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

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

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

#rockbox log for 2009-04-18

00:00:25 Quit kushalone ("Leaving. I cannot promise to be back but most likely will.")
00:01:02 Join kkurbjun [0] (n=kkurbjun@rockbox/developer/kkurbjun)
00:03:11bluebrothermaybe we can construct two (or three) letter shortcuts for the parts? As in Archos -> ar, Apple -> ap
00:04:39rasherWhat does ipodmini1g shorten to?
00:04:48 Quit ender` (" error compiling committee.c: too many arguments to function")
00:04:56rasheror mini1g I guess
00:05:20bluebrotherap-min-2g, ap-vid-55, ap-vid-5g, sa-e2x-pp, co-m3, to-gig-f
00:05:35rasherNot sure if it should be appleipodmini1g applemini1g.
00:05:36bluebrotherand so on. A bit cryptic, I need to admit.
00:05:43*rasher doesn't see why we need this
00:06:22bluebrotherI'd add some separators between the parts. Of course we could start using CamelCase instead but simply putting it together feels kinda strange
00:07:06*rasher doesn't feel strongly about this - we've not used any separators so far
00:07:08n1splease no CamelCase
00:07:16rasherYeah, I don't like that either
00:07:47bluebrotherme neither, but not using separators feels ... well, strange. If we put in more information it'll get harder to read
00:15:11 Quit n1s ("Lämnar")
00:26:43amiconnI wonder what advantage Cygwin 1.7 might have for a rockbox build environment
00:27:08amiconnIt's currently in beta, and according to the website a lot has changed much that Cygwin 1.5.x and 1.7 can even be installed in parallel
00:31:40***Saving seen data "./dancer.seen"
00:34:51 Quit Lynx_ (" HydraIRC -> <- It'll be on slashdot one day...")
00:34:52 Quit mrkiko (Read error: 104 (Connection reset by peer))
00:38:40 Join pinguen [0] (
00:39:03amiconnkkurbjun: Laaate answer.... what kind of corruption did you observe?
00:39:43 Join mrkiko [0] (
00:50:51 Join cspotcode [0] (
00:51:41Unhelpfuldo we have anybody who knows a fair amount about DCT? i was wondering just how large a computational cost switch jpeg decoder to using optimized 1D IDCTs would have. this would let us use an 8x16 IDCT, for example, when decoding subsampled chroma channels, so that a separate scaling step would not be needed. jpeg-7 already has a full set of 2NxN and Nx2N IDCTs, but i suspect this would be quite a bit too much code...
00:52:32 Quit midijunkie (Read error: 104 (Connection reset by peer))
00:53:04 Quit JdGordon| (" ajax IRC Client")
00:53:22bertrikI think saratoga knows a lot about various transforms
00:53:29 Quit itcheg (" ajax IRC Client")
00:53:42 Quit bluebrother ("zzz")
00:54:00 Join _lifeless [0] (n=lifeless@
00:54:19Unhelpfulright, i'll prod him about it later then. :)
00:54:30 Join Reptile211 [0] (
00:54:58 Quit __lifeless (Remote closed the connection)
00:55:48 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.8/2009032609]")
00:55:56 Quit _lifeless (Remote closed the connection)
00:56:44 Quit Reptile211 (Read error: 104 (Connection reset by peer))
00:57:24 Join Reptile211 [0] (
01:03:08 Join itcheg [0] (i=62db4767@gateway/web/ajax/
01:03:56 Quit bertrik ("Leaving")
01:06:16 Quit cspotcode ("Leaving.")
01:07:29 Quit Reptile211 ("ChatZilla 0.9.84 [Firefox 3.0.8/2009032609]")
01:12:16 Quit itcheg (" ajax IRC Client")
01:12:58 Join japc [0] (
01:21:04rasherBagder, bluebrother:">
01:21:35 Quit pinguen (Read error: 110 (Connection timed out))
01:21:54rasherSome of those only make sense for builds, and some even only for bootloader builds.
01:22:21rasherI'm not sure how to handle this best without ending up in a "multiple lists" situation
01:32:34 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
01:32:51 Join saratoga [0] (n=41becb3b@gateway/web/cgi-irc/
01:33:42kkurbjunamiconn, the scrolling lines would get offset each time it went down a line
01:34:06kkurbjunwhen I say line, I mean line of pixels
01:34:37kkurbjunit would be offset by some amount - the problem was just in the math for the rectangle update
01:36:10JdGordonkkurbjun: i've put a request for images in the forums to help the port out :) depending on my weekend ill try and finish the touchscreen stuff for the wps
01:38:02kkurbjunJdGordon: nice, hopefully someone can help out :)
01:38:12JdGordonhehe yeah
01:38:28JdGordonif you want to add any icons to the requested list, go for it :)
01:38:46rasherIs this for the mrobe500?
01:38:53*JdGordon requests the arty people in the channel check out
01:39:05rasherThe sim still doesn't seem to work quite right (windows anyway)
01:39:25kkurbjunrasher, what's the problem?
01:39:58rasherThe window looked like it isn't the right size, unless you fixed it in the last few weeks
01:40:18kkurbjunthe image for the sim isn't sized right
01:40:32kkurbjunI scaled the screen down from 640x480 to 320x240
01:41:01kkurbjunsince for now at least it would make the port much more usable and presentable
01:45:32rasherOkay, looks "right" now
01:49:49 Quit amiconn (Nick collision from services.)
01:49:53 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
01:49:59 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
01:50:10 Join pixelma_ [50] (n=pixelma@rockbox/staff/pixelma)
01:50:10 Quit pixelma (Nick collision from services.)
01:50:12PaulJam__I'm just curious, could the software USB stuff (like FS #10116) also work on the H300 via the secondary USB 1.1 host port?
01:50:30 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
01:51:27kkurbjunPaulJam__: I believe it could, but it would take some work
01:53:20 Join bubsy [0] (i=Bubsy@
01:53:30 Join bs66_1 [0] (
01:55:25 Quit DerPapst ("Leaving.")
01:55:38 Part toffe82
01:56:10PaulJam__kkurbjun: thanks
01:56:29 Nick JdGordon is now known as JdGordon|afk (n=jonno@rockbox/developer/JdGordon)
01:57:49 Quit jordan`` (Read error: 60 (Operation timed out))
01:57:57 Quit japc (Read error: 60 (Operation timed out))
02:00:31 Join jordan` [0] (
02:01:09 Quit Thundercloud (Remote closed the connection)
02:12:46 Quit bs66_ (Read error: 110 (Connection timed out))
02:17:17 Quit mrkiko ("leaving")
02:22:20 Quit BlakeJohnson86 (Remote closed the connection)
02:26:55 Join BlakeJohnson86 [0] (
02:27:21 Join hillshum [0] (n=hillshum@unaffiliated/hillshum)
02:31:42***Saving seen data "./dancer.seen"
02:35:19hillshumWhat is the best way to describe the rockbox version of gcc? (version?) (compiler?)
02:39:32Unhelpfulthe "rockbox version" is just gcc built as a cross-compiler with a few patches that help it to produce better code for our targets.
02:40:30hillshum I know, I'm writing a wiki page on the different options for windows developers, what should I call it?
02:41:09BigBambiI don't really get the question
02:41:15hillshumHere is my early work
02:41:23hillshumWhat term should I use?
02:41:23BigBambi"The cross compiler" ?
02:41:37hillshumO yes. Thanks
02:41:58BigBambiAre you sure that those are emulation platforms?
02:42:25saratogatechnically they're not emulators
02:42:32saratogaand anyway, we already have a wiki page for this
02:42:45BigBambisaratoga: That's my point :)
02:43:10BigBambiAnd we don't have a "Rockbox compiler"
02:46:53 Quit fyrestorm ("( :: NoNameScript 4.22 :: )")
02:47:12 Join fyrestorm [0] (
02:47:48BigBambihillshum: If you are going to edit the wiki, please take some care to e.g. leave spaces between words
02:53:21 Quit moos ("Rockbox rules the DAP world")
02:59:34 Join AndyI [0] (i=AndyI@
03:03:14 Quit DataGhost (Nick collision from services.)
03:03:21 Join DataGhost [0] (i=dataghos@unaffiliated/dataghost)
03:09:07 Quit saratoga ("CGI:IRC (Ping timeout)")
03:09:12*hillshum finds confusion among new devs on windows
03:09:52 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
03:11:26 Quit MT (SendQ exceeded)
03:11:32 Quit PaulJam__ (".")
03:12:19 Join cmwslw [0] (
03:12:57 Join MT [0] (n=chatzill@
03:13:15hillshumBigBambi: some of those names don't have spaces
03:14:09 Part cmwslw ("Ex-Chat")
03:14:51 Join midijunkie [0] (
03:18:47 Quit midijunkie (Client Quit)
03:21:33 Quit hillshum ("Leaving")
03:42:24 Join Strife89 [0] (n=nds@
03:52:14 Join joshwood [0] (n=794fd262@gateway/web/cgi-irc/
03:54:13joshwoodis there anyone here who can pls help with the wmware image for compiling rockbox? I need to get internet access
04:12:33 Quit joshwood ("CGI:IRC (EOF)")
04:13:20 Join joshwood [0] (n=794fd262@gateway/web/cgi-irc/
04:18:01 Quit joshwood (Client Quit)
04:18:59 Join g54pcys [0] (n=chatzill@
04:29:48g54pcysdo you know anything about VMWare and the rockbox image?
