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 2008-03-26

00:00:06disorganizerwe could rotate viewports dynamically on the rwps
00:00:14linuxstbdisorganizer: Remember Rockbox runs on DAPs, not graphic workstations...
00:00:25linuxstbAnd battery life is valuable ;)
00:00:40*disorganizer wants a 3d wps
00:00:53amiconndisorganizer: You can't easily rotate graphics, *especially* not on low-depth displays
00:00:57Lloreanlinuxstb: You must admit it'd solve the problem for people wanting portrait vs landscape displays. :-P
00:01:39disorganizerwell, to be serious: 90deg rotation could be easy by just swapping x+y values in the viewport drawing routines
00:01:56 Join cbr|w [0] (n=cbr@
00:02:24 Quit matsl ("Leaving")
00:02:26linuxstbThere are no "viewport drawing functions" - just LCD drawing functions. And they're designed to draw in the native direction.
00:02:35disorganizerit may be needed to do a lcdwidth-y and lcdheight-x on some target depending on usability
00:02:47 Quit moos (Read error: 110 (Connection timed out))
00:03:03amiconndisorganizer: Check the drawing functions. It is *far* more complicated than you might think to rotate stuff
00:03:41amiconnDrawing isn't done pixel-by-pixel - that would be far too slow on practically any target
00:04:09*disorganizer admits he has no clue about coding and just talks as a user here :-) hoping to bring in new ideas
00:05:07disorganizersounds so easy :-( on first thought ( @jdgordon: like making a backup copy of the config )
00:05:14 Quit tvelocity (Remote closed the connection)
00:05:16 Quit jas6180 (Connection timed out)
00:05:43amiconndisorganizer: The idea isn't new at all, and was discussed numerous times
00:05:50 Quit desowin ()
00:06:21disorganizeramiconn: it was new to me :-)
00:07:39*disorganizer likes to go on people's nerves
00:08:02linuxstbI think runtime rotation could be done via the screens API (replacing the drawing functions with their rotated equivalents) - if all of Rockbox fully used the screens API. But I'm not convinced the feature is worth the effort - at least I won't attempt it...
00:08:43amiconnThat means duplicating all drawing code - *and* all native bitmaps
00:09:08amiconn(unless you're willing to accept a serious performance drop in the rotated case)
00:09:11Bagderarbitrary rotation!
00:09:13BigBambidisorganizer: Why?
00:09:16linuxstbYes, you would duplicate the drawing code. The bitmaps wouldn't need to be duplicated though - just less efficient in one direction...
00:09:34amiconns/less/far less/
00:09:39disorganizernevertheless it would be an easy way to allow vertical progress bars and peakmeters :-)
00:09:45disorganizerBigBambi: why what?
00:09:55linuxstbdisorganizer: No, it would be the hard way to implement those...
00:09:57amiconnDid you have a look at the bitmap drawing optimisations in the packed pixel case?
00:10:01BigBambiWhy do you like to annoy people? What does it achieve?
00:10:44disorganizerBigBambi: hopefully i dont annoy them. going on other peoples nerves is not exactly annoying them ;-) just a tiny bit different
00:10:54*disorganizer hopes he doesnt annoy any1 here?
00:10:55BigBambiNo, it is the same thing
00:11:09BigBambiGetting on someone's nerves means annoying them
00:11:15scorche|shdisorganizer: by saying "any1", i can now say that you most certainly do :)
00:11:38linuxstbamiconn: OK, I accept it's far less efficient. But that could be dealt with, e.g. by possibly rotating any loaded bitmaps in-place when rotating the screen.
00:11:45disorganizerscorche: that was my personal ping to you ;-) like the cul8r when i left work
00:12:14BigBambiIt annoys many more people than scorche|sh
00:12:21*linuxstb isn't sure how many bitmaps we would be dealing with, or if performance would be an issue (outside the fonts)
00:12:23amiconnlinuxstb: That would require a little slack ram at the end of each bitmap. The rotated bitmap might need more ram depending on its x and y dimension
00:12:35amiconnAnd what about the compiled-in bitmaps?
00:12:59linuxstbamiconn: Is performance an issue for those? e.g. the usb logo?
00:13:13amiconnbuilt-in icons, ....
00:13:13*disorganizer didnt know fun was not on topic in here ;-)
00:13:35BigBambidisorganizer: Rockbox is on topic
00:13:55BigBambi#rockbox-community is for fun
00:14:05linuxstbamiconn: These problems are solvable though IMO. But as I said, I don't think the feature is worth the effort...
00:14:19linuxstbSo there's no need to convince me not to do it ;)
00:15:18 Join stripwax [0] (
00:16:11*linuxstb wonders how stripwax's cube hacking is coming along
00:16:31 Join framo [0] (
00:16:47preglowcube hacking?
00:16:50preglowgourad shading? :>
00:17:10 Quit miepchen^schlaf ()
00:17:26linuxstbHe was talking about the ships from Elite last time I spoke to him...
00:17:28*stripwax has a cobra mk3 spinning on his ipod :)
00:17:46stripwaxflat shading still though. hardly magic..
00:17:59 Join LambdaCalculus37 [0] (
00:19:01 Quit CyBergRind|w (Connection timed out)
00:19:34stripwaxbut (if I ever have time) I'd add a simple (flat) light source shader . gouraud would be .. a bit more work ..
00:19:45 Quit barrywardell ()
00:20:14ShaidI have a feeling the lack of floating points would make it a little harder.
00:20:39stripwaxI was looking at some interesting arm routines for computing 1/sqrt using fixed point math..
00:21:11stripwaxwasn't Descent written using fixed point math or did I dream that?
00:21:20Shaidno idea
00:21:21 Join miepchen^schlaf [0] (
00:21:23Shaidgo grab the source and see
00:21:51stripwaxIs USB storage enabled in a standard build?
00:22:09disorganizergot to go, good night to all of you
00:22:29stripwaxOn my local build I can see that my ipod 5g is requesting 500ma, I think I saw someone asking about that suggesting that it wasn't the case
00:22:47preglowstripwax: well, if you have lambert shading with a proper light source, gourad shading isn't much more work
00:23:08 Part disorganizer
00:23:23stripwaxpreglow - indeed. I haven't got lambert shading :)
00:23:47preglowaha :)
00:23:55preglowand then, after that, phong!
00:24:32*stripwax chuckles
00:25:19 Join `Tanner [0] (
00:25:24 Join LambdaCalculus3 [0] (n=1800dac6@gateway/web/cgi-irc/
00:25:41shotofaddsstripwax: there has been a fixed-point Descent, because I remember an Acorn (RISC OS) version being released some years ago. wonder if there's an open-source version available....
00:25:44jhMikeSwhat, no ray tracing?
00:25:56`TannerCan i play music that i purchased from the iTunes store while running Rockbox?
00:26:08LambdaCalculus3`Tanner: Nope.
00:26:15stripwaxTanner - are they in unprotected mp3 format?
00:26:27stripwaxor some other unprotected (no DRM) format?
00:26:32`TannerJust a sec.
00:27:33`TannerMeh. Mpeg-4 Audio file(Protected) is what i get... Dam
00:28:20 Join cool_walking_ [0] (
00:28:26stripwaxTanner - yep - I'm afraid you'll be out of luck in Rockbox with those
00:29:01`TannerAlright. But MP3 will work fine, right
00:29:13LambdaCalculus3As long as they're DRM-free.
00:29:22`TannerWhat does DRM stand for
00:29:38LambdaCalculus3Digital Rights (Restrictions) Management.
00:30:24`TannerWhen i boot into Rockbox, i won't see itunes purchased music., But when i boot back into standard iPod firmware, they will play?
00:30:52LambdaCalculus3They'll still play in the iPod OF.
00:30:59`Tannerok, Thanks.
00:31:03LambdaCalculus3No problem.
00:31:07 Quit `Tanner ("ChatZilla 0.9.81 [Firefox]")
00:32:32 Quit shotofadds (Read error: 104 (Connection reset by peer))
00:33:38 Join `Tanner [0] (n=4a809077@gateway/web/cgi-irc/
00:33:38 Quit `Tanner (Client Quit)
00:34:32 Quit LambdaCalculus37 (Read error: 110 (Connection timed out))
00:34:36 Join MethoS-- [0] (
00:35:34*linuxstb looks for a volunteer to implement target-specific %V tag validation into checkwps
00:35:44 Nick LambdaCalculus3 is now known as LambdaCalculus37 (n=1800dac6@gateway/web/cgi-irc/
00:36:21linuxstb(and loading bitmaps to validate them with Rockbox's bmp loader and image buffer size...)
00:36:29 Join [1]JordanG [0] (n=JordanG@
00:36:49*linuxstb suspects building 28 versions of checkwps will be simpler...
00:37:13*LambdaCalculus37 looks for a volunteer to help test FS #8595
00:38:18 Quit [1]JordanG (Client Quit)
00:41:02 Quit waldo (Remote closed the connection)
00:42:45 Join [2]JordanG [0] (n=JordanG@
00:43:35 Quit JordanG (Nick collision from services.)
00:43:51 Nick [2]JordanG is now known as JordanG (n=JordanG@
00:44:38Nico_Plinuxstb: can't you just accept screen dimensions as an argument to it?
00:44:58linuxstbNico_P: Yes, but the parser is full of #ifdefs, not if()s...
00:45:21linuxstbSo I think it would get very messy...
00:45:30Nico_Poh yes indeed
00:47:27linuxstbAnd looking at it, it's more than just the viewport tags - other parts of the parser are #ifdef'd out - e.g. for mono screens.
00:47:32 Quit ender` (" If the art world was ran by the RIAA, you would have to lease special glasses if you wanted to see a painting. -- Rodney Ca")
00:48:32linuxstbSo I think the only way to get 100% accurate parsing for is to build target-specific versions...
00:49:25linuxstbDoes anyone have any views on the WPS being rejected if bmps can't be loaded?
00:49:45linuxstbIIUC, currently missing bitmaps will just not be drawn.
00:51:22 Quit mirak (Remote closed the connection)
00:51:55Nico_Pyes, you understand correctly
00:52:31Nico_Pmaybe a warning in checkwps would be good, but I don't think plainly rejecting the WPS on target is very nice
00:53:03linuxstbWhy not? IMO that's just as broken as anything else being wrong...
00:54:40Nico_Pit's not quite the same, because e.g. a missing end conditional could cause unexpected behaviour... in case of a missing picture there won't be any big problems
00:54:57linuxstbBut it's still broken...
00:55:12 Join Chaotic [0] (n=0cb22419@gateway/web/cgi-irc/
00:55:13linuxstbi.e. not what the theme author intended.
00:55:19 Quit bluebrother ("leaving")
00:55:27 Join stripwax_ [0] (
00:55:53*Llorean likes "if it's broken, reject it"
00:55:59LloreanI don't like silent failures.
00:56:28ChaoticOk, so I just got back from my meeting, and was wondering if anyone would mind helping me out with my iPod 4G Photo power problem
00:56:59linuxstbChaotic: I wasn't paying close attention when you were first here, but do you get problems when running the Apple firmware?
00:57:23Chaoticlinuxstb: No, it runs fine with Apple firmware
00:57:38 Quit gregzx ("ChatZilla 0.9.81 [Firefox]")
00:58:37Chaoticlinuxstb: I can recap the problem if needed
00:59:20 Quit DerDome ("Leaving.")
01:00:34Chaoticit's actually charging now, and seems to be ok. I'm worried that it won't charge when I listen to music, which makes all the useful features of Rockbox out of my grasp. I almost exclusively listen to my ipod in the car whilst driving, plugged into a Monster radio transmitter/charger
01:01:09 Quit ol_schoola_ ()
01:01:31***Saving seen data "./dancer.seen"
01:01:40 Quit Thundercloud (Read error: 110 (Connection timed out))
01:02:28 Join piga [0] (
01:03:06linuxstbChaotic: I don't think it's anything you're doing wrong - Rockbox is known to not charge correctly from USB on the ipods.
01:03:19 Join ol_schoola [0] (
01:03:59Chaoticlinuxstx: I'm actually charging with the apple wall outlet adapter. Would that have the same problem?
01:04:26 Quit ol_schoola (Client Quit)
01:04:46Chaoticlinuxstb: I'm actually charging with the apple wall outlet adapter. Would that have the same problem?
01:05:24linuxstbChaotic: It depends if it connects to the ipod's USB or Firewire power pins.
01:06:46Chaoticlinuxstb: ok. It is a USB cable that plugs into the square power plug. So that may be the issue. I'll have to check the power in the car adapter before and after a commute
01:07:21Chaoticlinuxstb: Thanks for your help on that, it's certainly a drawback, but probably a livable one
01:08:02 Join tvelocity [0] (
01:08:11Chaoticlinuxstb: I'm going to ask one more question to the main channel, I don't want to monopolize your time. Thanks again.
01:10:22ChaoticSo one last newbie question: When I plug my 4G iPod Photo in, it shows a charging icon. If I try to shut it down while plugged in, it will turn off, then immediately show the non-backlit apple logo, then the backlit apple logo, and boot back into Rockbox
01:11:18Chaoticit seems that there's no way for me to shut it down and still have it charge. If I unplug it, I can shut it down, but then it boots again the moment I plug it in. Is this expected behavior? I couldn't find the answer in the manual nor forum
01:12:08Shaidthat's expected behaviour
01:12:31 Join RexDart [0] (n=cd9cbcfe@gateway/web/cgi-irc/
01:12:36Chaoticgreat. That explains a lot. Thanks Shaid & linuxstb!
01:12:50Shaidif you boot into the original firmware you can do what you're after
01:13:00Shaidbut then to get back to rockbox you'll need to do a reboot
01:13:01ChaoticI can't seem to do that though
01:13:44ChaoticI slide the hold button when I see the apple logo, but it just says "cleared" on the rockbox main page but doesn't reboot to the apple firmware
01:13:45 Quit stripwax (Read error: 110 (Connection timed out))
01:13:59Shaidyou're doing it too late
01:14:16Shaidand that's the old method to boot into the OF
01:14:24Shaidnowdays you hold down menu straight after powering on
01:14:42Shaidor possibly I'm confused
01:14:47ChaoticAh! Ok, was confused from the forum then. I'll give that a shot later
01:15:02 Quit stripwax_ (Read error: 110 (Connection timed out))
01:15:30Chaoticyeah, I must be doing it too late
01:16:52 Join fasmaie [0] (
01:16:57Chaoticyep, that worked
01:17:20ChaoticThanks again! Most heplful IRC channel I've been on in years!
01:17:27 Part toffe82
01:17:47 Join barrywardell [0] (n=barrywar@rockbox/developer/barrywardell)
01:19:26ChaoticI'm going to take off. Thanks again Shaid and linuxstb
01:19:39ShaidYou're welcome, Chaotic
01:19:53 Quit Chaotic ("CGI:IRC")
01:21:23Nico_Plinuxstb: I don't feel very strongly about the parser rejecting or not rejecting the WPS if there are too many pictures. It makes just as much sense to me to reject it... feel free to implement it that way ;)
01:21:33 Join midgey [0] (
01:22:01 Join moos [0] (
01:22:26 Quit ompaul (Client Quit)
01:23:40linuxstbNico_P: I agree (for Rockbox itself), but I think needs to reject such themes. So it's easier to just make Rockbox do the same thing.
01:24:55linuxstbWe'll need to take care of case sensitivity somehow though - assuming is going to a Unix server...
01:25:03linuxstb(in filenames) there some sort of problem with rockbox when switching between codecs for different filetypes in a playlist?
01:25:28linuxstbBut I guess theme authors should respect that anyway - for people using sims on Unix
01:25:33RexDartMy 5thG keeps crashing
01:25:45RexDartSongs stutter and it locks up totally
01:26:12linuxstbRexDart: How old is your Rockbox install?
01:26:17 Quit barrywardell ()
01:26:20RexDartlet me ask it
01:26:31 Join barrywardell [0] (n=barrywar@rockbox/developer/barrywardell)
01:26:43RexDartquoth the rockbox: "16667-080314"
01:27:04scorche|shRexDart: what other channel were you talkng about earlier?
01:27:17linuxstbThat's quite old in Rockbox terms - can you update to the latest current build and test if it still happens?
01:27:46RexDartWill do.
01:28:24linuxstbRexDart: Something else to check is if your filesystem is free of errors (e.g. with chkdsk or fsck)
01:28:30 Quit LambdaCalculus37 ("CGI:IRC 0.5.9 (2006/06/06)")
01:28:42RexDartalso good to know
01:30:22 Quit moos ("Rockbox rules the DAP world")
01:31:19RexDartI wonder if rockbox will ever progress to the point where it has user-selectable navigation styles, like you could choose for it to behave like a rio carbon or something
01:31:26Nico_Plinuxstb: checkwps loads files case-insensitively on windows?
01:32:11 Quit amiconn (Nick collision from services.)
01:32:17 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
01:32:17linuxstbNico_P: No, Windows does...
01:32:41LloreanRexDart: That sounds like customizable buttons, something we don't want.
01:32:51Nico_Pwhat ways do we have of making it case sensitive there?
01:33:10RexDartCustomizable buttons?
01:33:18RexDartoh, input method?
01:33:27LloreanRexDart: Well, what do you mean by "navigation styles" other than "navigating the UI"?
01:33:47RexDartthe entire layout of the UI & how it's navigated
01:34:07LloreanSo, "which buttons are used to get around it"?
01:34:14linuxstbNico_P: I guess we could make it scan a directory and do its own filename comparison...
01:34:29LloreanThe "layout" as it were is just visual, not actually an aspect of navigation.
01:34:51LloreanBut customizing what you press to get from place to place isn't going to happen in the official biuld.
01:35:06RexDarttoo bd
01:35:10 Quit RexDart ("CGI:IRC")
01:35:32Nico_Phmm yeah, that would work I guess
01:35:55Nico_Pthe dir shouldn't be too big either
01:37:20 Quit XavierGr ()
01:37:44 Join Cillian [0] (
01:38:36CillianIs it possible to alter the backlight brightness on a 5g ipod video in current rockbox software, or would it be reasonably trivial for me to add code to do that?
01:39:38Cillian(I figure it's possible from software because of the backlight fade in/out options)
01:40:31linuxstbYes, there's a backlight brightness setting
01:41:30CillianWhere abouts? (I've never seen one, so either it's not in my build or I'm looking in teh wrong place)
01:41:43linuxstbUnder LCD settings I would guess
01:42:10linuxstbAssuming you're running a recent Rockbox build - I think it was added about 6 months ago...
01:42:42CillianMine might well be older than that
01:42:50CillianAnd I certainly don't have that in LCD settings
01:43:00amiconn4 months ago
01:43:13CillianOk, thanks for the help
01:43:16amiconn12 Nov 2007 to be precise
01:43:40*linuxstb was only about 6 weeks out...
01:44:08 Quit perrikwp (" ajax IRC Client")
01:44:15Shaidyou might need to reset your settings, too
01:44:24 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
01:44:30ShaidI had to before the backlight brightness setting took effect.
01:44:33linuxstbamiconn: As you're here, you can save me reading the code - does the bmp loader return an error if there is no room to load the bitmap, or does it truncate and return success?
01:45:13*linuxstb goes to read the code...
01:45:30 Quit Rincewind ("bye")
01:45:38 Join `Tanner [0] (
01:45:48amiconnIt returns various errors depending on what goes wrong
01:46:01`TannerCan i play .FLV video files on my ipod while running rock box?
01:46:09*Cillian doubts it
01:46:17`TannerAlright, thought so.
01:46:20linuxstbHmm... I can't make checkwps fail when loading too many bitmaps...
01:46:27CillianIs there any video playback other than raw mpg on ipods?
01:46:28amiconnWorks similar to read() and such - on success, it returns the amount of memory taken. rc <0 means error
01:46:58barrywardellhmmm. the current svn e200 bootloader no longer boots the of for some reason
01:47:51amiconnlinuxstb: What does 'too many' have to do with the loader?
01:47:53barrywardellanyone have any suggestions as to which commit might have caused it?
01:48:03 Join _caotic_ [0] (n=caotic@
01:48:06Nico_Plinuxstb: I'm pretty sure the loading won't even be attempted ig there isn't enough space to store the bitmap
01:48:36 Part _caotic_
01:48:42 Quit JordanG (Connection timed out)
01:49:15`TannerCan i play .WMV videos on my ipod while running RockBox ?
01:49:25Shaidjust mpgs
01:49:30 Join stripwax [0] (
01:49:37Llorean`Tanner: I would strongly recommend you check out the manual.
01:49:51linuxstbamiconn: I mean too much data for the image buffer...
01:49:54`Tannerlol, alright.
01:51:46amiconnlinuxstb: Did you take into account that the image is converted to the internal pixel format on load?
01:52:53linuxstbIt's my mistake - I thought I was loading 26 full-screen images, but the filename was misleading me, and they're not full-screen...
01:53:40amiconnlinuxstb: I just discovered a WPS redraw bug
01:54:12amiconnIt looks like that on track change, the whole display is cleared, so that the status bar disappears for a fraction of a second
01:54:43linuxstbDo you think that's new? I've a feeling I've seen that in the past (but may be mistaken)
01:54:45amiconnHappens on main & remote, in parallel. Observed on X5
01:54:55amiconnNever observed it before
01:55:48linuxstbNo, you're right, I think I've added an lcd_clear_viewport() for each viewport, and the default viewport includes the statusbar area (with a ymargin) - that needs fixing.
01:56:09`TannerIf an ipod nano is set up for a mac, is there any way to change that to pc ?
01:56:14amiconnYou didn't test a wps with the statusbar enabled, did you?
01:56:40Llorean`Tanner: Restore with a PC version of iTunes, or follow the conversion instructions in our wiki
01:57:00amiconnAlso, the wps code used to just redraw the parts which needed redraw, saving resources this way...
01:57:01`TannerThanks Llorean. You've been a great help.
01:57:02linuxstbI did, but not patiently enough to wait for track changes. Early versions of my patch had bugs when the statusbar was enabled/disabled.
01:58:43linuxstbamiconn: I don't really understand the drawing code in the WPS, but will need to if I continue with my statusbar patch. So I'll definitely try and make it more efficient if I continue work on it.
01:59:34linuxstbAny idea when the partial redraw was removed? i.e. was it at the time of any particular change/new feature?
02:01:12 Quit `Tanner ("ChatZilla 0.9.81 [Firefox]")
02:05:07Cillianlatest daily rockbox, and my ipod sits on the loader2 loading screen with the bar full, spinning the harddrive
02:05:13Cilliandoing not much else
02:05:26 Join MethoS- [0] (
02:05:26 Quit jas6180_ (Read error: 104 (Connection reset by peer))
02:06:27 Join jas6180_ [0] (
02:06:29LloreanCillian: Well, loader2 isn't our software. If it's stopping while in loader2, I suggest you try our loader and see if everything works.
02:07:22CillianI realise that, I was just wondering if it was something someone else had come across (Since it's the rockbox update that triggered it)
02:09:04linuxstbCillian: If you're running a very old loader2 (and were also running an older version of Rockbox), you should update loader2 as well. Or as Llorean suggested, just use the Rockbox bootloader.
02:09:42 Quit Nico_P (Remote closed the connection)
02:09:57CillianIt is indeed an ancient loader2
02:10:46Cillian* has an idea *
02:10:49linuxstbThen that's very likely the problem.
02:10:56Cillianrockbox can reload it's kernely .ipod file, right?
02:11:00 Join stripwax_ [0] (
02:11:27CillianWhich has nothing to do with teh bootloader, when it does that?
02:11:28 Quit jas6180_ (Remote closed the connection)
02:11:57 Quit stripwax (Read error: 104 (Connection reset by peer))
02:12:09linuxstbThe problem is that the old loader2 doesn't initialise the second CPU core correctly when loading Rockbox - and there's nothing Rockbox can do about that later.
02:12:31 Join jas6180_ [0] (
02:12:47CillianOk, cheers
02:13:17linuxstbOlder versions of Rockbox didn't use the second core, so it didn't matter. New versions do (a little)
02:13:33 Nick fxb is now known as fxb__ (
02:14:18 Quit tvelocity (Remote closed the connection)
02:14:42 Quit jas6180_ (Remote closed the connection)
02:15:21 Join jas6180_ [0] (
02:19:15 Quit MethoS-- (Read error: 110 (Connection timed out))
02:27:54 Quit kugel ("ChatZilla 0.9.81 [Firefox]")
02:29:50barrywardelllooks like the bootloader problem came about with r16259 - disabling the hardware in PP502x
02:30:25linuxstbWhat bootloader problem?
02:31:51CillianGod, downloading 400KB files is painful
02:34:38barrywardelllinuxstb: the sansa bootloaders don't load the of any more
02:35:03barrywardellI've tracked it down to that commit
02:35:34*linuxstb recalls asking if anyone had tested the bootloaders when that commit was made...
02:37:40barrywardellthe hardware should probably not be disabled in the bootloader. much better to do it when rockbox starts up
02:38:01barrywardellor is that what it does already? I'm only just catching up with the changes now
02:38:14 Quit stripwax_ (Read error: 110 (Connection timed out))
02:38:52CillianGuh, now loader2 fails to load even the old rockbox
02:39:51JdGordonbarrywardell: that would explain my sansa problems... i installed rockbox svn on it the other day and cant get into the OF at all...
02:39:59linuxstbCillian: I can't remember if that's expected, but why would you want to load an old Rockbox?
