IRC log for #brlcad on 20100409

00:15.10*** join/#brlcad Ralith (~ralith@69.90.48.97)
00:54.20brlcadwishes this cold/flu thing would dissipate faster
01:57.14pooliobrlcad: yeah that's been no fun... I got it too :(
01:57.40poolioalso my raytracer is done for now, submitted it but I'll probably come back to it this summer. I'm curious to play around with some of the distributed ray tracing effects
02:02.05CIA-73BRL-CAD: 03brlcad * r38395 10/brlcad/trunk/include/ (bu.h common.h): reverse the logic on GNUC_PREREQ to match glibc's __GNUC_PREREQ() macro (and the implied meaning), but rename it to GCC_PREREQ() too. make __attribute__ a common.h provision.
02:06.48CIA-73BRL-CAD: 03starseeker * r38396 10/brlcad/trunk/src/tclscripts/mged/ (Makefile.am helpbrowse.tcl):
02:06.48CIA-73BRL-CAD: Add example tcl script that successfully displayed html with images using tkhtml
02:06.48CIA-73BRL-CAD: (needs tkpng, exactly correct directory placement currently) - crude, fragile,
02:06.48CIA-73BRL-CAD: and not general, it will be removed once Archer has an html displaying browser
02:06.48CIA-73BRL-CAD: that can load images. Just being stuck in to have handy as an illustration.
02:12.04CIA-73BRL-CAD: 03starseeker * r38397 10/brlcad/trunk/src/tclscripts/archer/ (Archer.tcl LoadArcherLibs.tcl): Doesn't do anything yet except pop up an empty html window, but it doesn't crash - since there was no working Help of any sort hanging off of this menu option to start with, go ahead and commit.
02:24.39CIA-73BRL-CAD: 03brlcad * r38398 10/brlcad/trunk/include/common.h: provide a similar block for the intel compiler for testing version
02:40.04*** join/#brlcad IriX64 (~Warlock@bas2-sudbury98-1177593152.dsl.bell.ca)
02:40.47CIA-73BRL-CAD: 03brlcad * r38399 10/brlcad/trunk/include/common.h:
02:40.47CIA-73BRL-CAD: add similar LIKELY and UNLIKELY macros for branch prediction hinting. be VERY
02:40.47CIA-73BRL-CAD: brazen in making it clear that the new GCC_PREREQ and ICC_PREREQ macros are not
02:40.47CIA-73BRL-CAD: intended to be used outside of the common.h header file (as client code should
02:40.47CIA-73BRL-CAD: not be based on compilers, they should be based on tested features).
02:46.51CIA-73BRL-CAD: 03brlcad * r38400 10/brlcad/trunk/include/common.h: doxygen commens on the new UNUSED/LIKELY/UNLIKELY macros.
02:53.06CIA-73BRL-CAD: 03brlcad * r38401 10/brlcad/trunk/include/common.h: little more consistency if we somehow happen to run into a conflicting macro
02:53.55CIA-73BRL-CAD: 03brlcad * r38402 10/brlcad/trunk/include/common.h: oops, premature commit
03:16.02CIA-73BRL-CAD: 03brlcad * r38403 10/brlcad/trunk/include/common.h: add one more for a DEPRECATED marker that can be left on public API to warn if they're used. fix a typo on UNLIKELY too.
03:21.45CIA-73BRL-CAD: 03brlcad * r38404 10/brlcad/trunk/include/bu.h: replace __BU_ATTR_DEPRECATED with a simple DEPRECATED and use that before the function name as part of the signature.
03:31.21*** join/#brlcad Nohla (~jesica@201.255.244.77)
04:15.14*** join/#brlcad talcite (~matthew@76-10-151-95.dsl.teksavvy.com)
05:59.06*** join/#brlcad Stattrav (~Stattrav@202.3.77.233)
06:50.56*** join/#brlcad talcite (~matthew@75-119-229-153.dsl.teksavvy.com)
12:26.07starseekernotes the Cocoa AquaTk is apparently known not to work on 10.4
12:26.10starseekermutter...
