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 2009-02-15

00:00:36Mouser_XIn other words, there's really nothing (aside from coding) I can do to "fix" this?
00:00:47Mouser_XOr work-around it?
00:01:24Mouser_XI really don't use the DB much, but I do have it enabled. Pretty much all I use it for though is to keep track of ratings.
00:01:32Mouser_X(That is, the user-ratings.)
00:01:41Mouser_X(It makes it easier to find the songs I like.)
00:02:11*gevaerts is still confused about keeping songs you don't like on your player
00:02:19Mouser_XAnother DB question: Is there any way to get it to display file names, when there's no tags?
00:02:49Mouser_Xgevaerts: I still have 3 GB free on my player. I see no reason to get rid of them.
00:02:58*Mouser_X has a Gigabeat S60)
00:03:28Mouser_XNot only that, I never said I didn't like them. I *did* say there are those I like more.
00:03:46gevaertsYou didn't say "more"!
00:03:48Mouser_X(There's very few I don't actually like. Those ones I usually do delete.)
00:04:29n1sMouser_X: i don't think you can make the database display filenames without coding either
00:04:41Mouser_XThat really stinks.
00:04:59Mouser_XI have 30 GB of ADXs, and the ADX format doesn't support tags.
00:05:16Mouser_XThus making it extremely difficult to recognize which file the DB is referencing.
00:06:16 Quit amiconn (Nick collision from services.)
00:06:17 Quit pixelma (Nick collision from services.)
00:06:17 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
00:06:18 Join pixelma_ [50] (n=pixelma@rockbox/staff/pixelma)
00:06:31 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
00:06:37 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
00:06:39Mouser_XI was hoping that perhaps there would be some way to create a custom display that could do file names instead of tags.
00:06:52Mouser_XI've never used those, so I have no idea.
00:06:57 Quit __lifeless (Remote closed the connection)
00:07:08n1swell, i don't think you can, i've never used the database though
00:07:14 Join __lifeless [0] (n=lifeless@
00:08:10n1salthough i can't imagine codign some kind of fallback to use the filename as track title if no metadata is found for a track when building the databse should be too difficult
00:08:44Mouser_XThe biggest problem with that, is that I really know very little programming.
00:09:08Mouser_XI'm taking a C++ course, in the hopes that it will give me some foundation upon which I can start from though.
00:09:25Mouser_XI'd like to get the GBS codec working in rockbox (fs#7331(
00:09:36Mouser_X... *)
00:10:12n1syeah the current way isn't too nice for formats that don't have metadata
00:12:32saratogabertrik: if you want to measure overhead, I recommend updating the buffering thread screen to only update once every 30 seconds or so
00:13:05saratogabesides giving you a very nice running average, this also prevents LCD updates from using a lot of CPU time
00:13:29saratogaBuschel did this and reported something like 2.5MHz of overhead during playback
00:14:01amiconnJdGordon: Eh? The database does support chunked browsing. That used to work, and it still did when I checked last
00:14:08saratogasince servicing interrupts is probably one of the largest parts of this [when dsp isn't doing much], a savings of about a MHz sounds about right
00:14:18JdGordonit does?
00:14:42amiconnYes, otherwise the "all tracks" list would be basically unusable
00:14:57Mouser_XI sort of thought it used to work...
00:15:14Mouser_XBut it's been a very long time ago, and I don't use the DB very often, so I wasn't sure.
00:15:25Mouser_XI'm using r20002.
00:16:42amiconnStill works here. I can browse my 3000+ all-tracks list with my standard "max files in dir" setting (300)
00:17:18amiconnThis is r19933
00:17:30Mouser_XHuh... Can you view the History section?
00:17:39Mouser_XRecently played/most played?
00:17:44amiconnNever used that one....
00:17:47Mouser_XThose are the ones that don't work for me.
00:18:58amiconn*yes. All my tracks seem to be in "Never played" though, as I don't let it collect stats
00:19:26amiconnIt's a bit slow when it fetches the next chunk (using my old trusty recorder), but it works
00:20:02Mouser_XOdd... I have my dir buffer at 7000 (at least, for a brief moment I saw 6000+ files found), and nothing shows up in either of those.
00:20:07Mouser_XAnd I do gather runtime data.
00:20:34amiconnSetting "max files in dir" to very high values used to cause weird problems in the past
00:20:47Mouser_XWhen I go to "Never played" I have about 1000 results.
00:20:55Mouser_XI'll try reducing it and see what happens.
00:21:10amiconnImo anything above 1000 is insane. I usually set it to 300 on hdd targets, and 150 or 200 on flash targets
00:21:53amiconnIt might mean that the weird behaviour (aka bug) with high settings is still there
00:21:54Mouser_XWould that setting be under "Max entries in file browser"?
00:21:59Mouser_XBecause I have that at 3000.
00:22:08Mouser_X(Just looking right now.)
00:22:13amiconnRemember to reboot to activate the changed value
00:22:17Mouser_XI thought it was higher.
00:22:25Mouser_XNo, that's what it's at already.
00:22:49Mouser_X(I have it at 3000 because of my OC ReMix directory. Though, I could get away with 2000 for that.)
00:23:52 Join sarixe [0] (
00:23:53Mouser_XI'll set it to 500 and see what happens.
00:24:20amiconnJdGordon: I just found a status bar bug while checking the database...
00:24:53JdGordonoh goody
00:25:23amiconnGo to "Database->Same as current...". Select an arbitrary subcategory. The vkeyboard will appear, without statusbar (of course). Now cancel the vkeyboard
00:25:46amiconn-> The statusbar doesn't reappear, and the top 8 pixels of the vkeyboard stay there
00:26:27JdGordonok, put it on flyspray incase i dont get aroudn to fixing it today
00:26:42Mouser_XSo I set it to 500, rebooted (check the OCR directory to be sure), and the DB says "Dir buffer full!" after about 6000 files.
00:27:00amiconnI never saw the database display that...
00:27:04 Part gartral
00:27:11*amiconn wonders what's going on there
00:27:24Mouser_XIt does it every time i try to access the "Most played" or "Recently played" sections.
00:27:48Mouser_XWorks fine on "Never played" though.
00:27:50amiconnThe dir browser does display that if the dirctory doesn't fit, but the db never should
00:27:59amiconnHmm, interesting
00:28:15amiconnI can't verify this because both sections show no entries for me
00:28:15Mouser_XAlso, when accessing "Artist" and "All files", that works as well.
00:28:27***Saving seen data "./dancer.seen"
00:29:19n1samiconn: they require the runtime data gathering to be enabled i think
00:29:26Mouser_XI suppose the next best thing would be to export my database stuff, wipe the database, reinitialize it, fiddle with it a bit and see if the results return?
00:29:28amiconnn1s: I know
00:29:38amiconnI don't collect stats. I don't use the database except for (1) occasional checking for tag spelling errors and (2) to make pf happy
00:29:38Mouser_XIf they don't, then import my changes and see if that effects it?
00:30:12amiconn(of course (2) doesn't apply to hwcodec, so just (1) on my recorder)
00:30:47Mouser_XI don't use the DB all that often. As I said, I listen to ADX files a lot, and those display really poorly in the DB.
00:30:57Mouser_X(Due to no tags.)
00:31:35 Quit {phoenix} (Remote closed the connection)
00:35:07 Join bs66_ [0] (
00:39:09Mouser_XI know that "Recently played" used to work. There's been a few times when Rockbox crashed (for whatever reason. Usually some unhandled IRQ? It was awhile ago, shortly after they go audio working on the GigabeaSt), and I'd use it to see what file I played last before it crashed.
00:41:56 Quit n17ikh ()
00:45:36 Quit rakata (Client Quit)
00:51:25 Join n17ikh [0] (n=n17ikh@
00:52:19Mouser_XOf interesting note: I'm reading the DataBase entry in the wiki, and I think you can do a conditional statement so that filenames are used when tags aren't present. However, since the tagnavi_custom.config thing confuses me, I'm not sure about this.
00:56:18 Quit bs66_1 (Read error: 110 (Connection timed out))
01:02:38 Join AndyIL [0] (i=AndyI@
01:09:57 Quit yhuang (Remote closed the connection)
01:14:20 Join miepchen^schlaf [0] (
01:15:45Mouser_XI'm wondering if the database (specifically the History section) has a difficult time handling an excessive amount of tracks? 7000+ is quite a few.
01:23:07 Quit AndyI (Read error: 113 (No route to host))
01:24:49 Quit SoapWork ("CGI:IRC")
01:26:59saratogan1s: [for the logs] I don't see any reason not to commit 8778, its absolutely trivial and can be reverted if it causes any problems, and it apparently fixes a serious problem
01:28:46 Quit dfkt (Read error: 104 (Connection reset by peer))
01:28:54amiconnIt works around it.... it's not a proper fix
01:32:10saratogaamiconn: does it break something else though?
01:33:40 Quit bertrik ("Leaving")
01:33:58amiconnIt might break on some G2's. Obviously we have a problem in our revision detection, and it needs some RE to find out how to do it properly
01:34:57amiconnActually I know what goes wrong on the problematic G1's - they don't have the expected data in the rom in the location we're checking
01:35:20amiconnBut I don't know yet what the proper way is
01:37:18n1samiconn: couldn't we disable the wheel poweroff thing untill that's done then, crappy runtime is better than no wheel
01:38:12amiconnG1/g2 aren't released yet
01:38:35amiconnOther targets have problems as well. X5/M5 are only charging slowly in rockbox, and M3 doesn't charge at all
01:38:45amiconnThat's a case of "not yet implemented"
01:39:13amiconnDo you know how much the wheel draws? Wheel poweroff is a necessity
01:39:13n1samiconn: g1/2 are listed on the "release" download page...
01:39:21 Join CaptainKwel [0] (
01:39:54amiconnOh... they weren't in 3.0 for that reason
01:44:23 Quit sarixe ("Bye")
01:45:50n1sI think that if the patch is not going to be committed, it should be closed as rejected
01:49:51 Quit CaptainKewl (Read error: 110 (Connection timed out))
01:52:02 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.6/2009011913]")
01:53:20 Join sarixe [0] (
01:55:57 Join jordoex [0] (
01:58:31 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
02:06:29 Part piotrek2234
02:08:29 Quit tessarakt (Read error: 113 (No route to host))
02:08:32 Join tessarakt [0] (
02:09:18 Join `Assassin [0] (
02:10:13`AssassinHow much better is this firmware than the default firmware for an iPod?
02:10:46`AssassinIs there a possibility of bricking my iPod while installing this?
02:11:03soapoutside using a hammer as part of your install process? No.
02:13:13 Quit gromit` (Read error: 110 (Connection timed out))
02:15:50 Quit moos ("Rockbox rules the DAP world")
02:16:36`AssassinWhat's the most annoying bug (if any) that I might easily experience?
02:19:42n1sthat depends on a) what you do with your ipod b) what you find annoying c) a lot of unknowns...
02:22:17`AssassinDoes the firmware get updates very regularly?
02:22:36 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
02:23:22saratoga5 or 10 times a day typically
02:23:38saratogathough lots of them are only for one or a few devices
02:24:24`AssassinIs my iPod video (30gb) one of those more popularly updated devices?
02:27:18 Quit n1s ()
02:27:28saratogai don't think i've ever bothered to think about that
02:28:31***Saving seen data "./dancer.seen"
02:31:35 Join kugel [0] (n=kugel@rockbox/developer/kugel)
02:39:18 Join markun [50] (n=markun@rockbox/developer/markun)
02:40:23 Quit tessarakt ("Client exiting")
02:41:10 Quit bmbl ("Woah!")
02:46:15 Quit thegeek (Read error: 60 (Operation timed out))
02:48:44 Quit kugel ("ChatZilla 0.9.84-rdmsoft [XULRunner]")
02:49:06 Quit _Auron_ ("Infinity repeatedly denies rumours of plotting with zero to bring down the Universe.")
02:49:38 Join thegeek [0] (
02:50:58 Join ucchan [0] (
02:59:39jordoexhey saratoga, thanks for the info
03:01:51 Join Darksair [0] (n=user@
03:02:18 Join kugel [0] (n=kugel@rockbox/developer/kugel)
03:10:04 Quit jordoex (Connection reset by peer)
03:11:24 Join jordoex [0] (
03:15:24 Quit Darksair ("Everything that has a beginning has an end.")