02:40:40CillianSince the new one doesn't work either
02:40:58barrywardellJdGordon: reverting to r16258 makes everything work fine again. there's a lot of changes to look at it r16259 though
02:41:23CillianLooks like the new loader2 doesn't use 8.3 filesnames anymore - old one did
02:41:45 Join kugel [0] (n=kugel@unaffiliated/kugel)
02:45:47 Join MethoS-- [0] (
02:45:56 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
02:46:07Beta2KCJRL 89.5 Mix FM
02:46:07Beta2K128 Main Street South
02:46:07DBUGSent KICK Beta2K to server
02:46:07Beta2KKenora, ON P9N 1S9
02:46:08Kick(#rockbox Beta2K :No flooding!) by logbot!n=bjst@gateway/web/cgi-irc/
02:46:08 Join Beta2K [0] (
02:46:28Beta2KI really need to make paste a different key
02:46:32Beta2KSorry all
02:46:45CillianOh, wow, new rockbox is *sexy*
02:46:50CillianAwesome work :)
02:47:12 Join saratoga22 [0] (n=9803c6dd@gateway/web/cgi-irc/
02:47:20 Quit saratoga22 (Client Quit)
02:47:24 Join saratoga22 [0] (n=9803c6dd@gateway/web/cgi-irc/
02:48:43 Join saratoga2 [0] (n=9803c6dd@gateway/web/cgi-irc/
02:49:04saratoga2barrywardell: I've not been able to update my sansa bootloader to try your new patch
02:49:19saratoga2sansapatcher always failed
02:49:25 Join confetti [0] (
02:49:33barrywardellwhich patch? and what error did sansapatcher give?
02:49:49confettiJust a quick question. Is rockbox anywhere near runnable on zune?
02:50:14confettiwill it be?
02:50:22scorcheit wont be with no one working on it
02:50:30confettinoone is?
02:50:43confettisad. :( I don't know how or I probably would be ;)
02:50:56 Quit saratoga ("CGI:IRC (Ping timeout)") one currently is actively working on the zunbe
02:51:06linuxstbNot that sad - there are 28 other devices Rockbox runs on, and more on the way...
02:51:58confettino, not sad as a whole. rockbox isn't sad. it is sad that zune doesn't support it :(
02:52:24 Quit saratoga22 ("CGI:IRC (Ping timeout)")
02:52:29 Quit confetti (Client Quit)
02:52:58saratoga2barrywardell: unfortunately i don't have the sansa in front of me so i can't double check, but when I last tried it I got some unspecific error about short write faililng if I remember correctly
02:53:09saratoga2i looked through the source but couldn't see anything obviously wrong
02:53:27saratoga2and i was able to flash it once before to the rockbox bootloader
02:54:17saratoga2i wouldn't have asked without it in front of me, but i haven't seen you online in a while
02:56:09barrywardellsaratoga2: which os were you using?
02:56:36barrywardellsaratoga2: and was that complied from svn? or the one on the download server?
02:56:38saratoga2i compiled the bootloader under Linux, but ran sansapatcher (downloaded from the website) under windows
02:56:41kugelLlorean: Why don't you want a reset list dimensions setting? There's a reset colours option.
02:58:44barrywardellsaratoga2: that was with the -bl flag?
02:58:54saratoga2barrywardell: yes
02:59:19barrywardellhmmm. I've never had that problem, but I never use windows so maybe that's a problem?
02:59:31saratoga2i will find a linux box with USB tomorrow and try it there
02:59:41saratoga2i just assumed it didn't matter
02:59:41 Quit MethoS- (Read error: 113 (No route to host))
02:59:52saratoga2great work on that patch though
03:00:03saratoga2i never understood why it made the OF take 2x as long to boot the way it is now
03:00:04barrywardellgreat. make sure to use a bl compiled before 16258 though
03:00:46saratoga2are you looking into fixing that issue already?
03:01:02barrywardellI think I've found the problem
03:01:32***Saving seen data "./dancer.seen"
03:01:38barrywardellit's the usb init/detect in main-pp.c. removing it lets the of boot again!
03:01:40saratoga2the USB thing?
03:01:44saratoga2ah yeah!
03:01:48saratoga2the OF turns it on by default
03:01:58saratoga2so thats probably bad if it boots without it
03:02:19 Join csc` [0] (n=csc@archlinux/user/csc)
03:02:45barrywardellyeah, I'll see if I can isolate it further
03:03:04 Join MethoS- [0] (
03:06:32saratoga2barrywardell: also, the power measurements taken for that patch were in part done on my sansa using the rockbox bootloader
03:06:47saratoga2i didn't even think of it at the time, but i wonder if it makes a difference if I use the Sandisk one
03:07:51 Quit MethoS- (Remote closed the connection)
03:09:47 Join rdeleon580 [0] (
03:09:51 Part rdeleon580
03:10:36barrywardellsaratoga2: it would be an interesting test to do if you have time some time
03:15:40 Quit fasmaie (Read error: 113 (No route to host))
03:17:30barrywardellI'm too tired to look at this further tonight. the problem gets fixed by removing the usb detection section in main-pp.c. I'll try to dig further tomorrow if nobody beats me to it.
03:19:25 Join Synx_hm [0] (
03:19:53 Join jcollie [0] (
03:19:58 Quit MethoS-- (Read error: 110 (Connection timed out))
03:20:08Synx_hmAnyone having continued issues with rockbox on a 30gb IPOD video... over the course of an hour or two ill have to reboot my ipod 2 or 3 times from hardlocks, and sound issues
03:20:17Synx_hmthat and my database keeps currupting
03:20:38 Quit barrywardell ()
03:20:48 Join Thundercloud [0] (
03:21:16saratoga2Synx_hm: never heard of that, but i don't have an ipod
03:21:18 Quit saratoga2 ("CGI:IRC")
03:22:04Synx_hmya i am not really sure whats going on, but it has been an ongoing problem :( And i have been updating to the latest build to see if that fix's it :(
03:24:07Synx_hmive also had the misfortune of blowing out the firewire charging circuit on my ipod, so i can use all those nifty ipod battery packs, so i think it may be time to go back to stock firmware so i can get some more use out of this little guy
03:25:27 Join MobileShaid [0] (
03:33:21 Quit Isolinear (Read error: 113 (No route to host))
03:38:13 Join Shaid3 [0] (
03:40:16 Quit Shaid (Read error: 113 (No route to host))
03:40:17 Nick Shaid3 is now known as Shaid (
03:45:41 Quit |AhIoRoS| (Remote closed the connection)
03:47:22 Part pixelma
03:47:38 Quit Zarggg ()
03:52:06 Join jas6180 [0] (
03:53:18 Join Zarggg [0] (
03:53:24 Join Shaid3 [0] (
03:55:27 Quit MobileShaid (Read error: 113 (No route to host))
04:03:35 Quit Cillian (Read error: 113 (No route to host))
04:07:52 Join miepchen^schlaf_ [0] (
04:11:18 Quit Shaid (Read error: 113 (No route to host))
04:11:19 Nick Shaid3 is now known as Shaid (
04:12:46 Quit jas6180_ (Read error: 110 (Connection timed out))
04:13:42 Join Davide-NYC [0] (
04:14:18 Part Synx_hm
04:14:22 Quit midgey ()
04:14:30Davide-NYCJdGordon: ping
04:15:06Davide-NYCHey there. Any news on the WRS front?
04:15:23JdGordonno, that got pushed to the side im afraid
04:15:32JdGordonactually, totally forgot about it untill jkust now :p
04:15:49*Davide-NYC falls off of his chair
04:16:12Davide-NYCI've been making a concerted effort to leave you alone about it.
04:16:19 Join axionix [0] (
04:17:12JdGordonarg.. this is odd... I'm sure there was a theme problem with this patch last night.. seems to work fine now
04:17:25*JdGordon fucking retarted moron
04:17:32JdGordonforgot to repatch my tree :p
04:18:12Davide-NYCIf you get back to working on the recording screens and could use a minion to test and/or do other small tasks let me know.
04:18:36Davide-NYCI'm willing to patch 'n' test on the targets available to me.
04:19:23JdGordoni tihnk the rec code need more splitting up like was mentioned in the thread before a proper effort can be done
04:20:23Davide-NYCoh yeah, I forgot about that and that's probably why you forgot about the rewrite.
04:20:29Davide-NYCjhMikeS: ping
04:20:35*Davide-NYC ducks
04:22:23Davide-NYCgeneral question: to run a text file like a batch file in cygwin what do I do:?
04:22:54jas6180just do ./
04:22:55 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
04:23:15jas6180and if it tells you permission denied, chmod u+x first
04:26:19Davide-NYCBad line endings! Saved as unix filetype and viola!
04:28:33 Join Jayden|AWAY [0] (
04:28:43Jayden|AWAYAnyone here using foobar2000 with a rockboxed iPod?
04:32:13 Join Fuxxer [0] (n=45e98cc2@gateway/web/cgi-irc/
04:32:50 Join the_bdquick [0] (
04:32:55krazykitJayden|AWAY, people might not always be around to answer right away. however, that question is offtopic, most likely, as no rockbox code is running when you'd be syncing with foobar
04:33:05 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
04:34:34Jayden|AWAYOk sorry, then how can I copy over FLAC files?
04:34:45 Quit Fuxxer (Client Quit)
04:34:50 Join Fuxxer [0] (n=45e98cc2@gateway/web/cgi-irc/
04:34:59jas6180the same way you copy over other kinds of files, most likely
04:35:02 Quit Fuxxer (Client Quit)
04:35:24Jayden|AWAYWithout it being in disk mode...
04:36:15scorchewhy would it not be in disk mode?
04:36:32scorcheyour computer will not even see it if you are not in disk mode...
04:37:19Jayden|AWAYNvm. I phrased the question poorly. I found my answer on a forum, discussing how to take it out of MTP
04:37:22 Part Jayden|AWAY
04:38:12scorcheipods dont have MTP
04:38:17 Quit argumentD (Read error: 113 (No route to host))
04:39:50 Quit Thundercloud (Read error: 110 (Connection timed out))
04:44:16JdGordonis the delta table broken?
04:44:34JdGordongroan, very red :p
04:47:55 Quit jas6180 (Remote closed the connection)
04:48:09 Join jas6180 [0] (
04:51:25 Join elinenbe_ [0] (
04:52:08 Quit elinenbe (Read error: 104 (Connection reset by peer))
04:52:08 Nick elinenbe_ is now known as elinenbe (
04:52:20 Quit XavierGr (Read error: 110 (Connection timed out))
04:53:57 Quit Davide-NYC (
04:53:57 Quit simonrvn (
04:53:57 Quit Llorean (
04:53:57 Quit HellDragon (
04:53:57 Quit crashmatrix (
04:53:57 Quit Rick (
04:53:57 Quit ChanServ (
04:55:14 Quit jas6180 (Remote closed the connection)
04:55:37NJoinChanServ [0] (ChanServ@services.)
04:55:37 Join Garbanian [0] (
04:55:37NJoinsimonrvn [0] (i=simon@unaffiliated/simonrvn)
04:55:37 Join Llorean [0] (n=DarkkOne@rockbox/administrator/Llorean)
04:55:37NJoinHellDragon [0] (i=jd@unaffiliated/helldragon)
04:55:37NJoincrashmatrix [0] (
04:55:37 Join Rick [0] (i=rick@unaffiliated/rick)
04:55:37Mode"#rockbox +o ChanServ " by
04:55:39***Server message 505: 'logbot :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
04:55:39***Server message 505: 'logbot :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
04:55:43 Quit ze (Read error: 104 (Connection reset by peer))
04:55:46 Join ze [0] (
04:55:46Garbaniananyone alive?
04:56:24 Quit joshin (Read error: 104 (Connection reset by peer))
04:56:27Mode"#rockbox -o logbot " by ChanServ (ChanServ@services.)
04:57:28Garbanianwhats up?
04:57:46cool_walking_The social channel is #rockbox-community
04:58:09GarbanianAnd this one is?
04:58:16cool_walking_Official, serious business.
04:58:28GarbanianAs in Dxxx
04:58:52cool_walking_As in, programming talk, and tech support of Rockbox.
04:58:59Garbaniancause all im doing is looking for some help. Rofl.
04:59:07cool_walking_Yeah then this is the place.
04:59:12scorchethen ask a question,,
04:59:21Garbanianstory time.
04:59:33GarbanianI installed it as the site said, put the text ect in.
04:59:39GarbanianAnd it looked right.
04:59:46BHSPitMonkeyit = rockbox?
04:59:48cool_walking_"the text"?
05:00:05GarbanianThthe text file that came with it, basically the same thing as the main file.
05:00:18Garbanianyea BHSPitMonkey
05:00:28Garbaniani rebooted and it looked retarded.
05:00:34GarbanianLike...blank >_<
05:00:39 Join joshin [0] (n=joshin@VDSL-130-13-252-76.PHNX.QWEST.NET)
05:00:42GarbanianSo i uninstalled as the guide said
05:01:00Garbanianit reverted to the original iTunes, and attempted a redownload/install
05:01:09Garbanianand im still getting the weird theme thing
05:01:19 Quit piga ("Leaving")
05:01:25scorchea few things:
05:01:35***Saving seen data "./dancer.seen"
05:02:02scorchefirst off...please write you complete is much harder to read when you have 7 different lines in a row
05:02:26GarbanianSorry >_>
05:02:36scorchesecondly...please try and explain things in greater detail and referencing what actually happens instead of "the weird thing
05:02:44scorchethirdly, which guide are you referring to?
05:04:33GarbanianOkay....Well i downloaded the 30 gig rockbox, from the site. Unzipped it to the desktop, and went to the guide at which i follwed the manual install. I copied and pasted the .rockbox to the drive in which my iPod is located. Then copied the voices to the iPod and then executed the bootloader. I rebooted the iPod, and rockbox-
05:05:42scorcheit got cut off at "the iPod, and rockbox-", but thanks a bunch for writing it out :)
05:06:17Garbanianloaded fine. The skin for the ipod looked amazing. Then i got the songs in there, it told me to reboot again to do something (not sure), so i did, and when this happened the skin went to like...(ill take a pick just read your comment, i put the - to continue it) a skin that...basically shouldnt be there.
05:07:01scorchewhat did this "skin" look like? in what made it stand out?
05:07:13Garbanianim taking a pic now
05:07:26Garbaniancause even reinstalling wont get it back to the rockbox original
05:07:55A-4and changing the theme in the settings doesnt work either?
05:08:20GarbanianLemme take a pic of all the themes then
05:09:14scorcheGarbanian: first, could you shut the ipod off, turn it on, and then flip the hold switch on while it boots?
05:09:58cool_walking_scorche: maybe expand a little on that, since that could get you into the Apple firmware, or Rockbox with cleared settings.
05:10:53cool_walking_Depending on when you flick the switch.
05:10:58Garbaniansaving them now
05:11:43 Part the_bdquick ("Leaving")
05:12:06GarbanianYay slow interwebz
05:13:22Garbanianoh snapp these are the big files from the cam
05:14:18krazykitGarbanian, please, this is an on-topic channel. take non-rockbox chatter elsewhere.
05:15:34GarbanianI am here for help
05:15:44 Join AJG983 [0] (n=209594fd@gateway/web/cgi-irc/
05:15:51Garbanianthey've been helping me, i just said that as an update for why it was taking so long.
05:15:59cool_walking_Yes, but unnecessary chatter clogs the logs (this channel is logged).
05:16:58GarbanianWell either way, i found an alternate way, ssed the pics, all them basically look the same, but 1 is yellow, not white.
05:17:32AJG983If I am running rockbox on a sansa e260 and i go to plug it in via usb to my pc while playing a song to charge the battery, is it supposed to turn off and just have the wheel light up blue while the screen remains black? I thought it was possible to plug the player in while running rockbox and not have it automatically power off.
05:19:00scorcheGarbanian: does the theme change when you click on the items under browse themes?
05:19:07A-4Garbanian: select cabbiev2 as theme, it's the default
05:19:23Garbanianeach one of those is a diff theme, only 1 changes color, to yellow, but that didnt save
05:20:59scorcheGarbanian: could you try as i said before? with the hold button?
05:21:33Garbanianturn it on like hold the play pause button, or hold menu + select?
05:21:43cool_walking_scorche: you didn't say _when_ to flick the hold switch on. flicking it on at different times will get you either into the Apple firmware, or cleared settings in Rockbox.
05:22:10scorchecool_walking_: i am well aware, but it is simple to say "try it again a bit later"
05:22:49scorcheGarbanian: turn it off by holding play/pause; turn it on; flick the hold switch on
05:23:20 Quit AJG983 ("CGI:IRC (EOF)")
05:24:01GarbanianI got to the normal iTunes look
05:25:20scorcheok...then reset with holding menu and select, and then turn it on again a bit later...right when you see the rockbox name if you can
05:25:55Garbanianshould it say cleared?
05:26:26Garbanianthen i think i win, should i just boot it up as normal?
05:26:52scorchedo you get the "regular look" again?
05:26:56Garbanianooo hey!!! its sexy!
05:28:02GarbanianThank you scorche!!!!<3
05:28:38scorchethanks for following instructions :)
05:29:09GarbanianLol, i take it that doesnt happen ofter?
05:29:23GarbanianWell....gtg actuall...getting yelled at
05:29:26 Quit Garbanian ()
05:30:13 Join Prikoki [0] (
05:30:47 Join midgey [0] (
05:33:47 Quit Zom (Remote closed the connection)
05:33:58 Join Zom [0] (
05:35:23 Part Prikoki
05:40:34 Quit Horscht ("I got raided by the FBI and all i got is this lousy quit message")
05:46:39 Join ol_schoola [0] (
05:47:19 Quit kugel ("ChatZilla 0.9.81 [Firefox]")
05:52:09 Join alienbiker99 [0] (
05:54:27 Join Jon-Kha [0] (
05:56:10 Join JordanG [0] (n=JordanG@
05:56:16 Quit JordanG (Remote closed the connection)
05:56:56 Join JordanG [0] (n=JordanG@
06:29:07 Quit amiconn (Nick collision from services.)
06:29:13 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
06:36:52 Join Par[4] [0] (n=Par_4_@
06:37:46Par[4]what version of rockbos shoud i use if i have ipod classic 160?
06:37:56scorchenone of them
06:38:33Par[4]thers no vesrion for the 160? it says on the front page
06:39:38 Quit Par[4] (Client Quit)
06:42:04 Join hwguru [0] (n=442830a6@gateway/web/cgi-irc/
06:44:08 Quit hwguru (Client Quit)
06:45:45 Quit axionix (Read error: 110 (Connection timed out))
07:01:37***Saving seen data "./dancer.seen"
07:02:13 Quit JordanG (" I love my HydraIRC -> <-")
07:06:09 Nick miepchen^schlaf_ is now known as miepchen^schlaf (
07:14:18 Quit midgey ()
07:18:21 Join mrjenkins [0] (n=765a72ef@gateway/web/cgi-irc/
07:22:06mrjenkinsI can now confirm that RockboxUtility works successfully on Windows Vista x64 and Windows Server 2008 x64 if thats useful
07:26:37 Quit perrikwp (" ajax IRC Client")
07:29:57 Quit mrjenkins ("CGI:IRC")
07:32:31 Quit miepchen^schlaf ()
07:33:28 Quit EspeonEefi ("さよなら")
07:45:44 Join axionix [0] (
07:50:23JdGordonso much of the list vp patch do we eventually want?
07:51:38JdGordonI mean, we can now reallly easily setup the menus and browsers to sit in an arbitary spot... bars arnt ready yet though
07:51:53LloreanLike, the scroll bar, or do you mean like the button bar?
07:52:33JdGordonstatus/button bars
07:53:17Lloreanlinuxstb is working on a solution for the status bar, I think.
07:53:18 Join Phreadde [0] (n=ca242d01@gateway/web/cgi-irc/
07:53:47LloreanBut honestly, I don't mind if they're locked in place for now, while the list itself is movable.
07:54:09LloreanUsers who want to use up a lot of the screen with backdrops will turn off the status bar (or incorporate it), and will be on a non-button-bar target anyway
07:55:38*JdGordon would rather a statusbar "wps" seperate from the wps "wps" than a single .wps for both
07:56:01JdGordonfix up the parser to handle that, then making the fm wpsable should be simple
07:56:29 Quit Shaid (Read error: 110 (Connection timed out))
07:57:30PhreaddeI've recently registered on the Twiki and need to be write-enabled so I can upload a modified WgtnNZ.fmr file
07:58:26Phreaddemy id is PeterDawson, btw
08:02:01LloreanJdGordon: Yeah, but ignoring that, is there a reason against linuxstb's method? I really like the fact that it means you can incorporate playback elements into the List (though it's still lacking for conditional viewports)
08:02:47JdGordonno, that screeny I saw looked great
08:03:25JdGordonits gonna be a fairly big ram hit though (doing it the way id rather)
08:04:21 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
08:04:41LloreanI mean I'd want a "status bar" up top, and a "playback status" at the bottom, which disappears and returns the space to the list if playback is stopped. But that's just me and how I'd design my list viewports, and illustrates a couple extra features it'd need to have to meet my "ideal"
08:04:47JdGordonjhMikeS: have you got any plans for radio.c? seems you were the last to fiddle in it (with a big TODO added...)
08:05:09JdGordonoh, you want both? that would be good too
08:05:28LloreanUp top would display "Time, Date, Battery, Volume" and maybe other things.
08:05:46LloreanAt the bottom would be "Now Playing" alternating between Artist and Title and "Playlist Position"
08:06:00LloreanThen the Now Playing stuff would be conditional, if playback was stopped no need to display "Nothing"
08:09:21*JdGordon wonders why swcodec doesnt have peakmeters on the fm screen
08:12:51PhreaddeJdGordon: Thanks - I've confirmed success with a new browser session.
08:17:36LloreanJdGordon: I would guess because nobody's bothered? In playback, it tells you if you're clipping (something you can fix), in Recording, it tells you if you're clipping (something you can fix). In FM, it tells you the station operators are clueless (not much you can do about it).
08:18:13GodEaterhas anyone complained about further breakage to the playback on ipods (g5.5 specifically) sometime between late last week and now ?
08:18:32pondlifeGodEater: What kind of breakage?
08:18:54JdGordonLlorean: hehe, well i was more asking because the screen is boooring
08:19:04LloreanJdGordon: VERY boring, yes.
08:19:08GodEaterlike it was for a while after the spinlock change went in originally
08:19:18GodEateri.e. lots of cutting out when you first turn on
08:19:27LloreanGodEater: Does it go away, though, and never come back?
08:19:33GodEateronly now, it does that on every folder change, not just while dircache
08:19:38pondlifeHmm, I've not seen that on my targets recently. I did before.
08:19:41GodEater+is building
08:19:56GodEaterLlorean: not really no
08:19:57 Quit Phreadde ("CGI:IRC (EOF)")
08:19:59pondlifeSo, not directly disk access related?
08:20:15pondlifei.e. once dircache has built, folder changes should need disk access...
08:20:25LloreanGodEater: jhMikeS made a recentish change on how priority is assigned for playback, which made my UI more responsive on my Sansa (and suggests then that playback gets less juice)
08:20:43GodEaterLlorean: I wondered if that might be the change that did it
08:20:54GodEaterunfortunately I no longer have svn access from work
08:21:02GodEaterso I can't test a build immediately prior
08:21:11pondlifeThat *is* unfortunate.
08:21:26pondlifeI can make a build for you, if you want.
08:21:31LloreanIt does sound kinda odd though, if it's happening even when dircache is built, etc.
08:21:56GodEaterpondlife: I'd be happy to test if you did
08:22:25pondlifeGive me a while to test/revert some stuff, should have a build by 9am
08:22:33GodEaterno problem
08:27:31amiconnJdGordon: That's rather obvious. Swcodec doesn't record from the fm screen, hence can't have peakmeters in that screen. In order to show peakmeters, you need to record the data
08:27:52JdGordonarnt they tied to line out?
08:28:00JdGordonlike the wps....
08:28:20amiconnThere is no CPU visible line out data with fm
08:28:45amiconnThe fm radio signal is just mixed into the analog signal path
08:28:59 Join Rob2222 [0] (
08:29:03*GodEater discovers the "Introduce yourself" mentors thread in his inbox, and wonders how long it will take to read.
08:29:19amiconnThe archoses with radio (fm recorder and Ondio fm) put the MAS into recording mode as soon as the fm screen is enetred
08:31:54amiconnSo if you also have prerecording enabled, you can always record your favourite songs. The drawback is that the MAS sucks more battery power this way, so if you really just want to listen to the radio, battery runtime is shorter than it could be
08:32:42 Quit BigBambi (Remote closed the connection)
08:32:52amiconnI will change that eventually, so that recording is only enabled in the fm screen if prerecording is enabled (maybe with a separate "prerecord fm" setting?)
08:34:12amiconnWe can even save more battery power on archos this way. While the MAS is a single chip, its components can be switched on and off separately, and for fm radio, we only need the audiocodec. The dsp core can be switched off
08:38:15amiconnJdGordon: Btw, should your (red-delta-) commit fix the button bar problem?
08:42:12jhMikeSJdGordon: I don't even remember what the TODO: was
08:44:56 Quit cool_walking_ (Remote closed the connection)
08:46:56 Quit Rob2223 (Read error: 110 (Connection timed out))
08:49:16amiconnJdGordon: Now the button bar is even more broken :( The menu still reserves the space for the button bar even when it's disabled, and the browser has lost its button bar completely...
08:50:06jhMikeSGodEater: I did have someone test the mutex after a small but deep bug was corrected an the overall buffering speed was just as good as the rollback but it was mentioned that initial playback could break up. I think any real fix has to be at the buffering level.
08:50:49 Join ender` [0] (
08:51:22 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
08:52:22GodEaterjhMikeS: yep - I saw your note on the commit
08:52:24jhMikeSJdGordon: Ah yes, moving the basic tuner control out of the GUI
08:54:28 Join Mauritz [0] (
08:54:55*jhMikeS thinks the buffering chunk should go back to 32KB since the codec dominates too much with it yielding less in dsp.c
08:56:17MauritzAnybody else seeing a complete lockup the moment you start playing any music on an IPod (Both my 5.5G 80GB and my Nano have the same problem). This started happening when I updated from Subversion today (Yesterday the build was fine).
08:56:58MauritzI need to do a full reset to get out of it.
