| 01:02.46 | *** join/#brlcad R0b0t1 (~Enigma@unaffiliated/r0b0t1) | |
| 01:46.53 | *** join/#brlcad louipc (~louipc@archlinux/trusteduser/louipc) | |
| 02:20.33 | *** join/#brlcad Ralith (~ralith@S010600221561996a.vc.shawcable.net) | |
| 03:17.42 | *** join/#brlcad juanman (~quassel@unaffiliated/juanman) | |
| 04:35.58 | CIA-53 | BRL-CAD: 03starseeker * r42816 10/brlcad/branches/cmake/CMakeLists.txt: |
| 04:35.58 | CIA-53 | BRL-CAD: Going to have to use a global property for the build targets lists - functions |
| 04:35.58 | CIA-53 | BRL-CAD: have local variable scope, not sure why it ever worked originally. This needs |
| 04:35.58 | CIA-53 | BRL-CAD: more testing, not sure it stays constant over multiple cmake runs and need to |
| 04:35.58 | CIA-53 | BRL-CAD: verify it's doing 'the right things'. Also, test noprod rule. |
| 05:07.48 | CIA-53 | BRL-CAD: 03brlcad * r42817 10/brlcad/trunk/src/librtserver/rtserver.c: add missing footer |
| 05:08.39 | CIA-53 | BRL-CAD: 03brlcad * r42818 10/brlcad/trunk/src/librtserver/rtserver.c: ws/indent consistency cleanup |
| 05:33.13 | CIA-53 | BRL-CAD: 03starseeker * r42819 10/brlcad/branches/cmake/CMakeLists.txt: |
| 05:33.14 | CIA-53 | BRL-CAD: This is a long shot, but see if a per-directory noprod can be set up. Just |
| 05:33.14 | CIA-53 | BRL-CAD: exploring the preliminary steps as yet - if this does work it will involve a |
| 05:33.14 | CIA-53 | BRL-CAD: hideous abuse of the old CMP0002 policy allowing non-unique target names, so not |
| 05:33.14 | CIA-53 | BRL-CAD: ideal even if it does work... |
| 05:35.51 | *** join/#brlcad cjdevlin (~devlin@d118-75-252-178.try.wideopenwest.com) | |
| 05:57.58 | *** join/#brlcad yukonbob (~bch@S010600235a187d92.ok.shawcable.net) | |
| 05:59.49 | *** join/#brlcad Stattrav (~Stattrav@122.172.33.253) | |
| 05:59.50 | *** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav) | |
| 06:44.58 | CIA-53 | BRL-CAD: 03jordisayol * r42820 10/brlcad/trunk/ (misc/debian/control misc/debian/rules sh/make_deb.sh): add the option to build debian source packages and change debian configuration. |
| 09:14.39 | CIA-53 | BRL-CAD: 03jordisayol * r42821 10/brlcad/trunk/ (4 files in 2 dirs): add two new doc menu entries |
| 09:58.29 | cjdevlin | !info xubuntu-desktop |
| 11:19.43 | dloman | Well, i must admit that this storm is rather disappointing :/ |
| 11:20.11 | dloman | snow was supposed to start at 1700 yesterday and accumulate 5 inches.. so far, just a dusting. |
| 11:20.34 | dloman | But the ice has finally started..... yay. |
| 12:38.49 | *** join/#brlcad juanman (~quassel@unaffiliated/juanman) | |
| 13:29.54 | *** join/#brlcad Stattrav (~Stattrav@117.192.135.87) | |
| 13:29.54 | *** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav) | |
| 13:36.57 | *** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav) | |
| 13:38.13 | ``Erik | hm, the cul de sac is a sheet of ice, no one else in my cul de sac has tried venturing out, one of my neighbors said he heard about lots of accidents on tv and was going to wait until it warms up later to try |
| 13:38.17 | ``Erik | O.O wee |
| 13:48.35 | dloman | time to take the M3 out for a 'spin' eh? =D |
| 13:51.29 | ``Erik | heh, it wouldn't be able to get out of the driveway, still too much snow and ice |
| 13:59.09 | dloman | fun fun! |
| 14:19.47 | *** join/#brlcad epileg (~epileg@unaffiliated/epileg) | |
