IRC log for #stepcode on 20130828

17:38.39*** join/#stepcode kesha (~kesha@49.249.17.122)
18:33.01*** join/#stepcode kesha_ (~kesha@49.249.17.122)
19:24.55*** join/#stepcode kesha_ (~kesha@49.249.17.122)
19:38.52kesha_hi brlcad
19:53.46kesha_brlcad: there /?
20:00.42*** join/#stepcode mpictor (~mark@2601:d:b280:3d4:d63d:7eff:fe2d:2505)
20:27.16kesha_hi mpictor
20:27.39kesha_https://github.com/stepcode/stepcode/pull/254
20:27.57mpictorhi kesha_
20:28.04mpictorI was just looking at that
20:28.19kesha_cool :)
20:29.17mpictorlines like "DESTINATION include/stepcode/exppp)" should use a variable instead of include
20:29.22mpictorexample: "DESTINATION ${INCLUDE_INSTALL_DIR}/stepcode/base)"
20:30.59mpictorlooks like it's only missing in 2 places, both in include/CMakeLists.txt
20:31.12mpictorI think that's the only thing
20:32.01kesha_Ya, that was there before the same. I just included them in /stepcode/ . Will rebase it
20:37.57mpictorok, let me know when you're done
20:38.07kesha_okay
20:43.52mpictorkesha_: you'll get a notification from github soon about a comment I just left on starseeker's pull request
20:44.03mpictorhe had to undo a few of your scope reduction changes
20:44.13kesha_done
20:44.25mpictorthe reason is that we have to target C89 so that it will compile with MSVC
20:44.53mpictorC89 doesn't allow variable declarations after any statements
20:44.59mpictorI'll take a look
20:48.11*** join/#stepcode kesha__ (~kesha@1.38.26.53)
20:49.25kesha__Ohh.. does declaration-after-statements violate at all places or just those over https://github.com/stepcode/stepcode/commit/f413047 ? :O
20:49.51mpictorI *assume* starseeker fixed all of them
20:49.59mpictorI think that most of your changes were in c++
20:50.47mpictorI had completely forgotten that old versions of C had that restriction
20:51.18mpictorunfortunately, microsoft has put little effort into updating their C compiler
21:06.05mpictorstarseeker: do you know if the man pages get installed to the correct place? https://github.com/stepcode/stepcode/blob/master/doc/CMakeLists.txt#L9
21:06.38starseekermpictor: well, I think they do for BRL-CAD
21:07.04starseekeryou may want to change that default for a "generic" Linux system install
21:07.08mpictorI was wondering if that should use ${CMAKE_INSTALL_PREFIX}
21:07.13starseekerno
21:07.18starseekerit needs to be a relative path
21:07.23mpictorno?
21:07.27mpictoroh ok
21:08.31starseekerfully qualified paths are critical for things like include directories (CMake 2.8.12 is adding that as a policy, IIRC) but for certain other situations they mess up things royally (like, say, re-creating the install layout in the build directory as part of the build ;-)
21:08.51mpictorah
21:08.58starseekeralso, one of the principles of the BRL-CAD build is to leave the install location untouched until someone actually says "make install"
21:09.11mpictorok, I'll leave it
21:09.16starseekerthanks :-)
21:09.55starseekermpictor: you wouldn't believe how hair-pulling those details can be...
21:10.03mpictorheh
21:10.26mpictorwould you mind looking over kesha_'s pull request then?
21:11.02mpictorhttps://github.com/stepcode/stepcode/pull/254
21:32.40*** join/#stepcode kesha__ (~kesha@49.249.0.218)
21:46.36mpictorkesha__: looks good to me. if starseeker is ok with it, I'll merge
22:01.36*** join/#stepcode travis-ci (~travis-ci@ec2-54-225-5-139.compute-1.amazonaws.com)
22:01.36travis-ci[travis-ci] stepcode/stepcode#36 (master - 76e2b2b : Mark): The build is still failing.
22:01.36travis-ci[travis-ci] Change view : https://github.com/stepcode/stepcode/compare/da3fe4543f86...76e2b2bcb258
22:01.36travis-ci[travis-ci] Build details : http://travis-ci.org/stepcode/stepcode/builds/10731296
22:01.36*** part/#stepcode travis-ci (~travis-ci@ec2-54-225-5-139.compute-1.amazonaws.com)
23:08.56*** join/#stepcode kesha__ (~kesha@49.249.0.218)
23:38.39starseekermpictor: looking...