08:57:19amiconnDid you do a full rebuild? jhMikeS' commit of yesterday seems to require that
08:57:39*amiconn got an IllInstr yesterday when forgetting to do that for X5
08:57:56MauritzOh, ok. No I didn't I just did a make. Thanks for the tip. I'll get back to you after I've tried it.
08:58:00*jhMikeS updated the codec api version so why's it not just failing?
08:59:39 Quit GodEater (Remote closed the connection)
09:00:20LinusNjhMikeS: maybe it is failing, with a "Codec failure", but if the playlist is wrong, it could take a while to fail on each track
09:01:04LinusNjust a guess
09:01:41***Saving seen data "./dancer.seen"
09:02:46JdGordonjhMikeS: did you have plans to do it soonish?
09:03:13jhMikeSI think the particular rearrangement of variables might have actually made it impossible to get to the point where it can figure that out.
09:03:45 Join GodEater [0] (i=c2cbc962@rockbox/staff/GodEater)
09:04:54jhMikeSJdGordon: I hadn't thought about it. Is it critical to something? It's pretty much just a "paste it in another file" kind of thing.
09:05:32JdGordonI was looking for something to do and thought about fixing up the fm screen and saw it
09:12:32 Join petur [50] (n=petur@rockbox/developer/petur)
09:13:48JdGordonah crap... /me thinks a line got commited which shouldnt have
09:14:19JdGordonhmm... maybe not
09:16:42amiconnLinusN: I observed a weird effect regarding the iaudio LCD remote.
09:17:28amiconnWhen I pull my M3 remote from the M3 while music is playing, and the plug it back in, everything works like it should (both in rockbox and the OF)
09:18:19amiconnBut when I use the X5 remote (which should be electrically identical) on my M3, the M3 powers off immediately when I plug it back in
09:18:47amiconnWhen the X5 remote is plugged from the start, everything is okay. That also happens both in rockbox and the M3 OF.
09:18:56JdGordonamiconn: do _any_ lists not have the button bar?
09:19:00amiconnOn the X5, I can use either remote without such effects
09:20:02jhMikeSa slight difference in connector shape?
09:20:04amiconnJdGordon: Iirc the debug menu didn't have a button bar. It could as well have one, but then the button labels would need to be adjusted of course
09:20:39LinusNamiconn: interesting indeed
09:21:18amiconnThe button bar labels are something that needs rework anyway. The labels should reflect what the buttons actually do (that's what they are for), but somethimes they don't
09:21:56amiconnI think this is because there aren't many devs with a soft-button target nowadays :/
09:22:40amiconnjhMikeS: The remote connectors are exactly the same among M3, M5 and X5
09:23:02peturmaybe the soft-button concept will come back with touch-screen devices...
09:23:10*amiconn didn't test re-plug behaviour on M5 yet
09:23:21amiconnpetur: That might in fact be a good idea...
09:23:46amiconnDoesn't the mr500 have a touch screen?
09:23:54peturiirc, it does
09:24:34JdGordonit does.. but it wont while its in this broken state :D
09:25:07 Join n1s [0] (
09:25:30 Join stripwax [0] (
09:26:06amiconnJdGordon: Also, it would probably be good to not show a button bar when calling the menu code from plugins. We could do that, but then plugin authors would need to specify the correct button labels
09:26:17MauritzBtw, has anybody done any work on getting remotes for IPod:s to work? Anyone have any tip on where I should start looking if I'm interested in it?
09:27:22JdGordonamiconn: they can, i added that in the commit
09:27:30JdGordonalthough its hide/show both bars...
09:32:21 Join quitte [0] (
09:33:51n1sMauritz: LinusN has posted a patch in the tracker with initial (far from complete) support for the accessory protocol
09:35:03 Quit stripwax ("Miranda IM! Smaller, Faster, Easier.")
09:35:05Mauritzn1s: Awesome. I'll check it out and see if I can help out.
09:35:28Mauritzamiconn: Your tip about the rebuild worked (At least for the 5.5G one). Thanks!
09:37:15 Join OlivierBorowski [0] (
09:47:48 Join DerDome [0] (
09:53:22 Quit rasher (Read error: 60 (Operation timed out))
09:55:07JdGordonalmost fixed completly... (well mostly :p ) just gotta get the damn menu to resize when the bb is en/disabled
09:55:14quittemorning. is someone here working on the zvm port? the forum thread is giving me a headache. where do i find the zenutils?
09:55:33 Part Mauritz
09:56:41 Join B4gder [0] (
10:01:28*pondlife wants the Cowan D2 to be his touchscreen target.
10:02:48JdGordonna, its too small
10:07:17peturhmmm iriver created another player with line-in: E100
10:08:45 Join Mathiasdm [0] (
10:14:22 Quit DerDome (Read error: 110 (Connection timed out))
10:16:01 Quit advcomp2019 (Read error: 104 (Connection reset by peer))
10:16:37 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
10:16:38peturinteresting firmware site:
10:17:05GodEaterdoes anyone know why the dircache thread screen on my ipod is no longer updating in real time ?
10:21:26pondlifeThe tarballs seem to be broken more often than not... :/
10:21:47pondlifeCould the tarball building process just be automated straight off SVN?
10:22:01pondlifei.e. no reliance on FILES
10:22:05n1syeps, there doesn't seem to be a nice way to automate the FILES thingy...
10:22:38n1spondlife: I think the idea is to not be dependent on svn, i don't understand why we need that though...
10:23:14n1sand adding too wide wildcards defeats the purpose too (as with the current files issue)
10:23:15pondlifeSeems like we are going to be dependent on the master source control system anyway.
10:23:41 Quit Mathiasdm ("Yuuw!")
10:23:44 Quit sneakums ("There shouldn't be this reign of silence, but what are the options when someone great is gone?")
10:23:48B4gderyeps, we should probably just make it suck everying that isn't .svn into a tarball
10:24:34*n1s didn't see LinusN's commit :)
10:25:56peturBagder: a 'hex' file starting with 11 22 33 44 55 66 77 88 99 AA BB CC DD EE FF ring any bell?
10:26:17peturit is probably some kind of an archive
10:26:26B4gdernot really, but it looks like a magic marker
10:26:35peturyeah ;)
10:26:40 Join Mathiasdm [0] (
10:26:47JdGordonits missing 00!
10:27:17*petur has a quick look at the E100 firmware upgrade file
10:27:42B4gderit's probably a custom magic way to identify their file format
10:28:21peturI was hoping it really was some standard archive format - maybe they just changed the marker
10:29:00*B4gder goes to check
10:30:00 Join sneakums [0] (
10:31:08GodEater16791 has definitely rendered my ipod a pain in the ass to use again
10:32:09B4gderpetur: it certainly looks scrambled or compressed
10:32:20peturyes, no strings at all
10:33:43B4gderand the hex file didn't compress much in the zip, meaning it has almost no long sequences of similarities
10:33:58B4gderwhich a hexdump also shows
10:34:02 Quit Mathiasdm ("Yuuw!")
10:34:13 Join Mathiasdm [0] (
10:34:52B4gderfile size at offet 0x10, little endian
10:35:28 Join davina [0] (
10:39:00 Nick fxb__ is now known as fxb (
10:39:30 Quit advcomp2019 (Read error: 104 (Connection reset by peer))
10:39:43 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
10:40:18*petur swears at the gsoc mentor list introduction thread
10:42:03 Quit ChanServ (Shutting Down)
10:42:29markunpetur: what's wrong with it?
10:42:34*GodEater has muted it already ;)
10:42:43GodEatermarkun: it's enormous ?
10:42:44peturtoo many mails?
10:43:07GodEaterpetur: you getting it to a gmail address, or elsewhere ?
10:43:33 Join gevaerts [0] (
10:43:35 Join ChanServ [0] (ChanServ@services.)
10:43:35Mode"#rockbox +o ChanServ " by
10:43:35peturgmail and then forwarded
10:43:51LloreanI just set a filter in gmail for it to skip the inbox and go straight to being archived.
10:44:05GodEaterLlorean: you mean "mute", but more complicated.
10:44:28GodEaterand won't wake it up again in the (unlikely) event you got addressed directly in it
10:44:28JdGordonhey gevaerts, should the "reconnect usb mode" debug option be in the svn builds?
10:44:30*gevaerts doesn't use gmail at all, and quickly set up a procmail filter this morning
10:45:06LloreanGodEater: I don't want someone mentioning me on it, then waking up in the morning to find my phone is freaking out again because it has over 400 emails on it.
10:45:15gevaertsJdGordon: probably not. I'll remove it some time today (I guess the "keep out" label is enough to make this non-urgent)
10:45:16LloreanAt about the 480 mark, my mobile email client starts giving up, I think
10:46:16GodEaterI don't think I've ever tried a thread that large on the gmail mobile client - so no idea how well my phone would cope.
10:46:55*gevaerts wonders if he's the only one who noticed that gsoc _doesn't_ require a gmail account
10:48:24GodEaterI don't use anything other than my gmail account these days
10:48:31GodEaterso I wouldn't have noticed/cared
10:48:52*JdGordon too
10:49:01*JdGordon also not getting any gsoc emails
10:54:14 Quit scorche (Read error: 104 (Connection reset by peer))
10:54:50 Join scorche [0] (i=Blah@rockbox/administrator/scorche)
10:57:50B4gder75 UKP shipping for a 13 UKP device is a bit too much...
10:58:01B4gder(that sansa c240 deal)
10:58:24pondlifeB4gder: Where was that?
10:58:37GodEaterpondlife: FS 8807
10:58:39JdGordongroup buy to make it worth while?
10:58:46 Join ze0 [0] (
10:59:06 Quit ze (Read error: 104 (Connection reset by peer))
10:59:11 Nick ze0 is now known as ze (
10:59:15B4gderit's just the Swedish shipping that seems absurdely expensive from there
10:59:25 Join kbblak [0] (
10:59:26pondlifeI could post it to you?
10:59:40pondlifeUK shipping is £3.29
10:59:53*gevaerts notes that their sales still seem slow, and now really considers to offer them 100 or 200 euro for the lot
11:00:14pondlifeDo the C200's have a v2 ?
11:00:32pondlifeSo that's not a guaranteed target?
11:00:33JdGordongevaerts: in the hope one of them is a v2?
11:00:39 Quit cbr|w (Read error: 104 (Connection reset by peer))
11:00:46B4gderlinuxstb just said the one he got from there was a v1
11:00:47gevaertsJdGordon: no. To build a huge Tower of Rockbox
11:01:04JdGordonyou wont beat the seattle one.. no matter how hard you try!
11:01:18*Llorean is slowly collecting to make the champion tower on his own. :-P
11:01:19*gevaerts tries to find it
11:01:43***Saving seen data "./dancer.seen"
11:01:43pondlife£9.99 for a C240 without radio!
11:01:53 Join cbr|w [0] (n=cbr@
11:01:57pondlifeThat's worth a punt.
11:02:22gevaertsJdGordon: If I have 250 c200's I'll beat it easily :)
11:02:48JdGordonna, mp3 players are useless stackers
11:03:03LloreanJdGordon: Gigabeat Fs stack really well.
11:03:08JdGordonstupid rounded edges
11:03:12LloreanNo protrusions, so you can just use 'em like bricks.
11:03:14JdGordonand protruding buttons
11:03:54gevaertsAre audio cables allowed in the structure ?
11:04:15GodEaterI don't see why not
11:04:18*Llorean would imagine "only what the device came with" as a fair restriction
11:04:33GodEaterthat would include docks then
11:05:10gevaertsThere is a Tower on the wiki that uses docks
11:05:24B4gderpeople do anything to cheat!
11:05:47B4gderI would claim a clean tower to be nothing but main players, no remote, no cables, no cradles
11:06:06gevaertsI want two towers, and an audio-cable suspension bridge between them
11:06:16JdGordonremotes and docks are fine
11:06:18GodEaternow *thats* ambition
11:06:30 Part kbblak
11:06:57*Llorean is sad that the Tango-style icons don't cover all the Rockbox targets.
11:06:58JdGordongevaerts: if you can suspend a dap on the cable as the bridge then that would be freeking sweet
11:07:28LloreanI was just thinking of adding them as icons in the forums, so people could put a very tiny picture of their player in their signature.
11:07:35*Llorean *tries* to do something friendly every now and then.l
11:07:41 Join barrywardell [0] (n=barrywar@rockbox/developer/barrywardell)
11:07:49JdGordonplease dont
11:07:54JdGordonpics in sigs are painful
11:08:11*GodEater recommends elinks to JdGordon
11:08:30LloreanJdGordon: 16x16 icons aren't exactly *bad* :-P
11:09:45JdGordonoh, that small is fine
11:10:10pondliferemotes/docks/accessories can only be used it they are fully supported by Rockbox...
11:10:27pondlife(which is trivial on many of them)
11:10:42B4gderwe should get rockbox to a DVD player soon, *that* will make the most awesome towers!
11:10:43JdGordonexcept plenty of towers have then unsupported targets
11:10:48JdGordonor unfisninhed targets
11:11:05LloreanJdGordon: Yeah, nothing big at all. 16x16 or maybe 24x24, but they'd be quite tiny, and no larger than the "smilies" currently used.
11:11:55barrywardelldoes anyone know what DEV_USB0 really is? that seems to be what's preventing the current sansa bootloader from loading the OF
11:18:45*gevaerts only knows that it needs to be set for usb to work
11:21:23barrywardellspecifically, unsetting it causes the of to not load
11:23:28 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
11:27:18markunid3.c is compiled when building a bootloader. Should it?
11:28:20peturfor the startup sound
11:30:33*markun would like to make a copy of firmware/id3.c to apps/metadata and use that one for swcodec targets
11:30:55markunand add support for itunes gapless
11:36:35barrywardellgrrr. the H10 bootloader isn't loading the OF any more either. It's not the same problem as with the e200 though
11:40:19 Quit leox (Read error: 110 (Connection timed out))
11:40:25 Join fasmaie [0] (
11:42:22 Join Shaid [0] (
11:43:20 Join leox [0] (
11:45:05 Join webguest71 [0] (n=535b52f2@gateway/web/cgi-irc/
11:45:10 Join MartinR [0] (n=5848d651@gateway/web/cgi-irc/
11:45:54 Quit MartinR (Client Quit)
11:46:17amiconnmarkun: Duplication is bad...
11:47:11linuxstb_markun: Does the fact that id3.c is in firmware prevent you doing itunes gapless?
11:52:01 Quit barrywardell ()
11:52:09*preglow blinks at the new mod player patch
11:55:44*amiconn wonders what itunes gapless is
11:57:31linuxstb_Just some metadata itunes writes to files, similar to the Lame header IIUC
11:58:05 Join barrywardell [0] (n=barrywar@rockbox/developer/barrywardell)
11:58:06markunlinuxstb_: the format of the id3 tag is basically the same as the one used in itunes aac files and I don't want to duplicate or move that code to system.
11:58:50markunamiconn: itunes stores the exact length of the audio file in a tag, just like lame
11:59:09pondlifeBut a different tag?
11:59:09markunoops, that's what linuxstb_ wrote
11:59:29JdGordonhas the samsung t10 come under anyones radar?
11:59:41linuxstb_markun: I know it's not ideal, but I think moving those functions to id3.c is cleaner than duplicating the file.
12:00:03markuncan't the hwcodec targets use apps/metatada/id3.c?
12:00:18linuxstb_Unless you want to attempt to move the hwcodec engine into apps/ - which I attempted to do once, and think it's feasible...
12:00:38markunsounds like a better option
12:00:41*pondlife does a little unification jig.
12:00:43 Join Shaid3 [0] (
12:00:49markunlinuxstb_: what made you stop?
12:01:03linuxstb_I got stuck... I emailed the dev list for some help, but I don't think anyone replied.
12:02:14*amiconn would rather keep the current hwcodec engine where it is
12:02:18markunJdGordon: it uses a telechips soc?
12:02:27markunamiconn: why?
12:02:50JdGordonthe reason id3.c is in firmware is because mpeg.c needs it right? well.. as a quick hack to move it, couldnt apps set the required function pointers to get around the firware->apps problem?
12:03:02linuxstb_markun: Here is my email -
12:03:08JdGordonmarkun: does it? trying to find out... might get one if its a possibile target
12:04:22 Join MethoS [0] (
12:04:30amiconnmarkun: Note that I said *current*. When you look at mpeg.c, you'll see that it's tied to the hardware, i.e. it is firmware code
12:04:50markunJdGordon: maybe not. I just googled a bit and found this post, but didn't read it carefully:
12:04:54amiconnThe proper solution would be to attempt engine unification - which requires new code at the firmware layer
12:05:08amiconnAlso note that mpeg.c handles *both* playback and recording
12:05:57linuxstb_amiconn: But you can equally argue that the id3 code belongs in apps... IMO a better compromise would be to have some MAS-specific code (if it has to be) in apps/, rather than id3 metadata parsing in firmware/.
12:06:25markunJdGordon: yes, it's the TCC7801 just like the D2
12:07:16linuxstb_amiconn: And it could be a start towards unification - i.e. abstracting the hardware playback/recording code a little.
12:07:39 Quit ColinT (Remote closed the connection)
12:07:51 Join ColinT [0] (
12:09:24JdGordonmarkun: thanks'
12:11:48 Quit webguest71 ("CGI:IRC 0.5.9 (2006/06/06)")
12:11:54 Join MethoS- [0] (
12:13:21 Join gevaerts_ [0] (
12:14:18barrywardellthe H10 bootloader problem (not loading the OF) happened somewhere in between 14004 and 14015.
12:16:20barrywardellIt looks like somewhere in amiconn's clock frequency cleanup(14005), or in the H10 LCD speedup(14006)
12:18:40 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
12:19:16 Quit Shaid (Read error: 110 (Connection timed out))
12:19:16 Nick Shaid3 is now known as Shaid (
12:20:48barrywardellamiconn: the problem is in setting CLCD_CLOCK_SRC in lcd_init_device. Can I safely #ifndef BOOTLOADER that line?
12:21:56barrywardellamiconn: and do you know does it also need to be done for the small H10 lcd driver?
12:23:26 Quit gevaerts (Nick collision from services.)
12:23:30 Nick gevaerts_ is now known as gevaerts (
12:24:00 Join tvelocity [0] (
12:26:02Nico_PGodEater: I've experienced FS #8807 just now
12:27:03GodEaterhurrah finally!!!!
12:27:07GodEaterit's not just me
12:28:16linuxstb_DIdn't jhMikeS mention there might be issues on the 5.5g?
12:31:03 Quit maraz (Remote closed the connection)
12:31:21 Quit MethoS (Read error: 113 (No route to host))
12:33:19GodEaterhe did
12:33:26GodEaterbut when I mentioned that he was right this morning to him
12:33:43GodEaterhe thinks the right place to fix the issue is firmly in buffering
12:36:36linuxstbYes, I recall him saying that as well...
12:36:44JdGordonis there any real reason we cant move the hwcodec fm recording into the rec screen like swcodec?
12:38:04linuxstbBeing able to record from the FM screen sounds like a useful feature to me - one to implement for swcodec...
12:38:45pixelmaJdGordon: no, I'd rather want it the other way round
12:38:49*linuxstb would even quite like a "timelapse" type feature for FM...
12:39:12linuxstb(being able to pause/rewind live radio)
12:39:20GodEatertivo for your ears
12:39:25JdGordonnow your being rediculous! :p
12:39:32GodEaterthat's a damn good idea actually
12:39:51*GodEater thinks it should get moved to the GSoC project ideas list immediately
12:40:04linuxstbRecording to RAM only, with MP3 compression would be nice for those targets with lots of RAM
12:40:48pixelmaand I'd also like the way the prerecording time is shown in the hwcodec FM screen everywhere else (counting up as it's buffering until it reached the set prerecording time, so you know when it reached "full" state)
12:41:39*petur wonders if it would be usefull to increase the pre-rec time
12:41:42linuxstbGodEater: I think it would make a decent SoC project - it's far from trivial...
12:41:47pixelmaif you do FM recordings it's much more convenient to do so from the FM screen - you can even change stations
12:41:56JdGordonis the fmrecorder the only archos with fm?
12:42:12GodEater <−− no more battery woes ?
12:42:14linuxstbI think some recv2 models have FM
12:42:21pixelmano, the OndioFM too (and some V2-Recorders
12:42:54JdGordonis there enough room under the current screen on them to stick the presets list if its in preset mode?
12:42:59*JdGordon going viewport crazy :p
12:43:10linuxstb_s/viewport// ?
12:44:38pixelmadepends on the user font
12:45:07JdGordondoesnt look like it :(
12:45:47pixelmabut I guess you can't see a list (maybe just one) even if you chose a tiny font
12:46:37 Quit barrywardell ()
12:49:38 Quit fasmaie (Read error: 113 (No route to host))
12:50:55 Join linuxstb__ [0] (
12:50:58Nico_PGodEater: have you tried increasing the buffering chunksize?
12:52:36GodEaterNico_P: I have no svn at work anymore
12:52:37GodEaterso no
12:52:44Nico_Pah. I'll try it
12:55:02 Join linuxstb___ [0] (
12:55:06 Quit linuxstb (Nick collision from services.)
12:55:12 Quit linuxstb_ (Nick collision from services.)
12:55:21 Nick linuxstb___ is now known as linuxstb (
12:55:36 Quit linuxstb__ (Remote closed the connection)
12:56:18Nico_PBagder: "we should probably just make it suck everying that isn't .svn into a tarball" => do you know of svn export?
12:56:46linuxstbNico_P: That sounds sensible.
12:56:54B4gderyes I know about that, but it seems like a wasteful thing to run when there's a complete checkout around all the time
12:57:25markunif we switch to git you just have to ignore 1 .git folder ;)
12:57:52*linuxstb thought the gsoc mentors list was for private discussions amongst mentors, and gets annoyed by all this off-topic spam
12:58:32Nico_PB4gder: makes sense... I guess if you know there won't be any cruft in the checkout you can just as well filter .svn out
12:58:59B4gderthat's my thinking, yes
12:59:15B4gderI have this single checkout that I base all my scripting against
12:59:36B4gder... well, almost all at least.. and it should be free of junk
13:00:27Nico_Pin case you want to be sure to get rid of the junk, maybe svn ls could come in handy
13:01:48***Saving seen data "./dancer.seen"
13:02:45B4gderah yes
13:03:58linuxstbYes, "svn ls -R" seems a nice command...
13:04:55B4gderbeats find . \! -path "*.svn*"
13:05:29*linuxstb looks forward to B4gder commit to remove FILES... ;)
13:05:38B4gderstay tuned!
13:10:06B4gderalmost 22MB uncompressed...
13:11:12n1sNico_P: have you looked at fs#8601 (disk spinning at trackchange with dircache disabled) ? I looked at it but don't quite understand the difference between playlist_end and end_of_playlist :(
13:13:15Nico_Pn1s: yes I've looked at it... It makes me want to rewrite the track skipping handling
13:13:39 Quit jcollie ("Ex-Chat")
13:14:02n1splease do :)
13:15:34pondlifeNico_P: And keep adding comments - I think I got stuck about the same place as n1s.
13:16:14Nico_Pyeah the difference is rather subtle
13:16:36pondlifeTiming, or checked/unchecked?
13:16:46 Join disorganizer [0] (n=c2785409@gateway/web/cgi-irc/
13:17:18Nico_Pend_of_playlist is local to audio_check_new_track, whereas playlist_end is global to playback.c (true meaning the last track of the playlist is buffered IIRC)
13:17:45 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
13:19:40B4gderis it bad to include the manual dir in the tarball?
13:19:49B4gderor should we take that as a bonus?
13:20:01B4gderI think so too
13:20:04LloreanPeople working on the source should be aware of the manual anyway.
13:20:29B4gderls -l test.tar.bz2
13:20:29B4gder-rw-rw-r−− 1 dast dast 17911983 Mar 26 13:17 test.tar.bz2
13:20:57B4gderI'm thinking of allowing it to be rsynced
13:21:06markunB4gder: can you try 7zip?
13:21:16Nico_PBagder: the source or the bzip2?
13:21:18B4gderI'll try
13:21:24Nico_Pand what's that in -h?
13:22:15markunB4gder: anything important I forgot?
13:22:18B4gderNico_P: I guess both the source and the bzip2, but when I think about it I guess that people that can do other ports than HTTP can just as well svn it
13:23:07Nico_Prsync could be nice, but would it be efficient on the compressed file?
13:23:16B4gdermarkun: I think you can spell out the (potential) DFU-problems to fix
13:24:57B4gderls -l test.tar.7z
13:24:57B4gder-rw−−−−−−- 1 dast dast 15992402 Mar 26 13:24 test.tar.7z
13:24:58markunB4gder: there is a winodws tool which does work, but I'll add it to the todo
13:25:20B4gderyes, as using windows will be painful...
13:25:31markunindeed ;)
13:26:04B4gderalmost 2MB smaller
13:28:15markunquite a bit
13:29:04B4gderbut at this size, the compression takes quite some time
13:29:48linuxstbpixelma: (and others...) I don't know if you read the discussion last night, but do you have any objections to me making Rockbox's WPS parser reject a WPS if the bmps fail to load?
13:29:54A-4rtc is as good as done, needs some cleaning and a final check tho. Also copied i2c from the gigabit port and added a read function
13:29:58linuxstb(this is so checkwps can do the same)
13:30:11markunA-4: nice
13:30:24pondlifelinuxstb: I think it's good to be strict
13:31:11linuxstbpondlife: Yes, me too...
13:31:21pondlifeGo for it
13:32:02markunA-4: anything on the todo you want to look at next?
13:32:55markunA-4: I'm adding some stubs to get the bootloader to compile and then we can use it as the basis for some test programs
13:33:05A-4doesnt realy mather 2 me
13:33:25markunmaybe a driver for the button/touch controller?
13:33:28A-4cool, but i dont have my m6 yet =[
13:33:33markunah yes :)
13:34:31A-4i could give it a try
13:35:31 Join dionoea_ [0] (
13:36:14*Nico_P 's S is only displaying the statusbar :/
13:36:30linuxstbpondlife: I have the patch, but it needs a little more testing, so I'll probably commit tonight if no-one shouts.
