| 00:02.01 | brlcad | ah, so I think I see how it's tied in yet .. don't think anything else is using ged_reopen just yet so it very well could be incomplete |
| 00:03.16 | brlcad | mged does it's own thing (see f_opendb() in src/mged/mged.c) and archer goes through the old display object interface (src/libged/dg_obj.c) |
| 00:03.35 | brlcad | so might want to reconcile what f_opendb() is doing with what ged_reopen() is doing |
| 00:03.50 | brlcad | s/might want/you need/ :) |
| 00:04.38 | brlcad | talcite: pongish |
| 00:05.38 | brlcad | talcite: what rpath "problem"? rpaths are set by libtool, which are managed primarily by automake built-in macro expansions |
| 00:06.05 | brlcad | you shouldn't be mucking with the rpaths else you will likely encounter dragons |
| 00:07.50 | brlcad | ``Erik: you can register a bu_bomb handler to tell it to keep going (if you haven't figured that out by now) |
| 00:08.02 | ``Erik | hrm? whu? |
| 00:08.06 | brlcad | (related to earlier tessellation talkage) |
| 00:08.32 | ``Erik | <-- doesn't recall which tessellation talkage |
| 00:08.59 | brlcad | tk+libfb ftw |
| 00:09.06 | ``Erik | heh |
| 00:10.03 | ``Erik | tk is making me a sad panda |
| 00:10.27 | ``Erik | I can't rotate objects in mged anymore, it executes on mouse down instead of mouse up, so it reads it as a zoom in command |
| 00:16.13 | brlcad | wonders if jdoliner wins when he defends himself to his goldfish |
| 00:16.37 | ``Erik | obviously you haven't talked to joe much ;> *duck* |
| 00:17.06 | ``Erik | it is gettin' to be wrapup time, though |
| 00:17.40 | brlcad | there aren't any release stoppers that I know of, so we should be good to release .. it has just been coma/sinus recouping .. *ship it!* |
| 00:18.30 | brlcad | ughs at the automatic search paths being added for X content instead of specifying or detecting |
| 00:19.05 | brlcad | finishes spewing from recent backlog |
| 00:19.26 | brlcad | ``Erik: you're on 10.5 right? |
| 00:19.36 | ``Erik | yeah, but I'm seeing the behavior on both .5 and .4 |
| 00:19.53 | ``Erik | (if you're asking about the mouse rotate issue) |
| 00:20.04 | ``Erik | also; opendb /path/to/castle.g ; E all.g |
| 00:20.07 | ``Erik | :D |
| 00:20.10 | brlcad | nick found a way to hit keys before/during the mouse events that lets him rotate smooth |
| 00:20.13 | brlcad | tricking it |
| 00:20.30 | ``Erik | eh? |
| 00:20.35 | brlcad | or there's hitting the xyzXYZ keys |
| 00:21.20 | brlcad | and yeah, just a couple more days before pens down |
| 00:21.27 | ``Erik | I don't fire up mged often enough to worry, and almost never e anything up... was just something I kinda noticed, not sure if it's due to something funky on my stuff or an actual issue |
| 00:21.44 | *** join/#brlcad Patmcc19_ (n=chatzill@174-17-172-168.phnx.qwest.net) |
| 00:21.53 | brlcad | yeah, I had a regress test that took each of our example .g files and would tessellate each one .. it was too painful to commit |
| 00:22.06 | brlcad | the crash is new |
| 00:22.07 | ``Erik | hehehe |
| 00:22.37 | brlcad | i think that input problem is probably a release stopper for mac binaries |
| 00:22.51 | brlcad | i fixed all the other mac problems |
| 00:22.59 | brlcad | but hadn't got to that one before siggraph |
| 00:23.12 | brlcad | not enough to stop a source release though, nothing new |
| 00:24.16 | ``Erik | hm, the null hack for linux finally got on smacksnot |
| 00:29.10 | ``Erik | huh, a retrograde orbiting exoplanet (amazing how little we know about exoplanets... we only know the orbit direction for a dozen or so?) |
| 00:40.01 | CIA-38 | BRL-CAD: 03johnranderson * r35543 10/brlcad/trunk/src/libged/bigE.c: "E" command was always failing because it was not adding solids to the "ged_display_list". Now adds the solids. |
| 00:41.14 | ``Erik | hah |
| 00:41.18 | ``Erik | that'd be the castle crash, I bet |
| 00:46.48 | brlcad | go go gadget anderson |
| 01:07.10 | ``Erik | they're making a game... knocking off a movie... knocking off an 80's cartoon... |
| 01:07.20 | ``Erik | knocking off a 50's toy |
| 01:07.47 | ``Erik | 60's, sorry |
| 01:08.24 | ``Erik | (knocking off a 60's tv show I'd never heard of) |
| 01:17.50 | *** join/#brlcad BigAToo (n=BigAToo@pool-96-230-124-48.sbndin.btas.verizon.net) |
| 03:24.22 | ``Erik | ls |
| 03:24.25 | ``Erik | doh |
| 04:24.42 | *** join/#brlcad Axman6 (n=Axman6@pdpc/supporter/student/Axman6) |
