irclog2html for #brlcad on 20100114

02:31.16starseekerwell, at least the tk framebuffer works on gentoo...
02:36.42*** join/#brlcad akafubu (n=akafubu@unaffiliated/akafubu)
02:47.40*** join/#brlcad R0b0t1 (n=Enigma@unaffiliated/r0b0t1)
03:11.00brlcadhmm.. interesting, http://www.explain.com.au/oss/docbook/
07:18.11brlcadlooks like the shaders test has been broken for more than 3000 commits
09:12.47*** join/#brlcad PrezKennedyII (i=Matthew@whitecalf.net)
10:48.41*** join/#brlcad mafm (n=mafm@79.Red-83-37-7.dynamicIP.rima-tde.net)
13:03.58*** join/#brlcad mafm_ (n=mafm@79.Red-83-37-7.dynamicIP.rima-tde.net)
13:48.45``Erikneat
14:32.39CIA-38BRL-CAD: 03irpguardian * r37249 10/brlcad/trunk/src/rt/view.c:
14:32.39CIA-38BRL-CAD: Made the heat-graph make an inverted image for testing of values
14:32.39CIA-38BRL-CAD: that need to be changed.
14:57.21*** join/#brlcad archivist (n=archivis@87-194-141-154.bethere.co.uk)
15:40.46*** join/#brlcad archivist (n=archivis@87-194-141-154.bethere.co.uk)
15:41.18starseekerthe Tk_Canvas call doesn't apply - can't find a canvas object in this mix anywhere
15:41.50starseekerthe photo stuff is all hidden - Tk_ImageChanged I can't call directly, and presumably it's being called anyway
15:42.18starseekerthe 8.6 build for tcl/tk seems to have changed significantly
15:42.30brlcadif you can get at the Tk_Window, I believe you can get at the canvas
15:42.33brlcadthrough another call
15:42.51brlcadwe use the tkwin all over the place, so should be derivable
15:43.15brlcador maybe not, but something to check
15:43.42starseekeris it in the libs?  don't see Canvas in mged
15:47.17starseekeroh wait, .fb_tk_canvas
15:47.24starseekermaybe I can use that
15:54.34starseekergrrrrrrrr
15:54.52starseekerno obvious way to get at the canvas object in C land
15:59.45brlcadwhat did you need the canvas for?
16:00.35brlcadif_tk creates a canvas via tcl, and manages it that way
16:01.00brlcadif you need a C hook, you'll probably have to replace those calls with the corresponding C
16:01.39brlcadwhich probably means reading the source of the tk 'canvas' command to see what it's doing during Tcl_Eval
16:04.36starseekerI'm trying to find an explicit refresh command I can invoke from this level, but the sense I'm getting is that I'm not supposed to be doing that
16:05.05brlcadah
16:05.28starseekerthe DoOneEvent call SHOULD work, and on X11 systems does work
16:07.32starseekerwhat do you think - I'm a little leery of dumping a ton of work up front into 8.6 tcl/tk when it may or may not work - might be better to finish our parts of the tcl/tk coding using the X11 systems and then it's ready when Tcl/Tk is ready
16:11.59brlcadmm.. http://books.google.com/books?id=X8TT0W7Wo0sC&lpg=PA315&ots=R6n0UWCQ_9&dq=tk%20refresh%20canvas%20-perl%20-ruby&pg=PA315#v=onepage&q=tk%20refresh%20canvas%20-perl%20-ruby&f=false
16:12.10brlcadI think I have my copy  of that book somewhere around there
16:12.30brlcadsure, get it working on X11
16:12.36brlcadshould be able to verify it works on Windows too
16:12.51brlcadif it works on both of those, that's a pretty safe bet that it's just some aquatk issue
16:13.14starseekerI'll toss an email to the tcl-mac list, see if anyone has tried this type of low level stuff with Aqua 8.6
16:13.41brlcad34000 == working && 34120 == broken .. lil closer
16:14.06starseekershould get 'em thinking about it, if nothing else
16:14.08starseekercool!
