IRC log for #brlcad on 20090718

00:04.25Ralithhm, I guess there are transparency issues.
00:04.27Ralithdamn.
00:06.11brlcadthe "pop-up" one isn't really a console as much as it is a interface to quick actions
00:07.54brlcadwhere the command line console is like mged's prompt where you can interact with the system, see results returned, script things out, etc; the on-demand interface is more for stateless quick commands "display object34", "help on rotations", "open http://brlcad.org", etc
00:08.19CIA-30BRL-CAD: 03ralith * r35183 10/rt^3/trunk/src/g3d/mainwindow.ui: Added a text label on top of the GL context, causing garbage to appear in transparent areas, a bug almost certainly related to the underlying OpenGL context.
00:08.48brlcadfor starters, it could be a simple one-shot command interface like the console, but the long-term intent is more of a hot-key command-line
00:09.25Ralithencountered our first (and probably only) bug resulting from drawing on top of, rather than into, Ogre.
00:09.31Ralithpretty ugly one, though
00:09.43brlcadoh?
00:10.02``Erikpix? :D
00:10.16Ralithhopefully avoidable by not using anything that does partial transparency directly on top of the context.
00:10.22Ralith(in this case, a bit of antialiased text)
00:10.24Ralith``Erik: sure, sec
00:11.07``Erik(btw, weirdnesses like that tend to be platform specific... qt can help, but it ain't a silver bullet... :)
00:11.21Ralithnah this is an OpenGL issue
00:11.24Ralithnothing platformy
00:11.26``Erikum
00:11.32``Erikopengl issues tend to be platformy :D
00:11.38Ralithalthough different video hardware might indeed have different effects
00:11.46``Erikit's the ogl context going weird against the display context
00:11.46Ralith``Erik: this is an opengl issue that's inherently opengly, and which I predicted :P
00:12.00Raliththat is to say, Qt can't do anything about it when used this way
00:12.09Ralithand I doubt any platform can either
00:12.15``Erikunder the hood, it's glueing itself with glxMakeContext() or wglMakeContext() or something
00:12.17Ralithperhaps a composited WM might be a special case
00:12.18``Erikiirc
00:12.20``Erikattachcontext?
00:12.24``Erikbeen a long time
00:13.12Ralithhttp://isitloadedyet.com/~ralith/gl_weirdness.png
00:13.27brlcadRalith: you should just keep a screenshot hot-key handy, we're just going to keep saying "pics!" ;)
00:13.29``Erikeven the venerable SDL had platform specific weirdness when things like ogl were in the picture :) *shrug*
00:13.44brlcadvenerable?!
00:13.47brlcadheh
00:13.57Ralithstill, this makes me want to go back and try the QGraphicsView approach again
00:13.59``ErikHEY!
00:14.02``Erikhugs his sdl
00:14.13brlcadremember fighting SDL to no end to make it behave when we ported BZ to it, had to sort out a half-dozen bugs
00:14.26``Erikneat, when was that?
00:14.43brlcadlike five years ago or something
00:14.56``Erikwas coding ground up with sdl in '98-02, didn't try porting existing stff to it
00:17.51``Erikso, yeah, I hadn't made any progress on swiss cheese, something in the plate negatives is screwed up. indianlarry mentioned something about some of those having air codes
00:19.34Ralithooh
00:19.36Ralithrich text labelling
00:19.57Ralithfantasizes about log messages with clickable object names
00:22.54Ralithyep
00:22.59Ralithit was partial transparency that was responsible
00:23.15Ralithgiving the label a solid-fill background killed the error
00:40.53brlcadthe painting algorithm does have some bug in it
00:40.58brlcador the things that call the painting do
00:41.05brlcadbecause it is exceeding that index
00:41.26brlcadwhich shouldn't be happening, I believe that's a count of how many edges are comging together iirc
00:41.34brlcad256 is nonsense
00:42.18``Erikgiven that it's an unsigned char, 256 is a big wtf o.O
00:42.39brlcadwell, that too, but the fact that it'd count anywhere near that high
00:44.04brlcadthinks he's going to pop in Ghost in the Shell 2 here in a few minutes, party at my house
00:44.11``ErikI told russ that my guess was that it's a longstanding bug that we haven't been able to get to :/
00:44.51``Erikheh
00:45.54``Erikstacking the crap at the end of my driveway got my wrist hurtin', I'm not gonna go move it again :D I'm trapped for the night
00:46.07``Erik</whine>
00:48.41brlcad~cheese ``Erik
00:48.42ibotACTION hands ``Erik a tray of sliced sharp cheddar and pineapple
00:49.32``Erikheh
00:50.27brlcadthis week was so off-kilter that I can't even bring myself to a glass of scotch
00:50.58brlcadsips some irish whiskey instead
00:52.51``Eriksips his cranberry juice o.O
00:52.53``Eriktop THAT! HA!
