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 2010-10-26

00:04:09 Join evilnick_ [0] (
00:07:13 Join Topy44 [0] (
00:10:44kugelJdGordon: shouldn't the sim still use sim_open()?
00:11:15 Quit hebz0rl (Quit: Leaving)
00:11:45kugelyou need to fix the path in lc_open_from_mem() for android
00:13:58kugelI think the sim doesn't need the wrapping since it's setting up a virtual root directly from the current dir anyway
00:15:30kugelI think there could be a application_opendir in the same way as application_open; you added the same code to both android's and sdl's opendir
00:16:16kugelNEED_WRITE is also needed for O_WRONLY and O_TRUNC
00:18:20kugelthe changes to file.h look a bit fishy to me
00:18:26rasherkugel: does kinetic scrolling add language strings? It would be helpful if no new strings are added until after the release
00:18:26kugelotherwise it looks fine I'd say
00:18:42kugelrasher: no
00:19:00kugelwe could make it configurable but I haven't done it yet
00:19:12kugelsince it works well for me :)
00:21:13n1sBuschel: i'll look at it tomorrow afternoon
00:21:26 Quit n1s (Quit: Lämnar)
00:28:10 Quit GeekShadow (Read error: Connection reset by peer)
00:29:21 Join xxcv [0] (
00:29:31 Quit ender` (Quit: Sex is dirty - if you do it right. -- Isaac Asimov)
00:29:33 Join JdGord [0] (
00:32:31 Quit saratoga (Quit: Page closed)
00:33:52 Join T44 [0] (
00:37:34 Quit Topy44 (Ping timeout: 252 seconds)
00:37:44JdGordKugel: the similar does still use sim_open
00:38:03kugelbut not directly?
00:38:31JdGordThat would be a misttake then...
00:38:43kugeli just looked at the diff, I may be wrong
00:39:08kugelI see #define open(x, ...) application_open(x, __VA_ARGS__) in the PLATFORM_SDL section
00:40:03kugelbtw, what happens about creat() etc calls?
00:40:54 Join antil33t [0] (
00:41:26JdGordCreat is a wrapper for open isn't it?
00:42:41 Join jdgord| [0] (
00:42:42 Quit JdGord (Read error: Connection reset by peer)
00:44:52kugelJdGordon: no. there are also other calls, remove(), rename()
00:45:10kugelI think you're going to need to wrap all of them
00:47:36jdgord|Ok ill look into it.
00:47:42jdgord|Damn :p
00:49:09 Quit linuxguy3 (Remote host closed the connection)
00:49:25 Join linuxguy3 [0] (
00:53:34 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
00:53:38 Quit jdgord| (Quit: Bye)
00:56:19 Quit linuxguy3 (Ping timeout: 240 seconds)
00:57:20 Join linuxguy3 [0] (
00:59:16JdGordon|kugel: as for the release being sucky, that is bassically because noone has put their hand up and said they'll be "release manager"
01:00:00kugelwe never needed a dedicated release manager, but maybe we do now
01:00:44JdGordon|not officialy but we have always had someone who has done the builds and pushed everything along
01:03:28 Quit linuxguy3 (Ping timeout: 250 seconds)
01:04:14 Join linuxguy3 [0] (
01:08:05 Quit antil33t (Ping timeout: 255 seconds)
01:12:30 Join anewuser [0] (anewuser@unaffiliated/anewuser)
01:14:51 Join antil33t [0] (
01:15:55 Quit JesusFreak316 (Remote host closed the connection)
01:19:07 Quit factor (Read error: Connection reset by peer)
01:20:18 Join factor [0] (
01:22:22 Part toffe82
01:25:01 Nick aexin is now known as aevin (eivindsy@unaffiliated/aevin)
01:28:59[Saint]JdGordon: When CustomWPS says "touchscreen areas - WPS only" that actually still true?
01:29:28*[Saint] edits
01:30:40[Saint]can you check and see if the touchscreen tags/keywords are up to date on that page for me?
01:30:52JdGordon|not now
01:30:55[Saint]I would have no idea if anything was missing.
01:31:01[Saint]Fair enough.
01:32:07JdGordon| search for "touchactions" shows the accepted names
01:32:31***Saving seen data "./dancer.seen"
01:38:04[Saint]I just remembered that you added a few recently, and wasn't sure if they were added to the wiki or not.
01:39:35 Quit xxcv (Ping timeout: 240 seconds)
01:40:20[Saint]can %TI be used conditionally? As in %?TI<%T(0,0,0,0,foo)|%T(1,1,1,1,bar)>?
01:40:24 Quit krazykit (Read error: Operation timed out)
01:41:29 Quit antil33t (Ping timeout: 264 seconds)
01:43:46[Saint]"if screen is X for <timeout>"
01:44:55JdGordon|I dont tihnk that will work
01:45:07JdGordon|more thoguh I dont tihnk %T can be in a conditional
01:45:45JdGordon|have a look at the 320x480 wps though, it has a touch regions which turn on and off
01:48:12[Saint]are the touch regions VP relative?
01:48:33[Saint]I'm trying to "touchify" iLike Video...and it's a PITfuckingA
01:49:12 Join antil33t [0] (
01:49:13JdGordon|wait, you tihnk someone who uses rockbox on a d2 would like the ipod look? :D
01:49:25[Saint]why not ;)
01:49:31JdGordon|fair enough
01:50:14[Saint]If there was an %(is it touchscreen) tag I could even change the title from "iPod" to "Cowon" ;)
01:51:28JdGordon|you want a target name tag, not a touchscreen one
01:51:36[Saint]I figure some nutter out there will at least try it out on the D2, and as a themer it's my duty to make sure people don't have to use that terrible bloody pox of a thing that is grid mode
01:52:20[Saint]Hmmmm....that would be interesting..go on ;)
01:52:36[Saint]then I could specify between Nano 1 and 2 Gen also.
01:53:17JdGordon|That isnt likely to happen :) I don't tihnk the target name is in the binary at all
01:53:44[Saint]could the %(is it touch enabled) be added?
01:53:45JdGordon|you could of course do it with the theme setting tag though
01:53:54[Saint]I could actually use that.
01:54:08JdGordon|you sure there isnt one?
01:54:21[Saint]'s not listed anywhere.
01:55:00[Saint]I'd expect it to be in with the touch tags, and it isn', "not 100% sure, but about 98% certain"
01:56:05JdGordon|yeah, there isnt one
01:57:38[Saint]I think it could be quite usefull, only really for Video/D2 themes, but still.
01:57:43[Saint]s/ I think it could be quite usefull/ I'd use it myself if there was one/
01:58:03JdGordon|%Sl(device name, freetext) .... %?Sl(device name)<%Sl(device name)>
01:59:23 Quit chattr (Read error: Connection reset by peer)
01:59:54JdGordon|grr.. those last 2 should be %Sv
01:59:59 Join fdinel [0] (
02:00:26JdGordon|1) setup the variable the user can add, freetext means anything. %Sv is "did the user set one? yes? ok so disoplay it"
02:01:20[Saint]Aha...right. You suggesting that as a suggestion for my "is touch available" query? Or just putting it out there?
02:01:35 Quit anewuser (Ping timeout: 240 seconds)
02:01:46JdGordon|for the getting the device name string, instead of adding a new tag for it
02:02:56 Quit parafin (Read error: Operation timed out)
02:03:23[Saint]Ah, right...well for now I could just do it with a "%(is it touch capable?)" tag, as with that resolution it's only going to be one of two targets, and it's a true/false (or yes/no whatever way you want to put it) condition so it works perfectly.
02:07:17 Join parafin [0] (
02:07:22 Join soap [0] (~soap@rockbox/staff/soap)
02:08:41JdGordon|it should be an easy addition, but because %T() on non touch targets doesnt affect the display (usually) it isnt really needed (unlike the "is rtc capable?" tag_
02:08:51JdGordon|although there is defintly a case to add it
02:09:09*JdGordon| can walk you through the addition if oyu're game?
02:09:11 Join anewuser [0] (anewuser@
02:09:13 Quit anewuser (Changing host)
02:09:13 Join anewuser [0] (anewuser@unaffiliated/anewuser)
02:09:28[Saint] "walk through the addition"?
02:09:39[Saint]that was weird
02:10:13JdGordon|its one line in tag_table.c and half a dozen in skin_parser.c
02:10:36 Join krazykit [0] (
02:10:40[Saint]I could have a look at how it's done I guess.
02:10:47[Saint]Couldn't hurt me to try.
02:12:21JdGordon|look how SKIN_TOKEN_HAVE_RECORDING is done in skin_parser.c
02:12:35JdGordon|in check_feature_tag
02:12:36[Saint]Ok, thanks.
02:14:12 Join chattr [0] (
02:15:09 Quit Kiwi_Cam (Ping timeout: 255 seconds)
02:16:40 Quit chattr (Remote host closed the connection)
02:17:16 Quit antil33t (Ping timeout: 272 seconds)
02:19:43 Quit soap (Read error: Connection reset by peer)
02:20:08 Join soap [0] (~soap@rockbox/staff/soap)
02:20:52 Join antil33t [0] (
02:22:24 Join chattr [0] (
02:27:42 Quit chattr (Remote host closed the connection)
02:28:01 Quit DerPapst (Quit: Leaving.)
02:30:46kugel^ cowond2 build with the kinetic scrolling patch
02:31:00JdGordon|ok cool, I'll try it tonight
02:38:39*JdGordon| even downloads it now so he cant blame his shitty home internet connection!
02:39:21kugelselecting with the track ball is a fishy thing
02:39:29JdGordon|how do you mean?
02:40:00kugelwell, at least on my legend, it's easy to accidentally select 2 or 3 times in a row
02:40:08kugel(it has this optical track ball thingy)
02:40:30JdGordon|ah ok
02:40:43JdGordon|is there a way to know if there is a real ball or optical?
02:40:47kugelperhaps we should only do up and down
02:41:14JdGordon|no thanks
02:41:21JdGordon|ball select works great on the n1
02:41:52JdGordon|maybe play with the TIME_BEFORE line in the button-android.c?
