IRC log for #brlcad on 20110719

01:44.54CIA-62BRL-CAD: 03kunigami * r45535 10/brlcad/trunk/src/other/ (4 files in 3 dirs): Added cmakefile to compile the .osl shaders into .oso ones. This is done only if the OSL flag is enabled
02:03.23CIA-62BRL-CAD: 03kunigami * r45536 10/brlcad/trunk/src/liboptical/sh_osl.cpp: changed sh_osl to read shaders from ../shaders instead of ./
02:17.17CIA-62BRL-CAD: 03bhinesley * r45537 10/brlcad/trunk/src/libged/edit.c: adding generic subargument handling to ged_edit; it's a work in progress. incomplete sections are disabled
02:30.37CIA-62BRL-CAD: 0399.125.86.110 07http://brlcad.org * r3012 10/wiki/User:Bhinesley: /* Log */ wrong month on a few dates
02:33.00CIA-62BRL-CAD: 0399.125.86.110 07http://brlcad.org * r3013 10/wiki/User:Bhinesley: /* Log */ overwrote a couple dates last change
02:53.22CIA-62BRL-CAD: 0399.125.86.110 07http://brlcad.org * r3014 10/wiki/User:Bhinesley: /* Log */ catching up
03:04.46CIA-62BRL-CAD: 0399.125.86.110 07http://brlcad.org * r3015 10/wiki/User:Bhinesley: /* Log */ cross out completed items
03:57.06*** join/#brlcad louipc (~louipc@archlinux/trusteduser/louipc)
04:38.51*** join/#brlcad epileg (~epileg@unaffiliated/epileg)
05:53.12*** join/#brlcad poolio (~poolio@BZ.BZFLAG.BZ)
07:14.06*** join/#brlcad merzo (~merzo@193.254.217.44)
07:57.17*** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav)
08:34.05*** join/#brlcad yAdam (~chatzilla@188.147.179.132.nat.umts.dynamic.t-mobile.pl)
10:09.28*** join/#brlcad yAdam (~chatzilla@188.147.201.90.nat.umts.dynamic.t-mobile.pl)
10:15.46*** join/#brlcad yAdam (~chatzilla@188.147.201.90.nat.umts.dynamic.t-mobile.pl)
12:52.35*** join/#brlcad yAdam (~chatzilla@188.146.24.10.nat.umts.dynamic.t-mobile.pl)
13:03.29*** part/#brlcad yAdam (~chatzilla@188.146.24.10.nat.umts.dynamic.t-mobile.pl)
13:20.35CIA-62BRL-CAD: 03erikgreenwald * r45538 10/brlcad/trunk/src/libged/edit.c: #if0 some unused funcs marked HIDDEN, causes compile failure when debugging is disabled
13:41.09CIA-62BRL-CAD: 03brlcad * r45539 10/brlcad/trunk/ (include/bu.h src/libbu/file.c):
13:41.09CIA-62BRL-CAD: initial implementation of bu_file_delete() for removing files. performs a
13:41.09CIA-62BRL-CAD: simple remove() but then will try harder by relaxing the file permissions on a
13:41.09CIA-62BRL-CAD: second pass attempt if the first fails. untested on windows but remove() is c90
13:41.09CIA-62BRL-CAD: so we should be able to rely on it. callers will just have to make sure the
13:41.10CIA-62BRL-CAD: file isn't opened.
13:49.45*** join/#brlcad yAdam (~chatzilla@188.146.24.10.nat.umts.dynamic.t-mobile.pl)
13:49.58*** part/#brlcad yAdam (~chatzilla@188.146.24.10.nat.umts.dynamic.t-mobile.pl)
14:16.51*** topic/#brlcad by brlcad -> BRL-CAD Open Source Solid Modeling || http://brlcad.org || http://sf.net/projects/brlcad || #brlcad logs: http://ibot.rikers.org/%23brlcad/ || BRL-CAD release 7.20.2 is posted (20110701) || BRL-CAD is BRL-CAD Open Source Solid Modeling || http://brlcad.org || http://sf.net/projects/brlcad || #brlcad logs: http://ibot.rikers.org/%23brlcad/ || BRL-CAD release 7.20.2 is posted (20110701) ||
14:16.58brlcadbah
14:17.34*** topic/#brlcad by brlcad -> BRL-CAD Open Source Solid Modeling || http://brlcad.org || http://sf.net/projects/brlcad || #brlcad logs: http://ibot.rikers.org/%23brlcad/ || BRL-CAD release 7.20.2 is posted (20110701) || BRL-CAD is participating in the ESA Summer of Code in Space!