13:38:13 Quit dionoea (Read error: 104 (Connection reset by peer))
13:38:57disorganizerlinuxstb: would the wps parser give out an error? or just default to the default-wps like now?
13:40:02linuxstbJust default like now.
13:40:03pixelmalinuxstb: I don't know, guess it could help if people have too many images and don't know that they do. If anything can make a WPS fall back to the built-in one it's sometimes harder to find the real reason(s). I know those WPSs are broken but it helped me more to see a broken image (or the broken WPS before this fall-back mechanism was invented). Could be that it's a bit different for me in comparison to those who hit the buffer limit and that wa
13:40:41linuxstbpixelma: Then you need checkwps ;) The intention is for that to give you more useful feedback about the problems.
13:40:49disorganizerlinuxstb: would it be hard to make it give out some kind of error message instead of defaulting? like "missing conditional" or in this case "too many images"
13:41:15linuxstbdisorganizer: Yes - we don't want to have to store all those error strings in Rockbox (and translate/voice them...)
13:41:42disorganizeryou got a point. maybe instead of defaulting just display a text "error in wps, use checkwps to verify"
13:41:57Lloreandisorganizer: There's a PC-side tool that uses the exact same code Rockbox does. If Rockbox won't display the WPS, then the PC side tool will give you a more helpful error message.
13:42:14pixelmadisorganizer has a point. It would be real nice if there was at least *some* notice that loading the WPS failed (look at all the questions in the forums or here of people who try to load broken themes/WPSs and don't know what's going on)
13:42:42linuxstbYes, a basic "WPS error" message could be nice.
13:42:50Nico_PI agree
13:42:52disorganizerproblem imho is that people do come here often with the question "the default wps shows if i load this and that theme". if they get any error message pointing in the right direction this could pevent channel-floods :-)
13:43:33disorganizeri wouldnt even show any default wps then. or add the text to the default wps any ONLY show it on error :-)
13:43:40disorganizera kind of "error-wps"
13:43:59pondlifeGood idea..
13:44:05linuxstbdisorganizer: That's because of all these nasty unofficial builds and unofficial themes everywhere.... ;) i.e. things others have done to make life hard for users.
13:44:33*linuxstb is looking forward to the new
13:44:45disorganizeryes, but also many changes in the wps syntax lately :-)
13:44:46Nico_PI don't think we need an "error WPS". the default is already just that. a splash should be good enough IMHO
13:45:28LloreanJust an "WPS Broken" splash?
13:45:41Lloreanthen business as usual
13:45:47disorganizernico_p: in fact even a better idea than a wps. a "working" wps on error implies to the user that something is working and just things arent shown. if we just show a splash-screen this will be a definite indication of failure
13:46:24linuxstbdisorganizer: The point is that this shouldn't happen - i.e. users should be using (the new) and all themes there will be guaranteed to work with official Rockbox.
13:46:26Nico_PI was suggesting showing a splash screen, then the default WPS
13:47:09*linuxstb is happy with a splash, then the current default.
13:47:14disorganizerok. but please let them press a button to show the default wps. and the splash should say: error in wps, use checkwps to verify. press any key to continue with the default wps. or something like that
13:47:15pixelmalinuxstb: re. "then you need checkwps" - does it run on the target? I sometimes did minor fixes to a .wps file directly on target using the text editor...
13:47:39linuxstbpixelma: No, it doesn't. Although it probably wouldn't be too hard to make it a viewer - a nice idea...
13:48:13Nico_Plinuxstb: how would that be done?
13:48:29linuxstbAlthough the errors are DEBUGF statements, so wouldn't be compiled on target....
13:49:16linuxstbNico_P: I'm not sure it could be now...
13:49:44Nico_Pyeah I don't think it would be trivial unfortunately
13:49:56linuxstbNico_P: My initial thought was simply that tools/checkwps.c could become a plugin. But that won't work as it relies on DEBUGF statements for the useful error messages...
13:50:22Nico_Pyes, and all the rb-> in front of "system" calls
13:50:36pixelmamy checkwps is "staring at the code of the wps" most of the times anyways... ;)
13:51:11linuxstbNico_P: Yes, it would be a program based on checkwps.c, not the same .c file.
13:51:31 Nick dionoea_ is now known as dionoea (
13:52:33linuxstbpixelma: Anyway, if you want to try it, just type "make checkwps" in the tools/ directory of the source, then do something like (if you're in tools/) ./checkwps ../simdir/archos/.rockbox/wps/mywps.wps
13:55:54 Join CaptainKewl [0] (
13:56:00*disorganizer had the linuxstb-statusbar-hack running today on his way to work, and it is indeed nice :-)
13:57:30 Quit Shaid (Read error: 110 (Connection timed out))
13:58:54 Join MethoS-- [0] (
14:02:57amiconnlinuxstb: checkwps.c as a plugin could just redefine DEBUGF()
14:03:17 Join Thundercloud [0] (
14:03:51disorganizerlinuxstb: did you see the list-size patch? will the screen-clearing conflict with the possibility to have a custom statusbar in general or is it just the method you hack works that causes the screen clearing to disturb sb display?
14:06:10*Nico_P voluteers to add the "WPS load error" splash after he eats his lunch
14:06:23linuxstbamiconn: To what? The DEBUGFs are in the core code (wps parser), so making them do something on target would affect the core. But maybe they could just encode the error in a few ints somewhere...
14:07:12 Join Rincewind [0] (
14:07:30amiconnI thought checkwps would have the parser compiled in?
14:07:47linuxstbdisorganizer: No, I haven't looked at that patch. My solution (hack?) so far is to make the status bar contain viewports which completely fill the screen (along with the list viewport), so the status bar in effect clears the unused parts of the screen.
14:08:07 Join piga [0] (n=leonardo@
14:08:24linuxstbamiconn: Yes, it does. But I was thinking of the plugin version of checkwps using the core code (via rb->), rather than compiling it in. But compiling it into the plugin is a better idea...
14:09:47 Quit piga (Client Quit)
14:09:57 Join nplus [0] (n=nplus@141.25.Globcom.Net)
14:10:17 Join piga [0] (n=leonardo@
14:10:33linuxstbSo yes, I think that could work...
14:11:07*pixelma still wonders about the long(er) delay between showing the bootlogo and the start screen (leaving a blank screen for a while and make Rockbox feel sluggish)
14:11:16*amiconn wonders whether disorganizer is aware of the variety of targets rockbox is running on when suggesting splash texts
14:11:31linuxstbsplash essays... ;)
14:12:04*B4gder whistles
14:12:45*linuxstb does the death of FILES dance
14:13:55linuxstbB4gder: Did you include the manual?
14:14:14B4gderthis is a plain 'svn ls -R' in the svn root
14:14:30B4gderuh trunk
14:14:35 Join LambdaCalculus37 [0] (n=44a0430f@gateway/web/cgi-irc/
14:15:35 Quit MethoS- (Read error: 110 (Connection timed out))
14:16:01*disorganizer wonders if it would be possible to let checkwps correct common errors. like auto-converting from older (official) wps syntax (vp without font to with font, for example)
14:16:32amiconnB4gder: where did go?
14:16:51linuxstbdisorganizer: What was "vp without font" ?
14:16:59B4gderamiconn: I guess it wasn't copied to svn
14:17:02disorganizeramiconn: no :-) we dont even need to display a graphic, just some kind of text.
14:17:22amiconndisorganizer: It was the text (specifically its length) I was referring to
14:17:44disorganizerlinuxstb: wasnt there a time when the viewports used %v without the font? (im not sure though whether it was still in the patch stage back then)
14:17:48linuxstbamiconn: I think it's in SVN though - the www module
14:17:49disorganizeramiconn: let it scroll?
14:17:55amiconnB4gder: Could we get that back, please?
14:17:56 Quit MethoS-- (Read error: 113 (No route to host))
14:18:11amiconndisorganizer: Check how many *screens* that text would fill on archos Player...
14:18:13B4gderI'm sure we can, we need to poke zagor
14:18:13linuxstbamiconn: Ignore me...
14:18:25*amiconn likes to have grepable logs
14:19:07disorganizeramiconn: hmmm. what about "wps brkn:chkwps!" then ? ;-)
14:19:29*Nico_P was thinking of a plain "WPS load error"
14:20:04linuxstb"Invalid WPS" ?
14:20:11LloreanNico_P: "WPS Broken" or "Invalid WPS" is better I think
14:20:18LloreanLoad Error makes me think the load failed, rather than the WPS is bad.
14:20:48Nico_Pgood point
14:21:05disorganizerwould "wps error" for the targets with small "screen" fit? and a pointing to checkwps for bigger screens?
14:21:36 Join midgey [0] (
14:21:58*Llorean thinks pointing to checkwps can be reserved for the manual
14:22:18amiconn"Invalid WPS file" would work
14:23:32linuxstbDo we want different messages for wps and rwps?
14:26:07JdGordonno need
14:27:04Nico_PWPS and RWPS are loaded with different code, so having disting messages is quite easy if we don't localise them
14:29:37Nico_Phow come colors don't change immediately on a theme change? a list-viewports issue?
14:30:03JdGordonwhich ones?
14:30:18JdGordonthey shuold, and in my (obviously crappy) testing they did
14:30:29linuxstbJdGordon: Did you see the latest FS task?
14:30:39*linuxstb just noticed his reply...
14:31:18 Join japc [0] (n=japc@
14:31:47JdGordonok, wtf? the colours were working when i checked in :'(
14:31:50Nico_Pfg and bg
14:31:56*JdGordon has a look
14:32:12*JdGordon tihnks the BB commit may have rebroekn it
14:33:57JdGordonyep, accidently reverted a bit i meant to commit
14:34:09 Quit midgey ()
14:36:06 Quit japc (Client Quit)
14:36:31 Join japc [0] (n=japc@
14:36:37 Quit japc (Read error: 104 (Connection reset by peer))
14:36:52 Join japc [0] (n=japc@
14:36:55JdGordonthose mr100 and m3 pics looks odd amongst the other ones :p
14:37:24pixelmaI'm working on "fixing" that ;)
14:38:15JdGordonwhile your at it... wanna add some height to the c200 image so its not up in the air?
14:39:01 Quit japc (Client Quit)
14:39:24 Quit CaptainKewl (Read error: 110 (Connection timed out))
14:39:29 Join japc [0] (n=japc@
14:39:42pixelmaI believe that was discussed when adding the picture (and some things were tried) but I don't remember why and how this way was chosen
14:40:12 Quit piga (Read error: 110 (Connection timed out))
14:40:40 Join piga [0] (
14:41:01 Quit japc (Client Quit)
14:41:19 Join japc [0] (n=japc@
14:41:25*disorganizer wonders if it would be possible to add an indication of the backlight status to the simulator somehow. for example drawing a white border around the lcd if bl is on.
14:42:12*petur thought background color changed
14:42:17Nico_Phmm the "invalid WPS" splash is hidden by "settings lodaded" when the WPS is loaded as part of a theme
14:42:22Nico_Ppetur: not on color targets
14:42:24JdGordonwe used to and it does
14:42:51pixelmawe once had a "backlight: on" notice in the console...
14:43:01peturNico_P: right... remembered seeing it for the remote of h300, not the main lcd :/
14:43:30 Quit japc (Client Quit)
14:43:47 Join Cricri [0] (n=84f4f619@gateway/web/cgi-irc/
14:43:48 Join japc [0] (n=japc@
14:44:44 Quit japc (Client Quit)
14:45:12 Join japc [0] (n=japc@
14:47:04 Join kugel [0] (n=kugel@unaffiliated/kugel)
14:47:49disorganizer@pixelma: no backlight message. just checked
14:48:08JdGordonit was removed months ago
14:48:27pixelmaI know, it was removed. That's why I said "once had"
14:48:34Rincewindabout the gsoc proposal, I see that there are two text fields, one for the abstact and one for the longer proposal, write these independently from each other or is it ok to put an overview in the abstract and details in the main field, without paraphrasing the abstract again?
14:48:47 Quit Cricri ("CGI:IRC (Ping timeout)")
14:49:02Rincewindoh, there is a " should I" missing
14:50:08 Join Cricri [0] (n=84f4f619@gateway/web/cgi-irc/
14:51:36 Quit japc ("Ex-Chat")
14:51:41JdGordonRincewind: either or
14:51:43B4gdertarball now as 7z tarball
14:51:58 Join japc [0] (n=japc@
14:52:09B4gderI'll just need to tweak the daily build script too and some more
14:52:27 Join clinx [0] (
14:52:39 Join TMM [0] (
14:53:13RincewindJdGordon: I suppose the question is, do you recieve the abstract and the proposal independantly from each other or do you read the abstract first and then the proposal?
14:53:15clinxcan i use a modified rockbox logo for my own program
14:53:34kugelI get this compiling error again:
14:54:54kugel2 days ago, I installed a teamspeak, which installed few libasound packages. Compiling sim still worked.
14:54:54 Quit Cricri (Client Quit)
14:55:39JdGordonRincewind: I dunno
14:56:02kugelYesterday, afer I booted my PC up, I had this issue. So I unstalled teamspeak and the packages again (without reboot) and compiling worked again. And today, there is this again.
15:00:47 Nick JdGordon is now known as JdGordon|zzz (n=jonno@rockbox/developer/JdGordon)
15:01:51***Saving seen data "./dancer.seen"
15:02:03kugelHmm, I reinsatlled libasound2 and it seems to work (until next reboot probably)
15:03:08linuxstbRincewind: If I remember correctly from last year, we view the applications via the google website, and see the abstract first, followed by the detailed description.
15:04:07B4gderamiconn: is back
15:04:38Rincewindlinuxstb: that's good, then I don't have to repeat myself. I want to put the general idea in the abstract (which should be understandable without knowing Rockbox), this wouldn't make much sense in the detailed description
15:05:10 Join jgarvey [0] (
15:06:29 Join midgey [0] (
15:06:33Rincewinddoes anyone here know a quick way to count the characters of a text? (in linux commands, or emacs)
15:06:54Aveunicode chars?
15:06:57Avethats tougher
15:07:16Avebut wc will do, or simply "ls -al" or so
15:07:32Aveassuming ascii or isolatin encoding
15:07:34B4gdercount-lines-region in emacs speak
15:08:20pixelmaB4gder: do you remember why the small c200.png is used as is? Should I bother adding some transparent space above or below and make it the same overall height as the others?
15:08:33RincewindB4gder: thanks, that's perfect
15:09:12B4gderpixelma: nah, i think it works fine as it is
15:09:26 Join BobShield [0] (
15:10:12kugeldisorganizer: I think I fixed the issue
15:11:04markunpixelma: We're making some progress with the meizu m6/m3 port. Do you feel like making a few vector drawings of the various models?
15:11:43markunthe m6sp and m6sl look the same, except for the label on one of the buttons
15:12:03pixelmaI first feel like finishing the M3 drawing and then making an M:Robe 100 one...
15:12:33markunof course. We don't have any code running on the meizu anyway
15:13:26pixelmayou only get a drawing once there is a build ore even a manual available for it :P
15:14:16markuna good motivation to hurry up then :)
15:15:23gevaertsWe could finish the manual ;)
15:16:03disorganizer@ all who are intersted regarding mf: i found the page from 05/2007. wouldnt that solve the font-glyphcaching "problem" with multifont?
15:16:14disorganizer@ kugel: will test asap
15:17:17disorganizer@kugel : please put it on fs, if possible as modification of custom_list_vp_v3.diff
15:17:45markundisorganizer: I started to work on (hopefully) better font system with support for multiple fonts, but lost my interest a bit
15:18:00 Quit midgey ()
15:18:01kugeldisorganizer: No, I have removed the hack in my version too, and added something new
15:19:20 Join mf0102 [0] (n=michi@
15:19:26disorganizer@markun : and if i say please ? ;-) could you share your thoughts in the multifont fs entry? maybe someone else could take over from there
15:19:52markundisorganizer: ok, but not now
15:20:11 Quit B4gder ("Lämnar")
15:20:21disorganizer@markun: np, tnx
15:23:01 Join midgey [0] (
15:23:33 Quit XavierGr (Read error: 110 (Connection timed out))
15:26:08linuxstbdisorganizer: I don't think anyone wants freetype in Rockbox...
15:27:59disorganizerwhy not? at least we could try to copy the caching algorythm :-)
15:28:13markungo ahead
15:28:30disorganizer@kugel : if you upload to fs i can test
15:28:43linuxstbdisorganizer: For all the reasons listed at the bottom of that wiki page...
15:28:49kugeldisorganizer: Yeah, I'm at it.
15:28:56kugelI just need to post a comment
15:29:52disorganizer@linuxstb: i missed the m-word in there. so propably the page should be moved to the garbage ?
15:30:32linuxstbIt's just a proposal - it's not doing much harm there...
15:32:02kugeldisorganizer: done
15:32:27amiconnmarkun: fyi - the rockbox font system was not written as a whole by that chinese contributor, only the lru caching was
15:32:52amiconnThe font system itself is based on microwindows (or whatever it's called today)
15:33:28disorganizerlinuxstb: ok. just thought that it would make more sense in the museum as it is would never be implemented
15:34:31 Join johnny7 [0] (n=57a773a0@gateway/web/cgi-irc/
15:35:19kugelWhat means "NIH" (listed in the disadvantages)
15:35:20 Quit johnny7 (Client Quit)
15:36:02LinusNNot Invented Here
15:36:54kugelAh ok
15:37:28 Quit clinx ("Verlassend")
15:37:45 Join webguest1234 [0] (n=57a773a0@gateway/web/cgi-irc/
15:39:21 Join markun_ [0] (n=54f18042@rockbox/developer/markun)
15:39:35markun_amiconn: ah, thanks
15:41:23Nico_Pkugel: what page are you referring to?
15:41:55 Quit markun_ (Client Quit)
15:42:36disorganizer@kugel: is it on purpose that it scrolls up on the line before the last?
15:42:55 Quit webguest1234 ("CGI:IRC")
15:43:10kugeldisorganizer: You mean in the settings?
15:43:28Lloreankugel: Please, respect the idea of trying to save binsize where possible, and not cluttering the UI with "reset" options for every submenu.
15:43:36kugelPlease look at apps/gui/bitmap/list.c, I have made a comment regarding that
15:43:52kugelLlorean: Then remove reset colors from SVN please
15:44:17disorganizerif i see a list, and scroll down (like in fileview) then it scrolls up in the line before the last line. (lets say i have 10 lines in the displayed vp, but 15 lines to display, i go down until i reach line 9 and then it scrolls instead of when i selected line 10)
15:44:24kugelLlorean: I have stated why I want that option
15:44:26Lloreankugel: "They did it so I get to" is not a valid argument
15:44:30Lloreankugel: You stated you want it for testing.
15:44:39LloreanA .cfg file works for testing.
15:44:45kugelLlorean: That's 1 reason, not the only one
15:44:50LloreanThat's the reason you listed.
15:45:45kugeldisorganizer: That's normal behavior
15:46:01kugeldisorganizer: This happens w/o my patch too
15:46:20disorganizer@kugel: ok then. didnt notice before, maybe didnt pay attention enough :-)
15:46:58Lloreankugel: In the official build, where themes come from the official repository (so, they're non-broken) what purpose does that option serve?
15:47:43kugelLlorean: "They did it too" is valid for me, since it shows that the bin size increase is worth a reset option at times
15:48:36 Join Cricri [0] (n=84f4f619@gateway/web/cgi-irc/
15:48:44 Quit Cricri (Client Quit)
15:48:51 Join Cricri [0] (n=84f4f619@gateway/web/cgi-irc/
15:49:07Lloreankugel: That doesn't answer my question regarding its purpose.
15:49:31LloreanIn fact I have no idea why you added menu options for resizing the list viewport in the first place, honestly
15:49:45LloreanIt doesn't seem like a thing that should be manipulated interactively.
15:49:46linuxstbIMO, the user can simply load a theme to reset all theme settings back to a sensible state - so even "reset colours" can go.
15:49:53Lloreanlinuxstb: Agreed.
15:50:04kugelLlorean: I allready told you, that the purpose is to have a quick way to do 4 jobs in one. Especially since this 4 jobs might be painful when the list is messed up extremly
15:50:22Lloreankugel: If the list is messed up too extremely then you can't get to the option either. Argument is null.
15:50:44kugelI could, since the icons where still visible when I tried it
15:50:50*gevaerts wants to know how Bagder managed to get 7zip compress better than 7zip
15:51:02LambdaCalculus37Lack of coffee?
15:51:03kugeland if it's null, what's the reason for having the reset color option?
15:51:06Lloreankugel: Then you had enough information to use the other options.
15:51:16Lloreankugel: there is no reason for the reset color option, really. It should be removed.
15:51:33Lloreanlinuxstb: I really, really don't think you need to be able to resize the list from within the rockbox menu system anyway.
15:51:51kugelSo you like hidden cfg values?
15:52:36Lloreankugel: Where in the menu system can you describe WPSes?
15:52:37linuxstbFor themes, I think it's inevitable - as we make more and more things customisable, we don't want to bloat the UI with code to edit everything
15:53:05disorganizeri must admit i can see that point of llorean. the viewport-sizing is extremely theme dependant as the wps and other parts of the viewport system needs to fit. so setting this will propably in the future break the themes, this it should not be settable by the user
15:53:06LloreanMost aspects of themeing don't need, or explicitly should not have, menu options.
15:53:24kugelThe options also offer to otimize the list settings quickly, without the painfull editting the theme.cfg with the rockbox text editor and reloading all the time
15:53:50linuxstbBut the question then becomes what is a theme setting, and what isn't? e.g. is the status bar visibility setting part of the theme? Custom list positions and status bar visibility are interlinked...
15:53:52 Quit Cricri ("CGI:IRC (Ping timeout)")
15:54:03disorganizer@kugel: the user should not be able to edit that for the reasons i stated above. viewports have to fit against each other, especially when we talk about statusbar-viewportification
15:54:26Lloreankugel: A customized list position will be part of a theme. Since the theme designer knows the backdrop dimensions, and the location of additional viewports, there's no "trial and error", there's simple, objective numbers.
15:55:08Lloreanlinuxstb: I think "if it can conflict with non-interactively settable settings, it probably shouldn't be interactively settable" is at least a good basic guideline.
15:55:19LloreanExceptions will always happen though
15:55:47kugeldisorganizer: I don't get your point actually. Breaking themes isn't an issue as we know
15:56:36disorganizerit will be. imagine we have custom statusbars. so the statusbar viewport is using y-values 0-15 and 200-250. now the list viewport should use 16-199, and the user can not know this.
15:56:48Lloreankugel: Except you're trying to put in a "break this working theme" option.
15:56:54disorganizerif the user is able to change those values, he can easily break the theme
15:57:01 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
15:57:35kugeldisorganizer: Sorry, but I don't take a possibility, which is far away from being in svn into account
15:57:37disorganizerand the user normally wont need to change it anyways, as it will be part of a theme with visuals fitting together (hopefully)
15:58:09disorganizeractually i think its as far away from svn as any custom build :-)
15:58:12Lloreankugel: you HAVE to take it into account. It's a planned feature, and you're writing a patch with the intent to go alongside that planned feature.
15:58:39kugelI don't since there's no way to know how to take it into account
15:58:52 Join sarixe [0] (
15:59:04Lloreankugel: Not allowing users to dynamically resize the list is a good way to take it into account
15:59:31 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
15:59:55 Join Cricri [0] (n=84f4f619@gateway/web/cgi-irc/
16:00:12Lloreankugel: I don't understand what your problem is. There's certain aspects of your patch that make it unsuitable for inclusion, and you seem explicitly disinterested in trying to resolve these.
16:00:17LloreanThey cause usability problems.
16:01:00kugelI just don't see what's wrong about having this options, honestly
16:01:05disorganizerwhy not split it up into 2 patch-files. one with the function and not the settings, and one to add the settings :-)
16:01:13 Join DerDome [0] (
16:02:09Lloreankugel: 1) It provides an easy opportunity for a user to mess things up, 2) It wastes bin size while doing so, 3) It's completely unnecessary.
16:02:19 Quit TMM (Read error: 104 (Connection reset by peer))
16:02:49Lloreankugel: Whether you can see what's wrong with it or not doesn't change the fact that those aspects of the patch are unlikely to be accepted, so you might as well fix them now
16:02:53kugelLlorean: Why do we have settings anyway? We can edit the config, so why do we waste bin size on that?
16:03:11Lloreankugel: Now you're just being pointlessly argumentative.
16:03:24LloreanViewport size is not something you need, or have any good reason, to change while using the player.
16:03:45LloreanIt should always be set properly by the theme, and every theme should change it to a valid setting. It's not a "user configurable" option.
16:03:49linuxstbkugel: There has long been a view amongst the devs that Rockbox has too many settings - so we think long and hard about any new ones. I agree that manual editing of the list position doesn't see a very useful feature.
16:04:00LloreanJust like we don't let users directly set the playback buffer size. There are values determined by other values that don't need direct interaction.
16:04:11 Join toffe82 [0] (
16:04:58 Quit Cricri ("CGI:IRC (Ping timeout)")
16:05:03linuxstbkugel: So at the very least, I think you should make that part of the patch separate, so it can be discussed separately.
16:05:19linuxstb(and not distract from the rest of your patch)
16:06:21kugelI'd probably do it. But I'd rather focussing on the remaining TODOs for now before I mess around with 2 patches
16:08:33scorche|shit would take 5 min...if that...
16:09:02 Quit sarixe ("(EE) Failed to load "quit" module")
16:10:37Lloreankugel: Then why not remove the code, and create the second patch later?
16:11:18kugelBecause I want that for testing while I fix the other issues
16:11:45kugelIf you don't want it in the svn, ok. but I want that for testing
16:11:52LloreanThen you can have it locally?
16:12:16LloreanIf you want a reset option that's not meant to be there for users, put it in the debug menu.
