| 00:59.35 | ``Erik | hah http://www.youtube.com/watch?v=fzza-ZbEY70 |
| 01:04.42 | brlcad | hehe |
| 01:23.53 | ``Erik | "Twilight's like soccer. They run around for 2 hours, nobody scores, and it's billion fans insist you just don't understand" heh |
| 01:42.45 | dtidrow | rofl |
| 01:58.01 | starseeker | brlcad: I've started pondering a regex based solution to red, and I'll probably need to discuss a few specifics... |
| 01:58.12 | starseeker | one question - is regex better than lex/yacc in this situation? |
| 01:58.22 | starseeker | (not that lex/yacc would be any easier, just curious) |
| 01:59.43 | ``Erik | does red take a CFG? O.o |
| 02:06.35 | starseeker | ``Erik: technically probably not, but we're sorta looking for the "this is all this command will ever need, period, dammit" solution |
| 02:07.28 | starseeker | brlcad wondered what would happen if someone pasted the mged binary contents into the temp file and tried calling it a comb - I'm not brave enough to try but I'd expect something rather... odd |
| 02:07.53 | starseeker | I don't think I've got sanity checking working on a level to cope with that :-/ |
| 02:08.54 | starseeker | right now it's extremely line centric, which brlcad pointed out isn't really correct - matricies previously were allowed funky whitespace between numbers and multiple lines per matrix |
| 02:09.19 | starseeker | ('course, this one does have the merit of working...) |
| 02:10.32 | brlcad | starseeker: I'd expect regex to be better just because there's not really anything to "compile" |
| 02:11.41 | brlcad | interesting thought, but I think it's more an idea that could work, but would be using a chainsaw where you needed a hacksaw |
| 02:16.12 | starseeker | nods - kinda figured |
| 02:40.07 | starseeker | knows the lex/yacc namespace question has to be resolved for libgcv, but is not looking forward to it... |
| 03:32.03 | brlcad | strict compilation failures in librt, warnings about mismatched sign |
| 03:43.43 | brlcad | fixed |
| 03:45.50 | CIA-93 | BRL-CAD: 03brlcad * r39771 10/brlcad/trunk/ (include/raytrace.h src/librt/db5_types.c): make db5_update_std_attributes() and db5_apply_std_attributes() return void since neither return anything. quell other strict warnings about type mismatches, unused vars, and constant expressions. |
| 03:51.19 | brlcad | starseeker: should quell all your warnings on red.c |
| 04:03.20 | starseeker | ah, thanks :-) |
| 04:03.36 | starseeker | blinks - didn't realize I had strict off on the Mac |
| 04:06.35 | CIA-93 | BRL-CAD: 03brlcad * r39772 10/brlcad/trunk/src/libged/red.c: declare the unpublished functions that we're using. looks like there are two still in raytrace.h too (i.e., db5_apply_std_attributes() and db5_update_std_attributes()) |
| 04:08.49 | CIA-93 | BRL-CAD: 03brlcad * r39773 10/brlcad/trunk/ (include/raytrace.h src/librt/db5_types.c): declare with const members to quell warnings. also update ws indent and style for consistency. |
| 04:08.53 | brlcad | er, you sould quell them .. I only quelled a few |
| 04:10.55 | brlcad | ahhh, and that quellage indicates a type error in red.c |
| 04:11.10 | brlcad | where you pass a const, but the func needs to be non-const as presently written |
| 04:13.33 | CIA-93 | BRL-CAD: 03brlcad * r39774 10/brlcad/trunk/ (include/raytrace.h src/librt/db5_types.c): ah, cannot be const since that's the comb we are updating. that means red libged function has bad logic, passing a const to a non-const function. |
| 04:22.48 | starseeker | trys on his gentoo box... |
| 04:23.33 | starseeker | I can't recall - do I have to enable strict explicitly now? |
| 04:25.24 | starseeker | oh, wait - I'll be I never updated write_comb and friends to be non-const after adding the standardize stuff... |
| 04:27.11 | CIA-93 | BRL-CAD: 03starseeker * r39775 10/brlcad/trunk/src/libged/red.c: write_comb shouldn't be declaring comb const |
| 04:37.58 | starseeker | explicitly turns on strict this time... |
| 04:49.44 | starseeker | erm. Got to docs with strict on here... |
| 06:02.28 | CIA-93 | BRL-CAD: 03d_rossberg * r39776 10/rt^3/trunk/include/: ignore generated file libcoreinterface.h |
| 06:34.36 | *** join/#brlcad alex_joni (~alex_joni@emc/board-of-directors/alexjoni) | |
