Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

#rockbox log for 2009-12-24

00:00:49 Quit Kitar|st (Read error: 60 (Operation timed out))
00:00:54 Join Kitar|st [0] (i=Kitarist@
00:01:32Unhelpfulamiconn: coldfire would just use extension words? and coldfire takes 3 ops to do the construction, i think?
00:03:45kugelsaratoga: that's so analog :)
00:04:09saratogait was a good excuse to understand how the code works :)
00:04:12saratogaplus its not very complicated
00:09:34 Quit sagemfreak_ ("Verlassend")
00:09:36***Saving seen data "./dancer.seen"
00:09:59 Quit Jaykay ("ChatZilla 0.9.86 [Firefox 3.5.6/20091201220228]")
00:12:37 Join Jaykay [0] (
00:13:06Jaykaywhy did r24099 increase binsize on all targets? because of CREDITS?
00:13:07 Join sagemfreak [0] (
00:15:03bertrikJaykay, IIRC, CREDITS is used by the credits plugin, so it should affect binsize
00:15:10Jaykay...and could the credits screen in rockbox be a plugin to save some kb? :)
00:15:23gevaertsJaykay: it is
00:15:38*bertrik adds a NOT in there somewhere
00:16:07JdGordonCasainho: ping?
00:16:27CasainhoJdGordon: pong
00:16:29Jaykaybut the revision also increased ram usage on all targets... why?
00:16:41 Quit relentless ("Leaving")
00:16:52JdGordonCasainho: I'm finally trying to get rockbox on my mini24340 and I'm not having much lcuk with openocd...
00:17:21 Join Actimehl [0] (
00:18:23CasainhoJdGordon: did you follow the instructions here?:
00:18:49CasainhoJdGordon: me and Bob_C, we got it... and I guess Domonoky also...
00:19:02JdGordonthats what I'm working through.... I've got openocd built, but cant get it to connect
00:19:16JdGordonthe board/mini2440.cfg file has errors apparently
00:19:30JdGordonError: target 'jedec_probe' not defined
00:20:29CasainhoJdGordon: did you used openocd before?
00:20:55Jaykaybertrik, gevaerts: ?
00:21:14bertrikJaykay, I don't know why it increased ram usage
00:21:16CasainhoJdGordon: what is that jedec_probe?
00:21:36Casainhowhat JTAG cable/dongle are you using?
00:21:50JdGordonno idea.. its in the flash bank init code
00:21:53JdGordonthe one supplied
00:22:59gevaertsJaykay: it's because r24097 didn't get successfully autobuilt
00:23:05CasainhoJdGordon: nor me nor Bob_C used that JTAG cable... looks like that one may work on Windows or something.
00:24:41Jaykaygevaerts: that means..?
00:24:55CasainhoJdGordon: "If you are using Linux or prefer to use OpenOCD, the Mini JTAG adaptor can also be used with OpenOCD. It should be configured as a "triton" adaptor type. Scripts for OpenOCD are in the Hg repository under "tools/OpenOCD". "
00:24:56 Quit faemir (Read error: 104 (Connection reset by peer))
00:25:12 Join faemir [0] (n=faemir@
00:25:27CasainhoJdGordon: did you make what Bob_C worte on that phrase?
00:27:01 Quit saratoga ("Page closed")
00:27:34gevaertsJaykay: it means the delta you see isn't for only r24099
00:28:37Jaykaygevaerts: would you please explain that for a clueless user? :)
00:28:48 Quit Omlet ("( :: NoNameScript 4.22 :: )")
00:32:22amiconnHmm, "Show track info" has quite a bit of information, but the codec format is missing
00:32:23gevaertsJaykay: the delta you see isn't for r24099 compared to r24098. It's for r24099 compared to the previous revision built
00:33:44 Quit merbanan (Read error: 60 (Operation timed out))
00:33:49 Quit Strife89 ("If you hold a Unix shell to your ear, you can hear the C.")
00:33:58Jaykaygevaerts: thanks, no i got it :)
00:34:06 Part Actimehl
00:34:49 Quit Jaykay ("ChatZilla 0.9.86 [Firefox 3.5.6/20091201220228]")
00:35:30flybackyou guys know anything about HQRP brand lithium batteries
00:36:50evilnickflyback: Sounds like #rockbox-community talk there
00:38:38flybacki'll probably get laid before I get an answer
00:38:40*flyback is 35
00:43:10 Quit Sajber^1 (Read error: 104 (Connection reset by peer))
00:43:48 Join stripwax [0] (
00:43:52 Join Sajber^ [0] (
00:44:57CIA-6New commit by amiconn (r24107): Assembler optimised bitreverse for ARM, again a little bit faster and smaller. * Separated 'x' and 'ret' gives gcc a bit more freedom regarding ...
00:47:58n1samiconn: nice, first commit in a long time :)
00:48:13n1show does the separation of x and ret help?
00:48:43*kugel thinks amiconn could've stuck to "mask" instead of "tmp", but well
00:49:10amiconnIt's not the mask...
