Nightwulf|work: hi all
Nightwulf|work: hmm...actual git isn't compilable o.O
obv1us: Wow, with some hacks I got it to compile on Feisty.
obv1us: (sorry, this is from the git I pulled yesterday.)
obv1us: It produced a file xf86-video-radeonhd/src/.libs/radeonhd_drv.so
obv1us: How do I install it? Do I just copy it to where the other driver .so's are ?
egbert: obv1us: if prefix and libdir were set correct: make install
egbert: obv1us: what hacks?
obv1us: oh, forgot about that
obv1us: where do I set prefix and libdir ?
egbert: obv1us: what hacks did you need? it should build just like that.
Nightwulf|work: can't even compile git with a pull from this morning
obv1us: The hacks: I had to go get some header files from Gutsy, then I modified radeon_accel.h (defining uint*_t temporarily), and dangerously commented out a line in radeon_textured_video.c
egbert: obv1us: configure --libdir=
obv1us: egbert: thanks I will try
egbert: ok, i see the uint problem. going to prelare a fix.
obv1us: cool :)
egbert: obv1us: what did you comment out in radeon_texured_video.c?
obv1us: line 236: exaMoveInPixmap(pPriv->pPixmap);
obv1us: Looks very dangerous to comment this - you didn't see me do it, ok?
egbert: obv1us: ok ;)
egbert: but what error message did you get?
obv1us: I've got notes on all I've done, so can clean up shortly.
egbert: obv1us: i'm asking because only this way we can fix the build. it's hard to test with several versions of x on several distros.
obv1us: I'm ready to help 100%
obv1us: I've never done this before though, compiling video drivers etc.
obv1us: Do you think this should be correct?
obv1us: libdir=/lib/modules/2.6.20-lowlatency/kernel/drivers/video/
obv1us: oops ..2.6.20-16...
obv1us: does that look right?
obv1us: and prefix=/usr ?
egbert: obv1us: it's not a kernel driver. therefore:
obv1us: ok
egbert: libdir=/usr/lib??/xorg/modules/drivers
egbert: ?? depends on what system is running.
obv1us: how about /usr/lib/xorg/modules/drivers/ ?
egbert: saves bet: check if you have /usr/lib64
obv1us: it contains voodoo_drv.so etc..
egbert: if you do use lib64 else use lib
egbert: yeah, /usr/lib/xorg/modules/drivers/
obv1us: My OS is 32bit (running on 64bit hardware)
egbert: ok. then lib
obv1us: and prefix=/usr ?
egbert: yes.
obv1us: right, so I should issue to the command line...
obv1us: ./configure --libdir=/usr/lib/xorg/modules/drivers/ --prefix=/usr
egbert: ah, no
egbert: --libdir=/usr/lib/
egbert: the rest will be added for you.
egbert: sorry for the confusion.
obv1us: That's ok. Seemed to work alright (but "DRI support is disabled")
obv1us: make
obv1us: ok
egbert: Option "DRI" "on"
obv1us: THat's a xorg.conf device section option isn't it?
libvde: default is on on r5xx
obv1us: Oh... it means "DRI support is disabled by *default*" right ?
obv1us: it doesn't mean that it's not compiled in, which is what it looks like on the surface.
obv1us: am I right?
obv1us: make install
obv1us: ok, it placed radeonhd_drv.so into /usr/lib/xorg/modules/drivers
obv1us: :)
libvde: obv1us: which hardware is this?
telmich: hello
telmich: is any radeonhd dev in here? the current git master branch fails to build again
telmich: with and without dri enabled
libvde: telmich: and the errors are what this time again?
libvde: more uint* stuff?
telmich: libvde: no
telmich: In file included from radeon_accel.c:68:
telmich: /usr/include/xorg/dri.h:199: error: expected ‘)’ before ‘pScrn’
telmich: and many more
telmich: I am copying it right now..
obv1us: libvde: my card is Asus EAH3450 (RV620)
libvde: obv1us: there is of course no DRI on this hardware
obv1us: Isn't there? gosh.
telmich: libvde: https://nopaste.eof.name/750
obv1us: well, after make install, I've edited xorg.conf, set driver to 'radeonhd', then tried startx, and I have black screen - a hard lock it looks like.
telmich: is a bit pissed of, thought the master tree is the one that should always build (as found in an answer on the ml)
obv1us: Sorry to disappoint. (I was kind of expecting this, after mixing in source files from Gutsy like that.)
obv1us: ... not quite a hard lock, I can ping my machine.
obv1us: .. and can ssh in.
