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

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

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

#rockbox log for 2011-01-10

00:02:07 Join factor [0] (~factor@
00:05:01 Part b0hoon ("GTG, bye.")
00:13:03soapsorry, I've been real busy today. Am I needed or not?
00:13:18soapLooks like not, is that correct?
00:16:21 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
00:22:48 Quit dfkt (Quit: -= SysReset 2.53=- Sic gorgiamus allos subjectatos nunc.)
00:23:17[Saint]soap: If you mean testing that LCD patch for Buschel, then if you could it would be cool...but don't stress it.
00:23:25[Saint]I can do it in a few hours.
00:27:37***Saving seen data "./dancer.seen"
00:31:25 Join krabador [0] (
00:31:38 Quit crwl (Read error: Operation timed out)
00:31:52 Quit maraz (Ping timeout: 250 seconds)
00:31:56 Join JdGordon1 [0] (
00:32:53 Quit aevin (Ping timeout: 272 seconds)
00:33:18 Join aevin [0] (eivindsy@unaffiliated/aevin)
00:41:45 Join crwl [0] (
00:42:21 Quit Buschel (Ping timeout: 240 seconds)
00:42:38 Join maraz [0] (
00:42:57 Quit Keripo (Quit: Leaving.)
00:45:06 Quit sasquatch (Quit: WeeChat 0.3.2)
00:45:31 Join sasquatch [0] (
00:51:16 Quit ender` (Quit: Debugging is twice as hard as writing the code in the first place. Therefore, if you write the code as cleverly as possible, you are, by definition, not smart enough to debug it. -- Brian W. Kernighan)
01:05:10mudd1Why did I empty my "Recent Bookmarks" list by calling bookmark_autobookmark()?
01:06:02mudd1or to put it differently: how *do* I just bookmark the current state to the recent bookmarks?
01:06:44*mudd1 definitely hasn't got his head around the whole bookmarking concept of Rockbox :/
01:12:04 Join Keripo [0] (
01:15:28 Quit pamaury (Remote host closed the connection)
01:17:51 Quit BHSPitMonkey (Ping timeout: 240 seconds)
01:27:51 Part sideral
01:28:40mudd1got to go to sleep, will ask again tomorrow
01:28:44mudd1good night
01:28:57 Quit mudd1 (Quit: Ex-Chat)
01:30:50 Quit simon_ (Quit: leaving)
01:40:21 Quit t0rc (Ping timeout: 264 seconds)
01:41:32 Join Derreck [0] (
01:42:11 Quit Derreck (Client Quit)
01:42:25 Join Derreck [0] (
01:43:54 Quit Derreck (Client Quit)
01:44:07 Join Derreck [0] (
01:44:28 Quit kugel (Remote host closed the connection)
01:56:45 Quit krabador (Ping timeout: 240 seconds)
02:08:00 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
02:09:12 Join krabador [0] (
02:12:24 Quit t0rc (Ping timeout: 260 seconds)
02:26:51 Quit Keripo (Quit: Leaving.)
02:27:17 Join Keripo [0] (
02:27:39***Saving seen data "./dancer.seen"
02:30:28 Join GuySoft [0] (
02:31:25GuySofthi all, i seem to be getting an unreported bug - sansa fuze device keeps freezing when transferring files over USB - How should i report the problem? is there any output that would help detect the cause of this?
02:36:25kisak is it a fuze v1 or a fuze v2, and what version of rockbox?
02:36:28 Quit Keripo (Quit: Leaving.)
02:36:52 Join Keripo [0] (
02:42:04 Quit user890104 ()
02:44:31 Quit DerPapst (Quit: Leaving.)
02:44:31 Quit DerPapst1 (Quit: Leaving.)
02:57:46 Quit GeekShadow (Quit: The cake is a lie !)
03:08:43 Quit krabador (Read error: Connection reset by peer)
03:10:13 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
03:25:34 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
03:32:20 Join simon_ [0] (
03:36:00 Quit simon_ (Client Quit)
03:37:37 Join krabador [0] (
03:39:29 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
03:42:11 Quit krabador (Client Quit)
03:44:01 Join krabador [0] (
03:53:41CIA-7New commit by rmenes (r29019): Add progress bar graphic for CabbieV2 for 128x160x16 LCDs. ...
03:56:47CIA-7r29019 build result: All green
03:56:55LambdaCalculus37Whew! :)
03:58:35GuySoftkisak, fuze V1 , sorry didn't see the reply because you didnt mention my user
03:58:57GuySoftkisak, the recent version 3.7,1
03:59:36LambdaCalculus37And now the SA9200 has plugins and CabbieV2. :)
04:04:34LambdaCalculus37Although we need artwork for bubbles and rockblox; the 128x128 artwork is being used currently and garbage is being displayed in the unused portion of the screen.
04:11:26 Join Barahir [0] (
04:14:31 Quit Barahir_ (Read error: Operation timed out)
04:16:22 Quit krabador (Ping timeout: 272 seconds)
04:22:47 Join Strife89_ [0] (
04:23:53 Quit BHSPitMonkey (Remote host closed the connection)
04:26:09 Nick Strife89_ is now known as Strife89 (
04:27:40***Saving seen data "./dancer.seen"
04:30:30LambdaCalculus37Anyone have any objections to promoting the SA9200 to Unstable?
04:39:57 Quit amiconn (Disconnected by services)
04:39:58 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:39:58 Quit pixelma (Disconnected by services)
04:40:00 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:40:02 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:40:15 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:41:02 Quit Judas_PhD (Quit: This is a quitting message)
04:53:04 Quit TheSeven (Ping timeout: 265 seconds)
04:57:48 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
05:03:54 Join Rob2223 [0] (
05:06:38 Quit factor (Ping timeout: 260 seconds)
05:07:06 Quit Rob2222 (Ping timeout: 240 seconds)
05:09:34 Quit Derreck (Read error: Operation timed out)
05:13:16 Quit Rob2223 (Ping timeout: 264 seconds)
05:21:24 Quit MethoS- (Read error: Connection reset by peer)
05:24:19 Join Horscht [0] (
05:24:20 Quit Horscht (Changing host)
05:24:20 Join Horscht [0] (~Horscht@xbmc/user/horscht)
05:26:28 Quit Horschti (Ping timeout: 264 seconds)
05:26:48 Join Judas_PhD [0] (
05:28:41 Join factor [0] (~factor@
05:34:19 Quit Horscht (Quit: Verlassend)
05:40:37 Quit LambdaCalculus37 (Quit: ZZZZZZZZzzzzzzzzzz.....)
05:42:23 Join Keripo1 [0] (
05:43:29jhMikeSLambdaCalculus37: (for the logs) I'm working on the graphics right now. Bubbles is done.
05:45:08 Quit Keripo (Ping timeout: 260 seconds)
05:51:35 Quit JesusFreak316 (Ping timeout: 240 seconds)
06:27:42 Quit panni_ (Quit: ( :: NoNameScript 3.81 :: ))
06:27:43***Saving seen data "./dancer.seen"
06:35:28 Quit Zarggg_ (Quit: Zarggg_)
06:39:17 Quit GuySoft (Remote host closed the connection)
06:43:19 Quit Strife89 (Quit: Bed! ZZZZZZZZZzzzzzzzzzzzzzz ....)
06:48:54 Join Rob2222 [0] (
07:09:54 Join Quex01 [0] (
07:18:43 Quit t0rc (Ping timeout: 276 seconds)
07:28:31 Join kugel [0] (~kugel@
07:28:32 Quit kugel (Changing host)
07:28:32 Join kugel [0] (~kugel@rockbox/developer/kugel)
07:29:02 Join Buschel [0] (
07:30:10 Quit Buschel (Client Quit)
07:33:20 Join Buschel [0] (
07:40:28 Quit Quex01 (Quit: Quex01)
07:45:54 Quit factor (Read error: Connection reset by peer)
07:48:38 Join factor [0] (~factor@
08:00:08 Quit JdGordon1 (Quit: leaving)
08:00:38 Join stoffel [0] (
08:05:20 Join kevku [0] (~kevku@2001:7d0:0:f000::135d)
08:08:07 Quit Buschel (Quit: ChatZilla 0.9.86 [Firefox 3.6.13/20101203075014])
08:09:43 Quit Judas_PhD (Quit: This is a quitting message)
08:25:37 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
08:27:45***Saving seen data "./dancer.seen"
08:30:01 Join T44 [0] (
08:31:26 Quit TheSeven (Ping timeout: 255 seconds)
08:32:32CIA-7New commit by jethead71 (r29020): Add backgrounds for 128x160 displays to bubbles and rockblox. Set the coordinates in the code.
