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).

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2009-10-21

00:01:31 Part janne_oksanen
00:12:57saratogano
00:15:03 Quit jgarvey ("Leaving")
00:15:13 Quit domonoky (Read error: 104 (Connection reset by peer))
00:16:17*amiconn slaps kugel with red builds
00:16:41kugelpatience...
00:17:15CIA-85New commit by 03kugel (r23303): Fix red caused by ancient incorrect #ifdefs.
00:18:05amiconnkugel: Btw, if you pass extra parameters to a function, no harm is done (other than the waste for preparing those parameters)
00:18:23amiconnIt's still bad style of course
00:19:57kugelI think so too
00:20:47kugelI was curious whether it would cause problems, but I think fixing bad style is good enough
00:21:04amiconnIf they're regparms it is obvious. The function just doesn't care about the register value
00:22:12amiconnExtraneous stackparms cause a bit of extra stack consumption, but don't cause badness either, because the caller is responsible for stack cleanup - and the caller "knows" it put those extra parms there
00:22:37kugelif they were stack params it could cause an overflow couldn't it?
00:22:40 Join jsj [0] (n=jsj@h-88-157.A163.priv.bahnhof.se)
00:23:00kugeloh, the caller needs to clean it up as well?
00:25:40amiconnOn our target archs, yes
00:26:36amiconnx86 can apparently be either. 'stdcall' means the callee does the cleanup, 'cdecl' means the caller does the cleanup.
00:27:22 Join liar [0] (n=liar@83.175.83.185)
00:27:52kugelso it the callee pops the param, then pushes it back before returning? that seems a bit wasteful to me
00:28:14kugelor does it just adjust the sp?
00:28:25amiconnThe callee 'pops' nothing if the calling convention is that the caller does the cleanup
00:29:20amiconnThe callee reads the parameters from the stack without changing the sp. Caller just adjusts sp after the function returns
00:29:29kugelah yea, I see
00:29:42 Join froggyman [0] (n=sopgenor@pool-72-69-220-194.chi01.dsl-w.verizon.net)
00:31:38 Quit LambdaCalculus37 ("Fwump")
00:33:06 Join fdinel [0] (n=Miranda@modemcable204.232-203-24.mc.videotron.ca)
00:33:13 Quit kugel (Remote closed the connection)
00:37:35 Join shai [0] (n=Shai@l192-117-110-233.cable.actcom.net.il)
00:38:01 Quit shai (Read error: 104 (Connection reset by peer))
00:40:03 Join Strife1989 [0] (n=michael@adsl-154-3-172.mcn.bellsouth.net)
00:41:11 Quit shotofadds ("Leaving")
00:47:03 Quit ender` (" PCMCIA: People Can't Memorize Computer Industry Acronyms")
00:47:59 Quit gevaerts (Nick collision from services.)
00:48:11 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
00:48:37 Part froggyman
00:49:37***Saving seen data "./dancer.seen"
00:50:15 Join Thundercloud [0] (i=thunderc@persistence.flat.devzero.co.uk)
00:50:39 Part JimOz
00:54:17 Quit Strife89 (Read error: 113 (No route to host))
01:00
01:02:43 Quit fdinel (Read error: 110 (Connection timed out))
01:12:22CIA-85New commit by 03kugel (r23304): Fix a few potential redraw problems with the custom statusbar and wps fighting for the same full redraw variable. ...
01:13:34 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
01:15:18CIA-85New commit by 03kugel (r23305): No need to have this variable global anymore.
01:17:22 Quit ercani ("Leaving")
01:28:08 Quit Thundercloud (Remote closed the connection)
01:29:15CIA-85New commit by 03kugel (r23306): Remove all update event callbacks if the custom statusbar is unused.
01:37:03 Quit MethoS- (Remote closed the connection)
01:42:38 Join phanboy4 [0] (n=benji@c-24-98-43-198.hsd1.ga.comcast.net)
01:49:27 Quit efyx_ (Client Quit)
01:51:32 Join andrewRB [0] (n=andrewth@81.170.115.71)
01:51:37andrewRBhey
01:51:44andrewRBany comments on FS #10697?
01:56:14kkurbjunandrewRB: looks like a nice feature - I don't know the list code or whether it's appropriate there, offhand it looks like the right way to do it
02:00
02:00:42JdGordon|going by the title it sounds like it could be good..
02:00:50JdGordon|going to have a look at it on the bus (in about 15min)
02:00:55andrewRBta
02:01:27JdGordon|custom statusbar + touchscreen + playback controls in the lists could make for some fun :p
02:01:31 Quit JdGordon| ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
02:02:13andrewRB^^ -> Cowon D2 port =P
02:07:11 Join fdinel [0] (n=Miranda@modemcable204.232-203-24.mc.videotron.ca)
02:07:23 Join peter__b [0] (n=peter_b@93.133.209.150)
02:07:47 Quit Stephen_ (Read error: 110 (Connection timed out))
02:15:13 Quit peter-b (Read error: 145 (Connection timed out))
02:20:25 Join JdGordon| [0] (n=Miranda@m6b0e36d0.tmodns.net)
02:21:48JdGordon|andrewRB: ah, you didnt add any logic to actually make it work :(
02:22:49andrewRBhmm? modify the keymap of whichever device you are talking about?
02:23:37andrewRBfor the D2, i'm actually going to submit a patch to do that very soon, but I know that is something which is disliked by many people here, so I seperated the functionality
02:24:43JdGordon|yeah, that functionality is the interesting bit (for me anyway :p )
02:25:34andrewRBsame
02:25:48andrewRBbut check #10683
02:26:11andrewRB(that isn't my patch, i don't endorse it, but the arguments in the comments still apply)
02:26:22 Quit Zarggg (Read error: 104 (Connection reset by peer))
02:26:23 Join Zarggg_ [0] (n=zarggg@65-78-69-194.c3-0.eas-ubr6.atw-eas.pa.cable.rcn.com)
02:26:55JdGordon|it doesnt always make sense to have all those actions in the same #define
02:27:24andrewRBsorry, which?
02:27:47JdGordon|your atch
02:28:21JdGordon|for 10683.. I'm against hold changing a button behaviour, but am completly or customisable buttons
02:28:28andrewRBhmmm, true
02:28:47andrewRBI figured it would be overkill to seperate them right now when there are no players to which it applies
02:29:38JdGordon|it would be... if you want "inspirtion" look at recordser/radio.c... pretty much each action is surrounded by a #if
02:29:59JdGordon|while in general i tihnk thats horrible.. it would be useful here
02:30:24andrewRBhuh
02:30:25andrewRByeah
02:30:56JdGordon|this keyboard+bumpy road makes baby jesus not able to read :p
02:31:39*andrewRB wonders if JdGordon| has an identity crises or an overinflated ego
02:32:05JdGordon|cant I have both?
02:32:59*andrewRB adds "greed" to JdGordon|'s list of personality disorders
02:38:17 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
02:38:47 Quit w1ll14m|_ (Read error: 110 (Connection timed out))
02:40:09 Quit JdGordon| ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
02:44:34 Join w1ll14m| [0] (n=nobody@84-104-81-17.cable.quicknet.nl)
02:45:51 Quit Zarggg_ (Read error: 54 (Connection reset by peer))
02:46:32Unhelpfulthat's more a *flaw* than a disorder imo
02:46:34 Join Zarggg [0] (n=zarggg@65-78-69-194.c3-0.eas-ubr6.atw-eas.pa.cable.rcn.com)
02:48:59*andrewRB thinks ^^ contributes to the 25% obesity in the USA
02:49:38 Quit Rondom (Nick collision from services.)
02:49:42***Saving seen data "./dancer.seen"
02:49:48 Join Rondom [0] (n=Rondom@dslb-084-057-189-187.pools.arcor-ip.net)
02:58:51 Quit fdinel ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
03:00
03:00:57 Quit toffe82 (Read error: 110 (Connection timed out))
03:03:49 Quit pixelma (Nick collision from services.)
03:03:51 Join pixelma_ [0] (i=quassel@rockbox/staff/pixelma)
03:04:11 Nick pixelma_ is now known as pixelma (i=quassel@rockbox/staff/pixelma)
03:04:17 Quit amiconn (Nick collision from services.)
03:04:20 Join amiconn_ [0] (i=quassel@rockbox/developer/amiconn)
03:04:39 Nick amiconn_ is now known as amiconn (i=quassel@rockbox/developer/amiconn)
03:14:19andrewRBugh
03:14:35andrewRBdoes anyone actually use the radio on the D2 in rockbox?
03:15:55 Quit thegeek (Read error: 104 (Connection reset by peer))
03:16:01andrewRBnvm
03:24:28 Join boxy [0] (n=556408bd@giant.haxx.se)
03:25:14 Join webguest13 [0] (n=556408bd@giant.haxx.se)
03:25:25webguest13hi people
03:27:29 Join webguest34 [0] (n=556408bd@giant.haxx.se)
03:27:30 Quit webguest13 (Client Quit)
03:27:50 Join webguest07 [0] (n=556408bd@giant.haxx.se)
03:27:58 Quit webguest34 (Client Quit)
03:28:22 Quit Topy44 (Read error: 104 (Connection reset by peer))
03:28:35 Quit webguest07 (Client Quit)
03:29:08 Quit Kopfgeldjaeger ("Serverwechsel")
03:29:30 Quit boxy ("CGI:IRC (Ping timeout)")
03:30:29 Quit bubsy ("Party time, EXCELLENT! *wring wring wring!!* Party on, Wayne! Party on, Garth!")
03:30:49 Join toffe82 [0] (n=chatzill@adsl-70-137-198-87.dsl.frs2ca.sbcglobal.net)
03:31:08 Join Topy44 [0] (n=Topy44@g227169068.adsl.alicedsl.de)
03:42:48 Join midgey [0] (n=tjross@rockbox/developer/midgey)
03:43:01 Nick hueafk is now known as hue (n=hue@c-98-247-116-47.hsd1.wa.comcast.net)
03:56:50 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/freenode/x-merhzysfppzvacam)
03:57:39 Quit Strife1989 ("Bed.")
03:58:07cool_walking_What's the policy on swearing in the forums? There doesn't appear to be anything about it in the guidelines, and there's quite a bit of it spread around on there, but someone reported a post.
04:00
04:04:00 Join bubsy [0] (n=bubsy@94.139.72.137)
04:12:52midgeyEh, so I don't have a problem with swearing. I think as devs we should obviously watch our language in mixed company, but for users I feel that as long as it isn't demeaning towards another user, we should let it go.
04:13:10cool_walking_I think I'll just PM the person who reported it saying we're still deciding our policy. Not enough people on right now to reach a consensus, I think. I agree with midgey, anyway.
04:16:20cool_walking_okay, sent
04:17:53 Quit TheSeven (Nick collision from services.)
04:18:14 Join The_Seven [0] (n=theseven@rockbox/developer/TheSeven)
04:18:26 Nick The_Seven is now known as TheSeven (n=theseven@rockbox/developer/TheSeven)
04:26:24 Quit Hillshum_ (Read error: 145 (Connection timed out))
04:27:03 Quit bubsy ("Party time, EXCELLENT! *wring wring wring!!* Party on, Wayne! Party on, Garth!")
04:32:51 Join froggyman [0] (n=sopgenor@pool-72-69-220-194.chi01.dsl-w.verizon.net)
04:33:15 Quit JdGordon ("Leaving.")
04:35:29 Quit Hillshum (Read error: 110 (Connection timed out))
04:49:46***Saving seen data "./dancer.seen"
04:51:47 Join dys` [0] (n=andreas@krlh-5f7070c7.pool.mediaWays.net)
04:59:49 Part froggyman
05:00
05:00:58 Join AndyI [0] (n=pasha_in@212.14.205.32)
05:03:32 Quit dys (Connection timed out)
05:12:05 Quit AndyIL (Read error: 110 (Connection timed out))
05:17:31 Join bubsy [0] (n=bubsy@94.139.72.137)
05:20:23 Quit StealthyXIIGer (Read error: 104 (Connection reset by peer))
05:20:42 Join StealthyXIIGer [0] (n=stealthy@c-68-62-19-6.hsd1.mi.comcast.net)
05:20:55 Nick linuxguy4 is now known as linuxguy3 (n=timj@adsl-75-58-36-229.dsl.emhril.sbcglobal.net)
05:24:15 Join lifeless_ [0] (n=lifeless@94.50.188.8)
05:28:23 Quit bubsy ("Party time, EXCELLENT! *wring wring wring!!* Party on, Wayne! Party on, Garth!")
05:34:49 Quit panni_ ("( www.nnscript.de :: NoNameScript 3.81 :: www.XLhost.de )")
05:39:04 Join JdGordon [0] (n=Miranda@c-24-22-210-83.hsd1.wa.comcast.net)
05:41:18 Quit Horscht ("Verlassend")
05:45:43 Quit StealthyXIIGer (Read error: 110 (Connection timed out))
05:45:58 Join StealthyXIIGer [0] (n=stealthy@c-68-62-19-6.hsd1.mi.comcast.net)
06:00
06:00:03 Join Crackerizerizer [0] (n=7135aa15@giant.haxx.se)
06:06:16 Quit cool_walking_ ("Page closed")
06:15:56andrewRBJdGordon: still there?
06:16:05andrewRByou may want to check out FS #10698
06:23:29JdGordonwhat the heck is the BUTTON_HOLD stuff?
06:23:55JdGordonit should be BUTTON_MINUS|BUTTON_REPEAT which is done for you magically
06:24:28JdGordonoh... this is the hold button switch?
06:24:32andrewRBnot following you there, sorry?
06:25:02JdGordonwe dont want buttons to work when the hold is enabled...
06:25:16andrewRBnot at all, ever?
06:25:30JdGordonyep
06:25:43JdGordonthe whole point of that switch is so things dont happen accidently
06:26:25andrewRBthe tactile buttons really can't be pressed accidentally, and without this functionality you have to remove the player from your pocket to change track or volume =/
06:26:49JdGordonyes
06:27:36andrewRBoh well. so what is it you are looking for? track skipping by holding +/- buttons when not in hold mode?
06:27:59JdGordontouchscreen controls
06:28:20andrewRBdespite the fact that you have to *see* it to use it?
