| 00:47.26 | *** mode/#brlcad [+o brlcad] by ChanServ |
| 04:46.33 | *** join/#brlcad Stattrav (~Stattrav@202.3.77.161) |
| 08:04.02 | *** join/#brlcad SWPadnos_ (~Me@dsl107.esjtvtli.sover.net) |
| 11:09.27 | CIA-43 | BRL-CAD: 03starseeker * r38120 10/brlcad/trunk/src/libgcv/obj/ (Makefile.am obj_grammar.y obj_parser.c obj_parser.h): First stab at stashing values in a data structure in obj |
| 11:22.39 | *** join/#brlcad R0b0t1 (~Enigma@unaffiliated/r0b0t1) |
| 11:28.19 | CIA-43 | BRL-CAD: 03bob1961 * r38121 10/brlcad/trunk/src/tclscripts/mged/ (asc2g.tcl dbupgrade.tcl g2asc.tcl points.tcl rt.tcl): Fixed several places that break on windows (i.e. needed to conditionally add a .exe to executable names). |
| 11:38.01 | CIA-43 | BRL-CAD: 03bob1961 * r38122 10/brlcad/trunk/src/tclscripts/lib/Ged.tcl: Spit out the length of both legs. |
| 12:29.39 | *** join/#brlcad mafm (~mafm@193.153.52.54) |
| 12:32.57 | *** join/#brlcad SWPadnos (~Me@emc/developer/SWPadnos) |
| 12:35.56 | CIA-43 | BRL-CAD: 03starseeker * r38123 10/brlcad/trunk/src/libgcv/obj/ (obj_grammar.y obj_parser.c): generalize obj_add_vertex a bit - this may not be the best approach here... |
| 12:48.17 | CIA-43 | BRL-CAD: 03starseeker * r38124 10/brlcad/trunk/src/libgcv/obj/obj_parser.c: print after parsing to make sure everything is really retained. |
| 12:53.58 | CIA-43 | BRL-CAD: 03davidloman * r38125 10/rt^3/trunk/ (include/GE/Logger.h src/GE/Logger.cxx): Clean up some mutex logic issues. Fixed Singleton implementation. |
| 13:04.50 | *** join/#brlcad mafm (~mafm@193.153.52.54) |
| 13:19.20 | CIA-43 | BRL-CAD: 03davidloman * r38126 10/rt^3/trunk/src/GE/Logger.cxx: Moved object lock to later in the log function. Locking that early was not needed. Also added a 'newline' to the end of each log entry so that the output is actually readable. |
| 13:20.23 | *** join/#brlcad d_rossberg (~rossberg@BZ.BZFLAG.BZ) |
| 13:32.28 | CIA-43 | BRL-CAD: 03davidloman * r38127 10/rt^3/trunk/src/iBME/: Modified SVN:IGNORE. Ignoring netMsgSerialTest binary. |
| 13:44.40 | CIA-43 | BRL-CAD: 03davidloman * r38128 10/rt^3/trunk/ (7 files in 3 dirs): Lots of bug fixes all towards making netPortalManagerTest work correctly. |
| 13:46.50 | d-lo_ | can anyone speak to the usefuleness of unit testing? Also, what are some recommended C++ Unit test suites? |
| 13:53.56 | brlcad | a tool like any other technique, one that can be abused, wasteful, unnecessary, useful, beneficial, necessary, ... |
| 13:55.34 | brlcad | often a crutch when it's simply hard/complicated to get your mind wrapped around a problem, but more a maintenance burden if it's a developing interface |
| 13:55.47 | brlcad | you double/triple your code complexity by forcing design contracts up front |
| 13:55.55 | brlcad | sometimes that complexity pays off, sometimes it does not |
| 13:57.02 | d-lo | okay, if not unit tests, then how would you recommend approaching 'testing' code as it is developed? keep simple test apps in a tests/ dir? |
| 14:09.10 | ``Erik | I've heard some people espouse TDD to help define interfaces between objects before developing them, that seems to be the big 'sell' usually |
| 14:09.52 | ``Erik | (kinda smells like people who can't or don't want to seriously think about the architecture to do SOMETHING to prevent design thrashing) |
| 14:10.22 | ``Erik | yeah, design contracts... should read the entire buffer before commenting heh |
| 14:15.43 | brlcad | d-lo: ideally there should always be system integration tests (answers high-level questions, does the tool work the way you're telling your user it should, do all the pieces work together), unit tests for 3rd party public API (e.g., GS net protocol, GE core object API), and regression tests for expected outcomes on sample inputs for nightly testing |
| 14:16.16 | brlcad | that's close to "ideal" .. and leaves room for lots of bloat and misdirection, but approaches ideal nonetheless |