08:32:35 Quit Topy (Ping timeout: 240 seconds)
08:35:05CIA-7r29020 build result: All green
08:35:10 Join factor_ [0] (~factor@
08:35:25 Quit factor (Ping timeout: 272 seconds)
08:40:34[Saint]Hmmm...database info in Debug menu. xxxxxx/yyyyyy
08:40:43Stummigood morning
08:40:52[Saint]yyyyy seems to be maximum ram available, but what is xxxxx?
08:40:58[Saint]used or remaining?
08:42:23 Join logiclost [0] (~lostlogic@
08:46:11 Quit lostlogic (Ping timeout: 272 seconds)
08:47:50 Quit logiclost (Quit: leaving)
08:53:22 Join ender` [0] (
08:53:59 Join Bagder [0] (
08:56:43 Quit bertrik (Ping timeout: 276 seconds)
09:01:39 Join lostlogic [0] (~lostlogic@rockbox/developer/lostlogic)
09:02:29 Quit bluebroth3r (Ping timeout: 265 seconds)
09:04:04 Join bluebrother [0] (
09:04:04 Quit bluebrother (Changing host)
09:04:04 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
09:06:00 Join AlexP_ [0] (~alex@rockbox/staff/AlexP)
09:09:04 Quit AlexP (Ping timeout: 276 seconds)
09:10:59 Join timccc [0] (~timccc@
09:11:02 Quit timccc1 (Read error: Connection reset by peer)
09:17:40 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
09:26:43 Quit lostlogic (Ping timeout: 255 seconds)
09:30:16 Quit kugel (Quit: Yaaic - Yet another Android IRC client -
09:45:27 Join petur [0] (d408b802@rockbox/developer/petur)
09:46:07 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
09:55:27*[Saint] wonders why his iPod Colors keep consistently screwing their file-systems up.
09:55:55[Saint]running chckdsk every day is getting somewhat tiring.
10:17:42 Join JdGord [0] (~jonno@
10:17:55 Join efyx [0] (
10:21:11 Join The_Pwny [0] (
10:27:49***Saving seen data "./dancer.seen"
10:28:03pixelmaZagor: (since you are here now) there was someone having problems with the wiki registration, probably because he registered with a nickname first and then with his real name. He also said he sent a mail, did you have time to look into it?
10:30:15pixelmaah, nice
10:34:26 Join pamaury [0] (
10:34:26 Quit pamaury (Changing host)
10:34:26 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
10:37:49 Quit Bagder (Changing host)
10:37:49 Join Bagder [0] (~danielx@rockbox/developer/bagder)
11:01:54 Join LinusN [0] (~linus@rockbox/developer/LinusN)
11:02:35 Quit parafin (Ping timeout: 260 seconds)
11:09:26 Join parafin [0] (
11:13:18 Join DerPapst [0] (
11:13:19 Join DerPapst1 [0] (
11:34:54 Join n1s [0] (~n1s@rockbox/developer/n1s)
11:37:36 Join cheesy [0] (
11:38:15 Join TheLemonMan [0] (
11:38:32cheesyhi, anyone help me here maybe?
11:39:05[Saint]only if you ask a question frst ;)
11:39:13cheesyhehe, mom
11:40:13cheesyi cannot select my sandisk sansa clip+ in rockbox utility because win7 doesnt show me the player as a drive
11:40:20cheesyonly my hdds
11:40:53cheesywhat do`? :D
11:41:14cheesywin7 shows it as a "portable player"
11:41:28 Join sideral [0] (~sideral@unaffiliated/sideral)
11:41:46gevaertsSet it to MSC
11:41:58gevaertsAs explained in the manual,
11:42:07cheesyok thx
11:43:54 Quit sideral (Remote host closed the connection)
11:44:34 Quit Keripo1 (Ping timeout: 276 seconds)
11:44:35 Join sideral [0] (~sideral@unaffiliated/sideral)
11:49:18sideralSaint: I've also had regular FS inconsistency recently again, on the ClipV2 w/ USB patch (FS #11664). I haven't had time to look into this with any depth yet, but apparently I can reproduce FS inconsistency (detected by Linux's fsck.vfat) directly after I umount the player (after having written some large files). This doesn't happen with the OF.
11:49:50[Saint]Same here.
11:49:54sideralCare to open an FS tracker to collect settings / revisions / patches?
11:50:10[Saint]I was going to once I got home, yes.
11:50:20[Saint]I will ping you once I have made it.
11:50:29sideralGreat, thanks!
11:50:59 Join lostlogic [0] (
11:50:59 Quit lostlogic (Changing host)
11:50:59 Join lostlogic [0] (~lostlogic@rockbox/developer/lostlogic)
11:54:10 Quit sideral (Remote host closed the connection)
11:55:14 Join sideral [0] (~sideral@unaffiliated/sideral)
12:03:58 Join dfkt [0] (dfkt@unaffiliated/dfkt)
12:04:25 Quit sideral (Remote host closed the connection)
12:05:14 Join sideral [0] (~sideral@unaffiliated/sideral)
12:14:34 Quit timccc (Ping timeout: 255 seconds)
12:22:12 Quit The_Pwny (Quit: Light travels faster then sound, which is why some people appear bright, until you hear them speak)
12:24:57 Join user890104 [0] (
12:27:51***Saving seen data "./dancer.seen"
12:32:15 Join timccc [0] (~timccc@
12:38:01 Join LambdaCalculus37 [0] (
12:38:01 Quit LambdaCalculus37 (Changing host)
12:38:01 Join LambdaCalculus37 [0] (~rmenes@rockbox/staff/LambdaCalculus37)
12:50:42LambdaCalculus37jhMikeS: (for the logs) On your SA9200, what do you see when you go to System > Debug > View HW Info? Mine says that there's a PP50222C in it.
12:51:50LambdaCalculus37That doesn't sound right; I know that the SA9200 has a PP5024B in it like the e200v1 does.
12:53:56pixelmaI believe there was something about this chip that lets it identify this way - just checked and my c200v1 says the same
12:54:50n1syeah the 5024 was a 5022 with a built in ams codec iirc
12:55:59LambdaCalculus37n1s: Ahh, that explains it. :)
12:58:30LambdaCalculus37jhMikeS: (also for the logs) Let me know if you're going to do any new bootloader work for the SA9200; I'm aiming to tag a v1.0 bootloader release Real Soon Now.
12:59:31 Quit LambdaCalculus37 (Quit: Fwump)
13:05:51 Quit kevku (Quit: KVIrc 4.0.2 Insomnia
13:12:28 Join Kitar|st [0] (
13:30:55gevaertsZagor: can you have a look at ?
13:32:46[Saint]should the topic of,26902.0.html not be changed?
13:33:02[Saint]AFAIK, the target doesn't meet the definition of unstable
13:33:19*Zagor was just offered a Creative Zen V
13:38:54Zagorgevaerts: sure, I guess we could do a more thorough check with −−version instead of merely 'which'.
13:39:08Zagorand we can check binutils too, just run "arm-elf-eabi-ar −−version" for example
13:39:23n1sgevaerts: i had a thought about that, should we make configure use the gcc-x.y.z to be sure we use the correct one or just tell people to get rid of the other toolchain first?
13:39:23gevaertsZagor: that would be best of course, but it requires someone who speaks fluent perl :)
13:40:03Bagderor someone who learns how to do it! =)
13:42:38gevaertsn1s: I think I prefer the plain unversioned name, but I can't really say why
13:42:50Bagderthat VP8870 is truly a monster
13:42:58Bagder2.5" disk and 800x480 lcd
13:43:34n1sgevaerts: then i prefer checking with −−version over checking for the versioned filename since it seems safer
13:44:23n1sdoes anyone want to do this or should i try to remember how to do perl? :)
13:45:04gevaertsFeel free!
13:45:07Zagorn1s: feel free to start. I can review/cheer on
13:45:16Zagor /taunt ;)
13:45:18*n1s feels free then
13:45:21gevaertsBagder: to be honest, I don't really see the point of rockbox on it
13:45:37Bagderthey claim the OF is crap
13:45:54Bagderand the ARM9 core is probably underpowered for anything "real"
13:46:41gevaertsMaybe it is, but I don't see a big overlap between what that device is built for and what rockbox supports
13:46:43n1sooh, sata disk
13:47:03[Saint]wow...that thing is a bit of a beast.
