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

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

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

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

#rockbox log for 2008-11-19

00:00:00darrenbownIt's closed. The fix appeared to be 'I downloaded it again and it worked.' Looking for the ref now.
00:00:29Lloreandarrenbown: If it's closed, then maybe you should describe your problem.
00:00:52LloreanBelieve it or not, just because it's not working for you doesn't mean it's not working for everyone. We need to figure out what's unique in your situation.
00:01:43BigBambiI haven't heard of a specific rbutil/Mac problem, but it may just be not many people use that combination
00:02:05BigBambiI imagine enough people do that if there were a general problem we would have heard though :)
00:02:18LambdaCalculus37darrenbown: What's the FS number?
00:02:20darrenbownok, I ran rbutilqt on mac OS X Leopard and it crashed with the following error:
00:02:20darrenbownDyld Error Message:
00:03:45 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
00:04:49darrenbown Library not loaded: /usr/local/Trolltech/Qt-4.4.0/lib/QtGui.framework/Versions/4/QtGui
00:04:53 Quit bertrik ("Leaving")
00:05:04darrenbown Referenced from: /Applications/
00:05:39 Quit ender` (" Trying to establish voice contact ... please yell into keyboard.")
00:05:41darrenbown Reason: image not found
00:05:44darrenbownexecutable was 1.0.7
00:06:45LambdaCalculus37Have you tried downloading it again? Your first download may have been corrupted.
00:06:59darrenbownTried three times
00:07:21funmandoes the rbutilqt download ship with dynamic libqt ?
00:07:37gevaertsDid you install that Qt in /usr/local yourself? I doubt if that's present on a typical installation
00:07:41darrenbownI even tried the exe on wine. It ran but (predictably) didn't work.
00:07:44funman/usr/local looks like the path of a non standard installation
00:07:45darrenbownAh. Dependencies.
00:09:42amiconnIsn't rbutil statically linked?
00:10:07LloreanIt should be.
00:10:11n1sit should be
00:10:14LloreanMaybe the latest MacOS build isn't?
00:10:16*n1s slow
00:11:13 Quit kugel (Nick collision from services.)
00:11:16 Join kugel_ [0] (
00:11:23 Nick kugel_ is now known as kugel (
00:12:07funmandomonoky: the OF sets the peripheral as the flow controller for SD slot and embedded SD
00:12:13darrenbownOk, I'm off to try to solve the actual problem I was looking at.
00:12:35funmandarrenbown: this is definitely a bug!
00:13:30LambdaCalculus37RButilqt 1.0.7 is working fine here (PowerPC G4 running OS X 10.4.11).
00:14:38darrenbownHmm. Well, I work as a software tester but have limited experience of either Rockbox or Flyspray. Should I report this and if so how?
00:15:58funmanand the DMA controller is the flow controller for i2sout
00:16:23funman(which makes a lot of sense, since the i2sout can't tell if data transfer is finished)
00:16:27darrenbownI was actually hoping that rbutil would prevent the pain-in-the-arse folder merge I have to perform every time I update the RB daily build
00:16:34Lloreandarrenbown: Just post to flyspray a task about Rbutil with a description/title similar to "RButil 1.0.7 for MacOSX does not appear to be statically linked"
00:16:48*Llorean wonders why darrenbown uses dailies instead of current builds.
00:17:21darrenbownbecause I'm frantically bleeding-edge ;)
00:17:27Llorean"current build" is newer.
00:17:34gevaertsYou aren't :)
00:17:42LloreanDailies are only updated daily, while the current build is literally "current" since it's made every time a source change is committed.
00:18:14gevaertsWhat folder merge issue do you have?
00:18:40funmani have the correct settings for DMA, tomorrow I will try to really use them !
00:19:51Lloreangevaerts: OS-X's file browser overwrites folders in such a way that it's equivalent to rming the folder, then copying the new one in its place. Themes, settings, etc will not be preserved.
00:20:01darrenbownOh, just the usual recently-converted windows-to-mac thing about folder replace not merge
00:20:04darrenbownI know I can use cp but it's a pain
00:20:15darrenbownYeah, exactly that
00:20:34gevaertsIs there a command-line unzip? Maybe that works better?
00:20:47Lloreandarrenbown: Are there perhaps better archive managers that would let you extract directly to the device, and do it non-destructively?
00:21:11LambdaCalculus37gevaerts: There is a version of 7-Zip for OS X for the CLI.
00:21:27saratogathe unix unzip command should work well enough i think
00:21:34LambdaCalculus37And it's there as well.
00:21:35darrenbownAnd also about scrobbler upload not working because of hidden . folders, but that's a whole other issue.
00:24:02darrenbownOk, will submit the bug tomorrow. Looking forward to my first day off for three weeks. Thanks for your help. Later.
00:26:20 Quit TheSphinX^ ("XChat@Linux")
00:28:55 Quit funman ("leaving")
00:30:24 Quit herrwaldo ("Konversation terminated!")
00:31:54 Quit DerDome ("Leaving.")
00:31:56 Quit n1s ()
00:35:55linuxstbOS X comes with "unzip" as standard.
00:35:58 Nick JdGordon|zzz is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
00:41:11 Quit captainkewl (" ajax IRC Client")
00:46:06 Quit Zagor ("Client exiting")
00:47:28 Quit MethoS (Remote closed the connection)
00:47:48 Join MethoS [0] (
00:48:29 Quit Genre9mp3 ()
00:51:10 Quit culture_ (Read error: 110 (Connection timed out))
00:52:39 Quit LambdaCalculus37 ("Ka-chunka")
00:53:47 Quit darrenbown (Read error: 110 (Connection timed out))
00:53:58 Quit mofux (Remote closed the connection)
00:58:33 Quit m0f0x ()
01:10:07 Join Thundercloud [0] (
01:10:19*amiconn has a suspicion regarding the usb resets on PP
01:11:44saratogado share
01:12:22amiconnIt could be ram timing. Afaik usb uses dma, and our ram timing might be too fast on PP5022
01:13:09saratogais it known how to adjust DRAM timings?
01:13:32amiconnPP5022 sees much more usb resets than PP5020 - and PP5020 seems to have a slower ram access according to my ape tests
01:14:11amiconnIt's not known, but it *might* be that we're doing it properly on PP5020 (only)
01:16:04*amiconn should try to measure RAM access speed on the various PP targets
01:16:27Unhelpfuli don't supposed arm has a dword*dword->hidword instruction?
01:16:28 Quit Nico_P (Remote closed the connection)
01:17:28 Nick super_ is now known as super (
01:17:34amiconnDepeds on what 'dword' means. For arm this would mean 64 bit, and then there's no such insn
01:18:12Unhelpfuli'm thinking too x86, it seems. 32x32->hi64
01:18:41amiconnThat you can do, using smull, and ditching the low32 word
01:18:58amiconn(or umull, depending on the signedness you want)
01:19:06 Quit lasser (Read error: 60 (Operation timed out))
01:19:32*amiconn knows sh1, coldfire, arm, and z80 asm, but next to nothing about x86 asm
01:19:34Unhelpfuli'm not planning on assembly, but i suspect i can get gcc to do the right thing... assembly can come later.
01:20:06 Quit tessarakt ("Client exiting")
01:20:24 Quit XavierGr ()
01:20:30Unhelpfulprobably ((uint64_t)v1*v2)>>32 ?
01:21:14amiconnMight work if you're lucky. Try it and check the disassembly
01:21:35amiconnOtherwise you could use a small macro or inline function using inline assembly
01:22:28Unhelpfulthat could work, too. either way, it makes it much easier to do division-by-a-constant without having to worry *terribly* much about my overhead
01:23:23 Part toffe82
01:23:29amiconnBeware of the roundoff errror of this method.
01:24:32amiconnI was thinking about an optimisation for small divisors - using a lookup table for the reciprocal. If the dividend is sufficiently large, the result won't be correct anymore
01:25:24amiconnIf you need to divide by a power of 2, that's easy to optimise, of course
01:29:07Unhelpfulamiconn: i need to divide by the width of the input image :/
01:29:39amiconnBut surely not once per pixel?
01:29:40 Quit jhMikeS (Nick collision from services.)
01:29:46 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
01:30:31 Join CaptainKewl [0] (
01:30:56Unhelpfulamiconn: yes, actually, because the scale factors that i'm getting for free from bresenham's are in that range. but since it's a constant, i can generate a reciprocal * some scalefactor, once, and use multiplies.
01:31:42amiconnFor sufficiently small multiples this will probably work
01:42:28 Join sarixe [0] (
01:44:59 Join fdinel [0] (
01:45:11 Quit Thundercloud (Remote closed the connection)
01:47:44***Saving seen data "./dancer.seen"
01:48:02 Quit MethoS (Remote closed the connection)
02:01:00 Join LambdaCalculus37 [0] (n=rmenes@
02:01:29 Quit LambdaCalculus37 (Client Quit)
02:01:47 Join LambdaCalculus37 [0] (
02:04:35 Quit moos ("Rockbox rules the DAP world")
02:06:29*LambdaCalculus37 is mildly amused by the fact that APE is faster on Coldfire than it is on PP :)
02:08:13kugelah LambdaCalculus37 is back :)
02:12:29LambdaCalculus37kugel: Yes, and I'm going to try jhMikeS's patch out now. :)
02:12:57*JdGordon cant tihnk of quiet a few patches he would like to see commited before pointless fading
02:15:04*jhMikeS is still too busy finishing halloween leftovers
02:15:28JdGordonthats not code for gigabeat s power mamangement is it? :D
02:15:56LambdaCalculus37JdGordon: Ummm... it's actually for pointless fading. :P
02:15:57*jhMikeS often speaks in code so perhaps :)
02:15:59*LambdaCalculus37 hides
02:16:40jhMikeSthat's like taunting sasquatch :)
02:17:10Lloreankugel: Does the fading patch have the same kinds of options that fading has on other targets?
02:17:26 Nick fxb is now known as fxb__ (
02:18:33jhMikeScan you really have much in the way of options when the step count is so limited? on a related note, beast HW fading is always a 1 sec ramp and nothing can be done with that.
02:19:32kugelLlorean: no, it'll get options to a) turn it off b) only ad fading in, c) only at fading off, d) both
02:20:40Lloreankugel: Well then why the interest in getting it committed if it's not done?
02:21:05pixelmathe other seem to have - fade in duration and fade out duration settings (if I understand correctly)
02:21:12 Quit jhulst (Read error: 104 (Connection reset by peer))
02:21:23kugelLlorean: because I didn't plan to have a setting
02:21:23Lloreanpixelma: Yes, that's what the older targets have.