02:43:25kugelso, with the pitchscreen I suppose we can safely enable absolute point mode by default
02:43:35kugelwithout losing core functionality
02:45:01 Join Topy [0] (
02:45:25JdGordon|I still think the sbs is a nasty blocker also
02:45:29JdGordon|also the lack of plugins
02:45:33JdGordon|(not game ones)
02:45:49kugellack of plugins has nothing to do with the absolute point mode
02:46:03*JdGordon| somehow read releasing into that
02:47:41kugelI've some plugins enabled locally, but it's just a shitload of work
02:48:05kugelalso the get_user_file_path() thingy is untested/unimplemented for plugins
02:48:19kugelwell, *all* plugins only work with the grid mode
02:48:29 Quit T44 (Ping timeout: 252 seconds)
02:48:38JdGordon|arg, yeah
02:48:57JdGordon|I was tihnking the basic ones (especially the ones with lists as the ui) shouldnt
02:49:24 Join chattr [0] (
02:49:44 Quit chattr (Read error: Connection reset by peer)
02:50:11 Join T44 [0] (
02:50:17 Join chattr [0] (
02:50:29 Quit soap (Read error: Connection reset by peer)
02:50:43kugelJdGordon|: was there a reason you made OK on the left and CANCEL on the right in the yesno screen?
02:51:01kugelI was under the impression that OK is commonly the right-most button
02:51:09 Join soap [0] (~soap@rockbox/staff/soap)
02:51:17JdGordon|thats the way the grid mode had it already
02:51:20JdGordon|isnt it?
02:51:42kugelno you changed the grid mode accordingly (if I remember your ci message right)
02:51:47 Quit Topy (Ping timeout: 252 seconds)
02:53:15 Quit antil33t (Ping timeout: 240 seconds)
02:54:08 Join BHSPitMonkey [0] (
02:54:09JdGordon|ah hmm...
02:54:15 Quit BHSPitMonkey (Changing host)
02:54:16 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
02:55:31JdGordon|That yes/no screen should be replaced with a OS yes/no if we can figure out how to do it
02:56:55 Join Topy [0] (
02:58:23 Quit T44 (Ping timeout: 252 seconds)
02:59:23 Join antil33t [0] (
03:02:11kugeluseless context menu. the menu button should go to the main menu instead
03:10:56JdGordon|need to get long ball or long menu working then
03:10:57 Quit BHSPitMonkey (Ping timeout: 255 seconds)
03:12:14JdGordon|or, menu pops up a menu with a few buttons
03:12:22JdGordon|a android menu that is
03:13:02kugelthe context menu via long-press on touchscreen doesn't work
03:13:06kugelgood enough?
03:15:17JdGordon|isnt it??
03:15:26JdGordon|-isnt it?
03:20:41 Join T44 [0] (
03:21:42 Quit T44 (Client Quit)
03:22:35 Quit advcomp2019_ (Ping timeout: 252 seconds)
03:24:14 Quit Topy (Ping timeout: 252 seconds)
03:25:24 Quit MagusG (Read error: Connection reset by peer)
03:25:57 Join MagusG [0] (
03:26:16 Quit kugel (Remote host closed the connection)
03:27:42 Quit antil33t (Ping timeout: 276 seconds)
03:32:32***Saving seen data "./dancer.seen"
03:32:47 Join antil33t [0] (
03:40:26 Join advcomp2019_ [0] (~advcomp20@unaffiliated/advcomp2019)
03:44:08 Quit antil33t (Ping timeout: 240 seconds)
03:45:51 Join soap__ [0] (
03:46:55 Quit soap (Ping timeout: 240 seconds)
03:49:11 Join antil33t [0] (
03:58:01JdGordon|[Saint]: how'd you go?
04:12:18 Quit _s1gma (Quit: Leaving)
04:35:19 Join edboyer93 [0] (
04:38:27 Join Barahir_ [0] (
04:41:43 Quit Barahir (Ping timeout: 252 seconds)
04:42:24 Quit krazykit (Ping timeout: 252 seconds)
04:42:27 Quit edboyer93 ()
04:44:14 Join krazykit [0] (
04:48:59 Quit krazykit (Ping timeout: 255 seconds)
04:50:40 Quit amiconn (Disconnected by services)
04:50:40 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:50:55 Quit bluebrother (Disconnected by services)
04:50:56 Join bluebroth3r [0] (~dom@rockbox/developer/bluebrother)
04:51:00 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:51:18 Quit pixelma (Disconnected by services)
04:51:21 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:51:23 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:55:23 Quit TheSeven (Ping timeout: 276 seconds)
04:58:27 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
05:05:02 Join krazykit [0] (
05:20:46 Quit soap__ (Ping timeout: 252 seconds)
05:23:50 Quit merbanan (Ping timeout: 276 seconds)
05:24:00 Join soap [0] (~soap@rockbox/staff/soap)
05:31:37 Join xxcv [0] (
05:32:35***Saving seen data "./dancer.seen"
05:33:02 Quit Horscht (Quit: Verlassend)
05:34:14 Quit ps-auxw (Ping timeout: 276 seconds)
05:36:12 Join merbanan [0] (
05:41:34 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
05:45:14 Join ps-auxw [0] (
05:47:55 Quit soap (Ping timeout: 240 seconds)
05:52:01 Join soap [0] (~soap@rockbox/staff/soap)
05:53:46 Quit xxcv (Ping timeout: 252 seconds)
05:57:51 Join xxcv [0] (
05:59:02 Quit sasquatch (Quit: WeeChat 0.3.2)
05:59:27 Join sasquatch [0] (
06:08:31 Quit soap (Ping timeout: 240 seconds)
06:08:44 Quit fdinel (Quit: Miranda IM! Smaller, Faster, Easier.
06:09:27[Saint]JdGordon: with?
06:09:36[Saint]re: "how did I go?"
06:09:44JdGordon|time touchscreen available tag
06:09:56JdGordon|s/time/the/ wtf
06:10:38[Saint]Oh, I haven't even had a chance to look yet. Had a Dr's appointment shortly after I talked to you, and just got home now.
06:10:38 Quit merbanan (Read error: Connection reset by peer)
06:11:57 Quit S00row (Read error: Connection reset by peer)
06:14:11 Quit xxcv ()
06:14:22 Join S00row [0] (
06:17:29 Quit [Saint] (Ping timeout: 255 seconds)
06:18:48 Join S_a_i_n_t [0] (S_a_i_n_t@
06:28:02 Join merbanan [0] (
06:33:24 Quit panni_ (Quit: ( :: NoNameScript 3.81 :: ))
06:43:05 Quit anewuser ()
07:15:44 Quit mc2739 (Ping timeout: 240 seconds)
07:16:49 Quit S00row (Read error: Connection reset by peer)
07:17:59 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
07:18:38 Join S00row [0] (
07:24:05 Quit krazykit (Ping timeout: 276 seconds)
07:29:31 Join krazykit [0] (
07:32:26 Join Buschel [0] (
07:32:38***Saving seen data "./dancer.seen"
07:46:11 Quit krazykit (Ping timeout: 276 seconds)
07:47:46 Join MethoS-- [0] (~clemens@
07:51:47 Join krazykit [0] (
07:54:13 Quit linuxguy3 (*.net *.split)
07:54:13 Quit MethoS- (*.net *.split)
07:54:13 Quit KiwiCam (*.net *.split)
07:54:13 Quit simonrvn (*.net *.split)
07:54:13 Quit literal (*.net *.split)
07:55:17 Join KiwiCam [0] (
07:55:18 Quit S00row (Read error: Connection reset by peer)
07:56:31 Join S00row [0] (
08:10:38 Join kevku [0] (~kevku@2001:7d0:0:f000::135d)
08:14:10 Join linuxguy3 [0] (
08:14:14 Quit Buschel (Quit: ChatZilla 0.9.86 [Firefox 3.6.11/20101012113537])
08:15:32 Quit BlakeJohnson86 (Ping timeout: 240 seconds)
08:16:10 Join simonrvn [0] (
08:16:31 Quit S00row (Read error: Connection reset by peer)
08:16:49 Nick simonrvn is now known as Guest75780 (
08:17:15 Join literal [0] (
08:17:57 Join S00row [0] (
08:21:35 Quit BHSPitMonkey (Remote host closed the connection)
08:23:52 Nick Guest75780 is now known as simonrvn (
08:23:53 Quit S00row (Read error: Connection reset by peer)
08:26:32 Join S00row [0] (
08:44:07 Join Tchno [0] (
08:44:27 Join drizztbsd_ [0] (~quassel@unaffiliated/drizztbsd)
08:45:33 Join ender` [0] (
08:46:05 Quit drizztbsd (Ping timeout: 272 seconds)
08:47:20TchnoHello, I have been having problems getting past the title screen on my Sansa V2 device, it just stays on there. I have already tried troubleshooting before, refer to: and
08:48:59 Join hebz0rl [0] (
08:49:45 Quit hebz0rl (Max SendQ exceeded)
08:50:33 Join hebz0rl [0] (
08:53:51 Part JdGordon|
08:58:27TchnoI tried posting on the forums as well, although no advice that I haven't tried before was presented.
09:04:35 Join JdGord [0] (~jd@
09:07:04 Join Simon14 [0] (~chatzilla@unaffiliated/simon14)
09:09:12Simon14hey, i'm running rockbox on my clip+ with a SD card. when I plug it into my computer (loads up Sansa fw) it says there are only 2gb/7.36GB remaining but when i look through the folders there are no files. whats going on?
09:11:07 Join soap [0] (~soap@rockbox/staff/soap)
09:12:07JdGordFilesyatem might be stuffed
09:12:19Simon14i can still play music fine
09:12:28ender`chkdsk /f /x x:
09:12:35Simon14also.. when i plug it in, i don't see the .rockbox folder anywhere
09:13:57Simon14ender`: x: being the internal memory drive?
09:15:51Simon14under windows 7, it doesn't really have a drive letter, the paths are "Computer\Sansa Clip+ 8GB\Internal Memory"
09:16:20ender`then put that path instead of x:
09:16:28ender`hm, no
09:16:31linuxstbDoesn't that mean it's in MTP mode?.
09:16:37ender`put the player to MSC mode
09:17:05Simon14sorry.. how do i do that?