16:12:32LloreanIt's a way of saying "I need this, but it's not meant to be there for the user version"
16:13:02kugelI'd want it for the user too, you know. You don't want to
16:13:18scorche|shit isnt just him...
16:14:04Lloreankugel: The patch tracker is "for things intending to be included in SVN." What part of this is confusing?
16:14:14kugel"You" was meant to mean Llorean exclusivly
16:15:15kugelLlorean: I come to that, when the patch reaches a committable state
16:15:46scorche|shkugel: you are missing the point...
16:15:54Lloreankugel: What's the harm in fixing it now, so that testers aren't mislead?
16:16:22LloreanYou know, at this point, that several committers including the one most likely to be the one who commits it lean in favour of not having it in the menus.
16:16:47scorche|shin fact, are there any committers around who would rather have the options in?
16:17:11LloreanOn an unrelated note, anyone here us RBUtility for more than one player?
16:19:44pondlifeI dp
16:19:51pondlifeGigabeat and H300
16:20:06Lloreanpondlife: Do you use the auto-detect feature ever?
16:20:17pondlifeYes, every time I switch
16:20:44LloreanWhich SVN revision's yours built from?
16:21:17pondlife1.0.4 off the wiki
16:21:19LloreanYeah, same here
16:21:27LloreanMine seems to not like autodetecting.
16:21:29pondlifeI can't build RbUtil here
16:21:35pondlifeWhich devices?
16:21:37 Quit piga (Read error: 110 (Connection timed out))
16:21:52 Quit japc (Remote closed the connection)
16:22:31Lloreanpondlife: And this test it worked. Gigabeat, e200, c200
16:22:44LloreanI'm wondering if it's just a case of "every time I've tested it, I was updating from a very old build"
16:22:53pondlifeI've never had it fail...
16:23:20kugelWhat do you guys think about customlist on remotes? I don't own a player with a remote, and honestly I don't know in which aspects the remotes are themable
16:23:32Lloreanpondlife: It fails silently for me. Just doesn't change which player is displayed.
16:24:05 Join japc [0] (n=japc@
16:24:15Lloreankugel: Remotes have WPSs, they currently use the same font as the main screen (likely to change), and from a user perspective are essentially a second screen for their player that is as affected by themes as the main screen.
16:25:59kugelCan have remotes backdrops?
16:26:10*linuxstb doesn't understand the C240's keymap...
16:26:35*gevaerts tells linuxstb to implement customizable keymaps
16:26:44*gevaerts hides
16:26:48Lloreanlinuxstb: I've got a couple fixes in mind for it, once I'm compiling again actually
16:26:57Lloreanbut you can't do much from how it's laid out.
16:27:28LloreanAgain, I'll put them on a tracker for 6 months, when nobody comments, commit them, then get yelled at, of course.
16:27:31LloreanOnce my laptop is back.
16:27:51kugelDoes anyone know why the statusbar is turned on the in the radio screen?
16:27:52pondlifeRemotes should ideally be equal to main displays in all ways - colour vs greyscale vs mono will be relevant, of course.
16:28:14Lloreankugel: Remotes should be allowed to have backdrops, but I don't think there are any suitable for them yet
16:28:28LloreanSo I don't think anyone's bothered with the functionality to do so
16:29:12Lloreankugel: Because both volume and battery level are still valuable in the Radio screen.
16:29:49kugelAh, yea, volume/battery isn't shown in the screen itselfs.
16:31:36linuxstbThe iaudio (greyscale) remote LCD can have a backdrop
16:32:08Lloreanlinuxstb: Is there a function for menu backdrops, or just WPS backdrop?
16:32:18kugelSo, remotes should be able to use custom list dimensions too I guess
16:32:34pondlifeOf course, one per screen
16:32:55pondlifeOne day we'll find a target with three displays....
16:33:11*linuxstb notices huge delays with the statusbar when changing volume on the c240
16:33:16pixelmalinuxstb: I changed the c200 keymap on my own builds, I'm still not 100% satisfied with it but always wanted to put it in the tracker for discussion...
16:33:31Lloreanpixelma: Do so. :)
16:33:38Lloreanlinuxstb: using the volume buttons while in the list?
16:33:53linuxstbpixelma: I would be tempted to make the volume buttons scroll - i.e. equivalent to fwd/back on a wheel.
16:33:53pondlifepixelma: Yes, it's your turn to take the keymap heat :)
16:34:00LambdaCalculus37pondlife: Update on my disktidy.c modifications... I patched my SVN table and compiled for my iPod color, and casually browsed it in Dolphin. Disktidy found and deleted all .dolphinview files found.
16:34:06gevaertspondlife: I'll start work on my c200-as-USB-remote ASAP
16:34:32pondlifehehe, can we get a USB hub or 2 in there...
16:34:40pixelmathe delayed statusbar update has become more annoying in general
16:34:42Lloreanlinuxstb: I'd like to make long-select be context menu in the WPS, and I really wouldn't mind the volume buttons scrolling (I'd use that personally, but wouldn't have put that in my suggested revision of the keymap)
16:34:58kugelHmm, quickscreen isn't using vp yet
16:34:59linuxstbpixelma: Then the keymap would be identical to ipods, with the exception of the menu/play being swapped, a dedicated "stop/off" button (power) and the record button.
16:35:09 Part LinusN
16:35:19pixelmalinuxstb: that's what I'd like to avoid (scrolling with up/down) because it forces me to two handed operation...
16:35:29Lloreanlinuxstb: You're basically talking about the e200 keymap, but with the "wheel" on the e200 being the volume buttons on the c200.
16:35:33gevaertspondlife: of course. I'll attach 120 c200-remotes to my gigabeat once I'm bored from building Towers
16:35:51pondlifeA fully functional Web Of Rockbox
16:36:14linuxstbpixelma: Yes, I can see the disadvantage in it as well...
16:36:55linuxstbHow do I return from the file browser to the WPS? i.e. the "play" button?
16:36:55pixelmavolume up/down, of course
16:37:19gevaertspondlife: it will provide a tiled 1320x960 screen
16:37:19pixelmaahem... it's some weird combo (rec + up, I think)
16:37:42linuxstbLong-press on rec started a recording ;)
16:37:55Lloreanlinuxstb: Select goes to the browser.
16:37:56linuxstbOr rather, took me to the recording screen.
16:38:05Lloreanlinuxstb: "power" is the main menu
16:38:07linuxstbLlorean: And then how do I get back?
16:38:09 Quit corevette ("Leaving")
16:38:13LambdaCalculus37pixelma: Not rec + up.
16:38:14pixelmaI changed that to "volume up" in my own build (hence sacrificing adjusting volume in lists) but I've been thinking about a short rec
16:38:19Lloreanlinuxstb: Oh, got what you wanted backward.
16:38:28Lloreanlinuxstb: I've been using the "Now Playing" menu option...
16:38:41LambdaCalculus37Me too.
16:38:59linuxstbSo it's missing a "play" button?
16:39:20pixelma("short rec" currently is "enter recording screen" but I could live with that being on "long rec"...)
16:39:31pixelmayes, there is no "resume" button currently
16:39:32*Llorean supposes someone could check /apps/keymaps_c200.c or appropriate almost that named file.
16:39:52Lloreanpixelma: I think recording should always require a long press to get there. Make it very explicit.
16:39:59*linuxstb looks forward to pixelma's cleanup, and won't use his c240 again to avoid confusion ;)
16:40:20pondlifeIs the H300 using short REC, or long REC?
16:40:24pixelmaLlorean: what do you think about short rec being resume then?
16:40:55peturbah, more REC button madness
16:41:14*linuxstb wonders if Sansa designed the c240's buttons just to annoy us
16:41:21*petur votes the usb charging option be moved to the hold switch *again*
16:41:24disorganizeron many sansas i know the rec button is hard to operate.
16:41:44 Quit DerDome (Remote closed the connection)
16:42:01disorganizerbut pressing play should work for the e to go to wps
16:42:34*petur also wonders why recording needs a dedicated button. And why radio not? Why is it hard to pick recording from the menu?
16:42:38pixelmaLlorean: as I said - currently I have resume on "volume up" and stop (in the tree/menus) on "volume down"... though
16:43:11pixelmabut this way I can't set volume in the tree/menu of course
16:43:51pixelmadisorganizer: that's the problem on the c - the "play" button is also used as "up"...
16:44:17 Join DerDome [0] (
16:44:29Lloreanpixelma: I wouldn't object to that. Most of the changes I'd like to make apply to the WPS (similar to the changes I made on e200).
16:44:47LloreanWhile I like trying to match button labels, sometimes usability is also important. :)
16:45:24LloreanI'd rather use the volume buttons than short-rec, honestly, though.
16:46:05pixelmamy changes don't touch the WPS controls (resume, radio, recording and virtual keyboard annoyed me more...)
16:47:07pixelmaI should just post the WIP (just need to make sure it still applies)
16:47:09disorganizer@pixelma: sorry, i found out you are talking c instead of e :-) i am ashame
16:48:53disorganizerbut shouldnt it be like with the e-series? volupdown for scrolling, play for wps, ... ?
16:49:16linuxstbCopying the ipod keymap, we could have the up/down/left/right/select on the C240 acting the same as scroll/left/right/center on the wheel, the menu button as menu, and the REC button as play/pause
16:49:41Lloreandisorganizer: The vol up/down are awkward to use on the c200, they are on "top" and face left/right rather than up/down.
16:49:46linuxstb(long REC = stop, very long REC = off)
16:49:58pixelmathe thing is that you lose overview if you're working at to many things at one :\
16:52:00*pixelma should just start ignoring her typos... ;)
16:52:17linuxstbAnyway, there seems to be a huge lag in the status bar in the file browser on the C240, but not in the menus... (when changing volume)
16:52:42 Quit XavierGr (Nick collision from services.)
16:52:49markunSoon we'll have 2 M3 players, will add even more confusion.
16:52:53 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
16:53:07linuxstbmarkun: I told you, you're not allowed to port to another M3 ;)
16:53:08gevaertspixelma: just blame your keyboard
16:54:07linuxstbAnyone have a gigabeat handy to see if there is a similar bug when changing volume there? (compare menu with file browser)
16:55:27LambdaCalculus37linuxstb: No Gigabeat, but I also have noticed the lag in the status bar in the file browser on my c240 as well.
16:55:41amiconnlinuxstb: The huge status bar lag was introduced by list viewports, I think
16:56:16linuxstbamiconn: Have you noticed it in both the file browser and menus? It seems to be significantly worse in the file browser
16:56:45amiconnI didn't notice it in the menus, but then I've set my targets to start up in the browser
16:57:04linuxstbIt seems OK in the menus.
16:57:10amiconnYou can do an experiment, if you have a target with soft disk led:
16:57:29amiconnDisable dircache, database auto-update etc.. Then reboot.
16:58:03amiconnYou'll notice that the disk activity icon stays for 10 seconds or more after the disk stopped spinning...
16:58:18kugelDoes anyone know what happened about viewportifying quickscreen? JdGordon did some start (which was reverted quickly)
16:58:20linuxstbYes, the delay seems similar to that when changing volume in the lists on the c240
16:59:37linuxstbIn fact it's not that, it seems the status bar isn't being updated as long as there are button events - i.e. the volume will change as soon as I release the volume button
16:59:48Lloreanlinuxstb: On my gigabeat, the status bar volume changes from within lists also lag (though last time i looked it wasn't as much as my c200 is lagging right now)
17:00:10linuxstbs/as soon as/about a second after/
17:01:32pixelmalinuxstb: I noticed yesterday that the statusbar lag from going WPS -> browsers/menus became more obvious on the Ondio because now it still shows the part of the WPS in that area until the statusbar is updated (before it was blanked)
17:01:53***Saving seen data "./dancer.seen"
17:05:37pixelmaand a longer delay from clearing the bootloago to showing the start screeen (file browser here)
17:07:32 Quit disorganizer ("CGI:IRC")
17:07:48 Join ZincAlloy [0] (n=d9eee2f2@gateway/web/cgi-irc/
17:08:35ZincAlloyhi everybody!
17:09:51pixelmahowdy :)
17:10:25ZincAlloyhow's it going?
17:11:32pixelmagood, but I must point at the channel topic...
17:11:57ZincAlloyoh, I see
17:13:15*linuxstb wonders why the Sansa bootloader becomes verbose when USB is inserted
17:14:13linuxstbAnyone object to me fixing that?
17:21:51linuxstbAnyone know why the M:Robe 100 bootloader uses "Power" to go to the OF, rather than Left (like all the other PP bootloaders) ?
17:22:44LloreanNon-iPod PP rather?
17:22:46ZincAlloyspeaking of bootloaders. the h300 bootloader is behaving oddly. after unplugging the power supply I need to push the stop button for it to display the "charging complete" screen and shutting down. play/on will start up the original firmware
17:23:13linuxstbLlorean: Yes - the bootloaders in "main-pp.c"
17:23:50linuxstbAnd on other bootloaders, only one button is used to make the messages verbose, the PP bootloaders were verbose if any button was pressed (including USB insertion)
17:24:02linuxstbe.g. RIGHT on the ipods
17:27:25linuxstbIs it something I've just done, or is bootloader USB detection broken in the Sansa bootloaders in SVN?
17:30:06linuxstbYes, reverting to a plain SVN bootloader, USB detection is broken...
17:33:21pixelmaUSB detection when?
17:34:37pixelmaI'm running quite an old bootloader of the time barrywardell tested disabling the refresh database bit on the e200
17:34:51 Quit Nevtus ("Gone")
17:36:32linuxstbUSB detection in the bootloader - with a bootloader built from current SVN
17:37:20linuxstbSo inserting USB in Rockbox will cause a reboot loop (holding LEFT will stop it and enter the OF)
17:41:57*linuxstb tries to drag on-topic conversations from #rockbox-community
17:42:26scorche|shso then...would we rather have our GSoC application discussions private in a channel such as #rockbox-mentors, or have it all out in here?
17:43:44linuxstbThe SoC page has a comment facility for us to comment directly on applications, but I believe general chat, or chat where we don't get too personal about a specific application should be open.
17:44:27gevaertsI agree, but I guess "too personal" will be different depending on who looks at it.
17:44:51pondlifeIt's useful to get input from non-mentors, generally...
17:44:55GodEaterwe don't need to name names for a student to work out we're talking about their application
17:45:18GodEaterpondlife: I'd include project old-hands too, not just official mentors
17:45:24GodEaterit would make the most sense
17:45:33linuxstbBut is it a bad thing for a student to know what we think of their application, as long as we're not insulting?
17:45:49pondlifeWe shouldn't be insulting, even in private
17:45:58 Join Mathiasdm2 [0] (
17:46:18GodEaterlinuxstb: what would you say about our applications so far then ?
17:46:22pondlifeIf a student is not responding, or progressing, then it's reasonable to discuss that in the open.
17:46:27 Quit Mathiasdm (Read error: 104 (Connection reset by peer))
17:46:48linuxstbGodEater: That we haven't had any...
17:47:02GodEaterlinuxstb: yes we have....
17:47:08gevaertslinuxstb: look again :)
17:47:15linuxstbI have looked....
17:47:15scorche|shwe have...which brought this whole thing on
17:47:59*GodEater thinks linuxstb possibly hasn't clicked on the "Applications" link
17:48:01scorche|shand we started discussing that application in -community
17:48:02linuxstbI meant that the "application" we have received doesn't include enough detail and we should ask the student to provide a fuller application.
17:48:06 Join Mathiasdm [0] (
17:48:11*linuxstb has read it.
17:48:26GodEateryou're being too obscure for us :)
17:48:28*gevaerts accuses linuxstb of being too subtle for us
17:48:29 Quit Mathiasdm2 (Read error: 104 (Connection reset by peer))
17:49:13GodEaterwell I'd like to bring up scorche|sh's point as well - which is "does a usability study, REALLY belong in the GSoC ideas list" ?
17:49:38scorche|shmarkun: since it seems you added the usability study idea onto the list, what were your thoughts on how a student would approach that in a plannable way guaranteeing that a decent bit of actual code work be done?
17:49:40linuxstbIt's quite hard for us to remove it now though...
17:50:25GodEaternot really - they are, after all, only ideas. Students are encouraged to bring their own...
17:50:53linuxstbI think any application for that project would need to convince us it was a significant project, and that Rockbox could be improved by it.
17:51:46 Join RexDart [0] (n=cd9cbcfe@gateway/web/cgi-irc/
17:51:58scorche|shi could see a usability study working out if the person applying has a good bit of history in such things, but it isnt really something a first-year uni student without experience in such things can properly do
17:52:16scorche|shespecially without a proper plan
17:52:39gevaertsI think _any_ project is acceptable as long as (a) the application is detailed enough to convey what will be done and how, (b) it is about the right amount of work, and (c) at least one mentor is interested in following it closely
17:53:11RexDartHmm...does rockbox have known issues with switching between codecs when playing files of different types in a playlist? Mine crashes like a kennedy.
17:53:15GodEaterI still think a summer of code project should produce actual code...
17:53:17gevaertsThat doesn't mean that just about any project gets to the top of course...
17:53:26RexDartOr does it perhaps have allergies
17:53:30linuxstbI think there's also d) Is generally considered a useful thing to do - i.e. that a significant number of Rockbox users will benefit from it.
17:53:38scorche|shand this one certainly isnt "a" and it remains to be seen about "b"
17:53:51scorche|shGodEater: it needs to
17:53:57 Quit ZincAlloy ("CGI:IRC (Ping timeout)")
17:54:12 Join midgey|web [0] (n=8dd3a843@gateway/web/cgi-irc/
17:54:30GodEaterscorche|sh: that's actually mandated somewhere by google is it ?
17:54:46scorche|shwe could verify that the student doing such a project will produce code if he already has ideas on things to fix
17:54:55 Join Nevtus [0] (n=Nevtus@unaffiliated/nevtus)
17:54:56scorche|shGodEater: well, if it the google summer of CODE
17:55:09linuxstbAnd they have to upload their code to google at the end of the summer.
17:55:47gevaertsI can imagine a project that (a) looks at usability issues, and (b) implements a new GUI based on what was found in (a)
17:56:03scorcheGodEater: the most relevant FAQ would be
17:56:19linuxstbgevaerts: But what if a) concludes that the current GUI is perfect? ;)
17:56:23scorche|shgevaerts: and like i said, they alreayd have a few ideas
17:56:31scorche|shfor the reason linuxstb said
17:56:53*pondlife suspects it's not.... ;)
17:57:04gevaertsBut I'm not sure if a usability study without access to lots of different DAPs is doable (the simulators don't provide the feel of the actual devices)
17:58:43scorche|sheither way, the usability study idea is quite touchy and so far, it does not seem like this student would do it justice...what we need to do is figure out *exactly* what we want from such a proposal and communicate that to him and the wiki
17:58:57RexDartHey scorche|sh , was it you yesterday who I was talking to about rockbox crashing with error numbers top-left when I play songs of different formats in a playlist?
17:59:12RexDartAnd you'd suggested I update rockbox from 16667 or so to a current build
17:59:18midgey|webhas any student expressed interest in the plugin localisations or real audio support?
17:59:20markunscorche|sh: I'm not so sure about the usability stuff myself anymore
17:59:24gevaertsI also think that a usability related project is much harder to evaluate for us than some clearly defined feature like a codec, and therefore needs a stronger proposal
17:59:39linuxstbmidgey: Someone mentioned an interest in RA
17:59:50scorche|shmarkun: i think it is a great idea, but one that requires past experience and existing idea of what we are doing wrong
18:00:25 Quit toffe82 (Read error: 104 (Connection reset by peer))
18:01:10midgey|weblinuxstb: what happens if two student pick the same project? is one rejected or can the rejected student be transfered over to another proposal?
18:01:52scorche|shmidgey|web: that is entirely up to us...past orgs have had students working on their own implementation of ideas and they picked the best one although that is not an encouraged way to do things
18:02:17 Join toffe82 [0] (
18:02:20linuxstbmidgey|web: I think that happened last year, and we (the mentors) voted on which applications to recommend to Google. So in effect we picked the best of the two competing applications, and the second student got nothing.
18:02:48Nico_Pif the second student made other applications he still has his chance though
18:02:51linuxstbmidgey|web: Are you also thinking of applying for RA?
18:03:13midgey|webboth projects sounded interesting to me
18:03:33linuxstbPlugin localisation is another project I'm not sure about - is it big enough for a summer of work?
18:03:55Nico_Pis multifont a SoC idea?
18:04:00scorche|shwe really should have had this discussion a week ago =/
18:04:03linuxstbI thought it had been designed, just not implemented
18:04:17midgey|webpossibly, i was planning on adding voice to plugins if i completed the project early
18:04:18linuxstbscorche|sh: Yes... But better now than in 1 weeks time...
18:04:57linuxstbThere are probably other voice-related things that could be included - e.g. voicing the database.
18:04:57 Join ZincAlloy [0] (n=d9eee2f2@gateway/web/cgi-irc/
18:04:58midgey|webmultifont is probably not big enough for a summer either
18:05:14scorche|shmidgey|web: well, get crackin on those applications! ;) doesnt hurt you in any way to apply for more than one project
18:05:29linuxstbSo the project could be "extend the voice and localisation support in Rockbox'
18:05:33 Quit Mouser_X (Read error: 110 (Connection timed out))
18:05:35midgey|webscorche|sh: after this exam tonight I will :)
18:05:55 Quit toffe82 (Read error: 104 (Connection reset by peer))
18:06:20*Nico_P probably missed some multifont discussions
18:06:26ZincAlloylinuxstb, why can't images be displayed underneath viewports anymore? is this a bug or a feature?
18:07:09linuxstbZincAlloy: What do you mean by "anymore" ? Viewports should overwrite what's underneath.
18:07:23ZincAlloyit used to work on the simulator
18:07:26linuxstbYou simply need to put images in their own viewport.
18:07:33Nico_PGodEater: increasing the buffering chunksize seems to help with the responsivity problems
18:07:39linuxstbWhen did it stop working? Viewports have only been in SVN for a few days.
18:08:01ZincAlloyyes, it was an inofficial sim...
18:08:02gevaertsAnyway I think someone should reply to this application with our doubts (both about the application itself and about the experience required for a usability study). If nobody beats me to it I'll try to think of something later tonight
18:08:08ZincAlloyI thought it was rather useful....
18:08:10 Join toffe82 [0] (
18:08:20ZincAlloysee this thread
18:08:25linuxstbZincAlloy: Did you read my description of how viewports should be used on the forum?
18:08:50ZincAlloyoh, there was a description?
18:08:59linuxstbThere have been many descriptions...
18:09:33linuxstbI've tried to describe them on the CustomWPS page. Plus I posted a comment here -
18:09:59pondlifeNico_P: What's the downsize of a larger chunksize? Granularity of buffer waste?
18:10:57ZincAlloyok.. this will make my wps programming rather complicated....
18:11:10linuxstbZincAlloy: Why? The intention is to simplify things.
18:11:22Nico_Ppondlife: honestly I'm not really sure. I left it at 32K when I did the buffering work, and jhMikeS later changed it down to 16K. I guess it mainly involves scheduling matters I'm not familiar with
18:11:26scorche|shso to try and get back to the earlier discussion, what do people think needs to be said specifically to this applicant?
18:12:01gevaertsscorche|sh: (a) proposal is much too sparse, and (b) we think usability needs more experience
18:12:05jhMikeSNico_P: I'm message around right now with it and that includes a 32K buffer chunk
18:12:16ZincAlloyI think I need to use conditionals for the viewports now...
18:12:23ZincAlloythough I'm not sure.
18:12:31Nico_Ppondlife: AFAIK, file I/O doesn't sleep or yield, so a bigger chunksize means less frequent yielding
18:12:36gevaerts(but written in longer sentences)
18:12:45scorche|shand (c) applicant needs to have some existing ideas of things that are "broken" to help guarantee code
18:12:47linuxstbscorche|sh: That an application needs to include a plan for exactly what the applicant is planning to do, and evidence that the student is capable of doing it.
18:13:13 Join roolku [0] (
18:13:23linuxstbZincAlloy: Conditionals in viewports aren't implemented yet - so you will need to avoid the need for them...
18:13:31 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
18:13:36scorche|shalright...i will write a public comment to them when i get off work if no one else gets to it first (which they can, of course...)
18:13:41gevaertsscorche|sh: (b) might include that we will consider it, but applicant needs to convince us that he/she is up to it
18:14:30gevaertsscorche|sh: if I get to it first, I'll post a proposal here first
18:14:40ZincAlloythat's why I need to display images underneath a viewport :D
18:14:57roolkulinuxstb: because power is the only physical button - the interrupt-drivern touchpad driver is not compiled for the bootloader
18:14:58linuxstbZincAlloy: Viewports not clearing their content was a bug I remember fixing.
18:14:59scorche|shi would like to try to get an actual template form for applying up as we discussed in the Dev-ml as well
18:15:22linuxstbroolku: OK ;) I'll add a comment to say that. So there's no way to make the bootloader verbose?
18:15:45Nico_PjhMikeS: it seems to work better for me
18:16:00Nico_PjhMikeS: what other changes are you planning?
18:16:05roolkulinuxstb: I don't think so - maybe if you get the timing just right with the power button
18:16:13ZincAlloylinuxstb: are you sure it was a bug and not a feature? I think it was really useful.
18:16:56linuxstbZincAlloy: Yes, it's a bug. It would cause lots of other issues if a viewport didn't clear its contents.
18:17:13ZincAlloyI see. I didn't notice any
18:17:46gevaertsscorche|sh: I'll try to make a proposal for that as well
18:18:05roolkulinuxstb: hm, now that I think about it, it may always be verbose? I'll have to check when I get home.
18:18:10 Quit gevaerts ("going home before I promise too much")
18:18:20*scorche|sh puts his whip made out of USB cables away for now
18:18:21LambdaCalculus37domonoky: Hey there!