04:30:28scorcheg54pcys: assuming you are joshwood, that sounds like a VMware configuration issue
04:30:51g54pcysyeah I am
04:31:12g54pcysI can ping my VM from my host OS ok...
04:31:17g54pcysI setup a static ip
04:31:37 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
04:31:44***Saving seen data "./dancer.seen"
04:35:06g54pcysis there much work to setting up the build environment from scratch in linux? I might wait a few days until the new ubuntu is released, it's a nice linux version for beginners like me :)
04:35:47Strife89g54pcys: See the simple Linux Guide to compiling on the Wiki.
04:35:59Strife89I wrote it as an Ubuntu user.
04:37:48g54pcysah sweet, I missed that guide. Easy as, I'm fine with the command line etc and general linux stuff, but don't know everything off top of my head
04:40:01 Quit g54pcys ("ChatZilla 0.9.84 [Firefox 3.0.6/2009011913]")
04:42:12 Join Tekno_Cowboy [0] (
04:45:03Tekno_CowboyHello, might anyone here be able to give me a reason why my Ipod video 30gb might think it has a 109GB drive?
04:46:56krazykitfilesystem corruption could be one cause
04:49:16davidfg4I had that problem once
04:49:32Tekno_Cowboyhow did you fix it davidfg4?
04:49:39davidfg4Can't remember what I did, but a reformat would probably fix it
04:51:16Tekno_CowboyLooks like I'm going to have some fun :D
04:51:37Tekno_CowboyThanks for the advice!
04:55:31krazykityou could try doing a fsck first (in windows, chkdisk or chkdsk or something, i'm not sure of the name)
04:56:27Tekno_CowboyI was just searching for what to use in Fedora 10
04:59:00Tekno_CowboyThanks, now I just need to figure out which drive to run it on.
04:59:53 Quit BXCracer (Remote closed the connection)
05:00:09krazykitpoking around in dmesg might help with that
05:01:14Strife89Back up first!
05:02:30Tekno_CowboyAlways good advice to back up. I only have 3 drives, so it was easy to figure out.
05:03:50 Quit miepchen^schlaf (Read error: 101 (Network is unreachable))
05:03:57 Quit froggyman ("CGI:IRC (Ping timeout)")
05:05:14 Join beta_ [0] (
05:05:26 Nick beta_ is now known as beta2k (
05:09:16 Quit beta2k (Client Quit)
05:09:47Tekno_Cowboywell, fsck is trying to read non-existent sectors, looks like a reformat is needed
05:16:35 Quit tvelocity (Remote closed the connection)
05:20:29 Quit perrikwp (" ajax IRC Client")
05:20:48 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
05:30:37 Quit Carnif3X ()
05:40:23 Quit Strife89 ("Bed.")
05:49:18 Join xavieran [0] (
05:50:04xavieranHi guys, I've had to remove rockbox from my ipod mini 2g as it is faulty and I need to send it back to the manufacturer, but I'm having some issues.
05:51:35xavieranI get the folder with ! on it when I reboot the ipod after putting the mbr back (I had to do it manually using dd) and if I run ipodpatcher to try and write my firmware partition I get: Firmware partition doesn't contain Apple copyright, aborting.
05:51:59xavieranm running linux and have access to a windows computer.
05:54:40 Quit perrikwp (" ajax IRC Client")
05:55:24 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
05:57:00xavieranDoes itune work in wine?
05:57:07xavieranI might have to use that.
05:57:25scorchexavieran: have you seen the IpodManualRestore wiki page?
06:00:03xavieranscorche: yes, I followed the instructions but the ipod gives me the folder w/ !
06:00:35xavieranI think I see my problem. I didn't unzip the .ipsw folder!
06:02:00xavieranYes, I will just have to connect my wall charger :) Thanks
06:03:18xavieranHow long does it typically take to restore?
06:03:35scorchehowever long it takes ;)
06:04:34xavieranOkay :)
06:04:55 Join Riku [0] (
06:05:03 Quit Lss (Read error: 104 (Connection reset by peer))
06:05:12xavieranbtw, I have to give you guys a really big congrats (and thanks) for making rockbox.
06:05:28xavieranI'm sure you've heard it all before, but I really like it :)
06:07:16 Part xavieran
06:11:31 Quit Seed ("cu, Andre")
06:17:45 Part Keripo
06:19:48 Quit perrikwp (" ajax IRC Client")
06:20:09 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
06:21:15 Quit MT (Connection reset by peer)
06:21:43 Join MT [0] (n=chatzill@
06:25:34 Quit kadoban (Remote closed the connection)
06:31:47***Saving seen data "./dancer.seen"
06:32:22 Join Tomers [0] (
06:38:24 Quit fyrestorm (Read error: 104 (Connection reset by peer))
06:39:51 Join jason_ [0] (
06:39:57 Nick jason_ is now known as CaptainKwel (
06:40:26 Part Tekno_Cowboy ("Konversation terminated!")
06:41:10 Join fyrestorm [0] (
06:45:28 Nick Riku is now known as Lss (
06:58:36Unhelpful*should* core JPEG reading support direct (ie, unscaled aside from reduced/enlarged decode) output?
07:01:36linuxstbWhat do you mean? Isn't that just "scaled" 1:1?
07:03:45Unhelpfullinuxstb: yes, but the bmp reader has a separate code path for "unscaled" output without calling resize_on_load. this saves a few multiplies per output pixel, and also doesn't require that the bmp buffer be any larger than needed for the output, while the scaler requires an extra twelve rows in the output width for scratch space.
07:04:30Unhelpfulif we're reading a file that is *exactly* the size we want, the jpeg reader could benefit from the same thing... it doesn't seem to me that that will happen terribly often, though.
07:07:22Unhelpfuli suspect a jpeg reader in core will mean that people who use it for album art will likely copy their AA at whatever size they use on their PC, so that such a code path would be unused the vast majority of the time
07:12:18 Join webguest08 [0] (n=185129de@gateway/web/cgi-irc/
07:13:13 Join AndyIL [0] (i=AndyI@
07:14:27Unhelpfulanyway, it's probably not a big deal to write, or a big deal to remove later if it's deemed too useless. just thought i'd put the question out in case anybody wants to weigh in on it while i'm doing other work on this...
07:14:56 Join l403 [0] (n=l@
07:17:44 Join jeffdameth [0] (
07:17:46 Quit webguest08 ("CGI:IRC (Ping timeout)")
07:18:07 Quit jeffdameth1 (Read error: 60 (Operation timed out))
07:23:59 Quit CaptainKwel (Remote closed the connection)
07:24:46soapIf a jpeg reader was in core, all my album art would be scaled to the smallest dimension of my player's screen - thus all could be displayed 1:1...
07:24:47 Quit AndyI (Connection timed out)
07:26:27Unhelpfulsoap: if that's how you store the art on your player, yes, it could be loaded 1:1, assuming your WPS displays at that size. that's a vote in favor of output without the scaler backend, then?
07:33:34soapNo, not a vote either way. I do not understand the issue enough (binsize, code complexity, likelyhood of use outside my example) to feel I should have a vote.
07:34:35linuxstbUnhelpful: Unless it's a lot of extra code, I think it's worth optimising the 1:1 case - I can imagine some people will want to use a PC-side scaler. We may also want to support jpeg for other things, like backdrops, where the size is always the same (more so than AA).
07:34:44soapJust bringing up an anecdote which applies.
07:35:08linuxstbUnhelpful: Although what is the downside of not optimising? Simply slightly slower loading?
07:35:18*linuxstb thinks perhaps he hasn't woken up yet...
07:35:37soapHow does/will the jpeg scaling compare speed-wise to the current bmp scaling? Faster?
07:35:59Unhelpfullinuxstb: more memory needed on load, since the scaler needs its buffers, and some per-pixel multiplies while the linear scaler puts pixels through its math and returns the very same pixels :)
07:36:29soapcurrently I scale my bitmaps, because I don't want to take the CPU hit on my non-gigabeats, but if jpeg was faster perhaps I wouldn't...
07:37:04 Join claydoh [0] (
07:37:16linuxstbsoap: Does that make any practical difference?
07:38:21 Join daurnimator [0] (n=daurnima@unaffiliated/daurnimator)
07:38:28linuxstbUnhelpful: How big are the scaler's buffers? Where are you planning to put them?
07:38:52Unhelpfulsoap: to hit the exact output size requested, it'll use the same scaler that bmp loads use now. *however*, it is possible with jpeg to decode each block partially and directly produce a scaled version of it. this allows scaling to any N/8, with better (because they're basically lowpass-filtered) and faster results than doing the same scale via the bitmap scaler.
07:39:39UnhelpfulN/8 will not always get you to exactly the size you want, though, so we'll sometimes have to feed the output from the decode stage to the scaler anyway.
07:41:28soapI didn't follow the question, linuxstb.
07:42:00linuxstbsoap: I mean do you notice the CPU usage - either by it introducing a delay, or reduced battery life?
07:42:10Unhelpfullinuxstb: 36B * output width on color targets, 12B * output width on greyscale. the caller is expected to call the bmp loader with a buffer large enough to load the bitmap *and* provide space for the scaler's buffers, and i'll be treating jpeg pretty much the same way, the jpeg decoder's data (which can easily exceed reasonable on-stack sizes), as well as scaler buffers if the scaler is called, will come out of the load buffer.