| 06:31.09 | *** join/#brlcad d_rossberg (n=rossberg@bz.bzflag.bz) |
| 06:37.01 | *** join/#brlcad talcite_ (n=matthew@75-119-230-162.dsl.teksavvy.com) |
| 06:57.27 | *** join/#brlcad elena (n=elena@89.136.118.141) |
| 06:58.46 | elena | ~log |
| 07:12.54 | brlcad | ~logs |
| 07:12.55 | ibot | All conversations are logged to http://ibot.rikers.org/channel, where "channel" is replaced by the URL-encoded channel name, such as %23freenode for #freenode. Lines starting with spaces are not logged. |
| 07:52.31 | *** join/#brlcad talcite__ (n=matthew@76-10-171-69.dsl.teksavvy.com) |
| 08:44.01 | CIA-38 | BRL-CAD: 03ebautu * r35544 10/web/trunk/htdocs/more/sites/all/ (30 files in 3 dirs): Implemented links for sharing services. |
| 09:08.29 | *** join/#brlcad talcite_ (n=matthew@69-165-133-209.dsl.teksavvy.com) |
| 09:10.09 | CIA-38 | BRL-CAD: 03admin 07http://more.brlcad.org * r21 10Model repository/: Axis example (update model: BRLCAD processing completed.) |
| 09:10.21 | CIA-38 | BRL-CAD: 03admin 07http://more.brlcad.org * r22 10Model repository/: Boolean operations (update model: BRLCAD processing completed.) |
| 09:10.24 | CIA-38 | BRL-CAD: 03admin 07http://more.brlcad.org * r23 10Model repository/: bldg391 (update model: BRLCAD processing completed.) |
| 09:10.37 | CIA-38 | BRL-CAD: 03admin 07http://more.brlcad.org * r24 10Model repository/: Havoc (update model: BRLCAD processing completed.) |
| 09:10.44 | CIA-38 | BRL-CAD: 03admin 07http://more.brlcad.org * r25 10Model repository/: Tank car (update model: BRLCAD processing completed.) |
| 09:12.15 | CIA-38 | BRL-CAD: 03ebautu * r35545 10/web/trunk/htdocs/more/sites/all/themes/fireflystreamcom/node-model.tpl.php: Fix links themeing for anonymous users. |
| 09:13.24 | CIA-38 | BRL-CAD: 03admin 07http://more.brlcad.org * r25 10Model repository/: Tank car (update model: ) |
| 09:14.09 | CIA-38 | BRL-CAD: 03admin 07http://more.brlcad.org * r24 10Model repository/: Havoc (update model: ) |
| 09:14.18 | CIA-38 | BRL-CAD: 03admin 07http://more.brlcad.org * r23 10Model repository/: bldg391 (update model: ) |
| 09:14.29 | CIA-38 | BRL-CAD: 03admin 07http://more.brlcad.org * r22 10Model repository/: Boolean operations (update model: ) |
| 09:14.38 | CIA-38 | BRL-CAD: 03admin 07http://more.brlcad.org * r21 10Model repository/: Axis example (update model: ) |
| 09:15.18 | CIA-38 | BRL-CAD: 03admin 07http://more.brlcad.org * r25 10Model repository/: Tank car (update model: BRLCAD processing completed.) |
| 09:15.53 | CIA-38 | BRL-CAD: 03admin 07http://more.brlcad.org * r21 10Model repository/: Axis example (update model: BRLCAD processing completed.) |
| 09:16.02 | CIA-38 | BRL-CAD: 03admin 07http://more.brlcad.org * r22 10Model repository/: Boolean operations (update model: BRLCAD processing completed.) |
| 09:16.11 | CIA-38 | BRL-CAD: 03admin 07http://more.brlcad.org * r23 10Model repository/: bldg391 (update model: BRLCAD processing completed.) |
| 09:16.21 | CIA-38 | BRL-CAD: 03admin 07http://more.brlcad.org * r24 10Model repository/: Havoc (update model: BRLCAD processing completed.) |
| 09:47.56 | *** join/#brlcad ornitorrincos (n=ilcra198@archlinux/trusteduser/ornitorrincos) |
| 10:53.00 | *** join/#brlcad BigAToo (n=BigAToo@pool-96-230-124-48.sbndin.btas.verizon.net) |
| 12:27.35 | *** join/#brlcad BigAToo (n=BigAToo@pool-96-230-124-48.sbndin.btas.verizon.net) |
| 12:37.38 | brlcad | oh that's cool |
| 13:11.46 | *** join/#brlcad Elrohir (n=kvirc@p5B14FACE.dip.t-dialin.net) |
| 13:13.04 | ``Erik | ? |
| 14:23.59 | *** join/#brlcad _clock_ (n=_sushi_@80-218-244-105.dclient.hispeed.ch) |
| 15:04.26 | brlcad | the model notifications |
| 15:04.59 | brlcad | noficiations when it's added and when processing is completed .. which looks like was about 5 minues for those tiny models :) |
| 15:15.09 | CIA-38 | BRL-CAD: 03brlcad * r35546 10/brlcad/trunk/src/rt/viewedge.c: minor tweaks for syncing antialias work |
| 15:26.08 | CIA-38 | BRL-CAD: 03erikgreenwald * r35547 10/isst/trunk/src/ (gui.c isst.h load_g.c main.c): specify db/regions from command line. |
| 16:41.59 | brlcad | haha |
| 16:42.05 | *** join/#brlcad jdoliner (n=jdoliner@c-68-51-75-169.hsd1.il.comcast.net) |
| 16:42.11 | brlcad | howdy joe! |
| 16:42.16 | jdoliner | hi |
| 16:42.30 | jdoliner | I haven't seen you in a while sean |
| 16:42.35 | brlcad | was just going through some of your commits yesterday |
| 16:42.40 | brlcad | yeah, was away on travel |
