| 01:48.37 | Notify | 03BRL-CAD:starseeker * 56496 brlcad/trunk/doc/docbook/system/mann/en/search.xml: No dash in front of exclamation mark for NOT in search |
| 02:00.19 | starseeker | tofu: if I want to dramatically change the librt interface to search, do I need to deprecate the old calls? I very seriously doubt anyone besides the search command used it, given how ackward it was/is to use... |
| 02:00.42 | starseeker | is willing to do it, but would feel a little silly... |
| 02:03.09 | starseeker | also, the dbfindtree tcl script command... I can probably make that a wrapper around search, but should we just deprecate it? |
| 02:10.27 | starseeker | hrm - 'make /test.s sph' causes some interesting problems |
| 03:31.18 | *** join/#brlcad PrezKennedy (~DarkCalf@173.231.40.99) |
| 03:36.38 | *** join/#brlcad ejno_ (~ejno@unaffiliated/kazaik) |
| 04:02.42 | *** join/#brlcad kesha (~kesha@14.139.122.114) |
| 04:44.31 | *** join/#brlcad Ch3ck (~Ch3ck@66-118-151-70.static.sagonet.net) |
| 06:16.38 | Notify | 03BRL-CAD Wiki:Harman052 * 5911 /wiki/User:Harman052/GSoc2013/Logs: |
| 08:35.03 | *** join/#brlcad evgeny (~Miranda@109.194.34.184) |
| 08:35.11 | *** part/#brlcad evgeny (~Miranda@109.194.34.184) |
| 08:45.03 | Notify | 03BRL-CAD:phoenixyjll * 56497 brlcad/trunk/src/libbrep/intersect.cpp: More powerful error handling - try another two options if the mid point is not sufficient. |
| 08:45.23 | Notify | 03BRL-CAD Wiki:Phoenix * 5912 /wiki/User:Phoenix/GSoc2013/Reports: /* Week 7 */ |
| 08:46.46 | Notify | 03BRL-CAD Wiki:Phoenix * 5913 /wiki/User:Phoenix/GSoc2013/Reports: /* Week 7 */ |
| 08:48.30 | Notify | 03BRL-CAD Wiki:Phoenix * 5914 /wiki/User:Phoenix/GSoc2013/Reports: /* Test Results */ |
| 08:49.37 | Notify | 03BRL-CAD Wiki:Phoenix * 0 /wiki/File:Extreme_final.png: |
| 09:16.23 | *** join/#brlcad vladbogo (~vlad@188.25.239.5) |
| 09:32.21 | *** join/#brlcad Ch3ck_ (~Ch3ck@195.24.220.16) |
| 09:48.32 | Notify | 03BRL-CAD Wiki:KeshaSShah * 5916 /wiki/User:KeshaSShah/GSoC13/Reports: /* Week 7 */ |
| 10:01.44 | Notify | 03BRL-CAD:mohitdaga * 56498 brlcad/trunk/src/libicv/CMakeLists.txt: Adding libicv/decimate.c. This will contain all the srinking and decimating utilities. Also adding fuction shrink_image. This function uses box averaging to shrink an image. |
| 10:24.34 | *** join/#brlcad Izak_ (~Izak@66-118-151-70.static.sagonet.net) |
| 10:46.31 | Notify | 03BRL-CAD:mohitdaga * 56499 brlcad/trunk/src/libicv/decimate.c: Add under_sample function. This will be one of the methods to decimate an image. |
| 11:29.41 | *** join/#brlcad avneet (~avneet@202.164.53.122) |
| 11:31.07 | Ch3ck_ | starseeker: saw the bug in all the patches that made them to return only correct results. I forgot to call the poly_init() function in the main() and this left all the array initialized to zero.. |
| 11:31.24 | Ch3ck_ | starseeker: reworking all of them and testing with wrong values to make sure everything is fine. |
| 12:49.40 | starseeker | Ch3ck_: that's looking better |
| 13:03.34 | Notify | 03BRL-CAD:starseeker * 56500 brlcad/trunk/src/libbn/tests/CMakeLists.txt: Add patch #221 from Nyah Check (Ch3ck) implementing a libbn test for bn_poly_multiply. |
| 13:04.22 | starseeker | Ch3ck_: remember if you do need to use functions like strcmp in the future, the HACKING file lists a number of functions for which we supply more robust replacements |
| 13:04.51 | starseeker | Ch3ck_: I suggest using the regression test 'make regress-repository' which will automatically check for a number of such issues (not all, but some) |
| 13:09.52 | zero_level | starseeker can u explain how regress works ? |