18:18:28linuxstbroolku: Yes, you're right. Checking common.c, it is always verbose...
18:18:34*domonoky smiles.. :-)
18:19:16LambdaCalculus37domonoky: I posted my disktidy modifications to FS #8595. I patched my SVN build and compiled for iPod color, then tested disktidy on it.
18:19:24linuxstbroolku: BTW, do you know if bootloader USB detection works on the M:Robe 100 with an SVN bootloader?
18:19:27ZincAlloyI'll need to wait for viewport conditionals then.. what I want to do is impossible with the current viewport implementation
18:19:35LambdaCalculus37It found and deleted all of the .dolphinview files I had in it.
18:19:53linuxstbZincAlloy: Do you just need to make viewports conditional on album-art?
18:20:05ZincAlloyor can viewports overlap?
18:20:19roolkulinuxstb: I haven't tried recently, but will when I am at home
18:20:34linuxstbZincAlloy: They can, but shouldn't... I won't try to predict what happens if they do.
18:20:49linuxstbroolku: OK, I just noticed that it's broken for the C240.
18:21:00ZincAlloylinuxstb: so there is a chance it might work.. I'm gonna give it a try...
18:21:01 Quit toffe82 (Read error: 104 (Connection reset by peer))
18:21:09linuxstbZincAlloy: Do you just need to make viewports conditional on album-art?
18:21:50ZincAlloylinuxstb: it's not just the album art. the text beneath it needs a conditional as well
18:22:14linuxstbWhat is the text conditional upon?
18:22:26jhMikeSNico_P: I don't know. On H10 sometimes things take their sweet time adding the next handle so the CPU ends up unboosted even though things haven't really finished.
18:22:40ZincAlloylinuxstb: oh, I see what you mean
18:23:03 Join toffe82 [0] (
18:23:05linuxstbZincAlloy: I mean that if viewports could be marked "only display if there is album art" or "don't display if there is album art", would that solve your problem?
18:23:12ZincAlloylinuxstb: yeah, it's all dependent on album art. but I can't just put everything into a viewport. the text would be misaligned
18:23:27Nico_Phmm actually I seem to have run in to the responsivity problem again. audio also paused for a while
18:23:50ZincAlloylinuxstb: yes, I think that would solve it
18:23:52Nico_Perr disregard the last part... it was supposed to pause
18:24:05jhMikeSI also took out the whole bit to defer to the codec with sleep() which works well on H10 and that's a pretty slow disk
18:25:45Nico_PI can test on the video if that's needed
18:26:08 Quit Mathiasdm ("Yuuw!")
18:27:08jhMikeSI put in a delay that has to have a few second with no handle request before unboosting CPU (though a bit messily atm :)
18:28:42roolkulinuxstb: BTW what do you expect to happen with USB detection in bootloader? Considering that it doesn't get into the rockbox bootloader with USB inserted, quick USB insertion during boot is not something that I would normally do
18:29:12roolkulinuxstb: looking at main-pp.c all USB activity seems to be sansa related ?
18:29:43*domonoky sees the first application to rockbox gsoc, and wants to see better ones.. :-)
18:30:19*scorche|sh thinks domonoky should read the things said befor ehe joins ;)
18:30:34*domonoky turns to the logs..
18:30:38scorche|shs/befor ehe joins/before he joined
18:30:40*scorche|sh sighs
18:30:58kugelLambdaCalculus37: What would you think of including windows files (thumbs.db, desktop.ini)?
18:31:17LambdaCalculus37kugel: thumbs.db is already covered by disktidy.
18:31:26LambdaCalculus37I can add in desktop.ini, though.
18:31:45LambdaCalculus37And call me crazy, but I'm also adding in FINDER.DAT files that are created by Mac OS 9.
18:31:47kugelLambdaCalculus37: Ah ok, it's not noted in task though
18:32:44*LambdaCalculus37 has to add in FINDER.DAT since he still uses an iMac with Mac OS 9.2.2 and hooks his JBRv1 to it from time to time
18:32:51linuxstbroolku: Ignore me ;) I didn't spot those Sansa #ifdefs...
18:32:59Nico_PjhMikeS: what sort of difference should it make?
18:33:05kugelOh, ok. Obviously I was wrong about disk tidy at all
18:33:32LambdaCalculus37kugel: If you want to try my patch out, it's FS #8595.
18:33:59 Quit petur ("work->home")
18:36:29Nico_PjhMikeS: buffering seems much slower with the patch
18:41:21jhMikeShow much is "much" time-wise? are you getting delays in adding handles like me?
18:41:55linuxstbBagder: Could you add &pathrev=xxxxx (where xxxxx is the same as r2) to the diff URLs generated by the SVN script for the front page (and elsewhere I guess). This fixes links if files are subsequently renamed or deleted (e.g. Linus's changes to FILES earlier today).
18:42:07n1slinuxstb: you mentioned that a system for localisable plugins was designed. know where i can find that?
18:42:58linuxstbn1s: I'm not sure.. I just have a recollection of reading about it, or maybe it was just vague discussions in IRC.
18:43:00RexDartHmm...I assume rockbox has some error logging feature I could enable to find out why it hates SNES music, right?
18:44:22Nico_PjhMikeS: it seems there are delays in adding handles, but my main observation was that the levels fill much more slowly
18:45:24Nico_PjhMikeS: is the "view OS stacks" screen useful for such debugging?
18:45:36jhMikeSI'm wonder how that is possible since the chunksize is back to 32K
18:45:39 Join miepchen^schlaf [0] (
18:45:55 Join bluebrother [0] (n=dom@rockbox/staff/bluebrother)
18:47:06linuxstbn1s: Hmm, I can't find anything - so maybe it was just IRC discussions...
18:47:13jhMikeSNico_P: perhaps. is this with no other disk access like a database update?
18:47:19Nico_PI think so
18:47:43n1slinuxstb: I will do some irc googling then :)
18:48:10linuxstbn1s: I'm almost certain amiconn had some ideas...
18:48:43Nico_Pwhat is it that's using the disk at startup?
18:49:04linuxstbdircache, database?
18:49:18Nico_Pprobably both in my case, but why do they need to?
18:49:25Nico_Pthey scan for new entries?
18:50:12linuxstbI think dircache does a full scan, the database will scan for new/deleted files, and deal with them appropriately.
18:50:55orsonjany status change on Rio Karma since last forum post at end of January?
18:51:12Nico_Pif they do that both at the same time, no wonder resuming playback on startup is so sluggish
18:51:20Nico_Pwith the ipod's slow disk...
18:51:24linuxstbNico_P: I've no idea how they interact...
18:51:29n1sorsonj: if noone's posted in the forum, then no
18:51:31*Nico_P goes to look
18:51:58jhMikeSthree threads needing disk access while running playback is alot to do
18:52:09 Join Davide-NYC [0] (
18:52:12Nico_Pespecially on a slow disk
18:52:48bluebrotherroolku: around?
18:53:01Davide-NYCmarkun: ping
18:53:06roolkubluebrother: yes
18:53:28bluebrotherI've got a problem installing Rockbox on an m:robe 100
18:53:50Nico_PjhMikeS: when dircache and DB aren't updating there is no UI sluggisheness but initial buffering still seems slow (even with the PCM buffer rather full), but then accelerates and finishes quite fast
18:53:50bluebrotherused rbutil, worked fine so far. Player rebooted, now it's stuck at the bootloader with "no partition found"
18:53:51roolkubluebrother: with rbutil? or manual?
18:54:12 Join gevaerts [0] (
18:54:30markunDavide-NYC: pong
18:54:38jhMikeSthe codec is running full-time with no sleeping until the buffer is actually topped off, then it does sleep()
18:54:39bluebrotherdiscovered a bug in rbutil during installation ;-)
18:54:44LambdaCalculus37Davide-NYC: Ahoy!
18:54:50roolkubluebrother: how old is your bootloader? Do you know if your m:robe is partitioned or formatted as superfloppy (the default)
18:55:03Nico_PjhMikeS: which buffer? the PCM buffer?
18:55:23bluebrotherthe bootloader if the released one from the website −− r12345-080305
18:55:29bluebrotherfunny revision number ;-)
18:55:34Nico_Pwell the difference between these two buffering phases doesn't seem to be the PCM buffer levels
18:55:35jhMikeSit only actually goes off the run list when it has to wait for buffer space
18:56:12amiconnNico_P: The ipod's disk isn't slower than on any other target for reading
18:56:21linuxstbn1s: There's some discussion on 2007-05-24 - search for "sectioned"...
18:56:36amiconnAnd for writing, it's only slower if the writes consist of lots of single-sector transfers
18:56:39roolkubluebrother: *blush* maybe I should have taken out my speedup patch for svnversion
18:56:49n1slinuxstb: thanks, found one from 2005 too
18:56:52jhMikeSI did run H10 with large sector access and it seemed to be about the same speed there
18:56:55Nico_Pamiconn: really?
18:57:00amiconnYes, really
18:57:10bluebrotherroolku: well, how can I reboot the thing and get it connected to the PC again?
18:57:13Nico_Pamiconn: then the main difference is the RAM size?
18:57:17linuxstbn1s: Putting those ideas in the wiki would be useful...
18:57:28amiconnWhenever transferring in 1024-byte or larger chunks, there is no difference at all
18:57:40Davide-NYCOi oi! was AFK
18:57:42roolkubluebrother: there is a reset hole at the bottom, so you need a paper clip or similar
18:57:51bluebrotheraaah :)
18:58:03*gevaerts somewhat regrets volunteering to write the gsoc application template
18:58:19amiconnWhen reading in 512-byte chunks, the ata driver reads a full 1024-byte phyical sector on the first call, and returns half of it. The next call, it just copies the other half from the sector cache
18:58:22jhMikeSmaybe I can emulate on H10 by making it do all this junk at once. The database and tagcache threads can be lowered one step (priority + 1)
18:58:33n1slinuxstb: I agree I thought about it a bit over the easter weekend and will try to give it a go after my next exam probably (well unless it is taken by a gsoc student :P )
18:58:57Nico_PjhMikeS: can't you just resume audio on startup?
18:59:12amiconnJust when writing 512-byte chunks, it gets messy: The driver reads a 1024-byte sector into the cache, replaces the respective half with the 512 bytes to be written, and writes that 1024-byte sector again
18:59:33jhMikeSI don't have database or dircache on though
18:59:34amiconnThe immediate write is necessary in order not to loose data
18:59:47bluebrotherit seems my player likes to hang. Just resetted it and flipped the hold switch. Now it stays at "shutting down"
18:59:53Nico_PjhMikeS: I suggest you enable them to add some strain :)
19:00:50jhMikeSwhat do you add to the database.ignore file to get it to ignore a directory? The HVSC will kill it.
19:00:59amiconnjhMikeS: It should also be possible to force the ata driver to use its sector cache mechanism
19:01:00bluebrotherwhat's the button for booting the of on the m:r?
19:01:05linuxstbNothing - just create the file.
19:01:06roolkubluebrother: "shutting down" the rockbox message?
19:01:16linuxstb^To jhMikeS
19:01:31bluebrotherroolku: yes
19:01:32amiconnDisk addressing doesn't change at all from this
19:01:39jhMikeSah. cool
19:01:43linuxstbjhMikeS: But since markun's commit to add a "database.unignore", it will still scan all the files...
19:01:53roolkubluebrother: if you reset while hold is on it will turn off instead of restarting
19:01:57***Saving seen data "./dancer.seen"
19:02:11*amiconn should probably add a #define to force the ata driver into 1024-byte mode for testing
19:02:26roolkubluebrother: hm, so you booted into rockbox and the "couldn't find partition error" occured there?
19:02:27Nico_PjhMikeS: I just went back to SVN for comparison and it's much worse than with your patch
19:02:34jhMikeSlinuxstb: ??
19:02:46bluebrotherno, the error occured in the bootloader.
19:02:57gevaertsDoes anyone happen to have a reasonable gsoc application available ? Maybe one from 2007. I need soe inspiration for this template
19:03:05amiconnlinuxstb: Afaik the unignore was reverted...
19:03:15bluebrotherand I haven't found the button to boot into the OF again ...
19:03:37Slasheriamiconn: it's still there
19:03:40jhMikeSNico_P: try adding one or two to the database and dircache thread priorities. they could be a bit closer than they ought to playback.
19:03:56Nico_PjhMikeS: ok
19:04:24Nico_PSlasheri: can you brief me on what dircache and the DB do on a normal strartup?
19:04:29roolkubluebrother: you press power again, when the headphone logo shows and hold until it says booting OF
19:04:33 Quit XavierGr (Nick collision from services.)
19:04:33jhMikeSactually the differential is rather slight now
19:04:43linuxstbjhMikeS: I put that wrongly - it won't "scan all files", but will still traverse the entire tree. Before that patch, it would stop as soon as it saw database.ignore
19:04:46 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
19:04:58bluebrotherroolku: doesn't work :(
19:05:32 Join Arathis [0] (
19:05:34roolkubluebrother: what happens?
19:05:36bluebrotherI only get the "no partition found" error
19:05:59SlasheriNico_P: on normal startup, dircache will scan through the entire file tree (using direct fat_* methods for fast access) and just create a linked list containing all files and directories (and attributes of those)
19:06:10RexDartHmmm....which source file contains crash=1?
19:07:07roolkubluebrother: hm, I suppose that is fair enough that it can't load the OF if it doesn't see the partition...
19:07:52bluebrotherroolku: so ... what can I do now?
19:08:01SlasheriNico_P: and tagcache checks at first if there is a temporary db on disk waiting to commit, and will commit that if found. After that it waits dircache to became ready (if enabled) and starts searching for new tracks. Finally there will be a search for removed tracks (that phase will be greatly optimized soon)
19:08:10roolkubluebrother: connect USB and reset
19:08:24Nico_PSlasheri: right. and what's happening when there is "scanning disk" splash? is that the DB?
19:08:30roolkubluebrother: or switch hold on and then reset
19:08:38SlasheriNico_P: that is dircache
19:08:51Nico_Pwhy doesn't it appear on all startups?
19:08:57bluebrotherah, so there is some kind of emergency disc mode? Nice.
19:09:12SlasheriNico_P: normally that is done in background
19:09:32Slasherithat splash appears only when a foreground scan occurs
19:09:38roolkubluebrother: this mechanism is in the OF's bootloader
19:10:11roolkubluebrother: I assume you can mount the drive on your pc ?
19:10:13bluebrotherhmm, doesn't look like my player is a superfloppy. Just dumped the mbr
19:10:21Nico_PSlasheri: what's the difference between fg and bg scans?
19:10:40SlasheriNico_P: foreground scan is initialized when the last cache size is unknown or directory cached failed for some reason
19:10:44bluebrotheryep, I can now access the player.
19:10:48*n1s creates a wikipage
19:10:54Slasheribg scan already knows the necessary amount of memory to allocate
19:11:03*jhMikeS knows there probably a retuning stage needed here
19:11:13Slasherifg doesn't and that finds it out
19:11:36Nico_PSlasheri: ah and I guess the known cache size is erased by an USB operation?
19:11:54Slasherino, it shouldn't be
19:12:22Slasheriit is probably a "new" bug if that happens
19:12:43roolkubluebrother: can you try the bootloader i am DCC sending right now?
19:12:45Slasherii almost never use the rockbox usb mode
19:12:56Nico_PI'm not quite sure what I did, there's probably no cause for concern. what kind of thing makes a fg scan necessary?
19:12:57bluebrotherroolku: just installed an svn bootloader. Now it boots fine
19:13:08*bluebrother has no idea how dcc works in irssi
19:14:31SlasheriNico_P: for example initializing dircache for the first time or cache reserve allocation is exeeded when performing "live" cache operations (such as adding new files to the cache after a usb connection)
19:14:36roolkubluebrother: Hm, I better verify the bootloader in the server then...
19:15:00roolkubluebrother: maybe it is an old one
19:15:09SlasheriNico_P: dircache allocated 64 KiB of extra every time to the last known real cache size during startup
19:15:29 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
19:15:37bluebrothera FAQ page telling how to enter disc mode on the m:r100 would also be a good idea now it's supported
19:15:55Nico_PSlasheri: thanks, things are much clearer now :)
19:16:02bluebrotheronce I'm finished with playing around with the thing a bit I think I could start one
19:16:30 Quit toffe82 (Read error: 104 (Connection reset by peer))
19:16:35roolkubluebrother: you mean rockbox USB? I thought that still needed a define to be enabled?
19:17:05 Join toffe82 [0] (
19:17:08bluebrotherroolku: no, the usb-and-reset trick
19:17:31 Join bertrik [0] (
19:17:40 Quit toffe82 (Read error: 104 (Connection reset by peer))
19:18:28 Join toffe82 [0] (
19:18:39roolkubluebrother: I do that with all my targets. :)
19:19:51bluebrotherhmm, wasn't rockbox usb safe on non-flash targets?
19:20:30amiconnNot entirely
19:20:46Nico_PjhMikeS: both the tagcache and dircache threads have PRIORITY_BACKGROUND whereas audio has PRIORITY_USER_INTERFACE
19:20:50roolkubluebrother: not sure what you mean, but it is faster to use the bootloader USB instead of waiting for rockbox to boot
19:20:55amiconnThere is the (slight) chance of data corruption due to the unstable connection itself
19:21:28roolkubluebrother: (bootloader USB is OF or rockbox depending on target)
19:22:11*amiconn wants the ipod bootloader to detect usb too, and start diskmode directly :/
19:22:17jhMikeSNico_P: yes, since they need to defer to audio. So I'd try PRIORITY_BACKGROUND+1 or +2 as a check since the specific mean of that has changed.
19:22:34linuxstbamiconn: I did try that, but it gives a nasty reboot...
19:22:39amiconnRight now, the bootloader always boots rockbox first, which will then detect usb and *try to* reboot into diskmode
19:22:48Nico_PjhMikeS: ok, I'll try that
19:23:17*linuxstb must be extremely lucky with ipod usb mode - it very rarely fails
19:23:22amiconn...which does only work in around 30% of all builds...
19:23:59amiconnThen I have to use Menu+Play to enter diskmode, which is also nasty because it doesn't auto-reboot on unplug
19:24:33linuxstbamiconn: Have you tried rolo'ing the diskmode app?
19:24:41amiconnYes. Doesn't work
19:24:56amiconnRoloing diagmode works, but diskmode just freezes
19:25:28bluebrotherroolku: well, the insert-usb-and-reset thingy is something I consider comparable to emergency disc mode on Ipods.
19:25:51 Join PaulJam [0] (
19:25:52 Quit ZincAlloy ("CGI:IRC (EOF)")
19:26:58amiconnlinuxstb: What nasty reboot did you observe in your experiments? Did the bootloader set the "diskmode, hotstuff" cookies?
19:28:27linuxstbamiconn: Yes, it set those cookies and reset. I can't remember exactly what it was that I didn't like though - maybe the disk needing to be spun down, then back up immediately...
19:28:37Nico_PjhMikeS: the priority change doesn't seem to cause any noticeable difference
19:29:28amiconnlinuxstb: Yeah, the bootloader would need to do a proper disk shutdown, like the core usb handling does
19:29:32linuxstbamiconn: I've just tested rolo'ing diskmode on my Color, and it works.
19:30:01linuxstbamiconn: But I seem to remember it not always being reliable...
19:30:09amiconnYes, I remember that you reported that both on Color and Video, roloing diskmode works, and roloing diagmode doesn't
19:30:17amiconnOn Mini G2 it's the opposite
19:30:20linuxstbAh yes...
19:31:06 Quit midgey|web ("CGI:IRC (EOF)")
19:31:25bluebrotherroolku: seems I can't leave the battery debug screen
19:31:46linuxstbamiconn: But in theory we _should_ be able to just copy the diskmode app from flash and branch to it. I posted a patch to flyspray that did that, but it was only as reliable as rolo'ing was...
19:32:17linuxstbI'm not sure if that's worth spending any more time on though...
19:32:27jhMikeSwhat the hell runs on 5g that suck so much CPU? I can do dircache+database refresh on H10 without any trouble.
19:32:43linuxstbMaybe restoring the sysinfo struct to IRAM will help diskmode...
19:33:02Nico_PjhMikeS: I have the 80 GB disk almost full, that has to make a difference
19:33:31 Join dantje_ [0] (
19:34:13roolkubluebrother: until we get rockbox USB this is the only diskmode there is on the m:robe
19:34:33bertrikbluebrother: try this patch it makes the battery debug menu behave more like the other debug menus w.r.t. key handling
19:34:43roolkubluebrother: can't remember offhand what button it is - power doesn't work?
19:34:46jhMikeSit doesn't seem tagcache is even accessing the disk at all at startup. I only the the priority of dircache being bumped to 16 occasionally, not both.
19:34:53bluebrotherhaven't tried that one :o
19:35:11RexDartjhmikes: hell if I know but my battery lasts all of about 45min
19:35:14Nico_PjhMikeS: where do you see that? stacks debug screen?
19:35:34jhMikeSNico_P: yes, while buffering and having the background update
19:36:01Nico_Pmaybe you could quickly brief me on the meaning of all the data?
19:36:29*roolku goes home now
19:36:37bluebrotherhmm, resuming playback uses the display button instead of play? That's awkward.
19:37:42jhMikeS(0) = core, RTBS = state, first number = base priority, second number = running priority (may be a higher inherited from another thread if a higher priority one is waiting for a lock) xx% = max percent of stack actually used
19:37:44SlasherijhMikeS: maybe it's still the codec and buffering fighting?
19:38:13n1samiconn, linuxstb, Bagder: the plan for making plugins localizable as far as i have understood, comments, suggestions, corrections?
19:38:23Nico_PjhMikeS: thanks. and what do R,T,B and S mean?
19:38:25jhMikeSthe codec yields much less now so the buffering can more easily be dominated
19:38:43jhMikeSR = running, T = blocked with timeout, B = blocked, S = sleeping
19:39:38amiconnHmm, that reminds me
19:39:44jhMikeSoh, and + = thread has called trigger_cpu_boost
19:39:50amiconnDoom still crashes if voice is enabled
19:40:00Nico_PjhMikeS: and what does the star mean?
19:40:21jhMikeSit's just always combined with "R" (I guess for emphasis)
19:40:24amiconnMaybe there's a buffer clash? Voice file buffer not properly reserved so that it's outside the returned audio buffer?
19:40:43SlasheriNico_P: it makes easier to spot the running status
19:41:40 Quit Davide-NYC ("ChatZilla 0.9.81 [Firefox]")
19:41:56 Join linuxstb_ [0] (
19:42:01 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
19:42:05pixelmathat reminds me that I once saw that voice file present or not makes no difference in the reported buffer size in the Rockbox info screen
19:42:09Nico_Pwhat is the "main" thread?
19:42:13*linuxstb_ spots yellow and deals with it...
19:42:16SlasheriNico_P: the ui thread
19:42:41pixelmalinuxstb: you want to sell it? ;)
19:42:50n1spixelma: afaik the voice file is loaded if present regardless of the setting
19:43:07linuxstb_pixelma: ;)
19:43:08 Quit japc (Read error: 110 (Connection timed out))
19:43:12Nico_Phmm then why is it always, running even others are too?
19:43:34SlasheriNico_P: you can also test and kill those threads by using the select :)
19:43:41pixelman1s: I said "voice file present or not" - not whether I have the option set or not...
19:43:53SlasheriNico_P: that thread displays the debug menu
19:43:55n1sah, sorry misread...
19:44:03Slasheriso when the menu updates, that thread is always running
19:44:07Nico_PSlasheri: I can't select anything , and I guess I'll abstain :)
19:44:11amiconnlinuxstb: The success/failure check for bmp loading in wps_parser.c is wrong, as I explained yesterday
19:44:17Nico_Pmakes sense
19:44:43linuxstb_amiconn: Sorry, I misunderstood you then... What's the issue?
19:44:44amiconnUmm, ignore that
19:45:04*amiconn didn't recognise that there is a wrapper
19:45:15linuxstb_Yes, that confused me as well...
19:47:10 Join Mouser_X [0] (
19:47:10Nico_PjhMikeS: isn't there anything running on the second core in audio playback?
19:47:38jhMikeSkilling threads arbitrarily can leave mutexes and other things abandoned though but I've resisted the temptation to add extensive cleanup procedure for that
19:47:43*linuxstb_ gets tangled up in some #ifdef spaghetti
19:47:46jhMikeSNico_P: only with SPC
19:48:04*gevaerts sends linuxstb_ some #endifs to help him get out
19:48:08jhMikeSA thread on COP will have a one (1)
19:48:40*linuxstb_ points gevaerts to line 1433 of apps/gui/wps_parser.c
19:48:50jhMikeSNico_P: was that to me?
19:49:10Nico_PjhMikeS: yes
19:49:18 Quit XavierGr (Nick collision from services.)
19:49:23SlasheriNico_P: it would be great to run the codec only on the second core, but that makes big shared memory caching issues and things like that
19:49:29 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
19:49:54SlasheriNico_P: i had that semi working a long time ago and there was plenty of problems
19:50:07*gevaerts hands linuxstb_ some dynamite. That will be more useful to deal with this code
19:50:32Nico_Pyeah I thought it had to be something like that... would it be improved if the audio and codec threads communicated through message queues?
19:51:05Nico_Plinuxstb_: oh yes sorry about that one... my fault :/
19:51:07jhMikeSNico_P: hasn't been done yet and there's a bunch of gotchas that have to be thought though. I think the most benefit would come from decoding being on COP and DSP on CPU
19:51:16linuxstb_Nico_P: ;) I like a challenge...
19:51:28linuxstb_That stray ; is nice...
19:51:38Nico_Phehe thanks ;)
19:51:44jhMikeSof course if the codec is lightweight, like WAV having it all on COP would be better.