07:43:05Unhelpfulif we want to load jpeg backdrops in core, i would suggest stealing the plugin buffer as a load target, then copying the backdrop out into a screen-sized buffer (i'm assuming we *have* a static one already for the backdrop?).
07:43:21linuxstbSo 36*320 for the current largest case - 11520 bytes.
07:43:50soaplinuxstb: delay is noticeable on large bmps.
07:44:24soaphaven't battery benched in a long time, but scaling in bulk from the pc is easy, so I do.
07:44:27linuxstbsoap: What would you call "large" ?
07:44:47soapmy 250 dpi scans.
07:45:36*linuxstb hopes that isn't a scan of 12" vinyl album covers...
07:48:35Unhelpfulthe album art loading just passes a big buffer (i believe whatever space is available) to read_bmp_file. there doesn't ever seem to be trouble getting the scaler buffers from this space. really, the jpeg reader is going to need a pretty big chunk of buffer *anyway*, there's a large-ish struct that data about the file needs to be stored in, and the scaler only handles one line at a time, while the jpeg decoder may need to decode as many
07:48:35Unhelpfulas 16 lines (in the case of vertical chroma subsampling)
07:51:28linuxstbHow far have you got with implementing this?
07:54:53Unhelpfulin my editor? not very. in my head? a bit farther, i'm pretty sure i know *how* i'll do most of it.
07:55:34 Join webguest67 [0] (n=185129de@gateway/web/cgi-irc/
07:55:37 Quit webguest67 (Client Quit)
07:55:40 Join webguest32 [0] (n=185129de@gateway/web/cgi-irc/
07:56:00 Quit webguest32 (Client Quit)
07:56:03 Join webguest57 [0] (n=185129de@gateway/web/cgi-irc/
07:56:17 Quit webguest57 (Client Quit)
07:59:47 Quit FlynDice (Remote closed the connection)
08:02:52 Join FlynDice [0] (
08:03:21 Join JayHaru [0] (n=jayharu@
08:06:12Unhelpfula pretty large part of the work is moving local data used by the decoder into a structure, because the scaler expects to read data in via a callback. the other big part is converting the decoder from fetching data directly from memory to reading it from the file as needed - the jpeg viewer plugin reads the entire file into memory before calling the decoder.
08:22:24 Join n1s [0] (n=n1s@rockbox/developer/n1s)
08:26:00 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
08:31:48 Join bertrik [0] (
08:31:51***Saving seen data "./dancer.seen"
08:32:20 Quit linuxstb (Read error: 113 (No route to host))
08:34:20 Quit ntfn (Read error: 60 (Operation timed out))
08:35:11 Quit bmbl (Client Quit)
08:38:12Unhelpfulhow much buffering is done on file reads? in particular, is reading only a few bytes at a time going to have a huge overhead (beyond the function call overhead)?
08:39:03 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
08:46:58 Join Rob2223 [0] (
08:48:30 Join J-23 [0] (
08:50:25amiconnUnhelpful: What I'm missing most in the jpeg viewer are two things: Scaling large jpegs (those where even 1/8 is larger than the screen) down to screen size, and a proper fit-to-screen scaling step for the smaller ones
08:51:58amiconn(i.e. scale in the loader to the lowest N/8 larger than the screen and then scale down exactly to screen size, instead of loading the highest N/8 smaller than the screen, which is sometimes rather tiny)
08:52:01Unhelpfulamiconn: my first steps are getting the reader in core, using an interface similar to the one used by the bmp reader, and handling the bits needed to make it a scaler frontend.
08:52:48amiconnJpeg in core is not for all targets though
08:54:03Unhelpfuli understand that it can't be... for targets where it's not reasonable to put it in core, the existing loader can be used, or the "core" loader can be made available via pluginlib, as the scaler is already on mono targets.
08:54:39 Quit parafin (Read error: 110 (Connection timed out))
09:00:44 Join flydutch [0] (
09:04:34 Quit bertrik (
09:04:34 Quit yosafbridge (
09:04:37 Quit Rob2222 (Read error: 110 (Connection timed out))
09:08:44NJoinyosafbridge [0] (
09:10:17NJoinbertrik [0] (
09:11:05 Quit MT (SendQ exceeded)
09:13:18 Join MT [0] (n=chatzill@
09:21:07 Join fragilematter [0] (n=fragilem@
09:22:24 Join stoffel [0] (
09:29:14amiconnkkurbjun: If you'd restore the <= 0 checks like they were in r20674, you could also go back to do{...}while() loops
09:29:35amiconnAlso, some pointer maths in the current driver are weird...
09:30:13kkurbjunyeah, I was wondering that, would it be faster to do a do while versus include a different check in the beginning
09:30:21kkurbjunwhat math are you referring to?
09:31:01amiconnE.g. dst=dst+width+(LCD_NATIVE_WIDTH-x-width)+LCD_FUDGE; can be shortened to dst += LCD_NATIVE_WIDTH - x + LCD_FUDGE;
09:31:12amiconn+width-width is just nothing...
09:32:51kkurbjunin the palette blit?
09:33:35kkurbjunyeah, I missed changing that - that code isn't currently used since we are building in landscape mode
09:34:36 Join ender` [0] (
09:35:26 Quit Tomers (Read error: 113 (No route to host))
09:35:27amiconnA do{...}while loop is a bit faster and smaller than a while{...} loop. The latter checks the loop condition one time more often. In machine code, the first check either needs to be separate, or there needs to be a branch to the condition check at the end of the loop. The latter does at least need the branch in addition to what a do{...}while loop needs, the former is a little faster but needs several extra instructions
09:38:10amiconnSo in this case, you'd trade an extra if() (two instead of the one with combining width and height) for the smaller loops.
09:38:25tmztwhich method does optimation usually use?
09:38:52amiconnProbably doesn't matter much on the m:robe remote though
09:39:18kkurbjunwell, these are on the main lcd
09:39:55kkurbjunI'll change it to a do while with the check in the beginning changed
09:40:09amiconnAH, yes, I confused the lcds
09:40:32kkurbjunthanks for the feedback
09:40:41amiconnOn a related matter, am I reading right that the remote lcd_update_rect() now takes x and width into account, but not yet y and height?
09:41:14amiconn(i.e. the update rectangle is always 16px high)
09:41:55kkurbjuncorrect, since it's vertically packed, there are only two effective Y heights, I just havn't written the code to do a half update only since in all of the cases that I use the remote LCD it has to do a full update anyway
09:42:37kkurbjunit would only need the half update if an 8 px high font is used which I think is becoming rarer
09:43:04amiconnI think an 8px font would be a good idea on the mrobe remote
09:43:33amiconnThat will need basic multifont support of course - I guess 8px fonts are next to unreadable on the main lcd...
09:43:48kkurbjunI agree, but until multifont it's not really that practical on the m:robe especially if you are doing 640x480 builds
09:44:07kkurbjunat 320x240 they aren't too bad
09:44:38amiconnUpdating only half height will save half of the transfers, and since urat isn't very fast... what's the transfer speed?
09:45:43kkurbjun19200, but it's interupt driven and the hardware has a 32 byte fifo
09:45:46 Quit bmbl ("Woah!")
09:45:50kkurbjunso not too much cpu time is spent on it
09:46:36kkurbjunright now it just sets up the interupt handler with a large buffer and that takes care of filling the fifo when it runs out
09:47:02 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
09:47:30kkurbjunthe transmit handling leaves room for improvement, it's not a circular buffer and blocks if you have an existing transfer in queue, but since the transmits happen once every 10 ms it shouldn't be blocking
09:47:54kkurbjun79 bytes plus some header information is the largest transfer per 10 ms
09:49:02amiconnIf my calculation is correct, a full screen update takes ~90ms
09:49:31kkurbjunon the LCD, but the CPU doesn't spend nearly that amount of time
09:49:36kkurbjunremote LCD
09:50:07amiconnIt limits the speed of animated content though (like scrolling text)
09:50:09kkurbjunthe pause states are necessary because if you send the LCD update command too often it doesn't work
09:51:30amiconnI guess if you set the remote scroll speed higher than 9, it'll get jumpy
09:53:21kkurbjunYeah, probably at some point - I haven't tried anything but the default setting.
09:56:00amiconnDefault is 9.
09:56:59 Quit stoffel (Read error: 113 (No route to host))
09:57:41kkurbjunI guess it is a bit choppy looking at it - I need to try the non-immediate update mode to see if I can reduce the pause states
09:58:02 Join gregzx [0] (
09:58:42amiconnI guess there's not much you can do about it. The half-height update would help only in case of 8px fonts (and smaller, but then not for all lines)
09:58:43kkurbjunThere's a microcontroller in the remote - maybe that code can be improved :-D
09:59:23amiconnThe scroll speed setting should probably be limited for slow displays like the mrobe remote
10:00:42 Join bimbel [0] (n=Miranda@unaffiliated/bmbl)
10:01:07kkurbjunamiconn, I was looking at the hardware for the m:robe and there's a idct accelerator that I was thinking about messing with
10:01:52 Join stoffel [0] (
10:02:35kkurbjunbut I'm not too sure on how the intermediate buffers would be used that it references.. does that sound familiar with anything you've worked on. There's an input pointer, an intermediate pointer, and an output pointer that it takes
10:02:47kkurbjunwell, and a secondary output pointer
10:03:13Unhelpfulamiconn: it also looks like the vertical and horizontal IDCTs as we currently use them are pretty easily separable. splitting them would let us do 8x8, 8x4, or 4x8 IDCT so that we can handle subsampled chroma when downscaling, without a chroma scaling step. we could do the same when not downscaling, if we add a 16-point IDCT - those shouldn't be hard to copy from the jpeg-7 sources, it looks like our existing IDCT are mostly borrowed fr
10:03:14Unhelpfulom there.