| 13:09.57 | zero_level | and what it checks ? |
| 13:10.06 | zero_level | Is there any document i can read about |
| 13:10.22 | starseeker | overall, regress is a broad series of integration tests that check whether functionality is working correctly |
| 13:10.31 | starseeker | raytracing, analysis codes, converters, etc. |
| 13:11.11 | starseeker | it works by launching shell scripts (currently) that in turn launch the programs and check that their output is as expected |
| 13:11.49 | starseeker | regress-repository is a bit different - it looks through the source code to make sure that certain coding conventions (for example, using bu_strcmp instead of strcmp) are followed |
| 13:12.40 | starseeker | any failure in those tests (like the previous issue of rtweight not doing what was expected) is a "regression" in functionality and hence constitutes a problem that must be fixed |
| 13:13.07 | starseeker | to run *all* the regression tests, you use the toplevel regress build target: make regress |
| 13:13.20 | starseeker | (the parallel -j flag will work, but makes the output harder to interpret) |
| 13:14.04 | starseeker | to run an individual test, you can run 'make regress-<testname>' |
| 13:14.40 | starseeker | to see what tests are available, check the CMakeLists.txt file in the regress repository |
| 13:14.59 | starseeker | for more details on what each tests actually does, check the .sh scripts in the regress directory |
| 13:15.36 | starseeker | zero_level: so the specific test that caught the rtweight issue was regress-weight |
| 13:15.46 | starseeker | there are several others that test the raytracer |
| 13:16.05 | zero_level | yeah. i see in my parrallel screen. There are lots of stuff in regeress |
| 13:16.29 | zero_level | do u think we will need regress testing for icv |
| 13:16.33 | zero_level | i see for gcv |
| 13:17.32 | starseeker | Probably not a bad idea - have small test images in .pix, .png, etc. format and ensure that the image converters using icv can reproduce the correct outputs given known inputs |
| 13:17.46 | starseeker | zero_level: I doubt you're quite there yet though |
| 13:18.08 | zero_level | alright. |
| 13:18.13 | starseeker | zero_level: once you reach the point where you have the icv design and implementation where you think it is in good shape, then we can come up with some tests |
| 13:18.49 | zero_level | alright. I will get back to you in october then. |
| 13:26.12 | starseeker | lots of possibilities - empty images, small single color images, automatically generating extremely large images (don't want to commit those to the repository, but there may be a way to automatically create some test inputs on the fly...) |
| 13:27.44 | starseeker | we'll probably lose some information (like transparency from PNG inputs) unless we move to something like openexr as the internal "pivot" format |
| 13:27.46 | ``Erik | icv should probably have unit tests wired into the ctest stuff |
| 13:28.52 | starseeker | nods - if that makes sense, then absolutely. Those can be developed in tandem with the library then |
| 13:29.02 | *** join/#brlcad avneet (~avneet@202.164.53.122) |
| 13:30.29 | zero_level | well i wrote in my trunk. |
| 13:30.42 | zero_level | But they are not very automatic. |
| 13:30.59 | ``Erik | libbu has examples of ctest style 'unit' tests being run, I'd argue that each function should have it's own add_test() line (maybe using the same executable with an argument to say which test to run) |
| 13:32.49 | zero_level | I mean for eg. icv_crop.c crops an image. but not automatically. there are two many values hardwired. |
