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 2009-07-25

00:00:41 Quit HellDragon (Read error: 104 (Connection reset by peer))
00:00:53 Join HellDragon [0] (
00:03:12pixelmais english.lang not compiled in time now on multicore machines (re. red) or so?
00:03:48ZagorI haven't figured out what the bug is yet.
00:05:05*pixelma grumbles, why are builds working ubtil they are flashed :\\
00:05:15pixelmauntil too
00:07:17JdGordon|Zagor: woo, nice commit... what sort of speedup does it give?
00:07:37gevaertsJdGordon|: I'm working on some stats :)
00:07:51JdGordon|work harder :D
00:07:55saratogarerunning make a second time after a make clean reduces build times by quite a lot on my machine
00:08:20saratogabut the new changes take build time from 69 to 54 seconds on my 4x Opteron
00:08:34saratogaafter a few make, make clean cycles
00:13:37gevaertsJdGordon|: yesterday, now. Gigabeat S, 8-core machine, no ccache, warm VFS cache
00:13:43Zagorit gave most fast machines in the round a healthy speed bump, too. I don't think we've had a 400+ speed before
00:14:28Zagorgevaerts: whoa, that is a nice boost
00:15:20JdGordon|still shows no difference with more than 6 jobs though
00:15:42JdGordon|or is the depgen maxed up to the number or cores?
00:15:42gevaertsBefore this, -j8 used 3.8 cores on average, now it uses 6.14
00:16:03ZagorJdGordon|: 8 is faster than 6
00:16:26JdGordon|then starts getting slower again
00:16:47Zagoryes, because 8 is the number of cores. this is expected.
00:17:26 Quit stripwax ("")
00:19:13 Quit bmbl ("Bye!")
00:19:49 Join derekja [0] (
00:19:58 Part derekja
00:25:24 Quit saratoga (
00:25:24 Quit evilnick (
00:25:24 Quit CaptainKwel (
00:27:36bertriklinuxstb, can you check what value the LCD_CON register has on the nano2g?
00:29:38 Quit aaron424 (Remote closed the connection)
00:29:58 Join arohtar [0] (n=faemir@
00:32:29linuxstbbertrik: Sure. (I assume you mean in the OF, before Rockbox touches it?)
00:33:03 Quit FlynDice_ (Remote closed the connection)
00:39:52 Quit jgarvey ("Leaving")
00:41:44 Quit arohtar (Client Quit)
00:41:48 Quit sinthetek (Read error: 60 (Operation timed out))
00:42:33Zagorlooks like there's a bug in my array slicing
00:44:00gevaerts shows some interesting numbers.
00:45:04Zagorlooks like my stab with $cores+1 isn't so bad
00:45:19 Join faemir [0] (n=faemir@
00:45:57gevaertsthis is with a warm cache of course. On shared machines things are going to look a lot less neat
00:46:19gevaertsit looks to me like I'd be best off running to -j4 clients
00:47:15gevaertsaccording to these numbers, that should be 20% faster than a single -j8 client
00:47:40Zagorjust keep in mind that a 4-core client will run -j5
00:48:12gevaertsI know, yes. These pesky smart things :)
00:49:38gevaertsok, this thing should get very close to 100% CPU use overall
00:49:49bertrikcan't get anything out of the LCD yet, not even corrupt it or clear it
00:50:22CIA-7New commit by zagor (r22022): Fixed buggy array slicing.
00:50:39*gevaerts waits eagerly
00:50:41bertrikay, maybe I should configure the LCD pins first ...
00:50:55*bertrik slaps forehead
00:53:19linuxstbbertrik: Hmm, with my current (working) code, the Nano2g still gives audio with the codec set to slave mode, but with the wrong frequency (i.e. speeded up). In master mode, it sounds right...
00:53:32 Quit Ridayah (Read error: 110 (Connection timed out))
00:54:08 Join Ridayah [0] (
00:55:48CIA-7New commit by wincent (r22023): Added myself to the list of committers.
00:56:06 Join wincent [0] (
00:57:23gevaertswincent: congratulations!
00:57:39wincentThank you!
01:01:13Zagorgevaerts: I suppose we should do a new score benchmark now...
01:02:09gevaertshm, yes. This could have changed some things
01:02:41gevaertsAlso, I'm going to go back to a single no-explicit-cores client. My experiments don't speed things up
01:03:08***Saving seen data "./dancer.seen"
01:08:51 Nick wincent is now known as wincent_ (
01:09:29 Nick wincent_ is now known as wincent (
01:10:35gevaertsZagor: building new scores now
01:11:39 Quit faemir ("Leaving")
01:12:05ZagorBagder: can you re-add the source 7z building cronjob? it's too big a job to be done by the build server after each round. (it takes nearly 2 minutes!)
01:13:28Bagderit's already there, built daily
01:14:13Zagorwhere? is not being updated
01:14:43Bagderthose are not daily
01:15:19 Quit Thundercloud (Read error: 104 (Connection reset by peer))
01:15:26Bagderif we can't keep up to do them every commit, then I think daily is the only sensible option
01:15:38 Join Thundercloud [0] (
01:16:13Zagorthat's fine with me
01:16:13gevaertsWhy should it be done after the round? Can't it be done during?
01:16:32Bagderit could be made as a separate build
01:16:51Bagderbut I don't think we need it
01:17:01Zagorgevaerts: yes it can. but the server is single-thread and runs a post-round shellscript that updates the web pages etc. it would fit there if it wasn't so big.
01:19:18*gevaerts thinks about it, and he thinks he's also in favour of a daily source file
01:19:44 Join funman [0] (n=fun@rockbox/developer/funman)
01:19:50 Part wincent ("Kopete 0.12.7 :")
01:20:05*Hillshum thought there already were daily tarballs
01:20:18Bagderyes, there were
01:20:23Bagderor there are
01:21:09ZagorHillshum: we're discussion the "bleeding edge" tarball, remade each commit.
01:21:32Hillshum* gevaerts thinks about it, and he thinks he's also in favour of a daily source file
01:21:52gevaertsyes, that wasn't phrased very well
01:23:32gevaertshm, doesn't viewvc have an archive download option?
01:23:48Unhelpfulwhy couldn't it be done as a commit hook?
01:24:23ZagorUnhelpful: because the web server and svn server are different machines
01:25:25Zagorgevaerts: we don't have that enabled. it causes too much load, and most users abort before it's done
01:25:28Unhelpfulbut the commit hook could upload the file, perhaps? it just seems silly to have some random client on the internet do make the source archive - i'd imagine even as separate machines the svn and web servers are more closely connected than either is to the build clients
01:26:00gevaertsseriously though, per-revision builds are useful because they find bugs and compiling lots of builds isn't that convenient. per-revision source doesn't give you anything that a normal svn checkout doesn't give yoy
01:26:11 Join wincent [0] (
01:26:16Zagorwe'll simply skip the bleeding edge source file. if you want bleeding edge, use svn. daily source is good enough.
01:26:38Unhelpfulgevaerts: other than "sources to people with blocked ports"... but i think they could use git-http?