12:29.50starseekerchuckles at this line: "Being an open source developer at Microsoft is like being a preacher in Vegas"
13:16.12brlcaddebates trying to be mobile
13:17.22CIA-73BRL-CAD: 03brlcad * r38405 10/brlcad/trunk/src/ (91 files in 43 dirs): use the new UNUSED() macro. e.g.: int main(int UNUSED(argc), char *UNUSED(argv)[]);
14:32.38CIA-73BRL-CAD: 03starseeker * r38406 10/brlcad/trunk/src/tclscripts/archer/Archer.tcl: Successful display of image containing html in Archer help, but scrollbar not working
14:40.27CIA-73BRL-CAD: 03erikgreenwald * r38407 10/brlcad/trunk/src/librt/primitives/nmg/nmg_tri_mc.c: fire the primary rays in the gridding function
14:45.22CIA-73BRL-CAD: 03starseeker * r38408 10/brlcad/trunk/src/tclscripts/archer/Archer.tcl: OK, scrolling is working now.
14:47.15CIA-73BRL-CAD: 03starseeker * r38409 10/brlcad/trunk/src/tclscripts/mged/ (Makefile.am helpbrowse.tcl): Shouldn't need the helpbrowse example any more.
15:35.41*** join/#brlcad Stattrav (~Stattrav@202.3.77.233)
17:04.35CIA-73BRL-CAD: 03brlcad * r38410 10/brlcad/trunk/src/librt/primitives/nmg/nmg_pr.c: fix a subtle bug reported via an sf bug report from Fred ( breakfastfish ) when strict warnings was failing for him due to array subscript being out of bounds.
17:33.48CIA-73BRL-CAD: 03brlcad * r38411 10/brlcad/trunk/src/libbn/sphmap.c: move spm_free up since spm_init calls it, even if spm.h has the decls.
17:36.32CIA-73BRL-CAD: 03brlcad * r38412 10/brlcad/trunk/src/libtermio/termio.c: fix the 'real prototype' warning/error reproted by Auch Scelsi ( scelsi ) in sf report 2984176. since the signature is a bit complex, just move it earlier to avoid the need for a forward decl.
17:47.33*** join/#brlcad Elrohir (~kvirc@p5B14AAAD.dip.t-dialin.net)
18:11.24CIA-73BRL-CAD: 03starseeker * r38413 10/brlcad/trunk/src/tclscripts/archer/Archer.tcl: Figured out how to get the title string from the html document - now to do something useful with it...
18:23.49CIA-73BRL-CAD: 03brlcad * r38414 10/brlcad/trunk/src/mged/titles.c:
18:23.49CIA-73BRL-CAD: add some additional validation checks related to sf report 1586998 (mged
18:23.49CIA-73BRL-CAD: segfault) from Karel Kulhavy ( clock3 ). was unable to reproduce his crash, but
18:23.49CIA-73BRL-CAD: suspect that there's some obscure case wehere illump is either ending up invalid
18:23.49CIA-73BRL-CAD: or illump->s_fullpath is invalid yet non-null, causing problems while we
18:23.50CIA-73BRL-CAD: iterate, print, and free the vls.
18:51.51CIA-73BRL-CAD: 03brlcad * r38415 10/brlcad/trunk/TODO: see if converting arbs with non-planar faces to brep arbs would suffice, allowing the removal of that validation restriction (at least in terms of ray-tracing). would be a great lil exercise for a new dev to play with.
18:52.58CIA-73BRL-CAD: 03starseeker * r38416 10/brlcad/trunk/src/tclscripts/archer/Archer.tcl: No links working, but explore using two html instances to provide a table of contents ability.
19:05.02brlcadalright.. bug count back down below 50
19:09.39starseekernice
19:12.29brlcadtom and wim's bugs should get some attention, prime users
19:12.46brlcadvery interesting tessellation failure bug report too, simpel geometry
19:47.18CIA-73BRL-CAD: 03brlcad * r38417 10/brlcad/trunk/include/brlcad_version.h: add a -0 to the version literals so that even if the file is empty, it'll result in the version getting set to zero. this should fix a peculiar case on windows where the COUNT file is empty.
19:55.22CIA-73BRL-CAD: 03brlcad * r38418 10/brlcad/trunk/src/libbu/cmd.c: quell Tcl_CmdProc type warning. bu_cmdtab's callback signature is rather limited for arbitrary callbacks. should be revisited.
20:00.47CIA-73BRL-CAD: 03starseeker * r38419 10/brlcad/trunk/src/tclscripts/archer/Archer.tcl: Adding hyperlinking support looks to be a bit of a challenge - might be able to do what is needed for help browser with a tree widget.
20:05.10CIA-73BRL-CAD: 03brlcad * r38420 10/brlcad/trunk/src/libfb/if_wgl.c:
20:05.10CIA-73BRL-CAD: reorder functions to avoid most of the forward declarations (which were mostly
20:05.10CIA-73BRL-CAD: all k&r style anyways causing msvc grief. move wgl_interface to the bottom
20:05.10CIA-73BRL-CAD: accordingly so it can reference those funcs properly without decls.
