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

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

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

#rockbox log for 2009-12-07

00:01:08***Saving seen data "./dancer.seen"
00:01:57 Quit angelwolf71885 ("ChatZilla 0.9.85 [Firefox 3.5.5/20091102152451]")
00:02:16 Join angelwolf71885 [0] (
00:06:06 Quit alex` (Remote closed the connection)
00:07:39 Join liar [0] (n=liar@
00:12:09 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
00:17:32 Quit domonoky1 (Read error: 104 (Connection reset by peer))
00:35:00 Join chrism [0] (
00:36:03 Join GeekShado_ [0] (
00:38:20 Join GeekShad__ [0] (
00:48:53 Quit angelwolf71885 ("ChatZilla 0.9.85 [Firefox 3.5.5/20091102152451]")
00:57:12 Quit GeekShadow (Read error: 110 (Connection timed out))
00:57:51 Quit GeekShado_ (Read error: 110 (Connection timed out))
01:01:11 Join froggyman [0] (
01:02:10kugeln1s: (for the logs) are you sure the taking the result is correct in your last commit? remember that the result is what would've been written if the buffer was big enough
01:02:17kugelnot what it actually wrote
01:09:50 Quit flydutch ("/* empty */")
01:19:43 Quit klapaucjusz ("Bedtime")
01:30:50 Quit Thundercloud (Remote closed the connection)
01:36:43kkurbjunJdGordon, I'm logging those notes about the new artifacts I'm seeing
01:36:57kkurbjunturns out you can see the sbs updating in the plasma demo on the sim too
01:38:14 Join angelwolf71885 [0] (
01:41:37 Quit FlynDice (Remote closed the connection)
01:42:11 Quit angelwolf71885 (Client Quit)
01:42:30 Join angelwolf71885 [0] (
01:50:48 Join tony__ [0] (
01:50:58tony__Hey guys.. I've installed iloader on 2nd gen nano. I also unzipped the latest RockBox for nano 2g into iloader directory and renamed rockbox.ipod to rockbox.bin. When I get to iloader screen and choose RockBox, the logo shows up and stays there (i.e., hangs). Any idea?
01:50:58tony__ anyone?
01:52:15tony__I'm not sure whether this question should go to #rockbox-community instead. Any insight?
01:52:44Torneit goes here, but you will probably have to wait
01:53:14tony__ok thx
01:54:56 Quit MethoS- (Remote closed the connection)
01:59:14Lloreantony__: If it happens during iLoader or with iLoader it's something you should be talking to the people who write iLoader about
01:59:24LloreanHave you tried starting Rockbox the way our instructions say to do it?
02:00:30angelwolf71885he was just on linux4nan but i sent him here because it seemed like a RB issue
02:00:49tony__I tried the linux4nano wiki tutorial
02:00:56liartony__: you are doing it wrong, you have to unzip the zip file direct on the ipod(not in a subdirectory)
02:01:09***Saving seen data "./dancer.seen"
02:01:31liartony__: and you dont have to rename rockbox.ipod to rockbox.bin, iloader will find your rockbox.ipod
02:01:46liari think it even does not work if you rename it to rockbox.bin
02:02:02tony__liar: I tried that too, but it didn't work.. tried both rockbox.ipod and renamed it to rockbox.bin.. when I go to RockBox from bootloader it just restarts the iloader
02:02:03Lloreantony__: Did you try *our* instructions
02:02:31liartony__: dont rename the file
02:02:33LloreanIf you're trying to use other tools or software than our own and it's failing, you need to talk with them. If you try it our way and it still fails, then we can start working on it.
02:02:42tony__Llorean, not yet
02:02:48Lloreanliar: Please don't talk him through how to use some other project's software here.
02:02:54 Join phanboy4 [0] (
02:02:54tony__Llorean, you're talking about
02:02:55LloreanKeep this channel on the topic of Rockobx.
02:02:58Lloreantony__: Yes.
02:03:15tony__Ok, I will try that now. Thanks!
02:04:46linuxstbtony__: The official manual links are down - you can get the Nano2G manual here:
02:05:27tony__Thanks, I'm on it.
02:06:02linuxstbtony__: But before you start you should remove iloader - i.e. restore your ipod so it just has the Apple firmware.
02:06:50tony__linuxstb, Any idea on how to do it?
02:07:09tony__Sorry, I am a newb when it comes to this.
02:07:21linuxstbThen I guess the easiest way is to restore your ipod with itunes.
02:11:34 Nick fxb is now known as fxb__ (
02:15:24tony__It works! Thanks guys! I extracted the directory contents of .rockbox into nano's root the first time. The linux4nano tutorial is a bit ambiguous (they talk about rockbox.bin copied into /iloader). Anyway, it works. Thanks!
02:15:48angelwolf71885glad to hear it
02:16:26 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
02:16:54 Quit froggyman (Read error: 54 (Connection reset by peer))
02:28:57 Join arohtar [0] (n=faemir@
02:29:24 Join GeekShado_ [0] (
02:32:39 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
02:33:44 Join dmb [0] (n=Dmb@unaffiliated/dmb)
02:37:18 Quit faemir (Read error: 110 (Connection timed out))
02:47:14 Quit GeekShad__ (Read error: 113 (No route to host))
02:47:42 Quit tony__ ("Ex-Chat")
02:49:04 Join froggyman [0] (
02:49:29 Quit GeekShado_ (Read error: 113 (No route to host))
02:57:20 Quit GeekShadow (Read error: 113 (No route to host))
03:10:13 Join KBH [0] (
03:12:04Unhelpfulgeneric bitscan function for codec lib, supporting log2- and clz-like output, optionally mapping 0 input to 0 output:
03:12:18 Part froggyman
03:13:18 Quit angelwolf71885 ("ChatZilla 0.9.85 [Firefox 3.5.5/20091102152451]")
03:13:35Unhelpfuland patches to replace av_log2 with a #define using it, as well as floor_log2 and wl_min_lzc in libfaad and count_leading_zeros in libalac. planning to commit once benchmark and crc results are verified for aac-he and alac, unless there are some style suggests or other objections...
03:13:36 Join hebz0rl_ [0] (
03:15:16 Quit hebz0rl (Read error: 60 (Operation timed out))
03:24:17 Quit HBK (Read error: 110 (Connection timed out))
03:24:27 Nick KBH is now known as HBK (
03:24:46 Quit kugel ("exit(0);")
03:37:40 Join angelwolf71885 [0] (
04:01:12***Saving seen data "./dancer.seen"
04:15:03 Quit angelwolf71885 ("ChatZilla 0.9.85 [Firefox 3.5.5/20091102152451]")
04:15:49 Quit hebz0rl_ ("Ex-Chat")
04:20:28 Join Rondom_ [0] (
04:34:14 Quit Rondom (Read error: 110 (Connection timed out))
04:48:08 Join rvvs89 [0] (
04:55:54 Join DerPapst [0] (
04:59:43 Quit DerPapst (Client Quit)
05:13:47 Quit DerPapst1 (Read error: 110 (Connection timed out))
05:14:08 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
05:19:44Unhelpfulanybody able to guess why gcc would generate code like this?
05:20:55Unhelpfulam i somehow wrong, or does it do the compare, then the shift, and then r2 = ip & 0xff00, but with the last op split into two opposite conditions, so it's done either way but wastes an instruction...
05:32:45JdGordonkkurbjun: thats a damn nice theme you've got there :p
05:33:33 Join funman [0] (n=fun@rockbox/developer/funman)
05:38:08kkurbjunI was trying to test the sbs
05:40:24JdGordonI tihnk the delay at boot is because there is lots of shit to do, and there is a new lcd clear which is now happening...
