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 2010-09-29

00:01:04brickshello, the 3.6 firmware for sansa clip v1 is giving me a 404 error.
00:02:33bricksIs there a mirror somewhere, or should I just go for the current build (I'm doing a new install)?
00:04:40ender`when 3.6 was released, clip wasn't considered stable yet, so the build doesn't actually exist - use the current build instead
00:04:58brickscool, thanks
00:07:16 Quit MagusG (Ping timeout: 265 seconds)
00:07:33 Quit Staphylo (Quit: Bye les gens =))
00:11:19 Quit bricks (Quit: CGI:IRC 0.5.9 (2006/06/06))
00:13:14 Quit jgarvey (Quit: Leaving)
00:14:18wodzpixelma: (for logs) I uploaded new version of the patch to FS #11641. This time it should apply cleanly.
00:19:10 Quit liar (Ping timeout: 240 seconds)
00:20:37 Quit wodz (Quit: Leaving)
00:23:05 Join liar [0] (
00:29:45 Join drizztbsd_ [0] (~quassel@
00:31:00 Join drizztbsd__ [0] (
00:33:01 Quit drizztbsd (Ping timeout: 272 seconds)
00:34:55 Quit drizztbsd_ (Ping timeout: 272 seconds)
00:35:17 Quit domonoky1 (Read error: Connection reset by peer)
01:01:16 Quit ender` (Quit: printk("; corrupted filesystem mounted read/write - your computer will explode within 20 seconds ... but you wanted it so!\n"); -- /usr/src/linux/fs/hpfs/super.c)
01:03:47 Join Strife89TX [0] (
01:05:15 Join Topy [0] (
01:05:24saratoga_we should release 3.7 so people stop asking about 3.6 on the clip!
01:08:27 Nick fxb is now known as fxb__ (
01:08:31 Quit T44 (Ping timeout: 245 seconds)
01:14:51 Quit kugel (Ping timeout: 240 seconds)
01:21:41 Quit esperegu (Read error: Connection reset by peer)
01:22:19 Join edboyer93 [0] (
01:31:58 Join freedrull [0] (
01:35:41 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
01:40:09 Quit Nausicaa (Ping timeout: 272 seconds)
01:47:55 Part freedrull
01:49:22 Quit DerPapst (Quit: Leaving.)
01:52:38***Saving seen data "./dancer.seen"
02:04:35 Quit t0rc (Ping timeout: 276 seconds)
02:06:09 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
02:06:56 Join shai__ [0] (
02:09:27 Quit _s1gma (Quit: Leaving)
02:09:52 Join JdGordon [0] (7bf38c1f@gateway/web/freenode/ip.
02:10:33 Quit shai_ (Ping timeout: 272 seconds)
02:16:22 Quit mt (Remote host closed the connection)
02:20:38 Join ntryon [0] (
02:21:13 Join ntryon_ [0] (
02:25:35 Quit ntryon (Ping timeout: 265 seconds)
02:27:18 Join Barahir [0] (
02:33:28 Join Tux2 [0] (
02:38:54 Join komputes [0] (~komputes@ubuntu/member/komputes)
02:46:28JdGordonfrom last nights buffering talk... I'm scared to think what would happen to the skin engine if it went to handle allocations instead of simple pointers
02:47:24JdGordonthe whole thing is linked trees, pointers to alloced arrrays, pointers to structs.. if even the block could move around it would be a nightmare
02:48:16JdGordonbut if the initial block could be compacted once we know how much has been used, and guarenteed to not be moved (without an exensive reload) then it might be ok
02:48:31JdGordon^ saratoga_, Torne, Unhelpful
02:49:26 Quit stripwax (Read error: Connection reset by peer)
02:50:14 Quit simonrvn (Read error: Connection reset by peer)
02:50:46Unhelpful@Torne: yes, i realize that special cases could potentially be expensive, my expectation was that they would also be very rarely *used*. the idea is that this sort of thing only happens on things like skin change, maybe a user changes one of the resizeable buffers that we can currently only size on boot, etc.
02:51:29Unhelpfulif there's a playback stall it won't surprise anybody much as it immediately follows a user interaction - the same thing already happens if you change to a theme w/ different AA size, and i have seen almost no complaint about it.
02:59:06 Quit t0rc (Quit: Give someone code, help them with one project. Teach someone to code, help them rule the world.)
02:59:19 Quit factor (Ping timeout: 272 seconds)
03:00:24 Quit Ramsey[LC] (Remote host closed the connection)
03:00:53 Join Ramsey[LC] [0] (~RamseyLC]@
03:05:15 Join clone4crw [0] (
03:06:01 Quit Judas_PhD (Quit: This is a quitting message)
03:15:36 Quit Xerion (Read error: Connection reset by peer)
03:32:25 Join Judas_PhD [0] (
03:32:29 Quit komputes (Remote host closed the connection)
03:44:35 Join Nausicaa [0] (
03:50:05 Join Strife89 [0] (
03:52:41***Saving seen data "./dancer.seen"
03:56:58 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
04:14:01 Quit engwan_ (Ping timeout: 276 seconds)
04:14:59 Quit clone4crw (Quit: leaving)
04:15:36 Join simonrvn [0] (
04:21:59 Quit amiconn (Disconnected by services)
04:21:59 Quit pixelma (Disconnected by services)
04:22:00 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:22:00 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:22:16 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:22:17 Quit Strife89 (Quit: Leaving)
04:22:20 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:25:44 Join anewuser [0] (anewuser@unaffiliated/anewuser)
04:30:34 Join clone4crw [0] (
04:36:51 Join engwan_ [0] (~engwan@
04:40:34 Quit Barahir (Read error: Operation timed out)
04:45:01 Quit fdinel (Quit: Miranda IM! Smaller, Faster, Easier.
04:56:55 Quit engwan_ (Ping timeout: 255 seconds)
04:57:34 Quit t0rc (Quit: Give someone code, help them with one project. Teach someone to code, help them rule the world.)
04:59:08 Join Keripo [0] (
04:59:44 Part Keripo
05:01:57 Join MagusG [0] (
05:14:30 Quit MethoS- (Read error: Connection reset by peer)
05:27:16 Join Barahir [0] (
05:30:43 Join froggyman [0] (~seth@unaffiliated/froggyman)
05:32:05 Quit ntryon_ (Ping timeout: 265 seconds)
05:33:52 Quit ps-auxw (Ping timeout: 245 seconds)
05:41:03 Join factor [0] (~factor@
05:45:32 Join ps-auxw [0] (
05:46:16 Quit edboyer93 ()
05:46:25 Quit clone4crw (Ping timeout: 255 seconds)
05:52:44***Saving seen data "./dancer.seen"
06:11:11 Quit Horscht (Quit: Verlassend)
06:18:53 Quit bluebrother (Ping timeout: 265 seconds)
06:20:13 Join bluebrother [0] (
06:20:13 Quit bluebrother (Changing host)
06:20:13 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
06:44:11 Quit Tux2 (Quit: I can't stay forever... but I will be back!)
