| 01:44.54 | CIA-62 | BRL-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 |
| 02:03.23 | CIA-62 | BRL-CAD: 03kunigami * r45536 10/brlcad/trunk/src/liboptical/sh_osl.cpp: changed sh_osl to read shaders from ../shaders instead of ./ |
| 02:17.17 | CIA-62 | BRL-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 |
| 02:30.37 | CIA-62 | BRL-CAD: 0399.125.86.110 07http://brlcad.org * r3012 10/wiki/User:Bhinesley: /* Log */ wrong month on a few dates |
| 02:33.00 | CIA-62 | BRL-CAD: 0399.125.86.110 07http://brlcad.org * r3013 10/wiki/User:Bhinesley: /* Log */ overwrote a couple dates last change |
| 02:53.22 | CIA-62 | BRL-CAD: 0399.125.86.110 07http://brlcad.org * r3014 10/wiki/User:Bhinesley: /* Log */ catching up |
| 03:04.46 | CIA-62 | BRL-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) | |
| 13:20.35 | CIA-62 | BRL-CAD: 03erikgreenwald * r45538 10/brlcad/trunk/src/libged/edit.c: #if0 some unused funcs marked HIDDEN, causes compile failure when debugging is disabled |
| 13:41.09 | CIA-62 | BRL-CAD: 03brlcad * r45539 10/brlcad/trunk/ (include/bu.h src/libbu/file.c): |
| 13:41.09 | CIA-62 | BRL-CAD: initial implementation of bu_file_delete() for removing files. performs a |
| 13:41.09 | CIA-62 | BRL-CAD: simple remove() but then will try harder by relaxing the file permissions on a |
| 13:41.09 | CIA-62 | BRL-CAD: second pass attempt if the first fails. untested on windows but remove() is c90 |
| 13:41.09 | CIA-62 | BRL-CAD: so we should be able to rely on it. callers will just have to make sure the |
| 13:41.10 | CIA-62 | BRL-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) || | |
| 14:16.58 | brlcad | bah |
| 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! | |
| 14:17.58 | brlcad | all orgs get just one slot |
| 14:18.48 | __name__ | shouldn't this have been announced yesterday? |
| 14:19.56 | brlcad | they were a little late due to more submissions than they anticipated |
| 14:20.17 | brlcad | it was just announced a little bit ago |
| 14:22.09 | __name__ | Their FAQ still are not complete either. |
| 14:24.37 | brlcad | is a faq ever complete? |
| 14:24.52 | __name__ | They have questions without answers isted. |
| 14:24.55 | brlcad | unless you define the frequency, there will always be someone with a question |
| 14:24.55 | __name__ | *listed even |
| 14:25.01 | brlcad | ah, meh |
| 14:25.03 | brlcad | not really concerning -- they're pretty closely mirrored after gsoc |
| 14:25.21 | brlcad | limited time and resources, building it up as they go .. it is a pilot program after all |
| 14:25.40 | brlcad | any info is a plus, the whole thing could be done over private e-mail exchanges |
| 14:25.41 | __name__ | Yeah. |
| 14:25.54 | brlcad | or (shudder) phone calls |
| 14:26.34 | brlcad | I'm impressed they got 20 slots funded by their management |
| 14:27.02 | brlcad | roughly 100-200k pilot |
| 14:28.00 | ``Erik | heh, 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) | |
| 15:10.02 | bhinesley | ``Erik: I wonder why it doesn't fail for me when there are unused functions... I have debugging enabled |
| 15:10.37 | bhinesley | I don't even get a warning |
| 15:11.41 | bhinesley | smacks self awake |
| 15:12.22 | *** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav) | |
| 16:05.21 | brlcad | bhinesley: different versions of the same compiler will report different warnings, plus other compilation settings (such as optimized) can affect warnings |
| 16:07.46 | bhinesley | do you always build with strict on? It never works for me |
| 16:07.58 | ``Erik | with debugging OFF, the NDEBUG flag gets set, which turns HIDDEN into "static", otherwise it's defined as /* */ |
| 16:08.24 | ``Erik | that's where that issue came from |
| 16:08.52 | bhinesley | hmm |
| 16:11.09 | bhinesley | the 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 | ``Erik | stop making a mess? :D |
| 16:11.51 | bhinesley | I have to figure out how to identify a mess first |