02:21:36Lloreankugel: Why did you plan to do it differently than it's done in the completed targets?
02:22:13pixelmait's also a different way of fading
02:22:30LloreanEven if it's different, just "it should be enabled always" is silly.
02:22:49LloreanIf we were okay with it always being on, we wouldn't have the option to turn it off on older targets.
02:23:26pixelmasure, but his way is probably not so flexible for a duration setting
02:23:36LloreanMy concern is on/off, not so much duration options.
02:23:50kugelpixelma: it's not, true
02:23:53LloreanI'm just curious why he didn't look at how it was done previously.
02:24:18*Llorean thinks this way leads to the path of "Rockbox works differently on every player"
02:25:23jhMikeSLlorean: so that means I can't use the hardware on the beast because I can't set duration?
02:25:35pixelmaI know and I'd be against it if I can't turn it off...
02:26:03jhMikeSHW fading can be turned off or used only for one of in or out
02:26:21LloreanjhMikeS: No, it means you can't drop options that you *can* do (the ability to disable/enable fading entirely, for example, if it allows you to)
02:26:50pixelmajhMikeS: I meant kugel's patch
02:27:01kugeljhMikeS: you could be use the options I'll create, not sure how easy that'll be
02:27:07LloreanjhMikeS: With kugel's proposed patch, fading would permanently be enabled even though it's possible by hardware to turn it off.
02:27:18jhMikeSkugel: if it's what you said, then it really is no problem.
02:27:36LloreanHe's updated it, but my concern is the philosophy of "not trying to do it like it's done elsewhere, until people complain"
02:27:46LloreanWell, updated the proposal, not the patch I think
02:28:14jhMikeSLlorean: "<kugel> Llorean: no, it'll get options to a) turn it off b) only ad fading in, c) only at fading off, d) both"
02:28:30LloreanjhMikeS: Yes, it'll *get* those.
02:28:48LloreanHe hadn't really talked about them before, and seemed interested in trying to get the patch committed without them previously.
02:29:23kugelLlorean: I didn't do that on purpose. To be honest I didn't you can configure it in pwm fading (I failed at researching properly...ok). But that's mainly because I didn't look even once into apps/ code while doing the work
02:29:24jhMikeSHmmmm...I could whip those into the mix and commit the S HW fading and then those very ones can be used.
02:29:29LloreanBut as I said, I'm more curious why he didn't check to see if there was already a set of options associated with backlight fading.
02:29:58*jhMikeS looks at his H120 which should allow every possible thing
02:31:01JdGordonjhMikeS: can you tihnk of any nice way to send the main thread a SYS_ message every HZ/2 or so toher than creating a tick task to send the event?
02:31:20Lloreankugel: Ah, okay. As a future note: If you're adding something other players have, you should probably check how they handle presenting it to the users. ;)
02:31:27jhMikeSYeah, this would have to be different, Off/In/Out/Both but not ramping times
02:31:55jhMikeSJdGordon: for what now?
02:32:00*Llorean thinks as long as it has the minimum "Off/In/Out/Both" it's okay, though when the hardware allows there should be duration options.
02:33:22*jhMikeS asks JdGordon to explain what he's aiming at in that it might make a difference
02:33:26JdGordonjhMikeS: just a twisted idea I have
02:34:00jhMikeSThere is a way to have a callback called every HZ/2
02:34:06JdGordonit would be used to send redraw messages to parts of the screen which arnt actually active
02:34:35jhMikeSA something done constantly or for a short time?
02:34:40kugeljhMikeS: yes, it's what I'll offer as settings.
02:34:59pixelmawouldn't a seperate fade in and fade out enable/disable would do the same as the other four options and be closer to the other targets? It's one level deeper though...
02:35:07kugelgiving a fade time doesn't make much sense to me, since it's already dependent on the backlight brightness
02:35:36jhMikeSpixelma: it would be closer and can probably reuse existing strings except just have on/off
02:36:30Lloreanpixelma, jhMikeS: Yes, I'd prefer it like that as well, honestly.
02:36:43LloreanI think it's a little less confusing too, since it's a little more explicit.
02:36:59kugelbut this way I only have 1 var and can use a simple bitmask. Splitting it into seperate screens for just a total of 4 possible options doesn't make much sense to me
02:38:21jhMikeSkugel: for HW it makes no difference to me, bitmask or 2 vars :\
02:38:46LloreanI think keeping .cfg files as compatible as possible cross target would be nice.
02:39:04JdGordonand if they arnt then dont use the same cfg name
02:39:17LloreanBut in this case, we can use the same options pretty readily.
02:39:52jhMikeSthis is similar to samplerates, it doesn't always translate from one to another
02:40:05 Quit JdGordon (Remote closed the connection)
02:40:17*kugel goes to bed as he sees he's gotta to the settings stuff all again
02:40:54 Quit kugel ("ChatZilla 0.9.84 [Firefox 3.0.4/2008111318]")
02:41:43 Join JdGordon [0] (
02:42:19jhMikeSPerhaps in the S case 1000ms should be the "on" setting for the in/out since that is what it is...if you really want it to translate.
02:43:04LloreanjhMikeS: So the options are "Off" and "1000ms"
02:43:15LloreanThat's nice and explicit, so users know what to expect when they choose it, too.
02:44:17jhMikeSLlorean: Sure. But for thread fading it varies with the brightness level so what then?
02:44:56LloreanFor those we just have "On" I guess, since we can't tell users what it actually will be?
02:45:29LloreanUnknown/unusable values for settings get the default value anyway, right? (Which, for fading on targets with only off/on choices will probably be 'on')
02:45:31jhMikeSLlorean: that's the case without getting a bit crazy trying to calculate it
02:45:48*Llorean thinks we don't need to get crazy.
02:45:57jhMikeSAnything not off=on?
02:46:25jhMikeSYeah, seems the logical course to me too
02:47:03 Quit JdGordon (Read error: 104 (Connection reset by peer))
02:52:51 Quit saratoga ("CGI:IRC (Ping timeout)")
02:54:17 Join JdGordon [0] (
02:58:08 Quit Dark_Apostrophe (Remote closed the connection)
02:58:17Unhelpfulamiconn: gcc does indeed emit umull for that. there seem to be four registers involved? is it actually a 64*64->64?
02:58:30jhMikeSLambdaCalculus37: you made sure the patch actually faded? did it?
02:59:47LambdaCalculus37jhMikeS: Nope, your patch gave me this:
03:00:48jhMikeSLambdaCalculus37: you applied it with kugel's? it should be done by itself.
03:01:20LambdaCalculus37jhMikeS: This was done by itself.
03:03:02*jhMikeS sees no "backlight_brightnes" in backlight-target.h
03:04:50 Quit JdGordon (Read error: 54 (Connection reset by peer))
03:06:25*jhMikeS tossed a new patch up there
03:07:48 Quit CaptainKewl (Remote closed the connection)
03:07:58 Join JdGordon [0] (
03:11:43 Join Strife89 [0] (n=michael@
03:17:17LambdaCalculus37jhMikeS: Trying again now.
03:17:49 Join neddy [0] (n=john@nat/sun/x-260131d8c9d6dd50)
03:17:51jhMikeSIf you get that conflict again, I'd suggest 'svn revert' to make sure
03:20:17 Quit Strife89 ("Bye, guys!")
03:21:02 Quit JdGordon (Read error: 104 (Connection reset by peer))
03:32:32 Part pixelma
03:32:59 Join pixelma2 [0] (n=marianne@rockbox/staff/pixelma)
03:37:26 Join Darksair [0] (n=user@
03:40:05 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
03:43:59 Join miepchen^schlaf_ [0] (
03:45:54 Join unL33T [0] (n=4801d5c5@gateway/web/cgi-irc/
03:47:43unL33TIs it possible to disable USB data somehow and only use the USB cable for charging? I modded my iRiver H120 with a single wire to allow charging through USB but it's a pain if you want to use the player while it's plugged in...
03:47:45***Saving seen data "./dancer.seen"
03:52:08 Quit miepchen^schlaf (Connection timed out)
03:56:09 Quit unL33T ("CGI:IRC")
03:56:17 Join unL33T [0] (n=4801d5c5@gateway/web/cgi-irc/
04:01:54 Join dghohens [0] (n=dghohens@
04:02:33 Quit Bensawsome ("The awsome is gone :(")
04:04:00dghohensHey guys, I'm having a problem installing Rockbox on my 5th gen 30 GB iPod. Using the Rockbox utility, the bootloader downloads as normal, then after the download finishes, I get a message saying "could not open ipod". My desktop is Ubuntu. Can someone here help me?
04:06:30 Join miepchen^schlaf [0] (
04:09:31Lloreandghohens: Are you running it with sudo?
04:16:29 Join moos [0] (
04:18:11dghohensLlorean: I'm using the utility. If I ran the utility from console, would that help?
04:18:46 Part dghohens
04:19:20 Join dghohens [0] (n=dghohens@
04:19:43dghohensSorry, my IM screwed up, I may not have received your last message.
04:21:17LloreanYou can run it however you want, it just may need to be run with specific privileges that your user account doesn't have, and this is often easiest to achieve with sudo
04:21:46 Join blkhawk- [0] (
04:22:18 Quit miepchen^schlaf_ (Read error: 110 (Connection timed out))
04:22:22dghohensI'm administrator on my account, though, so I wouldn't think it'd be a problem. I'll give it a shot.
04:27:39 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
04:28:58dghohensIt doesn't seem to be working... what command should I use?
04:30:22 Join BHSPitMonkey [0] (
04:33:30Llorean'sudo ./rbutilqt' or perhaps 'gksudo ./rbutilqt' wherever the file is located.
04:33:59dghohensahh... I forgot the ./
04:35:56dghohensAwesome, it's working now. Thank you very much!
04:37:45 Quit blkhawk (Read error: 110 (Connection timed out))
04:37:52 Part dghohens
04:38:44 Nick blkhawk- is now known as blkhawk (
04:54:50 Join ajonat [0] (n=ajonat@
04:55:48 Quit tarbo (Read error: 60 (Operation timed out))
04:58:57 Join Horschti [0] (
04:59:41 Quit Horscht (Nick collision from services.)
