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 2008-11-20

00:00:19 Quit sarixe ("Ex-Chat")
00:01:48 Join BlakeJohnson86 [0] (
00:02:56jhMikeSkugel: re ping
00:05:35 Quit funman ("leaving")
00:06:35gevaertspreglow: I think FS #9545 is ready for more eyes now, so if you want to try your sd driver for real with it, go ahead!
00:08:10 Quit balou (Remote closed the connection)
00:08:13 Join balou [0] (
00:08:35 Quit Thundercloud (Remote closed the connection)
00:09:17 Join ajonat [0] (n=ajonat@
00:09:25 Quit saratoga ("CGI:IRC")
00:09:33 Quit bertrik ("Leaving")
00:09:33 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
00:14:07*gevaerts now has both a ramdisk and a CF card in his mini. Unfortunately rockbox doesn't want to mount the ramdisk yet
00:15:15gevaertsUSB sees both
00:16:51 Quit ender` (" How long a minute is depends on which side of the bathroom door you are on. -- Theory of relativity")
00:23:04LarsMoeller how to get these hex values in the last table? (Kingston Elite Pro 133X 16GB)
00:24:32 Join Am0nXz [0] (n=normanz@
00:25:40 Join culture_ [0] (
00:29:34kugeljhMikeS: hm, my fading ain't longer too
00:31:11kugeljhMikeS: btw: I implemented your propose (it works), but I'm not entirely beware of the advantages. I think the code is mainly less readable, as I wouldn't connect SYS_TIMEOUT to fading, maybe you can clear it up for me?
00:32:41 Quit lasser (Read error: 145 (Connection timed out))
00:33:20 Part Am0nXz
00:38:56 Quit culture_ (Read error: 104 (Connection reset by peer))
00:39:07 Join culture [0] (
00:39:38jhMikeSkugel: that queue gets pretty busy with messages as it is and so lowering the traffic there seemed a good idea
00:40:56 Quit LarsMoeller ("ChatZilla 0.9.84 [Firefox]")
00:41:40 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
00:42:28 Quit bmbl (Client Quit)
00:44:06 Quit Nico_P (Read error: 104 (Connection reset by peer))
00:44:53 Quit haikubear ("CGI:IRC (EOF)")
00:45:00 Quit LambdaCalculus37 ("Ka-chunka")
00:45:59 Quit faemir (Remote closed the connection)
00:56:11amiconnlinuxstb_: ping
00:56:13kugeljhMikeS: do you have a preference which fading to use? I suppose your patch?
00:56:21kugelon the beast that is
00:56:34linuxstb_amiconn: Hello.
00:56:56 Quit shotofadds ("Leaving")
00:56:58amiconnlinuxstb_: Do you know how large 'absres' in the 3.98 filter can be?
00:57:04jhMikeSkugel: for beast, the hardware looks much better since it's always smooth as glass no matter the brightness
00:57:40amiconnThere's a possible optimisation (saving the division by three in that line), but only if absres is guaranteed not to become > INT_MAX / 3
00:57:58kugeljhMikeS: yep, that's what I think too, although my patch is (for me) already smooth at brightness of 3. At least it's very noticeable
00:58:31jhMikeSkugel: the way the increased brightness range works, changing current ranges causes brightness glitches so those also show in fading
00:58:31 Join groundup [0] (n=chatzill@
00:58:53linuxstb_amiconn: Sorry, I've no idea. I never looked deeply into the codec.
00:59:12amiconnOn ARM, gcc uses 'smull' for this / 3, but (absres * 3) would still be faster
01:00:52 Part toffe82_
01:01:16groundupI am very new to using Rockbox. In fact, I just installed it about 3 minutes ago. What is the best resource for finding out everything I can on customizing it? In particular, I want to know if I can use a picture as a background
01:02:04kugeljhMikeS: ah ok. I can just consider from what Lambda told me and he said it's quite smooth (but also that yours is a little smoother). Anyway, I'd be in favor of the hardware fading too.
01:02:08scorchegroundup: we like to recommend the manual
01:02:36kugeljhMikeS: have you looked at the settings I made? They should work for the gigabeat too with a proper #define
01:02:49linuxstb_amiconn: Yes, I see what you mean. It's hard to know if that's going to be safe though.
01:03:58jhMikeSkugel: I was looking and it seems compatible.
01:04:47 Part captainkwel
01:05:12 Quit herrwaldo ("Konversation terminated!")
01:05:41kugeljhMikeS: I had to reimplement some of the settings, since OFFON setting uses bool and the pwm TABLESETTING int. Is there a way to use int's with OFFON setting? A flag maybe?
01:05:58jhMikeSamiconn: voice on the beast should be working fine now
01:06:09amiconnjhMikeS: Yes, almost
01:06:11jhMikeSkugel: you can use true/false with an int
01:06:24jhMikeSamiconn: what else is an issue? I fixed the PCM driver.
01:06:40amiconnThe silence after boot is gone, and the delay at the beginning seems to be fixed too
01:07:27kugeljhMikeS: yes, but it gave warnings with incompatible pointers, isn't bool 1byte wide only?
01:07:51amiconnBut sometimes the end of a voice clip gets swallowed. This end doesn't vanish then though, it's played before the next clip, or next music
01:07:52jhMikeSkugel: bool is 32 bits
01:08:15kugelok, then the issue was at another point, I'll look into again
01:08:27amiconnbool can be different width depending on platform and even optimisation level
01:09:09jhMikeSamiconn: That sounds strange. I'm not experiencing that issue.
01:09:42amiconnNever assume that a bool* is compatible to an int* or similar without special precautions
01:11:23jhMikeSamiconn: Is there a recipe to get that to happen? Is it only .talk clips or all menus?
01:12:06amiconnI get it when playing a file and 'talk files' is set to 'numbered'
01:12:46amiconnAlso, the swallowed beginning seems to be not completely fixed.
01:13:18amiconnHmm, wait
01:14:01jhMikeS"talk files" set to "numbered" seems to be ok here.
01:16:26 Quit DerDome (Read error: 110 (Connection timed out))
01:22:14 Join Thundercloud [0] (
01:24:07 Join Slack [0] (
01:25:20 Quit tyfoo (Client Quit)
01:25:47 Join hillshum [0] (
01:29:54amiconnjhMikeS: I didn't have the fix in my build, even though I thought I did. It's working nicely now
01:30:31jhMikeSamiconn: hehe. I thought it should.
01:31:26jhMikeSIt was trying to write the FIFO with the whole module disabled which doesn't really work well.
01:31:51 Nick JdGordon|zzz is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
01:40:37 Join grndslm [0] (
01:45:17 Join LambdaCalculus37 [0] (
01:45:19 Quit MethoS- (Remote closed the connection)
01:45:27 Join sin613 [0] (
01:46:27kugeljhMikeS: What has to be done to get fading committed?
01:46:33 Quit mofux (Read error: 104 (Connection reset by peer))
01:46:47 Quit Thundercloud (Remote closed the connection)
01:48:14***Saving seen data "./dancer.seen"
01:53:12jhMikeSkugel: For the thread stuff? I don't know. I'll do 'beast HW stuff separately.
01:53:16 Join CaptainKewl [0] (
01:56:37kugeljhMikeS: ams sansas (at least e200v2,fuze and likely c200v2) will be able to use the thread fading too
01:58:49 Quit neddy (Read error: 104 (Connection reset by peer))
01:59:19jhMikeSMake sure the addition to each driver at the low level is minimal for each target?
01:59:57kugeljhMikeS: yea, the needed modifications are minimal. I factored most of the code out to backlight_thread_fading.c/h
02:00:58 Join neddy [0] (n=john@nat/sun/x-dd5fc4d49564cbd0)
02:01:30kugelbut that could probably go into backlight.c as well
02:01:39 Quit culture (Read error: 110 (Connection timed out))
02:02:47jhMikeSeek, isn't backlight.c enough of a maze already?
02:03:19kugelyea, that's what I initially thought too (thus I used the extra file).
02:03:39kugelI meant, if it's wanted, it could go into that
02:06:03jhMikeSstuff for the sim should be factored away into "sim drivers" or something.
02:07:12kugelor uisimulator/backlight.c
02:07:49kugeluisimulator/sdl rather
02:08:12kugeljhMikeS: Seems that I can use F_T_INT flag for OFFON to use int vars?
02:09:19jhMikeSI don't think true/false is incompatible with ints, you just get 0 or 1 for the value.
02:10:17jhMikeSThough for HW fading, since it's fixed, off/1000ms seems like a good idea since you know what you get.
02:10:57kugeljhMikeS: that might apply to beast, but not to thread fade targets
02:11:28UnhelpfuljhMikeS: you can't assume that all "true" values are 1 - any non-zero integer evaluates as true
02:11:41 Quit balou (Remote closed the connection)
02:11:44 Join balou [0] (
02:13:43kugeljhMikeS: you cannot know the fading time without calclulating it from the current brightness
02:14:17jhMikeSUnhelpful: I know that is true. I usually take context into account.
02:14:37UnhelpfuljhMikeS: sorry, then... the context thing occurred to me afterward :/
02:15:26jhMikeSkugel: For thread yes, I know. Llorean suggested that but I wonder if it's a good idea unless you really can set the time.
02:18:04Unhelpfulyou could have a thread fading that fades over a set time, it just wouldn't actually set the brightness each time it's called. probably more complicated than we'd want?
02:21:31kugeljhMikeS: it won't be much more than a "nice-to-know" information
02:21:31jhMikeSIf you do sleep delay = time/brightness then it would be constant (or close enough)
02:22:40JdGordon[12:08:13] <kugel> jhMikeS: Seems that I can use F_T_INT flag for OFFON to use int vars? <- that might not work
02:22:58jhMikeSkugel: I think I can get along with simple on/off just fine
02:23:00kugelJdGordon: yea, I already tried and it didnt
02:24:09kugelJdGordon: I guess the comment "variable type for the setting" confused me
02:37:10 Nick fxb is now known as fxb__ (
02:38:37*kugel might just not use a macro for that
02:47:20 Quit sin613 ("Leaving.")