| 14:57.07 | dloman | Icy here, icy there...but at least post is open! lol |
| 15:34.16 | ``Erik | once I got out of my neighborhood (I think about a mile?), it was all good |
| 15:34.43 | ``Erik | I don't think they salt back in my development, or the one I have to go through to get from mine to the highway |
| 15:35.43 | ``Erik | 0.7 miles according to google maps heh |
| 15:41.40 | CIA-53 | BRL-CAD: 03starseeker * r42822 10/brlcad/branches/cmake/CMakeLists.txt: (log message trimmed) |
| 15:41.40 | CIA-53 | BRL-CAD: Eeek. This appears to be a working per-directory noprod solution. Done the |
| 15:41.40 | CIA-53 | BRL-CAD: best I can for convenience vs. policy - CMP0002 forbids non-unique target names, |
| 15:41.40 | CIA-53 | BRL-CAD: and hence excludes a 'make noprod' rule for each directory. Even forcing that |
| 15:41.40 | CIA-53 | BRL-CAD: to OLD, get some odd behavior with the toplevel 'make noprod' rule (not |
| 15:41.41 | CIA-53 | BRL-CAD: surprisingly, the policy is there for a reason). Hence, there is a toplevel |
| 15:41.42 | CIA-53 | BRL-CAD: make noprod command and in subdirectories 'make noprod' becomes 'cmake -P |
| 15:42.56 | starseeker | brlcad: that appears to be the best I can do for noprod - you can achieve (I think) the effects of make noprod in a directory but you'll have to do the cmake -P thing instead of make noprod |
| 16:11.14 | CIA-53 | BRL-CAD: 03starseeker * r42823 10/brlcad/branches/cmake/ (9 files in 5 dirs): Fix step build properly - surprise, surprise using scl_cf.h correctly actually helped |
| 16:20.15 | CIA-53 | BRL-CAD: 03starseeker * r42824 10/brlcad/branches/cmake/src/conv/step/CMakeLists.txt: Add the output include dir for step |
| 17:05.17 | CIA-53 | BRL-CAD: 03starseeker * r42825 10/brlcad/branches/cmake/src/other/tcl/CMakeLists.txt: Nevermind the dash - this symlink logic is too fragile, so just go with CMake standard practice |
| 17:08.52 | CIA-53 | BRL-CAD: 03jordisayol * r42826 10/brlcad/trunk/misc/Makefile.am: add two debian doc menu entries |
| 17:53.08 | CIA-53 | BRL-CAD: 03starseeker * r42827 10/brlcad/branches/cmake/ (5 files in 5 dirs): Ah - remove files based on version number properties as well for noprod. Start using the exec list to strip executables for CPack |
| 18:28.15 | *** join/#brlcad andrecastelo (~andre@187.115.160.108) | |
| 18:33.10 | CIA-53 | BRL-CAD: 03erikgreenwald * r42828 10/brlcad/trunk/src/librt/primitives/nmg/nmg_junk.c: Remove the HIDDEN from these junk functions to avoid "unused symbol" failures when debugging is off. |
| 18:43.21 | CIA-53 | BRL-CAD: 03starseeker * r42829 10/brlcad/branches/cmake/src/other/step/CMake/SCL_Utils.cmake: Don't use absolute install paths |
| 18:54.28 | CIA-53 | BRL-CAD: 03starseeker * r42830 10/brlcad/branches/cmake/ (4 files in 4 dirs): another absolute install path |
| 19:12.52 | ``Erik | http://punditkitchen.files.wordpress.com/2011/01/1fe680f4-203b-47af-bb23-8bb0e0f2a623.jpg |
| 19:20.16 | CIA-53 | BRL-CAD: 03starseeker * r42831 10/brlcad/branches/cmake/ (3 files in 2 dirs): Little more tweaking to mark variables as advanced |
| 19:25.14 | CIA-53 | BRL-CAD: 03starseeker * r42832 10/brlcad/branches/cmake/src/tclscripts/ampi.tcl: verbose flag is intentional to highlight problems - restore |
| 19:29.26 | CIA-53 | BRL-CAD: 03johnranderson * r42833 10/jbrlcad/trunk/ (10 files in 2 dirs): Point and Vector3 now have a common base class (Triple) |
