| 00:13.08 | CIA-43 | BRL-CAD: 03starseeker * r42051 10/brlcad/branches/cmake/doc/docbook/system/mann/en/CMakeLists.txt: |
| 00:13.08 | CIA-43 | BRL-CAD: Start prototyping a system to use the GLOB feature of CMake to test whether our |
| 00:13.09 | CIA-43 | BRL-CAD: lists of files are complete - may be able to work out an EXTRA_DIST like |
| 00:13.20 | CIA-43 | BRL-CAD: feature, although this should probably only be run when a specific target is |
| 00:13.20 | CIA-43 | BRL-CAD: called and needs to handle all the various directories... may or may not be |
| 00:13.20 | CIA-43 | BRL-CAD: worth it, we could just do a clean svn checkout prior to packaging up for |
| 00:13.21 | CIA-43 | BRL-CAD: distribution. |
| 00:13.40 | brlcad | mid-building here, update with caution (may break build) |
| 00:28.17 | starseeker | brlcad: np - been putting off a sync to tree |
| 00:28.45 | starseeker | just trying to decide whether I have to attempt implementation of an EXTRA_DIST-like mechanism |
| 00:34.00 | starseeker | what I've done now is a little more specific - it haults the makefile generation if it spots a stray xml file not added to the CMakeLists.txt list - intended to catch man pages added to repository but not to build logic |
| 00:38.01 | starseeker | assuming I can coax the build into a state where a full build of BRL-CAD leaves a pristine source tree behind if built out-of-dir, would that alleviate the need for the EXTRA_DIST stuff? |
| 00:50.58 | DX^ | I posted a bug about IGES a long while ago |
| 00:51.03 | DX^ | but I don't think anyone cares any more :) |
| 00:55.20 | brlcad | DX^: caring doesn't imply time available, or more specifically mean that it'll get fixed quickly |
| 00:55.58 | brlcad | moreover, there's been literally months of time and effort going into that particular problem |
| 01:02.46 | *** join/#brlcad crazy_imp (~mj@a89-183-80-79.net-htp.de) | |
| 04:38.44 | *** join/#brlcad Ralith (~ralith@S010600221561996a.vc.shawcable.net) | |
| 06:57.37 | *** join/#brlcad Stattrav (~Stattrav@122.172.16.143) | |
| 06:57.37 | *** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav) | |
| 10:16.13 | *** join/#brlcad mafm (~mafm@231.Red-83-38-34.dynamicIP.rima-tde.net) | |
| 11:12.04 | CIA-43 | BRL-CAD: 03davidloman * r42052 10/rt^3/trunk/HACKING: Fixt up some Engrish issues in Hacking. |
| 11:34.34 | *** join/#brlcad mafm (~mafm@231.Red-83-38-34.dynamicIP.rima-tde.net) | |
| 12:08.43 | *** join/#brlcad Stattrav (~Stattrav@122.172.16.143) | |
| 12:08.43 | *** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav) | |
| 14:40.59 | *** join/#brlcad Zaebos (~irc@pd95b7f5e.dip0.t-ipconnect.de) | |
| 15:22.56 | *** join/#brlcad merzo (~merzo@193.254.217.44) | |
| 15:24.31 | *** join/#brlcad d_rossberg (~rossberg@BZ.BZFLAG.BZ) | |
| 15:28.23 | CIA-43 | BRL-CAD: 03d_rossberg * r42053 10/rt^3/trunk/src/coreInterface/Arb8.cpp: the VAPPROXEQUAL() macro was renamed to VNEAR_EQUAL(), so it was changed here too |
| 15:30.36 | CIA-43 | BRL-CAD: 03d_rossberg * r42054 10/brlcad/trunk/ (2 files in 2 dirs): |
| 15:30.36 | CIA-43 | BRL-CAD: libz is called zlib now |
| 15:30.37 | CIA-43 | BRL-CAD: (and opennurbs_ext.h a private header) |
| 15:42.11 | CIA-43 | BRL-CAD: 03d_rossberg * r42055 10/brlcad/trunk/src/conv/iges/brlcad_brep.cpp: another case of pullback_curve() usage and therefore opennurbs_ext.h include |
| 15:43.21 | CIA-43 | BRL-CAD: 03d_rossberg * r42056 10/brlcad/trunk/src/libbn/CMakeLists.txt: there is no screengrab.c in libbn |
| 15:48.53 | CIA-43 | BRL-CAD: 03d_rossberg * r42057 10/brlcad/trunk/src/libbu/CMakeLists.txt: |
| 15:48.53 | CIA-43 | BRL-CAD: synced with Makefile.am (timer.c) |
| 15:48.54 | CIA-43 | BRL-CAD: compiles fine with MSVC 2008 |
| 16:11.57 | CIA-43 | BRL-CAD: 03d_rossberg * r42058 10/brlcad/trunk/src/other/libz/CMakeLists.txt: |
| 16:11.57 | CIA-43 | BRL-CAD: the zconf.h has to lie in the same directory as zlib.h no matter where the binaries are build to be visible by the other libraries |
| 16:11.58 | CIA-43 | BRL-CAD: the original CMakeLists.txt makes sense for a standalone project but causes problems as a sub-project configuration in larger programs |
| 16:11.58 | CIA-43 | BRL-CAD: (nevertheless it's a good starting point) |
| 16:18.09 | starseeker | d_rossberg: I don't think we want to alter the libz CMakeLists.txt if we can help it |
| 16:18.58 | starseeker | d_rossberg: I'm using the vanilla file in the cmake branch with one tweak for MAN pages and it works fine - the trick is to include_directories both the source and the binary dirs to get both zlib.h and zconf.h (unless I'm missing something) |