09:18:09Simon14err wait i think i found it
09:18:40Simon14i think that did it
09:18:41Simon14thanks :)
09:25:59 Quit S00row (Read error: Connection reset by peer)
09:27:56 Join S00row [0] (
09:31:01 Join jdgord| [0] (~jd@
09:31:16 Quit JdGord (Read error: Connection reset by peer)
09:32:16 Join petur [0] (d408b802@rockbox/developer/petur)
09:32:42***Saving seen data "./dancer.seen"
09:35:10 Quit soap (Ping timeout: 265 seconds)
09:35:12 Quit factor (Ping timeout: 250 seconds)
09:39:59 Quit Tchno (Quit: CGI:IRC)
09:43:14 Join soap [0] (~soap@rockbox/staff/soap)
09:44:58 Quit hebz0rl (Quit: Leaving)
09:47:41 Quit soap (Ping timeout: 252 seconds)
09:49:02 Join soap [0] (~soap@rockbox/staff/soap)
09:54:24 Quit soap (Ping timeout: 276 seconds)
09:54:24 Quit S00row (Read error: Connection reset by peer)
09:55:28 Join Rob2222 [0] (
09:56:49 Join S00row [0] (
09:59:27 Quit Rob2223 (Ping timeout: 264 seconds)
10:10:44 Quit drizztbsd_ (Ping timeout: 250 seconds)
10:22:23 Join drizztbsd [0] (~quassel@unaffiliated/drizztbsd)
10:25:27 Quit powell14ski_ (Ping timeout: 276 seconds)
10:28:15 Quit jdgord| (Ping timeout: 264 seconds)
10:35:22 Join JdGord [0] (~jd@
10:35:57 Quit TheSeven (Ping timeout: 255 seconds)
10:37:52 Join DerPapst [0] (
10:39:29 Join casainho [0] (
10:45:26 Quit Adublaptop ()
10:46:55pixelmaI wonder why some themes on the theme site say "Works with current build" in the overview table but have "Doesn't work with current build" on the details page (e.g. the first in X5 themes - Rockbox Modern)
10:47:50 Join xxcv [0] (
10:49:20pixelmanoticed while trying to figure out why the difference in number of greyscale and colour160x128 themes is so big, it wasn't that much (maybe 4-5 themes) last I looked
10:50:40 Join wodz [0] (
10:50:45wodzpixelma: ping
10:55:41wodzI suppose you didn't try imageviewer_png_v6.diff - it is broke in the same way MAS patch was.
10:56:36wodzand did latest MAS patch fix sim build for you?
10:56:41wodzpixelma: ^
10:57:08wodzleave me the answer in log, have to go
10:57:12 Quit wodz (Quit: Leaving)
10:59:42 Join slater [0] (
11:00:10slaterhey i do have a short question. how can i communicate with the mp3 player via usb with busybox?
11:00:17slateror is that not possible at all?
11:15:06 Join evilnick__ [0] (
11:16:53 Join powell14ski_ [0] (
11:18:23 Quit MagusG (Ping timeout: 255 seconds)
11:18:23 Quit evilnick_ (Ping timeout: 255 seconds)
11:19:15 Join MagusG [0] (
11:23:13 Quit dionoea (Read error: Operation timed out)
11:23:18 Join dionoea [0] (
11:25:10 Quit krazykit (Remote host closed the connection)
11:26:06 Join krazykit [0] (
11:26:32 Join AlexP_ [0] (~alex@rockbox/staff/AlexP)
11:27:05 Quit linuxstb (Ping timeout: 252 seconds)
11:27:05 Quit AlexP (Ping timeout: 252 seconds)
11:27:22 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
11:31:58 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
11:32:43***Saving seen data "./dancer.seen"
11:33:39 Quit casainho (Remote host closed the connection)
11:35:28 Quit JdGord (Quit: Bye)
11:42:36slaterhey i do have a short question. how can i communicate with the mp3 player via usb with busybox?
11:42:38slateror is that not possible at all?
11:46:50JdGordonto do what?
11:49:30slaterto check the root syste, of the original firmware
11:49:42slatercould somebody add me to the wiki user groups
11:49:54slateri am currently working on porting YP-R0
11:50:02slaterand I made some progress so far
11:50:18JdGordonyou're AntonHerzfeld ?
11:51:18slaterthank you
11:52:01slaterso just out of curiosty is it possible to reach the player with busybox via usb?
11:52:33JdGordonbusybox is just a cut down version of core utils... so if it has mount you're set
11:53:11slaterplayer only mounted flash partion not root partion
11:53:53slaterbut i recall working on my phone there was a way to check if busybox is running on the device via usb wire
11:54:09slaterand if is then be able to use it
11:54:58Tornebusybox doesn't "run"
11:55:05JdGordonwhat makes you tihnk it is running linux?
11:55:08Torneit's just a bunch of utilities
11:55:10TorneJdGordon: it is
11:56:09Torneslater: there is no standard way to get access to a linux device, at all
11:56:09slaterwith which i would be able to use cat.
11:56:27Torneif it only exposes part of its filesystem over usb then that's tough
11:56:35Torneyour phone might have usb networking and an ssh server
11:56:40Tornebut there's no reason to assume that any othe rdevice does
11:57:06slaterthats what i was not hoping ;)
11:58:07S_a_i_n_tteru: The patch you gave me to test that fixes icons in plugins works 100%
11:59:10S_a_i_n_tI will let you know if I find any more plugins that require icons to be anabled, but for now I think it is just disktidy.rock and shopper.rock
11:59:26S_a_i_n_tthankyou again for doing that for me, are you going to commit it?
11:59:31S_a_i_n_tteru: ^
12:01:21JdGordonhas anyone tried putting %X or %xl inside the "has rtc?" or "has recording" skin tags?
12:02:07S_a_i_n_tnot personally...only %Vd(blah)
12:02:17S_a_i_n_twhy? is it borked?
12:02:28JdGordonI dont know, I'm wondering if it works as expected
12:02:36JdGordonbut i've just got the call so im heading out again :)
12:02:50JdGordonIf you end up hgaving a few min and are bored could you try it please?
12:03:58JdGordonthe way it used to work (pre new parser) was the the code in the true branch would never be read on the targets without that feature so you should have been able to conditionally load images
12:03:59 Join einhirn [0] (
12:04:04JdGordonwanna know if that still works
12:06:28 Quit xxcv (Ping timeout: 240 seconds)
12:09:01JdGordonactually that may not be true, but it should be... (i.e i'll work to fix that if it isnt)
12:11:36S_a_i_n_ta .icons file should be named <main iconset (the 0-31 one) name>.icons .....yes?
12:14:06S_a_i_n_tso if the main iconset was "iconset.bmp", the .icons file would be "iconset.icons"?
12:14:37S_a_i_n_tOr is it <viewers iconset name>.icons?
12:14:57 Join swilde [0] (
12:15:07S_a_i_n_tI'm confused...what I thought, and what the wiki says are completely different, but I know enough to not trust the wiki 100%
12:20:20 Quit kevku (Quit: KVIrc 4.0.2 Insomnia
12:32:51 Join JdGord [0] (~jd@
12:34:51 Join n1s [0] (~n1s@rockbox/developer/n1s)
12:42:23 Nick evilnick__ is now known as evilnick (
12:42:35 Quit evilnick (Changing host)
12:42:36 Join evilnick [0] (~evilnick@rockbox/staff/evilnick)
12:45:08 Join BlakeJohnson86 [0] (
12:51:18 Join kugel [0] (~kugel@rockbox/developer/kugel)
12:51:40kugelJdGordon: tested scrolling on d2 yet?
12:55:43 Part slater
12:58:23 Quit kugel (Ping timeout: 252 seconds)
13:01:26 Join kugel [0] (~kugel@rockbox/developer/kugel)
13:03:17kugelJdGord: ^
13:04:01JdGordYeah the build didn't work
13:04:39JdGordDidn't load. Ill play tonight (current out)
13:04:41kugeldoes svn even work?
13:06:24 Join xxcv [0] (
13:06:40 Join jdgord| [0] (~jd@
13:06:45 Quit JdGord (Read error: Connection reset by peer)
13:13:20 Join liar [0] (
13:14:38 Quit krazykit (Ping timeout: 240 seconds)
13:15:30 Join krazykit [0] (
13:16:08 Join kugel_ [0] (
13:16:36 Quit kugel (Disconnected by services)
13:18:32 Quit kugel_ (Remote host closed the connection)
13:22:17 Quit swilde (Quit: ERC Version 5.3 (IRC client for Emacs))
13:22:49 Quit n1s (Read error: Connection timed out)
13:24:32 Join kevku [0] (~kevku@2001:7d0:0:f000::135d)
13:25:15 Quit jdgord| (Ping timeout: 264 seconds)
13:26:56S_a_i_n_tJdGordon: jsyk...I've looked at tag_table and skin_parser...and it's all voodo to me.
13:27:00S_a_i_n_tsorry :-/
13:27:09S_a_i_n_tI wish it wasn't, but it is :-/
13:28:15pixelmaAFAIK there are still glitches with the D2 after the eabi switch, sometimes working, sometimes sound not working etc.
13:28:29 Join kugel [0] (~kugel@rockbox/developer/kugel)
13:32:44***Saving seen data "./dancer.seen"
13:34:05 Join factor [0] (
13:34:11 Quit krazykit (Ping timeout: 240 seconds)
13:35:40 Join krazykit [0] (
13:38:06S_a_i_n_tTheSeven: liar: I added "#ifnder BOOTLOADER -> #endif" around HAVE_LCD_SLEEP and HAVE_LCD_SLEEP_SETTING and the SVN bootloader plays nice with me now.
13:38:31S_a_i_n_tIt seems the bootloader down't like these defines when LCD sleep is in the main build.
13:39:11S_a_i_n_tWith that small addition, and TheSeven's bootloader commit from the other day, all is well with the SVN Nano2G bootloader
13:40:03S_a_i_n_tI would ay the LCD sleep is ready to be committed, and that a new Nano2G bootloader (with those changes I mention above) should be released for 3.7
13:40:24S_a_i_n_ts/should be/it would be cool if it was/
13:41:24AlexP_bootloaders aren't generally released as part of releases
13:41:35AlexP_Of course there is no reason not to, but it doesn't really matter either way
13:42:13S_a_i_n_tIn this case though, there's functionality in the core that is useless without the it kinda makes sense to get it out there.