06:29:30JdGordonyep
06:29:45andrewRBokies. well i guess that patch may as well be closed ;)
06:37:40 Quit andrewRB ("bye peeps")
06:39:12 Quit JdGordon (Read error: 104 (Connection reset by peer))
06:42:56 Join matsl [0] (n=matsl@dhcp126.contactor.se)
06:49:49***Saving seen data "./dancer.seen"
06:59:45 Quit StealthyXIIGer (Read error: 110 (Connection timed out))
07:00
07:00:21 Join thegeek [0] (n=nnscript@129.241.123.168)
07:01:59 Quit Lss (Read error: 110 (Connection timed out))
07:13:05 Quit Crackerizerizer ("CGI:IRC (Ping timeout)")
07:14:09 Quit liar (Client Quit)
07:18:50pixelmaonly four old daily builds are available now?
07:18:53 Part toffe82
07:20:08pixelmaand only three of the corresponding voices
07:21:56 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
07:24:41 Join dionoea_ [0] (n=dionoea@yop.chewa.net)
07:27:39 Quit dionoea (Read error: 104 (Connection reset by peer))
07:33:06 Join webguest74 [0] (n=47bd5122@giant.haxx.se)
07:34:23 Quit KBH (Read error: 110 (Connection timed out))
07:39:08 Quit webguest74 ("CGI:IRC (Ping timeout)")
07:48:07 Quit antil33t1 (Read error: 104 (Connection reset by peer))
07:49:49 Join uflops [0] (n=yogurt@90-231-195-226-no112.tbcn.telia.com)
07:55:16 Join antil33t [0] (n=Mudkips@119.224.12.185)
08:00
08:01:57 Quit mikroflops (Read error: 110 (Connection timed out))
08:05:07 Join esperegu [0] (n=quassel@145.116.15.244)
08:13:36 Quit antil33t (Read error: 104 (Connection reset by peer))
08:16:37 Quit midgey ()
08:17:35 Join JdGordon| [0] (n=Miranda@c-24-22-210-83.hsd1.wa.comcast.net)
08:20:40 Join antil33t [0] (n=Mudkips@119.224.12.185)
08:21:47 Join Rob2222 [0] (n=Miranda@p4FDCD504.dip.t-dialin.net)
08:25:02 Join ender` [0] (i=krneki@foo.eternallybored.org)
08:26:18 Join Utchybann [0] (n=lolo@ede67-1-81-56-102-26.fbx.proxad.net)
08:29:14 Quit Rob2223 (Read error: 60 (Operation timed out))
08:31:16 Join HBK [0] (n=hbk@rrcs-97-77-51-170.sw.biz.rr.com)
08:35:25 Join Thundercloud [0] (i=thunderc@persistence.flat.devzero.co.uk)
08:39:02 Join MethoS- [0] (n=clemens@134.102.106.250)
08:40:25 Join KBH [0] (n=hbk@rrcs-97-77-51-170.sw.biz.rr.com)
08:40:53 Quit JdGordon| ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
08:41:20 Nick dionoea_ is now known as dionoea (n=dionoea@yop.chewa.net)
08:45:43 Quit pixelma (".")
08:46:11 Join JdGordon| [0] (n=Miranda@c-24-22-210-83.hsd1.wa.comcast.net)
08:46:46 Join dys`` [0] (n=andreas@krlh-5f707906.pool.mediaWays.net)
08:48:03 Join pixelma [0] (i=quassel@rockbox/staff/pixelma)
08:49:53***Saving seen data "./dancer.seen"
08:51:47 Join dys``` [0] (n=andreas@krlh-5f72ef57.pool.mediaWays.net)
08:52:17 Quit MethoS- (Remote closed the connection)
08:52:48 Join Zagor [242] (n=bjorn@rockbox/developer/Zagor)
08:53:02 Join flydutch [0] (n=flydutch@host132-153-dynamic.15-87-r.retail.telecomitalia.it)
08:55:37 Quit HBK (Read error: 110 (Connection timed out))
08:56:46 Join dys```` [0] (n=andreas@krlh-5f7342a2.pool.mediaWays.net)
08:56:46***Alert Mode level 1
08:56:46DBUGEnqueued KICK dys`
08:56:46DBUGEnqueued KICK dys``
08:56:46***Alert Mode level 2
08:56:46DBUGEnqueued KICK dys```
08:56:46DBUGEnqueued KICK dys````
08:56:46***Alert Mode level 3
09:00
09:00:50 Quit dys` (Connection timed out)
09:06:04 Join petur [50] (n=petur@rockbox/developer/petur)
09:06:19 Quit dys`` (Connection timed out)
09:06:47***Alert Mode OFF
09:07:41 Quit dys``` (Connection timed out)
09:07:46 Join Rand_Althor [0] (n=chatzill@adsl-76-235-53-60.dsl.dytnoh.sbcglobal.net)
09:07:56Rand_AlthorDoes RB have any problems with exFAT?
09:14:12 Quit FOAD (Read error: 145 (Connection timed out))
09:15:24 Quit HellDragon (Client Quit)
09:18:05 Join maruk [0] (n=papier@titanium.sdv.fr)
09:18:43 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
09:20:21 Part LinusN
09:20:23 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
09:23:45 Quit parafin (Read error: 60 (Operation timed out))
09:23:53 Quit Rand_Althor ("ChatZilla 0.9.85 [Firefox 3.5.3/20090824101458]")
09:26:18 Join Grahack [0] (n=chri@82.216.222.222)
09:34:16 Quit maruk ("Leaving.")
09:36:05 Join maruk [0] (n=papier@titanium.sdv.fr)
09:42:13 Quit phanboy4 (Read error: 104 (Connection reset by peer))
09:42:33 Join phanboy4 [0] (n=benji@c-24-98-43-198.hsd1.ga.comcast.net)
09:43:13CtcpIgnored 4 channel CTCP requests in 0 seconds at the last flood
09:43:13*pixelma points petur to the most recent post in the recordings subforum
09:43:30*petur goes looky
09:43:32pixelmaor thread
09:45:33JdGordonis support problems still the biggest argument against any customisable menus/buttons?
09:46:13peturpixelma: I really have no time this week. Looks like I should take my laptop to the US to be able to work on it :/
09:46:26 Quit Thundercloud (Remote closed the connection)
09:47:12pixelmaok, maybe if I have time sometime this week or during the weekend - revert shouldn't be too hard and keymap things aren't either
09:48:24peturI would first only revert recording.c and see what that does on the ipoos
09:50:16 Join DerPapst [0] (n=DerPapst@wlan-nat-24.fh-friedberg.de)
09:51:13 Quit faemir ("Leaving")
10:00
10:01:46 Quit JdGordon| (Read error: 110 (Connection timed out))
10:05:56 Join kugel [0] (i=kugel@rockbox/developer/kugel)
10:06:00 Join shotofadds [0] (n=rob@rockbox/developer/shotofadds)
10:06:59shotofaddsJdGordon: please don't decide that something is "never" going to happen without any discussion (hold buttons on D2)
10:07:07shotofaddshow can you decide that for yourself?
10:07:14 Quit BHSPitMonkey (Read error: 60 (Operation timed out))
10:07:25JdGordonshotofadds: its not really device specific....
10:07:41JdGordonplenty f targets have software hold and this has been requested before
10:07:54 Join DerPapst1 [0] (n=DerPapst@wlan-nat-24.fh-friedberg.de)
10:08:09shotofaddsI didn't know it had come up before, sorry. Can you point me to a discussion of it?
10:08:37JdGordonnot really.. its either here or the ml's...
10:09:04shotofaddsThe reason it's come up for the D2 is that it cannot be used "in pocket" without controls such as these. The OF has this feature, and Rockbox is significantly less functional because of it.
10:09:07JdGordonI'm not completely against it... but others sure will
10:10:11JdGordonit does sound completyl wrong to me though... the hold switch is supposed to stop you being able to press buttons... not change their behaviour
10:10:22shotofaddsJdGordon: sure. I think the approach in fs#10683 in ideal - things won't magically happen when hold is enabled unless the user specifically enables it.
10:11:02shotofaddsI think the fact we now have 3 separate patches for this on Flyspray already speaks volumes
10:11:11kugelshotofadds: I'm thinking it's out of question that the hold button enables other actions. It's may possibly be a question if it's a user setting
10:11:39kugelbut we have discussed configurable buttons a dozen times already, always rejecting the idea (generaly...)
10:12:15shotofaddsit's not configurable buttons
10:12:35JdGordonsure it is
10:12:53kugelwe also had a number of patches that do something magical with the REC button on the e200, all are rejected
10:13:15shotofaddsok fine. I'll go with "Rockbox, significantly less functional than the OF" as a strapline :/
10:13:28JdGordonthere you go! everyone is happy :)
10:13:37shotofaddstime I went to work...
10:13:39 Join efyx_ [0] (n=efyx@lap34-1-82-225-185-146.fbx.proxad.net)
10:13:40 Quit shotofadds ("Leaving")
10:13:48kugelthe volume it speaks is simply that touchscreens have many disadvantages in usability, not the fact that people are keen on doing whatever with the hold button
10:14:05JdGordonis support problems still the biggest argument against any customisable menus/buttons?
10:16:32kugelI have no problem with these patches in general. but if it's accepted, we need to renew our decision for *all* targets
10:17:05 Quit Grahack (Client Quit)
10:17:25JdGordonyep
10:17:59JdGordonI would like to see some of these in also... I wouldnt even mind if they were disabled by default in #ifdef's which means anyone with that build wont get support anyway
10:20:34kugelJdGordon: any plans to get back to wfms?
10:20:55JdGordonthe radio code scares me...
10:23:33hueanyone know if the 2nd gen nano can scan ff and backwards on a track?
10:26:00 Quit DerPapst (Read error: 110 (Connection timed out))
10:27:46hueah you gotta hold forward or back down
10:27:47hue-_-
10:30:22 Quit Zagor ("Don't panic")
10:32:59*TheSeven just plumbed another pool block leak
10:33:06 Join Zagor [242] (n=bjorn@rockbox/developer/Zagor)
10:33:45 Quit DerPapst1 (Read error: 110 (Connection timed out))
10:37:37 Join einhirn [0] (n=Miranda@bsod.rz.tu-clausthal.de)
10:38:08topikeach time i'm about to update my nano you fix some new things :)
10:39:38TheSevenyes, i spotted another critical FTL issue tonight
10:40:55topikthe one from r23296 or later?
10:41:09 Quit shaggy-h (Read error: 104 (Connection reset by peer))
10:41:25TheSeventopik: the latest one still isn't fixed in svn, i'm right now testing it
10:41:40TheSevenbut it looks like stability has once again improved a lot
10:41:45topikthen i'll hold on updating just yet
10:41:53*TheSeven will now try copying c:\windows\system32 to his ipod as a stress test :-)
10:42:32topiki couldn't do that, my nano is only 2GB
10:42:59topikthis FTL issues, are they mostly in writing routines or is reading affected too?
10:43:15TheSevenmy system32 folder is ~1GB, however this will still need an hour (if it doesn't fail before that) at the current ftl performance
10:44:02TheSeventopik: reading seems to be perfectly stable, and if something fails while reading, that isn't dangerous either, the worst thing is that it could crash and you need to reboot it
10:44:31TheSevenhowever, i managed to get my FTL in unrecoverable (without complete wiping) states for about 20 times yesterday...
10:45:00topikyeah, that's my experience too. filling my nano in diskmode and playing from it works pretty solid
10:45:02hueto delete the itunes version of my library, just show hidden files and delete the music directory?
10:45:20topikdelete the ipod_contol dir
10:45:20TheSevenno, that's where your music is
10:45:29TheSevenipod_control\itunes is the one to delete
10:45:40TheSeventopik: the music is in ipod_control, too!
10:45:44hueyea
10:46:06hueso the whole ipod_control folder, oh ok\
10:46:09topikwhen i cut my ties to itunes, i wiped all itunes related stuff and put the music back on the proper way
10:46:18hueim gonna do the same thing
10:46:31*TheSeven loves to see people ditch the apple firmware altogether
10:46:34topikexcept of course when i fucked up my nano and needed itunes to restore it
10:46:53topiki would have done that, but i kept getting an iloader startup screen :)
10:47:17huei havent been following the latest updates.. havent had any issues so far. expect a legit update this weekend?
10:47:33topikupdates come as they come
10:48:02hueoic, but you guys are fixing stuff left and right
10:48:10topikTheSeven is
10:48:17huejust download and overwrite?
10:48:21topikyep
10:48:23hueeasy
10:48:40 Join kyle6513 [0] (n=kyle6513@58.174.128.189)
10:49:08 Quit kyle6513 (Client Quit)
10:49:55***Saving seen data "./dancer.seen"
10:50:37TheSevenmy transfer of system32 just failed again with a weird "file name is too long" message
10:50:59TheSevenbut after reconnecting the ipod (without rebooting it) it seems to be fine again, so this time it was an usb problem?
10:51:10gevaertsTheSeven: what was the filename?
10:51:33topik"too long"
10:51:40TheSevenc:\windows\system32\comsnap.dll => e:\system32\comsnap.dll
10:52:15TheSeven"file name is too long" seems to be the default message of windows if something weird is going on, like apple is showing a "dead battery" logo on the ipod's screen if the ftl is broken
10:52:18gevaertsok, that does indeed sounds like a bug somewhere then
10:53:05TheSeventhe sad thing is that this happened after copying only 46MB
10:53:22TheSeven(in 291 files)
10:55:58TheSevenhm, another 432 files (159 MB) copied without issues now. weird.
10:56:40*TheSeven curses the "safe, slow, stupid" algorithm he's using in the FTL and NAND driver right now, which is slowing things down massively
10:57:02TheSevenI really need to work on parallelizing things once it's working fine
10:58:36TheSeven839 files, 243 MB, still running fine...
11:00
11:02:54TheSevenlocked up again :-/
11:03:25*TheSeven thinks that this may be related to moving his ipod around a little
11:03:42TheSevengevaerts: how stable are the transfer error recovery procedures in the usb core and storage?