01:27:05 Quit JdGordon| ("Miranda IM! Smaller, Faster, Easier.")
01:27:09ZagorI'm pretty sure they don't want to grab a new 22MB tarball on every commit anyway
01:27:11gevaertsZagor: I was hoping that it would cache the thing, but it's probably much too flexible for that to be possible
01:29:49 Quit PaulJam (Nick collision from services.)
01:29:58 Join PaulJam_ [0] (
01:32:10ej0rgefunman: Do you know, is the EQ on the fuze (and other AMS models) done in hardware or software?
01:32:47linuxstbej0rge: I don't think any target has hardware EQ (apart from bass/treble)
01:33:06 Quit Zagor ("Clint excited")
01:33:15 Quit wincent ("Kopete 0.12.7 :")
01:33:49ej0rgelinuxstb: Some codec chips claim to have 5-band eq in hardware, i admit i haven't checked the relevant AMS datasheet
01:35:02ej0rgelinuxstb: so maybe i am confused in believing that rockbox on some targets may be using hardware EQ
01:36:39funmanI think the as3514 chip can only modify the volume, not separate bass/trebbles
01:37:01 Quit PaulJam_ (Nick collision from services.)
01:37:02ej0rgefunman: ok, thanks
01:37:04 Join PaulJam [0] (
01:37:47 Join JdGordon| [0] (i=ad808ab0@gateway/web/freenode/x-36ed1d043524355b)
01:37:47funmanat least i can't find anything showing that it can in the datasheet
01:38:02 Quit ender` (" Some people like my advice so much that they frame it upon the wall instead of using it. -- Gordon R. Dickson")
01:40:17bertrikyay, I'm slowly getting there, I just drew a black rectangle over the existing logo
01:40:23 Join Telazorn [0] (
01:41:22ej0rgefunman: it must be the wm8978 in the beast that i'm thinking of
01:41:48ej0rgefunman: which has hardware eq, but this isn't used in rockbox (yet?)
01:42:41bertrikquite a few of the wm codecs seem to support treble/bass in hardware
01:42:44JdGordon|anyone got a good name to diferenciate between the wps engine and the wps screen?
01:43:05JdGordon|or even a mediocre or bad name
01:43:06CIA-7New commit by gevaerts (r22024): re-measure all scores after the dependency generation rework
01:43:08funmanej0rge: i think bertrik is your man for using hardware eq in rockbox ;)
01:43:28ej0rgebertrik: wolfson's product page for the 8978 boasts of 5-band eq for either (both?) input or output
01:43:30gevaertsJdGordon|: the wps engine will be used for similar screens later, right?
01:43:41linuxstbJdGordon|: What do you mean? What to call the "engine" and what to call the current WPS screen?
01:43:52JdGordon|gevaerts: yes
01:43:55ej0rgefunman: Mostly i was curious because i've heard some people complain that the fuze OF distorts badly when using the EQ
01:43:55JdGordon|linuxstb: both
01:44:58JdGordon|actually, I'm tihnking about creating apps/screens/ and then just putting music.c in there and leave the engine named wps
01:45:31linuxstbThere is no music on the radio, or no spoken word etc in files?
01:46:30gevaertsI think the screen should keep the wps name, and the engine something else
01:47:06JdGordon|there is less work involved if the engine stays wps :)
01:47:36gevaertsmaybe, but that will be confusing in the long run
01:47:51 Join wincent [0] (n=wincent@rockbox/developer/wincent)
01:48:00JdGordon|and on the topic of cleaning apps... would anyone like to volanteer to move all the playback related stuff into a playback folder?
01:48:25*Hillshum volunteers MrSomeone
01:48:33gevaertsThat sounds easy!
01:48:44gevaertswhich probably means it isn't :)
01:48:58 Part wincent ("Kopete 0.12.7 :")
01:49:29JdGordon|it shold be dead simple
01:49:29 Join wincent [0] (n=wincent@rockbox/developer/wincent)
01:49:43*gevaerts refuses to bite
01:50:02*JdGordon| is thinking "theme" or something for the engine.. but thats a lie...
01:50:52*gevaerts doesn't believe in short names
01:51:19JdGordon|its not java or c#
01:51:24JdGordon|maybe just "a"?
01:51:47JdGordon|screener isnt nice because there is already screens and screen_access
01:58:12 Join bzed_ [0] (
01:59:12*bertrik sees some red and green rectangles on his display
01:59:53gevaertsdid you port the build status table to that player?
02:00:22 Quit bzed (Read error: 104 (Connection reset by peer))
02:00:24 Nick bzed_ is now known as bzed (
02:04:17 Quit JdGordon| (Ping timeout: 180 seconds)
02:04:39 Quit PaulJam (".")
02:05:05Unhelpfulanybody have opinions on FS #4817? markun vaguely remembered amiconn being against such a merge, but i can't find logs regarding that.
02:07:50bertrikyay, 2 tiny rockbox logos and a lot of garbage
02:08:43funmanUnhelpful: can you somehow benchmark scrolling code?
02:09:28funmanif you arer not in a hurry, you could wait for amiconn's return from holidays or write to the mailing list (i am not sure if he's reading it ..)
02:09:58Unhelpfulfunman: not too easily. scroll_fn in both old and new does nothing if no items require updates
02:10:37funmananyway scrolling doesn't need to be optimised for speed ..
02:12:30Unhelpfulthat was my thought. it adds one extra call depth to scroll_fn, with the callee basically making all of the calls that scroll_fn used to do itself. this, and merging the two gradient drawing routines, cuts out about 600B on color targets. it's pretty much a wash on coldfire greyscale, and saves about 50B on arm greyscale
02:12:55funmanabout Sansa AMS: using 48MHz PCLK / 240MHz PLLA, and 24MHz MCLK in SD module, I get different results on successive boots: "no .rockbox directory", "no partition found", DATA TIMEOUT (SD bank -1 : error 0x8)
02:14:07funmanUnhelpful: an effect you didn't mention : does it make the code more readable ?
02:14:32HillshumOh, by the way, my 2GB sandisk card sometimes works in my e200v2
02:14:45Unhelpfulfunman: it consolidates identical and nearly-identical puts* functions that were in each bitmap LCD driver :)
02:16:03funmanHillshum: "sometimes works" ? did it use to never work ?
02:16:43UnhelpfulTorne: i was thinking, it would be fairly safe to test stuffing rockbox bootloader bits into the blank areas of beast bootrom if we replace the jump at the start with a test (against the lock switch, which is gpio-accessible) and jump
02:18:26TorneUnhelpful: yes, if the gpios are accessible that early
02:18:37Torneit would need to be somewhat further down really i think
02:18:52Torneclocking and muxing needs to be set up
02:20:16Unhelpfulah... so you'd need some setup before we could read any buttons. :/
02:21:54 Quit GeekShado_ ("The cake is a lie !")
02:22:35Hillshumfunman: I can't find it right now, but iirc whenever I went to the filebrowser, there was no <microsd> entry. Reinserting the card usually fixed that
02:23:13funmanHillshum: i think it was a problem caused by gevaerts
02:23:19funmannot something AMS-specific
02:23:27Hillshumis it fixed?
