| 00:34.10 | Notify | 03BRL-CAD:starseeker * 62996 brlcad/trunk/regress/repository.sh: Update repository regression script. |
| 00:34.33 | Notify | 03BRL-CAD:brlcad * 62997 brlcad/trunk/src/libfb/if_wgl.c: no WSA or other networking calls, so should only need bselect.h |
| 01:48.47 | Notify | 03BRL-CAD:brlcad * 62998 brlcad/trunk/include/bsocket.h: catch order inclusion issues even when not compiling on windows |
| 01:49.06 | Notify | 03BRL-CAD:brlcad * 62999 brlcad/trunk/include/bnetwork.h: reword, clarity |
| 02:23.15 | Notify | 03BRL-CAD:starseeker * 63000 brlcad/trunk/src/libfb/if_wgl.c: rename |
| 02:33.16 | Notify | 03BRL-CAD:starseeker * 63001 brlcad/trunk/src/bwish/cadAppInit.c: tweak cadAppInit.c headers |
| 02:38.04 | Notify | 03BRL-CAD:starseeker * 63002 (brlcad/trunk/src/mged/dm-generic.c brlcad/trunk/src/mged/fbserv.c): Tweak some mged header includes - not sure if these are right yet... |
| 02:42.52 | Notify | 03BRL-CAD:starseeker * 63003 brlcad/trunk/src/mged/dm-generic.c: reorder includes for Windows |
| 02:48.35 | Notify | 03BRL-CAD:starseeker * 63004 brlcad/trunk/include/bu/version.h: Need defines or BU_EXPORT won't function on Windows. |
| 02:50.19 | starseeker | grr |
| 02:50.33 | starseeker | tries to figure out what he's still missing for version.h... |
| 02:52.44 | Notify | 03BRL-CAD:starseeker * 63005 brlcad/trunk/src/libbu/vers.c: add the header to vers.c |
| 03:16.37 | Notify | 03BRL-CAD:starseeker * 63006 (brlcad/trunk/src/adrt/isst_tcltk.c brlcad/trunk/src/conv/asc/asc2g.c): Linux header tweaks |
| 03:22.28 | starseeker | urk. http://sourceforge.net/p/tcl/mailman/message/32865888/ |
| 03:24.40 | starseeker | think that's got it... Windows build progressing now |
| 03:33.25 | starseeker | bingo |
| 03:58.51 | brlcad | starseeker: what was up with src/adrt/isst_tcltk.c ? |
| 03:59.40 | brlcad | header rules should put dm.h after all system headers, including gl.h and the bu headers etc |
| 04:24.42 | Notify | 03BRL-CAD:brlcad * 63007 brlcad/trunk/src/adrt/isst_tcltk.c: ws |
| 05:36.32 | brlcad | starseeker: quick test on my laptop with just libbu with and without header breakout, ~8% build time improvement (wallclock) best case |
| 05:38.41 | brlcad | takes about 10sec to build libbu, so good bit of noise in the numbers, but it is consistently faster (whew!) |
| 05:38.53 | brlcad | so probably warrants checking the overall build time |
| 05:39.29 | brlcad | here's a script that converts from breakout to bu.h: |
| 05:42.54 | brlcad | cd include ; for file in ../src/[^o]*/*.c ../src/[^o]*/*/*.c ../src/[^o]*/*/*.c ../src/[^o]*/*/*/*.c ; do second=0 ; for header in bu/*.h ; do if grep --quiet -E "include \"$header\"" $file ; then if test $second -eq 0 ; then echo "$file: SWAP $header for bu.h" ; perl -pi -e "s,include \"$header\",include \"bu.h\"," $file ; second=1 ; else echo "$file: KILL $header" ; perl -pi -e "s,.*include \"$header\",," $file ; fi ; fi ; done ; done ; cd |
| 05:43.06 | brlcad | meant to be run from the top-level source dir |
| 05:44.46 | brlcad | you'll want to comment out the header order check in bsocket.h and bnetwork.h too |
| 06:00.04 | brlcad | does a full build test on old mac |
| 06:00.39 | brlcad | (made sure to disable docs and prebuild src/other before "time make") |
| 06:04.52 | Notify | 03BRL-CAD:brlcad * 63008 brlcad/trunk/src/librt/primitives/pipe/pipe.c: pipe shadows pipe() when unistd is included |
| 06:06.54 | Notify | 03BRL-CAD:brlcad * 63009 brlcad/trunk/src/librt/primitives/pipe/pipe.c: didn't intend to inject bu.h |
| 07:11.05 | *** join/#brlcad mpictor (~mark@c-68-58-38-45.hsd1.in.comcast.net) | |
| 07:51.16 | *** join/#brlcad ries (~ries@D979EA84.cm-3-2d.dynamic.ziggo.nl) | |
