IRC log for #brlcad on 20140702

02:07.05Notify03BRL-CAD:zhaoanqing * 61514 (brlcad/branches/nmgreorg/AUTHORS brlcad/branches/nmgreorg/CHANGES and 528 others): Merge the changed of trunk to this branch. This status is not able to compile, I will fix them then.
02:21.45starseekerponders whether making tkhtml into a toolkit-agnostic render-to-image html library is in order - rather surprisingly, it seems to be one of the very best "small" HTML rendering libraries out there, as well as perhaps the most liberally licensed of them all
06:18.24*** join/#brlcad albertcoder (~albertcod@101.208.178.149)
06:53.31*** join/#brlcad d_rossberg (~rossberg@66-118-151-70.static.sagonet.net)
07:04.03*** join/#brlcad albertcoder (~albertcod@202.164.53.117)
07:09.15*** join/#brlcad albertcoder (~albertcod@202.164.53.117)
08:57.47*** join/#brlcad teepee- (bc5c2133@gateway/web/freenode/ip.188.92.33.51)
09:03.18Notify03BRL-CAD:d_rossberg * 61515 rt^3/trunk/src/coreInterface/Database.cpp: bugfix: copy all attributes, not only the first one
10:01.00*** join/#brlcad teepee- (bc5c2133@gateway/web/freenode/ip.188.92.33.51)
10:33.18*** join/#brlcad teepee- (bc5c2133@gateway/web/freenode/ip.188.92.33.51)
10:47.30*** join/#brlcad andrei_ (~IceChat77@5-12-112-86.residential.rdsnet.ro)
11:35.01*** join/#brlcad Ch3ck (~Ch3ck@66-118-151-70.static.sagonet.net)
11:56.27*** join/#brlcad teepee- (bc5c2133@gateway/web/freenode/ip.188.92.33.51)
11:57.48Notify03BRL-CAD:bob1961 * 61516 (brlcad/trunk/include/ged.h brlcad/trunk/src/libged/CMakeLists.txt and 2 others): Added an fbclear command to libged/libtclcad/Ged.tcl.
12:03.24Notify03BRL-CAD:d_rossberg * 61517 NIL: tag the C++ core interface with the corresponding BRL-CAD version (i.e. 7.24.2)
12:13.03*** join/#brlcad Zhao_Anqing (clouddrift@222.205.2.250)
12:39.07*** join/#brlcad Izakee (~Isaac@195.24.220.134)
12:41.55*** join/#brlcad teepee- (bc5c2133@gateway/web/freenode/ip.188.92.33.51)
12:43.13IzakeeDue to some pesky Internet Connection, my svn update locked svn. When I try to clean up, I get "svn: 'misc/tools/xmltools/libxml' is not a working copy directory". What should I do  brlcad ?
12:57.14mihaineacsuIzakee: have you tried so? http://bit.ly/1onPmPu
12:57.34*** join/#brlcad teepee- (bc5c2133@gateway/web/freenode/ip.188.92.33.51)
12:58.03Izakeeopening stackoverflow
13:10.10IzakeeThanks mihaineacsu
13:23.38Notify03BRL-CAD:starseeker * 61518 (brlcad/branches/openscenegraph/doc/docbook/system/man1/en/pix-fb.xml brlcad/branches/openscenegraph/include/ged.h and 4 others): Sync through trunk r61517
13:31.10Notify03BRL-CAD:starseeker * 61519 brlcad/trunk/src/libged/fbclear.c: Provide clearColor with memory
13:31.58Notify03BRL-CAD:starseeker * 61520 (brlcad/branches/openscenegraph/src/libged/fbclear.c Property Changed: and 3 others): Sync through trunk r61519
13:33.46Notify03BRL-CAD:starseeker * 61521 (brlcad/branches/gecode/CMakeLists.txt brlcad/branches/gecode/INSTALL and 41 others): Sync through trunk r61517
13:36.06Notify03BRL-CAD:starseeker * 61522 (brlcad/branches/gecode/src/libged/fbclear.c Property Changed: and 3 others): Sync through trunk r61519
13:39.44Notify03BRL-CAD:starseeker * 61523 (brlcad/branches/bullet/CMakeLists.txt brlcad/branches/bullet/INSTALL and 41 others): Sync through trunk r61517
13:45.07Notify03BRL-CAD:starseeker * 61524 (brlcad/branches/bullet/src/libged/fbclear.c Property Changed: and 3 others): Sync through trunk r61519
13:54.42Notify03BRL-CAD:d_rossberg * 61525 (rt^3/tags/rel-7-24-2/src/coreInterface/Arb8.cpp rt^3/tags/rel-7-24-2/src/coreInterface/Combination.cpp and 16 others): back port: bu.h wasn't broken out in this release, reverted the changes from revision 60116
14:11.27*** join/#brlcad clock_ (~clock@77-58-143-135.dclient.hispeed.ch)
14:15.03Notify03BRL-CAD:zhaoanqing * 61526 (brlcad/branches/nmgreorg/include/raytrace.h brlcad/branches/nmgreorg/src/librt/primitives/nmg/nmg_copy.c): fix conflicts after merge operation, now librt can be compiled.