02:24:06TorneUnhelpful: probably. the state of the chip at reset is very basic.
02:24:34TorneUnhelpful: it goes through something like a thousand instructions before it even has relocated the bootlaoder to ram and turned the mmu on
02:25:57Unhelpfuli was hoping to have a relatively safe "check the lock switch" test before the experimental parts, so that it would be reasonable to test flash bootloaders without a jtag handy :/
02:27:30Tornewell it might not come to needing a flash bootloader
02:28:45Unhelpfulwell, i'd really just like to even have that option, in order to speed things up. but if you can figure out how to patch out the tests that cause the problems in the "official" bootrom, that would be just about as good. :)
02:29:38funmanHillshum: i think it was fixed
02:30:01Tornetbh if we did modify the initial bootloader we'd want to do it *after* the bootloader did its initial startup
02:30:05Tornebut just before disk init
02:30:20Torneor during disk init, possiblt
02:30:45Tornebasically at some point where the code has already had the chance to jump to the recovery image if the flash is broken
02:31:07Tornebeing able to trigger that as a recovery option could be good
02:31:08 Join barrywardell [0] (n=barrywar@
02:31:23Unhelpfulthe bootrom is much, much too picky about deciding the player is broken, though. :/
02:31:33TorneYah, but i mean before it's even powered the disk up
02:31:49Torneit has a number of checks that cause reformat -> mtp recovery before ATA is even turned on
02:32:38Torneif we were to patch in code somewhere there we could probably manage a fairly easy button combinatino totrigger its own recovery logic if nothing else
02:32:48Unhelpfulah... yes, i suppose it would be good not to bypass that.
02:32:51Tornesince it's already in a state to jump to OSFatalError
02:38:30 Quit mt (No route to host)
02:40:51 Quit barrywardell ()
02:42:56 Quit darkless ("Leaving")
02:47:04 Quit Hillshum (Read error: 104 (Connection reset by peer))
02:49:18CIA-7New commit by dave (r22025): Correct the memory sizes (IRAM and DRAM) for the Nano2G.
02:49:36linuxstbOops. A few too many files....
02:52:12 Quit bertrik ("De groeten")
02:54:34CIA-7New commit by dave (r22026): Oops, too many files were committed in r22025, revert this one. The other changes committed as r22025 were initial button and audio drivers for the ...
02:54:59 Join TheCoolGman [0] (n=446e5e7d@gateway/web/cgi-irc/
02:56:39TheCoolGmanSo what is with the Current Build? The Source seems to be gone :O
02:58:20 Quit funman ("free(random());")
02:59:57linuxstbTheCoolGman: I think there was a decision earlier to stop offering source code downloads after every commit - just the daily.
03:00:30TheCoolGmanHmm, Well I tried that and I got build 21932
03:02:16linuxstbBagder and Zagor are the people you need to talk to, and they left (for bed I assume) a couple of hours ago.
03:02:39linuxstbYou can't use SVN?
03:02:50TheCoolGmanAhh then I know what to do.
03:03:01 Quit TheCoolGman ("CGI:IRC")
03:03:12***Saving seen data "./dancer.seen"
03:15:35 Nick YpsyZNC is now known as Ypsy (
03:39:29 Nick Ypsy is now known as YpsyZNC (
03:41:41 Part wincent ("Kopete 0.12.7 :")
03:47:12tmztlinuxstb: you can always svn -r right?
03:53:41linuxstbtmzt: ?
03:54:14tmztsvn co -r
03:54:20tmztto get a particular revision
03:59:39 Quit goffa__ (Remote closed the connection)
04:05:56 Quit Thundercloud (Remote closed the connection)
04:06:51 Quit TheSeven (Nick collision from services.)
04:07:07 Join The_Seven [0] (
04:07:11 Nick The_Seven is now known as TheSeven (
04:10:12 Join goffa_ [0] (n=goffa@
04:16:43 Join CaptainKwel [0] (
04:35:01soapgrepped CURRENT_BACKLIGHT to see how it is used. A shocker is that the mrobe 500's CURRENT_NORMAL = 100? abd CURRENT_BACKLIGHT is another 100. That is one dang inefficient player.
04:37:14soapand is it my understanding that CURRENT_BACKLIGHT is _only_ used if you have the backlight set to always on?
04:37:50soap(though, I guess it would be hard to use it in some other way - at least I can't think of a good way to use it)
04:43:57 Quit daurn (Read error: 113 (No route to host))
05:03:17***Saving seen data "./dancer.seen"
05:10:31 Join n00b2Hack [0] (n=Administ@
05:13:06 Join android6011_ [0] (n=chatzill@
05:16:02 Join AndyIL [0] (i=AndyI@
05:16:12 Part n00b2Hack ("Saliendo")
05:29:26 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
05:29:40 Quit AndyI (Read error: 110 (Connection timed out))
05:29:43 Join Horschti [0] (n=Horscht2@xbmc/user/horscht)
05:47:52 Quit Horscht (Read error: 110 (Connection timed out))
05:53:49 Quit AndyIL ()
05:56:13 Join AndyI [0] (i=AndyI@
06:06:06 Quit jfc (Read error: 104 (Connection reset by peer))
06:21:23 Join jfc [0] (
06:46:15 Quit CaptainKwel (Remote closed the connection)
07:03:21***Saving seen data "./dancer.seen"
07:13:59 Quit android6011_ ("ChatZilla 0.9.85 [Firefox 3.5.1/20090715094852]")
07:36:56 Join Horscht86 [0] (
07:55:26 Quit Horschti (Read error: 110 (Connection timed out))
07:56:44 Join BdN3504 [0] (n=5ce225de@gateway/web/cgi-irc/
08:01:11 Join n1s [0] (n=n1s@rockbox/developer/n1s)
08:01:26BdN3504which version of the gigabeat s supports dualbooting? i have 2.11 on the device
08:02:06 Join Horschti [0] (n=Horscht2@xbmc/user/horscht)
08:20:03 Join bertrik [0] (
08:20:40 Quit Horscht86 (Read error: 110 (Connection timed out))
08:36:36 Join Rob2223 [0] (
08:38:32BdN3504how do i compile mknkboot?
08:38:39BdN3504do i have to add it to sources? if so where?
08:40:08 Quit BdN3504 ("CGI:IRC")
08:45:08 Quit pixelma (Nick collision from services.)
08:45:10 Join pixelma_ [0] (i=quassel@rockbox/staff/pixelma)
08:45:30 Nick pixelma_ is now known as pixelma (i=quassel@rockbox/staff/pixelma)
08:45:53 Quit amiconn (Nick collision from services.)