05:41:22kkurbjunyeah, the clear seems to be happening too fast, the player used to show the rockbox logo while it was loading stuff up, but now it clears the logo and shows a partially rendered sbs with no menu,
05:43:01kkurbjunon the target it looks more strange because it hasn't loaded the sbs or graphics, so it clears and shows the default statusbar
05:43:05JdGordonhow bad would it be to add a fwe more show_splash()'s in main()?
05:43:10kkurbjunon a black screen
05:44:42kkurbjunI think that should be avoided if possible - it would slow down the boot process (although it would only be by a small amount)
05:44:59JdGordonyeah I know... nice and dirty solution thouhg :)
05:45:41JdGordonah the sbs updating is just the scrolling lines?
05:45:54kkurbjunyeah, you only see the scrolling lines
05:46:34JdGordonright.. ok... at least I know what the issue is :)
05:48:10 Quit Horschti ("Verlassend")
05:58:04 Join angelwolf71885 [0] (
06:01:16***Saving seen data "./dancer.seen"
06:19:17 Quit angelwolf71885 ("ChatZilla 0.9.85 [Firefox 3.5.5/20091102152451]")
06:21:03 Quit phanboy4 (Read error: 54 (Connection reset by peer))
06:21:24 Join phanboy4 [0] (
06:23:12 Quit panni_ ("( :: NoNameScript 3.81 :: )")
06:25:54 Quit tchan (Read error: 60 (Operation timed out))
06:28:22 Join tchan [0] (n=tchan@lunar-linux/developer/tchan)
06:28:52 Quit pixelma (Read error: 60 (Operation timed out))
06:28:54 Join pixelma [0] (i=quassel@rockbox/staff/pixelma)
07:15:33 Join angelwolf71885 [0] (
07:17:04 Quit CaptainKewl (Remote closed the connection)
07:18:53 Quit angelwolf71885 (Client Quit)
07:51:29 Part toffe82
07:54:04 Join stoffel [0] (
07:59:02 Quit efyx_ (Remote closed the connection)
08:01:18***Saving seen data "./dancer.seen"
08:22:25 Join Tomis [0] (n=Tomis@
08:28:28 Join Zagor [242] (n=bjorn@rockbox/developer/Zagor)
08:29:41 Join einhirn [0] (
08:34:07 Join Rob2223 [0] (
08:37:13 Quit rvvs89 (Read error: 110 (Connection timed out))
08:40:31 Quit Rob2222 (Read error: 60 (Operation timed out))
08:49:46linuxstbZagor: Have you changed the daily builds to only keep three days worth of builds?
08:50:48Zagorin mid-september
08:51:57linuxstbI guess that proves they're not needed then...
08:57:20 Join flydutch [0] (
08:59:59pixelmaI wanted a few more back when early USB on Archos was broken again and none of the dailies reached as far back (and in 3.4 it was broken too)
09:00:19pixelmamaybe 4 weeks ago
09:00:53pixelmaa 3.4.1 release would have been a nice alternative
09:04:06 Quit funman ("free(random());")
09:05:10JdGordonwould anyone have problems with this in the LCD drivers?
09:08:21amiconnHackaround for buggy app layer :(
09:08:46JdGordonnot a buggy app layer... merely a hack
09:09:12JdGordonthat fixes my gui artefacts problems.. but doesnt stop the display flickering sometimes :(
09:09:38amiconnImo this hack clearly works around a buggy app layer
09:09:48amiconnDon't update at the wrong time to start with
09:09:50 Join KBH [0] (
09:10:25JdGordonthe problem is I want to clear the whole screen but not update untill something is ready to, but when it is, it only updates a viewport
09:10:32*JdGordon is very open to suggestions
09:13:48JdGordonhack, flickers, or artefacts.. whats the less evil?
09:15:26 Quit Tomis ()
09:15:34amiconnNo hack, no flicker, and no artifacts
09:15:41amiconnThis all used to work...
09:16:10JdGordonyes, its amazing how stupidly sdimple things work, and complicated stuff doesnt....
09:17:43JdGordonof course we could just go back to doing a full screen update every time instead of a viewport only...
09:19:34amiconnIf complicated stuff doesn't work, it's usually because it is too complicated. Viewports are meant as a method to simplify apps/ ...
09:21:36JdGordonsimple -> simpler+betterer -> complicated and awesomer
09:21:53amiconnThe last step is a wrong one imo
09:22:35amiconnIt leads to bloatedness and bugginess
09:23:00amiconnThe various viewported gui elements are all drawn from a single thread. I don't see why it shouldn't be possible to synchronize them nicely
09:23:11*JdGordon doesnt bother with this argument... we know you dont want any visual niceties (except the animation in the statusbar!)
09:23:24amiconnFundamentally wrong
09:24:30amiconnOf course I want my rockbox (and other things) to look nice & clear.
09:24:48JdGordonyou have the sbs skin updating at XHz... the GUI for whatever screen you are in updating at YHz, and dead area never updating...
09:25:01JdGordonif you have a workable idea how to do it then go ahead of let us know...
09:25:06amiconnWhy does the sbs update itself?
09:25:25amiconnImo the main screen should trigger statusbar updates
09:25:41 Join Bagder [0] (
09:25:46JdGordonit updates in the action handler... once per loop (usually)
09:26:23 Quit HBK (Read error: 110 (Connection timed out))
09:26:28 Nick KBH is now known as HBK (
09:26:45JdGordonwhich is not the problem anyway... the problem is dead area doesnt get updated because we dont do full screen updates if we can avoid it
09:27:47amiconnMy focus regarding looks is on usability though. That means I don't squeeze useless graphics on a tiny screen, but rather want my track info nice and readable, accompanied by information about the state of playback and device status.
09:27:56JdGordonthe only place this really happens is when the splash is shown
09:28:22amiconnArtifacts like status bars going invisible, or statusbars overlaying the splash screen are quite annoying.
09:28:38JdGordonand when re-enabling the theme after it was disabled
09:28:45amiconnBut this didn't happen before .sbs, so the problem is in there
09:29:51 Join Thundercloud [0] (
09:30:08amiconnSo you need to do a full screen update if the main part changes. The main part knows when that happens, and if it would also trigger sbs updates, it would automatically be in sync
09:30:34 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
09:31:05JdGordonit wouldnt update often enough, and you end up having to do full screen updates
09:32:23JdGordonscreens only need to worry about updating themselves... they cant know if a sbs or statusbar is being displayed or not
09:32:55amiconnThey can just call a function that does the update... or nothing if there's nothing there
09:33:21*amiconn doesn't really understand the dead area problem though (and doesn't observe it either)
09:33:39amiconnWho would throw away parts of that already tiny screen?
09:33:48JdGordonnot all screens are tiny
09:34:17JdGordoneven the beast and ipods have a big enough screen where a 5 pixel margin is nice
09:35:10JdGordonamiconn: They can just call a function that does the update... or nothing if there's nothing there <- yes thats happening now.... but the calling screen still doesnt want to be told if it needs to do a full screen update or a partial one
09:35:19*amiconn has several ipods, but they all have tiny screens
09:35:36amiconnI do know the iPod Video screen - not really big either
09:36:00amiconnThe calling screen doesn't need to be told... it should know by itself
09:36:13JdGordonno it shouldnt!
09:36:40JdGordonit only cares about its rectangle that its drawing in
09:36:52amiconnIt knows whether it was freshly opened or whether it's already running for some time
09:37:08JdGordonyes, and that adds complexity
09:37:21amiconnI.e. just make the first update a full one, and be done
09:37:30JdGordonlots of it in the case of the wps where it has multiple "exit/entry" points
09:38:25amiconnAnd if parts like sbs are updated from within the main loop (with a handler call), it shouldn't happen that the sbs updates before the main part is ready
09:39:16JdGordonactually the sbs is almost always drawn before the main loop... its called in get_action()
09:39:51JdGordonwhich is just about always at the start of the loop
09:39:59 Quit phanboy4 (Read error: 113 (No route to host))
09:41:07amiconnAnd that's very likely a reason for many of those nasty artifacts
09:41:25JdGordonthats been there for a long time already
09:41:34JdGordonand not likely to change
09:42:07amiconnSo why didn't it cause those artifacts before sbs then?