16:16.55CIA-38BRL-CAD: 03erikgreenwald * r37250 10/brlcad/trunk/src/librt/primitives/ (metaball/metaball_tri.c nmg/nmg_tri_mc.c): some (untested) edge solving for marching cubes
16:16.57brlcadjust goes to show, the longer a bug goes unfixed, the more expensive it is to fix it
16:17.10``Erik"technical debt", wee
16:17.20``Erikjust goes to show, cron that shit :D
16:17.23brlcadit was very likely some minor "we'll get to that later" issue
16:18.30brlcadit's not been a matter of the test not running, cron wouldn't have helped
16:19.10brlcadit's been running, regression suite has been failing on it since at least March 2009
16:19.41brlcadand we've had at least a half-dozen releases since then with everyone running distcheck actively ignoring it
16:20.24``Erikregress isn't mentioned in the release procedure in HACKING
16:20.42``Erikdaily emails from a cron woulda annoyed someone into fixing it back in march :D
16:20.47brlcaddistcheck runs regress
16:26.15*** join/#brlcad talcite (n=matthew@dhcp-143-147.mcme-students.carleton.ca)
16:28.48CIA-38BRL-CAD: 03brlcad * r37251 10/brlcad/trunk/NEWS:
16:28.48CIA-38BRL-CAD: bob added support to the bot_dump command to allow combinations to be specified.
16:28.48CIA-38BRL-CAD: the combination is walked and all bots encountered are exported. (don't know
16:28.48CIA-38BRL-CAD: what it'll do with subtractions and intersects..but will probably export the
16:28.48CIA-38BRL-CAD: leaves as-is)
17:08.49brlcadstarseeker: is the display manager toggling stable now?
17:33.37*** join/#brlcad talcite_ (n=matthew@dhcp-143-147.mcme-students.carleton.ca)
17:38.50brlcadhuh, sourceforge acquired ohloh last summer
17:38.54brlcadmissed that tid bit
17:39.10brlcadmigrated to sf.net infrastructure just a couple months ago
17:52.47brlcaddamnits, had the range wrong, off by 500
17:58.36*** join/#brlcad akafubu (n=akafubu@unaffiliated/akafubu)
18:17.16*** join/#brlcad Elrohir (n=kvirc@p5B149393.dip.t-dialin.net)
18:20.44``Erikfind where starseeker broke the test yet? :D
18:21.10brlcadnot yet, still narrowing
18:21.34``Erik<-- did some Ren-esque fingerpointing and saying "YOUUUUUUuuuuu"
18:21.57brlcadgetting faster to update and retest the closer I get, fewer commits
18:22.58brlcadreally don't want to pick up shop and head in till I finish, though .. rather infuriating to not pinpoint it more quickly
18:23.19``Erikfor rev in `jot 20 34000` ; do xterm -c "svn co -r$REV https://... rev$rev && cd $rev && autogen.sh && ..." ;d one
18:23.22``Erik:D
18:24.23brlcadnot quite that simple
18:25.02brlcadsome revs require .Plo file edits, others are conflicted updates or require minor edits to compile
18:25.11``Erikeck
18:26.26brlcadnarrowed back down to a range of 50 commits
18:31.27starseekerbrlcad: that toggling should be stable
18:31.37starseekerif it breaks I'd like to hear about it
18:31.51starseekernot tried on Windows, obviously
18:32.32brlcadstarseeker: reason I ask is news file
18:32.37brlcadnever got an entry
18:33.14starseekeroh, ooops
18:33.24starseekerwant me to get it?
18:33.45brlcadnaturally :)
18:36.16CIA-38BRL-CAD: 03starseeker * r37252 10/brlcad/trunk/NEWS: Add support for runtime toggling between different display managers in MGED.
18:40.40*** join/#brlcad Ralith (n=ralith@69.90.48.97)
19:26.18brlcad*drumroll*
19:28.31brlcadstarseeker: you're safe!