03:16:34 Join Darksair [0] (n=user@
03:20:24kugelsaratoga: can you enable charging on the fuze? bertrik said the current while charging is needed to be messured (and I don't have the tools for that)
03:25:46 Join FlynDice [0] (
03:27:32kugelFlynDice: hi
03:28:28kugelsetting DBOP_DOUT = 0xf<<12 somwhere before somewhere before /* start read */ helps
03:28:56FlynDiceHey there that was quick... I'll try
03:29:36kugelwell, but you can imagine, that this is visible
03:29:58kugelI haven't worked out how to make it unvisible, or at least fix that to 1pixel without messing up the rest
03:30:34kugelFlynDice: you may try 0xff<<8 if the scrollwheel is still buggy
03:31:18FlynDiceI just got in from work, I'll try that out later tonight. Thanks for keeping me in mind on that!
03:33:56kugelI'm not entirely sure how to set dbop without affecting the display. I hoped it wouldn't be visible when setting it after the /* disable output */ line, but it was still visible
03:35:23 Quit __lifeless (Remote closed the connection)
03:35:39 Join __lifeless [0] (n=lifeless@
03:36:26 Quit kugel ("ChatZilla 0.9.84-rdmsoft [XULRunner]")
03:41:46 Quit Seed ("cu, Andre")
03:49:20 Quit balou (Remote closed the connection)
03:50:54 Join Seed [0] (
03:54:26 Quit jordoex (Read error: 110 (Connection timed out))
03:57:55 Quit Seed ("cu, Andre")
03:59:03 Join balou [0] (
04:05:50 Join Seed [0] (
04:08:12 Join blkhawk- [0] (
04:11:33 Quit Seed ("cu, Andre")
04:12:10 Join Seed [0] (
04:15:39 Join AndyI [0] (i=AndyI@
04:15:43 Join kugel [0] (n=kugel@rockbox/developer/kugel)
04:23:20 Quit |mr (Read error: 110 (Connection timed out))
04:25:01 Quit blkhawk (Read error: 110 (Connection timed out))
04:25:11 Nick blkhawk- is now known as blkhawk (
04:26:05 Quit AndyIL (Read error: 110 (Connection timed out))
04:28:33***Saving seen data "./dancer.seen"
04:30:41 Quit Seed ("cu, Andre")
04:31:03 Quit kugel (Remote closed the connection)
04:39:52 Quit fdinel (Read error: 104 (Connection reset by peer))
04:40:06 Quit midijunkie41 (Read error: 110 (Connection timed out))
04:40:07 Join Seed [0] (
04:40:28 Quit perrikwp (" ajax IRC Client")
04:40:28 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
04:42:53 Join Barahir_ [0] (
04:45:39 Join xSlack [0] (
04:47:00 Quit Darksair ("(define (add-1 n) (lambda (f) (lambda (x) (f ((n f) x)))))")
04:51:58 Join rocko [0] (
04:55:46 Join jordoex [0] (
04:59:51 Quit Barahir (Read error: 113 (No route to host))
05:00:49 Part Aurix_Lexico
05:10:37 Quit Marie ("CGI:IRC (EOF)")
05:25:55 Quit `Assassin (Read error: 104 (Connection reset by peer))
05:29:37 Join `Assassin [0] (
05:36:48`AssassinAre there any plans for a Rockbox application on the iPhone or iPod touch?
05:37:50krazykitnothing is really "planned". interested people do the work needed. that said, no one seems to be working on that
05:41:01 Quit xSlack ("Ex-Chat")
05:50:51 Join oksv [0] (
05:54:12 Part oksv ("Leaving")
06:06:59 Quit Seed ("cu, Andre")
06:07:25 Join Seed [0] (
06:13:15 Join film42 [0] (
06:28:35***Saving seen data "./dancer.seen"
06:39:23*JdGordon fears he has hit a bit of a wall with the wps patch :/
06:40:19 Quit Mouser_X (" ajax IRC Client")
06:40:53JdGordonthe way the sub/lines are stored.. keeping them the way they are is the simplest, but not very nie with the rest of the patch
06:41:12*JdGordon got rid of most of the waste at least
06:43:42 Quit sarixe (Read error: 104 (Connection reset by peer))
06:47:03LloreanTrimmed it down any?
06:48:17JdGordonyes and no... atm its using a static buffer the same size as svn, but its definetly overkill...
06:49:02*JdGordon gets out the pen and paper and does actual planning! :O
06:54:00 Join Verzel [0] (
06:57:51 Quit film42 ()
07:16:11 Quit BHSPitLappy (Remote closed the connection)
07:17:18 Join tchan1 [0] (
07:19:39 Quit tchan (Nick collision from services.)
07:19:43 Nick tchan1 is now known as tchan (
07:28:12saratogai get about a savings of 700kHz using the DMA playback patch
07:28:16 Quit tvelocity[a] (Read error: 110 (Connection timed out))
07:28:27saratogaon PP5024
07:28:38saratogabe interesting to see if anyone else can replicate my results
07:31:18 Quit Verzel ("Leaving")
07:31:35 Join sarixe [0] (
07:34:58 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
07:35:58 Quit sarixe (Client Quit)
07:36:14 Join sarixe [0] (
07:47:03 Join Darksair [0] (n=user@
07:51:03 Join radec [0] (
07:56:06 Join jessemallory [0] (
07:56:11jessemalloryHey guys
07:56:50jessemalloryAnyone around?
07:58:53 Quit fyrestorm (Read error: 54 (Connection reset by peer))
08:10:02 Quit jessemallory ("ChatZilla 0.9.84 [Firefox 3.0.6/2009011913]")
08:11:24 Join EspeonEefi [0] (i=eefi@SAFFRONCITY.MIT.EDU)
08:12:18 Quit radec (Read error: 110 (Connection timed out))
08:18:20 Part toffe82
08:23:17 Quit Llorean (Read error: 104 (Connection reset by peer))
08:23:39 Join Llorean [0] (
08:28:38***Saving seen data "./dancer.seen"
08:35:39 Quit ucchan ("Leaving...")
08:41:41 Join Rob2222 [0] (
08:41:50JdGordonwhat should I be using for printf for size_T ?
08:41:54JdGordonsize_t even
08:49:42JdGordonoh FFS... the wps debug info is displayed before the images are loaded.... no wonder I'm getting such small values :(
08:50:38JdGordoncabbiev2 on e200 uses about 15KB which sounds more normal than 940 bytes :p
08:51:33LloreanStill probably a lot less than we reserve.
09:00:11 Quit Rob2223 (Read error: 110 (Connection timed out))
09:10:07JdGordonoh yeah... much less
09:18:54 Quit timc ("Leaving")
09:39:45 Join n1s [0] (n=nils@rockbox/developer/n1s)
09:41:40 Quit JdGordon (Remote closed the connection)
09:43:38 Join JdGordon [0] (
09:43:49 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
09:45:33 Quit CaptainKwel (Read error: 110 (Connection timed out))
09:47:44 Join flydutch [0] (
10:23:21 Quit JdGordon (Read error: 104 (Connection reset by peer))
10:25:16 Quit Darksair ("Everything that has a beginning has an end.")
10:26:16 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
10:27:07 Quit nuonguy ("This computer has gone to sleep")
10:27:30amiconnjhMikeS: ping
10:27:54 Nick Beaver`alszik is now known as Beaver (
10:28:39***Saving seen data "./dancer.seen"
10:28:50 Join Deenoo [0] (
10:29:56 Quit z35 ("Leaving")
10:36:47 Join Deenoo1982 [0] (
10:37:18Deenoo1982can anyone help me?
10:41:37advcomp2019Deenoo1982, yes but we can not read your mind
10:53:01 Quit Deenoo (Read error: 110 (Connection timed out))
10:59:04 Quit Deenoo1982 (Read error: 110 (Connection timed out))
11:01:05gibbon_is there any pin.compatiblity between CF and embedded-ATA? in other words.. is it possible to replace an iPod 3rd or 4th generation harddrive with a CF-Card?
11:01:30BigBambiyou need an adaptor
11:01:37BigBambiThere is a wiki page about this
11:04:12gibbon_du you mean CFModGuide?
11:04:31BigBambiThat sounds right, I couldn't remember the name
11:04:33 Join {phoenix} [0] (
11:05:46`AssassinAnyone care to help me with an album artwork problem? I think I've tried everything. I made sure the wps and file name were correct a million times.
11:06:21BigBambiFirst, try using with the default theme, cabbie v2
11:06:28BigBambiAs that definately works
11:06:33gibbon_BigBambi: ah ok... i didn't know what the connector was called... helped me a lot, thanks
11:06:45BigBambigibbon_: No problem
11:07:02`Assassinoh wait I thought the default one was the theme that SAYS default.
11:07:18`AssassinI'll go back and delete my changes to the other default one. lol
11:07:39BigBambiThat's the old default - the name is confusing I admit
11:08:24amiconnPerhaps it should be renamed to bultin or fallback or similar
11:08:33BigBambiamiconn: Yes, I think so
11:08:38gevaerts"rockbox classic" !
11:10:32 Join ender` [0] (
11:10:46`AssassinThanks so much. It's all working now.
11:11:12BigBambiSo if you are writing a new wps, you can copy what is in cabbiev2
11:11:21`AssassinI just assumed all the wps files had no album artwork support by default.
11:11:34BigBambiDespite all the screenshots in the manual?
11:11:46`Assassinlol Who reads that?
11:11:53BigBambiYou, if you want support
11:14:40`Assassin You wouldn't happen to know where the download for this plugin is located, would you?
11:14:55BigBambiIt is included, as with all plugins
11:15:05BigBambiThe manual states this, please read it
11:15:21gibbon_:) the ipod 3G 40G chassis works on the iPod color (without modding)
11:15:30gibbon_lets hope the harddrive does, too
11:15:42`AssassinI thought I just looked for that plugin on my ipod a second ago though >_>
11:16:00gevaertsIt is there...
11:16:58 Join nplus [0] (n=nplus@243.131.Globcom.Net)
11:17:12`AssassinLol, now that I find it, it says "plugin returned error." Awesome, right?
11:17:28BigBambiAnd have you initialised the database?
11:17:43BigBambiRead PluginPictureFlow on the wiki
11:17:44`AssassinI tried again, and now I have a loading bar. :O
11:18:07BigBambiI hope to get picture flow into the manual fairly soon
11:18:13`AssassinI'll reinitialize it.
11:18:23 Quit nplus (Client Quit)
11:18:36 Join bertrik [0] (
11:19:00`AssassinIt returned an error again, but I'm not going to worry until after I reinitialize the database.
11:21:09`AssassinRockbox makes my iPod so accessible. iTunes had a deathgrip on it. :(
11:22:28`AssassinNow PictureFlow is saying "Could not read bmp" over and over. Now it's closed with the same "plugin returned error."
11:23:46`AssassinI think my pictures are too big. :(
11:24:11BigBambiDid you read the page I said on the wiki?
11:24:21BigBambiAs that tells you the limitations
11:24:27`AssassinI've had it open the whole time.
11:24:40`AssassinIt says the maximum size for pictures is the size of the screen.
11:24:47`AssassinI don't know what size my iPod's screen is.
11:25:04`AssassinAll my bitmaps are 600x600 lol
11:25:15BigBambihahaha, that is funny! :)
11:25:28gevaertsIs that still true? Doesn't PF use scale on load these days?
11:25:50BigBambigevaerts: don't know
11:26:00BigBambiUnhelpful probably des
11:27:06BigBambi`Assassin: It is probably best to check with a smaller cover and see if it works
11:27:16BigBambi`Assassin: That way we can narrow down the problem
11:27:32`AssassinThe worst thing about this is that I have waay too many albums with bmps to resize if this is indeed the problem.
11:27:45scorchebatch processing...
11:27:47BigBambi`Assassin: Batch conversion is easy
11:28:05`AssassinI have irfanview. Does that program do batch resizing?
11:28:12scorchethat goes beyond the scope of this channel, but yes it does
11:28:26`AssassinI used it for batch jpg to bmp conversion earlier.