| 16:42.47 | brlcad | been a busy couple weeks |
| 16:42.48 | jdoliner | yes, thoughts plz |
| 16:42.58 | brlcad | seen the chatter in here and commits, though |
| 16:43.25 | brlcad | well, how're things going? |
| 16:43.39 | brlcad | i noticed you were working on curve/face crossings for the face-face intersections |
| 16:44.02 | jdoliner | yes |
| 16:44.22 | jdoliner | you mean in the FindStartPoints_Internal? |
| 16:46.53 | jdoliner | sry that function is actualy GetStartPointsInternal |
| 16:56.59 | brlcad | mm, i'd have to dig back through the commits |
| 16:57.50 | brlcad | Curve_X_Profile() is what I had in mind |
| 16:57.56 | jdoliner | don't bother |
| 16:57.58 | brlcad | and FaceFaceIntersect() |
| 16:59.07 | jdoliner | yeah, Curve_X_Profile was actually an idea that I never wound up using |
| 16:59.11 | jdoliner | I'll remove that soon |
| 16:59.21 | jdoliner | but facface intersect is the workhorse |
| 17:00.50 | jdoliner | FaceFaceIntersect does all the work of creating the Face_X_Events |
| 17:01.07 | brlcad | cool |
| 17:01.15 | jdoliner | which are basically ON_X_Events but for two faces |
| 17:01.22 | brlcad | that does sound like the most important piece :) |
| 17:01.29 | jdoliner | yeah :) |
| 17:01.44 | jdoliner | it uses the numeric method of walking the intersection curve |
| 17:01.54 | brlcad | were talking yesterday about evaluating a given CSG operation as just the resulting surface intersections, finding those curves, attaching them as trims, and just feeding the result to opengl for visualization |
| 17:02.38 | brlcad | as a 'first step' of sorts, for visualization of arbitrary objects via shaded displays/opengl |
| 17:03.12 | jdoliner | I see, |
| 17:03.37 | jdoliner | ultimately would we like to just 'freeze' faces to have only 1 external trim? |
| 17:09.58 | brlcad | what do you mean? |
| 17:11.00 | brlcad | multiple trimming curves are normal |
| 17:11.07 | brlcad | or do you mean one trimming loop? |
| 17:11.09 | jdoliner | well at the end we're going to end up with faces with a whole bunch of trims |
| 17:11.22 | jdoliner | yeah i mean loops |
| 17:11.58 | brlcad | yeah, ideally minimal loops but even that shouldn't be a problem to have multiple ones |
| 17:12.22 | jdoliner | k |
| 17:12.49 | brlcad | will probably want some means to simplify, though -- if a loop is fully contained within another, eliminate it; if it intersects, weave it in, etc |
| 17:13.19 | jdoliner | yes, I actually already have it setup to weave intersecting loops in |
| 17:14.21 | jdoliner | also if it has 2 distinct external loops that should be an easy simplification |
| 17:14.22 | brlcad | awesome |
| 17:34.43 | CIA-38 | BRL-CAD: 03erikgreenwald * r35548 10/isst/trunk/src/ (Makefile.am gui.c isst.h load_g.c main.c): break worker functions out into their own files |
| 17:36.47 | CIA-38 | BRL-CAD: 03erikgreenwald * r35549 10/isst/trunk/src/test.c: rm dead file |
| 17:38.01 | CIA-38 | BRL-CAD: 03erikgreenwald * r35550 10/isst/trunk/src/ (local_worker.c net_worker.c): break worker functions out into their own files |
| 17:38.28 | CIA-38 | BRL-CAD: 03erikgreenwald * r35551 10/isst/trunk/src/poo.g: rm dead file |
| 17:39.36 | *** join/#brlcad talcite_ (n=matthew@69-165-133-209.dsl.teksavvy.com) |
| 17:45.19 | CIA-38 | BRL-CAD: 03erikgreenwald * r35552 10/isst/trunk/src/net_worker.c: bcopy->memcpy |
| 17:47.55 | talcite_ | hey brlcad, can I get your help in changing the configure script from using libtool runpaths to convenience libraries? |
| 17:54.21 | CIA-38 | BRL-CAD: 03erikgreenwald * r35553 10/isst/trunk/src/ (gui.c isst.h local_worker.c net_worker.c): remove gtk requirement from local_worker |
| 18:07.12 | brlcad | talcite_: not sure what you mean by that exactly |
| 18:07.27 | brlcad | talcite_: the convenience libraries aren't installed |
| 18:08.56 | talcite_ | brlcad: yeah. I'm just trying to remove all of the libtool runpaths and create partially linked executables so they're compliant with fedora guidelines |
| 18:09.17 | talcite_ | brlcad: I'm not sure what you mean by the convenience libraries aren't installed |
| 18:10.12 | brlcad | libtool "convenience libraries" are, by definition, not installed :) |
| 18:10.21 | brlcad | http://sources.redhat.com/autobook/autobook/autobook_92.html |
| 18:10.31 | brlcad | note the "noinst" |
| 18:12.52 | brlcad | can you give a reference to those fedora guidelines? 'partially linked executables' can mean a lot of things or nothing at all |