| 19:36.29 | CIA-53 | BRL-CAD: 03starseeker * r42834 10/brlcad/branches/cmake/ (25 files in 11 dirs): MFC 42833 |
| 19:43.05 | CIA-53 | BRL-CAD: 03starseeker * r42835 10/brlcad/trunk/src/libbu/brlcad_path.c: realpath tweak, add ifdef wrapped Windows logic |
| 19:45.49 | CIA-53 | BRL-CAD: 03starseeker * r42836 10/brlcad/trunk/src/libbn/anim.c: initialize dir (quiet a warning) |
| 19:57.47 | starseeker | um... |
| 19:57.49 | starseeker | cc1plus: warnings being treated as errors |
| 19:57.49 | starseeker | src/other/openNURBS/opennurbs_array.h: In function ?ON_Curve* brlcad::pullback_curve(ON_BrepFace*, const ON_Curve*, brlcad::SurfaceTree*, double, double)?: |
| 19:57.52 | starseeker | src/other/openNURBS/opennurbs_array.h:353: error: inlining failed in call to ?virtual ON_2dPointArray::~ON_2dPointArray()?: call is unlikely and code size would grow |
| 19:57.55 | starseeker | src/librt/opennurbs_ext.cpp:2400: error: called from here |
| 19:58.12 | starseeker | huh? |
| 20:11.00 | brlcad | c++ should not be compiling with strict |
| 20:11.29 | brlcad | though that warning is a valid "issue" |
| 20:12.01 | brlcad | destructors shouldn't be inline anyways |
| 20:12.27 | starseeker | doesn't see where it's being inlined |
| 20:33.08 | CIA-53 | BRL-CAD: 03brlcad * r42837 10/brlcad/trunk/misc/debian/ (7 files): switch to text/plain and set eol-style to native |
| 20:42.24 | CIA-53 | BRL-CAD: 03brlcad * r42838 10/brlcad/trunk/src/libbn/anim.c: VINIT_ZERO instead of {} |
| 20:50.00 | CIA-53 | BRL-CAD: 03starseeker * r42839 10/brlcad/branches/cmake/src/libfft/CMakeLists.txt: Yow - optimized is what kills fft - don't go above 256 until there's a demonstrated need |
| 20:51.28 | brlcad | hehe |
| 20:53.02 | brlcad | the autotools build only goes up to 256 |
| 20:53.10 | CIA-53 | BRL-CAD: 03starseeker * r42840 10/brlcad/branches/cmake/src/libtclcad/CMakeLists.txt: Make libtclcad strict |
| 20:53.31 | brlcad | those generated sources get big fast |
| 20:53.37 | starseeker | indeed |
| 20:54.26 | brlcad | but it's clever, iirc one of the fastest implementation methods |
| 20:54.43 | starseeker | neat |
| 20:54.54 | starseeker | hmm |
| 20:54.57 | starseeker | src/liboptical/sh_noise.c:174: error: ânoise_renderâ declared âstaticâ but never defined |
| 20:54.58 | brlcad | would be ineteresting to revisit some comparisons to see if we're still one of the best |
| 20:55.56 | brlcad | the curse of forward declarations |
| 20:56.17 | brlcad | there shouldn't be any forward declarations |
| 20:59.26 | starseeker | must have missed an update |
| 21:01.47 | CIA-53 | BRL-CAD: 03starseeker * r42841 10/brlcad/trunk/src/liboptical/sh_noise.c: Forward declarations bad. - move mfuncs down and get rid of them |
| 21:01.58 | starseeker | sh_points.c:79: error: useless storage class specifier in empty declaration |
| 21:04.09 | CIA-53 | BRL-CAD: 03starseeker * r42842 10/brlcad/trunk/src/liboptical/sh_points.c: Hidden causes a warning here about useless storage class specification |
| 21:05.11 | *** join/#brlcad R0b0t1 (~Enigma@64-136-219-55.dyn.everestkc.net) | |
| 21:05.11 | *** join/#brlcad R0b0t1 (~Enigma@unaffiliated/r0b0t1) | |
| 21:10.51 | epileg | I can not create pdf files in Linux. is this normal? |
| 21:11.06 | starseeker | epileg: you mean the build won't make them? |