11:28:44BigBambiThen I'd guess it will :)
11:30:14`AssassinOk so 320 is the width and 240 is the height right? It would be terrible to get that confused later. >_>
11:32:53BigBambi`Assassin: Is your screen wider than it is heigh?
11:33:07BigBambiand is 320 a bigger number than 240?
11:33:38`Assassinlol Yup. Sorry I'm just stressing over the details right now...
11:36:15`AssassinOk PictureFlow is loading up again... I hope it works this time.
11:37:16BigBambi`Assassin: You do know picture flow is just a eye candy demo - it doesn't actually do anything?
11:38:14`AssassinLOL. This time it loaded up NOTHING but placeholder pictures with question marks on them.
11:38:42`AssassinOh there we go
11:38:53`AssassinI finally spotted the one album cover I resized.
11:38:57`AssassinWorks great.
11:39:00BigBambi`Assassin: It'll give a placeholder for the others
11:39:16BigBambiAs it displays all your albums, but if it can't find art for them...
11:41:16`AssassinNow here's another problem... I made the pictuere the right size and everything for my iPod but PictureFlow still crops it off at the top. I would consider resizing just for pictureflow but what if a plugin comes along that uses my pictures with a fullscreen view?
11:41:44`AssassinHmm no wait
11:41:56`AssassinIt looks like pictureflow will stretch out the picture regardless.
11:42:13`AssassinI'm not sure.
11:42:43`AssassinNo, I'm looking at the wiki and those pictures don't get cropped. :(
11:43:40`AssassinPictureflow needs updates now :(
11:43:46`AssassinThanks for your help Bambi.
11:43:55BigBambino problem
11:45:18`AssassinHuh that's odd
11:45:52`AssassinFor the first time tonight my computer said "Detecting new hardware. Rockbox media player."
11:46:04`AssassinI wonder why it didn't say that the other hundred times I connected it.
11:47:35`Assassinlol I think it broke because I can't access my iPod's drive now. Time for another reset >_>
11:48:50`AssassinOh, my battery is pretty much dead. That figures.
11:58:31 Join grdxyxy [0] (n=eric@
11:59:16`AssassinIs there a way to attach lyrics to the songs in the way the album artwork is? I know I could just put some text files on my ipod but I really like the idea of hitting a button while the song is on screen playing to make the lyrics appear.
11:59:18scorche`Assassin: what is the purpose of that "hmm"? are being spammy enough without the unnecessary lines...also please try and not use the enter key as much by placing things on one line...
12:00:08BigBambi`Assassin: No, but there is a lyric viewer patch on the tracker. No idea if it is up-to-date or not, and even if it wa you need to compile to use it
12:01:54 Join rakata [0] (i=rakata@gateway/tor/x-ce8f1834113adba2)
12:02:12`AssassinYou know how when you hit the center button on an ipod a couple times while the song is playing and it zooms in on the album artwork to make it take up most of the screen? Is there a feature like that currently available? lol
12:02:33BigBambihahaha, no, lol
12:04:43 Join Lear [0] (n=chatzill@rockbox/developer/lear)
12:07:29JdGordonhow would people feel about losing the ability to load a [r]wps file from the filebrowser?
12:08:01 Quit {phoenix} (Remote closed the connection)
12:08:20BigBambibut still be abe to load a theme from the browser (and by extension a .wps)?
12:09:12JdGordonso it would be the same as the .colours and .icons files
12:10:05BigBambiWhat is the difference between selecting a wps directly, and selecting a cfg that then "selects" the wps?
12:10:13BigBambii.e. why does it need to disappear?
12:12:43JdGordonthere is a problem with FS #9886 (my wps ram reworking) on multi-lcd targets where loading from .wps files end up runnig out ot room because the buffer cant be reinitialised before loading one.. loading a rwps would clear the loaded wps (and vice verca)... the simple fix is to only do this when the .cfg is loaded
12:12:52*JdGordon hopes that makes sense
12:14:16BigBambiSort of :)
12:15:05LearSo you need to reload both WPS:es then. But what if a theme doesn't define a remote WPS? Making it revert to the default in that case could work...
12:15:55JdGordonwhen .cfg's are loaded everything gets reapplied, so even if it doesnt specify a new rwps the previous one will get reloaded anyway
12:18:44 Join tessarakt [0] (
12:18:58LearAnd reapplying everything when loading a wps isn't possible?
12:19:46JdGordonthats alot of extra stuff needing to be done to do that, LOTS of disk access also
12:20:46 Join robin0800 [0] (
12:24:33JdGordonthis of course doesnt have to happen... but it might end up a little easier support wise
12:26:44 Quit __lifeless (Remote closed the connection)
12:27:01 Join __lifeless [0] (n=lifeless@
12:28:44***Saving seen data "./dancer.seen"
12:29:48 Join moos [0] (i=Mustapha@rockbox/staff/moos)
12:39:35 Join Barahir [0] (
12:43:23 Quit `Assassin (Read error: 104 (Connection reset by peer))
12:51:10 Join tvelocity [0] (
12:54:56 Quit Barahir_ (Read error: 110 (Connection timed out))
13:05:16 Join piotrek2234 [0] (
13:10:17 Quit gevaerts (Nick collision from services.)
13:10:26 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
13:18:41 Quit grdxyxy (Connection reset by peer)
13:19:01 Join MethoS [0] (
13:20:07 Quit jon-kha (
13:20:07 Quit Slasheri (
13:22:21 Join MethoS- [0] (
13:24:39 Quit tmzt (Read error: 104 (Connection reset by peer))
13:33:22 Join tmzt [0] (
13:33:59 Quit MrDuck (Read error: 60 (Operation timed out))
13:37:17BigBambiany like to commit a two line change to correct the spelling in Pictureflow? It doesn't seem worth putting on flyspray:
13:37:30 Quit undertakingyou (Remote closed the connection)
13:37:41 Join undertakingyou [0] (
13:38:46 Quit MethoS (Read error: 113 (No route to host))
13:39:13gevaertsBigBambi: aren't those names of settings or something like that?
13:39:33BigBambigevaerts: It is what is displayed in the settings menu
13:39:42BigBambigevaerts: I haven't changed any actual code
13:40:00BigBambibut center is incorrect in UK English, (which rockbox uses)
13:40:31BigBambiand I did test build it :)
13:41:10gevaertsI don't doubt that it builds, but I'm not entirely convinced that some of those strings don't end up in a settings file
13:41:20BigBambiah, I see
13:41:26BigBambiIn that case, whoops
13:41:41BigBambiHowever, it should be changed
13:42:05 Join kugel [0] (n=kugel@rockbox/developer/kugel)
13:42:54BigBambiyes, forget that patch
13:43:00BigBambiI failed in my searching
13:43:13gevaertsThat set_int() is definitely settings code
13:43:23*BigBambi sees that now :(
13:43:38*gevaerts agrees that it should be changed, but he isn't sure if and how we should provide backward compatibility
13:44:37BigBambiIs there a way to just change what is displayed and leave the rest?
13:44:53BigBambii.e. it doesn't matter much in a settings file
13:45:05*BigBambi sticks to the manual
13:45:06kugelJdGordon: I'd hate to lose the ability to select wpses in the filebrowser
13:45:36gevaertsI *think* that if you just change the MENUITEM_STRINGLIST and not the set_int(), it will do that. I'm not too familiar with this though
13:45:44BigBambiI'll try
13:46:09kugelBigBambi: I think your patch is correct, isn't it?
13:46:20BigBambikugel: I think not now
13:46:30BigBambikugel: See what gevaerts said above
13:46:58kugelthe first string is the one shown in the main settings menu (before clicking center margin)
13:47:00BigBambikugel: Either way, I don't know and am going to leave it to those that do :)
13:47:15kugelthe 2nd string is the title of the int menu (_after_ clicking center margin)
13:47:24BigBambiin that case, great
13:47:53gevaertsthe set_int one?
13:48:03BigBambiThat is what I thought, but I bow to greater minds :)
13:48:07kugelgevaerts: yea
13:48:24*gevaerts decides that he doesn't know
13:48:33kugelit's a menu, which has a title, and many many int values to select 1 from
13:49:15BigBambikugel, well if you are sure I'd appreciate the change
13:50:49 Quit MethoS- (Remote closed the connection)
13:51:20 Join MethoS [0] (
13:52:06kugelgevaerts: line 220+ is what you'll see in the pf config file
13:52:42gevaertsah ok. That makes sense
13:53:45kugelbut should that changed too? I would think so, requires invalidating existing configs though
13:53:48 Join stripwax [0] (
13:54:21 Quit stripwax (Client Quit)
13:54:41BigBambiIt isn't uer visable though (normally)
13:54:51BigBambiand thus not so vital
13:54:55gevaertsAnyway that means tha the original patch is definitely fine. I'll commit it
13:55:05BigBambimaybe not worth breaking configs for
13:55:07BigBambigevaerts: thanks
13:55:14kugelgevaerts: ok!
13:55:57 Join stripwax [0] (
13:57:20Unhelpfulgevaerts: pf does scale on load, unless you disable that feature, mostly because screen-sized cover can look good in WPS on some targets, but will look bad in pf.
13:57:54gevaertsUnhelpful: I'm now starting to think that this was a 3.1 vs current issue
13:58:18BigBambiUnhelpful: So the "it musn't be bigger than scren size" is no longer valid?
14:00:19UnhelpfulBigBambi: we can scale down from well about screen size on most targets, but the scaler mostly favors accuracy over speed, so there's a cost. many targets can probably scale from screen size to 1/2 screen size in under a second, but when you talk about starting from, say, twice screen size, it gets more expensive.
14:00:33 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
14:00:53BigBambiUnhelpful: OK. I'm just writing the manual entry for it - perhaps I should put a note in about that
14:01:30 Join MethoS- [0] (
14:01:30 Quit MethoS- (Remote closed the connection)
14:01:32BigBambiSomething like "Picture Flow will accept album art larger than the dimensions of the screen, but the bigger it is the longer it will take to scale"
14:01:53 Quit rakata (Remote closed the connection)
14:02:06Unhelpfulindeed. if you like waiting at that splash screen, by all means, put 1024x1024 album art on your device, it will work fine.
14:02:18BigBambiOK, cool
14:04:27BigBambiUnhelpful: I asume the preferring cover.100x100.bmp is no longer true either, it scales to the size you select in Zoom
14:06:10bertrikthere are plenty of other places where center is used instead of centre
14:06:20BigBambiUser visable?
14:06:22Unhelpfulnot precisely. it scales to a size determined based on display size, and it doesn't check for a specially labeled file first. zoom doesn't change the display size, it changes the distance at which slides are rendered from the camera.
14:06:34BigBambiUnhelpful: OK, thanks
14:07:53Unhelpfulthe "center margin" is the distance, in screen pixels, with zoom at 100, between the center and side slides. the margin zooms, too, so at zoom=200 you're doubling the margin.
14:08:15bertrikBigBambi, in jewels, blackjack, sliding puzzle, sokoban, robotfindskitten, snake2, pictureflow, pegbox, star, equaliser
14:08:25 Quit kugel (Remote closed the connection)
14:08:27Unhelpfulslide spacing is the distance between the front edge of the side slides, and that scales with zoom, as well.
14:08:27BigBambibertrik: OK, I'll look at those in a bit
14:08:39bertriklang.c has "Center frequency" for example
14:08:51BigBambiyes, that should be changed too
14:09:36 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
14:09:53 Join MrDuck [0] (
14:11:29 Join MethoS- [0] (
14:14:10 Join kachna|lappy [0] (
14:14:44 Quit MethoS (Read error: 110 (Connection timed out))
14:16:49 Join kugel [0] (n=kugel@rockbox/developer/kugel)
14:17:43JdGordonanyone feel like doing some patch review on 9886? I'd like to start working on getting it commitable
14:20:55 Join MethoS-- [0] (
14:21:26BigBambiUnhelpful: Is rebuild cache needed when you add or remove albums/art, or just if you change a cover?
14:22:27Unhelpfuli think we still have some issues with the cache being "wrong" if albums are added or removed, and in any case, it won't pick up covers for new albums without a rebuild.