13:42:44AlexP_Already committed to the core?
13:43:09AlexP_It should be released ASAP then, not dependent on 3.7
13:43:54S_a_i_n_tYeah, I've mentioned it a few times...I thought it ties in well with 3.7 though. Obviously I can't fo it myself.
13:45:01TheSevenS_a_i_n_t: what kind of core functionality are you referring to that needs support from the bootloader?
13:45:40S_a_i_n_tthe "boot OF with MENU" thing, didn;t you mention that required changes to the core? or did I imagine that?
13:46:35AlexP_That sound slike bootloader only to me
13:48:51kugelthe nano2g is not a 3.7 device is it?
13:49:18S_a_i_n_tI thought it was going in to stable this release?
13:49:48kugelI don't know
13:49:52AlexP_kugel: I think TheSeven rates it as stable, so it is going to be promoted
13:49:58kugelare some of the unstable sansas going to be released?
13:50:14AlexP_If the maintainers consider them to meet the criteria :)
13:50:15kugelI think they're generally ready
13:50:34kugelusb is the only thing missing but usb isn't a blocker
13:51:02S_a_i_n_tAlexP_: There was a change in core, phew...glad I didn't go insane ;)
13:51:14S_a_i_n_tit needed a change to the clickwheel driver to make it work.
13:51:21AlexP_S_a_i_n_t: If you and the other involved people think they are ready, then make it so
13:51:27AlexP_er, kugel for the last one
13:51:38TheSevenS_a_i_n_t: that doesn't mean that there is a dependency between bootloader and main binary
13:51:39AlexP_S_a_i_n_t: OK, but that isn't what you said
13:51:41S_a_i_n_twas the commit: ^
13:51:59TheSeventhey just share the clickwheel driver, which needs to support that feature for the bootloader to be able to use it
13:52:01AlexP_S_a_i_n_t: You said there was functionality in the core that needs a new bootloader, which isn't true
13:52:26S_a_i_n_tyeah, it's the reverse...sorry.
13:52:33AlexP_no it isn't
13:52:33 Quit Amec (Read error: Connection reset by peer)
13:52:46 Join teru [0] (
13:52:52TheSevenconcerning nano2g stability: there are still some small quirks i'd like to see removed, but those aren't really user-visible
13:53:00AlexP_The bootloader and the core share the driver, so making changes necessary for the bootloader incidently changed the one used in the core
13:53:25 Join Amec [0] (
13:53:33AlexP_TheSeven: Well, your really hte only one that can decide, so IMO it is up to you :)
13:53:37S_a_i_n_tteru: I left a note for you in the logs
13:53:47kugelTheSeven: well, we have the 3 tier system to get a relatively high bar for the stable tier. what quirks do you mean?
13:54:24teruS_a_i_n_t: oh, thank you.
13:55:35TheSevenkugel: the insane way we need to do audio DMA currently, which indirectly decreases battery runtime and might lead to clicks in mpegplayer
13:55:46TheSevensome stacks that shouldn't need to be as large as they are...
13:56:24TheSeveni don't know of any stkovs left, but only because we increased stack sizes, not because we reduced stack usage
13:56:47TheSevenbut this kind of things is hardly a blocker from a user's perspective
13:56:55S_a_i_n_tiirc you squashed the last by "fixing" the screendump issue
13:57:27kugelTheSeven: wasn't the dma thing a hw limitation?
13:57:43S_a_i_n_tthat's the only reproducable stkov I knew of left on Nano2G
13:57:49kugeliirc only a tiny amount of samples can be transfered at a time?
13:57:57TheSevenkugel: i had a look at apple's irq handler, and this one just can't be meeting the insane latency requirements that we need to meet
13:58:14TheSevenso they must have a different way to do it, i just haven't figured out how exactly yet
13:58:40TheSevenand we still have a problem with increased battery usage after playing and stopping audio for some reason
13:58:56TheSevenbut that's only some mA, nothing compared to the USB one I fixed recently
13:59:13kugelisn't the battery life better than with the apple os anyway?
13:59:33S_a_i_n_t'tis for me.
14:00:03TheSevenit isn't in that situation
14:00:23TheSeven~21-23mA with a 400mAh battery will hardly reach 24 hours
14:05:27TheSevenalso, the runtime estimates aren't calibrated really good
14:05:57TheSevenwe should probably commit that adaptive algorithm for flash-based players one day
14:07:25TheSevenhm. battery current is higher when the player is idle than when it's playing audio
14:07:34TheSeven21 vs. 17 mA
14:08:07TheSevenstays at 17 this time when stopping playback
14:13:08S_a_i_n_t<TheSeven> we should probably commit that adaptive algorithm for flash-based players one day == FS #10890 (Dynamic Runtime Estimation)?
14:15:07TheSevenIIRC yes
14:15:51 Quit xxcv (Ping timeout: 240 seconds)
14:18:30TheSevenoh, and there's another bug remaining: USB sometimes locks up for some reason. This only appears to happen from time to time, I can usually transfer several gigabytes of data before it happens, but if the HID is enabled, it happens way too often
14:19:48 Quit BlakeJohnson86 (Ping timeout: 252 seconds)
14:20:08kugelI'd consider this as a blocker for stable actually
14:20:24AlexP_or disable HID for the release
14:20:30S_a_i_n_tfwiw, I have never seen that.
14:20:30AlexP_Does it happen without?
14:20:34 Nick AlexP_ is now known as AlexP (~alex@rockbox/staff/AlexP)
14:20:36TheSevenkugel: well, currently HID is disabled even on trunk for exactly that reason
14:20:37S_a_i_n_tHID is disabled
14:20:43S_a_i_n_thas been for some time
14:20:44TheSevenwithout HID it only happens once in a blue moon
14:20:46 Join BlakeJohnson86 [0] (
14:20:48 Join hebz0rl [0] (
14:21:05*TheSeven has seen it 3 or 4 times total with HID disabled
14:21:37*S_a_i_n_t never, but guesses he may be lucky, or puts his devices through less torture
14:21:57*TheSeven wonders if CIA is broken
14:22:21TheSevenor is there any reason why it wouldn#t announce r28362?
14:22:43AlexPTheSeven: Is it related to amount transferred? i.e. do you only see it occasionaly because you rarely transfer lots of data?
14:22:59TheSevenwhen it happened, it usually happened shortly after connecting
14:23:25TheSevenwith HID enabled, the trigger seems to be HID and MSC attempting to transfer things at the same time
14:24:58TheSeveni wrote a raw nand dumping tool yesterday and ran into the same issue there, after amounts from 200MB to 1.5GB, with no concurrent transfers, so it might also be just a flaky cable or something
14:25:25AlexPwell if it is very rare and only you, it is probaby OK
14:25:36 Quit krazykit (Remote host closed the connection)
14:25:37AlexPIf it were more widespread I think we would have heard something
14:25:51S_a_i_n_tperhaps, as I regularly transfer ~2-3-4GB, and have not seen this.
14:25:56 Join krazykit [0] (
14:26:28S_a_i_n_tUSB was terrible when HID was still enabled, but has been fine since then
14:26:41S_a_i_n_tHID suddenly broke inexplicably one day :/
14:26:54*TheSeven isn't sure if it ever worked...
14:27:03S_a_i_n_tit did, for a short time.
14:27:11 Join mcuelenaere [0] (~quassel@rockbox/developer/mcuelenaere)
14:27:23TheSevendid you try to use it while copying files back then?
14:27:51S_a_i_n_tyeah, that almost always aborted the transfer and left corrupt filesystem
14:28:01S_a_i_n_ta whole barrel of fun.
14:28:19mcuelenaerehmm CIA bot is slow?
14:28:31S_a_i_n_tdead I think
14:28:33TheSevennot just slow apparently
14:28:35pixelmaseems it's on strike
14:29:04*AlexP kicks CIA-6
14:29:07 Join JdGord [0] (~jd@
14:29:17AlexPthat usually gets a response I think
14:29:29pixelmayes, an "ow"
14:29:31TheSevenhm, the nand dumper crashed again with just 160MB transferred with a different cable on a different port
14:29:40TheSeveni managed to transfer >1.5GB yesterday
14:29:58AlexPwow :)
14:29:59TheSevenwow, that's slow seems to be slow too..
14:30:35pixelmadepends. Maybe AlexP's kick was slow too ;)
14:30:45AlexPIt was rapier swift :)
14:32:12CIA-6New commit by teru (r28362): fix that disktidy plugin and shopper plugin is not usable when Show Icons setting is turned off.
14:32:46AlexPkugel: Can you (or anyone else) check out for me? It reverts r28000 (or so I hope) for the 3.7 branch as it causes broken playback for a few people.
14:34:41kugelAlexP: if it's a simple revert go for it. I cannot test the patch because i'm also unaffected by the problem.
14:34:52kugelI could test if it compiles&boots if you mean that
14:34:58AlexPI meant more that I hadn't cocked up the revert :)
14:35:22AlexPIt compiles OK, I checked that - boots and works might be handy if youi have a moment
14:35:22kugelI haven't got a checkout of the 3.7 branch, do I need it?
14:35:34kugelcan you give me the build then? :)
14:35:35AlexPno, that's against trunk
14:35:54AlexPyep, one mo - it'll take me a minute to do some convoluted uploading
14:35:58AlexPfuze yeah?
14:36:32CIA-6r28362 build result: All green
14:38:20AlexPkugel: Actually, just to make sure (shouldn't change anything), I'll build the 3.7 branch with this
14:38:30AlexPIt'll take a few mins, compiling on atom atm :)
14:38:35kugelgood idea
14:38:39kugeloh, I'm sorry
14:38:57AlexPIt's my little home server :)
14:39:05AlexPThe only thing at home I have ssh access to
14:39:08 Quit BlakeJohnson86 (Ping timeout: 265 seconds)
14:39:39TheSevenoh, there's another remaining nano2g bug: some LCDs seem to have contrast problems
14:39:44 Join BlakeJohnson86 [0] (
14:40:08TheSevensomeone complained that the colors would be "greyish" on an ilitek LCD
14:40:18TheSevenon the leadis one it's fine
14:40:48S_a_i_n_tI have seen that. actually
14:40:55S_a_i_n_tI thought it was just the age of the LCD
14:40:58TheSevenhow bad is it?
