IRC log for #brlcad on 20120413

00:43.02CIA-128BRL-CAD: 03tbrowder2 * r49990 10/brlcad/trunk/regress/ (CMakeLists.txt comgeom/m35.g.gift5.txt comgeom.sh): tidy regress-comgeom test; remove old and now unused input test file
01:46.07CIA-128BRL-CAD: 03starseeker * r49991 10/brlcad/branches/STABLE/src/libdm/dm-ogl.c: Pull in bu_log statement change from trunk
02:04.24CIA-128BRL-CAD: 03starseeker * r49992 10/brlcad/branches/STABLE/src/libged/analyze.c: Backport fix for MSVC from trunk
02:12.52CIA-128BRL-CAD: 03starseeker * r49993 10/brlcad/branches/STABLE/src/rt/viewweight.c: Backport another MSVC fix from trunk
03:33.03*** join/#brlcad juan_man (~quassel@unaffiliated/juanman)
03:40.56*** join/#brlcad Neil___ (~chatzilla@117.228.195.122)
03:47.18*** join/#brlcad stevegt_1 (~stevegt@64-130-244-2-gracehospitalitygroup.bwg.phx.sparkplugbb.net)
05:03.16*** join/#brlcad stevegt_1 (~stevegt@64-130-244-2-gracehospitalitygroup.bwg.phx.sparkplugbb.net)
05:03.45*** join/#brlcad juanman (~quassel@unaffiliated/juanman)
06:44.17*** join/#brlcad andrei_ (~tuxilina@79.118.113.94)
06:44.56*** join/#brlcad andrei_ (~tuxilina@79.118.113.94)
08:25.40*** join/#brlcad stas (~stas@82.208.133.12)
08:32.23*** join/#brlcad jordisayol (~jordisayo@unaffiliated/jordisayol)
08:33.30*** join/#brlcad juan_man (~quassel@unaffiliated/juanman)
08:59.18*** join/#brlcad Neil___ (~chatzilla@117.229.46.253)
09:39.28*** join/#brlcad Neil___ (~chatzilla@117.229.107.190)
11:13.05*** join/#brlcad ``Erik (~erik@pool-108-3-159-149.bltmmd.fios.verizon.net)
11:39.30CIA-128BRL-CAD: 03tbrowder2 * r49994 10/brlcad/trunk/include/bu.h: update citation: expand author's first names, correct title, add note on availability of 3rd ed.
11:42.05CIA-128BRL-CAD: 03tbrowder2 * r49995 10/brlcad/trunk/include/bu.h: expand note
11:51.01*** join/#brlcad Al_Da_Best (~Al_Da_Bes@027e71f6.bb.sky.com)
11:58.27CIA-128BRL-CAD: 03tbrowder2 * r49996 10/brlcad/trunk/include/bu.h: expand note further highlighting possible errors in impementation
11:59.23CIA-128BRL-CAD: 03tbrowder2 * r49997 10/brlcad/trunk/TODO: add need for checking red-black tree implementation
12:22.24*** join/#brlcad Neil___ (~chatzilla@117.229.23.121)
13:52.29CIA-128BRL-CAD: 03203.110.247.221 07http://brlcad.org * r3504 10/wiki/Contributor_Quickies: /* HARD: Fix Bounding Box function for BoT primitive */
14:09.20*** join/#brlcad juanman (~quassel@unaffiliated/juanman)
14:17.03*** join/#brlcad juan_man (~quassel@unaffiliated/juanman)
14:24.34CIA-128BRL-CAD: 0368.34.98.23 07http://brlcad.org * r3505 10/wiki/Contributor_Quickies: Undo revision 3504 by [[Special:Contributions/203.110.247.221|203.110.247.221]] ([[User talk:203.110.247.221|Talk]])
14:26.58CIA-128BRL-CAD: 03Sean 07http://brlcad.org * r0 10/wiki/Special:Log/block: blocked [[User:203.110.247.221]] with an expiry time of 1 year (anonymous users only, account creation disabled): Removing content from pages
14:54.18brlcadstarseeker: there's a slim chance that the same attribute bug may have affected the other schemas
14:54.27brlcadsee if 203e2 works
16:15.06*** join/#brlcad jordisayol (~jordisayo@unaffiliated/jordisayol)
16:16.03CIA-128BRL-CAD: 03tbrowder2 * r49998 10/brlcad/trunk/include/bu.h: fix typo in quote
17:55.47*** join/#brlcad andrei_ (~andrei_@79.119.90.25)
17:57.02andrei_hello
18:23.25CIA-128BRL-CAD: 03erikgreenwald * r49999 10/brlcad/trunk/src/librt/shoot.c: start filling in a little bit of the rt_shootray_simple hit func
18:23.49CIA-128BRL-CAD: 03erikgreenwald * r50000 10/brlcad/trunk/TODO: bump bottie fixes out of this release cycle, they shouldn't be in any critical paths just yet
18:25.43brlcad``Erik: er, using libbu structures without libbu memory management?  that can cause memory problems
18:26.33brlcadrather, librt structures
18:28.05brlcadyou can free() a bu_malloc() as it just messes up the book-keeping, but I don't think you can safely bu_free() a malloc()
18:29.39``Erikyeh, still needs more filling, was just a quick commit to get 50k O:-) *duck*
18:30.04brlcadahhhh... woo hoo
18:30.14brlcadfeels like there should be a party or something
18:31.10brlcadmaybe a good time to recreate the gource visualization and post an announcement
18:31.18``Erikthere was a party, ya missed it
18:31.21``Erikputs his pants back on
18:31.23brlcaduh huh
18:31.29CIA-128BRL-CAD: 03erikgreenwald * r50001 10/brlcad/trunk/src/librt/shoot.c: malloc->bu_malloc. Mention that it adds malloc syscalls, so is not fast.