00:54.38brlcaddrops doggie poop in the juice, there topped it!
00:59.33CIA-30BRL-CAD: 03ralith * r35184 10/rt^3/trunk/src/g3d/ (5 files): Installed an intermediary MainWindow class on top of the autogenerated Ui::MainWindow to allow functionality to be attached to the GUI.
01:30.41Ralithwow
01:30.41Ralithabstracting the console out into its own widget was remarkably easy
01:30.41RalithQt was a great choice :]
01:30.41CIA-30BRL-CAD: 03ralith * r35185 10/rt^3/trunk/src/g3d/ (CMakeLists.txt Console.cxx Console.h mainwindow.ui): Made the console its own widget.
01:32.48``Erikhttp://www.collegehumor.com/video:1917186  we totally need EM trains here
01:45.15Ralithargh
01:45.24RalithI guess the GL error doesn't just manifest when you use text like that :|
01:47.15CIA-30BRL-CAD: 03ralith * r35186 10/rt^3/trunk/src/g3d/ (Console.cxx Console.h): Added autohiding of the console output label. Strangely, this seems to have caused the OpenGL error formerly assumed to be related to partial transparency to return.
01:48.01starseekerRalith: you think the QGraphicsView approach would avoid those issues if it could be made to work with Ogre?
01:48.14Ralithstarseeker: I'm just about certain it would.
01:48.29Ralithconsidering that the modelview example used 80% transparent everything.
01:52.39Ralithbrbs
02:03.03Ralithunbrbs
02:10.23starseekersobs as he sees that the "improved" Apollo 11 footage is not the original tapes'
02:10.45starseekergood to seem them preserving what they can, but ARRRRRRRRGH
02:11.12starseekersomething only possible at a large buerocracy
02:11.35starseekerreuse best quality tapes of historic human milestone to save a few bucks
02:20.02starseekeroddly fitting monument to America's reaction to space exploration post-moon-landing, though...
02:31.06starseekerdownloads the restored copies anyway...
02:40.59CIA-30BRL-CAD: 03ralith * r35187 10/rt^3/trunk/src/g3d/ (Console.cxx Console.h): Improved console output autohide: output can now be moused over, and the OpenGL error has once again dissapeared.
02:50.21CIA-30BRL-CAD: 03ralith * r35188 10/rt^3/trunk/src/g3d/Console.cxx: Removed unnecessary margin and spacing (which tended to not get redrawn, leading to uglyness) from Console's layout.
02:55.02CIA-30BRL-CAD: 03ralith * r35189 10/rt^3/trunk/src/g3d/mainwindow.ui: Removed some more unnecessary margin, this time from the options bar along the top.
03:22.48CIA-30BRL-CAD: 03bharder * r35190 10/brlcad/trunk/AUTHORS: reflect name change, incorporation
03:32.43Ralithbrlcad: at what point should I s/ogretest/g3d/?
03:32.56Ralithwhen I reach feature-equivalence?
03:33.09*** join/#brlcad stevegt_1 (n=stevegt@229.sub-75-211-192.myvzw.com)
03:33.14Ralithor now, since it's no longer just a testcase?
04:07.57CIA-30BRL-CAD: 03ralith * r35191 10/rt^3/trunk/src/g3d/Console.cxx:
04:07.59CIA-30BRL-CAD: Made Console output text selectable/copyable (and any links that might appear in
04:08.01CIA-30BRL-CAD: the future clickable). This should probably be done in any case where a user
04:08.03CIA-30BRL-CAD: might conceivably want to, for example, copy an error message into a request for
04:08.05CIA-30BRL-CAD: help or bug report.
05:27.12Ralithokay, at this point the GUI is tweaked such that the visibility of that OpenGL bug is minor
05:27.23Ralitheven though it keeps recurring
05:33.53Ralithstarseeker: something possibly of interest:
05:34.06Ralith"Qt normally erases the widget's area before the paintEvent() call. If the WRepaintNoErase widget flag is set, the widget is responsible for painting all its pixels itself."
05:34.30Ralithi.e. setting WRepaintNoErase may be a good idea
05:56.14CIA-30BRL-CAD: 03ralith * r35192 10/rt^3/trunk/src/g3d/OgreGLWidget.cxx: Ensure that Ogre rendering occurs frequently (targetting 100fps).
