| 00:00.32 | Notify | 03BRL-CAD:brlcad * 68682 (brlcad/trunk/src/libdm/dm-X.c brlcad/trunk/src/libdm/dm-generic.c and 9 others): reduce the number of Tcl symbols in libdm. don't need to be using TCL_OK/ERROR when we have equivalent API. |
| 00:44.20 | *** join/#brlcad dzbybkubocwmtqdy (~armin@dslb-088-066-144-194.088.066.pools.vodafone-ip.de) | |
| 01:00.12 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) | |
| 04:54.22 | *** join/#brlcad tandoorichick (b64b2d01@gateway/web/freenode/ip.182.75.45.1) | |
| 06:47.24 | *** join/#brlcad teepee_ (~teepee@unaffiliated/teepee) | |
| 08:09.41 | *** join/#brlcad sniok (~sniok@89.252.2.135) | |
| 08:18.15 | *** join/#brlcad Caterpillar2 (~caterpill@unaffiliated/caterpillar) | |
| 08:53.04 | *** join/#brlcad tandoorichick (3d0c28b1@gateway/web/freenode/ip.61.12.40.177) | |
| 09:35.01 | *** join/#brlcad merzo (~merzo@92.60.189.225) | |
| 09:35.01 | *** join/#brlcad teepee] (bc5c2134@gateway/web/freenode/ip.188.92.33.52) | |
| 10:58.26 | *** join/#brlcad asad_ (~asad00@host10-2.natpool.mwn.de) | |
| 12:13.23 | *** join/#brlcad sniok (~sniok@89.252.2.135) | |
| 12:58.34 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) | |
| 14:04.27 | *** join/#brlcad amarjeet (~amarjeet@101.211.212.152) | |
| 14:35.26 | Notify | 03BRL-CAD:starseeker * 68683 brlcad/trunk/regress/repository.cmake: set up so we can run all tests with a single variable define. |
| 14:48.02 | Notify | 03BRL-CAD:starseeker * 68684 brlcad/trunk/regress/repository.cmake: printing tweaks |
| 14:49.13 | Caterpillar2 | [16:06] <starseeker> fyi - there was apparently some interest in an openNURBS package back in 2011: https://lists.fedoraproject.org/archives/list/legal@lists.fedoraproject.org/thread/JP2RTIBREMZ5SIFMJ35AHXXQDJHQPQLX/ |
| 14:50.02 | Caterpillar2 | hi, I have just contacted Jiri about openNURBS. He told me he gave up due opennurbs shipping zlib library. I think it is quite easy to unbundle it, I had a similar experience while packaging darktable software |
| 14:51.45 | Caterpillar2 | (zlib is already shipped in Fedora and its version is 1.2.8) |
| 14:56.39 | Notify | 03BRL-CAD:starseeker * 68685 brlcad/trunk/regress/repository.cmake: More logic tweaks. |
| 14:56.49 | Caterpillar2 | if you have free time, / if you want, we could work together on packaging all libraries and finally brl-cad |
| 14:57.29 | Caterpillar2 | if not, I can do it, but it will require an undefinite amount of time |
| 15:01.30 | Notify | 03BRL-CAD:brlcad * 68686 brlcad/trunk/regress/repository.sh: all hial hiaku |
| 15:03.23 | brlcad | Caterpillar2: happy to support you in any way we can |
| 15:04.06 | brlcad | Caterpillar2: regarding opennurbs, is it not enough to simply ensure that opennurbs' bundled zlib is not compiled -- that it's using the system/package-installed zlib? |
| 15:04.43 | brlcad | that said, our bundling of openNURBS has zlib stripped out (because we already bundle it) ;) |
| 15:05.29 | brlcad | you could probably just turn our src/other/openNURBS directory into a package |
| 15:05.46 | brlcad | that would solve two problems |
| 15:06.22 | Caterpillar2 | [17:05] <brlcad> you could probably just turn our src/other/openNURBS directory into a package |
| 15:06.35 | Caterpillar2 | no the Fedora policy says that you must take the upstream package |
| 15:06.56 | Caterpillar2 | [17:04] <brlcad> Caterpillar2: regarding opennurbs, is it not enough to simply ensure that opennurbs' bundled zlib is not compiled -- that it's using the system/package-installed zlib? |
| 15:07.00 | Caterpillar2 | yeah it should work |
| 15:09.28 | brlcad | I'm suggesting that there is no upstream, that you are defining upstream |
| 15:10.14 | Caterpillar2 | brlcad: mmh can you rewrite your statement? Sorry, English is not my first language ;-) |
| 15:10.16 | brlcad | dumping a tarball on a website without any support hardly makes for your traditional "upstream" supplier |
| 15:11.52 | brlcad | Caterpillar2: you're talking about pulling a tarball from a URL, applying a set of patches (to remove zlib, apply other fixes) |
| 15:13.00 | brlcad | all I was suggesting is changing the URL (e.g., to a fork we can put on github) so you don't have to patch |
| 15:13.21 | brlcad | it doesn't matter, just offering options :) |
| 15:13.45 | Caterpillar2 | brlcad: is https://www.rhino3d.com/en/opennurbs the upstream developer of openNURBS? |
| 15:14.44 | brlcad | depends how you define upstream, but sure -- that's certainly where we forked from |