09:42:30JdGordonbecause the statusbar viewport was always known
09:42:36JdGordonthe list viewport was always known
09:42:45JdGordoneverything is known at compile time
09:43:10JdGordonsbs and ui viewport change all that
09:43:14amiconnI'M *not* talking about dead areas
09:44:04JdGordonartefacts being what then?
09:46:34 Quit Thundercloud (Read error: 54 (Connection reset by peer))
09:54:09 Quit jae ("leaving")
09:54:53 Join maruk [0] (
09:58:23Zagortoday is target rename day
09:58:32 Join Tomis [0] (n=Tomis@
09:58:40JdGordonhip hip hooray!?
09:59:57 Quit parafin (
09:59:57 Quit ps-auxw (
09:59:57 Quit crwll (
09:59:57 Quit blithe_ (
09:59:57 Quit Hadaka (
09:59:57 Quit LinusN (
10:00:06 Join blithe [0] (
10:00:09NJoinLinusN [0] (n=linus@rockbox/developer/LinusN)
10:00:27NJoincrwll [0] (
10:01:19***Saving seen data "./dancer.seen"
10:02:47CIA-6New commit by zagor (r23880): Rename targets to be consistent in all places. (FS #10819)
10:03:18CIA-6New commit by zagor (r23881): Rename targets to be consistent in all places. (FS #10819)
10:04:59 Join Naked [0] (
10:05:10 Nick Naked is now known as Hadaka (
10:05:23BagderZagor: I fiddled with the manual build script tonight and it should now log the full round outputs so that we can figure out why lots of manuals don't show up anymore
10:12:18linuxstbZagor: Why the 3.4 branch?
10:17:08linuxstbZagor: is this as bad as it seems? i.e. is it a lot of work to do the same thing in trunk?
10:17:31ZagorI meant to do it in trunk. the branch was a mistake. reverting now.
10:17:55CIA-6New commit by zagor (r23882): Reverted r23881 mistakenly committed to branch.
10:18:04linuxstbYes, I guessed that. I'm just wondering how hard it will be to do the same to trunk.
10:18:22Zagorhopefully not too bad
10:18:41Bagderhm is broken now...
10:19:18Zagoryes, due to the branch snafu
10:19:20Bagderit's probably one of the perl things
10:28:50Zagorblah, who is good with svn merging? I want to merge change 23881 onto trunk. apparently "svn merge -c 23881 ." isn't the right thing to do.
10:34:08Bagdercan't you just do a svn diff | (cd [path] && patch -p0) ?
10:34:27Zagorno, svn diff doesn't handle renamed/moved files
10:51:14amiconnJdGordon: (1) Statusbar appearing on the splash screen. (2) Statusbar disappearing in the menu
10:52:41amiconn(3) Statusbar drawing over lists (space not accounted for)
10:52:57amiconnThings like these... I on't have any dead areas
10:54:20 Join FlynDice [0] (
10:55:54 Join rvvs89 [0] (
10:57:53Zagorpixelma: what is the \playerman macro in the manuals? player name?
10:58:32Zagorand \playertype?
10:59:12Zagorlooks like my script went a little overzealous with the manuals...
10:59:27Zagor"Connect your \archosplayertype{} to the computer"
11:01:09pixelmacan't remember seeing this macro, would have to look for that myself.
11:01:37pixelmaand I'm at work, so it's a bit hard for me to do now...
11:03:02 Quit chrism (Read error: 104 (Connection reset by peer))
11:26:32linuxstbZagor: You've probably worked it out by now, but \playerman looks like the manufacturer (or "Ipod"), and \playertype is the model
11:28:26Zagoryeah. thanks anyway.
11:31:37pixelmabut \archosplayertype looks odd, more like a leftover
11:32:21Zagorno, it was my name changes sed that did too much. it changes the \opt{player} lines to \opt{archosplayer}. but it wasn't supposed to change \playertype
11:32:36ZagorI've fixed it now
11:33:06pixelmaoh, then I misunderstood and could have answered the question earlier :/
11:33:14Zagor:-) no problem
11:43:12 Join ps-auxw [0] (
11:47:06 Quit Tomis (Read error: 110 (Connection timed out))
11:50:51BagderZagor: are you working on fixing the site? is it pending your merge/renaming things?
11:51:31Zagoryes I do. the updated site uses a new target list from trunk/tools/ which isn't there yet.
11:52:40ZagorI'm considering reverting the web change though. the merge takes a while.
11:55:00 Quit z35 (Read error: 60 (Operation timed out))
11:55:18 Join z35 [0] (
11:58:56 Join watto [0] (n=watto@
12:01:23***Saving seen data "./dancer.seen"
12:16:53 Quit BHSPitLappy (Read error: 110 (Connection timed out))
12:23:18 Join Lynx_ [0] (
12:36:33 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
12:36:59 Quit stoffel (Remote closed the connection)
12:38:31 Join Guest23293 [0] (
12:39:01 Quit n17ikh (Nick collision from services.)
12:39:09 Quit Guest23293 (Client Quit)
12:39:32 Join n17ikh [0] (
12:52:02 Join Tomis2 [0] (n=Tomis@
13:02:40 Part Bagder
13:03:05 Quit liar (Read error: 104 (Connection reset by peer))
13:03:16 Join liar [0] (n=liar@
13:09:13 Join Bagder [0] (
13:19:15CIA-6New commit by zagor (r23883): Rename targets to be consistent in all places. (FS #10819). Take 2.
13:20:59CIA-6New commit by zagor (r23884): Added missing file.
13:24:44topikwill that chance bring back the current build page Zagor ?
13:24:48topikchange even
13:25:02Zagortopik: yeah, in a minute. some fixes remain.
13:26:19topikno rush, was just looking around for info on AMS usb and then wanted to look at sansa c200/e200 pics
13:29:48 Part Bagder
13:34:41 Join Bagder [0] (
13:38:16 Join tomers [0] (
13:38:26tomerszagor: Have you considered s/iriver/iRiver/ ?
13:38:47Bagderthat's not how they use it
13:38:53Bagderthey vary
13:38:59Zagoralso, all target names are lowercase. mixing case just adds trouble.
13:40:50tomerszagor: According to, it is produced by Cowon, so why not s/iAudio M3/Cowon iAudio M3/ and such?
13:41:08tomerszagor, Bagder: I'm refering to which shows the model names
13:42:33tomersbagder: Agreed. (on the 'iriver')
13:43:08Bagderwe call them ipods even though apple makes them
13:43:13 Join froggyman [0] (
13:43:16Bagdercowon's iaudios are only called iaudio
13:44:09BagderI think the important thing here is consistency anyway
13:44:10 Join kugel [0] (n=kugel@rockbox/developer/kugel)
13:44:24BagderI mean between different scripts
13:44:31kugelZagor: looks strange
13:45:12kugelit seems the first half is duplicated
13:45:42Zagorheh, yeah. fixing.
13:45:50 Quit tomers ("CGI:IRC")
13:47:06kugel\o/ for moving config-<target>.h into a separate dir btw :)
13:47:47Bagderthen they could even become config/<target>.h ...
13:48:07ZagorBagder: they are
13:48:20*Bagder pretends he was aware of that
13:51:36CIA-6New commit by zagor (r23885): Sort names case insensitive. Remove duplication.