19:28.41brlcadheh, it was old faithful ;)
19:29.21CIA-38BRL-CAD: 03irpguardian * r37253 10/brlcad/trunk/src/rt/view.c:
19:29.21CIA-38BRL-CAD: Now able to pull current working pixel, which will eventually lead to file/array
19:29.21CIA-38BRL-CAD: input for time graph.
19:30.38brlcad``Erik: you're good at whining .. and since I can't get him an e-mail...
19:31.23brlcadusing the wrong brace style, c++/c99-style comments, and should have a space after his astericks in block comments
19:38.06``ErikI let him know, he's fixing things up and looking into the email thing (he hadn't tried sending one to his sf account, he's forwarding it to gmail supposedly)
19:38.21``Erik(good at whining? why I oughta....
19:41.22CIA-38BRL-CAD: 03irpguardian * r37254 10/brlcad/trunk/src/rt/view.c: Various changes to comments / blocking style to make it more 'right'
19:52.42``Erik(is there anything he'd have to do special to get M-x butterfly-mode for indenting?)
19:56.10CIA-38BRL-CAD: 03brlcad * r37255 10/brlcad/trunk/src/libged/ged.c:
19:56.10CIA-38BRL-CAD: either the math is wrong in _ged_mat_aet(), or there is some bad initialization
19:56.10CIA-38BRL-CAD: juju in here. don't call _ged_mat_aet() to set the rotation matrix here
19:56.10CIA-38BRL-CAD: otherwise we end up horking the shaders.sh regression test. this gets the
19:56.11CIA-38BRL-CAD: regression suite passing again, busted after r33883.
19:56.35``ErikO.O
19:56.46``ErikI thought you said it worked in r34000
20:03.31CIA-38BRL-CAD: 03brlcad * r37256 10/brlcad/trunk/regress/ (mged_test.sh repository.sh): remove trailing ws
20:04.42CIA-38BRL-CAD: 03brlcad * r37257 10/brlcad/trunk/src/libged/vutil.c: document that _ged_mat_aet() needs to be investigated.
20:05.59brlcadthen I later said I went the wrong way
20:06.03brlcadoff by 500
20:06.05``Erikoh, missed that :D
20:06.51``Erikirpguardian asked if you're happy now... (looks like excessive whitespace... ")  {"?)
20:07.31brlcadsubdivided by 2k's, then 1k's, then 500, then .. went the wrong way all the way down to about 10 commits before realzing it
20:08.07brlcadif i'm happy now?
20:08.18brlcadahh
20:08.20``Erikyeh, r37254, he tried fixing the stuff
20:08.54brlcadit's better, but incomplete
20:09.41``Erikwhat's the emacs equivalent of ggvG= ?
20:09.48``Erik(or :%!indent)
20:10.03brlcadM-x indent-region
20:10.14``Erikis there an indent-file
20:10.15``Erik?
20:10.31brlcadit uses a built-in style that are footer defines
20:10.37``Erikmeh, region should be good 'nuff, I'll go tell him about it
20:10.43brlcaddoesn't redo braces, it's just basic indentation setup
20:11.29starseekergetting XAllocColor and other such things being undefined when I try the 8.6 beta in aqua, but NOT when I do it with X11... see stub definitions in the macosx directory that must not be getting built... may have to platform conditionalize the Makefile.am, looks like the unix dir won't cut it on macosx any more in the non-X11 case
20:12.10brlcad``Erik: the problem is more that he's following the existing style in the file, that file hasn't been cleaned up yet
20:12.21``Erikok, he's dorking with figuring out how to mark now
20:12.31``Erikah, 'k, I'll tell him that
20:14.19``Erikpointed him to the hacking file, *shrug* :)
20:18.42starseekeryeah, compile made it to the docs stage when using X11 - gonna have to conditinally go for macosx when doign the aqua build
20:22.36brlcadctrl-space is mark
20:22.43CIA-38BRL-CAD: 03brlcad * r37258 10/brlcad/trunk/src/rt/view.c: style cleanup. k&r brace style, eliminate space paddings, and more ws. style and comment consistency cleanup.