| 21:11.14 | epileg | yes, this is |
| 21:11.21 | starseeker | what is the error? |
| 21:11.35 | epileg | no error, on config says... |
| 21:11.39 | epileg | just a moment |
| 21:11.46 | starseeker | oh - you need to install Apache FOP |
| 21:13.03 | epileg | starseeker: thanks! I'll check |
| 21:14.54 | starseeker | brlcad: uh, this one has me puzzled... |
| 21:14.58 | starseeker | src/conv/g-vrml.c:821: error: variable â__streamâ might be clobbered by âlongjmpâ or âvfor |
| 21:25.46 | brlcad | starseeker: you didn't see all of those longjmp commits last week? |
| 21:26.20 | brlcad | or at least read the commit messages .. I blathered quite a bit about that :) |
| 21:26.27 | starseeker | thought he recalled something |
| 21:26.30 | starseeker | checks... |
| 21:30.28 | starseeker | brlcad: r42544? |
| 22:00.51 | starseeker | cannot follow what this logic is doing |
| 22:01.06 | starseeker | q |
| 22:12.34 | CIA-53 | BRL-CAD: 03starseeker * r42844 10/brlcad/trunk/src/conv/dem-g.c: Compiler doesn't like this inline |
| 22:19.40 | CIA-53 | BRL-CAD: 03starseeker * r42843 10/brlcad/trunk/src/liboptical/sh_text.c: Need a little more careful look here, but this removes enough forward decls to get it building |
| 22:20.39 | CIA-53 | BRL-CAD: 03brlcad * r42845 10/brlcad/trunk/ (4 files in 3 dirs): check for GetFullPathName() and conditionalize on HAVE_GETFULLPATHNAME instead of _WIN32 accordingly |
| 22:21.44 | CIA-53 | BRL-CAD: 03brlcad * r42846 10/brlcad/trunk/src/libbu/ (brlcad_path.c timer.c): include bio.h instead of direclty including windows.h; keying off _WIN32 is bad form |
| 22:21.44 | CIA-53 | BRL-CAD: 03brlcad * r42847 10/brlcad/trunk/src/mged/mged.c: ERR... really? somehow this commit didn't make it through. give mged the ability to forcibly flip the version number even if librt doesn't do it automatically. |
| 22:21.44 | CIA-53 | BRL-CAD: 03brlcad * r42848 10/brlcad/trunk/src/bwish/ (consoleMain.c winMain.c): another bio.h inclusion instead of windows.h |
| 22:21.45 | CIA-53 | BRL-CAD: 03erikgreenwald * r42849 10/brlcad/branches/bottie/ (4 files in 3 dirs): mimic rt_bot_minpieces for TIE algorithm |
| 22:27.20 | CIA-53 | BRL-CAD: 03starseeker * r42850 10/brlcad/trunk/src/conv/g-vrml.c: work around potential clobber in g-vrml |
| 22:27.20 | CIA-53 | BRL-CAD: 03starseeker * r42851 10/brlcad/trunk/src/conv/g-x3d.c: Same issue with g-x3d |
| 22:28.37 | CIA-53 | BRL-CAD: 03starseeker * r42852 10/brlcad/trunk/src/libdm/ (dm-plot.c dm-ps.c dm-tk.c): reorder to remove forward declarations |
| 22:39.41 | CIA-53 | BRL-CAD: 03starseeker * r42853 10/brlcad/branches/cmake/misc/CMake/BRLCAD_Util.cmake: Turn off Werror when we see cpp files in a strict library, unless CXX_STRICT is on. Leave the warnings unless the warnings option is off. |
| 22:42.31 | CIA-53 | BRL-CAD: 03starseeker * r42854 10/brlcad/branches/cmake/CMakeLists.txt: Add the BRLCAD-ENABLE_CXX_STRICT option |
| 22:44.30 | CIA-53 | BRL-CAD: 03brlcad * r42855 10/brlcad/trunk/src/mged/mged.c: tweak the message to let the user know whether -f had any affect. only mark as read-only and flip if we need to. |
| 22:45.16 | CIA-53 | BRL-CAD: 03starseeker * r42856 10/brlcad/trunk/src/ (4 files in 2 dirs): Few misc tweaks - turn off some unused libged code, UNUSED in brep.cpp |