02:49:28 Join Darksair [0] (n=user@
03:01:11 Join z35 [0] (
03:09:37*jhMikeS goes to work on other matters while kugel works out settings stuff :)
03:11:24kugeljhMikeS: doesn't look like I'll get it that easy. I think I'm just gonna keep it like it is
03:11:50kugelI would need to create a new struct bool_int_setting or something like that
03:13:25jhMikeSisn't there one for arrays of values that have other text and labels?
03:18:20kugelI didn't spot such one
03:18:50kugelbool_setting would just be perfect, except for the callback which obviously requires the var to be a bool
03:21:11kugeljhMikeS: well, the settings are working, it's just that I have to do this at 2 places
03:22:10*kugel leaves an updated patch on the tracker and goes to bad
03:22:13kugelbed even
03:22:31 Quit kugel ("ChatZilla 0.9.84 [Firefox 3.0.4/2008111318]")
03:22:46 Quit aarcane ("Leaving")
03:25:17 Part pixelma
03:25:40 Join pixelma2 [0] (n=marianne@rockbox/staff/pixelma)
03:37:43 Quit neddy (Read error: 110 (Connection timed out))
03:38:46 Join reacocard [0] (n=reacocar@WL-112.CINE.HMC.Edu)
03:46:39 Quit JdGordon (Remote closed the connection)
03:46:55 Join neddy [0] (n=john@nat/sun/x-214d213f74b4d691)
03:48:17***Saving seen data "./dancer.seen"
03:48:53 Join JdGordon [0] (
03:51:11 Join sarixe [0] (
03:51:17 Quit jhMikeS (Nick collision from services.)
03:51:23 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
03:52:05 Quit JdGordon (Read error: 54 (Connection reset by peer))
03:55:09 Join JdGordon [0] (
03:58:14 Quit LambdaCalculus37 ("Ka-chunka")
04:06:45 Join miepchen^schlaf_ [0] (
04:06:51 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
04:11:38 Quit saratoga ("CGI:IRC (EOF)")
04:19:07 Quit reacocard (".")
04:19:34 Join reacocard [0] (n=reacocar@
04:19:45 Quit jhulst (Remote closed the connection)
04:20:13 Join blkhawk- [0] (
04:21:40 Quit JdGordon (Read error: 54 (Connection reset by peer))
04:21:40 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
04:22:58 Join _Kopfgeldjaeger [0] (
04:23:48 Quit Kopfgeldjaeger (Read error: 104 (Connection reset by peer))
04:23:51 Nick _Kopfgeldjaeger is now known as Kopfgeldjaeger (
04:23:55 Quit tvelocity ("Αποχώρησε")
04:26:44 Nick Darksair is now known as Awaysair (n=user@
04:30:52 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
04:35:47 Quit blkhawk (Read error: 113 (No route to host))
04:36:11 Nick blkhawk- is now known as blkhawk (
04:39:15 Join aarcane [0] (
04:42:32 Quit aarcane (Client Quit)
04:46:22 Quit jhMikeS (Nick collision from services.)
04:46:28 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
04:46:52 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
04:47:58 Quit Bensawsome ("The awsome is gone :(")
04:49:33 Join aarcane [0] (
04:54:57 Join HBK- [0] (
04:54:57 Quit HBK1 (Read error: 54 (Connection reset by peer))
04:58:46 Nick Awaysair is now known as Darksair (n=user@
04:59:13 Join Horschti [0] (
04:59:44 Nick HBK- is now known as HBK (
04:59:59 Quit Horscht (Nick collision from services.)
05:01:58 Quit aarcane (Remote closed the connection)
05:02:12 Join aarcane [0] (
05:04:37 Join Bensawsome [0] (n=Bensawso@unaffiliated/bensawsome)
05:42:42 Quit aarcane (Client Quit)
05:45:51 Quit Horschti ("electromagnetic radiation from satellite debris")
05:48:20***Saving seen data "./dancer.seen"
05:49:40 Quit Bensawsome ("The awsome is gone :(")
05:50:14 Quit Seed ("cu, Andre")
05:59:19 Join aarcane [0] (
06:09:19 Quit aarcane (Client Quit)
06:12:40 Quit jhulst (Remote closed the connection)
06:15:34 Join aarcane [0] (
06:19:49 Quit aarcane (Client Quit)
06:20:37 Join aarcane [0] (
06:27:51 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
06:27:59 Quit groundup ("ChatZilla 0.9.84 [Firefox 3.0.4/2008111317]")
06:37:51JdGordondoes anyone know if you can redefine the format declerations in tagnavi_custom.config or if you can only add new ones?
06:39:07 Quit aarcane (Read error: 60 (Operation timed out))
06:40:27 Join aarcane [0] (
06:40:31LloreanJdGordon: No clue, but it'd take you like 10 seconds to test.
06:40:56JdGordonasking is easier
06:41:45LloreanYeah, but now there's nobody to answer
06:47:36 Quit Zarggg ()
06:48:10Unhelpfulprinciple of least surprise says that if it's in there, you can modify or remove it. not that that proves anything.
06:50:21Unhelpfulbetween 1/2 and 1/3 of the way to having a testable scaler... mostly depending on how much harder the vertical scalers turn out to be than the horizontal ones.
06:51:18Unhelpfulare there a macros somewhere for the screen pixel dimensions? right now, i'm just using w=240, which is fine for anything i have around to test on...
06:53:33JdGordonLCD_WIDTH and LCD_HEIGHT
06:53:55JdGordoncryptic names aye? :D
06:55:33 Quit hillshum ("Leaving")
06:57:11Unhelpfulonly cryptic if i have to start by guessing where they might be defined. i'm setting LCD_WIDTH as the limit for scaler output width. i don't see any reason to ever scale a bitmap to a size larger than the display :P
06:58:26JdGordonshould be defined in config.h I tihnk
06:59:08 Join nuonguy [0] (
07:08:25Unhelpfuli'm really not sure at all what some of the stuff in the existing scaler is doing. :/
07:14:18 Quit aarcane (Remote closed the connection)
07:14:57 Join aarcane [0] (
07:34:34 Join Bagderr [241] (n=daniel@rockbox/developer/bagder)
07:35:01 Nick Bagderr is now known as B4gder (n=daniel@rockbox/developer/bagder)
07:37:47 Quit aarcane (Remote closed the connection)
07:39:00 Join aarcane [0] (
07:40:30 Quit ajonat ()
07:45:45 Nick fxb__ is now known as fxb (
07:46:34 Nick fxb is now known as fxb__ (
07:48:24***Saving seen data "./dancer.seen"
07:49:08 Quit aarcane (Read error: 104 (Connection reset by peer))
07:50:06JdGordonis there any reason not to close 6069? apart from being really outdated, the latest version adds alot of stuff which probalky isnt needed
07:53:29LloreanJdGordon: Isn't that just going to be completely useless with Zagor's changes anyway?
07:53:44JdGordonall the makefile changes yes
07:53:57JdGordondo we bother much about making sure DEBUG builds work?
07:54:11*JdGordon is trying to keep <200 open bugs :p
07:55:50LloreanI dunno
07:56:02JdGordonmeh, it can always be reopened later
07:58:36 Join ender` [0] (
08:05:05 Quit miepchen^schlaf_ ()
08:07:01JdGordonI assume 5230 is still a bug, but can someone with an ipod and/or e200 check? seems to work on my e200
08:14:31 Quit planetbeing ("leaving")
08:14:43 Quit courtc (Read error: 60 (Operation timed out))
08:21:12 Join courtc [0] (n=court@unaffiliated/courtc)
08:27:52 Quit reacocard (".")
08:36:47 Quit BigBambi (Read error: 113 (No route to host))
08:43:07LloreanJdGordon: It used to be pretty irregular, so it'd be hard to confirm whether it's gone or not, sadly
08:46:51 Join bluebrother [0] (n=dom@rockbox/staff/bluebrother)
08:48:47 Join Rob2223 [0] (
08:56:14 Quit bluebrother (Read error: 60 (Operation timed out))
08:57:26 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
09:03:06 Quit Llorean (Read error: 104 (Connection reset by peer))
09:03:26 Join Llorean [0] (
09:04:13Unhelpfuli *do* still think that the resize_nearest ought to be nuked, except perhaps as an option for B&W-only targets? i don't think we want to be scaling icons, or theme images.
09:05:01 Join fyre^OS [0] (
09:05:01 Quit fyrestorm (Read error: 104 (Connection reset by peer))
09:05:07B4gderI too think we should avoid resizing such images
09:05:08LloreanTheme images won't really work anyway, unless we get the ability to position things as a percentage of the screen rather than absolutely
09:05:14LloreanSo, in general, I'm for avoiding resizing them.
09:05:55B4gderalso, resize means more CPU usage and in these cases rather pointless and we should avoid wasting resources
09:06:10Unhelpfulalso, about the dynamic buffer usage part... some people want to use the scaler outside of album art, possibly in plugins. perhaps the scaler should provide a function that returns the needed buffer, and then plugins could allocate that from their remaining plugin buffer, if they can, or by stealing audio buffer, whatever. it still gets a static buffer done away with.
09:06:33B4gderUnhelpful: yes, sounds like a plan
09:06:54 Quit Rob2222 (Read error: 110 (Connection timed out))
09:07:55Unhelpfuli'm really pretty much inclined to say that smooth upscaling and downscaling of 24-bit rgb and 8-bit grey images covers every reasonable use, cut anything else out, and call it done
09:08:13amiconnUnhelpful: B&W targets don't have album art, hence won't have scaling in the core. Also those are usually the most binsize & memory restricted targets
09:08:38Unhelpfulamiconn: don't build scaler at all on B&W targets?
