00:11.49 |
yukonbob |
Ralith: re: source -- yes. It's in DocBook
format, suitable for html if you like. |
00:12.06 |
yukonbob |
grab the distro from svn and checkout
/doc |
00:12.16 |
yukonbob |
s/checkout/look in/ |
00:12.20 |
yukonbob |
~svn |
00:12.20 |
ibot |
Subversion (aka SVN) is version control
software that aims to be a better CVS than CVS. See http://subversion.tigris.org/. |
00:12.55 |
yukonbob |
~distribution |
00:12.55 |
ibot |
distribution is probably a collection of
goodies built around a kernel, also a whole lot of infrastructure
and people. |
00:13.00 |
yukonbob |
gah |
00:13.19 |
pacman87 |
~cadsvn |
00:13.20 |
ibot |
To obtain BRL-CAD from Subversion: svn
checkout https://brlcad.svn.sourceforge.net/svnroot/brlcad/brlcad/trunk
brlcad |
00:13.39 |
yukonbob |
ya, what he said. |
00:13.43 |
yukonbob |
:P |
00:13.48 |
yukonbob |
pacman87: hi :) |
00:13.53 |
pacman87 |
yukonbob: howdy |
00:14.02 |
yukonbob |
happy friday |
00:14.25 |
CIA-60 |
BRL-CAD: 03andrecastelo * r31803
10/brlcad/trunk/misc/win32-msvc9/libged/libged.vcproj: Added files
to libged.vcproj. |
00:17.49 |
Ralith |
yukonbob: cool, tanks. Guess you wouldn't know
where to find technical docs on the file format? |
00:22.45 |
yukonbob |
well, there's the source, for one --
otherwise, I don't think I've personally come across a technical
paper describing it, no... |
00:22.58 |
pacman87 |
g2asc? |
00:25.58 |
Ralith |
what about it? |
00:25.59 |
Ralith |
also, brb |
00:26.16 |
Ralith |
yukonbob: so long as the relevant code's well
commented that's not too bad. |
00:36.24 |
brlcad |
Ralith: no, v5 is the current .g file format
revision |
00:36.36 |
brlcad |
the database format is versioned separately
from the rest of brl-cad |
00:37.01 |
brlcad |
v4 and v5 are all that you'd ever
encounter |
00:37.26 |
brlcad |
it's been v5 since about release 5.0 (just
coincidence they match) |
00:37.41 |
brlcad |
andrecastelo: he's on vacation, gets back
tomorrow |
00:38.34 |
brlcad |
Ralith: and docs for the .g format are on the
website, under documentation iirc |
00:38.46 |
brlcad |
best reference are the headers and sources,
though |
00:39.00 |
brlcad |
there are a few minor items out of
sync |
01:22.32 |
Ralith |
Ah, I thought the versioning thing might be
the case. |
01:30.18 |
andrecastelo |
brlcad: hm ok ok :D |
01:30.36 |
andrecastelo |
I was kind of worried since the mid term
evaluations end monday |
01:57.00 |
*** join/#brlcad thing0
(n=ric@203-59-111-122.dyn.iinet.net.au) |
02:00.48 |
brlcad |
starseeker: other huge payoff I forgot to
mention, decoupling from X11 on Mac -- having a "BRL-CAD Shell" and
MGED that work identical and natively on Mac and Windows |
02:01.15 |
brlcad |
andrecastelo: no worries on your part, the
admins back the mentors up if needed |
02:01.41 |
Ralith |
that reminds me |
02:02.06 |
Ralith |
why does brl-cad have its own command window
thing instead of just opening the shell in the terminal it was
launched from? |
02:02.30 |
pacman87 |
Ralith: that's classic mode |
02:02.33 |
pacman87 |
mged -c |
02:02.52 |
Ralith |
oh, cool. |
02:03.42 |
brlcad |
you can run both simultaneously too |
02:03.57 |
brlcad |
mged -c, select nu or X, then run
"gui" |
02:04.50 |
brlcad |
another fun way, mged -c, select nu or X, then
run "attach X" for as many window views as you want |
02:07.31 |
Ralith |
fun! |
02:09.08 |
Ralith |
how hard would it be to get the command
equivalent of arbitrary GUI actions automatically? |
02:15.28 |
brlcad |
depends which gui actions, what actions
they're taking |
02:15.48 |
brlcad |
some of the tcl logic is actually processed in
tcl for the tcl/tk interface |
02:15.54 |
brlcad |
most of it isn't though |
02:16.21 |
Ralith |
was imagining that it might be educational to