14:22:36BigBambiOK, thanks
14:22:51Unhelpfuli want to get rid of the cache build on startup at some point, but that's definitely in the future.
14:23:07 Nick fxb__ is now known as fxb (
14:23:41BigBambiUnhelpful: And last question for the minute - is number of slides number of each side, or total?
14:24:41Unhelpfulyou'll love the answer, i can't think how to state it both briefly and clearly. it's the number of slides on a side, including the center.
14:25:05BigBambiI could just put total number of slides shown
14:25:33Unhelpfulit's not that, though... the total slides shown is that number, times two, plus one :/
14:26:00BigBambiOK, so it is the number on each side?
14:26:18BigBambii.e. if it is set to 4, there are 4 on the left, then the centre one, then 4 on the left?
14:26:41Unhelpfulno, the center is included, so set to four means three on each side.
14:27:03BigBambiSo set to 4, it is 3 then centre then 3?
14:27:26BigBambiSo total slides are that number - 1
14:27:36BigBambi*2 times that number - 1
14:28:03 Quit MrDuck (Read error: 110 (Connection timed out))
14:28:47***Saving seen data "./dancer.seen"
14:30:37Unhelpfulmight be more sensible to change it to "number of slides on each side"
14:30:54 Join jaykay [0] (
14:30:58 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
14:31:10BigBambi I was thinking that :)
14:31:47JdGordonNico_P: hey, got a min (some time) to checkout my abomination of the wps code? (fs#9886)
14:31:55BigBambiUnhelpful: If I right the manual as number of slides on each side, could you change the plugin? Or will that cock up previously saved configs?
14:32:01BigBambier, s/right/write/
14:32:11jaykayjhmikes: would a battery bench with FS #9910 with an e200 be helpful?
14:32:17 Join jon-kha [0] (
14:32:20Nico_PJdGordon: ok :)
14:32:42Unhelpfulif i change what the value in config means, yes, it will change the meaning of previously saved configs. it's not like we've not done that, or forced reset to defaults, before :)
14:32:51BigBambiUnhelpful: true :)
14:33:02BigBambiI can write it up as is, it isn't too bad with an example
14:33:23BigBambiI doubt many people will notice that setting changing however
14:33:35BigBambiby 1 anyway
14:33:40Nico_PJdGordon: you're getting awefully close to malloc ;)
14:33:43 Join Slasheri [0] (i=miipekk@rockbox/developer/Slasheri)
14:34:00 Quit MethoS- (Connection timed out)
14:34:13JdGordonno im not :)
14:34:20JdGordonno closer than we already are
14:34:28UnhelpfulNico_P: speaking of which, right now, metadata, including the actual bitmap image for album art, are just tacked on the front of the file in buffer, yes?
14:37:18Unhelpfulwhat about bufalloc'ing the image data, and putting the handle in the metadata? are there obvious problems with this? it would make it possible to change the image size without a rebuffer, if we can keep bufalloc from failing (which is the next question)
14:38:47 Quit Rob2222 (Read error: 104 (Connection reset by peer))
14:38:53 Join Rob2222 [0] (
14:39:31 Join Aurix_Lexico [0] (
14:40:24Nico_PUnhelpful: would the image data still be stored before the audio data? I don't see how what you propose would allow changing the image size without rebuffering
14:42:04Nico_PJdGordon: is the tokens list always read in order?
14:42:16 Join einhirn [0] (
14:42:44JdGordonwhy? or?
14:42:56Unhelpfulno, only the handle for the image data would be stored with the metadata. to change album art size, you'd walk through the buffered tracks, bufclose their AA handle, bufalloc new ones, and reload load the image data into those.
14:43:26Nico_PJdGordon: because if it is, it's probably no more expensive than an array. but how do you handle stuff like conditionals?
14:43:40amiconnUnhelpful: I just tried test_greylib_bitmap_scale on my recorder v1. Unlike on pixelma's Ondio, it doesn't crash (even with memory access check enabled), but it doesn't work properly either
14:43:40BigBambiWhere are plugin actions defined for each target? i.e. where do
14:43:41Nico_P(expensive in time)
14:43:57Unhelpfulamiconn: doesn't crash... does it work? :)
14:43:58amiconnIt just shows a black square of the size the image should have
14:44:00BigBambiI look up which actual button PLA_MENU refers to?
14:44:35pixelmapluginlib something directory
14:44:57JdGordonNico_P: yeah, right now tokens are stored in a static array, same as sub/lines... I'm not entirely sure how to move them to a list.. they dont really need to be moved but it could make things smaller (RAM-wise)
14:45:05JdGordonand artificial limits are always bad
14:45:11Unhelpfulhrm... maybe i need to test the >>24 scaling math on sim, which is not what's happening now, since the scaler math is determined by CPU_*
14:45:49Nico_PUnhelpful: I think that can already be done. the AA handle ID isn't stored as the metadata but as part of struct track_info
14:46:11Unhelpfulthe AA is already behind a handle?
14:46:34Nico_PJdGordon: oh, I saw struct wpsll and assumed you had converted the array to a linked list
14:46:40Nico_PUnhelpful: yes
14:47:58pixelmaBigBambi: sorry apps/plugins/lib/pluginlib_actions.c
14:48:14BigBambipixelma: Thanks very much
14:48:49kugelpixelma: have you had time to test the patch I uploaded to ?
14:49:01 Quit rocko ("Leaving")
14:50:26Nico_PJdGordon: right now, bitmaps, strings and viewports are in linked lists, right?
14:50:53pixelmayes, didn't comment on it yet. With my test directory the crash is really gone and I don't see CPU staying boosted either but I remember there were similar bugs in the past
14:51:01pixelmakugel ^
14:52:32 Join denes [0] (
14:53:51pixelmaBigBambi: have fun with the mixed contexts though :\ e.g. controls are slightly broken on c200 currently anyway (you can't move down in the track list if you selected an album, as "down" is used to call the menu)
14:54:05BigBambipixelma: nice
14:54:18kugelpixelma: you mean that it keeps boosting on my setup could be another (possibly unrelated) bug?
14:55:43Unhelpfulis there an easy way to figure out if we're building sim for a SH device?
14:55:54Nico_PJdGordon: I sure like the idea. I don't think converting the token array to a ll is worth the added 4 bytes per token
14:56:59JdGordonNico_P: well think of it this way.. currently its using 4KB regardless of how many tokens are used... changing it could use much less (or more for really really really complex WPS') its more flexible
14:58:16pixelmakugel: not sure at all. It's not easy because as you experienced before, slight changes in the settings/available RAM whatever can make the crash disappear...
14:58:26rasherUnhelpful: where?
14:59:09Nico_PJdGordon: I think the variability in the token array size is negligible compared to the total size the WPS buffer will be
14:59:17Unhelpfulrasher: in the source. sh CPUs use different scaler math, and i want the same math used on the sims for those models.
14:59:51Nico_PJdGordon: even if you make it big enough to accomodate for a really complex WPS, it's still going to be nothing compared to the space needed for bitmaps
14:59:56pixelmakugel: btw. I tried your patch against a revision I got the crash with before
15:00:08JdGordonNico_P: well, really the only tokens which use any space are the bitmap ones
15:00:14kugelso I did ;)
15:00:32Nico_PJdGordon: why is that?
15:01:22JdGordonviewports use 30 odd bytes each, strings are less than that (avg of course), images are the big buffer users
15:03:23kugelpixelma: I'm just curious what gareth reported there
15:03:40Nico_Pfor viewports and strings I think the linked list makes sense because there are less of them than tokens and their size is (or can be) far more than the cost of the pointer. For tokens adding the pointer is doubling the cost of each token and there are usually a lot of them
15:04:06kugelHe said he crashes directly when the buffers are full; if that's delayed that probably something entirely different
15:04:17Nico_Palso the linked list structure might complicate the rendering
15:04:31BigBambigrrr, can't find my e200. In pictureflow for a scroll wheel target I assume that the scroll wheel goes up and down the track list
15:05:15JdGordonok, although I'm not so sure the current line mechanism is very efficient... maybe it could be fixed up using lists... but I'm also happy to call it "done" as it is
15:05:28 Join Horschti [0] (n=Horscht@xbmc/user/horscht)
15:05:49kugelNico_P: any idea for ?
15:05:51Nico_PI agree there are probably some improvements to be made
15:06:34Nico_PJdGordon: have you started collecting data on the actual buffer size for some WPSs?
15:07:03JdGordononly cabbiev2 and a random image-intensive one i found on the wiki
15:07:16Unhelpfulamiconn: bad news, making the default SC_SHIFT 24 still leaves the plugin working on sim. how big is the bitmap you're testing?
15:07:22JdGordoncv2 on e200 is 15KB and 18KB on h300 (both displays together)
15:07:38Nico_Pkugel: that's an ugly one... I'd like to try debugging it but I'm a bit short on time
15:07:46amiconnUnhelpful: I tried both 200x200 and 112x64
15:08:30Nico_Pkugel: have you tried it in the sim?
15:08:48kugelNico_P: it's not reproducable in the sim
15:09:14 Quit stripwax ("Miranda IM! Smaller, Faster, Easier.")
15:09:40amiconnUnhelpful: Here even an unmodified sim behaves different. It opens the file and immediately exits again
15:09:46kugelNico_P: well, it seems to be an alignment issue on ARM, as it the patch I attached fixes it for me and pixelma. However, after the buffering of the very album it keeps boosting for me (not for pixelma). Also, I've no idea how to understand the latest comment by gareth
15:10:13amiconnHmm, that seems to be a button/action problem
15:10:48amiconnIt only happens when using Right to open the plugin from the 'Open with...' context menu, not when using Play
15:11:07pixelmakugel: unfortunately you can't be 100% sure if the patch really fixes it because of what it does - it could also fix it because of changing binary size whatever
15:11:58amiconnHmm, now I got "failed to load bitmap" once.... but it worked when I tried again
15:12:28kugelpixelma: the patch adds maybe 8 or 12 bytes. the binsize of the sansas has changed more since the task is opened
15:12:42kugel8 or 16*
15:13:07kugelwell, shouldn't be more than 8
15:14:08amiconnThe button problem is obvious, and easily fixable. I wonder about the "failed to load bitmap" thing....
15:14:39 Quit piotrek2234 (Connection timed out)
15:14:45kugelNico_P: any idea why it doesn't unboost due to this change, or if this is unrelated?
15:15:02JdGordonNico_P: also, what do you think about moving the 4 wps_* files into apps/gui/wps/ and splitting wps_parser and gwps.h up into multiple smaller files?
15:15:35JdGordonand moving all the AA related vars in the wps_data struct into a seperate wps_aa struct?
15:15:55Nico_Pkugel: I'm having trouble believing it's related, but you never know
15:16:11kugelSo I do, actually
15:16:26Nico_PJdGordon: why not. it would certainly make things cleaner
15:17:10 Join MethoS [0] (
15:18:12kugelNico_P: Toni just added a comment.
15:21:23kugelI'm going to reboot, then try that out
15:21:25 Quit kugel ("ChatZilla 0.9.84-rdmsoft [XULRunner]")
15:23:21 Quit Horscht (Read error: 110 (Connection timed out))
15:24:48 Join kugel [0] (n=kugel@rockbox/developer/kugel)
15:24:56BigBambi{PLA_QUIT, BUTTON_MENU|BUTTON_SELECT, BUTTON_NONE}, is saying menu + select together right?
15:25:58BigBambiI was sure it was, it just seemed a slightly odd combo so wanted to check :)
15:26:45pixelmait's not an odd combo for leaving plugins on Ipods I believe...
15:27:17BigBambiOK, that's cool
15:28:12amiconnIt's the standard combo if plain Menu is already taken
15:29:52 Quit denes ("bye")
15:32:35 Quit MethoS-- (Read error: 113 (No route to host))
15:34:23amiconnUnhelpful: If I change SC_SHIFT to 24, I *do* get the black rectangle in the sim...
15:35:22 Quit tessarakt ("Client exiting")
15:38:51*amiconn has an idea
15:41:42 Join MethoS- [0] (
15:42:25 Join miepchen^schlaf [0] (
15:42:44 Quit miepchen^schlaf (Remote closed the connection)
15:45:09 Join tessarakt [0] (
15:47:36amiconnHmpf, can't test this idea right now
15:47:57amiconnUnhelpful: Are you building your sims on a 64 bit host?