13:48:27Bagdergevaerts: I agree, but still if the OF sucks and they want to work on getting something else on it, Rockbox is a good bet since we have lots of reverse engineering enthusiasts, while getting Linux on it might be senseible you'll find less people that have done that
13:49:30Bagder270MHz arm9 is not a lot to boast with for that screen and intended purposes, I wonder if they have graphics stuff in the fgpa
13:49:57Bagderhm, and that TI chip has a DSP core too I bet
13:50:00ZagorBagder: surely the fpga is for decoding video. you don't do 720x480@30fps mpeg4 on an arm9
13:50:20Bagderright, dsp and fpga areas indeed
13:50:48[Saint]"VP8870 provides infinite storage expansion via a customer swappable internal 2.5" hard drive (SATA Interface, 40~500GB). "
13:50:55[Saint]500GB is infinite? ;)
13:51:13gevaerts[Saint]: "swappable" is
13:51:14Bagderprobably since the forum user said he has 640 which is then beyond that limit B)
13:51:14Zagorpresumably 500GB was the largest available disk at the time of writing
13:54:36[Saint]It's not the prettiest of things, but it seems pretty versatile.
13:54:51[Saint]an actual OS wold be a nice touch, and a touchscreen.
13:56:19ZagorI agree with gevaerts that the existing owners are not likely to want a rockbox port as much as they want everything they have today, in a rockbox port
13:56:32Zagor...which is a much larger task
14:00:39n1sso, should just store a list of required versions and look for one of those in the output of foo-elf-gcc −−version ?
14:02:53gevaertsn1s: one version per name I'd say, yes (with name being things like arm-eabi-gcc444)
14:04:01gevaertsSo the %which table would get a version entry as well (at least conceptually, I don't know how that would work in practice)
14:04:47n1sthat's what i'm looking at now but i need to google a bit to get how it matches the archlist entries to the entries in %which
14:05:09 Join kugel [0] (~kugel@
14:05:10 Quit kugel (Changing host)
14:05:10 Join kugel [0] (~kugel@rockbox/developer/kugel)
14:06:08 Quit parafin (Read error: Operation timed out)
14:06:52 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
14:07:54 Join parafin [0] (
14:09:37 Quit JdGordon| (Quit: leaving)
14:11:35 Quit JdGord (Quit: Bye)
14:11:44 Quit stoffel (Ping timeout: 272 seconds)
14:20:19 Quit kugel (Read error: Connection reset by peer)
14:24:03 Join Keripo [0] (
14:27:54***Saving seen data "./dancer.seen"
14:35:59 Join Strife89_ [0] (
14:38:38 Join stoffel [0] (
14:41:09n1shere's what i have
14:41:30n1sit seems to work but doesn't check binutils at all
14:41:55n1signore the sdl comment
14:42:11n1snot that checks for binutils at all currently
14:42:29gevaertsrbclient assumes reasonably sane setups
14:43:17gevaertsI think that's reasonable. If we find the right compiler, with the right version, what are the chances of it having the wrong binutils?
14:43:51n1sif you build with, virtually 0
14:45:04 Quit GeekShadow (Read error: Connection reset by peer)
14:45:20 Join GeekShad0w [0] (
14:51:40 Quit TheLemonMan (Quit: free(me))
15:00:42 Quit cjcopi (Ping timeout: 264 seconds)
15:01:05 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
15:03:50 Join kugel [0] (~kugel@
15:03:51 Quit kugel (Changing host)
15:03:51 Join kugel [0] (~kugel@rockbox/developer/kugel)
15:04:55 Quit parafin (Read error: Operation timed out)
15:06:12 Join wodz [0] (
15:06:59 Join parafin [0] (
15:09:40 Join TheLemonMan [0] (
15:11:12wodzMy benches on HD300 shows ~17h of runtime. MPIO advertised 16h and independent measurements from the age of the player shows something like 13h in OF :-)
15:11:15[Saint]restore using itunes, install rb bootloader/build, unmount, test things for a while, re-plug the device, now Windows thinks the FS is "RAW" :/
15:19:10CIA-7New commit by wodz (r29021): HD300 - adjust default battery capacity (based on battery benches)
15:22:02CIA-7r29021 build result: All green
15:23:07 Quit DerPapst1 (Quit: Leaving.)
15:23:07 Quit DerPapst (Quit: Leaving.)
15:25:00 Quit kugel (Ping timeout: 260 seconds)
15:27:08n1swodz: did ypur battery bench show any difference with the new compiler?
15:27:19n1sZagor: should i commit that patch?
15:30:09wodzn1s: I didn't run bench with new compiler build yet
15:30:18 Part LinusN
15:31:28 Join nurdys [0] (~nurdys@2a01:e34:ef6e:c5d0:21d:baff:fe1b:15ec)
15:32:36wodzn1s: Could you perhaps build current svn for me with new compiler? I would like to see how my last tweaks to discharge curve and default battery capacity look on the target. With new build I will have to run only one battery_bench instead of two.
15:33:22 Join komputes [0] (~komputes@ubuntu/member/komputes)
15:33:51n1swodz: ok
15:42:35n1swodz: With new build I will have to run only one battery_bench instead of two.
15:43:48sideralSaint, gevaerts: Re filesystem corruption: I have the theory that my unmount fix in r28693 is broken. It does prevent invalid writes after remounting the device, but it seems to unmount the filesystems without flushing buffers to disk first (as it's layered on the code that handles SD-card ejects). If the FS gets written to during USB insert (e.g., for a DB or bookmark update when playback stops), the FS may be inconsistent
15:44:15sideralso backing out that change may be a good experiment to try.
15:44:20wodzn1s: I'll run bench when charging finishes.
15:44:28n1swodz: cool
15:44:42gevaertssideral: hm, so it would trade one possible corruption path for another?
15:45:04sideralgevaerts: Yes, that's my theory
15:45:28sideralI'll try to come up with a better unmount-before-USB-mode handling within a couple of days
15:45:44 Quit Kitar|st (Ping timeout: 276 seconds)
15:46:40[Saint]I have noticed vastly increased times before the drive is accessable after it is detected by the OS in RB too.
15:47:10[Saint]in the OF, it is accessable instantly, in RB it can take anything up to a minute before it will show as a mounted drive
15:47:34sideralSaint: Hmm, my theory cannot explain that; and I haven't seen that myself
15:47:57[Saint]but, the "safely disconnect hardware" shows up immediately on plugging the device, however the drive can't be accessed for up to a minute
15:48:28gevaerts[Saint]: hard drive or flash?
15:48:48[Saint]CF'd and HDD Colour
15:48:55gevaertscan you see or hear disk activity while it's doing that?
15:49:25*gevaerts will try to have a look later
15:49:38[Saint]Don't stress, but, thanks.
15:49:47[Saint]I need to add a few FS tasks when I get home.
15:49:48sideraldoes windows have an fs consistency check upon mount? perhaps it notices immediately that the two FATs don't match (one type of corruption I see frequently)
15:50:04 Join Kitar|st [0] (
15:50:11sideraland takes some time to "repair" it
15:50:12[Saint]sideral: Aha...I think you might be onto something.
15:52:48wodzDischarge curves posted on OndaRuntime are plain wrong. The voltage is not scaled correctly.
15:53:11[Saint]I can't actually seem to find if XP Pro *does* check disk consistence on mount.
15:53:18[Saint]sideral: ^
15:53:45wodzI don't think windows do any consistency checks
15:53:54[Saint]I would asume so, but, cannot confirm it.
15:55:07wodztry linux than It doesn't do any checks on mount definitely.
15:55:07sideralTo reinforce my theory: The new disk_unmount_all calls disk_unmount, which indeed removes the volumes rather roughly (FDs are not closed, but simply marked invalid; calls fat_unmount(flush=false))
15:55:24[Saint]sideral: gevaerts: One thing that may/may not be important with my issues with the iPod Color and delayed mount, is when I plug USB the "USB Connected" noise fires twice.
15:55:39TornePossibly it's the free space info in the extended bpb?
15:56:03Torneif it thinks fsinfo is invalid it will recalculate it
15:56:07Tornewhich involves reading the entire FAT
15:56:08gevaerts[Saint]: is HID enabled?
15:56:13[Saint]Torne: Noidea, you're a better disk expert than I ;)
15:56:19[Saint]I can only speculate.
15:56:23sideralTorne: sounds plausible!
15:56:28[Saint]gevaerts: Yes, is this normal?
