is it to add new data without rebuilding the whole file?
15:31:37*Bger is so stupidddddddd
15:31:48linuxstbI'm not trying to suggest that a plugin should be the normal way people update their databases, but it would be useful in some circumstances - e.g. if you want to quickly add some files, but not wait for the tagdb to update before leaving the PC.
15:32:23linuxstbOr if you wanted to delete some files, or add your own recordings.
15:32:32BgerBagder if you read the log: i'm really stupid. the problem is that the devine's aes_set_key routine takes the key size in bits, and the kernel's one takes it in bytes ....
15:33:08Bgerof course, the kernel's aes_set_key rejects the key ...
15:33:53Paul_The_NerdYeah, a way to apply some changes to the tagdb on target is almost necessary.
15:34:35ashridahI've got to agree. define a quick append section to the end of the database.
15:35:22ashridahhave the routines in rockbox process most of the db normally, and then apply the 'append' section after the fact. horrible performance if the append section grows, but potentially we should be able to do the small amoutn of sorting to include a few extras
15:35:55ashridahjust have it bitch that the append section is getting too long at some point
15:36:05*Bger cannot slallow his stupidness
15:37:55linuxstbI think an append section will just complicate things. The tagdb plugin could have two modes - a "build new database" mode which does everything from scratch, and an "update db" mode which uses the existing database for the metadata for existing files, and only parses new files.
15:38:13linuxstb"update db" will still need to scan the whole disk, but that can be done quickly using the dircache (if enabled).
15:38:53ashridahwell, true, i'm making assumptions about what's more complicated.
15:39:01linuxstbBut we need a working C database generator before we can think about a plugin.
15:40:09linuxstbashridah: I'm thinking that changes to core Rockbox (to deal with the append section) are the complications we want to avoid.
15:40:19ashridahhopefully i'll be at loose ends in two weeks. only got 4 exams over the course of five weeks, and a few assignments to mark.
15:40:27ashridahlinuxstb: fair enough.
15:40:50*ashridah could well be underestimating the sorting/juggling necessary
15:44:46linuxstbamiconn: Where in CVS is the db creation C code you talked about?
15:45:29amiconnapps/tagdb
15:46:21 Join gddd [0] (i=comar666@bji49.neoplus.adsl.tpnet.pl)
15:46:26gdddHi
15:47:06linuxstbamiconn: So none of the code in apps/tagdb is actually used in Rockbox at the moment?
15:47:07 Part gddd
15:47:13amiconnnope
15:47:17 Join ze__ [0] (i=ze@ca-dstreet-cuda1-c6a-130.snbrca.adelphia.net)
15:47:24Bgerquick way of getting time interval in C under linux (with better precision than sec) ?
15:47:26 Quit ze (Read error: 104 (Connection reset by peer))
15:47:31 Nick ze__ is now known as ze (i=ze@ca-dstreet-cuda1-c6a-130.snbrca.adelphia.net)
15:49:13ashridahBger: man clock_gettime
15:49:29ashridahdepends tho, some systems cannot return decent resolution, it's system dependent
15:49:55Bger10x
15:50:14ashridahpart of SUSv2/POSIX 1003.1 so it should exist in most modern unicies
15:50:36ashridahthere's possibly also the RTC stuff
15:51:03ashridahi don't know how accurate that is tho. most people end up using it for high resolution timing for syncing things like a/v, from what i've seen
15:51:14_FireFly_Bger: gettimeofday
15:51:18ashridahbut that's almost certainly linux specific
15:51:43ashridahgettimeofday only goes to microseconds :)
15:52:22 Join tucoz [0] (n=81b17b04@labb.contactor.se)
15:52:30Bgergettimeofday is enough :)
15:52:37ashridahbah
15:52:45ashridahgive me nanoseconds or give me death
15:53:02tucozhi, grrrrr. What is it with macs and their will to put hidden files all over the plays. My player is infested with .DS_store files
15:53:08ashridah(or at least give me a profil()alike interface that has better resolution than 10ms sample period
15:53:15tucozs/plays/place
15:53:25ashridahtucoz: the same reason windows uses Thumbs.db ?