15:50:24 Quit SirFunk (Read error: 110 (Connection timed out))
15:50:53 Join SirFunk [0] (
15:51:19 Quit sarixe ("Leaving")
15:53:36 Quit kugel (Nick collision from services.)
15:53:39 Join kugel_ [0] (n=kugel@
15:53:49 Nick kugel_ is now known as kugel (n=kugel@
15:53:58 Quit tessarakt ("Client exiting")
15:54:16 Join midijunkie [0] (
15:54:20 Join Buschel [0] (
15:54:36 Join Guest83079 [0] (
15:55:44*Buschel just reached 19:30h runtime on his iPod 5.5G 30GB (stock battery)
15:56:08 Join rocko [0] (
15:56:12kugelBuschel: with the lcd enable patch?
15:56:16BuschelI am impressed! With fs#9890 this was
15:56:26kugelbtw, what do you think about the points I mentioned?
15:57:17Buschelthe issues about the consistency? those are good points. the introductionmof some special behaviour should not be needed
15:57:34Buschelhmm, too many typos...
15:58:08 Quit MethoS (Read error: 110 (Connection timed out))
15:58:32amiconnUnhelpful: ping
15:58:32n1sBuschel: how does that compare to the OF?
15:58:41 Quit shadearg (Read error: 113 (No route to host))
15:59:29Buscheln1s: I never used the OF, without this patch I have reached ~15h
16:00:15 Nick JdGordon is now known as JdGordon|zzz (n=jonno@rockbox/developer/JdGordon)
16:05:19n1sBuschel: wow, that is an impressive improvement
16:08:40Buscheln1s: yep, I am just writing a short report to flyspray right now
16:09:13 Join SirFunk_ [0] (
16:22:09 Join MethoS-- [0] (
16:22:27 Quit balou (Read error: 104 (Connection reset by peer))
16:22:32 Join balou [0] (
16:24:11 Quit SirFunk (Read error: 110 (Connection timed out))
16:24:33 Join Mortiffer [0] (
16:24:59MortifferAny one know what everything from is downloading at like 4kBs ?
16:25:32 Join mipfi [0] (
16:25:39Mortifferor does any one have a mirror for this stuff
16:26:52MortifferHerro ?
16:28:51***Saving seen data "./dancer.seen"
16:29:09 Quit krazykit ("brb")
16:30:44Unhelpfulamiconn: i'm using a 32-bit host. i wonder if i have a deps issue, or something...
16:32:24Unhelpfuli think i got it, actually, i didn't do make install, and the scaler is in pluginlib ;)
16:33:39 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
16:33:49 Part Guest83079 ("Leaving")
16:34:22Unhelpfulok, yes, i'm getting it here, as well, now. i'll have to figure out what, exactly, is wrong with the 24-bit scaler math... or else come up with an efficient 40-bit multiply for sh.
16:34:58 Join krazykit [0] (n=kkit@
16:38:25 Quit MethoS- (Read error: 110 (Connection timed out))
16:39:26 Join morloi [0] (n=x1jmp@
16:42:41n1sBigBambi: yes, some kind of automated screenshot feature for the sim would be really nice
16:43:24BigBambiFor reference for others, I was bitching about taking pictureflow screenshots :)
16:44:11*Unhelpful will wait until you're done to start working on PF rotation :)
16:44:33*BigBambi hits Unhelpful
16:46:54 Join fdinel [0] (
16:48:23 Join bluebrother [0] (i=5ce0ec25@gateway/web/ajax/
16:49:04 Join saratoga_home [0] (n=41becb3b@gateway/web/cgi-irc/
16:51:00Unhelpfulamiconn: got it, grey_draw.c:output_row_grey was using a straight multiply + >>32, instead of the SC_MUL macro. committing :)
16:51:36 Quit robin0800 (Read error: 104 (Connection reset by peer))
16:51:56 Quit kugel ("ChatZilla 0.9.84-rdmsoft [XULRunner]")
16:52:06 Join robin0800 [0] (
16:53:18saratoga_homecould someone take a quick look at the iram_malloc function in FS #9882 - Speed-up mod for large block (Vorbis)
16:53:28saratoga_homei don't see how that doesn't overwrite other stuff in IRAM
16:54:30 Join kugel [0] (n=kugel@rockbox/developer/kugel)
16:57:01soapBuschel, re: FS 9708. Does one need to use BOTH pp5020-dma-v0.3 AND align_buffer-v0.3?
16:58:08Learsaratoga_home: It uses a specific malloc buffer, so it should be fine, I think...
16:59:06saratoga_homeLear: but isn't the buffer the entire 48k of IRAM? so on first alloc whatever was first in IRAM is overwritten?
17:00:09Learsaratoga_home: No, it is using iram_buff, which is declared in oggmalloc.c.
17:00:11 Quit SirFunk_ (Read error: 60 (Operation timed out))
17:00:31saratoga_homeLear: static unsigned char iram_buff[IRAM_IBSS_SIZE] IBSS_ATTR LINE_ATTR;
17:00:40Buschelsoap: I used both. Boris mentioned that this rises the probability of data being loaded via dma. But from my tests on the dma-patch I could not see a difference... In fact I already had a patched version with both −− so, I used this basis.
17:00:41saratoga_homethats 48k on most systems, no?
17:01:21LearNo, see config-tremor.h, which defines IRAM_IBSS_SIZE.
17:01:30soapthank you Buschel
17:01:53saratoga_homeLear: thanks, assumed that was a rockbox define
17:02:08Mortifferi guess i'm kinda late but did any one address the slow downloads ?
17:02:50saratoga_homestill that seems kind of dangerous, because anyone allocating to IRAM will have to update that define or else corrupt memory
17:03:10 Join miepchen^schlaf [0] (
17:03:42soapif you got 19h on a 30GB, that = >25 hours on a 60/80GB. Disgusting
17:03:53 Join SirFunk [0] (
17:03:53saratoga_homei wonder if we could just malloc from the back of IRAM since that wouldn't overlap unless someone tried to use more iram then we had
17:03:59LearBut it is only within that codec, so it shouldn't be a big problem, right?
17:04:52Buschelbtw, as there seem to be some progress on PP502x DMA: is there a chance somebody with deeper knowledge of ata.c can check fs#9703 (ATA UDMA for PP502x)?
17:04:55saratoga_homeLear: yeah its only in vorbis, but its not a nice thing to do to anyone else who might want to tweak it
17:06:21Buschelsoap: yes, the saving when disabling the LCD is massive. nevertheless you'll need to also use 24MHz normal clock −− otherwise you have too high power consumption for high efficiency codecs (flac, mp3, mpc)
17:06:30saratoga_homeBuschel: 9708 right?
17:06:50 Join jaykay_ [0] (
17:06:56Buschelsaratoga_home: ooops, typo -> yes, fs#9708
17:07:02saratoga_homeI don't get why the LCD uses so much power, is it by design or do we not init the broadcom chip correctly
17:07:47LearThe linker will still complain if someone uses too much iram. So isn't the biggest problem that it's a bit uncertain if the windowing tables fit in iram?
17:08:19saratoga_homeLear: the problem is that different tables are used for different files
17:08:28soaphow is list scrolling performance with that 24Mhz clock, Buschel?
17:08:34saratoga_homeso this patch attempts to make it dynamic what gets allocated in IRAM
17:08:55Buschelsaratoga_home: the savings are >6mA. maybe there is some kind of eco mode for the BCM?
17:09:33Learsaratoga_home: Yes, and if the table in question fits in iram, it will be copied there.
17:09:52saratoga_homeBuschel: thats what I was thinking, 6mA is an enormous amount of power for LCD
17:09:53Buschelsoap: bad. you'll need to also add a gui boost patch (e.g. fs#8668)
17:10:27 Quit jaykay (Read error: 110 (Connection timed out))
17:10:28amiconnUnhelpful: Do you think it would be better to use 32 bit precision on SH1 as well?
17:10:33 Nick jaykay_ is now known as jaykay (
17:10:38saratoga_homeLear: I just mean that it only gets copied if the define is correct, so if someone else adds something to IRAM, they'll corrupt tremor's memory without any warning
17:10:48amiconnA 32x32->64 multiply wouldn't be much slower than the 32x32->32 bit used now
17:11:02saratoga_homeBuschel: or we could bug Zagor to finish his boosting patch
17:11:20 Join CaptainKewl [0] (
17:12:02Unhelpfulamiconn: it would be more precise, and more consistent with other targets, but even at 400 input pixels error should be around 1 grey level
17:12:04Learsaratoga_home: No, that shouldn't happen. The iram buffer is allocated as any other data. Try to allocate too much and the linker complains.
17:12:41amiconnUnhelpful: Isn't the bmp size limit lowered as well by using 24 instead of 32 bit precision?
17:13:09saratoga_homeLear: as I understand it, the buffer is allocated after the last static data in IRAM based on the value of IRAM_IBSS_SIZE, if you allocate something else to IRAM statically, but don't know to update that define, you overwrite your data, right?
17:13:35 Part LinusN
17:13:57mipfihe it get the error this is a macpod but i installed a new firmware with windows... filesystem ist fat32
17:14:22amiconnBuschel: Imo we shouldn't disable the lcd on targets with a transflective display, even if it saves a bit of power
17:14:22Unhelpfulnot much, actually, the scaler will overflow at (2**32)/255, which places the effective limit at barely over 2**24 pixels, anyway.
17:15:58saratoga_homeamiconn: the thing is it seems to save an enormous amount of power
17:16:07Buschelamiconn: well, disabling BCM and LCD saves a vast amount of power (~1/4 of total current consumption). the patch allows to configure between normal behavior (keeping LCD on, but shutting off the backlight) and power-saving (shutting off LCD and BCM)
17:16:50Unhelpfulit's really only a reduction of precision, which will begin showing around 2**16 pixels of input, when the reciprocal starts to only have 8 bits left.
17:17:01Buschelamiconn: for example I am mostly using the ipod when going to bed. so, I do not need the LCD −− I am just listening.
17:17:01amiconnIt may be that the bcm is busy polling if there is nothing to do
17:17:08kugelthat's why I proposed to not do lcd_enable (so to keep it translective), but only implementing lcd_sleep
17:17:41amiconnUnhelpful: 2^16 pixels isn't much - that's just 256*256 for square images
17:17:41Learsaratoga_home: No, irambuff is just bss data that ends up in a particular linker section. In that sense, it is just like the ipcm_buff that is there now. The patch basically removes the most common windowing tables from iram, then makes ipcm_buff bigger and allocates the pcm buffer dynamically from it. Windowing tables are copied there if they fit, otherwise they stay in normal ram.
17:17:42kugel(and lcd_sleep will turn everything off, but is configurable and delayed after backlight off)
17:18:11 Join tvelocity[a] [0] (
17:19:04amiconnUnhelpful: I could whip up the required sh1 asm routine for SC_SHIFT == 32, but the function should be non-inline. Shall I put it in resize.c?
17:19:39saratoga_homeLear: maybe I don't understand. what happens if IRAM_IBSS_SIZE ends up being wrong ?
17:21:49Buschelkugel: I am not into the lcd/backlight framework. is it possible to put the BCM on/off into the lcd_sleep-functionality? so, that there is no need for the additional lcd_enable?
17:23:09Learsaratoga_home: If made much too small, iram_malloc will fail and the codec will fail (or more likely crash). If a little too small, the common case windowing table won't be copied, which would make it slower. If too big, linker error.
17:23:56Lear(the iram_malloc failure I mean is the allocation of the pcm buffers)
17:24:26 Quit goffa (Read error: 104 (Connection reset by peer))
17:26:08Unhelpfulamiconn: if you want to. i didn't think i stood much chance of doing it well in sh1 asm, without a device to test it on, or much knowledge of the language. if 32x32->32 is a function call *anyway*, i would imagine a 32x32->40 could be done quite cheaply, still.