15:56:30 Quit parafin (Read error: Operation timed out)
15:56:33gevaertsWe could logf() some disk access things quite easily actually
15:56:37Tornesideral: i forget hwo OSes normally update fsinfo
15:56:44Torneiirc they invalidate the magic in there on mount
15:56:47Torneand maintain free space in ram
15:56:53Tornethen write it out and mark it valid on unmount
15:56:55gevaerts[Saint]: I wouldn't expect it. Just trying to collect all information that has a slight chance of being relevant
15:56:56Tornebut ic ould be mistaken
15:57:25Torneyou could check that quite easily if you look upt eh format of the fsinfo record and just look at the disk on a linux machine
15:57:30Tornehexdump the relevant sector without mounting the fs
15:57:46sideralTorne: AFAIR it takes a long time on FAT to calculate the free disk space if you don't trust the FATs (which Windows may not do if the FATs don't match)
15:58:04 Quit Bagder (Remote host closed the connection)
15:58:07Tornesideral: it's not about trusting the FATs
15:58:13Torneit's about whether the fsinfo sector is up to date
15:58:23Torneit takes a long time to calculate free space from the FAT :)
15:58:26TorneYou can't do it any other way
15:58:35 Join parafin [0] (
15:58:41Tornethe direntries don't have enough info to work it out without rebuilding the whole fs in the process
15:58:51Tornesee the end of
15:59:08Torneas the page says it's normally stored in sector 1 of the fs
15:59:13Torneso plug it into a linux machinet hat doesn't automount
15:59:21Tornedd out sector one of the partition
15:59:22Torneand ehxdump it
15:59:26 Quit CaptainKewl (Ping timeout: 260 seconds)
15:59:36[Saint]bbs, restart.
15:59:44Torneif the free cluster count is 0xffffffff then the previous user of the FS didn't bother to update it
15:59:56Torneand windows will almost certainly go read the entire FAT from beginning to end to count the free clusters for itself
15:59:57sideralah, so the update to the FSinfo sector may get lost if we unmount too quickly, forcing expensive recalculation on remount
15:59:59 Quit [Saint] (Quit: I'm only going to Heaven if it feels like Hell, I'm only going to Heaven if it tastes like caramel...)
16:00:04Tornesideral: exactly
16:00:20Tornesideral: if the fsinfo signature is valid and the count is anything but -1 then it assumes it's right
16:00:24Torneotherwise it has to scan the entire FAT
16:00:40sideralok, another piece of evidence for my theory
16:00:58Tornebut yeah, it's easily tested if you have a computer that can read the raw sectors without automounting the FS :)
16:02:59gevaertshm, why doesn't disk_unmount() call fat_unmount() with flush=true?
16:04:04sideralgevaerts: because it's normally called after the SD card has been forcibly removed by the user
16:04:52gevaertsIt might be worth testing if changing that fixes the issue. Of course a proper fix will take both cases into account
16:05:15gevaertsSetting it to true for ipod builds (so Saint can test) should be safe though
16:05:19sideralwe might as well close all FDs properly while we are at it
16:05:24 Join [Saint] [0] (S_a_i_n_t@
16:06:36Torneah, if we're treating it the same as an SD card pull then fsinfo seems very likely to be the culprit
16:07:09Zagorn1s: yes, please commit
16:09:29gevaertssideral: I guess the easiest solution is to declare your original patch to be better, i.e. do everything in disk_unmount_all() instead of disk_unmount()
16:10:08 Join slooopy [0] (
16:10:45sideralgevaerts: if you really want to commit a quick temporary fix, you could set flush to true for all HAVE_HOTSWAP targets. But I suggest we do a complete fix even if it takes a day or two longer
16:11:20sideralgevaerts: Yeah, that would be a fine approach.
16:11:41 Quit parafin (Read error: Operation timed out)
16:11:42sideralBut we have to fix my first patch as well to properly close FDs and unmount with flushing.
16:12:05 Join parafin [0] (
16:12:13[Saint]Another thing that may be AV also will not scan the Colors upon mount
16:12:21[Saint](it's set to scan removable devices)
16:12:50Tornewhat "removable" means in this context is fantastically complicated and weird
16:12:51sideral(re HAVE_HOTSWAP: I meant "for all !defined(HAVE_HOTFLUSH) targets")
16:13:24gevaertsTorne: it's not, it's just that nearly everyone gets it wrong :)
16:13:27[Saint]Torne: Sorry..I fucked that up massively
16:13:38[Saint]it *does* scan the OF, but not RB.
16:13:49Tornegevaerts: that's efectively the same :)
16:14:36gevaertssideral, [Saint]: quick patch (totally untested, may not compile):
16:16:22[Saint]Hmmm...HID is fucking up the scan on the Colors.
16:16:27[Saint]but not on the Nanos
16:16:58[Saint]this is fantastically hard to debug.
16:19:51 Quit wodz (Ping timeout: 240 seconds)
16:22:47sideralgevaerts: Indeed, does not compile. ;) Here's a better patch (hot-edited, may not even apply ;)
16:23:38gevaertsAh, yes :)
16:24:30[Saint]HID is definitely messing up the AV scan, and responsible for the delayed mount.
16:24:45 Quit factor_ (Quit: Leaving)
16:25:03[Saint](just tested *many* combinations of HID/plugging)
16:25:05 Join factor [0] (~factor@
16:25:29[Saint]However, HID behaves fine on all my other targets.
16:25:35[Saint]just the Colors suffer.
16:25:49[Saint](ana Nano2G, where it is disabled)
16:25:55gevaerts[Saint]: all your other targets are nano1G?
16:25:56 Quit nurdys (Quit: Quitte)
16:26:17[Saint]gevaerts: Nano1G, FuzeV1
16:26:44gevaertsColor is PP5020, Nano1G is PP5022. Maybe that makes a difference
16:27:22[Saint]perhaps...could read/write speed be an issue here? and the Nano1G and Fuze are just faster?
16:27:48[Saint]I'm thinking HID may be causing the scan to time out while it tries to enumerate
16:27:55***Saving seen data "./dancer.seen"
16:28:08 Join MethoS- [0] (~clemens@
16:28:21[Saint]Nope...that can't be it or I wouldn't see iton the CF'd Color too...gah!
16:32:19[Saint]*aaaaaannnnnnddd*...disk corruption.
16:32:49 Quit cheesy ()
16:32:52 Nick GeekShad0w is now known as GeekShadow (
16:33:16[Saint]chkdsk found ~1000 screwed files.
16:33:17 Quit GeekShadow (Changing host)
16:33:18 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
16:33:24[Saint]yay :/
16:34:22sideralSaint: are you running with that patch yet? (
16:35:31[Saint]I'll build it while I reformat my Colour for the umpttenth time
16:36:04[Saint]It seems to get so badly messed up that chkdsk can't recover the files.
16:37:43 Join enthdegree [0] (
16:38:04enthdegreeDid I brick my C200V1? When I turn it on the blue light comes on and the screen stays blank
16:38:16enthdegreeAFAIK no hardware buttons respond
16:38:29sideralI've put the patch on my ClipV2, and my first test was successful
16:39:24gevaertsenthdegree: have a look at When something is unclear, don't guess, ask (doing things wrong in this state can lead to a *really* unrecoverable state)
16:39:29[Saint]sideral: Good to know
16:39:44enthdegreegevaerts: It's been runninc robo for a while fine now
16:40:00enthdegreethis just happened one day
16:41:02[Saint]sideral: For me, the acid test seems to be trying to access the DB after boot.
16:41:09[Saint]if I can't, corruption.
16:41:26[Saint].tcd files seem to be really prone to disk corruption, and .fnt files
16:41:44sideralI'll run with the patch a few days and see what happens. closing all FDs may be a nice addition to this, but unneeded in the case of unmount before USB mode
16:42:41sideralSaint: For me, it's fsck detecting corruption after copying a large file + umount. This is what I just did, and there was no corruption −− yet. :)
16:42:46[Saint]sideral: You use the DB?
16:42:56[Saint]do you get the same issues with .tcd corruption?
16:43:46sideralSaint: Yes, I do. I have seen the DB being rebuild after reboot a couple of times. This may have been an indication of corruption.
16:44:09[Saint]sideral: I believe it is, yes.
16:44:49 Quit Strife89_ (Quit: Reboot.)
16:45:58 Quit jhMikeS (Ping timeout: 250 seconds)
16:46:24 Join Strife89_ [0] (
16:46:45 Join jhMikeS [0] (~jethead71@rockbox/developer/jhMikeS)
16:48:56sideralBTW, the recent reports of {flac, other} files being skipped randomly (FS #11501) also may have been an indication of FS corruption
16:50:37sideralanyway, need to get some work done how. talk to you later!