have that sort of thing dumped to stdout or somewhere every time an
action was performed |
02:16.26 |
brlcad |
for editing, you can get most of the options
on the faceplate with the 'press' command |
02:16.40 |
brlcad |
the geometry browser does exactly
that |
02:17.03 |
brlcad |
it'll output every command that corresponds to
what it's showing |
02:17.24 |
brlcad |
the main menu actions, though, are a hybrid
mix |
02:17.30 |
Ralith |
neat |
02:17.35 |
Ralith |
kk |
02:17.57 |
brlcad |
"press help" to get a list of built-in actions
press supports |
02:18.56 |
brlcad |
if you "sed some_primitive" .. it goes into
edit mode, with the faceplate gui enabled, you get a menu in the
graphics context. you can run: press "Menu Option" to press most
of the edit operations and run the "p" command or mouse edit from
there |
02:19.46 |
Ralith |
sounds useful |
02:19.50 |
Ralith |
especially if I ever have to script
something |
02:25.22 |
brlcad |
yep |
02:25.51 |
brlcad |
the mged quick reference card is a good
starting point for learning what commands are available, and their
categorization |
02:25.56 |
brlcad |
(tis on the website) |
02:26.38 |
brlcad |
~botmail for mafm:
http://www.ogre3d.org/docs/api/html/classOgre_1_1ManualObject.html
.. looks like what we'd need for rendering wireframes/points
manually |
02:27.50 |
Ralith |
Hm. In following the tutorials in the
introduction to mged, it's apparent that the raytracer makes
visible objects which are arguably hidden, such as the intersection
between a two objects rendered along with the difference between
two objects taken in both directions. |
02:27.54 |
Ralith |
Is this intentional? |
02:31.11 |
brlcad |
hm? |
02:32.37 |
brlcad |
what section/page? |
02:32.51 |
Ralith |
when you create a rcc, then a sph with a
centerpoint on the center of the top of the rcc and a radius equal
to that of the rcc, then render sph + rcc (an inverted dome) and
sph - rcc (an upright dome) and rcc - sph (a cylinder with an
inverted dome cutout) you can see a sliver of the sph +
rcc. |
02:32.58 |
Ralith |
sec |
02:33.17 |
brlcad |
or screenshot |
02:33.17 |
Ralith |
pdf page 57 == real page 45 |
02:33.28 |
Ralith |
said sliver is visible both in the image in
the pdf and in the actual rendering |
02:38.30 |
brlcad |
ah! I see what you're getting at |
02:39.27 |
brlcad |
they're not hidden, for that particular
intersection it just numerically happens to work out that there is
overlap on the tangency (which naturally extends all the way to the
surface) |
02:40.12 |
Ralith |
tangency? |
02:40.46 |
brlcad |
where the cynlinder meets the sphere, it's
perfectly tangent to the sphere's surface |
02:41.21 |
brlcad |
so there is a thin sliver of material there
where they both exist on the exterior |
02:41.34 |
Ralith |
isn't that infinitely thin, though? |
02:41.40 |
Ralith |
and therefore invisible? |
02:41.51 |
Ralith |
or does the raytracer make approximations
necessary for performance reasons? |
02:42.37 |
brlcad |
sure, that sliver's thickness numerically
approaches being infinitely thin with inifinite precision
arithmetic |
02:42.52 |
*** join/#brlcad thing0
(n=ric@203-59-111-122.dyn.iinet.net.au) |
02:43.15 |
brlcad |
but yes, performing infininte precision or
even fixed precision is two-to-three orders of magnitude slower
than floating point |
02:43.16 |
Ralith |
ahh. |
02:43.36 |
Ralith |
so it's proportional to the actual way the
math is being performed. |
02:43.37 |
Ralith |
neat! |
02:43.50 |
Ralith |
thanks for the in-depth answer |
02:44.23 |
brlcad |
and for the sake of being a *solid* modeling
engine and a solid model ray-tracer, you want all sements and
objects have some real non-infinite thickness when sampled for
analytic purposes |
02:44.34 |
Ralith |
I had no idea that that was the sort of place