06:53:06 Quit Nausicaa (Ping timeout: 252 seconds)
07:02:03 Join powell14ski_ [0] (
07:05:51 Quit krazykit (Quit: awe yeeeeeee)
07:12:21 Quit togetic (Ping timeout: 252 seconds)
07:33:40 Join Staphylo [0] (
07:39:51 Quit Judas_PhD (Quit: This is a quitting message)
07:47:06 Join Judas_PhD [0] (
07:52:47***Saving seen data "./dancer.seen"
08:01:21 Quit anewuser (Ping timeout: 272 seconds)
08:10:19 Quit Staphylo (Quit: Bye les gens =))
08:45:43 Join einhirn [0] (
08:46:01 Quit einhirn (Client Quit)
08:50:32 Join einhirn [0] (
08:55:33 Join ender` [0] (
08:56:01 Join einhirn_ [0] (
08:59:18 Join LinusN [0] (~linus@rockbox/developer/LinusN)
08:59:22 Quit einhirn (Ping timeout: 240 seconds)
09:01:16 Quit einhirn_ (Ping timeout: 240 seconds)
09:03:31 Quit GodEater (Read error: Operation timed out)
09:07:05 Join GodEater [0] (~bibble@rockbox/staff/GodEater)
09:08:08 Join esperegu [0] (~quassel@
09:11:42CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
09:11:42*pixelma wonders about the number of reports (in the forum) of problems with installation, seemingly bootloader installation, on PP Ipods
09:12:15pixelmaI believe there are 4 since yesterday
09:18:09 Quit Judas_PhD (Quit: This is a quitting message)
09:22:23 Join GibbaTheHutt [0] (
09:28:31 Join einhirn [0] (
09:28:31 Join bertrik [0] (
09:28:31 Quit bertrik (Changing host)
09:28:31 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
09:30:24 Quit einhirn (Read error: Connection reset by peer)
09:30:34 Join swilde [0] (
09:39:57 Quit JdGordon (Ping timeout: 252 seconds)
09:43:12 Join wodz [0] (
09:44:45wodzpixelma: could you try FS #11641 on Your Ondio? It *should* apply cleanly
09:46:35uberushaximusSo I mentioned this yesterday or the day before but I'm having a weird problem on the last few revisions with the clipv1
09:48:11 Join einhirn [0] (
09:49:06uberushaximuson newer operating systems with identical hardware (think windows 7 or ubuntu don't work vista/xp do) I get "*PANIC* SD Xfer read err:0x8 Disk0"
09:49:27 Quit einhirn (Client Quit)
09:49:42 Join einhirn [0] (
09:50:01uberushaximusand by "don't work" I mean "the error is reproducible"
09:50:55uberushaximusthe build I installed ~2-3 days ago didn't have any problem
09:52:48***Saving seen data "./dancer.seen"
09:53:16wodztry bisecting
09:53:49wodz2-3 days means only a couple commits
09:57:48 Join Rob2222 [0] (
09:59:27 Join utanapischti [0] (
10:01:18 Quit Rob2223 (Ping timeout: 252 seconds)
10:03:28 Quit sasquatch (Ping timeout: 245 seconds)
10:05:02gevaertsuberushaximus: that's *really* weird. That panic is in the SD code. That really shouldn't be influenced by the OS you use...
10:06:49gevaertsI also don't see anything in the svn log during the last two weeks that looks even vaguely related
10:11:07 Quit antil33t (Read error: Connection reset by peer)
10:11:16 Join antil33t [0] (
10:15:28 Join Judas_PhD [0] (
10:19:11AlexPwodz: Sorry, I was asleep last night - I can test tonight
10:19:45wodzok, thx
10:22:26 Quit Guest30770 (Quit: Leaving.)
10:23:47 Join {phoenix} [0] (
10:24:36bug2000Silly question, why I don't wish to be trolled for asking, but as I've noticed, Rockbox has a 3 month release cycle. I'd like to know if there is anything extremily wrong or lots of belated bug fixes which delay the release or there is just no apperant reason or what so ever.
10:26:37gevaertsThere's a combination of things going on I'd say
10:27:00pixelmain my experience the changes to the skin engine still aren't settled, e.g. you can experience screen cuorruption sometimes and in different ways on at least the targets with monochrome screens
10:27:37gevaerts(a) someone has to take the initiative to start the release process, and nobody has yet, and (b) several people (I'm one of them) feel that the recent skin changes aren't stable enough yet
10:27:50gevaertsPossibly (a) is caused by (b)
10:28:29bug2000I see, well, I can't help with that, I only use one spesific skin on a spesific device.
10:28:30 Quit Judas_PhD (Quit: This is a quitting message)
10:37:23 Join DerPapst [0] (
10:40:31 Join Judas_PhD [0] (
10:53:41bug2000 −− Seem to be 9/2009 is that right?
10:54:42gevaertssounds very possible
10:55:05 Quit wodz (Quit: Leaving)
10:55:28bug2000gevaerts: Except it seems the normal rockbox build got newer font files in it.
10:58:40bug2000gevaerts: Also, according to, rockfont.bdf supports Hebrew, yet with either of the files [latest / fonts build] Hebrew isn't working correctly.
10:59:10gevaertsrasher's page is auto-generated as far as I know
10:59:19*gevaerts isn't a font specialist
11:05:35TorneUnhelpful: if the special cases are very rarely used that makes them even *more* expensive in terms of binsize, code complexity and space for potential bugs to hide :)
11:06:25UnhelpfulTorne: i suppose. very cheap in terms of run time, and not hard to test as we already have lists of things we can't change without reboot right now. ;)
11:08:11Tornebut we don't test stuff particularly methodically a lot of the time. :)
11:08:44Tornei am just, mostly, waving my paw a bit and assuming that a special purpose malloc can be made to work fine :)
11:08:54Tornei have the vaguest of vague ideas about what that would be
11:08:58Tornei may well be wrong :)
11:09:25Unhelpfulcould just say "we have a real malloc, only for MMU devices" ;P
11:11:04Tornethere's not really a significnat benefit to the MMU here
11:11:13Torneunless you promote all allocations to multiples of 4kb
11:11:15Tornewhich would suck
11:11:23Torneand we'd need to start having actual pagetables :)
11:12:18UnhelpfulTorne: wouldn't a benefit w/ an MMU be that we could compact and remap existing allocations, while still using the classic malloc pointer interface instead of some sort of handle-based thing?
11:13:01Tornebut you can't.
11:13:16Torneyou can sort of half do it in a very limited way for chunks of 4kb, at the cost of loads of overhead and extra memory used
11:13:29Tornevirtual memory is not all that ;)
11:15:03*Torne thinks a useful thing to do, at this point, would be to throw some logf's in there so someone can look and see how many buffer_allocs there actually are and what sizes they are and why they exist
11:15:55Tornemaybe later ;)
11:16:37bug2000gevaerts: Where can I get the real latest 8-rockfont.fnt?
11:16:56UnhelpfulTorne: useful things? is that how development works?
11:18:43Torneoccasionally :)
11:20:05Unhelpfulit's more fun to debate how best to implement things i believe to be gross misfeatures :)
11:21:35 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
11:32:31 Quit Strife89TX (Quit: Leaving)
11:52:50***Saving seen data "./dancer.seen"
11:53:23 Join einhirn_ [0] (
11:56:48 Quit einhirn (Ping timeout: 240 seconds)
12:05:43 Join Barahir_ [0] (
12:07:42 Join wodz [0] (
12:08:48 Quit Barahir (Ping timeout: 240 seconds)
12:09:03AlexPgevaerts: Yes, I'm quite happy to do the release stuff, but have been waiting as I picked up that people weren't entirely happy with the state of some things
12:10:57Tornelike what?
12:11:06Torne(not questioning you, just curious)
12:11:23AlexPmainly skin stuff
12:11:45AlexPI only mean unhappy as in some bugs need squashing :)
12:11:57AlexPBut I may well have got the wrong end of the stick
12:15:25gevaertsTorne: until one or two weeks ago there were regular crash reports on the forums. Things have been improving since then, but there are still issues for some people
12:16:24Tornei'm concerned about the recent spate of people having issues installing on ipod
12:16:44Tornebut maybe it's just a coincidence and they all have knackered drives :)
12:17:09LloreanSeveral of them seem to be 1st gen Nanos, right?