| 15:15.41 | brlcad | Caterpillar2: hablas español? what's your native language? |
| 15:15.49 | Caterpillar2 | brlcad: Italian |
| 15:17.07 | brlcad | ah, molto bene ... tranne il mio italiano é terrible :) |
| 15:17.36 | Caterpillar2 | brlcad: ehhe |
| 15:18.18 | Caterpillar2 | https://fedoraproject.org/wiki/Packaging:Guidelines#Bundling_and_Duplication_of_system_libraries |
| 15:18.23 | brlcad | ho studiato per un paio di anni |
| 15:18.39 | brlcad | ma ho dimenticato .. tutti |
| 15:18.40 | Caterpillar2 | di dove sei? |
| 15:19.27 | brlcad | depende lo che cuenta |
| 15:19.43 | Caterpillar2 | ? |
| 15:19.43 | brlcad | vivo in us |
| 15:19.46 | Caterpillar2 | ah ok |
| 15:20.36 | brlcad | fully fluent in spanish, but also rusty writing it |
| 15:21.44 | Caterpillar2 | nice |
| 15:21.57 | brlcad | anyways... so yeah, rhino3d are the main opennurbs developers but we have a fork that is improved, in ways they will not integrate |
| 15:22.52 | brlcad | we have fixes that they don't care about (we've pushed them upstream), we have other changes they specifically exclude for business reasons |
| 15:23.06 | brlcad | they sell a product containing the same modifications we make |
| 15:24.07 | Caterpillar2 | brlcad: ok, so I have to ask to Fedora Packaging Committee if they give me an exception to the rules. There is a specific ticket system for such problems |
| 15:24.09 | brlcad | so we could treat our version as a completely separate fork (e.g., Xorg vs X11 or ECGS vs GCC or thunderbird vs firefox, etc) |
| 15:24.31 | Caterpillar2 | brlcad: interesting |
| 15:24.53 | brlcad | either way, we need the patches we made -- geometry fails to render correctly without at least one of them |
| 15:25.14 | brlcad | rendering is something they don't support |
| 15:26.16 | brlcad | if needed, we can set up a proper repository just for our fork, or we can revisit discussions with the rhino folks (we don't really want to maintain a fork) |
| 15:27.15 | Caterpillar2 | brlcad: if you revisit discussions with the rhino folks it would be perfect |
| 15:28.37 | brlcad | we can try, but it will probably take a while to sort out succinct patches again for discussion |
| 15:28.58 | brlcad | starseeker: care to try upgrading opennurbs next? :) |
| 15:29.09 | Caterpillar2 | brlcad: yeah I can imagine |
| 15:32.35 | Notify | 03BRL-CAD:starseeker * 68687 brlcad/trunk/regress/repository.cmake: start working on new regex match |
| 15:38.17 | Caterpillar2 | brlcad: I have to go out home in a few minutes, for any question you can send me an e-mail or just wait for me to get back later :-)) |
| 15:45.33 | *** join/#brlcad asad_ (~asad00@host10-2.natpool.mwn.de) | |
| 15:56.04 | brlcad | starseeker: I note that [# ] != [[:space:]#] ... you want something like [ \t\r\n\f#] if it doesn't support the posix classes |
| 15:59.04 | Notify | 03BRL-CAD:starseeker * 68688 brlcad/trunk/regress/repository.cmake: Make the cmake outputs a bit closer to the repository.sh output. CMake matching is slower than repository.h, but we're also catching some cases like src/mged/qray.h:38: #ifndef _WIN32 that I'm not seeing in the repository.sh output. Maybe it's time to see if the unifdef program can be used to do the symbol extraction piece, rather than cobbling |
| 15:59.06 | Notify | together regex patterns... |
| 15:59.08 | Notify | ... |
| 16:05.35 | *** join/#brlcad amarjeet (~amarjeet@101.211.209.70) | |
| 16:14.47 | brlcad | starseeker: matching that qray.h line is technically an incorrect match for that regular expression |
| 16:17.55 | brlcad | obviously a desirable match in this particular instance, but not what the regular expression says -- it says match platform symbol followed by a non A-Z character (and there are no more chars on that line) |
| 16:18.53 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) | |
| 16:24.56 | starseeker | brlcad: no argument |
| 16:25.35 | starseeker | brlcad: I can take a run at openNURBS, sure |
| 16:25.59 | starseeker | be a little bit - want to get this repository symbol stuff sorted if we can... |
| 16:27.09 | starseeker | brlcad: I'm trying an experiment with unifdef, which is showing a bit of promise |
| 17:00.20 | *** join/#brlcad ickby (~stefan@x5d84498e.dyn.telefonica.de) | |
| 17:27.25 | *** join/#brlcad sniok (~sniok@89.252.2.135) | |
| 17:36.53 | *** join/#brlcad ickby (~stefan@x5d84498e.dyn.telefonica.de) | |
