| 03:41.47 | Notify | 03BRL-CAD:brlcad * 55051 brlcad/trunk/src/libbu/heap.c: duh, race condition on initializing the per-cpu structure dynamically. statically allocate for the compile-time maximum number of cpus. wastefully consumes 32MB (32KB per cpu with 1024 bins), but avoids a runtime branch and fixes the problem of how to release the dynamic memory so not all bad. |
| 03:48.21 | Notify | 03BRL-CAD:brlcad * 55052 brlcad/trunk/src/libbu/heap.c: the printing function need not be made public api, tighten the scope. also make sure it only prints once since it can conceivably get registered multiple times |
| 03:51.43 | Notify | 03BRL-CAD:brlcad * 55053 brlcad/trunk/src/libbu/malloc.c: ws cleanup |
| 03:52.26 | Notify | 03BRL-CAD:brlcad * 55054 brlcad/trunk/src/libbu/heap.c: few comments, still need to track and reuse freed memory |
| 04:20.19 | Notify | 03BRL-CAD:brlcad * 55055 brlcad/trunk/src/libbu/heap.c: update the comments to reflect the new structure layout which also helps simplify the explanations. remove the old containers (assumes we'll get better cache coherence grouping them together, but that still needs to be tested). |
| 05:17.44 | Notify | 03BRL-CAD:brlcad * 55056 brlcad/trunk/src/libbu/heap.c: save a few more instructions and reduce per-real-allocation overhead slightly by not tracking per page usage. just keep track of the last page being filled since we know the rest are full. |
| 05:29.38 | Notify | 03BRL-CAD:brlcad * 55057 brlcad/trunk/src/libbu/heap.c: rewicker the terminology so that it might actually make some sense to someone else too |
| 05:37.45 | Notify | 03BRL-CAD:brlcad * 55058 brlcad/trunk/src/libbu/heap.c: because it's awesome when structure elements all line up perfectly with the same length |
| 05:59.32 | *** join/#brlcad jordisayol (~jordisayo@unaffiliated/jordisayol) | |
| 07:11.16 | *** join/#brlcad hsrai_ (~hsrai@202.164.53.116) | |
| 07:17.17 | *** part/#brlcad jordisayol (~jordisayo@unaffiliated/jordisayol) | |
| 08:59.25 | *** join/#brlcad crdueck (~cdk@24.212.219.10) | |
| 09:44.17 | *** join/#brlcad crdueck (~cdk@24-212-219-10.cable.teksavvy.com) | |
| 12:00.02 | *** join/#brlcad pepca (~Pepa@ip-78-102-220-106.net.upcbroadband.cz) | |
| 12:15.35 | *** join/#brlcad caen23 (~cezar@92.81.210.84) | |
| 12:37.22 | pepca | Hello again, is there a Git mirror of the SVN repository ? |
| 12:43.33 | brlcad | pepca: there is, but I'm not sure it's up to date |
| 12:45.33 | brlcad | ``Erik: you have it syncing? |
| 12:45.38 | brlcad | pepca: it's http://brlcad.org/brlcad.git |
| 12:48.12 | pepca | hmm... 30 years of history via http transport ... it will take a while |
| 12:51.03 | Notify | 03BRL-CAD:bob1961 * 55059 brlcad/trunk/src/tclscripts/archer/ShaderEdit.tcl: Added support for the camo shader in Archer's comb edit panel. |
| 12:53.22 | ``Erik | just getting a copy of /usr/ports/distfiles to another machine... |
| 12:54.28 | brlcad | ``Erik: nah, I meant the git mirror ;) |
| 12:55.59 | ``Erik | ohhh, yeah, uhm, that should be clicking hourly |
| 12:56.15 | brlcad | there ya go pepca |
| 12:59.21 | brlcad | pepca: fortunately compresses very well, about 300MB |
| 12:59.30 | ``Erik | that'd be a huge clone, so we have a tarball of a checkout at http://brlcad.org/brlcad.git.tar.gz .. let me update it to the most recent (took it off of the hourly as it chewed a bit of cpu and io time) |