09:09:02amiconnNo, just plain bmp loading (for icons, wps bitmaps etc)
09:09:22amiconnThe scaler might still be useful in the pluginlib though
09:09:38 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
09:09:58 Quit BHSPitMonkey ("Ex-Chat")
09:12:45 Quit CaptainKewl (Read error: 110 (Connection timed out))
09:13:18Unhelpfulnow, if only i could get the darn thing to work. for some reason, the bmp reader is failing before it's passed as many rows as it claims are in the file...
09:13:28 Join petur [0] (
09:16:01 Join Nibbler [0] (
09:31:35 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
09:33:23 Join lasser [0] (
09:34:05Zagorhmm, with a bit of post processing we might get dependency generation down from 1500 forks to... 5?
09:34:56Zagorgcc can take any number of input files, but it always strips the path from the target it outputs. so we need to put that back.
09:35:20Zagorbut that could be done with a single sed/awk/perl call
09:35:37B4gderthere's also -MT -MQ
09:36:00B4gderbut I've found them inadequate in the past
09:36:11Zagoryes but multiple MT puts multiple targets before a single :, which is not what we want
09:38:07 Quit bmbl (Read error: 60 (Operation timed out))
09:38:47 Join Thundercloud [0] (
09:48:27***Saving seen data "./dancer.seen"
09:54:23 Join DerDome [0] (
09:56:27 Quit Nibbler (Read error: 113 (No route to host))
09:57:41 Quit pixelma2 ("-")
09:57:51 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
10:02:17 Quit DerDome ("Leaving.")
10:03:10 Join tvelocity [0] (
10:08:48 Join spiorf [0] (
10:12:27 Join jeffdameth1 [0] (
10:13:44 Quit jhulst (Read error: 113 (No route to host))
10:15:56 Quit jeffdameth (Read error: 60 (Operation timed out))
10:20:47 Quit jhMikeS ()
10:28:33 Join JdGordon_ [0] (
10:28:36 Quit Thundercloud (Remote closed the connection)
10:28:40 Quit JdGordon_ (Read error: 104 (Connection reset by peer))
10:37:59 Join Juxta [0] (
10:38:19JuxtaHello, any admin i can ask a question?
10:38:28GodEateradmin ?
10:38:35GodEateradmin of what ?
10:38:57JuxtaI want to use a figure from and i want to ask for permission ? I don't think we own that.
10:39:19GodEaterwe have
10:39:19Juxtai know
10:39:26Juxtathats what i ment :p
10:39:36JuxtaCopyright 1999-2008 by the contributing authors.
10:39:43ZagorJuxta: what figure?
10:39:59Juxtarockbox kernel and a few screens
10:40:09 Join mofux [0] (
10:40:20*GodEater guesses at more homework.
10:40:21ZagorJuxta: and where do you want to use it?
10:40:46Zagorah right you are the one who asked here about the kernel here a couple of days ago?
10:40:57Juxtayea lol you remember hehe
10:41:06Juxtayea.. and i just wanted "proper" permission
10:41:32Juxtamy teacher will demand it anyways, so "a written" agreement from an author would be good enough
10:41:46JdGordondont give it to him so he fails the assingment :D
10:42:05Juxtahehe then i will have to manipulate a conversation ^^
10:42:34Zagoras I said then you need to contact the author of that image, Miika Pekkarinen
10:44:29Zagorwith "screens" do you mean screenshots?
10:50:07Zagorfyi: I plan to commit #9534 today
10:51:47*B4gder buckles up
10:53:13B4gderwe have have an extra beer tonight and celebrate this!
10:53:22B4gderuhm, replace one have with 'can'
10:53:29Zagorcan can? ;)
10:53:40B4gderno cans, just pints!
11:04:39 Join DerDome [0] (n=DerDome@
11:04:55LinusNZagor: while you're at it, remove the depencendy for make zip
11:05:11 Nick Darksair is now known as Awaysair (n=user@
11:25:04rasherZagor: have you considered crosscompiling sims? (using mingw32 on Linux, I think is the only current, and possible, use)
11:25:16 Join PaulJam [0] (i=PaulJam_@
11:28:01 Join culture [0] (
11:37:04Zagorrasher: that could be interesting
11:42:41rasherZagor: it appears to be working at least a good deal through the compile
11:43:28 Quit tvelocity ("Αποχώρησε")
11:46:25 Nick Awaysair is now known as Darksair (n=user@
11:48:30***Saving seen data "./dancer.seen"
11:52:34rasherZagor: it "pretty much" works. UI256.bmp doesn't end up in the correct dir, and the executable doesn't get a pretty icon, but the output does in fact appear to work
11:53:45*Zagor spots some sneakily-named bitmaps
11:53:54Zagorrasher: nice
11:54:23ZagorI never understood the UI256 thing. does it have to be in the root? how is it used?
11:54:48rasherZagor: Actually, the .rocks might be wrong, I didn't manage to run a plugin, but I've no time to test further (and codecs might be broken as well, I assume)
11:55:00rasherZagor: It has to be in the root, and it's used if you start rockboxui with −−background
11:55:11rasherthat is, the same folder as rockboxui(.exe)
11:55:23Zagorok. I'll put it there then.
11:55:47ZagorI've moved a number of files out of there because I like it relatively clean.
11:56:14rasherZagor: if you want to test yourself, simple build instructions are here:
11:56:36rasherAnd the sim runs in wine, so you don't even need to touch Windows at all
11:59:35 Join LarsMoeller [0] (
12:01:01 Quit linuxstb_ (Read error: 110 (Connection timed out))
12:01:05 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
12:02:15 Join pvbcharon [0] (n=charon@
12:04:18 Join tyfoo [0] (
12:07:30 Join MegafEee [0] (n=Linux@unaffiliated/megaf)
12:16:46 Nick Darksair is now known as Awaysair (n=user@
12:17:26 Join n1s [0] (n=nils@rockbox/developer/n1s)
12:21:07Zagorbrace for commit
12:21:37*JdGordon is taking bets on the score
12:24:38*n1s hides under his desk
12:25:17*GodEater grabs his desk with both hands
12:25:34 Quit MegafEee (Read error: 145 (Connection timed out))
12:25:45*gevaerts expects a new record
12:26:32Zagordoubters :)
12:26:42*GodEater wonders if anyone's told dev.cgi to expect the sudden hammering it's about to receive
12:26:50 Join Lynx_ [0] (
12:27:29*gevaerts tweaks his build server to help the record along ;)
12:27:49JdGordonremoving the compiler doesnt give it a big enough scrore
12:28:03 Quit havien (Read error: 60 (Operation timed out))
12:28:42preglowthat's what i call a commit
12:29:33preglowwhy, this looks golden
12:29:35 Join pvbcharon_ [0] (
12:30:23Zagorthe commit message has one of the longest subject lines I've seen
12:30:38 Join kugel [0] (n=chatzill@unaffiliated/kugel)
12:30:50*GodEater watches the countdown on dev.cgi
12:31:08 Nick Awaysair is now known as Darksair (n=user@
12:32:21GodEater"Build should have been done (n) seconds ago" :)
12:34:03 Join havien [0] (
12:34:35 Quit pvbcharon_ ()
12:35:36*GodEater suspects around 26 suddenly hung build jobs around the globe.
12:36:11ZagorOh Ye of Little Faith
12:36:29*pixelma guesse that the first compile in a new way just takes longer
12:36:51preglowcheckout probably takes a while too, maybe
12:37:06preglowwhat with umpteen files changed
12:37:26 Quit lasser (Remote closed the connection)
12:38:11*GodEater hands in his doubters badge
12:38:41*pixelma checks if everything that should be is in the zip ;)
12:39:16Zagorwe also need testing that the binaries work as they should. plugins & codecs etc.
12:40:36ZagorI don't expect problems but as they say: faith is good. control is better. :-)
12:41:35 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
12:43:12ZagorB4gder: the file size delta table isn't working. it says 0/0 in every column.
12:43:20preglowlet the users do that!
12:43:55kugelZagor: yay nice
12:46:13 Quit pvbcharon (Read error: 110 (Connection timed out))
12:47:13 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
12:47:21n1sZagor: the location of changed and i think that is what the delta table script it looking for
12:47:40Zagorn1s: ok
12:47:45kugelZagor: what's the specific reason to CC apps/ code after plugins and codecs?
12:48:45Nico_PZagor: congrats for the commit :)
12:48:52 Join LarsMoeller_ [0] (
12:49:02ZagorNico_P: thanks
12:49:09Nico_PI shall test with a sansa build :)
12:49:14 Join rvvs89 [0] (
12:49:30Zagorkugel: there is no reason. make tries to build apps first, but things in there depend on other things like bitmaps and speex voice codec and so those must be built first.
12:49:50n1shmm, my newly built sim tells me this: "failed to load archos/_temp_codec0.dll
12:49:50n1sdlopen(archos/_temp_codec0.dll): archos/_temp_codec0.dll: invalid ELF header
12:50:00n1swhen trying to play a file
12:50:29kugelZagor: ah ok, makes sense
12:50:33Zagorn1s: ok, I'll look at that
12:50:35n1ssame for plugins, "invalid ELF header"
12:50:48n1sZagor: x86_64 if that matters
12:52:21 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
12:52:43*pondlife pops in to congratulate Zagor a bit more
12:53:49pondlifeNo sarcasm, but I'm also getting the sim playback failure.
12:54:00pondlife"failed to load archos/_temp_codec0.dll"
12:54:09Zagoryup, sounds like a sim plugin/code link issue
12:54:50Zagorhehe: archos/.rockbox/rocks/games/blackjack.rock: VMS Alpha executable
12:54:55Zagorumm, not quite what I had in mind...
12:57:20kugelZagor: "make bin" followed by "make bin" doesn't work? do I need to rerun configure or something?
12:57:52Zagordefine "doesn't work"
12:58:05kugelor doesn't it want to compile because I didn't change anything?
