00:13.34 |
CIA-4 |
BRL-CAD: 03brlcad * 10brlcad/TODO: eliminate
vararg and non-ansi cray-specific support code |
00:21.54 |
brlcad |
FYI, if you run screen on .bz, I've made a
simple 'rx' script that reattaches you to your existing screen
session |
00:31.11 |
CIA-4 |
BRL-CAD: 03brlcad *
10brlcad/src/rt/viewedge.c: ws expand for readability |
00:33.45 |
``Erik |
O.o "screen -r" is that tricky? |
00:40.22 |
brlcad |
not in the least |
00:40.37 |
brlcad |
it's just a shorthand that saves the typing
(for what is a fairly frequent command for me) |
00:41.04 |
brlcad |
two chars vs 12-22 chars |
00:42.13 |
brlcad |
it does an -x instead of an -r, I like em in
parallel myself |
00:48.48 |
brlcad |
dear god man |
00:49.03 |
brlcad |
it's a toolbox essential |
00:49.18 |
brlcad |
almost as necessary as learning a decent
editor |
00:55.25 |
starseeker |
heh - and I thought using fluxbox was going
minimalist on the graphical desktop side... ;-) |
00:55.50 |
CIA-4 |
BRL-CAD: 03brlcad * 10brlcad/src/libbu/log.c:
don't call va_start until after making sure it's not a null
format |
00:58.48 |
CIA-4 |
BRL-CAD: 03brlcad * 10brlcad/
(src/libbu/bomb.c include/bu.h): |
00:58.48 |
CIA-4 |
BRL-CAD: add a new bu_exit() function that
never returns. this function differs from |
00:58.48 |
CIA-4 |
BRL-CAD: bu_bomb() in that it supports an
error code and printf style arguments. it |
00:58.48 |
CIA-4 |
BRL-CAD: should not be used for 'exceptional'
exit conditions where stability and |
00:58.48 |
CIA-4 |
BRL-CAD: validity are in question; it should
be used when you simply want to politely |
00:58.51 |
CIA-4 |
BRL-CAD: exit the application after printing a
message (like printing application usage |
00:58.53 |
CIA-4 |
BRL-CAD: and exiting). |
00:59.24 |
CIA-4 |
BRL-CAD: 03brlcad *
10brlcad/src/rt/viewedge.c: oops, remove debugging |
01:02.11 |
starseeker |
brlcad: If you can remember this far back,
how long did it take you for the key bindings to become second
nature on screen? |
01:03.27 |
brlcad |
hm, a few days iirc |
01:03.40 |
CIA-4 |
BRL-CAD: 03brlcad * 10brlcad/src/libbu/bomb.c:
stfu about the longjmp handler being taken. the user doesn't care
and the message is rather misleading. |
01:03.59 |
brlcad |
getting used to ctrl-a a and ctrl-a e were
the tricky ones as I use them all the time |
01:04.09 |
brlcad |
but now it's pretty fluid whether i'm in
screen or not |
01:04.31 |
starseeker |
Cool. |
01:04.56 |
brlcad |
you'll want to learn ctrl-a f for flow control
and ctrl-a F to "resize" the screen window (if you reattach from a
terminal of a different size) |
01:05.12 |
brlcad |
it's basically persistent shell sessions, so
long as the machine stays up |
01:05.46 |
brlcad |
so I can be on irc 24/7 or have instant access
to various remote servers that I need, dedicated windows to watch
various log files, etc |
01:06.56 |
starseeker |
Ah, OK. Very nice! |
01:07.39 |
starseeker |
I can definitely see where this would help
avoid the need to use the mouse... |
01:07.56 |
brlcad |
particularly if you run screen on a server
that's always up (e.g. .bz) then you can even ssh out from there to
various hosts and keep windows open to them |
01:08.38 |
brlcad |
you can then attach to that session on the
remote server from anywhere and get back to your "console
workspace" |
01:09.22 |
starseeker |
So a screen "session" would consist of all the
windows open in that particular instance of screen? |
01:09.23 |
brlcad |
other useful commands are creating a new
screen (ctrl-a c) and navigating (ctrl-a n or p) |
01:09.39 |
brlcad |
yeah |
01:09.42 |
brlcad |
they all persist |
01:10.09 |
starseeker |
That's nifty. |
01:10.44 |
starseeker |
On a machine without an X server, this has to
be a "must" |
01:10.53 |
brlcad |
so, for example, I log into .bz where my
screen session consists of one context that has irssi always
running, another has e-mail, another has various logfiles being
watched, and then others are created/destroyed as needed depending
on things I'm working on |
01:11.39 |
brlcad |
even with an X server .. it's low overhead,
mouseless, efficient console/workspace management |
01:12.28 |
starseeker |
Interesting - like the command line itself, it
gives a considerable efficiency increase for a small initial
educational investment |
01:13.52 |
starseeker |
Which email client do you use -
Mutt? |
01:14.53 |
brlcad |
heh, I'm still a pine die hard |
01:15.11 |
starseeker |
Hehe - haven't used pine since my undergrad
days |
01:15.19 |
CIA-4 |
BRL-CAD: 03brlcad * 10brlcad/NEWS: |
01:15.19 |
CIA-4 |
BRL-CAD: user-visible change.. removed longjmp
warning message from applications that was |
01:15.19 |
CIA-4 |
BRL-CAD: called during bu_bomb's where an
application exception handler was registered. |
01:15.19 |
CIA-4 |
BRL-CAD: this was a common cause for confusion
in the converters that use the jump |
01:15.19 |
CIA-4 |
BRL-CAD: handler during normal operations
quite frequently -- made it look like the |
01:15.22 |
CIA-4 |
BRL-CAD: application was repeatedly 'crashing'
during conversion yet the application |
01:15.24 |
CIA-4 |
BRL-CAD: would keep processing. |
01:15.44 |
starseeker |
I should try Mutt - right now I'm using
thunderbird and it gets on my nerves sometimes... |
01:15.47 |
brlcad |
I've used mutt, it's fine .. but I rather like
pine's no-nonesense simplicity (which obviously removes a lot of
functionality) |
01:16.22 |
brlcad |
a list of messages, message folders, basic
attachment support, I'm good |
01:17.12 |
brlcad |
I rather like Apple Mail as well, incremental