14:26.47*** join/#brlcad teepee (~teepee@gateway/tor-sasl/teepee)
14:32.17*** join/#brlcad teepee- (bc5c2133@gateway/web/freenode/ip.188.92.33.51)
14:45.47Notify03BRL-CAD:starseeker * 61527 brlcad/trunk/CMakeLists.txt: It matters in the lrint test on some systems whether or not we are using the C99 flag - the test will succeed with gnu89 and fail with gnu99.
15:04.05Notify03BRL-CAD:starseeker * 61528 brlcad/trunk/CMakeLists.txt: Changes weren't right to begin with, but further experimentation shows the situation to be even worse - when this failure occurs, we can't successfully defined lrint until *after* math.h is included. Currently, lrint is defined in common.h, which always comes before math.h
15:04.23Notify03BRL-CAD:ejno * 61529 (brlcad/trunk/src/conv/3dm/conv3dm-g.cpp brlcad/trunk/src/conv/3dm/conv3dm-g.hpp): fix performance when converting large models
15:06.56Notify03BRL-CAD:ejno * 61530 brlcad/trunk/src/conv/3dm/conv3dm-g.cpp: make unique_name() static
15:45.23Notify03BRL-CAD:carlmoore * 61531 (brlcad/trunk/misc/tools/lemon/README brlcad/trunk/misc/tools/lemon/lemon.c and 3 others): fix spellings, and remove trailing blanks/tabs
15:52.11Notify03BRL-CAD:starseeker * 61532 (brlcad/trunk/regress/CMakeLists.txt brlcad/trunk/regress/fastgen.sh brlcad/trunk/regress/gqa.sh): Gtools no longer use underscores.
15:54.17Notify03BRL-CAD:carlmoore * 61533 (brlcad/trunk/misc/tools/re2c/htdocs/manual.html.in brlcad/trunk/misc/tools/re2c/re2c.1.in): fix 2 more spellings
16:01.41Notify03BRL-CAD:carlmoore * 61534 (brlcad/trunk/misc/tools/re2c/htdocs/manual.html.in brlcad/trunk/misc/tools/re2c/re2c.1.in): and yet another variable-name fix
16:11.06Notify03BRL-CAD:starseeker * 61535 (brlcad/trunk/CMakeLists.txt brlcad/trunk/include/common.h and 9 others): Ugh. Do what we seem to have to do to get balky lrint going on problematic Linux systems.
16:15.39Notify03BRL-CAD:starseeker * 61536 brlcad/trunk/include/common.h: minor formatting fix.
16:20.02*** join/#brlcad harmanpreet (~harmanpre@124.253.51.93)
16:20.30Notify03BRL-CAD:carlmoore * 61537 (brlcad/trunk/misc/tools/perplex/scanner.re brlcad/trunk/misc/tools/re2c/lessons/002_strip_comments/strip_002.s.re and 2 others): fix spellings and remove trailing blanks/tabs
16:22.21Notify03BRL-CAD:ejno * 61538 brlcad/trunk/src/conv/3dm/conv3dm-g.cpp: create a root layer
16:28.10Notify03BRL-CAD:zhaoanqing * 61539 (brlcad/branches/nmgreorg/src/libged/bev.c brlcad/branches/nmgreorg/src/libged/decompose.c and 7 others): fix conflicts after merge, now libged can be compiled.
