--- Log opened Сбт Июл 22 00:00:56 2006
00:35 -!- pmdata [i=patrice@ANantes-154-1-21-130.w81-53.abo.wanadoo.fr] has quit [Remote closed the connection]
00:51 -!- Lumag [n=Lumag@chimpanzee.school.ioffe.ru] has joined #nouveau
02:52 -!- tuXXX [n=damien@dyn-83-152-190-234.ppp.tiscali.fr] has quit [Read error: 104 (Connection reset by peer)]
03:40 -!- allix [n=allix@host81-158-40-25.range81-158.btcentralplus.com] has joined #nouveau
04:14 -!- xaid|work [n=zaid@d150-132-196.home.cgocable.net] has left #nouveau []
04:40 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has quit [Read error: 110 (Connection timed out)]
04:42 -!- xaid [n=Xaid@d199-126-153-40.abhsia.telus.net] has joined #nouveau
04:56 -!- Lumag [n=Lumag@chimpanzee.school.ioffe.ru] has left #nouveau []
05:56 -!- allix [n=allix@host81-158-40-25.range81-158.btcentralplus.com] has quit ["Leaving"]
06:36 -!- xaid [n=Xaid@d199-126-153-40.abhsia.telus.net] has quit ["leaving"]
09:54 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has joined #nouveau
10:21 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has quit [Read error: 110 (Connection timed out)]
11:25 -!- blx [n=x@h132n1c1o1028.bredband.skanova.com] has quit [Read error: 104 (Connection reset by peer)]
11:32 < Colbert> I think the following is glClearColor()
11:32 < Colbert> 185f   0x00000000   0xff00ff00                       NvType004a      [0x05fc/4] = 0xff000ff000ff00 | UNKNOWN = ff
11:33 -!- marcheu [n=marcheu@lattice.u-strasbg.fr] has quit [Read error: 104 (Connection reset by peer)]
11:34 < Colbert> It seem to convert to A8R8G8B8 from rgba.
11:35 < Colbert> at least on nv 34.
11:38 -!- pmdata [n=patrice@ANantes-154-1-70-172.w86-195.abo.wanadoo.fr] has joined #nouveau
11:40 < Colbert> hi
11:40 < pmdata> hi colbert
11:41 < Colbert> decided to go at the frambuffer and color buffers from another direction.
11:42 < pmdata> which is...?
11:42 < Colbert> trying to figure out how clears work
11:42 < pmdata> I think it uses the 2d blitting engine to define a rectangle from framebuffer, then zero it
11:43 < pmdata> can you create stereo buffers?
11:44 < Colbert> not yet, having problems with autocrud
11:44 < pmdata> you have a nv17, right?
11:44 < Colbert> and a nv34
11:45 < pmdata> ah, I don't think you can create stereo buffers no nv17
11:45 < Colbert> got the 34 in right now
11:46 < pmdata> you see some difference in dumps between single/stereo buffers?
11:47 < Colbert> i need to compile the new nvclock and autoheader is giving me trouble. cant find X11
11:47 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has joined #nouveau
11:48 < pmdata> you can compile without gtk/qt/whatever x11 gui, and keep the console version
11:49 < Colbert> i'm tempted to go back to our version,
11:50 < pmdata> nvtype004a is 	{__(NV04_GDI_RECTANGLE_TEXT), NV04|NV10|NV15, {
11:50 < pmdata> add |NV30 to the mask, and see if the info displayed is correct
11:51 < Colbert> is that new, i just upgraded crom cvs and added my code back in
11:52 < pmdata> no, see objects.c, this object was only defined for nv10 and older cards, but seems supported also by nv30
11:52 < pmdata> if it's ok, I'll commit the change in cvs
11:53 < Colbert> i'll paste my code
11:54 < pmdata> do you have any object type still displayed as nvtype00xx in your dumps?
11:54 < Colbert> http://rafb.net/paste/results/S3QbYk80.html
11:55 < Colbert> i'll look
11:55 -!- shenki [n=shenki@ppp192-59.lns1.adl4.internode.on.net] has joined #nouveau
11:56 < Colbert> see a nvtype0062
11:57 < pmdata> it is also known, see http://nouveau.freedesktop.org/wiki/NvObjectTypes
11:57 < pmdata> in objects.c, also add a NV30 mask for it
11:57 < pmdata> or I can do it in cvs, and you check it out?