12:17:26Tornei think so
12:17:39Tornethere was also one guy reporting the "buttons don't work after hold" thing on ipodvideo
12:17:46Tornewhich we thought we fixed ages ago
12:18:00Tornebut then a couple of people complained it still happened on the mini
12:18:10Tornehadn't heard anyone mention it on any other model until now
12:18:43gevaertsI wonder if this is just a sign of a new large group of people discovering rockbox
12:18:53gevaertsWe have been in the news in some circles recently
12:19:13Torneoh, it's a good job nwe didn't do anything to host the psgroove thing it seems
12:19:21Tornesince sony are now suing the crap out of eveyrone who is
12:19:41pixelmaah, I wanted to look up when the fix got committed exactly, because the guy with the buttons not working after hold said he would be using r3.6
12:19:49Tornepixelma: i'm pretty sure it was long before 3.6
12:19:53pixelmaas in "release"
12:19:55Tornelemme check, it's on my watchlist
12:20:32LloreanWhat skin change seems to have caused all the instability?
12:20:36pixelmathere are also these instability issues on some Fuzes
12:20:54gevaertsLlorean: basically the entire skin handling was rewritten since 3.6
12:21:06pixelmaLlorean: there were quite a few in quick succession :\
12:21:18gevaertsSome instability is not really unexpected in those conditions
12:21:41Lloreangevaerts: True. I was just curious if it was done right after 3.6, or like... within the last month or so.
12:21:54LloreanI mean it seems like 3 months ought to be enough time to get stability at least reasonably recovered.
12:22:03Tornethe fix for the wheel thing was committed in july 2009 so it should be in 3.6
12:22:14gevaertsLlorean: part of it was gsoc
12:22:20Torneunless subsequent changes to the wheel driver (nano2g being merged with it) broke it again :)
12:24:20*Llorean will try to dig up his 1G Nano tomorrow.
12:24:43Torneyeah i should actually update my ipodvideo and test it for a bit
12:25:05LloreanThough if I recall mine tended to be one of the ones that didn't show issues the last few times we had "only some players have this problem" types of issues too.
12:25:37*Llorean does wish Rockbox were releasing more consistently.
12:27:17TorneI'm still pondering if we should maybe provide *both* shutdown workarounds as alternatives
12:32:54 Quit petur (Quit: *plop*)
12:37:10LloreanTorne: What are they? How?
12:37:45TorneLlorean: currently the ipods clear the latter part of IRAM on shutdown to try and make startup not fail
12:37:56Tornethis doesn't seem to work for everyone, but it has no negative effect i'm aware of
12:38:16Tornethe older workaround of shutting down by rebooting to the OF made lots of people's players behave weirdly but fixes the problem for more people
12:39:37LloreanExcept those it makes behave weirdly?
12:40:07Torneno, some of those people are the same people :)
12:40:26LloreanWell new weird behavior doesn't really make it a "fix" does it?
12:41:06*Torne shrugs :)
12:42:22 Join timc [0] (~tim@
12:42:25 Nick timc is now known as Guest75106 (~tim@
12:46:54 Part LinusN
12:47:06 Join LinusN [0] (~linus@rockbox/developer/LinusN)
12:48:05 Part LinusN
12:49:25 Join MethoS- [0] (~clemens@
12:50:32 Join LinusN [0] (~linus@rockbox/developer/LinusN)
12:54:49 Part LinusN
12:55:03 Part Zagor
12:55:12 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
13:00:18 Quit Hadaka (Ping timeout: 240 seconds)
13:02:38Lloreangevaerts: Another possibility is that the psgroove build of Rockbox has other unexpected bugs for whatever reason, and that's why these people are having additional difficulties?
13:03:20gevaertsLlorean: I'd hope they try a standard build first
13:03:40LloreanHah. One can hope for anything, I suppose.
13:10:07 Join LinusN [0] (~linus@rockbox/developer/LinusN)
13:10:20 Join Hadaka [0] (
13:21:29 Part LinusN
13:25:55 Join teru [0] (
13:35:16 Join krazykit [0] (
13:36:53 Quit wodz (Quit: CGI:IRC (Ping timeout))
13:50:12 Join wodz [0] (
13:50:24wodzteru: ping
13:50:41teruwodz: pong
13:51:07wodzI updated patch for png plugin
13:51:33teruI saw it.
13:51:39wodzI refactored code so decoder is selfstanding and do not depend on global vars
13:51:58wodzI have fixes for mrobe scaling in my tree also
13:52:54***Saving seen data "./dancer.seen"
13:53:03wodzI fixed simple_resize_bitmap() and realy do not wan't to touch smooth_resize_bitmap()
13:56:07wodzwhat is the point of different draw_image_rect() separate for every format?
13:56:58terujpeg doesn't create native bitmap.
13:57:23wodzI see
13:59:12teruwodz: I tried v2 patch on my gigabeat but I only get "Invalid CRC". I guess *(uint32_t *)(in + N) doesn't work but not sure.
13:59:21 Join n1s [0] (
13:59:21 Quit n1s (Changing host)
13:59:21 Join n1s [0] (~n1s@rockbox/developer/n1s)
13:59:47wodzdo we look for magic values to determine decoder or we simply relay on file extension?
14:00:07terujust only file extention.
14:04:23 Join dfkt [0] (dfkt@unaffiliated/dfkt)
14:06:38wodzteru: could You prove this by changing *(uint32_t *)(in + N) to in[N]<<24 | in[N+1]<<16|in[N+2]<<8|in[N+3] and see if this passes CRC check?
14:09:01teruI'll try.
14:09:50n1swodz: most pc's are little endian so testing that in the sim on a pc should work
14:09:59wodzit works on sim
14:10:37 Quit einhirn_ (Ping timeout: 240 seconds)
14:10:41n1s(also casting a char pointer to a larger type and dereferencing it can cause data aborts on arm)
14:10:49n1sif the char pointer is not aligned
14:11:52wodzthis is probably the case :/
14:12:48n1steru: what gigabeat is that?
14:13:05terugigebeat X30
14:13:44n1sthen it would throw an exception on a misalinged access which it isn't doing if it complains about invalid crc
14:15:20gevaertsDoesn't the gigabeat F/X have some unusual behaviour with unaligned accesses?
14:15:56n1soh, i thought any of our arm's other than the beast aborted
14:16:14wodzanyway I'll change back to byte-by-byte png interpretation
14:16:18n1sif it doesn't abort it probably gives back rotated data or something
14:17:09 Quit bertrik (Read error: Connection timed out)
14:17:12n1swodz: there are functions for reading from a char array and translating into integer of chosen endianness in many places in rockbox
14:17:42 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
14:17:45gevaertsI know my F usually freezes with code that makes other ARM DAPs give data aborts
14:18:20n1shmm, i think that is configurable for at least some cores, maybe it's just turned off for some reason on the f/x?
14:19:30wodz? I am aware of betoh16(), betoh32() and little endian variants but this takes value as argument not ptr to array
14:20:44n1sfor example apps/metadata/metadata_common.c get_(long|short)_(be|le)
14:21:38wodzare they any faster than 'get-byte-shift' stuff?
14:22:02Tornewhat processor is the gigabeat f?
14:22:03n1sno, they do exactly that but since you seem to have trouble tested functions could be nice
14:22:17n1sTorne: arm9tdmi
14:22:30Tornemaybe their bootcode enables unaligned
14:23:20Tornehm, except a tdmi wouldn't have the register
14:23:49 Quit wodz (Quit: CGI:IRC)
14:27:20teruchanged betoh32(*(uint32_t) (X)) to (((uint32_t) (X)[0] << 24) | ((uint32_t) (X)[1] << 16) |((uint32_t) (X)[2] << 8)|((uint32_t) (X)[3])) and it worked.
14:27:36TorneOh right, yes :)
14:27:46TorneARMs without CP15 don't genreate alignment faults
14:27:51Tornethey just return rotated data, always
14:28:10Torneso ARM7TDMI and ARM9TDMI will just fuck up
14:28:34TorneARMs with CP15 default to generating alignment faults, but it can be turned off (until ARMv7 where it defaults to allowing them)
14:29:26Tornen1s: the beast probably aborts, ARMv6 still doesn't allow it by default even though it does the right thing
14:31:39n1sTorne: according to this strict alignment is disabled, not that i know why though
14:32:32Torneit probably shouldn't be
14:32:49Torneunless we have ARMv6 specific code which is using it specifically
14:33:19n1sis a misaligned long load faster than individual byte loads on v6?