| 22:46.55 | CIA-53 | BRL-CAD: 03starseeker * r42857 10/brlcad/branches/cmake/ (26 files in 10 dirs): Put various fixes into CMake branch |
| 22:46.57 | CIA-53 | BRL-CAD: 03brlcad * r42858 10/brlcad/trunk/include/config_win.h: file wasn't saved for r42845. set HAVE_GETFULLPATHNAME |
| 22:49.11 | CIA-53 | BRL-CAD: 03starseeker * r42859 10/brlcad/trunk/src/libtclcad/ged_obj.c: not defined |
| 22:50.05 | CIA-53 | BRL-CAD: 03starseeker * r42860 10/brlcad/branches/cmake/src/bwish/ (consoleMain.c winMain.c): More fixes from trunk |
| 22:50.36 | CIA-53 | BRL-CAD: 03starseeker * r42861 10/brlcad/branches/cmake/src/libtclcad/ged_obj.c: remove this on cmake branch too |
| 22:50.54 | brlcad | should delete unused code |
| 22:51.13 | starseeker | brlcad: thought I'd check with Bob/Richard before nuking |
| 22:51.42 | brlcad | it's in revision history |
| 22:51.49 | starseeker | true |
| 22:52.07 | brlcad | if there's no comment, then there's definitely no reason to keep it |
| 22:52.07 | CIA-53 | BRL-CAD: 03starseeker * r42862 10/brlcad/branches/cmake/src/mged/mged.c: update from trunk |
| 22:52.47 | brlcad | commented out code, #if 0'd code |
| 22:53.41 | brlcad | it's a maintenance burden and just makes the files messy and often error prone, the code falls out of sync quickly and can cause bugs if re-enabled |
| 22:57.02 | CIA-53 | BRL-CAD: 03starseeker * r42863 10/brlcad/trunk/src/libged/ (bigE.c human.c): Go ahead and remove the code from libged, instead of commenting itout |
| 22:57.45 | brlcad | ah yeah, covered by HACKING under the refactoring section too :) |
| 23:00.50 | CIA-53 | BRL-CAD: 03starseeker * r42864 10/brlcad/trunk/src/ (4 files in 4 dirs): |
| 23:00.50 | CIA-53 | BRL-CAD: Let's try this again - see if we can do without tclcad_tcl_library. Using |
| 23:00.50 | CIA-53 | BRL-CAD: private Tcl functions is Not Good, particularly when it's for debug printing. |
| 23:00.50 | CIA-53 | BRL-CAD: also removing the found_init_tcl call to TclSetLibraryPath - same basic problem |
| 23:00.50 | CIA-53 | BRL-CAD: and the comment says it doesn't do what we might expect it to anyway. |
| 23:02.20 | CIA-53 | BRL-CAD: 03starseeker * r42865 10/brlcad/branches/cmake/src/ (4 files in 4 dirs): remove tclcad_tcl_library in CMake branch too |
| 23:02.37 | starseeker | brlcad: is that alternative way to do the make noprod thing acceptable? |
| 23:04.18 | brlcad | doing a "make noprod" from the top-level would even probably work if you can relink a given binary (e.g. "make mged") with one step |
| 23:05.57 | brlcad | hard to say without trying it in practice really, wasn't clear what the final step(s) needed were going to be |
| 23:06.29 | starseeker | uh... you just do cmake -P ./cmake_noprod.cmake instead of make noprod |
| 23:07.12 | brlcad | heh, might just end up with an sh/noprod.sh script then that runs that |
| 23:08.13 | starseeker | cmake will relink the libraries required by mged before linking mged |
| 23:08.23 | starseeker | following its dependency logic |
| 23:09.13 | brlcad | so what about a simple rule that wipes out all products? |
| 23:09.23 | starseeker | that's the toplevel make noprod |
| 23:09.25 | starseeker | that works |
| 23:09.41 | starseeker | it's just in the subdirectories you need to run the specific script |
| 23:09.48 | starseeker | I can't define a noprod target for every directory |