14:17.58brlcadall orgs get just one slot
14:18.48__name__shouldn't this have been announced yesterday?
14:19.56brlcadthey were a little late due to more submissions than they anticipated
14:20.17brlcadit was just announced a little bit ago
14:22.09__name__Their FAQ still are not complete either.
14:24.37brlcadis a faq ever complete?
14:24.52__name__They have questions without answers isted.
14:24.55brlcadunless you define the frequency, there will always be someone with a question
14:24.55__name__*listed even
14:25.01brlcadah, meh
14:25.03brlcadnot really concerning -- they're pretty closely mirrored after gsoc
14:25.21brlcadlimited time and resources, building it up as they go .. it is a pilot program after all
14:25.40brlcadany info is a plus, the whole thing could be done over private e-mail exchanges
14:25.41__name__Yeah.
14:25.54brlcador (shudder) phone calls
14:26.34brlcadI'm impressed they got 20 slots funded by their management
14:27.02brlcadroughly 100-200k pilot
14:28.00``Erikheh, a faq with no answers would be awesome "look, it's a faq, not an atfaq! just the questions!" :D
14:29.16__name__Hah
14:33.33*** join/#brlcad starseeker (~starseeke@BZ.BZFLAG.BZ)
15:10.02bhinesley``Erik: I wonder why it doesn't fail for me when there are unused functions... I have debugging enabled
15:10.37bhinesleyI don't even get a warning
15:11.41bhinesleysmacks self awake
15:12.22*** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav)
16:05.21brlcadbhinesley: different versions of the same compiler will report different warnings, plus other compilation settings (such as optimized) can affect warnings
16:07.46bhinesleydo you always build with strict on? It never works for me
16:07.58``Erikwith debugging OFF, the NDEBUG flag gets set, which turns HIDDEN into "static", otherwise it's defined as /* */
16:08.24``Erikthat's where that issue came from
16:08.52bhinesleyhmm
16:11.09bhinesleythe question I really mean to ask is: what should I be doing differently, so that you guys don't have to follow me around to clean up my mess? :)
16:11.27``Erikstop making a mess? :D
16:11.51bhinesleyI have to figure out how to identify a mess first
16:11.54``ErikI like to have a 'build' directory with several subdirs for different configurations to spot those
16:12.47bhinesleyokay, so you build with various options as a matter of course
16:13.37``Erikyeah, and a variety of os/arch's, to boot
16:14.35bhinesleyok
16:21.27brlcadbhinesley: strict should always work -- if it doesn't, that's something that needs to be addressed
16:21.48brlcadshould be building with strict enabled
16:22.20brlcaddebug enabled and strict enabled -- optimized can be on or off
16:23.22bhinesleythere are thousands upon thousands of warnings... I'm guessing that only certain types are treated as errors
16:23.26bhinesley?
16:24.13bhinesleyI will take a look at what is stopping me from building strict
16:35.24brlcadwhat are the first few?
16:35.48brlcadi'm guessing that it's something like printf specifier conversions
16:38.04*** join/#brlcad Stattrav (~Stattrav@117.192.157.143)
16:38.04*** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav)
16:54.02*** join/#brlcad alex_jon1 (~alex_joni@81.196.65.201)
16:59.28bhinesleybrlcad: lots of printf specifier conversions, yes. Also, a lot of variables not set or not used (these are stopping me from building strict).
16:59.43*** join/#brlcad kunigami_ (~kunigami@201.53.206.27)
17:01.57brlcadbhinesley: what os and version of gcc?
17:02.03brlcadand cmake or autotools build?
17:02.35brlcadif you can produce a full build log, I can look into whether there is a categoric fix
17:03.03bhinesleyFedora 15, gcc 4.6.0 20110530, cmake
17:03.08brlcadI suspect you're just using a version of gcc newer than everyone else or on a curious platform with odd 32-bit/64-bit size mixtures
17:03.17bhinesley64-bit
17:03.23brlcaddefinitely a pretty new gcc
17:04.21brlcadmake -k 2>&1 | tee build.log
17:05.01``Erikinstalls gcc47 on crit O.o
17:05.19bhinesleyOkay, I'm running that. It seems that the unused variables are primarily what is preventing build, so I'll see how far the rabbit hole goes in the meantime.
