| 00:31.40 | *** join/#brlcad dvketuyyvnwbheux (~armin@dslb-088-066-148-076.088.066.pools.vodafone-ip.de) | |
| 01:05.44 | *** join/#brlcad merzo (~merzo@167-8-132-95.pool.ukrtel.net) | |
| 03:42.11 | *** join/#brlcad gabbar1947 (uid205515@gateway/web/irccloud.com/x-iyjhfqpvsftxqrol) | |
| 05:30.23 | Notify | 03BRL-CAD:brlcad * 70052 brlcad/trunk/include/vmath.h: include an implementation of linear interpolation (lerp) between two vectors/points. take care to manage floating point error if we used the mathematically simpler form. lerp is useful for interpolating between two points, alpha blending, animation, and more. |
| 06:44.32 | *** join/#brlcad Caterpillar (~caterpill@unaffiliated/caterpillar) | |
| 07:30.55 | *** join/#brlcad teepee (~teepee@unaffiliated/teepee) | |
| 07:48.55 | Notify | 03BRL-CAD Wiki:Mariomeissner * 10174 /wiki/User:Mariomeissner/logs: |
| 08:00.12 | *** join/#brlcad vasc (~vasc@bl4-144-10.dsl.telepac.pt) | |
| 08:00.32 | vasc | fun fact: look at https://svn.code.sf.net/p/brlcad/code/brlcad/trunk/src/librt/bool.c |
| 08:00.48 | vasc | bool_eval() is a HIDDEN i.e. static function. |
| 08:01.48 | vasc | it's only called in rt_boolfinal(). like this: bool_eval(regp->reg_treetop, pp, TrueRg, ap->a_resource) |
| 08:02.51 | vasc | TrueRg is declared as "struct region *TrueRg[2]" and is used to return some values. but TrueRg isn't used anywhere! eliminating this parameter from bool_eval would remove a lot of branches as well. |
| 08:03.57 | *** join/#brlcad teepee (~teepee@unaffiliated/teepee) | |
| 08:04.03 | vasc | so... why was it here in the first place? leftovers from and older code implementation? so why is this appendix here? |
| 08:04.13 | vasc | s/and/an/ |
| 08:04.28 | vasc | questions, questions... |
| 08:29.01 | *** join/#brlcad merzo (~merzo@user-94-45-58-139.skif.com.ua) | |
| 08:58.32 | Notify | 03BRL-CAD Wiki:Mariomeissner * 10175 /wiki/User:Mariomeissner/logs: |
| 09:19.03 | *** join/#brlcad teepee (~teepee@unaffiliated/teepee) | |
| 09:42.05 | *** join/#brlcad teepee (~teepee@unaffiliated/teepee) | |
| 11:00.12 | *** join/#brlcad teepee (~teepee@unaffiliated/teepee) | |
| 12:35.37 | *** join/#brlcad vasc (~vasc@bl4-144-10.dsl.telepac.pt) | |
| 13:20.49 | *** join/#brlcad yorik (~yorik@2804:431:f721:ff6:290:f5ff:fedc:3bb2) | |
| 13:34.59 | *** join/#brlcad merzo (~merzo@171-17-203-46.pool.ukrtel.net) | |
| 14:11.30 | *** join/#brlcad merzo (~merzo@user-94-45-58-139.skif.com.ua) | |
| 15:24.52 | Notify | 03BRL-CAD:starseeker * 70053 (brlcad/trunk/CMakeLists.txt brlcad/trunk/include/analyze.h and 26 others): Try a mechanism to 'override' the dll import/export mechanism when we're doing static builds with MSVC. I think this will prove simpler than trying to apply or not apply the DLL import/export definitions based on the build type, if it works. (not tested yet with MSVC) |
| 15:30.06 | *** join/#brlcad vasc (~vasc@bl4-144-10.dsl.telepac.pt) | |
| 15:38.20 | Notify | 03BRL-CAD:starseeker * 70054 brlcad/trunk/CMakeLists.txt: simplify test |
| 16:27.05 | Notify | 03BRL-CAD:starseeker * 70055 (brlcad/trunk/CMakeLists.txt brlcad/trunk/misc/CMake/BRLCAD_Targets.cmake): Sigh. OBJECT libraries + MSVC isn't happy - y2038-obj genrates y2038-obj.lib, which is successfully used by y2038 to make a dll. However, y2038 doesn't in turn make a y2038.lib file, which means libbu can't link successfully. |
| 16:30.51 | *** join/#brlcad djkonro (~djkonro@129.0.42.236) | |
| 18:22.55 | Notify | 03BRL-CAD:starseeker * 70056 (brlcad/trunk/misc/CMake/CMakeLists.txt brlcad/trunk/src/conv/step/step-g/CMakeLists.txt and 2 others): cotire is an interesting idea, but probably not something to turn on just yet... |
| 19:00.52 | *** join/#brlcad merzo (~merzo@56-50-133-95.pool.ukrtel.net) | |
| 20:20.24 | *** join/#brlcad merzo (~merzo@56-50-133-95.pool.ukrtel.net) | |
| 20:33.49 | *** join/#brlcad yorik (~yorik@2804:431:f721:ff6:290:f5ff:fedc:3bb2) | |
| 21:53.20 | *** join/#brlcad vasc (~vasc@bl4-144-10.dsl.telepac.pt) | |
| 21:57.37 | vasc | mdtwenty's been working on the new OpenCL bool_eval() prototype. the performance prolly can still be improved, been talking with him about it, but it's faster than the code i wrote and only slightly slower (like 1% slower) than the existing code in trunk/. |
| 21:58.15 | vasc | still i expected it to be slower, it's probably because of the stack management code, i've suggested some changes for him to try. |
| 21:58.24 | vasc | s/slower/faster/ |
| 22:14.17 | *** join/#brlcad merzo (~merzo@56-50-133-95.pool.ukrtel.net) | |
| 22:23.21 | Notify | 03BRL-CAD:Marco-domingues * 0 /wiki/File:Comparison.png: |
| 22:23.50 | Notify | 03BRL-CAD:Marco-domingues * 10177 /wiki/User:Marco-domingues/GSoC17/Log: /* 7 August */ |