| 07:54.26 | *** join/#brlcad teepee- (bc5c2133@gateway/web/freenode/ip.188.92.33.51) | |
| 09:37.51 | *** join/#brlcad clock (~clock@77-58-143-135.dclient.hispeed.ch) | |
| 09:41.05 | *** join/#brlcad teepee- (bc5c2133@gateway/web/freenode/ip.188.92.33.51) | |
| 10:01.34 | *** join/#brlcad clock (~clock@77-58-143-135.dclient.hispeed.ch) | |
| 10:42.10 | starseeker | brlcad: on Windows, gl.h needs some other stuff defined |
| 10:42.34 | starseeker | or rather, dm_xvars.h does |
| 10:43.50 | starseeker | no, sorry that one was GL/gl.h |
| 10:44.03 | starseeker | is looking at diff |
| 10:48.19 | starseeker | huh |
| 10:48.30 | starseeker | moves the header and tries again... |
| 10:52.01 | Notify | 03BRL-CAD:starseeker * 63010 brlcad/trunk/src/adrt/isst_tcltk.c: try moving gl.h back with the rest of the system headers - it should be there. |
| 10:58.18 | starseeker | I swear, sometimes these compiles feel like quantum computing... |
| 11:00.05 | ``Erik | iirc, gl.h on windows needed something weird like windows.h included before |
| 11:03.32 | starseeker | that's my recollection as well - but now it seems to be building with r63010 |
| 11:03.42 | starseeker | will have to try another clean build tonight |
| 11:05.23 | ``Erik | is common.h the first thing included? I'd think that might cover the windows required header issue |
| 11:05.59 | starseeker | it is, but it doesn't include windows.h |
| 11:23.24 | starseeker | reflects that if we *really* want to increase our build speed we'll have to port a lot of C++ to C... |
| 11:26.49 | starseeker | ``Erik: feel like a project? C versions of openNURBS and stepcode would be just the ticket :-P |
| 11:51.11 | *** join/#brlcad teepee (~teepee@gateway/tor-sasl/teepee) | |
| 12:16.14 | *** join/#brlcad yiyus (1242712427@je.je.je) | |
| 12:34.47 | *** join/#brlcad teepee- (bc5c2133@gateway/web/freenode/ip.188.92.33.51) | |
| 12:53.08 | *** join/#brlcad teepee- (bc5c2133@gateway/web/freenode/ip.188.92.33.51) | |
| 12:55.11 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) | |
| 13:27.31 | *** join/#brlcad teepee- (bc5c2133@gateway/web/freenode/ip.188.92.33.51) | |
| 14:08.57 | *** join/#brlcad teepee- (bc5c2133@gateway/web/freenode/ip.188.92.33.51) | |
| 14:19.30 | Notify | 03BRL-CAD:starseeker * 63011 (brlcad/trunk/CMakeLists.txt brlcad/trunk/src/other/CMakeLists.txt): Merge gct into main branch - doesn't look like working with this code will be disruptive enough to warrant keeping in its own branch. |
| 14:20.38 | Notify | 03BRL-CAD:starseeker * 63012 NIL: remove gct branch |
| 14:21.35 | Notify | 03BRL-CAD:starseeker * 63013 brlcad/trunk/misc/svn2git/archive_branches.sh: Add gct to git archive list |
| 14:40.57 | Notify | 03BRL-CAD:starseeker * 63014 brlcad/trunk/src/mged/fbserv.c: Fix bsd build of mged |
| 14:46.56 | Notify | 03BRL-CAD:starseeker * 63015 brlcad/trunk/src/libdm/fontstash/fontstash.h: clear warnings |
| 14:50.34 | Notify | 03BRL-CAD:n_reed * 63016 brlcad/branches/brep-debug/src/libbrep/boolean.cpp: fix leak |
| 14:56.18 | Notify | 03BRL-CAD:starseeker * 63017 brlcad/trunk/CMakeLists.txt: whoops, OSG enablement slipped in |
| 14:58.38 | *** join/#brlcad clock (~clock@77-58-143-135.dclient.hispeed.ch) | |
| 15:13.51 | brlcad | oof, that's actually why the header was namged bselect.h instead of bsocket.h |
| 15:16.01 | brlcad | sys/socket is both networking and non-networking sockets whereas the select api is socket agnostic (just works on descriptors) |
| 15:16.30 | brlcad | is a little torn |
| 15:20.30 | starseeker | brlcad: can leave the explicit socket include there - just wanted to check if it should be added or not |