searching of all mailboxes in realtime is shibby cool |
01:17.35 |
starseeker |
sweet |
01:17.44 |
brlcad |
nice to see that applications are finally
starting to learn from emacs' incremental searching that was done
20 years ago :) |
01:18.03 |
starseeker |
heh - some things do seem to catch on slow,
don't they? |
01:18.27 |
brlcad |
yeah, like brl-cad and a nice gui ;) |
01:18.34 |
starseeker |
hehe |
01:19.31 |
starseeker |
at the very least, it mixes VTK and
QT |
01:19.35 |
brlcad |
hm, I wonder what kind of response I'd get if
I started a project where I'd pay people to find and fix
bugs |
01:20.07 |
brlcad |
something like $1 to find them and $10 if they
provide a fix |
01:20.39 |
brlcad |
the bugs would have to be approved beforehand
so someone doesn't just dump 1000 bugs down and expect 10000 in
payment of course, but I think something like that might
work |
01:20.52 |
``Erik |
some 'bugs' are opinion :/ |
01:20.58 |
starseeker |
That was my thought |
01:21.07 |
brlcad |
yep, exactly why they'd have to be
approved |
01:21.42 |
starseeker |
It might have some potential, both for utility
and as an advertising mechanism |
01:22.18 |
brlcad |
could use the tracker system .. the poster of
the message would get the $1, and the first to attach a patch
(after the tracker state went from submitted to acknowledged) would
get a paypal |
01:22.43 |
starseeker |
well, first patch that was verified to fix the
bug |
01:23.27 |
brlcad |
I could see putting 1k towards something like
this |
01:23.36 |
brlcad |
get 100 bugs fixed |
01:24.42 |
starseeker |
Might be worth a shot |
01:25.23 |
brlcad |
the idea would be to only approve bugs that
would likely take about an hour to find the bug, create a patch,
and document it |
01:25.57 |
starseeker |
Right. Spelling errors in comments need not
apply |
01:26.15 |
brlcad |
would have to quantify that into the
instructions of course -- don't care if the bug is something like a
documentation bug, "manual page is missing -g option" |
01:26.27 |
brlcad |
yeah, though I certainly wouldn't call that a
bug |
01:26.51 |
starseeker |
me either, but people can have surprisingly
"elastic" definitions at times... |
01:27.26 |
brlcad |
a bug is a run-time flaw that either causes a
crash/corruption, or is run-time behavior that contradicts the
documented and expected behavior |
01:27.53 |
starseeker |
Oh, I was wondering earlier - has the use of
unit testing in BRL-CAD ever been considered? |
01:27.58 |
starseeker |
That's a good definition |
01:28.39 |
brlcad |
there are also compilation bugs and
documentation bugs, but those rarely ever take more than a couple
minutes |
01:28.45 |
brlcad |
(to identify and fix) |
01:28.50 |
starseeker |
Yep. |
01:29.43 |
brlcad |
yeah, unit testing has been extensively
considered.. |
01:30.03 |
brlcad |
right now, it's just a bunch of simple
integration tests that tests overall functionality |
01:30.27 |
brlcad |
with the idea that the tests would be expanded
to even more and more tests, particularly as issues are
found |
01:31.11 |
starseeker |
OK. |
01:31.14 |
brlcad |
now as to doing actual code unit testing, the
issue there has always been that there's about a million lines of
existing code, thousands of "units" that one would want to test,
and nobody wants to write/maintain that code at the
moment |
01:31.27 |
starseeker |
Figured :-) |
01:31.56 |
brlcad |
it's enough work as it is just to ensure that
all the public functions are even documented, much less write code
that properly tests each one |
01:32.35 |
brlcad |
so, we have the minimal maintenance approach
-- tests at a higher level, integration tests that make sure apps
end up doing what they're expected to do |
01:32.46 |
starseeker |
Makes sense. |
01:33.47 |
brlcad |
still could be way better than what we have,
and if someone pulled up a test generator that wrote a test for
each routine, it could be a laundry list to fill in the guts for
each test as people feel like it |
01:34.52 |
brlcad |
did something "vaguely" related with the new
flawfinder 'test' that is in the regression suite -- it runs
flawfinder and if there are any items reported at/above the level
specified, it means that someone made a function call that wasn't
allowed (like bob a few days ago) |
01:35.26 |
starseeker |
Ah :-) |
01:35.35 |
brlcad |
had to change a few hundred instances of code
just to get it down to level 5, but that was good/necessary code
cleanup |
01:36.04 |
brlcad |
I should have level 4 done soon, fixed about
700 issues with about 400 remaining |
01:36.15 |
starseeker |
Impressive. |
01:36.32 |
brlcad |
that's probably a good balance.. level 3 is
just flawfinder being picky :) |
01:36.43 |
starseeker |
Most of these were never triggered by "normal"
BRL-CAD use? |
01:37.05 |
brlcad |
oh, flawfinder just does a static code
analysis |
01:37.12 |
starseeker |
Oh, OK. |
01:37.24 |
brlcad |
it looks that you're using function BLAH, and
BLAH is in it's level N list |
01:37.51 |
brlcad |
e.g. you're using strcmp, which is generally
unsafe .. should use strncmp to make sure you don't read past your
memory buffer |
01:38.17 |
starseeker |
Gotcha. |
01:38.26 |
brlcad |
so whether they're actually security holes or
not depends massively on how the logic is structured |
01:39.13 |
brlcad |
but they're certainly "potentially"
exploitable situations that are almost never absolutely necessary,
so we can eliminate the ones we have and keep it at a given level
(e.g. level 4) |
01:39.35 |
brlcad |
splint blew chunks on brl-cad's code last I
tried it |
01:39.39 |
brlcad |