14:41:05TheSeveni don't have an ilitek one to test
14:41:06S_a_i_n_tnot very.
14:41:42S_a_i_n_tit can be brought up the approximately the same colour as the other screen by increasing the brightness by 3~4 points
14:41:58TheSevenS_a_i_n_t: can you try reverting r28190?
14:42:20*preglow wishes someone would design a new eq screen
14:42:22S_a_i_n_tit's not a recent thing.
14:42:27S_a_i_n_tit;s always been this way.
14:44:16TheSevengrawity said that this revision (which reduces the LCD supply voltage) triggered it on his device
14:45:03S_a_i_n_tI call BS on that.
14:45:28CIA-6New commit by mcuelenaere (r28363): Ramdisk: add hotswap support
14:45:40kugelpreglow: you could be the one!
14:45:47S_a_i_n_tthat commit is actually a very good thing, the LCD was getting way too much power...and the high pitched scream it was emitting was bloody annoying
14:45:55S_a_i_n_tit is almost silent now.
14:45:59preglowkugel: you kidding me? i had to make someone else make the current one for me in the first place :P
14:46:14*TheSeven never heard sounds from an LCD
14:46:23preglowui work makes me sad
14:46:54S_a_i_n_tTheSeven: Really?
14:47:06TheSeveni have a leadis one though
14:47:13S_a_i_n_tprior to that patch, mine, and Buschel's I assume screamed like a banshee
14:47:15TheSevenso with all these minor issues brought up, what's the general opinion on promoting it to stable?
14:47:16CIA-6r28363 build result: All green
14:48:14JdGordAdd the needed tags and make the eq skunnable :p
14:49:29kugelthat alone doesn't solve the problem. We still have the crappy fm radio screen by default even though it's skinnable
14:49:42 Join Brainbug [0] (
14:51:38S_a_i_n_tTheSeven: Well, I think you know my views...I use RB and the RB bootloader on a daily basis without, I say go for it.
14:52:31AlexPEvanCarroll: Can you also try and see if playback works for you?
14:53:41kugelideally the change would've been made during the freeze and I could've simply downloaded the zip from the official RC download page :)
14:54:30AlexPwell yes
14:54:48AlexPBut the release process hasn't exactly been model
14:56:28kugelAlexP: boots and plays music fine
14:56:38AlexPkugel: great, thanks
14:57:00AlexPI'll wait for EvanCarroll to test, as he is one of the people with a fuze that got broken playback with r28000
14:57:24mc2739AlexP: EvanCarroll has an e280v2, not a fuze
14:57:36AlexPI'll build that then too :)
14:58:04kugelhm, sometimes an item in the playlist viewer has no text. does that only happen for me?
14:58:19kugelI haven't seen that on my phone
14:58:42AlexPI've never really used it
15:00:49 Join komputes [0] (~komputes@ubuntu/member/komputes)
15:04:30kugelcan r28362 go into the 3.7 branch?
15:04:30kugelTheSeven: if you make the nano2g stable the screendump fix needs to be backported
15:04:30AlexPyeah, should do
15:04:30AlexPlooks like a bug fix to me :)
15:04:30TheSevenkugel: the fix won't be backported as it affects other targets
15:04:30TheSeveni have increased the stack size to prevent the stack overflow on the branch though
15:04:51TheSeventhis isn't really nice, but i don't want to risk touching the file system code on the branch now
15:04:57kugelso the screen dump is fixed?
15:05:03AlexPI still think r28341 (bertrik) should be backported too :)
15:05:18TheSevenkugel: i haven't tested it yet, but it should be
15:05:43TheSevenlet me do a checkout on my build client...
15:06:00kugelr28341 has c++ style comments :/
15:06:19Brainbugwhile installation, the utility is unable to load can i download it manually?
15:07:21AlexPIt doesn't exist
15:07:27AlexPThat's a slight bug
15:07:38AlexPThe clip wasn't included in the 3.6 release
15:07:53AlexPGet RBUtil to install a current build instead
15:07:54Brainbugthen what can i do to get over the "filenot found"
15:08:21Brainbugi installed with the utility. still doesnt work
15:08:33AlexPneed more info
15:08:53TheSevenBrainbug: try choosing "current build" instead of "stable" in the utility
15:09:02Brainbugi see, my original thext wasn't sent
15:09:03BrainbugI got a problem with my Sansa Clip v1 on win7 x64. After installing with the utility, my player tells me "file not found"
15:09:15Brainbugill try
15:09:16AlexPDid you install a current build?
15:09:31AlexPIf not, then you didn't install Rockbox, so the bootloader of course can't find it
15:09:33 Join Topy44 [0] (
15:10:33 Quit soap_ (Ping timeout: 265 seconds)
15:10:46 Quit chattr (Ping timeout: 255 seconds)
15:10:56Brainbugthx. works
15:14:25 Quit krazykit (Ping timeout: 265 seconds)
15:14:56 Join soap_ [0] (~soap@rockbox/staff/soap)
15:18:40AlexPEvanCarroll: Please try instead, last link I gave you was for a fuze
15:20:14 Join krazykit [0] (
15:23:47 Join chattr [0] (
15:29:31 Quit Amec (Read error: Connection reset by peer)
15:30:33 Join Amec [0] (
15:32:47***Saving seen data "./dancer.seen"
15:40:06 Join FlynDice [0] (
15:41:09 Quit FlynDice (Remote host closed the connection)
15:44:06 Quit mcuelenaere (Ping timeout: 272 seconds)
15:50:16 Quit Amec (Read error: Connection reset by peer)
15:50:23 Join Amec [0] (
15:51:15 Join _s1gma [0] (~d.d.derp@
15:53:34 Quit JdGord (Quit: Bye)
15:54:31 Join slater [0] (
15:55:53slaterhello everybody, i am trying to port rockbox on yp-r0, i was able to decrypt firmware and I mounted the firmware rootfs.
15:56:30slaterthe thing is i am kinda stuck, were does linux usually tell the mp3 player which folders are shared when connected with usb
15:56:49slateri would like to share all 5 partions when connecting with usb
16:00:11 Quit JdGordon (Ping timeout: 240 seconds)
16:01:23kugel /etc/export maybe?
16:01:39 Join mcuelenaere [0] (
16:01:39 Quit mcuelenaere (Changing host)
16:01:39 Join mcuelenaere [0] (~quassel@rockbox/developer/mcuelenaere)
16:01:45kugelcould be very well be hacked into the kernel or some obscure config file as well though
16:01:48 Quit mcuelenaere (Remote host closed the connection)
16:02:59 Quit teru (Quit: Quit)
16:03:31slaterhm that would be really bad
16:03:44 Quit krazykit (Ping timeout: 272 seconds)
16:04:08slaterif [ "$1" = "stop" ]
16:04:12slater echo Unmounting filesystems
16:04:14slater umount -a -r
16:04:16slater mount -o remount -r %root% /
16:04:18slater [ -x /sbin/swapoff ] && swapoff -a
16:04:28slaterif [ "$1" = "start" ]
16:04:30slater echo Mounting filesystems
16:04:35slater # /proc is mounted by the mount-proc-sys script
16:04:38slater # mount -t proc proc /proc
16:04:40slater if [ -z "$(mount | grep "/ type nfs (rw")" ] && [ -n "$RAMDIRS" ]
16:04:42slater then
16:04:45slater echo Rootfs is read-only, mounting rw filesystem
16:04:47slater mount -t $TMPFS rwfs /mnt/rwfs -o size=$TMPFS_SIZE
16:04:49slater for i in $RAMDIRS
16:04:51slater do
16:04:53slater if [ ! -e /mnt/rwfs/$i ]
16:04:56slater then
16:04:58slater cp -a $i /mnt/rwfs/
16:05:00slater mount -o bind /mnt/rwfs/$i $i
16:05:02slater fi
16:05:04slater done
16:05:06slater fi
16:05:07Mode"#rockbox +o TheSeven" by ChanServ (ChanServ@services.)
16:05:08slater #if [ "$TMPFS" = "tmpfs" ]
16:05:10slater #then
16:05:12slater #mount -t $TMPFS shm /dev/shm
16:05:14slater #fi
16:05:16slater if [ ! -e /etc/mtab ]
16:05:18slater then
16:05:19Kick(#rockbox slater :please use a pastebin for that) by TheSeven!~TheSeven@rockbox/developer/TheSeven
16:05:27 Join krazykit [0] (
16:05:44Mode"#rockbox -o TheSeven" by ChanServ (ChanServ@services.)
16:05:58 Join slater [0] (
16:06:15slatersorry about the earlier post
16:08:37 Join Jaykay [0] (
16:10:57AlexPslater: pastebin is a wonderful invention :)
16:11:49 Quit krazykit (Ping timeout: 240 seconds)
16:12:10slateryeah i didnt think about it
16:13:13 Join n1s [0] (~n1s@rockbox/developer/n1s)
16:16:46 Join anewuser [0] (anewuser@unaffiliated/anewuser)
16:18:47 Join krazykit [0] (
16:28:17 Join T44 [0] (
16:32:21 Quit Topy44 (Ping timeout: 252 seconds)
16:34:11 Quit kugel (Remote host closed the connection)
16:38:49 Quit Brainbug (Quit: CGI:IRC (EOF))
16:42:23 Quit Amec (Read error: Connection reset by peer)
16:42:48 Join Amec [0] (
16:47:40 Quit T44 (Read error: Connection reset by peer)
16:51:32 Join Topy44 [0] (
16:54:02 Join toffe82 [0] (
17:01:02 Part slater
17:03:07 Join jfc^3 [0] (
17:04:57 Quit Dreamxtreme (Read error: Connection reset by peer)
17:06:21 Quit jfc^2 (Ping timeout: 240 seconds)
17:10:14 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
17:11:55EvanCarrollAlexP: last link worked through
17:12:18AlexPEvanCarroll: Great, so I'll commit that to the release branch
17:12:58AlexPmusic etc. was fine?
17:12:58 Quit krazykit (Ping timeout: 240 seconds)
17:15:14EvanCarrollEverything was excellent.