16:48.33Notify03BRL-CAD:starseeker * 61540 (brlcad/trunk/CMakeLists.txt brlcad/trunk/include/common.h and 9 others): Back up, take another run at 'proper' lrint handling
16:49.47Notify03BRL-CAD:carlmoore * 61541 brlcad/trunk/misc/tools/xmltools/libxml/src/xmlregexp.c: remove trailing blanks/tabs, and fix spellings
17:00.56Notify03BRL-CAD:carlmoore * 61542 brlcad/trunk/misc/tools/xmltools/libxslt/src/xslt.c: fix spellings; remove trailing blanks/tabs
17:13.38Notify03BRL-CAD:starseeker * 61543 brlcad/trunk/CMakeLists.txt: lrint needed the math library for the function test, not just the subsequent macro foo. Looks like we need to expand BRLCAD_FUNCTION_EXISTS's options
17:17.12Notify03BRL-CAD:n_reed * 61544 brlcad/branches/brep-debug/src/libbrep/intersect.cpp: InPlane is not a predicate function, replace it with the same test we use elsewhere. Also, fix a case where single-point line-plane intersections were erroneously reported as overlap intersections. Alas, these changes don't fix any test conversions.
17:18.12Notify03BRL-CAD:carlmoore * 61545 brlcad/trunk/misc/tools/xmltools/libxml/src/xmlreader.c: fix spellings; remove trailing blanks/tabs
17:30.29Notify03BRL-CAD:carlmoore * 61546 (brlcad/trunk/misc/tools/re2c/lessons/001_upn_calculator/calc_002.re brlcad/trunk/misc/tools/xmltools/libxml/src/uri.c and 2 others): remove trailing blanks/tabs; fix spellings
17:41.26*** join/#brlcad clock (~clock@77-58-143-135.dclient.hispeed.ch)
17:45.06Notify03BRL-CAD:carlmoore * 61547 (brlcad/trunk/misc/tools/xmltools/libxml/src/debugXML.c brlcad/trunk/misc/tools/xmltools/libxml/src/xmlreader.c): remove trailing blanks/tabs; fix spellings
17:57.05*** join/#brlcad albertcoder (~albertcod@202.164.53.117)
18:12.41Notify03BRL-CAD:brlcad * 61548 brlcad/trunk/BUGS: rtweight/rtarea/gqa issues computing volumes
18:15.01*** join/#brlcad clock (~clock@77-58-143-135.dclient.hispeed.ch)
18:17.59Notify03BRL-CAD:ejno * 61549 brlcad/trunk/src/conv/3dm/conv3dm-g.cpp: fix windows build issue
18:28.13*** join/#brlcad mihaineacsu (~mihaineac@92.81.55.66)
18:29.33*** join/#brlcad piyushparkash (~piyushpar@117.205.68.122)
18:33.48*** join/#brlcad kintel (~kintel@unaffiliated/kintel)
18:46.55*** join/#brlcad albertcoder (~albertcod@101.208.170.244)
19:03.21*** join/#brlcad kintel (~kintel@unaffiliated/kintel)
19:28.58Notify03BRL-CAD:brlcad * 61550 brlcad/trunk/src/libged/fbclear.c: we shouldn't be allocating dynamic memory for data whose initial size is known and doesn't change. simplify the logic.
19:30.01Notify03BRL-CAD:starseeker * 61551 (brlcad/trunk/CMakeLists.txt brlcad/trunk/misc/CMake/BRLCAD_CheckFunctions.cmake): Perform what should be a fairly major upgrade to the capabilities of BRLCAD_FUNCTION_EXISTS - can now specify required libraries, include dirs, definitions and compile flags, as well as provide one or more snippits of source code which must be compiled successfully for the macro to report a success.