15:53:26linuxstbtucoz: I can simplify. And the ._filename.ext files containing the resource forks.
15:53:32ashridahand System Information Whateveritis
15:53:46tucozit is irritating
15:53:49linuxstbs/simplify/empathise/ - not sure where that typo came from.
15:54:08ashridahthe keys are like right next to each other
15:54:17tucozis it possible to turn that "feature" off?
15:54:55linuxstbI don't know. It's annoying to do a "cp" in bash and have two files created on the target filesystem.
15:55:01tucozhehe
15:56:00tucozOr having files created just by browsing the iriver
15:56:07Paul_The_NerdIsn't there a plugin in the patches that cleans up all that extra crud?
15:56:40tucozashridah: I do not think that windows creates Thumbs.db if view is set to something other than view thumbnails
15:57:22tucozprobably tons of scripts that take care of this though.
15:58:29amiconnThumbs.db is created if you use the bultin image viewer and don't enable "Don't cache thumbnails" (or whatever it is called in english Windows)
15:58:51amiconnProbably rockbox should have a cleanup plugin, which deletes all windows and mac clutter
15:59:20amiconn...like Recycler/recycled, System Volume Information, thumbs.db files for windows
15:59:24tucozI do not think that Thumbs.db is as bad as that mac clutter.
15:59:34Paul_The_NerdThere is a cleanup plugin, rasher has it in his build.
15:59:38Paul_The_NerdLemme see if I can find it
15:59:55tucozah, sounds nice
16:00
16:00:02linuxstbI agree - the Mac clutter more than doubles the number of files on your disk. And they are all playable as well because they have the same extensions.
16:00:15tucozoh, really? sounds bad
16:00:22Paul_The_Nerdhttp://sourceforge.net/tracker/index.php?func=detail&aid=1243966&group_id=44306&atid=439120
16:00:53amiconnSystem Volume Information can be rather big as well. Recently I deleted that on my archos from within rockbox. Took more than a minute...
16:00:56tucozPaul: thanks
16:00:58linuxstbA file "mysong.mp3" will end up as two files on a fat32 disk - "mysong.mp3" and "._mysong.mp3"
16:01:26tucozlinuxstb: do you know why osx handles it that way?
16:01:34amiconnI already told my windows to disable system restore on these volumes, but windows is stubborn
16:01:38linuxstbBecause it wants to keep the resource forks.
16:02:18tucozwhich is?
16:02:51_FireFly_I have a question: Should i put all remote dependent options, which aren't in the menu "Remote LCD Settings" in a seperate menu or simply add to the same place where the same option for the main lcd resists ??
16:03:17linuxstbtucoz: No idea :). You have called my bluff and discovered my lack of Mac knowledge.
16:04:14tucozhehe
16:04:24LinusN_FireFly_: i dunno, do what you feel is best
16:05:58_FireFly_ok
16:07:16amiconn_FireFly_: Imho some options don't make sense to have them separated by display (like the bidi text support)
16:07:55_FireFly_ok i wil do it the same as i did it already for scroll/statusbar :)
16:12:03amiconnHmm, currently the remote lcd drive doesn't support bidi text, but that's easy to fix
16:12:13amiconn*driver
16:15:34XavierGrscrolling and statusbar should be saperate imho.
16:15:51_FireFly_it seems that i have only add this if (bidi_support_enabled)
16:15:51_FireFly_ str = bidi_l2v(str, 1); to lcd_putsxyofs()
16:16:04_FireFly_scoll and statusbar is seperate
16:16:25_FireFly_only timeformat and bidi will be not seperate
16:16:28novimonwhat does the scrolling bar do anyway?