11:04:20gevaertsTheSeven: a bus reset should clear everything
11:04:38TheSeven...in theory, yes. so the host is responsible for recovering from that
11:04:47gevaertsyes
11:06:36TheSevenbut the fact that the hid died at the same time and works again after reconnecting clearly points towards USB, not the FTL
11:06:36gevaertsTheSeven: something that I've done in the past is comment out the storage_write calls in usb_storage.c, and use dd to write large amounts of data
11:06:36gevaertshow long did you wait to see if it recovers?
11:06:49TheSevenuntil windows failed the transfer and showed the drive as empty, after the transfer was stuck for like a minute
11:07:31gevaertssounds likely to be USB, yes
11:08:54TheSevenbah. looks like one of the failing transfers damaged the FS a little
11:09:32TheSevenwindows complains that the file or directory could not be *read* while trying to continue copying
11:09:32TheSevenlet's try a chkdsk /f
11:10:48TheSevenok, that ran fine and it caught damaged directory entries in the system32 folder, so it probably was a consequence of the failed connection, not FTL issues
11:11:12TheSevenand it tells me that the .rockbox\eqs folder is damaged in a non-recoverable way
11:12:01TheSeven...and themes and codepages and wps\boxes and wps\cabbiev2 and lots of interconnected clusterchains
11:12:13TheSevenlooks like the fat suffered quite a lot as well
11:13:12TheSevenfrom chkdsk's behavior i'd say a block of the FAT was just zeroed out, which may point to the FTL again
11:17:28 Quit kugel (Read error: 54 (Connection reset by peer))
11:17:28 Nick hue is now known as hueafk (n=hue@c-98-247-116-47.hsd1.wa.comcast.net)
11:21:58 Join tomers [0] (n=86bfe845@giant.haxx.se)
11:22:14tomersIs there any binary build of tools for win32 anywhere?
11:22:18 Join swilde [0] (n=wilde@212.95.126.10)
11:23:44linuxstbtomers: No
11:24:29pixelmatomers: the MacOS 10.4 USB problem is not fixed
11:25:11tomerspixelma: So we should close this issue as WON'T FIX or even INVALID, as it is a non-RB bug
11:25:29gevaertstomers: I wouldn't do that until we have a setting to disable HID
11:26:22tomerslinuxstb: Don't you think it should be built somewhere (build system)? I want to install Cowon, not at home, so I don't have a build system, and I am on Windows... What should I do (it is nat part of rbutil yet)
11:26:23tomers?
11:26:58tomersgevaerts: So I guess I have to implement such feature, right? (tried to avoid that, hoping the OSX issue will be fixed)
11:27:00pixelmaI agree with gevaerts, otherwise people won't be able to access their devices on that OS at all without being able to compile their own build without HID
11:27:34tomersgevaerts: I could implement that in the next couple of days
11:27:37linuxstbtomers: I don't understand your question... Are you saying that the required binaries to install the Cowon D2 build are not available on the download server?
11:28:17tomerslinuxstb: I think they're not there. I need tools/mktccboot compiled to Windows. That was my original question
11:28:37*TheSeven is experiencing weird problems like winrar being unable to extract rockbox-full.zip to the ipod directly but working if you just open the zip file and drag'n'dropping the files
11:32:33linuxstbtomers: No, your original question was about "tools" in general. But yes, if that's required to install on the Cowon D2, it should be on the download server (or the D2 should be relegated back down to Unusable)
11:34:34tomerslinuxstb: Are you going to fix it? Will it be part of the current build?
11:34:47 Join shaggy-h [0] (n=kiwi@host-87-74-127-193.dslgb.com)
11:34:48 Join Grahack [0] (n=chri@82.216.222.222)
11:34:51linuxstbThe easy fix is to relegate the D2 back down to unusable...
11:35:08linuxstb(as it isn't meeting the critera for unstable)
11:38:20linuxstbtomers: Bootloaders (and the tools needed to install them) are not built automatically - someone needs to make a release and ask Zagor/Bagder to put the files on download.rockbox.org
11:43:11tomerslinuxstb: Don't you think these should be build like current build is built? (daily / every commit)?
11:43:38TheSevenuh oh.... it looks like data can be read back fine as long as it's in scattered page vbs, but sometimes things are getting lost while trying to commit them?
11:44:36tomerslinuxstb: How much work is expected from getting rbutil support the D2? I might do that... But I guess we need to have bootloader and tools be on download.rockbox.org
11:46:06 Nick uflops is now known as mikroflops (n=yogurt@90-231-195-226-no112.tbcn.telia.com)
11:46:18 Quit phanboy4 ("Leaving")
11:47:37linuxstbtomers: No - on many targets (D2 included I think), a bad bootloader can brick the device. So we just want known-good releases.
11:47:38swildeHi *, I'm having an strange problem with my Sansa C240:
11:47:38swildeWhen connection to computer A (GNU/Linux Debian Lenny) I get two USB MSC devices: sda for the internal flash and sdb for the micro sd card inserted in the dap. On computer B (GNU/Linux Debian Lenny, too) I only get one USB MSC device: sda for the internal flash. The only known differences are: system A has USB2.0 system B only 1.x and the kernels differ (one 2.6.29.x one 2.6.30.x). Any Ideas what might be the cause of the problem?
11:50:14tomerslinuxstb: Is the D2 bootloader not safe at the moment? (don't want to brick mine..)
11:52:04gevaertsthe D2 is pretty safe IIRC. You can boot using tcctool to recover
11:52:14TheSevenIIRC gevaerts said that D2s aren't very easy to brick
11:52:19TheSevenah yes.
11:52:37polobricoloTheSeven: with the usb enabled i can'
11:52:41gevaertsswilde: B doesn't have the CONFIG_SCSI_MULTI_LUN option in the kernel
11:52:46polobricolo*can't load some plugins
11:52:58polobricoloundefined instruction
11:53:23polobricoloeg with mpegplayer
11:53:37tomersgevaerts: If it's safe, what is preventing it from being in rbutil?
11:54:08swildegevaerts: Oh, that sounds like a logical explanation! Thanks a lot.
11:54:22gevaertstomers: someone needs to do the work? I don't know...
11:55:29tomersgevaerts: I was suggesting trying to work on it (provided the BL and tools are saef enough). Asked linuxstb what is the scope of such thing (time) and he said it's not that safe... I'm confuzed here.
11:56:08tomerss/confuzed/confused
11:57:00amiconnTheSeven: Does test_disk.rock "Write & Verify" work on Nano2g without reporting errors?
11:57:49TheSevenamiconn: the problem is that the storage layer has a cache-like thing of ~20MB which makes these things hard to test
11:58:10amiconntest_disk.rock writes and then reads a total of 300MB (except on Ondio)
11:59:02amiconnIt uses the prng to generate this data, so comparing it needs no stored value, just the same seed
11:59:05*TheSeven will first need to recover his file system before trying that
11:59:17gevaertstomers: linuxstb said that using an autobuilt bootloader is not a good idea
11:59:21amiconnIt also uses (pseudo-) random blocksizes
12:00
12:00:20polobricolowhat does "incompatible version" mean ? i get that when loading a file (txt or mpg)
12:00:32tomersgevaerts: OK, I get it. So that's why we're tagging BLs, right?
12:01:17gevaertsno, that's why we're *releasing* bootloaders. Tagging is to be able to find the release source easily :)
12:02:09tomerstomers: same same for me :-) It's like saying - this one is good at the moment. If anyone screws things later, no harm done
12:02:30tomerss/tomers/gevaerts/ :-)
12:05:42*polobricolo just realized he updated his rockbox bin but not his dir
12:07:12*linuxstb wonders what the phrase "a compiled bootloader" means on the TargetClassifications page
12:10:52TheSevenhow do i compile that test_disk.rock?
12:11:06 Join parafin [0] (i=parafin@paraf.in)
12:11:57linuxstbTheSeven: Add test_disk.c to apps/plugins/SOURCECS
12:16:09 Join kugel [0] (i=kugel@rockbox/developer/kugel)
12:27:52TheSevenam i correct that this test_disk thing only does sequential writes?
12:29:39 Quit JackWinter (Remote closed the connection)
12:30:31 Join JackWinter [0] (n=jack@vodsl-10804.vo.lu)
12:36:33*TheSeven can't reproduce any problems using test_disk.rock
12:47:35 Join Crackerizerizer [0] (n=7135a017@giant.haxx.se)
12:49:58***Saving seen data "./dancer.seen"
12:50:30Crackerizerizerhello
12:50:54 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
12:51:15CrackerizerizerI'd like to ask something about font and graphics.
12:51:29Tornethen as it
12:51:31Torneask, even
12:51:44 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
12:52:03CrackerizerizerI found out that latest svn code has a lot of graphic api changed.
12:52:06pondlifeI haven't started an investigation yet, but a new simulator built today keeps pausing itself, every few mins. Any idea why that might be - I'd guess some kind of HID thing...
12:52:57tomerspondlife: HID is not enabled in simulator
12:53:09 Quit tomers ("CGI:IRC")
12:53:09Crackerizerizermy old patch for thai font doesnt work anymore and i'm a little confusing.
12:53:13pondlifeOK - one theory out then
12:55:07*TheSeven just caught another worm creeping out of the can
12:55:53CrackerizerizerHow are fonts drawn now?
12:56:45TorneCrackerizerizer: what revision was the patch for?
12:57:21CrackerizerizerTorne: I'm not sure. may be before 3.2 released.
12:57:36Torneis it on FS?
12:57:47CrackerizerizerI saw a lot of change in the graphic api.
12:57:52Tornelike what?
12:58:21CrackerizerizerI mean in the lcd driver.
12:58:33Crackerizerizermy patch was applied to the lcd driver.
12:58:50CrackerizerizerI'm not so sure.
12:58:53Torneyou mean changes to the implementation, not the API?
12:59:00Crackerizerizeryes.
12:59:06Tornedon't say API then :)
12:59:07TorneA whole bunch of new ifdefs have been added for vertical packed LCDs
12:59:11Torneit may be that?
12:59:14Crackerizerizer:) sorry.
12:59:24Torneis the patch in flyspray?
12:59:30Crackerizerizeryes
12:59:44Torne..which number?
12:59:46Crackerizerizerit corrects vowel and tonal maker position.
12:59:55Crackerizerizergive me a second.
13:00
13:00:51CrackerizerizerI also see that there is an arabjoin library in the source code.
13:01:07Torneyes, people have been working on RTL support lately
13:01:12Crackerizerizerwhich i didnt see in my previous check out.
13:01:19Tornewhich may also be relevant
13:01:34Crackerizerizeremm..this is interesting.
13:02:01Crackerizerizerbecause i may write a new patch for thai.
13:02:23CrackerizerizerTorne: 9595
13:03:29Tornem
13:03:32TorneHm, even
13:03:47TorneSo your patch is doing combining characters?
13:03:54Crackerizerizeryes
13:03:58Tornebut only for your specific requirements :)
13:04:06Crackerizerizeryes :)
13:04:11Tornethat shouldn't be too difficult to bring up to date with svn
13:04:16Tornebut it's not going tobe committed like that, for sure :)
13:04:27CrackerizerizerI'd like to make it more generic.
13:05:06TorneIs it really sufficient to do combining chars like that in all fonts (by just overprinting in foreground?)
13:05:11 Quit dfkt (Read error: 145 (Connection timed out))
13:06:15CrackerizerizerI think so, if that font has thai.
13:06:17pondlifeWow - this is really annoying ... the sim pauses (i.e. just like I pressed the PLAY/PAUSE button) every minute or so. Time to svn up -r, I guess...
13:06:50CrackerizerizerI use that patch for almost a year now without any problem.
13:07:12TorneCrackerizerizer: yes, but my point i syou have only used it for thai
13:07:21Tornethere are lots of languages that need combining chars.
13:07:26TorneI guess it's reasonable, though.
13:07:38Crackerizerizeroh, yes.
13:07:58TorneI think a proper patch to support combining character rendering in rockbox, for unicode in general, would have a reasonable chance of being looked at
13:08:02Crackerizerizerbut i dont know if other language has the same combination as thai.
13:08:30Torneother languages won't use the same combinations, no
13:08:45Tornebut the mechanism works similarly/the same in all languages in unicode, afaik
13:10:01CrackerizerizerTorne: how is that?
13:10:26TheSeventhe absolute killer FTL test seems to be extracting rockbox-full.zip twice (overwriting), then copying some 350MB of music (multiple folders) and then copying over system32 :-P
13:10:32Tornehttp://en.wikipedia.org/wiki/Combining_character
13:11:40Zagorsystem32, that's cruel
13:12:01Zagor;)
13:12:35CrackerizerizerI'm reading. let's talk about how fonts are drawn in rockbox.
13:13:20TheSevenlooks like a USB fluke killed the transfer once again
13:13:24CrackerizerizerTorne: can you give me an over view.
13:16:07Crackerizerizeror at least, a direction.
13:26:54 Join pamaury [0] (n=pamaury@140.77.26.148)
13:31:18 Quit kugel ("exit(0);")
13:32:03 Part JackWinter ("Konversation terminated!")
13:33:31polobricolo where is the wakeup struct defined ?
13:40:19 Quit TheSeven (Nick collision from services.)
13:40:38 Join The_Seven [0] (n=theseven@rockbox/developer/TheSeven)
13:40:50 Nick The_Seven is now known as TheSeven (n=theseven@rockbox/developer/TheSeven)
13:41:41 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
13:49:05 Join DerPapst [0] (n=DerPapst@p4FE8F237.dip.t-dialin.net)
13:53:10amiconnTheSeven: test_disk does filesystem writes, which are not just sequential. The data is written sequentially, but fat updates and directory updates are out of sequence, as well as possible fragmentation
13:53:42pondlifehttp://www.rockbox.org/since34.html isn't displaying much...
13:54:32topikit does here
13:55:02pondlifehttp://www.rockbox.org/since33.html is ok, but since34 has no table
13:55:18 Part LinusN
13:55:20pondlifeIE8 (sadly - at work)
13:55:51 Join ps-auxw [0] (n=arneb@dyn37.ps-auxw.de)
13:56:04topikonly 969 validation errors. perhaps that explains it
13:56:14pondlifeAh, ctrl-refresh fixed it, ignore me
14:00
14:00:04 Join daurn| [0] (n=daurnima@freenode/staff/daurnimator)
14:02:10markunCrackerizerizer: did you download the source?
