| 00:42.58 | *** join/#brlcad IriX64 (~IriX64@bas2-sudbury98-1128565211.dsl.bell.ca) | |
| 00:59.37 | *** join/#brlcad Ralith (~ralith@S010600221561996a.vc.shawcable.net) | |
| 01:19.07 | starseeker | realizes he hasn't delt with OpenGL yet |
| 01:19.10 | starseeker | arrgh |
| 01:25.15 | Ralith | in what context? |
| 01:25.22 | starseeker | CMake build |
| 01:25.33 | Ralith | linking? |
| 01:25.38 | starseeker | detecting |
| 01:25.43 | Ralith | iirc, that's quite easy |
| 01:25.56 | Ralith | I may have the code lying around somewhere. |
| 01:25.58 | Ralith | checks |
| 01:26.04 | starseeker | Apple OpenGL is complicated - X11 OpenGL, AGL - which do we want, which headers go with which, etc |
| 01:26.11 | starseeker | need to be able to pick one or the other at compile time |
| 01:26.17 | starseeker | or configure time rather |
| 01:26.34 | Ralith | oh, right, OSX. |
| 01:27.00 | starseeker | OSX is a pain - it's what complicated the Tcl/Tk find logic so bad |
| 01:27.41 | starseeker | someone may have solved it already - I'll look around, and OpenGL is more popular than Tcl/Tk, but the fine grained control we need is somewhat unusual |
| 01:28.11 | starseeker | It's not immediately clear to me how to tell whether an OGL is compiled for X11 or Aqua |
| 01:30.34 | Ralith | the different ones can be resolved by the same search? |
| 01:30.52 | starseeker | unknown |
| 01:31.03 | Ralith | well, that seems like it'd save a lot of time to test. |
| 01:31.09 | starseeker | I essentially wrapped many searches into one script with tcl/tk |
| 01:31.31 | Ralith | my intuition suggests that the shipped cmake opengl search will only find the aqua one |
| 01:31.42 | starseeker | the idea is to be able to set options (X11/Aqua/Win32) and have FindOGL.cmake return a valid result if there is one |
| 01:31.55 | starseeker | yeah, I believe that's what happened when I tried it |
| 01:33.38 | Ralith | so perhaps all you need to do is build another tweaked FindOGL.cmake (renamed to -X11 or something) that'll find X11, then use whichever's appropriate |
| 01:34.17 | starseeker | yep |
| 01:34.47 | starseeker | devil's in the details though |
| 01:35.01 | Ralith | last I checked, those're quite simple files to author |
| 01:35.05 | Ralith | simpler still if you're just editing one. |
| 01:36.54 | starseeker | <snort> http://brlcad.svn.sf.net/viewvc/brlcad/brlcad/branches/cmake/misc/CMake/FindTCL.cmake?revision=40415&view=markup |
| 01:40.27 | starseeker | some of 'em are easy, but keeping straight multiple versions and featuresets in a single search... |
| 01:51.58 | starseeker | 'course, I make no claim that that file is optimal, but you'd be surprised how varied the places are that distros store tclConfig.sh |
| 03:24.27 | CIA-14 | BRL-CAD: 03starseeker * r40527 10/brlcad/branches/cmake/ (20 files in 20 dirs): A variety of fixes and cleanups - still need the step data, libbrlcad, librtserver and a few misc items but closer. |
| 04:52.49 | Ralith | hm |
| 04:52.50 | Ralith | well |
| 04:52.57 | Ralith | OpenGL's a hell of a lot more standard, fwiw |
| 04:53.02 | Ralith | which may not be much |
| 06:58.50 | *** join/#brlcad merzo (~merzo@smartbussiness.mobicom.net.ua) | |
| 07:33.43 | *** join/#brlcad cosurgi (~cosurgi@atak.bl.pg.gda.pl) | |
| 09:08.20 | *** part/#brlcad raininja (~raijin@unaffiliated/raijin) | |
| 09:50.56 | *** join/#brlcad mafm (~mafm@81.37.118.230) | |
| 11:49.23 | *** join/#brlcad archivist_ (~archivist@host217-34-113-62.in-addr.btopenworld.com) | |
| 13:20.13 | *** join/#brlcad Elrohir (~kvirc@p4FC5A818.dip.t-dialin.net) | |
| 14:28.29 | *** join/#brlcad ``Erik (Here@c-69-140-109-104.hsd1.md.comcast.net) | |
| 15:03.37 | *** join/#brlcad ibot (~ibot@rikers.org) | |
| 15:03.37 | *** topic/#brlcad 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 7.16.10 is posted! (20100805) | |
| 15:26.21 | CIA-14 | BRL-CAD: 03starseeker * r40528 10/brlcad/branches/cmake/src/other/step/ (CMakeLists.txt data/CMakeLists.txt): Get the data from step installing, but shouldn't be using a BRLCAD variable here - need to think about it a little. |
| 17:09.31 | *** join/#brlcad mafm_ (~mafm@81.37.118.230) | |
| 17:47.36 | CIA-14 | BRL-CAD: 03starseeker * r40529 10/brlcad/branches/cmake/src/other/URToolkit/tools/ (CMakeLists.txt clock/CMakeLists.txt): Add in rleClock |
| 18:43.02 | CIA-14 | BRL-CAD: 03starseeker * r40530 10/brlcad/branches/cmake/ (CMakeLists.txt src/librtserver/CMakeLists.txt): Add in librtserver, although it's not confirmed working yet and the option logic for it isn't there yet either. |