14:33:26Tornenot if it's not cached
14:33:27Torneunaligned cache fills are slower and it only works for LDR, not LDM (or LDREX/etc)
14:33:48Tornei forget the exact cycle timing of doing the load unaligned
14:34:08Tornebut generally the code would have to be *really* heavily dependent on accessing unaligned data for it to be faster to not fix it
14:34:30n1swe do use unaligned loads explicitely on coldfire in some place(s) as it's faster there
14:34:50Tornewell, the core->cache interface only knows aligned accesses
14:35:02Torneso the unaligned accesses are emulated on core by rotates and stuff
14:36:02Torneif we do actually do it anywhere i'd be curious to see benchmarks
14:36:03n1shmm, the case i remember is bitstream reading in codecs so if it fetches subsequent memory to the cache probably doesn't hurt
14:36:16Tornen1s: well the cache->core transfer is more expensive too, is my point
14:36:21Torneit's not just potentially having to bring in more cachelines.
14:37:07*Torne shrugs
14:37:23Tornesince we run on armv5 there must be an aligned versoin of any code that does it too
14:37:27Torneso it would be easy to bench :)
14:37:32Tornewell, easysh
14:37:50n1si could bench on my beast, although test_codec runs are abit skewed i think, since when the core is running at full speed memory accesses will be relatively slower
14:38:04 Join Jaykay [0] (
14:38:08TorneIt was always traditional to leave it disabled so that you didn't accidentally write code that wouldn't work on ARMv5
14:38:15n1sunless i misunderstood all that
14:38:19Tornebut ARM themselves now allow it by default in v7
14:38:25Torneso, who knows :)
14:38:32*n1s tests
14:39:49Tornealso i guess the armv5 and armv6 versions of a given bit of code are not necessarily comparable anyway
14:42:35 Nick drizztbsd__ is now known as drizztbsd (
14:54:51n1shmm, usb inside rockbox is super screwy on my beast
14:55:12n1skeeps disconnecting and reconnecting and then died with a black screen and shut off
14:55:29n1sbootloader usb is still stable as a rock
14:58:24 Join anewuser [0] (anewuser@unaffiliated/anewuser)
14:58:42 Nick fxb__ is now known as fxb (
14:59:35n1sTorne: doing a straing long load + byteswap instead of 4 individual byteloads in the ffmpeg bitstream reading code speeds up flac decoding by 3.5% on the beast
15:00:07Tornewell, ok then
15:00:33n1sso now the flac 8 test file decodes in 3.56s instead of 3.69 :)
15:00:34Tornei assume it's already hidden behind a macro
15:00:57Torneis it already ifdef'ed for v6, then, or are you adding that?
15:01:58 Part Zagor
15:02:17n1si'd have to add that, it's just #ifdef for coldfire ATM
15:02:57Tornewell, sounds worthwhile :)
15:04:23n1sit could even be generalized into htobe32(*(const uint32_t*)(x)) as htobe will not do anything on coldfire
15:06:19n1sthere's also ROCKBOX_STRICT_ALIGN which is defined for all ARM cpus (as well as sh and mips) but this is checked in only one place
15:07:15n1sin, dircache.c
15:15:55 Join {-phoenix-} [0] (
15:19:50 Join evilnick_B [0] (0c140464@rockbox/staff/evilnick)
15:20:05CIA-81New commit by nls (r28183): ARMv6 supports unaligned memory accesses and they are enabled on our only ARMv6 target so we might as well use them. Speeds up decoding of a flac8 ...
15:20:17 Quit {phoenix} (Ping timeout: 272 seconds)
15:21:00n1shmm, my checkin seems to have hung
15:21:58CIA-81r28183 build result: All green
15:23:26CIA-81New commit by teru (r28184): explicitly set img->using_preloaded_icons in parse_image_load() and don't rely on parse_image_display().
15:23:28n1sthe command still hasn't returned...
15:24:03 Join einhirn [0] (
15:25:04CIA-81r28184 build result: All green
15:47:08amiconngevaerts: At least on SH the C string "Game" needs 8 bytes
15:50:10n1samiconn: because of alignment or something else?
15:50:24n1sbtw, no sh targets have these codecs anyway
15:52:58***Saving seen data "./dancer.seen"
15:55:54amiconnYes. For some reason, sh-elf-gcc aligns all strings at 4 byte boundaries
15:56:36amiconnWe would save quite a bit of binsize if we could change that. Unfortunately I never found out whether that's possible
15:57:30n1sthat seems stupid if the alignment isn't required by something
15:59:32amiconnIt *may* be required by gcc's inline string/mem copying stuff
16:15:51 Join Nausicaa [0] (
16:16:21 Quit antil33t ()
16:19:25 Join sudoman [0] (d8ecfce7@gateway/web/freenode/ip.
16:36:43 Quit {-phoenix-} (Remote host closed the connection)
16:37:19amiconn[14:27:45] <Torne> ARMs without CP15 don't genreate alignment faults <== At least that's not true for PP
16:38:35 Join engwan_ [0] (~engwan@
16:38:54 Join kevku [0] (
16:39:01amiconn[14:55:11] <n1s> keeps disconnecting and reconnecting and then died with a black screen and shut off
16:39:19amiconn^ that's probably due to USB charging. I do have the same problem when using my hub
16:39:40amiconnMy hub does proper overcurrent shutdown, and the charging code draws more than what's allowed
16:39:52n1samiconn: ah, i'll test disabling that
16:40:36Torneamiconn: the cache cell probably does it as an external abort
16:41:17n1samiconn: that was it, thanks
16:41:17Torneamiconn: there's no such thing as an alignment fault without CP15, anyway; all aborts are external. they might still be caused by wrong alignment, but it's up to the external device
16:44:04 Join HeWhoMustNotBeNa [0] (~pbbbhhttt@
16:44:54HeWhoMustNotBeNaawfully quiet in here isnt it?
16:46:38Tornenobody talked for almost sixty whole seconds
16:47:19Lloreann1s: So, bringing the sample rate topic in here.
16:47:25HeWhoMustNotBeNaits a conspiracy against me, a conspiracy of silence
16:47:30LloreanA setting for "preferred sample rate" that resamples everything to it?
16:47:39amiconnTorne: Okay, regardless of how it's generated, PP fires data abort on unaligned accesses
16:48:04*Llorean would prefer, then, if there was also an option of "native" that disabled features incompatible with native sample rates, but played things back at their native ones, too, possibly
16:48:08Torneright, but htat's unrelated to whether any other ARM does
16:48:14Tornebecause it's external.
16:48:16n1sLlorean: i'm no expert but that's the approach i think seemed the most feasible from previous discussions on the subject
16:48:22amiconnn1s: It's kinda funny that my laptop doesn't do overcurrent shutdown, but my external (powered) hub does
16:48:27Tornea normal ARM7TDMI wouldn't.
16:48:34HeWhoMustNotBeNahey guys, what are the chances of getting flamed (or worse) for asking a truly noob question in here
16:48:56amiconnI usually connect the charger in order to avoid the effect. It also doesn't happen if the beast is fully charged
16:49:19n1samiconn: i used the frontside ports on my computer... i wonder why our usb charging does this, shouldn't it negotiate the current?
16:49:29 Join togetic [0] (~togetic@unaffiliated/ibuffy)
16:49:35amiconnIt does, but somehow the calculation seems to be off
16:49:42GodEaterHeWhoMustNotBeNa: virtually none, provided you're polite about it :)
16:50:06amiconnI.e. it calculates how much the charger is allowed to draw for a certain amount of total current, but that calculation is a bit off
16:50:18HeWhoMustNotBeNai should be able to accomplish that I think, being polite that is
16:50:19LloreanHeWhoMustNotBeNa: Though, if it's *truly* noob-ey, you might check the manual first to avoid being referred to it.