10:04:46 Quit bertrik (
10:06:25kkurbjunUnhelpful: I don't know how much this would effect your work, but the hardware IDCT function that the m:robe has access to only supports 8x8 blocks
10:06:49kkurbjunIt would be nice ot be able to integrate it with any jpeg/mpeg code we have
10:09:35Unhelpfulkkurbjun: we could certainly special-case the 8x8 case. whether it would pay off to use that, or a reduced or expanded software IDCT, when scaling would be a bit harder to say.
10:10:06kkurbjunaccording to the datasheet it can do the idct in 144 CPU cycles
10:10:46 Quit bmbl (Read error: 110 (Connection timed out))
10:15:12amiconnUnhelpful: Afaik there are 4 cases of chroma subsampling in jpeg - no subsampling (1x1), and 1x2, 2x1 and 2x2 pixels
10:15:31NJoinbertrik [0] (
10:16:27amiconnThat means if we want to handle all possible sizes without having to scale in the yuv blit, we'd need 16x16, 16x8, and 8x16 in addition to the existing 8x8 idct for the unscaled jpeg
10:16:50amiconnBut there's actually a disadvantage in doing this - the decoded data becomes larger
10:17:55amiconnIt's probably not an issue for a core jpeg loader that passes its output to the fractional scaler. For the jpeg viewer it is.
10:19:51amiconnkkurbjun: I know the fascination of new stuff vs. fixing/improving existing things...
10:20:14*amiconn really needs to do those mmc driver fixes/ improvements
10:20:24Unhelpfula core jpeg loader, if it will use the scaler, has to convert all image data to RGB before passing it, anyway. my plan was that we reserve enough space for one block worth of ints, which will hold the intermediate values between the two passes, and enough space for one row of blocks (or two if there is vertical subsampling).
10:21:06kkurbjunamiconn, fortunately, to my knowledge, only 3 people are using the M:robe 500 port :-D
10:22:21Unhelpfulthe IDCT would be split into separate vertical and horizontal IDCT functions - they have different ranges for output values anyway, so we can't use one and fiddle with step sizes unless we have a third function that will copy from a working space and scale the values down to their output ranges.
10:23:04Unhelpfulwe could *do* that, but i didn't think that writing that data back to a block of memory before reading it back in and scaling it down would be a good idea (for performance)
10:24:37amiconnkkurbjun: In case you're talking about the idct in libmpeg (mpegplayer) - armv5 and armv6 should allow for a faster idct on cpu anyway. Our current idct is optimized for armv4 and its slow multiplier
10:25:23kkurbjunamiconn, faster than the hardware idct implementation?
10:25:26amiconnIdct for armv5 and higher could work very similar to the coldfire version
10:25:32amiconnI don't know yet
10:26:07kkurbjunoh, I missed your meaning
10:26:36kkurbjunI thought you were saying that it would be faster on the CPU, not that there is room for improvement on v5/6
10:26:46kkurbjunI see what you mean now though
10:31:01Unhelpfulamiconn: an 8x8 IDCT is *quite* a few multiplies, isn't it? if i count correctly, the one in our jpeg decoder is doing 192 multiplies per block, and when you factor in the other math it needs to do, i'm guessing that we'd to better with that hardware IDCT...
10:31:56***Saving seen data "./dancer.seen"
10:32:12 Join gregzx_ [0] (
10:32:53 Quit gregzx (Nick collision from services.)
10:32:55 Nick gregzx_ is now known as gregzx (
10:33:09 Quit l403 (Read error: 148 (No route to host))
10:34:04amiconnYeah, probably. This doesn't mean the idct shouldn't be optimised for armv5+. Not all armv5+ targets have hardware idct support...
10:38:59 Join l403 [0] (n=l@
10:43:06Unhelpfulhrm... i haven't taken a close look at the yuv2rgb part yet, but i *believe* a fixed-size, fixed-ratio version of the scaler could be done for quite cheap. all of the division, and much of the multiplication, would be by fixed values, and convenient ones at that.
10:45:19 Join planetbeing [0] (
10:54:48 Quit MT (SendQ exceeded)
10:56:19 Join MT [0] (n=chatzill@
11:03:06 Join evilnick [0] (
11:03:30 Quit kachna (
11:03:30 Quit tchan (
11:03:30 Quit Ridayah (
11:03:30 Quit kubzior (
11:03:30 Quit synergist (
11:03:30 Quit evilnick1 (
11:03:30 Quit Foxx (
11:03:30 Quit shadearg (
11:03:56 Quit planetbeing (Connection reset by peer)
11:04:04NJoinkachna [0] (
11:04:04NJointchan [0] (n=tchan@lunar-linux/developer/tchan)
11:04:04NJoinRidayah [0] (
11:04:04NJoinkubzior [0] (n=kubz@unaffiliated/kubz)
11:04:04NJoinsynergist [0] (
11:04:04NJoinevilnick1 [0] (
11:04:04NJoinFoxx [0] (n=Foxx@
11:04:04NJoinshadearg [0] (
11:04:18 Join planetbeing [0] (
11:04:25LloreanWhat causes the inaccuracies in seeking in very large MP3 files?
11:06:22 Quit Llorean ("Leaving.")
11:08:50 Join stripwax [0] (
11:09:07 Join Llorean [0] (n=DarkkOne@
11:11:01 Quit stripwax (Client Quit)
11:11:46LloreanSeeking seems to be _very_ broken actually
11:14:32LloreanIf I seek one second forward, it hops me to a time ~2:30 forward (in a 21 hour file). This isn't irregular, this is the inaccuracy that used to be there (you could get around it by just figuring out roughly how much the offset was that late in the file, and approximating an earlier point to seek to) but now it seems to be playing someplace from ~5.5 hours in the file, when the point I seeked to was ~17.5 hours into the file
11:17:30amiconncbr or vbr? with or without xing header if the latter?
11:18:52amiconnCBR seeking should be exact.
11:19:00LloreanHow would I check? It's encoded fairly recently, so I would assume it's whichever is "better" unless there's no real standard for it.
11:19:07amiconnVBR amkes use of the xing header if present, which has 100 seek points
11:19:10LloreanBeing off by 12 hours is hardly simply "inexact"
11:19:42amiconnSo the error should not exceed 12 minutes (worst case) for your file
11:20:09Lloreanthe progress bar doesn't update to the proper time either. it says I'm at approximately 17.5 hours into the file, when it's playing audio from 5.5 hours in. if I stop playback and resume, though, it's fine.
11:20:12amiconnWithout xing header, the error can be huge if bitrate is very unbalanced across the file
11:20:16Llorean"fine" being, the progress bar updates to the 5.5 hour mark
11:20:34 Join Thundercloud [0] (
11:21:23 Quit evilnick1 (Read error: 110 (Connection timed out))
11:22:02amiconnIs it just the progress bar or also the current time? Maybe there's an error in progressbar position calculation
11:22:10LloreanCurrent time is off too
11:22:49LloreanIf I stop and resume, both are correct. Until I do, it's playing audio from 5.5 hours in, and displaying 17.5 hours in (since I simply seeked one second forward from my bookmark at about that point)
11:22:50amiconnHmm. This used to work?
11:23:30LloreanI'm not certain. I've never experienced something like this before, though I was using a half-month old build when I just ran into it, and it's still in current SVN.
11:23:32*amiconn wonders whether this has to do with the recent change in current_id3 handling in playback.c
11:23:47LloreanI listen to audiobooks pretty regularly, so I'd *think* it's something I'd have encountered before, but I can't say for certain.
11:24:18amiconnI don't have such a long file. For the longest mp3 file I have, seeking is working fine (that file is a little longer than an hour and is also cbr iirc)
11:24:49 Join DerPapst [0] (
11:27:25 Quit fragilematter ("leaving")
11:28:32LloreanAfter running VBRfix on the file, I get the same issue. I assume this means that it's an issue that occurs even with the proper headers.
11:28:47 Quit planetbeing ()
11:28:51*Llorean assumed the file had them anyway, but didn't know how to test.
11:30:51 Join planetbeing [0] (
11:36:12 Join jaykay [0] (
11:44:33 Quit stoffel (Read error: 113 (No route to host))
11:55:44 Nick fxb__ is now known as fxb (
12:07:38 Quit MT (SendQ exceeded)
12:10:11 Join MT [0] (n=chatzill@
12:15:45 Quit FOAD ("I'll be back")
12:16:01 Join FOAD [0] (
12:19:15 Quit planetbeing ()
12:21:06 Quit Thundercloud (Remote closed the connection)
12:25:42 Join stripwax [0] (
12:26:03 Quit JayHaru (Read error: 110 (Connection timed out))
12:29:50Lloreanamiconn: One thing that seems odd to me. Bookmarks work pretty much fine, but seeking doesn't.
12:31:59***Saving seen data "./dancer.seen"
12:36:45 Join stripwax5443 [0] (
12:36:49 Quit stripwax5443 (Client Quit)
12:48:22pixelmaLlorean: maybe it has to do with the changes in r20633 (and following), it caused some bugs which were ironed out since then but there could be more. Daily builds from before should still be available (April 4th), perhaps a quick way to test?