| 15:20.38 | Notify | 03BRL-CAD:brlcad * 63018 (brlcad/trunk/include/bnetwork.h brlcad/trunk/include/bsocket.h): sys/socket is both networking and non-networking, which is part why it wasn't originally included but now with 'network' being the nomenclature, it does embiggen inclusion. it's consequently also why bsocket was named bselect, to avoid having headers with overlapping api (select.h is socket agnostic, sys/socket.h is far more |
| 15:20.40 | Notify | general). go ahead an include in both for now to see how things shake out but this probably either warrants combining them or going back to the bselect distinction. |
| 15:27.22 | brlcad | starseeker: it's more about having headers that fit expected intent implied by their name |
| 15:29.43 | brlcad | it's the difference between having a header named myselect.h and mysocket.h ... I would certainly expect the latter to provide sys/socket.h if I was told it's meant to be a wrapper |
| 15:30.33 | brlcad | could argue that it already belonged in myin.h vs mynetwork.h as socket.h is very much internet-centric (it just does much more too) |
| 15:35.38 | Notify | 03BRL-CAD:n_reed * 63019 brlcad/branches/brep-debug/src/libbrep/boolean.cpp: fix leak |
| 15:36.40 | Notify | 03BRL-CAD:starseeker * 63020 brlcad/trunk/regress/repository.sh: For in-src-dir build, don't do bio.h test on obj_rules.cpp - it incorporates the perplex scanner template, which we don't want to tie to BRL-CAD's specialized headers. |
| 15:44.06 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) | |
| 15:44.42 | Notify | 03BRL-CAD:starseeker * 63021 (brlcad/trunk/CMakeLists.txt brlcad/trunk/misc/CMakeLists.txt and 2 others): Merge in the OpenCL work from opencl branch, using the BRLCAD_ENABLE_OPENCL flag to turn it on/off. Need to actually try this and make sure it still runs. |
| 15:47.00 | Notify | 03BRL-CAD:starseeker * 63022 NIL: Delete opencl branch - work merged into trunk, can be enabled with BRLCAD_ENABLE_OPENCL |
| 15:47.57 | *** join/#brlcad bhattigurjot (~scott@117.214.202.39) | |
| 15:48.19 | Notify | 03BRL-CAD:starseeker * 63023 brlcad/trunk/misc/svn2git/archive_branches.sh: Add opencl to the archived branches list |
| 16:00.41 | bhattigurjot | Hi, I compiled qtged branch on my system. I have using Qt version 4.8.1 |
| 16:01.02 | bhattigurjot | should I compile against the latest qt? |
| 16:01.05 | bhattigurjot | 5+? |
| 16:01.20 | brlcad | bhattigurjot: did you run the tool successfully? |
| 16:01.37 | bhattigurjot | brlcad: yes |
| 16:02.03 | brlcad | part of your goal should be to become familiar with the current state of the tools, run mged and/or archer and use the qt display manager (dm) and framebuffer (fb) |
| 16:02.51 | brlcad | are you yet familiar with mged? |
| 16:02.51 | bhattigurjot | okay |
| 16:03.30 | bhattigurjot | not extensively.. I guess I should go through it thoroughly |
| 16:04.17 | brlcad | yeah, I suggest doing a couple of the exercises (if not all of them) on the Intro to MGED here: http://brlcad.org/wiki/Documentation |
| 16:04.58 | bhattigurjot | alright |
| 16:05.09 | brlcad | there's about 16 lessons, you don't need to do them all, but it's a good idea |
| 16:05.26 | brlcad | they take about 10-20 minutes each |
| 16:05.48 | bhattigurjot | yeah I did 1 or 2 few months back |
| 16:06.33 | bhattigurjot | I'll be doing them again |
| 16:06.35 | brlcad | okay, so maybe just a refresher |
| 16:06.46 | bhattigurjot | yeah |
| 16:06.48 | brlcad | a really quick way to test the Qt interface: |
| 16:07.06 | brlcad | the "attach" command and the "rt" command |
| 16:07.25 | brlcad | for example, start your qt mged: mged test.g |
| 16:07.32 | brlcad | create a sphere: make sph sph |