14:02:11 Join kugel [0] (n=kugel@rockbox/developer/kugel)
14:02:45Crackerizerizeryes, i checked out from svn
14:03:04Crackerizerizeryesterday
14:03:12markunok, let me see where you could start
14:03:34 Quit daurn (Read error: 60 (Operation timed out))
14:04:17Crackerizerizermarkun: thank you.
14:05:23Crackerizerizermy previous patch combines charecters in the lcd driver.
14:05:40Crackerizerizerbut i think there should be a better way.
14:06:12markunin firmware/drivers/lcd-bitmap-common.c?
14:06:38markuncan I see your patch?
14:07:09CrackerizerizerIt doesnt work any more in the latest code.
14:07:22markunstill, can you give me a link?
14:07:29Crackerizerizerpatch no: 9595
14:07:43markunthanks
14:07:44Crackerizerizerhttp://www.rockbox.org/tracker/task/9595?string=thai&project=1&type%5B0%5D=4&sev%5B0%5D=&pri%5B0%5D=&due%5B0%5D=&reported%5B0%5D=&cat%5B0%5D=&status%5B0%5D=open&percent%5B0%5D=&opened=&dev=&closed=&duedatefrom=&duedateto=&changedfrom=&changedto=&openedfrom=&openedto=&closedfrom=&closedto=
14:08:01CrackerizerizerIt is dirty but works
14:08:06Crackerizerizeronly for thai.
14:08:48Crackerizerizerand i've been thinking about make it more generic.
14:09:10*kugel wonders about rasher's statement that rockbox would be C89
14:09:54kugelCrackerizerizer: looks like the patch has 3 identical hunks
14:10:04Crackerizerizeryes
14:10:05markunkugel: he knows that already :)
14:10:08kugelseems to be enough to copy&paste 1 of them into lcd-bitmap-common.c
14:10:51Crackerizerizerwhat about making rockbox to support character combining for the other languages.
14:11:03Crackerizerizer?
14:11:19markunof course we would want that
14:11:28CrackerizerizerI'm reading how the combination works..
14:11:47markunwe had a request for (I think) ababic combining chars as well
14:11:51markunarabic
14:12:59 Join robin0800 [0] (n=robin080@cpc3-brig8-0-0-cust436.brig.cable.ntl.com)
14:13:01markunCrackerizerizer: http://forums.rockbox.org/index.php?topic=22388.msg156182#msg156182
14:13:08markun:)
14:15:49Crackerizerizermarkun: thanks
14:16:30Crackerizerizerit seems that a string need to be preprocessed before drawn.
14:16:55markunwhat do you mean?
14:18:13markunthe characters that get combined would have to be ignored when calculating the string width and skipped when scrolling
14:18:39markunis the combining glyph added before or after?
14:19:47Crackerizerizermarkun: in thai?
14:19:59markunCrackerizerizer: also in thai, yes
14:20:08markunI assume it's the same with all of them, no?
14:20:38markundid you have problems with scrolling with your patch?
14:20:43CrackerizerizerI think so, it is after in thai.
14:20:55Crackerizerizernot at all.
14:21:10Crackerizerizerit works great for supported font.
14:21:21 Join Yadda [0] (n=62b0859e@giant.haxx.se)
14:22:25 Join webguest94 [0] (n=5d8489a1@giant.haxx.se)
14:22:34YaddaHey, quick question. I'm modding an iPod 5G, and want to make a matching theme. Can someone give me pointers on making rockbox themes?
14:23:01Yaddaer, can I HAVE pointers on it. I don't want to set myself up for the 'yes' answer.
14:23:09Yadda...oh god, I did it again.
14:23:19Yaddawhat are some pointers, then.
14:23:30topikpointers are sort of tips
14:23:39Yadda*trying to be upyeah.
14:23:41 Quit webguest94 (Client Quit)
14:23:42Yaddadammit
14:23:45Yaddadidn't backspace all the way
14:23:55Yaddayeah, I want tips on theme-makinging. :P
14:24:30topiktake a look at existing themes
14:24:35topikand use http://www.rockbox.org/wiki/CustomWPS as a reference
14:24:45linuxstbYadda: Plus http://www.rockbox.org/wiki/SimpleGuideToWPSMaking
14:24:46Crackerizerizerand in thai, there is second level diacritic called tonal maker.
14:25:06topikand use the simulator to test your theme-in-progress
14:25:11YaddaAHA. Those links are exactly what I was looking for.
14:25:37YaddaI'm making a Half-Life 2 themed iPod, and figured that if I want to go the whole nine yards, why not make a rockbox theme.
14:25:43markunCrackerizerizer: so what is the challenge? Shouldn't we just draw the all the combining glyphs over the previous one?
14:26:21topikthat's a rhetorical question, right?
14:26:59YaddaSo yeah, the main problem is the cost of parts.
14:27:04Crackerizerizermarkun: i'm checking if other languages use the same trick.
14:27:09YaddaDamn iPod parts vendors are insane.
14:27:37 Join dfkt_ [0] (i=dfkt@unaffiliated/dfkt)
14:27:56Crackerizerizerif they are, the patch should be easy.
14:28:23markunyes, just make a lookup table to see which characters have a zero-width advance
14:29:43YaddaAny tips on how to get a LED in the back of the iPod turn on when the iPod comes out of sleep mode? >>;
14:30:04Crackerizerizeris rockbox supported any character combination?
14:30:42markunWhat do you mean exactly?
14:31:01 Quit robin0800 (Remote closed the connection)
14:31:21 Join robin0800 [0] (n=robin080@cpc3-brig8-0-0-cust436.brig.cable.ntl.com)
14:31:56*TheSeven didn't know that iPods had LEDs, besides the backlight
14:32:32markunCrackerizerizer: did you see http://rishida.net/docs/unicode-tutorial/part3 ?
14:32:40markunnot that we can support all of that..
14:33:13kugelTheSeven: any plans about frequency scaling on nano2g?
14:33:59TheSevenkugel: it doesn't seem to yield much additional battery life
14:34:04linuxstbkugel: Have you read the IpodNano2GPort wiki page?
14:34:08TheSeveni haven't tried undervolting yet, though
14:34:25kugellinuxstb: yes, not thoughoutly though
14:34:59 Quit Topy44 (Read error: 104 (Connection reset by peer))
14:35:02 Join T44 [0] (n=Topy44@g227169068.adsl.alicedsl.de)
14:36:23kugelTheSeven: did you messure the current?
14:37:50TheSevenyes, scaling down to half speed seemed to reduce it by ~10%, but only if you also clock down the PLL, which means that the APB clock won't be stable during that
14:38:25*TheSeven wonders where all the other power consumption is coming from
14:39:37 Quit DerPapst ("Leaving.")
14:41:39 Join JackWinter [0] (n=jack@vodsl-10804.vo.lu)
14:43:28 Join faemir [0] (n=faemir@78.33.109.163)
14:43:44kugelit sounds strange that frequency has so little impact. of course, unless there's another factor that makes all other facters having a little influence
14:47:15 Quit dfkt (Read error: 110 (Connection timed out))
14:47:59 Join dfkt [0] (n=dfkt@unaffiliated/dfkt)
14:49:08 Quit SUSaiyan (Read error: 104 (Connection reset by peer))
14:50:00***Saving seen data "./dancer.seen"
14:50:32Crackerizerizermarkun: thanks
14:52:40TheSevenkugel: the main advantage of speed scaling will probably be being able to undervolt to an even lower level
14:53:01Crackerizerizermarkun: are fonts in rockbox fixed font?
14:53:45TheSevenyou mean fixed width? then no.
14:53:58 Join _zic [0] (n=user@91-165-241-175.rev.libertysurf.net)
14:54:03kugelwow, why is the scrolling going so crazy with the wheel acceleration?
14:55:08 Join Casainho [0] (n=chatzill@87.196.39.27)
14:55:08kugelis it polled or interrupt driven?
14:55:27 Join SUSaiyan [0] (n=SUSaiyan@cc84863-b.zwoll1.ov.home.nl)
14:55:38CrackerizerizerTheSeven: Thanks
14:55:55TheSevenkugel: interupt-driven
14:56:08kugelI got a state where it actually accelerates, but it doesn't scroll fluently
14:56:22 Quit Casainho (Client Quit)
14:56:27kugelinstead it does a big jump every 100ms
14:56:44TorneCrackerizerizer: Back, soryr, work internet went down
14:56:52kugelI'm not sure the interrupt works as expected
14:57:27TheSevenkugel: i would rather blame that on accelleration schemes used for button-based scrolling targets
14:57:38TorneCrackerizerizer: as far as I know all combining chars work the same in unicode; certain characters are just always interpreted as being combined with the one before (sometimes multiple ones in sequence)
14:57:48kugelTheSeven: huh?
14:57:58kugelthe code is different for button-based acceleration
14:57:59 Quit dfkt_ (Read error: 110 (Connection timed out))
14:58:01Torneso it's literally just a matter of having a lookup table and not advancing the drawing position for those
14:58:13TheSevenand i think the rockbox driver reduces the wheel accuracy a great deal, by reducing it to 16 or such steps instead of 96
14:58:14Tornebut yes, it probably does affect scrolling/etc as well
14:58:20 Join dfkt_ [0] (n=dfkt@unaffiliated/dfkt)
14:58:29TheSeven(as a single step usually means scrolling a single item)
14:58:43Tornei expect we'd need to also skip them when counting the width of a line etc
14:59:47TheSevenkugel: is the scrolling accelleration done by firing button events more quickly from the button driver, or is there also something makes it scroll more steps per button event?
14:59:49kugelTheSeven: the wheel responds very slowly, especially if you move the finger faster. that's not what you experience on e200 or fuze for example where it's basically instant
15:00
15:00:39*TheSeven doesn't have any other targets, so he can't judge on this
15:00:55CrackerizerizerI have checked what thai font look like.
15:01:06kugelI can, and there's defintiely something fishy
15:01:15TheSevenbut the wheel and lowlevel driver code shouldn't be the issue here, that's responding very fast in an ibugger-based testing app
15:01:31TheSevenso it must be hiding in some more highlevel place like acceleration code
15:01:43TorneCrackerizerizer: http://unicode.org/notes/tn2/ <- this is a technote on how to render combining chars :)
15:01:54Tornethough unicode.org seems to be timing out for me, but it's in google cache.
15:02:25TorneI dunno that it's really appropriate to us though.
15:02:26Crackerizerizeras i mentioned before, vowel and tonal maker have different vertical position. So, my patch works.
15:02:45kugelTheSeven: I doubt the higher level code is on fault
15:02:54TorneCrackerizerizer: the problem is things like horizontal width..
15:03:00kugelit works fine for a dozen other scrollwheel targets
15:03:11TorneCrackerizerizer: where to place the combining char horizontally relative to the character it's being combined with
15:03:23TheSevenkugel: I'm not saying it's outside the clickwheel driver, but it's definitely not that ints aren't coming through
15:03:30Tornee.g. if the main character is 9 pixels wide and the combining character is 3 pixels wide
15:03:40Torneyou probably don't want to put them both at the same X coord.
15:03:45Torneas that's likely to look weird
15:04:31CrackerizerizerTorne: let's me check my previous patch.
15:05:02TorneYou do draw them all at the same left position
15:05:12Tornewhich means that narrow combining chars on wide base chars will be too far to the left
15:05:17Torne(for many languages)
15:05:25TorneIt may be that it doesn't look too bad for thai
15:05:34Tornebut it would look weird for, say, regular Roman characters with accents
15:05:48kugelTheSeven: it's strange, I notice that I have a working acceleration. but the UI is responding so slow
15:06:18TheSevenyes, that's my impression, too
15:06:18CrackerizerizerTorne: Emm, correct.
15:06:42TheSevenand i somehow tend to blame it on highlevel UI code not being able to cope with the events coming in fast enough
15:06:53kugelTheSeven: ah, it's actually only slow in the filebrowser
15:07:24Crackerizerizerwhat about applying the patch based on language/charset setting in rockbox?
15:07:36TheSevenhm, my filebrowser is responding like every other menu, but the behavior of e.g. logf is driving me mad
15:08:02kugelscrolling through the volume levels in the setting is blazing fast compared to filebrowser
15:08:13TorneCrackerizerizer: I dunno how happy people would be to have a hacky solution that kinda looks ok for just one language
15:08:21Torne:)
15:08:26Torneeven if it was only applied to that language.
15:08:29topikdatabase screen scrolls slower too for me
15:08:42Tornewhich is already the case for your patch without a setting.. nobody else is going to be using Thai combining marks
15:09:07Torne(the same is not true of the more general combining chars though: acute/grave/umlaut/macron/etc
15:09:31kugelbrowsing tracks in the database is most horrible
15:10:20 Quit FlynDice (Remote closed the connection)
15:10:46CrackerizerizerTorne: I'm thinking about applying the patch based on language/charset setting.
15:11:02TorneCrackerizerizer: yes, i don't think that's going to be considered, is what i'm saying.
15:11:19TorneCrackerizerizer: I would guess that the 'minimum' to do, which would probably be fine for your case and for many others, is to render the combining char in the way you do, but center it horizontally with respect to the base character
15:11:26 Join MethoS- [0] (n=clemens@134.102.106.250)
15:11:42Tornethat works about as well as we can reasonably manage in our font system, for almost all combining chars commonly used in many languages
15:11:43kugelthe jumpiness using ipodnano1g defines is probably due to the usage of current_tick instead of a usec timer
15:13:06TorneCrackerizerizer: if you implemented that, with a full list of which characters are combining chars, and the binsize impact was not too much then that could probably get committed
15:13:22TorneCrackerizerizer: the only dodgy case there that I can think of is combining chars which are *wider* than the base char they are applied to
15:13:33Tornee.g. if you shove an accent on an 'i' in a font where 'i' is only 1 pixel wide
15:13:46kugelTheSeven: I think button-clickwheel can't work properly with current_tick. the driver is designed for calculating the speed from wheel changes from 1 us to another
15:14:19CrackerizerizerTorne: I got your point now.