12:49:34 Join Sedgewick [0] (
12:49:40Lloreanpixelma: I'm _pretty_ sure my previous build was from before that and still experiencing the problem. But not absolutely. I'll see if I can track down a revision when I can later.
12:58:19 Join parafin [0] (
13:01:21*GodEater wonders why the hell his PC is refusing to mount his sansa
13:01:47GodEater....and of course it bloody does as soon as I mention it
13:02:04 Join stoffel [0] (
13:07:37*pixelma wonders where the interest in FS #8824 (c200 keymap) went
13:09:33rasherpixelma: not enough devs with c200s around?
13:10:25pixelmaI put the patch there about a week ago (and people showed interest in it before, here in IRC) :)
13:12:20GodEateris usb still not enabled in the release c200 bootloaders ?
13:13:38pixelmawouldn't make much sense if it isn't even enabled in the last release build... :)
13:14:12rasherUSB isn't enabled in any release bootloaders
13:15:14GodEateryou know, I think I've had brain failure this morning
13:15:24GodEaterI'd best not touch anything complicated - I might break it
13:16:32 Join schrottplatz [0] (
13:16:44 Join tvelocity [0] (
13:17:15schrottplatzwhen comes the pictureflow out?
13:17:20schrottplatzare there any plans?
13:17:40stripwaxschrottplatz - pictureflow is already part of rockbox, and I don't understand your question
13:17:54schrottplatzbut i cant play the songs..
13:18:00schrottplatzout of pictureflow
13:18:24stripwaxpictureflow does not yet let you select the songs to play, that is true. it only shows the albumart pictures..
13:18:33LloreanIt's in the demo category for this very reason.
13:18:55 Join miepchen^schlaf [0] (
13:19:09schrottplatzand when will it be released?
13:19:41LloreanThere's nothing to "release"
13:19:44LloreanEverything we have is there.
13:20:10schrottplatzwill it ever work?
13:20:15stripwaxIf you mean 'when will it be possible to select the tracks directly from pictureflow itself' - I don't think there are any immediate plans
13:20:21stripwaxpictureflow works. it shows pictures.
13:20:47Lloreanschrottplatz: Rockbox is entirely volunteer effort. There are no timelines or plans, things only happen if interested people do the work.
13:21:01schrottplatzi know
13:21:10stripwaxschrottplatz - if you're interested, you can take a look at the code and help extend it to enable song selection from pictureflow?
13:21:35schrottplatzhmm i can only code php...
13:21:53schrottplatzi could make a better design
13:22:36schrottplatzbut it seems like you are happy with it :/
13:23:30 Join einhirn [0] (
13:25:33GodEatera better design for pictureflow ?
13:26:40 Join wincent [0] (
13:29:19stripwaxI think he means a better rockbox web design?
13:29:37 Join fdinel [0] (
13:30:06GodEaterif he does, he's not been paying attention to the "redesign the www" thread in the forums
13:33:01 Quit wincent (Read error: 60 (Operation timed out))
13:33:30 Join cmwslw [0] (
13:38:42 Join einhirn_ [0] (
13:38:42schrottplatzGodEater: yes the webdesign
13:38:52 Part cmwslw ("Ex-Chat")
13:38:54schrottplatzwhich forum abi?
13:39:13LloreanThe Rockbox forum.
13:44:22stripwaxHm, running test_codec a couple times seems to give sizeable variation in timings (e.g. +- 1.5MHz ish on coldfire h120, which seems to be of the same approx order of magnitude as the effect I'm trying to test for). Is there another way to measure codec performance other than watching the buffering debug screen?
13:45:20LloreanIs there a possibility there's a bug in test_codec causing that?
13:45:52schrottplatzLlorean: are there already subs?
13:46:07Lloreanschrottplatz: "subs"?
13:46:51Lloreanschrottplatz: There's a forum thread with ideas and mock-ups.
13:47:18schrottplatzi ve an idea
13:47:52schrottplatzi made already something similar
13:47:58stripwaxLlorean - it's possible, I guess... do you know of any bug in test_codec?
13:48:25Lloreanschrottplatz: If you wish to discuss it, the best place is the thread in the forum.
13:49:09Lloreanstripwax: No, but I thought the purpose of it was to remove as many external factors - to be as close as deterministic as possible. If it's giving different results on repeats of the same input, it sounds to me like it's either not doing its job, or it's just not very good at it.
13:50:18Lloreanschrottplatz: This kind of thing needs to be discussed somewhere that discussion can be asynchronous, so that many can view, comment, and reply over a period of time. IRC is not really suitable.
13:50:54schrottplatztell me what do you think of it, pls
13:51:15rasherIt's not exactly a website design..
13:51:53schrottplatzyes not yet
13:52:01schrottplatzbut the style...
13:52:27LloreanIt's rather hard to infer from that image what you'd actually make website navigation look and function like.
13:53:04schrottplatz <- this is aaaaaaawwwwwwesome
13:53:38LloreanYes, it's from the thread we referenced you to.
13:53:51 Quit Sedgewick ("off")
13:54:49schrottplatzbut the yellow at the to should be a bit brighter
13:55:09scorche[04:48:21] <Llorean> schrottplatz: If you wish to discuss it, the best place is the thread in the forum.
13:55:11stripwaxschrottplatz - it would be very helpful for everyone involved if you added your comments to the thread in question, rather than here
13:58:14stripwaxschrottplatz - that way, the people involve can find your comments. if they're just in an irc log somewhere, they will never read what you have to say ..
13:58:26 Quit einhirn (Read error: 110 (Connection timed out))
14:00:33 Quit stoffel (Read error: 113 (No route to host))
14:01:19 Join moos [0] (i=mustapha@rockbox/staff/moos)
14:04:06 Join stoffel [0] (
14:05:35schrottplatzis there a good tool to sync the newest podcast with my sansa? im using rhythmbox...
14:06:16scorchethis is #rockbox...not #rhythmbox...
14:10:47 Quit flydutch ("/* empty */")
14:17:24 Quit kps00000 (Read error: 110 (Connection timed out))
14:18:11 Join PaulJam [0] (
14:32:00***Saving seen data "./dancer.seen"
14:34:18 Join __lifeless [0] (n=lifeless@
14:39:42 Join midijunkie [0] (
14:56:01 Join hittudiv [0] (n=hittudiv@
14:56:10 Quit n1s ("Lämnar")
14:57:01 Quit SirFunk (Read error: 110 (Connection timed out))
14:57:56 Quit stoffel (Read error: 113 (No route to host))
15:01:00 Join PaulJam_ [0] (
15:01:09 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
15:02:29 Quit PaulJam_ (Read error: 113 (No route to host))
15:03:37 Join PaulJam_ [0] (
15:04:20 Join SirFunk [0] (n=Sir@
15:08:18 Join Horscht86 [0] (
15:11:03 Quit antil33t (Read error: 54 (Connection reset by peer))
15:16:38 Join antil33t [0] (
15:17:19 Quit MT (SendQ exceeded)
15:17:24 Join SirFunk_ [0] (
15:18:21 Quit PaulJam (Read error: 113 (No route to host))
15:21:26 Join MT [0] (n=chatzill@
15:21:49 Join Seed [0] (
15:25:37 Quit Horschti (Read error: 110 (Connection timed out))
15:26:49 Join funman [0] (i=53c2ba56@rockbox/developer/funman)
15:29:44funmanFlynDice: yesterday I tried to write from the lcd framebuffer using a buffered only memory area
15:30:02 Join robin0800 [0] (
15:30:27funmanit didn't show any difference, I think the other devices use a framebuffer for DMA, separated from rockbox' "lcd_framebuffer"
15:31:52 Quit SirFunk (Connection timed out)
15:42:51 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
15:45:50funmani'm getting some documents on USB to start having a look at Sansa AMS.
15:46:09funmanDo you have some advices ? (I have USB spec 2.0 and the OnTheGo Supplement)
15:47:02mcuelenaerefunman: regarding implementing a Rockbox USB driver: look at usb-drv-arc.c for an example
15:47:48mcuelenaere(usb-drv-tcc.c & usb-jz4740.c are other USB drivers)
15:48:17 Nick Horscht86 is now known as Horscht (
15:50:11funmani had a quick look but i need some vocabulary first ("endpoint" for example)
15:50:47mcuelenaerehave you seen ?
15:50:54mcuelenaereit's quite good
15:54:25funmanthanks, that looks clear!
15:54:46*mcuelenaere wonders whether that link shouldn't be on some wiki page
16:08:17 Join aapo [0] (
16:08:37 Quit shadearg ("goodbye.")
16:08:57 Quit hittudiv (Read error: 110 (Connection timed out))
16:09:57aapohi, I tryied preparing to compile rockbpx on Linux. I started sudo ./ and chose all targets, it doesn't get end, it stucked on " ar: regex.o: No such file or directory"
16:10:13 Quit BlakeJohnson86 (Remote closed the connection)
16:10:16aapomaking "/tmp/rbdev-build/build-gcc-m/libiberty"
16:11:53aapocd /tmp/rbdev-build/build-gcc-m/libiberty & sudo make -> I got regex.o
16:12:41 Join shadearg [0] (
16:12:42rasheraapo: making all targets is currently rather broken, do them one at a time
16:12:52gevaertsjust a wild guess, but do you have things like $MAKEFLAGS or other bits that may influence building?