| 18:14.01 | brlcad | the default build is shared libraries with dynamic linkage |
| 18:15.09 | brlcad | all i've ever heard fedora guideline-wise was that they (rightfully) prefer non-static compilation/linkage so dependencies can be properly updated and managed |
| 18:17.11 | talcite_ | brlcad: http://fedoraproject.org/wiki/Packaging:Guidelines#Beware_of_Rpath |
| 18:17.16 | brlcad | yeah, just found that |
| 18:17.49 | talcite_ | sorry for the slow reply, I was just reading over the autotools book page you sent me. I haven't seen that one before |
| 18:17.55 | brlcad | that's not what I'd consider having much to do with "partially linked executables" ;) |
| 18:18.13 | talcite_ | ahh, sorry for the confusing terminology |
| 18:18.19 | brlcad | yeah, I think you're just calling them the wrong thing |
| 18:18.36 | brlcad | I think you just mean our installed libraries |
| 18:18.46 | brlcad | so did you try --disable-rpath? |
| 18:18.56 | talcite_ | yup, it didn't do anything |
| 18:19.19 | brlcad | did you already have a libtool script generated? |
| 18:19.30 | talcite_ | also, removing the hardcode_libdir_flag_spec didn't do anything |
| 18:19.50 | talcite_ | brlcad: libtool script generated? |
| 18:20.06 | brlcad | no used gnu autotools much I take it? :) |
| 18:20.32 | talcite_ | I'm just running everything from sh autogen.sh to make install in the spec file |
| 18:20.40 | talcite_ | brlcad: no, I haven't heh |
| 18:20.56 | brlcad | when you run our configure, it spits out a script called 'libtool' |
| 18:21.23 | brlcad | that script is wired into the Makefiles automatically by automake |
| 18:21.53 | brlcad | so during make, the build invokes that libtool script for all linkage |
| 18:22.01 | *** join/#brlcad docelic (n=docelic@78.134.200.109) |
| 18:22.20 | talcite_ | yeah, I've seen that script being used |
| 18:22.36 | talcite_ | I haven't done anything like generate my own though |
| 18:22.48 | brlcad | that script encapsulates all logic on how to build libraries and link binaries (static and dynamic), and is otherwise mostly a "black box" that you just run .. minus a few knobs you can manually tweak to override what the gnu folks think is best |
| 18:24.03 | brlcad | the "sh autogen.sh" step turns the configure.ac file into the venerable configure script and all the Makefile.am automake template files into Makefile.in autoconf template files |
| 18:24.20 | talcite_ | yup |
| 18:24.44 | brlcad | when you run "./configure --whatever..." that turns all the Makefile.in files into Makefile files and generates the libtool script if it's a libtool-enabled project (which we obviously are) |
| 18:25.06 | brlcad | okay, so back to the problem at hand.. post up your libtool script somewhere |
| 18:25.30 | brlcad | their guideline are either out of date with the latest libtool, or ya did something wrong |
| 18:26.22 | talcite_ | brlcad: http://fpaste.org/NZS3/ |
| 18:27.27 | talcite_ | brlcad: well going through the build logs shows the -rpath argument being passed to libtool several times |
| 18:29.12 | brlcad | well the good news is the libtool script doesn't force it |
| 18:31.16 | talcite_ | brlcad: would the configure script help? |
| 18:31.26 | brlcad | note yet |
| 18:32.59 | *** join/#brlcad SWPadnos (n=Me@dsl107.esjtvtli.sover.net) |
| 18:33.49 | brlcad | did you verify that it's actually generating installed binaries with an rpath set? |
| 18:34.18 | talcite_ | brlcad: yeah. Fedora has a check_rpaths tool built into rpmbuild |
| 18:34.51 | brlcad | I know, I'm saying you've checked them explicitly since having ran those regex's on the libtool script, not just noticed the --rpath cmdline option |
| 18:35.00 | brlcad | more specifically, noticed post-install |
| 18:35.10 | brlcad | libtool is a multiphase interface, it knows the difference between compilation, preinstall, and postinstall |
| 18:35.51 | brlcad | not saying that'll do it, but worth checking before going further down the rabbit hole |
| 18:35.52 | talcite_ | brlcad: yeah. check-rpaths runs after the rpm is generated, then it goes over the binaries |
| 18:43.31 | brlcad | talcite_: grep -A1 "checking how to hardcode library paths into programs" brlcad/config.log | grep result |
| 18:44.21 | talcite_ | configure:20803: result: immediate |
| 18:44.21 | talcite_ | configure:24617: result: immediate |
| 18:46.04 | brlcad | hm |