12:58:22Zagor"Nothing to be done" means precisely that
12:59:16*Nico_P 's sansa doesn't seem to be recognized by linux
12:59:37Nico_Pdmesg reads "config 128 interface 0 altsetting 0 endpoint 0x81 has an invalid bInterval 100, changing to 10" and no device seems to appear
12:59:38 Join LarsMoeller__ [0] (
13:00:17 Quit nuonguy ("This computer has gone to sleep")
13:00:26 Quit Juxta ()
13:01:23kugelNico_P: have you trued fs#6800 already (on a h300)?
13:01:45Nico_Pkugel: I have no H300 at hand
13:02:02n1sZagor: a target build on my gigabeast also displays "Codec failure" when trying to play a file, plugins work there though
13:02:31*Nico_P was being stupid about the sansa... forgot to set it to MSC instead of MTP
13:02:56Zagorn1s: yeah I found a mistake in plugins.make and codecs.make
13:03:38Zagorn1s: try changing $(CODECFLAGS) to $(CODECLDFLAGS) in the %.codec pattern rule at the bottom of codecs.make
13:03:56Zagorno, wrong.
13:04:09n1sWhich also shows an unrelated bug, the "Codec failure" splash comes from a different thread than ui, it enven splashes on top of plugins...
13:04:58 Join LarsMoeller___ [0] (
13:04:58***Alert Mode level 1
13:04:58DBUGEnqueued KICK LarsMoeller
13:04:58DBUGEnqueued KICK LarsMoeller_
13:04:58***Alert Mode level 2
13:04:58DBUGEnqueued KICK LarsMoeller__
13:04:58DBUGEnqueued KICK LarsMoeller___
13:04:58***Alert Mode level 3
13:05:31Nico_Pwoah woah my sansa's screen is fading to white
13:06:29kugelNico_P: same here
13:07:39 Quit LarsMoeller (Read error: 110 (Connection timed out))
13:07:40 Nick LarsMoeller___ is now known as LarsMoeller (
13:07:40DBUGEnqueued KICK LarsMoeller
13:07:40***Alert Mode level 4
13:07:44Nico_P3.0 is ok... phew :)
13:07:46kugelZagor: sansa build doesn't seem to boot
13:08:16Zagorkugel: which sansa? It ran fine on my c200 yesterday
13:08:25 Join lasser [0] (
13:09:30pixelmaI remember the "screen melting bug" reported on e200s in the beginnings of the port
13:09:42kugelZagor: the screen fades out slowly after the bootloader says "calculating crc32"
13:10:00Zagorsim plugin/codec fix committed
13:10:03pixelmanot sure what there was done to solve it but I guess one can find something in the commit logs
13:11:55 Join LarsMoeller___ [0] (
13:11:55***Alert Mode level 5
13:11:55***Alert Mode level 6
13:11:55DBUGEnqueued KICK LarsMoeller___
13:11:55***Alert Mode level 7
13:11:55***Alert Mode level 8
13:11:55***Alert Mode level 9
13:12:06 Join einhirn [0] (
13:12:09Zagorn1s: can you try different codecs on the gigabeat s? do they all fail?
13:12:33preglowseems like the bug where you need to make codecs twice is gone too
13:13:45n1sZagor: bot vorbis and mp3 fail, those are the only fiels on it ATM
13:14:36CtcpIgnored 6 CTCP requests in 6 minutes and 57 seconds at the last flood
13:14:36CtcpVersion from J-23!
13:15:03 Quit Lynx_ (" Try HydraIRC -> <-")
13:15:32kugelso let's iron out the sansa build bug
13:15:52n1sZagor: with your latest fix, my sim build spits out a "cp: kan inte ta status på "/home/nils/rockbox/h300sim/adx.elf": Filen eller katalogen finns inte" for each codec, and codecs aren't linked...
13:17:18Zagorcopy/paste mistake
13:17:26 Quit LarsMoeller_ (Read error: 110 (Connection timed out))
13:17:28n1splugins are fine though :)
13:17:48Zagortry now
13:18:27 Quit LarsMoeller___ ("ChatZilla 0.9.84 [Firefox]")
13:18:50kugelpixelma: could you meanr13531
13:19:02kugellol, nice number
13:21:17n1sZagor: now i get the same cp error for plugins too, and codecs still :/
13:21:44Zagorn1s: eh? I didn't change plugins.make
13:21:56***Alert Mode OFF
13:22:27 Quit LarsMoeller__ (Read error: 110 (Connection timed out))
13:23:17n1sZagor: I don't know, i find make scary and confusing... anyway it seems to be looking for all plugins in subdirs, so those which have subdirs work...
13:27:13Zagorshow me one of the error lines
13:27:52 Quit LarsMoeller (Read error: 110 (Connection timed out))
13:29:26n1shmm, i was mistaken, it looks in the right place it seems, and the file is there... "cp: kan inte ta status på "/home/nils/rockbox/h300sim/apps/plugins/flipit.elf": Filen eller katalogen finns inte"
13:30:02n1sif i run make a second time they are linked
13:30:14n1sor cp'ed
13:30:38Zagoryou have flipit.o in the same directory, right?
13:31:10Zagorthen that error message is very strange.
13:31:46Zagorremove one rock and run make again
13:33:19n1sit recreates the .rock just fine
13:33:25Zagorand no error?
13:33:46Zagorwhat if you rm -rf apps/plugins?
13:33:48kugelZagor: do I need to rebuilt the bootloader? my svn-of-a-few-days-ago bootloader doesn't boot into rockbox
13:34:33Zagorkugel: no there should be no functional differences
13:35:03pixelmakugel: did you try with bootloader and a daily build? Maybe this helps tracking down where the error is...
13:35:23kugelpixelma: it worked after jhMikeS'
13:35:26kugellatest commit
13:35:42kugelit's Zagor's commit. Nico_P also has the problem
13:36:07Zagorscreen fading doesn't mean it doesn't load. it just means the screen is fading.
13:36:32Zagoryou need to be very specific when describing problems, so we are not chasing ghosts
13:36:37kugelyea, true. I cannot get to the main menu
13:36:47Zagorcannot get, or cannot see?
13:37:15kugelZagor: I already told you: the screen fades out after the bootloader says "calculating crc32"
13:37:35kugelscreen fades out as in turns all black slowly over time
13:37:58Zagorthat sounds like rockbox loading fine, but does something wrong in the lcd setup
13:38:19n1sZagor: if i delete the whole apps/plugins dir the cp errors are back
13:38:37Zagorn1s: and you don't get any .rock files?
13:39:00n1sexactly, a missing dependency on the elf files, maybe?
13:40:36Zagorno, I think it is a make problem I've ran into before. try removing the $(shell ) around the cp line
13:40:59n1s$(SILENT)cp $(BUILDDIR)/$*.elf $@ ?
13:41:07pixelmakugel: "fading black" as in backlight fades out or the LCD turns black - and is this with your backlight fading changes, Nico_P mentioned fading to white...
13:41:35kugelpixelma: the lcd turns black, the backlight is on all the time
13:42:02kugelpixelma: well, it has very different colors when that fading starts, but in the end it's black
13:42:28n1sZagor: that fixed the plugins :)
13:42:41ZagorI wonder why I don't get the problem
13:42:42kugelnow it started with beeing white, then going over to orange then brown then black
13:43:16 Quit culture (Connection timed out)
13:43:24Zagorn1s: yeah I've found that make has a bug in that it doesn't run all commands sequentially(!). make function calls are executed before "normal" exec lines
13:43:41Zagorso the $(shell) call was called before the gcc linker
13:45:28 Join Bensawsome [0] (n=Bensawso@unaffiliated/bensawsome)
13:46:03n1show fun, same fix in codecs.make works for codecs too, now only one small peculiarity left on the sim, midi.rock ends up in .rockbox/rocks while it should be in .rockbox/rocks/viewers when making install
13:46:20kugelpixelma: it looks like lcd_enable(false) without clearing the lcd before
13:47:56Zagorn1s: ok, checking that.
13:48:31***Saving seen data "./dancer.seen"
13:49:57kugelZagor: could that be a linking error or some kind? I can't imagine how your commit breaks the build
13:51:19Zagorwe simply have to be more imaginative...
13:51:48Zagorlinking errors are usually caught in the linker since we have a relatively complex linker script
13:52:02Zagorbut obviously there is something wrong
13:52:15kugelZagor: could it be related to the mi4-format?
13:52:40Zagorcan we get some more people testing sansa, please? c200, c100 etc. does it happen on all?
13:53:20Zagorkugel: my guess based on your description is that it loads and starts. if mi4 was wrong it wouldn't be able to start, as I understand it.
13:53:57Nico_Pkugel: have you tried the rev before Zagor's commit?
13:54:33kugelah yea. but it's not responsive too. the backlight and the button light stays on, and you can only turn it off with holding power 10s (normally you only need to hold it 3s)
13:54:37kugelNico_P: yes
13:55:12 Join what [0] (n=79b45a2d@gateway/web/cgi-irc/
13:55:44pixelmagoing to test on my c200
13:57:50n1stesting on c200 here, it freezes on a black screen with backlight on directly after the bootloader
13:58:07 Quit spiorf (Remote closed the connection)
13:58:37 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
13:58:37 Quit HBK (Read error: 104 (Connection reset by peer))
13:58:48 Join HBK [0] (
14:02:00n1sZagor these are the last few lines of a "make bin V=1" and the whole compression thing looks fishy,
14:03:15Zagoreww, that's not supposed to be done for sansas. well spotted.
14:05:52 Quit what ("CGI:IRC (EOF)")
14:06:08Zagoroh but it isn't done either. that's just the output of the line. but compression is never done.
14:07:10 Quit Darksair (Remote closed the connection)
14:07:22LinusNZagor: are the files linked in the same oerder?
14:07:31Zagorif it did, you'd see "Image too big, making a compressed version!"
14:07:50 Join Darksair [0] (n=user@
14:07:54ZagorLinusN: every single object file? most should be, but probably not all.
14:08:11LinusNZagor: crt0.S might be important
14:08:28Zagorthat one is loaded by the linker script, so that works
14:08:41Zagoralso the linker complains loudly if it doesn't find the start symbol. been there... :)
14:09:03LinusNwould it make sense to compare md5 sums?