16:51:03HeWhoMustNotBeNaLlorean, i am trying to read as much as possible.....but my 3 year old daughter wants me to vacate the pc so time is of the essence
16:51:28n1sLlorean: it's not only a problem of dsp things but switching the actual DAC to a new samplerate needs, to be done and at an exact time and this switching introduces glitches AFAIU
16:52:12HeWhoMustNotBeNais it safe / advisable to use the rockbox utility to install rockbox for a noob like me?
16:52:24Lloreann1s: If you don't care about gapless, couldn't you let the PCM buffer empty of 44.1 content, switch, then start decoding 48 content?
16:52:42LloreanHeWhoMustNotBeNa: Rockbox Utility is intended to be very safe.
16:53:09n1sLlorean: probably
16:53:13HeWhoMustNotBeNaexcellent. thanks Llorean
16:53:38TorneLlorean: yeah, that'd be the easiest way
16:53:40Lloreann1s: I'd say that sort of thing is reasonable enough for people who want multiple sample rates mixed.
16:54:05LloreanI'd *suspect* that at least within a single album they'll always be the same sample rate, and that's (to my experience at least) where gapless is generally important.
16:54:46n1sLlorean: i'm cynical so i expect people will complain whatever we do :)
16:55:00LloreanYou could also just not bother fixing any of the DSP stuff to work with 48khz, and just require people to use resampling if they want DSP, since that affects audio quality anyway. :-P
16:55:11HeWhoMustNotBeNaLlorean, i was reading different instructions elsewhere and there are 2 files for the installation there 1. Rockbox ipod nano 1g 2. ipod patcher 1g nano
16:55:28TorneHeWhoMustNotBeNa: that's to do it manually; just use the utility
16:55:33LloreanHeWhoMustNotBeNa: Just use our manual. Other instructions aren't supported.
16:55:34HeWhoMustNotBeNaif I use the utility would, it have the ptcher builtin?
16:56:19HeWhoMustNotBeNaLlorean, many thanks.....i had to read those instructions as I want to port psgroove on the nano later
16:56:23Tornethe utility downloads everything it needs automatically
16:57:01HeWhoMustNotBeNathen I would like to thank all the people who have made this utility
16:57:16HeWhoMustNotBeNai shall go the utility way now
16:57:19LloreanHeWhoMustNotBeNa: The psgroove patch isn't really supported. You can get help installing normal Rockbox in here (which is dead simple on the Nano and pretty much automated) but for help with the psgroove stuff you should contact its author.
16:57:51HeWhoMustNotBeNai know Llorean, i wasnt going to ask about psgroove in here in the first place
16:58:14HeWhoMustNotBeNai'm very thankful for all of your help.....i shall bother you guys again if i run into any problems
16:58:22 Quit einhirn (Ping timeout: 265 seconds)
17:01:46 Join Xerion [0] (
17:02:29 Quit teru (Quit: Quit)
17:14:42soapWould not an interim solution be to keep the all-44 playback and introduce an "audiophile grade" resampler and tell the audiophools to stuff themselves if they think 48->44 is bad because "it is not integral"?
17:15:53evilnick_BIsn't it that a better resampler is too expensive in terms of cpu cycles?
17:16:02LloreanWould adding a higher quality resampler be easier than enabling native playback in a "no DSP, no crossfade, no gapless" way?
17:16:19soapthe expense is only born by those who choose to use it.
17:16:49soapLlorean, I would think a higher quality resampler would be /preferable/ to "no DSP, etc"
17:16:57Lloreanevilnick_B: Yeah, it's like saying "aren't the constant spinups of FLAC too expensive on disk targets?", people can deal with it if that's what they really want.
17:17:20Lloreansoap: Which DSP are we really talking about here? How many of them are things "audiophiles" are actually going to turn on?
17:17:53LloreanI'm not talking about disabling them entirely, just disabling them when someone switches from "resample everything to 44.1" to "native playback"
17:19:19 Quit AlexP (Remote host closed the connection)
17:19:36soapI don't care which DSP is affected. If we consider DVD audio rips there is a source of legit 48K material which should not be burdened by either forceable PC-side resampling or lack-of-DSP. A good resampler should have identical perceptual quality to native 48k, so I don't see the advantage of native 48k playback if it would be burdensome on the rest of the Rockbox audio chain.
17:20:23soap(and when I say "I don't care" I mean "I don't think it matters")
17:20:44LloreanPeople aren't going to stop asking for native support just because there's a good resampler.
17:21:12LloreanWhat's better? Saying "no, you can't have it" or "you can have it, but when you're using it some other features may not be available"?
17:21:28soappeople aren't going to stop asking for all sorts of indefensible things.
17:21:40 Join AlexP [0] (~alex@rockbox/staff/AlexP)
17:21:47LloreanBut this is one that could not only be provided, but falls into the category of "better making use of the hardware's features"
17:22:11soapIf you could tell them "Rockbox's resampler is just as good as native 48k" (which you can't now)...
17:22:53soapbut a better resampler serves more people than native 48k on the hardware which can do 48k.
17:23:14saratoga_i think native 48k is probably not that hard to do
17:23:18LloreanHow bad is our resampler?
17:23:22saratoga_really bad
17:23:27soapit allows all the crazy sampling rates, and supports all the (powerful enough) software decoding targets.
17:23:33LloreanFor 48->44.1 is it noticeable?
17:23:35saratoga_most of the stuff will work with 48k more or less, since its so close to 44.1k
17:23:48bertrikLlorean, saratoga_ why do you think it's really bad? Has anyone ever measured it?
17:23:52saratoga_could probably use the EQ as is and just shift the center frequency values a couple percent
17:24:00saratoga_bertrik: its linear interpolation
17:24:04Lloreanbertrik: I don't know how bad it is, that's why I'm asking. I've just been told it's bad.
17:24:30bertriksaratoga_, so, that doesn't have to be so bad
17:24:45LloreanIs linear interpolation really that bad a way to resample audio? I don't know much about it, but it seems it'd be semi-reasonable.
17:24:56saratoga_theres only one way to do linear interpolation, so it pretty much has to be bad
17:25:52saratoga_give me a sec, i'll make a plot
17:25:56saratoga_i've got matlab open
17:28:12bertrikIt would be nice to have some hard data on the various methods (and their computational impact)
17:29:07HeWhoMustNotBeNaLlorean, it worked...i have rockbox on my nano now.........many thanks
17:29:30HeWhoMustNotBeNabut how do i switch between rockbox and the default os?
17:29:41gevaertsThat's described in the manual
17:30:03HeWhoMustNotBeNathere now
17:30:58saratoga_bertrik: at FS/3, resampling via linear interpolation down by 20% results in an aliased copy of the signal thats attenuated by 7dB relative to the peak
17:31:15saratoga_source peak i mean
17:32:29HeWhoMustNotBeNai havent used this thing for ages and i forgot what the select button was
17:32:34 Quit tchan (Quit: WeeChat 0.3.3-dev)
17:32:36sudomani'm looking at rockbox/firmware/target/arm/ipod/button-clickwheel.c line 357: "static bool hold_button = false;". is that correct? and does it have anything to do with this: ?
17:33:18gevaertssudoman: how can a variable declaration on its own be incorrect?
17:34:07Lloreansaratoga_: We're primarily talking about 48->44.1 though
17:34:13HeWhoMustNotBeNagevaerts, it says shutting down and then it restarts :(
17:34:20HeWhoMustNotBeNaor Llorean for that matter
17:34:38saratoga_i'll try 9%
17:35:14pixelmaLlorean: I could make out an additional hiss when playing back a really bad quality 11.025kHz file in swcodec Rockbox compared to it being played back in foobar2000 but that's an extreme example (and since it was bad quality to start with...)