17:26:47amiconnWell 32*32->32 needs 3 mulu instructions, 32*32->64 needs 4
17:27:00kugelBuschel: yes
17:27:22saratoga_homeLear: ah I see
17:27:30saratoga_homenot sure what that was so difficult to grasp
17:27:37kugellcd_sleep turns everything off, that can also mean it does the same as lcd_enable + more
17:27:52amiconnUnhelpful: The fix works on target and sim here
17:28:18 Join jaykay_ [0] (
17:29:47Buschelkugel: so, that's the main change this patch needs to get near to a committable state
17:32:20soapBuschel, little bit confused = #8668 is _both_ 24MHz clock AND boost on button?
17:32:27 Join stoffel [0] (
17:34:29Buschelsoap: yes, use v5.
17:34:30soapand firmware/target/arm/system-pp502x.c kicks boost to 100, yet the #define in firmware/target/arm/system-target.h is still 80? Or does the #define not trigger anything else we care about?
17:34:48soap(not challenging the patch - just trying to understand)
17:34:52Buschelsoap: the patch uses 24/80Mhz.
17:35:03 Quit tvelocity (Connection timed out)
17:35:08 Join nuonguy [0] (
17:35:16soapahh, the stuff in system-pp502x.c is commented out is it not?
17:35:46Buschelsoap: yes, the settings for 100MHz are commented out in system-pp502x.c
17:37:08Buschelsoap: if you want to use them, you will also need to set CPUFREQ_MAX to 100000000 in system-target.h −− the runtime will increase with lowering the default clock to 24MHz. so, changing the max clock to 100MHz is not needed.
17:37:37 Quit jaykay (Read error: 110 (Connection timed out))
17:37:40Unhelpfulhrm, i also think i can see how we might do full-precision reciprocals instead of 31-bit. instead of (((1<<31)+(d/2))/d)<<1, we could use (((1<<32)-1)-(d/2)+1)/d+1
17:37:42Buschel(nevertheless my iPod runs @100MHz for >1 year now)
17:39:11 Join {-phoenix-} [0] (
17:39:14soapJust that all your comments on #9708 appear to be measuring the DMA performance @ 24/100. That's why I was curious about the other patch being at 24/80.
17:40:22Buschelsoap: I never posted the 24/100-patch. maybe I should update the patch with this option...
17:40:53Buschelsoap: buffering is a bit faster with 24/100 compared to 24/80
17:40:57 Join jaykay [0] (
17:41:43soapBuschel, it's an easy change - I was just curious about the "disparity" - that's all.
17:42:16 Quit midijunkie (Read error: 104 (Connection reset by peer))
17:42:24Buschelsoap: ok
17:42:48 Join robin0800_ [0] (
17:43:44 Quit robin0800 (Read error: 104 (Connection reset by peer))
17:43:51 Join merbanan [0] (
17:43:56jaykaywhat are the disadvantages of setting the normal cpu-clock lower than 30mhz?
17:44:06jaykaythere are some patches changing that, without any disadvantages
17:44:18saratoga_homethe GUI gets slower
17:44:33saratoga_homefor now at least
17:44:56 Quit __lifeless (Read error: 113 (No route to host))
17:45:15 Join goffa [0] (n=goffa@
17:45:41jaykayi didnt see that with the zero-wait-boost... means with 24mhz
17:46:27 Join faemir [0] (
17:47:38 Quit robin0800_ (Read error: 104 (Connection reset by peer))
17:52:06*amiconn noticed that the 32*32->64 bit SH1 multiply already exists...
17:52:27amiconnIn mandelbrot.c - just that is signed, and I need to leave out some stuff
17:53:55amiconn20 instructions - maybe that's even small enough for inlining
17:54:07Unhelpfuli would, there's really only one caller.
17:57:26 Quit jaykay_ (Read error: 110 (Connection timed out))
17:59:17Unhelpfulthe main reason the definition itself isn't inlined in output_row_native is so that it's accessible to plugins for use in custom output plugins
18:00:36amiconnThe plugin becomes smaller by inlining
18:01:34amiconnIt works :)
18:02:20 Quit Nico_P (Remote closed the connection)
18:05:41 Quit rocko ("Leaving")
18:06:35 Join rocko [0] (
18:08:13 Quit moos ("Rockbox rules the DAP world")
18:08:30soapBuschel, last question.
18:09:14soapDo I also want to uncomment DEV_TIMING1 = 0x00000a0a in system-pp502x.c and comment out DEV_TIMING1 = 0x00000808?
18:09:33Buschelyes, you want ;-)
18:11:16 Join jaykay_ [0] (
18:13:21 Join nibbler_ [0] (
18:13:42 Quit kugel (Nick collision from services.)
18:13:45 Join kugel_ [0] (
18:13:55 Nick kugel_ is now known as kugel (
18:15:25 Quit jaykay (Connection timed out)
18:15:35 Nick jaykay_ is now known as jaykay (
18:17:09 Quit saratoga_home ("CGI:IRC (EOF)")
18:20:20soapand the sea of green suffers a red tide ;)
18:25:08 Quit mcuelenaere (Read error: 113 (No route to host))
18:28:52***Saving seen data "./dancer.seen"
18:29:00 Quit kugel ("ChatZilla 0.9.84-rdmsoft [XULRunner]")
18:29:48 Quit einhirn (Read error: 104 (Connection reset by peer))
18:30:20 Join jaykay_ [0] (
18:30:32 Quit Mortiffer ()
18:32:30amiconnugh seems to have a perl problem
18:34:17amiconnjhMikeS: ping
18:36:01 Join kugel [0] (n=kugel@rockbox/developer/kugel)
18:42:20 Join gromit` [0] (
18:44:19*Unhelpful thinks perhaps a failed build ought to be worth one fixed amount of points, rather than per-line-of-errors
18:45:14 Quit jaykay (Read error: 110 (Connection timed out))
18:45:16gevaertsYou mean you want to impose an absolute maximum score? Where's the fun in that?
18:46:27mipfihe i have some trouble with rockbock.. you have a macpod but when i show under prefences it shows me windows :( anyone a idea howto fix this error
18:47:59amiconnUnhelpful: test_greylib_bitmap_scale is now working properly on my recorder. Not sure what the crash problem on Ondio is, but I can't test that now (don't have my Ondio with me)
18:52:33 Quit nuonguy ("This computer has gone to sleep")
18:52:42 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
18:54:01bluebrothermipfi: what exactly is the issue and what did you first?
18:54:47 Join radec [0] (
18:57:06mipfibluebrother: this is a macpod, rockbox only running on winpods.
18:57:29jhMikeSamiconn: what's up?
18:57:57bluebrotherI know. Did you try to convert the Ipod and how? Are you on OS X? Have you already tried installing Rockbox?
18:58:49amiconnjhMikeS: Over a year ago, you changed something in the sim kernel, which meant moving the sim screendump call to a kernel thread
18:58:50mipfibluebrother: after this error i checked wikipedia and used the easy way (
18:59:26amiconnYou chose the backlight thread, which means that screendump doesn't work in Ondio sims (no backlight - no backlight thread)
18:59:41mipfibluebrother: yes im on osx
18:59:46jhMikeSamiconn: yes, because the sim file calls had to be done on a proper thread
19:00:01amiconnNow I am thinking about the best way to fix this. I could move the call to the scroll thread, but that means quite a bit of additional ifdefing
19:00:28amiconnIt might be better to add an extra kernel thread just for sims. What do you think?
19:01:28jhMikeSit's possible to just create a properly registered sim thread to do odds and ends.
19:01:44jhMikeSheh, guess we're thinking the same.
19:01:55bluebrothermipfi: please tell details on your own −− I'm not willing to ask you about every detail multiple times. So you restored the Ipod on a windows machine using Itunes. Have you tried installing Rockbox after that? Does that give you an error? What exactly _is_ the issue _now_=
19:03:00mipfibluebrother: yes i have tryed installing rockbox after this step but i got the error that i use a macpod :(
19:03:54bluebrotherand how did you try installing Rockbox? Have you used Rockbox Utility? Which version? Have you closed Itunes before trying that?
19:04:07 Quit faemir ("Lost terminal")
19:05:20mipfii have closed i tunes and then i started rockbox utillity version m 1.0.9
19:06:10bluebrotherok, so it tells you that your ipod is a macpod all the time. Please try version 1.0.7 and check if the issue persists.
19:10:20mipfibluebrother: something is wrong with the new installier now i used 1.0.7 and i think it works :)
19:10:30jaykay_jhmikes: should i do a bench with your (latest) patch?
19:10:47 Nick jaykay_ is now known as jaykay (
19:11:09bluebrotheryou *think* or it *did* work?
19:12:07 Quit kadoban (Remote closed the connection)
19:13:04 Join faemir [0] (
19:15:02mipfibluebrother: its processing :)
19:15:16mipfiim going to eat something see u later :)
19:15:21bluebrotherand what exactly is "processing"?
19:20:50jhMikeSjaykay_: sure. There's some glitches with regard to reading pcm progress (mpegplay uses that heavily) but general playback should be fine.
19:23:59 Quit radec ("later")
19:26:11 Quit perrikwp (" ajax IRC Client")
19:28:33 Quit Lear ("ChatZilla 0.9.84 [Firefox 3.1b3pre/20090213033101]")
19:29:19 Quit bluebrother (" bbl")
19:34:33 Nick fxb is now known as fxb__ (
19:40:31 Join toffe82 [0] (
19:42:39 Join MethoS [0] (
19:46:29Unhelpfulamiconn: do we know who, if anyone, has done the serial mod for gdb on their ondio? ;)
19:46:51 Quit MethoS-- (Read error: 60 (Operation timed out))
19:47:05amiconnIt is not possible on the Ondio
19:47:10 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
19:48:22 Quit stoffel (Read error: 113 (No route to host))
19:48:40Unhelpfulah... that'd make it a real pain if this turns out to happen only on ondio. how does the aspect ratio correction work out in test_greylib_bitmap_scale? i've only tested by stretching sim screenshots.
19:49:00amiconnThe SH7034 has two serial interfaces, which can work both synchronously (spi) and asynchronously (rs232). One is taken for the MAS audio data transfer in all archoses. The other is used for the serial remote in the Player and recorders, and can be used for gdb
19:49:27amiconnBut in the Ondio, the other serial interface is used for the MMC spi....
19:50:00amiconnUnhelpful: Aspect correction works properly. A 200x200 cover.bmp looks square on my Recorder
19:52:31 Join pyro_maniac [0] (
19:53:04jhMikeSI guess it's possible to get past the glitch by splitting up transfers so the pcm tick runs fast enough for video playback and leave it at that (a few times lcd refresh)
19:55:11mipfibluebrother: now it works :)
19:55:32 Join MethoS- [0] (
19:56:22 Quit jordoex (Read error: 110 (Connection timed out))
19:57:48 Join stoffel [0] (
20:00:04jaykayjhmikes: is oqq or mp3 better (because of boosting)
20:00:09jaykay or shouldnt it make anydifference
20:05:05 Quit faemir ("Lost terminal")
20:06:21 Part Buschel
20:07:02 Quit BigBambi (Read error: 104 (Connection reset by peer))
20:07:23bluebrothermipfi: did you try restarting rbutil when you tried first (i.e. with 1.0.9)?
20:07:40 Join matsl [0] (
20:08:01 Join faemir [0] (
20:08:05 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
20:08:41 Quit courtc (Read error: 113 (No route to host))
20:08:57mipfibluebrother: i installed it with 1.0.7
20:09:00 Quit MethoS- (Remote closed the connection)
20:09:08 Join BigBambi_ [0] (
20:09:32bluebrotherwell, you previously said that you tried with 1.0.9 first and it failed. Did you restart Rockbox Utility and try again when you did that?
20:10:31saratogajaykay: Mp3 won't show a difference because it doesn't boost anyway
20:11:02 Quit morloi (Remote closed the connection)
20:11:27jaykayok, ill use ogg
20:12:06 Quit MethoS (Read error: 113 (No route to host))
20:12:43Unhelpfulvorbis will have better quality/size than mp3, mostly
20:12:44 Join nuonguy [0] (
20:15:59saratogaUnhelpful: I meant for a battery bench
20:16:27 Quit EspeonEefi ("さよăȘら")
20:23:50 Join piotrek2234 [0] (
20:24:03 Quit nuonguy ("This computer has gone to sleep")
20:28:57***Saving seen data "./dancer.seen"
20:32:05 Join courtc [0] (n=court@unaffiliated/courtc)
20:32:51n1sbluebrother: are you going to commit fs#9911 or should i?
