| 01:34.12 | *** join/#brlcad cad911 (~44176e6f@bz.bzflag.bz) | |
| 02:04.15 | brlcad | hello cad911 |
| 03:32.08 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/sh/ (make_dmg.sh make_pkg.sh): ensure the environment by overriding the path, the locale, and the umask. search for a suitable tmpdir. |
| 04:01.20 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/Makefile.am: add openUp as a dmg target dependency, automatically edit the ReadMe and Welcome rich text Installer documents to set their version automatically. |
| 04:13.14 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/sh/ (make_dmg.sh make_pkg.sh): search /usr/sbin too |
| 04:15.31 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/sh/make_dmg.sh: try to fit on a 250 zip |
| 04:40.21 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/sh/make_dmg.sh: the engrish was suckish, so make it slightly less suckish |
| 05:31.55 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/sh/make_dmg.sh: run openUp on the volume so that OS X 10.2 will automatically open the directory in Finder when mounted. while we're at it, make sure a disk with the same name we're trying is not already mounted |
| 05:52.14 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/sh/make_dmg.sh: screw the .background .. there can be only one |
| 10:16.39 | *** join/#brlcad Fikrann (~fikrann@zmt.zinc.eu.org) | |
| 10:18.14 | Fikrann | Hello |
| 12:05.17 | *** join/#brlcad EricWilhelm (~ewilhelm@c-67-171-188-54.hsd1.or.comcast.net) | |
| 16:05.28 | narnia | brlcad, in interfacing the express language library and the part21 step files with brl-cad, what should the end-user be working with; the part21 step file entity instances or the brl-cad database equivalent? |
| 16:46.56 | *** join/#brlcad Fikrann (~fikrann@zmt.zinc.eu.org) | |
| 16:49.47 | narnia | Fikrann, hello again. |
| 16:52.05 | narnia | brlcad, even though i have express2html working i do not think it is needed. exppp (express langauge pretty print) could be used as the bases for a new express language cross-reference program. |
| 16:56.15 | Fikrann | brlcad, btw, there are some issues with openGL and configure. configure script seems not to set up compiler flags properly, it doesn't add -lGL to link list. |
| 17:00.17 | Fikrann | I'm using brlcad 7.2.2 on Ubuntu .. I'll tell you what where exactly the problem shows up when I get make rolling again. |
| 17:03.07 | Fikrann | From ./configure: |
| 17:03.10 | Fikrann | Open GL support ......................: yes |
| 17:15.25 | Fikrann | Interesting... I unpacked the program again, run ./configure && make .. and this time it went okay. o.o |
| 17:31.23 | *** join/#brlcad narnia (~terrylr@johann.blauedonau.com) [NETSPLIT VICTIM] | |
| 17:31.53 | Fikrann | wb narnia |
| 17:46.35 | narnia | brlcad, the reason i say that is that exppp prints out an express schema without re-parsing the file. express2html, on the other hand uses both the libexpress parser and a second parser. so basically, express2html is parsing the express language file twice. exppp, only parses the express language file once. |
| 17:47.52 | narnia | microwave links and storms do not play well together. |
| 17:49.03 | Fikrann | Ouch |
| 17:59.52 | narnia | when it rains hard the microwave link drops out. |
| 19:31.08 | Axios | Im beginning to like brlcad |
| 19:37.28 | narnia | Axios, that is a 'good' thing. ;-) |
| 19:42.17 | Axios | ;) |
| 19:42.27 | Axios | does alt + click has a funktion? |
| 19:43.01 | Axios | because in my windows manager that makes me pull the window around (totally nice feature) |
| 19:45.11 | Fikrann | Axios, if it has, I haven't found it. |
| 19:45.38 | Axios | ok |
| 19:45.45 | Axios | just wanted to hear |
| 19:56.18 | Axios | delete key is strande |
| 19:56.21 | Axios | strange |
| 20:12.55 | *** join/#brlcad Fikrann (~fikrann@zmt.zinc.eu.org) | |
| 20:14.36 | narnia | brlcad, have you had any time and/or chance to scan in any of 10303 parts? i am interested specifically in part203. |
| 20:43.34 | brlcad | Fikrann: hrm.. that's rather odd .. did the libGL test fail? if you send me your config.log (or take a peek) I can look into that |
| 20:44.31 | brlcad | Fikrann: automatic opengl detection is still being worked on.. but it should at least be adding the -lGL now.. ;-) maybe a priority issue |
| 20:45.27 | brlcad | narnia, I have not yet, though I did schedule some time to test the pdf scanner .. i'll make 203 my test case then (or at least some of it) |
| 20:45.30 | Fikrann | brlcad, no, it went nicely |