| 12:59.41 | Notify | 03BRL-CAD:jordisayol * 55060 brlcad/trunk/misc/debian/brlcad.sh: Properly add/modify the environment variables PATH and MANPATH on deb and rpm packages. |
| 13:00.16 | brlcad | could renice it into oblivion |
| 13:02.05 | ``Erik | could, but I'm hoping to integrate the sync into notify so it syncs on commit instead of hourly, then figure out something less lame for the tarball generation (maybe just keep an anonymous checkout, so I can update and tar quicker |
| 13:02.55 | ``Erik | the jumpstart tarball finished updating |
| 13:03.32 | brlcad | why is the .tar.gz bigger than the .git ? |
| 13:04.00 | ``Erik | because it has the .git and the checkout (and the .git is already compressed) |
| 13:04.22 | ``Erik | (cd /usr/web/brlcad.org && /bin/mkdir brlcad.git.tgz$$ && cd brlcad.git.tgz$$ && /usr/local/bin/git clone -q http://brlcad.org/brlcad.git brlcad.git && /usr/bin/tar zcf brlcad.git.tar.gz brlcad.git && /bin/rm -f ../brlcad.git.tar.gz && /bin/mv brlcad.git.tar.gz ../brlcad.git.tar.gz && cd .. && /bin/rm -rf brlcad.git.tgz$$) |
| 13:04.25 | *** join/#brlcad hsrai (~hsrai@202.164.53.116) | |
| 13:05.01 | brlcad | really worth the extra 100MB then? :) |
| 13:05.15 | brlcad | once you have 300 you can probably extract the checkout faster than downloading it |
| 13:06.36 | ``Erik | *shrug* I wanted to make a trivial 'batteries included' download with the minimum number of steps... untar and go, pull when you want... if you want to tweak the tarball and write up docs for it, knock yourself out :) |
| 13:10.07 | n_reed | hmm, I've heard about consumer level 3d printers for years, but consumer level 3d scanners not so much until now |
| 13:10.15 | n_reed | http://www.kickstarter.com/projects/621838643/desktop-3d-scanner |
| 13:10.24 | n_reed | http://www.engadget.com/2013/03/08/makerbot-3d-scanner/ |
| 13:10.32 | n_reed | http://www.matterform.net/index.php |
| 13:10.54 | pepca | awsome ... i have cloned everything since 1983-12-16 to 2013-04-05 05:37:31 :) |
| 13:11.10 | pepca | but there are no tags in this clone :( |
| 13:11.47 | brlcad | pepca: it wouldn't be 300MB if it included all tags and branches |
| 13:12.00 | brlcad | several GB iirc |
| 13:12.48 | pepca | yeah :) |
| 13:13.06 | pepca | all branches are not important, but tags would be nice |
| 13:15.30 | pepca | just to make sure... master in http://brlcad.org/brlcad.git == trunk in http://brlcad.svn.sourceforge.net/viewvc/brlcad/brlcad/ -- is it correct ? |
| 13:16.46 | ``Erik | so git pull --tags |
| 13:17.02 | ``Erik | yeah, the git is trunk |
| 13:17.15 | pepca | ok |
| 13:17.48 | pepca | tags are not in http://brlcad.org/brlcad.git, cannot pull something that is not there ;) |
| 13:18.14 | ``Erik | ah, hm, guess they didn't come through in the git svn checkout |
| 13:18.32 | pepca | yes |
| 14:22.17 | Notify | 03BRL-CAD:brlcad * 55061 brlcad/trunk/src/libged/inside.c: be consistent, follow the face winding |
| 14:27.29 | Notify | 03BRL-CAD:mendesr * 55062 NIL: Tag version for 1.4 release |
| 14:29.24 | Notify | 03BRL-CAD:carlmoore * 55063 brlcad/trunk/src/conv/dem-g.c: remove trailing blanks I accidentally put it; put in note about cannot use stdin; remove (redundant) 'Exiting.' |
| 14:29.36 | Notify | 03BRL-CAD:mendesr * 55064 NIL: Tag version for 1.4 release |
| 14:40.34 | *** join/#brlcad ncsaba (~ncsaba@p54981CF4.dip.t-dialin.net) | |