00:49:32amiconnThe mask doesn't use a variable in the C version (but it does in the asm)
00:50:11 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
00:50:34stripwaxamiconn - how much time does tremor decode spend in that particular bitreverse ?
00:50:44 Quit evilnick ("Page closed")
00:54:03amiconn[23:11:15] <n1s> amiconn: gives a speedup of 0.1% for 96kbps going up for higher bitrates to 0.7% for 500kbps on c200 over the version i committed earlier today
00:56:35amiconnJdGordon: The status bar still appears too early
01:01:49 Join Sajber^1 [0] (
01:08:53n1swow, with only 3 to 4 small changes rockbox built with gcc 4.4 boots and plays on h300
01:09:37 Quit Sajber^ (Read error: 113 (No route to host))
01:10:03n1si'll do some benching in the next few days
01:10:15kugelplease do!
01:11:08 Quit n1s ("Lämnar")
01:28:10 Join Sajber^ [0] (
01:29:09 Join Sajber^2 [0] (
01:35:08 Quit DerPapst ("Leaving.")
01:35:18 Join Sajber^3 [0] (
01:35:18***Alert Mode level 1
01:35:18DBUGEnqueued KICK Sajber^1
01:35:18DBUGEnqueued KICK Sajber^
01:35:18***Alert Mode level 2
01:35:18DBUGEnqueued KICK Sajber^2
01:35:18DBUGEnqueued KICK Sajber^3
01:35:18***Alert Mode level 3
01:37:17 Quit stripwax ("")
01:45:20***Alert Mode OFF
01:48:37 Quit Sajber^1 (No route to host)
01:49:44 Quit bertrik ("De groeten")
01:49:48 Quit Res1 (Read error: 60 (Operation timed out))
01:49:59 Join Res1 [0] (
01:51:09 Quit Sajber^ (Connection timed out)
01:51:40 Quit Sajber^3 (Read error: 104 (Connection reset by peer))
01:52:25CtcpIgnored 4 channel CTCP requests in 0 seconds at the last flood
01:52:25*flyback decides to give the xmas tree animation hack one last chance to get done in time before giving up
01:52:26 Quit Sajber^2 (Read error: 113 (No route to host))
01:53:03*flyback bbl
01:54:42 Part pamaury ("exit(*(int *)0 / 0);")
01:55:08 Join Sajber^ [0] (
01:57:57 Quit matsl (Read error: 110 (Connection timed out))
02:04:11 Join sagemfreak_ [0] (
02:06:05 Quit sagemfreak (Read error: 60 (Operation timed out))
02:07:34 Quit Sajber^ ("Leaving.")
02:09:40***Saving seen data "./dancer.seen"
02:14:32 Quit GeekShadow ("The cake is a lie !")
02:24:01 Quit Casainho ("ChatZilla 0.9.86 [Firefox 3.5.6/20091215231400]")
02:26:20 Quit bluebrother (Nick collision from services.)
02:26:23 Join bluebroth3r [0] (n=dom@rockbox/developer/bluebrother)
02:36:52 Join toffe82 [0] (
02:40:37 Quit mt (Read error: 104 (Connection reset by peer))
02:41:06 Join mt [0] (n=mtee@rockbox/developer/mt)
02:41:28 Quit liar (Read error: 113 (No route to host))
02:47:25 Quit Topy44 (Read error: 60 (Operation timed out))
02:47:27 Join Topy44 [0] (
02:58:14 Part froggyman
03:10:25 Quit kugel ("exit(0);")
03:51:41 Join CaptainKewl [0] (
03:52:27 Quit FlynDice (Remote closed the connection)
04:09:41***Saving seen data "./dancer.seen"
04:23:41 Quit Hillshum (Read error: 60 (Operation timed out))
04:49:06 Join darkham [0] (
04:54:42 Join Hillshum [0] (
05:03:56 Quit thegeek (Read error: 54 (Connection reset by peer))
05:04:00 Join thegeek_ [0] (
05:08:26 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
05:26:27 Quit Hillshum (Read error: 60 (Operation timed out))
06:00:40 Quit darkham ("Sto andando via")
06:07:03 Join xavieran [0] (
06:09:43***Saving seen data "./dancer.seen"
06:27:55 Quit StealthyXIIGer (Read error: 110 (Connection timed out))
06:34:08 Quit CaptainKewl (Read error: 110 (Connection timed out))
06:40:38 Join webguest08 [0] (
06:40:57webguest08are any developers on?
06:42:09 Quit webguest08 (Client Quit)
06:43:58 Join Brendo [0] (
06:44:38BrendoAre any developers online?
06:46:27 Quit Brendo (Client Quit)
07:09:13 Part flyback ("Leaving")
07:33:27 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
07:51:15 Quit Horschti (Read error: 110 (Connection timed out))
07:59:03GodEateranyone got any ideas about committing #10832?
08:02:28JdGordonwho knows the AAC code?
08:02:40GodEatersaratoga I think
08:03:01GodEatermaybe linuxstb too?