had trouble parsing several files |
01:39.50 |
starseeker |
hmm. pity |
01:40.04 |
starseeker |
their parser isn't up to snuff? |
01:40.06 |
brlcad |
would be worth trying again, that was years
ago |
01:40.34 |
brlcad |
would make for yet another good test to add
after getting flawfinder down to level 3 issues only |
01:41.02 |
brlcad |
flawfinder was nice because I could just add
it to the sourcebase as it's just a few thousand line perl
script |
01:41.14 |
brlcad |
s/perl/python/ |
01:41.17 |
starseeker |
Ah, that is nice. |
01:41.24 |
brlcad |
1877 lines |
01:41.28 |
brlcad |
er 1866 |
01:43.01 |
starseeker |
IIRC someone had a one line perl script to
decode the Matrix DVD back when the whole DeCSS thing was all the
rage... |
01:43.16 |
brlcad |
yeah, you can do some really crazy
stuff |
01:44.07 |
brlcad |
I have several perl programs that I've gone
back to after many many years have passed, with absolutely no idea
what the code does other than what the comments say |
01:44.50 |
brlcad |
they sure can be fast and compact to write
though ... great for little cleanup and temporary processing
tasks |
01:45.17 |
starseeker |
If there's some way to measure how much like
line noise legal program source code can be made to look as a
function of the language, I'll bet Perl wins |
01:45.20 |
starseeker |
true |
01:46.05 |
brlcad |
that's how the majority of BRL-CAD's files got
a standard header and footer that were auto-populated with existing
documentation pulled from within the file's existing contents
including doxygen labels |
01:46.20 |
starseeker |
:-) |
01:46.29 |
brlcad |
as well as a few other processing tasks I've
had to do over the years |
01:46.54 |
brlcad |
things that awk/sed/find -exec/and friends
couldn't easily handle |
01:48.20 |
starseeker |
Just curious - did you pick up all the various
tools as you worked on BRL-CAD, or did you learn them before
that? |
01:48.28 |
brlcad |
both |
01:48.43 |
brlcad |
some before, some after |
01:49.36 |
starseeker |
Impressive. |
01:49.37 |
brlcad |
over the years, just learning the tools as
needed depending on what is generally best for the job instead of
trying to make the tools I know fit the job |
01:49.51 |
starseeker |
That's the way to do it. |
01:50.00 |
starseeker |
I wish that approach was more common |
01:50.40 |
brlcad |
it really helps to know/learn several
different styles of programming languages too, just gives you a
better fundamental understanding |
01:50.55 |
brlcad |
and not just procedural and
object-oriented |
01:51.19 |
starseeker |
Indeed. |
01:51.31 |
brlcad |
though those are certainly really great
starts |
01:51.52 |
starseeker |
We had a semester in college where they went
over a few different languages, but I think they need to do a lot
more of it. |
01:53.53 |
brlcad |
this "liberal arts" that you speak of.. it
confuses me |
01:54.03 |
brlcad |
what engineering department do they teach that
in? :) |
01:54.07 |
starseeker |
hehe |
01:54.32 |
starseeker |
Guess I should have said "typical
undergraduate education: |
01:54.35 |
brlcad |
j/k, I loved some of my arts and crafts
classes :) |
01:54.49 |
Maloeran |
Liberal arts, that must be part of web
design |
01:54.54 |
brlcad |
the breadth of knowledge is
impressive |
01:55.14 |
starseeker |
Actually, my advanced English class was a lot
of fun. (Don't tell any of my physics peers I said so
though...) |
01:55.32 |
starseeker |
Turns out it really sucked to be a
contemporary of Shakespeare |
01:56.52 |
starseeker |
Everyone else kinda got swamped in the history
books, despite some decent writers being around... |
01:56.53 |
brlcad |
yeah, I can see that .. my advanced Spanish
class was a lot of fun too .. read a book a week and write an
in-depth report on it that everyone discusses |
01:57.24 |
starseeker |
Cool :-) |
01:58.05 |
starseeker |
or "muy bien" as it were |
01:58.36 |
brlcad |
took the placement test for kicks, but then
found out I almost aced it and placed out of two years of the
full-time undergrad courses .. so my courses actually counted
towards my engineering "language requirements" at that level just
as if it were an english course |
01:59.01 |
starseeker |
Sweet |
01:59.04 |
poolio |
college. :) |
01:59.11 |
brlcad |
was 4 classes away from a degree in
spanish |
01:59.18 |
brlcad |
dunno wtf I would have ever done with that
though :) |
01:59.48 |
starseeker |
I'm convinced the second AP test I took in
English is proof that essay grading is impossible to do
correctly... |
02:00.17 |
starseeker |
brlcad: Well, you could translate the BRL-CAD
docs :-) |
02:00.33 |
brlcad |
and those last 4 classes would have been utter
hell .. learning the etimology and history of the language, reading
books in "old spanish" and matching them to the context of the
times they were written in so that language analyses could be
performed, etc .. bleh |
02:00.44 |
starseeker |
Ouch |
02:01.18 |
poolio |
I'm only...like 30 classes away from any
degree |
02:01.31 |
starseeker |
poolio: Hang in there :-) |
02:01.42 |
brlcad |
once you get past the 'basic comprehension'
phase and get into the "study of spanish" as a language, the fun
level drops fast :) |
02:01.48 |
brlcad |
poolio: hehe |
02:01.51 |
poolio |
And I'm having second thoughts on my major, so
yeah... |
02:02.00 |
starseeker |
poolio: What's your current major? |
02:02.10 |
poolio |
I'm currently ECE--Electrical and Computer
Engineering |
02:02.36 |
starseeker |
That can be cool, especially if you like the
lower level work. |
02:03.06 |
brlcad |
I remember having that thought my first year..