20:09.35CIA-73BRL-CAD: 03erikgreenwald * r38421 10/brlcad/trunk/src/libgcv/region_end_mc.c: fuse the vertices and edges before the regions/shells are killed.
20:14.09*** join/#brlcad R0b0t1 (~Enigma@unaffiliated/r0b0t1)
20:17.18CIA-73BRL-CAD: 03brlcad * r38422 10/brlcad/trunk/src/libfb/if_wgl.c: style, comment, indent cleanup
20:23.21CIA-73BRL-CAD: 03brlcad * r38423 10/brlcad/trunk/src/libfb/if_wgl.c: more cleanup and de-k&rification. remove register keyword.
20:28.34CIA-73BRL-CAD: 03bob1961 * r38424 10/brlcad/trunk/src/tclscripts/archer/ (Archer.tcl ArcherCore.tcl): Disable the old tree viewer. Added functionality to colorize tree nodes depending on whether or not they're displayed. Also added (currently disabled) functionality to colorize tree nodes impacted by current edit state.
20:34.43CIA-73BRL-CAD: 03brlcad * r38425 10/brlcad/trunk/src/libfb/fbserv_obj.c: quell msvc warnings about Tcl params not matching.
20:38.11CIA-73BRL-CAD: 03brlcad * r38426 10/brlcad/trunk/NEWS: bob added colorization to archer's tree view. indicates what objects are presently selected.
20:57.41starseekerbrlcad: what do you think is the better course of action - work MGED's browser launching abilities into Archer, or  pursue a tkhtml based viewer?
20:58.23starseekerI might be able to get something very simple working quickly with tkhtml, but hyperlinking will be a trick
20:59.05starseekerbarring something like sucking in hv3
20:59.14starseeker(which requires sqlite)
21:02.46CIA-73BRL-CAD: 03erikgreenwald * r38427 10/brlcad/trunk/src/librt/primitives/metaball/metaball_tri.c: decimate if possible.
21:19.58CIA-73BRL-CAD: 03brlcad * r38428 10/brlcad/trunk/src/conv/iges/makedir.c: wow, using write(1,... to send to stdout. ballsy stupid.
21:20.50CIA-73BRL-CAD: 03brlcad * r38429 10/brlcad/trunk/src/conv/iges/makedir.c: cleanup indent, style, comments, remove authors.
21:24.21``Erikhmmmmm
21:24.43``Erikstarts thinking that nmg_model_face_fuse() simply doesn't work O.o
21:25.31CIA-73BRL-CAD: 03brlcad * r38430 10/brlcad/trunk/src/mged/fbserv.c: use explicit Clientdata vars to try and appease msvc
21:26.17``Erikalmost looks like the typical "inci(int i){ i++; return; }" issue
21:29.12CIA-73BRL-CAD: 03starseeker * r38431 10/brlcad/trunk/src/tclscripts/archer/Archer.tcl: Start working on the man page viewer - close, but can't get it to update the html display with the new data yet.
21:29.16CIA-73BRL-CAD: 03brlcad * r38432 10/brlcad/trunk/src/mged/fbserv.c: reorder to avoid forward decls
21:31.45CIA-73BRL-CAD: 03brlcad * r38433 10/brlcad/trunk/src/mged/fbserv.c: oops, missed one reorder. quell all warnings.
21:32.22brlcadcould be
21:32.58brlcadthat would be f'ing hilarious if most of the tessellation problems turned out to be a simple off-by-one code bug in an obscure routine
21:33.21brlcadand by fixing it, all problems go away
21:33.33brlcadsupremely unlikely, but it would be slap dead hilarious
21:35.11``Eriknot on off by one
21:35.45CIA-73BRL-CAD: 03starseeker * r38434 10/brlcad/trunk/src/tclscripts/archer/Archer.tcl: Ah, there we go - got it working. We now have man page viewing in Archer.
21:36.01``Erikit looks like the various functions start out by creating a ptbl, wiring the pointers in the ptbl to the content in the NMG struct... do their magic on the ptbl without altering the NMG... then delete teh ptbl and return the # of items supposedly  removed.