13:55:58kugelZagor: I don't think we want m200v4 on the download page
13:57:17kugelI think the philips and ondas aren't right there too
13:57:31Zagorthey're not supposed to be there
13:57:40 Part froggyman
13:58:35kugelah gone :)
13:58:40*kugel is slow
14:01:25***Saving seen data "./dancer.seen"
14:04:59 Join teru [0] (
14:06:48kugelZagor: ipodnano2g and e200v2 wrongly appear on the release download page
14:06:49Unhelpfulwould somebody with a coldfire target mind giving a try? main things to look for is that a codec that uses the log2 or clz functions 1) produces the same CRC in test_codec checksum mode 2) is faster with the patch than without. alac or flac are probably the easiest to test.
14:08:03CIA-6New commit by zagor (r23886): sansae200v2 and iponano2g are unstable targets
14:08:07Zagorkugel: thanks
14:09:17BagderZagor: maybe you can make variables like $stable and $unstable to set in that hash to make it more readable than 2 and 3
14:09:57Bagderor even use the actual strings
14:10:22ZagorI like integers, to be able to quickly sort >1 for instance
14:10:31mc2739Zagor: also e200r is on daily and release pages - e200r uses e200 build
14:11:43Unhelpfulgevaerts or AlexP maybe if you have the chance...?
14:14:02 Join parafin [0] (
14:15:30Zagormc2739: hmm, it shouldn't be a separate build at all then should it? it's only ever used in tools/configure to build the boot loader.
14:16:38mc2739Zagor: That is correct.
14:17:53kugelwell, they're separate models in theory, it would make sense if it appears separately on the download page too (but linking to the same binary)
14:24:37gevaertsUnhelpful: maybe tonight
14:28:24kugelZagor: so, what if I wanted to make the fuze stable. would I only need to change and does the page still need a manual update?
14:28:37Zagoryes and yes
14:32:00 Join webguest74 [0] (
14:32:44 Quit webguest74 (Client Quit)
14:33:55*kugel thinks we should decide on whether to (generally) accept the/a default sbs for cabbiv2
14:34:00kugelI'll vote for it
14:34:29Unhelpfulgevaerts: thanks. no hurry. :)
14:38:49pixelmakugel: if it hasn't changed, the one in the forum thread is missing some important info in the recording screen (and displays repeat and shuffle info instead which is not necessary there)
14:39:48kugelright :/
14:40:10CIA-6New commit by zagor (r23887): A few more target rename fixes
14:40:16pixelmaand there's also only one port so far - just my thought that it isn't ready yet
14:42:06gevaertsI'm not sure if we should require it to be ready for all screen sizes, but it should be complete for all screen sizes that are ready
14:44:55kugelpixelma: well, I was just saying that we should decide on the general idea of a default sbs not on that particular one
14:53:59CIA-6New commit by zagor (r23888): Corrected ondio language wildcards.
15:14:04CIA-6New commit by zagor (r23889): Corrected boot output filenames
15:18:45Zagorhmm, gigabeatfx reports plugin ram full on doom.rock
15:28:00CIA-6New commit by zagor (r23890): Target name fixes
15:38:33AlexPUnhelpful: Tonight sure, but I'm at work atm
15:38:53 Join Omlet [0] (
15:38:59AlexPkugel: In principle yes, when the info is there as it is now
15:52:46kugelZagor: "corrected"? I think the boot filenames are wrong now :/
15:53:11kugeland inconsistnet it appears
15:53:33kugeloh no, not inconsistent
15:53:37Zagorthey are now consistent with tools/configure
15:54:00ZagorI have not renamed all boot files, since I don't know if any of them have outside dependencies
15:54:52kugelwell, the samsa ones don't
15:55:49Zagor(I meant I have not renamed _any_ boot files)
15:55:59Zagorat least I think not :)
16:01:27***Saving seen data "./dancer.seen"
16:05:26 Join roolku [0] (
16:06:12roolkuZagor: this line in configure looks suspicious: if test "$modelname" != "gigabeatf" -a "$t_manufacturer" != "as3525" -a "$ARG_ARM_EABI" != "1"; then
16:07:53CIA-6New commit by zagor (r23891): Gigabeat f => fx fix
16:15:56 Join jgarvey [0] (
16:21:48 Quit teru ("Quit")
16:29:36 Quit Sajber^ ("Leaving.")
16:29:45 Quit roolku ()
16:40:42 Join DerPapst [0] (
16:41:28 Join Strife89DS [0] (n=nds@
16:48:29 Quit DerPapst ("Leaving.")
16:49:48 Join toffe82 [0] (n=chatzill@
16:54:00 Quit Strife89DS ("ClIRC - IRC client for Nintendo DS")
17:00:53 Part LinusN
17:01:30 Quit Zagor ("Don't panic")
17:04:25 Join roolku [0] (
17:05:49 Quit kkurbjun (Read error: 110 (Connection timed out))
17:09:16roolkuZagor: the file tools/checkwps/targets.txt still contains the old modelnames - is this correct?
17:13:00CIA-6New commit by roolku (r23892): make tpj1022 bootloader compile again
17:13:16kugelroolku: I don't think so
17:17:18 Join calman_ [0] (
17:18:13 Join phanboy4 [0] (
17:18:44kugelroolku: targets.txt is for
17:19:28 Join panni_ [0] (
17:20:39roolkukugel: yes. It's probably easier for Zagor to run his sed script instead of fixing it manually
17:21:17kugelroolku: I think it should be changed so that it's not needed at all
17:21:28kugeli.e. what the release scripts have now
17:21:57*linuxstb wonders about - shouldn't the names be bootloader-$modelname.sansa ?
17:21:59roolkukugel: alternatively discontinue since checkwps is a build target now
17:25:35 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
17:25:46 Join DerPapst [0] (
17:26:46linuxstbZagor: (for the logs) "ipod4gray" is still in rbutil.ini - is that right? Also, you haven't symlinked bootloader-ipodnano1g.ipod on the download server.
17:27:02 Join kkurbjun [0] (
17:28:49 Join funman [0] (n=fun@rockbox/developer/funman)
17:30:54kugelroolku: works by doing configure && make for each tarfget already
17:32:33 Quit antil33t (Read error: 104 (Connection reset by peer))
17:32:39 Join antil33t [0] (
17:43:48 Quit Galois (Remote closed the connection)
17:44:07 Join Galois [0] (
17:47:14 Quit roolku ()
17:47:36 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
17:48:08 Part Bagder
17:48:33 Part pondlife
17:59:19 Join ded [0] (n=user@luscinia.Princeton.EDU)
18:01:29***Saving seen data "./dancer.seen"
18:02:49dedI'm trying to help someone who is in rural Ethiopia and hence is extremely limited in size of files he can download (and will soon be out of internet access). He has rockbox v3.4 installed but is getting "Connect to your computer. Use iTunes to restore." What would his best options be?
18:07:26 Nick Tomis2 is now known as Tomis (n=Tomis@
18:10:33 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
18:14:47linuxstbded: That normally means the disk is corrupted in some way that the Apple bootloader (stored in ROM) is not happy about. This page may help -
18:14:58linuxstbI meant
18:19:37 Quit Topy44 (Read error: 104 (Connection reset by peer))
18:19:38 Join Topy [0] (
18:21:13 Join Kitar|st [0] (
18:22:25 Quit calman_ (Client Quit)
18:23:31dedlinuxstb, thanks. Unfortunately he only has access to Windows machines. Is using rbutil to reinstall the bootloader likely to help? On my ipod I have completely removed the Apple firmware (under linux). How would one do that under Windows?
18:26:19 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
18:30:20linuxstbded: It's not straightforward under windows (and I'm not the best person to ask for help there...) You either need to use itunes, or find a way to do those steps on a Windows computer.