08:45:57 Join amiconn_ [0] (i=quassel@rockbox/developer/amiconn)
08:46:04 Nick amiconn_ is now known as amiconn (i=quassel@rockbox/developer/amiconn)
08:54:39 Quit Rob2222 (Read error: 110 (Connection timed out))
09:03:24***Saving seen data "./dancer.seen"
09:16:23 Join stoffel [0] (
09:18:43 Quit thegeek (Read error: 54 (Connection reset by peer))
09:30:02 Join DarkDefender [0] (
09:35:15 Join thegeek [0] (
09:41:00 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
09:48:32 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
09:50:55 Quit advcomp2019_ (Read error: 104 (Connection reset by peer))
09:51:57 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
09:56:16 Join Grahack [0] (
09:59:02 Quit advcomp2019 (Read error: 113 (No route to host))
10:04:34 Quit DarkDefender ("Leaving")
10:06:35 Join mt [0] (n=mt@
10:06:43 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
10:11:14 Quit Grahack ("Leaving.")
10:17:18 Quit TheSeven (Read error: 104 (Connection reset by peer))
10:17:42 Join TheSeven [0] (n=theseven@
10:19:27 Quit thegeek (Read error: 104 (Connection reset by peer))
10:23:28 Join thegeek [0] (
10:27:04 Join DarkDefender [0] (n=rob@
10:33:25 Nick Horschti is now known as Horscht (n=Horscht2@xbmc/user/horscht)
10:34:50 Join advcomp2019__ [0] (n=advcomp2@unaffiliated/advcomp2019)
10:46:13 Quit advcomp2019 (Nick collision from services.)
10:46:15 Nick advcomp2019__ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
10:47:15 Quit thegeek (Read error: 110 (Connection timed out))
10:49:11 Join Thundercloud [0] (
11:02:58 Join kugel [0] (n=kugel@rockbox/developer/kugel)
11:03:25***Saving seen data "./dancer.seen"
11:19:00 Join stripwax [0] (
11:37:19 Join Lear [0] (n=chatzill@rockbox/developer/lear)
11:38:13 Join ender` [0] (
11:38:42 Join daurn [0] (n=daurnima@unaffiliated/daurnimator)
11:45:46stripwaxmt - in libcook decode_vectors, how often do we hit the 'index>=7' part, do you think? Rather than memset the subband vectors to zero, could we just avoid/simplify the entire call to scalar_dequant in that case
11:48:14stripwaxalso, (although I won't pretend to understand cook codec), that for loop to initialize category[band+j] to 7 should probably be a memset
11:53:22stripwaxmt - did you see the profiles by the way? scalar_dequant_math seems to take the most cycles after the mdct itself.
11:54:53mtstripwax: Yes I did. It's just that I'm currently focusing on adding AAC rather then optimizing cook.
11:55:40stripwax:) no worries, just wondering if you had any thoughts on whether the amount of time spent by each part made sense or not, anything stand out
11:56:57mtThanks for making those profiles by the way. :)
12:08:56 Quit stoffel (Read error: 113 (No route to host))
12:10:17 Quit stripwax ("")
12:21:49 Quit kugel (Read error: 110 (Connection timed out))
12:33:11 Quit linuxstb (Read error: 113 (No route to host))
12:38:31 Join wincent [0] (n=wincent@rockbox/developer/wincent)
12:42:05 Join stoffel [0] (
13:03:29***Saving seen data "./dancer.seen"
13:07:47 Join Lear_ [0] (i=chatzill@rockbox/developer/lear)
13:12:23bertrikI made a little more progress with the samsung yp-s3
13:14:17bertrikBeing able to display debug info on the screen helps alot. I'm able to communicate with the RTC, reading the WM codec somehow still results in a hard hang. Can't read the fmradio either yet. ADC does work now and I think I can read the USB input voltage and battery voltage
13:15:14bertrikThe backlight enable GPIO is also a timer output, maybe I can put a PWM signal on that for backlight dimming.
13:15:34Lear_Darn, there's a small memory leak in the vorbis codec...
13:20:23 Join barrywardell [0] (
13:25:51 Quit Lear (Read error: 113 (No route to host))
13:31:18 Quit Thundercloud (Remote closed the connection)
13:34:26TheSevenbertrik: are interrupts working?
13:34:50TheSevenif not, everything accessing i2c (e.g. wmcodec and fmradio) will lock up
13:39:26 Quit stoffel (Read error: 113 (No route to host))
13:43:27bertrikTheSeven, no interrupts are not working yet, so I by bypassed the interrupt stuff
13:44:02TheSevenok, so you are polling that flag?
13:54:08 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
13:56:58 Nick YpsyZNC is now known as Ypsy (
14:07:09bertrikI'm polling the interrupt flag in the device (i2c or adc), not in the interrupt controller
14:10:49 Part wincent ("Kopete 0.12.7 :")
14:11:45bertrikI can read the rtc over the i2c bus, but somehow I can't read other devices on the same bus yet. Reading wmcodec locks up, reading the si470x tuner just returns 0xFF bytes
14:12:10bertrikWe don't check any ACKs in the current i2c driver, so it could be that it simply does not respond at all
14:13:17bertrikI guess there are some enable pins that actually power up some of the peripherals. I don't know if this indeed the case, and if so, what pins these are on.
14:14:22 Nick fxb__ is now known as fxb (
14:26:30 Quit linuxstb (Read error: 110 (Connection timed out))
14:27:31 Join BdN3504 [0] (n=5ce22508@gateway/web/cgi-irc/
14:27:57 Join einhirn [0] (
14:29:57BdN3504i am a bit puzzled by the gigabeatsinfo wiki page.
14:30:21BdN3504it reads: Patched Gigabeat V updater (download below)
14:30:40BdN3504but if you follow that link
14:31:03 Join benutzer [0] (
14:31:11BdN3504it reads: Unpatched version of the Gigabeat V firmware updater:
14:32:04 Join teru [0] (
14:32:11BdN3504so it first reads "patched" and then "unpatched". Do i simply have to get the version provided there
14:36:04 Quit einhirn (Read error: 54 (Connection reset by peer))
14:39:05CIA-7New commit by teru (r22027): fix some bugs reported in FS #9707 (some of fix is from the patch by Yoshihisa Uchida). see the task for more details. ...
14:41:00 Quit BdN3504 ("CGI:IRC")
14:56:02 Join shodanX_ [0] (
14:56:43 Join borges_ [0] (
15:01:00 Quit shodanX (Read error: 110 (Connection timed out))
15:03:08 Quit bmbl (Read error: 104 (Connection reset by peer))
15:03:31***Saving seen data "./dancer.seen"
15:05:19 Join efyx_ [0] (
15:05:28 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
15:08:41 Join stoffel [0] (
15:14:28 Quit borges (Read error: 110 (Connection timed out))
15:17:10 Quit coney (Read error: 110 (Connection timed out))
15:22:12 Join BdN3504 [0] (n=5ce22508@gateway/web/cgi-irc/
15:25:29 Join einhirn [0] (
15:26:17BdN3504i really need your help... i don't know where to get the nk_orig.bin for the gigabeat s. do i have to simply rename it to that and take the nk.bin form the /b unpatched /b gigabeat updater provided on the wiki?
15:27:18n1syeah, pretty sure
15:28:59BdN3504ok so now i have that and patch it with mknkboot, right? how do i know if the patching has been successful?
15:29:14n1stry it? ;)
15:29:47n1sdon't remember really, maybe someone who does can fill in?