18:31.38brlcadI wouldn't say I "missed" it
18:32.10brlcadcomment is wrong ;)
18:32.12CIA-128BRL-CAD: 03starseeker * r50002 10/brlcad/trunk/ (5 files in 5 dirs):
18:32.12CIA-128BRL-CAD: Per user request, at a way to indicate to CMake that it should build either a)
18:32.12CIA-128BRL-CAD: just what is needed for librt b) just what is needed for the libraries sans
18:32.12CIA-128BRL-CAD: executables or c) everything (default) via the BRLCAD_ENABLE_TARGETS variable
18:32.44starseekerbah - s/at/add/
18:32.53``Erikeh?
18:33.37brlcadmalloc()
18:33.42starseeker50002 -> the commit that *should* have been 50000 :-P
18:35.05``Erikshould have been 49999, you mean
18:35.21starseekeroh, i would have found something small to commit first :-P
18:39.16*** join/#brlcad juanman (~quassel@unaffiliated/juanman)
18:42.06CIA-128BRL-CAD: 03erikgreenwald * r50003 10/brlcad/trunk/src/librt/shoot.c: add some more copy bits to the shootray_simple hit func
18:43.45andrei_brlcad , as I you have mentioned I did spend some time looking into the rb tree api
18:43.51andrei_is there any way I could help?
18:45.55CIA-128BRL-CAD: 03starseeker * r50004 10/brlcad/trunk/CMakeLists.txt: fix it so the documentation summary line doesn't lie when we're only building libraries
18:46.00*** join/#brlcad juan_man (~quassel@unaffiliated/juanman)
18:46.13*** join/#brlcad stas (~stas@188.24.35.114)
18:47.50brlcadandrei_: sure!
18:48.19brlcadyou could read the link browder posted to see what the issue was, then devise a way to test whether our implementation has that flaw
18:49.06brlcadit's almost certainly not as simple as "delete didn't work" .. but more likely didn't work under some given conditions
18:49.41andrei_alright, I will look into it tomorrow
18:49.45brlcadtheir update or errata may say what those conditions were, and a test could be devised
18:49.50brlcadawesome
18:50.14``Erikheh, src/libbu/test_rbtree.c seems strangely absent :D
18:57.13andrei_Erik, as far as I remember there wasn't a test_rbtree
18:57.26andrei_at least not in the source I downloaded
18:57.35brlcad``Erik: patches tracker
18:57.48brlcadit's not yet committed
18:58.45brlcadandrei_'s unit test included testing delete and it succeeded, so presumably the failure is either some specific condition or our implementation doesn't have the flaw
18:59.26andrei_perhaps this is also an occasion to improve the unit test
18:59.30``Eriki've started using ctest in the libwa project I'm doing, cmake's ctest ENABLE_TESTING() ADD_TEST() stuff, would require some reworking of our test programs to use it well, though... (relies solely on program return value)
19:02.50``Erikmight be worth discussing if that's the angle we want to take
19:04.48brlcadshrug, I'm not enamored by ctest
19:05.09brlcadas far as testing frameworks go, that was is rather mediocre
19:05.44brlcadit's fine, but not much value (to me) vs the simplicity of just writing a program that tests what you want
19:05.52``Erikcmake is wired for it, though, I wasn't able to override the 'test' target to do my own (using libcheck)
19:06.05brlcadso
19:06.11brlcadthat wiring is dirt cheap
19:09.18brlcadthe setup and maintenance complexity offsets the nominal benefits
19:10.00brlcadmaybe if we had a lot more tests or if it just had some other feature(s) to make it actually save time vs the simplicity of the alternative (what we have)
19:11.00brlcadnobody has to learn anything to write a test right now, you just write a program as simple/complex as you like/need/want
19:11.45``Erikkinda the same for ctest, it really is int main() { if(test_succeeds()) return 0; else return 1; }
19:12.42``Erikmy approach has been to pass test values in via argc and define all the test cases in the CMakeLists.txt