18:31:52krazykitdd is available for windows (dd.exe) and fat32format will do as the name suggests. i can't really support either software, as i've never used them
18:33:34 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
18:37:04 Quit phanboy4 ("Leaving")
18:39:23 Quit Kitr88 (Connection timed out)
18:39:25 Join Kitr88 [0] (
18:40:28 Quit Kitar|st (Read error: 110 (Connection timed out))
18:44:07dedOK, thanks.
18:46:28 Join tomers [0] (
18:46:39 Join Lynx [0] (
18:47:06 Nick Lynx is now known as Guest47578 (
18:47:34 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
18:54:27 Part pondlife
18:55:29*LambdaCalculus37 is close to finishing the GoGear SA9200 plugin keymaps
18:57:40LambdaCalculus37Still need 128x160 cabbiev2 theme, though. :)
19:00:04 Quit funman ("free(random());")
19:00:45 Join archivator [0] (n=archivat@
19:01:59LambdaCalculus37pixelma: Can you help me get a cabbiev2 theme for the GoGear SA9200?
19:02:54 Quit maruk ("Leaving.")
19:03:21 Quit Lynx_ (Connection timed out)
19:13:11 Quit HellDragon (Read error: 104 (Connection reset by peer))
19:13:29 Join HellDragon [0] (
19:21:08 Join einhirn [0] (
19:23:27CIA-6New commit by rmenes (r23893): Yet more GoGear SA9200 plugin keymaps. Almost done!
19:24:30 Quit LambdaCalculus37 ("Fwump")
19:25:04 Quit pixelma (Nick collision from services.)
19:25:06 Join pixelma_ [0] (i=quassel@rockbox/staff/pixelma)
19:25:25 Nick pixelma_ is now known as pixelma (i=quassel@rockbox/staff/pixelma)
19:25:32 Part ded ("ERC Version 5.3 (IRC client for Emacs)")
19:25:44 Join amiconn_ [0] (
19:26:13 Quit amiconn (Nick collision from services.)
19:26:49 Quit einhirn (Read error: 104 (Connection reset by peer))
19:27:07 Join einhirn [0] (
19:32:18 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
19:32:26 Quit archivator (Remote closed the connection)
19:36:47 Quit tomers ("ChatZilla 0.9.85 [Firefox 3.5.5/20091109125225]")
19:49:09 Join JdGordon| [0] (n=Miranda@nat/microsoft/x-naikfzinunyjruqg)
19:49:41 Quit parafin (Nick collision from services.)
19:49:51 Join parafin [0] (
19:49:58 Join PaulJam [0] (
19:52:53 Join Strife89 [0] (n=michael@
19:56:14 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
19:58:06 Join Omlet05 [0] (n=omlet05@
20:01:34***Saving seen data "./dancer.seen"
20:10:32 Join mrtok1 [0] (
20:12:16mrtok1Q: Is someone here to help troubleshooting profiling setup? followed these instuctions ( but still get no meaningful results ...
20:15:05 Quit Omlet (Read error: 110 (Connection timed out))
20:21:33 Quit bmbl ("Bye!")
20:42:31*Llorean thinks is interesting
20:42:42 Join solexx [0] (
20:42:46LloreanIt might explain those occasional reports by people of it taking dozens of seconds or minutes before they can use their player
20:43:44 Join domonoky1 [0] (
20:44:25*JdGordon| doesnt rememeber any of those reports
20:44:57GodEaterJdGordon| hasn't got nearly as many hours logged in the forums as Llorean though ;)
20:45:17LloreanJdGordon|: There have been several people who've mentioned having 10-60 second times before they can start browsing
20:45:35LloreanUsually we recommend enabling dircache, and often that clears it up, but >10 seconds before you can browse without dircache seems somewhat ridiculous
20:47:14gevaertsLlorean: you're saying that maybe the thing that blocks is maybe not just listing the file, but looking up the font file over and over again?
20:47:43LloreanI just always felt that >10 seconds was really out of scale for "just waiting for the disk to spin up"
20:48:02JdGordon|that delay should only ever happen if there is no .fontcache file though
20:48:02 Quit domonoky (Read error: 60 (Operation timed out))
20:48:23gevaertsJdGordon|: or if there are new glyphs
20:48:27LloreanJdGordon|: Or if the root directory already contains more glyphs than the fontcache can deal with?
20:48:39LloreanOr more uncached glyphs than reasonable.
20:49:06gevaertshm, if that's the case, and enabling dircache by default remains impossible because of the various limit issues, maybe a .rockbox-only dircache would be a good idea?
20:49:41JdGordon|isnt the issue loading glyphs?
20:49:44JdGordon|not dircache?
20:50:12gevaertsJdGordon|: loading glyphs requires opening the font file, which requires finding it on disk
20:50:26gevaertsso this sounds like a plausible hypothesis to me
20:50:46 Join phanboy4 [0] (
20:50:52JdGordon|I would have expected the glyph loader to keep the fd open...
20:50:52 Quit phanboy4 (Client Quit)
20:51:11gevaertsI'm not sure. We don't have very many fds
20:51:20*gevaerts tries to check
20:51:29JdGordon|also, does dircache actually help with opening files? or just listing them?
20:51:57gevaertsit should help with both I'd say
20:52:52gevaertshm, you're right. It keeps the file open while caching
20:53:20JdGordon|the mass seeking sounds like a possible issue... but I cant tihnk how else you'd do it..
20:53:33JdGordon|you cant really load 25 glyphs in one hit
20:54:09JdGordon|i mean, sequentially load glyph 0->25... we might not want them all
20:55:04JdGordon|ignore that :p
20:58:50 Quit solexx_ (Read error: 110 (Connection timed out))
20:58:56 Quit BHSPitLappy (Read error: 110 (Connection timed out))
21:00:23gevaertsfonts that are small enough are loaded in one go, without using the cache. I'm a bit surprised about them not working though. The limit is 60000 bytes, and (I thought) popular fonts like 14-Nimbus.bdf are under that
21:01:02gevaertsJdGordon|: by the way, unifont seems to be 8.5MB, not 2 :)
21:01:38JdGordon|hehe slight woops :p
21:02:53*gevaerts wonders if a logging ata driver would be a good idea
21:03:27Dhraakellianwould it be possible for Car Adapter Mode to disable the reboot to OF on USB insertion?
21:04:11Dhraakellianon a target that doesn't have USB in Rockbox, obviously
21:05:09Dhraakellianis that what does?
21:05:54gevaertsDhraakellian: seems so
21:07:09gevaertsit doesn't exactly look like the cleanest way to do this though
21:07:41 Join pamaury [0] (
21:08:12gevaertsthere's actually a related problem. Bootloader USB is very annoying with car adapter mode
21:09:21Dhraakellianwell, making sure my fuze is on before turning the key isn't so bad
21:10:21*gevaerts isn't sure what to think
21:10:45Dhraakellianbut pressing select with gloves on is a wee bit more difficult
21:11:16gevaertsoh, I definitely agree that there is a problem, I'm just not at all sure about what the best solution is
21:12:18DhraakellianI'm guessing the best solution would be to have USB working properly
21:12:29gevaertsit's not that simple I think
21:12:58gevaertsThere are several variables : (a) is the "usb" cable connected to a host complete with USB data, or does it just provide 5V? (b) can we detect the difference between those cables on the target in question?
21:13:23*gevaerts thought there was a (c) as well, but he can't remember now :\
21:13:24DhraakellianI seem to remember my e200v1 being able to tell the difference
21:13:31DhraakellianI could possibly be misremembering
21:13:56gevaertsit does, by going to charging mode and only switching to MSC mode if something happens on the bus
21:14:23gevaertsI can't remember now if all hardware USB bridge targets can see the difference. I wouldn't be surprised if some can't
21:14:33gevaerts(or even all)
21:15:59 Join polobricolo [0] (
21:17:03 Part watto
21:17:41Lloreangevaerts: It'd also be solved by the choice between default USB behaviour.