15:30:12BdN3504well if i take the nk.bin created with mknkboot and the send it via sendfirm, the player reboots
15:30:30BdN3504but it doesn''t load the rockbox bl
15:30:49BdN3504but gets tuck at the error message 3 of the original fw
15:31:36 Join IuDeX [0] (n=52a0f8f7@gateway/web/cgi-irc/
15:31:37IuDeXtake a look on this;topicseen#msg153813
15:31:46IuDeXlook on this ;);topicseen#msg153813
15:31:48n1snever got that does it work with a clean OF?
15:32:01IuDeXouch sorry for twice msg
15:32:05BdN3504lemme try
15:33:05BdN3504can i simly use the nk_orig from the wiki or will this do something bad?
15:33:44 Join fdinel [0] (
15:34:11gevaertsBdN3504: the very worst those things will do is asking to be restored again
15:34:49BdN3504that's what's happening :) but instead of error message 3 i'm stuck at 2: "connect to pc"
15:35:41BdN3504if i try to sendfirm an unpatched version of the rockbox bl, i don't have dual boot, right?
15:36:47BdN3504and that means i cannot charge the device as of now, right?
15:37:05gevaertsrockbox does charge the S
15:37:48 Quit IuDeX ("CGI:IRC")
15:38:24BdN3504oh, so i don't need dual boot actually? i thought charging doesn't work?
15:39:25gevaertsyou thought wrong
15:40:32BdN3504cool thanks!
15:42:25BdN3504killer, i am now in rb usb mode. now i have to fix the partition table and i'm ready to go. yippie!
15:47:18 Join Typhoon31 [0] (i=5dddee85@gateway/web/freenode/x-9a88bc31a1d7fdbd)
15:47:32 Quit Zarggg ()
15:48:53 Quit borges_ (Read error: 110 (Connection timed out))
15:49:29 Part Typhoon31
15:49:35BdN3504could you help me define the mountpo9int of the device? i can't make anything out of the dmesg output...
15:55:17 Quit TheSeven ("ChatZilla 0.9.85 [Firefox 3.5.1/20090715094852]")
15:56:11 Join TheSeven [0] (
15:57:27gevaertshm, are you using ubuntu?
15:58:48Lear_I do believe I have a working malloced Vorbis decoder now. Quick tests in the simulator look good at least. :)
15:58:58BdN3504nope, i am using the vmware image provided on the wiki
15:59:17Lear_Any suggestions for the location of TLSF?
15:59:49BdN3504but i mounted sda to a dir and now have a different dmesg output:
15:59:57BdN3504guess i can now simply use sda rightß
16:00:31BdN3504fdisk /dev/sda
16:00:45BdN3504fdisk /dev/sda1
16:01:24gevaertsAlso, if this is a recent bootloader, you shouldn't need to fix the partition table anymore
16:01:44BdN3504hm, i just compiled it with a recent revision
16:02:15gevaerts/dev/sda is the disk, /dev/sda1 is the firmware partition where the bootloader lives, /dev/sda2 is the main partition where .rockbox lives
16:02:43BdN3504ah ok
16:04:25BdN3504right, it shows up normal as two devices in windows. can i delete all the files on the big partition (content and firmware directory and a bunch of nondescript files) and just put rb on it, or do i need these files?
16:11:31 Join mcuelenaere [0] (
16:13:56 Quit BdN3504 ("CGI:IRC")
16:45:10 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
16:50:35 Join saratogahome [0] (n=41becb3b@gateway/web/cgi-irc/
16:50:56saratogahomeLear_: which malloc did you end up using?
16:53:15 Quit mcuelenaere ()
16:53:55 Join w1ll14m [0] (
16:54:09Lear_saratogahome: TLFS, since PDBox used it...
16:56:03bertrikI'm not really into this, but what did we use before and how has it improved now?
16:58:39bertrikah, you must be Mr Someone! :P I just read on his todo list : "Introduce a real malloc for tremor so that we don't waste hundreds of KB on unfree'ed buffers."
17:03:35***Saving seen data "./dancer.seen"
17:04:48n1svorbis used a pretty primitive kind-of-malloc before
17:04:50Lear_It wasn't as bad as several hundreds when I tested it now though...
17:05:35n1sLear_: how big are the savings?
17:08:05Lear_What kind of savings? But before there were 50-80 kB of unfreed buffers.
17:08:25 Join arohtar [0] (n=faemir@
17:08:41Lear_And TLSF reported about 380 kB of peak usage, for a floor0 file (I think...)
17:08:43n1smemory savings, i suppose speed should be pretty much unaffected?
17:10:29Lear_It (almost) only allocates during init, so any additional time ought to be really small in comparison.
17:10:45gevaertscan someone who understands perl have a look at and see if I'm being totally stupid, or if this actually has a chance to work?
17:12:38saratogahomeLear_: I can test again using an extreme example file
17:12:48saratogahomeit was several hundred KB more then typical IIRC
17:13:35Lear_Ok, I can make a patch.
17:13:53CIA-7New commit by mt (r22028): Compile librm separately in preparation for addition of more codecs.
17:14:19Lear_Should I add a (simulator only) debug with peak memory usage perhaps?
17:15:10n1sLear_: i owndered about memory savings since iirc a motivation for a real malloc was to be able to shrink the 1MB codec buffer
17:15:41n1sI think a peak memory diagnostic would be good
17:21:32 Quit teru ("Quit")
17:24:46 Join Zarggg [0] (
17:25:13bertrikyes, nice idea
17:25:53CIA-7New commit by gevaerts (r22029): Add the -msgscript parameter, which allows to specify a script to run whenever a server message is received.
17:26:22gevaertsok, now let's hope that Bagder and Zagor don't dislike this too much :)
17:27:22 Join n00b81 [0] (n=n00b81@unaffiliated/n00b81)
17:27:30 Part n00b81 ("Leaving")
17:27:42 Quit barrywardell ()
17:32:11gevaertshm, that's probably not good enough...
17:36:38CIA-7New commit by gevaerts (r22030): rename msgscript to commandhook, and run it for all server commands instead of just for messages
17:37:23gevaertsrasher: that should allow you to do this xterm title script
17:39:53 Join BdN3504 [0] (n=5ce22508@gateway/web/cgi-irc/
17:41:17BdN3504argh another problem with the gigabeat s: i have succesfully used the sendfirm tool to copy a bootloader compiled from the latest build onto the device under windows.
17:41:49BdN3504but the device does not restart automatically after the success. instead i have to disconnect it and then it boots into rockbox.
17:42:53BdN3504but it only boots into rockbox, when it is supplied with power from the adapter. if i detach the adapter, the error screen 3 is shown, asking me to be patient
17:43:55BdN3504then i can reconnect the device, after i pulled the plug and reset the battery. then i have to use sendfirm again to boot into rockbox... what to do?
17:47:05 Join CaptainKwel [0] (
17:47:17CIA-7New commit by mt (r22031): Adding support for playback of aac audio in rm container, with seeking.