08:05:10JdGordoncodecs really shouldn't be touched without someone that actually knows the code giving it the ok :)
08:05:28*JdGordon is inspired by FS #8844 but wants to do it better
08:05:54JdGordonany ideas what the token for smart image exansion shuold be? I want it to be able to let you choose which image and which start and end subimages to add...
08:06:03GodEaterindeed, but it looks ready to go - so I don't want it to linger in patch-hell, so I thought I'd bring it here for the logs :)
08:06:34*JdGordon thinks %xd|a|a|z| doesnt look very good
08:06:50 Part toffe82
08:09:46***Saving seen data "./dancer.seen"
08:13:08CIA-6New commit by jdgordon (r24108): Accept FS #8321 by Lee Kang Hyuk ...
08:13:43JdGordonmore than 2 years on the tracker that one :p
08:17:11GodEaterwe should celebrate
08:25:08GodEaterwhat's the deal with patches that are *never* going to make it into Rockbox core?
08:25:19GodEatershould they get to hang around in the tracker, or should we be closing them?
08:25:53GodEaterfor example #2653 - it's changing the behaviour of the select button in the file viewer
08:27:01JdGordonI think the outcome last time the question came up was close them
08:27:45GodEaterI agree, I don't see the point of leaving them there
08:27:51GodEatershall I?
08:29:32JdGordonI wouldnt mind one click insert functinoality
08:29:40JdGordonbut yeah, its never going to go in
08:29:58GodEaterthere's another later patch which is still open which tries to do the same thing
08:30:04GodEaterso you can have a go at that one ;)
08:31:09JdGordonI think the way it could be done is a "set select button action" option in the context menu which would last for the duration of the browser, the options would be all the ones in the playlist context menu
08:31:24GodEaterthat seems reasonable
08:32:07GodEaterI wish there was a column in flyspray that showed when a task was last edited
08:32:30JdGordonI of course wont waste my time though because it will inevitably turn into a "you're trying to bring customisability in blaa blaa blaa"
08:50:27 Join stoffel [0] (
08:56:09 Join flydutch [0] (
09:00:25 Quit faemir (Read error: 54 (Connection reset by peer))
09:00:47 Join faemir [0] (n=faemir@
09:21:24 Join liar [0] (n=liar@
09:32:48 Quit AaronM ("g'night to all and merry christmas eve")
09:39:16 Join petur [50] (n=petur@rockbox/developer/petur)
09:43:17 Join Tomis2 [0] (n=Tomis@
09:58:50 Join MethoS- [0] (n=clemens@
10:02:06 Quit Tomis (Read error: 110 (Connection timed out))
10:02:06 Nick Tomis2 is now known as Tomis (n=Tomis@
10:09:49***Saving seen data "./dancer.seen"
10:12:37 Join n1s [0] (n=n1s@rockbox/developer/n1s)
10:17:21 Quit BHSPitLappy (Read error: 60 (Operation timed out))
10:17:23 Join Omlet [0] (
10:19:37 Quit faemir (Read error: 110 (Connection timed out))
10:20:20 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
10:29:13 Join pamaury [0] (
10:45:16 Quit Omlet (Read error: 54 (Connection reset by peer))
10:47:24 Join Omlet [0] (
10:58:24 Join q0_0p [0] (
10:59:19q0_0panyone here know how to revive archos studio 20 after formatting the hdd?
10:59:49 Join evilnick [0] (i=5752157c@rockbox/staff/evilnick)
10:59:50q0_0pthe archos website has the AJZ file but i have no clue what to do with it and i've tried rockbox installer
11:00:16q0_0pjust clueless what to do right now. the hdd works like a hdd but when turn it on cannot play music so i assume firmware is not running good
11:02:37q0_0pis it critical to have the AJZ file?
11:02:53 Join faemir [0] (n=faemir@
11:03:58evilnickq0_0p: Have you searched for a way to reinstall the Archos firmware?
11:04:08q0_0pevilnick, could not find one online
11:04:17q0_0pevilnick, i have the AJZ file from archos site
11:04:24q0_0pevilnick, but not sure what to do with it
11:04:34 Nick Omlet is now known as Omlet^away (
11:04:47q0_0pevilnick, i've tried putting the AJZ file in the root of the drive
11:08:18amiconnThe archos Studio's firmware file isn't called ajbrec.ajz, but archos.mod
11:08:32amiconnIf it really is a Studio and you have an .ajz, you have the wrong file
11:09:03amiconnBut the archoses have their firmware built-in, the on-disk version is just an optional upgrade
11:11:30amiconnRight file, wrong name. In order for a Studio to pick it up, it must be named archos.mod, and reside in the root of the drive
11:11:50q0_0pamiconn, so all i have to do is rename the file to archos.mod?
11:12:07q0_0pamiconn, that would be cool if that was the answer :)
11:12:12q0_0pamiconn, will try
11:12:17*amiconn wonders why archos have this obviously wrong filename (even with wrong extension) on their site
11:13:20q0_0pamiconn, i've successfully got rockbox to work once but then somehow it disappeared and stopped working
11:13:42 Join merbanan [0] (
11:14:20pamauryDoes someone know if the rockbox loader resolves symbols at load time for plugins ?