| 16:08.12 | brlcad | if you run the rt command, it'll create a 2d framebuffer window: rt |
| 16:08.55 | brlcad | if you run the attach command, it'll create (another) 3d display manager window: attach qt |
| 16:09.02 | Notify | 03BRL-CAD:starseeker * 63024 (brlcad/branches/qtged/CHANGES brlcad/branches/qtged/CMakeLists.txt and 1088 others): Sync qtged with trunk up to r63023. |
| 16:09.46 | brlcad | you can create different types of framebuffer and display manager windows by specifying other types, for example: |
| 16:09.50 | brlcad | rt -F/dev/X |
| 16:09.53 | brlcad | rt -F/dev/ogl |
| 16:09.57 | brlcad | attach X |
| 16:09.59 | brlcad | attach ogl |
| 16:10.00 | brlcad | etc |
| 16:10.30 | brlcad | compare them, get comfortable with using them |
| 16:11.02 | brlcad | then yes, the next useful step is probably to upgrade to qt 5 and make sure everything is still working at least as well as it is now |
| 16:11.03 | bhattigurjot | ok |
| 16:11.29 | starseeker | bhattigurjot: er, you might want to try updating qtged branch again - I just did a big sync with trunk :-/ |
| 16:11.35 | brlcad | if you see any differences between creating an X, ogl, and qt display manager or framebuffer, that would also be a perfect place to get working |
| 16:12.13 | bhattigurjot | starseeker: ok |
| 16:12.23 | brlcad | yeah, make sure the HEAD sources are still working :) notice commit r63024 |
| 16:13.08 | bhattigurjot | okay |
| 16:13.17 | starseeker | sorry, didn't realize that sync was landing right in the middle of someone working with the branch |
| 16:13.55 | starseeker | would have needed to be done at some point, but I wasn't trying to make life harder on purpose ;-) |
| 16:13.55 | brlcad | i thought that's why you did it ;) |
| 16:14.04 | starseeker | heh |
| 16:14.33 | starseeker | is thoroughly in clean-up mode... might as well ride the momentum while it lasts |
| 16:15.29 | starseeker | brlcad: do you know if the work in attr-extension-mods was merged? |
| 16:16.14 | starseeker | spose I should ask Tom... |
| 16:17.11 | starseeker | need to warn him he needs to sync his d-bindings branch soon anyway if he wants to keep it compatible |
| 16:19.39 | *** join/#brlcad kamal__ (~kamal@14.98.39.73) | |
| 16:21.38 | brlcad | it most definitely was not (and probably should not, lots of changes that need careful reviewing) |
| 16:23.25 | brlcad | i've got a nice isolated compile-test going on right now to check the header impact |
| 16:23.32 | starseeker | nods |
| 16:24.37 | starseeker | I'll leave attr-extension-mods alone then |
| 16:25.14 | brlcad | in theory, should get the benefit per lib additive, so it should be a good general measure of how it will shake out ... just working on getting stable numbers now without other influences (src/other, c++, etc) |
| 16:26.31 | brlcad | basically doing a full build, then running this: for i in lib* fb gtools mged nirt rt sig util ; do echo $i ; cd $i && make clean && make .. ; done && time make |
| 16:26.54 | brlcad | then running the script I pasted last night, rebuilding full, and re-running the above |
| 16:27.11 | brlcad | should give a reasonable picture |
| 16:27.44 | brlcad | oops |
| 16:28.02 | brlcad | for i in lib* fb gtools mged nirt rt sig util ; do echo $i ; cd $i && make clean && cd .. ; done && time make |
| 16:29.07 | starseeker | nods - should tell the story |
| 16:29.46 | starseeker | why the make .. after make clean? |
| 16:33.52 | brlcad | that was the oops... |
| 16:34.18 | starseeker | ah :-) |
| 16:34.20 | starseeker | sorry |
| 16:35.17 | brlcad | you try the script that converts all the bu's back to a single bu.h? |