12:00 -!- EdB|w [n=EdB@212.234.68.206] has joined #nouveau
12:02 < pmdata> colbert> I changed the cvs, can you checkout from it, and redump?
12:03 < Colbert> sure. 
12:04 < Colbert> do you see anything obviously wrong with my code
12:06 < pmdata> will have a look
12:07 < pmdata> seems ok, do you have a dump?
12:07 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has quit [Remote closed the connection]
12:07 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has joined #nouveau
12:16 < Colbert> hold on, having toruble pasting into firefox now.
12:21 < Colbert> I'll be right back.
12:21 -!- Colbert [n=colby@69.211.27.8] has left #nouveau []
12:23 -!- Colbert [n=colby@adsl-69-211-27-8.dsl.chcgil.ameritech.net] has joined #nouveau
12:25 -!- Colbert [n=colby@adsl-69-211-27-8.dsl.chcgil.ameritech.net] has quit [Remote closed the connection]
12:27 -!- Colbert [n=colby@adsl-69-211-27-8.dsl.chcgil.ameritech.net] has joined #nouveau
12:28 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has quit [Remote closed the connection]
12:28 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has joined #nouveau
12:32 -!- shenki [n=shenki@ppp192-59.lns1.adl4.internode.on.net] has quit ["Leaving"]
12:36 < Colbert> part 1
12:36 < Colbert> http://rafb.net/paste/results/5bJrYb83.html
12:38 < Colbert> part 2
12:38 < Colbert> http://rafb.net/paste/results/YzKlEa70.html
12:41 < Colbert> and part 3
12:41 < Colbert> http://rafb.net/paste/results/Ki2fCW77.html
12:44 < Colbert> sorry about having to break it up, but nopaste was also giving me trouble. and dont know what is wrong with firefox now.
12:48 < Colbert> almost forgot, heres my error handling code called in test_draw_buffers
12:48 < Colbert> http://rafb.net/paste/results/nGg4pn49.html
12:51 < pmdata> did you use a stereo buffer for your dumps? I see depth buffer setup twice
12:52 < pmdata> oops, it was stencil
12:56 -!- Lumag [n=Lumag@chimpanzee.school.ioffe.ru] has joined #nouveau
12:57 < Colbert> nope, not on this dump, tried it on earlier iterations
12:57 < Colbert> before i started doing the glClear
12:58 < Lumag> hi all
12:58 < Colbert> hi
12:59 < Colbert> need coffee, brb
13:00 < pmdata> ah, no, later hw has 2-sided stencil! one is front, second is back
13:02 < Colbert> where? what should i look for
13:06 < pmdata> checkout renouveau cvs, you'll understand
13:06 < Colbert> lumag just did a upload?
13:07 < pmdata> no, me
13:07 < pmdata> I saw it from your dumps
13:10 < Colbert> all those changes? Eureka! lol
13:13 < pmdata> stencil setup has a very recognizable pattern: mask (ff), function (207,0,ff) mask (1e00,1e00,1e00)
13:13 < pmdata> 1e00 = GL_KEEP
13:14 < pmdata> I just added nv30 front face
13:15 < Colbert> it does make sense now.
13:18 < Colbert> pmdata: how do you look at the results, I have been using kdiff to compare the output of diffferent runs, or breaking long dumps up into smaller parts and diff those, but the line numbering keeps getting in the way
13:19 < pmdata> yep, I also use diff, but only on very small dumps, so line numbering is not a problem
13:19 < pmdata> or you can 'cut' the line number out
13:20 < pmdata> hm, can not reach sourceforge cvs now, I still have some stuff to update
13:20 < Colbert> takes a long time on long dumps, i wonder if I can regexp out the line number in kompare (not kdiff)
13:21 < pmdata> ah it's ok
13:21 < pmdata> recheckout the cvs, more stuff
13:22 < Colbert> I kept hoping I would start to understand this, but you're going to figure it all out before i do
13:23 < pmdata> can you make me a dump for the test_nv_register_combiners function?