16:55:14[Saint]gevaerts: sideral: patch: **** malformed patch at line 36: return unmounted;
16:55:43[Saint] was the patch
16:55:45gevaerts[Saint]: sounds like the usual pastebin issue. Add an linefeed at the end
16:56:17[Saint]a blank line?
16:56:26[Saint]I did so.
16:56:38gevaertsThen remove one :)
16:57:37pixelmaline 36 is not the last line
16:57:51*[Saint] realises this
16:58:08[Saint]it is the "malformed" line
16:58:53gevaertsoh, remove all carriage returns
16:59:31gevaertshm, no
17:00:01[Saint]same error
17:00:15pixelmathere is something weird about the lines with and without the " - " at the beginning there
17:02:47[Saint]how so?
17:02:57 Part Zagor
17:03:50gevaerts[Saint]: try
17:05:39 Join sideral1 [0] (~sideral@unaffiliated/sideral)
17:06:29CIA-7New commit by nls (r29022): check for the correct version of crosscompilers.
17:06:38 Quit sideral1 (Remote host closed the connection)
17:06:40[Saint]gevaerts: success...thankyou
17:06:42 Quit sideral (Ping timeout: 255 seconds)
17:07:11 Join sideral [0] (~sideral@unaffiliated/sideral)
17:14:56n1sok, so i think we are pretty much ready to switch the cf toolchain now, so I think 1) commit change, and get someone with a buildserver to upgrade 2) commit change and get an admin to push an update and then commit all the rockbox changes last
17:16:05n1sone thing i wondered about is how the buildsystem will cope, as if is updated to the new version of gcc, all clients that specify the m68k target but have the old tollchain will error out so will they not build anything then?
17:16:33gevaertsn1s: shouldn't be "updated to the new version of gcc", it should be updated to also support the new version
17:16:45 Join sbmull [0] (
17:16:53 Quit sideral (Remote host closed the connection)
17:17:06n1sgevaerts: ah, hmm
17:17:31gevaertsUsing "m68k-gcc451" instead of "m68k" as the new name
17:17:35 Join sideral [0] (~sideral@unaffiliated/sideral)
17:17:37gevaerts(or whatever your version is)
17:17:47sbmullIs it possible at this point to play a vorbis file in LUA? I would like to incorporate this into a LUA application for RB.
17:17:50n1sah, right, like arm-eabi-gcc444
17:18:04n1ssbmull: no
17:18:39n1sgevaerts: right, now i remember you already explained this, sorry
17:19:11gevaertsn1s: yes. And then (when some people have build clients with the new toolchain), we change the builds file to require this new m68k-gcc451 instead of m68k
17:19:44gevaertsOh, and we change and configure more or less simultaneously as well of course
17:20:13 Quit sbmull (Client Quit)
17:20:17n1sok, now i get it, i'll whip up that change for then
17:22:45CIA-7New commit by nls (r29023): add m68k-gcc452 arch
17:23:47gevaertsn1s: is the patch online somewhere?
17:24:21n1syes, last patch in FS #7832
17:24:37n1salso does configure
17:25:24gevaertsok. I'll build a new toolchain for my client then, so we have at least one that's ready to go
17:26:58n1swe need an admin to push the chane
17:27:08 Quit sideral (Remote host closed the connection)
17:27:47 Join sideral [0] (~sideral@unaffiliated/sideral)
17:29:26gevaertsYes, *before* people start updating toolchains preferably. I'll update mine by hand
17:29:32 Join wodz [0] (
17:31:29 Join sideral1 [0] (~sideral@unaffiliated/sideral)
17:31:49 Quit sideral (Ping timeout: 240 seconds)
17:35:17 Join DerPapst [0] (
17:35:40 Join DerPapst1 [0] (
17:38:57 Quit sideral1 (Remote host closed the connection)
17:39:16 Join sideral [0] (~sideral@unaffiliated/sideral)
17:43:30 Part Llorean
17:49:08 Quit TheLemonMan (Quit: free(me))
17:49:12 Quit sideral (Remote host closed the connection)
17:49:14 Quit GeekShadow (Ping timeout: 250 seconds)
17:50:04 Join sideral [0] (~sideral@unaffiliated/sideral)
17:50:33 Quit factor (Ping timeout: 276 seconds)
17:50:40 Part sideral
17:55:11 Join panni_ [0] (
17:57:20 Quit petur (Quit: Page closed)
18:02:00 Join factor [0] (~factor@
18:13:03 Join WonTu [0] (
18:13:17 Part WonTu
18:16:33 Quit Keripo (Quit: Leaving.)
18:17:50 Quit TheSeven (Ping timeout: 250 seconds)
18:23:31 Quit Strife89_ (Quit: Putting away the laptop.)
18:27:14 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
18:27:58***Saving seen data "./dancer.seen"
18:35:31 Join swilde [0] (
18:39:28 Join sideral [0] (~sideral@unaffiliated/sideral)
18:48:31 Join kugel [0] (
18:48:35 Quit kugel (Changing host)
18:48:36 Join kugel [0] (~kugel@rockbox/developer/kugel)
18:50:06 Join Keripo [0] (
18:53:15kugeln1s, gevaerts: I'm also updating my client, but I'm having trouble at downloading gmp
18:54:10kugeldownloaded it by hand (with wget) but still tries to download (with curl)
18:54:32n1skugel: any of the mirrors here should work i haven't found a generic adress that redirects to a suitable mirror automatically
18:54:54[Saint]I have both curl and wget installed, and favours wget here
18:55:18kugelit doesn't according to the code
18:56:12n1skugel: if you download the packeges by hand and put them in /tmp/rbdev-dl/ shoudln't dl them again
18:57:30kugel"/home/kugel//makeinfo: Ungültige Option −− L" :(
19:01:08 Quit swilde (Remote host closed the connection)
19:01:28 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
19:02:28[Saint]n1s: I'm going back a ways to my grandparents...but, it's "invalid", or "incorrect" IIRC
19:02:47*[Saint] looks to kugel for comfirmation
19:02:47n1sthat makes no sense since 1) we don't call makeinfo the tool's builds do so they are using an invalid option ??? 2) we configure with −−disable-docs so the docs shouldn't be built in the first place
19:03:06n1s[Saint]: it's not the German that confused me :)
19:03:49kugeliirc makeinfo is used by the build process, i needed to compile it for the other toolchains
19:04:24n1sthe gcc docs say it isn't needed for releases unless you modify the docs source
19:04:40n1sor build an svn version
19:04:52[Saint]I needed to pull makeinfo for CygWin also
19:04:54n1skugel: do you have makeinfo?
19:07:36n1saccording to the various manuals there doesn't seem to be an -L or −−L option so i wonder where that came from
19:08:00n1sgevaerts: did your toolchain builld ok?
19:08:34kugelmeh, I'm unable to compile this
19:08:41kugelit always fails at downloading gmp
19:09:14[Saint]download it by hand and put it in tmp/rbdev-dl
19:09:44kugeli tried that
19:09:48kugelsee above
19:09:50n1s<n1s> kugel: any of the mirrors here should work i haven't found a generic adress that redirects to a suitable mirror automatically
19:10:06kugelI tried a few mirrors now
19:11:00 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
19:11:15 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
19:11:56kugelalways the same error
19:13:35n1sthere's one too many "gcc" in that adress
19:14:58kugelI'm just pasting links from that mirror site
19:14:59 Join Buschel [0] (
19:15:19kugelit's the same error i get with only your patch applied (i.e. whatever the hardcoded mirror is)
19:16:19 Join mystica555_ [0] (
19:16:42kugelsame without the extra gcc. i also don't understand why downloading by hand doesn't work
19:17:13n1smeh, those mirrors don't have the same contents
19:17:36kugelthe mirrors i tried do have gmp-4.3.2.tar.bz2
19:17:47n1syes, that is what we want
19:19:12n1sthe reason for the swithc was to use the infrastructure dir they set up that contains the recomented versions of gmp, mpfr and mpc, since the regular gnu mirrors dont' have mpc
19:19:53[Saint] works for me
19:19:59n1s*but some* of the gcc mirrors don't have anything else than gcc apparently so no binutils :(
19:20:03[Saint](by hand)
19:20:54n1si guess i'll have to revert to the old gnu mirror scheme and hardcoding the mpc addr, hold on and i'll whip up a patch
19:21:07kugeloh, works now
19:21:17kugelapparently confused by relative paths for RBDEV_*
19:21:37 Quit stoffel (Remote host closed the connection)
19:21:46n1sah, ok, but still the mirror has problems sometimes
19:25:04gevaertsn1s: no. Still downloading stuff
19:25:15gevaertsI think I'm going to stop it and try another mirror
19:25:16n1sgevaerts: that slow, eh
19:25:25gevaerts500 bytes per second now
19:25:36n1shold on for a moment
19:25:47Buschel[Saint], soap: did the nano 1g work well with the LCD patch?