17:06.59bhinesleya lot of "{int ret; ret = somefunction();}" -> "{(void)somefunction();}"
17:13.33bhinesleyhttp://db.tt/rITDTCH
17:13.54bhinesleybrlcad ^ build log
17:18.31CIA-62BRL-CAD: 03bhinesley * r45540 10/brlcad/trunk/src/libbu/ (bomb.c crashreport.c): removed unused variables and quiet compiler
17:20.23bhinesleymore complex to fix: http://pastebin.mozilla.org/1276121
17:21.11bhinesleyI'm assuming we want to keep BU_LIST_APPEND's BU_ASSERT's
17:27.52CIA-62BRL-CAD: 03bhinesley * r45541 10/brlcad/trunk/src/libbu/vlb.c: remove unused variable missed last commit
17:32.20bhinesleyvery quietly disables strict and goes back to work for now
17:50.20brlcadbhinesley: don't go too far down that route
17:50.56brlcadfunctions like read()/write() will thrown warnings with different versions of the compiler if you cast the return to (void)
17:51.10bhinesleyah
17:51.35brlcadthe fact that the return value is being stashed in a var was to quiet earlier warnings
17:51.46bhinesleyalright, I'll revert
17:52.30brlcadnote that r45540 unveils two issues
17:53.07brlcadfwrite doesn't return an int, it returns a size_t
17:54.43bhinesleyokay, so I'll fix it
17:54.45brlcadwrite and fwrite return values are compared against the number of values written, write is an int, fwrite a size_t, both if (ret != nvals) perror("fwrite/write failed");
17:55.14brlcadjust calling perror() preserves current behavior, just printing the issue
17:55.51brlcad(as a two-liner, not one-liner)
17:57.15*** join/#brlcad merzo (~merzo@48-81-132-95.pool.ukrtel.net)
18:06.42*** join/#brlcad Stattrav (~Stattrav@117.192.157.143)
18:06.42*** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav)
18:33.35*** join/#brlcad KimK (~Kim__@209.248.147.2.nw.nuvox.net)
19:23.15CIA-62BRL-CAD: 03bhinesley * r45542 10/brlcad/trunk/src/libbu/ (bomb.c crashreport.c vlb.c): resolved issues regarding fwrite/write return value validation, unveiled by r45540/r45541 per conversation with Sean
19:42.26*** join/#brlcad KimK_afk (~Kim__@209.248.147.2.nw.nuvox.net)
19:46.08brlcadbhinesley: thought you said there were thousands?
19:46.29brlcadonly see a dozen or so files in your log
19:56.57bhinesleyI must have different options set. I did a `wc` of lines containing "error" a couple weeks ago, and there were thousands.
19:57.07bhinesleyer "warning"
19:59.31bhinesleythat's how I found the 511 warnings that I fixed with r45238
20:00.02bhinesley(all regarding printf specifiers)
20:03.15bhinesleyanyways, it seems that I misunderstood the meaning of strict; I thought that *all* warnings were treated as errors. When I saw so many warnings, I assumed that compiling w/ strict was a goal, not a reality. Then I noticed that people were finding problems with my code a little too fast >.<
20:06.22brlcadyeah, nope .. you're just ahead of us with your fancy new compilerness ;)
20:06.33bhinesleyhah
20:06.45brlcadsuprised starseeker hadn't hit the issues on gentoo
20:09.06CIA-62BRL-CAD: 03brlcad * r45543 10/brlcad/trunk/src/libbu/crashreport.c: quell warning on || && logic, wrap latter in parens
20:15.08CIA-62BRL-CAD: 03128.63.32.74 07http://brlcad.org * r3016 10/wiki/Community_Publication_Portal: stub in ESA SOCIS announcement
20:19.56CIA-62BRL-CAD: 03brlcad * r45544 10/brlcad/trunk/ (NEWS src/mged/mged.c src/mged/setup.c): now that the ged struct is fully initialized during ged_init(), it exposed a bug where code wasn't initializing the output handler properly after opening a database. this restores rt command output.
20:47.30brlcadstarseeker: the csgbrep bomb is valid -- the arbn block creates an nmgmodel and tries to pass it forward as an rt_arbn_internal, so it bombs
20:47.54brlcadthe magic check was probably what changed, no idea how it ever would have worked as it is now
20:48.28brlcadcsgbrep.cpp:261 is where it goes wrong
20:52.41CIA-62BRL-CAD: 03brlcad * r45545 10/brlcad/trunk/src/proc-db/csgbrep.cpp:
20:52.41CIA-62BRL-CAD: looks like this is calling the wrong function table entry. it's setting an NMG
20:52.41CIA-62BRL-CAD: as the object pointer, but was trying to call the ARBN functab. tripped a bomb
20:52.41CIA-62BRL-CAD: magic detection. calling the NMG converter seems to work in a better non-crashy
20:52.41CIA-62BRL-CAD: way.