23:39.34starseekerow
23:39.44starseekeryeah, that'll cause some trouble...
23:41.17starseekerwhy not just set(INCLUDE_INSTALL_DIR include)  ?
23:44.22mpictor<PROTECTED>
23:44.42mpictorI was thinking that absolute paths were needed
23:45.10mpictorit will work without cmake_install_prefix?
23:45.16starseekermpictor: in BRL-CAD we basically duplicated the hooks autotools originally provided - see the toplevel CMakeLists.txt around line 220
23:45.23starseekermpictor: it should, yes
23:45.27mpictorok
23:45.29mpictoroops
23:45.48mpictorkesha__: did you see that?
23:46.20starseekerSee http://sourceforge.net/p/brlcad/code/57247/tree/brlcad/trunk/include/CMakeLists.txt for an example of what we're doing
23:46.43starseekerline 78 in that file is the install line
23:47.29starseekerDESTINATION is already relative to CMAKE_INSTALL_PREFIX
23:48.43starseekerabsolute paths are a problem when you build (for example) binary tar.gz distributions - CPack overrides the output paths in that case to put everything in the location to be tarred up, which (IIRC) is usually different from the CMAKE_INSTALL_PREFIX location
23:48.54mpictoroh yea
23:49.06starseekerif you use absolute paths in install commands, that kinda doesn't work :-)
23:49.20starseekerit's a relatively obscure issue in some ways - I initially made the same mistake
23:49.28mpictorthat line shows DESTINATION ${INCLUDE_DIR}/brlcad). So INCLUDE_DIR is relative?
23:49.36starseeker(like I said, hair-pulling ;-)
23:49.38starseekerright
23:49.40mpictorthat would cause trouble!
23:50.06starseekerINCLUDE_DIR is defined in this file:  http://sourceforge.net/p/brlcad/code/57247/tree/brlcad/trunk/CMakeLists.txt
23:50.10starseekerline 232
23:51.04starseekernotice we don't set it if someone else already has (I suppose properly speaking those tests should be NOT DEFINED - haven't fully worked out that subtlty yet)
23:51.30starseekerthat way, if someone wants to include BRL-CAD in their own src/other build, they can override things to their own liking
23:51.57starseekerwhich is why you've occasionally seen me submit patches to stepcode allowing similar overrides - had to do the same for libpng, IIRC
23:52.23mpictorI see
23:52.25starseekerCMake is way better at sub-builds than autotools, but you do have to be aware of the possibility sometimes
23:52.43starseeker(sorry, probably way more detail than you wanted/needed)
23:52.55mpictorI just commented on kesha__'s commit on github - https://github.com/stepcode/stepcode/commit/40541f68fd39ad8e1d844e36db15c01489945dca#commitcomment-3968380
23:53.00mpictorno problem
23:53.10mpictorI've looked at autotools in the past
23:53.14mpictorretreated in horror
23:53.21starseekerhehe
23:53.40mpictorhey, are you on a computer that's locked down, or can you install client certificates?
23:53.52mpictorwe were talking about IDEs before, and I found something cool
23:54.11starseekerah - github integration?
23:54.16mpictorno
23:54.35mpictorhttps://raw.github.com/mb0/lab/master/screenshot.png
23:54.51mpictorit's a collaborative ide
23:55.31mpictorfewer features than https://code.google.com/p/collide/ , but I haven't encountered any bugs yet
23:55.38starseekercool
23:56.25starseekerfor the Go language?
23:56.35mpictorI'm running it on a dockstar (approximately equivalent to a 500MHz pentium 3) http://archlinuxarm.org/platforms/armv5/seagate-dockstar
23:56.37mpictoryea
23:57.00mpictorit uses ace code editor, which has highlighting for many languages
23:57.47mpictorwith a client certificate installed, I can use it from a computer or tablet anywhere
23:57.53mpictorhttps://maplest.redirectme.net:49989/#file/share/source/src/brlcad/src/conv/step/g-step
23:58.37mpictorit would be nice if it told you what was wrong when it doesn't accept the certificate; for now, it just gives you a blank page
23:58.38starseekerah - yeah, gives me the invalid signature
23:58.48mpictorone sec
23:58.53starseekerper firefox:  Peer's certificate has an invalid signature.
23:58.59starseekerso it did tell me :-)
23:59.20mpictornot quite what I mean :)
23:59.40starseekerso you're watching my g-step thrashing eh? :-)

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