| 14:18.15 | brlcad | priority is probably, #1 system integration, #2 regressions, #3 units .. unless you're working with a team intently decided on and doing test-driven-development in which case that'd be #1 system, #2 unit, and #3 regression |
| 14:21.44 | brlcad | decidedly departs out towards the mythed daystar |
| 14:49.49 | *** join/#brlcad mafm (~mafm@193.153.52.54) |
| 14:53.33 | CIA-43 | BRL-CAD: 03davidloman * r38129 10/rt^3/trunk/ (include/GE/Logger.h src/GE/Logger.cxx): Change logging calls to separate functions for the individual logging categories. |
| 14:58.17 | CIA-43 | BRL-CAD: 03davidloman * r38130 10/rt^3/trunk/src/GS/ (3 files in 2 dirs): Update classes that use Logger to use the new fns. |
| 15:12.49 | CIA-43 | BRL-CAD: 03davidloman * r38131 10/rt^3/trunk/tests/ (12 files in 4 dirs): New Dir heirarchy for tests. Moved over tests that were starting to accumulate in /src/tests. Added to cmake build system. |
| 15:15.47 | CIA-43 | BRL-CAD: 03davidloman * r38132 10/rt^3/trunk/tests/ (. GE/ GS/ GS/libNetwork/): Modified SVN:IGNORE to ignore cmake build byproducts |
| 15:19.00 | CIA-43 | BRL-CAD: 03davidloman * r38133 10/rt^3/trunk/src/ (iBME/CMakeLists.txt tests/): Removed old /src/tests dir and the cmake references to it. |
| 15:28.54 | *** join/#brlcad mafm (~mafm@83.45.253.170) |
| 16:29.07 | CIA-43 | BRL-CAD: 03davidloman * r38134 10/rt^3/trunk/ (3 files in 2 dirs): Removal of debug printing statements. |
| 16:30.38 | CIA-43 | BRL-CAD: 03davidloman * r38135 10/rt^3/trunk/tests/GS/libNetwork/: Modified SVN:IGNORE to ignore qmake and build byproducts |
| 16:31.30 | CIA-43 | BRL-CAD: 03davidloman * r38136 10/rt^3/trunk/tests/GS/libNetwork/ (NetPortalManagerTester.cxx NetPortalManagerTester.h): Implement NetPortalManagerTester as a thin wrapper for testing NetPortalManager. |
| 16:31.55 | CIA-43 | BRL-CAD: 03davidloman * r38137 10/rt^3/trunk/tests/GS/libNetwork/ (CMakeLists.txt netPortalManagerTest.cxx): Modify test to utilize new wrapper. |
| 16:32.45 | CIA-43 | BRL-CAD: 03davidloman * r38138 10/rt^3/trunk/CMakeLists.txt: Forgot to commit top level CMAKE changes for new test/ dir |
| 16:57.27 | CIA-43 | BRL-CAD: 03bob1961 * r38139 10/brlcad/trunk/src/mged/mged.c: Mods to read from stdin on Windows. |
| 17:10.15 | *** join/#brlcad parigaudi (~quassel@pd95b7f5e.dip0.t-ipconnect.de) |
| 17:46.13 | CIA-43 | BRL-CAD: 03starseeker * r38140 10/brlcad/trunk/NEWS: Bob fixed how Windows invokes fbclear on Windows, issue reported by Diane Carabetta. Also adjusted a few other calls which might have similar problems on Windows. |
| 18:02.58 | CIA-43 | BRL-CAD: 03bob1961 * r38141 10/brlcad/trunk/ (3 files in 3 dirs): Added a declaration for a rt_nmg_mc function in raytrace.h |
| 18:10.29 | CIA-43 | BRL-CAD: 03indianlarry * r38142 10/brlcad/trunk/include/opennurbs_ext.h: |
| 18:10.29 | CIA-43 | BRL-CAD: Function getCurveEstimateOfV() wasn't always converging to a solution before |
| 18:10.29 | CIA-43 | BRL-CAD: hitting iteration bound. Now uses binary subdivision of curve 't' parameter |
| 18:10.29 | CIA-43 | BRL-CAD: until derivatives at end points in 'u' are within 5 percent before using the |
| 18:10.29 | CIA-43 | BRL-CAD: original linear estimating method. Also increased the iteration bound to 1000. |
| 18:30.24 | ``Erik | heh |
| 20:14.22 | *** join/#brlcad R0b0t1 (~Enigma@unaffiliated/r0b0t1) |
| 20:42.40 | CIA-43 | BRL-CAD: 03erikgreenwald * r38143 10/brlcad/trunk/src/librt/primitives/nmg/nmg_tri_mc.c: use tables instead of magic numbers... some... |
| 23:44.52 | brlcad | ts: hard to say with the server out of date but sounds like a relatively new issue |
| 23:45.00 | brlcad | oops, disregard :) |
| 23:47.06 | ``Erik | heh |
| 23:48.46 | CIA-43 | BRL-CAD: 03brlcad * r38144 10/brlcad/trunk/include/ged.h: quell compilation warning, change abort to aborted to not hide abort(). |