17:52:13saratogahomeLear_: I have company in town this weekend but i'll try to get a look at it soon
17:52:26saratogahomeregardless I think this is a big improvement to vorbis and should go in when its ready
17:52:38saratogahomei expect it will be very helpful for small memory targets like AMS
17:54:08CIA-7New commit by mt (r22032): Fix yellow - (Value in an enumeration not handled in a switch).
17:58:31saratogahomeinteresting to see that Palm has borrowed our idea of identifying as an ipod via software USB stack
17:58:44saratogahomespeaking of which did that work ever get commited?
17:59:22gevaertsno. It was still too fragile
18:00:31CIA-7New commit by gevaerts (r22033): use system() to run the commandhook script, so it can interract with stdout and stderr (needed for e.g. xttitle to work)
18:01:36saratogahomeis it on the tracker somewhere?
18:02:59gevaertsthe patches are on FS #10023. There's also information in the wiki (IpodItunesCommunication)
18:03:47gevaertsI'd mainly like to see the last patch there to be finished (clock setting). The gtkpod people have a tool to set the clock from linux that we can then borrow
18:05:14 Join darkless [0] (
18:09:03n1swe (well, Mr Someone) could hack that functionality into rbutil!
18:09:31 Join Thundercloud [0] (
18:10:03gevaertsexactly. What stopped me from finishing this was the lack of an easy way to actually set the clock based on the data we get from itunes
18:10:04 Quit saratogahome ("CGI:IRC (EOF)")
18:13:47 Join r0b- [0] (
18:15:46 Quit Thundercloud (Read error: 104 (Connection reset by peer))
18:15:52 Quit einhirn (Read error: 101 (Network is unreachable))
18:17:19 Join shotofadds [0] (n=rob@rockbox/developer/shotofadds)
18:25:16shotofaddsI want to add a line to the panic and UIE handlers to turn on the backlight for LCDs that are unreadable without it (eg. with a direct call to _backlight_on()). Obviously this would require an #ifdef, but the best I can come up with is LCD_REQUIRES_BACKLIGHT. Also, which targets apart from the D2 need this?
18:26:31gevaertsthe sansas are unreadable as well. Not sure about the ipods, the 5G is readable, but I don't know about nano and color
18:27:30shotofaddsis that the case for v1s and v2s?
18:27:36shotofaddssansas, that is
18:27:50gevaertsat least the v1. I suspect the v2s as well though
18:28:30TheSevennano2 is, well, kind of readable if you hold it into direct light, but not really good
18:28:52shotofaddsWell, I'll add it for the targets we know need it and others can be added later.
18:28:58n1swell, most color lcd's are unreadable without backlight and it doesn't hurt on the rest so i think it should be unconditional for players that have backlight
18:29:17shotofaddsI know there was some discussion of LCD_IS_KINDA_UNREADABLE_WITHOUT_REALLY_STRONG_DIRECT_SUNLIGHT before ;-)
18:29:33n1sbut why not do it for all players?
18:29:34shotofaddsn1s: well, that would be even easier
18:29:47gevaertswell, those that do have a backlight :)
18:29:51shotofaddsI can't really think of any reason not to
18:32:57 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
18:35:40 Quit Lear_ ("ChatZilla 0.9.85 [Firefox 3.5.1/20090715094852]")
18:42:46shotofaddsn1s: of course the UIE handler is arch-specific, and I can't see the point of adding this for coldfire/SH if those targets don't need it, other than for consistency. panic is common to all, though
18:46:07 Join infzy [0] (
18:46:39infzyany recommendations for a good PMP that runs rockbox?
18:49:22 Join thegeek [0] (
18:57:09 Join darkhamm [0] (
19:00:49 Quit w1ll14m (Read error: 110 (Connection timed out))
19:02:59BdN3504aaah. now i can't even boot rb anymore. i always get the same please be patient error message 3 on my gigabeat s...
19:03:24*gevaerts thinks that BdN3504 will soon understand why we don't call the gigabeat s supported yet
19:03:29BdN3504i am using a freshly compiled bootloader
19:03:39***Saving seen data "./dancer.seen"
19:03:48BdN3504gevaerts: true, true
19:07:20 Join daurnimator_ [0] (
19:08:07 Quit BdN3504 ("CGI:IRC")
19:09:51 Quit dfkt (Read error: 104 (Connection reset by peer))
19:11:35bertrikIf it weren't for this proprietary NAND flash translation layer, I think a lot of the s5l8700 based players would be easy rockbox targets
19:12:39 Join w1ll14m [0] (
19:12:43 Join daurnimator__ [0] (
19:13:27 Join derekja [0] (
19:13:32 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
19:13:35 Part derekja
19:13:48 Join n00b81 [0] (n=n00b81@unaffiliated/n00b81)
19:13:56 Part n00b81 ("Leaving")
19:18:30 Quit daurn (Read error: 110 (Connection timed out))
19:26:16 Quit daurnimator_ (Connection timed out)
19:26:37 Join mcuelenaere [0] (
19:34:39 Quit stoffel (Read error: 113 (No route to host))
19:35:34 Join mcuelenaere_ [0] (
19:35:34 Quit mcuelenaere_ (Client Quit)
19:36:29 Join mcuelenaere_ [0] (n=mcuelena@rockbox/developer/mcuelenaere)
19:36:53 Quit mcuelenaere (Nick collision from services.)
19:37:07 Nick mcuelenaere_ is now known as mcuelenaere (n=mcuelena@rockbox/developer/mcuelenaere)
19:39:28 Join BryanJacobs [0] (
19:40:38 Quit gevaerts (Read error: 60 (Operation timed out))
19:40:38 Quit arohtar (Client Quit)
19:40:48 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
19:42:05 Quit benutzer ("Verlassend")
19:46:43 Join Guest84819 [0] (
19:47:53 Join stoffel [0] (
19:49:20Guest84819hey! the dude dave.b that made the x5 cloak theme is god!
19:50:10Guest84819i've been thinking of do one just like it, and now i don't have to ;)
19:50:30 Part Guest84819
19:54:38 Nick Ypsy is now known as YpsyZNC (
19:58:39 Join daurnimator_ [0] (
19:59:39 Quit bertrik ("De groeten")
20:03:39 Quit daurnimator__ (Read error: 60 (Operation timed out))
20:06:23r0b-for a microsd card will rockbox only read Fat32 or does it not matter?
20:07:08gevaertsfat32 or fat16
20:07:26r0b-oh ok
20:12:56mcuelenaeredoesn't that depend on whether Rockbox is compiled with FAT16 support?
20:14:23gevaertsyes, but as far as I know all targets that have a microsd slot are
20:14:36gevaertsand if they aren't, I think that would be a bug
20:15:24 Quit darkhamm ("Sto andando via")
20:16:16 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
20:20:11w1ll14mdamn... that bin2note exploit is awesome
20:23:09CIA-7New commit by rob (r22034): Make the panic and UIE handlers turn on the backlight (needed for various colour LCDs that are unreadable without it). UIE is only changed on ARM for ...