13:23 < pmdata> 900-914,8ec-8fc seems to be them
13:29 < Colbert> sure, one moment
13:32 < Colbert> http://rafb.net/paste/results/lHS6Vp41.html
13:42 -!- marcheu [i=marchesi@pellinore.u-strasbg.fr] has joined #nouveau
13:52 -!- EdB|w_ [n=EdB@212.234.68.206] has joined #nouveau
13:54 < pmdata> colbert> recheckout from cvs, and in tests.c, change gl_combiner0_nv to gl_combiner1_nv, and make a dump
13:57 < Colbert> done
13:57 < Colbert> http://rafb.net/paste/results/DSlTdd29.html
13:58 < Colbert> le me redo this, one second, didnt make the change.
14:00 < pmdata> if not already done, recheckout from cvs before, I added fog stuff
14:01 -!- EdB|w_ [n=EdB@212.234.68.206] has quit [Read error: 104 (Connection reset by peer)]
14:02 < Colbert> waiting on cvs, do you waant 2 dumps one without the change to gl_combiner0_nv?
14:03 < pmdata> no, I want the one with the change
14:04 < Colbert> checking. this line? glCombinerInputNV(GL_COMBINER1_NV, GL_RGB,
14:05 < pmdata> yep
14:06 < Colbert> http://rafb.net/paste/results/5dDgYJ24.html
14:09 -!- EdB|w [n=EdB@212.234.68.206] has quit [Read error: 110 (Connection timed out)]
14:11 < pmdata> hum, no change, like for nv20
14:18 < Colbert> has anyone looked for differences between driver versions? 
14:19 < Colbert> some of the older slower drivers might give us better data
14:25 < pmdata> I think we have enough work with current driver :)
14:27 -!- blx [n=x@h29n11c1o1028.bredband.skanova.com] has joined #nouveau
14:27 < pmdata> colbert> does glxinfo list NV_texture_shader as extension?
14:29 < Colbert> I was asking because I dont trust certain things with my current driver, mutlitexturing seems wrong for example
14:30 < Colbert> yep, GL_NV_texture_shader, GL_NV_texture_shader2, GL_NV_texture_shader3
14:30 < pmdata> woot
14:31 < pmdata> I think 908,90c is nv_texture_shader, but I must write a test function to be sure
14:33 < Colbert> got to take the dogs out,brb
14:33  * Colbert is back
14:33  * Colbert is back
14:33  * Colbert is away: Colbert
14:35 < pmdata> oops, I made the output combiner wrong for nv20 and nv30
14:47 -!- marcheu_ [n=marcheu@AStrasbourg-252-1-113-30.w90-6.abo.wanadoo.fr] has joined #nouveau
14:55  * Colbert is back
15:25 < pmdata> I can't find how to use *_BLEND_COLOR, any hint?
15:27 < pmdata> ah, glblendcolor, /me dump
15:27 < pmdata> dumb
15:56 -!- Lumag [n=Lumag@chimpanzee.school.ioffe.ru] has left #nouveau []
15:59 -!- marcheu_ [n=marcheu@AStrasbourg-252-1-113-30.w90-6.abo.wanadoo.fr] has quit ["leaving"]
16:20 -!- EdB|w [n=EdB@212.234.68.206] has joined #nouveau
16:40 -!- megaloutre [n=megalout@modemcable040.100-70-69.mc.videotron.ca] has joined #nouveau
16:59 < Colbert> pmdata do you have the time to look at another dump?
16:59 < pmdata> yep, what stuff?
17:00 < pmdata> grab renouveau cvs before
17:00 < Colbert> I  removed the second color change and just dumped on the fron buffer with double buffer off and then on
17:01 < Colbert> http://rafb.net/paste/results/sVPh3332.html
17:02 < Colbert> http://rafb.net/paste/results/eravw395.html
17:02 < pmdata> could you also make a dump of your glxinfo?