19:26:36[Saint]Buschel: Yes, but, it adds a white flash to the shutdown routine which was not there presently on the Nano1G
19:26:48[Saint](Colour does this also)
19:27:08n1sgevaerts: this applied to clean svn should work better,
19:27:33Buschelhmmm, this happens on the iPod Video as well...
19:27:36n1susing the geolocated and hardcoding the mpc url instead
19:28:41*gevaerts tries
19:29:14kugeln1s: I'm finished now. errors out though
19:29:49kugelI assume needs to be updated
19:29:59n1skugel: yes
19:30:06n1sand we need an admin for that
19:30:37n1sgevaerts: sorry but that may or may not work, it looks like the mirrors don't have a frigging consistent dir structure or i am going mad :(
19:30:37gevaertsyes and no
19:30:55gevaertskugel: you can get a new by hand
19:31:30gevaertsn1s: we'll see
19:32:34kugeldone that
19:33:32 Join GeekShadow [0] (
19:33:34 Quit GeekShadow (Changing host)
19:33:34 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
19:34:24Buschel[Saint]: can you try to add "backlight_off(); sleep(HZ/5);" before "lcd_shutdown();" (line 791, firmware/powermgmt.c)?
19:35:23[Saint]will do, just waiting for a set of builds to finish.
19:35:37[Saint]there's something seriously wrong with mounting/unmounting.
19:35:42Buschelok, just let know if it fixes the issue
19:35:49[Saint]will do.
19:36:04 Quit benedikt93 (Quit: Bye ;))
19:54:03 Join Parsi [0] (~Maysam@
19:54:09 Part Parsi
19:56:10gevaertsn1s: client running again with 4.5.2
19:58:09Buschel[Saint]: the change I posted above does not solve the issue on my iPod Video. the flash is introduced with the SLEEP command itself −− independent of the backlight
19:58:20Buschelso, not much hope it will fix the issue for you
19:59:11[Saint]there's a comment in the original patch about the white flash on waking, and increasing the wait should fix it...does this not work for sleeping?
19:59:49 Join Horscht [0] (~Horscht@xbmc/user/horscht)
20:00:50Buschelwill make a short test on my Video again
20:01:29 Quit lostlogic (Quit: leaving)
20:02:26 Join TheLemonMan [0] (
20:04:19 Join esperegu [0] (~quassel@
20:04:47n1sgevaerts: good
20:06:26 Quit slooopy (Ping timeout: 240 seconds)
20:08:02 Quit Horscht (Quit: Verlassend)
20:10:07n1sgevaerts, kugel: I'll switch back to the kernel mirror
20:13:41CIA-7New commit by nls (r29024): switch back to using the mirrors as the mirrors were unreliable and inconsistent. Hardcode the mpc url as mpc is not ...
20:14:10 Join Horscht [0] (~Horscht@xbmc/user/horscht)
20:14:35 Join lostlogic [0] (
20:14:36 Quit lostlogic (Changing host)
20:14:36 Join lostlogic [0] (~lostlogic@rockbox/developer/lostlogic)
20:15:13 Quit liar (Ping timeout: 255 seconds)
20:16:13CIA-7r29024 build result: All green
20:17:56 Quit antil33t ()
20:19:52 Join slooopy [0] (
20:20:00 Quit t0rc (Quit: Give someone code, help them with one project. Teach someone to code, help them rule the world.)
20:21:20*[Saint] wonders if there is a way he can check if configure is newer than the makefile, and run reconfigure automatically at compile time.
20:22:18 Join liar [0] (
20:28:00***Saving seen data "./dancer.seen"
20:28:22 Quit factor (Read error: Connection reset by peer)
20:29:45 Join cjcopi [0] (~craig@
20:29:52 Join antil33t [0] (
20:30:27 Join factor [0] (~factor@
20:45:16kugel[Saint]: the first part should be working already
20:49:57 Join evilnick_B [0] (0c140464@rockbox/staff/evilnick)
20:53:42Buschelhmm, the white flash when shutting the iPod Video can be shortened but not avoided. right now is HZ/4 longer than possible. but the single SLEEP-cmd itself does the flashing, even when swithcing off the power supply instantly after this command
21:00:26 Quit sideral (Ping timeout: 260 seconds)
21:04:57TheSevenBuschel: isn't the LCD white in its unpowered state?
21:05:04 Quit enthdegree (Quit: leaving)
21:05:16TheSevenso isn't it mainly a matter of turning the backlight on and off at the correct points?
21:05:27 Join newClipUser [0] (
21:05:33newClipUserhello hello
21:05:42Buschelno. you can swithc off the bl directly before the SLEEP command and it will kep flashing
21:05:58TheSevenhow can it possibly flash if there is no backlight?
21:06:23TheSevennewClipUser: if you have a question just go ahead and ask it. someone who knows an answer might respond.
21:06:51*TheSeven has no flashing at all on ipod classic
21:07:07Buschelusing power down code or not?
21:07:24TheSeveni'm shutting it down properly, yes (like apple does)
21:07:25newClipUserCan you tell me something about the power consumption of MP3 voice recording with the rockbox on sansa clip+?
21:07:36TheSevenbut i'm powering off the backlight immediately before doing so
21:07:51[Saint]kugel: Yes, it does warn if configure is newer than the makefile...I'm wondering how I can hook into that warning to fire a script though.
21:08:04Buschelat least calling backlight_off() does not solve it... maybe I should check this implementation
21:08:18TheSevenmaybe there is some delay before the backlight actually powers off?
21:08:20TheSevendoes it fade?
21:08:33gevaertsnewClipUser: I don't think we have much data about that
21:08:34TheSeveni set the fading speed to instantaneous before powering it off in that case
21:08:48kugel[Saint]: wrap it in another script. reconfiguring automatically was enabled but reverted because it's annoying
21:09:15gevaerts[Saint]: configure rewrites the makefile, which means that while you could run configure automatically, you can't immediately build
21:09:31CIA-7New commit by bluebrother (r29025): Add missing ) to manual.
21:09:55TheSeven(but that's specific to the PCF50633 IIUC)
21:10:23newClipUserI have a sansa clip+ running with external (soldered) 3xAA battery pack (about 7500mAh), It was about half charged and the batteries died after 7hours
21:10:39Buschelbl_off(); sleep(HZ); before the SLEEP cmd still does not solve it...
21:11:04TheSevennewClipUser: three batteries in series? you shouldn't add up the mAh values in that case
21:11:57newClipUser@gevaerts ok, I'll have some test runs with fullcharded batteries then.
21:12:01TheSevenalso 3xAA might possibly make it shut down before the batteries are actually empty (depending on what the shutoff threshold on that device is)
21:12:13CIA-7r29025 build result: All green
21:12:32gevaertsAlso, what type are they? NiMH is 1.2V IIRC, so that would be 3.6V. That's on the low side
21:13:01TheSevengevaerts: I don't think you can safely use more as the maximum voltage after charging is rather high with 4.5V
21:13:23TheSevenit might be worthwhile to decrease the shutoff threshold though
21:13:28gevaertsTheSeven: right
21:13:41TheSeven2.7V would be perfectly fine for the batteries, but I'm not sure if the player still runs stable at that voltage
21:13:44gevaertsIt depends on how they are connected though
21:13:46newClipUser@The Seven... you are ritght.. hehehe
21:14:15TheSevenalso the charging method used by the player won't really work well with NiMH
21:14:34[Saint]gevaerts: I'm not concerned about the delayed build due to re-running configure...I just want to automate it as much as possible. Having to run make reconf manually is annoying.
21:14:34TheSevenit shouldn't be able to damage the cells, but it probably won't charge them to full capacity
21:14:53TheSeven[Saint]: why not just always run make reconf then?
21:14:58[Saint]I suppose I will just look for the commit that reverted the behaviour I'm looking for.
21:15:20[Saint]TheSeven: *always* running it is a bit excessive.