17:35:16bertriksaratoga_, the signal is a 16 kHz sine I assume?
17:35:31sudomangevaerts: well i'm not very experienced in this, but in context, it looks like the the test: "if (hold_button)" is only being performed if the "hold_button" in the previous test is true, but it should be peformed if there's any difference to "hold_button_old"
17:36:12sudomanwhy is "hold_button_old" automatically false? what if it was on before?
17:36:25sudoman* what if it was on hold?
17:37:17 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
17:37:59Lloreanpixelma: I'm kinda surprised that upsampling from a fraction of 44.1 causes noticeable sound problems.
17:38:08*Llorean doesn't know why he would expect it not to though, just does.
17:39:51 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
17:40:08pixelmaoh, just looked it up... it's even an 8kHz mono file
17:40:21sudomangevaerts: what do you think?
17:41:48saratoga_bass frequency obviously work a lot better, higher ones a lot worse
17:43:06pixelmaI believe preglow also had a distinct opinion about our resampler ;)
17:43:39pixelmajust thinking he could add some facts to the discussion
17:43:40saratoga_didn't he write it?
17:44:47bertriksaratoga_, I'm not sure how to interpret that. Isn't the blue peak at around 1050 the same as the red peak at around 1250 the same, except resampled?
17:45:12saratoga_bertrik: yes, when you resample the size of each frequency bin changes, so the FFT peaks won't line up
17:45:20 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
17:45:50bertrikThe next highest blue peak is about 25 dB down
17:46:29saratoga_thats not very good
17:47:02bertrikI agree, but far from the 7 dB you mentioned earlier :)
17:47:27HeWhoMustNotBeNathanks for all your help guys!!!
17:47:33HeWhoMustNotBeNabye from Pakistan
17:47:36 Quit HeWhoMustNotBeNa ()
17:48:55 Join Staphylo [0] (
17:50:29saratoga_bertrik: like I said it depends on frequency
17:50:41saratoga_so as you decrease the test tone frequency you will decrease the aliasing
17:51:33saratoga_and unfortunately as you increase the frequency it aliases into lower frequencies
17:52:42 Join komputes [0] (~komputes@ubuntu/member/komputes)
17:53:00***Saving seen data "./dancer.seen"
17:53:16bertrikIs this 9% frequency change, or 20% ?
17:54:01saratoga_do you have matlab? i can give you the code
17:54:03saratoga_its just 2 lines long
17:54:13bertrikno, I don't
17:54:27saratoga_ah ok
17:55:00saratoga_well heres what i did:
17:55:05saratoga_should give you an idea what you're looking
17:56:16saratoga_the basic problem here is that linear interpolation takes quiet high frequency noises and turns them into (perceptually louder) mid frequency noises
17:57:59n1show much more complex are better resampling algos?
17:58:34n1slinear needs, like a an add and a shift per sample, right?
17:59:18gevaertssudoman: yes, isn't that what happens?
18:00:40bertrikn1s, it uses a multiply for each input sample plus some adding
18:01:18n1sbertrik, oh
18:01:49saratoga_its only a shift if you're resamplilng by a factor of 2 :)
18:03:08 Join hebz0rl [0] (
18:03:26sudomangevaerts: it looks like "if (hold_button)" on line 368 won't be tested if hold_button is false because it is nested by the previous condition wich assumes that hold_button_old is false. so if the hold button is turned off then the input won't be unlocked.
18:04:26n1sright, i wasn't thinking :)
18:04:31sudomangevaerts: maybe line 368 shouldn't be nested by line 364
18:04:44saratoga_cubic interpolation would be the next simplest thing, i think it needs something like a dozen multiplies per output sample
18:04:45bertrikHas anyone currently online ever done an ams sansa recovery?
18:05:10gevaertssudoman: "which assumes that hold_button_old is false"? Where does it do that?
18:05:11saratoga_though maybe non-polynomial resampling would be more efficient
18:05:32gevaertssudoman: except for the first time after boot
18:05:37saratoga_i really don't know much about the various algorithms other then linear < cubic spline < sinc
18:05:46bertrikProbably somebody has already designed 48->44.1 khz algorithms
18:05:51sudoman"hold_button_old = hold_button; hold_button = button_hold();" right above
18:06:09gevaertssudoman: That sets hold_button_old to whatever hold_button is, not false
18:06:26sudomanoops, i meant: " static bool hold_button = false; bool hold_button_old; /* normal buttons */ hold_button_old = hold_button;"
18:06:29saratoga_yeah theres hundreds on PC, but most of them use 50-100MHz on an Intel machine since they insist on having >120dB accuacry or something stupid
18:07:17gevaertssudoman: did you miss the "static" there?
18:07:24n1ssaratoga_: "like a dozen multiplies per output sample" assuming 44.1kHz output in stereo would mean about 1 million multiplies per second, which *should* be doable
18:07:27bertrikWhat accuracy would be considered reasonable for rockbox, something like 60 dB?
18:07:56saratoga_i'm not sure
18:08:01sudomannoob question: what does static do?
18:08:12n1sit makes the variable static
18:08:27saratoga_i guess we should probably run the rockbox resampler (on the Sim) through RMAA and see what that says
18:08:31n1s(it survives after the function exits)
18:08:56 Quit n1s (Quit: Lämnar)
18:09:00sudomanoh, so it only runs once?
18:09:11gevaerts's only initialised once, yes
18:09:24saratoga_i wonder if its just easier to support 48kHz though
18:09:29sudomanok, ic sorry for the confusion
18:09:53saratoga_i think the EQ wouldn't care at all since its only a 9% difference, so unless you want a 20kHz EQ band its going to be the same coefficients anyway
18:10:30saratoga_probably just a matter of fishing out all those nasty built in assumptions about how long a sample lasts in the PCM system
18:11:22bertrikAFAIU, the sample frequency is important basically only in apps/dsp.c where we now have NATIVE_FREQUENCY, so we could make NATIVE_FREQUENCY a variable instead of a constant
18:12:29saratoga_apply_crossfeed would need slightly different delay constants, but i doubt anyones going to notice
18:12:44bertrikah, yes indeed
18:13:46saratoga_we should probably ask Blue_Dude
18:13:53saratoga_he was the last person to really work on the PCM code
18:14:04saratoga_but i don't see anything in there thats all that tied to the exact value of sample rate
18:14:46saratoga_EQ will fail badly if you try to use it on say 32kHz audio, but its probably reasonable to just disable it for really low sample rates or else force resampling in that case
18:16:18saratoga_cross fade could be tricker to deal with, since i'm not sure how easy it is to see what the sample rate of a track is before you start playing it
18:22:50 Join [sko] [0] (~sko]
18:30:07 Quit kevku (Remote host closed the connection)
18:37:32 Join kevku [0] (~kevku@2001:7d0:0:f000::135d)
18:38:48 Join _s1gma [0] (~d.d.derp@
18:38:55 Quit togetic (Ping timeout: 245 seconds)
18:39:38saratoga_"fast" FIR based resampler on ARM
18:39:53saratoga_although at 13 taps I'm not sure if its fast in the sense we think of :)
18:39:54 Join Jerom [0] (~jerome@
18:41:39saratoga_looking online, a lot of arm stuff seems to use hermite splines for interpolation, which is basically just fitting a 3 order polynomial to the input data and then reading out the value at the output sample positions
18:43:23 Quit swilde (Quit: ERC Version 5.3 (IRC client for Emacs))
18:44:29 Join togetic [0] (~togetic@unaffiliated/ibuffy)
18:51:42 Join {phoenix} [0] (
18:52:44 Quit sudoman ()
18:58:27 Quit komputes (Quit: I haven't slept for ten days, because that would be too long.)
18:58:29 Join crow [0] (
19:11:05 Quit yorick (Read error: Operation timed out)
19:13:08 Join robin0800 [0] (
19:13:18 Quit DerPapst (Quit: Leaving.)