| 07:01.25 | *** join/#brlcad SWPadnos_ (~Me@dsl107.esjtvtli.sover.net) | |
| 07:04.34 | *** join/#brlcad indianlarry (~indianlar@BZ.BZFLAG.BZ) | |
| 07:04.34 | *** join/#brlcad cosurgi (~cosurgi@atak.bl.pg.gda.pl) | |
| 07:06.26 | *** join/#brlcad SWPadnos (~Me@dsl107.esjtvtli.sover.net) | |
| 07:56.57 | *** join/#brlcad Stattrav (~Stattrav@117.192.130.253) | |
| 09:53.16 | *** join/#brlcad mafm (~mafm@245.Red-88-23-77.staticIP.rima-tde.net) | |
| 12:20.04 | *** join/#brlcad PrezKennedy (~Prez@2002:601f:5460::601f:5460) | |
| 12:40.36 | brlcad | starseeker: different versions of the compiler do better at warnings, add --enable-warnings too |
| 12:40.53 | brlcad | that will make sure they are verbose |
| 12:41.02 | brlcad | can also make sure you try at least one other platform |
| 12:46.20 | starseeker | gcc version 4.3.4... |
| 12:46.35 | starseeker | adds enable warnings and tries again... |
| 12:47.49 | starseeker | wonders if there is still any hope of the symbolics and macsyma stuff ever seeing the light of day again... |
| 12:50.53 | starseeker | OK, THAT worked, but it tripped up on metaballs |
| 12:53.14 | ``Erik | hm http://www.caelinux.com/CMS/ |
| 12:53.33 | ``Erik | what trip-up? O.o |
| 12:53.42 | CIA-93 | BRL-CAD: 03starseeker * r39777 10/brlcad/trunk/src/librt/primitives/metaball/metaball.c: Hmm - gentoo was complaining about a global being shadowed... make it mb_stat |
| 12:54.25 | ``Erik | ah, stat |
| 13:01.38 | CIA-93 | BRL-CAD: 03starseeker * r39778 10/brlcad/trunk/src/libged/red.c: Take a stab at quelling a few warnings in red.c |
| 13:02.06 | starseeker | anybody know what time the meeting is? |
| 13:02.41 | ``Erik | which, pbr? I think that's like 2:30 or so? |
| 13:02.47 | starseeker | ah, cool |
| 13:03.10 | starseeker | can see if is squashage of an if_ogl complaint works then |
| 13:03.21 | starseeker | s/is/his |
| 13:15.04 | starseeker | O.o uh, %p wants a (void **) cast in printf, not (void *)?? |
| 13:15.14 | starseeker | wonders if his compiler has gone loopy |
| 13:20.07 | ``Erik | yeah, trying for 2:30, "if enough people bother showing up" |
| 13:20.37 | ``Erik | %p is vprintf for "pointer", it should be void* O.o |
| 13:23.50 | starseeker | ../../../brlcad/src/libfb/if_ogl.c:1471: error: format â%pâ expects type âvoid **â, but argument 3 has type âvoid *â |
| 13:25.05 | ``Erik | typing for vprintf family stuff has to be a speecial case, what if you fake it by calling it (int *) or something and see if that works? *shrug* |
| 13:25.50 | starseeker | oh, and also ../../../brlcad/src/libfb/if_ogl.c:1878: error: comparison between signed and unsigned |
| 13:26.07 | ``Erik | there're lots of those that msvc shows :/ |
| 13:27.17 | starseeker | count is size_t there - do the libfb data structures need updating? |
| 13:33.53 | brlcad | that's not in printf |
| 13:34.00 | brlcad | that's in sscanf |
| 13:34.05 | brlcad | so the void ** is right |
| 13:34.13 | brlcad | it takes a pointer to the thing you want to set |
| 13:34.32 | brlcad | so if you're setting a void * address, you need a pointer to that void *, i.e. a void ** |
| 13:35.02 | brlcad | the argument is right (which is why it works), but the cast is wrong |
| 13:35.45 | CIA-93 | BRL-CAD: 03brlcad * r39779 10/brlcad/trunk/src/libfb/if_ogl.c: cast is wrong, it should be void ** |
| 13:38.04 | brlcad | yay, got red to crash |
| 13:39.45 | brlcad | starseeker: http://brlcad.org/tmp/red_crash.log |
| 13:39.47 | ``Erik | heh, noticed that, but ed walked in before I could say anythin |
| 13:42.36 | brlcad | looks like several sanity checks are missing down that stack (should test pointers before strcmp, shouldn't pass bad params from parent, etc |
| 13:49.29 | ``Erik | pbr's deferred until next week |
| 13:55.52 | starseeker | ``Erik: uh... I'm coming in... |
| 13:56.08 | ``Erik | yeah, but others aren't |
| 13:59.10 | starseeker | brlcad: auugh. OK... |
| 13:59.37 | ``Erik | <-- was looking forward to killing a couple cards and adding a couple more O.o |
| 14:19.42 | starseeker | would be satisfied to kill red at this point... |
| 14:19.45 | starseeker | alright, heading in |