14:09:17Zagorno, they will inevitably differ
14:09:19LinusNi mean, your makefile changes should produce identical binaries
14:09:29LinusNhmm, except the version string
14:09:44Zagorand only if the libraries and object files are linked in the exact same order
14:10:02LinusNwell, shouldn't they?
14:11:53Zagorif we have such a fragile system that moving a bitmap struct from one place in the image to the other, we have problems
14:12:03 Join Schmogel [0] (
14:12:32Zagorinsert "crashes" at a suitable place :)
14:16:45B4gderis rockbox-info.txt done correctly?
14:17:07B4gderindicates it isn't
14:17:21Zagorno some fields seem to be missing
14:17:41Zagor"Actual size:" is empty and "RAM usage:" is 0 moved to the root. I guess that's it?
14:18:03B4gderprobably, yes
14:20:15 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
14:25:08Zagor.text segment size are nearly identical, suggesting no bug chunks of unlinked code: 0x72c60 now vs 0x72cf8 for last rev
14:27:26kugelshouldn't they be exactly identical?
14:27:44JdGordongcc adds some randomness just to be fun
14:27:52 Join LambdaCalculus37 [0] (i=44a04303@gateway/web/ajax/
14:28:38n1scould anyone try a different mi4 target perhaps? might give some clues
14:29:26Zagorand lcd_init for example is in nearly the same location
14:29:43pixelmawould there be a point in trying a cygwin compiled c200 build?
14:29:59Zagorpixelma: can't hurt. it might give some clues.
14:30:40JdGordone200v1 crashes also
14:30:49kugelwe know :)
14:31:07pixelmacompiling went fine btw. (I think it was faster but am not sure if I remember build times correctly for the target build.)
14:31:17*JdGordon just saw the call for mi4 targets and only the c100 mentioned
14:31:18kugelisn't that philips the only other mi4 target?
14:31:28n1sh10 too iirc
14:31:32pixelmano, h10 too
14:31:40*pixelma too slow
14:31:47kugeloh, h10 and mrobe100 also
14:32:24n1sso let's hope the mrobe user comes here :)
14:32:26 Quit bmbl ("Woah!")
14:32:34 Join TheSphinX^ [0] (
14:32:38kugelyes, both :P
14:32:47JdGordonI'm sure if there are any and its broken we will hear from them
14:33:57kugelbut maybe it's related to PP. then ipods would also experience that
14:34:01 Quit DerDome (Read error: 110 (Connection timed out))
14:34:13kugeldid anyone test on an ipod?
14:34:27fyre^OSwhat needs testing on an ipod
14:34:35kugelcurrent build, if it boots
14:34:40fyre^OSill check right now.
14:34:44fyre^OSusing 5g
14:35:13 Quit amiconn (Nick collision from services.)
14:35:20 Join amiconn [0] (
14:35:43LambdaCalculus37I'll test too on my iPod color.
14:37:15*LambdaCalculus37 backs up his current build first
14:37:24JdGordonbroekn ipod reported in flyspray
14:37:36fyre^OSoh nice
14:37:44LambdaCalculus37JdGordon: What's the tracker number?
14:37:54JdGordon4min ago
14:38:19 Join CaptainKewl [0] (
14:38:59fyre^OSnot booting
14:39:03LambdaCalculus37Not booting.
14:39:13pixelmacygwin compiled build gives the backlit black screen too, right after the SanDisk logo. I see nothing of the Rockbox bootloader but I don't have the latest released, I think
14:39:17LambdaCalculus37fyre^OS: Is yours stuck at the Apple logo?
14:39:19pixelmaon c200 that is
14:39:22n1sZagor: H300 is the same as the Beast, codec failure (mp3, vorbis, flac, sid) but plugins working
14:39:25LambdaCalculus37Same here.
14:39:40Zagorn1s: ok, thanks
14:39:42*LambdaCalculus37 boots back into the OF
14:39:50kugelhaha, good that Zagor broke the targets with the most user :)
14:40:03B4gderthat should teach'em!
14:40:06Zagorit seems I broke all targets
14:40:18fyre^OSits good that i saw this too
14:40:24fyre^OSi had some kind of weird cache bug
14:40:26Zagornot bad for an all-green commit, eh? ;)
14:40:32JdGordonZagor: its nice to see the full paths now :)
14:40:32LambdaCalculus37iPod 3G has a red!
14:40:38fyre^OSi found files in my .rockbox dir that were over 2 gb
14:41:07JdGordonlinux e200 sim works (incase noone else tested yet)
14:41:55pixelmaa nice round score ;)
14:42:20ZagorLambdaCalculus37: that's odd. it's not supposed to run the linker until that .link file is created
14:42:26JdGordonBagder: untill this is all sorted should you maybe still a link on the build page to the last daily?
14:42:47LambdaCalculus37JdGordon: Good idea.
14:42:58Zagorah, /bin/sh: cannot create /home/rbclient/build-ipod3g/apps/codecs/ Directory nonexistent
14:43:42LambdaCalculus37Zagor: The iPod 3G is reported "broken?" on the current build page.
14:43:55LambdaCalculus37What's it linking to?
14:44:11JdGordonI just did a 3g build and it built fine.. might just be a server qwerk?
14:44:30ZagorJdGordon: it happens with massive -j (paralellism)
14:46:56B4gderwhy is $(shell) needed for cases like that?
14:47:15Zagorto avoid the bug where make runs $(call) lines before "simple" exec lines
14:47:29Zagorat I least I consider that to be a bug
14:47:57 Quit amiconn (Nick collision from services.)
14:48:04 Join amiconn_ [0] (
14:48:17LambdaCalculus37Should we announce on the ML that current builds may be broken for a little bit?
14:48:24Zagorthree functions are missing from the new e200 build, compared to the previous revision: speex_mode_query, statusbar_icon_play_mode and statusbar_icon_shuffle
14:48:41 Quit kugel (Nick collision from services.)
14:48:42 Join kugel_ [0] (
14:48:44ZagorLambdaCalculus37: probably a good idea
14:48:46 Nick kugel_ is now known as kugel (
14:48:58*Zagor sends a mail
14:50:06Zagorperhaps a note on the download page too
14:50:10Zagorinstead, even
14:50:14pixelmawhy would the absence of these function already let booting fail?
14:50:41Zagorpixelma: I don't think they are the cause of this
14:51:00Zagorrather they are not used and the new link line removes them where the old didn't
14:51:12Zagorif they were called and not linked, the linker would scream
14:51:42 Nick JdGordon is now known as JdGordon|zzz (n=jonno@rockbox/developer/JdGordon)
14:51:55pixelmaaha, can't imagine these as the cause either
14:51:59kugelI din't think they're not used. I know there's a shuffle icon in the e200 statusbar
14:53:08kugelZagor: maybe we get an undefined instruction somewhere, but we cannot see it if it is before lcd init
14:53:20Zagorcould be
14:54:43 Join kharo [0] (
14:55:18kugelstatusbar init is way after lcd init, so we would probably see it if this 3 functions are missing
14:57:45*Zagor added a big red note to
14:57:59kugelZagor: so I moved lcd init up a bit, and it doesn't fade out anymore (but it still doesn't boot)
14:59:48kugelZagor: now that I moved i2c init before lcd init it fades out again
15:00:21kugeli assume the bug is somewhere in i2c_init()?
15:02:13kugelcommented out i2c_init removes the fade out again
15:05:46n1sZagor: the plugins that have their own foo.make file are still objcopied when building a sim
15:06:17Zagorn1s: ah
15:07:17Zagorumm, not all
15:08:01n1sat least midi and doom, "file" tells me they are "VMS Alpha executable"
15:08:35Zagorright. all but two, it turns out..
15:11:49LambdaCalculus37n1s: "VMS Alpha executable"? Strange. :)
15:12:29Zagoryeah that's what an objcopied ELF file becomes
15:17:16 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
15:18:57B4gderspeaking of nothing at all, I wonder if we'd gain anything significant by adopting the linux kernel's likely() and unlikely() usage
15:19:10Zagorwhat is that?
15:19:15B4gderif(likely(wrong)) { do this; }
15:19:24n1scompiler hints?
15:19:29B4gderusing a gcc builtin thing to hint
15:19:36B4gder#define likely(x)__builtin_expect(!!(x), 1)
15:19:52n1swe do use that in the tremor codec in a few places
15:20:29B4gder"GCC hacks in the Linux kernel" => has some good ideas to borrow
15:21:31Zagorohh, switch ranges. I didn't know that.
15:21:44B4gderZagor: possibly OC and OCFLAGS could be variables and then OC could be set to cp for sims
15:21:56B4gderswitch ranges look pretty neat
15:23:30Zagorlikely() looks like something one should get used to using
15:23:53 Quit Darksair (Remote closed the connection)
15:24:13ZagorI don't know if our current architectures/gcc versions benefit from it, but it's definitely a good programming practice imho
15:24:16 Join Darksair [0] (n=user@
15:24:18B4gderand it's not very disturbing in the code either
15:24:39Zagorexactly. one could actually argue it makes the code easier to follow.
15:24:41B4gderit's also actually a hint to the casual reader of the code
15:25:14B4gderit's almost like you're my brother!
15:29:35kugelswitch ranges is basically case 1: case 2: case 3: ?
15:29:56kugelwith fallthrough that is
15:30:17Zagorkugel: yes
15:30:31kugelnot sure if using a non-standard extension is worth it if it could be done as easily with standard stuff
15:33:36 Quit amiconn_ (Read error: 113 (No route to host))
15:35:43n1skugel: a) we already use some gcc extensions and b) are pretty unlikely to start using a different compiler any time soon
15:35:53Zagorit could be rather nice if you have a value range of, say, 10000 and a dozen fields within those.
15:36:17Zagora thousand case lines isn't quite the same :)
15:36:26kugelyea, true too
15:36:36amiconnBagder, Zagor: I tried the type-checking min/max macros a while ago. They would be nice, as they're also producing better code, but they cause a ton of warnings which need fixing...