19:13:30 Join yorick [0] (yorick@unaffiliated/yorick)
19:24:15 Quit kevku (Quit: KVIrc 4.0.2 Insomnia
19:24:23 Join leachim6 [0] (
19:24:35leachim6does USB mode work at all in rockbox on the Fuze2?
19:24:50leachim6like, wile using rockbox, can I plug in the cable to have it charge in the car?
19:24:53leachim6no data, just charging
19:25:13 Join Maggux [0] (
19:27:40leachim6nevermind, I got my answer.
19:28:35 Quit bmbl (Quit: Bye!)
19:34:53 Join Horscht [0] (
19:34:53 Quit Horscht (Changing host)
19:34:53 Join Horscht [0] (~Horscht@xbmc/user/horscht)
19:38:28 Join Buschel [0] (
19:40:09Buschelliar: you there?
19:45:29 Quit crow (Quit: ( :: NoNameScript 4.22 :: ))
19:47:36 Nick fxb is now known as fxb__ (
19:48:04 Join crow [0] (
19:53:04***Saving seen data "./dancer.seen"
19:57:30Buschelliar: (for the logs) from your nano2g patch I understood that you own an LCD type (2) ILI9320. it would be great if you could test FS #11646. this patch was already verified on LCD type (7) and needs test on the LCD type that you own.
19:59:01crowseems that patch I was missing for my nano 2g :D
19:59:12 Quit Jerom (Ping timeout: 252 seconds)
20:01:16Buschelcrow: do you own an nano2 2g with ILI9320 ?
20:02:27 Join hadoukenx [0] (83d2648b@gateway/web/freenode/ip.
20:08:34hadoukenxI've searched the forum and know that most accessories are unable to interact with rockbox, but is rockbox capable of interpreting input signals over its serial port currently?
20:09:08bertrikhadoukenx, that depends on the target
20:09:16bertrikthe specific player I mean
20:09:17hadoukenxsorry, sansa specifically
20:09:27hadoukenxc200 or e200 series
20:09:56hadoukenxv2 for both
20:10:17bertrikI think we have don't even have a serial driver for the sansas
20:10:27crowBuschel hmm how to know that?
20:11:17bertrikI don't know if there is currently anything known about the protocol used for sansa accessories
20:11:29 Join kevku [0] (
20:12:49 Quit kevku (Client Quit)
20:12:50Buschelcrow: just go to the debug menu, enter the battery debug screen's second screen (right-scrolling the nano's wheel) and there it is
20:12:52amiconnsaratoga_ (and others): There's another problem wrt native 48kHz (and other sample rates) support: Not all targets can do that
20:13:42amiconnI.e. most targets can do several sample rates, but no target can do all 9 (or 12, if you allow up to 96kHz) standard rates
20:13:59 Join Jerom [0] (
20:14:25amiconnMost notably, coldfire can only do 11.025, 22.05, 44.1, 88.2
20:14:39amiconnSo we still need a good resampler
20:15:35 Join kevku [0] (
20:15:48amiconnpreglow wanted to look into this, as he asked for a recording of the iriver H1x0 OF resampling result for 48 -> 44.1 about 4 weeks ago
20:16:00amiconnUnfortunately he seems to have disappeared again :\
20:16:58 Join wodz [0] (
20:18:58 Join DerPapst [0] (
20:25:19 Quit leachim6 (Ping timeout: 255 seconds)
20:31:50pixelmahe posted in the forum today or yesterday ;)
20:32:04pixelmabut about recording for D2
20:33:50crowBuschel when i go to Debug Menu i see battery 4.104V ; 4.104-4.118V (5mV)
20:34:05 Join komputes [0] (~komputes@ubuntu/member/komputes)
20:34:05Buschelcrow: now scroll to the right
20:34:18Buschelcrow: you will see another screen with lots of text then
20:36:58crowBuschel yea i see that screen, which option do i am looking for
20:37:25Buschelcrow: LCD type
20:38:56crowBuschel dont see that option.
20:39:02Buschelcrow: must be either ILI9320 or LDS176
20:39:50crowBuschel LDS176 :)
20:40:12crowBuschel it was under Debug Menu> View HW info :)
20:40:30Buschelcrow: good news for you, bad news for me... good = my patch will work for you, bad = still need a tester with the other type
20:40:59crowBuschel type: 1, (7) LDS176
20:41:14Buschelcrow: really? damn, my memory (brain, not RAM ;-) is tricking me
20:41:34crowBuschel well i am with rockbox just fev daysso i will not see much on battery drain..
20:42:00 Quit hadoukenx (Quit: Page closed)
20:43:16Buschelcrow: with the patch your LCD should make this wheeeee sound anymore
20:43:34Buschelcrow: of course "should _not_ make..."
20:48:05 Quit CGL (Remote host closed the connection)
20:52:41crowBuschel didnt sow that jet :)
20:57:22Buschelcrow: then you're a lucky one :)
20:58:28crowBuschel well as i wrote, i have rockbox just fev days..
20:58:31 Part Guest75106 ("Leaving.")
20:58:51crowinstalled with rockbox installed which picked up for my nano 2g latest rockbox version
21:07:13 Nick fxb__ is now known as fxb (
21:13:54 Join webguest76 [0] (
21:15:10 Quit webguest76 (Client Quit)
21:24:32 Join kinky [0] (
21:24:48 Nick fxb is now known as fxb__ (
21:26:46kinkyI've installed the rockbox daily build (r28184-100929) about an hour ago, and now I realise that the sound output is "wobbering". Every 1-2 seconds, the sound is going down and high again, it's really annoying. On the OF, the problem is nonexistant.
21:27:30saratoga_what device do you have
21:27:53kinkysaratoga_: a sandisk clip+
21:28:07 Join stripwax [0] (
21:28:14 Quit stripwax (Client Quit)
21:29:07 Join stripwax [0] (
21:29:26 Quit stripwax (Client Quit)
21:33:48 Join stripwax [0] (
21:37:43kinkyso, should I replace my .rockbox folder with an older version? maybe this one? :
21:39:16 Quit togetic (Quit: WeeChat 0.3.0)
21:39:31 Quit anewuser ()
21:40:02evilnick_Bkinky: It'd be much more use to check recent changes to svn that affect your target/playback and revert to a build before any likely change happened
21:41:07evilnick_BThe Daily builds are just whichever svn revision happens to be checked in each day at a certain time. i.e. they're not special.
21:42:18*gevaerts suggests waiting for a developer who's familiar with the clip+
21:42:24 Join panni_ [0] (
21:42:33kinkyevilnick_B: I will look around a lil bit in the repo, regarding major changes
21:42:44kinkygot my clip+ just yesterday :o
21:46:54 Quit [sko] (Quit: Leaving.)
21:53:05 Quit stripwax (Ping timeout: 245 seconds)
21:53:08***Saving seen data "./dancer.seen"
21:55:34 Join jgarvey [0] (
22:00:59MagguxAre there anywhere more levels for Bubbles? or just the 7
22:04:06soapamiconn, can I assume the iRiver H3x0 OF resampling would be equivelent to the H1x0 OF resampling?
22:04:21amiconnI don't know
22:04:46amiconnThe reason why preglow asked for the H1x0 output is that it can be recorded digitally
22:04:46 Join simonrvn_ [0] (simon@
22:04:53soapI could record a loop of the RMAA test sweeps as resampled by the H3x0 if that would be of any service.
22:05:03*amiconn did this using his archos recorder and the wavrecord plugin
22:05:25amiconnAs well as an optical->coax s/pdif converter
22:06:49 Join al4711 [0] (
22:07:21 Quit al4711 (Client Quit)
22:07:50 Quit simonrvn (Ping timeout: 240 seconds)
22:07:50 Nick simonrvn_ is now known as simonrvn (simon@
22:07:52 Join al4711 [0] (
22:09:28kinkyevilnick_B: took a look at the svn, nothing worth mentioning regarding my case there
22:09:54evilnick_Bkinky: What does "wobbering" mean?