| 20:13.49 | CIA-14 | BRL-CAD: 03starseeker * r40531 10/brlcad/branches/cmake/src/ (8 files in 8 dirs): Don't install some things not being installed by autotools, add step man pages. |
| 20:36.17 | CIA-14 | BRL-CAD: 03bob1961 * r40532 10/brlcad/trunk/src/tclscripts/lib/RtControl.tcl: Added special handling for Windows to replace the backslashes with forward slashes before being evaluated again by Tcl in RtControl::clear. |
| 20:41.45 | CIA-14 | BRL-CAD: 03bob1961 * r40533 10/brlcad/trunk/misc/nsis/brlcad.nsi: |
| 20:41.46 | CIA-14 | BRL-CAD: Added a call to RequestExecutionLevel to set the execution level to user to get |
| 20:41.47 | CIA-14 | BRL-CAD: the failed CreateShortCut calls to work. Added a call to insertmacro for |
| 20:41.48 | CIA-14 | BRL-CAD: WordFind3X. Commented out the shortcut for the Archer documentation. |
| 20:45.55 | CIA-14 | BRL-CAD: 03bob1961 * r40534 10/brlcad/trunk/misc/win32-msvc8/tktable/tktable.vcproj: Quote a few preprocessor defs for the x64 build. |
| 21:21.43 | CIA-14 | BRL-CAD: 03starseeker * r40535 10/brlcad/branches/cmake/misc/enigma/CMakeLists.txt: If we have a crypt library use it, but don't turn it off - need to see what autotools is checking for here. |
| 21:22.49 | brlcad | enigma is a crypt replacement |
| 21:22.56 | brlcad | if you have crypt, you don't need enigma |
| 21:23.22 | starseeker | brlcad: when we compile on Linux, we use -lcrypt for enigma |
| 21:23.29 | starseeker | apparently need it |
| 21:25.14 | brlcad | yes, because enigma calls crypt() |
| 21:25.22 | brlcad | which on linux is in libcrypt |
| 21:25.59 | brlcad | (you could have libcrypt and not have crypt) |
| 21:26.03 | starseeker | right - all I was doing there was setting up the build to use the crypt library if found |
| 21:26.20 | starseeker | sorry, comment wasn't clear |
| 21:26.56 | starseeker | figures out whether proc-db/terrain or util/terrain ends up installed... |
| 21:27.48 | starseeker | ah, the util one |
| 21:27.53 | brlcad | interesting that there are two... |
| 21:28.01 | brlcad | pdb should be renamed |
| 21:28.15 | starseeker | CMake complained that there were non-unique target names, and sure enough... |
| 21:28.22 | brlcad | yep |
| 21:28.25 | brlcad | I didn't realize that |
| 21:28.52 | starseeker | kinda a handy feature, really... |
| 21:30.35 | brlcad | not much different than a non-recursive Makefile.am |
| 21:30.45 | brlcad | which is basically what cmake is producing |
| 21:31.14 | brlcad | (i.e., a non-recursive Makefile) |
| 21:31.28 | starseeker | nods |
| 21:32.09 | CIA-14 | BRL-CAD: 03starseeker * r40536 10/brlcad/branches/cmake/src/ (proc-db/CMakeLists.txt util/CMakeLists.txt): The util terrain is the actual terrain exec - call the proc-db one terraingen for now. |
| 21:52.14 | CIA-14 | BRL-CAD: 03starseeker * r40537 10/brlcad/branches/cmake/ (5 files in 5 dirs): More cleanups driven by diffs between autotools and cmake installs. Getting close to complete - virtually all remaining diff are explained by expected differences. |
| 22:31.33 | *** join/#brlcad velociostrich (~nsd@c-68-37-119-2.hsd1.nj.comcast.net) | |
| 23:04.26 | starseeker | ah that's why it didn't want to run - I was sticking db in share/7.17.0/db |
| 23:05.28 | starseeker | pix too |
| 23:30.03 | *** join/#brlcad velociostrich (~nsd@c-68-37-119-2.hsd1.nj.comcast.net) | |
| 23:42.16 | CIA-14 | BRL-CAD: 03starseeker * r40538 10/brlcad/branches/cmake/ (CMakeLists.txt bench/CMakeLists.txt sh/CMakeLists.txt): Get benchmark running - needed a slightly different target definition. |
| 23:51.18 | CIA-14 | BRL-CAD: 03starseeker * r40539 10/brlcad/branches/cmake/bench/CMakeLists.txt: |
| 23:51.18 | CIA-14 | BRL-CAD: Interestingly, it looks like the make clean command can be overridden in the |
| 23:51.18 | CIA-14 | BRL-CAD: benchmark directory. This need to be carefully tested to ensure it's not |
| 23:51.18 | CIA-14 | BRL-CAD: messing with the overall make clean setup, but this may be a viable approach. |
| 23:53.05 | CIA-14 | BRL-CAD: 03starseeker * r40540 10/brlcad/branches/cmake/bench/CMakeLists.txt: Ooo, yeah, not good - warnings at the toplevel makefile level. |
| 23:54.20 | starseeker | brlcad: http://www.cmake.org/pipermail/cmake/2006-October/011477.html |
| 23:57.06 | starseeker | Brad King's comment may be an idea... http://www.cmake.org/Bug/view.php?id=6348 do our own custom subdir target for bench, but I don't know what other consequence that might have |
| 23:57.27 | starseeker | i.e. whether we'd still get non-recursive make speed/performance |
| 23:58.01 | starseeker | plus, that's a make-only solution |
| 23:59.18 | starseeker | http://public.kitware.com/Bug/view.php?id=8438 |