15:36:40kugelthe un/likely looks pretty interesting to me too
15:36:57 Join amigan [0] (
15:38:05Zagorkugel: did you do anymore boot tests? I would do it myself but I don't have a target here... :-(
15:38:44kugelZagor: no, I haven't time right now
15:38:50n1skugel: btw, in the likely/unlikely case it is easy to define an empty macro depending on __GNUC__
15:39:22kugeln1s: that's why I find it interesting :)
15:42:26 Join nplus [0] (n=nplus@141.25.Globcom.Net)
15:44:21Zagoranyone who can run some target tests for me?
15:44:51Zagorfirst test: remove option -Wl,−−gc-sections from root.make line 115
15:45:16n1sZagor: sure, sansa or booting target?
15:48:07Zagorwith that option removed, the .text segment is exactly the same size as before
15:48:35***Saving seen data "./dancer.seen"
15:51:08n1syes, it boots with that option removed, same deal with codecs as the other targets too
15:51:23Zagorah, great
15:51:42n1syes :)
15:52:06amiconnZagor: Is there a reason why -Wl,−−gc-sections is used for normal builds now? Previously it was only used for bootloaders
15:52:42ZagorI assumed it would remove unused code without harming anything. I was wrong.
15:52:57amiconnImho using that is unwanted for normal builds, because it causes laziness in cleaning up unused stuff. But if it is wanted, some sections need to be marked using KEEP() in the linker script
15:53:11Zagorwhat sections?
15:53:14amiconnCompare and
15:53:39amiconnvectors, for instance
15:55:20ZagorI think we should strive towards making it work with that option, but I'll disable it now
15:55:42Zagorit is very difficult to manually keep track of exactly which functions are used and not in every build
15:56:40Zagorn1s: please try removing the same option from codecs.make line 93
15:56:56 Quit sarixe ("Ex-Chat")
15:57:24Zagorn1s: did you say some plugins worked?
15:57:48n1syes, i tried some random ones
15:59:33amiconnWere there any non-arm targets affected by that? I guess not...
16:01:01Zagorgigabeat s even worked
16:01:10n1sZagor: i did rm-rf apps/codecs and rebuilt but get the same codec failure, would a make clean be needed in between?
16:01:27Zagorn1s: no that won't make a difference
16:01:53Zagorn1s: can you do a thorough plugin check?
16:03:45amiconnSH and coldfire seem to have proper KEEP() protection in both and
16:03:56amiconnARM has no .vectors but there must be an equivalent
16:04:39amiconnZagor: Look at the gigabeat S .lds files...
16:07:10 Nick Darksair is now known as Awaysair (n=user@
16:07:36 Join evilnick [0] (i=0c140464@gateway/web/ajax/
16:08:07 Quit CaptainKewl (Read error: 110 (Connection timed out))
16:09:30n1sZagor: tested all games on c200 and they seem to work as well (or poorly) as they used to :)
16:10:06Zagorn1s: great, thanks. now why doesn't codecs work...
16:11:37 Part linuxstb_ ("Leaving")
16:16:34 Join obo [0] (n=obo@rockbox/developer/obo)
16:17:28kugelZagor: regarding your commit message, it wasn't only sansa which didn't boot
16:17:40kugelit seemed to affect pp targets
16:17:52 Quit ap0 ("Ba")
16:17:59 Join ap0 [0] (
16:20:07kugelB4gder: I would've a use for likey(x) if we decide to use it
16:20:43 Quit ap0 (Client Quit)
16:21:28 Quit kharo (Read error: 110 (Connection timed out))
16:25:39 Join kharo [0] (
16:25:51 Nick fxb__ is now known as fxb (
16:29:52Zagorthe codecs are quite a bit smaller
16:30:04n1sZagor: this might be a clue, in an old file i found codec_start is the first function in the text section (from codec_crt0.o) and in one from a new build it's second after codec_main (from vorbis.o)
16:30:32B4gdera good clue I'd say
16:30:33 Join toffe82 [0] (
16:31:14Zagorin the map files I'm comparing now, they are in the exact same order
16:31:29B4gderthen perhaps not so good clue ;-)
16:31:35Zagorbut the new has a .text section of 0x83f4 bytes while the old is 0x8534
16:32:17 Nick Awaysair is now known as Darksair (n=user@
16:32:30Zagorhaha, that is because the old codec includes libcodec twice...
16:33:15Zagorso dual memcpy, memset, codec_malloc etc
16:33:50ZagorI'm surprised ld does that
16:33:55B4gderdual is much better than single
16:35:21 Join captainkwel [0] (i=2669ecc2@gateway/web/ajax/
16:35:38 Join robin0800 [0] (
16:37:14 Quit TheSphinX^ (Read error: 145 (Connection timed out))
16:38:11n1sI want to try changing the order the object files are linked into the elf, how would i do that?
16:39:41 Join dany_21a_ [0] (
16:43:23Zagorn1s: the codecs are typically just one object file and a library
16:44:35Zagorand the order is the same as before
16:44:55Zagorthere are actually two object files: codec.o and crt.o
16:48:23ZagorPLUGIN_RAM is defined differently in the lds file
16:48:30 Join {phoenix} [0] (
16:49:19n1sat least for an x5 build the order of crt0.o and codec.o in the codecs i checked is reversed
16:49:31n1sso crt0 is no longer first
16:49:44 Quit mofux (Read error: 113 (No route to host))
16:49:53n1s(x5 is the only semi recent build from before your changes i have still)
16:49:59Zagorn1s: really? the old build I'm comparing with links the crt0.o after codec.o
16:50:31 Join mofux [0] (
16:50:31Zagorin any case the code maps turn out the same. I'm pretty sure the plugin_ram thing is the problem.
16:53:13Zagoryes, it preprocesses without -DCODEC
16:54:59Zagorupdate and try now
16:56:24 Quit robin0800 (Read error: 104 (Connection reset by peer))
16:57:33n1sah, yes working now (h300)
16:58:24ZagorI'll remove the big red sign on the download page then
17:01:15ZagorB4gder: what is needed to fix dev.cgi to show size deltas again?
17:02:08B4gderrockbox-info.txt needs to be fine
17:02:18Zagorah right, that was the problem
17:07:13Zagormeh, bootloaders still need gc-sections
17:07:58n1sZagor: where should individual compiler flags for plugins and codecs be set now? IIUC most of them were lost/removed?
17:09:13Zagorthey should still be there, in their respective .make file. files (directories) which want special flags define a new pattern rule that matches only that directory
17:09:27Zagorsee doom.make for example
17:13:59Zagoramiconn: do we want gc-sections on all bootloaders? it looks like only bootbox needs it
17:18:58Zagorumm, the h100 and player deltas look a bit strange
17:19:20Zagoranyone have a h100 to test on?
17:19:58 Join nuonguy [0] (
17:21:40 Join robin0800 [0] (
17:22:18pixelmawill this change also fix the problem with garbled bitmaps when the size of a certain bitmap changes?
17:22:18*petur looks at LinusN
17:22:55ZagorI haven't heard about that problem or the cause of it, so I don't know.
17:23:37pixelmawell the bitmaps look as if the old data is used and stuffed into the new size. I'll try that then
17:25:18Zagorwas that plugin bitmaps or the core ones?
17:26:33ZagorI tried touching some bmp files, and it seems to work.
17:26:52Zagorplugin bitmaps trigger a full rebuild of all plugins, but better safe than sorry :-)
17:27:01 Part B4gder
17:27:02pixelmanice :)
17:31:39 Join Seed [0] (
17:31:46Zagoroh dear we have a new fun "comes and goes" delta game. only this time it's 100k up and down instead of 26 bytes...
17:33:14n1sit's also a pure ram delta no bin difference
17:34:11amiconnThe bitmap-of-different-size problem is most probably not fixed by this rework
17:35:32amiconnIt is an object naming problem within the .a
17:36:03Zagorok, what IS the problem?
17:36:25amiconnCurrently the bitmap objects names contain the size, but the provided variable is the same
17:37:06amiconnSo if you change the size of a bitmap, the compiler will create an .o with a different name, which is then added to the .a, i.e. doesn't replace the old one
17:37:29amiconnBut the linker will choose the first object that provides the variable it is looking for
17:37:29Zagornot at all
17:37:49Zagorthe .a contains a specific number of files that are listed in SOURCES
17:38:50amiconnYes of course. But it will also contain objects from a previous run
17:38:52Zagorand in fact the bitmaps aren't even in a library anymore
17:39:14amiconnAha, then it might be fixed. Why not, btw?
17:39:31Zagorwrong, the plugin bitmaps are
17:40:11Zagorbecause making a library is an unnecessary step for objects that we know will be included in the elf
17:41:07Zagorthe problem you describe is easily solved by removing the .a before running ar
17:41:58ZagorI must say I find it odd that ar does not have an option to create a new library
17:42:09 Quit perrikwp (" ajax IRC Client")
17:43:41amiconnZagor: You need to remove both the .a and the old objects, otherwise ar will grab them again
17:44:00Zagorthe old objects will only be grabbed if they are still listed in SOURCES
17:44:02amiconn(well, unless you specify them explicitly instead of a wildcard)
17:44:07ZagorI do
17:48:34 Quit pixelma (Read error: 110 (Connection timed out))
17:48:38***Saving seen data "./dancer.seen"
17:50:10 Join herrwaldo [0] (
17:53:48 Quit Zagor ("Client exiting")
18:01:24 Quit petur ("work->home")
18:04:02 Part pondlife
18:05:27 Join MethoS [0] (
18:06:01 Join bertrik [0] (
18:09:25 Quit synergist (Remote closed the connection)
18:10:44 Quit Darksair (Read error: 60 (Operation timed out))
18:11:21 Quit shodanX (Read error: 104 (Connection reset by peer))
18:12:21 Join shodanX [0] (
18:14:14 Quit Kohlrabi_ ("leaving")
18:18:54 Quit tyfoo (Client Quit)
18:20:41 Join netzwurm [0] (
18:20:55netzwurmis there a linux program to update the rockbox db while connected via usb?