Nightwulf|work: thinks, *everything* committed to a repository should compile...not nessecarily work but at leat compile ;-)
Nightwulf|work: +s
obv1us: that would be ideal
libvde: telmich: our standards apparently did slip a bit recently :)
telmich: libvde: yep, looks like... I added non dri version compile errors to https://nopaste.eof.name/751
telmich: *grummel*
libvde: ah, bleh... there's massive attack playing today half an hour from my previous home in .be :(
libvde: nobody ever comes to nue to play a gig :(
telmich: has the problem, when running xrandr, that the primary display is set off
libvde: heh, my machine here is unable to support DRI, but that build bombs as well
telmich: well, with and without dri
telmich: err
telmich: no, without dri
telmich: 1.2.1, built from git branch master, commit abccd6ff
libvde: yeah
libvde: it's really terrible
obv1us: My black screen is not a hard lock. I can still get the machine to do stuff by keyboard.
obv1us: The Xorg.0.log : http://pastebin.com/m741385ff
libvde: unsurprisingly, CS works :)
libvde: but there's some logic issues in there that egbert pointed out yesterday that i still need to go and fix :)
libvde: will hopefully get to that today :)
telmich: well, I can do some bugfixing and trying to get it running
telmich: libvde: did someone of you decide what todo with the broken uint_32t thingies?
telmich: libvde: perhaps I can apply the uint32_t rewrites I put to http://unix.schottelius.org/cgi-bin/gitweb.cgi?p=foreign/radeonhd;a=summary again onto the master branch
libvde: telmich: i'm not sure the -ati developers will like this, and this then makes our technology partner unhappy
telmich: libvde: ok, so what's the way to solve it?
telmich: libvde: I am currently running a quite bad mode on my xorg, so I want to fix it _now_ :-)
libvde: try CS :) the fixes that are needed won't take long, just need to get to doing them
telmich: *checkout* *configure* *build* ...
telmich: ahh, very less errors! https://nopaste.eof.name/752
libvde: telmich: ./autogen.sh :)
telmich: libvde: argh
telmich: libvde: thanks :-)
libvde: :)
libvde: well, this is the easiest of the autogen fixable errors
telmich: *recompile*
libvde: some other times an object isn't included, and then the X server just stops no message no nothing
telmich: woooooot? it builds
telmich: *makeinstall*
libvde: :)
telmich: hmpf:
telmich: [12:12] denkbrett:~% xrandr --output PANEL --auto
telmich: xrandr: Configure crtc 1 failed
telmich: and the panel is black again
libvde: telmich: oh, is this the hd3xxx laptop?
telmich: libvde: ibm t60
libvde: ah, X1400 also broke?
telmich: btw, for reference: (II) RADEONHD: version 1.2.1, built from git branch CS, commit b05d54ca
libvde: yuck!
telmich: libvde: yep
libvde: libv: when you finally get to the office, drag out the m56 laptop and test this
telmich: libv: you are having multiple personalities!
telmich: libv: a hint on how to fix that?
libvde: telmich: yeah, i should find out how to vpn into the company and then use screen on the irrsi client there :p
telmich: may be a bit annoying, but I am totally motivated today, as I fixed the "no cursor problem" in gpm-1.99.7 :-)
libvde: :)
libvde: thing is, m54 should've had its display enabled properly for a long time already
telmich: the "who the fuck does 7 ifs in a row - rewrote-bug"
libvde: maybe it pays to find a working version, and bisect?