| 16:19.38 | starseeker | I do that in a lot of places - my goal is to have a complete BRL-CAD build take place without the build process altering the source tree in any way |
| 16:20.08 | starseeker | I'm pretty close - the main issue is a tcl script that insists on generating files in src - I need to provide it with an output dir target |
| 16:26.57 | d_rossberg | starseeker: i'm afraid zlib's CMakeLists.txt has to be modified anyway, either as i did or it has to export its include directories |
| 16:27.31 | starseeker | but if you're doing a subproject, you know where both its src dir and its build dir are? |
| 16:28.06 | starseeker | I'm building openNURBS in the cmake branch successfully... |
| 16:29.00 | d_rossberg | i don't want to guess where zlib's include files could be all around |
| 16:29.38 | starseeker | that's just it - you don't. There are only two possibilities |
| 16:29.48 | starseeker | if you include both, you should be good |
| 16:30.33 | d_rossberg | i don't even know where zlibs binary directory is (in other libs than zlib) |
| 16:31.51 | starseeker | ah - that's probably my THIRD_PARTY_SUBDIR macro |
| 16:33.15 | d_rossberg | no, the is cmake 2.6 (?) standard behavior |
| 16:33.47 | starseeker | yeah, you're right that the default behavior works that way |
| 16:34.10 | starseeker | but we have a macro in misc/CMake in the cmake branch that lets us avoid those issues |
| 16:34.26 | starseeker | have you tried the cmake branch lately? |
| 16:35.01 | d_rossberg | no, today is my first day back in the office |
| 16:35.04 | d_rossberg | we can talk about it tomorrow, i've to catch my bus ;) |
| 16:35.09 | starseeker | cool |
| 16:35.38 | d_rossberg | ... i'll try the cmake branch later ... |
| 17:04.42 | CIA-43 | BRL-CAD: 03starseeker * r42059 10/brlcad/branches/cmake/src/other/ (6 files in 6 dirs): Generalize the logic for writing out pkgIndex files |
| 17:07.19 | brlcad | gone way down the rabbit hole with these latest size_t changes.. so much stuff! |
| 19:27.23 | *** join/#brlcad Ralith (~ralith@d142-058-093-144.wireless.sfu.ca) | |
| 19:48.19 | CIA-43 | BRL-CAD: 03starseeker * r42060 10/brlcad/branches/cmake/src/other/tk/CMakeLists.txt: Hmm - put the stubs line inside quotes, and just tclstub on Win32... |
| 19:48.41 | *** join/#brlcad Ralith (~ralith@d142-058-093-144.wireless.sfu.ca) | |
| 20:14.23 | starseeker | WOOOOT |
| 20:14.57 | starseeker | first successful run of MGED on Windows from a CMake build |
| 20:15.48 | brlcad | congratulations |
| 20:18.45 | *** join/#brlcad Ralith (~ralith@d142-058-093-144.wireless.sfu.ca) | |
| 20:24.27 | CIA-43 | BRL-CAD: 03brlcad * r42061 10/brlcad/trunk/ (24 files in 7 dirs): |
| 20:24.27 | CIA-43 | BRL-CAD: revert libtie migration until the problems are fixed. there was no build logic |
| 20:24.28 | CIA-43 | BRL-CAD: in src/libtie/ (so the build was broken) and the tie.h header included private |
| 20:24.28 | CIA-43 | BRL-CAD: implementation details not appropriate for promotion into the include/ |
| 20:24.29 | CIA-43 | BRL-CAD: directory. this reverts the remaining portion of commit r42032 not already |
| 20:24.29 | CIA-43 | BRL-CAD: reverted by r42035. |
| 21:29.12 | CIA-43 | BRL-CAD: 03johnranderson * r42062 10/brlcad/trunk/src/conv/asc/asc2g.c: |
| 21:29.14 | CIA-43 | BRL-CAD: Now accepts initial comments at start of file |
| 21:29.15 | CIA-43 | BRL-CAD: First non-comment non-blank line must still be a "title" or "put" command, |
| 21:29.25 | brlcad | woot |
| 21:47.35 | CIA-43 | BRL-CAD: 03starseeker * r42063 10/brlcad/branches/cmake/src/other/tkhtml/CMakeLists.txt: Tweak building of tkhtml for Windows |
| 22:56.41 | *** join/#brlcad Ralith (~ralith@S010600221561996a.vc.shawcable.net) | |
| 23:03.18 | CIA-43 | BRL-CAD: 03starseeker * r42064 10/brlcad/branches/cmake/CMakeLists.txt: Start setting up logic for CPack. Start slowly - this is just the tar.gz, tar.bz2 and .zip files, to match what make distcheck produces (that will need to be confirmed). |
| 23:36.40 | CIA-43 | BRL-CAD: 03starseeker * r42065 10/brlcad/branches/cmake/src/tclscripts/CMakeLists.txt: |
| 23:36.41 | CIA-43 | BRL-CAD: Try moving the generated script files into the bin dir - this should, in theory, |
| 23:36.41 | CIA-43 | BRL-CAD: be the last thing that was writing back to the source tree during build. Will |
| 23:36.42 | CIA-43 | BRL-CAD: need to try in an svn or git repository with no ignore settings. |
| 23:51.17 | starseeker | woo-hoo, that looks like it may be it - untarred one of the cpack tarballs, built using the untarred result as the src dir, untarred again into another dir, and a diff -r of the two directories reported no differences |