16:16:30Bgerstupid q.: clock_gettime's man page says i need to link with librt ... so, how do i make it ? gcc -llibrt ??
16:16:45_FireFly_gcc -lrt
16:17:00_FireFly_when the lib name is librt.so
16:17:14_FireFly_the lib filename
16:17:17amiconn_FireFly_: Yes, just these 2 lines are missing, and an #include "bidi.h" at the top of the file
16:17:26Bger_FireFly_ 10x
16:17:29_FireFly_ok i will add
16:17:56Bgerand if it's not libxxx.so ?
16:18:36 Part tucoz
16:19:09_FireFly_Bger: the standard nameing schema for libs is libxxx.so for dynamic libs and libxxx.a for static libs
16:19:35Bgeri just ask
16:24:25preglowlinuxstb: know if that flac implementation is any good?
16:24:39novimonI'm about to try it out
16:24:51novimonripping my cds to flac
16:25:12novimongot my alessandro MS-1 today :)
16:26:14_FireFly_i have some flac files and i have no problems to hear to these files
16:26:24preglowit's _tiny_ compared to libflac
16:27:24preglowthere's always the problem of having to rip it out of libavcodec, of course
16:27:30 Quit ashridah ("Leaving")
16:31:05 Part LinusN
16:34:35novimonhmm
16:34:44amiconnpreglow: I hope it's not using the pretent-we-have-C++ style
16:34:46novimonanyone know how to get rid of the recylebin?
16:34:48amiconn*pretend
16:34:54novimonit alaways comes back
16:35:38preglowamiconn: it's not
16:36:34amiconnnovimon: Yes. Click the desktop recycle bin with the right mouse button, and select 'properties'
16:36:48novimonamiconn, did that
16:37:02novimonit keep coming back
16:37:10linuxstbpreglow: I only just saw it today for the first time. I don't know how good it is.
16:37:12amiconnThen set 'configure drives independently' (dunno exactly what it's called
16:37:13amiconn)
16:37:27novimonmaybe the file "recyle" must be there, but it doesnt actually use it
16:37:49amiconnHmm, possible
16:45:31 Quit t0mas (" Like VS.net's GUI? Then try HydraIRC -> http://www.hydrairc.com <-")
16:48:31 Quit Maxime ()
16:52:27 Join Maxime [0] (n=flemmard@fbx.flemmard.net)
16:55:38 Quit Paul_The_Nerd ("Chatzilla 0.9.68a [Firefox 1.0.7/20050915]")
16:56:23 Join keithhub [0] (n=keithhub@cpe-066-057-064-152.nc.res.rr.com)
17:00
17:04:40 Quit keithhub ()
17:06:48 Join dpassen1 [0] (n=dpassen1@resnet-233-61.resnet.umbc.edu)
17:11:04 Join keithhub [0] (n=keithhub@cpe-066-057-064-152.nc.res.rr.com)
17:16:18***Saving seen data "./dancer.seen"
17:17:12 Quit einhirn ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
17:17:21 Join einhirn [0] (i=Miranda@bsod.rz.tu-clausthal.de)
17:36:18 Quit ebone_ ("leaving")
17:52:10 Quit tvelocity (Connection timed out)
17:53:04 Quit keithhub ()
17:56:53 Quit einhirn ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
17:59:42 Join einhirn [0] (i=Miranda@bsod.rz.tu-clausthal.de)
18:00
18:14:35 Join Moos [0] (i=DrMoos@m14.net81-66-158.noos.fr)
18:32:24 Join t0mas [0] (n=Tomas@unaffiliated/t0mas)
18:37:58 Quit linuxstb ("CGI:IRC (EOF)")
18:42:40t0mashi
18:42:49t0masBagder of Zagor around now?