telmich: libvde: it _has_ been good some time ago
telmich: libvde: I know there's a working version, I mailed it about two weeks ago to the ml
telmich: 1.2.1-187-g4eebf4b is good for instance
telmich: there must be some bisect output on the ml
telmich: I did quite a long bisection run, but I do not see it on the ml
libvde: great :) first thing i have to do when i get to work is finish up the driver porting for opensuse, then i'll get to this issue
libvde: m54 panel regressing is _very_ bad
telmich: ahh, it was dropped by the ml software
telmich: *repost*
libvde: driver porting takes an hour or two, as it's just testing and removing some test code left :)
telmich: libvde: the latest answer to "Re: [radeonhd] Primary screen black and cannot be reenabled"
telmich: "the last working version is 7fd3760ec6144dd3d5c2e69d497454d949b48b7c", "the first broken commit is indeed d09223bdda031503fa3ea04324fe85bf4cb73478"
telmich: I think I'll reset my radeonhd driver to 7fd3760ec6144dd3d5c2e69d497454d949b48b7c again
obv1us: my black screen was accompanied by usplash eating 100% cpu
telmich: *build*
ndim: Hmm.
ndim: All those "static const OptionInfoRec RHDOptions[]" options are for the 'Section "Device"' in xorg.conf, right?
telmich: ah, it works
telmich: *xrestart*
telmich: libvde: btw, any plans to get console working again, after one started radeonhd? and the brightness stuff?
obv1us: egbert: sorry, didn't answer you before; this is the error I got before commenting out the line:
obv1us: radeon_textured_video.c:236: error: implicit declaration of function 'exaMoveInPixmap'
libvde: ndim: yeah
libvde: telmich: we're still trying to get our hands on somebodies "old" m54 in the company
telmich: libvde: hmm, ok
libvde: i think he ordered something new already for his company laptop, and then we are supposed to get his old one :)
libvde: m54 still has some restore issues we haven't been able to fully iron out yet
libvde: but i'll try to reproduce the blank panel thing on an m56 first
telmich: libvde: hmm, ok...I do not have any other laptop with m54, otherwise I could do more intensive testing
telmich: libvde: what does CS stand for, btw?
ndim: telmich: command submission
telmich: ndim: thanks
Nightwulf|work: ndim: are the compile errors in the actual git master already known?
ndim: In current masteR?
Nightwulf|work: yes
Nightwulf|work: e.g.
Nightwulf|work: r5xx_xaa.c: In Funktion »R5xxXAASetClippingRectangle«:
Nightwulf|work: r5xx_xaa.c:132: Fehler: »struct RHDRec« hat kein Element namens »TwoDPrivate«
Nightwulf|work: and d some more of them
Nightwulf|work: -d
Nightwulf|work: uhm...i should have done an LANG=C make instead of make :P
ndim: Nightwulf|work: I'm trying to reproduce.
ndim: Nightwulf|work: I am German, so I can read this, even though that German message stuff is quite unfamiliar :)
Nightwulf|work: ndim: yes, i know...but here are not german speeking people too...so it's just unpolite i think :)
telmich: Nightwulf|work: yes, known
Nightwulf|work: k, thx
telmich: Nightwulf|work: I am flamming around the whole day, because they annoy me -- libvde pointed me to the CS branch, that compiles
ndim: Well... a default compile of master works for me. --disable-dri breaks it with that stupid uint32_t definitions.
ndim: I can't reproduce Nightwulf|work's problem yet.
telmich: ndim: saw above the compile errors I've?
Nightwulf|work: ndim: it's just a configure --prefix=/usr && make, nothing special...compiled all git builds that way for several weeks
telmich: hmm. I also have this one here: https://bugs.freedesktop.org/show_bug.cgi?id=13405
ndim: telmich: Yupp, and they do look very strange.
ndim: I fear someone will need to dig deep into the #include mess to properly fix all that stuff.