15:14:31*TheSeven needs to implement the usec timer then
15:14:46TorneCrackerizerizer: A neat patch that did that could probably go in, even if it's not perfect for every language
15:14:54Torneit sounds like that would be sufficient for you, at least
15:14:59Torneand it would probably work fine for some others.
15:15:17Torneif some language finds that's not sufficient then someone who knows baout rendering that language would probably have to propose a fix.
15:15:43kugelTheSeven: basically, it uses the time difference to messure the accel factor. but with current_tick the time difference is 0 in most of the cases.
15:16:13 Quit antil33t (Read error: 145 (Connection timed out))
15:16:29CrackerizerizerTorne: Emm, then my previous patch with a little modification should be ok? at least for me :)
15:16:39TorneCrackerizerizer: Hm?
15:17:02TorneCrackerizerizer: I mean one that works for all combining chars in unicode
15:17:03Tornenot just thai
15:17:06kugelTheSeven: it doesn't necessarily have to be a usec timer
15:17:20Torneit doesn't necessarily have to be the optimal way to render it for all languages
15:17:43kugelwe could use the TIMER1 value that is used for the tick interrupt in the early samsa days without actually implementing it as a usec_timer or even having a udelay()
15:17:44TheSevenwe don't have anything more accurate than the tick timer right now, but i know that the HW must have a usec one
15:17:46Tornei think it's fair to assume that "render combining chars centered over the base char" is acceptable for our purposes
15:17:56Torneat least until someone tells us otherwise
15:18:12CrackerizerizerTorne: then, i need to find more knowledge.
15:18:23TorneCrackerizerizer: hm? how so?
15:18:43Crackerizerizerreading wikipedia and other sources.
15:18:58kugelTheSeven: the timer used for tick task interrupts is probably accurate enough. you may need to do some scaling if it's way off from "1 value change = 1 usec"
15:19:14TorneCrackerizerizer: Well yes, but that shouldn't be that difficult
15:19:34 Quit dfkt_ ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
15:19:34TheSeven1 value change = 1ms I think
15:19:37TheSevenor was it 100µS?
15:19:44kugelwhat's the frequency?
15:20:22TheSevenjust looked it up, it's 100µS = 10kHz
15:20:31CrackerizerizerTorne: yes, just making sure for other language.
15:20:39kugel100us seems sufficient to me, I'll have a play
15:20:42TheSevenin fact implementing the usec timer won't be any harder i think
15:20:45TorneCrackerizerizer: think of it this way: if you do a little work now to make it generally useful then you won't have to keep updating your patch when we change the code :)
15:20:54kugelI have the ipod for another 3 hours
15:21:35CrackerizerizerTorne and others: thanks for your help, hopefully i can get it done tonight.
15:21:38 Join antil33t [0] (n=Mudkips@119.224.12.185)
15:21:47 Join liar [0] (n=liar@213.162.70.99)
15:21:48gevaertskugel: we'll tell you how to brick it in 175 minutes :)
15:21:49 Quit pamaury (Remote closed the connection)
15:22:05TorneCrackerizerizer: no problem
15:22:39 Join dfkt_ [0] (n=dfkt@d86-33-57-45.cust.tele2.at)
15:22:50TorneCrackerizerizer: we like people who are willing to rework patches for acceptance! :)
15:23:41TheSevenkugel: http://pastie.org/663436
15:23:57TheSeventhat's how apple is accessing something that seems to be a usec timer at least
15:23:58 Quit antil33t (Read error: 54 (Connection reset by peer))
15:28:15Crackerizerizerit seems that rockbox's recent code is better organized. :)
15:29:55 Join Topy [0] (n=Topy44@f054182193.adsl.alicedsl.de)
15:31:52 Join antil33t [0] (n=Mudkips@119.224.12.185)
15:34:00 Quit dfkt (Nick collision from services.)
15:38:28kugelTheSeven: I don't have the ds, any idea what 0x3C700084 is?
15:40:31kugela 5th timer?
15:40:55TheSevenit's undocumented, but beneath the timer regs
15:41:30TheSevenif it would be a regular timer, 80/84 would be the control/command regs, so that's probably not true
15:41:49 Quit Crackerizerizer ("CGI:IRC (Ping timeout)")
15:41:51 Join evilnick [0] (i=0c140464@rockbox/staff/evilnick)
15:41:52TheSevenwhat about just showing their contents on the dbg_ports screen?
15:42:26kugelgood plan
15:43:41 Nick dfkt_ is now known as dfkt (n=dfkt@unaffiliated/dfkt)
15:48:49 Quit T44 (Read error: 110 (Connection timed out))
15:51:45 Quit jsj (Read error: 60 (Operation timed out))
15:53:17kugelhrm, the cursor moves by just touching the wheel slightly without rotating
15:53:19 Join DerPapst [0] (n=DerPapst@asr-nat2.its.fh-giessen.de)
16:00
16:02:57TheSevensounds a little hyper-sensitive...
16:03:52 Quit Zagor ("Don't panic")
16:04:56kugelI think I have found a reasonable acceleration, but the respnsiveness is still really slow
16:05:48kugelit feels like the caches were turned off
16:06:04TheSevenyes, but they aren't
16:06:28 Quit MethoS- (Remote closed the connection)
16:07:53 Quit liar (Remote closed the connection)
16:10:06kugeltest_fps shows 50fps, so that's not the problem. the ui sluggishness is strange
16:10:37TheSevenhm. my test_fps is around 70 for full screen RGB
16:15:06 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
16:15:30 Quit antil33t ()
16:16:48 Join T44 [0] (n=Topy44@f054182193.adsl.alicedsl.de)
16:17:26 Quit gitster (Remote closed the connection)
16:17:26 Quit lifeless_ ("Ex-Chat")
16:17:31 Join gitster [0] (n=user@ip68-225-240-211.oc.oc.cox.net)
16:17:46kugelit seems the more and longer the strings to display are, the slower is the ui
16:18:17 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
16:19:20 Quit Topy (Read error: 104 (Connection reset by peer))
16:22:26 Join JdGordon| [0] (n=Miranda@c-24-22-210-83.hsd1.wa.comcast.net)
16:22:49linuxstbTheSeven: That's possibly the different LCDs - IIRC, I got exactly 54fps for both RGB and YUV full-screen updates.
16:28:33rasherkugel: I'm pretty sure that's what I was told way back when (C89)
16:31:17 Join esalazioneHg [0] (n=nay@151.61.174.72)
16:31:42*TheSeven wonders if there are cans that can contain infinite amounts of worms.
16:32:29esalazioneHghi!i'm trying to put rockbox on my ipod nano 2gen, but i have trouble with the ipodpatcher
16:33:11esalazioneHgafter lunching ./ipodpatcher it shows some infos and terminate with "2gen nano unsupported by ipodpatcher"
16:33:54linuxstbesalazioneHg: Sounds like you're using an old version of ipodpatcher.
16:34:07linuxstbWhere did you download it from? (and when?)
16:34:26 Quit matsl (Remote closed the connection)
16:34:32esalazioneHgi've just downloaded it from http://download.rockbox.org/bootloader/ipod/ipodpatcher/linux32x86/
16:34:46esalazioneHgas said in the guide for ipod 2gen nano
16:34:52 Join lifeless_ [0] (n=lifeless@94.50.188.8)
16:35:06kugeltry with sudo
16:35:12esalazioneHgi'm root
16:35:37 Join StealthyXIIGer [0] (n=stealthy@c-68-62-19-6.hsd1.mi.comcast.net)
16:35:41linuxstbesalazioneHg: What version number is reported by ipodpatcher?
16:35:41esalazioneHgwhen i lunch ./ipodpatcher it sais
16:35:41esalazioneHgipodpatcher 4.0 with v3.0 bootloaders (v1.0 for 2nd Gen Nano)
16:35:53gevaertswhat does lsusb output for the ipod?
16:36:12esalazioneHgBus 001 Device 006: ID 05ac:1260 Apple, Inc. iPod Nano 2.Gen
16:36:40TheSevenlooks sane...
16:36:52 Join dfkt_ [0] (i=dfkt@chello062178002170.1.11.univie.teleweb.at)
16:37:10amiconnmarkun: Simpyl drawing combining marks over the base char won't work properly, due to the draw modes
16:37:10esalazioneHgdo you want i paste the output of ipodpatcher?
16:37:46TheSevencan't hurt, but please use a pastebin for anything more than a few lines
16:38:02esalazioneHgok :)
16:38:12amiconnIt could be made to work for DRMODE_SOLID (combining would then need DRMODE_FG) and plain DRMODE_FG and for the actual inverse of each
16:38:22amiconnSo half of the modes wouldn't work.
16:39:50amiconnFor combining to work properly, we'd need an intermediate bitmap. One of my (too many :\) ideas is to add such intermediate bitmap for text drawing. It would also allow for easy algorithmic emboldening
16:39:52esalazioneHghttp://pastebin.com/d2f92ea7f
16:39:52 Join dfkt__ [0] (i=dfkt@chello062178002170.1.11.univie.teleweb.at)
16:40:04 Quit dfkt (Nick collision from services.)
16:40:13 Quit dfkt_ (Nick collision from services.)
16:40:17 Nick dfkt__ is now known as dfkt (i=dfkt@unaffiliated/dfkt)
16:40:19TheSevenbtw, is any nano2g user still experiencing "ATA error: 1"?
16:40:33 Join Strife89 [0] (n=michael@168.16.239.253)
16:40:50linuxstbesalazioneHg: That's very weird... Let me test that binary...
16:41:21amiconnIt might even speed up text drawing, if the intermediate bitmap resides in iram, because there would be fewer calls to lcd_mono_bitmap(), drawing multiple chars each
16:42:09TheSevenlinuxstb: this binary does at least contain the string "[ERR] *** ipodpatcher does not support the 2nd Generation Nano! ***"
16:42:32*linuxstb has found the bug...
16:42:56Torneamiconn: that's a good point, hrm
16:43:12esalazioneHgheheheh
16:43:15 Quit JdGordon| (Read error: 113 (No route to host))
16:43:16linuxstbesalazioneHg: The problem seems to be that ipodpatcher still has an old check for the Nano2g, but that only succeeds with an older version of the Apple firmware.
16:44:02TheSevenwow, there are still nano2gs with old firmwares? i would have thought itunes would have caught these by now.
16:44:59esalazioneHghehehe, i don't use iTunes, so i never updated the firmware
16:45:22esalazioneHgi don't know if gtkpod could update a firmware too
16:45:57linuxstbesalazioneHg: Can you try linuxstb.cream.org/rockbox/ipodpatcher">http://linuxstb.cream.org/rockbox/ipodpatcher ?
16:46:41esalazioneHgi'll try it now
16:48:29esalazioneHgit seems to work... now i write "i" and try to install...i'll let you know the result ;)
16:48:55esalazioneHg[INFO] Bootloader installed successfully.
16:49:25esalazioneHgyahoooo! i'll try to switch on the ipod..but i think it's all ok
16:50:02amiconnTheSeven: Error returns in rockbox should *always* be negative
16:50:03***Saving seen data "./dancer.seen"
16:50:48esalazioneHgok, it works! thanks guys!!
16:53:03TheSevenamiconn: that's something to add to CONTRIBUTING then?
16:53:18TheSeven(or however that guideline file is called)
16:54:36 Join toffe82 [0] (n=chatzill@12.169.218.14)
16:56:34 Join FOAD [0] (n=dok@dinah.blub.net)
17:00
17:00:44 Part esalazioneHg
17:04:33 Quit JackWinter (Read error: 54 (Connection reset by peer))
17:07:09 Join JackWinter [0] (n=jack@vodsl-10804.vo.lu)
17:09:15 Join Hillshum [0] (n=hillshum@75-165-234-241.slkc.qwest.net)
17:15:58 Join jgarvey [0] (n=jgarvey@cpe-098-026-065-013.nc.res.rr.com)
17:18:14 Join pamaury [0] (n=pamaury@140.77.26.96)
17:19:34pondlifeOK, I've found the revision that breaks the H300 sim button handling a bit. See http://www.rockbox.org/tracker/task/10702...
17:19:52pondlifer22799 is the first version that shows the problem.
17:20:29polobricolois there an curses version ofthe simulator ?
17:20:39 Join Casainho [0] (n=chatzill@87-196-39-27.net.novis.pt)
17:21:06Tornepolobricolo: no.
17:23:43pondlifekugel: ping
17:26:52pamaurygevaerts: ping
17:28:14CIA-85New commit by 03theseven (r23307): Fix another bunch of iPod Nano 2G FTL bugs. It seems to work fine for me now, but this doesn't necessarily mean that it's bug-free.
17:28:45 Join HellDragon [0] (i=jd@modemcable178.248-201-24.mc.videotron.ca)
17:30:18TheSevenbtw, what about a build system IRC bot that reminds you if you've caused a red?
17:30:29 Quit DerPapst ("Leaving.")
17:30:46pondlifeGood idea (or a yellow, indeed)...
17:31:10kugelpondlife: pong
17:31:56HillshumTheSeven: This has come up before. I think Bagder implemented a mail blamer
17:32:17pondlifekugel: I'm having problems with the SDL button driver (I think). Any idea why I might get spurious buton presses on the H300 sim since you redid the driver way back in r22799?
17:32:27TheSevenit would be noticed more quickly if it would ping the one to blame on IRC
17:32:53kugelpondlife: no idea sorry
17:33:01HillshumTheSeven: Yeah, and more convenient in general
17:33:13pondlifeNo problem, just thought you might - and save me working it out!
17:33:15kugelhave you tracked it down to that revision? kkurbjun changed it a bit after too
17:33:46pondlifeYes, all ok in 22798, broken in 22799
17:34:34pondlifeTotally reproducible here - I've just put some printf stuff in the SDL driver and am awaiting a failure
17:35:48pondlifeAha - no + button press, yet still it paused...
17:37:26kugeljust by printf'ing?
17:37:50pondlifeJust an "OUCH" in the case..
17:38:28kugelhm, I was running the h300 sim a lot since then for testing custom statusbar on remotes, I haven't noticed that
17:39:40amiconnpondlife, kugel: There are other button problems in some sims. E.g. the 1st/2nd Gen iPod sim always posts 2 wheel events per cursor up/down
17:40:02amiconnVery nasty for lists with even number of entries, as every other entry is unreachable
17:40:04kugelamiconn: that's since kkurbjun change
17:43:18pondlifeHmm, it happens without a call to button_event().