20:23.54brlcadnow it's clean
20:24.47starseekergrowl
20:24.48``Erikhe's updated and going now
20:25.10starseekerget the Tcl_WaitForEvent: CFRunLoop finished behavor even with X11 on 8.6 beta...
20:33.49``Erikwow, the tk fb works
20:36.47``Erik(middle and right click don't work in the tk fb)
20:36.59starseekerno surprise
20:37.32``Erikjust sharing my findings :) (or, "whining", according to some)
20:37.49starseekerwhat's right click supposed to do?
20:38.08``Erikclose the window
20:38.13starseekerah, k
20:38.16``Erikmiddle click gives you x,y,r,g,b
20:38.18starseekerjust need to add another binding
20:38.48starseekertake a look at if_tk.c as compared to the other if_*.c files, it looks... a bit short at the moment :-P
20:39.19``ErikI'd hope it'd be significantly shorter than the others even when it's complete... :D TK abstracts a lot of the footwork...
20:39.31``Erikif_sdl.c might be even shorter still O.O
20:39.35starseekeroh, absolutely
20:40.13starseekerthat's one of the main reasons to hope that the tk framebuffer can become the "one true MGED/Archer framebuffer" - big old code simplicifation
20:40.31starseekerbut right now there are a lot of stub functions with nothin ;-)
21:01.08*** join/#brlcad talcite_ (n=matthew@dhcp-143-147.mcme-students.carleton.ca)
21:07.51CIA-38BRL-CAD: 03erikgreenwald * r37259 10/brlcad/trunk/src/librt/primitives/metaball/metaball_tri.c: document the vertex/edge topology of the cube. Return the error code on failure (instead of -1)
21:24.44starseekeruh...
21:26.06brlcadgcc/g++ should work on mac
21:28.06starseekerI'm getting a trigger of #if !__OBJC__
21:29.04starseekerweird
21:29.16starseekergcc does have objc support, according to gcc -v
21:29.41brlcadif !defined(__OBJC__)
21:31.06starseekerstill triggering
21:31.15starseekerand followed immediately by a slew of parse errors
21:31.20starseekerin all cases
21:32.05starseekerah, well
21:35.08CIA-38BRL-CAD: 03brlcad * r37260 10/brlcad/trunk/src/libged/gqa.c:
21:35.08CIA-38BRL-CAD: fix gqa crash on 64-bit linux where strsep() is not defined (due to strict
21:35.08CIA-38BRL-CAD: compilation, it's a bsd extension). the 32-bit int return value getting cast to
21:35.08CIA-38BRL-CAD: a 64-bit pointer was badness. instead rewrite the units parsing to use strtok()
21:35.09CIA-38BRL-CAD: since it's c89, also taking the opportunity to get rid of the unnecessary goto
21:35.11CIA-38BRL-CAD: logic. while we're at it, take care of fstat/fileno warnings along with a
21:35.13CIA-38BRL-CAD: strdup() -> bu_strdup() fix.
21:37.09CIA-38BRL-CAD: 03brlcad * r37261 10/brlcad/trunk/NEWS: diagnosed and fixed a bug in 64-bit linux where strsep getting run without a prototype was causing a segfault (due to 32-bit int to 64-bit pointer conversion). observed on RHEL5.
21:37.15``Erikum, linux doesn't ship gobjc by default, it's an addon package for most distros
21:37.20``Eriksame with bsd
21:37.46``Erikbut it is part of the gcc suite and the gcc frontend can point towards it with the right flags, I think :)
21:38.04``Erikautomake should see .m and assume it needs to run gobjc instead of gcc
21:38.21brlcadthey build it in on the mac
21:38.38brlcadsome other juju is wrong
21:39.52starseekerfirst order of business is to get committing to dmtogl - there are some configure.ac and other changes to be made here
21:40.03brlcadwoo hoo, the regression suite now passes save for one false-positive on a common.h header check
21:40.09starseekersweeet!
21:40.14starseekergood work brlcad
21:41.00``Erikthe points lexer?