11:15:30amiconnPlugins are linked to a fixed address
11:16:27amiconnq0_0p: The archos Player/Studio's built-in original firmware sometimes fails to find an archos.mod in the root
11:16:54q0_0pamiconn, u think its a goner?
11:17:06q0_0pamiconn, seems like original firmware fails
11:17:10amiconnIt depends on the state of the root directory (fragmentation etc)
11:17:24q0_0pamiconn, tried rockbox and it worked for a little bit
11:17:33pamauryDoes someone knows thing about the zxbox plugin ? I'm trying to determine if we can safely export errno to plugins via rb and #define errno *rb->errno but it seems like zxbox (and others) is using errno and strerror whereas it's not accessible
11:17:50 Quit merbanan (Read error: 60 (Operation timed out))
11:17:54amiconnIf you manage to make it load archos.mod again, rockbox should work too
11:18:25pamaury(perhaps it's dead code, I checked with a grep)
11:18:35amiconnThen you could check whether this archos is flashable (not all are). If it is, put rockbox in flash, and you won't have to worry about the OF not finding archos.mod anymre
11:19:01q0_0pamiconn, cool
11:19:33q0_0ppamaury, how do i check with a grep
11:19:59q0_0ppamaury, i only know how to grep when for example ls -la | grep 'something i want'
11:20:13amiconnOne method that will probably work is to format the drive, then put archos.mod on it as the first (and only) file. Better still, install rockbox directly after formatting
11:20:30q0_0pamiconn, will do that
11:20:33amiconnBe sure to format it as fat32, of course
11:20:43q0_0pamiconn, but when rockbox installs it also puts a foldoer called .rockbox
11:20:52q0_0pamiconn, should i rid that too? or keep it
11:21:10amiconnKeep it, it belongs to rockbox
11:21:21amiconnIt shouldn't do harm.
11:21:38q0_0pamiconn, flashing the whole thing sounds like fun
11:21:45amiconnThe main thing about a clean install after formatting is that the root directory won't contain deleted entries
11:22:31pamauryq0_0p: grep -R "errno" apps/plugins
11:22:59q0_0ppamaury, where is apps/plugins?
11:23:15amiconnI don't think anyone tried to track down *why* the OF sometimes fails to load archos.mod; my guess is that deleted files (maybe only earlier version of archos.mod) or even just a too large root directory confuse it
11:23:31 Join merbanan [0] (
11:23:47pamauryin the source code of rb but in a more general way, grep -R bla dir will search recursively for bla in files under directory dir
11:28:01q0_0pamiconn, how do i put rockbox in flash
11:28:09q0_0pamiconn, using rockbox utility?
11:28:22q0_0pamiconn, would that be the bootloader?
11:29:24 Quit merbanan (Read error: 60 (Operation timed out))
11:30:20evilnickq0_0p: Check the RB manual for that
11:31:13evilnickAnd it seems to be only for certain models of Archos players
11:32:11q0_0pevilnick, thx will look
11:36:37evilnick q0_0p: Check Advanced Topics then Rockbox in Flash - it took me a while to find, so hope that helps you!
11:37:35q0_0pevilnick, thx will read
11:38:34q0_0pevilnick, cool
11:39:09q0_0pevilnick, when i play the file can i play it in vlc or any media player?
11:39:32q0_0pevilnick, Play the file rockbox.ucl, or preferably rombox.ucl, and follow the instructions. The plugin handling this is rockbox_flash, a viewer plugin.
11:39:46evilnickIt means to play it on the archos itself
11:39:50q0_0pevilnick, oh
11:40:01q0_0pevilnick, the problem with my archos is
11:40:06q0_0pevilnick, it wont even let me play music
11:40:13q0_0pevilnick, so i'm not able to file brose
11:40:19q0_0pevilnick, not able to file browse
11:40:31q0_0pevilnick, im stuck with a screen that says archos with the version number
11:40:34 Nick Omlet^away is now known as Omlet^away^away (
11:40:35 Nick Omlet^away^away is now known as Omlet (
11:40:38evilnickYes, so this is something you can do *after* fixing that problem
11:40:46q0_0pevilnick, ah
11:40:50q0_0pevilnick, make sense
11:42:38pamauryDoes someone know if rocbox uses/can use host-side headers or if it's self-contained ?
11:50:53 Join bertrik [0] (
11:54:56q0_0pso once im able to load rockbox i can flash it
11:55:11q0_0pi pretty much have to beg the device to load it lol
12:00:05 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
12:00:21q0_0pholy crap it works
12:00:42q0_0pi chose recorder 20 gig
12:00:51q0_0pstrange cause i have jukebox studio 20
12:01:03q0_0pi chose jukebox recorder v2 20 gig
12:02:00n1spamaury: afaik it's supposed to be self contained
12:03:06 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
12:04:11pamauryThat's what I would also say :) This errno problem could be more tricky than it seems. It would require to clean all plugins code to properly use errno. Or to modify errno.h to redirect to rb->errno when PLUGIN is defined.