"Hmm.. 6 classes + electives and I have a BA in Spanish... or 20+
classes + electives and I have a BS" |
02:03.08 |
poolio |
Thing is I don't, I'm more of an algorithms
dude. But I'm not really interested in pure CS. I'm thinking of
just doing biology or some such |
02:03.32 |
starseeker |
(me thinks the CADR Lisp Machines are neat but
prefers to work with software - fewer burned fingers and mistakes
are easier to fix.) |
02:03.49 |
brlcad |
poolio: applied mathematics? :) |
02:03.51 |
starseeker |
Biology is a good field to go into - lots of
growth |
02:03.56 |
poolio |
Always :) |
02:04.21 |
poolio |
brlcad: eh, I'm not that hardcore into math. I
thought I was really into AI for a while but I'm not so sure
anymore. |
02:04.37 |
brlcad |
heh |
02:04.42 |
poolio |
The issue is my school is mainly
CS/Engineering based so majoring in something else here isn't going
to look very good in the future |
02:04.48 |
brlcad |
well, that's the great part of college, find
what you really love |
02:05.16 |
poolio |
Hope to |
02:05.21 |
starseeker |
My advice would be to follow what you want to
do - forcing yourself to do something you don't like usually
doesn't pay off in the long run |
02:05.23 |
brlcad |
there was one guy in my graduating class that
came in dead-set on some engineering discipline iirc, or maybe
history |
02:06.03 |
starseeker |
Transferring schools isn't unheard of by any
means, if your school isn't strong in what you want |
02:06.07 |
brlcad |
he went away one summer and got a job at a
state park after the first semester, and after that he was dead set
on a carrier in being a state park ranger |
02:06.30 |
starseeker |
:-) |
02:06.40 |
starseeker |
I can see where that would have it's
perks. |
02:06.42 |
brlcad |
and all of his classes afterwards were
directly in support of that (and they of course hired him on the
spot, he was dead passionate about it) |
02:07.28 |
poolio |
that's awesome |
02:07.34 |
poolio |
I've been mountain biking a ton in a local
state park |
02:09.04 |
starseeker |
Well, I'd better hit the hay - this next week
is going to be a doozie. |
02:09.09 |
starseeker |
Thanks all - good stuff! |
02:11.39 |
poolio |
nite |
02:24.55 |
brlcad |
poolio: awesome |
02:25.04 |
brlcad |
see ya guys |
03:57.23 |
poolio |
brlcad: you still around? I"m getting a mildly
confusing linker error |
03:57.45 |
poolio |
/home/poolio/brlcad/src/conv/.libs/lt-asc2g:
error while loading shared libraries: librt.so.19: cannot open
shared object file: No such file or directory |
03:57.59 |
poolio |
(during build) |
04:29.38 |
brlcad |
jeez |
04:30.06 |
brlcad |
wth is causing that .. you're not the first to
report that issue, yet I don't have a single system that shows that
problem |
04:30.50 |
louipc |
what's the test? |
04:30.54 |
brlcad |
try removing the two LD_LIB lines in
db/Makefile.am |
04:31.26 |
louipc |
oh build error |
04:32.15 |
brlcad |
it's trying to run the asc2g converter in db
to generate the example .g files -- to do that, libtool creates an
lt-asc2g that is supposed to work with the as-yet-not-installed
libraries .. "some" change in the linking logic seems to be
provoking libtool to failure on some subset of systems |
04:34.03 |
brlcad |
presumably some consistent (bugged) version of
libtool or some distro with the same config setup |
04:35.50 |
louipc |
why not just distribute .g files instead of
converting? |
04:36.33 |
brlcad |
have thought about it |
04:36.47 |
brlcad |
in part so that the contents of the files can
be tracked in revision control |
04:36.47 |
CIA-4 |
BRL-CAD: 03brlcad *
10brlcad/regress/solids.sh: more consistent echo output, use rt
suffix for the solids.rt render script |
04:36.53 |
brlcad |
instead of being binary files |
04:37.00 |
louipc |
ah right |
04:37.02 |
CIA-4 |
BRL-CAD: 03brlcad * 10brlcad/regress/ (14
files): more consistent echo output |
04:37.55 |
brlcad |
other part is in part to verify that
compilation succeeded prior to installation |
04:38.40 |
brlcad |
just a pita to fight libtool issues with what
it tries to keep track of |
04:39.22 |
brlcad |
if I could get a system to reproduce the
error, I could probably find a work-around fix |
04:39.22 |
louipc |
yes I've heard of libtool issues heh |
04:39.35 |
brlcad |
yeah, don't get me started on
libtool |
04:39.51 |
brlcad |
it's the weakest of the (already flaky)
autotools build chain |
04:40.11 |
brlcad |
still useful overall, but sometimes a royal
pita |
04:45.06 |
