irclog2html for #brlcad on 20070314

00:09.06*** join/#brlcad IriX64 (n=mario_du@bas3-sudbury98-1168051043.dsl.bell.ca)
00:22.48MaloeranHence the "very often" part
00:25.12brlcaddon't care about the rate, care about the prevalence :P
00:26.34brlcadaligning to 8 is fine anyways, the problems would be alignments > 16 or non % 8
00:35.34Maloeran*nods* Overlooking SSE, the cache memory can prefer getting proper alignment for chunks of 16,32,64 bytes
00:38.03brlcadwhat I really need is a configure test that compares 8, 16, 32, 64 to see which is fastest for some simple operation
01:12.50*** join/#brlcad IriX64_ (n=mario_du@bas3-sudbury98-1168056388.dsl.bell.ca)
01:24.45*** join/#brlcad IriX64 (n=mario_du@bas3-sudbury98-1168056388.dsl.bell.ca)
06:08.38``ErikI put in a ticket to get that del 2450 offically recognized and permitted... but the hw seems to have issues...
07:13.48*** join/#brlcad clock_ (i=clock@84-72-60-125.dclient.hispeed.ch)
07:18.18CIA-7BRL-CAD: 03brlcad * 10brlcad/src/other/libtermlib/ (termcap.src Makefile.am): add a termcap source file, version 9.13.25, pulled from NetBSD. avoids license issues and works with existing library sources.
07:18.50MaloeranDarn. You make me check in the source, and fbsd's malloc really does align on 16 bytes
07:19.53MaloeranIt seems a bit weaker than Linux's though, and of course suffers the same problem of a global mutex lock
07:20.27MaloeranPosix threads could have included proper scalable memory management as part of the standard
07:32.24brlcadyou would probably appreciate the results from a raw malloc benchmark that I wrote/ran several years ago
07:33.09brlcadit allocated memory across the entire spectrum of allocation sizes, looking at the raw allocation performance numbers
07:33.32brlcadran the test on most operating systems available at the time, with some wildly surprising results in some cases
07:34.58brlcadyou could see the see various allocation size optimizations that some libraries catered too, pathalogical horror allocation sizes, performance scalability
07:55.28CIA-7BRL-CAD: 03brlcad * 10brlcad/src/other/libtermlib/ (10 files):
07:55.28CIA-7BRL-CAD: update from the old vi termcap sources to the more modern NetBSD implementation.
07:55.29CIA-7BRL-CAD: this version doesn't suffer the assumption that there needs to be an
07:55.29CIA-7BRL-CAD: /etc/termcap among many other improvements. it's under 3-clause BSD and seems
07:55.29CIA-7BRL-CAD: to be version 0.6 dated December 17, 2006.
08:12.26CIA-7BRL-CAD: 03brlcad * 10brlcad/misc/enigma/configure.ac: use newer autoconf/automake symantics
08:12.56CIA-7BRL-CAD: 03brlcad * 10brlcad/misc/enigma/Makefile.am: include the manpage in the dist, clean up after junk. makes a local make distcheck work for enigma.
08:36.29CIA-7BRL-CAD: 03brlcad * 10brlcad/src/other/libtermlib/termcap.c: also search in BRLCAD_DATA/etc for a termcap file. quell warning on cgetent() constness.
08:36.52CIA-7BRL-CAD: 03brlcad * 10brlcad/src/other/libtermlib/Makefile.am: install the termcap.src file into BRLCAD_DATA/etc
08:38.16CIA-7BRL-CAD: 03brlcad * 10brlcad/misc/enigma/Makefile.am: just use dist_mans
08:42.38CIA-7BRL-CAD: 03brlcad * 10brlcad/NEWS: fixed jove/termcap issues on Mac OS X
08:44.58CIA-7BRL-CAD: 03brlcad * 10brlcad/NEWS: updated bundled libtermcap to NetBSD version 0.6
08:48.07CIA-7BRL-CAD: 03brlcad * 10brlcad/src/ (libbu/vfont.c librt/vshoot.c): bu_malloc is guaranteed to never return NULL, so remove the unnecessary checks
10:02.41*** join/#brlcad clock_ (n=clock@zux221-122-143.adsl.green.ch)
11:59.31CIA-7BRL-CAD: 03d_rossberg * 10brlcad/src/librt/g_brep.cpp:
11:59.31CIA-7BRL-CAD: missing return statements added
11:59.31CIA-7BRL-CAD: min and max initialized
11:59.50clock_brlcad: did you manage to repair some of the buggers I reported?