| 18:51.04 | brlcad | lets try some more aggressive libtool mods.. |
| 18:51.14 | brlcad | try this: |
| 18:51.46 | brlcad | sed -e s/^hardcode_direct.*$/hardcode_direct=yes/g libtool || sed -e s/^hardcode_minus_L.*$/hardcode_minus_L=yes/g || sed -e s/^hardcode_shlibpath_var.*$/hardcode_shlibpath_var=no/g > libtool.new |
| 18:52.04 | brlcad | mv libtool libtool.old && mv libtool.new libtool |
| 18:52.23 | brlcad | diff libtool libtool.old .. should see at least three lines changed |
| 18:53.54 | brlcad | can either add that sed and mv to configure.ac after BC_PATCH_LIBTOOL (after making sure it works manually first), or make it a step after ./configure |
| 18:54.36 | talcite_ | man that sed was huge |
| 18:56.11 | CIA-38 | BRL-CAD: 03starseeker * r35554 10/brlcad/trunk/src/proc-db/ (Makefile.am csgbrep.cpp): |
| 18:56.11 | CIA-38 | BRL-CAD: Add the skeleton for a csgbrep proc-db - eventually this will be used to do |
| 18:56.11 | CIA-38 | BRL-CAD: examples of all of the primitives going from an implicit to a NURBS |
| 18:56.11 | CIA-38 | BRL-CAD: representation. At the moment it looks like the ability to do this is not yet |
| 18:56.11 | CIA-38 | BRL-CAD: set up, so need to wire in the rt_*_brep ability to some sort of public API. |
| 18:56.25 | talcite_ | brlcad: your i/o redirection isn't working |
| 18:56.31 | brlcad | oh, might want to add one more, sed -e 's/(hardcode_into_libs)=.*$/\1=no/' |
| 18:56.58 | brlcad | that was three seds |
| 18:57.04 | brlcad | so four with the last one |
| 18:57.36 | talcite_ | brlcad: I'm not really sure what you did with the || , but the sed stuff just outputs to stdout |
| 18:57.39 | brlcad | and the ||'s should be just |'s .. typo |
| 18:57.43 | talcite_ | it isn't getting redirected to libtool.new |
| 18:57.47 | talcite_ | oh ok |
| 18:58.07 | brlcad | they were pre-escaped for m4 |
| 18:58.12 | talcite_ | oh... |
| 18:59.17 | brlcad | might just make that last/fourth one be like the rest: s/^hardcode_into_libs.*$/hardcode_into_libs=no/g |
| 19:01.04 | talcite_ | brlcad: lots of changes |
| 19:01.18 | talcite_ | I can try building now if you'd like |
| 19:01.26 | brlcad | no, paste the diff |
| 19:01.30 | talcite_ | k |
| 19:01.31 | brlcad | diff -u |
| 19:02.07 | talcite_ | http://fpaste.org/h4EL/ |
| 19:04.05 | brlcad | hm, almost |
| 19:04.44 | brlcad | make the first one, s/^hardcode_direct=.*$/hardcode_direct=yes/g |
| 19:04.50 | brlcad | repaste |
| 19:08.19 | talcite_ | http://fpaste.org/ENoS/ |
| 19:09.10 | starseeker | brlcad: I'm slightly twisted up - the rt_sph_brep and rt_ell_brep functions aren't accessible directly, and upon reflection should probably be called through rt_functab like all the other primitive functions. But what about situations like wdb where we have (say) the sphere parameters and want to generate the resulting nurbs without detouring through creating the implicit version of the primitive? (We could, of course, but that would mean a write AND |
| 19:13.04 | CIA-38 | BRL-CAD: 03jdoliner * r35555 10/brlcad/trunk/src/proc-db/surfaceintersect.cpp: BrepBrepIntersect now cycles through all the pairings of faces and gets their intersections accurately |
| 19:15.14 | brlcad | talcite_: that looks great, give that a test |
| 19:15.19 | brlcad | with a clean build |
| 19:15.31 | talcite_ | brlcad: sure |
| 19:15.43 | brlcad | it'll still pass --rpath on the command line, but it should think that it doesn't need to do anything |
| 19:16.12 | brlcad | if that doesn't work, we can probably just trick up the --rpath option in the script |
| 19:17.02 | brlcad | starseeker: they weren't added to the functab just because it's woefully incomplete -- but they are directly accessible |
| 19:17.39 | brlcad | and your message was too long, cut off after AND |
| 19:19.28 | talcite_ | brlcad: sounds good. I've started the build process. It'll be 17 minutes or something |
| 19:19.44 | brlcad | you can just call rt_*_brep() for testing purposes to make sure things are working .. once they're working or as they're working, can add them to functab |
| 19:19.49 | brlcad | talcite_: k |
| 19:21.21 | brlcad | starseeker: I also wouldn't want to double-up the API just for a representation type -- the idea would probably be to create an in-memory-only object and then make the functab call on it for a given representation type |
| 19:21.48 | brlcad | or they create a nurbs object via opennurbs and write out using mk_brep() |