17:15:36EvanCarrollI thought it was someone else that made my distro through
17:16:44EvanCarrollthat's the one I'musing now that's workinggreat
17:18:13AlexPWell it wasn't me
17:18:30AlexPthat made that one I mean
17:18:54AlexPI just wanted you to test the reversion I did before I committed it, so cheers
17:19:22EvanCarrollah. I can do that and get back to you tomorrow, want me to try it?
17:19:50AlexPI thought you did?
17:19:59AlexPWhat I meant was - please test :)
17:20:21 Join krazykit [0] (
17:20:33EvanCarrollNo, I used, I'll try tonight
17:20:42AlexPcool, thanks
17:21:08AlexPI don't know what the previous one is - I suspect it is identical, but it'd be good to have confirmation on this one
17:21:11EvanCarrolldid you revert r28000 ?
17:21:20AlexPyeah, but on the 3.7 branch
17:21:39AlexPOnce you've confirmed it working I'll commit it :)
17:21:44EvanCarrollcool I think his was on trunk
17:21:51AlexPIt should be identical
17:21:58AlexPbut it is always best top be safe
17:26:09 Join esperegu [0] (~quassel@
17:29:20 Join Dreamxtreme [0] (~Dre@
17:32:49***Saving seen data "./dancer.seen"
17:34:47 Join LambdaCalculus37 [0] (~3f74f70d@rockbox/staff/LambdaCalculus37)
17:44:40 Quit TheSeven (Ping timeout: 255 seconds)
17:54:54 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
18:12:35 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
18:21:15 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
18:22:39 Quit petur (Quit: now sports!)
18:23:55 Join DSStrife89 [0] (~Strife89@
18:27:52 Join T44 [0] (
18:29:37 Quit LambdaCalculus37 (Quit: food)
18:31:09 Quit Topy44 (Ping timeout: 252 seconds)
18:33:28 Join slooopy [0] (
18:33:36 Quit toffe82 (Remote host closed the connection)
18:40:12 Join LambdaCalculus37 [0] (~rmenes@rockbox/staff/LambdaCalculus37)
18:52:11 Join TheLemonMan [0] (~lemonboy@
18:58:50 Join saratoga [0] (9803c6dd@gateway/web/freenode/ip.
18:59:10saratogacan rockbox on android install to the SD card? my old G1 has hardly any disk space on the internal memory
18:59:16 Join toffe82 [0] (
18:59:28 Quit froggyman (Ping timeout: 264 seconds)
19:01:17 Quit DSStrife89 (Quit: Low battery!)
19:02:17 Join DSStrife89 [0] (~Strife89@
19:08:01 Join Topy [0] (
19:09:31 Quit hebz0rl (Quit: Leaving)
19:09:54LambdaCalculus37saratoga: I installed Rockbox on my HTC Magic's SD card via apps2sd and it works.
19:11:34 Quit T44 (Ping timeout: 240 seconds)
19:17:11 Quit DerPapst (Quit: Leaving.)
19:20:48 Part Shapeshifter
19:25:22 Join bertrik [0] (
19:25:22 Quit bertrik (Changing host)
19:25:22 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
19:25:26 Quit LambdaCalculus37 (Quit: back to work)
19:26:01 Join kugel [0] (~kugel@rockbox/developer/kugel)
19:28:14 Quit DSStrife89 ()
19:31:14 Join Strife89TX [0] (~cstrife89@
19:31:56saratogaso i need to get apps2sd?
19:32:24 Join Horscht [0] (~Horscht@xbmc/user/horscht)
19:32:32Strife89TXQuick question, what's the current revision number?
19:32:51***Saving seen data "./dancer.seen"
19:35:16Strife89TXBest to update before I file a bug report then.
19:36:02*Strife89TX is on r27988
19:37:21 Join DSStrife89 [0] (~Strife89@
19:37:23 Quit Strife89TX (Quit: Switching to Blazer.)
19:37:51kugelsaratoga: froyo has it built in
19:38:06saratogayeah, but does it work?
19:38:13saratogaIIRC an app has to actually support it
19:38:13kugelmy plain cyanogenmod install has it
19:38:14 Join LambdaCalculus37 [0] (~3f74f70d@rockbox/staff/LambdaCalculus37)
19:38:32kugelyou could just try it?
19:38:48kugelmculenaere added support for it but I don't know what is moved. everything cannot be moved because some paths are hardcoded
19:38:56saratogayeah i'll figure out how to install it
19:39:13kugelinstall what? i thought you run cyanogenmod with froyo
19:41:06saratogais there a prebuilt somewhere for android?
19:41:13saratogaor is the apk all i need?
19:42:04scorchekugel: suggested debugging procedure for rockbox that freezes on boot on an android device?
19:42:12kugelthe apk is enough
19:42:46kugelscorche: freeze on booz?
19:42:59scorcheon booting rockbox, yes
19:44:33scorcheit apparently freezes on the rockbox logo
19:44:48kugelit takes a while to unzip
19:45:09 Join bmbl [0] (~bmbl@unaffiliated/bmbl)
19:46:09kugeltry adb logcat
19:47:15 Quit TheSeven (Ping timeout: 252 seconds)
19:47:26 Join T44 [0] (
19:47:51saratogawhats the command to install an apk from the terminal emulator?
19:47:59 Quit Topy (Read error: Connection reset by peer)
19:48:08scorcheterminal emulator on the device ot terminal emulator from adb?
19:48:45saratogai don't have adb setup
19:49:40saratogathe obvious "install rockbox-320x480.apk" doesn't work
19:49:44scorchewhy do you need to? your browser likely popped up with a downloads screen after it finished - just click on the apk and it should try to install
19:50:02saratogadoes android have a file browser?
19:50:08saratogai can only see the file via terminal
19:50:40saratogaoh apparently it does
19:51:36scorcheopen up the browser > press menu > more > downloads > the apk
19:51:58saratogayeah i grabbed it via the gui off the sd card
19:53:00kugelsaratoga: browse the apk with a file browser
19:53:18saratogagrid view is the worst thing ever
19:53:56scorcheyou can change it to point in the config file - make sure to have a decently sized font though
19:54:11saratogai switched to point, but how do i go back a level in the menus?
19:54:23 Quit DSStrife89 (Quit: ClIRC v0.05 (homebrew IRC client for Nintendo DS))
19:54:29scorchetouch the top level
19:54:40kugelback button works in svn
19:54:44saratogaah, thats not very intuitive
19:55:47saratogai don't seem to be able to access the Files menu option
19:55:52saratogait just pauses forever if i click it
19:56:46kugelwe cant fix that
19:57:10saratogawhats the issue?
19:57:11 Join fritschy [0] (
19:57:36kugelsaratoga: the slowness of the device
19:57:47saratogait'll move about 2.5MB to the SD card
19:59:21saratogaweird if i relaunch it the files menu works immediately
19:59:29kugelbecause linux caches
19:59:30saratogaprobably some weirdness on my crappy g1
19:59:36 Quit komputes (Remote host closed the connection)
19:59:48kugelI observe the same, just not as bad as you probably
20:00:13fritschyHow common are crashes+fs-corruption? I have a sansa-clip-v1 and after some hours of playback it stops (hangs, no display, not off either) and afterwards the fs (music files only so far) is corrupted... this happened with a (the?) daily build from 20101019.
20:00:27saratogaare the builds on your page behind SVN?
20:00:50kugelprobably. you can see the revision number as usual in the debug menu
20:00:53saratogaritschy: i've put a couple hundred hours on mine over the year and never had corruption
20:01:23kugelscorche: still no worky for you?
20:01:47fritschysaratoga: i thought of hwardware trouble already...
20:01:54kugelsaratoga: can you see what has been moved?
20:02:02fritschysaratoga: which rockbox build are you using?
20:02:13kugelif it's the .so files then app2sd will probably break future updates
20:02:30saratogacurrent build from a couple days ago
20:02:50saratogakugel: perhaps we should include a larger default font for android, the current one is a bit smaller and theres no larger font in the apk
20:02:57fritschysaratoga: thank you, which OF version anyway?
20:03:08kugel27-adobe right? I find it sufficiently large
20:03:17kugelit could just need some spacing
20:03:22saratogayeah that would help
20:04:23saratogabut theres plenty of unused screen space at 320x480, so might as well go a little bigger too
20:04:23saratogafritschy: i don't remember, i don't use the OF
20:04:23kugelyou don't have album art do you?
20:04:32saratogakugel: i don't have any music, just thinking of the menus
20:04:34fritschysaratoga: isn't that needed to install rockbox? ;) sry, pretty new stuff for me.
20:05:02saratogayes but i don't use the OF so I don't care what version is on the device
20:05:55fritschyi meant which version you used for rockbox install :)
20:06:04kugelany version works
20:06:12saratogalike i said i have no idea
20:06:20saratogakugel: where do i put the fonts?
20:06:32kugel /sdcard/rockbox
20:06:33scorche|shkugel: apparently it now boots, but no sound - i do not have the device in front of me, but am trying to help another person
20:06:47saratogain a .rockbox folder or should I remove that?
20:06:59kugelremove it
20:07:11saratogaso /sdcard/rockbox/fonts?
20:07:15scorche|shso you will have /sdcard/rockbox/fonts - yes
20:07:47saratogawhy isn't the stock font shown there ?
20:08:32kugelit installs everything in the zip into the internal memory, but you can add stuff onto the sdcard
20:09:02scorche|shkugel: is there a 400x240 build anywhere?
20:09:42kugelbut you can try to compile a 240x400, I think it even has cabbie since that's the ondas resolution
20:09:51kugelthere's no landscape RaaA yet
20:10:27 Quit TheLemonMan (Quit: Help me, i got shot! *DIES*)
20:10:35saratogaanother nice UI tweak would be to make it a little harder to get into context menus
20:10:52saratogathe touch screen on my G1 seems to interpret a lot of normal presses as long presses
20:11:56saratogathe touch buttons on the WPS are quite nice though
20:12:09saratogaand Cabbie seems to work with 32 point font, at least nothing is missing that i can see
20:12:15scorche|shwell, yeah...240x400 works too - I assume this information is still current for building?
20:13:01scorche|shwell, meant to link the AndroidPort wikipage, but i assume that is the one you are thinking of as well?