20:35:02 Join gregzx [0] (
20:35:29 Join _lifeless [0] (n=lifeless@
20:36:11jhMikeSjaykay: whatever gives you the best comparison for what you've benched before
20:36:30jaykayi benched both..
20:38:45jhMikeSsaratoga: the core gets woken up less frequently even if not boosting at all
20:40:00jaykayso... is unboosted or bosted better...?
20:40:14 Join stripwax [0] (
20:41:01jhMikeSjaykay: flip a coin, I can't really say
20:42:04n1sBigBambi_: 2 small nitpicks about fs#9911, the first mention of pricture flow is written as "PictureFlow" while it is written as "Picture Flow" in the remaining places, maybe just a typo? The first item in the itemize is missing a period at the end. Looks nice otherwise :)
20:45:30 Join DrSexitude [0] (i=80e6d823@gateway/web/ajax/
20:45:42DrSexitudehey there
20:45:50DrSexitudeif i have rockbox on an ipod4gB&W
20:45:56DrSexitudewill it have a normally directory structure
20:46:01DrSexitudenone of that apple bullshit
20:46:29gevaertsWhichever way you like
20:46:37DrSexitudeno kidding
20:46:39DrSexitudethank god
20:46:45DrSexitudewhat license is rockbox under again
20:47:03n1sGPL v2 (or later)
20:47:28DrSexitudehow's the battery use
20:48:13n1sdepends on the target and what you do with it but generally pretty good, there are battery benchmarks in the wiki
20:49:11DrSexitudelisten to music real loud
20:49:25saratogajhMikeS: waking the core costs very, very little power though
20:49:35bluebrothern1s: feel free to go ahead −− I'm a bit busy tonight
20:50:09bluebrotherDrSexitude: you've noticed we have a FAQ on the website?
20:50:21saratogai doubt a change of 700kHz is measureable if its the difference between awake and idle, rather then awake and boosted
20:52:16DrSexitudei didn't even go
20:52:16DrSexitudeinstall rbutilqt through CLI :)
20:52:40bluebrotherrbutil runs without installation.
20:52:53DrSexitudedefine 'installation'
20:53:07bluebrothermore than "unzip and run".
20:53:11saratogaisn't this in the manual
20:54:43DrSexitudewell the script does tar -xf and moves it to /usr/bin then
20:54:43DrSexitudewho cares
20:54:49DrSexitudeit's still simple as hell
20:55:03saratogawho cares indeed
20:55:39bluebrotherwhich script?
20:55:43Unhelpfulwhat script? i don't think we provide such a script. and seriously, go look at the web site, and the manual.
20:55:48 Join z35 [0] (
20:56:21DrSexitudethe PKGBUILD
20:56:45bluebrotherthere is no pkgbuild script
20:57:00DrSexitudethere is!
20:57:07bluebrotherno. Nothing official.
20:57:14Unhelpfulthere is "the" pkgbuild script.
20:57:15saratogais ffmpeg flac the normal embedded flac decoder or is there another one in use?
20:57:58bluebrotherdefinitely not. We provide tarballs of prebuilt binaries, that's it.
20:58:11Unhelpfulsaratoga: i imagine the reference would likely work almost anywhere, wouldn't it? no float math, etc...
20:58:20DrSexitudewe're going in circles here
20:58:24BigBambi_n1s: Thanks. Yes, the first is a typo, the second an oversight
20:58:35rasherDrSexitude: which script are you talking about? Where did you get this from?
20:58:45bluebrotherno. I know what we (i.e. provide. There is no pkgbuild script or whatever.
20:58:48*kugel would rather call it Pictureflow anyway :/
20:58:58DrSexitudefrom the AUR
20:59:02kugelI don't like this CamelCase thing
20:59:15bluebrotherso if there is a script around its not official, and thus nothing we care about or support
20:59:26BigBambi_kugel: I agree, but I was striving for consistency
20:59:43kugelwhat consistency? :>
20:59:48rasherApparently AUR is some sort of Arch Linux thing
20:59:50DrSexitudesome friendly fellow put together an installation for the tarball
20:59:50DrSexitude./configure etc or not
20:59:51bluebrotherwhat's the problem with CamelCase in wikis? It makes it easy to spot WikiWikiWords ...
21:00:05DrSexitudeit's a conspiracy
21:00:32bluebrotherwell, it may be. But we do neither provide nor support something like that.
21:00:35rasherDrSexitude: would it have killed you to provide this information? We have no knowledge of what you did
21:00:37saratogahmm google says theres also closed source flac codecs too
21:00:52BigBambi_DrSexitude: We do not care what other people do. If it isn't from the Rockbox website it is unofficial and nothing to do with us
21:00:52linuxstbsaratoga: What do you mean by "normal embedded flac decoder" ? The ffmpeg decoder is LGPL'd, so I doubt any official firmwares will use it, if that's what you mean.
21:01:05DrSexitudeit's still nifty though, isn't it?
21:01:16saratogalinuxstb: thanks thats what I was wondering
21:01:19BigBambi_I don't care, please stop talking about it
21:01:25*gevaerts wonders what people are actually talking about. Did DrSexitude complain about problems with this unofficial thing?
21:01:26bluebrotherDrSexitude: -we- -don't- -care-
21:01:43linuxstbsaratoga: The ffmpeg FLAC decoder seems quite a well-kept secret. Everything seems to just use libFLAC.
21:01:50BigBambi_gevaerts: He just won't shut the hell up about it, and confused people in the meantime
21:01:50DrSexitudeyou act like it's painful to think about
21:02:04DrSexitudeperhaps we should delve deeper
21:02:12DrSexitudeyou seem to associate some sort of trauma with unofficial rockbox utilities
21:02:21BigBambi_or perhaps you should stay on topic
21:02:22BigBambi_or leave
21:02:23kugelBigBambi_: consistency with?
21:02:25saratogalinuxstb: why do we use ffmpeg again? its faster I guess?
21:02:34BigBambi_kugel: Other stuff in the manual
21:02:38DrSexitudei hear that a lot
21:02:38UnhelpfulDrSexitude: no, the issue is that you were deliberately unclear in what you were talking about, and made a point of not clarifying, and then acted like everybody *else* was ignorant.
21:02:46DrSexitudeit gets a little less friendly every time
21:02:56rasherDrSexitude: Sounds like you should work on that
21:02:59BigBambi_kugel: or the wiki page - I forget now
21:03:20BigBambi_kugel: It is easy to change if a consensus is reached
21:03:32DrSexitudenobody is ignorant, we have search engines now
21:03:32DrSexitudeyou stop being ignorant the minute you learn how to access any type of knowldge within seconds
21:03:41linuxstbsaratoga: Speed is the main reason, but it's also simple and small. libFLAC is a nightmare of layer-upon-layer of APIs. IIRC, the entire FLAC codec (bss and code) fits in 48KB (and hence IRAM)
21:03:57linuxstb(ffmpeg FLAC codec)
21:04:00saratogaah ok
21:04:17gevaertsDrSexitude: exactly. Why did you ask about battery runtime and file structure again?
21:04:27saratogaDrSexitude: maybe you could search then instead of talking
21:04:38DrSexitudei asked about battery runtime cause apple did such a terrible job with batteries
21:04:39saratogaits kind of annoying when you use words and ideas
21:05:06Unhelpfuldoes the ffmpeg one support the rice2, or whatever, coding that was added somewhat recently to improve compression ratios in 24-bit flac?
21:05:17DrSexitudei asked about file structure to see if it was worthwhile to install rockbox so that i could, say, drag mp3s onto the ipod
21:05:17DrSexitudewhich is an important goal for any non-terrible mp3 player
21:05:37BigBambi_So do we want Picture Flow, Pictureflow or Picture flow?
21:05:48BigBambi_I think on balance I prefer the 2nd
21:06:14 Quit nibbler_ (Remote closed the connection)
21:06:18Unhelpful"pictureflow", maybe?
21:06:37amiconnsaratoga: reference libflac is oo-in-C to the extreme, meaning painful to follow, and dog slow compared to ffmpeg
21:06:52BigBambi_Unhelpful: Depends if we think of it as a proper noun
21:08:23Unhelpfulamiconn: does that include virtual functions in C? i can't imagine calling via pointer is good to do on ARM any more than absolutely necessary
21:08:23 Join MethoS [0] (
21:08:39amiconnJust have a look....
21:08:56amiconnlibflac is still somewhere in the rockbox svn history
21:09:24kugelBigBambi_: just do PictureFlow. I looked, and the image has a uppercase F too (the image you see when it builds the cache)
21:09:24amiconnAnd it's slow on coldfire *and* arm
21:09:41BigBambi_kugel: In that case, OK
21:09:50 Quit faemir ("Lost terminal")
21:09:56 Quit bluebrother (Nick collision from services.)
21:10:01 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
21:11:02 Join faemir [0] (
21:11:28Unhelpfulheh, i didn't know it was awful on coldfire, as well. would it be worthwhile special-casing the call to output_row_native in the scaler, instead of always using a pointer (which points to that function almost all of the time, anyway)?
21:11:37 Join MethoS- [0] (
21:12:42 Quit MethoS (Read error: 60 (Operation timed out))
21:14:10 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
21:16:08BigBambi_OK, new one there
21:16:13 Part DrSexitude
21:16:21 Join SoapWork [0] (n=42c07542@gateway/web/cgi-irc/
21:17:46 Join MethoS-- [0] (
21:19:59linuxstbUnhelpful: The libFLAC author submitted a patch to ffmpeg to add that, but I'm not sure it's working correctly - we applied it to Rockbox, but I think there is an open bug regarding 24-bit FLAC files encoded by recent libFLAC encoders.
21:21:20*linuxstb got 1000s of lines of CVS activity credit by adding libFLAC to Rockbox, and then removing it...
21:21:49 Quit SoapWork ("CGI:IRC (Ping timeout)")
21:23:08 Join guest001 [0] (
21:24:18 Join MethoS [0] (
21:26:58BigBambi_urgh. I made it PictureFlow and I really don't like it. Even though the plugin text when you start is PictureFlow, would anyoen object to Pictureflow?
21:27:30bluebrotherwhy not PluginPictureflow, like other plugins?
21:27:50BigBambi_bluebrother: This is for the manual
21:28:27bluebrotherah, ok. That removes the need for CamelCase completely of course. Thought you were talking about the wiki ...
21:30:57 Part pyro_maniac
21:31:05 Quit MethoS- (No route to host)
21:32:01 Quit kugel (Nick collision from services.)
21:32:04 Join kugel_ [0] (
21:32:07 Nick Beaver is now known as Beaver`away (
21:32:15 Nick kugel_ is now known as kugel (
21:32:40 Quit guest001 ()
21:33:44kugelbluebrother: also, PluginPictureflow already exists (PluginPictureFlow even)
21:34:33BigBambi_well I made it Pictureflow for now. If the person who commits it wants to change it then that is always an option :)
21:35:01 Join gregorovius [0] (
21:35:35 Join SoapWork [0] (n=42c07542@gateway/web/cgi-irc/
21:36:12 Quit SoapWork (Client Quit)
21:37:33linuxstbIsn't it loosely based on a Qt app?
21:37:48 Quit MethoS-- (Connection timed out)
21:38:01kugelBigBambi_: I think I changed my mind (I searched for consistency). I still like Pictureflow more, but the settings menu is also entitled "PictureFlow Settings"
21:38:12kugellinuxstb: not anymore ;)
21:38:31merbanansaratoga: I guess you are mike (regarding the lowbitrate wma files) ?
21:38:31linuxstbkugel: ?
21:38:33kugelUnhelpful stripped everything loose away
21:38:38 Nick BigBambi_ is now known as BigBambi (n=Alex@rockbox/staff/BigBambi)
21:39:39Unhelpfulleaving only things, what, closely based on a Qt app?
21:40:32kugeldamn, BigBambi beats me
21:40:34linuxstbWhich sadly calls itself "PictureFlow"...