18:42:52t0mas*or
18:45:53XavierGrhi
18:46:05thegeek_heh
18:46:09thegeek_Bagder of Zagor
18:46:18t0masyeah... of = dutch for OR
18:46:26thegeek_ah;)
18:46:29t0masand the F is close to the R :)
18:46:39t0masso 2 reasons to make it of in english :P
18:46:40thegeek_just sounds kinda... medieval
18:46:43thegeek_mhm
18:46:46t0masghehe
18:46:53t0masor startrek like...
18:46:56t0mas7 of 9
18:47:11thegeek_;)
18:47:38t0mas(I typed of again btw...)
18:47:41t0masof startrek
18:47:42t0masghehe
18:48:10*t0mas should setup auto-correction
18:48:20 Join matsl_ [0] (n=matsl@1-1-4-2a.mal.sth.bostream.se)
18:48:42XavierGrin mirc?
18:49:02t0masno... Xchat
18:49:07t0masoh... wait... HydraIRC now
18:49:09t0masI'm on windows
19:00
19:01:21 Join Paul_The_Nerd [0] (n=paulthen@cpe-66-68-93-2.austin.res.rr.com)
19:16:23***Saving seen data "./dancer.seen"
19:25:01 Join Lear [0] (n=chatzill@h73n11c1o285.bredband.skanova.com)
19:25:32Paul_The_NerdNew Apple product to be announced on the 12th apparently. Rumors fly it may be iPod video.
19:25:43LearSlasheri: Would you mind if I commited a removal of playlist_delete from playback.c?
19:29:04amiconnLear: You are going to fix that rockbox messes with my playlist behind my back?
19:29:11amiconnIf so, I'm all for it :)
19:29:23LearThat was the idea, yes...
19:29:43Paul_The_NerdSo, what's it going to do?
19:29:49LearI have code that I've been running for a while now, but I haven't tested it much with "known bad" playlists...
19:30:05LearJust skip tracks it can't find, rather than call playlist_delete() on them...
19:31:06Paul_The_NerdAh
19:31:45MoosSlasheri: ...and what's about your dircache ?
19:32:14 Quit einhirn ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
19:33:28amiconnLear: Does your code handle skipping forward & back?
19:33:57LearYes, that should work...
19:34:31amiconnreboot, brb
19:34:33 Part amiconn
19:35:40SlasheriLear: please do so, but keep in mind that i have a real fix for that waiting.. i just haven't had time yet to finish it
19:36:46LearWell, my change only deals with one function... You have any idea when you might get the time? If it's fairly soon, there's not much point in me doing it. :)
19:36:57Slasheriit marks bad songs instead of delete and the playlist_keep skips them automatically. That gives a correct behaviour with all skipping
19:37:15Slasherii can try to do it soon.. :)
19:37:25Slasheri*playlist_peek
19:37:37LearAh, yes, my solution is simpler than that...
19:37:52Slasheriif you want to do it the other way, you must handle also playlist_next() etc.
19:38:03Slasherior you will end up with wrong metadata
19:39:47Slasherihmm, or in fact incorrect skipping or loading of songs
19:40:33 Join amiconn [0] (n=jens@p54BD4CFE.dip.t-dialin.net)
19:40:38XavierGrSlasheri: Any idea when your dircache patch will be comitted?
19:40:45XavierGrDid you fix the rename bug?
19:41:13SlasheriXavierGr: i am working on it but i don't have much time at the moment
19:41:28LearOnly real problem I've noticed is that availability of next track metadata can be delayed...
19:41:33Slasherii think i can commit it this week unless anybody is against that
19:41:54SlasheriLear: how have you done the skipping?
19:41:59LearThough I've noticed slightly odd behaviour a few times, but I don't know if my changes are causing that...
19:42:18Slasherithere is some global variables to correct the track peek index?
19:42:24LearStep peek_offset in audio_load_track if file isn't available.
19:43:23SlasheriHmm.. what about if the invalid entry is middle of the current tracks being buffered? Or if you skip tracks back/next and the next track would be the invalid one
19:43:31Slasherii don't think that is enough
19:44:30LearIt just skips them...