17:04 < pmdata> tell me what you would expect from these dumps
17:05 < Colbert> http://rafb.net/paste/results/mtbU2O97.html
17:07 -!- johill_ [n=johannes@p5487D207.dip.t-dialin.net] has joined #nouveau
17:08 < pmdata> I can't guess what you want to search
17:09 < Colbert> 2df   0x00000000   0x0014220c             {size: 0x5   channel: 0x1   obj: beef3097 opcode: METHOD }
17:09 < Colbert> 2e0   0x00000000   0x0a000a00            NV30_TCL_PRIMITIVE_3D      [0x020c/4] = 0x0a000a00 | UNKNOWN = 0a000a00
17:09 < Colbert> 2e1   0x00000000   0x004b4000            NV30_TCL_PRIMITIVE_3D      [0x0210/4] = 0x004b4000 | UNKNOWN = 004b4000
17:09 < Colbert> 2e2   0x00000000   0x07d28000            NV30_TCL_PRIMITIVE_3D      [0x0214/4] = 0x07d28000 | UNKNOWN = 07d28000
17:09 < Colbert> 2e3   0x00000000   0x004b4000            NV30_TCL_PRIMITIVE_3D      [0x0218/4] = 0x004b4000 | UNKNOWN = 004b4000
17:09 < Colbert> 2e4   0x00000000   0x00000a00            NV30_TCL_PRIMITIVE_3D      [0x021c/4] = 0x00000a00 | UNKNOWN = 00000a00
17:09 < Colbert> 2e5   0x00000000   0x0008222c             {size: 0x2   channel: 0x1   obj: beef3097 opcode: METHOD }
17:09 < Colbert> 2e6   0x00000000   0x00000280            NV30_TCL_PRIMITIVE_3D      [0x022c/4] = 0x00000280 | UNKNOWN = 00000280
17:09 < Colbert> 2e7   0x00000000   0x07cc4000            NV30_TCL_PRIMITIVE_3D      [0x0230/4] = 0x07cc4000 | UNKNOWN = 07cc4000
17:10 < Colbert> these seem to change between the runs
17:13 < Colbert> 381   0x00000000   0x00000000            NV30_TCL_PRIMITIVE_3D      [0x1434/4] = 0x00000000
17:13 < pmdata> 20c seems to be the pitch for 2 buffers
17:13 < pmdata> 210,214,218 seems to be offset in memory (maybe also 21c)
17:13 < pmdata> 230 seems to also be some offset
17:14 < pmdata> if you try a glClear() command, you should know which buffer (color, stencil/depth, or other) is at which offset
17:15 < pmdata> don't forget at least a small rendering with tri() between clears, or the driver will only make the last clear before rendering
17:15 < Colbert> should i change my code to do that?
17:16 < pmdata> try the test_clear() command
17:16 < Colbert> i recompiled between the dumps, so i dont know if i need to do the tri()
17:18 < Colbert> how can i see the offset for the buffer?
17:18 < pmdata> find were/how the clear is done, you should see same offset as the dump you did there
17:20 < Colbert> this is where i always get lost, analyizing the dumps.
17:21 < pmdata> on my nv15, glClear() is done with NV04_GDI_RECTANGLE_TEXT object
17:22 < pmdata> and also a NV10_CONTEXT_SURFACES_2D is setup before
17:24 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has quit [Read error: 110 (Connection timed out)]
17:24 < Colbert> am I trying to anaylize something thats obvious to you and the others? or is this useful
17:25 < pmdata> well, can you make a dump of test_clear() ?
17:25 < Colbert> one moment
17:25 < pmdata> at least, we know which objects are used for a clear on nv10 hw, but I don't know if it is the same on nv20 or nv30
17:29 < Colbert> is test_clear new? dont see it
17:31 < pmdata> it is in tests.c, but not listed in main.c, you must add it
17:32 < pmdata> I fixed it in cvs if you want
17:32 < pmdata> .o0O0o.
17:36 < Colbert> never mind i found it
17:37 < Colbert> http://rafb.net/paste/results/9ulOy288.html
17:38 < Colbert> never noticed that before, Im duplicating effort already done
17:39 < pmdata> I don't think so, because on different hw, different object/commands may have been used
17:39 < pmdata> now we know it uses same stuff on nv30
17:40 < pmdata> the first glClear() is done on depth buffer
17:40 < pmdata> lines 551-552 you see the offset setup for the clear
17:41 < pmdata> lines 106-113 you should find same offset, which is the depth buffer offset set for the opengl context
17:42 < pmdata> I let you search some minutes :)
17:44 < Colbert> line 551-552 real line in an editor or the dump #
17:45 < Colbert> the NV10_CONTEXT_SURFACES_2D_OFFSET_SRC      right?