19:51:56*linuxstb_ hands Nico_P the obfuscation of the year award
19:52:20gevaertsCan people interested in gsoc have a quick look at ? It's a short summary of an application template, and I'd like to know if there's anything missing or superfluous
19:53:20Nico_Plinuxstb_: thanks so much. I'm very moved
19:54:16Nico_PjhMikeS: I'd love to try to help out
19:54:45 Quit RexDart ("CGI:IRC (EOF)")
19:55:14Nico_Plinuxstb_: the logic is actually rather straightforward
19:55:23jhMikeSNico_P: there's some advantage in having certain variables simply polled though you have to beware of race conditions when doing that.
19:55:38linuxstb_Nico_P: Maybe I'm just tired...
19:55:40 Join stripwax [0] (
19:55:48 Quit stripwax (Client Quit)
19:55:58 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
19:55:59 Join DerPapst [0] (
19:56:19*DerPapst wonders if LambdaCalculus37 has seen FS #8637
19:57:00Nico_Plinuxstb_: if there is a remote LCD but it's only 1bit, the if's body needs to be empty
19:57:03jhMikeSthe worst thing with PP actually is the cache handling (or lack of it) accross cores.
19:57:28 Join rasher [50] (n=rasher@rockbox/developer/rasher)
19:57:45Nico_PjhMikeS: and that's a problem with shared variables?
19:58:05 Join piga [0] (n=leonardo@
19:58:12jhMikeSif they're in cached ram, quite.
19:58:21jhMikeSeach core keeps it own cache
19:58:39Nico_PI can imagine the problems
19:59:35 Quit DerPapst (Nick collision from services.)
19:59:37 Join DerPapst [0] (
19:59:51*DerPapst grmbls at his ISP
20:00:38linuxstb_Nico_P: I think this does the same thing -
20:01:13jhMikeSNico_P: mpegplayer has the entire file buffer accessed as not cached to get around this but fortunately RAM is pretty fast on ARM.
20:01:28 Join jazznazzz [0] (n=4d46139c@gateway/web/cgi-irc/
20:02:12 Quit jazznazzz (Client Quit)
20:02:30markunI can't find what I need to change for sysfont.h to be built..
20:02:48markunfor the meizu
20:03:42linuxstb_Are you compiling anything that has sysfont.h as a dependency?
20:04:08 Join jazznazzz [0] (n=4d46139c@gateway/web/cgi-irc/
20:04:21markunI'm trying to build a bootloader (with a lot of stubs)
20:04:39markunfont.h includes sysfont.h
20:05:01Nico_Plinuxstb_: looks fine to me... and much clearer too
20:05:30jhMikeSNico_P: all decoding there is done from local cached copies of the elementary streams.
20:05:44 Join fml [0] (n=4fd3ef13@gateway/web/cgi-irc/
20:06:16Nico_PjhMikeS: how is the decoding spread accross cores? audio on one and video on the other?
20:06:37markunlinuxstb_: when I build a bootloader for the nano for example, sysfont.h is the first thing which gets created, but I can't figure out why.
20:06:55LambdaCalculus37DerPapst: I have seen FS #8637.
20:07:20gevaertsmarkun: maybe it's needed for printf to work ?
20:07:22fmlHello. I've made a patch for the WPS tags in the manual ( but couldn't actually build it since I don't have TeX. Could somebody please try it?
20:07:51markungevaerts: yes, but I can't find out what I need to change to compile it
20:08:19jhMikeSNico_P: on mpegplayer yes. for SPC, only the emu runs on COP and passes the samples to the CPU codec thread for DSP and pcm playback.
20:08:58markunmake V=1 doesn't tell me much either why it's built for the nano
20:09:07linuxstb_markun: Have you looked at firmware/Makefile ?
20:10:05linuxstb_The dependencies seem clear - sysfont.h should always be created when firmware (librockbox.a) is built.... So I don't know why it wouldn't work.
20:10:52markunthe files from firmware/ are being built
20:11:14markunmust be something really stupid I forgot
20:11:40linuxstb_Have you tried the normal "delete build dir and reconfigure" ?
20:12:30jazznazzzHi everyone, i cant delete the songs from my sansa c200 - "delete ....mp3 - YES", then "deleting..." and nothing happens, the song continues playing and is still there
20:12:39markunthis is frustrating..
20:13:24bluebrotherfml: will try ...
20:13:24 Quit toffe82 (Read error: 104 (Connection reset by peer))
20:13:28linuxstb_jazznazzz: When you say "still there", where is "there" ? If you delete files, they won't (I think) immediately disappear from the database, and if they are playing, it will be buffered in RAM, so will continue playing.
20:13:34gevaertsI have put a more detailed template at Please comment on it. I plan to put it on the wiki soon (with the mention that it's still a draft ?). I'd like to have this page more or less ready before I add a comment to the application that's currently there
20:13:34 Quit jazznazzz ("CGI:IRC (EOF)")
20:13:52 Join jazznazzz [0] (n=4d46139c@gateway/web/cgi-irc/
20:14:06 Join toffe82 [0] (
20:14:21 Quit DerPapst (Nick collision from services.)
20:15:44 Join DerPapst [0] (
20:15:48n1smarkun: are all the various fields filled in in configure for your target?
20:16:12n1sbottool, bootoutput etc?
20:16:12markunn1s: I think so. Any field in particular you are thinking of?
20:17:03*jhMikeS notices a marked increase in handle request speed with dircache on with H10 (no delay between them) and so buffering is faster with dircache and database updating the background than with none of those enabled. :\
20:17:20fmlbluebrother: ok, I'll wait for the feedback. Would you write in in the FS entry?
20:17:34markunn1s: boottool wasn't
20:18:03Nico_PjhMikeS: a handle request requires asking for the filesize, which is much faster with dircache
20:18:11DerPapstLambdaCalculus37: i do like that approach more. It avoids future patches to support more different filetypes. Also some may want to keep files that disktidy deletes otherwise.
20:18:26n1smarkun: I'm not sure if it makes a difference but most of the tests in configure seem to make a lot of assumptiojns..
20:18:34jhMikeSwhat else? does it involve asking tagcache for anything?
20:18:54markunn1s: doesn't seem to help
20:19:03jhMikeSor playlist cachectrl ?
20:19:05*gevaerts notices that apparently nobody here is interested in gsoc
20:19:29n1smarkun: maybe you can pastebin your generat4ed Makefile in the bootloader build dir?
20:19:35*DerPapst is ;-)
20:19:54linuxstb_gevaerts: I've read your pastebin...
20:20:22n1smarkun or maybe you have hit FS #7583
20:20:28gevaertslinuxstb_: any comments ?
20:21:30Nico_PjhMikeS: it doesn't use tagcache directly. things like open use it when it is enabled
20:22:30Rincewindgevaerts: I'm interested in gsoc
20:22:47linuxstb_gevaerts: No, I think it's good.
20:23:07markunn1s: that fixes it!
20:23:08gevaertslinuxstb_: OK. I'll put it on the wiki, so DerPapst and Rincewind can get to work ;)
20:23:13markunls: target/arm/crt0.S: No such file or directory
20:23:15linuxstb_Rincewind: As a student, do you have any feedback on this?
20:23:22n1smarkun grat :)
20:23:27RincewindI just started reading it
20:24:01markunn1s: I have to go now. Maybe I can fix this bug so others will not have the same problem
20:24:09markunany idea why it got reverted?
20:24:19DerPapstLambdaCalculus37: though adding linux as a third option is nice :-)
20:24:54markunn1s: Ah, I see it in the commit log..
20:25:27n1sI guess we can poke Bagder then :)
20:25:44linuxstb_He's had a lot of thinking time... ;)
20:26:38RincewindFeedback: I don't know if I can submit a reasonable timeline in advance. Project milestones is ok, but putting time estimates to the milestones is guesswork at best
20:27:28 Quit fml ("CGI:IRC (EOF)")
20:27:31gevaertsRincewind: I'll add something to the effect that this subdivision is approximate
20:27:41Rincewindapart from that, it covers everything I would have told in the application anyway.
20:28:25jhMikeSNico_P: did the skipping stop with the patch? I don't remember you saying it still skipped.
20:28:58Nico_PjhMikeS: I never experienced any skipping, with or without the patch
20:29:10*gevaerts now fights the wiki to format this thing correctly
20:29:55jazznazzzHi everyone, i cant delete the songs from my sansa c200 - "delete ....mp3 - YES", then "deleting..." and nothing happens, the song continues playing and is still there
20:30:05Rincewindgevaerts: this is what I have written yet: (there is much missing obviosly)
20:30:34linuxstb_jazznazzz: Please read the IRC logs - I answered you earlier...
20:30:51DerPapstlinuxstb_: i think it's quiet good what you posted on pastebin. Though i haven't followed the entire gsoc discussion.
20:31:11linuxstb_DerPapst: It's gevaerts' post.
20:31:29 Quit toffe82 (Read error: 104 (Connection reset by peer))
20:31:31DerPapstAh. then it's all bad of course ;-)
20:31:32jhMikeS"12:26:02 Nico_P GodEater: I've experienced FS #8807 just now "
20:33:00RincewindDerPapst: what project do you want to apply for?
20:33:30 Join toffe82 [0] (
20:33:50DerPapstummm... acually none because i'm not good enough yet.
20:34:04Nico_PjhMikeS: oh actually I had only seen the mention of the unresponsiveness in that report :/
20:34:29Nico_PI don't know why but I never get skips on the ipod I have
20:34:30*gevaerts hits DerPapst on the head
20:36:01*DerPapst goes away with a big headace....
20:36:39 Join brunocardoso [0] (n=bruno@
20:36:46brunocardosohi all
20:37:12brunocardoso i was looking at rockbox projects for Google Summer of Code
20:37:35 Join argumentD [0] (
20:37:37brunocardosoare you really interested in compiling rockbox with LLVM ?
20:38:06DerPapst^ see? i'm not even needed as a gsoc appliant :-P
20:38:06 Quit bertrik (Remote closed the connection)
20:38:20*DerPapst hides
20:38:37 Join portavales [0] (n=portaval@
20:38:43brunocardosoi wrote the Mips backend for LLVM, and maybe i'm eligible for this project
20:38:55brunocardosoi can tweak the ARM backend if needed
20:39:06 Quit toffe82 (Read error: 104 (Connection reset by peer))
20:39:07brunocardosofor llvm to compile rockbox stuff
20:39:31 Join bertrik [0] (
20:39:58 Join toffe82 [0] (
20:40:10linuxstb_brunocardoso: What do you think? Could it be a useful project for Rockbox?
20:41:11jhMikeSthe only reason I could see for unresponsiveness is for the codec to be chronically boosted because PCM is draining and the buffering isn't keeping pace
20:41:26thegeekI think it would be awsome, llvm seems like a really interesting project
20:41:49 Join Arathis2 [0] (
20:41:57jhMikeSNico_P: what happened with the response with the patch though?
20:42:18amiconnWhat would be the advantage of using llvm?
20:43:22 Join EspeonEefi [0] (i=espeonee@CERULEANCITY.MIT.EDU)
20:43:27Nico_PjhMikeS: responsiveness improved but it's not perfect
20:43:53brunocardosolinuxstb_: i think it would be very nice, apple uses llvm to compile stuff for iphone, the ARM backend is very stable, so i guess optimizations are always arriving for this target!
20:44:41jhMikeSNico_P: did you trying skipping from the buffering screen to see if PCM hangs around at the low end?
20:44:53Nico_PI'll try
20:45:21jhMikeSabsolutely nothing else is capable of overriding UI like that
20:45:33*gevaerts finished the wiki work. Feel free to improve on it
20:47:36gevaertsbrunocardoso: how good is the arm optimisation of llvm as compared to gcc ? Also, (much less important, but good to know) are there plans to support coldfire and sh in the future ?
20:48:30Nico_PjhMikeS: PCM seems to remain at quite a healthy level
20:49:42bluebrotherwhat is the use of llvm for Rockbox? I just read the wikipedia entry but I still don't get it
20:50:30 Quit Arathis2 ("Bye, bye")
20:50:35n1sbluebrother: it can work as a static compiler and is supposedly better than gcc
20:51:03jhMikeSNico_P: unfortunately you can't go to watch PCM while the UI is being overridden. :\
20:51:18bluebrotherso it's a complete compiler replacement? I'm somewhat puzzled by this intermediate language thing
20:51:45gevaertsbluebrother: from what I understand, yes.
20:51:53n1sbluebrother: has a quite nice summary on the frontpage
20:52:07Nico_PjhMikeS: probably not :)
20:52:07 Quit jazznazzz ("CGI:IRC (EOF)")
20:52:11*bluebrother should have tried that page in the first place
20:52:29*jhMikeS is fishing for clues that are hard to come by without actually having a device handy
20:52:40 Join PaulJam_ [0] (
20:53:43 Join mib_x5t9yvr187 [0] (i=d2d43dfb@gateway/web/ajax/
20:53:58 Quit toffe82 (Read error: 104 (Connection reset by peer))
20:54:13gevaertsCan students actually revise their application ?
20:54:49 Part mib_x5t9yvr187
20:55:02 Join toffe82 [0] (
20:55:11bluebrotherhmm. So the C code gets transferred to this LLVM language and afterwards compiled into the actual binary?
20:55:43gevaertsbluebrother: actually something similar happens with gcc (and I guess these days with most compilers)
20:55:47 Quit LambdaCalculus37 ("CGI:IRC 0.5.9 (2006/06/06)")
20:55:53 Join RubberDucky132 [0] (n=chatzill@
20:55:58Nico_Pyeah it's just an abstract tree, isn't it?
20:56:02 Quit toffe82 (Read error: 104 (Connection reset by peer))
20:56:08brunocardosogevaerts: sometimes llvm is better sometimes worse than gcc! never heard coldfire and sh plans, but i can check with the mans behind ARM backend
20:56:15RubberDucky132Quick question: can Rockbox output on iPod 5G 30GB through USB connection?
20:56:35 Join toffe82 [0] (
20:56:46bluebrothergevaerts: I thought I remembered something like that (in regards of gcc). So the main difference for llvm is its optimizations?
20:56:56gevaertsRubberDucky132: I don't really understand that question
20:57:02bluebrotherRubberDucky132: what do you mean by "output through usb"?
20:57:07gevaertsbluebrother, meet brunocardoso. brunocardoso, meet bluebrother
20:57:12RubberDucky132sorry, output *audio*
20:57:23thegeekllvm is a kind of unified framework for compiling I think
20:57:23RubberDucky132through the USB connector at the bottom
20:57:29RubberDucky132for iPod Hi-Fi docks
20:57:40n1sbrunocardoso: do you think the gain for rockbox compiler with llvm will be substantial over compiling with gcc?
20:57:58bluebrothergevaerts: hehe, you're right ;-)
20:58:09brunocardosohi bluebrother
20:58:26thegeekI don't think it the gain would be _substantial_ right now, but seems to be llvm growing fast, and it's getting a lot of attention
20:58:36gevaertsRubberDucky132: ok, I see. No it can't. From what I understand this is actually a hardware limitation (i.e. the ipod does not have USB host support)
20:58:49thegeekwow that sentence was horrible;P
20:59:17 Quit Arathis (Read error: 110 (Connection timed out))
20:59:26RubberDucky132gevaerts: perhaps my understanding of hardware limitation is false, but surely it isn't since the iPod's firmware can do it?
20:59:27bluebrotherthe hifi could use the Ipod as storage device (i.e. decode the music itself). In that case it would use disc mode and therefore work with Rockbox too.
20:59:48scorche|shRubberDucky132: are you sure it takes USB?...i think the hi-fi just uses the line out
20:59:54bluebrotherRubberDucky132: you haven't explained how exactly it works ;-)
20:59:56Nico_Pgevaerts: did you get anywhere with USB audio? (hoping that's not the same question...)
21:00:01n1sRubberDucky132: that connector has far more pins than usb and the output to those accessories is throu some of the non-usb pins
21:00:25RubberDucky132Alright maybe its not traditional USB
21:00:33scorche|shnamely line out which is used by most of those sorts...
21:01:02scorche|shit *is* traditional USB...the dock connector ont he ipod is just much more than USB....hence why it is a "dock connector" and not "usb connector"
21:01:13 Join gregzx [0] (
21:01:23RubberDucky132Well you seem to have decoded my babble
21:01:29gevaertsNico_P: I have "working" usb audio here. The quotes mean that it doesn't actually do sound
21:01:44RubberDucky132so can Rockbox output through the iPods "dock connector"?
21:01:51Nico_Pgevaerts: would it take a lot more to do sound?
21:01:54scorche|shthrough the line out, yes
21:02:01***Saving seen data "./dancer.seen"
21:02:20 Join toffe82_ [0] (
21:02:26RubberDucky132that being a selection of pins on the dock connector? or the headphone socket
21:02:32bluebrotherbrunocardoso: so llvm is basically a new architecture for a compiler?
21:02:33scorche|shmost people with devices like the hi-fi get sound and cna control the volume, but not be able to skip songs and such since that is using the serial connection
21:02:35 Join DerPapst748 [0] (
21:02:39 Quit DerPapst (Nick collision from services.)
21:02:40scorche|shthat is on the dock connector
21:02:41 Nick DerPapst748 is now known as DerPapst (
21:02:44gevaertsNico_P: it would need quite a lot of work yes.
21:02:48RubberDucky132wonderful :)
21:03:01scorche|shRubberDucky132: try the AppleAccessories wiki page and see if your device is on there
21:03:06RubberDucky132because my dad invested in a hifi grade iPod doc so I've been using Rockbox to use FLAC on the iPod
21:03:10brunocardoson1s: i really dont know, we must compile and run to see! that's why i'm wondering if this is a eligible as a GSoC projet! ;)
21:03:31RubberDucky132scorche|sh: ok :)
21:03:34thegeekI think llvm is mentioned in the gsoc wiki-page?
21:03:42 Quit PaulJam_ (".")
21:03:49scorche|shhence why he is asking about it
21:03:49gevaertsNico_P: but I'll only do that work once I'm confident that we can solve the current usb issues
21:04:01brunocardosobluebrother: think about llvm as a alternative for gcc! although its the name of the compile IR too!
21:04:01thegeekthen clearly it is of some interest;P
21:04:17n1sbrunocardoso: understood, do you think it would be enough work for gsoc to make it compile rockbox ok (with all the associated build system adjustments etc.)?
21:05:08bluebrotherbrunocardoso: ok, that somewhat answers my question. And in case it gets a SoC project I guess I'll learn more about it in one way or another ;-)
21:05:26brunocardosobluebrother: ;)
21:05:39jhMikeSNico_P: think it's a good idea to make the filechunk 32KB again? it's definitely a little better on anything I have to test.
21:05:50thegeekas a user with a lot of interest in rockbox development(I've been following the project for some years) I think it would be very interesting
21:06:06gevaertsbrunocardoso: do you think that getting rockbox to compile (and work...) with llvm is enough work for SoC ?
21:06:08RubberDucky132it's not listed as working or not working, but it's not a fancy dock / speaker / remote combo it's just a hifi grade "pickup" than turns the iPod line out into dual phono out, so prosumably it'll work? I mean it isn't even using serial at all I guess
21:06:15Nico_PjhMikeS: it looks better to me too... I couldn't really say why it's that way though ;)
21:06:40brunocardoson1s, gevaerts: yeah, i think the time is enough
21:07:04thegeekI think he meant is it too easy (not enough work for an entire summer)
21:07:09jhMikeSNico_P: heck, I'll even try larger ones if that won't break anything
21:07:27Nico_Pwhy not... I'll try too
21:07:29bluebrotherwhat do people think about FS #8751? IMO it should get rejected as it doesn't give much benefit
21:07:33brunocardoson1s, gevaerts: and like i said, if we get in trouble with some ARM backend issues, i can send patches to fix it! yeah, i think the time is enough
21:07:54 Join tessarakt [0] (
21:07:59brunocardososorry for the duplicated time enough message ;)
21:08:03amiconnIf llvm is about better optimisation, I think that coldfire could profit more from it than arm
21:08:30 Quit toffe82 (Read error: 104 (Connection reset by peer))
21:08:42n1samiconn: writing a coldfire backend seems more like an llvm project than a rockbox one :)
21:08:49scorche|shllvm doesnt have coldfire or sh support yet
21:08:50amiconngcc optimises reasonably well for arm in most cases. For coldfire, gcc compilations are sometimes horrible
21:08:58 Quit toffe82_ (Read error: 104 (Connection reset by peer))
21:08:59 Join toffe82__ [0] (
21:09:08 Nick toffe82__ is now known as toffe82 (
21:09:40*amiconn is still wondering why gcc -O2 and especially -O3 often make slower code for coldfire than just -O
21:09:46amiconnn1s: true
21:10:48Nico_Pbluebrother: I say close it, yes
21:11:08 Quit PaulJam (Read error: 110 (Connection timed out))
21:11:14brunocardosoamiconn: if it's just an ARM extension and if its very important to rockbox, that could be a rockbox project i think!
21:12:18n1s-O3 turns on a lot more inlining so it probably kills the icache, but generally I believe that the state of the m68k backend in gcc makes the rest of gcc take stupid decisions (m68k is used as an example of outdated backend in need of cleanup in the gcc wiki...)
21:12:35domonokybrunocardoso: coldfire is an m68k variant..
21:12:49brunocardosodomonoky: =/
21:13:08 Join techhelper1 [0] (
21:13:10Nico_PjhMikeS: a clue might be that the UI stays unresponsive even when audio is paused... shouldn't the codec be idle then?
21:13:18brunocardosodomonoky: not eligible then
21:15:56n1sheh, googling 'llvm m68k' gives our gsoc wikipage as nr 9
21:16:17bertrikI also wonder sometimes how the compiler can make decisions regarding speed if it doesn't know how fast memory accesses are (e.g. internal RAM could be faster than external RAM and a lot faster than flash)
21:16:38 Quit toffe82 (Read error: 104 (Connection reset by peer))
21:16:38jhMikeSNico_P: It's not really "idle" when paused per se but should be doing sleep() enough to be effectively so. btw, with 64K chunks the H10 buffers about as fast as e200.
21:16:58Nico_Pquite impressive I guess?
21:17:21gevaertsscorche|sh, other mentors: can you review ? It's my comment on the usability study application
21:18:24brunocardoson1s, so do you think i must send a proposal or leave the idea?
21:19:11 Join toffe82__ [0] (
21:19:19 Nick toffe82__ is now known as toffe82 (
21:19:38n1sI'd say go ahead and send in the application but I'm not the one deciding which are accepted (I don't know exactly how that works)
21:19:39jhMikeSNico_P: though it does seem more affected by codec PCM filling and speeds up more when it finally tops the PCM off
21:20:40brunocardoson1s: ok then! if anyone has any ideas regarding llvm use, i'm here to discuss! thanks!
21:20:57 Quit toffe82 (Read error: 104 (Connection reset by peer))
21:21:00scorchegevaerts: add a bit about having some previous ideas of things that are currently wrong with our "usability"...also change the thing about experience from a first-year student to someone without past expereience in doing such usability studies/indepth UI design
21:21:06 Join toffe82__ [0] (
21:21:14 Nick toffe82__ is now known as toffe82 (
21:22:18gevaertsscorche: I'm not sure what you mean by your first point
21:23:04 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
21:23:07bluebrotherwrite something up why the current usability approach needs changing?
21:23:10Nico_PjhMikeS: when resuming playback with the player essentially idle, there seems to be two phases. at first buffering is rather slow, then past approx. mid buffer it gets much faster... any idea why that could be? codec boosting less maybe?
21:23:35Nico_Pall the time the PCM buffer appears essentially full to me though
21:23:38scorche|shgevaerts: in other words, i would like to see the person applying for that project list a few things they think could be done better/is done wrong that they could fix
21:23:44saratogadid anyone still want to see my old GSOC app
21:24:14scorche|shas in, they already have some ideas of where they could go and what they could do to improve our usability even before the "study"
21:24:38gevaertssaratoga: I've finished the template now. Feel free to review and improve it though
21:25:35saratogamaybe include something about feasibility
21:25:55 Quit toffe82 (Read error: 104 (Connection reset by peer))
21:26:35gevaertssaratoga: could you suggest a sentence ?
21:26:49*gevaerts is not good at making up text
21:27:02scorche|shWhy do you think this project would be beneficial to Rockbox?
21:27:05 Join toffe82__ [0] (
21:27:08jhMikeSNico_P: codec only boosts when under 1/2 of audio otherwise it's normal priority
21:27:13 Nick toffe82__ is now known as toffe82 (
21:27:14jhMikeS*1/2 second
21:27:38scorche|shsaratoga: were you referring to something like that, or something more along the lines of how well they htink they can do the project
21:27:42Nico_Phow much does the PCM buffer hold?
21:27:45Rincewindsaratoga: if you don't mind, I would like to have look at your application
21:28:04*Nico_P can also paste his application if it is of interest
21:28:23jhMikeSNico_P: 3 seconds without crossfade enabled
21:28:44Nico_PjhMikeS: ok, so it definitely didn't go below 2 secs
21:28:56Nico_Pas far as I could see anyway
21:29:00*Rincewind is stuck at formulating his ideas in english at the moment
21:29:11gevaertsscorche|sh: current draft response is at
21:29:17 Join robin0800 [0] (
21:29:43RincewindNico_P: I would like to see your's, too.
21:30:15Nico_PRincewind: as soon as I find it :)
21:30:42scorche|shRincewind: planning on applying? ;)
21:30:49jhMikeSNico_P: was it the case that "access to the menus seems unaffected"? The statement in 8807 seems to indicate the unresponsiveness doesn't happen to the menus which would rule out any sort of priority override.
21:30:53Rincewindscorche|sh: yes
21:31:51*bluebrother wonders why the m:robe dock doesn't have a line out connector
21:31:52Nico_PjhMikeS: I had to wait to be able to get in the menu but once there it was ok
21:32:45jhMikeSNico_P: so you just entered the menu normally without changing playback state in any way and then no UI sticking?
21:33:22Nico_PjhMikeS: I think that's it, yes
21:33:43jhMikeSbtw, does it ever return to normal again without a reboot or anything?