19:44:35Slasheriprobably the easiest solution would be to do the marking on playlist level
19:44:49Learyes, that does avoid some problems...
19:48:32LearOtoh, playlist code is a bit hairy, so I went for what I hoped to be a simpler fix. :)
19:50:21LearAnyway, I can wait for a better fix from you.
19:53:12SlasheriBut lets think about this (V=valid track, I=invalid track): current_track = 1, buffer data: V0 V1 V2 I3 V4 V5. It will load tracks to V2 normally, then step the peek_offset and load to V5. When skipping to V2, decreasing peek_offset and calling playlist_next(1). Still ok. Then what about when V2 is finished? We continue to V4, call playlist_next(1) and decrease peek_offset again. When we are finally end of the current buffer, peek_offset would be one more ...
19:53:18Slasheri... than it normally would be (this shouldn't cause problems at this point but it will later, for example when skipping to next folder or probably when skipping back). Anyway, sooner or later the offsets will be messed up
19:59:08amiconnThe archos playback engine contains a mechanism to handle missing/invalid playlist entries. It does work well. Why not reuse it?
19:59:10LearAnd load V4 you mean? But no global peek_offset is decreased, so I don't see the problem really. The peek offset in the tracks array should still be right?
20:00
20:00:18LearHm.. Odd behaviour: In the middle (?) of a song, playback stopped unexpectedly, leaving the wps. I went to the audio thread debug screen and notice that the pcm buffer is empty but the codec buffer is full...
20:00:37LearI only think I've seen behaviour like that with Vorbis files, for some reason.
20:01:39LearSorry, the codec buffer isn't full, it's overflowing (like 39 MB...). :)
20:05:35Slasheriamiconn: we would need to reimplement that on iriver or we could the playlist and also simplify and unify the archos skipping system
20:05:54Slasheri+implement it to
20:06:10amiconnAdding data to each playlist entry would be a bad idea
20:06:31amiconnCurrently each playlist entry is just one index (4 byte)
20:06:38Slasheriamiconn: we wouldn't "add" anything, just using one more attribute bit
20:06:49amiconnRemember that playlists up to 20000 entries are supported
20:06:54Slasheriand reducing maximum playlist size to 256 MiB (should be enough for everybody)
20:07:06Slasheriyes, hmm..
20:08:31amiconnI think the archos way of handling it is quite straightforward. The buffering code is what notices when a file is missing, and so it is the buffering code that handles them
20:09:13amiconn(by setting the load_ahead_index field in the track info structure accordingly)
20:09:29LearSounds like my fix, really... :)
20:09:50SlasheriHmm, indeed.. It is possible to do it that way.. I just see it more complex than the playlist way.
20:09:59SlasheriLear: But please do it if you would like to :)
20:10:38LearHave you actually done any coding on your fix (which sounds more complete)?
20:11:02Slasherisome coding, but not so much you shouldn't try the other way
20:12:38 Join Miq1 [0] (n=Micha@p54B1CD4D.dip.t-dialin.net)
20:13:19 Part Miq1
20:13:29Slasherithere is also one good thing with the playlist way that it could be possible to mark the entries found to be incorrect also in the playlist viewer
20:13:49 Join Miq1 [0] (n=Micha@p54B1CD4D.dip.t-dialin.net)
20:15:48 Join arkascha [0] (n=arkascha@xdsl-213-168-118-134.netcologne.de)
20:15:55 Join _arkascha [0] (n=arkascha@xdsl-213-168-118-134.netcologne.de)
20:16:01 Quit arkascha (Remote closed the connection)
20:16:10 Quit _arkascha (Remote closed the connection)
20:16:28 Join arkascha [0] (n=arkascha@xdsl-213-168-118-134.netcologne.de)
20:16:32 Quit Paul_The_Nerd ("Chatzilla 0.9.68a [Firefox 1.0.7/20050915]")
20:19:30 Part Miq1
20:20:26 Join Link9064 [0] (n=41ae2315@labb.contactor.se)
20:22:58 Join Miq1 [0] (n=Micha@p54B1CD4D.dip.t-dialin.net)
20:23:43Link9064does anyone know what volt ac adaptor an FM recorder 20 use?