poolio |
brlcad: remove the LD_LIBRARY_PATH
lines? |
04:45.35 |
poolio |
and why in db/Makefile.am? Isn't the issue
with src/conv/.libs/lt-asc2g? |
04:46.27 |
CIA-4 |
BRL-CAD: 03brlcad *
10brlcad/src/conv/dxf-g.c: |
04:46.27 |
CIA-4 |
BRL-CAD: refactor and use the new bu_exit()
instead of bombing on normal exit conditions. |
04:46.27 |
CIA-4 |
BRL-CAD: also fix a bug where specifying the
the same filename for the dxf and g file |
04:46.27 |
CIA-4 |
BRL-CAD: would result in it actually 'working'
because it called wdb_open() to |
04:46.27 |
CIA-4 |
BRL-CAD: open/create the the .g output file
before checking if the dxf exists. really |
04:46.29 |
CIA-4 |
BRL-CAD: minor but happened to run into it
when 'dxf-g asdf asdf' surprisingly completed |
04:46.31 |
CIA-4 |
BRL-CAD: successfuly |
05:23.44 |
brlcad |
ugh, nearly 2000 instances of bu_bomb to
check |
05:23.58 |
poolio |
eek. |
05:24.09 |
poolio |
is this for flawfindger or some
such? |
05:24.29 |
poolio |
s/flawfindger/flawfinder |
05:31.52 |
brlcad |
no, I changed the interface |
05:32.27 |
poolio |
eek. I think that effects my tiny portion of
code too |
05:32.50 |
brlcad |
bu_bomb() was a general catch-all application
abort, initially only intended as a means to abort from generally
unrecoverable situations (like being out of memory or detecting
outright memory corruption) |
05:33.08 |
brlcad |
over the years, it grew to be a convenient
means to abort with a message |
05:34.07 |
brlcad |
will now that we have automatic crash-report
generation, it really needed to distinguish between the core
critical aborts (like corruption) and 'normal' application
exits |
05:34.40 |
poolio |
what automatic crash-report generation do you
have? |
05:34.49 |
brlcad |
so there's a new bu_exit() routine .. so I
have to go through the 2000 or so callers of bu_bomb() one at a
time and determine if it's something that would be potentially
useful to have a crash report of |
05:35.54 |
poolio |
Did you write this up somewhere or is this
brand new? |
05:35.59 |
brlcad |
brand new |
05:36.13 |
poolio |
And so bu_bomb does the stack trace and what
not, bu_exit() exits with a message? |
05:36.23 |
brlcad |
yeah |
05:36.51 |
poolio |
makes sense. have fun :) |
05:36.55 |
brlcad |
heh |
05:37.07 |
brlcad |
code gardening |
05:37.13 |
brlcad |
happens |
05:37.19 |
poolio |
that's what interns are for |
05:37.38 |
brlcad |
sometimes |
05:37.55 |
brlcad |
I generally like to put interns on stuff
that's actually fun |
05:38.24 |
poolio |
Heh. I'm very happy about that :) |
05:38.55 |
brlcad |
don't always succeed, but I certainly try
:) |
05:40.17 |
brlcad |
some kids just aren't motivated by code, or
aren't capable of touching/editing code safely |
05:42.33 |
poolio |
yeah...but don't the interns kind of choose
what they're interested in based on their skills? Like are you
given interns to put to work or is it some sort of selection
process? |
05:46.12 |
brlcad |
'yes' |
05:47.26 |
brlcad |
it's a mix, an exceptional mix of talent ..
and depends heavily on whether they are hs or college usually (or
at least how much experience they have thinking for
themselves) |
05:49.38 |
poolio |
well, I must sleep. 9:30 class tomorrow. ta
ta |
06:33.05 |
*** join/#brlcad ibot
(i=ibot@pdpc/supporter/active/TimRiker/bot/apt) |
06:33.05 |
*** topic/#brlcad is BRL-CAD
Open Source Solid Modeling || http://brlcad.org || http://sf.net/projects/brlcad
|| BRL-CAD is on scan.coverity.com but not completely set up yet,
more news TBD || Release 7.10.4 imminent |
07:33.49 |
*** join/#brlcad
MinuteElectron
(n=MinuteEl@silentflame/member/pdpc.base.minuteelectron) |
08:06.06 |
*** join/#brlcad Z80-Boy
(n=clock@zux221-122-143.adsl.green.ch) |
08:06.20 |
*** join/#brlcad Z80-Boy
(n=clock@zux221-122-143.adsl.green.ch) |
08:09.33 |
*** join/#brlcad Z80-Boy
(n=clock@zux221-122-143.adsl.green.ch) |
08:37.40 |
brlcad |
yeah, I shoudl probably do that too |
08:43.56 |
Z80-Boy |
brlcad: have you heard computer generated
singing? |
08:44.29 |
CIA-4 |
BRL-CAD: 03brlcad * 10brlcad/include/ (bu.h
nmg.h): tweak debug log messages for consistency |
08:44.36 |
Z80-Boy |
http://cslu.cse.ogi.edu/tts/flinger/jukebox/song10.mp3 |
08:44.43 |
Z80-Boy |
http://cslu.cse.ogi.edu/tts/flinger/jukebox/song1.mp3 |
08:46.21 |
Z80-Boy |
I wrote a string synthesizer a C program that
takes a guitar tab in ASCII on stdin and spits a .wav file on the
stdout |
08:46.34 |
Z80-Boy |
http://ronja.twibright.com/gut_feeling.mp3 |