16:13:10aapostarting sudo ./ again leads: ROCKBOXDEV: mkdir build-binu-m, can't make, directory exists
16:13:30rasheraapo: yeah, you need to clean out the dirs it created.. but build-binu-m ??
16:13:59 Join BlakeJohnson86 [0] (
16:14:09aapoI just follow the LinuxSimpleGuideToCompiling
16:14:34 Quit funman (" ajax IRC Client")
16:15:27rasheraapo: Clean out the dirs, start again and type only a single letter at the prompt
16:16:02aaporasher, ok, what dirs I clean? /tmp/rbdev-build?
16:16:16rasherAnything in /tmp starting with rbdev
16:16:43aaporasher, ok I will try
16:26:18 Quit robin0800 ("No Ping reply in 90 seconds.")
16:26:48 Join robin0800 [0] (
16:27:17 Join kps00000 [0] (
16:32:03***Saving seen data "./dancer.seen"
16:32:03 Quit linuxstb (Read error: 110 (Connection timed out))
16:38:05 Quit faemir (Read error: 104 (Connection reset by peer))
16:38:44 Join aum [0] (
16:39:25 Join faemir [0] (
16:40:01 Join FOAD_ [0] (
16:40:04 Part aum
16:55:28 Quit FOAD (Read error: 110 (Connection timed out))
16:55:28 Nick FOAD_ is now known as FOAD (
16:56:07 Quit robin0800 (Read error: 60 (Operation timed out))
17:03:01 Join FOAD_ [0] (
17:03:48 Join robin0800 [0] (
17:06:53 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
17:11:01 Quit parafin ("So long and thanks for all the fish")
17:11:06 Join parafin [0] (
17:18:54 Quit FOAD (Read error: 110 (Connection timed out))
17:18:54 Nick FOAD_ is now known as FOAD (
17:28:50 Join sdubois92 [0] (
17:29:14sdubois92When i start my ipod, it asks if i want to build the db. But i don't want to use the db, can i turn this message off?
17:30:04stripwaxsdubois92 - you can delete the db
17:31:21sdubois92stripwax: i did
17:33:16 Quit robin0800 (Remote closed the connection)
17:48:41 Join toffe82 [0] (
17:51:09 Quit l403 (Read error: 148 (No route to host))
17:52:55 Join PaulJam__ [0] (
17:57:45PaulJam__sdubois92: could it be that your start screen setting is set to database?
18:08:12 Join PaulJam [0] (
18:08:21 Join jason [0] (
18:08:42 Quit gevaerts (Nick collision from services.)
18:09:10 Join gevaerts [0] (n=fg@
18:10:53 Join tomers [0] (
18:12:27 Quit PaulJam_ (Read error: 110 (Connection timed out))
18:13:04 Quit PaulJam__ (Read error: 60 (Operation timed out))
18:28:44tomersI created "FS #10116 - USB HID class driver" and I would like your help committing this
18:29:25tomersIt can be separated into small change sets, which will make it easier to review and be understood
18:30:10gevaertsI've just been looking at FS #10142 and it got me wondering. What
18:30:22gevaertsWhat's the license of the linux ch9.h?
18:30:49gevaertsThat file doesn't have any license or contact information in it at all...
18:32:08***Saving seen data "./dancer.seen"
18:32:12tomersThat's really weird, indeed. But its a bit too late as its already in Rockbox's code.
18:32:38gevaertsI know, this isn't a new problem (if it is one at all)
18:32:56tomersI can assume it isn't any proprietary license, as it wouldn't be included there in the first place
18:33:25tomersHow do you suggest we work to commit my changes?
18:33:52gevaertsno, but without information I assume it's GPLv2 strict, and we should at least mark that in our copy
18:33:53 Quit bimbel ("Woah!")
18:34:14tomersWe should ask in the Linux mailing-list
18:35:25 Join PaulJam_ [0] (
18:35:32tomersOther files in /usr/src/linux-headers-2.6.27-11/include/linux/usb/ seem to have GPL v2 license
18:35:37gevaertsAnyway that isn't a blocker for FS #10142. Are there any real changes in that, or just some new values?
18:36:32 Quit schrottplatz ("o.O")
18:36:49gevaertsAlso, I'm not actually sure about the tabs in there. We do usually leave files with external origin in their original style
18:36:54tomersNo real changes, but do the diff yourself to verify. There are some parts that weren't integrated. I guessed they were not included earlier, either since they were not required, or they did not existed
18:37:29gevaertsThere seem to be some more comments as well
18:37:36 Nick fxb is now known as fxb__ (
18:37:54 Join robin0800 [0] (
18:37:58tomersI learned that Rockbox does not modify the original code styling, but I think we does change tabs to spaces, am I right?
18:38:27rashergevaerts: I think that's only the rule if there's any chance we're going to keep them in sync with upstream
18:38:56gevaertsrasher: well, tomers is doing just that right now :)
18:39:10 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
18:39:27rasherWell in that case yes, we should change as little as possible I believe
18:41:04tomersSo what is the verdict? keeping the original tabs?
18:41:27rasherI'd say so
18:41:30mcuelenaereWhat YUV format does Rockbox use?
18:42:12kkurbjunfor blitting?
18:42:24gevaertsOn the other hand, the types are different (uint8_t vs __u8 and things like that), so merging is not straightforward anyway
18:42:33kkurbjun4:2:0 I believe is the name for it
18:43:00kkurbjunmcuelenaere: did you see that the mr500 has yuv blitting in the lcd driver?
18:43:17mcuelenaerenot yet
18:43:17kkurbjunif you were thinking about that for the creative ZVM
18:43:26mcuelenaereno, I was talking about the Onda VX747
18:43:32kkurbjunoh, gotcha
18:43:33gevaertstomers: do you expect that merging from linux later on again is useful? We're not going to see USB3 anytime soon on a rockbox-supported DAP I guess, so we could just use the current file and maintain it ourself
18:43:36mcuelenaere(which also has HW-accelerated YUV)
18:44:23 Join l403 [0] (n=l@
18:44:23gevaertsAnd if we maintain it ourself, we get rid of the tabs...
18:46:36tomersgevaerts: I don't think this file changes often, and keeping it somewhat in sync with Linux, today, can only be good for us. I don't see the pitfalls. Also, I am in favor of converting tabs to spaces
18:47:52tomersgevaerts: It will make the difference between the two system more obvious.
18:48:13gevaertstomers: I'm certainly in favour of syncing now. I think I'll apply the patch as it is now
18:48:40tomersgevaerts: Thanks! I'll work on my next patch now.
18:48:52 Quit bmbl ("Woah!")
18:49:36LambdaCalculus37This is weird... I just build the m68k environment on my machine, but m68k-elf-gcc is missing from /usr/local/m68k-elf/bin
18:49:43*LambdaCalculus37 summons a Mac owner
18:50:20*gevaerts gives LambdaCalculus37 a mirror ;)
18:50:29LambdaCalculus37gevaerts: Another Mac owning dev? ;)
18:51:10gevaertsLambdaCalculus37: m68k mac OK? ;)
18:51:28rasherLambdaCalculus37: preglow reported being unable to build m68k
18:51:32LambdaCalculus37gevaerts: Can you build the dev environment on it? ;)
18:51:50LambdaCalculus37rasher: Same reason? Because of the missing m68k-elf-gcc binary?
18:52:12rasherLambdaCalculus37: I don't think so
18:52:31LambdaCalculus37rasher: All the other needed tools appear to be there.
18:52:35LambdaCalculus37Just missing that one.
18:54:58LambdaCalculus37rasher: IIRC Domonoky also has a Mac, right? As does JdGordon.
18:55:51rasherThat sounds correct
18:56:10rashermidgey as well, it seems
18:56:44LambdaCalculus37And I think lostlogic does, too.
18:56:55LambdaCalculus37Hopefully one of them has the file I need.
18:57:06 Quit robin0800 (Remote closed the connection)
18:57:20LambdaCalculus37I can build for all other platforms except m68k right now, and that's bad since I own a Coldfire target (H340).
18:59:34 Quit stripwax (Read error: 104 (Connection reset by peer))
19:01:19 Quit dmb ("Leaving")
19:02:17 Quit PaulJam (Read error: 113 (No route to host))
19:15:46 Join planetbeing [0] (
19:28:30 Join kenzo [0] (n=5430d510@gateway/web/cgi-irc/
19:29:22kenzoGuys, I really need some help in installing rockbox
19:30:00LambdaCalculus37Well, tell us what kind of help you need.
19:32:06kenzoIm trying to install it on a iPod 5,5 using a Mac. I have no clue whatsoever about compiling and what have you not;) The rockboxUtility does not work and I cant seem to get Mtools helping me either. I get error report that the ipod is still a Mac and not windows.
19:32:24 Quit jeffdameth (Read error: 113 (No route to host))
19:32:44LambdaCalculus37When what you need to do is format the iPod to use FAT32 instead of HFS+.
19:33:18LambdaCalculus37Restoring the iPod on a Windows machine with iTunes is the easiest way to do that.
19:33:35kenzoI really dont have that option
19:33:48kenzoI want to make it on my own machine, you know
19:33:51 Join jeffdameth [0] (
19:34:20kenzocan someone guide me through the mtools procedure?
19:34:28 Join stoffel [0] (
19:34:32LambdaCalculus37How about just follow the instructions here:
19:35:34kenzocant even find the config file), its driving me nuts.
19:37:00 Quit jhMikeS (Nick collision from services.)