18:23:27 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
18:24:10n1snetzwurm: there is a standalone app in the rockbox sources that can do it but it has not seen much love in quite a while so it probably doesn't work anymore...
18:24:48netzwurmn1s: that's too bad. for some reason auto-updating doesn't work for me ever and that means i have to restart my rockbox twice after transfering podcasts.
18:25:05bertrikwow, the generating dependencies step takes seriously long
18:25:25 Join perrikwp [0] (i=98212aaf@gateway/web/ajax/
18:25:52n1sbertrik: IIUC that is because dependencies for _all_ object files are gnereated in one go, earlier they were spread into one dep per dir so less noticable
18:25:52netzwurmn1s: thanks anyway.
18:26:01netzwurmn1s: what's the app called?
18:26:25n1snetzwurm: if you are able to hack some c you could probably make it work again. :)
18:26:38netzwurmn1s: i'll look into it.
18:27:17n1sthe main file is tools/database.c as you can see it relies heavily on the rockbox code
18:30:55n1srunning "make database" in the tools dir _should_ build it
18:32:33netzwurmyeah. or not :)
18:34:21n1shmm, someone moved some files around
18:34:58bertrikand sansa clip no longer builds
18:36:59 Join Strife89 [0] (n=michael@
18:39:12n1snetzwurm: even after fixing the paths it spits out a bunch of warnings and dies so it is probably quite a bit of work to get it working :)
18:42:20 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
18:45:03 Quit MethoS (Remote closed the connection)
18:45:59 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
19:01:17 Quit {phoenix} (Remote closed the connection)
19:02:53 Quit BigBambi (Read error: 104 (Connection reset by peer))
19:05:30 Join {phoenix} [0] (
19:06:32 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
19:10:39 Nick fxb is now known as fxb__ (
19:13:12 Quit dany_21a_ (Remote closed the connection)
19:13:31 Join miepchen^schlaf [0] (
19:13:40 Join dany_21a_ [0] (
19:13:51 Join aarcane [0] (
19:15:45 Quit herrwaldo (Remote closed the connection)
19:18:44 Quit robin0800 (Connection timed out)
19:19:12 Quit Strife89 ("Just removed and reseated my Belkin card. Switching computers.........")
19:19:47 Join Strife89 [0] (n=michael@
19:21:19 Join herrwaldo [0] (
19:32:10 Part LinusN
19:41:27 Quit n17ikh|Lappy ()
19:41:54 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
19:42:32 Quit tarbo (Client Quit)
19:42:44 Join Thundercloud [0] (
19:43:08 Join tarbo [0] (n=me@unaffiliated/tarbo)
19:46:51 Join mofux_ [0] (
19:47:59 Join Lear [0] (i=chatzill@rockbox/developer/lear)
19:48:40***Saving seen data "./dancer.seen"
19:48:57 Quit mofux (Read error: 60 (Operation timed out))
19:51:42 Quit miepchen^schlaf ()
19:52:23 Join petur [50] (n=petur@rockbox/developer/petur)
19:53:42 Join j-b [0] (n=jb@videolan/developer/j-b)
19:53:55 Join kharo1 [0] (
19:54:13 Join miepchen^schlaf [0] (
19:55:03 Quit kharo (Read error: 110 (Connection timed out))
19:55:16 Part j-b
19:56:24 Quit perrikwp (" ajax IRC Client")
19:56:42preglowwhat do we use the hi res timer for? backlight fading and... ?
19:58:44 Join reacocard [0] (n=reacocar@
19:59:31 Part dany_21a_
19:59:57 Join dany_21a_ [0] (
20:00:05 Part dany_21a_
20:01:08 Join mofux [0] (
20:02:43amiconnpreglow: The core uses it only for fading (on targets which can do that), but several plugins need it: metronome, everything using the greylib, video.rock, doom (on colour targets), ...
20:02:55 Join guessed [0] (
20:03:24 Join ap0 [0] (
20:04:11 Part guessed
20:07:29 Quit mofux (Read error: 60 (Operation timed out))
20:08:13 Quit mofux_ (Read error: 145 (Connection timed out))
20:08:45 Join mofux [0] (
20:13:21 Join mofux_ [0] (
20:13:39 Quit mofux (Read error: 60 (Operation timed out))
20:15:02Nico_Pcan FS #9564 be closed?
20:15:33*amiconn wonders what the heck causes this weird behaviour of test_codec with ape -c5000 on his PP5022 targets
20:16:19 Join faemir [0] (
20:17:45 Join mofux [0] (
20:19:53LambdaCalculus37Nico_P: I think it can.
20:19:57*LambdaCalculus37 goes to close it
20:24:01 Join Talyn0 [0] (n=817749e4@gateway/web/cgi-irc/
20:24:04kugelcheck your emails! I sent a mail to the dev-ml
20:24:19 Join bluebrother [0] (n=dom@rockbox/staff/bluebrother)
20:24:54 Quit Thundercloud (Remote closed the connection)
20:25:29Talyn0I installed Rockbox on an iPod v5.5 about a year and a half ago
20:25:54Talyn0and at some point the album art images started loosing a strip of pixels
20:26:15Talyn0a row about 20 pixels wide on every album
20:26:33Talyn0it shows up for a second or two when I first bring up an album
20:26:46Talyn0anyone have any ideas what this is?
20:28:23Talyn0anyone have any ideas what this is?
20:29:01 Join webguest78 [0] (n=817749e4@gateway/web/cgi-irc/
20:29:12 Join pixelma [0] (i=pixelma@
20:30:32 Quit mofux (Read error: 60 (Operation timed out))
20:31:16n1sBagder: using (un)likely() in the midi plugin actually gave a not insignificant improvement :)
20:31:29n1sso, where should the macros go?
20:32:05 Join mofux [0] (
20:32:30 Quit kharo1 (Read error: 110 (Connection timed out))
20:32:40 Quit faemir (Remote closed the connection)
20:33:23 Join kharo [0] (
20:34:20 Quit mofux_ (Connection timed out)
20:35:30 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
20:35:32 Join mofux_ [0] (
20:35:40kugeln1s: I suppose misc.h
20:35:57 Join DerDome [0] (
20:36:30*n1s was thinking more system.h this is not apps specific really (in fact it could probably be useful in firmware too)
20:36:51 Join faemir [0] (
20:39:17 Join funman [0] (
20:40:02kugeln1s: is systems.h #lncluded in apps/ code commonly? if not put it in both :)
20:40:19 Join tvelocity [0] (
20:40:21 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
20:41:34n1safaik it is, as it defines a bunch of useful macros, but having the same thing in several places is bad so i'll stick them in there and we'll just have to include it if we want them
20:43:00kugelsounds good
20:43:10kugel"put it in both" was more of a joke anyway
20:44:39pixelmathe jumping big deltas seem to still be for rombx targets (maybe depending on which server compiled and including some race conditions or so)
20:46:00pixelmaso instead of weird failing builds for those there are now weird binsize deltas
20:46:02 Quit FOAD (Remote closed the connection)
20:46:55 Quit Talyn0 ("CGI:IRC (EOF)")
20:47:41pixelmaor RAM usage
20:49:14 Quit mofux (Connection timed out)
20:49:58 Join AndyI [0] (i=AndyI@
20:50:09 Join FOAD [0] (
20:53:35*domonoky just got his first interrupt from the DMAC :-)
20:54:06funmandomonoky: nice ! ready to share your work?
20:54:09domonokyfunman: you forgot the interrup masks in DMAC_C0_CONFIGURATION...
20:54:18domonokybit 14 and 15..
20:54:38 Quit mofux_ (Read error: 110 (Connection timed out))
20:54:51domonokyi also move the dma_channele enable before asking the sd controller for data, but dont know if this matters..
20:55:00funmanoops ok
20:55:49 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
20:55:55*funman will only set bit 15 - errors can't happen O:-)
21:00:21funmandomonoky: i get fast booting after clearing the terminal count interrupt
21:00:45funmanand lang file loads fine :-)
21:02:04funmantest_codec remains stuck on 'Loading...' however
21:02:09 Quit AndyIL (Read error: 110 (Connection timed out))
21:02:14domonokyso DMA is now under our control.. :-)
21:02:36funmanand hopefully write support is a straight adaptation from sd_read_sectors
21:03:05 Join tyfoo [0] (
21:03:52funmannow the boring part : error checking, and put unique error codes all over the functions
21:05:03 Quit Lear (Read error: 104 (Connection reset by peer))
21:05:33bertrikfunman, can you still build a clip rockbox with recent svn? I get a TIMER_FREQ undeclared in grey_core.c
21:06:38funmanbertrik: a lot of plugins fail to build. I should take some time to calculate the exact timer frequency
21:07:12funmanI thought since the peripheral clock is at 64MHz, the timer would decrase 64M times per second but it seems it's quite less than that ..
21:07:42 Join merbanan [0] (n=banan@
21:08:08 Quit kharo (Read error: 60 (Operation timed out))
21:08:37funmanbertrik: my current diff with keymaps for some plugins and dma for sd
21:08:54 Join kharo [0] (
21:09:17bertrikfunman, oh I haven't done ../tools/configure in a while. I may have edited something to prevent plugins to build for clip and forgot about it.
21:09:32funmandomonoky: the api of dma-pl081.c needs to be changed a bit to support i2sin & i2sout, there is still quite a few assumptions (directions from/to peripheral, channel 0 only ..)
21:09:53funmanbertrik: well if it fails in greylib they are being built.
21:10:15domonokyyes, sure.. but we now know that it works the way we think.. :-)
21:10:46bertrikfunman, how are the clock frequencies for the ams sansas decided, i.e. why use 64 MHz and not some other frequency?