20:24:39 Join einhirn [0] (
20:27:27webmindw1ll14m, inded
20:27:59w1ll14mit's just to bad i have a nano with type2 lcd
20:29:05*webmind has no nano at all anymore
20:36:54 Quit daurnimator_ (Read error: 110 (Connection timed out))
20:37:37 Join daurnimator_ [0] (
20:37:50infzyany recommendations for a good PMP that runs rockbox?
20:38:41r0b-or a 1st gen nano
20:38:45infzyo rly?
20:38:57Tuplanolla+1 for e200
20:39:09r0b-Tuplanolla do oyu own the e200 or just saying that
20:39:21Tuplanollai've got a pair of e280s
20:39:29r0b-ah ok
20:39:40r0b-im happy with my single e250
20:39:56infzyso sansa's are thought to be pretty happy
20:41:06dze200 and c200 v1s, yes
20:42:27infzybut not the newer models
20:42:33CIA-7New commit by rob (r22035): Revert r22034 since the backlight handling functions are not consistently named between targets.
20:42:52infzyis that because noone's ported it yet, or because it can't be done?
20:42:53infzyjust curious :)
20:43:18r0b-the ports are work in progress
20:43:34infzyokay, cool.
20:45:01CIA-7New commit by rob (r22036): D2: Fix the POWER button check after an exception/panic. Make this power the player off, since we don't know how to force a reboot yet.
20:46:36infzyhas anyone tried a port to the visualland stuff?
20:46:40infzysorry for annoying noobie questions :)
20:47:38r0b-infzy check out
20:48:02infzyyeah, ok ;) srry :)
20:48:50*shotofadds thinks the red on r22035 looks a little iffy (blank log)
20:53:05gevaertsthat was on cygwin. I guess there's still something strange with those
20:54:52 Part BryanJacobs ("Leaving.")
20:58:08 Join flydutch [0] (
21:01:10 Quit mcuelenaere (Read error: 110 (Connection timed out))
21:03:42***Saving seen data "./dancer.seen"
21:05:17gevaertsIs it expected that you have to rebuild the entire pictureflow cache to add a few covers?
21:09:06 Join kugel [0] (n=kugel@rockbox/developer/kugel)
21:11:16TheSevenw1ll14m: rockbox can handle type-2 lcds
21:11:25TheSevenit's just ibugger that can't
21:11:40TheSevenso you'll only get no splashscreen, but the rest will work
21:12:41 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
21:13:30 Join tinker_ [0] (
21:14:05tinker_Hi. Will Rockbox work on a Sansa Sandisk m260 ?
21:15:17tinker_LOL. I guess that settles a lot of other questions
21:15:23tinker_Thanks much
21:15:32 Quit tinker_ (Client Quit)
21:16:52 Quit stoffel (Remote closed the connection)
21:17:17 Nick fxb is now known as fxb__ (
21:23:54w1ll14mTheSeven: i was just playing around with iBugger, not tried anything else yet
21:24:31TheSevenw1ll14m: everything besides the splash screen will work
21:25:03TheSevenand the updatelcd command, of course
21:25:25TheSevenw1ll14m: which version are you using?
21:29:58 Quit advcomp2019 (Read error: 113 (No route to host))
21:30:01 Quit flydutch (Read error: 104 (Connection reset by peer))
21:31:45 Join flydutch [0] (
21:32:00w1ll14mi had a strange screen once, but i cannot get it to reproduce
21:36:44w1ll14mTheSeven: upload 0x22000000 bootloader.bin then execute 0x22000000 0
21:37:05w1ll14mTheSeven: gives me +++ this is a very long i=xxx
21:38:39TheSevenyep, that's the current rockbox bootloader :-)
21:40:00w1ll14mok so it went well
21:41:41 Quit advcomp2019_ (Read error: 110 (Connection timed out))
21:45:09 Nick aidy_ is now known as aidy (
21:59:34 Quit n1s ("Lämnar")
22:13:41 Quit bmbl ("Bye!")
22:15:31 Join sinthetek [0] (
22:16:40sinthetekshould themes affect device configuration/settings? or appearance only?
22:18:49sinthetekfor some reason next and previous buttons won't change tracks. they only ff and rw the current track no matter how long i press them
22:21:02sinthetek(but they work fine with other themes. seems that only the 'free-state.amremix' theme is the only one with this issue
22:21:27gevaertssinthetek: they could change settings in theory, but they're not supposed to do that. Button mappings are not configurablt though
22:21:50gevaertsis that theme on the theme site?
22:22:46*gevaerts finds it
22:23:15gevaertsit doesn't do anything special
22:24:09sinthetekhmm... weird
22:24:22sinthetekwell, it's just a slight nuissance. i'll live
22:25:41gevaertsmaybe it's slightly more CPU intensive so it detects long press easier?
22:30:53sintheteksounds feasible i suppose
22:31:23sinthetekmaybe it is because i am using version 3.3 and it only purports to work with 3.2
22:32:38gevaertsit says 3.2 because of technical issues with the automated checking on the theme site. If it didn't work with 3.3, it wouldn't actually load
22:34:15sinthetekahh ok then. i guess your theory seems most plausible then
22:37:04 Join BdN3504 [0] (n=5ce22508@gateway/web/cgi-irc/
22:47:37 Join BeChris [0] (i=5c89f62b@gateway/web/freenode/x-4249be9ee2c9836a)
22:51:56BeChrisI'd like to know who could evaluate and include officialy what I did in FS #9493 (png viewer)
22:54:14gevaertsthere was a discussion about it a few days ago
22:55:18gevaertssome people want a unified image viewer first, others think that requiring that is madness
22:55:25BeChrisOK thanks I will search in the log
22:56:37Unhelpfulmy "vote" would be for inclusion
22:56:57gevaertsmine too. Let's do a quick review and stop this waiting :)
22:57:08Unhelpfulis the png viewer capable of using the scaler?
22:57:24gevaertsBeChris: I assume a lot of the patch is mostly-unmodified libpng?
22:57:45Unhelpfuli'd say the main questions are 1) can it use the scaler? 2) does it support greylib?
22:58:01Unhelpfulboth are fairly important if we want one integrated image viewer
22:58:27gevaertsyes, but atr devcon2008 we decided that "does not run on all targets" should not be a blocker
23:00:03BeChrisUnhelpful : I'm a quite newcomer in the rockbox world : what do you mean by the scaler ? I use bmp_slooth_scale function.
23:00:07BdN3504hey guys i want to point people who have problems with the gigabeat s installation in the right direction:
23:00:08Unhelpfulactually, #2 isn't so important, the decoder doesn't need to know about greylib.
23:00:39BdN3504after having patched the firmware provided on the wiki, i had all sorts of problems
23:01:07BdN3504what solved was: download the 12 us updater (llink provided in the manual)
23:01:07UnhelpfulBeChris: ideally, scaling should be done by providing a callback function that the core scaler can use to fetch rows. see apps/recorder/resize.c, bmp.c, jpeg_load.c
23:01:40BeChrisUnhelpful : then this is not the case :(
23:01:51gevaertsdo we need that before commit though?