floating point imprecision became visible, but I guess if it's
reasonable anywhere it's places like that. |
02:45.16 |
Ralith |
yeah, I thought it might actually be
intentional for those reasons, but sacrificing accuracy because of
that alone seemed strange. |
02:45.39 |
brlcad |
oh yeah, floating point precision can wreak
absolute hell if it's not managed consistently and robust -- our
librt library goes through extensive hoops to validate and do its
best to stabilize the numerics |
02:57.29 |
Ralith |
the primitive selection menu appears to be
broken in the latest release. |
02:57.44 |
Ralith |
when I click an item, the console gives this
error: |
02:57.45 |
Ralith |
db_lookup(-n) failed: -n does not
exist |
02:57.45 |
Ralith |
bgerror failed to handle background
error. |
02:57.45 |
Ralith |
Original error: |
02:57.45 |
Ralith |
Error in bgerror: window name
"_secErrorDialog" already exists in parent |
03:25.29 |
Ralith |
brlcad: any ideas? |
04:20.53 |
Ralith |
I can't work out how to do this on the command
line, so it's blocking my use of brl-cad just as I was learning
enough to do cool things :/ |
05:34.21 |
*** join/#brlcad thing1
(n=ric@203-59-111-122.dyn.iinet.net.au) |
07:51.43 |
*** join/#brlcad clock_
(n=clock@77-56-95-5.dclient.hispeed.ch) |
09:12.55 |
*** join/#brlcad vedge
(n=vedge@205-237-251-204.ilesdelamadeleine.ca) |
09:45.09 |
*** join/#brlcad thing0
(n=ric@203-59-111-122.dyn.iinet.net.au) |
09:53.35 |
*** join/#brlcad elite01
(n=elite01@unaffiliated/elite01) |
11:54.30 |
*** join/#brlcad Axman6
(n=Axman6@pdpc/supporter/student/Axman6) |
11:57.41 |
*** part/#brlcad Axman6
(n=Axman6@pdpc/supporter/student/Axman6) |
12:14.48 |
*** join/#brlcad dtidrow
(n=dtidrow@c-69-255-182-248.hsd1.va.comcast.net) |
14:30.27 |
*** join/#brlcad alex_joni
(n=juve@emc/board-of-directors/alexjoni) |
14:44.14 |
*** join/#brlcad
andrecastelo___ (n=chatzill@189.71.73.20) |
16:03.00 |
*** join/#brlcad Axman6
(n=Axman6@pdpc/supporter/student/Axman6) |
18:13.07 |
*** join/#brlcad
archivist_emc
(n=archivis@host81-149-119-172.in-addr.btopenworld.com) |
18:30.26 |
*** join/#brlcad esben
(n=esben@0x573ff382.boanqu1.static.dsl.tele.dk) |
18:34.00 |
*** join/#brlcad esben
(n=esben@87.63.243.130) |
18:45.33 |
*** join/#brlcad Ralith
(n=ralith@216.162.199.202) |
18:47.46 |
Ralith |
anybody about? |
18:49.24 |
Ralith |
I'm getting progress-halting errors when
progressing through the goblet tutorial |
18:52.06 |
Ralith |
Specifically, when I try to use the GUI
primitive selection tool, I get the following errors (one is
printed each time I click on a primitive in the list): http://codepad.org/YHA1WcK4 |
19:08.21 |
Ralith |
ah, *there's* the command line version. The
gui bug is still annoying, though. |
19:15.46 |
Ralith |
also, occasionally the command window opens at
a very small size. |
20:14.20 |
*** join/#brlcad thing0
(n=ric@203-59-111-122.dyn.iinet.net.au) |
20:32.57 |
``Erik |
O.o |
20:44.19 |
*** join/#brlcad thing0
(n=ric@203-59-111-122.dyn.iinet.net.au) |
20:57.18 |
brlcad |
Ralith: howdy |
20:58.53 |
brlcad |
yeah, that's a known bug that happened in the
latest release, you can either revert to a previous binary or
.. |
20:59.03 |
brlcad |
follow the suggestion I mention here: http://sourceforge.net/forum/message.php?msg_id=5088783 |
21:16.04 |
Ralith |
cool, thank |
21:16.04 |
Ralith |
s |
21:44.42 |
brlcad |
dives into obscene
quantities of sushi he picked up on the way home |
21:52.21 |
``Erik |
heh |
21:52.26 |
``Erik |
that tofi place? |
22:27.04 |
Ralith |
mm, sushi |
22:45.49 |
brlcad |
yeah, pretty decent |
22:46.06 |
brlcad |
huge thick pieces of sashimi |
22:46.48 |
Ralith |
I need to get more good sushi |
22:46.56 |
Ralith |
standard store stuff sucks |
23:14.33 |
*** join/#brlcad pacman87
(n=Timothy@71.170.63.120) |