19:35.17Notify03BRL-CAD Wiki:Albertcoder * 7450 /wiki/User:Albertcoder/GSoC2014/logs: /* Week 7 */
19:51.04Notify03BRL-CAD:starseeker * 61552 (brlcad/trunk/CMakeLists.txt brlcad/trunk/misc/CMake/BRLCAD_CheckFunctions.cmake): Confound it, that somehow brok the lrint testing again.
20:01.58Notify03BRL-CAD:starseeker * 61553 (brlcad/trunk/CMakeLists.txt brlcad/trunk/misc/CMake/BRLCAD_CheckFunctions.cmake): I think that's got it... toplevel compilation tests use CMAKE_C_FLAGS, so do the same thing in the macro.
20:04.25brlcadstarseeker: you could try to transcode tkhtml into a small C lib
20:08.09*** join/#brlcad pandrei (~pandrei@5-12-132-241.residential.rdsnet.ro)
20:12.36mihaineacsubrlcad: did you get a chance to look over the last patch I submitted?
20:13.55brlcadmihaineacsu: only briefly, let me take a look again now
20:19.22brlcadmihaineacsu: yeah... now I remember!  okay, so it looks good but there are a few issues
20:19.58mihaineacsuokay
20:20.20brlcadthe biggest is probably that you hooked into both tools at the "read file from disk" step of their processing
20:20.40brlcadinstead of AFTER the density file is in memory
20:21.41Notify03BRL-CAD:starseeker * 61554 (brlcad/trunk/CMakeLists.txt brlcad/trunk/misc/CMake/BRLCAD_CheckFunctions.cmake): Grr - *still* not working. Now it's ignoring gnu99 during release builds.
20:23.24brlcadto do that, you also created a temp file, using an interface that is not portable - tmpfile()
20:23.36mihaineacsuyeah, but wasn't I supposed to get the density data at this step?
20:23.57brlcadwell it's not that, just that it's better to not write to a file if we don't need to
20:24.29brlcadand we don't ... CURLOPT_WRITEFUNCTION will let you read it into memory directly
20:24.51brlcadthat avoids the portability issue (and the need to close and clean up that temp file)
20:24.58mihaineacsuright
20:25.14brlcadanother issue is that you basically added the exact same code to both
20:25.34brlcadwhich I would expect, but notice now in wanting to change it to use WRITEFUNCTION, you'll have to change it in two places
20:25.37mihaineacsuit will be different once I load the data
20:25.40mihaineacsudirectly in mem
20:26.04brlcadevery bug, every change .. we'll have to remember to edit both or a bug/change will go unnoticed, so we don't want the duplication
20:26.50brlcadthat means creating a general function that both tools can call
20:27.04pandreibrlcad: can we talk a bit in private ?
20:28.59brlcadmihainescsu: since one tool is in src/libged and another in src/rt, that implies creating a new interface in one of the other base libraries (like libbu, perhaps a bu_url_download() function)
20:29.16brlcadpandrei: yes, but in a few minutes
20:29.28pandreibrlcad: ok, let me know. Thanks!
20:30.17mihaineacsuokay, makes sense
20:32.40mihaineacsuthanks for the feedback
20:33.54Notify03BRL-CAD:n_reed * 61555 (brlcad/branches/brep-debug/src/libbrep/boolean.cpp brlcad/branches/brep-debug/src/libbrep/debug_plot.cpp and 4 others): add a dplot subcommand to show the face curves that result from clipping the original event curves
20:42.15brlcadmihaineacsu: yeah, think about that new function a little bit, try to make it as simple as possible
20:42.23brlcadeven the name I suggested could use some more thought
20:42.59brlcadmaybe bu_url_get() .. implying there might later be a bu_url_put() or bu_url_post() too
20:43.19brlcadthe point would be to completely encapsulate cURL under this API
20:43.27mihaineacsuyeah
20:43.30brlcadwhich was the last issue in the patch ..