12:14.41*** join/#brlcad rossberg (n=rossberg@bz.bzflag.bz)
13:31.33brlcadyou never saw it?
13:33.10brlcadit made Apple reimplement their malloc between 10.2 and 10.3 (and what an improvement they made..)
13:37.04``Erikdon't believe I've seen it, no... I know you've told me about it, but ...
13:38.22brlcadpretty simple code, came up with it back in college days as a way to take down the uni's servers
13:38.48``Erik:)
13:38.49brlcadcould outright lock everyone out of the system on-demand, even consolers
13:38.51archivistnaughty boy
13:39.27brlcadbut that was also just due to the "inexperience" of those sys admins to figure out how to counter it
13:40.09brlcadlater turned it into a basic stresser, added some timers, started making graphs .. found lots of surprising results quickly across implementations
13:40.18archivistuni has to be a good training for sysadmins
13:40.37brlcadindeed :)
13:40.57``Erikhrm, probably failry easy to bin os's by which allocator they use
13:41.22``Erikdl vs phk ...
13:48.02*** join/#brlcad ak_real (n=d4b0f9ad@bz.bzflag.bz)
14:32.06brlcadaha!
16:23.00*** join/#brlcad clock_ (n=clock@zux221-122-143.adsl.green.ch)
16:28.59*** join/#brlcad clock_ (n=clock@zux221-122-143.adsl.green.ch)
16:45.14IriX64./configure --for-sean built successfully, but without opengl sigh.
16:45.47IriX647.6.2 however opengl works yay  :)
16:46.25IriX64heh man the parts weren't made to fit ;)
16:46.50clock_IriX64: you mean overlap?
16:47.30IriX64_clock I mean as you and me unles you're a woman of course.
16:47.40IriX64oh man...
16:47.54IriX64modelling it? how would you get it built?
16:49.24*** join/#brlcad SWPadnos (n=Me@dsl245.esjtvtli.sover.net)
16:49.45IriX64there's a picture of the opengl window on my space, forget what number.
16:50.28IriX64www.spaces.live.com/irix64
16:51.44IriX64its called ogl
17:00.16IriX64mmmm x works, got a havoc uo on the screen.
17:00.22IriX64up too.
17:03.16IriX64that "x" pix i just uploaded is 7.8.4.
17:04.24IriX64before you wonder, I have nothing but time on my hands and i love playing with software.
17:06.25IriX64rt however says rt:  can't open frame buffer, no cuch device "/dev/ogll"/
17:07.04brlcadshould be using /dev/Xl
17:07.17IriX64i just typed rt
17:07.22IriX64no parameters
17:07.26brlcadthen you have FB_FILE set somewhere
17:07.35IriX64shouldnt it default to what you have up?
17:07.54IriX64syas opengl wasn't built.
17:08.05brlcadit defaults to what it was compiled to use, unless FB_FILE is set or unless you specify
17:08.14IriX64ty ill look
17:08.17brlcadthe "fbhelp" command lists the default and all others supported
17:08.25IriX64ty again
17:11.05IriX64christ man its used everywhere but its not in the ./include dir files at all unless im missing it.
17:11.25brlcadhuh?
17:11.42brlcadFB_FILE is an environment variable, nothing to do with the source code
17:11.58brlcadfbhelp is a command (outside of mged)
17:12.11IriX64my mistake thought it was a define somewhere :)
17:13.05brlcad"brlman fbhelp" explains
17:13.21IriX64ty ill read it.
17:18.20IriX64what dir plz brlcad i can't find it.
17:19.28IriX64looking in html docs dir.
17:20.21brlcadwhat dir what?
17:20.35IriX64ill look later, cot a compile to set in motion at the moment.
17:20.51brlcadi have no idea what you're asking for
17:20.59brlcadbrlman and fbhelp are *commands*
17:21.07brlcadlike mged and rt are commands
17:21.13brlcadthere are lots and lots of commands
17:21.23IriX64scuse while I eat egg :)
17:21.28brlcadyou don't look for them -- they should be in your path
17:23.25IriX64say whatever happened to -doxygen?
17:23.44IriX64deprecated?
17:29.33brlcadyour question continues to fail parsing
17:29.49brlcaddoxygen comments are used throughout the sources
17:37.47IriX64the switch is gone or at least it doesn't appear in the summary anymore.
17:53.39brlcadwhat switch?
17:54.40IriX64--enable-doxygen
17:55.20IriX64its in 7.6.2
17:56.20IriX64and there used to be a line in the summary, i just noticed in 7.8.4 trhat line isn't there anymore.