17:53:18gevaertspamaury: pong
17:55:31pamaurygevaerts: hello, I have a question; during the development of the MTP code (which will hopefully will soon be finished I hope) I dreamed about a simulator which could do virtual usb with the linux kernel. Of course it doesn't but I wondered if such a feature is worth implementing it or not ? It would make it possible to register the simulator as a virtual usb device (by using a kernel module, such a module already exists)
17:56:10 Quit Strife89 ("My number of files is OVER 9000!")
17:56:36 Quit _zic ("Ухожу")
17:57:10gevaertspamaury: it would certainly be fun, but is it worth the effort? If you think it is, maybe it's worth looking at http://usbip.sourceforge.net/ as well
17:57:56 Join dfkt__ [0] (i=dfkt@unaffiliated/dfkt)
17:58:14pamaurywell usbip is not the project I had in mind because it do usb->ip in the kernel but it's similar in the idea. I'm not sure but if I have time I will check whether it's feasable.
17:58:43gevaertsthe nice thing about usbip is that it won't require platform specific code
17:59:33*TheSeven wonders what to do next
17:59:51gevaertsmake the FTL faster?
18:00
18:00:01pamaurygevaerts: well it's true that's we could do usb->ip->rockbox but there is virtual hci hub project in Sourceforge which would allow to do usb->rb directly but it's linux specific you're right.
18:00:10pamauryI'll have a more precise look at usbip
18:01:47 Quit pamaury ("exit(*(int *)0 / 0);")
18:03:11 Quit petur ("work->...")
18:04:47TheSevenlinuxstb: what's your opinion on the pcm problem?
18:06:59TheSevengevaerts: making the FTL faster means another 6KB of additional BSS and another 6KB of additional stack usage
18:07:36TheSevenI'd rather have a look if we can get rid of some of the FATs sector buffers on the stack
18:11:46pixelmaI saw some reports (in the forums I believe) of Nano 2nd gen owners said they can hear that the channels switch after each pause and play, sounded familiar somehow. Wasn't there a similar problem on the 1st gen Nanos (or other Ipods) too once?
18:12:25pondlifekugel: It was down to having car adapter mode enabled - the newer button driver must allow events to be processed properly.
18:12:46pondlifeThs sim simulates a charger disconnection..
18:15:29kugelhaha
18:15:30 Join phanboy4 [0] (n=benji@c-24-98-43-198.hsd1.ga.comcast.net)
18:16:00kugelit fixed a number of thinks by using more on-target code
18:16:28pondlifeNo idea why I had car adapter mode enabled either!
18:16:35pondlifeAnyway, thanks
18:16:39 Quit dfkt (Read error: 110 (Connection timed out))
18:18:55 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
18:21:53 Join bertrik [0] (n=bertrik@87.211.49.117)
18:22:11TheSevenpixelma: if that happened on the 1g, too, this would be very interesting, as it's using the same codec
18:22:26kugelCasainho: ping
18:24:17 Part pondlife
18:24:25Casainhokugel: pong
18:25:04kugelCasainho: the mini2440 wiki page says button driver is at 100%. does it mean the touchscreen works?
18:26:44Casainhokugel: no, that means the buttons works. The touch screen not yet. I heard that domonoky is working on touchscreen and Bob_C on audio
18:27:11kugelcan I work on something too? :(
18:27:13 Nick dfkt__ is now known as dfkt_ (i=dfkt@unaffiliated/dfkt)
18:27:18 Nick dfkt_ is now known as dfkt (i=dfkt@unaffiliated/dfkt)
18:27:21*kugel thinks it shouldn't say 100% then
18:28:16TheSeven...or just add a touchscreen column
18:28:22TheSevenor rather row
18:29:32TheSevenkugel: did you use the µsec timer in the end?
18:30:11Casainhokugel: domonoky wrote that on Rockbox wiki page, go and change it if you think so. Ask to Bob_C what is being doing −− I am just being blogging, nothing more: http://lyre.sourceforge.net/?q=blog
18:31:09Casainhokugel: I remember you told me that you would not want the board for Rockbox.....
18:31:30kugelTheSeven: just TACNT, I also scaled the necessary values accordingly (I think the 1000000 in the v calculation needs scaled too)
18:31:55kugelor TBCNT, I don't remember. the one that's used for tick tasks
18:31:58Casainhokugel: I mean, I didn't know you were interested on it. I would like to know if you plan to use it or not... because some people are asking for HDD on it
18:32:03bertrikI can think of something that could cause flipping, e.g. accidentally losing 2 bytes somewhere
18:32:46kugelCasainho: I said I'm not planning to work on making it the lyre project (so that it can be selled)
18:33:20bertrikFor the samsung targets I think we should do something like #define values for PCLK and HCLK, so timers can be scaled independent of the actual frequency of the target
18:33:50 Join JdGordon1 [0] (n=jonno@173-128-233-66.pools.spcsdns.net)
18:35:12kugelTheSeven: the problem is that the ui isn't responsive enough to test it properly
18:36:55markunbertrik: are you talking about stereo flipping? We used to have that problem a lot with the gigabeat f
18:36:56JdGordon1kugel: probably a silly question... does %w[e/d] confuse the .sbs loader?
18:37:07bertrikThe kernel timer is now probably slightly off for other samsung targets
18:37:43bertrikmarkun, yes, sorry, I was indeed talking about stereo flipping
18:37:47kugelJdGordon1: it shouldn't
18:38:26CasainhoJdGordon1: hello :-) −− I would like to ask you if you already tested the Rockbox firmware on Mini2440?
18:38:46JdGordon1no
18:38:48TheSevenbertrik: you're talking about that 50MHz to 48MHz change?
18:38:51JdGordon1openocd is fucked on my desktop
18:39:10TheSevenI think this was 48MHz originally and we changed it to 50MHz for the nano, until we were proven wrong.
18:39:56bertrikTheSeven, yes I was talking about that
18:40:09JdGordon1kugel: and viewports are required? so completly ignore the default vp and just start with a new %V| line?
18:40:11TheSevenand there are define values for PCLK and FCLK_CPU (even though they aren't called like this)
18:40:46TheSevenit won't affect the codec anyways as that's driven by a different pll, so is there anything else that's timing-critical and not done using the tick or user timer?
18:40:53bertrikTheSeven, by the way, we don't really have to guess about it I think, since we know how the PLL is configured.
18:41:00kugelJdGordon1: you can do that yes. the only thing that works in the default viewport is loading images and showing other viewports (%Vda)
18:41:24TheSevenbertrik: I don't know the exact xtal freq of the nano
18:41:50bertrikWe have scans of the PCB, I hope that shows the crystal
18:42:54 Join bughunter2 [0] (n=bughunte@unaffiliated/bughunter2)
18:43:33 Quit polobricolo (Read error: 60 (Operation timed out))
18:43:54bertrik12, 24 or 48 MHz would make some sense as those are frequencies often used by USB controllers
18:44:35CasainhoJdGordon1: I have 2 computers and OpenOCD works great. Did you build it?
18:44:44TheSeventhere is a 24MHz crystal, yes, but the OSC freq is somewhere around 1.8MHz, and it's switchable to 32kHz by some means we don't know
18:45:26 Quit maruk ("Leaving.")
18:45:59markunbertrik: I can't find the gigabeat commit which fixed the stereo problem..
18:46:20TheSevenbertrik: calculating from PCM measurements, it seems to be 1843200Hz
18:46:26 Join Moiman [0] (n=harri@a91-153-34-172.elisa-laajakaista.fi)
18:46:48bertrikTheSeven, that is a common frequency for "communication" crystals
18:47:16bertrikexactly 16 times 115200
18:47:23TheSevencalculating from timer measurements I end up at 1846153 Hz
18:48:19JdGordon1did someone completly break the CustomWPS tables?
18:48:28TheSevenif it's a 1.8432MHz one, then FCLK_CPU will in fact be 191692800Hz
18:49:34*bertrik can't find nano2g pcb scans on the wiki
18:50:02TheSeventhere are some linked on the l4n wiki
18:50:04***Saving seen data "./dancer.seen"
18:51:48bertrikthe rockbox wiki should have that link too
18:53:41JdGordon1bah.. how did the word "Custom" get into the lang strings?
18:53:45JdGordon1they are redundant and silly
18:54:05JdGordon1it should either be just "statusbar" or "statusbar skin" or something
18:54:53TheSevenbertrik: and remember we also have a different pcm problem: interrupt latency
18:55:18bertrikTheSeven, I thought you found a fix for pcm that reduced it to 2 us? :)
18:55:31TheSeveni thought i'd flush dcache before passing the pointer to the DMA just in case something would still be stuck in the cache, but that was enough to increase latencies to a degree that sounds like dirty vinyl
18:56:11bertrikThe DMA controller can interrupt on half done transfers, but I don't know if we have some use for it yet
18:56:17TheSeventhat "fix" involves some questionable work in pcmbuf.c, to prevent it from overwriting "used" (but not entirely played yet) buffers too early
18:56:17 Quit phanboy4 ("Leaving")
18:56:24 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
18:56:48TheSeveni tried auto-refreshing transfers but they didn't work either for some reason (the dma controller just stopped it as usual)
18:57:21 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
18:57:42TheSevenand while the 2µs FIQ interrupt seems to fix most of it (and introduce some other bugs i still need to track down) i consider it a hack
18:57:47TheSeventhere must me a way to get rid of that realtime requirement
18:57:49bertrikI think the idea of the half-done interrupt is to tell the CPU which half of the DMA buffer can be updated with new data, but rockbox pcm doesn't work that way
18:58:10JdGordon1can #| and #> not be used inside conditionals?
18:58:32JdGordon1so a literal | and > gets displayed
19:00
19:00:48*JdGordon1 is a drongo
19:00:53JdGordon1time to wake up and start reading
19:02:38 Quit swilde (Remote closed the connection)
19:03:45JdGordon1is there going to be major objections to using http://pastebin.com/mde8f288 as the inbuilt statusbar?
19:04:16 Quit JackWinter (Read error: 104 (Connection reset by peer))
19:05:34JdGordon1excpet have it use the users 12/24hr setting for the clock
19:05:36 Quit evilnick (kubrick.freenode.net irc.freenode.net)
19:05:36NSplitkubrick.freenode.net irc.freenode.net
19:05:36 Quit saratoga (kubrick.freenode.net irc.freenode.net)
19:05:58 Quit kugel (Read error: 110 (Connection timed out))
19:06:21NHealkubrick.freenode.net irc.freenode.net
19:06:21NJoinevilnick [0] (i=0c140464@rockbox/staff/evilnick)
19:07:20 Join Lss [0] (n=Lss@cm46.delta91.maxonline.com.sg)
19:09:09TheSevenbertrik: what about making the pcm buffer work that way? is that feasible?
19:10:29 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
19:11:42 Quit domonoky (Client Quit)
19:12:30 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
19:15:32 Join JdGordon| [0] (n=Miranda@nat/microsoft/x-tdscixexavcfbibx)
19:15:44 Join JackWinter [0] (n=jack@vodsl-10804.vo.lu)
19:17:53bertrikTheSeven, I don't know, I'm not really a pcm expert, I'd have to ask the veteran developers
19:21:17bertrikwe could do a single DMA buffer used as a kind of circular buffer that we copy data into that is handed to use from the generic pcm layer, but the copying seems a bit wasteful
19:27:00domonokydoes anybody have a idea where to best put a header with generic mmc/sd definitions (copied from linux) ? at moment i would just put it as mmc-defs.h into the target tree. But maybe there is a better place ?
19:27:23JdGordon|firmware/include ?
19:27:38JdGordon|that seems to be where other sort of standard headers go?
19:28:10domonokyfirmware/include looks more like generic non-hardware depended headers..
19:28:41domonokythis is a header with definitions (currently) only needed by one target
19:29:01JdGordon|can it be merged with the current mmc.h?
19:30:01JdGordon|shotofads: http://forums.rockbox.org/index.php?topic=22983.0
19:30:39 Quit einhirn ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
19:30:46domonokyhm.. dont know if this is good. mmc.h seems to contain only the interface to access a mmc, the new defs are about mmc details only needed in the implementation (like mmc commands and answers)
19:31:26*domonoky wonders where other targets using mmc/sd places such info.
19:31:42TheSevenbtw, the usec timer seems to rather be a 5µsec timer
19:32:16TheSevenbertrik: can you have a quick check if that's also present on the S5L8700? Just read 0x3C700084 and see if it moves...
19:32:23 Quit JdGordon1 (Read error: 110 (Connection timed out))
19:32:36 Join HBK- [0] (n=hbk@rrcs-97-77-51-170.sw.biz.rr.com)
19:32:51bertrikwhat's at that address?
19:34:27domonokyhm, looks like other targets just embedd those definitions directly into the sd/mmc driver .c file...
19:35:09 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
19:35:31TheSevenbertrik: this seems to be a reg that increments at every 1000 cpu cycles
19:35:31 Quit KBH (Read error: 60 (Operation timed out))
19:36:05 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
19:36:16TheSevenbtw, the apple timeout code seems to fail for wrapping cases
19:37:03Bob_CJust to clarify, the definitions in this proposed mmc-defs.h file are taken from the SD/MMC spec, so should not be target specific, and only relate to SD/MMC protocol
19:38:49 Join Hillshum_ [0] (n=hillshum@75-165-234-241.slkc.qwest.net)
19:38:58 Quit Hillshum_ (Read error: 54 (Connection reset by peer))
19:39:28domonokymaybe we should put it as mmc-defs.h into firmware/export, so other targets can use it as well in the future.
19:40:35Bob_CIf that makes sense within the rockbox architecture, then it sounds like the simplest solution
19:40:52 Join Hillshum_ [0] (n=hillshum@75-165-234-241.slkc.qwest.net)
19:40:54domonokyBob_C: about the sdi-defs.h file. is this generic to s3c2410 ?
19:41:08amiconnTheSeven: The solution for switched channels on the earlier ipods was simply to send packed samples (L and R in one 32 bit word) instead of sending L and R individually
19:41:11domonokyor even generic to s3c2440 ?