| 14:27.47 | brlcad | configure: running /bin/sh ../misc/configure --prefix=/usr/brlcad/dev-7.16.9 '--disable-dependency-tracking' '--enable-all' '--enable-warnings' '--prefix=/usr/brlcad/dev-7.16.9' --enable-symbols --with-tcl="/vld/other/morrison/brlcad/.hermes//src/other/tcl/unix" --with-tk="/vld/other/morrison/brlcad/.hermes//src/other/tk/unix" --exec-prefix="/usr/brlcad/dev-7.16.9" --exec-prefix="/usr/brlcad/dev-7.16.9" --cache-file=../../../config.cache.linux-gnu.hermes.arl.ar |
| 14:27.55 | brlcad | /bin/sh: ../misc/configure: No such file or directory |
| 14:27.57 | brlcad | configure: error: /bin/sh ../misc/configure failed for src/other/tktable |
| 14:27.58 | brlcad | distcheck ALL+WARN |
| 14:55.38 | d-lo | I lol'd: http://www.theonion.com/articles/life-in-the-navy-rocks-even-harder-than-the-commer,11181/ |
| 15:10.02 | brlcad | heh, U.S.S. Abraham Linkin Park |
| 15:21.00 | ``Erik | huh, tron lightcycles for sale, 35k |
| 15:22.54 | starseeker | brlcad: arrgh - stop finding broken stuff that's my fault! :-P |
| 15:30.01 | *** join/#brlcad SWPadnos (~Me@emc/developer/SWPadnos) | |
| 15:30.33 | starseeker | ooooo - bad bug |
| 15:30.39 | starseeker | rendered the .g unloadable |
| 15:32.51 | CIA-93 | BRL-CAD: 03starseeker * r39780 10/brlcad/trunk/src/libged/red.c: eeep - bad programmer. Don't attempt to do things with empty strings in av pairs. |
| 15:53.45 | *** join/#brlcad parigaudi (~quassel@pd95b7f5e.dip0.t-ipconnect.de) | |
| 16:01.47 | brlcad | starseeker: very bad bug.. should investigate how specifically it's unloadable, that means there's a secondary bug in some other routine that left the DB in a bad state (which should NEVER happen) |
| 16:02.20 | brlcad | or if it's a bug in the scan where something is unloadable that should be loadable |
| 16:38.03 | *** join/#brlcad waprat (~packrator@99-67-225-40.lightspeed.livnmi.sbcglobal.net) | |
| 17:30.45 | starseeker | brlcad: crash was here: BU_ASSERT_PTR(cp+1==ep) failed, lhs=0xed9aea, rhs=0xed9b06, file ../../../brlcad/src/librt/attributes.c, line 73 |
| 17:31.17 | starseeker | I believe caused by an av pair with a value but no name |
| 17:33.55 | starseeker | ah - a "null" name string is supposed to indicate the end of the attribute list, which was then checked by BU_ASSERT_PTR - which of course failed, because a stray av pair with null name string was stuck in there |
| 17:35.19 | starseeker | hmm... possibly there should be a check i bu_avs_add to not add an av pair with zero length string for the name |
| 17:35.34 | starseeker | looks... kinda wonder why that's not already there... |
| 17:38.15 | starseeker | erm |
| 17:38.35 | starseeker | ok it's looking for !name - maybe should also check strlen? |
| 17:53.53 | CIA-93 | BRL-CAD: 03starseeker * r39781 10/brlcad/trunk/src/libbu/avs.c: Just checking for null in bu_avs_add may not be enough - if it's an empty non-null string, it's also meaningless. Try checking strlen |
| 17:59.10 | d-lo | brlcad: is there going to be a BRL-CAD BOF this year? |
| 18:11.48 | brlcad | d-lo: given the "delays", I hadn't decided yet, at least haven't registered one yet |
| 18:12.08 | brlcad | there's a lot on the advance program to start with |
| 18:13.01 | brlcad | lots of good justification to be found regardless |
| 19:04.16 | *** join/#brlcad PrezKennedy (~Prez@96.31.84.96) | |
| 19:17.37 | CIA-93 | BRL-CAD: 03erikgreenwald * r39782 10/brlcad/trunk/src/adrt/librender/camera.c: merge (copy) changes from rel8 r39741 |
| 19:19.00 | ``Erik | git, eh? O.o |
| 19:21.58 | ``Erik | http://vicclap.hu/static/media/201002/pic90054.jpg vuvuzela! |
| 20:05.02 | brlcad | wow, I actually finished |
| 20:05.10 | brlcad | I think my eyes are bleeding |
| 20:06.49 | starseeker | O.o |
| 20:20.53 | *** join/#brlcad R0b0t1 (~Enigma@unaffiliated/r0b0t1) | |
| 20:38.11 | *** join/#brlcad PrezKennedy (~Prez@96.31.84.96) | |
| 20:58.19 | ``Erik | hm, hotel reservation for siggraph: made. O.o |
| 21:02.10 | *** join/#brlcad Stattrav (~Stattrav@117.192.133.250) | |
| 21:30.59 | CIA-93 | BRL-CAD: 03r_weiss * r39783 10/brlcad/trunk/doc/docbook/system/man1/en/obj-g.xml: updating obj-g documentation |
| 21:40.39 | *** join/#brlcad Stattrav (~Stattrav@117.192.133.250) | |
| 21:45.56 | *** join/#brlcad R0b0t1 (~Enigma@unaffiliated/r0b0t1) | |
| 22:09.42 | *** join/#brlcad R0b0t1 (~Enigma@unaffiliated/r0b0t1) | |
| 23:47.30 | *** join/#brlcad yukonbob (~svs@S0106001125477e9c.ok.shawcable.net) | |