Nightwulf|work: the only real "bug" i have in radeonhd(git or non-git) with the HD3650 @work is a corruption of all white space characters after shutting down xorg
telmich: has the hardware-cursor, brightness, non-compiling and panel black problems (x1400 on the t60) :-)
mlankhorst: Fun ;-)
Diablo-D3: Nightwulf|work: that happens on my 4850 too
libv: for the germans: http://www.heise.de/newsticker/AMD-beschleunigt-Chipsatzgrafik-mit-neuem-790GX-Chipsatz--/meldung/113865
Nightwulf|work: Diablo-D3: the white space corruption?
libv: rs780 but 40% clock increase and sideport memory
Diablo-D3: yeah
libv: this thing will rock, but i don't think we can support sideport memory yet, so we will have to wait for hw again
Nightwulf|work: strange...my 3870x2 works perfectly
Diablo-D3: shrugs
Diablo-D3: the only other bugs I have are lack of 3D acceleration and xv
libv: rs780 already was mighty fast for igp... 40% increase in clock will be quite measurable
libv: Diablo-D3: those are features i think :p
Nightwulf|work: Diablo-D3: well, that's not a bug really
Diablo-D3: nyet, bugs
Honk:
mlankhorst: I'm probably going to play around with r7xxx, maybe I can get 3d accel working after a lot of work.
Diablo-D3: Honk: I already use it
Diablo-D3: but not gl2
Diablo-D3: vo=gl:ati-hack:yuv=4:rectangle=1:swapinterval=1:lscale=1:cscale=1
Honk: gl doesnt have osd, does it?
Diablo-D3: gl does, gl2 doesnt
Honk: errh
Honk: ;P
Diablo-D3: gl has screen resolution softsub rendering too
Diablo-D3: which is hilarious when you have soft subs combined with a really shitty video
Honk: I guess I should change my config sooner or later ;P
Diablo-D3: ultra sharp font rendering combined with blurry and blocky video
Diablo-D3: but yeah, the above looks fucking sweet
Honk: yeah, gl2 does that too
Diablo-D3: gl2 just inherited gl's code then
Diablo-D3: gl2 used to not have it at all
Honk: prolly
Diablo-D3: remembers when gl2 didnt even exist
Honk: well.. the manuals says gl2's gonna be discontinued
Diablo-D3: lawl
Honk: *shrug*
Diablo-D3: probably because it was rather pointless
Diablo-D3: gl was already as fast as its going to get
Honk: it surely didnt say that when I created my config ;P
Diablo-D3: the only thing they could add is shaders that do colorspace conversion
Honk: I've got no problem with performance anyway
Diablo-D3: (gl still does yuv->rgb in software)
Honk: (when the drivers do support opengl ;P)
Diablo-D3: but yeah, like, my settings use the bicubic spline shaders
Diablo-D3: so it looks really damned good
Diablo-D3: film on dvd looks like film instead of pixelated shit
Diablo-D3: just need a little optional -auto 100 -vf spp to fix it up
Diablo-D3: (do NOT enable that for h264, it does wrong things)
Honk: hehe
Honk: now all we need it opengl support =D
brot: does the fglrx module need to be unloaded while using the radeonhd driver ?
ndim: brot: yes, and you really really should reboot
brot: ok ;)
brot: gentoo-wiki doesnt mention that
brot: so i added the note here, and am waiting until the download is finished so i can try the driver.
brot: is already excited
Diablo-D3: so 3D support for 4xxx dead?
Diablo-D3: inserts an 'is' in there
marcheu: Diablo-D3: gl doesn't do yuv -> rgb in software, check your facts
ndim: Diablo-D3: 4waht? R4xx?
marcheu: Diablo-D3: if features a heap of conversion methods
ndim: Or HD4whatever?
Diablo-D3: ndim: 4xxx as in r700s
Diablo-D3: marcheu: it didnt used to
Diablo-D3: marcheu: when I say convert, I dont mean scale the chroma lanes
marcheu: "scale the chroma lanes" ?
Diablo-D3: yes, apparently my p key doesnt work always
marcheu: yes gl does full color space conversion, with many different techniques
Diablo-D3: it should, yes, but last time I looked it didnt
marcheu: right, so if last time you looked was 5 years ago, don't assume it's still the truth
Diablo-D3: (its not even that hard to convert between wildly different colorspaces, complete with white and black point compensation, using no shaders)
Diablo-D3: ndim: I was told someone was working on 3D support for hd4000s for radeonhd, but had an NDA from ATI so they couldnt release any of it
libv: Diablo-D3: eh?
emmes: Diablo-D3: you obviously have no freaking clue.
libv: Diablo-D3: who would that be then :)
emmes: sorry for being rude
Diablo-D3: this is what I was told, I dont know if its true or not
Diablo-D3: emmes: you attribute the blame wrongly
emmes: ok, then this guy has no freaking clue :]
emmes: I'm working on 3D, we have docs, yes, they will be public.