| 19:23.56 | talcite_ | brlcad: looks like we're getting somewhere. It's calling ld now |
| 19:24.04 | talcite_ | brlcad: but we have a make error |
| 19:24.29 | talcite_ | brlcad: http://fpaste.org/3jXm/ |
| 19:25.20 | brlcad | heh, well it certainly seems to have worked :) |
| 19:26.52 | brlcad | cd /home/matthew/rpmbuild/BUILD/brlcad-SVN_010809/src/other/URToolkit/cnv/rletoabA62 |
| 19:27.05 | brlcad | /bin/sh ../../../../../libtool --tag=CC --mode=link gcc -I../../../../../src/other/libutahrle/include -O2 -g -pipe -fno-strict-aliasing -fno-common -fexceptions -m64 -g -O3 -L/usr/local/lib64 -L/usr/local/lib -pipe -fno-strict-aliasing -fno-common -fexceptions -m64 -g -O3 -o rletoabA62 rletoabA62-rle.o rletoabA62-rletoabA62.o ../../../../../src/other/libutahrle/libutahrle.la |
| 19:27.11 | brlcad | (run that) |
| 19:27.26 | brlcad | paste the really long gcc line that it spits back |
| 19:27.39 | talcite_ | libtool: link: LD_LIBRARY_PATH="../../../../../src/other/libutahrle/.libs:" gcc -I../../../../../src/other/libutahrle/include -O2 -g -pipe -fno-strict-aliasing -fno-common -fexceptions -m64 -g -O3 -pipe -fno-strict-aliasing -fno-common -fexceptions -m64 -g -O3 -o .libs/rletoabA62 rletoabA62-rle.o rletoabA62-rletoabA62.o -L/usr/local/lib64 -L/usr/local/lib -lutahrle -lm -Wl,-rpath -Wl,/usr/lib64/brlcad |
| 19:27.39 | talcite_ | /usr/bin/ld: cannot find -lutahrle |
| 19:27.39 | talcite_ | collect2: ld returned 1 exit status |
| 19:28.20 | brlcad | huh, odd |
| 19:28.33 | brlcad | ls -la ../../../../../src/other/libutahrle/.libs/lib* |
| 19:29.17 | talcite_ | http://fpaste.org/61p6/ |
| 19:30.11 | brlcad | well that's stumpworthy |
| 19:30.30 | brlcad | there's libutahrle.so right there, and LD_LIBRARY_PATH points there |
| 19:30.44 | brlcad | ah, but no -L for it, hrm |
| 19:32.26 | brlcad | think we need to remove one of the sed's |
| 19:32.46 | brlcad | the hardcode_minus_L one |
| 19:35.27 | starseeker | brlcad: what do I #include to get them in directly? |
| 19:36.21 | brlcad | starseeker: they're not in a public header yet, you just have to declare them |
| 19:36.28 | starseeker | ok |
| 19:36.36 | brlcad | see table.c |
| 19:36.43 | brlcad | there is a declaration macro there |
| 19:36.55 | starseeker | excellent, thanks |
| 19:37.22 | starseeker | should the rt_functab for tnurbs morph into the brep one? |
| 19:38.02 | ``Erik | that'd wig out import/export |
| 19:38.16 | brlcad | you don't need to use the macro, but shows the basic form (or just declare them simple) .. it's temporary either way |
| 19:38.25 | starseeker | nods |
| 19:39.19 | starseeker | is figuring the rt_functab stuff for brep should be handled Sometime Soon Now... |
| 19:39.42 | brlcad | until it's obsoleted, I wouldn't touch the functab entries -- the guts to those functions need to change |
| 19:40.38 | brlcad | sure, you can add them now if you like -- just have to be careful you don't blow a loop somewhere |
| 19:41.04 | brlcad | i've been working on hiding the functab |
| 19:41.12 | starseeker | oh, OK |
| 19:41.17 | starseeker | that's different |
| 19:41.22 | brlcad | right now it's the only way in the api to get at primitives and their callbacks |
| 19:41.31 | starseeker | nods |
| 19:41.53 | brlcad | without calling the primitive-specific function directly, of course |
| 19:42.23 | brlcad | i started with mirror, and it just turned out to be a lot to chew on, still at it |
| 19:42.47 | starseeker | am I right that we're basically looking at needing most of the facetize type logic for "nurbize" as well? |
| 19:43.27 | brlcad | idea will be to have a corresponding rt_*() api call for each of the functab entries -- the rt_*() calls into the functab |
| 19:43.53 | brlcad | que? |
| 19:43.53 | brlcad | not necessarily |
| 19:43.53 | starseeker | given CSG primitives, we can currently run facetize, big E, etc. to get mesh |
| 19:44.04 | starseeker | don't we want the same to "get" nurbs? |
| 19:44.42 | *** join/#brlcad bobbens (i=bobbens@saw4ever.de) |
| 19:45.34 | brlcad | ah, at the command level .. yeah, probably |
| 19:45.47 | brlcad | probably an option for some commands, default for others |
| 19:45.55 | starseeker | nods |
| 19:46.18 | brlcad | bigE/ev's purpose is visualization, so they could just be updated to be nurbs-only if ogl is available |
| 19:46.42 | brlcad | evaluated visualization |
| 19:46.59 | starseeker | maybe - might want to have a bot fallback if someone's ogl isn't up to NURBS though |
