| 02:14.16 | *** join/#brlcad stevegt_ (~stevegt@c-69-181-134-76.hsd1.ca.comcast.net) | |
| 03:04.08 | brlcad | has the initial script working |
| 03:04.15 | brlcad | interesting success/failures |
| 03:15.52 | CIA-50 | BRL-CAD: 03brlcad * r41208 10/brlcad/trunk/sh/ (Makefile.am conversion.sh): (log message trimmed) |
| 03:15.52 | CIA-50 | BRL-CAD: add an initial, useful, and VERY informative script that walks over all objects |
| 03:15.52 | CIA-50 | BRL-CAD: in the geometry file(s) specified and reports which will successfully convert to |
| 03:15.53 | CIA-50 | BRL-CAD: nmg as well as bot. notably interesting that many nmg succeed while the |
| 03:15.53 | CIA-50 | BRL-CAD: supposedly trivial subsequent conversion to bot sometimes fails. also |
| 03:15.53 | CIA-50 | BRL-CAD: interesting that several debug messages seem to be escaping stdout/stderr |
| 03:15.54 | CIA-50 | BRL-CAD: redirect. work is performed on a temporary copy so original input is |
| 03:16.15 | brlcad | original input is unmodified. example use: sh conversion.sh db/*.g |
| 03:32.55 | brlcad | huh, that's odd and wrong .. somehow it's getting to the /dev/tty failsafe printing during bu_bomb instead of triggering the exception handling |
| 08:17.15 | *** join/#brlcad merzo (~merzo@smartbussiness.mobicom.net.ua) | |
| 08:59.24 | *** join/#brlcad CIA-48 (~CIA@208.69.182.149) | |
| 09:22.00 | *** join/#brlcad _clock_ (~sanook_ba@217-162-130-133.dclient.hispeed.ch) | |
| 09:46.05 | *** join/#brlcad mafm (~mafm@129.Red-81-43-146.staticIP.rima-tde.net) | |
| 10:11.45 | CIA-48 | BRL-CAD: 03brlcad * r41209 10/brlcad/trunk/TODO: noticed a couple problems already. facetize creates crap BoTs by only dumping first region, first shell. nmg_bot() and nmg_from_bot() both suck from an API perspective in that they only work with a shell too. |
| 10:20.57 | CIA-48 | BRL-CAD: 03brlcad * r41210 10/brlcad/trunk/src/libged/facetize.c: this is the cause of the spurious bu_bomb() /dev/tty print messages infesting my conversion script. need to catch exceptions thrown by nmg_bot(). |
| 10:23.22 | CIA-48 | BRL-CAD: 03brlcad * r41211 10/brlcad/trunk/NEWS: no more /dev/tty printing when running 'facetize' command when conversion to BoT fails. |
| 10:31.41 | CIA-48 | BRL-CAD: 03brlcad * r41212 10/brlcad/trunk/src/conv/nmg/nmg-bot.c: convert to try/catch block exception style, cleanup ws |
| 10:36.38 | CIA-48 | BRL-CAD: 03brlcad * r41213 10/brlcad/trunk/src/libged/facetize.c: |
| 10:36.38 | CIA-48 | BRL-CAD: convert our bu_setjump/bu_unsetjump blocks to more familiar try/catch exception |
| 10:36.38 | CIA-48 | BRL-CAD: handling layout. this should happen across the entire code base for clarity and |
| 10:36.38 | CIA-48 | BRL-CAD: consistency. I remember it taking me a while to wrap my head around jumps when |
| 10:36.38 | CIA-48 | BRL-CAD: I started and the inconsistent calling style didn't help. |
| 10:39.48 | CIA-48 | BRL-CAD: 03brlcad * r41214 10/brlcad/trunk/TODO: cleanup bu_setjump/bu_unsetjump blocks to more familiar try/catch style |
| 10:43.47 | d-lo | Mernin! |
| 10:47.38 | CIA-48 | BRL-CAD: 03brlcad * r41215 10/brlcad/trunk/src/libged/facetize.c: report if this is an NMG with more than one region or shell when going to BoT |
| 10:49.19 | CIA-48 | BRL-CAD: 03brlcad * r41216 10/brlcad/trunk/src/libged/facetize.c: db_free_tree() already tests the pointer and does nothing if null. simplify. |
| 10:52.11 | CIA-48 | BRL-CAD: 03brlcad * r41217 10/brlcad/trunk/src/conv/nmg/nmg-bot.c: wants a pointer |
| 11:10.34 | CIA-48 | BRL-CAD: 03brlcad * r41218 10/brlcad/trunk/src/librt/db_tree.c: |
| 11:10.34 | CIA-48 | BRL-CAD: need another pair of core eyes on this but this looks like an outright bug in |
| 11:10.34 | CIA-48 | BRL-CAD: the logic! checking the wrong left/right side during db_tree_parse(). the |