20:13:04saratogamight be nice to redesign cabbie a bit though so that the buttons are larger
20:13:14saratogathe volume scrolling is a bit tricky for instance
20:13:17saratogaotherwise it works quite nicely
20:13:39saratogaabsolute mode seems to work quite well, we should consider making it default on android
20:13:41kugelyou usually don't use rockbox' volume control anyway
20:14:05kugelsaratoga: see mailing list :)
20:14:09saratogain the current build the hardware volume buttons will work?
20:14:17saratogakugel: i try not to read the mailing list :)
20:14:28kugelthe hardware buttons do work globally for all apps
20:15:17saratogathey don't seem to do anything in rockbox for me, although that could be something wrong with my phone
20:15:57kugelrb doesn't hook into the global volume control
20:16:09saratogaah thats what you meant
20:16:30saratogaso what did you mean by not using rockbox's volume control?
20:16:54kugelyou can do it, but you usually don't because the global volume is more convinient and makes more sense to use
20:17:07kugelif anything you set the rb volume once then never touch it again
20:17:40saratogaoh i misunderstood, i thought you mean that global volume doesn't change rockbox
20:17:44 Join komputes [0] (~komputes@ubuntu/member/komputes)
20:17:53kugelit changes all apps, rockbox as well
20:18:06kugelbut rb has a seperate volume control for its own
20:20:15 Join DerPapst [0] (
20:22:09saratogaits probably something wrong with my phone then
20:24:22 Join Buschel [0] (
20:25:49Buschelgrmpff, my mpc filterbank resort seems to be slowing down my former optimizations... in total +/-0 on PP5022
20:26:36n1sBuschel: you saw my cf patch?
20:26:44saratogabuschel: i've been working a little on the mp3 filterbank
20:27:06Buscheln1s: yes, that was 0.9 MHz faster compared to svn?
20:27:15Buschelsaratoga: any patch at hand?
20:27:24n1sBuschel: yes
20:27:43Buscheln1s: good. so, I should finally comitt then
20:27:50saratogai think by unrolling it 4x and treating the %4 =0,1,2,3 cases separately I can pack the values such that each D coefficient load will load two values sequentially everytime
20:28:10saratogathe code will be huge though
20:28:50saratogai don't have a patch yet, just lots of dumping info from the current filterbank trying to see how it works
20:29:08n1sBuschel: since pretty much all regs were used already i couldn't use a big movem but the sequential access makes the emac's parallel loading faster
20:29:17saratoganext step will be to rearrange the D array and then the filterbank so its all sequential
20:29:23saratogathen try it with packed 16 bit values
20:30:06saratogathat should give a huge speed up, since i'll have more then enough register to work around the load latency, and arm5e has a single cycle throughput on 32x16 multiply accumulate instructions
20:31:26n1ssequential access would help cf on mp3 too for the exact same reason as it did on mpc
20:31:46saratogahow many registers does CF have?
20:32:06n1s8 dregs and 8 aregs of which one is sp
20:32:15saratogais it even a load/store system?
20:32:37n1snot really, most instrs can work on memory
20:32:49saratogaok so thats a little different
20:33:03saratogatheres no Dcache, but does it hold the latest memory access in memory somehow?
20:33:15saratogai guess the memory probably fetches more then 4 bytes per load
20:33:39 Quit slooopy (Ping timeout: 240 seconds)
20:33:42n1smemory access is slow and dram access is *slow*
20:33:53saratogaright now the filterbank steps through all the odd values, then comes back and does all the even values, so its probably not too bad on systems with L1 dcache
20:34:11n1sblock accesses help a lot for dram and the emac paralell loads help a bit in emac blocks
20:34:33saratogaok so this might be useful for cf too
20:34:46saratogaif i get it to work i'll add a c version of it too so people can look into it for CF
20:35:05Buschelnano2g behaves totall different than PP5022
20:35:13saratogawell yeah its arm9
20:35:16n1sthese parallell loads can use 16 bit offsets (which mpc and mp3 both do) but that requires an extension word which makes the whole thing slower than a nice postinc access
20:35:29saratogaslow RAM, faster load/store, and more pipeline interlocks
20:35:58saratogayou mean it can do 32 bit half world aligned loads?
20:36:01n1sthat's where (i think) the whole speedup of mpc comes from on cf
20:36:44saratogaPP is a little weird since it has fast DRAM and almost no pipeline interlocks, other arm chips are a lot different
20:37:21n1ssaratoga: i can try writing cf code if you write the c
20:37:22saratogathats why most of are ARM asm code is crap :(
20:37:41saratogan1s: cool, i'll try and get something together soonish, just busy with work and a little burned out on coding
20:38:28n1salso no i mean that the adress can be offset by a 16 bit value, but that requires an extension word so the whole instr becomes 6 bytes long which i think something chokes on because it is a lot slower
20:39:09n1sso in essence "sequential accesses are good" :)
20:39:20Buschelsome numbers (always PP/S5L in MHz): svn 22/36, patch v7 22/45, patch v8 23/43 (v7 = "dumb" ldm all coeffcients like c reference implementation, v8 = make us of symmetries)
20:42:25n1sBuschel: pp is also weird because of the dualcore thing
20:42:52Buschelmpc only uses 1 core, therefor this should not have a mjor effect
20:43:01n1sah, right, was thinking mp3
20:43:37Buschelbut saratoga is right. even though PP is not the most loved processor, it seems to be one of the fastest in terms of codec efficiency
20:43:45n1sBuschel: if you post your asm somewhere i can see if i can see any obvious interlocks, the arm asm code in svn sure has lot's of them
20:43:49Buschelvery fast iram
20:45:53saratogaPP has good efficiency because it has very fast DRAM and IRAM, so if you only look at performance per clock, its very good
20:46:19saratogaonly AMSv2 is better, and even then only if you look at unboosted performance (since boosted the IRAM gets a lot slower per clock)
20:46:59 Join slooopy [0] (
20:47:12saratogaFWIW if you want to PP fast, you can use the COP like MP3 does
20:47:25saratogado the DCT on one core, the synthesis filter on the other
20:48:37Buschelwell, more focussing on the nano these days ;o)
20:48:59saratogawell try and avoid all those weird strides that the mp3 decoder likes to do
20:49:32saratogathe nano2g isn't going to like those
20:49:57saratogaalso mind the pipeline interlocks when loading and then immediately using them in a multiply
20:51:32 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
20:53:43kugelsaratoga: the hw buttons on your phone should change the global media volume, whether in rockbox or not
20:56:05Buschelwow, a small change in the decoder (reducing the amount of data moved via memmove) had no impact on PP, but saves ~3 MHz on S5L
20:56:22BuschelS5L _does not_ like memory accesses...
20:56:43n1si thought about those memmoves, they are likely slow on cf too
20:56:57n1sso they can be reduced?
20:57:25Buschelagainst my patch, yes.
20:57:43Buschelwait a sec
20:58:25Buscheln1s: you need to exchange both memmove lines like here ->
20:59:36 Quit saratoga (Ping timeout: 265 seconds)
21:01:37n1si'll try it out
21:03:03 Quit Zambezi (Read error: Connection reset by peer)
21:03:28 Join Zambezi [0] (Zulu@
21:05:11n1sBuschel: that screws up the sound
21:05:30 Join krabador [0] (
21:05:41n1sunless i screwed something up
21:06:00n1sit sounds high pitched, tinny and faint
21:07:44Buscheloutput is identical here
21:08:40 Quit kevku (Read error: Operation timed out)
21:10:30n1sah, no something else is screwed up, sorry
21:12:16BuschelI have updated two patches with some numbers...
21:16:29n1ssounds fine with that memove change, running test codec now
21:20:29n1shmm, the mmemove changed slowed cf down ~0.3MHz
21:21:15Buschelthe best trade-off for now seems to be: keep svn :)
21:21:47 Quit markun (Read error: Connection reset by peer)
21:22:01n1sthe arm asm has a lot of ldr-multiply interlocks on newer arm cores
21:22:41Buschelthe 2nd (#else) path?
21:22:44 Quit scorche|sh (Ping timeout: 265 seconds)
21:22:56 Quit alexbobP (Ping timeout: 255 seconds)
21:22:58n1si meant it slows down 0.3 MHz versus keeping the big movem, it's still 0.6 faster than svn
21:23:24 Join scorche|sh [0] (
21:23:58Buscheln1s: regarding speed -> yes, but the decosing is slowed down by 4 MHz (patch vs. svn) for S5L
21:24:23 Join alexbobP [0] (
21:24:49n1sBuschel: ok, do you think that is cache or something else?
21:26:10Buschelcould be...
21:26:25 Join markun [0] (
21:26:25 Quit markun (Changing host)
21:26:25 Join markun [0] (~markun@rockbox/developer/markun)
21:26:30n1son arm11 ldr has a 2c result latency and multiplies have multiplicands as "early regs" so you get a 3c stall between ldr and smull/smlal etc if the loaded reg is used as a multiplicand, i think arm9 has a similar, shorter stall
21:27:10Buschelso, reordering the ldr/smlal a bit could help?
21:28:01n1syes, i'll try it out on the beast
21:32:54***Saving seen data "./dancer.seen"
21:35:13n1sso what is the difference between 9a and 9b?
21:35:42Buschelwhich path of the ARM asm is used
21:36:14Buschel9a = "dumb" like the c reference, 9b = more sophisticated
21:39:01Buschel9b only loads half of the D coefs due to internal symmetries
21:39:10Buschelthis scales a lot on S5L
21:39:45Buschelbut, 9b does not access the data consecutive and jumps a lot by large amounts
21:40:42 Quit AlexP (Remote host closed the connection)
21:40:44n1sbut if it's basically the same as in svn with other offsets the difference has to come from somewhere
21:41:05 Join AlexP [0] (~alex@rockbox/staff/AlexP)
21:41:18n1s9b is 4MHz slower than svn on the beast too
21:41:21Buschelbefore the re-ordering PP5022 was impacted like "9a and 9b do random access, but 9b does less loads -> faster". now it is like "9a does ldm, 9b does random access but on less data -> 9b is slower)
21:42:31Buschelhmm, the reordering has the effect that writing to RAM is more random like, it was consective before. but the writed data amount is small.