| 19:47.09 | brlcad | nurbize is kinda funky, don't see a direct need like there was for facetize outside of debugging yet |
| 19:47.35 | starseeker | for that matter, do we really need facetize? |
| 19:47.47 | brlcad | E/ev are rarely used these days on real geometry because of the robustness problems |
| 19:48.27 | starseeker | In the new GUI I'm assuming it will be hidden behind view modes rather than specific command line things like that? |
| 19:48.47 | brlcad | it being? |
| 19:49.23 | talcite_ | brlcad: build was successful, make install failed |
| 19:49.27 | starseeker | the functionality of switching between wireframe, shaded, etc |
| 19:49.44 | brlcad | yeah, those are just viewing options in the gui |
| 19:50.14 | talcite_ | http://fpaste.org/ttBy/ |
| 19:50.22 | brlcad | lots of potential viewing options that can be exposed through a panel or buttons or keys or what-have-you |
| 19:52.42 | starseeker | brlcad: what's the best way to create a bn_tol to toss into rt_*_brep? |
| 19:53.00 | brlcad | cd /home/matthew/rpmbuild/BUILD/brlcad-SVN_010809/src/libbn && /bin/sh ../../libtool --mode=install /usr/bin/install -c libbn.la '/home/matthew/rpmbuild/BUILDROOT/brlcad-SVN_010809-0.fc11.x86_64/usr/lib64/brlcad' |
| 19:53.43 | talcite_ | brlcad: http://fpaste.org/1RJZ/ |
| 19:54.05 | brlcad | starseeker: er, create a tol struct and pass a ref to it? :) |
| 19:54.14 | brlcad | init it with some values.. |
| 19:54.55 | starseeker | ok, so it won't much care whether it's the same as the database being written to? |
| 19:55.55 | brlcad | talcite_: hrm, now we're fighting libtool .. |
| 19:56.19 | brlcad | starseeker: databases don't have a tolerance |
| 19:56.52 | starseeker | oh, OK |
| 19:56.55 | brlcad | they're working tolerances |
| 19:57.18 | brlcad | you're telling it what computation tolerances it needs to use .. which is kinda odd for *brep() |
| 19:57.46 | brlcad | probably just from being started as a copy of tess() or the old tnurb() interface |
| 19:57.53 | starseeker | oh |
| 19:58.01 | starseeker | shall I clean it up? |
| 19:58.05 | brlcad | not harmful, though .. maybe important for some primitives where it will be some sort of approximation |
| 19:58.49 | brlcad | could see tol being important for dsp, ebm |
| 20:01.14 | starseeker | hmm, point. ok |
| 20:04.25 | talcite_ | brlcad: I'm going to switch locations. I'll be back in 20 minutes ok? |
| 20:04.29 | brlcad | aha, talcite .. one more sed |
| 20:04.33 | talcite_ | sure |
| 20:05.29 | brlcad | s/^hardcode_automatic=.*$/hardcode_automatic=yes/g |
| 20:05.52 | brlcad | that should prevent that relink rule from kicking off |
| 20:07.45 | brlcad | i'll be somewhat impressed if this actually works :) |
| 20:07.45 | talcite_ | alright, rebuild started. I'm switching locations now |
| 20:07.54 | brlcad | kcya |
| 20:08.53 | *** join/#brlcad SWPadnos (n=Me@dsl107.esjtvtli.sover.net) |
| 20:12.53 | starseeker | growls as he sees the brep.cpp files are all extradisted.... |
| 20:14.25 | starseeker | brlcad: do I link in just sph_brep.o or whatever to avoid depending on librt in a procdb? |
| 20:14.53 | *** join/#brlcad SWPadnos (n=Me@dsl107.esjtvtli.sover.net) |
| 20:17.38 | starseeker | grrrrowlll. Oh, lovely - sph_brep.cpp doesn't build |
| 20:19.41 | CIA-38 | BRL-CAD: 03n_reed * r35556 10/brlcad/trunk/ (include/dm-rtgl.h src/libdm/dm-rtgl.c): grouping jobs for decent job shuffling |
| 20:27.04 | *** join/#brlcad SWPadnos (n=Me@dsl107.esjtvtli.sover.net) |
| 20:29.59 | *** join/#brlcad SWPadnos_ (n=Me@dsl107.esjtvtli.sover.net) |
| 20:37.06 | talcite | brlcad, argh. It's still failing the check-rpaths test |
| 20:37.39 | talcite | brlcad, this tool has been known to give false positives in some cases though. Do you know of another method to check for rpaths? |
| 20:47.22 | brlcad | talcite: well if you don't set rpaths and haven't added the path to /etc, the binaries shouldn't work |
| 20:47.35 | brlcad | ldd /usr/bin/rt |
| 20:51.20 | talcite | brlcad, http://fpaste.org/Chsp/ |
| 20:56.30 | brlcad | that looks like a lack of rpaths |
| 20:56.56 | brlcad | assuming that is your installation prefix |
| 20:57.14 | starseeker | thanks ``Erik for noticing that rt_sph_brep is mangled and decides linking can wait 'til Monday... |
| 20:57.16 | brlcad | if it's chrooted, you'll have to try it again in the root |
| 20:57.58 | talcite | hmm it's just an installation prefix |