22:10:22kinkyevilnick_B: the sound-level falls and rises every 1-2 seconds
22:10:50kinkybut not on every song, just some of them
22:11:45kinkywell, I think I'll stay with this version until weekend and then try the newest one
22:11:53evilnick_BIs it the same songs each time? Or does one song sometimes do it and then sometimes not?
22:12:08evilnick_BBasically, can you rule out there being an error in those music files themselves?
22:12:10kinkyevilnick_B: happens on the same songs everytime
22:12:17 Quit Nausicaa (Read error: Connection reset by peer)
22:12:27kinkyevilnick_B: on the OF, the problem is nonexistant
22:12:39kinkyso it should be no file related issue
22:13:05gevaertsWhich codec is this?
22:13:06kinkychecked the same song on both RB and OF
22:13:10evilnick_BThat doesn't matter. If you can find out what the affected files have in common then it'll help a developer to reproduce the issue and hopefully fix it
22:13:37gevaertsAnd does it happen with different codecs?
22:13:41al4711Today I wanted to install rockbox on my IPod Classic 6st Generation but the RockboxUtility told me that this device is not supported. I'am willing to help to run rockbox on this device. I have tried to find some hint in the faq where I can start, I hope you can give me a hint. Thanks
22:14:16saratoga_al4711: check out the NewPort wiki page
22:16:38gevaertsal4711: possibly for the classic can help, or the people in #freemyipod
22:16:55evilnick_Bal4711: A word of warning, unless you have experience in coding/cracking encryption or a vast amount of spare time then it'll be an uphill struggle.
22:17:06evilnick_BActually, it will be anyway even if you do have those skills
22:17:52kinkyevilnick_B: seems like they're all mp3's with variable bitrate, 128kbps (v1)
22:18:23al4711Sorry, but I'am not sure if I understand you right evilnick_B. Do you mean the 6st Generation is in someway encrypted?!
22:18:55saratoga_it was, but someone cracked the encryption a year or two ago
22:19:17al4711Do you think it is possible to replace the current IPod SW with rockbox on this device?
22:19:29saratoga_yes of course its possible
22:19:30al4711Ah ok I will google for this issue
22:19:38evilnick_BAnything is possible, but it's a very difficult and time-consuming job
22:19:42saratoga_did you read the wikipage?
22:19:55al4711I'am started
22:20:08saratoga_well stop wasting time
22:20:20 Join edboyer93 [0] (
22:20:30evilnick_Bkinky: Your best bet would be to post on the forum/bug-tracker and attach a file that doesn't work if at all possible
22:20:56saratoga_or try one of the files in the sim
22:21:14kinkyevilnick_B: will do that tomorrow, need sleep now :o
22:21:15al4711What do you mean?!
22:21:30saratoga_what does who mean?
22:21:50al4711I refer to 'saratoga_: well stop wasting time'
22:23:11saratoga_stop wasting our time because you haven't bothered to read up on whats involved
22:24:41AlexPsomewhat harsh
22:24:53al4711Ok I have seen that it is a lot more then just write some code ;-( . You need also to dig into the HW. Thanks for the hint to the NewPort page I think I will follow your advise.
22:25:17 Quit esperegu (Read error: Connection reset by peer)
22:25:48 Part kinky
22:27:17saratoga_somewhat more direct
22:27:25saratoga_the less harsh version was less understandable
22:27:35al4711I'am sure there are some guys from time to time to want to port rockbox and after reading what does this means the knowledge or the time consuming the most guys quickly rethink about this idea
22:27:44al4711now I'am one of this guys ;-)
22:27:52saratoga_yeah someone does this every day or two
22:28:15al4711I will try to find a device on which rockbox works stable.
22:29:25al4711maybe there should be a new faq entry 'how can I port rockbox to my device' or something similar
22:29:38AlexPThere is, it is the NewPort wiki page
22:30:05al4711I have overread this point in the FAQ sorry
22:38:10CIA-81New commit by wodz (r28185): fix bitmap scallers smooth_resize_bitmap() and simple_resize_bitmap() to properly handle LCD_STRIDEFORMAT == VERTICAL_STRIDE case
22:38:18CIA-81New commit by wodz (r28186): imageviewer bmp - drop special case for grey bitmap scaller
22:39:51CIA-81r28185 build result: All green
22:40:17al4711thanks for the talk and rockbox. good luck. by aleks
22:41:37CIA-81r28186 build result: 51 errors, 0 warnings (wodz committed)
22:44:37pixelmaall monochrome targets
22:44:52 Quit al4711 (Quit: Bye all ;-))
22:45:01pixelmaMaggux: I believe there are 99 or 100 levels in bubbles
22:45:49wodzhmm but grep -rn _simple_resize_bitmap * gives me no hits
22:47:04 Join insp_ [0] (
22:47:15pixelma;type=mrobe100sim#prob1 says "apps/plugins/imageviewer/bmp/bmp.c:285: undefined reference to `simple_resize_bitmap'" without the _
22:47:28 Quit Jaykay (Quit: ChatZilla 0.9.86 [Firefox 4.0b6/20100914073604])
22:48:32wodz;type=archosfmrecorder has _
22:51:37wodzhmm I don't understand this error
22:52:16Torneisn't the leading _ just part of the C ABI on some platforms?
22:53:04Torneor in fact in the spec :)
22:54:52gevaertssimple_resize_bitmap() seems to be within #if LCD_DEPTH > 1
22:55:40 Quit crow (Remote host closed the connection)
22:56:09gevaertswodz: moving that #endif up a bit makes things link
22:58:29gevaertsIt probably should still be tested then of course
22:59:30 Join stripwax [0] (
22:59:33 Quit evilnick_B (Quit: Page closed)
23:00:04wodz but look at
23:00:21wodzI don't change #if #endif order
23:03:42wodznow I understand
23:03:57gevaertsYou removed grey_resize_bitmap, so things now use simple_resize_bitmap
23:04:42wodzyes but I changed simple_resize_bitmap to work in that case. I missed that it won't be compiled at all on monochrome
23:05:36 Quit stripwax (Ping timeout: 245 seconds)
23:06:27 Join stripwax [0] (
23:06:57Jeromare the strncat and strncpy functions not implemented because you guys don't care about buffer overflow or you are doing it some other way ?
23:07:01 Quit GibbaTheHutt (Ping timeout: 276 seconds)
23:07:47gevaertsIt was felt that strlcat and strlcpy offer better semantics
23:08:17gevaertsThe difference may be subtle, but the l variants are often what people meant anyway
23:08:51gevaertsIf you want the n semantics, use memcpy
23:09:05CIA-81New commit by wodz (r28187): fix red ...
23:10:15Jeromdidn't know these functions exist, thanks. I feel like an idiot now :p
23:10:38gevaertsThey're not universally available
23:10:42CIA-81r28187 build result: All green
23:11:26gevaertsSo people don't always know about them
23:12:06 Join GibbaTheHutt [0] (
23:13:57 Quit stripwax (Ping timeout: 265 seconds)
23:17:29 Join stripwax [0] (
23:25:38 Join clone4crw [0] (
23:32:28 Quit Maggux (Remote host closed the connection)
23:32:52 Quit Buschel (Quit: ChatZilla 0.9.86 [Firefox 3.6.10/20100914125854])
23:33:43 Quit panni_ (Read error: Connection reset by peer)
23:33:49saratoga_did sandisk ever contact anyone about a Fuze+?
23:34:04 Join panni_ [0] (
23:43:57 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
23:48:00 Quit jgarvey (Quit: Leaving)
23:48:24 Quit domonoky (Read error: Connection reset by peer)
23:53:10***Saving seen data "./dancer.seen"
23:56:45 Quit bertrik (Quit: sleeo)
23:57:57 Nick fxb__ is now known as fxb (

Previous day | Next day