21:18:01LloreanThough there's still no solution for bootloader USB mode.
21:18:10gevaertswell, that would provide a reasonable workaround
21:18:45gevaertsHaving to toggle two settings isn't that perfect, but it's not the end of the world either, and you can just make .cfg files for each
21:19:02*Dhraakellian ponders going back to rolling his own personal builds with that patch applied
21:19:14LloreanWell, "Car Adapter Mode" relates to pause/play when power is detected, USB mode just relates to USB mode (for car, or for home)
21:19:40*Llorean still also speaks in favor of changing it to *any* button refusing USB mode rather than any single one.
21:20:01Dhraakellianhow about hold?
21:20:17gevaertsLlorean: Car Adapter Mode also predates this power-over-usb nonsense
21:20:44LloreanDhraakellian: Not hold, it's a toggle.
21:21:12DhraakellianLlorean: if hold is set, no reboot vs. if a button is pressed, no reboot
21:21:34Lloreangevaerts: This is true, but having it change USB behaviour (much like having hold change it) would mean they could hook it up to their home PC and have no clue why it's not working because "Car adaptor mode" or even "Car mode" doesn't seem entirely obvious (and would affect non-usb behaviour on targets that say, have a mains charger and don't charge over USB)
21:21:40 Quit mrtok1 ()
21:21:46Dhraakellianmeans that I don't have to worry about whether inserting the cable makes my finger move ever so slightly in such a way that the button is no longer pressed
21:21:50 Join saratoga [0] (i=9803c6dd@gateway/web/freenode/x-dznrevpnclrqpocf)
21:21:53LloreanDhraakellian: I know what you meant, I'm very strongly against Hold doing it.
21:22:25saratogacar adapter mode will never work as expecte don the Fuze due to the way the mkamsboot process works
21:22:39Lloreansaratoga: How's that?
21:22:39saratogaand changing that would be fairly risky, since you'd lose the ability to get back to the OF if anything goes wrong
21:22:48 Quit Topy (Read error: 104 (Connection reset by peer))
21:22:49 Join Topy44 [0] (
21:22:54saratogaif you plug in USB with the device off you'll always get the OF
21:23:02saratogaand changing the bootloader or rockbox won't change that
21:23:12LloreanCan it be changed to "if Rockbox.sansa isn't found, and USB is detected, then go to the OF"?
21:23:22saratogai doubt it
21:23:24LloreanAh well
21:23:26saratogawe don't have disk access
21:23:33saratogaso we can't read the file system yet
21:23:40saratogathe bootloader is the earliest you can read the disk
21:23:40gevaertsLlorean: that assumes bugfree fat and disk drivers that are guaranteed to never crash
21:23:54Lloreansaratoga: Can we detect a charger (rather than proper USB connection) like we used to do on the iPods?
21:24:00Lloreangevaerts: Very true
21:24:12saratogayou'd have to issue a new mkamsboot and probably brick a few players in the process
21:24:29saratogaand i'm not sure if its possible to detect a charger at all in that situation
21:24:44*Llorean is okay with it only working after Rockbox is booted
21:24:48saratogaactually i'm not sure if the fuze can even detect a charger at all
21:24:53gevaertssaratoga: I suspect that this will have to happen eventually anyway. As soon as there is a stable rockbox usb driver, people will complain about this
21:25:05saratogathe OF seems to charge off anything plugged in without bothering to deal with the USb issues
21:25:12DhraakellianLlorean: ditto
21:25:17gevaertsbut yes, bootloader behaviour is a separate problem
21:25:34Lloreangevaerts: On AMS it's the prebootloader thing.
21:25:53gevaertsLlorean: it's the same binary in the end :)
21:26:07Dhraakellianlacing the boots before pulling up by the bootstraps?
21:26:22Dhraakellian</badly extended metaphor>
21:27:11saratogai guess eventually we could make mkamsboot check for a button press only instead of a USB insert
21:27:18gevaertsWe should probably revisit FS #10198. Can anyone remember why that didn't go further?
21:27:40Lloreansaratoga: That sounds fair
21:27:56saratogayou'd have to ask funman how safe that would be though
21:28:10saratogasince i think all of this is on GPIO it should be fairly safe
21:28:17saratogabut i don't know
21:28:23Lloreangevaerts: I really have no idea that stopped. It was pretty much done right? A setting to pick which is default, and any button causes the other case?
21:29:09saratogathat idea seems sensible enough to me (and useful)
21:29:53gevaertsIt interferes a bit with this future settings screen that we'll need when we have MTP, but that's not at all ready yet (pamaury?), so I don't think it should hold things up
21:30:06saratogathe current system is really annoying on the clip
21:30:35saratogasince the buttons are very small and accidentally hitting any other button while trying to force the stiff USB cable in will result in the opposite behavior
21:30:51saratogait actually took me a couple tries the first time to get into charge mode on that player
21:30:55gevaertssaratoga: you should plug in the other end then :)
21:30:57Lloreangevaerts: Future settings screen?
21:31:38Dhraakelliangevaerts: indeed. I've found that it's sometimes easier just to plug in my Fuze, hold select, and *then* turn the ignition
21:32:09Dhraakellianstill a bother to remember, though
21:32:49gevaertsLlorean: we discussed things a few weeks ago. At some point (and it seems that MTP would be that point) we'll need a way for people to choose between USB class drivers. Some controllers won't have enough endpoints anymore to run them all, and MTP and MSC are mutually exclusive in practice (in theory you could have them both on different drives, but that won't work well...)
21:33:08Lloreangevaerts: Wouldn't you do that before you plug in USB though?
21:33:11LloreanRather that at connect time.
21:33:18saratogaworry about that once you get MTP working :)
21:33:30LloreanJust Settings->System->USB->Classes or something
21:33:36gevaertsLlorean: yes. The interfering bit is the setting itself, not the any-handling
21:34:07gevaertssaratoga: exactly :)
21:34:12LloreanI guess I'm not seeing how it interferes.
21:34:27LloreanWouldn't holding a button deny (or allow) all drivers currently enabled in the setting?
21:35:03Dhraakellianset a first choice and a second choice?
21:35:26gevaertsThere always has to be one left, presumably HID. You may be right in that it's a separate thing though
21:35:45Lloreangevaerts: "one left"?
21:35:45gevaerts(HID as in the class driver, not as in actually sending button events)
21:35:59gevaertsLlorean: yes. You can't have a do-nothing USB device
21:36:20LloreanAh. Well, I mean, we do *something* when charging currently, right? just that, I imagine.
21:36:27gevaertswe use HID
21:36:42*pamaury has noticed gevaerts and says that MTP is not ready but seems be functional (understand lacking lots of features) (at least on linux)
21:37:09gevaertsor the old dummy CHARGE_ONLY thing on devices that don't do interrupt, which gives you this nice "now go and look for a driver" dialog
21:38:07LloreanBut yeah, it's something that can be addressed once MTP is in/active since really, charging shouldn't be any different then, you're just deciding what drivers to load for the non-charging case.
21:41:12 Join thegeek [0] (
21:41:54*domonoky1 thinks always using hid as charging mode is fine, but we still need a way to allow users to set some combination of usb classes for the normal mode.. (for example msc +serial)
21:42:46 Quit thegeek (Read error: 104 (Connection reset by peer))
21:42:48domonoky1or mtp+hid+serial if the hardware has enough endpoints.. or every other combination which is possible and sane :-)
21:42:56 Join thegeek [0] (
21:43:17JdGordon|mtp+msc on different partitions!