05:00:23 Quit rvvs89 (Read error: 60 (Operation timed out))
05:03:14 Join nuonguy [0] (
05:21:09 Quit unL33T ("CGI:IRC")
05:24:35 Join Photoguy [0] (n=Adams@
05:27:00 Quit LambdaCalculus37 ("Ka-chunka")
05:47:48***Saving seen data "./dancer.seen"
05:47:58 Quit Horschti ("Snak 5.3.3 Unregistered copy. Evaluation period is over. Program will now quit. Thanks for using Snak.")
05:51:34 Join Shaid [0] (
05:53:10 Nick Seedy is now known as Seed (
05:54:44 Quit Llorean ("Leaving.")
05:55:19 Join Llorean [0] (
05:55:47 Join jhulst [0] (
05:59:37 Quit Shaid ("*toink*")
06:10:44 Nick Darksair is now known as Awaysair (n=user@
06:17:12 Quit martian67 (Read error: 60 (Operation timed out))
06:24:33 Join martian67 [0] (i=lol3izer@about/linux/regular/martian67)
06:50:44 Quit jhulst (Remote closed the connection)
06:53:09 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
07:11:27 Join midkay [0] (n=midkay@rockbox/developer/midkay)
07:13:50 Quit midkay_ (Read error: 60 (Operation timed out))
07:29:19 Quit jhMikeS (Nick collision from services.)
07:29:25 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
07:38:36 Quit krazykit (Remote closed the connection)
07:43:39 Quit reacocard (Read error: 60 (Operation timed out))
07:47:50***Saving seen data "./dancer.seen"
07:51:16 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
07:59:05 Join reacocard [0] (n=reacocar@
07:59:57 Join n1s [0] (n=nils@rockbox/developer/n1s)
08:08:46 Quit perrikwp (" ajax IRC Client")
08:12:24 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
08:14:53J-23hmm, why (in newest SVN build) no buttons are recognized on e200v2?
08:15:08 Quit miepchen^schlaf ()
08:16:32LloreanJ-23: Were they recognized in previous builds?
08:16:47J-23I don't know, but probably no.
08:17:06LloreanSo the most likely answer is "because it's still in a very early state"...
08:21:50 Join stoffel_ [0] (
08:25:37 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
08:25:44 Join Bagderr [241] (n=daniel@rockbox/developer/bagder)
08:26:08 Nick Bagderr is now known as B4gder (n=daniel@rockbox/developer/bagder)
08:28:02 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
08:29:53 Quit skipper (Remote closed the connection)
08:30:23 Quit BigBambi (Read error: 113 (No route to host))
08:37:11pondlifeZagor: Morning
08:37:19pondlifeHow large should make.dep get would you say?
08:37:28Zagorquite large...
08:37:52Zagoralmost 2 MB
08:38:10pondlifeThe v14 unified make has taken about 12 mins so far "Generating dependencies" and make.dep is about 10MB
08:38:15B4gderdepends on the path to your src dir too quite alot, doesn't it?
08:38:43pondlifeIt might be something more fundamental -
08:39:10pondlifeI get lots of those warnings before generating dependencies.
08:39:13ZagorB4gder: very true
08:39:16n1spondlife: that looks like you applied the patch twice
08:39:32n1sor frogot to delete new files when svn reverting
08:39:39pondlifeProbably the latter
08:39:53pondlifeIs there an SVN command to delete new files
08:40:21B4gderremove all files and then svn revert brings back the ones from svn
08:40:21pondlife(or to find them, at least)
08:40:43B4gderif you just keep the .svn dirs that is
08:41:18pondlifeAh, I deleted apps and firmware, then did an svn up..
08:41:23pondlifeHopefully that will do
08:41:57B4gderthat works too, but doing a revert first and then up would use less network ;-)
08:42:51pondlifeHmm, still getting those warnings
08:43:05n1skill tools too
08:46:14 Join ender` [0] (
08:46:24pondlife...and uisimulator, and now no warnings
08:46:40*pondlife wishes patch -R was a bit cleverer
08:47:18 Quit ajonat ()
08:49:59 Join Rob2222 [0] (
08:51:42pondlifezagor: I've put my make.dep up at
08:52:00pondlifeLet me know if you'd like anything more, or any other Cygwin tests
08:52:10pondlifeNo problem
08:52:39Zagorrather good compression. 98%!
08:53:25Zagorpondlife: but you still get "No rule to make target `/home/Steve/rockbox/buildsim/apps/SDL_main.o"?
08:54:01pondlife"make: *** No rule to make target `/home/Steve/rockbox/buildsim/apps/SDL_main.o', needed by `/home/Steve/rockbox/buildsim/rockboxui.exe'. Stop."
08:58:03Zagorit's not in make.dep so it is one of the dynamic dependencies
08:58:07 Quit jhulst (Remote closed the connection)
08:59:35pondlifeZagor: The old make had -Dmain=SDL_main, does the new one?
08:59:52pondlifeI've just destroyed my local copy :/
09:00:10 Join skipper [0] (n=skipper@
09:00:29Zagorthat sounds like it could be the cause
09:01:02Zagorbut that effects gcc, and your error message is printed by make
09:01:22pondlifeI guess make needs to know about this switcheroo too
09:02:33Zagorwhere was that? I can't find it in the svn Makefiles
09:03:48pondlifeI get it in my sim Makefile, just searching
09:04:09Zagorin GCCOPTS?
09:04:21Zagorwow. I don't.
09:04:29pondlifeexport GCCOPTS=-W -Wall -g -fno-builtin -I/usr/local/include/SDL -Dmain=SDL_main -I/usr/include/mingw -DWIN32 -Uunix -mno-cygwin -I$(SIMDIR)
09:04:55Zagorsdl-config seems to output different things for us
09:05:16Zagorwhich version is yours? (sdl-config −−version)
09:05:31Zagor1.2.12 here
09:06:36pondlifeI can look for an update, but seeing as I can build SVN ok....
09:07:00pondlifeBest avoid moving the target too
09:07:21Zagoris anyone else (on linux) getting a -Dmain=SDL_main in their sim Makefile?
09:08:01 Quit Rob2223 (Read error: 110 (Connection timed out))
09:08:50B4gderis that from sdl-config −−cflags?
09:09:16pondlifeIndeed it is: -I/usr/local/include/SDL -Dmain=SDL_main -I/usr/include/mingw -DWIN32 -Uunix -mno-cygwin
09:11:02B4gder1.2.13 on linux here, no such flags
09:12:06Zagorpondlife: I saw that too, but that is a compiler error. in our case make is confused into thinking we want to build a SDL_main.o file
09:14:11Zagorah ah ah!
09:14:31Zagorgot it
09:14:43Zagorwe preprocess SOURCES with GCCOPTS
09:14:54Zagorso apps/main.c becomes apps/SDL_main.c
09:15:09pondlifeSo that got lost in the new version?
09:15:30Zagorno, the old didn't use GCCOPTS. the new does.
09:15:48 Quit n1s ()
09:16:39Zagorchange in tools/functions.make so preprocess and preprocess2file uses $(INCLUDES) $(DEFINES) instead of $(CFLAGS)
09:21:39pondlifeBuild in progress...
09:23:20pondlifeLots of errors, need to make clean, I suspect
09:24:20Zagorhmm. you shouldn't have to. but try it.
09:32:01pondlifeNope, still getting link errors -
09:32:17pondlifeMaybe something silly I did
09:32:38pondlifeI need to go to work now, back in a bit though
09:37:10 Join n1s [0] (
09:37:16Zagorwell that is much later, so I count it as progress :-)
09:38:27 Join Thundercloud [0] (
09:47:54***Saving seen data "./dancer.seen"
09:51:53 Quit n1s ()
09:54:57 Join lasser [0] (
10:07:59 Join n1s [0] (n=nils@rockbox/developer/n1s)
10:10:04 Join Nibbler [0] (
10:13:08Zagorof course it's not as simple as not using GCCOPTS...
10:14:48 Quit BHSPitMonkey (Remote closed the connection)
10:16:32 Quit super ("leaving")
10:18:29 Quit stoffel_ ("leaving")
10:27:27 Quit jeffdameth (Read error: 113 (No route to host))
10:27:46 Quit Thundercloud (Remote closed the connection)
10:28:30 Join jeffdameth [0] (
10:43:14 Quit n17ikh|Lappy ()
10:43:34 Quit moos ("Few days of vacations in Andorra :) c you soon")
10:46:40 Join _lifeless [0] (n=lifeless@
10:48:44 Quit faemir (Remote closed the connection)
10:49:33 Quit nuonguy ("This computer has gone to sleep")
10:50:56 Join stoffel_ [0] (
10:53:33 Quit pixelma2 ("-")
10:53:45 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
10:56:02 Quit Zarggg (Excess Flood)
10:56:41 Join Zarggg [0] (
10:58:38 Quit linuxstb ("Leaving")
10:59:18PhotoguyAnyone have some advice on a replacement for my E-200?
10:59:34PhotoguyI want something new, maybe not onther Sansa.
11:00:19LloreanPhotoguy: The list of supported targets is on the front page, and there's more information about them on the BuyersGuide wiki page
11:00:46PhotoguyOk, just wondering if anyone had some thoughts..
11:01:11n1sPhotoguy: if you want useful advice you need to tell us what you want the player to do
11:01:32PhotoguyI think I might go with the Cowon D2..
11:01:40PhotoguySince there is a port in development for it.
11:02:34PhotoguyDo you think the touch screen will be full compatible?
11:02:46PhotoguyProbably eventuall..
11:03:02LloreanThe D2 has some unique hurdles to overcome
11:03:12LloreanRight now the touchscreen already works pretty well
11:03:25PhotoguyWhat are the other issues?
11:03:26LloreanBut due to the way the internal memory is handled, it may be a considerable amount of time until we can write to it.
11:03:36LloreanMeaning you may need an SD card just to be able to use Rockbox regularly
11:04:19PhotoguyWhich do you think is further developed? The D2 or the Sansa V2?
11:04:21pixelmaLlorean: where did you see volume up/down buttons in the Clip sim (can't find a reference in the keymap either)? Also - if it was as easy to simply replace scroll forward/backward with presses of volume up/down, having a nice c200 keymap wouldn't be so hard
11:04:57Lloreanpixelma: They're on the right side, similar to where they are on a Gigabeat F.
11:05:18LloreanEven if they weren't suitable for up/down, the keymap could then just mirror the c200's instead.
11:06:06LloreanI didn't see them in the clip sim. I don't know what keys are currently mapped to them.
11:08:34pixelmafound a reference to them now, seems like they are not used in the keymap file at all...