20:24:27 Quit Link9064 (Client Quit)
20:26:08 Part Miq1
20:37:58 Join solexx_ [0] (n=jrschulz@d157200.adsl.hansenet.de)
20:42:43_FireFly_ok most of the menus are working on the remote-display
20:44:07_FireFly_i will upload a compiled version for h120/140 and ihp-100/110/115 in a few minutes
20:48:28_FireFly_ok the zip-files can be found on home.arcor.de/s.wezel
20:49:55XavierGrwhat is not working?
20:50:10_FireFly_only the debug-menu-entrys and the playlist-viewer and all plugins won't display somethin on the remote
20:50:13 Quit solexx (Read error: 113 (No route to host))
20:50:36_FireFly_the debug-menu-entrys show currently only a "not enought screen space"-message
20:50:52_FireFly_but the button to leave the entry works
20:51:06amiconnDuplicating the debug menu doesn't make sense, imho
20:51:38XavierGrcan you upload a patch too?
20:53:03_FireFly_yes in few minutes
20:53:48_FireFly_the zip-file will contain all patches, which are applied in the compiled versions
20:54:30XavierGrwhy not have 1 big patch?
20:54:41XavierGr1 patch with all changes I mean
20:55:57_FireFly_because the other patches are the wps-sb-patch RemoteLock-patch and remote-hold-fix-patch
20:56:05_FireFly_the main remote-patch is one file
20:56:45XavierGroh yes I mean the remote only patch.
20:57:46_FireFly_the other patches can be found on the tracker but these wouldn't apply cleanly when useing the remote-patch
20:58:30_FireFly_afaik
20:59:05 Join Miq1 [0] (n=Micha@p54B1CD4D.dip.t-dialin.net)
20:59:42_FireFly_XavierGr: so do you want the remote-patch itself or the other too
20:59:44_FireFly_?
21:00
21:00:29 Part Miq1
21:04:47_FireFly_the id3-viewer doesn't display something on the remote
21:04:50_FireFly_also
21:05:54 Join Chamois [0] (n=Florian@AStrasbourg-252-1-40-169.w83-196.abo.wanadoo.fr)
21:06:56_FireFly_ok the patch itself and the zip-file(if someone want the others too) are uploaded
21:07:10_FireFly_they can be found on home.arcor.de/s.wezel
21:16:26***Saving seen data "./dancer.seen"
21:22:52LearWell, fixing the id3 viewer ought to be easy...
21:24:00 Join Nixsos [0] (n=Nixsos@i155063.upc-i.chello.nl)
21:25:09_FireFly_oh some bugs found quick-menu doesn't react on some remote-buttons and enter the quickmenu form file-tree doesn't work
21:27:30 Quit arkascha (Remote closed the connection)
21:36:00 Join phaedrus96 [0] (n=bob@ppp-69-229-251-93.dsl.bkfd14.pacbell.net)
21:36:59 Join webguest34 [0] (n=53afb0c2@labb.contactor.se)
21:41:46_FireFly_ok bug fixed updated compiled version and the bugfix-patch can be fetched from my webspace in few minutes
21:42:16_FireFly_ok all uploaded
21:44:57 Quit phaedrus961 (Read error: 110 (Connection timed out))
21:45:49 Quit Lear ("Chatzilla 0.9.68.5.1 [Firefox 1.4.1/undefined]")
21:48:31 Join muesli- [0] (i=muesli_t@Bbc87.b.pppool.de)
21:49:21muesli-high
21:49:30_FireFly_hi :)
21:49:42muesli-hi _FireFly_ :)
21:49:46_FireFly_what is high ;)
21:49:57muesli-me..maybe ;)
21:49:57 Nick phaedrus96 is now known as phaedrus961 (n=bob@ppp-69-229-251-93.dsl.bkfd14.pacbell.net)
21:50:24 Quit Chamois (" Try HydraIRC -> http://www.hydrairc.com <-")
22:00
22:05:43muesli-_FireFly_ what have changed in your rls?