| 16:11.54 | ``Erik | I like to have a 'build' directory with several subdirs for different configurations to spot those |
| 16:12.47 | bhinesley | okay, so you build with various options as a matter of course |
| 16:13.37 | ``Erik | yeah, and a variety of os/arch's, to boot |
| 16:14.35 | bhinesley | ok |
| 16:21.27 | brlcad | bhinesley: strict should always work -- if it doesn't, that's something that needs to be addressed |
| 16:21.48 | brlcad | should be building with strict enabled |
| 16:22.20 | brlcad | debug enabled and strict enabled -- optimized can be on or off |
| 16:23.22 | bhinesley | there are thousands upon thousands of warnings... I'm guessing that only certain types are treated as errors |
| 16:23.26 | bhinesley | ? |
| 16:24.13 | bhinesley | I will take a look at what is stopping me from building strict |
| 16:35.24 | brlcad | what are the first few? |
| 16:35.48 | brlcad | i'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) | |
| 16:59.28 | bhinesley | brlcad: 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) | |
| 17:01.57 | brlcad | bhinesley: what os and version of gcc? |
| 17:02.03 | brlcad | and cmake or autotools build? |
| 17:02.35 | brlcad | if you can produce a full build log, I can look into whether there is a categoric fix |
| 17:03.03 | bhinesley | Fedora 15, gcc 4.6.0 20110530, cmake |
| 17:03.08 | brlcad | I 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 |
| 17:03.17 | bhinesley | 64-bit |
| 17:03.23 | brlcad | definitely a pretty new gcc |
| 17:04.21 | brlcad | make -k 2>&1 | tee build.log |
| 17:05.01 | ``Erik | installs gcc47 on crit O.o |
| 17:05.19 | bhinesley | Okay, 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. |
| 17:06.59 | bhinesley | a lot of "{int ret; ret = somefunction();}" -> "{(void)somefunction();}" |
| 17:13.33 | bhinesley | http://db.tt/rITDTCH |
| 17:13.54 | bhinesley | brlcad ^ build log |
| 17:18.31 | CIA-62 | BRL-CAD: 03bhinesley * r45540 10/brlcad/trunk/src/libbu/ (bomb.c crashreport.c): removed unused variables and quiet compiler |
| 17:20.23 | bhinesley | more complex to fix: http://pastebin.mozilla.org/1276121 |
| 17:21.11 | bhinesley | I'm assuming we want to keep BU_LIST_APPEND's BU_ASSERT's |
| 17:27.52 | CIA-62 | BRL-CAD: 03bhinesley * r45541 10/brlcad/trunk/src/libbu/vlb.c: remove unused variable missed last commit |
| 17:32.20 | bhinesley | very quietly disables strict and goes back to work for now |
| 17:50.20 | brlcad | bhinesley: don't go too far down that route |
| 17:50.56 | brlcad | functions like read()/write() will thrown warnings with different versions of the compiler if you cast the return to (void) |
| 17:51.10 | bhinesley | ah |
| 17:51.35 | brlcad | the fact that the return value is being stashed in a var was to quiet earlier warnings |
| 17:51.46 | bhinesley | alright, I'll revert |
| 17:52.30 | brlcad | note that r45540 unveils two issues |
| 17:53.07 | brlcad | fwrite doesn't return an int, it returns a size_t |
| 17:54.43 | bhinesley | okay, so I'll fix it |
| 17:54.45 | brlcad | write 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"); |
| 17:55.14 | brlcad | just calling perror() preserves current behavior, just printing the issue |
| 17:55.51 | brlcad | (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) | |
| 19:23.15 | CIA-62 | BRL-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) | |
| 19:46.08 | brlcad | bhinesley: thought you said there were thousands? |
| 19:46.29 | brlcad | only see a dozen or so files in your log |
| 19:56.57 | bhinesley | I must have different options set. I did a `wc` of lines containing "error" a couple weeks ago, and there were thousands. |
| 19:57.07 | bhinesley | er "warning" |
| 19:59.31 | bhinesley | that's how I found the 511 warnings that I fixed with r45238 |
| 20:00.02 | bhinesley | (all regarding printf specifiers) |
| 20:03.15 | bhinesley | anyways, 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 >.< |