| 23:11.00 | CIA-53 | BRL-CAD: 03bob1961 * r42866 10/brlcad/trunk/src/tclscripts/archer/Archer.tcl: Updated Archer::constructor to properly set the view of the command window (i.e. how much command window to display verses geometry window etc. |
| 23:13.04 | starseeker | O.o the toy jeep is giving out Bend radii errors |
| 23:15.10 | brlcad | the toplevel make noprod sounds like it might end up being more practical than cmake -P ./cmake_noprod.cmake |
| 23:15.31 | starseeker | well, they're both there |
| 23:15.45 | starseeker | it's not an either/or situation |
| 23:16.04 | brlcad | it just feels way too long and prone to being forgotten |
| 23:16.19 | starseeker | what about cmake -P noprod |
| 23:16.39 | starseeker | the last arg is a file name, I can just call it noprod if that's better |
| 23:16.40 | brlcad | that'd probably be better |
| 23:17.01 | brlcad | no way to inject custom make rules? |
| 23:17.26 | brlcad | some "if target is makefile, #include Makefile.inc" |
| 23:17.44 | brlcad | similar to what is done for "make fast" |
| 23:17.59 | starseeker | not that I've found so far... |
| 23:18.00 | brlcad | and make noprod for that matter, iirc |
| 23:20.00 | brlcad | my gut feeling is if all products are unique, then cd'ing down into the hierarchy isn't going to be as important |
| 23:20.31 | brlcad | since I can "make noprod && make rt" and it'll relink rt (along with all of rt's dependencies) |
| 23:20.37 | starseeker | nods |
| 23:20.50 | starseeker | well, the cmake -P option is there if it's needed |
| 23:21.12 | starseeker | was quite the operation to set up |
| 23:21.18 | brlcad | not quite as ideally targetted as *only* relinking rt, but possibly workable |
| 23:21.28 | brlcad | have to see in practice, but it's a minor point |
| 23:21.55 | brlcad | I don't doubt it |
| 23:22.56 | brlcad | unfortunately, cmake makes some things much harder to do |
| 23:23.09 | brlcad | a known limitation going into this |
| 23:23.29 | starseeker | a few maybe, but I haven't had too much trouble on the whole... |
| 23:23.35 | brlcad | heh |
| 23:23.44 | starseeker | main trouble was Tcl/Tk |
| 23:23.45 | brlcad | says you more than 6 months later :) |
| 23:24.19 | starseeker | <snort> If I'd been re-creating the autotools logic, I'd bet on a year easy (for me, anyway) |
| 23:25.58 | CIA-53 | BRL-CAD: 03starseeker * r42867 10/brlcad/branches/cmake/CMakeLists.txt: no reason to use cmake_noprod.cmake - just call it noprod so it's more like a make target |
| 23:27.01 | starseeker | you super scripting gurus are more comfortable with sh and friends - for me it might as well be perl |
| 23:28.40 | brlcad | autotools was a learning hurdle too |
| 23:29.03 | brlcad | both erik and I put several months whipping configure.ac and friends into shape |
| 23:30.55 | brlcad | the cmake effort is probably on-par, seems like even a bit more given it's not as mature and flexible |
| 23:31.29 | brlcad | but then higher short-term with the hope that maintenance is lower in the long-term |
| 23:31.48 | starseeker | nods |
| 23:33.11 | brlcad | realistically speaking, there's undoubtedly months of work still remaining to bring it on-par cross-platform in a low-maintenance fashion with all the bugs squeezed out |
| 23:34.06 | starseeker | I suppose. On the whole I'm rather happy with how it's turned out, but of course I haven't done extensive options testing and such |