22:07:24_FireFly_rls ?
22:07:31muesli-release
22:08:03_FireFly_look at http://forums.rockbox.org/index.php?topic=1549.90
22:09:09muesli-kewl, cheers mate :D
22:09:45 Join DangerousDan [0] (n=Miranda@newtpulsifer.campus.luth.se)
22:16:25 Join ]RowaN[ [0] (i=a2b0y@82-43-214-84.cable.ubr10.newm.blueyonder.co.uk)
22:18:09]RowaN[guys what do you think about having an option to specify a background audio file, to play in times that would usually be silent.. like the way my TV has ambient music in the EPG, u could specify your own unobtrusive mp3 (wind chimes.. whale song hehe).
22:19:37XavierGrwell what if someone wants to stop that.
22:19:48]RowaN[like i say, an option
22:19:52XavierGrAnother option?
22:20:03]RowaN[hide it away, like the others
22:20:26XavierGrI don't like the idea, but let's hear another opinion on this.
22:21:11]RowaN[i bet its one of those features that a lot of people poo-poo, but they would like it if they could try it
22:22:05 Join Zagor [0] (i=foobar@pdpc/supporter/sustaining/Zagor)
22:23:05muesli-sounds nice...
22:23:15muesli-if you ask me..useless bt nice ;)
22:23:18muesli-u
22:23:41]RowaN[useless compared to the other "cancer-cure" features of rockbox, yes
22:25:35muesli-a nice inconspicuous beeping gameboy sample would be nice
22:25:51XavierGris firefly's page down?
22:26:47muesli-nope
22:26:52muesli-http://home.arcor.de/s.wezel/
22:26:54muesli-worx
22:27:04muesli-hi XavierGr btw
22:27:13_FireFly_i have no page ;)
22:27:26_FireFly_only webspace
22:27:39muesli-haarspalter ;)
22:27:46 Quit XavierGr (Read error: 113 (No route to host))
22:28:02_FireFly_stimmt doch ;)
22:28:10muesli-;)
22:28:47muesli-just write an index.html and you got one ;)
22:28:52muesli-get
22:29:13_FireFly_why should i ??:)
22:30:08muesli-just 2 be proud of having a website :D
22:32:40_FireFly_wow my old page about Proxy+, which i had used years ago is still online
23:00
23:02:45 Quit Ismo (clarke.freenode.net irc.freenode.net)
23:02:45NSplitclarke.freenode.net irc.freenode.net
23:02:45 Quit pilophae (clarke.freenode.net irc.freenode.net)
23:03:55 Quit matsl_ (Remote closed the connection)
23:04:47NHealclarke.freenode.net irc.freenode.net
23:04:47NJoinpilophae [0] (i=pilophae@montezuma.acc.umu.se)
23:04:59NJoinIsmo [0] (i=laitinei@huippu.net)
23:16:30***Saving seen data "./dancer.seen"
23:18:43 Quit webguest34 ("CGI:IRC")
23:18:53 Quit _FireFly_ ("Leaving")
23:21:33 Quit muesli- (Read error: 110 (Connection timed out))
23:33:56 Quit ender` (Read error: 113 (No route to host))
23:39:27 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
23:48:39 Join ashridah [0] (i=ashridah@220-253-122-87.VIC.netspace.net.au)
23:48:48 Quit dpassen1 ()
23:53:53 Quit Zagor ("Client exiting")

Previous day | Next day