23:02:11BeChrisBut someone else could maybe implement it (or me I have time and skill to do so)
23:02:21BdN3504extract nk.bin from that exe, patch it with mknkboot and use sendfirm to put it on the device. Works like a charm. Thank you guys for making rockbox possible on such an outstanding piece of hardware
23:02:23BeChrisI meant If i have ...
23:02:59Unhelpfulgevaerts: not really. it's a bit of work. i've pondered turning that situation around a few times, so that only the *scaler* needs to encapsulate its state in a struct, and the readers can then call it to process each line they've decoded.
23:03:44gevaertsBeChris: the v8 patch seems to miss the apps/plugins/SOURCES bit. I assume it should be for HAVE_LCD_COLOR?
23:03:45***Saving seen data "./dancer.seen"
23:03:52 Quit infzy ("leaving")
23:03:54BeChrisThis is how the official libpng could be used but my tests have shown that it is slower than my implementation (a mix between LodePng and zlib)
23:04:04Unhelpfulit would make porting new decoders to the scaler much less ugly.
23:04:34BeChrisgevaerts : really ! I always forget files in my patch.
23:04:54BeChrisOf course it is for HAVE_LCD_COLOR (due to the fact I use bmp_smooth_scale)
23:08:38BeChrisgevaerts : no, there is no problem with the patch since apps/plugins/SOURCES is no more impacted. It's apps/plugins/SUBDIRS which is modified now.
23:08:48gevaertsyes, I see that now
23:09:37Unhelpfulencapsulating the scaler state in a sructure is tedious drudge-work, about on the same order as the core jpeg decoder was... but it would make supporting the scaler for other uses much easier, i think.
23:09:59 Join [1]w1ll14m [0] (
23:10:35 Join _zic [0] (
23:11:49 Join ademille [0] (
23:11:52 Quit _zic (Client Quit)
23:12:10BeChrisSo, do you think guys this patch has any chance to be officialy included leaving jpeg one untouched for the moment ?
23:12:32UnhelpfulBeChris: i think it should be.
23:12:37gevaertsunless someone speaks up soon, I plan to commit it tonight
23:13:00 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
23:13:00Unhelpfulthe scaler bit is something i need to do myself, as soon as i can work myself up to doing it. :)
23:13:08BeChrisYes ! Very happy. I will see my name in the credits soon :)
23:14:20*gevaerts wonders why he gets unrelated warnings
23:15:00BeChrisDoes it mean I will get write access to the svn repository in order to work on the unification ?
23:15:25Unhelpfuldepends, how do you feel about rewriting the scaler? ;)
23:16:11BeChrisMaybe too much chalenging for me at the moment :(
23:18:37BeChrisgevaerts : I think I should post another patch because I found that patch_png8 accidentaly modify the CFLAGS for all sources not only the png plugin
23:18:53gevaertsah, that would explain it
23:19:18gevaertscan you just tell me what to change?
23:19:23 Quit kugel (Read error: 110 (Connection timed out))
23:19:34Unhelpfuli'm kidding, and it's also not up to me :)
23:19:51BeChrisgevaerts : ok or can I send you the correct png.make file ?
23:20:11gevaertsyes, maybe on a pastebin
23:20:45BeChrisgevaerts : another tech I'm not aware of
23:21:03gevaertshave a look at and all shall become clear
23:21:41 Join kugel [0] (n=kugel@rockbox/developer/kugel)
23:24:26BeChrisgevaerts : I found sourcepod and you can find the file here :
23:25:00 Quit [1]w1ll14m (Read error: 60 (Operation timed out))
23:26:19 Quit w1ll14m (Read error: 110 (Connection timed out))
23:28:03gevaertsBeChris: that one seems buggy. I don't seem to find a way to get at the raw source... is excellent in that regard :)
23:29:02gevaertsit also has bugs
23:29:07gevaertsless, but still
23:29:34Unhelpfulit doesn't break patches like :)
23:29:53gevaertsah, wait, if I ask to edit the thing, I can copy it properly
23:30:25Unhelpfulhrm, actually, it seems i can properly download it raw
23:30:50BeChrisOK : I was studying
23:31:00gevaertsBeChris: do so anyway :)
23:31:13BeChrisruby rulez :)
23:32:49BeChrisIf I give you this URL :
23:32:53BeChrisIs it ok for you ?
23:33:22gevaertsit is, yes. just needs .txt a the end for the raw content
23:33:57 Join Zagor [242] (
23:34:01*gevaerts does a quick test
23:34:05BeChrisAh right. Interresting tool.
23:34:12 Join n00b81 [0] (n=n00b81@unaffiliated/n00b81)
23:36:15CIA-7New commit by gevaerts (r22037): Add a png viewer plugin ...
23:36:37*gevaerts hopes for all green
23:38:22BeChrisThanks gevaerts
23:39:10Bagdernice work BeChris
23:40:31 Quit BdN3504 ("CGI:IRC")
23:41:55BeChrisGlad to see that my first contribution to rockbox is officially integrated
23:42:00BeChrisThanks to you all
23:42:06gevaertsBeChris: there are some warnings
23:42:35gevaertsI guess 64bit sims
23:42:53BeChrislet me see
23:43:35Bagder51 clients
23:43:43kugelnice \o/
23:43:59kugelI mean, the png viewer. but the 51 clients are also nice :)
23:44:05gevaertsBagder: I don't know if you notices, but we've had a few sub-4-minute builds today :)
23:44:18BagderI spotted that, it's great!
23:44:56Bagder118 builds in less than 4 minutes is unbelievably cool
23:45:06kugelmulti-threading dependency generation really boosted things up
23:45:29CIA-7New commit by gevaerts (r22038): Fix some casts from pointer to int, now cast to intptr_t
23:46:23kugelBeChris: nice work
23:46:42pixelmasinthetek: did you change your track skip options accidentally maybe or turn on "prevent track skipping"? Not sure where this option is now but there is one called this way
23:47:58 Part n00b81 ("Leaving")
23:48:16kugelpixelma: I asked you yesterday but you probably were already gone. Do you have an idea what to do about the X5 w.r.t. to playlist viewer button. Also, does the M5 have the same problem?
23:49:05BeChrisI don't get the warning on my iMac when building a creativezen M30 simulator (gcc 4.0.1)
23:49:36gevaertsI think I fixed it anyway
23:50:24gevaertsZagor, Bagder: do you think this new -commandhook option should get a new client version?
23:50:56Bagdernah, all those you really want to play with it atm can easily just get it
23:51:31Bagderor has anyone suffered?
23:51:36gevaertsyou mean all two of us? :)
23:52:05Bagderwell, once you release your awesome cool scripts I expect the entire world to want them! ;-)
23:52:19pixelmakugel: answer is in the logs, somewhere between 23 and 24 yesterday
23:53:27gevaertsok, for posterity, I run
23:53:59 Quit flydutch ("/* empty */")
23:56:34sintheteki did enable prevent track skipping. i thought it was to help with distortions from file corruptions or rips from bad cds, heh
23:57:32pixelmaa dap is not a record player (I mean those vinyls) ;)

Previous day | Next day