11:09:03LloreanI think the clip should either duplicate the e200 or c200's keymap.
11:09:30LloreanI *think* as they physically face up and down the volume keys would do well enough to replace the scroll wheel, but I haven't handled one yet, and if they don't, the c200 keymap can do a 1:1 mapping too.
11:09:34pixelmathe c200 keymap could be a better base, I think, although this one still needs work itself :\
11:10:35*Llorean has no objections to that.
11:10:38pixelmayeah right, one can only tell after trying out with the target
11:11:00LloreanI just don't think it needs a completely unique keymap when there's two players that are very similar and have the same number of buttons and close layouts.
11:11:04LloreanIncluding labeling.
11:11:24*Llorean has ordered a low-capacity Clip, since he thinks it'll be useful anyway.
11:12:50pixelmaI was especially curious about the volume up/down buttons because I didn't see them and because I couldn't imagine that the thing that actually looked like a wheel wasn't one...
11:13:22LloreanYeah, I was a little surprised by that too.
11:13:49Zagorpondlife: when you get back, run "make SILENT=" and paste the full link command somewhere
11:13:55Zagoranyone else on cygwin here?
11:14:35*pixelma is considering getting a Clip for the graphic works ;)
11:14:57ZagorI would be intersted in seeing the output of sdlconfig −−libs on cygwin
11:15:58pixelmaZagor: me - do you mean with SVN or with your patch?
11:15:59pondlifeZagor: You mean with SVN, or patch 14?
11:16:21Zagorwith my patch, that gives you link errors
11:17:00pondlifeZagor: I don't have a copy here (of the modified one) - any chance you could upload it to FS?
11:17:38pondlifeJust to be sure I'm doing the right thing
11:17:39 Nick fxb__ is now known as fxb (
11:17:58Zagorpondlife: use this:
11:18:37Zagorbut you don't want SILENT= on the whole build. let it fail once first, and then rerun with SILENT=
11:19:09Zagorand before you do that, tell me the output of "sdlconfig −−libs"
11:19:11 Join mofux [0] (
11:19:34n1ssaratoga: ping
11:19:42pondlifehmm, sdlconfig isn't found on this PC at all!
11:19:50Zagoryes, sorry
11:20:02LinusN$ sdl-config −−libs
11:20:02LinusN-L/usr/local/lib -lmingw32 -lSDLmain -lSDL -mwindows -mno-cygwin
11:20:06pondlife-L/usr/local/lib -lmingw32 -lSDLmain -lSDL -mwindows -mno-cygwin
11:20:08n1sor anyone else who have played with profiling plugins
11:20:11LinusNi won!!!
11:20:21*B4gder hands LinusN the award
11:20:34pondlifeCygwin can't be involved in any speed contests ;)
11:21:28Zagorlooks ok. but I need to see the link line.
11:21:56pondlifeInitial build running
11:22:03pondlifeTakes about 5 mins, I think
11:22:41ZagorLinusN: if you have a few minutes, could you run a speed comparison btwn "make" and "make -r" on cygwin?
11:22:55LinusNwith your patch?
11:23:15LinusN"few minutes" - ha, i can see you are not using cygwin yourself :-)
11:23:51pondlifehmm, I'm back to the SDL_main.o make error :/
11:24:05Zagorpondlife: huh!?
11:24:32Zagorah, hang on
11:24:45LinusNZagor: i need to adjust the patch a little, right?
11:25:38ZagorLinusN: yes, wait a minute. my main=SDLmain fix wasn't correct
11:26:11 Join petur [50] (n=petur@rockbox/developer/petur)
11:27:18Zagorright, in line 11 of tools/root.make, add an underscore btwn SDL and main.
11:28:05Zagorline 10 even
11:28:45LinusNZagor: which target do you want me to build?
11:29:09ZagorLinusN: it doesn't really matter. pick any binary. e200 is a nice one :-)
11:29:24LinusNx5 it is
11:29:31pondlifeI'll stick with H300 sim
11:29:36Zagorpondlife: yup
11:30:03LinusNwoo, lots of override warnings
11:30:16LinusNmaybe i didn't apply it correctly
11:30:24ZagorLinusN: huh? sounds like you have a double-applied then
11:30:39pondlifeSorry about the wrapping
11:30:49LinusNwhat do i need to remove?
11:31:19n1sLinusN: delete new files added by the patch, svn revert and apply again
11:31:34Zagor...basically everything
11:32:01Zagorthe patch changes 160 files
11:32:34LinusNhow fun
11:34:25pixelmaZagor: what I wanted to ask - is "make manual" also affected by your changes?
11:34:52Zagorpixelma: no
11:35:59LinusNZagor: but the new files are only *.make, right?
11:36:11ZagorLinusN: yes
11:36:39LinusNi did a find . -name "*.make" -exec rm {} \; and it seemed to work
11:36:42 Quit stoffel_ ("leaving")
11:37:08pondlifeZagor: Was my link command enlightening?
11:37:16LinusNthe dependency generation takes ages....
11:37:31Zagorpondlife: a little. I'm digging into some issues from that.
11:37:49pondlifeLinusN: I let it run for 10 mins before cancelling - that was also caused by double-patching
11:37:55ZagorLinusN: yes, it builds all dependencies at once. the old one built them per-directory. and missed some..
11:37:59pondlifeLinusN: Does make clean run without warnings?
11:38:29ZagorLinusN: how large is your make.dep file?
11:38:41LinusN1.5MB so far
11:39:52LinusN2.5 now...
11:39:59Zagormy x5 make.dep is 2.1MB with a 32-char build dir path
11:40:57LinusNok, this looks wrong
11:41:12LinusN3.3MB now
11:41:24*pondlife suspects LinusN needs to go back to SVN source
11:41:26LinusNwhoa, it started to build
11:41:33Zagorhow long is your path?
11:41:34*pondlife might be wrong
11:42:02LinusNZagor: 25
11:42:26amiconnHmm, all-in-one dependencies might be bad for build time. If a few files change, it needs to rebuild the whole .dep
11:42:32Zagorso how the heck did you manage to create a 3.3MB dep file...? upload it somewhere
11:43:14Zagoramiconn: the dep file doesn't autorebuild
11:43:22 Join einhirn [0] (
11:43:49ZagorI was choosing between one big and one per source file, and settled with this. It's fairly easy to change if we want to.
11:45:39LinusNZagor: i guess you wanted the timing without the dependency generation, right?
11:45:47ZagorLinusN: yes
11:47:18 Nick fxb is now known as fxb__ (
11:47:48LinusNZagor: frustrating... make clean removes make.dep
11:47:58***Saving seen data "./dancer.seen"
11:48:19Zagorah right. run "make foo" (or whatever) to create dep and nothing else.
11:48:47LinusNhow about "make dep"? ;-)
11:49:12Zagorthat will make it twice :-)
11:50:42LinusNi must say that the dependency generation time is a royal pain
11:51:02Zagorwhat is not a pain in cygwin?
11:51:12LinusNand having "make clean" remove it is even worse
11:51:25ZagorI can move it to veryclean
11:51:31LinusNplease do
11:52:02B4gderI'm not sure that's a good idea
11:52:26B4gderI think dep problems are one of the common things people fix with "make clean && make"
11:53:11Zagordependency generation takes 18 seconds for an x5 build on giant
11:54:20B4gderyes, the total build time is what counts
11:54:30Zagordo we still have no clue why cygwin is so slow?
11:54:41Zagorit's not even an emulator
11:55:07rasherLaunching a process in cygwin is orders of magnitudes slower than on Linux.
11:55:34Zagorrasher: ok
11:55:54Zagorno wonder the dependency generation is slow then...
11:56:02rasherShell scripting obviously suffers greatly from this
11:56:13Zagorbut it would just as slow with one dep file per source file
11:56:25Zagoreven slower, since it would mean more file operations
11:57:24ZagorI experimented with calling gcc depgen with more than one source file per execution, but there is no way to solve the source/builddir handling
11:58:53rasherMoving voicefile generation from shell scripts to a perl script and employing a helper-script cut down build times on cygwin to about 10% iirc.
12:00:15Zagorpondlife: can you do a h300 simbuild in svn and post the link line for that too? I con't see anything wrong with that link line...
12:01:43Zagorthe double makes I can understand, but all the undefined SDL functions are really strange
12:02:28Zagoroh I bet it's order related again...
12:03:36Zagortry moving $(LDOPTS) to the end of the line in uisimulator/uisimulator.make:28 and see if that makes a difference
12:07:34 Join {phoenix} [0] (
12:10:53 Join kugel [0] (n=chatzill@unaffiliated/kugel)
12:14:51 Join ap0 [0] (
12:18:03 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
12:20:44pondlifeZagor: SVN make -
12:21:36Zagorpondlife: thanks. did you try moving LDOPTS as I instructed?
12:21:47pondlifeAh, no - missed that line
12:23:34kugelpondlife: did you test my latest fading patch?
12:24:02pondlifeNo - didn't see if you'd sorted the problem I reported (backlight stays off), but will do later
12:24:18pondlifeZagor has me at the moment ;)
12:27:46kugelpondlife: yes, it should be fixed
12:31:51 Nick Awaysair is now known as Darksair (n=user@
12:35:24pondlifeZagor: When you get back, I moved LDOPTS to the end, and now get "make: *** No rule to make target `os_types.h', needed by `/home/Steve/rockbox/buildsim/apps/codecs/libtremor/oggmalloc.o'. Stop."
12:35:48pondlifePaste of SILENT= output coming up...
12:36:07pondlifeAh no - it doesn't get that far!
12:36:31 Join kachna|lappy [0] (
12:57:53 Quit daurnimator ("Cyas later...")
13:02:34 Quit kugel (Remote closed the connection)
13:04:14Zagorpondlife: ?? and you changed nothing else?
13:04:28pondlifeAs far as I can tell
13:04:37pondlifeProbably best if you pop a new patch up
13:05:23ZagorI think I need to get a windows computer and install cygwin. this kind of remote debugging is rather frustrating.
13:06:04pondlifeI'm happy to try more - but if you supply a patch it's one step closer to local debugging
13:06:17pondlifeI'm currently on SVN and about to have lunch
13:06:40 Join HBK- [0] (
13:16:31 Part pondlife
13:22:16 Quit HBK (Read error: 110 (Connection timed out))
13:24:54 Join kugel [0] (n=chatzill@unaffiliated/kugel)
13:25:59 Quit kugel (Remote closed the connection)
13:30:40 Join LambdaCalculus37 [0] (
13:33:50 Join funman [0] (
13:34:57 Join kugel [0] (n=chatzill@unaffiliated/kugel)
13:37:15 Join jpt9 [0] (
13:37:18jpt9Hey guys. has 1GB refurb Sansa Clips.