| 11:10.34 | CIA-48 | BRL-CAD: intent seems clear so fix it so we do not return a corrupt tree pointer (with a |
| 11:10.34 | CIA-48 | BRL-CAD: non-null left and null right). no telling the damage that has caused. |
| 11:11.22 | d-lo | yikes, that's a nasty one (r41218) |
| 11:13.56 | CIA-48 | BRL-CAD: 03brlcad * r41219 10/brlcad/trunk/src/librt/db_tree.c: the db_free_tree for the left was intentional. free it up so we don't leak. |
| 11:24.53 | *** join/#brlcad Zaebos (~irc@pd95b7f5e.dip0.t-ipconnect.de) | |
| 11:26.19 | brlcad | that bug has been in there since the original verison 12 years and 10 months ago.. ouch |
| 11:26.26 | brlcad | copy/paste error |
| 11:27.10 | brlcad | given where that is at, it had to cause some really obscure problems |
| 11:28.51 | CIA-48 | BRL-CAD: 03brlcad * r41220 10/brlcad/trunk/src/other/tcl/generic/tclDecls.h: quell shadow on trace() |
| 11:29.20 | CIA-48 | BRL-CAD: 03brlcad * r41221 10/brlcad/trunk/src/libged/put_comb.c: simplify |
| 11:30.47 | CIA-48 | BRL-CAD: 03brlcad * r41222 10/brlcad/trunk/src/ (10 files in 3 dirs): don't need to check the pointer parameter to db_free_tree() as it will do a test for callers and return without action if null. simplify. |
| 11:33.22 | brlcad | hm.. "mged -c db/cube.g facetize cube.bot cube" .. me thinks thar be an infinite loop |
| 11:35.19 | brlcad | yep |
| 11:35.40 | brlcad | bad nmg_ptbl_vfuse(), no donut for you |
| 11:37.02 | d-lo | so you seem to be on a bug killing spree today :) |
| 11:37.24 | brlcad | didn't sleep well, figured why just lay there |
| 11:37.46 | d-lo | bummer man, sorry to hear that. kudos to making it productive though =D |
| 11:40.48 | ``Erik | I believe there are unlocked globals somewhere hiding in the nmg code |
| 11:41.25 | ``Erik | there is some serious attention that needs to be paid to that former library now part of librt |
| 11:42.25 | ``Erik | the threaded marching cubes thing I commited, I'm definite that I did everything right, but it goes infinite loop on a BU_FOR_EACH macro down in nmg land, and that ain't right |
| 11:42.34 | ``Erik | I mean, I locked around ALL nmg calls |
| 11:42.43 | ``Erik | aanyways |
| 11:43.17 | ``Erik | (ditch nmg for gts? argue for a week long offsite bugstomp event?) |
| 11:43.51 | brlcad | so you're saying you had a reliable test case that hit the problem but chose not to fix? nobody else to blame there... |
| 11:44.27 | brlcad | there's no reason to believe gts wouldn't be any better or worse, and I'd actually bet worse overall |
| 11:44.29 | ``Erik | I have a reliable test case that has just shown up and have added it into my priority queue appropriately and spoke to richard about some of the details |
| 11:44.33 | brlcad | especially for guaranteeing solidity |
| 11:45.20 | brlcad | at least, I'd bet the time it'd take to properly integrate and test gts is on order with or longer than the time it'd take to clean up nmg |
| 11:45.30 | ``Erik | that's what I'm thinking |
| 11:45.51 | brlcad | my test script is showing that the failures are pretty rare |
| 11:46.02 | brlcad | they just happen at a low level and cascade up the tree |
| 11:46.15 | ``Erik | I also think that doing marching cubes at all was less effective and more time consuming than fixing nmg, which is why I argued my ttttttttm task to be 'improve facetization' instead of marchine cubes specific things |
| 11:46.32 | ``Erik | ~3% rate on 'normal' geometry, right? :) |
| 11:47.06 | brlcad | don't have it calculating stats yet -- hit this infinite loop case first |
| 11:47.35 | ``Erik | ~3% feels roughly right for the geometry sets I've been working with |
| 11:47.37 | brlcad | this should have been done, oh, about 10 years ago |
| 11:47.51 | ``Erik | don't look at me, I wasn't here then :D |
