| 00:34.59 | *** join/#brlcad eqvvemnedeyjauos (~armin@dslc-082-083-185-155.pools.arcor-ip.net) |
| 02:38.01 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 04:56.43 | starseeker | nuts - our libregex doesn't build out of the box on openbsd - confliciting types for off_t |
| 05:01.46 | Stragus | o.O sys/types.h and what else would try to declare it? |
| 05:09.41 | *** join/#brlcad ickby (~stefan@x5d84595c.dyn.telefonica.de) |
| 05:27.17 | *** join/#brlcad ickby (~stefan@x5d84595c.dyn.telefonica.de) |
| 05:57.01 | *** join/#brlcad ickby_ (~stefan@x5d84595c.dyn.telefonica.de) |
| 06:12.42 | *** join/#brlcad ickby (~stefan@x5d84595c.dyn.telefonica.de) |
| 06:27.07 | *** join/#brlcad ickby_ (~stefan@x5d84595c.dyn.telefonica.de) |
| 07:02.26 | *** join/#brlcad shubham (71c18b41@gateway/web/freenode/ip.113.193.139.65) |
| 08:15.45 | *** join/#brlcad sniok (~sniok@89.252.2.135) |
| 08:26.30 | *** join/#brlcad boj_ (~boj@116.7.245.187) |
| 10:15.47 | *** join/#brlcad skat00sh (uid103741@gateway/web/irccloud.com/x-ljaeiisgnilaxekt) |
| 10:43.59 | *** join/#brlcad amarjeet (~Amarjeet@101.216.69.53) |
| 10:55.33 | *** join/#brlcad amarjeet (~Amarjeet@101.216.69.53) |
| 11:02.40 | *** join/#brlcad amarjeet (~Amarjeet@101.216.69.53) |
| 11:34.22 | *** join/#brlcad Ch3ck_ (~Ch3ck@154.70.108.196) |
| 11:38.34 | *** join/#brlcad Ch3ck_ (~Ch3ck@154.70.108.214) |
| 12:19.24 | *** join/#brlcad Ch3ck_ (~Ch3ck@154.70.108.210) |
| 12:48.25 | *** join/#brlcad Ch3ck_ (~Ch3ck@154.70.108.218) |
| 14:08.33 | *** join/#brlcad amarjeet (~Amarjeet@101.216.69.53) |
| 14:41.19 | *** join/#brlcad yorik (~yorik@187.35.19.227) |
| 14:57.48 | *** join/#brlcad Mathnerd314 (~quassel@supertux/Mathnerd314) |
| 15:33.21 | Notify | 03BRL-CAD:starseeker * 67810 brlcad/trunk/src/other/libregex/regex.h: OpenBSD defines off_t too |
| 15:36.06 | *** join/#brlcad ickby (~stefan@x5d84595c.dyn.telefonica.de) |
| 15:38.12 | Notify | 03BRL-CAD:starseeker * 67811 brlcad/trunk/include/common.h: OpenBSD actually defines __BEGIN_DECLS and __END_DECLS, so check before doing our own definition in common.h |
| 15:59.43 | Notify | 03BRL-CAD:starseeker * 67812 brlcad/trunk/src/other/libgdiam/gdiam.hpp: Floating point exact comparison warnings being triggered. |
| 16:00.32 | Notify | 03BRL-CAD:starseeker * 67813 brlcad/trunk/src/librt/primitives/bot/gct_decimation/auxiliary/mmthread.h: avoid conflict with signal definition in OpenBSD. |
| 16:04.39 | starseeker | OK, this one's got me stumped so far |
| 16:05.13 | starseeker | meshdecimation.c:3416: warning: ignoring alignment for stack allocated 'tdata' |
| 16:05.27 | starseeker | Stragus: does that ring any bells? |
| 16:06.54 | starseeker | I'm assuming it relates to the RF_ALIGN64 in the typdef for mdThreadData's struct, but I'm not sure what to do about it |
| 16:11.53 | Notify | 03BRL-CAD:starseeker * 67814 brlcad/trunk/src/librt/primitives/bot/gct_decimation/meshdecimation.c: For whatever reason, the aligned attribute appears to be trouble on OpenBSD (tdata allocated on stack is flagged as ignoring this attribute, which is enough to hault the build.) Until we sort out properly what to do, fall back to not using the aligned attribute on OpenBSD |
| 16:16.40 | Notify | 03BRL-CAD:starseeker * 67815 brlcad/trunk/src/librt/primitives/bot/gct_decimation/meshdecimationSSE2.c: These headers are also problematic on OpenBSD. |
| 16:31.54 | starseeker | openbsd agrees with the comment in clone.c:226 apparently |
| 16:35.09 | starseeker | does a make -k and reluctantly concludes that OpenBSD isn't going to happen this release |
| 17:11.17 | Stragus | starseeker, if I'm not mistaken, that RF_ALIGN64 was to ensure the per-thread chunks of data would all reside in different cache lines |
| 17:11.52 | Stragus | So one could just insert some padding to ensure that cores aren't trashing each other's cache constantly |
| 17:13.13 | Stragus | I guess the "right" solution would be to allocate that struct with alignment rather than putting it in the thread's main(), for compilers unable to align the stack |
| 17:15.07 | *** join/#brlcad ickby (~stefan@x5d84595c.dyn.telefonica.de) |
| 17:16.09 | Stragus | Looking at the code again, the impact of non-alignment would be insignificant |
| 17:28.23 | *** join/#brlcad ickby_ (~stefan@x5d84595c.dyn.telefonica.de) |
| 17:33.05 | *** join/#brlcad poxip (~poxip@unaffiliated/mrpoxipol) |
| 17:39.32 | *** join/#brlcad sniok (~sniok@89.252.2.135) |
| 18:51.51 | *** join/#brlcad amarjeet (~Amarjeet@101.216.69.53) |
| 19:07.28 | *** join/#brlcad amarjeet (~Amarjeet@101.216.69.53) |
| 20:29.11 | Notify | 03BRL-CAD:starseeker * 67816 brlcad/trunk/src/libgcv/plugins/obj/wfobj/obj_util.h: protect obj_util.h definitions - I think these are intended to allow wfobj to build stand-alone? |
| 20:42.16 | Notify | 03BRL-CAD:starseeker * 67817 brlcad/trunk/src/other/freetype/CMakeLists.txt: Set VERSION as well as SOVERSION |
| 21:26.12 | starseeker | OK, not as bad as I thought - that VERSION flag made a huge difference for some reason |
| 21:26.26 | starseeker | Now getting a lot of the "Array smaller" errors: |
| 21:26.36 | starseeker | http://brlcad.org/wiki/OpenBSD_Build_Log |