13:37:35jpt9Should I get one to help test Rockbox on it for you guys?
13:38:19jpt9they're $15 :-)
13:42:23 Join Bensawsome [0] (n=Bensawso@unaffiliated/bensawsome)
13:44:24kugeljpt9: we have enough clips, because they're so cheap
13:44:44kugeljpt9: so unless you're going to use it I don't think it'll be helpful
13:44:47jpt9But they're so freaking *cute*!
13:45:14LambdaCalculus37Well, we can't tell you what to do with your money. If you want one, go ahead. But as kugel said, we have plenty already.
13:47:22JdGordonZagor: I can give you remote desktop access to a windows box if you want? I dunno how usable it will be with my poor upload speed though
13:47:40funmanI was wondering if I should put the DMA PL081 driver in firmware/drivers or in target/arm/as3525
13:47:59***Saving seen data "./dancer.seen"
13:48:13funmanJdGordon: I think setting up a ssh server would take less bandwidth and be more practical
13:48:14JdGordonhow generic is it? stick it in drivers if its not really target specific
13:48:30ZagorJdGordon: thanks. it's definitely worth trying.
13:48:35funmanit's specific to pl081, which is only present in as3525 afaict
13:48:43JdGordonfunman has a poiont though... is cygwin+ssh enough?
13:49:05JdGordonokey dokey...
13:49:10*JdGordon installs cygwin
13:50:12jpt9So how goes the Clip porting effort, btw?
13:51:18LambdaCalculus37LCD works, buttons need work, storage works (?), no sound.
13:51:26funmanjpt9: the TargetStatus wiki page is up to date, and the forum thread has more precise information
13:52:15jpt9What about plugins? If I get one, even if it doesn't have sound, I could always read Ender's Game on it... :-)
13:52:51LambdaCalculus37I don't think plugins build yet. But I don't have a Clip (yet), so funman will be the man to ask.
13:52:56 Quit LambdaCalculus37 ("Ka-chunka")
13:53:41funmanmost plugins don't build because of missing keymap, if you already know rockbox you can definitely help here
13:54:24funmani have built some plugins with variations of the C200 keymap, because it looks quite similar
13:54:31jpt9Never coded on Rockbox before...
13:56:08funmani meant if you already used it, the keymap definition is quite simple '#define ACTION_XX BUTTON_zz'
13:56:12Zagorkeymaps are easy, you don't need to know the rockbox code much
13:57:49JdGordonZagor: installing the base cygwin system.... 64% done
13:57:58jpt9Of course I've used it!
13:58:51 Join HBK1 [0] (
13:58:51 Quit HBK- (Read error: 104 (Connection reset by peer))
14:08:59 Quit gevaerts (Nick collision from services.)
14:09:12 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
14:16:11 Join LambdaCalculus37 [0] (i=44a04303@gateway/web/ajax/
14:16:58 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
14:17:14saratogan1s: pong
14:17:32 Join robin0800 [0] (
14:23:57 Join tyfoo [0] (
14:24:21n1ssaratoga: did you play around with profiling, lately? i tried to profile the midi plugin on a sansa but it just freezes
14:25:32 Quit Bensawsome (Read error: 54 (Connection reset by peer))
14:26:30n1sany clues?
14:26:50 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
14:28:00saratogan1s: never for plugins, but it works nicely for codecs lately
14:28:21saratogamaybe double check that the plugin api has all the functions defined that the codec api does for profiling?
14:28:25saratogaanyway, got to run for a bit
14:28:45n1sok, i'll look
14:30:22n1syep, same functons exposed to both plugins ans codecs...
14:32:21 Join Strife89 [0] (i=a810ebb0@gateway/web/ajax/
14:32:21 Quit nplus (Read error: 104 (Connection reset by peer))
14:33:47 Join nplus [0] (n=nplus@141.25.Globcom.Net)
14:34:26 Join jpt9_ [0] (
14:37:04 Join Bensawsome [0] (n=Bensawso@unaffiliated/bensawsome)
14:39:14 Quit jpt9 (Read error: 110 (Connection timed out))
14:39:21 Nick JdGordon is now known as JdGordon|zzz (n=jonno@rockbox/developer/JdGordon)
14:41:12 Quit Nibbler (Read error: 113 (No route to host))
14:41:56 Join jpt9 [0] (
14:44:30 Join TheSphinX^ [0] (
14:47:10 Join jpt9__ [0] (
14:47:49 Quit amiconn (Nick collision from services.)
14:47:55 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
14:49:30 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
14:50:03 Quit jpt9 (Read error: 60 (Operation timed out))
14:51:18 Join darrenbown [0] (n=darrenbo@
14:51:26 Quit tyfoo (Read error: 104 (Connection reset by peer))
14:51:47 Part darrenbown
14:52:35kugelLambdaCalculus37: ping
14:52:55LambdaCalculus37kugel: Pong
14:53:07kugelhave you tried jhMikeS mikes patch already?
14:53:24LambdaCalculus37kugel: His first patch didn't compile. His second one does.
14:53:46kugeland does it fade? more or less nicely than my patch?
14:54:27LambdaCalculus37kugel: A little more nicely.
14:54:29 Quit Photoguy ("Leaving")
14:54:45kugelLambdaCalculus37: heh, cool
14:54:50*kugel expected that
14:55:10kugelLambdaCalculus37: looks like we did hours of coding&testing my patch for nothing :)
14:55:43LambdaCalculus37kugel: Not necessarily. We did hours of coding and testing to come down to a conclusion of how best to make fading work right.
14:56:05LambdaCalculus37There's some strange beauty in all of that work. :)
14:56:27Strife89Oh, there's a new fade patch?
14:56:46kugelStrife89: not really for c200, but of course you can test if you want
14:57:02Strife89I might when I get home.
14:57:29Strife89I should probably unpatch the old one though, correct?
14:57:36 Quit jpt9_ (Read error: 113 (No route to host))
14:57:46LambdaCalculus37Strife89: We were also discussing a patch jhMikeS did for the Gigabeast.
14:57:58LambdaCalculus37Strife89: Yep, remove the old patch first.
14:58:48 Quit Strife89 (" I'm off to class.")
14:59:57kugelLambdaCalculus37: I guess the the gigabeat s patch should be committed soon then. It'll need my settings work first though, which I finished just now
15:01:04kugelLambdaCalculus37: changing the delay between fade steps didn't work well
15:01:31kugelso my latest patch on the tracker is definitely not as nicely as possible
15:04:10pondlifeI still don't quite understand why different hardware requires different fading code. I'd think that the controlling/ timing code could be shared, only requiring a "set backlight (or button light) brightness" routine for each target.
15:04:38pondlife(Which we already have for most targets anyway, where backlight brightness can be configured.)
15:06:10 Quit Zagor ("Client exiting")
15:07:40kugelpondlife: some hardware does offer that, so you don't need to loop though the brightness steps
15:08:11pondlifeAh, ok. Just send a command for off/on?
15:08:55kugelso the hardware will fade solely. On targets not featuring that you implement a software controlled fading
15:09:02 Join jpt9_ [0] (
15:09:21 Quit kugel ("ChatZilla 0.9.84 [Firefox 3.0.4/2008111318]")
15:09:45 Quit jpt9_ (Client Quit)
15:09:45 Quit jpt9__ (Read error: 60 (Operation timed out))
15:11:58LambdaCalculus37kugel: I'll try your patch in a little while.
15:12:24pondlifeme too
15:15:37n1spondlife: some targets do fading with pwm too (those usually can't set brightness)
15:17:12n1shmm, if i remove the calls to profile_thread and profstop the midiplugin still freezes, maybe the actual -finstrument-functions is the problem...
15:20:23 Part LinusN
15:20:51Nico_PLambdaCalculus37: where can I find the beast fading patch?
15:21:35LambdaCalculus37Nico_P: FS #6800
15:21:43Nico_Pthanks :)
15:24:18 Join Schmogel [0] (
15:28:36 Join kugel [0] (n=chatzill@unaffiliated/kugel)
15:31:21*kugel is glad he found some tester
15:31:49LambdaCalculus37kugel: I'm going to try your patch out in a bit.
15:34:36Nico_Pkugel: I'm curious to know how good it looks on the H300
15:35:14Nico_Pin the early days, LinusN said the brightness controller was too slow to implement a proper looking fading
15:35:24 Quit ap0 (Read error: 113 (No route to host))
15:37:24kugelNico_P: I*m curious too
15:40:42 Nick fxb__ is now known as fxb (
15:43:09 Join culture_ [0] (
15:45:10 Join esthar [0] (
15:46:06n1sah, i think i got it working, turns out trying to call a function pointer set to NULL is a bad idea
15:47:17n1sbtw is there a reason for these small differences between codecs and plugins? codecs use the ci pointer directly while plugins requires a call to profile_init that sets a local copy of the rb pointer.
15:48:00***Saving seen data "./dancer.seen"
15:51:38saratogan1s: that sounds like a bug
15:52:27n1sif i changed the plugin version to work like the codec one it works at least
16:01:02n1sturns out my feeling was correct, the majority of time is spent in one function, it acually uses more than 100x as much cpu as the second in the list
16:03:06 Join krazykit [0] (
16:03:17B4gderI recommend using valgrind's −−tool=callgrind to get nice profiling data (obviously when running a sim)
16:04:08B4gdertogether with kcachegrind
16:06:49 Join tyfoo [0] (
16:08:32n1sone way that will help a bit is to basically copy the loop a couple of times and check a couple of things before and entering the right copy as that would get rid of quite a bit of conditionals inside it but it feels messy...
16:09:57saratogawhat function
16:10:55n1sthe synth.c:synthVoice()
16:12:09 Quit kugel (Nick collision from services.)
16:12:12 Join kugel_ [0] (n=chatzill@
16:12:19 Nick kugel_ is now known as kugel (n=chatzill@
16:12:45n1sas you can see it doesn't exactly look like a performance function...
16:12:58 Quit denes_ (Read error: 113 (No route to host))
16:13:02 Join denes_ [0] (
16:14:03kugelLambdaCalculus37: did you already test? you can compare nicely by rolo'ing
16:14:23LambdaCalculus37kugel: Not yet.