| 11:48.11 | brlcad | yeah, that's about what I'd guesstimate too if you count cascading up the hierarchy |
| 11:48.16 | brlcad | if you don't, it's less |
| 11:48.53 | ``Erik | vic and dwayne are the kinds of guys who this matters to, and they don't think about the hierarchy and cascade failures, they just know what comes out the end |
| 11:49.13 | ``Erik | they're the metric that matters :D |
| 11:49.50 | ``Erik | is waiting for the recycling truck so'z he can put away his bin before heading in, they seem to be running a bit slow today :/ |
| 12:25.48 | d-lo | thinks that recycling trucks should be electric powered, otherwise its a bit of an oxymoron :) |
| 12:25.50 | CIA-48 | BRL-CAD: 03brlcad * r41223 10/brlcad/trunk/NEWS: |
| 12:25.50 | CIA-48 | BRL-CAD: various commits by richard to improve the robustness of facetization and |
| 12:25.50 | CIA-48 | BRL-CAD: geometry export (to polygonal formats). fixed ell tess bug, coplanar validity |
| 12:25.50 | CIA-48 | BRL-CAD: test, tightness of orthogonal vectors, coplanar lines test, and more. |
| 12:28.20 | CIA-48 | BRL-CAD: 03brlcad * r41224 10/brlcad/trunk/NEWS: keith added (improved) ray bundle shooting capability to librt via new rt_shootrays() routine. also implemented ray pattern generators for circular and elliptical. functionality is presently exposed in the rtshot tool. |
| 12:37.10 | CIA-48 | BRL-CAD: 03brlcad * r41225 10/brlcad/trunk/NEWS: (log message trimmed) |
| 12:37.11 | CIA-48 | BRL-CAD: bob fixed a problem with the pro/e export plugin being unable to load on |
| 12:37.11 | CIA-48 | BRL-CAD: windows. the problem was that the prodevelop/protk libraries that we have to |
| 12:37.11 | CIA-48 | BRL-CAD: link against are statically compiled against an old version of the C runtime. |
| 12:37.11 | CIA-48 | BRL-CAD: that C runtime wouldn't necessarily match the newer version we compiled against |
| 12:41.55 | brlcad | richard will need to regenerate that first page |
| 12:47.25 | CIA-48 | BRL-CAD: 03brlcad * r41226 10/brlcad/trunk/NEWS: |
| 12:47.25 | CIA-48 | BRL-CAD: improved robustness of boolean tree parsing by fixing a 12 year 10 month old bug |
| 12:47.25 | CIA-48 | BRL-CAD: in db_tree.c:rt_tree_parse() where we were proceeding with a corrupt tree if the |
| 12:47.25 | CIA-48 | BRL-CAD: left tree was non-empty and right tree was empty. cause was a simple copy/paste |
| 12:47.25 | CIA-48 | BRL-CAD: error, fix was to check the right side for validity so it'd release and unset |
| 13:03.45 | *** join/#brlcad R0b0t1 (~Enigma@unaffiliated/r0b0t1) | |
| 13:09.09 | *** join/#brlcad roberthl (~robert@2001:ba8:1f1:f03d::2) | |
| 13:09.15 | ``Erik | I'll tell him |
| 13:09.26 | *** join/#brlcad roberthl (~robert@2001:ba8:1f1:f03d::2) | |
| 13:09.26 | *** join/#brlcad roberthl (~robert@mediawiki/RobertL) | |
| 13:39.04 | brlcad | just the news ones, not the sf ones |
| 14:28.29 | *** join/#brlcad d_rossberg (~rossberg@BZ.BZFLAG.BZ) | |
| 14:29.50 | d_rossberg | ah, still daylight saving time |
| 14:33.48 | CIA-48 | BRL-CAD: 03starseeker * r41227 10/brlcad/branches/cmake/src/other/tcl/CMakeLists.txt: |
| 14:33.48 | CIA-48 | BRL-CAD: Organize the darwin specific tests, add a few - gonna need to study the complex |
| 14:33.48 | CIA-48 | BRL-CAD: awk/grep lines being added to the LD flags and see what they're doing before |
| 14:33.48 | CIA-48 | BRL-CAD: it's clear how to do it with CMake - just adding the text from the configure.in |
| 14:33.48 | CIA-48 | BRL-CAD: does not work. |
| 14:34.58 | d-lo | d_rossberg: ya, we still have a few more days of DST. EU just switched off DST, right? |
| 14:36.40 | d_rossberg | right, last sunday morning we got the hour back |
| 14:38.12 | d_rossberg | i've a telco this week and have to kepp in mind that the time difference is 5h now |