Diablo-D3: on 3D for what exactly?
emmes: no, the chip isn't drawing a single triangle yet. :-(
Diablo-D3: I dont care if any card but my 4850 has 3D
Diablo-D3: has no clue why the hell ATI just doesn't release working code
emmes: me and 3-4 people at ATI are currently looking *why* the programming we have doesn't work the way it should
Diablo-D3: lol nice
Diablo-D3: (and for the record, fglrx is not working code)
z3ro: Diablo-D3, I don't know for sure, but I would guess most likely patent/license issues.
emmes: it's IP issues, lawyers, and more crap no developer ever wants to look at (me too)
z3ro: Diablo-D3, although we already have some code from amd.
brot: :)
Diablo-D3: z3ro: no, that prevents fglrx being released
brot: yesterdays radeonhd cs branch works really nice.
brot: with my 3850
Diablo-D3: z3ro: and that only prevents _all_ of fglrx
libv: brot: it won't gain you anything on 3850 :)
libv: brot: except maybe a working build :)
Diablo-D3: I dont care about the crap IP stuff prevents
Diablo-D3: I only care about the 3D engine
mlankhorst: Lawyers do :-/
emmes: Diablo-D3: actually you do. it prevents you from having the docs and more developers working on it :-/
Diablo-D3: emmes: see above
Diablo-D3: only care about the 3D engine
brot: libv: well, i read on phoronix that the cs branch has some nice progress the last few days, so i checked out that one :)
Diablo-D3: I dont care about UVD, I don't care about the more estoric gpgpu stuff, I don't care about anything but 3D
z3ro: Diablo-D3, there are still legal issues. you know everything released must be cleared by amd legal first. (at least from what I understand)
Diablo-D3: (I doubly dont care about UVD, ffmpeg will never support it)
Diablo-D3: z3ro: cleared, sure, but Im saying theres no non-ATI IP involved here
Diablo-D3: the lawyers cant outright prevent the stuff Im interested
Diablo-D3: (because it'd imply ATI doesn't even own their own 3D engine, which means someone has a 3D engine I could buy)
Diablo-D3: and, for the record, 2D stuff works
Diablo-D3: so they can't retroactively take that away from me
Diablo-D3: (mind you, its slow as shit, but it does infact work)
z3ro: also keep in mind writing the programming guides is no small task.
Diablo-D3: true
z3ro: r6xx+ does things quite differently to previous generations.
Diablo-D3: r6xx is unrelated to previous generations, thats why
Diablo-D3: same way r3xx took a huge break from r2xxx
Diablo-D3: they slapped the r200 2D engine on that company they bought's 3D engine and filled in the gaps with whatever they could find
z3ro: r3xx to r5xx are all pretty similar in they way they are programmed.
Diablo-D3: yup, they are
mlankhorst: What about r7xx ?
mlankhorst: Is that done differently from r6xx?
Diablo-D3: kind of
z3ro: from what I understand (I've only dumped r6xx) they should be similar, but some regs moved around.
z3ro: eg MC setup etc
Diablo-D3: the arch is massively differently designed
Diablo-D3: but the interface from what Im told is very similar
Diablo-D3: isn't AMD pushing ATI towards a forever unified interface for future GPUs?
mlankhorst: I just care about the driver level interface right now, I've ordered a rd 4850 and I would love to get it working with open source drivers. :-)
Diablo-D3: I have a 4850
mlankhorst: But I haven't seen specs for it yet.
Diablo-D3: it does work with radeonhd
Diablo-D3: its also an extremely fast card
Diablo-D3: its faster than a 3870x2
Diablo-D3: but uses about as much power as a 3850
Diablo-D3: the upcoming 4870x2 is faster than anything nvidia has
z3ro: I'm kind of annoyed that I didn't get to buy a r7xx at the SEG (or maybe SCG, I forget now) market. they are about half price.