12:06:08 Join GeekShado_ [0] (
12:08:34q0_0pmy archos cannot use mod extension?
12:09:50***Saving seen data "./dancer.seen"
12:09:57q0_0puses that weird ajz
12:10:10pamaurybertrik: I'm looking for advise about the errno issue in plugins. Do you have time for that ?
12:11:06 Quit stoffel (Read error: 60 (Operation timed out))
12:21:47 Join dfkt_ [0] (i=dfkt@unaffiliated/dfkt)
12:23:01 Quit GeekShadow (Read error: 110 (Connection timed out))
12:24:42bertrikpamaury, sorry I don't have much time for that today
12:26:12n1sgcc 4.4 seems to be doing as one might expect on coldfire, some codecs are faster, some slower, tremor still takes a pretty big hit...
12:27:31 Quit dfkt (Read error: 60 (Operation timed out))
12:28:24 Quit dfkt_ ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
12:36:01q0_0pdo i have a bootloader problem?
12:36:07q0_0psometimes it will load sometimes it will not
12:36:12q0_0pvery random
12:52:56 Join stooo [0] (
12:56:18 Join Sajber^ [0] (
12:56:57 Quit Sajber^ (Read error: 104 (Connection reset by peer))
12:58:04 Join Sajber^ [0] (
12:59:07 Join linuxguy3 [0] (
13:01:08 Quit stooo ("Leaving.")
13:13:14 Quit linuxguy4 (Read error: 110 (Connection timed out))
13:20:06 Quit q0_0p (Read error: 113 (No route to host))
13:31:59 Quit shaggy-h ()
13:35:12 Join shaggy-h [0] (
13:38:31 Join Casainho [0] (n=chatzill@
13:44:43 Quit MethoS- (Remote closed the connection)
13:46:38 Join DerPapst [0] (
13:46:42 Join MethoS- [0] (n=clemens@
13:50:29 Join Sajber^1 [0] (
13:55:53 Quit fyrestorm ("lamers envy me like they envy bill g -- main boot xp, just the way it should be!")
13:55:59 Join einhirn [0] (
13:58:24 Quit MethoS- (Remote closed the connection)
14:00:19 Join MethoS- [0] (n=clemens@
14:02:34 Quit Sajber^ (Read error: 110 (Connection timed out))
14:05:12 Quit Casainho (Read error: 60 (Operation timed out))
14:07:09 Join Casainho [0] (n=chatzill@
14:09:52***Saving seen data "./dancer.seen"
14:13:28 Quit MethoS- (Remote closed the connection)
14:15:42 Join MethoS- [0] (n=clemens@
14:29:02 Quit Casainho (Read error: 60 (Operation timed out))
14:30:09 Join efyx [0] (
14:42:36 Quit MethoS- (Remote closed the connection)
14:44:27 Join MethoS- [0] (n=clemens@
14:47:31 Quit MethoS- (Remote closed the connection)
14:51:05 Join MethoS- [0] (n=clemens@
14:59:50 Join FlynDice [0] (
15:00:39 Quit MethoS- (Remote closed the connection)
15:00:56 Join MethoS- [0] (n=clemens@
15:01:16 Quit einhirn (Read error: 104 (Connection reset by peer))
15:01:55 Join StealthyXIIGer [0] (n=stealthy@
15:05:28 Join mc2739_ [0] (n=mc2739@rockbox/developer/mc2739)
15:06:20 Join mt_ [0] (n=mtee@
15:14:23 Quit advcomp2019 (Read error: 113 (No route to host))
15:16:05 Quit MethoS- (Remote closed the connection)
15:16:56 Quit mc2739 (Read error: 110 (Connection timed out))
15:22:48 Quit mt (Read error: 113 (No route to host))
15:28:08 Join ded [0] (
15:30:50 Nick mc2739_ is now known as mc2739 (n=mc2739@rockbox/developer/mc2739)
15:33:58 Join Grahack [0] (
15:35:38n1shmm aacHE, a52 and mpc are faster, aac and mp3 are pretty neutral and everything else s slower :(
15:36:57n1swhere cook, vorbis and vw are in the 10-20% slower range :/
15:42:16 Quit DerPapst ("Leaving.")
15:45:58 Join fyrestorm [0] (
15:46:19 Join b1uebrother [0] (n=dom@
15:56:39 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
16:02:28 Quit petur ("*plop*")
16:03:30 Quit sagemfreak_ ("Verlassend")
16:09:55***Saving seen data "./dancer.seen"
16:11:03 Quit b1uebrother ("leaving")
16:12:25 Quit ded (Read error: 110 (Connection timed out))
16:13:49Unhelpfuln1s: try playing with -O levels... i saw much larger differences between 4.0 and 4.4 with -O1 than at higher levels.
16:16:42 Join Phyber0ptik [0] (
16:26:01pamaurygevaerts: ping
16:26:36gevaertspamaury: pon
16:28:34pamauryI had a look this morning at the errno issue we discuss yesterday and I got a fix but there are two things to note: first I had to modify errno.h because several plugins include it and I need to rename errno to *rb->errno. second, several plugins use or refer to errno, I have to check that such a would not break them. I checked and it compiles but could you have a look at it ? (see
16:29:55 Quit mc2739 (Read error: 60 (Operation timed out))
16:30:27gevaertswhy volatile?