19:41:20amiconnThis way losing a word doesn't cause switched channels
19:41:43 Quit Hillshum (Nick collision from services.)
19:41:49 Nick Hillshum_ is now known as Hillshum (n=hillshum@75-165-234-241.slkc.qwest.net)
19:41:51amiconn...and it also means fewer transfers, which translates to half the FIQ rate
19:42:10 Join Hillshum_ [0] (n=hillshum@75-165-234-241.slkc.qwest.net)
19:42:12Bob_Csdi-defs.g, or regs-sdi.h?
19:42:18gevaertsdomonoky, Bob_C: is there overlap with firmware/export/sd.h?
19:42:31*TheSeven wonders why we're losing a word on pausing at all
19:42:34Casainhome thinks
19:43:05domonokyBob_C: i meant regs-sdi.h
19:43:07Casainhome thinks that others targets already use SD Card, like Sansas... where is that definitions for that target?
19:43:50Bob_Cgevaerts: yes there is some
19:44:00domonokyBob_C: similar questions about some of the DMA definitions. if they are generic, they should go into the soc header.
19:44:20gevaertsalso, IIRC the various sd drivers share a .c file somewhere
19:45:02Bob_Cregs-sdi is specific to 2410/2440, the 2440 SD registers are virtually the same as 2410 with a couple of exceptions
19:45:38Bob_Cthe soc header being s3c2440.h ?
19:45:49domonokyBob_C: so i think the contents of regs-sdi.h should go into s3c2440.h
19:46:01 Join Strife89 [0] (n=Strife89@168.16.237.233)
19:46:11domonokyBob_C: yes
19:47:17domonokysd.h looks like the right place to merge mmc.h into. But then its contents should get renamed to SD_xyz
19:47:33Bob_CI have no problem with that, is the GPL licence on these files compatible with our usage?
19:48:27*domonoky takes a look about the licences...
19:50:46domonokymmc.h only says GPL (no version) so that seems fine. regs-sdi.h is "GPLv2 only" and rockbox is "GPLv2 and later" so thats a small clash. But dont know how important that is.
19:51:12 Quit Hillshum_ (Remote closed the connection)
19:51:52 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
19:52:37domonokyif we move sdi-regs.h into s3c2440.h, we also should rename the contents, so i think the licence is neglible for this, as we essentialy rewrite all those defines.
19:52:49TheSevenlinuxstb: IIRC rolo did work before? did somebody break it?
19:53:45bertrikTheSeven, I'll have a look at that register on my samsung yp-s3
19:54:01TheSeventhanks
19:54:04bertrikIs it really a separate register or could it be some alias of another timer register?
19:54:22TheSevenit doesn't behave like the other timers
19:54:33TheSevenit seems to be a 64bit timer represented as 2x32bit regs
19:54:50 Quit Horscht (Client Quit)
19:55:00 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
19:56:11 Quit Horscht (Client Quit)
19:56:31 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
19:56:47domonokyBob_C: so should i try to make this proposed modifications to your patch and commit it, or do you want todo the rework ?
19:57:03 Quit Horscht (Client Quit)
19:57:24Bob_CSure, I can do it
19:57:41 Quit evilnick ("Page closed")
19:58:14TheSevenamiconn: how do you send those packed samples?
19:58:22TheSevenwouldn't that screw lrck?
19:58:44domonokygood. i will continue to review your patch, trying to find more things to rework :-)
19:59:00Bob_CTo recap, I'll move SD/MMC defs into sd.h, and regs-sdi and dma defines into s3c2440.h
19:59:48domonokyBob_C: yes
19:59:50 Quit robin0800 (Read error: 110 (Connection timed out))
20:00
20:03:30amiconnTheSeven: Both the codec and the PP I2S have a mode that supports packed samples
20:03:39domonokyBob_C: from a first look, the rest of the patch looks fine. And your uart driver makes want todo logf-over-serial somewhere in the future :-)
20:03:47 Join bertrik_ [0] (n=bertrik@ip117-49-211-87.adsl2.static.versatel.nl)
20:04:43*TheSeven will try to fool the i2s controller
20:05:05 Join panni_ [0] (i=hannes@ip-95-222-21-143.unitymediagroup.de)
20:05:16Bob_Cok, i will work on those changes and do a new patch.
20:05:47 Part Dgby714
20:05:52 Join Dgby714 [0] (n=Dgby714@pool-173-78-91-222.tampfl.fios.verizon.net)
20:05:58domonokyBob_C: great, thanks for your work.
20:06:52 Quit Moiman (Read error: 60 (Operation timed out))
20:07:31TheSevendoes somebody have a datasheet of the wm8975?
20:10:24bertrik_meh, something changed in s5l8700 DRAM sections, can't compile the yp-s3 bootloader
20:12:05pixelmaJdGordon|: the CustomWPS tables break at the | character. It was the column seperator before but it seems TWiki was less picky when used without spaces right before and after
20:13:14JdGordon|ah ok
20:16:30bertrik_TheSeven, 0x3c700084 doesn't move on my samsung yp-s3, it shows value 1
20:16:47TheSevenso this is an 8701 feature...
20:18:34bertrik_I do see something moving at 0x3c700094 ! :) Could be another counter
20:19:07TheSeven...or a wrap of something
20:19:09bertrik_(and it's not an alias of one of the other counters)
20:19:26TheSevenhow fast is it moving?
20:20:06bertrik_"not very fast"
20:20:24bertrik_it looks like the inverse of counter B value
20:21:01TheSevenamiconn: do you know where they set up the codec appropriately? it seems to be the same codec, but i can't find the code doing that
20:22:53bertrik_TheSeven, oh no it's not
20:24:35bertrik_I'm getting curious how many more timers there are
20:25:58 Quit JackWinter (Read error: 104 (Connection reset by peer))
20:30:45 Join JackWinter [0] (n=jack@vodsl-10804.vo.lu)
20:31:30 Join evilnick [0] (i=0c140464@rockbox/staff/evilnick)
20:34:33 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
20:35:03bertrik_TheSeven, treating it like a "TIMER E" influences the TIMER A setting
20:35:25 Quit Horscht (Client Quit)
20:36:20TheSevensounds like some kind of wrap
20:37:18 Join mirak [0] (n=mirak@85-171-108-41.rev.numericable.fr)
20:37:38CasainhoJdGordon|: will you use just SD Card on Mini2440? or you plan to also add an HDD?
20:41:09bertrik_TheSeven, yes indeed, I was seeing different values between TIMER A and E but that is probably due to a delay in reading out the TxCNT register
20:41:19 Join liar [0] (n=liar@83.175.83.185)
20:41:54 Join antil33t [0] (n=Mudkips@119.224.12.185)
20:42:00JdGordon|Casainho: SD to start with... my pie-in-the-sky ulitmate goal is using the network for storage :)
20:45:37domonokyJdGordon|: if you volunteer to write a network fs driver (smb, nfs or alike), i volunteer to port lwip as a networkstack :-)
20:46:01JdGordon|the uipstack is the easy bit isnt it?
20:46:07JdGordon|relativly easy anyway
20:46:56 Join dfkt_ [0] (n=dfkt@unaffiliated/dfkt)
20:47:35domonokylwip should be pretty easy to port (i already ported it to vxworks for work), so only the network driver is real work. and of course the applications to use the net :-)
20:48:22JdGordon|nfs is presumably easier than samba?
20:48:34ender`definitely faster
20:48:44ender`as for easier, it depends
20:49:31Casainhohey guys, there are network radio stations, no? −− since some people likes to listen radio, a Rockbox device to listen radio stations looks a nice idea to me :-)
20:50:06***Saving seen data "./dancer.seen"
20:50:19Casainhoand this Mini2440 board would be great for a home web radio :-)
20:50:24JdGordon|rockbox is NOT linux!
20:50:25JdGordon|:p
20:50:46domonokyCasainho: that is a good idea, shouldnt be too hard when network runs.. :-)
20:51:46 Join DerPapst [0] (n=DerPapst@p4FE8F237.dip.t-dialin.net)
20:52:11CIA-85New commit by 03theseven (r23308): iPod Nano 2G: "Fix" the latest PCM issues, channel swap and vinyl popping. The latter is a revert of the cache coherency patch for now, we need to ...
20:52:40Casainhodomonoky: I mean a device like this one: http://www.watterott.net/projects/webradio-arm
20:52:41Bob_CI'd like to replace my MP101 which never worked properly. A mini2440 would fit neatly inside the MP101 box.
20:53:44CIA-85New commit by 03theseven (r23309): Add iPod Nano 2G µsec timer, and use it in the clickwheel driver.
20:54:05domonokythe integration of the streamed, encoded data into our playback system will get tricky :-)
20:54:22Casainhohave a home radio that plays normal web stations or our particular collection of music stored on a multimedia server :-)
20:55:31 Join bubsy [0] (n=bubsy@94.139.72.137)
20:57:17Bob_CI can see how a network drive would work, not so sure about streaming.
20:58:47gevaertsBob_C: streaming is easy. Rockbox already knows how to skip if there isn't enough data buffered, and that's the hard part :)
20:59:24TheSevengevaerts: then do it :-)
20:59:44Bob_CSee what I did there ;)
21:00
21:00:57*TheSeven is still looking for a better solution to this irq latency issue
21:00:57 Quit dfkt (Nick collision from services.)
21:01:05 Nick dfkt_ is now known as dfkt (n=dfkt@unaffiliated/dfkt)
21:01:12gevaertsit's all planned! someone does this network driver, then domonoky does the lwip stack, and I also have a promise from someone else to port curl as soon as lwip is there :)
21:01:49Bob_CI do network drivers for work, so that should be easy ...
21:03:21TheSevenare there objections agains an ASM FIQ handler for the PCM stuff?
21:05:21*domonoky thinks other targets also use (or used) FIQ handlers for PCM.. and if it needs ASM for speed, why not ?
21:06:36TheSevendomonoky: well, ASM allows me to set up the DMA controller within 3 instructions after the FIQ handler gets run
21:06:49TheSeventhe issue is just that the pcmbuf doesn't seem to like that
21:07:14 Join pamaury [0] (n=pamaury@88-123-144-142.rev.libertysurf.net)
21:07:33TheSevenit's overwriting buffers it's thinking that have already been played (but in fact only be asked for in order to pass them to the DMA when the next FIQ comes in) too early
21:07:40TheSevenso this needs a little hack
21:08:38 Join peter-b [0] (n=peter_b@mnch-5d8619be.pool.mediaWays.net)
21:11:38 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
21:16:45 Quit peter__b (Read error: 145 (Connection timed out))
21:17:39*TheSeven glares at the red
21:18:58TheSevenah, it's a sim, that explains quite a few things
21:23:44 Join jakethemighty [0] (n=jake@CPE000c6e43a872-CM001ac3101306.cpe.net.cable.rogers.com)
21:24:06jakethemightyHello #rockbox.
21:27:30jakethemightyI'm having an issue. Some files (mp3s) aren't showing up in the database. They run fine on my computer, and on other mp3 players, and they play fine if I play them from the files menu, but they won't show up in the database. I've tried rebuilding / initializing, with no change.
21:28:10CIA-85New commit by 03theseven (r23310): Fix red
21:28:16 Join shotofadds [0] (n=rob@80-44-101-34.dynamic.dsl.as9105.com)
21:28:18evilnickjakethemighty: Have they got album art in the tags? Or very large tags (such as Comments)?
21:28:30*TheSeven wonders what to do about the yellow
21:28:55jakethemightyTheSeven: I don't think so, but I will check.
21:29:06TheSevenhuh?
21:29:10jakethemightyErr, I mean evilnick
21:29:29jakethemightyRead the wron username :P
21:29:40evilnickjakethemighty: Which folder are they in, on your player?
21:30:22jakethemightyevilnick: main dir /artistname/album/
21:30:41*domonoky has a nearly working touchscreen driver for mini2440 :-) (it just need cleanup, calibration and some filtering)
21:31:27evilnickjakethemighty: And your other albums show up the database fine? So that rules out having a "database.ignore" file there.
21:32:49shotofaddsdomonoky: be aware that the touchscreen calibration setting isn't implemented quite right and the settings do not get persisted properly
21:32:59shotofaddsit caught me out a few times :/
21:33:12jakethemightyevilnick Oh yeah, everything else is fine. I also checked the tags, and they seem normal to me
21:34:58evilnickjakethemighty: Hmmm, if you can narrow it down to one album in particular then it might be helpful to upload a sample of one track somewhere so that the clever people here can check it.
21:35:10 Join dmb [0] (n=Dmb@unaffiliated/dmb)
21:35:50domonokyshotofadds: thanks for reminding me, for now i will try to get a basic calibration done in the driver itself, because the touch seems to be mounted sideways (100,100 is left-bottom)
21:35:59evilnickjakethemighty: Actually, have you tried checking for these albums in <Untagged>? It might be that the wps you're using falls back to displaying the folder names so well that it seems to be reading the tags
21:36:38CIA-85New commit by 03theseven (r23311): Fix yellow
21:37:15shotofaddsdomonoky: oh, that's weird - the driver will need to do that correction. I think the calibration screen is using global_settings wrongly, I'll see if I can fix it later.
21:39:14jakethemightyevilnick: not in untagged
21:39:27shotofaddslinuxstb: following on from your conversation with tomers earlier, you're quite right that a mktccboot binary and bootloader are indeed missing for the D2. I thought I already had those linked from the wiki :/
21:40:06shotofaddsbut in any case you can't seem to brick a D2 in any way, shape or form, so "known good" version aren't so critical as other targets
21:41:15 Quit Bob_C ("ChatZilla 0.9.85 [Firefox 3.5.3/20090824101458]")
21:41:38*shotofadds has a win32 mktccboot ready but no Swede to put it on the download server...
21:42:56 Quit shotofadds (Read error: 104 (Connection reset by peer))
21:43:02 Join kugel [0] (n=kugel@rockbox/developer/kugel)
21:43:22kugelTheSeven: I tried rolo on the nano2g I had today. it worked
21:43:26jakethemightyevilnick: does rockbox have an issue with any filename conventions?