Diablo-D3: I bought a 4850 at about $175, they're like $150 now
z3ro: iirc the price I got for a 4870 was 1500 yuan which is about 300NZD
z3ro: normally they are about 600NZD locally (in nz)
Diablo-D3: whats that in USD?
Nightwulf|work: bought an 3870x2 some weeks before the 4xxx came out...damn s**t
z3ro: 219USD according to xe.com/ucc, but I'm not sure thats right.
Diablo-D3: google says 600NZD = $432
Diablo-D3: thats fucking bullshit
Diablo-D3: the 4870s were $299 the day they came out, and they're below $250 now
Diablo-D3: I'd move out of NZD honestly
Diablo-D3: thats fucking bullshit
z3ro: yeah hopefully I could move to another country soon, but it's expensive.
z3ro: and then there is finding a job, etc etc
Diablo-D3: heh
Diablo-D3: why dont you start a store that sells hardware cheaper?
z3ro: bbl, food
mlankhorst: When will specs be available for r7xx?
GraveDigger: only god knows
GraveDigger: ... and the amd/ati folks, of course
Diablo-D3: even god isnt allowed to know, he refused to sign the nda
bridgman: Diablo-D3 - for 6xx/7xx 3d we're down to a couple of issues but they're non-trivial. I'm on vacation next week (officially 2 weeks but online for the 2nd week) so best guess is 2-3 weeks from now
bridgman: for 7xx display/modesetting there isn't much difference so we're treating that as lower priority
Honk: bridgman: sounds awesome ;P
Honk: I thought it'd be months away =)
Diablo-D3: 2D works for me, btw
Honk: radeonhd's software 2d already faster than fglrx's accelerated 2d for me ;P
Diablo-D3: not for me
Honk: I've got some trouble with my fglrx driver I think ;P
Diablo-D3: it may be a lack of something gtk apps use though
Diablo-D3: or, specifically, firefox
Honk: I'm talking about moving windows around
Diablo-D3: Im not sure if I should blame this on xrender, because fglrx iirc lacks it too
Honk: ;P
Diablo-D3: its slower than compiz, honk
Diablo-D3: since compiz can damage proof your windows as well
Honk: hurr
Honk: :]
Diablo-D3: hey, that uses time too
Diablo-D3: especially on apps that manage their own shit and arent smart enough to cache the canvas
Diablo-D3: (though technically thats a fix at the wrong level, compiz is the right answer)
Diablo-D3: (although compiz is the wrong answer for X, because X is ultimately the wrong answer)
Diablo-D3: ie, hardware compositing = good, X = bad
Honk: well.. but compiz isnt fluxbox ;P
Honk: and I'm not ready to switch my window manager because my gfx driver sucks
mlankhorst: :-)
Diablo-D3: thats only partially why X sucks
Honk: well.. X is doing a good enough job with proper drivers
Diablo-D3: X does a lot of shit wrong
Honk: sure
Diablo-D3: window managers is one of those problems
Honk: but still good enough for me ;P
Diablo-D3: as is allowing apps to do their own UI widgetry
Diablo-D3: and the combination of the two where the window decoration isn't integerated into the UI widgetry
Diablo-D3: and the fact it isn't already doing compositing on its own
Honk: :]
Diablo-D3: plus the fact you can't directly overlay the X protocol over GL
Diablo-D3: its just a horrid mess of crap
Honk: *nod*
Honk: still.. good enough ;P
Diablo-D3: not really
Honk: for me anyway ;P
Diablo-D3: its on my todo list to flush it all away
Diablo-D3: its shit, and in the sewer it belongs
brot: Diablo-D3: you sound very productive ;)
Honk: yeah, port everything to Y ;P
Diablo-D3: I have more important things on the todo list above replacing X
Diablo-D3: Y sucks, honk
Honk: I was kidding obviously ;)
Diablo-D3: actually, its funny, I may write this X replacement in perl
Diablo-D3: just because Im on a perl kick lately
Honk: just make it backwards compatible ;)
Honk: if it's not, noone's gonna use it
Diablo-D3: thats actually impossible
Honk: orly
ssam: will it run on Perl/linux http://perllinux.sourceforge.net/
Diablo-D3: and it'd be stupid
Honk: sure would be
Diablo-D3: I can provide a rootless X server though
Honk: but if it's not compatible, it'll die like Y did
Diablo-D3: or rather, someone can
Honk: easy like that :]
Diablo-D3: Honk: that isnt it
Diablo-D3: you cant use gtk or qt either