21:40:38BigBambiand it i PictureFlow there
21:40:49BigBambilinuxstb: sadly so
21:41:03BigBambiOK then, PictureFlow
21:41:38 Quit MethoS (Read error: 110 (Connection timed out))
21:42:00 Join tessarakt [0] (
21:42:10bluebrotherdomonoky: around?
21:42:32 Quit tessarakt (Client Quit)
21:42:44 Join jgarvey [0] (
21:42:57bluebrotherI've tried to fix the issue on OS X that the installation doesn't wait for the player to get remounted. Are you able testing the patch in any way?
21:43:00*kugel thinks the project has to backport many code from us
21:43:05 Part piotrek2234
21:43:06kugelbut it apparently isn
21:43:11kugelisn't active anymore
21:43:23bluebrotherI'd like to get that in and then try to release 1.0.10 the next days −− unless someone has objections to a new release
21:43:50domonokybluebrother: nope, i can not test on mac, i only can build it on some mac, on another continent :-)
21:44:14*linuxstb wonders how many more 1.0.x releases there will be...
21:44:19bluebrotherhmm. Anyone else around using a mac that can test?
21:44:41bluebrothermaybe we should write a roadmap draft on the wiki page :)
21:45:02Unhelpfulthe other pictureflow offers white backgrounds, and blurred reflections?
21:45:13kugelI just saw that too
21:45:33linuxstbbluebrother: rbutil is still doing the install backwards? i.e. bootloader first, then rockbox itself?
21:45:43bluebrotherlinuxstb: yes.
21:45:48bluebrotherwhy is that backwards?
21:46:16Unhelpfuli see that the desktop app offers bilinear filtering. wonder if it does it in software?
21:46:28linuxstbThe manual method is to install rockbox first, then the bootloader, so that the device is never in an "unbootable" state.
21:46:54linuxstb(i.e. how the manual describes the install process)
21:47:06bluebrotheriirc the manual always told to first install the bootloader. It's questionable if it would be better doing the other way round, but with rbutil we still have the issue that the player could be in an unmounted state
21:47:23bluebrother(f.e. if the user installs other stuff and the OS was too slow ...)
21:47:46amiconnJdGordon|zzz: r19747 is the cause for the null pointer access at end-of-playlist on hwcodec. However, a trackchange event based "follow playlist" has a fundamental flaw anyway: it does not, and cannot, work properly
21:48:33bluebrotherlinuxstb: do you have an idea how long remounting usually works on OS X?
21:49:04linuxstbI don't know if they've fixed the bug, but mounting a FAT formatted ipod could take a couple of minutes when I had a Mac (iBook).
21:49:37bluebrotherhmm, bad. I currently have a timeout of 15 seconds.
21:50:53bluebrotherwhich could be increased of course. But waiting several minutes just for remounting feels kinda like ages
21:51:12bluebrothercan't we initiate the remount with disktool? Or is that as slow as the system remount?
21:51:32 Quit kugel ("ChatZilla 0.9.84-rdmsoft [XULRunner]")
21:55:34linuxstbYes, I think it's just as slow. It seems that OS X does a full scan of the disk (I remember seeing the disk icon spinning whilst it was waiting to mount).
21:55:57linuxstbSo what exactly happens now? When the bootloader is installed, the ipod reboots itself immediately?
21:56:56bluebrothercurrently the automated install installs the bootloader, then immediately proceeds with the build. Which obviously will get installed in the wrong folder
21:57:27linuxstbYou could poll for the presence of the iPod_Control folder.
21:58:01bluebrotherI choose to check if the device ipodpatcher detected is mounted using mntent
21:58:34linuxstbOK, even beter.
21:58:34bluebrotherif you want to have a look at the current implementation:
21:58:43 Part domonoky
22:01:07linuxstbbluebrother: Does the ipod even reboot into disk mode? If it reboots after the bootloader is installed (but before Rockbox is installed), won't it just stop in the bootloader with the "rockbox.ipod not found" error.
22:02:57bluebrotherAFAIK it doesn't reboot, but I might be wrong.
22:04:36linuxstbYes, I think you're right. (it's slowly coming back to me...)
22:08:28 Join piotrek2234 [0] (
22:08:30bluebrotherfound it: initially the manual installation instructions had the bootloader first, but that was changed around two years ago. Somehow I missed that swap ...
22:09:51bluebrotherat least there is a reason why I thought it's the other way round ;-)
22:10:52 Join gregorovius_ [0] (
22:10:57 Join SoapWork [0] (n=42c07542@gateway/web/cgi-irc/
22:12:04 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
22:12:26 Quit domonoky (Read error: 104 (Connection reset by peer))
22:13:22 Join MethoS [0] (
22:16:00 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
22:17:48 Join MethoS- [0] (
22:18:26 Quit SoapWork ("CGI:IRC (Ping timeout)")
22:19:37 Quit jgarvey ("Leaving")
22:22:00 Quit domonoky (Read error: 104 (Connection reset by peer))
22:22:50 Join SoapWork [0] (n=42c07542@gateway/web/cgi-irc/
22:23:38 Quit stoffel ("leaving")
22:24:23 Quit gregorovius (Read error: 110 (Connection timed out))
22:24:53 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
22:25:36 Join kadoban [0] (
22:28:59***Saving seen data "./dancer.seen"
22:29:23 Join MethoS-- [0] (
22:29:56 Join kugel [0] (i=kugel@rockbox/developer/kugel)
22:34:19 Quit MethoS (Success)
22:34:21 Quit SoapWork ("CGI:IRC (Ping timeout)")
22:36:57 Join SoapWork [0] (n=42c07542@gateway/web/cgi-irc/
22:37:43 Quit SoapWork (Client Quit)
22:37:47 Join SoapWork [0] (n=42c07542@gateway/web/cgi-irc/
22:38:45 Quit SoapWork (Client Quit)
22:38:51 Join SoapWork [0] (n=42c07542@gateway/web/cgi-irc/
22:42:19 Quit domonoky (Read error: 104 (Connection reset by peer))
22:43:04 Quit SoapWork (Client Quit)
22:44:12 Join SoapWork [0] (n=42c07542@gateway/web/cgi-irc/
22:44:42 Quit SoapWork (Client Quit)
22:44:46 Join SoapWork [0] (n=42c07542@gateway/web/cgi-irc/
22:45:53 Quit MethoS- (No route to host)
22:46:24 Quit jaykay (Read error: 110 (Connection timed out))
22:46:48 Quit perrikwp (" ajax IRC Client")
22:47:42 Quit MethoS-- (Read error: 113 (No route to host))
22:48:53 Nick fxb__ is now known as fxb (
22:49:10pixelmatime for blaming perl for the failed V2 Recorder build?
22:49:11kugelFlynDice: hi
22:49:48*kugel wonders if amiconn didn't notice it
22:49:49 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
22:50:08kugelor, he finds his score so impressive that it needs to stay ;)
22:50:36pixelmakugel: looks like a server error to me. Does someone know who is running (wasn't it remotely related to GodEater)?
22:50:46 Quit SoapWork ("CGI:IRC (Ping timeout)")
22:51:14*pixelma adds a .co
22:53:28 Join jaykay [0] (
22:55:26rasherGodEater_ it is indeed
22:55:36kugelpixelma: although the server did 3 builds, the other two succeeded
22:56:57 Join tessarakt [0] (
22:58:27 Quit Tuplanolla (Read error: 60 (Operation timed out))
22:58:44amiconnkugel: Sure I noticed. See log @ 18:32...
22:59:25kugeloh, I overlooked
22:59:45FlynDiceKugel: Hi, tried the DBOP_DOUT= last night. Starts a red line at the top of the screen. More later no time now.
23:00:21 Quit Horschti ("Verlassend")
23:01:55pixelmakugel: the other two builds (Gigabeat F and H120 bootloader show perl warnings too already but didn't fail completely
23:05:53 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
23:20:23 Quit jaykay (Read error: 110 (Connection timed out))
23:22:53 Quit merbanan (Remote closed the connection)
23:33:11 Join itcheg [0] (i=62db4767@gateway/web/ajax/
23:34:30 Quit bluebrother ("leaving")
23:35:02 Join bs66_1 [0] (
23:35:25 Quit domonoky (Read error: 104 (Connection reset by peer))
23:39:58 Join Thundercloud [0] (
23:40:58 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
23:41:17 Join Linearity [0] (
23:42:10*pixelma can now see a greyscale Cover.bmp on her Ondio too :)
23:42:40pixelmawith the correct aspect ratio as well
23:43:46 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
23:44:19Unhelpfulpixelma: i'm glad it works for you, though it's a bit of a wtf that it was crashing before
23:44:51 Join ferric84 [0] (
23:45:30ferric84i just installed rockbox on my ipod video 5.5 30gb using rockbox utility. now it continuously restarts every 5 seconds. makes it to the apple screen and thats it
23:45:48ferric84is it possible to recover from this?
23:47:31pixelmaUnhelpful: this time I did a "make clean" before maybe there was something halfway recompiled or whatever. It also doesn't crash with the "catch mem accesses" activated so it's very hard to say what was wrong
23:48:22pixelmaferric84: did you try to reset the Ipod - and do you get any error messages?
23:48:31Linearityok, i've read the FAQ and manual and i still have problems
23:48:50saratogamerbanan: [for the logs] yes thats me
23:48:56ferric84pixelma: i can hold menu+the middle button forever but it just keeps restarting. it won't stay on long enough. no error messages
23:49:08Linearityfor some reason, the rockbox utility can find the sansa e280 fine on the autodetect, but it can't detect on the install
23:49:12ferric84after installation i just unplugged it from my computer and this is what happened
23:50:08*amiconn found a superfluous instruction in the 32*32->64 bit sh1 multiply
23:50:15ferric84after the battery dies it's a brick for sure so if anyone has any ideas, please let me know
23:50:18domonokyLinearity: thats known bug for sansas, use sansapatcher for the bootloader install for now
23:50:59Linearityok, thanks
23:51:17Unhelpfulamiconn: the new, unsigned one, the one you wrote for mandelbrot, or the gcc one?
23:51:29amiconnThe former two
23:51:41pixelmaferric84: unfortunately I don't know much about Ipods and installing with the Rockbox Utility, already out of ideas. Stay around for a while, guess there are people around who can help better (but not all the time)
23:51:44amiconngcc does something C based iirc - rather slow
23:52:56pixelmaferric84: and I don't think you'll get a "brick", that's very very unlikely with Ipods
23:53:01 Quit {-phoenix-} (Remote closed the connection)
23:53:16Unhelpfuli tried various ways to do it in C, gcc turned all of them into pages of asm. i'm kind of curious as to what scaler speed is like on sh1, it seems the slowest (by clock, anway) CPU on any target.
23:53:28amiconnUnhelpful: The unsigned one is effectively the core of the signed one. The signed one does an unsigned multiply first, then fixes up the top 32 bits depending on the sign of the two factors. And the output shift is different in mandelbrot, requiring a few shll/shlr instructions
23:53:40 Quit bertrik ("Leaving")
23:54:12linuxstbferric84: Is Rockbox starting, but then immediately rebooting?
23:54:34linuxstbferric84: Sorry, just re-read your first posts - it's rebooting at the apple logo, and going no further?
23:54:40Unhelpfulit's a bit annoying, given that multiply output can *be* twice as large as input, and that so many CPU have an instruction for doing that, that there's not a "nice" way to express it in C.
23:54:42ferric84linuxst: correct
23:54:59linuxstbferric84: What if you hold SELECT+PLAY? That should enter the emergency disk mode.
23:55:03 Quit gregorovius_ ()
23:55:08 Part Linearity
23:55:41ferric84linuxstb: nice - thank you
23:56:09linuxstbferric84: When your ipod was rebooting, was the backlight coming on?
23:56:27ferric84linuxstb: yeah
23:56:55 Part toffe82
23:57:11 Quit n1s ()
23:57:26linuxstbThen the Rockbox bootloader was running (that turns on the backlight). I don't recall anyone reporting a similar problem before...
23:58:04 Quit bs66_ (Read error: 110 (Connection timed out))
23:58:11linuxstbferric84: Was the Apple firmware working fine before you installed Rockbox?
23:58:35ferric84i just reinstalled rockbox manually instead of doing complete installation and it worked
23:58:52linuxstbHmm, OK...

Previous day | Next day