| 16:35.39 | brlcad | that was fun, took about 15 minutes |
| 16:39.55 | starseeker | brlcad: did try it on my home machine - was pretty quick there for some reason (suspiciously so) |
| 16:40.48 | starseeker | build seemed slightly faster, but I didn't do a proper isolated test so hard to say |
| 16:41.00 | brlcad | shouldn't be faster :) |
| 16:41.00 | starseeker | will give it the full treatment tonight |
| 16:41.17 | starseeker | no, I mean the breakout of bu includes was slightly faster |
| 16:41.27 | brlcad | ahh, right |
| 16:41.31 | starseeker | sorry, mean one thing and saying another |
| 16:42.01 | starseeker | is trying to make sure everything builds everywhere at the moment... |
| 16:42.16 | brlcad | I think i'm going to do one more run with just the libs too as I suspect app linking is still going to dominate |
| 16:42.32 | brlcad | yay for distcheck-full |
| 16:42.35 | starseeker | heh |
| 16:42.51 | starseeker | the handy "melt your computer in three easy steps" tool |
| 16:43.37 | starseeker | probably is a decent stress test for operating systems in a lot of ways |
| 16:43.42 | Notify | 03BRL-CAD:n_reed * 63025 brlcad/branches/brep-debug/src/libbrep/boolean.cpp: fix leak |
| 16:43.44 | brlcad | we seriously need to get buildbot going so this all just happens |
| 16:43.51 | starseeker | agrees |
| 16:44.08 | bhattigurjot | Is it negligence on my part that I am not able to run "attach qt"? It says there are only two types valid: X and ogl |
| 16:44.21 | starseeker | what's needed? is the existing buildbot setup working and someone needs to just write the proper config? |
| 16:44.28 | brlcad | bhattigurjot: that means you probably did not compile against qt |
| 16:44.34 | brlcad | check your cmake log |
| 16:44.34 | starseeker | bhattigurjot: um. Did your build detect system Qt? |
| 16:44.53 | starseeker | cmake configure, not build |
| 16:47.48 | starseeker | brlcad: now that I think about it, does it make sense to maintain distcheck-full vs. distcheck? I know I wouldn't be happy with a release that didn't satisfy distcheck-full... |
| 16:50.07 | brlcad | yeah, I think so .. I find myself using both regularly |
| 16:50.16 | brlcad | quick check here, full check there, etc |
| 16:50.41 | starseeker | ok, cool |
| 16:50.44 | brlcad | and distcheck is ever so slightly easier to debug |
| 16:51.02 | starseeker | as long as someone is using the basic distcheck :-) |
| 16:51.15 | starseeker | just blasts all of them typically... |
| 16:51.16 | Notify | 03BRL-CAD:starseeker * 63026 (brlcad/branches/gecode/src/other/CMakeLists.txt brlcad/branches/gecode/src/other/gecode/CMakeLists.txt and 32 others): Update gecode to version 4.3.0 |
| 16:51.22 | brlcad | don't have to read down a convoluted make -j100 log to find which one *actually* failed, find the dir, hunt down the problem, ... |
| 16:52.23 | starseeker | raises eyebrow - if a sub-component of distcheck-full fails, I just look at the .log file for the build error. Can even cd into the distcheck-whatever subdir and work directly on the problem there |
| 16:53.01 | starseeker | if any of the tests complete, the directory associated with that test is gone and it's just the log file |
| 16:54.14 | brlcad | yeah, but which log file? |
| 16:54.20 | brlcad | that's my point |
| 16:54.28 | starseeker | the one associated with the test that left a directory behind |
| 16:54.42 | starseeker | if all of them left a directory behind, most any will do - it's a common problem then |
| 16:54.56 | brlcad | eh? they're all still there usually |
| 16:55.09 | brlcad | but only one or two got to the point of failure |
| 16:55.28 | starseeker | really? I often get situations where the debugs will complete but the release don't, or tk or in-src-dir fail while all the rest succeed |