| 20:46.27 | Fikrann | I had all libraries and header files installed .. In fact I've been compiling oGL stuff just before I tried to install brl. |
| 20:48.08 | brlcad | Axios: glad to hear it =) questions? |
| 20:48.31 | brlcad | Axios: the keys are configurable (though I forget the .mgedrc magic that binds them) |
| 20:49.12 | brlcad | Fikrann: but you say it didn't add -lGL? |
| 20:49.27 | brlcad | that implies that the libGL test failed or a "better" library was chosen before it |
| 20:51.20 | Fikrann | No, it didn't add -lGL, had to force it with --with-libs=GL |
| 20:51.28 | narnia | brlcad, that will be fine. i am still testing and mulling over different ideas concerning the libexpress library. i still do not like the idea of writing 10303 ap specific parsers. |
| 20:52.14 | brlcad | Fikrann: can you look into your config.log file and see what happened during the -lGL test? |
| 20:52.26 | brlcad | without using --with-libs option |
| 20:52.46 | brlcad | narnia: what don't you like about that? |
| 20:53.03 | brlcad | isn't that the idea? you write a parser for 203 .. another for 214 |
| 20:55.51 | narnia | brlcad, in my mind part203 should just be treated like modules are in perl. instead of perl we are dealing with express language. |
| 20:56.17 | brlcad | so sort of a mega parser for all things step |
| 20:56.26 | brlcad | that could get very complicated.. :) |
| 20:56.34 | brlcad | very quickly |
| 20:57.45 | narnia | brlcad, we 'see' a part21 step file which is 10303-203. we pull in the part203 short or long form and parse the part21 step file. |
| 20:58.35 | brlcad | i still haven't quite grasped the point of having both a short and long form |
| 20:58.42 | brlcad | shouldn't the long form be sufficient? |
| 20:58.49 | narnia | brlcad, the current method is build a parser on top of a parser. the lowest level parser is the express language parser and next level is the ap parser. |
| 20:58.52 | brlcad | or is the short not a proper subset? |
| 20:59.28 | Fikrann | btw, you might want to drop -g when compiling with --enable-optimized configure option |
| 21:00.16 | brlcad | --disable-debug does that |
| 21:01.11 | brlcad | there's often a need to test optimized with debug symbols too, default is simply to "on" unless disabled |
| 21:01.14 | narnia | brlcad, concerning short and long forms for the ap parts i do not understand that either. the best i am able to come up with is if one of the underlying ir's changes using the short form would pull in those changes where as the long form would not. a new long form would have to be generated using shtolo (short-to-long). |
| 21:01.41 | brlcad | Fikrann: perhaps you could rerun configure without the with-libs option and send me your config.log file? |
| 21:02.41 | brlcad | narnia: hrm.. i didn't think the short/long had anything to do with changes |
| 21:03.07 | brlcad | the long/short are both available from iso in what appeared to be pretty "stable" forms |
| 21:03.08 | Fikrann | I'm checking if it'll crash. |
| 21:03.19 | brlcad | if what will crash? :) |
| 21:03.35 | Fikrann | the build process. |
| 21:03.44 | brlcad | k |
| 21:07.00 | narnia | brlcad, the short form of part203 has several 'use' and 'reference' which 'use' or 'reference' ir's and aic's (part 41,42,43,44,45, etc and 501, 502, etc). the long form of part203 does not. the long form is a 'frozen' snapshot of whatever the ir's and aic's at the time it was created. if part41 changes the part203 long form would have to be generated again. |
| 21:07.55 | *** join/#brlcad cad533 (~8106210e@bz.bzflag.bz) | |
| 21:08.22 | narnia | brlcad, the ir's and aic's are 'shared' between many part2xx. think of the ir's and aic's as include files. the part203 as a specific 'program'. |
| 21:09.00 | brlcad | yeah, I get |
| 21:09.33 | brlcad | sounds like a parser nightmare if the changes are not backwards compatible |
| 21:09.40 | brlcad | hello cad533 |
| 21:10.33 | narnia | brlcad, this is definitely true. |
| 21:10.58 | Fikrann | brlcad, it went okay this time .. I guess it was just a local glitch |
| 21:11.19 | *** part/#brlcad cad533 (~8106210e@bz.bzflag.bz) | |
| 21:11.39 | brlcad | Fikrann: you notice if it was a cached result? |
| 21:12.28 | brlcad | deleting the cache, it should be pretty much deterministic results |
| 21:12.56 | Fikrann | Hmmm... |
| 21:13.09 | narnia | brlcad, in theory, if any of the ir's or aic's change the specific part 2xx parser may have to change also. |