19:37:06 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
19:40:37 Quit moos ("Rockbox rules the DAP world")
19:41:00kenzoThe problem for me lies in "Format the main partition with FAT32 file system". Anyone, lease?
19:41:47LambdaCalculus37I just told you what to do.
19:41:59LambdaCalculus37And Rockbox Utility is telling you what to do.
19:42:13LambdaCalculus37You have to format your drive to use FAT32.
19:42:24kenzoRockbox Utility is not working for my iPod!
19:42:36kenzoIts stated clearly in the known issues
19:43:11LambdaCalculus37kenzo: Unless you want me to write it out in giant neon letters, please follow the instructions on the page I linked to earlier.
19:43:27rasherkenzo: you need to format it as fat32. If you don't know how, I suggest seeking out OS support somewhere
19:44:18kenzoI am fully aware that I need FAT32. The problem is that all the automated procedures of getting this job done is not working.
19:44:36kenzoNot the utility function, and the mtools is way over my head
19:44:47kenzowas just trying tog get some kind help
19:44:55LambdaCalculus37Then perhaps you should turn to Google. We didn't write mtools, and therefore we don't support it.
19:45:06LambdaCalculus37Ask the mtools authors for assistance.
19:45:49kenzoOr perhaps maybe the team could write a tutorial that actually works...
19:45:54rasherBagder: no icons in the menu - what gives?
19:45:59markunkenzo: perhaps
19:46:01rasherkenzo: many people have used it
19:46:05LambdaCalculus37The tutorial works just fine.
19:46:35Bagderrasher: and no themes...
19:46:44 Join flydutch [0] (
19:46:51rasherooh indeed
19:46:55markunkenzo: if you eventually find a good way to do it, maybe you can help to make the tutorial better.
19:47:16rasherBagder: did you see">
19:47:51Bagderoh nice
19:48:10*rasher notices that he hadn't removed the : from m:robe
19:48:13kenzoThis is not working: RockboxUtility....and this: For 5.5G iPods These iPods have their storage exported as 2048-byte sectors. Simply add an argument to the 'newfs_msdos' command:.... and on top of it you have to be a nerd to learn another program in order to install Rockbox. Great work guys.
19:48:34LambdaCalculus37kenzo: We'll be happy to refund any money you paid for Rockbox.
19:48:43kenzoI just love the attitude you guys are giving
19:49:25rasherkenzo: We *know* RockboxUtility doesn't work with HFS ipods. It's not supposed to. You're supposed to convert it to FAT32. We have a guide for doing that. You have failed to do so.
19:50:02rasherAlternatively, you can convert it using a Windows machine.
19:50:03LambdaCalculus37We've already told you that Rockbox does not work on HFS iPods.
19:50:06kenzoAre you guys totally deaf???????? Blind???? or what? The guide has FLAWS!!!!!!!!!!!!
19:50:09BagderI don't think its our fault your ipod is HFS
19:50:17kenzoI know it doesnt support HFS
19:50:19Bagderand now he's just rude
19:50:24rasherkenzo: No need to get abusive.
19:50:32rasherkenzo: You have yet to explain what your problem is
19:50:49LambdaCalculus37kenzo: I think you're just angry because you can't follow simple instructions.
19:51:20LambdaCalculus37We didn't write mtools, so how can we support it or give advice on it? That's best left to the people who *did* write it.
19:51:32kenzolambda: you are seriously stoned. I have clearly written my problems, yet you insist on refering to the guide which is the basic problem!!!!!!!!!!!!!!!!!!!!!!
19:51:41rasherkenzo: Which step of the IpodConversionToFAT32 page is giving you trouble?
19:51:48Bagderkenzo: please calm down and talk to us
19:51:48kenzo1 min
19:51:54LambdaCalculus37kenzo: I usually am not stoned on Saturdays. ;)
19:52:06DerPapstheh ;)
19:52:18kenzothis line: newfs_msdos -F 32 -S 2048 -v iPod /dev/rdiskNs2
19:52:23kenzocommand line
19:52:34kenzoand the Rockbox Utility
19:52:52rasherStop talking about the utility. It's not relevant until you've converted your ipod to FAT
19:52:57kenzowhich leaves me with the option of learning linux to play FLAC:D hahah
19:53:19LambdaCalculus37And is that a bad thing to learn it? What's wrong with learning how to do things?
19:53:34gevaertsWhy would you need to learn linux at all?
19:53:37rasherkenzo: What did you type for that command? Did you replace N with the number you found in the previous step?
19:53:51kenzoI agree - thats why I have been in fron t of my comp the last 10 hours! Im not kiddin
19:53:57LambdaCalculus37gevaerts: Because we said so? ;)
19:54:03kenzoyes, i replaced it like it says
19:54:14rasherSo what is the *exact command* you ran?
19:54:18kenzohaha, funny LAMDA
19:54:18rasherAnd what was the output?
19:54:55*LambdaCalculus37 wonders why people keep getting his handle wrong
19:55:01kenzowhere can I possibly find that now when I have tried for 10 hours straight?
19:55:16rasherkenzo: well you could try again
19:55:17LambdaCalculus37By pressing up at the command line a couple of times?
19:56:22Bagderif you've tried it for 10 hours, one would think you would have it remembered by now...
19:57:01kenzoor maybe not - Im outta here!
19:57:08DerPapstkenzo: connect your ipod and rerun that command again after making sure N is still the same number
19:57:12rasherStop that.
19:57:34Bagderkenzo: if you can't be nice and polite, you can just as well leave
19:57:34 Quit kenzo ("CGI:IRC (EOF)")
19:58:03DerPapst... silence
19:58:12LambdaCalculus37That was easy.
19:59:22rasherLambdaCalculus37: honestly, I don't think you were helping things..
19:59:52 Join FOAD_ [0] (
20:00:40 Quit FOAD (Read error: 60 (Operation timed out))
20:00:41 Nick FOAD_ is now known as FOAD (
20:01:31*LambdaCalculus37 will remain quiet for a while here
20:02:58 Join barrywardell [0] (n=barrywar@
20:07:45 Join FOAD_ [0] (
20:12:36 Quit stoffel ("leaving")
20:14:43 Quit flydutch (Read error: 110 (Connection timed out))
20:14:55 Join showstopper [0] (n=4db57f33@gateway/web/cgi-irc/
20:19:56 Quit FOAD (Read error: 110 (Connection timed out))
20:19:57 Nick FOAD_ is now known as FOAD (
20:20:23 Quit showstopper ("CGI:IRC (Ping timeout)")
20:22:07 Quit barrywardell ()
20:25:25 Join barrywardell [0] (n=barrywar@
20:32:10***Saving seen data "./dancer.seen"
20:34:30 Quit aapo (Remote closed the connection)
20:35:09 Quit bzed_ ("leaving")
20:35:12mcuelenaeregevaerts: shouldn't FS #10143 rather be a mail on rockbox-dev than a Flyspray item?
20:35:17 Quit bzed ("leaving")
20:35:24 Join bzed [0] (
20:37:31gevaertsmcuelenaere: mails tend to be forgotten about, flyspray tasks not so easily
20:55:48 Join merbanan [0] (
20:57:03 Join robin0800 [0] (
20:58:22 Join pyro_maniac [0] (
20:59:23 Quit pyro_maniac1 (Read error: 110 (Connection timed out))
21:00:09 Quit jhMikeS (Nick collision from services.)
21:00:11 Join _jhMikeS_ [50] (n=jethead7@rockbox/developer/jhMikeS)
21:01:26 Join archivator [0] (n=archivat@
21:01:54MTmerbanan : Hi :), in COOKContext , is the initial value of "random_state" always zero ?
21:02:13merbananMT: doesn't matter
21:02:30merbanancan be anything it is just a random seed
21:03:14 Join PaulJam__ [0] (
21:03:35MTmerbanan : I was just checking to see if the code is working properly (after the isolation of the codec an all that) .. after I call cook_decode_init I always get a random_state of zero
21:04:07 Join EternalRains [0] (
21:05:54 Join PaulJam [0] (
21:07:52 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
21:10:40 Quit PaulJam__ (Read error: 60 (Operation timed out))
21:15:50 Join dmb [0] (n=dmb@unaffiliated/dmb)
21:22:28 Quit PaulJam_ (Read error: 113 (No route to host))
21:31:34 Quit robin0800 (Read error: 110 (Connection timed out))
21:32:28mcuelenaerehmm is YUV data in Rockbox supposed to be displayed rotated?
21:32:51 Join robin0800 [0] (
21:32:57kkurbjunmcuelenaere: it depends on the natural orientation of the player
21:33:14kkurbjunif the screen is portrait format the intention is for it to be rotated
21:38:59 Join bluebrother [0] (
21:39:41 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
21:46:45*MT waves to linuxstb :)
21:54:06gevaertstomers: I'm not entirely sure if usb_class_driver.h is the right place for this PACK_DESCRIPTOR macro. On the other hand, I don't really know a better place either...
21:54:35 Join PaulJam_ [0] (i=PaulJam_@
21:55:06 Join froggyman [0] (n=47ba40e2@gateway/web/cgi-irc/
21:57:46 Quit linuxstb (Read error: 110 (Connection timed out))
21:58:39 Quit dmb (Read error: 104 (Connection reset by peer))
21:59:45 Quit PaulJam (Read error: 60 (Operation timed out))
22:02:47gevaertsOn the other hand, it's just one small macro so moving it later is not a big problem I guess...