| 16:55.49 | starseeker | in that situation (no-tk, for example) only the no-tk directory should be left |
| 16:56.08 | brlcad | usually all the dirs are there for me |
| 16:56.15 | brlcad | but not all fail |
| 16:56.23 | starseeker | O.o |
| 16:56.35 | starseeker | that's not good... |
| 16:57.00 | brlcad | anyways, the whole point is having to hunt and peck (i.e., randomly pick one) is not my debugging style either ... I want to see an error and trace it |
| 16:57.16 | starseeker | nods |
| 16:57.25 | brlcad | so I have to read the intermixed log output, which is a pita |
| 16:57.39 | brlcad | so distcheck wins if all I want is quick validation |
| 16:57.54 | starseeker | right |
| 16:57.56 | brlcad | which is probably 75% of the time unless it's time to release |
| 16:58.00 | brlcad | then it flips |
| 16:58.37 | starseeker | you'll have to show me if you get that case where only a few of the builds failed but extra directories got left behind - that's not supposed to happen |
| 16:58.57 | Stragus | Erik, it's terrible, the Elfga comics page is down! :) |
| 17:26.50 | *** join/#brlcad Notify (~notify@66-118-151-70.static.sagonet.net) | |
| 17:26.52 | Notify | 03BRL-CAD:n_reed * 63027 (brlcad/branches/brep-debug/CHANGES brlcad/branches/brep-debug/CMakeLists.txt and 842 others): sync from trunk through r63026 |
| 17:39.01 | *** join/#brlcad Notify (~notify@66-118-151-70.static.sagonet.net) | |
| 17:39.02 | Notify | 03BRL-CAD:n_reed * 63027 (brlcad/branches/brep-debug/CHANGES brlcad/branches/brep-debug/CMakeLists.txt and 842 others): sync from trunk through r63026 |
| 17:39.10 | Notify | 03BRL-CAD:n_reed * 63028 (brlcad/trunk/src/libbrep/intersect.cpp Property Changed: and 3 others): merge changes from brep-debug r61544; fixes bugs in brep intersection |
| 17:40.50 | Notify | 03BRL-CAD:n_reed * 63029 (brlcad/trunk/src/libbrep/boolean.cpp Property Changed: and 3 others): merge changes from brep-debug r62107; we shouldn't discard ssx curve intervals that lie on face boundaries |
| 17:40.55 | ``Erik | heh, it's up now |
| 17:55.48 | Notify | 03BRL-CAD:n_reed * 63030 (brlcad/trunk/src/libbrep/boolean.cpp brlcad/trunk/src/libbrep/intersect.cpp and 4 others): merge changes from brep-debug r62117, r62521; rename vars, quell warning |
| 18:09.44 | Notify | 03BRL-CAD:n_reed * 63031 (brlcad/trunk/src/libbrep/boolean.cpp Property Changed: and 3 others): merge changes from brep-debug r62416-r62519, r62522-62584; initial reworking of split_trimmed_face to perform boolean combinations of closed curves for trimmed face construction that's more robust and easier to debug |
| 18:15.21 | *** join/#brlcad clock (~clock@77-58-143-135.dclient.hispeed.ch) | |
| 18:15.33 | bhattigurjot | While compiling qtged branch against Qt, i'm getting the error regarding header file inclusion for bio.h to come after bsocket.h which is kinda surprising since bio.h is included after bsocket.h in include/fb.h |
| 18:15.56 | bhattigurjot | <PROTECTED> |
| 18:17.42 | Notify | 03BRL-CAD:carlmoore * 63032 brlcad/trunk/doc/docbook/system/man1/en/rt.xml: rewrite parts of rt |
| 18:18.49 | Notify | 03BRL-CAD:carlmoore * 63033 (brlcad/trunk/src/conv/asc/asc2g.c brlcad/trunk/src/libdm/osg-test.cpp): remove trailing blanks/tabs; fix spelling |
| 18:19.49 | Notify | 03BRL-CAD:n_reed * 63034 (brlcad/trunk/src/libbrep/boolean.cpp brlcad/trunk/src/libbrep/intersect.cpp): merge changes from brep-debug r62593-r62614; fix unhandled cases in the sub_curve routine and have it throw exceptions so callers can't ignore a NULL result |