08:48.19 |
CIA-4 |
BRL-CAD: 03brlcad * 10brlcad/src/conv/ (14
files): |
08:48.19 |
CIA-4 |
BRL-CAD: first 10% of the review of the
bu_bomb() instances. updating many to bu_exit() |
08:48.19 |
CIA-4 |
BRL-CAD: for the instances where the error or
exit condition aren't catastrophic errors |
08:48.19 |
CIA-4 |
BRL-CAD: (i.e. they wouldn't benefit from
having a stack trace of how we got into that |
08:48.20 |
CIA-4 |
BRL-CAD: state code-wise). this also makes it
easier to refactor usages of log & exit |
08:48.22 |
CIA-4 |
BRL-CAD: too to the new routine. |
08:52.44 |
*** join/#brlcad elite01
(n=elite01@dslc-082-082-082-070.pools.arcor-ip.net) |
08:52.58 |
CIA-4 |
BRL-CAD: 03brlcad *
10brlcad/src/canon/png-ipu.c: more bu_bomb to bu_exit
conversions |
09:07.15 |
*** join/#brlcad butti_
(n=butti@e178112163.adsl.alicedsl.de) |
09:07.27 |
butti_ |
he all |
09:07.54 |
butti_ |
brlcad: 20 mm must be something about 20
m..Blender has no units at all |
09:08.20 |
butti_ |
brlcad: just grids whit 1 unit |
09:09.50 |
*** join/#brlcad butti
(n=butti@e178112163.adsl.alicedsl.de) |
09:10.00 |
*** part/#brlcad butti
(n=butti@e178112163.adsl.alicedsl.de) |
09:10.31 |
*** join/#brlcad butti
(n=butti@e178112163.adsl.alicedsl.de) |
09:11.24 |
butti |
brlcad: mabe its enogh to scall it 1mm to
1m..thats all |
09:22.23 |
butti |
brlcad: I would be very intrested to learn
about this software and I will ask you for help not only one time.
For this moment, Im runnig already another taktic. No time for
brlcad :(. My Project needs to be finished (some objects must be
chenge and new will coming. So if plan to make something from this
file, which I gave you, pls dont, is not ready yet. I will draw in
qcad each view in 2d. I will take a shot from blender wires view
and take the |
09:26.52 |
*** part/#brlcad butti
(n=butti@e178112163.adsl.alicedsl.de) |
09:37.50 |
*** join/#brlcad butti
(n=butti@e178112163.adsl.alicedsl.de) |
09:42.32 |
butti |
brlcad: do you know the software javawiew? I
love to work with that. Lot of inspirations comes from this app to
me. It has almost the same gui. It is an app. which is working
with mathematical formel and its make quit the same what all 3d
progs doing. It is lovely. Take a look at that! http://www.javaview.de/ |
09:58.44 |
*** join/#brlcad louipc
(n=louipc@bas8-toronto63-1128543972.dsl.bell.ca) |
11:12.45 |
*** join/#brlcad minute
(n=MinuteEl@silentflame/member/pdpc.base.minuteelectron) |
13:40.21 |
*** join/#brlcad
MinuteElectron
(n=MinuteEl@silentflame/member/pdpc.base.minuteelectron) |
14:09.48 |
CIA-4 |
BRL-CAD: 03erikgreenwald *
10brlcad/src/conv/g-nmg.c: use bu_exit() instead of bu_log() on
unknown argument (I think that's what was intended... segfaults
using bu_log with given args) |
14:25.47 |
CIA-4 |
BRL-CAD: 03erikgreenwald *
10brlcad/src/libbu/semaphore.c: make i unsigned to match
nsemaphores |
14:46.28 |
CIA-4 |
BRL-CAD: 03erikgreenwald *
10brlcad/src/libbu/htester.c: |
14:46.28 |
CIA-4 |
BRL-CAD: Add missing include for bu.h. Fixed
sign issues. Rearranged and modernized. |
14:46.28 |
CIA-4 |
BRL-CAD: Added -v (version) flag and improved
argument checking. |
15:16.54 |
*** join/#brlcad Elperion
(n=Bary@p548743D2.dip.t-dialin.net) |
15:35.37 |
CIA-4 |
BRL-CAD: 03erikgreenwald *
10brlcad/src/mged/clone.c: start stubbing in the -c stuff |
15:36.44 |
*** join/#brlcad bpoole
(n=bpoole@UNIX33.andrew.cmu.edu) |
15:42.15 |
``Erik |
oi, hoolio |
15:58.28 |
bpoole |
mornin' |
16:13.27 |
brlcad |
butti: heh, blender does have units, just
doesn't expose them very well |
16:13.40 |
brlcad |
butti: no, I'm not familiar with
javawiew |
16:15.37 |
butti |
<PROTECTED> |
16:15.41 |
brlcad |
ahh, it's like mathematica's plugin, but in
java |
16:15.50 |
brlcad |
shame it's not open source |
16:17.02 |
butti |
brlcad: you cant change the units to meter or
zoll,,this is like that an im gald about that |
16:17.46 |
brlcad |
it wasn't a problem, it's trivial to scale it
up even after importing into brl-cad |
16:18.21 |
butti |
sure |
16:18.48 |
``Erik |
*shudder* |
16:18.53 |
brlcad |
I noticed because rtedge takes the absolute
size into account when determining whether to draw an edge --
several of the features were "too small" when I first rendered
it |
16:19.07 |
CIA-4 |
BRL-CAD: 03erikgreenwald *
10brlcad/src/mged/clone.c: crude -c functionality |
16:19.19 |
``Erik |
yes, rtedge should use like a fraction of
scene bounding sphere or something O.o |
16:19.44 |
``Erik |