21:42:39n1sarmv4 benefits a lot from ldm so that's not too surprising but how the newer arms can slow down so much surprises me a bit
21:42:51amiconn[20:43:34] <Buschel> but saratoga is right. even though PP is not the most loved processor, it seems to be one of the fastest in terms of codec efficiency <= Depends a lot on codec. For APE, not true at all
21:43:23Buschelamiconn: but for the most codecs which are dct or filterbank based it is true
21:44:02amiconnARMv4 is horrible when it comes to processing packed 16 bit data, so on that we use 32 bit integers in the filter
21:44:12amiconnThat means moving twice as much data
21:45:09amiconnStill better than fiddling with packed ints (which svn did at some point)
21:46:39 Join froggyman [0] (
21:46:39 Quit froggyman (Changing host)
21:46:39 Join froggyman [0] (~seth@unaffiliated/froggyman)
21:47:07n1sBuschel: the reference asm is 0.4MHz faster on the beast too but that chip has 64bit memory bus so loading twice as much more data with ldm versus ldr should actually be about the same speed
21:49:54amiconnI think only the internal bus (i.e. to/from cache) is 64 bits wide
21:52:53kugeln1s: (correct me if I'm wrong) I think armv4 benefits from ldm so much because each ldr is 3 cycles (always). that's not the case anymore on new arm (constant 3 cycles -> 1 cycle + 2 result latency), so ldrs mixed with other insns are faster than ldm
21:53:36kugelso ldm doesn't slow it down, but it's just not optimal anymore. and since gcc never uses ldm AFAIK, C code might be actually faster
21:54:32n1syou are right about armv4 but not quite about ldm on newer cores
21:56:49n1son arm11 ldm (not messing with the pc) takes only 1c but the memory remains busy so no other memory accesses may start before it's done and registers start becomming available after 2c so if you can fill that delay with something else you can load as much as you want in only 1c :)
21:57:36kugelwasn't that different from ldm on armv5 again?
21:58:07amiconnOn arm9 it doesn't matter whether you use ldr or ldm, as long as you keep an eye on latencies
21:58:36amiconnThat applies both to ARM9 v4 and v5
21:58:40n1son arm9 ldm takes n cycles when n > 1 else 2 and ldr takes 1 cycle and has 1 c result latency
21:59:02n1s^what amiconn said
21:59:52 Quit LambdaCalculus37 (Quit: CGI:IRC 0.5.9 (2006/06/06))
21:59:59kugelbut well, ldm is much more effective on arm7 than on later cores
22:00:46kugelon later code you can get faster code with ldr if you get the latencies right
22:00:51kugellater cores*
22:01:16 Quit krabador (Ping timeout: 252 seconds)
22:01:19amiconnldm might be the better choice on arm9 as well
22:01:44amiconnAlthough it doesn't reduce cycle count, it does reduce code size, possibly improving caching
22:02:24kugelyou could split ldms into ldrs to fill multiply latencies for example
22:03:01kugelso I'm not so surprised massive ldm use has not such a good effect on arm9+ as on arm7
22:03:07amiconn...and get ldr latencies instead
22:03:34amiconnldm use should never make things worse
22:04:06amiconnOn arm9 it may be indifferent, or *very slightly* better, on arm7 and arm11+ it's definitely better
22:04:35kugelbetter than sequential ldrs sure
22:07:38kugelthe libmad arm asm is hugely faster on arm7 (pressumable due to heavy ldm usage), but is in fact slower than C code on arm11
22:08:03n1skugel: it has a bunch of ldr stalls
22:08:03amiconnThat's almost for sure not caused by ldm use
22:09:17kugeln1s: the asm is also faster on arm9e (although not as much)
22:09:56n1skugel: yes but those stalls are much shorter there
22:11:27kugelbecause of the early regs in multiplies?
22:12:36n1syes the 2 cycler ldr latency and the early regs, the ldr latency is only 1 c on arm9 and it doesn't have early regs afaiu
22:13:01n1sso a ldr-smull stall is 3c on arm11 and 1c on arm9
22:13:03kugelI thought ldr latency is 2c on arm9 as well
22:13:09 Join krabador [0] (
22:13:29n1snot according to the system developers guide ata least
22:13:57n1shmm, this mpc code uses all registers available already
22:15:12Buscheln1s: you could split an ldm-4 to two ldm-2 and use the free 2 registers
22:15:50n1sBuschel: yes, but that feels like a cop-out :)
22:16:08n1si need to go to bed soon though so maybe i'll look closer tomorrow
22:16:16Buschelsame for me
22:16:25n1sit's too bad 4 regs are taken for accumulators
22:17:43Buschelwell, this give the speed up. you can re-use the D coefs for two multiply-adds at the same time and save 50% memory loads
22:18:23n1sworth a try
22:18:46Buschelhowever, good night! :)
22:19:00n1sgood night!
22:19:05 Quit Buschel (Quit: ChatZilla 0.9.86 [Firefox 3.6.11/20101012113537])
22:21:24n1smaybe stacking some pointers are a better idea
22:22:26 Quit Xerion (Ping timeout: 276 seconds)
22:33:18 Quit krabador (Ping timeout: 265 seconds)
22:33:43 Quit Jaykay (Quit: ChatZilla 0.9.86 [Firefox 3.6.11/20101012113537])
22:44:00 Join saratoga [0] (9803c6dd@gateway/web/freenode/ip.
22:45:17 Join krabador [0] (
22:45:36 Quit fritschy (Quit: Lost terminal)
22:46:45 Join Xerion [0] (
22:47:18 Quit Rondom (Quit: Bye)
22:47:47 Join Rondom [0] (
22:54:06saratogaBuschel: i don't know how mpc's filterbank works, but if you could make those strides a little more regular it would probably help a lot
22:54:26saratogaperhaps you could rearrange the coefficients in memory
22:55:23saratogathe V coefficients in particular look bad
22:56:21 Join wodz [0] (
22:58:49wodzAlexP: I cut down png decoding of 'United Kingdom flag' to about 5s on coldfire. I didn't tested what's the impact of this on other targets. I think there is room for farther optimizations.
22:59:07AlexPGood work!
23:00:49wodzUnfortunately I lost my patch for 'aspect correction' - I will have to implement this from scratch.
23:02:02pixelmawodz: sorry, didn't get around to test yet
23:02:20kugelwodz: isn't it anywhere in the irc logs?
23:03:17wodzkugel: the patch? no
23:04:00n1ssaratoga: that's what his patch in fs#11702 does
23:04:42wodzI thought I posted it to FS but the patch doesn't include the file of interest.
23:04:55kugelwodz: oh, but the latest patch on the task mentions it has the correction
23:05:29kugelok, I understand. I recalled (wrongly) you put it on pastebin or something
23:05:40wodzyes but I found out it is incomplete
23:05:56saratogaoh i only looked at the last
23:06:30wodzI wanted to produce patch from clean tree and forgot to locally commit new files :/
23:06:58saratogahmm the last has it too, but the ASM still looks ugly
23:07:57saratogaa lot of the steps are still bigger then a cache line
23:08:12 Join klapaucjusz [0] (
23:08:20n1syes i'm poking at it a little to see if i can rearrange it in a way that's more friendly to newer arm cores but not having much luck
23:08:57klapaucjuszI'm wondering if anyone is maintaining the mpegplayer plugin?
23:09:21klapaucjuszIt'd be nice if it could support MPEG 1/2 files muxed with AAC audio.
23:09:23n1ssaratoga: something that patch does slows it down about 4MHz on both the beast and nano2g
23:09:38n1sso likely something to do with memory or cache
23:09:42klapaucjuszThis would avoid transcoding audio when grabbing from Youtube.
23:09:46saratogaklapaucjusz: no one is
23:10:00saratogaand you'll have to transcode anyway, youtube doesn't use MPEG1/2 video
23:10:22klapaucjuszYeah, but only the video.
23:10:32saratogais that any better?
23:10:39klapaucjuszVideo artefacts are okay for me, audio not so.
23:10:42saratogai think converting audio is quick compared to video
23:10:45klapaucjusz(Especially when downscaling.)
23:11:13klapaucjuszYeah, but it creates very unpleasant artefacts. It's not like Youtube audio is very high quality in the first place...
23:11:24wodzklapaucjusz: I would say it is a matter of transcoding software
23:12:05AlexPklapaucjusz: Nobody is working on it, but patches are always welcome
23:12:17AlexPSo if someone were to work on it, great! :)
23:12:35*klapaucjusz is grateful to AlexP for a straight answer to a straight question.
23:13:47saratogaif converting from low bitrate AAC is adding artifiacts you're not doing it right
23:14:13wodzklapaucjusz: I was saying that if you have audio artefacts after transcoding than probably the transcoding process is to blame
23:15:01saratogagetting streaming AAC quality isn't exactly hard
23:15:30pixelmatranscoding lossy to lossy isn't nice. I recently wondered that too btw. since Rockbox can play AAC anyway
23:16:36 Join edboyer93 [0] (
23:17:28 Quit benedikt93 (Quit: Bye ;))
23:19:59klapaucjuszBy the way, nice to see how much the nano2g port has progressed since last time I checked :-)
23:26:20 Quit esperegu (Read error: Connection reset by peer)
23:31:02klapaucjuszStupid question: on Debian, I build the cross-compiler myself, right? Or do I download a binary somewhere?
23:32:57***Saving seen data "./dancer.seen"
23:33:35Bagderyou build it with the script
23:35:57saratogasee this:
23:36:57*klapaucjusz makes some coffee while git-svn runs.
23:45:01 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
23:51:06bertrikAlexP, r28341 was more of a new feature for targets with a wm8975 codec than really a fix for an existing feature
23:51:27 Quit domonoky (Read error: Connection reset by peer)
23:51:56bertrikAbout the c++ style comments in r28341, I noticed those after the commit, but didn't think it was important enough to do another commit just to fix those.
23:52:11bertrikI wouldn't mind some kind of compiler setting to warn me about those.
23:52:46 Quit bmbl (Quit: Verlassend)
23:56:01TheSevenbtw, what's so bad about them?
23:57:04bertrikdocs/CONTRIBUTING says to use only /* */
23:57:53saratogawe don't use them for some reason or another

Previous day | Next day