21:08.42starseekerbrlcad: ah, cool - thanks
21:31.14*** join/#brlcad merzo (~merzo@48-81-132-95.pool.ukrtel.net)
21:31.56CIA-62BRL-CAD: 03kunigami * r45546 10/brlcad/trunk/src/liboptical/ (liboslrend.cpp liboslrend.h sh_osl.cpp): Added support for setting string parameters to OSL shaders
21:41.03CIA-62BRL-CAD: 03Intmiti 07http://brlcad.org * r3017 10/wiki/Videopoker_is_definately_the_best_casino_games: New page: look for for online casinos on [http://www.google.com Google]
21:41.55*** join/#brlcad epileg (~epileg@unaffiliated/epileg)
22:03.48CIA-62BRL-CAD: 03bhinesley * r45547 10/brlcad/trunk/src/libged/edit.c: change a labeled block to a private function
22:27.16*** join/#brlcad merzo (~merzo@48-81-132-95.pool.ukrtel.net)
22:48.17CIA-62BRL-CAD: 03Antlipi 07http://www.solidgeometry.org * r3018 10/wiki/Blackjack_is_without_any_doubts_the_best_casino_games: New page: look for for ambling on [http://www.google.com Google]
23:14.21CIA-62BRL-CAD: 03Sean 07http://brlcad.org * r0 10/wiki/Special:Log/block: blocked [[User:Antlipi]] with an expiry time of infinite (account creation disabled, e-mail blocked): Spamming links to external sites
23:14.24CIA-62BRL-CAD: 03Sean 07http://brlcad.org * r0 10/wiki/Special:Log/delete: deleted "[[Blackjack is without any doubts the best casino games]]": content was: 'look for for ambling on [http://www.google.com Google]' (and the only contributor was '[[Special:Contributions/Antlipi|Antlipi]]')
23:14.28CIA-62BRL-CAD: 03Sean 07http://brlcad.org * r0 10/wiki/Special:Log/block: blocked [[User:Intmiti]] with an expiry time of infinite (account creation disabled, e-mail blocked): Spamming links to external sites
23:14.33CIA-62BRL-CAD: 03Sean 07http://brlcad.org * r0 10/wiki/Special:Log/delete: deleted "[[Videopoker is definately the best casino games]]": content was: 'look for for online casinos on [http://www.google.com Google]' (and the only contributor was '[[Special:Contributions/Intmiti|Intmiti]]')
23:24.26CIA-62BRL-CAD: 03kunigami * r45548 10/brlcad/trunk/src/other/osl/shaders/ (converter.osl sh_texture.osl): texture shader. this one was pretty straightforward since there's already an implemented internal function
23:25.09kunigami_Testing the texture shader: http://dl.dropbox.com/u/1399996/GSoC/osl_texture.png
23:39.04brlcadheh, nifty
23:39.22brlcadwhat happened to the light source, though?
23:40.26brlcadilluminating the left corner but not the right just from the reflective box?  if so, seems like there's some energy loss (I'd expect it to be brighter given previous images)
23:51.14kunigami_brlcad: in previous images I was multiplying the colors by a factor of 2 or 3 because the scene was too dark. Now I'm just using a very very bright light. I'm not sure if that's the cause of difference
23:51.30brlcadkunigami_: so you mentioned earlier that you have to hypersample 1000 times .. that just sounds wrong for many reasons.. how is hypersampling being used?
23:53.01kunigami_the image above was hypersamples 1000 times too. I'm just using -H 1000 -J3
23:53.13brlcadshould work without any hypersampling, just perhaps not as well converged global light (e.g., corners not quite as dark as they should be)
23:53.25brlcadright, but .. why? :)
23:53.38brlcadwhat does non -H1000 look like?
23:55.22kunigami_brlcad: http://dl.dropbox.com/u/1399996/GSoC/no-hypersampling.png
23:55.55brlcadso can you explain that?
23:56.43brlcadevery pixel has a ray being fired, so why wouldn't they all return a color for a primary hit?
23:57.07kunigami_the osl system returns a random reflection direction everytime I make a query (biased depending on the shader), so it's necessary
23:57.19``Erik-H only effects the primary ray, didn't the path tracer and photon mapper twingy did get a HUGE benefit by multiplying secondaries instead of primaries?
23:57.24kunigami_a large number of samples so the color converges
23:58.58brlcadkunigami_: returning a random reflection direction sounds like a controllable parameter though, no?
23:59.41brlcadit's the shader's job to determine whether there is reflection in the first place
23:59.50kunigami_hm not sure. I'd have to find out

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