16:31:29pamaurysame as current_ticks, unintelligent copy paste I guess :)
16:31:47gevaertscurrent_ticks is updated from an interrupt handler I think
16:31:56pamauryyes I know, I'll change that :)
16:34:00pamauryI believe this patch is not sufficient, we need to change or fix a few plugins. For example in apps/plugins/lua/rocklibc.h
16:34:03gevaertsit looks as if only lua and pdbox actually ever use errno
16:35:04pamauryactually the following plugins refers to errno: zxbox, pdbox, png and lua. But that's grep result, it can be dead or commented code
16:35:43gevaertszxbox is all commented out, png does refer to "extern int errno", but the variable is never read or assigned
16:35:57pamaurySo we should delete this reference
16:36:25 Join mc2739 [0] (n=mc2739@rockbox/developer/mc2739)
16:36:40pamaurypdbox seems to use errno for its own purpose.
16:36:54gevaertsthat's ugly :\
16:37:33gevaertsoh well, I don't think it will break anything
16:37:35pamauryBut I can't see a place where it's actually allocated because it can't be linked
16:38:03gevaertsmaybe also commented out?
16:38:43pamauryNot sure, there's a static int rb_errno but I don't a #define errno rb_errno, still looking at the grep output
16:39:36*pamaury is looking more precisely at the code...
16:40:25pamaury#ifndef ROCKBOX
16:40:28pamaury errno = 0;
16:41:13pamaurythat's the same everywhere, errno is not used when in rockbox. There is something to improve there but nothing will break
16:42:00pamaurySo pdbox seems safe.
16:42:59pamauryThat's ugly, lua defines a errno in rockaux.h and have a stub strerror
16:43:20pamauryUgly but seems safe also
16:45:48pamaurygevaerts: do you think it's safe enough so I commit it ?
16:49:26 Quit Zarggg ()
16:54:29gevaertspamaury: I think so
16:55:45 Quit Res1 (Read error: 60 (Operation timed out))
17:05:32 Quit xavieran (Read error: 54 (Connection reset by peer))
17:09:59 Join Zarggg [0] (
17:15:59 Join stooo [0] (
17:17:23 Join kugel [0] (n=kugel@rockbox/developer/kugel)
17:23:26 Quit Sajber^1 (Read error: 110 (Connection timed out))
17:24:59 Quit Phyber0ptik (Read error: 110 (Connection timed out))
17:48:11 Quit jasio (Remote closed the connection)
17:50:37n1sUnhelpful: yeah, i'll try that, the codeclib is built with the core O level (Os for coldfire) and 3 of the biggest loosers are using the mdctlib...
17:57:51kugelthe core's O level is Os for cf?
17:58:45 Quit evilnick (Ping timeout: 180 seconds)
18:02:02 Join evilnick [0] (i=5752157c@rockbox/staff/evilnick)
18:05:15 Join DerPapst [0] (
18:05:37 Quit linuxguy3 (Read error: 110 (Connection timed out))
18:07:22CIA-6New commit by pamaury (r24109): Add support for errno in plugins.
18:09:14 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
18:09:59***Saving seen data "./dancer.seen"
18:11:04kugelpamaury: you could also have plugin.h define errno
18:11:25kugelI think posix defines errno to be a macro anyway (but I'm not sure)
18:12:11kugelmaybe that wouldn't work with plugins including errno.h directly
18:12:11gevaertskugel: all code that uses errno includes errno.h anyway
18:13:32gevaertshm, you should probably have incremented the plugin API version
18:13:47 Quit GeekShado_ (Read error: 60 (Operation timed out))
18:15:34gevaertspamaury: can you bump PLUGIN_API_VERSION?
18:16:35kugelhm, dircache wastes a lot even if disabled
18:16:58*kugel is inspired by FS #9506
18:17:01kugelJdGordon: ping
18:17:12gevaertsgah, sims and errno make things difficult
18:20:02gevaertspamaury: maybe make the entire thing conditional on !defined(SIMULATOR)? In the sim, I suspect that the normal system errno is available to plugins anyway
18:22:33*gevaerts likes the idea of FS #9506
18:23:53kugelplayling with it right now
18:25:53 Quit fyrestorm ("Ur skills' fireproof like a wooden panel -- U got feds talking leet on your IRC channel!")
18:26:03kugelHAVE_DIRCACHE alone adds ~50k binsize (unless bload-o-meter detects ram usage too but I doubt). 40k seems to be buffers which aren't used if dircache is off
18:26:51pamaurygevaerts: perhaps, I see that it don't compile on most target, that's strange
18:27:03gevaertspamaury: I'm testing a fix now
18:27:15gevaertspamaury: it's only on the simulators
18:28:09pamauryDo you have a fix ready?
18:28:21 Quit kugel (Read error: 104 (Connection reset by peer))
18:28:50gevaertsI think so. Just checking some last details
18:29:13pamauryAnd why does it fails with simulator ?
