| 02:34:23 | CIA-11: | martinl * r31992 /grass/trunk/gui/wxpython/ ( 5 files in 2 dirs ): nviz2/wxGUI: initial support for vector line overlays added |
| 02:34:28 | sigq: | grassfeed: Changeset [31992]: nviz2/wxGUI: initial support for vector line overlays added <https://trac.osgeo.org/grass/changeset/31992> |
| 03:56:27 | epifanio: | hi, tring to build an updated grass from svn, i've that it can't find the gdal libraries ( i'm on osx 10.5 ) , i tried to run the same configure in the old svn directory and it works as aspected |
| 03:57:25 | epifanio: | http://rafb.net/p/BoUhO565.html |
| 03:57:31 | sigq: | Title: Nopaste - No description ( at rafb.net ) |
| 04:12:29 | CIA-11: | neteler * r31993 /grass/branches/develbranch_6/ ( 27 files in 9 dirs ): glynn: Add 'const' to vector functions; Remove 'const' from G__ls( ) etc; Miscellaneous clean-up ( merge from trunk ) |
| 07:19:08 | Manel: | manel_ ? |
| 07:19:34 | Manel: | manel_ ? |
| 07:20:26 | Manel: | test |
| 11:51:56 | CIA-11: | glynn * r31994 /grass/trunk/ ( 26 files in 3 dirs ): Add "const" to OGSF, G3D functions where appropriate |
| 11:51:58 | sigq: | grassfeed: Changeset [31994]: Add "const" to OGSF, G3D functions where appropriate <https://trac.osgeo.org/grass/changeset/31994> |
| 12:56:40 | CIA-11: | glynn * r31995 /grass/trunk/ ( 8 files in 8 dirs ): |
| 12:56:40 | CIA-11: | Elminate FILE* <-> struct Map_info* hack from sites library |
| 12:56:40 | CIA-11: | Use "const" where appropriate in sites library |
| 12:56:40 | CIA-11: | Disable raster/simwe as it's too badly broken |
| 13:08:40 | : | <_wolf_> hey RMatev how's it going? |
| 13:08:51 | RMatev: | hi |
| 13:09:15 | RMatev: | I'm trying to remove those extra loops |
| 13:09:56 | : | <_wolf_> how is that going? You had some promising examples |
| 13:10:19 | RMatev: | I'm borrowing an idea from radim's old code |
| 13:11:01 | : | <_wolf_> cool |
| 13:11:13 | RMatev: | I've generated the looped parallel line. Since I already have function which give inner contours, I can use it to split the looped line into non-looped parts |
| 13:11:41 | : | <_wolf_> ah nice |
| 13:11:52 | RMatev: | then I can check for every loop whether it should stay or not |
| 13:12:44 | : | <_wolf_> neat |
| 13:13:35 | RMatev: | the problem is that the old check is slow |
| 13:13:47 | : | <_wolf_> I can assume so |
| 13:14:15 | : | <_wolf_> what is its average running time? n^2? |
| 13:14:29 | RMatev: | mm, I'm thinking.. |
| 13:15:25 | RMatev: | it's asymptotic time should be n |
| 13:15:35 | RMatev: | hm... it's in fact fast |
| 13:16:26 | RMatev: | and when we add the fact that n is not number of vertexes in input line but number of vertexes only of the particular inner loop |
| 13:17:01 | : | <_wolf_> and you have m loops fo O( m*n ) ? |
| 13:17:12 | RMatev: | yep |
| 13:17:27 | : | <_wolf_> that's not too bad in fact It's quite ok |
| 13:17:37 | RMatev: | on general data and reasonable buffer distance m should be very small |
| 13:17:48 | RMatev: | most likely 1 |
| 13:17:50 | : | <_wolf_> yeah |
| 13:18:02 | : | <_wolf_> and n will be ~4 |
| 13:18:44 | RMatev: | there's only one issue I'm trying to figure out |
| 13:38:33 | RMatev: | the old check is unreliable. |
| 13:38:48 | : | <_wolf_> hmmh :( in what way? |
| 13:39:17 | RMatev: | well, I'll send you a picture to demonstrate |
| 13:39:29 | : | <_wolf_> ok |
| 13:49:46 | RMatev: | you need to zoom cause i made it miniature |
| 13:49:55 | : | <_wolf_> ok : ) |
| 13:51:21 | : | <_wolf_> hmmh, the whole top triangle should be removed |
| 13:51:37 | RMatev: | black is input line, red is the parallel line, we obviosly need to remove the upper triangle. Old method does the following: get a centroid of the triangle; calculate whether it is in D distance from the input line |
| 13:52:01 | RMatev: | but what if the centroid function returns the blue dot ? |
| 13:53:34 | : | <_wolf_> hmm... it should return the middle of the triangle... |
| 13:53:45 | : | <_wolf_> what is the green arc? |
| 13:55:01 | RMatev: | points from the red triangle above the arc are >D distance from black line |
| 13:56:25 | RMatev: | What about this : remove every loop which is outside input line |
| 13:56:34 | RMatev: | then use the old method |
| 13:57:05 | RMatev: | partly outside of course |
| 13:57:12 | : | <_wolf_> I suppose that would work |
| 13:58:10 | RMatev: | I cant find counterexample. I'm starting to code |
| 13:58:12 | RMatev: | : ) |
| 13:58:26 | : | <_wolf_> remember to sleep ; ) |
| 13:58:34 | : | * _wolf_ goes to bed |
| 13:58:46 | RMatev: | mm, already slept 5 hours in the afternoon |
| 13:58:51 | : | <_wolf_> :D |
| 13:58:58 | RMatev: | I've shifted my cylce |
| 13:59:01 | RMatev: | cycle |
| 13:59:18 | RMatev: | good night to you |
| 14:00:05 | : | <_wolf_> I've thought about that too. Sleeb between 16:00 and 00:00 Then stay up all night ( it is not so hot ), and then got to work in the "evening" from 8-16 |
| 14:00:20 | : | <_wolf_> happy coding |
| 20:37:28 | CIA-11: | martinl * r31996 /grass/trunk/gui/wxpython/ ( 5 files in 2 dirs ): wxGUI/vdigit: visualize line directions implemented, trac #209 |
| 20:39:56 | CIA-11: | martinl * r31997 /grass/branches/develbranch_6/gui/wxpython/ ( 5 files in 2 dirs ): wxGUI/vdigit: visualize line directions implemented, trac #209 ( merged from trunk, r31996 ) |
| 21:41:51 | CIA-11: | martinl * r31998 /grass/trunk/gui/wxpython/gui_modules/mapdisp.py: nviz2: virtual OnZoomToRaster method added |
| 21:43:37 | sigq: | grassfeed: Changeset [31998]: nviz2: virtual OnZoomToRaster? method added <https://trac.osgeo.org/grass/changeset/31998> |