21:43:24LloreanWhich hardware has the most endpoints?
21:43:37pamauryJdGordon: that would be tricky !
21:43:50 Quit parafin (
21:43:50 Quit shodanX_ (
21:43:50 Quit Utchybann_ (
21:43:50 Quit Tuplanolla (
21:43:50 Quit Torne (
21:43:50 Quit Kopfgeldjaeger (
21:43:50 Quit xavieran (
21:43:50 Quit Zambezi (
21:43:50 Quit jds2001 (
21:43:52JdGordon|pamaury: tricky but not impossible :)
21:44:41NJoinparafin [0] (
21:44:41NJoinjds2001 [0] (n=jds2001@fedora/jds2001)
21:44:41NJoinUtchybann_ [0] (
21:44:41NJoinTuplanolla [0] (n=jani@unaffiliated/tuplanolla)
21:44:41NJoinTorne [0] (i=torne@rockbox/developer/Torne)
21:44:41NJoinxavieran [0] (
21:44:41NJoinKopfgeldjaeger [0] (
21:44:41NJoinshodanX_ [0] (
21:44:41NJoinZambezi [0] (i=Zulu@unaffiliated/zambezi)
21:44:44domonoky1mtp on internal and msc on sdcard :-)
21:44:44gevaertsLlorean: gigabeat s
21:44:50 Join froggyman [0] (
21:44:54JdGordon|mtp on internal... msc on a ramdisk to update rockbox with!
21:45:14gevaertsJdGordon|: technically you can run MTP without storage if I understand things correctly :)
21:45:32gevaertsYou'd use it for battery reporting and things like that
21:46:05pamauryyes you could use MTP without storage
21:47:33gevaertsLlorean: that one has seven in each direction (possibly including control, not sure not). That's enough for serial+HID+MTP+MSC+audio+video and a bit left over I think
21:48:52LloreanIs there an HID keymap that maps 1:1 to the keys used for those buttons in the sim, btw?
21:49:37JdGordon|I cant decide if that is more pointless.. or more awesome
21:49:46gevaertsnot right now. 1:1 isn't really possible in all cases, because of hold and the button that's used to switch between keymaps though
21:51:00LloreanAh, right, hold especially
21:51:11LloreanThere's a button wasted on switching keymaps mid-HID mode?
21:52:39domonoky1combine 1:1 simulator keymaps with a simulator exe on the player and autostart, and it will look like we can export the player display to the pc display :-)
21:53:01Lloreandomonoky1: USB-Video class is for that. :-P
21:53:22domonoky1Llorean: that USB-Video for cheap :-)
21:53:27 Join fml [0] (
21:55:00fmlHello. What do you think about changing the font on the front page of the manual? To a sans serif font. All chapters are titled using such font, and I think that it better suits the document title as well.
21:56:05 Join gitster [0] (
21:56:20 Part gitster ("ERC Version 5.3 (IRC client for Emacs)")
21:56:59JdGordon|how do people feel about the issues in this comment? me and Karl both agree that they are still less bad than svn...
21:58:32kugelsvn isnt so bad
21:58:38kugeljust the plugins
21:59:13 Quit thegeek_ (Read error: 110 (Connection timed out))
22:00:56JdGordon|well how important are the plugins... this makes their menus much better
22:01:16kugelI mean it's only plugins in SVN
22:01:36***Saving seen data "./dancer.seen"
22:01:36gevaertsJdGordon|: I'll take your word for it. I use the thing for listening to music, and I only ever use the play, pause, stop and volume keys :)
22:01:38 Join krabador [0] (
22:02:03kugeland people many using the release are going to care about plugins
22:02:18JdGordon|so the onyl actual issues that people would see is splash artefacts in dead space
22:02:33LloreanAnd apparently wrong colors in plugins?
22:02:35JdGordon|havnt come up with a clean way of dealing with that yet... and/or flickering
22:02:37LloreanIs this capable of making them unusable?
22:02:42gevaerts"dead" means no viewports defined there?
22:02:44JdGordon|I didnt see that
22:02:50JdGordon|gevaerts: yes
22:02:56kugeljust keep the event (as function call) and the splashes won't be a real problem
22:03:17LloreanJdGordon|: Didn't see what?
22:03:25JdGordon|the wrong colours
22:03:34LloreanJdGordon|: You replied to it as "fix later" and nothing else
22:03:47 Nick YPSY is now known as Ypsy (
22:04:00krabadorhi, i'm using rockbox 3.4 in a sansa fuze v1. No hard problems, battery life good
22:04:13JdGordon|well, I didnt see it and didnt spend much time looking, but yes its fix later if its actually there
22:04:32kugelkrabador: I doubt that
22:04:32gevaertsJdGordon|: that sounds like release notes material :)
22:04:34LloreanJdGordon|: Can it make them unusable though?
22:04:50krabadorbut i've a "pop" when i power on, and when i power down rockbox
22:04:55LloreanJdGordon|: I don't really like the idea of plugins being black on black with some themes or anything, unpredictably
22:05:29krabadorand sometimes video screens fails, for 2-e seconds
22:05:32JdGordon|its menu colours, so that shouldnt affect anything
22:06:06JdGordon|kugel: remond me.. how did we decide the function would work?
22:06:06LloreanJdGordon|: If you can't read the menu it makes it a little hard to start the game...
22:06:10*pamaury hates Windows because when it doesn't work, it's impossible to debug !
22:06:23krabadorkugel, i run the battery benchmark, and i got 12 hours with flac playback
22:06:30Lloreankrabador: Rockbox 3.4 wasn't released for the Fuze.
22:06:34 Join captain_kewl [0] (i=2669ecc2@gateway/web/freenode/x-hzuvryhvovrvloix)
22:07:06krabadori only want report my problems
22:07:41gevaertskrabador: we know, but an exact revision number helps with that, and 3.4 isn't one for the fuze :)
22:07:53krabadori know the branvh of developing
22:08:45AlexPkrabador: I'm not sure what that means :)
22:09:10kugelJdGordon|: well, you said it's ok to keep that function
22:09:32krabadorAlexP, i only want to say that i know the unstability of fuze's developing
22:10:06krabadorgevaerts, i installed the one available the 3.4 release day
22:10:22kugeland it does the clear+update combo with much less (zero) flickering
22:11:01AlexPkrabador: There were probably a few that day
22:11:10JdGordon|crap.. I'll porbbaly just put the event back and a temp workaround
22:11:39AlexPkrabador: 3.4 wasn't released for the fuze - if you go into the system menu then Rockbox Info it will tell you the revision
22:11:39gevaertskrabador: there have been lots of changes since then
22:11:58AlexPkrabador: But as it is under heavy development you should always test with the newest version first
22:12:02*JdGordon| refuses to admit it.. but what amiconn_ was saying last night about knowing what needs to be updated does make sense :p
22:12:16AlexPkrabador: We don't want bug reports on things that have already been fixed :)
22:13:12AlexPyes, very old
22:13:18krabadorthis is my version
22:13:29krabadormust i install the current one?
22:13:37AlexPIf you want to report problems, yes
22:13:45AlexPYou need to try it on the current version
22:13:59AlexPWe don't want to waste time investigating things that may have been fixed
22:14:39krabadorAlexP, i only have fear to have problems like less battery life, and other type of crash...
22:14:56AlexPIt should get better, not worse
22:14:59JdGordon|the other option is clearing the 4 rectangles around the UI viewport instead of doing a fullscreen clear when the theme is toggled back on
22:15:02gevaertskrabador: you can always make a backup of your .rockbox directory
22:15:13JdGordon|that should eliminate the flicker and deadspace and be better
22:15:35 Quit fml ("CGI:IRC")
22:15:44gevaertsJdGordon|: a dirty workaround that might be harder to implement than the real thing could be to find all dead space and define viewports to fill it
22:16:22JdGordon|harder version of what I just said? :)
22:16:41gevaertsyes, but more general :)
22:16:57JdGordon|when the theme is toggled back on, clear the area outside the ui viewport, then force the sbs to redraw, then next main loop the ui area will update
22:17:06JdGordon|there shuold be no flicker at all
22:17:10JdGordon|... hopefully..