20:44.00brlcadwe can't just start using curl/curl.h and friends or our build will be basically broken on most platforms
20:45.03brlcadthere needs to be at least a header check in the top-level CMakeLists.txt file, then you can conditionalize all that code on HAVE_CURL_CURL_H -- maybe print a message in the tools if a URL is provided to let them know when that functionality is not compiled in
20:45.04mihaineacsuhow can we go about this?
20:45.44mihaineacsucool
20:50.30brlcadmihaineacsu: how expensive is curl_global_init()?
20:51.40mihaineacsuI don't know
20:51.44brlcadthis main() function is almost the bu function you need: http://curl.haxx.se/libcurl/c/getinmemory.html
20:52.11mihaineacsuwhy do you ask?
20:52.44brlcadwell, if it's really expensive, then your implementation will need to return a url context of some sort so that an application only runs init once
20:53.04brlcadif there's basically no cost, then you can init, download, and cleanup repeatedly as needed
20:53.57mihaineacsuthey suggest using just one init if multiple requests are made
20:54.27brlcadmihaineacsu: note that implementing this in libbu means using libbu for memory management too (not directly calling malloc()/realloc()/free(), etc)
20:54.35mihaineacsuyeah
20:54.36brlcadsure, but why :)
20:54.50brlcadif it's fast enough, I'm not sure it's worth it
20:55.14brlcadbasically, have to find out what the cost/downsides are
20:56.39mihaineacsuscratch that, it was about the handle. "Mind you, it is even preferred that you re-use an existing handle if you intend to make another transfer. libcurl will then attempt to re-use the previous connection."
20:59.10brlcadahh
20:59.41brlcadI'm not sure we care about that right now, but that would beg a slightly different interface
21:00.41brlcadlike a bu_url_init() + bu_url_get() + bu_url_destroy()
21:01.07mihaineacsuhmm
21:02.14mihaineacsuyeah, I'm not sure if it's worth it, performance wise
21:02.58mihaineacsuhttp://stackoverflow.com/questions/3787002/reusing-the-same-curl-handle-big-performance-increase
21:07.59brlcadgood thread, thanks
21:08.21brlcadmihaineacsu: yeah, so lets keep it simple for now .. just do it all behind one func
21:08.30mihaineacsuok, great
21:09.42brlcadwe can add a context api later if needed
21:10.37Notify03BRL-CAD:ejno * 61556 brlcad/trunk/src/conv/3dm/conv3dm-g.cpp: generate color/shader combs for stand-alone geometry at high levels of the hierarchy
21:18.02Notify03BRL-CAD:n_reed * 61557 brlcad/branches/brep-debug/src/libbrep/debug_plot.cpp: whoops, output order needs to match what the parser's expecting
21:25.03Notify03BRL-CAD:n_reed * 61558 brlcad/branches/brep-debug/src/libged/brep.c: tweak help message
21:54.17pandreibrlcad: ping
21:55.38Notify03BRL-CAD:starseeker * 61559 (brlcad/trunk/CMakeLists.txt brlcad/trunk/misc/CMake/BRLCAD_CheckFunctions.cmake): Needed the build type c flags, not just the standard set.
21:57.50starseekerhah.  finally.  time for lunch
21:59.32Notify03BRL-CAD:carlmoore * 61560 brlcad/trunk/doc/docbook/system/man1/en/pix-fb.xml: pix-fb manpage didn't have -p writeup; it does now
22:01.18clockno to je fakt
22:01.23clockja tam hlavne nevidim tu atraktivitu
22:26.36kanzureis there a pile of opennurbs examples anywhere?
22:27.37Notify03BRL-CAD Wiki:Popescu.andrei1991 * 7451 /wiki/User:Popescu.andrei1991/devlogs2014: /* Week 7 */
22:36.50brlcad~x from du ja tam hlavne nevidim tu atraktivitu
22:37.02brlcad~x from du no to je fakt
22:37.31brlcadkanzure: we have a few (all very similar) in src/proc-db
22:37.40brlcadwhere we construct objects manually
22:37.51brlcadthey'd be the c++ files
22:48.37kanzurehuh they do look similar yes

Generated by irclog2html.pl Modified by Tim Riker to work with infobot.