*shudder* don't look at that patch, I'm
fighting ugly with ugly to get something in their hands fast, and
am starting to fix it... |
16:19.57 |
``Erik |
oh, um, heh, that's right, I was gonna look at
that :D |
16:20.12 |
brlcad |
i won't, I"m on a bu_exit() frenzy probably
for the rest of the day |
16:20.40 |
``Erik |
yeah, I had to fix one of those in
src/conv/g-nmg.c O.o |
16:20.40 |
brlcad |
different semaphore bug |
16:21.00 |
``Erik |
bu_log(1, "blah", argv[0]); ? O.o |
16:21.27 |
brlcad |
there's "some" bug in rt at least where it'll
crash during raytrace, something is either not being protected or
something in semaphore.c has changed that breaks it |
16:21.50 |
``Erik |
ohhh, the src/libbu/semaphore.c ? heh, no, I'm
just fixing trivial Wall type errors there |
16:21.55 |
brlcad |
it's exceptionally random, and seems to only
happen a few % of the time, so I've not been able to provoke it to
find it |
16:22.09 |
``Erik |
hum, -p3000 help expose it? |
16:23.07 |
brlcad |
if I put rt in a while loop rendering to
/dev/null, it usually can happen anywhere from a few dozen to few
hundred iterations (and sometimes not even then) |
16:23.49 |
``Erik |
does the bomb code catch unexpected seggies
and bus faults? |
16:24.51 |
``Erik |
hum, tcl85b1 triggers portaudit, 'ReadImage'
has known buffer overflow issues O.o |
16:25.22 |
brlcad |
no, I didn't want to add the signal handlers
before release until all the issues were sorted out with
bu_bomb() |
16:26.22 |
brlcad |
plus wasn't positive about the scope of impact
and how it'd play if there were multiple signal handlers
installed |
16:26.28 |
``Erik |
I guess pick a fast machine multiproc and ask
rt to render 1e6 frames in gdb or dbx or something? :) |
16:26.52 |
``Erik |
without a stack trace or something, it'd all
be guessing |
16:27.12 |
brlcad |
I got stack traces.. they're just ..
bizarre |
16:27.42 |
``Erik |
url? |
16:28.34 |
brlcad |
http://pastebin.bzflag.bz/m41c678a9 |
16:28.54 |
brlcad |
the moncount was just because I made a profile
build, happens with/without it |
16:29.32 |
brlcad |
the *parent* thread crashed there.. which
should be just quietly waiting for the threads to exit |
16:29.38 |
``Erik |
O.o |
16:29.46 |
brlcad |
and that was probably 1/3 into the
ray-trace |
16:29.56 |
``Erik |
heh, g_arb.c:798 is a variable decl on my
source O.o |
16:30.22 |
brlcad |
that line changes from crash to
crash |
16:30.33 |
``Erik |
heh |
16:30.34 |
brlcad |
thread 0 is usually the one that crashes
though |
16:30.39 |
``Erik |
only tested on a hackintosh? |
16:30.53 |
``Erik |
<-- tends to favor bugchasing on bsd or
even leenewx to a mac :/ |
16:32.00 |
``Erik |
find . -name '*.[ch]' | xargs sed -i.bak
's/register //g' && cvs -z3 commit -m "damnit." |
16:32.01 |
``Erik |
*cough* |
16:32.02 |
``Erik |
O:-) |
16:32.54 |
``Erik |
make dist && open http://sf.net/projects/brlcad #
damnit. *cough* O:-) |
16:33.56 |
``Erik |
hum |
16:34.20 |
brlcad |
like I said, I can't reliably reproduce
it |
16:34.24 |
``Erik |
with the rtedge not quitting issue, click
closing the fb exist the program with $? 0 |
16:34.59 |
brlcad |
everytime I make a clean build with debug
symbols, it seems to work, or the bt in gdb is
meaningless |
16:35.46 |
brlcad |
I ran a test removing all the register
keywords many years ago .. it actually slowed down
significantly |
16:36.10 |
brlcad |
would be interesting to see if it's still the
case |
16:36.14 |
``Erik |
O.O "rtedge -F/dev/X havoc.g havoc" lingers
the window |
16:36.28 |
brlcad |
linger is default |
16:36.29 |
``Erik |
but the program exits immediately after the
performance stuff is printed |
16:36.36 |
``Erik |
yeah, but -F/dev/X should over-ride
that |
16:36.38 |
``Erik |
no? |
16:36.46 |
brlcad |
no, that's just X with defaults |
16:37.00 |
brlcad |
Xl would force linger on (if it weren't the
default) |
16:37.47 |
Z80-Boy |
brlcad: I should actually download the newest
code and try out if the arbn bug is now gone in my
situation |
16:37.57 |
``Erik |
and how do you force linger OFF,
then? |
16:38.07 |
brlcad |
and /dev/Xt |
16:38.15 |
brlcad |
transient vs lingering |
16:38.20 |
Z80-Boy |
The poll is missing the entry "I use it for so
long that I don't remember anymore" |
16:38.35 |
brlcad |
Z80-Boy: yeah, should be fixed now |
16:39.51 |
Z80-Boy |
COngrats to putting up the new
website |
16:41.33 |
brlcad |
it's still not quite ready, several issues to
sort out |
16:41.49 |
brlcad |