18:29:43gevaertsbecause then errno.h includes the system errno.h, which I suspect already defines errno as a macro
18:29:44pamauryAh I think I understand
18:30:43JdGordonkugel: late pong
18:31:34 Join kugel [0] (n=kugel@rockbox/developer/kugel)
18:31:46gevaertsDoes anyone have a cygwin and/or mingw sim building setup available?
18:32:21kugelyou mean if anyone in here is insane? :p
18:32:24CIA-6New commit by gevaerts (r24110): Fix building simulators (at least on linux. cygwin and mingw might need more work) ...
18:32:34 Quit kadoban (Remote closed the connection)
18:32:56JdGordonkugel: pong
18:33:21kugelsee a few lines up in the logs
18:33:42kugelI'm playing with that patch right now, I think it's a sensible thing to do
18:36:08pamaurygevaerts: and what about errno ? Perhaps there should be a SIMULATOR guard also.
18:36:38gevaertspamaury: look closely. It has one
18:37:57JdGordonkugel: yeah, I'm not so sure
18:38:15JdGordonit will make it harder to eventually be able to realloc all those buffers if we ever get there
18:38:46JdGordonand is 30K a noticable difference to anything?
18:39:04kugelwouldn't it make it easier?
18:39:11pamaurygevaerts: yes you're right
18:39:23kugelthan reallocing a static buffer ?
18:40:17gevaertspamaury: the bit I'm concerned about is that that errno.h guard also looks for cygwin and mingw
18:41:53 Quit kugel (Read error: 60 (Operation timed out))
18:46:35 Quit Grahack ("Tu m'as vu ?")
18:48:10 Join AaronM [0] (
18:48:47 Quit bertrik (Read error: 60 (Operation timed out))
18:49:01pamaurygevaerts: that's a bit strange you're right
18:49:22gevaertsrasher: ping
18:52:52 Join froggyman [0] (
19:02:24 Nick YPSY is now known as Ypsy (
19:17:06 Join Sajber^ [0] (
19:30:56Unhelpfuln1s: i have a LUT-free log2_int that works as nicely as the LUT version now. it does a lot of shift->add, which is not too nice on coldfire, but perhaps the fact that it works without loading anything from memory will be a help. :)
19:32:20n1sUnhelpful: nice, can the shift->add be done with a multiply on cf?
19:34:13 Part stooo
19:34:22Unhelpfuln1s: on the emac maybe, with the fractional mode. the shifts are rightward.
19:36:00Unhelpfulif we're going to go that way, things could be made even more ideal...
19:36:37 Quit amiconn (Nick collision from services.)
19:36:39n1sthat might require reconfiguring the emac though
19:36:40 Join amiconn_ [0] (i=quassel@rockbox/developer/amiconn)
19:37:00 Nick amiconn_ is now known as amiconn (i=quassel@rockbox/developer/amiconn)
19:37:07 Quit pixelma (Nick collision from services.)
19:37:08 Join pixelma_ [0] (i=quassel@rockbox/staff/pixelma)
19:37:23 Nick pixelma_ is now known as pixelma (i=quassel@rockbox/staff/pixelma)
19:37:40Unhelpfuldon't we keep it in signed fractional mode when not in use?
19:39:16n1sdunno, some codecs set the mode when they init and assume it will remain until they exit...
19:41:23n1smac and speex, seem to use integer mode, everything else fractional but different rounding modes
19:42:18Unhelpfuln1s: that should be fine, too... the mode is part of the thread state.
19:43:39n1swouldn't the log function be called by the codecs (in the codec thread) so share the same thread state as the codec (or am i way off as usual :))
19:44:27Unhelpfuln1s: ah, i'm referring to the specific fixed-point log2 used by aac :)
19:44:56n1saha :)
19:53:47Unhelpfulthe solution we have now is likely the best one for floor_log2... it times best on pretty much everything.
19:55:25Unhelpfulhere's the function:
19:56:43Unhelpfulif you use the emac, you may as well get rid of the 1.5, 1.25, 1.125 etc multiplier scheme. these are only approximations of the ideal values of 2^(2^-1), 2^(2^-2), 2^(2^-3), etc.
19:58:08Unhelpfulinstead you'd want to load an approximation of the ideal values and use that as the multiplier. the subtracted values would also changed - they're the base-2 logs of the multipliers.
19:58:46Unhelpfulessentially it's if (1.5*x < 2) {x *=1.5; result -= log2(1.5);} etc
20:00:13Unhelpfulthe final shift/subtract that's not part of a conditional sequence is approximating log(2/val), with val at this point being the remainder from the process done above.
20:00:24n1si can't say i understand the algo
20:02:36Unhelpfuln1s: you start with floor(log2(val)) as an estimate of the log. this is the log of (val*x) where x is some number less than 2... the rest of the algorithm is estimating that x, basically.
20:02:51 Join Strife89 [0] (
20:03:05Unhelpfulif you can multiply val by 1.5 and still have val <2, then do so, and subtract log2(1.5) from the result. then repeat for 1.25, 1.125, etc.