16:19:31 Join jgarvey [0] (
16:24:48saratogan1s: did you check if that divide by a power of 2 is switched for a shift?
16:26:01n1ssaratoga: no, i don't think it happens too often, but so->decay can be negative
16:26:22n1si mean i don't think the division happens too often
16:26:55 Quit advcomp2019 (Read error: 104 (Connection reset by peer))
16:27:09 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
16:27:35 Quit {phoenix} (Read error: 110 (Connection timed out))
16:28:05pondlifekugel: Your patch fails to compile for H300 here, but I wonder if I have some remnants of Zagor's work on my PC.
16:29:04pondlifeIf you can check an H300 build compiles, and let me know if it does for you.
16:29:37 Join kushal_12_27_200 [0] (n=kushal@
16:29:50 Join tic_toc [0] (n=4670e35d@gateway/web/cgi-irc/
16:32:37tic_tocgreetings. question. i noticed all the ipods show "no" in the USB column on the current status chart. what does this mean exactly? that it won't function as a USB device?
16:33:11B4gdertic_toc: it means you still need to use the OF's USB support as Rockbox has no usb of its own on these devices
16:33:29kugelpondlife: changle line 63 in firmware/target/coldfire/iriver/h300/backlight-h300.c to read "#ifndef USE_BACKLIGHT_THREAD_FADING"
16:33:41B4gdertic_toc: Rockbox's own usb stack is still work in progress
16:34:27tic_tocroger that, and thank you.
16:35:30*kugel thought I've fixed that yesterday
16:35:32 Join toffe82 [0] (
16:35:45pondlifekugel: I picked up the patch from Flyspray
16:36:03pondlifeThe v7 one
16:36:33kugelpondlife: well, I remember I already told you yesterday to change that line. Apparently I didn't change it in my tree
16:36:40 Quit lasser (Remote closed the connection)
16:36:56 Quit tic_toc ("CGI:IRC")
16:42:36 Join captainkwel [0] (i=2669ecc2@gateway/web/ajax/
16:42:55kugelpondlife: does it compile?
16:43:06pondlifeYes, about to copy it over
16:47:36pondlifeYes, works
16:48:09pondlifeLooks ok, but slighter faster fade in and out would be better, I think
16:48:22kugelhow long does it take?
16:48:28 Quit bmbl ("Woah!")
16:48:37pondlifeAbout 0.5s, I guess
16:48:54pondlifeAbout twice as fast would be better, I think
16:49:08pondlifeMy battery is almost dead though, so can't experiment right now
16:50:03 Quit kugel (Nick collision from services.)
16:50:06 Join kugel_ [0] (n=chatzill@
16:50:14 Nick kugel_ is now known as kugel (n=chatzill@
16:50:25pondlifeheh, the USB-charge button doesn't work on H300 if the quick screen appears.
16:51:49 Quit funman ("leaving")
16:52:37kugelpondlife: 0.5s sounds reasonable. I wouldn't make it faster. Also, keep in mind that it depends in your brightness setting
16:53:02pondlifeI was on full brightness
16:53:14kugellevel 15 is that right?
16:53:22kugelit takes about 1s for me at lvl12
16:53:48 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
16:54:03kugelpondlife: so you can make it shorter by reducing the brightness. I guess you don't run at full brightness usually anyway?
16:54:21pondlifeI do normally
16:54:48pondlifeI like the fade anyway
16:54:59pondlifeDoes it need a config setting? Or is that conventional?
16:55:15 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
16:56:04kugelpondlife: it seems to need it. I personally don't see a need, but from the discussion tonight it seems generally wanted
16:56:27kugelNico_P: you heard? fading seems to be nice on the h300
16:57:03pondlifeTwo options (fade in and out) seems overkill.
16:57:14pondlifeMaybe a fade time (0 for none)?
16:57:36kugelpondlife: see todays log
16:57:43pondlifeWill do
16:57:48 Part pondlife
16:57:56kugelthe 2 options are mainly because to keep it consistent with pwm fading targets
16:58:03kugeland to reuse strings from the lang
16:58:26 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
16:59:03kugelpondlife: the 2 options are mainly to keep it consistent with pwm fading targets, and to reuse strings from the lang
16:59:30pondlifeConsistency is good; I'll read the logs.
16:59:54pondlifeMy gut feeling is that no setting is needed - or if so, just one (speed)
17:02:21pondlifeI agree that consistency across targets is the overriding factor
17:02:36pondlifeBut am surprised that there are 2 options.
17:03:03pondlifeI guess somebody must be using "fade out, don't fade in"...
17:03:08kugelwith 2 options we can perfectly reuse strings from lang files
17:03:37kugeland pwm targets can set fade in/out differently too (even though there's more to configure about)
17:03:56pondlifeI was more thinking that we could merge those options on all targets - i.e. backlight fade = on/off. But I didn't realise that we already had 2 options
17:04:16LambdaCalculus37pondlife, kugel: FWIW the iPod video has Backlight Fade In/Out options under its LCD Settings. Is this a PWM target or no? (just curious, really)
17:04:38pondlifeNo idea
17:05:09 Quit kushal_12_27_200 ("This computer has gone to sleep")
17:05:20kugelLambdaCalculus37: yes
17:06:23*kugel goes home
17:06:25 Part kugel
17:10:02 Join funman [0] (
17:17:13 Part pondlife
17:20:32 Quit skipper (Remote closed the connection)
17:20:34 Quit dionoea (Read error: 104 (Connection reset by peer))
17:20:37 Join dionoea [0] (
17:20:41 Quit tyfoo (Client Quit)
17:21:08 Quit aarcane (Read error: 113 (No route to host))
17:21:19 Join hjh [0] (n=55dc2e80@gateway/web/cgi-irc/
17:23:59 Quit hjh (Client Quit)
17:34:10 Quit advcomp2019 (Read error: 104 (Connection reset by peer))
17:34:11 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
17:34:44 Part B4gder
17:38:02 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
17:44:52 Join Llorean1 [0] (
17:45:14 Quit esthar (Read error: 60 (Operation timed out))
17:45:51 Join toffe82_ [0] (
17:46:45 Join esthar [0] (
17:48:04***Saving seen data "./dancer.seen"
17:48:14 Quit bmbl (Client Quit)
17:51:32 Quit toffe82 (Read error: 60 (Operation timed out))
17:51:34 Nick toffe82_ is now known as toffe82 (
17:53:44 Quit Llorean (Read error: 145 (Connection timed out))
17:54:48 Quit Slack__ (Connection timed out)
17:55:29 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
17:56:37funmanhello domonoky, I have found that the Sansa AMS OF uses the SD controller as the flow controller in DMA transfers
17:57:42funmanso if I understand correctly we have to setup the SD controller and the DMA controller for DATA transfer, ask the SD controller to start it, and wait a DMA interrupt which would tell if it has finished or if an error happened
17:58:20domonokyyep.. setup the DMA controller, ask the Sd for data, and wait for the interrupt..
17:58:39funmansince the SD FIFO is quite small, I think we should use burst DMA ?
17:58:56 Join bertrik [0] (
17:59:56funmanI checked the dma driver for creative zvm, it uses kernel.h 'struct wakeup' to notify when a DMA transfer has finished
18:09:31 Join MethoS [0] (
18:11:33 Join kugel [0] (n=chatzill@unaffiliated/kugel)
18:12:49funmanI wonder if we have to set the transfer size, if the SD controller controls the data flow
18:13:53 Quit _lifeless (Remote closed the connection)
18:14:22domonokyhm, i dont we need to set the transfer size if the sd controller controls the flow
18:17:35domonokydatasheet says we have to set transfer size to 0 if the DMAC is not the flow controller
18:18:01domonokyotherwise the DMAC may get confused... so its really 0
18:18:56funmanok .. so nothing happens, the interrupt handler is not called
18:19:36kugelLambdaCalculus37: I updated the patch, it should be fading faster now, although I don't think we have to put further effort into the beast for this patch
18:20:06 Join dany_21a_ [0] (
18:20:49LambdaCalculus37kugel: I'll try it out in a bit.
18:31:21 Part dany_21a_
18:33:03funmandomonoky: i'm not sure if dst/src size and width in channel control register must reflect the FIFO size, or a sector size .. let's check which sizes the OF use
18:33:06 Join {phoenix} [0] (
18:36:00 Join haikubear [0] (n=46488ec8@gateway/web/cgi-irc/
18:37:04haikubearis this the proper place to ask an install-gone-awry question?
18:37:57 Join ap0 [0] (
18:38:22gevaertshaikubear: yes
18:38:47 Join dany_21a_ [0] (
18:38:49haikubeargreat. i'm installing on an ipod video 30G, i run debian on my home system.
18:39:28haikubearthe bootloader complains that it can't find rockbox.ipod, but i can see it in /.rockbox/
18:40:00haikubearthe permissions on /.rockbox/rockbox.ipod are 755
18:40:35n1shaikubear: did you format the partition as anything else than FAT32?
18:40:58gevaertsCan you check the output of dmesg for the sector size? (it should be either 512 or 2048 bytes)
18:41:27haikubearn1s: no
18:41:30haikubeargevaerts: checking
18:41:54haikubearsd 9:0:0:0: [sdc] 14651280 2048-byte hardware sectors (30006 MB)
18:43:40gevaertsCan you pastebin the output of sfdisk -l /dev/sdc somewhere?
18:46:41*gevaerts suspects that the "ghost filesystem" issue is still not solved on the 5.5g
18:46:49gevaertsLet me think for a bit
18:49:29gevaertsWhat's the output of "cat /proc/partitions |grep sdc" ?
18:50:21haikubear 8 32 29302560 sdc
18:50:26haikubear8 33 96264 sdc1
18:50:31haikubear8 34 29206170 sdc2
18:51:43haikubearshould i just repartition the damn thing?
18:51:49gevaertsWon't help
18:52:34gevaertsIt needs a bit of dd
18:53:20 Quit Darksair ("Everything that has a beginning has an end.")
18:53:27gevaertsFirst make a backup of /dev/sdc1 (in case something goes wrong later) : "dd if=/dev/sdc1 of=$HOME/firmware.img bs=64k"
18:53:33 Join faemir [0] (
18:54:32kugeljhMikeS: ping
18:54:33gevaertsThen "dd if=/dev/zero of=/dev/sdc seek=48190 count=10"
18:55:00gevaerts(for future log readers : don't type that in without checking if the numbers are right for you)
18:55:07 Join Photoguy [0] (n=Adams@
18:55:44 Quit dany_21a_ (Remote closed the connection)
18:55:58PhotoguyGood evening.