| 20:58.06 | talcite | well maybe seeing the check-rpaths output would help |
| 20:59.29 | brlcad | sure |
| 20:59.40 | talcite | http://fpaste.org/MoSZ/ |
| 21:03.24 | CIA-38 | BRL-CAD: 03brlcad * r35557 10/brlcad/trunk/src/librt/primitives/sph/sph_brep.cpp: unpooch it |
| 21:04.58 | brlcad | talcite: ah, so they do have an rpath, they have a preinstallation rpath so you can run them without installing them |
| 21:05.22 | brlcad | which is exactly why it tried to relink them earlier |
| 21:05.29 | brlcad | to get rid of that path |
| 21:05.37 | brlcad | starseeker: compiles now |
| 21:06.28 | talcite | brlcad, I see. So what are our options? |
| 21:08.13 | brlcad | talcite: i'm really liking the bail-out option :) |
| 21:08.25 | CIA-38 | BRL-CAD: 03brlcad * r35558 10/brlcad/trunk/ (3 files in 2 dirs): enable sph_brep.cpp compilation |
| 21:08.28 | talcite | brlcad, bail out? =S |
| 21:08.54 | brlcad | BuildRequires: chrpath |
| 21:09.36 | talcite | haha yeah, we could do that. I'm not sure what it actually does though. If you delete the files with rpaths in it, doesn't it break other stuff? |
| 21:09.43 | brlcad | my feeling is even if we get something hacking.. it's going to be fragile to libtool updates |
| 21:10.14 | brlcad | talcite: that tool simply strips out the rpath |
| 21:10.21 | brlcad | doesn't delete files |
| 21:10.24 | talcite | oh I see |
| 21:10.54 | talcite | sure, I wouldn't mind doing that. the libtool stuff is really finicky |
| 21:11.13 | brlcad | so you'd let it build and install, then set up a massive chrpath --delete on all 400+ binaries |
| 21:11.42 | talcite | brlcad, and after the rpaths are deleted, it would turn to ld to find the libs? |
| 21:12.20 | brlcad | fwiw, the claim that "the Linux dynamic linker is usually smarter than a hardcoded path" is a boatload of crap :) |
| 21:12.36 | brlcad | they're both flimsy |
| 21:12.56 | talcite | heh. Well I think there was one specific reason that distros avoided rpaths. I was reading about it on the debian mailing lists |
| 21:13.03 | brlcad | to insist on one over the other is a bit silly, but hey their system their rules |
| 21:13.18 | talcite | it was massively breaking some upgrades I think |
| 21:13.34 | brlcad | that usually indicates some other stupidity on someone's part |
| 21:13.51 | brlcad | debian and fedora are similarly managed *ahem* |
| 21:13.55 | talcite | anyways, debian also bans rpaths, so if the chrpaths method works, we can probably push to package this into debian as well =D |
| 21:13.57 | talcite | heh |
| 21:15.27 | brlcad | ports, fink, portage, and others all get along just fine not getting involved in whether they're set or not |
| 21:16.42 | brlcad | I suspect it just makes package management easier for the package management *system* developers, pushing the work onto the porters to customize most packages |
| 21:18.14 | talcite | oh... =/ |
| 21:18.21 | brlcad | the debian devs also insist of screwing with libtool directly to impose one of their requirements, which actually outright breaks packages that include public libraries with binaries |
| 21:18.49 | brlcad | that's the BC_PATCH_LIBTOOL I mentioned earlier .. reverts damage they make to the libtool script that causes failures |
| 21:21.52 | talcite | oh...yeah I think fedora also has patches applied to libtool |
| 21:31.26 | CIA-38 | BRL-CAD: 03erikgreenwald * r35559 10/isst/trunk/src/gui.c: thread monkeying. pass args to gtk_init() |
| 21:31.55 | CIA-38 | BRL-CAD: 03erikgreenwald * r35560 10/isst/trunk/src/ (isst.h load_g.c): disable "fast" loading for now. |
| 21:38.39 | talcite | alright, hopefully that build will work out. I need to head out for a bit. I will be back later |
| 21:49.03 | CIA-38 | BRL-CAD: 03erikgreenwald * r35561 10/isst/trunk/src/gui.c: get shotline working again. |
| 22:49.43 | ``Erik | hm |
| 22:50.25 | ``Erik | damn linux weenies, screwing up build systems and thinking it's a good thing |
| 22:50.28 | ``Erik | :D |
| 22:59.25 | CIA-38 | BRL-CAD: 03erikgreenwald * r35562 10/isst/trunk/src/ (local_worker.c net_worker.c): updates for osX.5 |
| 23:04.19 | brlcad | talcite_: thanks again for your efforts |
| 23:07.30 | ``Erik | thinks he'll put in a loader dialog then try to cook a winderz binary of isst O.o |
| 23:07.37 | ``Erik | and chuck it over the fence |
| 23:08.52 | ``Erik | (btw, facetize_all_regions or facetall.sh ... needs to be unsucked.) |