21:41.23CIA-38BRL-CAD: 03erikgreenwald * r37262 10/brlcad/trunk/src/librt/primitives/metaball/metaball_tri.c: add missing curly
21:41.48``Erikif I do 'DISPLAY="" make regress', is it going to freak out? :)
21:41.55brlcad``Erik: yep
21:42.58``Erik(is that the last gotcha before it's cron ready?)
21:42.58brlcader, yep to the first, dunno about the latter
21:44.05``Erikhm, it seems to pass without a valid display, neat
21:44.21``Erik(with the display, I thought I saw the mged tk gui flash)
21:44.34starseekeryeah, I saw that too
21:44.42starseekerI think some of the mged commands trigger it
21:44.57``Erikheh, "attach"? :D
21:48.52``Erikheh, "attach"? :D
21:48.56``Erikwoops
22:03.18CIA-38BRL-CAD: 03starseeker * r37263 10/brlcad/branches/dmtogl/ (3 files in 3 dirs):
22:03.18CIA-38BRL-CAD: Start working on an upgrade to the way the build logic handles tcl/tk on OSX -
22:03.18CIA-38BRL-CAD: looking ahead to 8.6 tcl/tk and Aqua support, we can no longer pretend OSX is
22:03.18CIA-38BRL-CAD: unix and build the unix directory of tcl/tk - will need to use the actual macosx
22:03.18CIA-38BRL-CAD: directory. Also, new work on the Tk Aqua backend is being done via Cocoa rather
22:03.21CIA-38BRL-CAD: than Carbon, so need to reexamine what is needed for the non-X11 build flags to
22:03.23CIA-38BRL-CAD: tcl/tk
22:04.08yukonbobhello, #brlcad
22:04.28starseekerhey yukonbob
22:04.49CIA-38BRL-CAD: 03starseeker * r37264 10/brlcad/branches/dmtogl/ (9 files in 6 dirs): Sync dmtogl trunk to r37263.
22:04.51starseekerthis Tcl/Tk stuff is tiring :-P
22:12.02CIA-38BRL-CAD: 03irpguardian * r37265 10/brlcad/trunk/src/rt/view.c:
22:12.02CIA-38BRL-CAD: Moved timetable_init as so it wouldn't cause warning messages if called in
22:12.02CIA-38BRL-CAD: color view. Also added funtionality to timetable_init as now it stores values
22:12.02CIA-38BRL-CAD: into the time_table array, as well as longest time and shortest time.
22:23.07starseekerhmm, looks like that Cocoa framework test doesn't work
22:23.17starseekerblegh
22:31.29CIA-38BRL-CAD: 03irpguardian * r37266 10/brlcad/trunk/src/rt/view.c:
22:31.29CIA-38BRL-CAD: Added a variable which increments whenever a new entry is added into time table,
22:31.29CIA-38BRL-CAD: will be used for calculating average.
22:32.37starseekerwhat about success that makes my brain hurt?
22:33.41``Eriksee, now you're just begging for a mean quip about any success making your brain hurt..
22:34.10starseeker<snort> after this week it would be hard to argue
22:37.37CIA-38BRL-CAD: 03starseeker * r37267 10/brlcad/branches/dmtogl/configure.ac: See what happens with the old 8.5 tcl/tk when this build logic is used.
22:43.34CIA-38BRL-CAD: 03erikgreenwald * r37268 10/brlcad/trunk/src/librt/primitives/nmg/nmg_tri_mc.c: start building triangles
23:19.21starseekerwhat's the point of having stub library files and linkages?
23:31.55louipcI wonder how hard it would be to make one of these in brl-cad: http://www.youtube.com/watch?v=qLJxMUw51N8
23:42.25CIA-38BRL-CAD: 03starseeker * r37269 10/brlcad/branches/dmtogl/configure.ac: Need at least one file from the unix dir for includes.
23:58.32``Erikpuh-puh-puh-puh-pokerface *sing*
23:58.39``Erik(cartman style)

Generated by irclog2html.pl by Jeff Waugh - find it at freshmeat.net! Modified by Tim Riker to work with infobot logs, split per channel, etc.