17:45 < pmdata> yep
17:45 < Colbert> ok
17:46 < Colbert> so my offset is 0x07d28000 
17:47 < pmdata> yep
17:47 < pmdata> so in lines 106-113, where is the depth buffer offset?
17:49 < Colbert> e2   0x00000000   0x07d28000            NV30_TCL_PRIMITIVE_3D      [0x0214/4] = 0x07d28000 | UNKNOWN = 07d28000
17:49 < Colbert> ok
17:50 < pmdata> \o/
17:51 < pmdata> the fact the command is named ' NV30_TCL_PRIMITIVE_3D      [0x0214/4]' means we don't know what it is
17:51 < pmdata> but now I will change it to depth_offset, and commit to cvs
17:52 < pmdata> I let you find the color buffer offset
17:52 < Colbert> ok, but i miss being smart
17:54 < Colbert> and I have a ton of catching up to on this before I'm going to see this quickly
17:55 < pmdata> I also started slowly
17:55 < EdB|w> yes and now you go more and more faster :o)
17:56 < pmdata> well, at least for nv10 tcl, what remains is hard, I hope it is not some quadro stuff I can't test
17:59 < pmdata> I added a test function for glColorMask()
17:59 < pmdata> nv10 does not support alpha component for it
18:01 < pmdata> colbert> so did you find the color buffer offset?
18:02 < Colbert> not yet
18:06 < Colbert> i thin
18:06 < Colbert> 64f   0x00000000   0x004b4000   NV10_CONTEXT_SURFACES_2D_OFFSET_SRC           = 0x004b4000
18:08 < pmdata> yep, the test_clear only clear depth, then color then depth and color
18:08 < Colbert>  2e1   0x00000000   0x004b4000            NV30_TCL_PRIMITIVE_3D      [0x0210/4] = 0x004b4000 | UNKNOWN = 004b4000
18:08 < pmdata> what about the [0x218/4] ?
18:09 < pmdata> it is same offset
18:10 < Colbert> it is the offset for the color buffer
18:10 < pmdata> do you run a stereo color buffer, or a double buffer?
18:11 < pmdata> either one is gl_left and the other gl_right, or you have gl_front and gl_back buffers
18:11 < Colbert> i think i had it set to double buffer 
18:12 < Colbert> in main.c
18:12 < pmdata> you did SDL_GL_SetAttribute(SDL_GL_DOUBLEBUFFER,1); ?
18:12 < Colbert> yep
18:13 < pmdata> then maybe they are stereo left & right buffers
18:13 < pmdata> anyway, I will mark them as color0 and color1
18:13 < Colbert> right fails, right now
18:13 < Colbert> get an error
18:14 < pmdata> hum, could you also try clearing stencil?
18:14 < Colbert> ok
18:14 < Colbert> do you want SDL_GL_DOUBLEBUFFER on or off
18:15 < pmdata> keep it on, hum, I don't know if sdl can create a stereo buffer
18:15 < Colbert> it wont
18:15 < Colbert> already tried, & not using nvclock -Q right now
18:18 < pmdata> hum maybe used for multisample or fsaa then
18:18 < pmdata> I also added the pitch values
18:19 < pmdata> you know what the pitch is?
18:23 < Colbert> http://rafb.net/paste/results/Lrgjre99.html
18:24 < Colbert> added a stencil clear to  test_clear() 
18:24 < Colbert> brb
18:25  * Colbert is away: Colbert
18:37  * Colbert is back
18:39 < Colbert> dont remember the term pitch, what is it?
18:43 < pmdata> the number of bytes per horizontal line in a buffer
18:57 < pmdata> colbert> I see from glxinfo output that you can try multisample stuff
18:58 < Colbert> the examples from glut dont seem to work, though. but those are sgi only i think
18:59 < Colbert> what do you have in mind?