| 18:33.11 | Notify | 03BRL-CAD:n_reed * 63035 (brlcad/trunk/src/libbrep/boolean.cpp Property Changed: and 3 others): merge changes from brep-debug r62623, r62660, r62674; more work on split_trimmed_face implementation |
| 18:39.24 | bhattigurjot | This is the complete cmake command I'm using: |
| 18:39.25 | bhattigurjot | "cmake ../brlcad-qtged -DBRLCAD_ENABLE_STRICT=NO -DBRLCAD_BUNDLED_LIBS=ON -DBRLCAD_ENABLE_QT=ON -DCMAKE_BUILD_TYPE=Debug" |
| 18:40.08 | bhattigurjot | I saw in the CMakeLists.txt file, that Qt5 is required. |
| 18:40.17 | Notify | 03BRL-CAD:n_reed * 63036 (brlcad/trunk/src/libbrep/intersect.cpp Property Changed: and 3 others): merge changes from brep-debug r62681-r62771; fixes bugs in brep intersection |
| 18:41.30 | bhattigurjot | Although I'm using Qt 4.8.1(checked using qmake -v) but my system does have qt5 files |
| 18:43.30 | Notify | 03BRL-CAD:starseeker * 63037 brlcad/trunk/src/libdm/CMakeLists.txt: add osg-test.cpp to ignore list |
| 18:44.59 | starseeker | bhattigurjot: the bio.h thing is most likely a consequence of recent header work |
| 18:45.53 | bhattigurjot | starseeker: ok.. so how do I resolve this? |
| 18:46.01 | starseeker | what file is throwing the error? |
| 18:46.24 | bhattigurjot | bsocket.h |
| 18:46.51 | starseeker | OK - what file is the build trying to compile when that error is generated? |
| 18:46.56 | Notify | 03BRL-CAD:n_reed * 63038 (brlcad/trunk/src/libbrep/boolean.cpp Property Changed: and 3 others): merge changes from brep-debug r62876; make linking of intersection curves from separate faces more robust |
| 18:46.58 | starseeker | bsocket.h is included by something else |
| 18:47.14 | bhattigurjot | dm-qt.cpp |
| 18:47.45 | starseeker | OK - try just removing the bio.h include from that file |
| 18:47.58 | starseeker | the other headers should take care of it, so you shouldn't need it explicitly |
| 18:48.12 | starseeker | (should be line 28) |
| 18:48.18 | bhattigurjot | yeah |
| 18:50.35 | bhattigurjot | starseeker: compiling now.. but is this "bio.h" required in dm-qt? |
| 18:51.21 | starseeker | it's most likely already included by other headers - that's why you were getting the conflict |
| 18:51.58 | bhattigurjot | okay,, btw this branch is it compatible with qt5? |
| 18:52.39 | starseeker | not sure - vlad was just starting work on this, so I'm not sure what version(s) he was assuming |
| 18:53.14 | bhattigurjot | I guess my cmake configuration picked qt5 files.. |
| 18:53.23 | starseeker | IIRC, Qt5 is needed primarily because of the need to embed Qt in Tk - we're trying to get away from that by using Qt instead of Tk, so either ought to work |
| 18:53.48 | bhattigurjot | yes |
| 18:53.59 | starseeker | Qt5 is the new version, so that's fine |
| 18:54.32 | starseeker | if Qt4 can be supported without too much work we will, but if Qt5 ends up a requirement we can live with that |
| 18:55.07 | bhattigurjot | alright |
| 18:56.15 | Notify | 03BRL-CAD:starseeker * 63039 brlcad/trunk/src/libdm/dm-qt.cpp: handled by other headers... |
| 18:57.53 | Notify | 03BRL-CAD:n_reed * 63040 (brlcad/trunk/src/libbrep/boolean.cpp Property Changed: and 3 others): merge changes from brep-debug r62901, r62907, r62910, r62925, r62931-r63025; more work on split_trimmed_face implementation |
| 19:02.45 | bhattigurjot | starseeker: on running "attach qt" I am getting two windows.. one titled 'dm_qt0' it is blank.. and the other one titled '(Unknown)' that shows the sphere I created |
| 19:03.10 | bhattigurjot | I guess it is not being embedded |
| 19:03.27 | starseeker | bhattigurjot: that's were vlad left off, IIRC - he had turned off the embedding of the Qt dm as a first step towards replacing the Tk window with a Qt window |