| 14:52.08 | Notify | 03BRL-CAD:brlcad * 55065 brlcad/trunk/doc/docbook/lessons/en/mged15_creating_a_toy_truck.xml: update the inside command parameters to match the new rpp vertex behavior that was made to match the ordering of an arb8 in 7.14.6 (r33973). the "right" parameters were ironically the ones already documented in a note when the lessons were written up in 2001 as having changed. tracing back through the commit history, it |
| 14:52.10 | Notify | looks like this however was entirely unintentional (i.e., it was a bug) that was documented here instead of getting fixed.what had happened: circa 1985 (r373) 'in' command directly implemented rpp (matching arb8), circa 1987 (r1518) wdb rpp routine added (with "natural origin" but not matching arb8), circa 2000 (r18885) 'in' command updated to use libwdb (inadvertently changing the order), circa 2009 (r33973) wdb |
| 14:52.12 | Notify | routine finally fixed to match arb8, and now the doc finally gets fixed to match. so there ya have it, took three devs scratching heads for a day to figure this all out.left a note on how to verify the right values ('l' command so as to not rely on graphical inspection) and the old command so the docs will still work with older releases. |
| 14:57.27 | Notify | 03BRL-CAD:brlcad * 55066 brlcad/trunk/NEWS: with bob and keiths help, we finally figured out what was wrong in lesson 15. it's a four-year old bug with a 30-year legacy. see previous commit log message for all the nitty gritty detais. basically had to update the ordering now that they match the change in 7.14.6 to make the rpp vertices match arb8. |
| 15:11.56 | *** join/#brlcad yiyus (1242712427@je.je.je) | |
| 15:17.43 | *** join/#brlcad luca79 (~luca@net-188-216-230-48.cust.dsl.vodafone.it) | |
| 15:49.59 | *** join/#brlcad jordisayol (~jordisayo@unaffiliated/jordisayol) | |
| 15:50.05 | jordisayol | hello |
| 15:51.44 | ncsaba | Hi Jordi |
| 16:09.03 | ncsaba | soo, I uploaded my first patch to sourceforge |
| 16:29.34 | Notify | 03BRL-CAD:indianlarry * 55067 (brlcad/trunk/include/raytrace.h brlcad/trunk/src/libged/brep.c and 4 others): Initial checkin of shaded displays for BREP NURBS. Ese the poly2tri software to generate CDT triangulations. Water tight edges to follow. Uses "rel","abs" and "norm" tolerance values when determining facet resolution(WIP). Also updated "brep" mged command with features for CDT debugging. |
| 17:15.29 | ncsaba | anybody around ? |
| 17:15.47 | ncsaba | I uploaded a patch, but found an error in it by now |
| 17:17.11 | ncsaba | I wonder if there are some ways to test refactored code other than just visually check if rendering looks the same ? |
| 17:55.49 | ncsaba | OK, after more code-reading I find my patch completely bogus - I will need some more discussion about it |
| 17:57.00 | Notify | 03BRL-CAD:starseeker * 55068 brlcad/trunk/include/CMakeLists.txt: Re-indent and simplify header CMakeLists.txt |
| 19:32.04 | Notify | 03BRL-CAD:jordisayol * 55069 (brlcad/trunk/misc/debian/brlcad.install brlcad/trunk/sh/make_deb.sh brlcad/trunk/sh/make_rpm.sh): Add "/usr/brlcad/version" file on deb/rpm packages. |
| 20:26.55 | Notify | 03BRL-CAD:carlmoore * 55070 brlcad/trunk/src/fb/cmap-fb.c: implement Usage & continue-running for no-argument/no-pipe situation; allow use of -?, omitting -h because that's in use for something else |
| 22:16.29 | Notify | 03BRL-CAD:jordisayol * 55071 brlcad/trunk/misc/debian/brlcad.sh: Verify if $BRLCAD_PATH and $BRLCAD_MANPATH are already in $PATH and $MANPATH environment variables (deb and rpm packages). |