19:00 < pmdata> maybe the extra color buffer is used for multisampling
19:00 < Colbert> hmm
19:01 < pmdata>     SDL_GL_MULTISAMPLEBUFFERS,
19:01 < pmdata>     SDL_GL_MULTISAMPLESAMPLES,
19:01 < pmdata> you can try to set those attributes before sdl_setvideomode
19:01 < Colbert> lets see
19:02 < pmdata> your system can support 1 buffer, and 1,2 or 4 samples
19:02  * pmdata wrong, 2 or 4 samples
19:02 < pmdata> and to note: I did not see anything new with the stencil clear, so it may resides in the depth buffer as for nv10
19:03 < pmdata> so try setting multisample stuff, and compare the glclear dumps
19:06 < Colbert> http://rafb.net/paste/results/rZJyCr52.html
19:07 < pmdata> hum, nothing new
19:08 < pmdata> you should try making 'diff -Naur' between the dumps
19:09 < Colbert> line 64 changed to 0x02000000 from 0x02000004
19:09 < pmdata> did not see that
19:09 < pmdata> how did you setup the multisample?
19:11 < Colbert> just added SDL_GL_SetAttribute(SDL_GL_MULTISAMPLEBUFFERS,1); SDL_GL_SetAttribute(SDL_GL_MULTISAMPLESAMPLES,1); below stencil in main.c
19:12 -!- johill_ [n=johannes@p5487D207.dip.t-dialin.net] has quit [Read error: 110 (Connection timed out)]
19:12 < pmdata> samples must be set to 2 or 4
19:13 < Colbert> hold on i think i pasted the wrong file
19:13 < pmdata> also update renouveau from cvs
19:14 < Colbert> http://rafb.net/paste/results/Y1XRrv95.html
19:14 < Colbert> sorry,
19:18 < Colbert> line 328 changed one bit
19:19 < pmdata> I thought of something, give that you have 2-sided stencil
19:19 < pmdata> it means there is an extra stencil buffer
19:19 < pmdata> stencil is stored in a byte-wise buffer
19:20 < Colbert> meant line 188 :328   0xffff0001   0xffff0001            NV30_TCL_PRIMITIVE_3D      [0x1d7c/4] = 0xffff0001 | UNKNOWN = ffff0001
19:20 < pmdata> and 22c is 0x200, and 230 is an offset, so maybe this is the other stencil buffer
19:20 < pmdata> you put 2 for multisamplessamples?
19:21 < Colbert> no, 1
19:21 < Colbert> shoul i try with 2
19:21 < pmdata> try using 2 then 4, and compare both dumps
19:23 < Colbert> nothing changes at 2
19:23 < Colbert> trying 4
19:24 < Colbert> well actually  -- program at 0x0xa688b000 - 114000 changed to -- program at 0x0xa6870000 - 114000
19:25 < pmdata> hm, program is at a different offset
19:25 < Colbert> should i change SDL_GL_MULTISAMPLEBUFFER also
19:26 < pmdata> no
19:26 < pmdata> we should find something else to work on
19:27 < pmdata> I wrote a test_color_mask(), could you run it?
19:28 < Colbert> mulisample 4 changed stuff
19:28 < pmdata> how much?
19:30 < Colbert> http://rafb.net/paste/results/LChOCc20.html
19:33 < Colbert> offset changed line 87, line 92 0x0230/4] = 0x07a3c000. MORE
19:33 < pmdata> checkout renouveau from cvs I added a test for 2sided stencil
19:34 < pmdata> the offset can change anytime, because the driver search for some free mem in the video card
19:40 < Colbert> NV04_GDI_RECTANGLE_TEXT_BLOCK_LEVEL1_BR chaned = right = 1024 | bottom = 51 to  = right = 1024 | bottom = 1024
19:40 < Colbert> doubled the size
19:41 < Colbert> checking out cvs
19:49 -!- _Demo_ [n=Demo@modemcable206.154-131-66.mc.videotron.ca] has joined #nouveau
19:49 < Colbert> have conflicts during merge
19:50 < Colbert> going to just co again
19:51 < Colbert> my mother just garbage picked a broken led monitor for me, new toys! ;)
19:52 < pmdata> I would like to find a nv2x agp video card to test nv20 stuff
19:53 < Colbert> Me too, probably faster than this 5200
19:54 -!- marcheu [i=marchesi@pellinore.u-strasbg.fr] has quit ["leaving"]
19:54 -!- marcheu [n=marcheu@lattice.u-strasbg.fr] has joined #nouveau
19:54 -!- EdB|w [n=EdB@212.234.68.206] has quit [Read error: 104 (Connection reset by peer)]
19:55 < pmdata> colbert> so can you try color_mask or ext_stencil_two_side?