21:33:44Nico_Pthe menu appeared after a delay, but then it wasn't sluggish
21:33:49Nico_Pyes it does
21:34:10Nico_Pafter a while it becomes perfectly fine again... with your patch atleast
21:34:29jhMikeSdoes the WPS in use matter?
21:36:06 Quit mf0102 ("Verlassend")
21:36:21gevaertsscorche: Can I submit ?
21:37:01 Quit n1s ()
21:37:15*gevaerts is too cowardly to submit this without approval
21:37:41Nico_PjhMikeS: I haven't checked that
21:37:44Nico_PI'll try
21:37:50gevaertslinuxstb_: you also expressed some doubts about this application earlier. Can you review as well ?
21:38:15saratogaregarding feasibility, i remember doing some profiling of the WMA codec in fp to explore how much optmization would be needed and in what code
21:38:40saratogait turned out that the fft used most of the time, so I knew that the fixed version would be feasible without too much optimization
21:38:55scorchegevaerts: just do some small things such as capitalizing Rockbox, say "Google Summer of *Code*" instead of "summer of CODE", etc
21:39:02 Join Zagor [242] (
21:39:33gevaertssaratoga: I added a sentence to the effect that they should include info to convince us that they will be able to complete the project
21:39:38gevaertsscorche: OK. I'll do that
21:40:11 Join silvio [0] (
21:40:31linuxstb_gevaerts: Seems fine to me
21:40:31scorchealso where you say "a more detailed proposal" say "a more detailed proposal with some specific tasks or a sort of rough plan"....well...something like that
21:40:32 Join xnyhps [0] (n=xnyhps@2001:470:1f08:142:1:3:3:7)
21:40:56 Quit RubberDucky132 ("ChatZilla 0.9.81 [Firefox 3.0b4/2008030714]")
21:40:58linuxstb_scorche: I think that's covered by the page gevaerts links to
21:41:14jhMikeSNico_P: how bout enabling the kernel checks in kernel.c? It's behaving like it could if that mutex is being corrupted by something which is more catastropic than if that struct spinlock got corrupted.
21:41:16bluebrotheris there a wiki page about the m:robe 500 port?
21:41:19scorchealright...that works too then
21:41:35Nico_PjhMikeS: what do they do?
21:41:42jhMikeSNico_P: might as well turn them in thread.c too
21:41:48*gevaerts submits this comment
21:42:05*linuxstb_ wonders about the best way to build 20 versions of checkwps
21:42:07jhMikeSthey check the queues and lock owners to be sure they're correct and will panic if something isn't right
21:44:18 Join gregzx_ [0] (
21:45:23Rincewind"they check the queues and lock owners to be sure they're correct and will panic if something isn't right" <−− sounds a lot like airport security to me :)
21:45:40 Quit gregzx (Nick collision from services.)
21:45:42 Nick gregzx_ is now known as gregzx (
21:45:52Nico_PjhMikeS: btw the term "UI response" might have been a bit wrong... what I'm doing is pausing/unpausing the audio, which of course requires the audio thread to run at some point
21:45:54saratogahere is the nonperson part of my application (preceeding this was a detailed list of the experience and coursework I'd done to qualify me for the proposal):
21:46:17jhMikeSthe terrorists might be getting be getting through in this case
21:46:23saratogahere is the public proposal:
21:46:42Nico_Psaratoga: are the peronal parts still available to us?
21:46:57gevaertsRincewind: anyway, I guess you're good at running if there is any rumor of panic ;)
21:47:00saratogaNico_P : presumably bagder, et al still have them
21:47:08 Quit toffe82 (Read error: 104 (Connection reset by peer))
21:47:13saratogai odn't mind emailing them, just don't want to pastebin them
21:47:28Nico_PjhMikeS: accessing the menu doesn't seem to lag if it's not waiting on an audio command's response
21:47:40Nico_Pso I guess it's the audio thread that's making things sluggish
21:47:52Rincewindsaratoga: thanks
21:48:05*gevaerts considers reading all these proposals to see if he can reject them easily ;)
21:48:05 Join toffe82__ [0] (
21:48:13 Nick toffe82__ is now known as toffe82 (
21:48:50Nico_PjhMikeS: also the WPS might have an influence but it could also just be my appreciation
21:49:08 Quit toffe82 (Read error: 104 (Connection reset by peer))
21:50:10 Join toffe82__ [0] (
21:50:19 Nick toffe82__ is now known as toffe82 (
21:51:35jhMikeSNico_P: all I know is something funny is happening on a particular device that doesn't ever seem to happen with others.
21:51:44 Quit techhelper1 ()
21:52:49 Quit portavales ("Leaving")
21:52:54Nico_PI've enabled the checks, I'll test with that now
21:53:47pixelmaBagder: you around?
21:55:46Nico_PjhMikeS: there appears to be no panics
21:56:54Nico_Pmaybe it's time to call usb serial to the rescue
21:57:10jhMikeSNico_P: this never occurs without the startup scans also running?
21:57:27*gevaerts hears someone mention usb and wakes up
21:57:39Nico_PjhMikeS: I haven't tried disabling the DB yet
22:00:40 Quit argumentD (Read error: 110 (Connection timed out))
22:02:08Nico_Pfunnily enough, things don't seem much better with DB and tagcache disabled... (still 64K chunksize)
22:03:48Bagderpixelma: yeps
22:03:58saratogaBagder: did you add my machine as a build server?
22:04:14jhMikeS?? and yet in that situation only the buffering thread should access the disk and the mutexes practically a nop in that case. hmmm.......
22:05:02Bagdersaratoga: uh no, I've missed that info. what's the host name?
22:05:08Nico_PjhMikeS: I'm surprised too.. could it be that I've left the checks enabled or are those light enough not to have an influence?
22:05:18 Join smiler777 [0] (n=smiler-7@
22:05:34smiler777hi evryone!
22:05:35Nico_Pfrom what I'm seeing I'm more and more convinced the lack of responsiveness is caused by the audio thread
22:05:38jhMikeSthey are quite lightweight
22:06:01Nico_Pmaybe its queue is flooded, or maybe it's not running enough
22:06:08smiler777i have 1 question about rockbox
22:06:17smiler777can U help me&
22:06:25bertrikask the question
22:06:27jhMikeSit's priority is the same as UI now but since it doesn't have too much to do that should be ok
22:06:33Nico_Pbut the pause/resume events take a damn long time before getting processed
22:06:52linuxstb_Bagder (or Zagor?) - can I point you to this?
22:07:26Nico_PjhMikeS: is there a PCM thread?
22:07:27 Quit toffe82 (Read error: 104 (Connection reset by peer))
22:08:19jhMikeSNico_P: I'll cook up a debug screen that displays the some important internal data that if messed up, well produce odd behvior but it was always correct on any target I have.
22:08:31 Join toffe82__ [0] (
22:08:34jhMikeSThere's no PCM thread. The codec thread feeds the buffer
22:08:40 Nick toffe82__ is now known as toffe82 (
22:08:47smiler777i from russia, so im sorry for bad english... I have ipod video 30 gb. 1 year ago i install to ipod rockbox version 16xxx. Everything allright, but battary life - only 3-4 hours. Does it fixed on latest rockbox and how long ipod can work without charging?
22:09:10Nico_Pthen I think I'll build a usb serial build and watch the audio queue
22:09:30 Join techhelper1 [0] (
22:09:31scorche|shsmiler777: i wouldnt say "fixed", but it does last a good deal longer now...try it yourself
22:10:13smiler777scorche|sh - how long ipod can work non-stop now?
22:10:24scorche|shthat depends on your battery
22:10:33scorche|shlike i said...try it yourself
22:11:29 Join FOAD_ [0] (
22:11:29 Quit toffe82 (Read error: 104 (Connection reset by peer))
22:11:46smiler777scorche|sh thx! ;) Is russian user here except me? ))
22:11:56*scorche|sh shrugs
22:12:12Bagderwe're all a happy family here from planet internet
22:12:35 Join toffe82__ [0] (
22:12:43 Nick toffe82__ is now known as toffe82 (
22:12:45smiler777Bagder okey , im in )
22:12:49 Quit tessarakt ("Client exiting")
22:13:14smiler777i go to reflash ipod.. wish me lusk ))
22:13:30scorche|shyou arent reflashing it
22:13:38Bagderpixelma: small m3 pic updated
22:13:54smiler777scorche|sh ?
22:14:27scorche|shwe dont touch the flash
22:14:44pixelmaBagder: already spotted it, thanks :)
22:16:04 Join crzyboyster [0] (n=6210f0db@gateway/web/cgi-irc/
22:16:53smiler777scorche|sh im russian, so my english dictonary is very small. "reflash" - mind install latest rockbox. xorry
22:16:59crzyboysterHas the CustomWPS page been updated to have the tags for viewportifying the menu/list?
22:17:37linuxstb_crzyboyster: What tags?
22:17:42 Quit silvio (Remote closed the connection)
22:18:19crzyboysterWhat exactly was the recent commit? I figured it allowed viewports in the menus now? That's the tags I was asking about.
22:19:36scorche|shsmiler777: well, reflash is a specific technical term..."flashing" is as well...being that we just change a few things on the hard drive and dont touch the flash, we arent "reflashing ipods"
22:21:29 Quit crzyboyster ("CGI:IRC (EOF)")
22:22:04pixelmaLlorean: do you mean "Party Mode" in your recent forum post? ;)
22:23:35Nico_Pgevaerts: I interrupted cu with CTRL-C and now when I try to restart it I get "/dev/ttyUSB0: Line in use"... any idea what I should do?
22:24:22gevaertsNico_P: are you doing this as root ?
22:24:30 Quit toffe82 (Read error: 104 (Connection reset by peer))
22:24:52Nico_Pshould I?
22:24:53Lloreanpixelma: Fixed, thanks.
22:25:18gevaertsNico_P: no. It's just that I sometimes get that error when running as root (but not otherwise)
22:25:27gevaertsDid you unplug and replug since then ?
22:25:33 Join toffe82__ [0] (
22:25:41 Nick toffe82__ is now known as toffe82 (
22:25:56Nico_Pah. did that and now I'm good
22:26:36gevaertsProbably the previous cu lock was not cleared. Try exitting by ~. (and waiting a few seconds after that)
22:27:17 Join mirak [0] (
22:28:41 Quit FOAD (Read error: 110 (Connection timed out))
22:28:42 Nick FOAD_ is now known as FOAD (
22:28:44 Quit toffe82 (Read error: 104 (Connection reset by peer))
22:28:48Nico_PjhMikeS: plenty of things seem to be happening between the moment the pause event is sent and the moment it is processed
22:29:49 Quit mirak (SendQ exceeded)
22:30:05 Join toffe82__ [0] (
22:30:13 Nick toffe82__ is now known as toffe82 (
22:31:20 Join mirak [0] (
22:31:27 Join soap_ [50] (n=soap@rockbox/staff/soap)
22:33:32 Quit toffe82 (Read error: 104 (Connection reset by peer))
22:34:23 Join toffe82__ [0] (
22:34:32 Nick toffe82__ is now known as toffe82 (
22:41:34 Quit saratoga ("CGI:IRC (EOF)")
22:42:40*gevaerts has good news and bad news
22:42:44 Join OlivierBorowski_ [0] (
22:43:35 Quit toffe82 (Read error: 104 (Connection reset by peer))
22:43:38gevaertsI just copied 7GB from my c250, 2GB of which were behind a hub, without a single message in dmesg
22:43:58 Join toffe82__ [0] (
22:44:06 Nick toffe82__ is now known as toffe82 (
22:44:10bluebrothernice. And for the bad news?
22:44:14*scorche|sh hopes that was the bad news
22:45:11gevaertsThe bad news is that I had to make sure that there is _nothing_ else running while the usb transfer is busy. I tested by disabling double buffering, and all non-usb interrupts
22:45:53gevaertsMy next test will try using IRAM (and interrupts enabled). Maybe that will be sufficient as well
22:46:22smiler777oh my god )) i install rockbox 16828xxx - the ipod baterry lifetime on display - 9 hours! never seen that digits ))
22:46:37scorche|shthat number isnt never was
22:46:44smiler777thx everybody
22:47:19 Quit soap (Read error: 110 (Connection timed out))
22:47:41smiler777scorche|sh i know, but that number near truth, vith turned off display, i think so
22:48:01scorche|shi wouldnt trust anything short of an actual battery bench
22:48:19bluebrotherBagder: do you think MarcGuay should be able closing FS tasks? He seems to be doing quite some nice cleanup work
22:48:43BagderI think he could use that, yes
22:49:08*gevaerts tries to find out how he can get 32kb IRAM
22:49:16bluebrotherhe usually isn't around here, right?
22:49:45 Quit toffe82 (Read error: 104 (Connection reset by peer))
22:50:50bluebrotherI'm going to give him a hint in a forum PM ;-)
22:50:59 Part xnyhps
22:52:00 Join toffe82__ [0] (
22:52:08 Nick toffe82__ is now known as toffe82 (
22:52:52Bagder1056 open tasks...
22:54:49bluebrotherurgh :(
22:54:50LloreanI've closed 5 or 6 in the last three days. If everyone else in the channel follows my lead, we'll be fine. :-P
22:55:02bluebrotherwe really need a new cleanup week ...
22:55:11*Bagder noticed 35 PM requests...
22:55:42LloreanPM requests?
22:56:09Bagderyes, in the tracker
22:56:15Bagderclose and re-open requests basically
22:56:24gevaertsCan anyone confirm that static unsigned char buffer[BUFFER_SIZE*2] IDATA_ATTR;
22:56:44LloreanBagder: Any of them on tasks I closed?
22:56:48gevaertswill end up in IRAM ? And will I need UNCACHED_ADDR with it ?
22:56:49*Llorean tries to be clear on 'why' when closing.
22:56:56ZagorI'm strongly considering removing the possibility of sending PM requests
22:57:12Bagdergevaerts: you can always check the map file after a build to verify
22:57:28*gevaerts should have thought of that
22:57:36BagderZagor: yes, it seems to not do much good
22:58:44 Quit toffe82 (Read error: 104 (Connection reset by peer))
22:58:51 Part smiler777
22:58:52Bagderwell, some of the close-requests are at least partly useful
22:59:04Zagoryes but they could just as well be made in mail or irc
22:59:13gevaertsregion IRAM is full :(
22:59:27BagderZagor: yes, or as plain comments in the entry
22:59:43 Quit OlivierBorowski (Read error: 110 (Connection timed out))
23:00:11 Join toffe82__ [0] (
23:00:19 Nick toffe82__ is now known as toffe82 (
23:00:47bluebrotherhmm, has a date for devcon been discussed at all?
23:00:51 Join mikus [0] (n=5316dbf1@gateway/web/cgi-irc/
23:01:06Bagderthere's too little devcon talk in general
23:01:30*scorche|sh nods
23:01:43*scorche|sh sees it on his list under GSoC and themes
23:02:04***Saving seen data "./dancer.seen"
23:02:39 Join haemmy [0] (n=haemmy@
23:03:42amiconngevaerts: If you declare an array like this, use IBSS_ATTR, not IDATA_ATTR
23:04:20amiconn(that won't help with the 'region blah full' problem though)
23:04:20 Quit ender` (" Theory: we know everything, but nothing works. Practice: things work, but we don't know why. Combine them: nothing wo")
23:04:52gevaertsamiconn: thanks. It's just for a test right now, but still
23:05:08 Quit toffe82 (Read error: 104 (Connection reset by peer))
23:05:41 Quit jgarvey ("Leaving")
23:06:05 Join toffe82__ [0] (
23:06:14 Nick toffe82__ is now known as toffe82 (
23:06:31 Join `Tanner [0] (
23:07:28*amiconn now gets 436fps on the H1x0 remote boosted :)
23:07:55Nico_Pamiconn: you should really stop now :)
23:08:25 Quit piga (Remote closed the connection)
23:08:49BagderI bet you can't reach 1000!
23:09:06amiconnIndeed I can't
23:09:32amiconnEven the fastest main LCDs don't reach 1000 (but the H1x0 and especially the M5 get close...)
23:09:49 Quit pondlife ("Leaving.")
23:10:22amiconnI just applied the things I've learned while squeezing the iaudio remote transfer back to the iriver remote
23:11:19amiconnThey can't be applied as-is though, because the iaudios have remote lcd CLK and DATA on the same GPIO port, while the irivers have it on different GPIO ports
23:11:43 Quit brunocardoso ("Leaving")
23:12:32 Quit Zarggg ()
23:12:45 Join matsl [0] (
23:13:56*Llorean wonders how much ticking you get with greylib running.
23:14:04 Quit toffe82 (Read error: 104 (Connection reset by peer))
23:14:44amiconnI won't port the greylib just now, but I think it would be more like humming than ticking
23:14:58 Join toffe82__ [0] (
23:15:07 Nick toffe82__ is now known as toffe82 (
23:15:16amiconnThe faster transfer should in fact help to reduce ticking a bit further on affected irivers
23:15:31LloreanThat's good, then.
23:17:18amiconnWhat we're doing to reduce ticking is to spread out the byte transfers, in order to reduce the energy density of the transfer. The faster transfer results in shorter transfers and longer pauses, given the same overall update speed
23:17:37amiconn(which aims at 60fps with ticking reduction enabled)
23:18:40 Join petur [50] (n=petur@rockbox/developer/petur)
23:18:41amiconnMy H180 only ticks very little, almost unnoticeable with my current earphones. My H340 doesn't tick at all
23:18:46LloreanDo we need 60? That seems a bit high to me.
23:19:16amiconn60 is already quite low
23:20:33 Join GameJerk [0] (
23:21:03LloreanWell, specifically for WPS actually, I'd imagine we could go a good deal lower?
23:21:04 Quit mikus ("CGI:IRC (EOF)")
23:21:09 Quit DerPapst (Read error: 110 (Connection timed out))
23:21:28GameJerki can't seem to find the page that lists what each version specifically does and does not do. I'm looking for specs on Rockbox for the Sansa e200 series.
23:21:31 Quit miepchen^schlaf ()
23:22:01LloreanGameJerk: There's not really differences in the versions, besides things the hardware simply can't do (no radio if your player doesn't have radio, etc).
23:22:55amiconnLlorean: Don't forget peakmeters etc...
23:23:31Lloreanamiconn: "etc"? I would think it's just peakmeters.
23:23:40GameJerkthx badger. Is there a way to boot the Original firmware while still having rockbox installed?
23:23:44 Join stripwax [0] (
23:23:47amiconnWell, in the wps itself, yes, atm
23:23:49BagderGameJerk: yes
23:24:17 Quit stripwax (Client Quit)
23:24:31 Join feindbild [0] (
23:24:33GameJerkbagder: how would i got about doing that on the sansa e200?
23:24:39feindbildhi =)
23:24:40amiconnBut maybe the idea of a plugin viewport gets implemented some day, and you'll run oscilloscope in one...
23:24:44 Join miepchen^schlaf [0] (
23:24:48scorche|shGameJerk: have you read the manual?
23:24:49BagderGameJerk: it's all detailed in the fine manual!
23:25:44GameJerksorry i got it now. I must have missed it
23:25:45 Quit `Tanner ("ChatZilla 0.9.80 [Firefox]")
23:25:52GameJerksorry for being a twat :)
23:26:06feindbildany sansa e200 users here using the latest builds? I keep get 'Undefined instruction at 001400C8 (0)' and was wondering If I'm the only one with this problem ...
23:26:14 Quit gromit`` (Read error: 110 (Connection timed out))
23:26:27GameJerki get that too feinbild
23:26:35GameJerkwhen i connect to USB
23:27:25GameJerki like rockbox and all but i wish i could record radio w/it. thats the only thing that is keeping me from using it
23:27:44feindbildGameJerk: =/ I get it without connecting to USB ...everytime rockbox boots ...
23:27:46gevaertsGameJerk: that should work
23:28:26pixelmaGameJerk: hmm? You can... (might be not so obvious that you have to do that from the radio screen though)
23:28:28bertrikI have a sansa e260 with a recent build but no undefined instruction
23:29:04pixelmaerrmm s/radio screen/recording screen
23:29:55GameJerkoh so i have to turn on the radio first and then go to the recording screen?
23:30:41pixelmano, you have to go to the recording screen and set the source to radio (sorry for the typo that causes confusion)
23:31:42 Quit domonoky (Read error: 104 (Connection reset by peer))
23:32:19GameJerkoh the recording settings screen. found it
23:32:30kugelpixelma: Did noone actually thought of implementing press/hold REC in radioscreen to record?
23:32:41kugelI have no radio, so I didn't :p
23:32:57 Quit alienbiker99 ("( :: NoNameScript 4.02 :: )")
23:33:27 Quit Rincewind ("bye")
23:33:29GameJerkkugel: ya that would be sweet. That way if i wanted voice notes i wouldn't have to go back and forth all the time to switch the input
23:34:01pixelmakugel: I think it's not that easy to just implement the interface for it on swcodec. On hwcodec it's done this way... :P
23:34:20kugelpixelma: ah ok
23:35:06GameJerkwhere is the proper placement for the radio preset file?
23:35:36 Join Shaid [0] (
23:37:33 Quit tvelocity (Remote closed the connection)
23:39:11linuxstb_GameJerk: .rockbox/fmpresets/ - see here -
23:40:48jhMikeSpixelma: you can get to the recording screen from the radio screen from the radio menu
23:42:08gevaertsIt seems that just putting the usb buffer in IRAM doesn't help (it actually makes things worse)
23:44:01pixelmajhMikeS: ah yes, forgot about that (only one of my two swcodec players have a radio and I very seldom use radio recording)
23:45:11GameJerkthere any hope for better video codec support or does the hardware just not capable of anything better?
23:45:41 Quit toffe82 (Read error: 104 (Connection reset by peer))
23:46:22bluebrotherGameJerk: define "better".
23:46:44Nico_PjhMikeS: seen my pastebin?
23:46:47 Join Zarggg [0] (
23:47:11feindbildxvid would be nice ...
23:47:22GameJerkmore codecs supported. like xvid etc..
23:47:55jhMikeSNico_P: check this OS stacks screen, one thing you should never see in there is any "FF" or the like for a value:">
23:48:12dionoeaYou'd need to resize the video to fit the screen anyway (else the CPU wouldn't be able to decode the video)
23:48:12bluebrotherGameJerk: those are too resource hungry.
23:48:22jhMikeSNico_P: looking now
23:48:43GameJerkhmm i load the FM preset file from the file tree just fine. but when i go back to FM radio on the main screen it says "no presets found.. autoscan?"
23:48:54Nico_PjhMikeS: there's a bunch of buffering and even audio stuff between the moment a pause event is sent and the moment it's processed?
23:49:01kugelfeindbild, GameJerk: Mostly the hardware isn't capable. Future targets may have enough power. But even if the targets are strong enough, there're needs to be someone who implements it (anyone who wants that too). I personally think mpeg2 does a good job
23:49:22GameJerkdionoea: what is a good program to reencode the vids?
23:49:35 Join DerPapst [0] (
23:49:50feindbildGameJerk: I use mencoder ...
23:49:52dionoeaGameJerk: check the MpegPlayer wiki page
23:50:05 Join jhulst [0] (
23:50:33*pixelma is surprised dionoea didn't recommend ... ;)
23:50:34markunfeindbild, GameJerk: maybe you can use this as a starting point for your xvid player:
23:50:51scorche|shpixelma: me too :)
23:51:14dionoeapixelma: I'd have to explain it if I did
23:51:20 Quit ompaul (Client Quit)
23:51:38dionoeaAnd I don't have time now :)
23:51:49 Join CaptainKewl [0] (
23:51:52feindbildGameJerk: mencoder infile.avi -of mpeg -oac mp3lame -lameopts vbr=2:mode=1:q=7 -af resample=44100:0:0 -ovc lavc -lavcopts vcodec=mpeg2video:dia=6:mbd=1:vqscale=5 -vf scale=220:176,harddup -ofps 15 -zoom -xy 220 -o outfile.mpg
23:52:10feindbildGameJerk: thats what I use for the sansa e200 ...
23:52:19jhMikeSNico_P: wouldn't it keep buffering during pause?
23:52:24amiconnYou don't need to resample anymore
23:53:14Nico_PjhMikeS: it would, yes. I'm referring to what happens between the sending of an event and its processing
23:53:27GameJerkthx fein: i might skip it. this player is for my brother. I'll just keep using my iphone for video
23:53:28Nico_Pthat's what induces the noticeable delay
23:53:29 Join stripwax [0] (
23:53:36 Quit stripwax (Client Quit)
23:53:41 Quit Zagor ("Client exiting")
23:55:16gevaertsOK. Short summary of my USB tests (on c250). If I (only) disable all non-usb interrupts, it works "better". If I (only) disable double buffering (= simultaneous disk and usb I/O), there are no real changes. If I disable both double buffering and non-usb interrupts, everything seems to work perfectly (I now did 7GB without a hub and 2GB with a hub)
23:55:36 Quit jhulst ("Konversation terminated!")
23:56:11gevaertsTrying to put the buffer in IRAM doesn't work at all
23:57:01GameJerkwhen i put the FM preset file in .rockbox/fmpresets it doesn't load when i reboot. And when i load the file from the file tree it opens up just fine but doesn't allow me to save the presets so when I go to "FM Radio" it just says "No preset found, autoscan?" any ideas on whats going wrong?
23:57:02kugellinuxstb_: Regarding your parse_list function. Needs to be there a traling seperator? i.e. |..|... or ...|...
23:57:10feindbilddeleting the .rockbox dir and freshly installing rockbox fixed the undefined instruction thingy ^^
23:57:45gevaertsMy interpretation of this is that the USB controller reads directly from RAM, without any buffering. If something else tries to access RAM simultaneously, the controller hiccups and either breaks of the packet or sends it with a (very) short interruption halfway, to which host and hub react badly.
23:58:05 Join toffe82__ [0] (
23:58:13 Nick toffe82__ is now known as toffe82 (

Previous day | Next day