21:36.38``Erik(this is all fusing stuff, the decimation type routines)
21:36.57CIA-73BRL-CAD: 03brlcad * r38435 10/brlcad/trunk/src/mged/fbserv.c: instead of going through long, use uint32_t stdint type for truncating/testing the clientdata fd
21:37.50``Erikemailed daytona, may do some serious code spelunking on tuesday
21:38.38CIA-73BRL-CAD: 03brlcad * r38436 10/brlcad/trunk/src/mged/titles.c: remove unused vls and subsequent tiny memory leak
21:39.32brlcadwas thinking to start hashing out nmg test cases for api validation
21:40.08``Erikheh, unit test style?
21:40.09brlcadgoing through the nmg api func by func, writing a little test case that verifies its behavior for a given set of inputs
21:40.12brlcadyeah
21:40.32brlcadstarting with the super basic low-level funcs
21:40.33starseekeris inclined to think nmg stuff is complex enough to justify it
21:40.54``Erikmight be interesting to run some coverage type tools, see how many of those functions aren't even used (by us)
21:40.58brlcadstarseeker: it is with our current dev set, nobody here is an nmg expert
21:41.12brlcadused to be several
21:41.28``Erikdaytona is probably the closest left
21:41.42brlcadbar far, certainly
21:41.58``Erikand he doesn't seem interested in remembering it... :D
21:42.18brlcadhe knows that stuff way better than he lets on
21:42.26``ErikI know, that's why I emailed him
21:43.24brlcadhe spent just a few weeks (years ago) and cleaned up years of instability, took it from 90% reliable to at least 95% reliable
21:43.40``Erika decade ago? :D
21:43.45brlcadnot quite
21:44.30brlcadby the end, I think he had it close to 99%
21:44.40``Erikhas to brush up on some bu stuff before he can get any deeper into nmg :/
21:44.49brlcadfew more weeks and I bet he could have added a few nines
21:47.41``Erikprobably
21:48.08``Erik<-- throws these 'kimchi flavored' vegetable dumplings in his cooler and heads home O.o
21:49.09CIA-73BRL-CAD: 03brlcad * r38437 10/brlcad/trunk/src/mged/utility1.c: const qualifier mismatch.
21:54.08CIA-73BRL-CAD: 03brlcad * r38438 10/brlcad/trunk/src/mged/utility1.c: more cleanup, reorder to avoid forward decls. remove slew of non-existent forward decl funcs.
22:05.42CIA-73BRL-CAD: 03brlcad * r38439 10/brlcad/trunk/src/mged/utility1.c: quell all remaining verbose compilation warnings (gcc401)
22:06.52starseekerO.o  anybody else getting a seg fault with Archer on Linux?
22:07.34CIA-73BRL-CAD: 03brlcad * r38440 10/brlcad/trunk/src/rt/view_bot_faces.c: casting quellage. stupid use of ptbl..
22:08.10starseekerlibbu/cmd.c line 75 - cannot access memory of at address of ctp
22:08.29CIA-73BRL-CAD: 03brlcad * r38441 10/brlcad/trunk/src/rt/view_bot_faces.c: cleanup, ws, indent, consistency
22:08.31starseekersame deal with mged
22:08.39brlcad``Erik: thanks anyways.. almost headed in just for that
22:08.55brlcadstill wasn't 100% though today seems to be turning around
22:09.40brlcadstarseeker: oops
22:10.58starseeker38418 I assume?
22:11.04CIA-73BRL-CAD: 03brlcad * r38442 10/brlcad/trunk/src/libbu/cmd.c: shame on you, gcc, for shaming me. should have warned about the wrong scope bad dereference.
22:11.15starseekerhehe
22:11.17starseekerthanks
22:11.18brlcadjust read the line, it's bogus
22:11.25brlcaddereferencing ctp
22:12.20starseekerah, right
22:18.47starseeker/brlcad/src/rt/view_bot_faces.c:239: error: 'fnp0' undeclared
22:20.08starseekeroh, I see it
22:21.23CIA-73BRL-CAD: 03starseeker * r38443 10/brlcad/trunk/src/rt/view_bot_faces.c: Fix typo
22:26.57CIA-73BRL-CAD: 03r_weiss * r38444 10/brlcad/trunk/src/conv/obj-g_new.c: some variable name cleanup and debugging nmg creation logic
22:30.27CIA-73BRL-CAD: 03starseeker * r38445 10/brlcad/trunk/NEWS: The html manual page viewer in MGED has now been ported to Archer.
22:32.31CIA-73BRL-CAD: 03starseeker * r38446 10/brlcad/trunk/src/archer/TODO: Mark man pages as done in Archer TODO
22:53.33``Erikmebbe I'll bring some back in next week *shrug*

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