06:50.55*** join/#brlcad docelic (n=docelic@78.134.200.176)
07:57.04*** join/#brlcad stevegt_ (n=stevegt@c-24-130-122-25.hsd1.ca.comcast.net)
07:59.07*** join/#brlcad stevegt_2 (n=stevegt@c-24-130-122-25.hsd1.ca.comcast.net)
08:48.08CIA-30BRL-CAD: 03ebautu * r35193 10/web/trunk/htdocs/more/sites/all/modules/brlcad/ (3 files in 2 dirs): Improved php-mged interface (all metadata extraction and raytracing is done in a single call to mged)
08:51.23CIA-30BRL-CAD: 03Ebautu 07http://brlcad.org * r1573 10/wiki/More_Changelog: July 15-17
08:53.18*** join/#brlcad stevegt_ (n=stevegt@c-24-130-122-25.hsd1.ca.comcast.net)
09:21.36*** join/#brlcad _sushi_ (n=_sushi_@84-73-207-71.dclient.hispeed.ch)
09:35.39``Erik*yawn*
11:51.43``Eriknifty, got X back up on the 'hack' box
12:53.10*** join/#brlcad docelic (n=docelic@78.134.196.35)
13:50.28*** join/#brlcad Don_ (n=Don@71.238.51.148)
14:37.57CIA-30BRL-CAD: 03ebautu * r35194 10/web/trunk/htdocs/more/sites/all/modules/multicolumn_select/ (. multicolumn_select.info multicolumn_select.module): Multicolumn_select - custom developed module for BRLCAD (used for selecting the object to be rendered)
14:54.00CIA-30BRL-CAD: 03ebautu * r35195 10/web/trunk/htdocs/more/sites/all/modules/brlcad/brlcad.module: Allow users to select which object to render.
15:42.14``Erikcompiles ogre3d on his fbsd box O.o
15:42.17CIA-30BRL-CAD: 03ebautu * r35196 10/web/trunk/htdocs/more/sites/all/modules/brlcad/brlcad.module: Queue model for BRLCAD processing when the user changes the object to render.
15:43.56*** join/#brlcad hippieindamakin8 (n=hippiein@210.212.55.3)
15:46.10CIA-30BRL-CAD: 03ebautu * r35197 10/web/trunk/htdocs/more/sites/all/modules/imagefield/ (imagefield.info imagefield.module): Allow Node ID to be used in file paths.
15:47.20CIA-30BRL-CAD: 03ebautu * r35198 10/web/trunk/htdocs/more/sites/all/modules/filefield/filefield.module: Allow Node ID to be used in file paths.
15:50.21*** join/#brlcad BigAToo (n=BigAToo@pool-96-230-124-27.sbndin.btas.verizon.net)
15:52.13CIA-30BRL-CAD: 03erikgreenwald * r35199 10/brlcad/trunk/src/adrt/libtie/tie.c: avoid puking of initial triangle buffer is 0 bytes
16:05.54CIA-30BRL-CAD: 03n_reed * r35200 10/brlcad/trunk/ (include/dm-rtgl.h src/libdm/dm-rtgl.c): storing dm-rtgl points in vertex arrays to improve performance
16:22.56*** join/#brlcad SWPadnos_ (n=Me@dsl107.esjtvtli.sover.net)
16:46.35*** join/#brlcad stevegt_ (n=stevegt@167.sub-75-211-20.myvzw.com)
18:24.39CIA-30BRL-CAD: 03johnranderson * r35201 10/brlcad/trunk/src/librt/primitives/bot/ (bot.c g_bot_include.c):
18:24.44CIA-30BRL-CAD: prep method will now fail if an index for a vertex is out of range.
18:24.50CIA-30BRL-CAD: describe method now points out faces that have vertex indices out of range.
18:24.54CIA-30BRL-CAD: This fixes bug ID: 1592074 "Bot does not raytrace"
19:34.51CIA-30BRL-CAD: 03Ebautu 07http://brlcad.org * r1574 10/wiki/More_Changelog: July 18
21:16.41*** join/#brlcad stevegt_ (n=stevegt@239.sub-75-210-95.myvzw.com)
21:23.20*** join/#brlcad BigAToo (n=BigAToo@pool-96-230-124-27.sbndin.btas.verizon.net)
22:36.18*** join/#brlcad Patmcc19 (n=chatzill@71-223-26-58.phnx.qwest.net)
23:39.12*** join/#brlcad BigAToo (n=BigAToo@pool-96-230-124-27.sbndin.btas.verizon.net)

Generated by irclog2html.pl Modified by Tim Riker to work with infobot.