17:57.19IriX64its also in ./configure --help, let me check 7.8.4
17:58.29IriX64not there anymore.
17:59.04*** join/#brlcad ak_real (i=ak_real@87.249.59.53)
17:59.05brlcadahh, configure option
17:59.13brlcadhelps to be specific
17:59.30brlcadthe switch didn't do anything in 7.6.2
18:00.03brlcadso it was removed until someone actually adds the functionality
18:00.04IriX64let generate a 7.6.2 configure i'll pastebin it for you.
18:00.08brlcadplease don't
18:00.11IriX64ok
18:00.57brlcadi'm quite familiar with what we did and do .. you just had originally said "-doxygen" without even mentioning configure and without that even being the option
18:01.10brlcadi know it was removed, i removed it
18:01.11IriX64sorry
18:01.13brlcadbecause it did nothing
18:01.38brlcadnow, if you're volunteering to make it actually do something .. like run doxygen, that would be sweet
18:01.58brlcadotherwise, "patches welcome"
18:06.25ak_realhello people!
18:06.36ak_reali'm a newcomer...
18:13.11ak_reali would like to take part in GSoC and became a BRL-CAD developer... who should i contact to discuss that?
18:21.34MaloeranIf you have features or improvements in mind to add, I'm sure brlcad will welcome good patches
18:26.21archivistlurk in here and join the mailing list
18:31.51ak_realoh thx! where can i find mailing list?
18:32.14``Eriksf.net/projects/brlcad
18:45.24IriX64brlcad ty -M -F0 does nicely.
18:53.26brlcadak_real: you're in the right place.. but have to wait until orgs are announced
18:54.04brlcadwe're not guaranteed a slot, and might not be approved .. we find out later today
18:54.58brlcadthat said, even if we're not approved, there are independent job opportunities available ;)
19:00.45ak_realit is interesting too, because i need some CAD-related practice for my school this summer...
19:02.08brlcadwe are really the only choice for open source CAD, but whether it's a direction google wants to support this year is of course up to them (we never applied before this year)
19:02.40brlcadak_real: what sort of project did you have in mind, and have you seen our ideas page?
19:10.06ak_realyeah, i read it a lot of times :)  i like the idea about BREP operations, because we had geometric modeling lectures at school and it's interesting to try them on practice
19:10.14``Erikhum, halfway down the page "The effectively amounts to..."
19:10.32``Erik(in dual-rep geometry support)
19:12.31brlcadak_real: ahh, excellent
19:12.45brlcadthat's actually one of the most interesting/appealing topics at the moment
19:14.03ak_realit's strong math skills requirement there... i had such skills about 3 years ago, but now i may take a look in some books, am i right?
19:14.13*** join/#brlcad IriX64 (n=mario_du@bas3-sudbury98-1168056670.dsl.bell.ca)
19:14.24``Erikplease to be chmod 775 on directory
19:14.53``Erikmeh, I'll just sudo
19:15.14brlcadall web content is owned by www
19:15.16brlcadsudo -u www
19:18.13brlcadmore than pondering apparently
19:18.55``Erikfor a in `jot 1000` ; do sudo -u www echo "eek, I'm being spied on!" ; done
19:19.32brlcadak_real: pretty strong math skills required .. but yeah, using a book to refesh wouldn't be uncalled for ;)
19:20.43brlcadhaving it evaluate as fast as possible (e.g. real-time) is going to be the hardest part I fear
19:21.11brlcadespecially for arbitrary operations
19:22.20``Erik"Kaleido" ?
19:23.00brlcadhuh?
19:24.42``Erikin doc/IDEAS
19:24.54``Eriklast one with numbered difficulty/time
19:29.10brlcadahhh
19:29.24brlcadjust a random idea
19:30.45brlcadKaleido is a nifty 3D geometry code that generates various mathematical shapes
19:31.50brlcadwould make for a sweet mged plugin for the math types
19:31.56brlcade.g. http://www.math.technion.ac.il/~rl/kaleido/data/
19:32.30brlcadhttp://www.math.technion.ac.il/~rl/kaleido/dual.html
19:32.57``Erikhm, neat
19:33.15brlcadwould be even better as a direct primitive that would take those same equations as the parameters
19:33.28brlcadto get all those shapes from one primitive.. sweet ;)
19:33.39ak_realit's time to go to bed for me... see you tommorow... big THX for this warm chat ;)
19:33.53``Eriknight, ak_real o.O
19:33.56brlcadak_real: nice talking to ya!
