| 00:00.23 | mpictor | fedex_plus crashes with some schemas, when built with any of the last 3 commits |
| 00:01.22 | mpictor | #read_lazy_cpp_AP214E3_2010_s1-c5-214 |
| 00:01.26 | mpictor | oops |
| 06:58.55 | *** join/#stepcode kesha_ (~kesha@49.249.18.23) |
| 09:24.01 | *** join/#stepcode kesha_ (~kesha@49.249.18.23) |
| 12:44.11 | *** join/#stepcode mpictor (~mark@99-93-104-202.lightspeed.iplsin.sbcglobal.net) |
| 14:15.55 | *** join/#stepcode kesha (~kesha@49.249.1.158) |
| 14:28.31 | *** join/#stepcode kesha (~kesha@49.249.1.158) |
| 15:02.23 | *** join/#stepcode kesha_ (~kesha@49.202.239.115) |
| 17:13.48 | mpictor | kesha_: I found at least part of the problem finally |
| 17:14.25 | mpictor | there are two different versions of complexSupport.h |
| 17:15.51 | mpictor | the files in clstepcore used that complexSupport header, while the files still in fedex_plus/ used that header |
| 17:16.04 | mpictor | s/used/use |
| 17:18.20 | *** join/#stepcode kesha_ (~kesha@49.202.239.115) |
| 17:18.59 | mpictor | from src/clstepcore, I ran the following |
| 17:19.01 | mpictor | for i in `grep clstep ../fedex_plus/CMakeLists.txt |sed 's/ //g;'`; do sed -ie 's|#include "complexSupport.h"|#ifndef FEDEX\n #include "complexSupport.h"\n#else\n #include "../fedex_plus/complexSupport.h"\n#endif|g;' $i; done |
| 17:19.54 | mpictor | that replaces the includes with two different includes. which one is used depends on whether FEDEX is defined or not |
| 17:20.48 | mpictor | adding "add_definitions( -DFEDEX )" to src/fedex_plus/CMakeLists.txt defines that while fedex_plus is being compiled |
| 17:21.42 | mpictor | then I had to change declarations in src/fedex_plus/complexSupport.h to match those in the clstepcore version |
| 17:23.15 | mpictor | that fixed most of the schema failures with fedex_plus, and I suspect that the other problems are similar |
| 17:23.35 | mpictor | the real fix would be to merge those two headers |
| 17:25.34 | mpictor | p21read failed on some files, and the fix there was related to https://github.com/stepcode/stepcode/commit/1f1747e768e0f1bf4d38fe3a95e75b20da2919c0#L0L35 |
| 17:26.02 | mpictor | fedex_plus was happy with case-sensitive compare (perhaps because all strings were already converted) |
| 17:26.33 | mpictor | but p21read needs case-insensitive comparison |
| 17:27.01 | mpictor | changing that from strcmp to strcmpins seemed to work |
| 17:31.05 | mpictor | another option there might be to retain strcmp but change the case of the strings - perhaps one is converted to uppercase while the other is converted to lowercase or something |
| 17:36.38 | mpictor | when you started on this issue, I intended to mention two copies of that header as a potential problem, but it completely slipped my mind :/ |
| 18:24.34 | starseeker | mpictor: does this look like stepcode errors or express file errors? http://paste.lisp.org/display/137286 |
| 18:30.24 | mpictor | starseeker: hard to tell |
| 18:31.09 | mpictor | it always prints many of the "Possibly ambiguous implicit downcast" warnings |
| 18:31.22 | mpictor | and the errors may or may not be legit |
| 18:56.49 | mpictor | oops didn't realize I continued the conversation on #brlcad |
| 19:15.59 | mpictor | ok, the second error is due to the definition of entity kinematic_joint |
| 19:16.28 | mpictor | apparently they went with the ap242 definition, which is different from the one in ap214 |
| 19:18.12 | mpictor | commenting out the where rule (12636-12638) eliminates the problem |
| 19:18.46 | mpictor | libexpress parses those rules, but fedex_plus ignores them :/ |
| 19:22.29 | mpictor | functions aren't used either, so comment out 29175-29180 and we're down to 5 errors |
| 19:29.37 | mpictor | I'll look at the others later - looks like they'll be fairly easy to fix |
| 19:30.02 | mpictor | off to the bike shop to look at saddle bags |
| 19:37.32 | *** join/#stepcode cstirk (~quassel@c-71-56-216-45.hsd1.co.comcast.net) |
| 20:09.08 | *** join/#stepcode Christer (50d8cef6@gateway/web/freenode/ip.80.216.206.246) |