22:17:52JdGordon|the other thing... do we want to force the backdrop image in with the theme? so plugin menus look exactly like core menus?
22:18:08krabadorgevaerts, to upgrade now, i can delete my old one, and put the newest, or i must follow the steps i followed when i installalled my old one?
22:18:42gevaertskrabador: no need to delete. The new version will just overwrite old files where needed
22:18:52gevaertskrabador: the bootloader should still be OK
22:22:16krabadorgevaerts, you mean by the rockboxutility'
22:22:49gevaertskrabador: either that, or download and unzip by hand
22:22:49kugelso what now?
22:23:45 Join archivator [0] (n=archivat@
22:26:53 Join petur [50] (n=petur@rockbox/developer/petur)
22:35:05krabadori installed the newest, and nothing happens for audio pops at start and stop playback
22:36:20krabadorbut i'm happy having a new rockbox version, after a look in the news from my old version
22:43:38 Join FOAD_ [0] (
22:44:12 Quit pamaury (Read error: 104 (Connection reset by peer))
22:45:20 Join pamaury [0] (
22:48:06 Quit krabador ("Sto andando via")
22:51:20 Join dfkt_ [0] (n=dfkt@unaffiliated/dfkt)
22:51:44 Join stooo [0] (
22:51:48 Part stooo
22:52:03 Join stooo [0] (
22:55:09 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
22:55:19 Quit FOAD (Read error: 110 (Connection timed out))
22:55:19 Nick FOAD_ is now known as FOAD (
22:56:47 Quit dfkt (Read error: 60 (Operation timed out))
22:58:09 Join LambdaCalculus37 [0] (n=LambdaCa@rockbox/staff/LambdaCalculus37)
22:59:19 Quit polobricolo (Read error: 110 (Connection timed out))
23:00:20 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
23:00:35*FlynDice thinks he has AMS Sansa15 MHz SD write CRC failure issue solved, looking for testers ;-)
23:00:57FlynDicetopik: ^^
23:03:49saratogaFlynDice: any chance you'll look at running DRAM at 2x PCLK next ;)
23:05:34FlynDicesaratoga: I've tried several times with no luck....
23:06:01saratogaFlynDice: on which target?
23:06:12FlynDicesaratoga: e280v2
23:07:08saratogathe DRAM chip on that is rated for 130MHz IIRC, but maybe its the SOC
23:07:25saratogadid you ever try 31 pclk and 62 DRAM clock?
23:08:52FlynDiceI tried lot's of things and I couldn't get any config with PCLK != DRAM to do anything but whitescreen on me
23:10:52saratogahmm thats too bad, the high pclk on ams is probably responsible for most of the difference between us and the OF
23:10:52saratogain battery life
23:11:06FlynDiceI'm pretty sure now that we need to change MCLK for the internal disk to IDECLK and I'm on to 4 bit bus for SD, then I'll look some more at the DRAM speed thing
23:12:00saratogasounds great!
23:12:03FlynDicebut I'm learning all this as I go and I haven't gotten to the chapter on DRAM yet.... ;-)
23:12:17saratogain the as3525 datasheet?
23:12:51 Quit pamaury ("exit(*(int *)0 / 0);")
23:12:51FlynDiceno,.... just in general ;)
23:13:06 Join stripwax [0] (
23:14:12saratogaah ok was wondering if there was a dram refernernece around here i missed :)
23:14:24 Join mrtok1 [0] (
23:16:12mrtok1Q: some profiling gurus here - ipod nano 1g ?
23:17:36mrtok1IŽd like to profile some functions in my code but all values in profile.out equals to 0000 - which means nothing to me and the perl script
23:20:23 Quit dfkt_ ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
23:28:47 Join GeekShado_ [0] (
23:30:58 Quit Strife89 ("Going home.")
23:35:18 Join Soap_ [0] (n=David_Ha@rockbox/staff/soap)
23:35:26 Part stooo
23:37:13stripwaxmrtok1 - mm, did you compile your codecs with profiling options enabled? that's an additional step, in addition to making a build with profiling enabled and wrapping in profile_thread/profstop
23:40:30saratogastripwax: did you ever get time to profile AAC-HE in rockbox?
23:42:18saratogawell if you're ever bored
23:42:34saratogawill probably take about 30 minutes just to decode the standard test file :)
23:42:42 Quit GeekShadow (Read error: 113 (No route to host))
23:43:08 Join Sajber^ [0] (
23:44:29 Quit Sajber^ (Read error: 104 (Connection reset by peer))
23:45:06 Join Sajber^ [0] (
23:46:49stripwaxmrtok1 - you need to ensure -D$(PROFILE_OPTS) is in your cflags somewhere e.g. put it in CODECFLAGS in codecs.make (iirc)
23:47:37stripwaxsaratoga - which codec is that - demac? for some reason I have profiling already set up in that; so presumably that's the last thing I tried profiling :)
23:47:53saratogastripwax: no AAC-HE
23:47:54stripwaxah. no. that's monkey
23:47:55saratogaits libfaad
23:49:33stripwaxhrm, doesn't libfaad compile with CODECFLAGS defined? libfaad.make seems a bit .. sparse. (compared to e.g. libmad.make)
23:50:24mrtok1atripwax: of course yes - i also get a profile.out
23:51:42stripwaxmrtok1 - if you don't compile your codecs with $(PROFILE_OPTS) defined in that codecs makefile [or similar], then gcc won't add any instrumentation to the function calls and you'll get zeros in your profile.out
23:51:59mrtok1stripwax: the makefile claims these profile settings and I would not get a profile.out if I hadnt enabled profiling ??
23:52:54mrtok1its not a codec its kind of eq - which is called in dsp_process
23:53:24stripwaxmrtok1 - you will get a profile.out if you created a build with Profiling mode enabled when you used ../tools/configure. But that is not enough to generate a profile.out with useful information. You need to put code, in place, in the rockbox source to make sure it catches calls to your functions in order to profile. for performance reasons, Profile builds still only catch profiles for the functions you tell it to profile.
23:53:28mrtok1I put around these function call the profile_thread() / profstop calls
23:53:55stripwaxwhich thread
23:54:16mrtok1I guess the audio thread - as i said in dsp_process
23:54:34stripwaxNo, I mean where did you put the profile_thread / profstop calls?
23:54:44stripwaxIn dsp_process itself?
23:55:04mrtok1whithin this function - do i have to put it somewhere else audio_init ?
23:56:25 Quit archivator ("Leaving")
23:56:30stripwaxFor example: If you wanted to profile a codec, you'd put profile_thread when the codec loads, and profstop when the codec exits (i.e. at start/end of codec_main). If you put profile_thread/profstop in a function that gets called many many times, I don't think you're going to catch reliable information. You really want to put profile_start and the start of some batch and profile_end at the end - but (necessarily) in the sam
23:56:30stripwaxe thread that you're trying to profile
23:56:44 Join archivator [0] (n=archivat@
23:56:55 Quit jgarvey ("Leaving")
23:57:34stripwaxI'm not sure which thread dsp_process runs from - possibly the codec thread - if so you'd need to build your codecs with -DPROFILE_OPTS and put the profile_thread/profstop in the codec_main function
23:58:20stripwax(I'm guessing that dsp_process gets called by pcmbuf_insert - if that's not the case at all, profile_thread/profstop needs to be elsewhere)

Previous day | Next day