21:43:50TheSevenkugel: it just locks up for me. reproducibly, and since lots of revisions.
21:44:10*TheSeven checks for related local modifications
21:44:19 Quit bubsy ("Party time, EXCELLENT! *wring wring wring!!* Party on, Wayne! Party on, Garth!")
21:45:08*TheSeven can't find anything related
21:46:20evilnickjakethemighty: Not at all, your files can be called anything (provided it's valid under FAT32)
21:46:49evilnickBut the database will only store tags, and doesn't have the "fallback" to folder/filenames that most wpses have.
21:47:37jakethemightyevilnick: Yeah, that can't be the problem anyway. Other albums in the same dir have the same conventions
21:48:33evilnickjakethemighty: What format are the files in? And what format are the tags?
21:48:53evilnicks/format/codec/
21:49:22jakethemightyWell, the files are just mp3, and I know nothing about tags
21:49:24*domonoky s Touchscreen driver now works nice, after i calibrated from the UI, luckily i also have normal buttons to navigate there :-) But i think we need some way to start in the calibrate screen, if the touch isnt calibrated.
21:49:30 Quit jfc (Read error: 104 (Connection reset by peer))
21:49:38 Join shotofadds [0] (n=rob@rockbox/developer/shotofadds)
21:50:49evilnickjakethemighty: Some desktop apps added ape tags (rather than id3 tags) to mp3 files which Rockbox won't read, so uploading one file would help
21:52:00kugelamiconn: have you seen my findings about the asm code the F/X has it his lcd driver?
21:52:43*kugel once again pokes bertrik about the dbop patch
21:52:51kugelwould you mind committing it as is?
21:53:15jakethemightyevilnick: I just downloaded a id3 tag editor, and it appears these files have no tags... Queer
21:54:20evilnickjakethemighty: So we've solved that problem then :)
21:56:04TheSevenI just did some measurements...
21:57:34kugelTheSeven: I used plain SVN from the current build page when I tried rolo (I tried that since I saw your notice from a few days)
22:00
22:00:30TheSeveniPod Nano 2G output level (1kHz sine, full PCM amplitude) at 0dB is 732mV RMS (1035mV amplitude, that's nice calibration!). 1dB still has no visible clipping and is at 821mV RMS (1161mV amplitude). Maximum (at +6dB with massive clipping) is 1072mV RMS (~1170mV amplitude), -10dB = 231mV RMS (327mV amplitude), -20dB = 71mV RMS (100mV amplitude)
22:01:11TheSevenwho was that guy who said that portable DAPs don't manage to go above 150mV peak?
22:01:19 Join jfc [0] (n=john@dpc6682208002.direcpc.com)
22:02:24kugelnot bad
22:02:35jakethemightyThanks evilnick. Pretty sure that'll work. Waiting for the db to initialize now
22:02:43TheSeventhe funny thing is that the scale seems to be off by factor 2
22:02:47 Quit jakethemighty ("Lost terminal")
22:02:56TheSeventhe -20dB setting is in fact -10dB
22:03:08kugelindeed
22:03:41*kugel suspects the scale isn't very accurate on targets too
22:04:15amiconnTheSeven: It's not off by a factor of 2...
22:04:46amiconnFor voltages, dB values are 20*log(V2/V1)
22:04:49TheSevenah right, forgot that dB are referring to the power, not the voltage
22:05:02amiconnyup
22:05:33*kugel was tricked by TheSeven !
22:05:45TheSeventhen I call that a nicely calibrated scale, and that +1dB still does not have clipping is nice.
22:07:21 Nick fxb__ is now known as fxb (n=felixbru@85.214.97.64)
22:07:47bertrikkugel, I think I can simplify it a bit more
22:08:17TheSevenAlexP, saratoga, topik: ping regarding this
22:08:19kugeldo it please, lcd corruption is getting annoying with custom statusbar :)
22:08:32 Quit Horscht ("Verlassend")
22:08:35AlexPTheSeven: eh?
22:08:51kugelI really wouldn't mind doing optimizing after it's in SVN though
22:09:20TheSevenAlexP: the nano2g seems to be a target that has an output that's nicely calibrated to an 1V scale
22:09:37TheSeventhe dB you see in the menu are true dBV
22:09:51AlexPTheSeven: Ah right. I don't remember saying anything other than 0 dB should be line-out level :)
22:10:14TheSevenyes, and this is the only player i know where this is really true
22:10:25AlexPcoolio :)
22:15:33TheSevenkugel: did you try rolo with an usb-enabled build?
22:15:35kugelno
22:15:45TheSeventhat may be the difference then
22:19:20kugelbertrik: any eta on when you might have a look? what ideas do you have for simplifying?
22:21:45CIA-85New commit by 03theseven (r23312): storage_flush() should be done before disabling interrupts in ROLO.
22:23:03TheSeven...but it still doesn't work
22:23:43 Quit thegeek (Read error: 131 (Connection reset by peer))
22:25:42JdGordon|domonoky: yes... got any ideas? maybe we force it to run on startup if some global_status setting isnt set?
22:26:20JdGordon|shotofadds: hey, do you reads logs? there is a forum thread which is exactly what I said... customisable buttons request using hold on every target
22:26:48domonokyJdGordon|: yes, something like this.. but i dont know this part of the code.. so no real ideas :-)
22:26:48TheSeven...it did work, just took longer than expected
22:26:49 Join thegeek [0] (n=nnscript@s168c.studby.ntnu.no)
22:27:02shotofaddsJdGordon|: yes i saw that and replied. I still don't see the problem though
22:28:20bertrikkugel, I think I can have a patch by friday evening
22:29:34JdGordon|kugel, others... any chance we can get rid of the word "custom" in the strings?
22:29:52CIA-85New commit by 03theseven (r23313): Add a "Flushing storage buffers" message to ROLO if it needs to do that.
22:29:53Casainhodomonoky: will you update your code to Rockbox SVN? −− I would love to test it ;-)
22:30:31kugelJdGordon|: yea, I saw you wanta the one statusbar option to be named statusbar
22:30:49domonokyCasainho: we still dont have all code from hg in SVN, so i have to wait :-)
22:30:53kugelwant to*
22:31:31shotofaddskugel: it makes sense to remove the "custom". WPS isn't called "custom WPS" is it? ;)
22:31:45 Join froggyman [0] (n=sopgenor@pool-72-69-220-194.chi01.dsl-w.verizon.net)
22:32:34Casainhodomonoky: ok, but can you post a patch on FS?
22:32:37kugelah you mean the custom in "Custom Statusbar" that leads to the filebrowser?
22:32:44domonokyCasainho: but i could send you a patch against hg if you want.
22:32:54JdGordon|kugel: yes
22:33:01Casainhodomonoky: ok, send by e-mail then.
22:33:06kugelwell, there's an existing statusbar setting that is still important and used, I don't want confusion
22:33:22JdGordon|also, the themes should get a submenu so all the remote options are together instead of being mixed in with the main lcd
22:34:02JdGordon|and maybe all of them go into a "skins" submenu under themes once a few more items get added
22:34:08kugelit could be statusbar skin yes, I don't really care
22:34:11*JdGordon| is giong for a sub-menu depth of 11!
22:34:33kugelwfms first please :)
22:35:25JdGordon|anyone want to volanteer to clean up the hardware and actual radio control part of recorder/radio.c?
22:35:55kugelno thanks :p
22:36:30JdGordon|MVC ftw!
22:36:38JdGordon|or.. thats where it needs to go anyway
22:36:50TheSevenETOOMANYTLAS
22:37:29domonokyCasainho: http://pastebin.com/m4263d918
22:37:42JdGordon|model-view-controller :)
22:37:58Casainhothanks ;-)
22:38:10domonokyCasainho: it still needs improvements, but works (if calibrated)
22:39:01Casainhodomonoky: how to calibrate?
22:39:48Casainhodomonoky: and by the way, I just need to build firmware, or the bootloader also?
22:40:02domonokyCasainho: use the normal buttons to navigate to: settings->global settings->display-touch->calibrate.. Main binary is enough
22:40:17kugeldomonoky: got the touchscreen to work?
22:41:06domonokykugel: yes, preliminary driver works fine. still needs filtering against spurious values and accidental double clicks and some code clean-up. but it works :-)
22:41:40kugelcool :>
22:42:56domonokywas pretty easy, after i found out that the normal adc reading interfered with the touch readings and i just disabled them :-)
22:46:01Casainhodomonoky: how to I aplly your patch file? −− I am following the wiki instructions and I can't...
22:46:46domonokyoh, may hg checkout is probably not uptodate..
22:47:15Casainhodomonoky: do you want hg write access?
22:47:48domonokyCasainho: not really needed. i want it in svn :-)
22:48:10Casainhodomonoky: ok. So, how can I test your code now?
22:49:02 Join petur [0] (n=peter@84.198.249.178)
22:49:24domonokyah, my hg checkout is 7 days old. dont know if it still works if i update :-) maybe you could checkout the older version, and then apply my patch ? (or just wait a few days till we have everthing in svn)
22:49:42 Quit Grahack (Client Quit)
22:49:56Casainhook, but, how do you know your hg revision version?
22:50:07***Saving seen data "./dancer.seen"
22:51:23 Quit Yadda ("CGI:IRC (Ping timeout)")
22:52:23Casainhodomonoky: can you tell me your revision number on hg?
22:53:23 Quit pamaury ("exit(*(int *)0 / 0);")
22:53:39domonokywith hg tip you can get your current changeset number. mine is: 17:7374ef589985
22:54:03domonokyor is that the server ?
22:54:39Casainhodomonoky: oh, I am not really expert on hg :-) −− but maybe discover here: http://lyre.hg.sourceforge.net/hgweb/lyre/lyre/summary
22:55:45 Quit bluebrother (Nick collision from services.)
22:55:48 Join bluebroth3r [0] (n=dom@rockbox/developer/bluebrother)
22:56:44TheSevenhttp://pastie.org/664176
22:57:16TheSevenlet's try which of them we can shut off
22:57:31TheSeven...and undervolt of course
22:58:53TheSevenkilled, LDO2, still running
22:59:11 Join bubsy [0] (n=bubsy@94.139.72.137)
22:59:32domonokyCasainho: from what i can read you can just clone the hg directory again with -r <changeset nummer above> as additional option to get the same revision as me.
23:00
23:00:30TheSevenLDO6 refuses to power down
23:02:58Casainhodomonoky: ok. Looks like is "changeset 177374ef589985"
23:04:15 Quit petur ("Zzzzz")
23:06:32TheSevenhm. the PMU is behaving weird, not performing the writes I send...
23:10:14Casainhodomonoky: you know that Bob already found a good plastic case from Hammond? http://www.hammondmfg.com/scpg.htm
23:11:08Casainhodomonoky: I would love to have a 3D printer to make my own box :-)
23:11:46*domonoky moves Casainho to #rockbox-community for off-topic talk :-)
23:12:07Casainhodomonoky: ok. sorry
23:16:31 Quit esperegu (Read error: 60 (Operation timed out))
23:18:31 Join Stephen_ [0] (n=S@86-40-191-208-dynamic.b-ras2.srl.dublin.eircom.net)
23:18:50 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
23:20:12*TheSeven wonders why rolo still fails occasionally
23:21:37 Join LambdaCalculus37 [0] (n=LambdaCa@rockbox/staff/LambdaCalculus37)
23:28:33 Join BdN3504 [0] (n=5ce227c2@giant.haxx.se)
23:29:51BdN3504hi guys. I am dealing with the vmware image of rockbox (again). i got a bit annoyed that i have to use vmwaretoolbox for simple copy and paste...
23:30:19TheSevenLDO2 seems to be clickwheel-related
23:30:33BdN3504and that i can't use special characters when i change the keymap with xf86cfg
23:30:58BdN3504so i figured, why not simply use putty and this wokrs perfect.
23:31:05BdN3504can i add this to the wiki?
23:31:24BdN3504works perfectly
23:32:11 Join dfkt_ [0] (i=dfkt@unaffiliated/dfkt)
23:32:36 Quit Strife89 ("Going home.")
23:32:46 Quit dfkt (Nick collision from services.)
23:32:49 Nick dfkt_ is now known as dfkt (i=dfkt@unaffiliated/dfkt)
23:34:37TheSevenLDO1 off => deadlock
23:35:32 Quit kugel (Remote closed the connection)
23:35:53 Quit flydutch ("/* empty */")
23:36:33TheSevenLDO3 = LCD
23:37:08 Quit evilnick ("Page closed")
23:37:13 Join Ubuntuxer [0] (n=johannes@dslb-094-221-092-095.pools.arcor-ip.net)
23:39:04TheSevenLDO4 = probably codec
23:39:07TheSevenLDO5 = NAND
23:39:14TheSevenLDO6 doesn't even power off
23:41:33 Quit mirak ("Ex-Chat")
23:42:24 Quit faemir ("Leaving")
23:43:22 Join faemir [0] (n=faemir@78.33.109.163)
23:43:53TheSevenlet's try undervolting
23:45:22 Quit BdN3504 ("CGI:IRC (EOF)")
23:46:17TheSevenwow! 950mV instead of 1075mV at full clock!
23:46:26TheSevenlet's see how stable that is
23:46:35 Quit bubsy ("Party time, EXCELLENT! *wring wring wring!!* Party on, Wayne! Party on, Garth!")
23:50:50UbuntuxerHi, I can't edit any wikipage of Rockbox. I get always the error message "access denied". How do I get permission ?
23:51:48LambdaCalculus37Ubuntuxer: By asking here. :)
23:52:11Ubuntuxerah,;-)
23:53:29mc2739Ubuntuxer: try now
23:53:51UbuntuxerPlease can you change the rights of my account ? (JohannesSchwarz)
23:54:18mc2739Ubuntuxer: I just did
23:54:22 Join Casainho_ [0] (n=chatzill@87.196.39.193)
23:54:28Ubuntuxerok, thanks
23:54:41mc2739you're welcome
23:55:42JdGordon|shotofadds: once again, I'm not entirely against the patch... but start a thread on the dev ml....
23:56:18 Join arohtar [0] (n=faemir@78.33.109.163)
23:57:30 Quit dys```` (Remote closed the connection)

Previous day | Next day