22:03:31tomersI also notices it is not the ideal place, but that was the one I can up with eventually
22:03:51gevaertsI'll commit it as-is. If we ever find a better place, we move it
22:06:52mcuelenaerewhat's the YUV test in test_fps.c supposed to show? some grey area?
22:09:05 Join CIA-19 [0] (
22:15:33Unhelpfulso, would it probably be best in the jpeg reader to implement a getc and putc that work from a small-ish buffer? maybe 16B, larger if we want to reduce read frequency a bit?
22:15:48Bagderrasher: about the target names, how about a column that uses "line-model" if there's a line, and "company-model" otherwise?
22:16:14 Quit jaykay ("ChatZilla 0.9.84 [Firefox 3.0.8/2009032609]")
22:16:55tomersgevaerts: Please verify that FS #10146 (USB: Pass to the USB classes a pointer to USB core buffer) compiles before committing. I have no time to checkout a new, clean, tree.
22:16:59Unhelpfuli'd started out with a small buffer that's read into at appropriate places, but that's getting to be painful, even while just processing the initial markers
22:17:32gevaertstomers: I always do that anyway. Let me read through it first though :)
22:17:53rasherBagder: done
22:18:17rasherThat's probably a good compromise between unambigous and short
22:18:47Bagderand reasonably close to what we have today
22:18:55mcuelenaererasher: have you seen FS #10137?
22:19:26gevaertstomers: it misses the actual call to those functions
22:20:09rashermcuelenaere: no, nice
22:20:47rashermcuelenaere: unclear license on the js though :|
22:21:02mcuelenaereheh, I guessed something like that would arise :)
22:21:05mcuelenaereI could rewrite it
22:22:08tomersgevaerts: Are you sure? They are callbacks. The are being called in control_request(). See usb_core_request_endpoint()
22:23:21gevaertstomers: I get
22:23:52*gevaerts wonders if mcuelenaere saw that he added some red earlier today
22:24:13 Join jmillikin [0] (
22:25:08gevaertsr20730 apparently
22:25:19CIA-19mcuelenaere r20734 trunk/firmware/target/mips/ingenic_jz47xx/onda_vx767/lcd-onda_vx767.c: Really fix red
22:25:59 Part sdubois92
22:28:10 Quit barrywardell ()
22:29:06tomersgevaerts: Thanks. I've uploaded an updated patch.
22:31:02 Quit FlynDice (Remote closed the connection)
22:32:03gevaertstomers: still not OK.
22:32:12***Saving seen data "./dancer.seen"
22:33:23tomersgevaerts: Fixed. I must be tired!
22:33:33 Join FlynDice [0] (
22:35:28gevaertstomers: it compiles now. Now I'm trying to understand what it's good for :)
22:36:24 Join schrottplatz [0] (
22:36:32gevaertsYou want to centralise control handling I guess?
22:36:39tomersgevaerts: I had to send data in the HID from the control request handler. I either had to implement a statically allocated buffer in the HID class, or use the one in the USB core.
22:37:27tomersThis will allow in the future for the class drivers not to allocate buffer statically.
22:37:29gevaertshm, ok. I guess HID is a bit special in that it needs more data to set things up than to actually function
22:37:59tomersBut do you think that this is the appropriate solution?
22:38:07gevaertsI mean, for storage we don't care because we use dozens of kilobytes anyway
22:38:44gevaertswell, that memory is there anyway, so we might as well use it
22:39:05 Join barrywardell [0] (n=barrywar@
22:39:14tomersGreat :-) So are you going to commit this?
22:40:20 Quit FlynDice (Remote closed the connection)
22:40:54CIA-19gevaerts r20735 trunk/firmware/usbstack/ (6 files): Allow class drivers to reuse the core data buffer for control transfers. This doesn't make much difference right now, but it should keep HID memory ...
22:43:48 Quit killan ("( :: NoNameScript 4.22 :: )")
22:45:23mcuelenaeregevaerts, tomers: did you try this patch on target ?
22:46:03gevaertsmcuelenaere: I didn't, but it's part of a larger patch that does work. Did it break anything?
22:46:37mcuelenaereI'm not sure, I'm getting device descriptor read errors
22:48:00 Quit midijunkie ("?(???~•~)?")
22:49:19 Join midijunkie [0] (
22:49:22gevaertsI don't really see how that can be caused by these changes. They just pass a pointer that's not even used for now. Let me get a target...
22:49:42 Join Harman [0] (n=chatzill@
22:49:45mcuelenaerehmm then this will probably be my buggy USB driver
22:49:48 Quit midijunkie (Client Quit)
22:49:54mcuelenaerereverting to r20734 didn't resolve it, so nvm
22:50:02 Join midijunkie [0] (
22:53:30 Join FOAD_ [0] (
22:53:35 Join evilnick1 [0] (
22:54:07schrottplatzsomehow i dont like the main menu
22:54:19 Join killan [0] (
22:54:29 Quit FOAD (Read error: 60 (Operation timed out))
22:54:30 Nick FOAD_ is now known as FOAD (
22:55:18 Quit _jhMikeS_ (Nick collision from services.)
22:55:22 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
23:00:05gevaertsmcuelenaere: I can confirm that the code still works properly on my mini
23:00:35CIA-19mcuelenaere r20736 trunk/firmware/target/mips/ingenic_jz47xx/lcd-jz4740.c: Fix some issues with YUV blitting on Onda VX747 (still not working properly)
23:01:52amiconnmcuelenaere: The yuv test in test_fps is supposed to show a two-dimensional colour gradient
23:02:16mcuelenaereis there any screenshot of that somewhere?
23:02:29amiconnOne axis varies the U component, the other varies the V component
23:02:47amiconnThe Y component is fixed
23:03:12amiconnAnd the format is indeed 4:2:0. This isn't chosen by rockbox, it's just the standard MPEG YUV format
23:03:31mcuelenaereYUV 4:2:0 ? (not YCbCr 4:2:0)
23:03:50amiconnThere is no target screenshot since this is blitted directly to the lcd. The sim should show it though iirc
23:03:56 Quit _Auron_ (Read error: 104 (Connection reset by peer))
23:04:16 Quit LambdaCalculus37 ("too nice to sit on my butt all day")
23:05:02amiconnYeah, YCbCr to be precise. YUV is just shorter (although not 100% precise naming)
23:06:00mcuelenaerehmm the IPU supports both YUV 4:2:0 and YCbCr 4:2:0 as input format
23:06:39tomersgevaerts: I uploaded yet another cosmetic patch. I'm starting to 'clean my desk up' before the HID patch. Maybe I'll do it tomorrow, as I am going to sleep soon...
23:07:12gevaertsthere's no hurry
23:07:29 Join Keripo [0] (
23:10:19 Join FlynDice [0] (
23:10:29amiconnmcuelenaere: The sim does indeed show how test_fps yuv output should look like
23:10:44froggymanis there a list of all the screen sizes somewhere, for all supported rockbox targets?
23:10:50mcuelenaereyes, I saw it (and on-target doesn't look like it at all)
23:11:33bluebrotherfroggyman: check the firmware/export/config-*.h files or rbutil/rbutilqt/rbutil.ini
23:11:47 Quit evilnick (Read error: 113 (No route to host))
23:11:50bluebrotherit's not a list of all screen sizes but they are present there.
23:11:57gevaertstomers: I haven't gone through it all yet, but you replace 0x80 by USB_DIR_IN somewhere in usb_core_ack_control(). While the USB_DIR_IN is indeed 0x80, I tend to see it more as an endpoint address kind of thing
23:12:16*gevaerts is probably being silly about it
23:14:22froggymanbut, there isnt a list on the wiki of the screen sizes
23:14:38rasherfroggyman: DeviceChart
23:16:12tomersgevaerts: I agree with you (for now). Can you please remove this, and continue with the reviewing process, and I'll look over it more seriously later?
23:16:49tomersMaybe there's more appropriate #define for that. Anyway, I dislike having hard-coded values in the code...
23:17:30gevaertsActually, usb_ch9.h has "It's also one of three fields in control requests bRequestType" in the comment that explains USB_DIR_IN
23:18:32froggymanrasher: where exactly do i find that?
23:18:51bluebrotherfroggyman: that's a wiki name. Just go to the wiki page with that name
23:18:59rasherfroggyman: You type in DeviceChart in the "Go" field
23:19:10rasheror the search field in the menu
23:19:33bluebrothersearch? What's that?
23:20:42froggymanok enough of the mockery, i found it
23:25:39 Quit Harman ("ChatZilla 0.9.84 [Firefox 3.0.8/2009032609]")
23:27:59gevaertstomers: I think I'll leave that USB_DIR_IN there. The comment in usb_ch9.h makes it clear that it applies to bRequestType as well
23:32:44CIA-19gevaerts r20737 trunk/firmware/ (6 files in 3 dirs): USB related Cosmetics, whitespace and readability fixes (FS #10147 by Tomer Shalev)
23:33:47Bagder <= now with the correct menu setup
23:35:45 Quit einhirn_ ("Miranda IM! Smaller, Faster, Easier.")
23:39:01 Join FOAD_ [0] (
23:48:34 Quit Horscht ("Verlassend")
23:48:55 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
23:51:01 Quit FOAD (Read error: 110 (Connection timed out))
23:51:01 Nick FOAD_ is now known as FOAD (
23:52:37 Quit Seed ("cu, Andre")
23:58:37 Quit archivator ()

Previous day | Next day