| 14:41.28 | ``Erik | brlcad: libged/facetize.c is busted, +256, 'm' is undefined.. not sure if ya mean nmg_model instead |
| 14:49.52 | starseeker | does anybody know what feeding a double dash to grep in this fashion does? grep -q -- '-prebind ' |
| 14:51.30 | ``Erik | double dash b y itself stops getopt parsing |
| 14:51.43 | ``Erik | otherwise, it'd try to read -prebind as an option instead of a symbol |
| 14:52.14 | ``Erik | (the single quote helps the shell, not the program executed by the shell) |
| 14:52.35 | starseeker | ah, thanks |
| 14:55.04 | CIA-48 | BRL-CAD: 03starseeker * r41228 10/brlcad/branches/cmake/src/other/tcl/CMakeLists.txt: |
| 14:55.04 | CIA-48 | BRL-CAD: Hmm... if I'm not mistaken, the awk logic isn't needed if we pass in the right |
| 14:55.04 | CIA-48 | BRL-CAD: pre-defined variables. Grep logic is a tad less clear but it seems to be |
| 14:55.04 | CIA-48 | BRL-CAD: checking for -prebind in LDFLAGS and passing that into the link flags if needed |
| 14:55.04 | CIA-48 | BRL-CAD: - can handle that elsewhere more cleanly I think. |
| 14:57.29 | CIA-48 | BRL-CAD: 03starseeker * r41229 10/brlcad/branches/cmake/src/other/tcl/CMakeLists.txt: Make sure we point to the generated plist files - need to confirm those are generating like they do in the autotools build. |
| 16:01.43 | CIA-48 | BRL-CAD: 0392.249.84.146 07http://brlcad.org * r2328 10/wiki/Main_Page: |
| 17:38.24 | CIA-48 | BRL-CAD: 03brlcad * r41230 10/brlcad/trunk/src/anim/ (7 files): quell verbose strict warnings on linux |
| 17:40.38 | CIA-48 | BRL-CAD: 03brlcad * r41231 10/brlcad/trunk/src/anim/ (cattrack.c chan_permute.c): minor ws consistency |
| 17:57.31 | CIA-48 | BRL-CAD: 03Sean 07http://brlcad.org * r2329 10/wiki/Main_Page: Reverted edits by [[Special:Contributions/92.249.84.146|92.249.84.146]] ([[User talk:92.249.84.146|Talk]]); changed back to last version by [[User:Dloman|Dloman]] |
| 17:57.45 | CIA-48 | BRL-CAD: 03Sean 07http://brlcad.org * r0 10/wiki/Special:Log/block: blocked [[User:92.249.84.146]] with an expiry time of infinite (anonymous users only, account creation disabled): Spamming links to external sites |
| 18:04.11 | CIA-48 | BRL-CAD: 03brlcad * r41232 10/brlcad/trunk/src/canon/canonize.c: track fwrite() results |
| 18:18.23 | CIA-48 | BRL-CAD: 03brlcad * r41233 10/brlcad/trunk/src/libged/facetize.c: copy paste diff with src/conv/nmg/nmg-bot.c code. here it's nmg_model, not m. |
| 19:10.58 | CIA-48 | BRL-CAD: 03starseeker * r41234 10/brlcad/branches/cmake/src/other/tk/ (8 files in 2 dirs): Start reworking Tk build to use CFLAGS too. |
| 19:12.22 | *** join/#brlcad mafm (~mafm@81.35.69.185) | |
| 19:46.37 | CIA-48 | BRL-CAD: 03starseeker * r41235 10/brlcad/branches/cmake/src/other/tk/CMakeLists.txt: |
| 19:46.37 | CIA-48 | BRL-CAD: Add some more stuff, replace a couple accidentally deleted files - showing signs |
| 19:46.37 | CIA-48 | BRL-CAD: of getting closer to being a functional Tk build, although Tcl vs. Tk CFLAGS |
| 19:46.37 | CIA-48 | BRL-CAD: still need straightening out and the plist logic for Tk is not yet present. |
| 20:08.34 | *** join/#brlcad mafm_ (~mafm@81.35.69.185) | |
| 21:59.52 | CIA-48 | BRL-CAD: 03starseeker * r41236 10/brlcad/branches/cmake/src/other/tk/ (CMake/tcl.cmake CMakeLists.txt): Switch TCL_CFLAGS to TK_CFLAGS to avoid conflicts - need to make this more generic in some fashion later. |
| 22:46.43 | *** join/#brlcad stevegt_1 (~stevegt@cislunar.TerraLuna.Org) | |
| 23:35.48 | CIA-48 | BRL-CAD: 03starseeker * r41237 10/brlcad/trunk/TODO: Toss in a few quick notes on ideas pertaining to NURBS and NMG - moving logic from old NURBS to new and using BoT raytracing to handle NMGs. |
| 23:38.27 | *** join/#brlcad Ralith (~ralith@d142-058-094-190.wireless.sfu.ca) | |
| 23:40.48 | CIA-48 | BRL-CAD: 03starseeker * r41238 10/brlcad/trunk/TODO: Typo |