| 20:06.22 | brlcad | yeah, nope .. you're just ahead of us with your fancy new compilerness ;) |
| 20:06.33 | bhinesley | hah |
| 20:06.45 | brlcad | suprised starseeker hadn't hit the issues on gentoo |
| 20:09.06 | CIA-62 | BRL-CAD: 03brlcad * r45543 10/brlcad/trunk/src/libbu/crashreport.c: quell warning on || && logic, wrap latter in parens |
| 20:15.08 | CIA-62 | BRL-CAD: 03128.63.32.74 07http://brlcad.org * r3016 10/wiki/Community_Publication_Portal: stub in ESA SOCIS announcement |
| 20:19.56 | CIA-62 | BRL-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. |
| 20:47.30 | brlcad | starseeker: 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 |
| 20:47.54 | brlcad | the magic check was probably what changed, no idea how it ever would have worked as it is now |
| 20:48.28 | brlcad | csgbrep.cpp:261 is where it goes wrong |
| 20:52.41 | CIA-62 | BRL-CAD: 03brlcad * r45545 10/brlcad/trunk/src/proc-db/csgbrep.cpp: |
| 20:52.41 | CIA-62 | BRL-CAD: looks like this is calling the wrong function table entry. it's setting an NMG |
| 20:52.41 | CIA-62 | BRL-CAD: as the object pointer, but was trying to call the ARBN functab. tripped a bomb |
| 20:52.41 | CIA-62 | BRL-CAD: magic detection. calling the NMG converter seems to work in a better non-crashy |
| 20:52.41 | CIA-62 | BRL-CAD: way. |
| 21:08.42 | starseeker | brlcad: ah, cool - thanks |
| 21:31.14 | *** join/#brlcad merzo (~merzo@48-81-132-95.pool.ukrtel.net) | |
| 21:31.56 | CIA-62 | BRL-CAD: 03kunigami * r45546 10/brlcad/trunk/src/liboptical/ (liboslrend.cpp liboslrend.h sh_osl.cpp): Added support for setting string parameters to OSL shaders |
| 21:41.03 | CIA-62 | BRL-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) | |
| 22:03.48 | CIA-62 | BRL-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) | |
| 22:48.17 | CIA-62 | BRL-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] |
| 23:14.21 | CIA-62 | BRL-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 |
| 23:14.24 | CIA-62 | BRL-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]]') |
| 23:14.28 | CIA-62 | BRL-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 |
| 23:14.33 | CIA-62 | BRL-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]]') |
| 23:24.26 | CIA-62 | BRL-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 |
| 23:25.09 | kunigami_ | Testing the texture shader: http://dl.dropbox.com/u/1399996/GSoC/osl_texture.png |
| 23:39.04 | brlcad | heh, nifty |
| 23:39.22 | brlcad | what happened to the light source, though? |
| 23:40.26 | brlcad | illuminating 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) |
| 23:51.14 | kunigami_ | 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 |
| 23:51.30 | brlcad | kunigami_: so you mentioned earlier that you have to hypersample 1000 times .. that just sounds wrong for many reasons.. how is hypersampling being used? |
| 23:53.01 | kunigami_ | the image above was hypersamples 1000 times too. I'm just using -H 1000 -J3 |
| 23:53.13 | brlcad | should work without any hypersampling, just perhaps not as well converged global light (e.g., corners not quite as dark as they should be) |
| 23:53.25 | brlcad | right, but .. why? :) |
| 23:53.38 | brlcad | what does non -H1000 look like? |
| 23:55.22 | kunigami_ | brlcad: http://dl.dropbox.com/u/1399996/GSoC/no-hypersampling.png |
| 23:55.55 | brlcad | so can you explain that? |
| 23:56.43 | brlcad | every pixel has a ray being fired, so why wouldn't they all return a color for a primary hit? |
| 23:57.07 | kunigami_ | 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? |
| 23:57.24 | kunigami_ | a large number of samples so the color converges |
| 23:58.58 | brlcad | kunigami_: returning a random reflection direction sounds like a controllable parameter though, no? |
| 23:59.41 | brlcad | it's the shader's job to determine whether there is reflection in the first place |
| 23:59.50 | kunigami_ | hm not sure. I'd have to find out |