19:55 < Colbert> ok
19:58 < Colbert> http://rafb.net/paste/results/xFiIzA56.html
19:58 < Colbert> for ext_stencil_two_side, color mask coming up
20:02 < Colbert> http://rafb.net/paste/results/n2ulzd15.html
20:07 < pmdata> it seems the 2 stencil buffers reside at same offset
20:11 < pmdata> anyway I renamed stencil0 as stencil_back and stencil1 as stencil_front
20:12 < Colbert> in my xorg.conf  Option         "TripleBuffer"               "true".
20:13 < Colbert> this is probably causing this. didnt remember doing that
20:13 < Colbert> brb
20:13 -!- Colbert [n=colby@adsl-69-211-27-8.dsl.chcgil.ameritech.net] has left #nouveau []
20:16 -!- EdB [n=EdB@ARennes-251-1-39-55.w81-250.abo.wanadoo.fr] has joined #nouveau
20:18 -!- Colbert [n=colby@adsl-69-211-27-8.dsl.chcgil.ameritech.net] has joined #nouveau
20:18 < Colbert> pmdata; Im going to run those again
20:21 < pmdata> also checkout renouveau cvs, I updated stuff
20:21 < Colbert> color_mask
20:21 < Colbert> http://rafb.net/paste/results/TiNlwx56.html
20:22 < pmdata> is the window still 512x512?
20:23 -!- EdB [n=EdB@ARennes-251-1-39-55.w81-250.abo.wanadoo.fr] has quit [Remote closed the connection]
20:23 -!- EdB [n=EdB@ARennes-251-1-39-55.w81-250.abo.wanadoo.fr] has joined #nouveau
20:25 < Colbert> http://rafb.net/paste/results/UMvwfU13.html
20:27 < pmdata> oops, I made a mistake
20:27 < Colbert> still 512x512
20:36 -!- EdB [n=EdB@ARennes-251-1-39-55.w81-250.abo.wanadoo.fr] has quit [Remote closed the connection]
20:37 -!- EdB [n=EdB@ARennes-251-1-39-55.w81-250.abo.wanadoo.fr] has joined #nouveau
20:52 < Colbert> Hmmm, lost some visuals with the change to xorg.conf
20:53 < Colbert> think I'll add AddARGBGLXVisuals back in
20:55 < pmdata> can you checkout cvs and test the new test_depth_range() ?
21:02 < Colbert> cant find it. going to update cvs again
21:07 < Colbert> http://rafb.net/paste/results/fTsmOB91.html
21:07 < Colbert> there's the dump 
21:13 < pmdata> did you find the depth range?
21:13 < pmdata> or I should have a look? :)
21:14 < Colbert> sorry had to go outside to get my dogs,
21:15 < pmdata> anyway, it just confirms what I was thinking, 394-398
21:21 < Colbert> what were you looking for?
21:25 < pmdata> the depth range, I already seen these when I compared against nv10, I just wanted to be sure
21:26 -!- johill_ [n=johannes@p5487D207.dip.t-dialin.net] has joined #nouveau
21:28 < Colbert> i'm fading here, bbl
21:29  * Colbert is away: Colbert
21:41 -!- pmdata is now known as pmdata_miam
21:42 < pmdata_miam> enough commits for  today
21:47 -!- johill_ [n=johannes@p5487D207.dip.t-dialin.net] has quit [Remote closed the connection]
22:12 -!- pmdata_miam is now known as pmdata
22:32 -!- pmdata is now known as pmdata_tv
--- Log closed Вск Июл 23 00:00:56 2006