Honk: that's my point
Diablo-D3: nor could you simply port apps
Honk: noone's gonna rewrite all their programs
Diablo-D3: they would when you realize Im providing things people always wanted
Honk: no they wouldnt
Diablo-D3: such as real task oriented computing
Diablo-D3: and "crash proof" apps
Honk: unless you get compatibility into the picture somewhere
Diablo-D3: how? the apps simply wont support the additional features
Honk: that'd be fine ;)
Diablo-D3: so if you add an X app to a task, and save the task and close it and open it, the X app wont load where it used to be
Diablo-D3: that makes it rather useless
Diablo-D3: I'm also adding the ability to send apps to other machines as well
Honk: well.. better than not having any usable system at all
Diablo-D3: X apps /really/ wont support that
Diablo-D3: well, it depends how you define usable
Diablo-D3: I use the same 4 apps day in and out
Honk: well.. having new stuff is fine, but if I cant fire up k3b at some time, it's not usable
Diablo-D3: a web browser, vim, and a xterm.... okay, 3 apps
Honk: (assuming I'd burn stuff regularly ;P)
Diablo-D3: Honk: k3b really sucks
Honk: *shrug*
Diablo-D3: I'll write my own cd burner
Honk: do that after you've got some wrapper for X apps ;)
Diablo-D3: that isnt up to me
Diablo-D3: I'll leave that task to someone else
Honk: wouldnt need to support any fancy stuff =)
Diablo-D3: what will be supported damned easily, though, are glx apps
Diablo-D3: since ogl will be basically the lowest level drawing API
Honk: =)
Honk: well.. good luck
Diablo-D3: btw, you obviously wont be using gnome on it
Diablo-D3: since thats a conflict of interest
Diablo-D3: the on screen part of the desktop environment will most obviously be built in
Honk: so you're gonna dictate how the windowmanager is gonna work? ;P
Diablo-D3: yup
Diablo-D3: thats going to be built into the display server
Honk: that sucks tbh :]
Diablo-D3: no, its actually the opposite of sucks
Diablo-D3: I'm forcing themes to also do UI logic
Honk: yeah, cuz your windowmanager will do everything anyone could ever want
Honk: oh
Diablo-D3: yup, my window manager _will_ do everything
Honk: well.. that's not really built in then
Diablo-D3: actually it _is_ built in
Honk:
Diablo-D3: its no different than using sawmill
Honk: wha?
Diablo-D3: sawmill themes are written in lisp
Diablo-D3: how they actually function is all controlled by the lisp
Honk: well, when there are themes that will change how the windowmanager works, it wont really be built in
Diablo-D3: except it will be
Honk: wtf
Diablo-D3: theres no seperate window manager program
Honk: there are themes that'll do the job
Diablo-D3: the themes wont be seperate programs though
Honk: that doesnt matter
Diablo-D3: they'll just change the behavior of the display server
Honk: it's not hardcoded ;)
Diablo-D3: why should it be hardcoded?
Diablo-D3: hard coding shit like that is fucktarded
Honk: I know
Diablo-D3: like, all window managers DO THE EXACT SAME THING
Diablo-D3: ALL
Diablo-D3: THE
Diablo-D3: EXACT
Honk:
Diablo-D3: SAME
Diablo-D3: THING
Honk:
Diablo-D3: yet theres like 500 fucking window managers
Honk: we had that discussion already ;P
Diablo-D3: this is not only a waste of time and effort, its fucking retarded
Honk: well.. go ahead and code it =D
Diablo-D3: its on my todo list
Honk: how long is that list?
Honk: *lol*
Diablo-D3: getting amazingly rich is higher on the list
Diablo-D3: the todo list is in order of whats most likely going to get done, least likely at the bottom
Diablo-D3: if Im rich, I can spend the next 40 years doing the rest of whats on it
Honk: you'll probably spend the rest of your life trying to get rich enough to do anything on that list at all =D
Diablo-D3: nope, unlikely
Temujin_: hello
Simon-: does anyone have i2c for ddc working in any form? :/
Simon-: I used to be able to use it via the radeon fb driver, but that doesn't support this newer card - it's annoying to have to use the front panel controls to change brightness/contrast
sgodsell: hey guys does the ati x700 card work with the radeonhd driver?