18:56:02 Join miepchen^schlaf [0] (
18:56:26gevaertshaikubear: ok. If my theory is correct, _and_ I didn't make a stupid mistake with the numbers, it should now work
18:56:33 Join dany_21a_ [0] (
18:56:45*haikubear tests
18:57:15haikubearbut the output changed somewhat
18:57:41haikubearnow: Partition 1: 0x08 7130 MB
18:58:05 Quit TheSphinX^ ("XChat@Linux")
18:58:15gevaertsBut still no luck?
18:58:33haikubearsame error message
18:59:09gevaertsLet me check the numbers
18:59:40 Join nuonguy [0] (
18:59:55nanokis there a way to charge a iaudio x5 via usb, while playing? (i know there is for sansa, ipods and a few others, can't figure it out for the iaudio. couldn't find anything in the manual..
19:01:17pixelmananok: I believe there currently is none (couldn't figure it out for my M5 either... ;) )
19:04:38 Join miepchen^schlaf_ [0] (
19:04:40 Quit haikubear ("CGI:IRC (Ping timeout)")
19:05:10funmanthe Sansa AMS OF doesn't seem to use peripheral-to-peripheral or memory-to-memory transfers
19:05:37funmanand happily writes 1 in bits described as 'should write 0' (not only in the DMA controller)
19:06:30 Join haikubear [0] (n=46488ec8@gateway/web/cgi-irc/
19:06:51haikubeari dropped connection, it seems
19:07:13gevaertshaikubear: could you get that firmware.img to me somehow? How big is it if you compress it?
19:07:13 Join Dieterbe_ [0] (
19:07:38 Quit miepchen^schlaf_ (Client Quit)
19:07:47*gevaerts is pretty sure that there's something in there that confuses the bootloader
19:07:50 Join miepchen^schlaf_ [0] (
19:09:48 Quit miepchen^schlaf (Read error: 60 (Operation timed out))
19:10:24haikubearit's 24m
19:10:37haikubearyou want it over http?
19:10:52gevaertsThat works
19:12:16 Join Hillshum [0] (
19:12:29haikubearit's going
19:15:13*gevaerts downloads
19:16:28domonokyfunman: any new findings ? .. and strange that it writes 1 to where only 0 should go..
19:16:31nanokpixelma: thanks. at least i know to not waste my time trying.
19:17:06funmandomonoky: i'm on it ^^ I think the programmers didn't check their code, and since "it worked" it didn't harm
19:17:21domonokythey are just lucky.. :-)
19:17:41funmanthe code passes structures, stack pointers, and offsets in structures around, so following it is not straight
19:17:48nanokmaybe there should be a section in the manual about charging (near "getting files to your device", maybe?), with some standard statements, one of them being usb charging, stating in the case of iaudio that charging via usb while playing is _not_ an option as of now (and for other targets stating how to do it)?
19:18:23 Quit saratoga ("CGI:IRC (Ping timeout)")
19:18:24funmanah I found a function which returns the content of the channel control register
19:18:52 Quit Dieterbe (Read error: 110 (Connection timed out))
19:20:43 Join stoffel_ [0] (
19:21:02 Part dany_21a_
19:21:28 Join dany_21a_ [0] (
19:21:55nanokwould enabling this feature for iaudio x5 be easy enough to do (charging via usb while playing); it seems to be quite a handy feature for using in the car for instance..
19:22:14gevaertshaikubear: I've got it
19:22:39haikubearwait, "it"=file or "it"=problem?
19:22:39*gevaerts means "the file"
19:24:15*gevaerts will look into this after dinner
19:25:12haikubearsweet, thanks so much!
19:25:21haikubearcan you email me w/ anything you find?
19:25:54 Quit robin0800 (Remote closed the connection)
19:27:45haikubearpm get through?
19:29:29 Part dany_21a_
19:31:51 Join herrwaldo [0] (
19:32:54pixelmananok: not a bad idea (IMO). Would give me a place where to put the instructions how to get the M3 to charge at all which is a bit tricky, mildly put. The only "problem" I see is that it could be forgotten and not updated when things change since a dev might not expect something there
19:33:38pixelmaI mean the need to update the manual for such a change
19:34:07 Join tarbo [0] (n=me@unaffiliated/tarbo)
19:35:52 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
19:36:07 Quit culture_ (Success)
19:36:35nanokpixelma: yes, that's another good point: in my (short, granted) experience, the only reliable way to a full charge is with the wall charger, and outside rockbox (iaudio "bootloader", or whatever that is). that's also not mentioned in the manual, if i am not mistaking
19:44:37 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
19:46:16 Join mofux_ [0] (
19:47:43 Quit mofux (Operation timed out)
19:48:05***Saving seen data "./dancer.seen"
19:49:37 Join Strife89 [0] (n=michael@
19:53:03 Join Thundercloud [0] (
19:59:30gevaertshaikubear: I can't see anything strange in there. Do you have files on your ipod that you need to keep, or would a full wipe be acceptable?
20:04:59 Quit nplus (Remote closed the connection)
20:05:46 Join ibseo [0] (
20:06:27 Join nplus [0] (n=nplus@141.25.Globcom.Net)
20:09:57 Quit nplus (Read error: 104 (Connection reset by peer))
20:10:22 Join nff [0] (n=d88256ab@gateway/web/cgi-irc/
20:10:49advcomp2019nff, hello
20:10:58nffah hello
20:11:06 Join nplus [0] (n=nplus@141.25.Globcom.Net)
20:11:37advcomp2019nff, you can just ask
20:11:37nffwas just wondering if the classic gen pods will be geting RB?
20:12:15advcomp2019nff, there is no time lines unless you are going to work on it
20:12:32nffah wasent asking for one
20:12:43nffjust wondering if it was planed
20:12:52LambdaCalculus37Nothing is planned.
20:13:04LambdaCalculus37Ports happen when someone wants to work on one.
20:13:27 Quit stoffel_ ("leaving")
20:15:06 Join Strife1989 [0] (n=michael@
20:15:23 Quit Strife89 (Nick collision from services.)
20:15:30 Nick Strife1989 is now known as Strife89 (n=michael@
20:18:00nffwell thanks for you answers
20:18:10 Join lasser [0] (
20:18:34 Quit nuonguy ("This computer has gone to sleep")
20:20:58nffgooday ppl
20:21:05 Quit nff ("CGI:IRC")
20:22:47 Part pixelma
20:23:13 Join pixelma2 [0] (n=marianne@rockbox/staff/pixelma)
20:24:33 Quit lasser ("ChatZilla 0.9.84 [Iceweasel 3.0.3/2008092816]")
20:25:58 Quit herrwaldo (Remote closed the connection)
20:26:13 Join herrwaldo [0] (
20:27:39 Quit Photoguy ("Leaving")
20:30:19 Join Zagor [242] (
20:31:08 Quit Hillshum ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
20:31:30 Join tessarakt [0] (
20:35:29funmandomonoky: i found the offsets in the main dma structure, now I 'just' have to find where they are set (it should be different for each dma line)
20:35:41 Join MethoS- [0] (
20:43:34funmanoh I found where it's set, the value is loaded from another offset in the struct :'(
20:44:01 Quit miepchen^schlaf_ ()
20:44:14 Join miepchen^schlaf [0] (
20:50:13 Join AndyIL [0] (i=AndyI@
20:50:25 Quit MethoS (Read error: 113 (No route to host))
20:58:18 Quit Strife89 ("Gotta go for now.")
21:00:03 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
21:00:26 Quit Thundercloud (Remote closed the connection)
21:01:16 Quit AndyI (Read error: 110 (Connection timed out))
21:02:15 Quit ibseo ("quit")
21:03:42domonoky:-/ sounds like a good maze..
21:05:20funmani found the setting for memory (source of destination): width is 32 bits, burst size is 4 (binary 001)
21:10:15domonokynice, thats the first part.. :-)
21:11:23funmanwell i believe we are not limited for memory transfers
21:11:33funmanbut we'll test that later :)
21:14:23 Quit Acksaw (Read error: 104 (Connection reset by peer))
21:14:38 Join Acksaw [0] (
21:22:01 Quit jhulst (Remote closed the connection)
21:22:22 Quit Rob2222 (Read error: 104 (Connection reset by peer))
21:22:48 Join Rob2222 [0] (
21:23:44 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
21:29:34 Quit Dieterbe_ (Client Quit)
21:29:46 Join Dieterbe [0] (
21:33:53 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
21:39:16 Quit nplus (Remote closed the connection)
21:40:13 Join Lear [0] (i=chatzill@rockbox/developer/lear)
21:45:13bertrikn1s, any luck yet with the si4700?
21:46:11n1sbertrik: i got it to work so i could listen to the radio, but have not really done the integration into rockbox yet
21:47:04n1sthe problem was that i was trying the wrong input on the dac :)
21:48:08***Saving seen data "./dancer.seen"
21:50:12 Join Thundercloud [0] (
21:56:56 Join lasser [0] (
21:58:37LambdaCalculus37n1s: Nice! :)
21:58:41*LambdaCalculus37 hands n1s a beer
21:58:55 Quit pixelma2 ("-")
21:59:00 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
21:59:20 Quit LambdaCalculus37 (" ajax IRC Client")
21:59:39Zagorthe oldest compiler we use is the native gcc on cygwin...
22:00:20Zagorand it processes dependencies differently from 3.4.6
22:01:17Zagorbut I think I have got it nailed now
22:02:38funmanbertrik: did you try to find the equivalent of the "bit-banged i2c" you found in the fuze firmware, on other models ?
22:04:51bertrikI tried to find it for the clip, but couldn't find it
22:05:32funmandid you look at the diagnosis menu code ? (iirc it's located in the otg_functio library block)
22:05:34bertrikwe do still have some unused GPIOs on the clip I think
22:05:48amiconnZagor: Ahem, Interix has gcc 3.3 ...
22:06:12funmanmost of them are on gpiod (only pin7 is used)
22:06:17Zagoramiconn: and someone is building rockbox with that?
22:06:29bertrikfunman, no so far I basically haven't looked any further than the first block
22:06:30 Quit sarixe ("Ex-Chat")
22:06:48amiconnWell, it's a documented method
22:07:02bertrikfunman, do you think the list of GPIOs on the sansav2hardwaremapping wiki is still correct/complete for the clip?
22:07:04amiconn(but no sims on interix yet - they need to be crosscompiled)
22:08:26 Nick Llorean1 is now known as Llorean (
22:08:34ZagorI see no point in supporting such old versions.