| 23:34.21 | starseeker | knows turning off X11 won't quite work yet, for example - it still builds Tk |
| 23:35.29 | brlcad | simple code metrics -- no code is perfect, there will be bugs even aside from incompleteness |
| 23:37.40 | brlcad | new code is pretty consistently reliable to have bugs on the order as bad as 1% to as good as 0.1% (on average) |
| 23:38.35 | starseeker | brlcad: what are the remaining hurdles for me to be able to merge into trunk? |
| 23:39.33 | brlcad | lines-of-code of course isn't really strictly reliable, but it's pretty reasonable to assume that if 10000 new lines of logic were written, that there's at least a dozen bugs in there if not more |
| 23:42.45 | brlcad | starseeker: well whenever you're confident that you're "done done", that should mean that someone else should be able to do a code review to see if there are any show-stopper code differences, things that were turned on/off, that a release can be produced and verified, that everything is properly documented (which is really part of release verification) |
| 23:45.10 | CIA-53 | BRL-CAD: 03starseeker * r42868 10/brlcad/branches/cmake/src/ (bwish/CMakeLists.txt other/CMakeLists.txt): Don't enable Tk even with ENABLE_ALL if we aren't supposed to build it |
| 23:45.39 | starseeker | brlcad: so I need the equalivent to our documentation of the configure options for the significant CMake variables |
| 23:46.27 | starseeker | ponders... hmm... |
| 23:46.41 | brlcad | ideally, yes -- but more specifically, README, INSTALL, HACKING, doc/* have to be updated |
| 23:47.24 | brlcad | places that refer to "autogen.sh", or "configure", "make", "Makefile.am", etc |
| 23:47.43 | brlcad | we have to be able to push out a source release |
| 23:47.59 | brlcad | compilation docs are part of that |
| 23:47.59 | starseeker | nods. |
| 23:52.04 | brlcad | notes that it's time to prepare a release |
| 23:53.00 | brlcad | how extensively did you test r42864 ? |
| 23:53.18 | brlcad | tclcad_tcl_library() removal |
| 23:53.45 | starseeker | I've had it gone in the CMake branch for a while, but less extensively than I'd like |
| 23:54.15 | starseeker | unless its documentation was wrong, it's a debug output function? |
| 23:54.26 | brlcad | it's not |
| 23:54.43 | brlcad | it prevents a runtime crash for some versions of tcl |
| 23:56.14 | brlcad | TclGetLibraryPath() initializes some things internal to the Tcl library that were not otherwise getting initialized by Tcl_Init or other calls and would result in a non-catchable run-time crash |
| 23:56.59 | brlcad | iirc, not visible during --enable-all but was reproducible if you used a system tcl/tk |
| 23:57.51 | brlcad | don't know if it was specific to any tcl version, 8.4 or 8.5 but it was a hard show-stopper crash of mged |
| 23:58.32 | brlcad | it was a while ago too, so if it works with --enable-all and --disable-all now with 8.5+ for example, then we can just move forward |
| 23:58.57 | starseeker | ok, I'll see what happens with a system tcl/tk |
| 23:59.15 | starseeker | We can put it back in if we have to, but I really don't like calling private Tcl/Tk functions |
| 23:59.26 | brlcad | putting it back isn't the issue |
| 23:59.34 | brlcad | given it's a hard crash, we can't do a release until verifying |
| 23:59.40 | starseeker | ah |
| 23:59.43 | brlcad | because it's otherwise DOA |