| 19:16.55 | Notify | 03BRL-CAD:starseeker * 68689 brlcad/trunk/regress/repository.cmake: Tweak regex, fix quick test for line |
| 19:23.13 | *** join/#brlcad sniok (~sniok@89.252.2.135) | |
| 19:36.35 | Notify | 03BRL-CAD:brlcad * 68690 (brlcad/trunk/regress/repository.cmake brlcad/trunk/regress/repository.sh): tweak the regex to match symbols ending on the end of line. remove MACH as it's really a hardware platform. |
| 19:36.42 | brlcad | starseeker: why in the world are there two platform lists already?? :) |
| 19:39.33 | *** join/#brlcad Caterpillar2 (~caterpill@unaffiliated/caterpillar) | |
| 19:39.38 | Caterpillar2 | back again |
| 19:40.08 | brlcad | seems like premature complication... |
| 19:40.12 | brlcad | hi Caterpillar2 |
| 19:44.53 | brlcad | starseeker: is doing the two non-regex MATCHES followed by the regex MATCHES actually faster? that's a bit surprising |
| 19:46.13 | brlcad | also, how can I manually test it? get error with this: |
| 19:46.14 | brlcad | agua:brlcad.trunk morrison$ cmake -P regress/repository.cmake |
| 19:46.14 | brlcad | CMake Error at regress/repository.cmake:97 (list): |
| 19:46.15 | brlcad | <PROTECTED> |
| 19:56.25 | Notify | 03BRL-CAD Wiki:Tandoorichick * 9815 /wiki/User:Tandoorichick/GSoC2016/Logs: /* Development Logs */ |
| 20:07.17 | Caterpillar2 | brlcad: so, I will wait for your news before doing anything with openNURBS |
| 20:09.03 | brlcad | well assume all goes perfectly or poorly, I'm not sure it changes much from a packaging perspective unless we host a fully managed fork ourselves |
| 20:09.24 | brlcad | and the only incentive for doing that right now would be for packaging |
| 20:09.53 | Caterpillar2 | ok |
| 20:11.39 | brlcad | so you're either waiting to hear if we need to fork (less work for you, more work for us) or package them as-is (more work for you, less work for us), yes? |
| 20:13.59 | Caterpillar2 | Rather than forking, we were talking about you revisiting discussions with Rhino guys |
| 20:14.05 | Caterpillar2 | I am not in a hurry, so we can take all the time we need. |
| 20:18.56 | brlcad | yes we revisit discussions with rhino guys, that wasn't in question :) |
| 20:19.31 | brlcad | you said you would wait ... I'm trying to understand why you'd need to wait as the outcomes of good or bad discussions may be the same (for you) |
| 20:19.43 | brlcad | it's only different if we fork |
| 20:27.46 | Caterpillar2 | brlcad: the difference for me is: if your openNURB patches will be accepted upstream, I can simply take the new & patched upstream openNURBS code and package it into Fedora. If those patches are not accepted by upstream, I will have to package BRL-CAD modified version of openNURBS, I will be trapped for some weeks into Fedora burocracy emails, explaining and asking permissions to Fedora Packaging |
| 20:27.48 | Caterpillar2 | Committee, etc. |
| 20:30.32 | Caterpillar2 | but the final result should be almost the same |
| 20:44.36 | Caterpillar2 | now I have to go to bed, see you tomorrow! |
| 20:46.47 | Stragus | ponders regex with JIT compilation of SSE 4.2 string operations |
| 21:06.44 | *** join/#brlcad infobot (ibot@rikers.org) | |
| 21:06.44 | *** topic/#brlcad is BRL-CAD release 7.26.0 is out! More than 150 user-visible changes including 6 major efforts! || GSoC 2016 is coming to a close, showcase forthcoming || Help needed reviewing and integrating 700+ GCI tasks || Logs: http://ibot.rikers.org/%23brlcad/ | |
| 21:15.52 | Notify | 03BRL-CAD:starseeker * 68691 (brlcad/trunk/misc/tools/CMakeLists.txt brlcad/trunk/regress/repository.cmake): Use a customized version of unifdef to identify and report symbols in the source code. Appears to be both fast and robust. |
| 21:19.48 | Notify | 03BRL-CAD:starseeker * 68692 brlcad/trunk/misc/tools/unifdef/platform_symbols.h: remove MACH from os platform symbols list |
| 21:21.29 | starseeker | brlcad: to manually test, it's cmake -DSOURCE_DIR=/path/to/brlcad -DUNIFDEF_EXEC=/path/to/customized/unifdef -DRUN_ALL_TESTS=1 -P repository.cmake |
| 21:22.34 | starseeker | (sorry, wasn't watching channel) |
| 23:13.43 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) | |
| 23:15.49 | brlcad | starseeker: I still get the subfilter error |
| 23:16.12 | brlcad | is that perhaps using some >3.3 cmake feature? |
| 23:52.31 | Notify | 03BRL-CAD Wiki:Sean * 9816 /wiki/Example_db_walk_tree: merge in the void* update from asad |
| 23:52.45 | Notify | 03BRL-CAD Wiki:Sean * 0 /wiki/Db_walk_tree: not needed, merged with main |