22:09:05 Quit kugel (Remote closed the connection)
22:09:25funmanbertrik: yes
22:09:47amiconnIf the dependencies are generated using the target/ sim gcc instead of native gcc, there will be no problem
22:10:19Zagorin simbuilds, native and target gcc are the same: 3.4.4
22:10:33amiconnOn Interix they won't be
22:11:16amiconnThe native gcc is 3.3, and there is no newer (working) one, but sims will need the mingw crosscompiler, as if you would build a win32 sim on linux
22:12:54ZagorI don't care about interix. I have no desire to support Microsoft software even Microsoft themselves refuses to support.
22:13:15bertrikfunman, I would expect i2c pins to have a pull-up, therefore read as high when switched as an input. Pins B4 and B7 on the clip could be i2c because they read high as input and do not seem to be used for anything else.
22:13:25 Quit Bensawsome ("The awsome is gone :(")
22:13:59 Join Bensawsome [0] (n=Bensawso@unaffiliated/bensawsome)
22:14:33Bagderinterix is also the only non-open environment that could be used
22:14:44Bagdernot to mention that amiconn is the only one who ever did ;-)
22:14:55amiconnThat's not true
22:15:16Bagderis there anyone using it and actually preferring it to cygwin?
22:16:21amiconnI don't know whether XavierGr uses it now. He did try it at least. And there was someone else who used it in the past, but I don't remember the nick right now
22:16:39 Join aarcane [0] (
22:17:25amiconnAs said, I don't expect problems if the same gcc is used for building the deps and for the actual build
22:17:54 Quit reacocard (".")
22:18:13Zagoramiconn: the problem isn't differing version for dep and compiles. the problem is old gcc versions don't build deps the same way, and hence the build will fail.
22:19:00amiconnYes I understand that. My point is that the build itself will *never* see the native gcc on Interix
22:19:14amiconn(well, except for tools of course)
22:19:24 Quit {phoenix} (Remote closed the connection)
22:27:45 Quit jeffdameth (Read error: 110 (Connection timed out))
22:29:30 Quit bmbl ("Woah!")
22:29:56 Join jeffdameth [0] (
22:29:58bertrikfunman, so far most of the low-level driver stuff seems to be in the first firmware block, I also would the radio communication code to be there somewhere
22:31:03 Join w23 [0] (
22:35:18funmanbertrik: i suggested that the diagnosis mode could point to where the FM code is located (in the first firmware block)
22:35:41gevaertshaikubear: still there?
22:36:05funmandomonoky: the OF uses 8 * 32bits for SD transfers (half of the FIFO, like expected)
22:36:59funman4 * 32 bits for i2sout and i2sin
22:37:20domonokyah, for both source and destination ?
22:38:13funmani think yes (i'm a bit lost in all this code)
22:38:50funmani only see one instance of burst size/width in the structure, and i also see evidence that the OF doesn't do peripheral-to-peripheral or memory-to-memory DMA, so I guess the setting is peripheral specific regardless of the direction
22:39:05 Quit n1s ()
22:40:12 Join FOAD_ [0] (
22:40:13domonokyso theoretically we should now have all info needed to get DMA running...
22:40:30*funman would emphasize 'theoretically'
22:41:31 Quit FOAD (Read error: 60 (Operation timed out))
22:41:39 Nick FOAD_ is now known as FOAD (
22:41:41 Quit tessarakt ("Client exiting")
22:41:52 Join DerDome [0] (
22:43:58 Join toffe82_ [0] (
22:44:49funmanfor now I don't read/write interrupt status (clear) registers, but the DMA interrupt isn't raised at all :/
22:47:01domonokyand the terminal count interrupt is enabled, right ? :-)
22:47:40gevaertsHow do you debug early-boot issues?
22:48:49domonokygevaerts: depends on how early... :-)
22:49:06gevaertsdomonoky: not early enough to get the bootloader printf()
22:49:15gevaertsBut not late enough to get logf
22:49:27domonokyif you dont have lcd or leds, delay loops are helpfull..
22:49:43gevaertsI think I can probably do splash()es
22:49:57 Quit herrwaldo (Remote closed the connection)
22:49:57gevaerts(from firmware/, urgh)
22:50:13domonokypanicf ?
22:50:27 Join herrwaldo [0] (
22:50:27gevaertsIt does that by itself :)
22:50:56 Join sarixe [0] (
22:52:17funmandomonoky: yes, i just enabled it. I wasn't sure of its exact meaning (whether _the current LLI_)
22:52:39domonokyfunman: if you could provide your DMA testcode, i would try my luck with it tomorrow.. :-)
22:53:39funmandomonoky: currently it's mixed with plugins modifications, just give me one minute to sort the essential
22:55:33 Quit herrwaldo (Remote closed the connection)
22:56:14 Join herrwaldo [0] (
22:58:30 Quit toffe82 (Read error: 110 (Connection timed out))
22:58:52funmandomonoky: + add target/arm/as3525/dma-pl081.c in firmware/SOURCES (forgotten in the patch sorry)
22:59:41domonokythanks, and no problem...
23:00:11*gevaerts wants to know why his latest patch in FS #9545 doesn't work properly for multi-driver storage
23:01:00 Join bluebrother [0] (n=dom@rockbox/staff/bluebrother)
23:03:32funmandomonoky: oh and you might want to remove _buttonlight_* from INT_DMAC(). The m200v4 has no button led, right?
23:03:57domonokyyes, no button_lights, i will use ther debug methods..
23:09:33 Join shotofadds [0] (n=rob@rockbox/developer/shotofadds)
23:10:49 Join tic_toc [0] (n=4670e35d@gateway/web/cgi-irc/
23:12:33tic_tocgreetings all. just wanted to pop in and say thank you to everyone who has worked on this project. i've been running my 5G iPod on Rockbox for 30 mins and have been very impressed. i'll never go back to the stock firmware. i appreciate all the hard work.
23:14:29gevaertsThanks :)
23:15:46 Quit tic_toc (Client Quit)
23:17:20 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
23:17:41 Join lars [0] (
23:18:04 Nick lars is now known as larsm (
23:18:55 Join m0f0x [0] (
23:19:09 Quit Lear ("ChatZilla 0.9.84 [Firefox 3.1b2pre/20081119072659]")
23:19:28domonokyfunman: perhaps we also have to enable DMA on the pl180 ?
23:19:29 Quit jhMikeS (Nick collision from services.)
23:19:35 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
23:20:22funmandomonoky: i just thought setting bit 3 of mcidatactrl was enough
23:20:53domonokyfunman: yes, that what i though of, you already did it ?
23:21:01larsmi want to add my cfcard to but i cant edit the page. i registered at twiki successfully but cant edit the page.
23:21:06 Join tyfoo [0] (
23:22:18funmandomonoky: yes, in pl180 code i only did this modification and remove status checking
23:26:43advcomp2019larsm, what is your wiki name?
23:27:09 Quit domonoky (Read error: 104 (Connection reset by peer))
23:27:29 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
23:27:33 Quit Zagor ("Client exiting")
23:27:37 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
23:27:59saratogayou need a proper name
23:28:41advcomp2019larsm, saratoga is right unless that is your full name
23:28:44larsmmy real name is Lars Moeller
23:28:55LloreanAnd the wiki signup says you need to use your real name.
23:29:01larsmmy nic is larsm everywhere
23:29:08LloreanThat doesn't matter.
23:29:31saratogaits not on the wiki!
23:30:21saratogadon't we have a note on the registration page saying you need a full name?
23:30:29LloreanWe used to, at least
23:31:01larsmok, how to change my name?
23:31:04saratogaah yes we do
23:31:14saratogaregister a new one
23:33:35 Join n17ikh|Lappy [0] (
23:33:38 Join kugel [0] (n=chatzill@unaffiliated/kugel)
23:33:56kugeljhMikeS: ping
23:35:12larsmLarsMoeller -> it doesn't help, still can't edit
23:40:07 Quit jgarvey ("Leaving")
23:42:39 Nick larsm is now known as LarsMoeller (
23:42:47 Quit mofux_ (Read error: 110 (Connection timed out))
23:43:06 Join mofux [0] (
23:43:12 Quit linuxstb (Read error: 110 (Connection timed out))
23:43:20 Quit Zarggg ()
23:43:49LarsMoellerplease help
23:45:26 Join LambdaCalculus37 [0] (
23:46:08pixelmanow it should :) Hope you don't plan to spam... ;)
23:46:10funmanLambdaCalculus37: just wait, did you know it's late in european timezone, where most of present people reside?
23:47:07LambdaCalculus37funman: I know that.
23:48:12***Saving seen data "./dancer.seen"
23:48:14 Quit Schmogel ("Miranda IM! Smaller, Faster, Easier.")
23:48:15 Quit petur ("Zzzz")
23:48:47 Quit bluebrother (Read error: 113 (No route to host))
23:50:48 Join Zarggg [0] (
23:52:42LloreanDo we know how to boost the Gigabeast and just aren't doing it, or is it a case of something we don't know yet?
23:53:48 Join jhulst [0] (
23:54:30BigBambiOr underclock it further (I'm thinking of battery life)
23:54:45BigBambiIt also does core voltage switching doesn't it?
23:55:09LloreanI don't know, but yeah, I was assuming we'd want to move it to switching between "fastest" and "rather slow" like all our other targets.
23:55:26LloreanI was also curious what codec benchmarks would look like at 532
23:55:35*linuxstb_ wants a turbo button ;)
23:55:51LloreanSince apparently APE is at 52% for the highest compression now
23:55:55gevaertsLlorean: you mean you want ape -c5000?
23:56:02linuxstb_That's insane.
23:56:05LloreanI don't *want* it
23:56:18LambdaCalculus37linuxstb_: Of course it's "insane". ;)
23:56:19*gevaerts would be happy if his mini showed the ramdisk in the file browser
23:56:21LloreanI just want to see it have a number higher than 100% even if still not realtime for actual use
23:56:52*LambdaCalculus37 bets that at 532MHz, the beast possibly *could* decode -c5000 APE files in realtime
23:57:41 Quit BlakeJohnson86 ("Leaving.")
23:57:43gevaertsIf it scales linearly, it would have 4% margin today, which is about 20MHz
23:57:52LloreanOn the iPods we needed to get to about 120% for realtime, but we probably need less on faster targets, since "core Rockbox stuff" takes away a much smaller percentage the more total power you have.

Previous day | Next day