| 13:33.07 | zero_level | and then i tested from matlab if it cropped at the said pixel. |
| 14:02.10 | *** join/#brlcad avneet (~avneet@202.164.53.122) |
| 14:38.00 | *** join/#brlcad avneet (~avneet@202.164.53.122) |
| 14:39.24 | Notify | 03BRL-CAD:mohitdaga * 56501 brlcad/trunk/src/libicv/decimate.c: Adding nintrep : nearest neighbour method to interpolate an image. |
| 15:06.45 | *** join/#brlcad vladbogo (~vlad@188.25.239.5) |
| 15:11.49 | *** join/#brlcad merzo_ (~merzo@253-199-133-95.pool.ukrtel.net) |
| 15:32.59 | *** join/#brlcad Ch3ck_ (~Ch3ck@195.24.220.16) |
| 15:33.06 | Ch3ck_ | starseeker:ok |
| 15:35.48 | Ch3ck_ | starseeker: i have also fixed 222 and 223 tickets |
| 16:17.05 | Ch3ck_ | starseeker: 224,225 |
| 16:17.08 | Ch3ck_ | fixed |
| 16:17.36 | *** join/#brlcad Ch3ck___ (~Ch3ck@195.24.220.16) |
| 16:29.47 | *** join/#brlcad Notify (~notify@66-118-151-70.static.sagonet.net) |
| 17:23.23 | Notify | 03BRL-CAD Wiki:NyahCh3ck20 * 5917 /wiki/User:NyahCh3ck20/GSoc2013/Coding_Repor: /* 29 July - 4 August */ |
| 17:46.02 | Notify | 03BRL-CAD:mohitdaga * 56502 brlcad/trunk/src/libicv/decimate.c: Bilinear interpolation binterp for up scaling of images. |
| 17:49.04 | *** join/#brlcad Izak_ (~Izak@66-118-151-70.static.sagonet.net) |
| 18:03.04 | Notify | 03BRL-CAD:starseeker * 56503 brlcad/trunk/src/libicv/decimate.c: fix typo |
| 18:04.41 | Notify | 03BRL-CAD:starseeker * 56504 brlcad/trunk/TODO: Add a TODO note that Archer needs to remember the command window size and placement when it is in detached mode. |
| 18:05.31 | starseeker | ``Erik: dumb question - if argv lists are null terminated, why is there a need to pass in argc? Why not just deduce it from the argv array? |
| 18:08.23 | zero_level | starseeker : Indded was just making a commit correcting that and other. |
| 18:08.32 | zero_level | anyways. Thanks |
| 18:08.44 | starseeker | zero_level: heh - sure. Just saw broken build, fixed it |
| 18:12.38 | zero_level | alright. |
| 18:16.37 | zero_level | starseeker : is there some complex process to change the file name in svn |
| 18:16.47 | zero_level | maintaining the log. |
| 18:17.56 | zero_level | or it is trivial mv command. |
| 18:18.00 | Notify | 03BRL-CAD:mohitdaga * 56505 brlcad/trunk/src/libicv/decimate.c: Fix Memory allocation issue and parameter. |
| 18:18.22 | zero_level | I searched, they recommend svn mv. |
| 18:18.39 | zero_level | but using this are the logs entacted ? |
| 18:27.23 | starseeker | svn mv |
| 18:27.39 | starseeker | there are ways to get the history |
| 19:01.09 | Ch3ck_ | starseeker: I have fixed almost all the patches on the last one |
| 19:01.20 | Ch3ck_ | you could check them and give me feedback ;) |
| 19:54.38 | ``Erik | starseeker: I dunno... is the argv array itself guaranteed to be null terminated? (could be that it didn't used to be null terminated, but so much stuff used argc, it was left in) *shrug* that'd be an interested project, digging up the early details of argv/argc :D |
| 19:55.16 | Notify | 03BRL-CAD Wiki:Level zero * 5918 /wiki/User:Level_zero/GSOC13/logs: /* Week 7 */ |
| 19:56.02 | ``Erik | starseeker: here's an interesting tidbit, if the program itself cannot determine it's name, argv[0] will be NULL even if there are valid arguments |
| 20:08.31 | Ch3ck_ | starseeker: finished updating all patches. They all should apply now.. waiting for feedback. |
| 20:22.30 | *** join/#brlcad Ch3ck_ (~Ch3ck@195.24.220.16) |
| 22:25.14 | Notify | 03BRL-CAD Wiki:Vladbogolin * 5919 /wiki/User:Vladbogolin/GSoC2013/Logs: /* Week 7 */ |