19:34.06ak_realthx )
19:34.37*** part/#brlcad ak_real (i=ak_real@87.249.59.53)
19:42.24``Erikheh, op-bw, neat
19:44.15``Erikyay, pi day
19:45.03``Erikjust like warm apple pi
19:46.10archivisthome made lemon merangue pie
20:32.31CIA-7BRL-CAD: 03brlcad * 10brlcad/src/other/libtermlib/ (termcap.c tgoto.c tputs.c tputws.c): we don't provide the __RCSID macro that netbsd provided, comment it out.
21:05.41*** join/#brlcad docelic (n=docelic@212.15.184.17)
21:06.10brlcadgaaaaaaaah
21:06.16brlcaddamn termlib
21:07.08brlcadthought I had a successful compile last night, but turns out that the impl is based on a slew of functionality that is in BSD libc, not gnu's
21:07.20brlcadso linux is giving major hissy fits
21:38.37louipcwow over 10K downloads / month?
21:38.40louipcwhere is everybody?
21:40.25brlcadthe vast majority aren't devs or otherwise interested in participating apparently
21:41.35brlcadthat's where I believe work on the website and modeler interface would do well to hook lots of folks' interest
21:42.10louipcyeah that's what I just thought. They ran away after a taste of mged.
21:42.20archivistsexy gui would make the real difference
21:45.23louipcand java applets to look at models
21:45.31louipchar har
21:54.31*** join/#brlcad louipc (n=louipc@bas8-toronto63-1096669545.dsl.bell.ca)
22:23.23brlcadlouipc: heh, the guy that (mostly) wrote mged actually has that working .. (java applet)
22:24.35brlcadnever took it to a functional state though, just playing around and got some geometry up iirc (like 4 years ago)
22:29.00``Erikbob?
22:31.21brlcadyeah
23:06.55IriX64that little 5-clicks thing does png.
23:08.23IriX64poor ValarQ ... I sent her image after image for a while :)
23:12.15brlcadwhy would you do that?
23:14.14IriX64she asked
23:14.28IriX64so i did (did i screw up?)
23:15.14IriX64you said those pix are gnu licence.
23:17.04brlcadshe probably said "sure" to one image I bet.. and you let the flood gates open
23:17.37IriX64heh a few but i stopped when she asked me why i kept spamming her, thought she was interested.
23:17.57brlcadsee, if it comes to that .. then yes, you screwed up
23:17.59IriX64she turned me on to irssi remember
23:18.17IriX64man shwe was smiling...
23:18.24brlcadi.e. being polite
23:18.41IriX64well whatever i got the message from her :)
23:19.34brlcadone or two images might be interesting, especially if the content was novel/unique/interesting.. beyond that, why would they want 5, 10, 15 images?
23:19.48brlcadespecially when the first ones weren't really interesting
23:20.08IriX64you speak for her do you
23:20.13brlcadi mean, i don't know what you sent.. they might have all been picaso's
23:20.22brlcadof course I don't
23:20.28brlcadi'm speaking in general
23:20.35brlcadjust curious why
23:20.48brlcadbecause I frankly can't see why
23:20.54IriX64she asked and gave me an email address to send it to.
23:20.56MaloeranI'm sorry to intrude, but who is the "her" being mentionned?
23:21.13brlcadValarQ
23:21.20IriX64ValarQ (not sure of gender)
23:21.53IriX64but i took it short for Valarie, although I shouldnt assume i guess.
23:33.52CIA-7BRL-CAD: 03brlcad * 10brlcad/src/other/libtermlib/.cvsignore: ignore the generated termcap file
23:34.31IriX64man don't you ever rest :)
23:35.49IriX64err wait what would that make me ;)
23:44.32*** join/#brlcad cad28 (n=52f40aab@bz.bzflag.bz)
23:54.49CIA-7BRL-CAD: 03brlcad * 10brlcad/src/other/libtermlib/ (9 files):
23:54.49CIA-7BRL-CAD: revert the NetBSD libtermcap library. there are too many dependency woes
23:54.49CIA-7BRL-CAD: (getcap, db, fgetln, strlcpy) to make it work cleanly without bringing in all of
23:54.49CIA-7BRL-CAD: db too. it was only working well under the osx/bsd's because everything
23:54.49CIA-7BRL-CAD: happened to be available.

Generated by irclog2html.pl by Jeff Waugh - find it at freshmeat.net! Modified by Tim Riker to work with blootbot logs, split per channel, etc.