MostAwesomeDude: sgodsell: Yep.
sgodsell: thanks MostAwesomeDude :-)
sgodsell: hmmm, will it support compiz?
MostAwesomeDude: sgodsell: No problem. Both radeon and radeonhd support it.
egbert: Simon-: you mean ddc ci, not ddc.
airlied: sgodsell: its -ati driver
airlied: MostAwesomeDude: radeonhd doesn't.
MostAwesomeDude: airlied: X1700?
airlied: MostAwesomeDude: he said X700
MostAwesomeDude: Oh, oops. X700. My bad.
MostAwesomeDude: XD
airlied: there is no x1700
MostAwesomeDude: ( >&)
sgodsell: hmmm, okay what driver does then?
MostAwesomeDude: sgodsell: Mesa 7.0.3, Xserver 1.4.2, xf86-video-ati 6.8, and you should get compiz.
airlied: sgodsell: -ati
MostAwesomeDude: sgodsell: xf86-video-ati supports all Radeons.
sgodsell: okay thanks guys
sgodsell: cya
matteo: does the CS branch supports EXA on a radeon 4870?
airlied: matteo: no
MostAwesomeDude: matteo: AFAIK there's no EXA on r6xx+ yet.
matteo: mine is r770
agd5f: matteo: r6xx and r7xx are in the same boat
surfer: i just built the latest repository version and it was fucked
surfer: there is text corruption and it doesn't load drm or even use it at all
surfer: i thought radeonhd was in a working state?
yangman: works for me
yangman: FMMV, which is why there's no release
surfer: what card?
yangman: mobility X1300
surfer: i have an x1800
surfer: so for u it's using drm?
surfer: and 3d acceleration works
yangman: yeah
surfer: can i see ur log?
surfer: and config
yangman: http://pastebin.ca/1094261
yangman: http://pastebin.ca/1094263
yangman: it's the first 900 lines or so
surfer: thanx
yangman: I'm running the CS branch currently, but haven't had any major issues with recent master either
surfer: hmm
surfer: option dri
surfer: is that required?
surfer: u dont need that for the radeon driver
surfer: yes
surfer: apparently option dri is required
surfer: that fixed it
surfer: no more corruption, time to test the rest of it
surfer: hmm
surfer: xvideo isnt working
surfer: is there a known problem playing videos with cmopiz?
surfer: compiz
agd5f: surfer: switch to exa
surfer: (**) RADEONHD(0): Selected EXA 2D acceleration.
surfer: (--) RADEONHD(0): Detected an R520 on a Powercolor X1800XT
surfer: when i play videos it's just solid black window
agd5f: surfer: what does xvinfo say?
surfer: X-Video Extension version 2.2
surfer: screen #0
surfer: Adaptor #0: "Radeon Textured Video"
surfer: number of ports: 16
surfer: port base: 91
surfer: operations supported: PutImage
surfer: supported visuals:
surfer: depth 24, visualID 0x21
surfer: no port attributes defined
surfer: maximum XvImage size: 4096 x 4096
surfer: vo x11 works ok
surfer: any ideas?
yangman: are you running the necessary versions in the stack as per http://wiki.x.org/wiki/radeonhd%3ADRI?
yangman: er.. http://wiki.x.org/wiki/radeonhd%3ADRI
surfer: yea
surfer: and xv worked on the other radeon driver
surfer: lol, i see the part about option dri on that wiki page though
surfer: i'm gona try it without compiz on
surfer: it works when compiz is not enabled
surfer: but black window when compiz is running
surfer: with the other radeon driver it worked either way
yangman: file a bug
surfer: ok
surfer: do u use compiz?
yangman: no
surfer: aight
yangman: although I'm thinking of giving it a try in a week
surfer: before i file a bug i'm going to upgrade the the latest compiz build
surfer: u should
surfer: it's awesome