19:13.23``Erikthe only real arguments for ctest are "it's the cmake way" and consistency
19:21.20brlcadexcept I had to learn test_succeeds() and whatever other prerequisites
19:22.54brlcadmuch simpler and less complexity to maintain to just write a program that does what I need
19:23.08brlcadno limitation of return values or data
19:27.48``Eriktest_succeeds() is the bit you write
19:28.33``Erikthe only thing in the C (or whatever) program to communicate back to the framework is the return value from main(), 0 succeeds and non-zero fails
19:28.46``Erik(so LOTS of forks and execs)
19:29.14brlcadsee already confusing ;)
19:29.22brlcadyou shouldn't have to explain it
19:30.33``Erikheh, ok, "int main() { return 0; }" is a test that passes. the part that tripped me up was that I was looking for something to learn :/
19:33.09``Erikecho "exit 0" > blah.tcl && echo "ADD_TEST(sometcltest bin/tclsh blah.tcl)" >> CMakeLists.txt && make all test
19:34.39``Erikthe fu of ctest is all in the wiring, it aggregates the results and generates a report at the end
19:34.56brlcadnods
19:35.16brlcadand that's my point .. that little tidbit doesn't seem to be "enough" .. doesn't pull its long-term weight
19:36.29brlcadthe main feature is basically the summary report/output
19:36.31``Erikat the moment, we have no aggregation, just 'all stop' if one thing fails
19:36.57brlcadthat's good, have to fix that failure ;)
19:37.27``ErikI'm under the impression (and starseeker might know something about this, if he wants to chime in), it's a short hop to get to a red/green dashboard webpage or email report
19:39.25brlcadit is
19:39.43brlcadstill, that's all so weak compared to some of the other frameworks
19:40.17brlcadit's basically just/only the report/output/summary that it offers (whether text/web or otherwise)
19:40.32andrei_I am not sure if it it's a good idea
19:40.45andrei_but wouldn't valgrind help on this issue?
19:41.26brlcadandrei_: what issue?
19:41.41andrei_detecting the delete bug?
19:42.09andrei_I haven't managed to read all up tho :(
19:42.12andrei_so I might be misinformed
19:43.19andrei_I ll read everything that has been said and come back after :)
19:44.05brlcadI doubt it'll help
19:45.59brlcadvalgrind would be a memory issue (i.e., an implementation detail issue)
19:46.17brlcadthe book implied it was an algorithmic problem
19:48.40``Erikstarseeker: vdeck-comgeom-g.sh
19:49.44starseekerah - yeah, Tom did some updates to filenames with the comgeom regression
19:49.56``Erikregenerating from a clean dir now
20:02.02``Erikstarseeker: http://paste.lisp.org/display/128945
20:10.26*** join/#brlcad Neil___ (~chatzilla@117.229.59.196)
20:16.16``Erik*grumble* very well might be an NFS issue, yet again
20:17.45starseekerheh
20:19.15``Erikhttp://www.masswerk.at/googleBBS/  google as an 80's BBS, awesome
20:28.24CIA-128BRL-CAD: 03erikgreenwald * r50005 10/brlcad/trunk/src/librt/shoot.c: initialize c to avoid the uninitialized warning on older gccs
21:24.42*** join/#brlcad juanman (~quassel@unaffiliated/juanman)
21:36.32*** join/#brlcad Neil___ (~chatzilla@117.229.121.43)
22:18.11brlcadstarseeker: looks like the INSTALL file validation isn't matching the exact option names
22:18.31*** join/#brlcad stevegt_1 (~stevegt@64-130-244-2-gracehospitalitygroup.bwg.phx.sparkplugbb.net)
22:58.49*** join/#brlcad stevegt_1 (~stevegt@64-130-244-2-gracehospitalitygroup.bwg.phx.sparkplugbb.net)
23:28.56CIA-128BRL-CAD: 03tbrowder2 * r50006 10/brlcad/trunk/src/conv/g-nff.c: ws, style
23:46.16*** join/#brlcad stas (~stas@188.24.35.114)
23:59.59brlcadactually, looks like some other issue .. just not enabling optimized

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