| 19:04.02 | *** join/#brlcad n_reed_ (~molto_cre@66-118-151-70.static.sagonet.net) | |
| 19:04.12 | bhattigurjot | ok |
| 19:04.35 | starseeker | so the things to do are 1) understand how the attach mechanism is working to create a Tk window and embed the dm, and how to do that with a Qt window |
| 19:05.05 | starseeker | 2) figure out how key bindings are set up for the Tk dm window and translate those binding to the Qt mechanism for key bindings |
| 19:05.19 | starseeker | 3) same thing for mouse bindings |
| 19:06.23 | starseeker | you want to get to the point where "attach qt" in the branch gives you the same functionality as "attach X" in trunk |
| 19:06.49 | bhattigurjot | okay |
| 19:07.39 | bhattigurjot | correct me if I am wrong.. what I have understood so far is that tcl script is being run to create a tk widget when we run "mged" |
| 19:08.02 | bhattigurjot | which I can see in mged.c |
| 19:08.07 | starseeker | in "gui" mode I believe that's correct |
| 19:08.21 | starseeker | "classic" mode - which is what this work is looking at - may be doing something more basic |
| 19:09.22 | starseeker | but the 'attach' command is a tcl command that will trigger other calls, since the MGED command prompt is a Tcl prompt (currently, anyway) |
| 19:10.35 | starseeker | I would send an email to the dev list and see if you can talk with vlad to find out what his design approach is |
| 19:11.06 | bhattigurjot | ok |
| 19:11.22 | bhattigurjot | I'll post on devel list |
| 19:21.19 | bhattigurjot | starseeker: from the points you listed above, for the 1st one, where should I begin looking into the code? |
| 19:21.37 | bhattigurjot | for the "attach" one |
| 19:35.25 | bhattigurjot | ok.. attach.c it is! |
| 19:43.46 | Notify | 03BRL-CAD:starseeker * 63041 brlcad/trunk/src/libbrep/CMakeLists.txt: Add brep_except.h to ignore list |
| 19:43.53 | brlcad | there's no point in worrying about qt4 .. by the time we're done with an alpha, qt6 will probably be out |
| 19:43.57 | brlcad | they're already on 5.3 |
| 19:44.16 | brlcad | we should target the latest only imho |
| 19:44.56 | starseeker | nods - works for me. I was figuring that if we stick to the C api (as opposed to using qml) then 4.8 might "just work" for a basic interface |
| 19:45.32 | starseeker | but the momentum is shifting to Qt5 now (KDE is either on it or close to it, IIRC) |
| 19:53.00 | Notify | 03BRL-CAD:n_reed * 63042 brlcad/trunk/src/libbrep/boolean.cpp: partial merge of changes from brep-debug r61367; don't forget the surface-surface overlap intersections; brep command now produces good output for all combs in librt/tests/brep_boolean_test.g that have non-empty results |
| 20:00.38 | Notify | 03BRL-CAD:starseeker * 63043 brlcad/trunk/src/libbrep/CMakeLists.txt: Add the surfacetree test code using queues as a patch for documentation purposes - needs to be rethought in light of subsequent work. |
| 20:02.21 | Notify | 03BRL-CAD:starseeker * 63044 NIL: saved work added here as a patch in trunk - this branch is no longer needed |
| 20:03.13 | Notify | 03BRL-CAD:starseeker * 63045 brlcad/trunk/misc/svn2git/archive_branches.sh: Add surfacetree to archived branches |
| 20:15.18 | Notify | 03BRL-CAD:starseeker * 63046 (brlcad/branches/qtged/doc/docbook/system/man1/en/rt.xml brlcad/branches/qtged/misc/svn2git/archive_branches.sh and 7 others): Update qtged to trunk r63045 |
| 20:35.02 | *** join/#brlcad clock (~clock@77-58-143-135.dclient.hispeed.ch) | |
| 22:40.28 | Notify | 03BRL-CAD:carlmoore * 63047 brlcad/trunk/doc/docbook/system/man1/en/rt.xml: Add 'This is' remark involving -J |
| 22:59.01 | Notify | 03BRL-CAD Wiki:Ernesto1955 * 0 /wiki/User:Ernesto1955: |