21:12:29funmanbertrik: i chose what the OF uses, but I may have made a mistake: I thought it used a 384MHz plla, but it might be 96MHz. I guess testing lower speeds would help in reducing battery life
21:15:25funmanif we want to use a fixed peripheral clock, and a variable cpu clock, we might need to enable the 2nd pll (pllb); i don't know what's the cost on battery
21:16:09funmanwhich means we would need to use the as3514 to read battery levels first .. :)
21:17:31 Quit Strife89 ("Gotta go.")
21:18:01funmandomonoky: do you want to work on cleaning up the dma code?
21:19:19domonokyfunman: my time i very limited at moment, i will not be online this weekend.. but if it isnt cleaned up till next week, i will make a try on it..
21:20:16funmanok, then if i have time i will work on it, and keep you informed by the forum if it's not ready to be committed by then
21:20:38funmanat least the functionality is there ;)
21:24:05bertrikfunman, I don't know yet what would be the best w.r.t. clocking, but I think the pll itself is not the biggest power user
21:24:45funmaniirc clocking the peripheral clock at 24MHz caused no problems, and needs no PLL
21:27:38 Join Acky [0] (
21:28:01 Join p3tur [50] (n=petur@rockbox/developer/petur)
21:28:21Nico_Pn1s: do you (or anyone) know why there is "!!" in the likelyness compiler hints?
21:29:04funmanif I try to use DMA for writing, I got a reboot after 1 or 2 seconds of uptime
21:29:07kugelNico_P: maybe to reduce every expression to 0 and 1
21:29:18n1sNico_P: afaik !!(true) == 1, !!(false) == 0
21:29:55n1sso !!4 == 1
21:30:01kugelyep, that's what I meant
21:30:01Nico_Pah, thanks
21:30:35*funman seds s/SD_READ/SD_WRITE/g
21:31:05 Quit petur (Nick collision from services.)
21:31:10 Nick p3tur is now known as petur (n=petur@rockbox/developer/petur)
21:31:21funmanyep writing support is fine, and I think read_ and write_ sectors() could be easily factorized
21:32:36 Join mofux [0] (
21:35:14funmanhopefully i2sout support will be quick to implement!
21:35:23*funman looks at gevaerts
21:37:03funmani was mentioning the sansa-meizu race ;)
21:38:54*gevaerts tries to think of ways to cheat
21:39:12*kugel hopes funman doesn't make porting work into a competition :(
21:39:19funmankugel: _friendly_ competition on how the 2 ports evolve
21:39:30 Quit Horscht ("electromagnetic radiation from satellite debris")
21:40:10*domonoky defines the gen
21:40:21funmanthe winner gets the right to buy a new device and help the other port ;)
21:40:26 Quit jhulst (Remote closed the connection)
21:40:32*domonoky defines the target of this race, that gentlemen mail...
21:40:41 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
21:40:49Nico_Pfunman: that would be with the rockbox fund then ;)
21:40:50domonokyand proposes a sansa for the winner :-)
21:40:52 Quit Acky (Read error: 60 (Operation timed out))
21:41:52 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
21:43:54 Quit Acksaw (Read error: 110 (Connection timed out))
21:44:26LambdaCalculus37Gentlemen, start your coding! ;)
21:44:34 Join Acksaw [0] ( abort!
21:45:20 Join massiveH [0] (
21:46:10funmanhm it's not perfect yet : the system deadlocks in read() in test_codec ..
21:47:38 Quit funman ("leaving")
21:48:25preglown1s: nice idea
21:48:43***Saving seen data "./dancer.seen"
21:49:34n1spreglow: wasn't my idea, Bagder and Zagor were talking about it before, but it actually helped a bit in the midi player and was pretty trivial to add so I'm happy :)
21:51:23preglowbeen thinking about it myself
21:51:33preglowshould start using it here and there in codecs
21:51:40n1sgo! :)
21:53:02n1sIt is still somehow baffeling that we can play ape files (c1000) on the pp players in realtime but not a f*ing midi >8
21:59:10 Quit LambdaCalculus37 (" ajax IRC Client")
21:59:55 Quit {phoenix} (Remote closed the connection)
22:06:40 Nick J-23 is now known as Pustelnik (
22:06:44 Nick Pustelnik is now known as J-23 (
22:06:58 Quit jhulst (Read error: 110 (Connection timed out))
22:07:14 Nick J-23 is now known as blablabla (
22:07:55 Nick blablabla is now known as J-23 (
22:09:43 Nick J-23 is now known as J-23___ (
22:10:19 Join culture [0] (
22:10:32 Nick J-23___ is now known as J-23 (
22:11:29 Quit nuonguy ("This computer has gone to sleep")
22:11:47 Join jeffdameth [0] (
22:13:19 Quit jeffdameth1 (Read error: 60 (Operation timed out))
22:14:44 Quit merbanan (Remote closed the connection)
22:15:50 Join kharo1 [0] (
22:16:40 Quit perrikwp (" ajax IRC Client")
22:16:46 Quit ap0 (Read error: 104 (Connection reset by peer))
22:16:56 Quit kharo (Read error: 110 (Connection timed out))
22:17:04 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
22:18:39preglown1s: heh, not really, a midi player is a whole synth, an ape decoder is just a stream processor
22:28:10 Join Thundercloud [0] (
22:31:39 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
22:46:20 Quit massiveH ("Leaving")
22:48:30gevaertsbluebrother, domonoky: At least on somewhat recent linux you can get from usb device to block device (and from there presumably to mountpoint) by walking /sys
22:52:11bluebrothergevaerts: interesting. Should give it a look ...
22:52:58 Quit nplus (Remote closed the connection)
22:54:32Lloreandomonoky: If gevaerts' method is correct, you could pair .rockbox folders with USB IDs and never have to worry about local HD "installations"
22:55:15domonokyLlorean: if there is already a .rockbox dir, we will find it...
22:55:22domonokythe problem is the first install.
22:56:01domonokyautodetection also uses rockbox-info.txt and many more detection methods :-)
22:56:06Lloreandomonoky: Yeah, but we want to *not* find .rockbox directories on local HDs.
22:56:27domonokybut i want that. its good for testing :-)
22:56:40LloreanWell, then we want to clearly mark them as "not real" devices.
22:56:47domonokyit better to find more,then too less..
22:56:52 Quit n1s ()
22:57:11LloreanAnother thing it'll help prevent, is accidentally installing the wrong build on a device.
22:57:28domonokywith our different detection methods, we could also assign a "liklyhood" of the result. but i thought that would be overkill..
22:57:37*Llorean has too many times forgotten to reset his configuration, and installed an iPod Nano build on his Gigabeat F, because he forgot he updated his Nano recently
22:58:05domonokyrbutil, does already complain about that..
22:58:23domonokyat least svn, dont know if it made it into the last release...
22:58:25LloreanClearly I haven't messed up since my last upgrade of RButil then.
22:58:36LloreanIt may be in 1.0.7, I haven't done it in a few months.
22:58:57domonokybefore installing it checks if the rockbox-info.txt matches you config, and complains if they are different..
22:59:15*bluebrother suggests a compile-time switch to enable local hdd support ;-)
22:59:55Lloreanbluebrother: I'd think displaying them after the "real" devices, and possibly in italics or otherwise visually distinguished (with a note somewhere else on screen as to what it means) would also work
23:00:08domonokythe problem is to integrate such a "local hd detection" into the detection system..
23:00:14 Join jhulst_ [0] (n=jhulst@unaffiliated/jhulst)
23:00:41bluebrotherthe main problem is someone to find the time ;-)
23:01:12 Quit jhulst (Read error: 104 (Connection reset by peer))
23:01:41domonokyLlorean: if you write a method to get the HD -> USB-Id relation on all three OS, i would try to integrate it :-)
23:02:15*bluebrother would prefer USB -> HD resolving ;-)
23:02:31domonokythe way doesnt really matter...
23:03:07domonokybut we dont have this, especially not for all three supported OS.... sadly.
23:03:18bluebrothertrue, but we need that way for detection
23:03:46domonokybluebrother: you could always iterate over the HDs and map it this way..
23:05:45bluebrotherwe're already doing this for UNC -> drive letter
23:06:21domonokyso the best we get at moment is to list all detected device.. maybe the presentation could be improved.. one target with multiple mountpoints (drop down perhaps) ?
23:06:21 Join MethoS [0] (
23:06:21bluebrotherwhile it works I somewhat dislike that it's somewhat inefficient −− but better than no resolving at all
23:09:12 Quit Jabone (Remote closed the connection)
23:09:15 Join Jabone [0] (
23:09:52gevaertsbluebrother: If you plug in a USB storage device, you'll get (e.g.) /sys/block/sdb. In there there is a "device" symlink, and if you get the target path from that you can extract usb bus number and device id, the same as shown by lsusb
23:10:12gevaertsYou can probably also go the other way. It's a bit of a maze
23:12:57gevaertsYou may be able to get more help via the linux usb people, at
23:15:32 Quit domonoky (Read error: 104 (Connection reset by peer))
23:18:07 Quit bertrik ("Leaving")
23:27:03 Quit mofux (Read error: 110 (Connection timed out))
23:27:21 Join mofux [0] (
23:28:10 Join jgarvey [0] (
23:31:34*gevaerts suddenly realises one possible reason why his multidriver tests didn't work properly
23:34:59 Quit obo (Read error: 110 (Connection timed out))
23:35:15 Quit kharo1 (Read error: 110 (Connection timed out))
23:35:44*gevaerts decides that he is less of an idiot today than he was yesterday
23:35:54 Join kharo [0] (
23:35:59 Join obo [0] (n=obo@rockbox/developer/obo)
23:36:16Llorean It's working now?
23:36:56gevaertsIf you format an 8mb filesystem as FAT16, you need to make sure the number of sectors per cluster is small enough
23:39:54*gevaerts does a few more test builds before committing
23:41:54 Join LambdaCalculus37 [0] (
23:45:31 Join mofux_ [0] (
23:48:21 Quit mofux (Operation timed out)
23:48:46***Saving seen data "./dancer.seen"
23:50:13 Join eightfold [0] (
23:52:03 Part eightfold

Previous day | Next day