21:15:21n1s[Saint]: make a script that always runs configure, then make
21:15:28 Quit TheLemonMan (Quit: free(me))
21:15:41n1sconfigure is hardly expensive to run
21:15:48 Quit newClipUser (Quit: CGI:IRC)
21:15:54[Saint]no, but time consuming.
21:16:06TheSevenn1s: won't it cause make to recompile basically everything because autoconf.h and the Makefile dates change?
21:16:15TheSevenbut most of that should be caught by ccache
21:16:22 Join newClipUser [0] (
21:17:03n1sTheSeven: yes, it probably will, forgot [Saint] uses cygwin ;)
21:17:03TheSeven[Saint]: actually you hardly ever really need to run make reconf if it tells you to do so
21:17:24n1s[Saint]: you can also just patch out the warning :)
21:17:29newClipUserSo I have to get an IC that keen the voltage constantly on 3,7V (like provided from the litiumion pack)
21:17:40TheSevenit might be worthwhile if there are changes to e.g. optimization flags, but if it's just stuff like new targets, you can just forget about it
21:18:06TheSevennewClipUser: 4.2V should be safe as well. but doing that reduces efficiency and thus battery runtime
21:18:49TheSevenwhat you might want to try is how low you can go with the shutoff threshold
21:19:01TheSevenif you get that down to like 2.7V it should use the batteries optimally
21:19:01newClipUser4.2V? what battery pack provides this voltage?
21:19:17TheSevennewClipUser: that's about the voltage of a fully charged li-ion/li-polymer battery
21:19:30newClipUseroh ic
21:20:19newClipUserok, back to the shut down treshold... can this be set per firmware?
21:20:19TheSevenso you might want to disable low-battery shutoff altogether, monitor the battery voltage, and check if there is a sudden drop starting at some point, and at which point the player starts to behave eratically
21:20:52newClipUserhmm ok, I will remember that
21:21:10TheSevennewClipUser: as i don't really know the sandisk players i can't be really sure, but usually there's both a software-based rather high threshold, and a lower last-resort hardware threshold
21:21:21TheSeventhe hardware threshold is 2.9V on an ipod for example
21:21:42TheSevenwhich would be fine for your batteries
21:21:54TheSevenas long as they are balanced properly NiMH doesn't care about relatively deep discharging
21:22:12newClipUserbut i think it would be much easier to get a voltage regulator IC
21:22:23TheSevenwhat do you want to achieve?
21:23:25newClipUserI use the sansa clip for longterm sound recording for different lowcost projects
21:23:42newClipUserlike ambient sound (streets, malls, public places)
21:23:54TheSevendo you care about efficiency?
21:23:55newClipUseror forest sound recording to identify birds
21:24:07TheSevenor can you just throw another dozen cells at it if you need more runtime?
21:24:24n1swouldn't it be easier to connect 4 cells to a 4.8V package and connect that to the charger input?
21:24:31newClipUseryes, so i need energy efficiency AND good compression (the clip i use got 8Gig)
21:24:48TheSeventhe shutoff threshold for the clip+ is 3.3V in rockbox. you may want to decrease that to 3.0V (as a comment says that's what sandisk uses)
21:25:12TheSevenn1s: that highly depends on the charging circuitry
21:25:23TheSevenif it can handle a 3.6-6.0V range properly that may work
21:25:30TheSevenbut the charging circuitry may waste power
21:25:40newClipUser@n1s I'm considering that too, but I don't know how it would behave
21:26:25TheSeveni think your best bet is probably to reduce the shutoff voltage to 3.0V or even lower and use three batteries in series at the battery terminals
21:26:33newClipUseri mean there might me powerlost when the external battery pack tries to charge the internal battery constantly
21:26:40TheSeven(and maybe several of them in parallel if needed)
21:27:01TheSevenyou might also want to think about just using a bigger li-ion/li-polymer battery
21:28:02newClipUserat some places the size of the recording device matters, I think I should stick with the 3xAA battery pack
21:28:24newClipUserok, the first thing i will try is to lower the shut off treshold
21:28:35TheSevenli-polymer will get more energy into the same space
21:28:58wodzI would recommend bigger li-po. This are dead chip this days
21:29:34BuschelTheSeven: you were right. the backlight was *not* switched off immediately. now searching for a good solution
21:29:35newClipUseryes, but it is for some lowbudget projects, for a li-po I will probably need an extra charger and the li-pos are also not cheap
21:30:03wodzyou don't need extra charger - clip serves as charger itself
21:30:09newClipUserNi-MH batteries are quite handy, and every household has a suitable charger
21:30:26TheSevenif you have enough time to charge them, charging them in the player will work
21:30:42TheSevenand Li-Polymer isn't that much more expensive than NiMH
21:31:32newClipUseryeah, the charging time is also something to consider: we want just to change the battery, get the audiofiles on a laptop and have the device running again
21:31:38wodznewClipUser: clip has 330mA battery almost any cheap cellphone battery is bigger
21:31:46newClipUserusing the clip to charge the battery would cost too much time
21:32:17TheSevenli-polymer can be charged in like 3 hours if you have a powerful-enough charger (which the clip won't be for a highly extended battery pack)
21:32:38TheSevenNiMH can be charged in like one hour, but needs an even more powerful charger and produces lots of heat
21:32:48newClipUserhmm, ok I still have a charger for my old smartphone (XDA mini/HTC magician)
21:32:48TheSevenalso the capacity will probably suffer from charging them that fast
21:33:10TheSevennewClipUser: that one probably isn't a charger but rather a 5V power supply
21:33:21TheSeventhe charger is usually integrated into the device
21:33:29newClipUser@TheSeven, you are rght again :(
21:33:40newClipUserno, wait
21:34:14TheSevenif charging time matters and size/weight doesn't matter that much, go for NiMH. otherwise go for li-polymer
21:34:17newClipUserit is like a cradle where i can charge a battery that is not put in the device, like an 2nd battery
21:35:01TheSevennewClipUser: these will probably want to check the battery temperature etc. and will refuse plain cells without the electronics
21:36:16newClipUser@TheSeven, I i use this charger I would also use li-ions batteries that fits into this charger, should be cheap on ebay by now
21:36:29newClipUser*if i use this charger
21:36:33TheSeventhat may of course be an option :)
21:36:46TheSevenhow much capacity do these have?
21:36:50newClipUserthe size would also be quite handy :)
21:36:57TheSeveni'd guess around 1500mAh, so you might want to use two of them in parallel
21:37:06newClipUserabout 1000mAh I guess
21:37:21newClipUseryes :D
21:37:44newClipUserI buy like 4 for every recorder and exchange them on the fly
21:37:59TheSeventhat might not be a good idea
21:38:13TheSevenyou shouldn't connect batteries with different charge state in parallel
21:38:25newClipUserno no...
21:38:46newClipUseras ou said, 2 li-ion packs in parallel at a time
21:39:28TheSevenyeah, i though you meant replacing one of that pair at a time to keep the player powered while you exchange them
21:39:59newClipUserno I don't need seemless recording that much
21:40:33newClipUserthe device can be switch off for the battery changing, and I have to get the files off the clip too
21:40:34TheSevenin that case it sounds like a plan :)
21:41:02TheSevenyou can also charge the battery using the player, it will just take longer than usual
21:41:50newClipUseryes... but as I said... we don't want to do that. changing batteries are fine :)
21:42:24newClipUserso, just for the fun of it: how do I decrease the shutdown voltage?
21:43:59newClipUserI haven't done anythin in the code of rockbox yet...
21:46:33 Nick amee2k is now known as amee2dead (
21:46:59 Join captainkewll [0] (2669ecde@gateway/web/freenode/ip.
21:47:28 Nick amee2dead is now known as amee2undead (
21:47:45newClipUserno help regarding this question? :)
21:48:10 Nick amee2undead is now known as amee2fries (
21:48:56Buschel[Saint]: still there and willing to test an anti-flash patch for nano1g and color?
21:49:12wodznewClipUser: edit firmware/target/arm/as3525/sansa-clip/powermgmt-clip.c
21:49:59TheSevenhe has a clip+, so that would be /trunk/firmware/target/arm/as3525/sansa-clipplus/powermgmt-clipplus.c
21:50:56newClipUserah, thank you. I will start looking into the code from there
21:51:35[Saint]Buschel: I'm still here, and, of course!
21:51:40[Saint]always willing.
21:51:55Buschel:o) wait a minute, I think I got it
21:52:13Buschel(famous last words of a sw-dev :)
21:52:46 Quit n1s (Quit: Lämnar)
21:53:41newClipUserThank you all for the help, especially TheSeven!