| 21:13.24 | brlcad | it will tell you, however, whether the cache was used or not |
| 21:13.57 | brlcad | narnia: to be honest, I would just stick to the long form pushed out by iso and update as/when they update |
| 21:14.49 | brlcad | unless you're looking to take over that project as well and become the defacto source of "the" step parsing suite ;) |
| 21:15.24 | brlcad | which is fine.. it's just my goal is pretty much limited to importing/exporting step geometry coming from other cad packages ;) |
| 21:17.10 | brlcad | which probably in theory is also limited to pro/e, unigraphics, acis, and solidworks for starters if not just the first two even |
| 21:17.46 | brlcad | for a first step goal, that is |
| 21:21.20 | brlcad | no pun intended |
| 21:35.28 | Fikrann | Hmmm .. it seems that gcc-3.4 is generating faster code than gcc-4.0 with the same compiler flags. |
| 21:54.59 | narnia | brlcad, understand. |
| 22:08.00 | Fikrann | brlcad, where I can find the meaning of the numbers showing up at the end of make benchmark? |
| 22:12.59 | brlcad | Fikrann: there's a couple places where the BRL-CAD benchmark is documented .. one of the most recent is on the website under the presentations link under Documents |
| 22:13.41 | brlcad | Fikrann: another is the doc/benchmark.tr troff file |
| 22:13.53 | brlcad | that's better for explaining it in more detail |
| 22:14.41 | brlcad | it's easy enough to describe too: the results are comprised of two lines, an Abs and vgr line |
| 22:15.59 | brlcad | the Abs line shows your raw performance number for each raytrace test in terms of a number that is rays/s (aka 'rtfm's) |
| 22:16.56 | brlcad | the vgr line shows you how much faster/slower you are from the base system, which is an old vax 11/780 (where the vgr count is 1.0) .. so you are N times faster than that machine |
| 22:17.33 | brlcad | the last number on each row is the important one, in particular the last vgr number, as that's your overall average performance metric |
| 22:18.59 | Fikrann | Yes.. |
| 22:19.06 | brlcad | you can equate that number directly from machine to machine, from version to version, etc to see exactly how well you compare in terms of real end-application performance. and there are historical values that will show how you compare through time (see how you compare to an 512 processor Origin 3000 for example) |
| 22:19.44 | Fikrann | heh |
| 22:20.05 | brlcad | we decommissioned our vax about 5 years ago now .. shame too, that beast was interesting |
| 22:21.16 | brlcad | the historic computing project has a vax in their virtual machine -- I want to set that up for historic regression testing (version to version comparisons) |
| 22:21.55 | brlcad | I got a basic install of NetBSD going on it a couple of years ago and got a compile started.. pretty cool |
| 22:22.33 | brlcad | getting data in/out of the virtual machine was tricky, though |
| 22:22.49 | Fikrann | I guess so .. |
| 22:23.17 | Fikrann | Even trickier than from the real thing, I guess. |
| 22:31.27 | brlcad | that was why the machine had to be decommisioned.. the last remaining disk controller failed and no reasonable means to repair it |
| 22:32.35 | brlcad | the virtual machine is pretty neat, though.. that got me excited for a bit.. attaching a file to the vm as a device, read from the raw device to transfer data into the vm's filesystem environment (no tcp/ip) |
| 22:33.27 | brlcad | we could lock-step the vm to the original cpu frequency and should get a vgr count of 1.0 |
| 22:36.32 | Fikrann | ahh |
| 22:37.50 | Fikrann | btw, looking at the benchmark.tr .. vapour, wilson and wax had 1024 GB of RAM? |
| 22:41.42 | brlcad | yes, those were the supercomputers of the day |
| 22:42.27 | brlcad | ack, GB |
| 22:43.35 | Fikrann | All three has G instead of M .. which makes the amount quite stunning, even in today's terms. |
| 22:47.44 | brlcad | yeah, heh, good eye.. |
| 22:47.53 | brlcad | that was a typo |
| 22:47.56 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/doc/benchmark.tr: heh, old old typo.. wax,wilson,vapor had a gig of memory, not a terabyte |
| 22:48.32 | brlcad | and those weren't one of the supercomputers.. they were just midsize |
| 22:53.08 | brlcad | thx for catching that |
| 23:20.28 | Fikrann | See you all tomorrow. |
| 23:30.35 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/ (INSTALL doc/README.IRIX doc/Makefile.am): separate out the irix notes into it's own readme in the doc directory. refer readers to the doc/README.* files in INSTALL |