20:04:15Unhelpfulyou might want to replace __builtin_clz with a shift/add sequence sinc that will produce both the estimate and the shifted value directly.
20:10:01***Saving seen data "./dancer.seen"
20:21:27Unhelpfuln1s: here it is without __builtin_clz:
20:23:42Unhelpfulthat might make the algorithm a bit more clear... treating the input as unsigned fixed-point with 2.30 bits, keep multiplying by various values to get the input as close to 2.0 as possible.
20:24:53 Quit Tomis ()
20:38:04 Quit martian67 (Read error: 110 (Connection timed out))
20:52:05 Quit Strife89 ("If you hold a Unix shell to your ear, you can hear the C.")
20:56:18 Join q0_0p [0] (
20:58:10q0_0pdo i need a program to flash my archos studio?
20:58:31q0_0pjust following these instructions
20:58:32q0_0pSelect Plugins → Applications, and run the firmware_flash plugin. It will tell you about your flash and which file it is going to program. After pressing Menu it will check the file. If the file is OK, pressing On will give you a big warning. If we still didn’t manage to scare you off, you need to press Plus to actually program and verify. The programming takes just a few seconds.
21:08:40Unhelpfuln1s: it's very slightly faster on armv4 w/o the benefit of clz, also... maybe just try the second version as-is on CF and commit if it's not slower, then worry about making it better w/ emac later?
21:09:09 Quit gevaerts (Nick collision from services.)
21:09:21 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
21:14:24Unhelpfulbenefit is: maximum error is 2^-14, the lowest representable error, no more use of large 64-value log2 table or bs_generic's table if we remove the unused log2_fix, and slight speed gain on at least two devices so far :)
21:15:24q0_0pUnhelpful, im not sure what i did but i have internalrom.bin files
21:15:42q0_0pUnhelpful, it seems to boot better, at first i was begging the archos to load rockbox
21:16:00q0_0pUnhelpful, i'm not sure what i did but i tried to flash it in the options?
21:16:43Unhelpfuloop, it's missing something... re-pasted with the zero-input check:
21:17:56Unhelpfulq0_0p: i don't know, i don't have an archos and have never seen one in person. you might need to wait a bit until somebody who knows what you're doing is around. the channel is logged, so if somebody gets back who knows they'll probably see that you asked :)
21:18:11q0_0pUnhelpful, cool
21:24:09 Join fyrestorm [0] (
21:33:12 Quit zMastaa (Read error: 60 (Operation timed out))
21:42:06 Nick Ypsy is now known as YPSY (
21:46:21 Join Strife89DS [0] (n=nds@
21:47:06 Join toffe82 [0] (
21:49:31 Quit Strife89DS (Success)
21:53:28 Quit bmbl ("Bye!")
22:00:13q0_0pi have a panic
22:00:18q0_0ppanic ata:-11
22:00:23q0_0pis it dead?
22:06:12 Quit q0_0p ("Leaving")
22:06:16 Join Casainho [0] (n=chatzill@
22:06:40 Join Strife89DS [0] (n=nds@
22:10:02***Saving seen data "./dancer.seen"
22:10:08Unhelpfuln1s: there's only *one* callsite for log2_int without a zero guard around the call... i removed the zero guard in the function and added one there, it's a bit cheaper than having it in the function. amusingly this version is *slooooooow* on x86, which now gets 50% realtime on my system :)
22:19:35 Join q0_0p [0] (
22:19:50q0_0pi'm able to mount my archos hdd
22:19:59q0_0pits able to go into usb mode
22:20:09q0_0pbut when i load it regularly i get panic
22:23:45 Quit grndslm (Read error: 110 (Connection timed out))
22:27:57n1sUnhelpful: yeah, no need to worry about emac now if it speeds it up on cf anyway, i'd say commit :)
22:28:57q0_0pis my archos dead?
22:29:41 Quit DerPapst ("Leaving.")
22:30:35q0_0pall i have is a screen that says rockbox
22:31:53gevaertsq0_0p: right now seems to be a really bad time to get archos experts...
22:32:11gevaertsToday's particular date might be a reason...
22:32:33q0_0pwas just wondering if its dead
22:39:48 Quit Strife89DS (Connection timed out)
22:48:03 Join Res1 [0] (
22:48:42 Quit fyrestorm ("Ur skills' fireproof like a wooden panel -- U got feds talking leet on your IRC channel!")
23:07:14 Join stooo [0] (
23:08:11 Quit stooo (Client Quit)
23:15:05 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
23:27:20 Quit q0_0p ("Leaving")
23:42:55Unhelpfulmaking log2_int(0) undef simplifies some other things as well
23:45:09Unhelpfulthe algorithm works without keeping a clear high bit that is equated to 2.0, for example... and that means that it can be coded so that it doesn't need to shift left if the high bit is set.
23:50:56 Join zMastaa [0] (
23:55:32 Join Kitr88 [0] (n=Kitarist@
23:56:37 Quit Casainho (Read error: 60 (Operation timed out))

Previous day | Next day