21:54:52Buschelsame works fine on my Video now
21:55:00 Join JesusFreak316 [0] (
21:55:08*[Saint] crosses his fingers.
21:55:28[Saint]It would also be very nice to get LCD-Sleep working.
21:56:17 Join enthdegree [0] (
21:56:25CIA-7New commit by Buschel (r29026): iPod Video LCD: Avoid white flash when entering sleep mode or shutting off.
21:58:55CIA-7r29026 build result: All green
21:59:06TheSevenBuschel: is there any particular reason for the backlight-nano_video.c change?
21:59:35 Join Strife89_ [0] (
22:00:22 Nick Strife89_ is now known as Strife89 (
22:02:24wodzTheSeven: what's your next target? nano3g?
22:02:29*Buschel got himself a new beer
22:02:35TheSevenwodz: not sure
22:02:39BuschelTheSeven: no. mistake. :/
22:02:45TheSevennano3g/4g or maybe fixing the cowon d2
22:05:03wodzTheSeven: you mean rev eng FTL on cowon d2 or add support for DSP core in d2?
22:05:49TheSevenmostly the FTL and IIRC there were some touchscreen problems
22:06:10TheSevenDSP also sounds interesting but i have no idea if i could manage to do that
22:07:57CIA-7r29027 build result: All green
22:07:57CIA-7New commit by Buschel (r29027): Revert unneeded change from r29026.
22:10:51wodzhmm I read it wrong probably - I thought d2 cop is different arch than main cpu
22:17:50 Quit komputes (Quit: I haven't slept for ten days, because that would be too long.)
22:21:17Buschel[Saint]: still building?
22:21:55[Saint]Buschel: Just about finished.
22:22:16*Buschel is curious
22:23:07 Quit slooopy (Ping timeout: 250 seconds)
22:26:30BuschelStrife89: do you have access to your iPod color? In FS #10034 you mentioned you have a lcd_type 2 LCD controller.
22:27:08[Saint]Buschel: Nope, sorry.
22:27:21[Saint]Still a full white screen before the LCD shuts down.
22:27:31Buschelcolor or nano?
22:28:04***Saving seen data "./dancer.seen"
22:28:17[Saint]haven't tested on Nano.
22:28:44Strife89Buschel: You'll need to talk to LambdaCalculus37 about it; but, it's probably dying, so it'll be difficult to run tests on it.
22:28:55[Saint]Buschel: Strife89's Color is dead.
22:29:15[Saint]ah, yeah...what Strife89 said.
22:31:39 Quit markun (Read error: Connection reset by peer)
22:32:38[Saint]It also seems to take longer to shut down, it's *great* that those horrible corrupt lines are gone now, but that white screen on shutdown is a little annoying.
22:32:58[Saint]Not even close to as annoying as the screen corruption though.
22:33:15Buschel[Saint]: can you test the nano1 g
22:34:37*Buschel is not sure why the _backlight_off() does not kill the flashing...
22:34:40 Quit GeekShadow (Read error: Connection reset by peer)
22:35:06 Join GeekShad0w [0] (
22:35:13 Quit wodz (Quit: Leaving)
22:35:23*Buschel should have read the comments
22:35:25 Nick amee2fries is now known as amee2k (
22:36:12 Join stripwax [0] (
22:36:43 Join slooopy [0] (
22:36:44Buschel[Saint]: can you please do the following change -> uncomment line 62 in firmware/ipod/backlight-4g_color.c ?
22:38:02Buschelthis will only change the iPod color LCD backlight handling. no clean build required
22:39:38[Saint]line 62?
22:40:15Buschelshould be something like "GPIO_CLEAR_BITWISE(GPIOB_OUTPUT_VAL, 0x08);"
22:40:28 Quit newClipUser (Ping timeout: 276 seconds)
22:40:43[Saint]ah...line 55 here.
22:41:03Buschelno, that's _backlight_on()
22:41:03 Join thomasjfox [0] (
22:41:44Buschelyou might uncomment both lines though
22:44:29[Saint]fuck it...I need to revert all my local changes.
22:44:42BuschelI might just give you a binary :)
22:46:58Buschelcolor -> nano1g ->
22:47:06Buschel(just the rockbox.ipod files)
22:47:17Buschelif you need the full install, let me know
22:47:53 Quit JesusFreak316 (Ping timeout: 240 seconds)
22:49:57 Quit thomasjfox (Remote host closed the connection)
22:50:36 Join thomasjfox [0] (
22:53:56 Quit [Saint] (Ping timeout: 240 seconds)
22:54:38 Join [Saint] [0] (S_a_i_n_t@
22:55:15[Saint]Buschel: Bad checksum...what target did you build for?
22:56:02Buschelcolor and nano1g
22:56:23[Saint]what one did you link just now?
22:57:27Buschel? there were 2 links
22:57:44Buschelcolor ->
22:57:55Buschelnano1g ->
22:58:06[Saint]aha...shit, sorry.
22:58:11[Saint]I missed the first.
23:02:22[Saint]VERY bad screen corruption
23:02:40[Saint]*EXTREMELY* bad screen corruption.
23:02:56Buschelin general or when shutting down?
23:03:25Buschelcolor or nano?
23:03:32[Saint]Oh,'s probably due to AA fonts
23:03:43[Saint]shutting down seems ok, I'll do some more testing.
23:04:17BuschelI might also update the full install zip's
23:04:37Buschel(if needed)
23:05:07[Saint]no, I just reset my .cfg by flicking on hold and it was the AA fonts
23:05:16[Saint]shutdown of the screen works perfectly.
23:05:30Buschel:o) on both targets?
23:05:45[Saint]One secong.
23:05:57[Saint]still need to test on Nano1g
23:08:32 Quit slooopy (Ping timeout: 255 seconds)
23:09:20[Saint]Buschel: Yep.
23:09:23[Saint]commit it.
23:09:37[Saint]and, might I say....*excellent* work my friend.
23:10:12Buschelthx :)
23:10:33Buschelno lcd_type 2 was tested...
23:10:38Buschelamiconn: you there?
23:12:25Buschel[Saint]: and thank you for your excelllent support.
23:12:35Buschelnext step will be the sleep mode
23:12:40[Saint]I wouldn't call it excellent ;)
23:12:44Buschelbut not today... ;)
23:13:06[Saint]No, you've achieved more than enough for one day.
23:13:16[Saint]I can't thank you enough, really.
23:13:45Buschelwell, not submitted yet :P
23:14:01[Saint]I am quite certain it would be safe to commit this.
23:14:13Buschelmaybe amiconn's iPod color uses a lcd_type 0 or 2 for a last check
23:14:37Buschelif there is none available for test I will submit
23:14:40[Saint]perhaps it's a timing thing, but I can never seem to get hold of him.
23:15:06[Saint](I wanted to confirm/deny a few Color issues I found early on)
23:17:29 Quit bertrik (Quit: :tiuQ)
23:17:34 Quit esperegu (Read error: Connection reset by peer)
23:20:36 Join slooopy [0] (
23:22:51[Saint]Buschel: If you're not going to commit it, could you add it to the FS task I added about screen issues?
23:23:00[Saint]s/it/a full patch/
23:23:12Buschelalready did this a few minutes ago
23:23:24[Saint]Aha, awesome.
23:24:53 Quit lostlogic (Quit: leaving)
23:27:54 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
23:29:23 Quit thomasjfox (Remote host closed the connection)
23:32:32Buschelgood night
23:32:35 Quit Buschel (Quit: ChatZilla 0.9.86 [Firefox 3.6.13/20101203075014])
23:34:38 Quit slooopy (Ping timeout: 240 seconds)
23:34:47 Join markun [0] (
23:34:47 Quit markun (Changing host)
23:34:47 Join markun [0] (~markun@rockbox/developer/markun)
23:37:04 Join kevku [0] (~kevku@2001:7d0:0:f000::135d)
23:42:25 Join JesusFreak316 [0] (
23:47:56 Join komputes [0] (~komputes@ubuntu/member/komputes)
23:47:56 Join slooopy [0] (
23:47:56 Quit JdGordon| (Quit: leaving)
23:48:28 Quit froggyman (Quit: Ex-Chat)
23:48:38 Join froggyman [0] (~seth@
23:48:38 Quit froggyman (Changing host)
23:48:38 Join froggyman [0] (~seth@unaffiliated/froggyman)
23:52:20 Quit cjcopi (Ping timeout: 264 seconds)

Previous day | Next day