but I'd like more eyes on it to see how it
renders |
16:43.06 |
Z80-Boy |
I use links so it doesn't render |
16:43.17 |
Z80-Boy |
It's just all gray |
16:44.34 |
``Erik |
that's neat |
16:45.22 |
``Erik |
oh |
16:45.25 |
``Erik |
rt does it for me, too |
16:46.34 |
``Erik |
"rt havoc.g havoc" doesn't exit, but if I give
it -F/dev/X -F/dev/Xl or -F/dev/Xt, it does exit when it
should |
17:36.33 |
brlcad |
Z80-Boy: screenshot? |
17:37.44 |
brlcad |
``Erik: using full cpu? |
17:38.09 |
*** join/#brlcad yukonbob
(n=yukonbob@99.233.193.42) |
17:38.13 |
brlcad |
rtedge was using full-cpu and never printed
the raytrace complete message |
17:38.28 |
yukonbob |
Hello, whirled. |
17:38.35 |
``Erik |
didn't look |
17:39.34 |
``Erik |
hum, it's holding the cpu load up against
1 |
17:42.49 |
``Erik |
same behavior with rt |
17:43.13 |
``Erik |
and right click releases the program |
17:43.16 |
CIA-4 |
BRL-CAD: 03brlcad * 10brlcad/src/conv/ (31
files in 5 dirs): |
17:43.16 |
CIA-4 |
BRL-CAD: whoosh, another 10% complete on the
bu_bomb to bu_exit conversion review. also |
17:43.16 |
CIA-4 |
BRL-CAD: includes a slew of custom exit()
calls that can use bu_exit() as well as toning |
17:43.16 |
CIA-4 |
BRL-CAD: down a slew of overzealous
exclamation point use. (20% complete overall) |
17:44.16 |
``Erik |
heh |
17:44.20 |
``Erik |
grep -r '!!!' src |
17:44.21 |
``Erik |
eeks |
17:45.05 |
``Erik |
hum |
17:45.11 |
``Erik |
and a lot looks like john |
17:45.20 |
``Erik |
png tools, pattern, ... |
17:51.22 |
brlcad |
it's just a style thing, but for messages
being sent to the user, the "emotion" it conveys is a bit wierd
sometimes :) |
17:51.48 |
``Erik |
<-- looking at printed strings, not code
comments |
17:52.13 |
``Erik |
:) a bit much, perhaps *shrug* but I'm not
gonna jump in and do anything right now, got other stuff to
do |
17:52.43 |
brlcad |
hm, so i'm apparently chunking through at
about 10% every 2 hours.. |
17:52.58 |
brlcad |
that means I have 16 hours to go .. oh
joy |
17:53.09 |
``Erik |
when rt and rtedge get wedged like that, the
kill is in a wait4() call |
17:53.23 |
poolio |
``Erik: a snippet of your code reminds me of
some commit you corrected in my code earlier.... 8) |
17:53.49 |
``Erik |
so I'm imagining that the fb process might not
be releasing all the parent fd's or something |
17:53.59 |
``Erik |
which snippet? O.o |
17:54.21 |
``Erik |
adding '#include "bu.h"' or something? O.o
:D |
17:55.18 |
brlcad |
heh |
17:55.19 |
brlcad |
bu_bomb("Choke! ack! gasp!
wheeeeeeze.\n"); |
17:55.37 |
brlcad |
bu_bomb("choke, gasp, *croak*\n"); |
17:55.44 |
brlcad |
funny lee |
17:56.57 |
``Erik |
heh, hopefully no one ever sees the output of
a bu_bomb() *shrug* :) |
17:57.35 |
``Erik |
hmmmm, liboptical/treetherm.c |
17:57.41 |
``Erik |
sh_treetherm.c rather |
17:57.51 |
brlcad |
hence the addition of bu_exit() .. bu_bomb()
are conditions that shouldn't ever be encountered |
17:58.45 |
brlcad |
yay, Heroes tonight |
17:59.22 |
``Erik |
so clone can have "-c [num]" |
17:59.59 |
``Erik |
if I can get my freakin' browser off of
http://www.primitiveways.com/
that is |
18:06.03 |
poolio |
``Erik: "silly coder, this is C! you can't
declare variables mid-scope" |
18:09.14 |
``Erik |
hehehe |
18:09.42 |
``Erik |
yes, in c90, it's illegal to declare a
variable anywhere but the beginning of the scope (or block, if you
like that word better) |
18:11.15 |
``Erik |
try compiling with your CFLAGS containing
something like '-Wall -W -Werror -ansi -pedantic' some time...
it'll hurt you... not as bad as, say, splint, but it will
:) |
18:13.23 |
poolio |
``Erik: I know, I'm just saying you did the
same thing with your last commit :) |
18:18.47 |
``Erik |
'crude c functionality'? |
18:18.49 |
``Erik |
-c |
18:19.14 |
poolio |
jah |
18:19.17 |
``Erik |
<-- not understanding |
18:19.22 |
``Erik |
-c is a flag to the clone command |
18:19.33 |
poolio |
errr |
18:20.14 |
poolio |
I was just giving you a hard time for
declaring a variable mid-scope |
18:20.41 |
``Erik |
um, I don't declare in mid scope |
18:21.00 |
``Erik |
curly brackets define new scope, I have
curly's around that block |
18:21.23 |
``Erik |
or, rather; show me where :) |
18:21.46 |
poolio |
ah durrrrr. |
18:22.04 |
``Erik |
hum, speaking of |
18:22.18 |
``Erik |
cmu, huh? do you know or know of a lee
salzman? |
18:26.38 |
``Erik |
!#~!@!~ fucker :( |
18:26.52 |
``Erik |
is the new season already running? |
18:27.01 |
brlcad |
heh, for like 4 weeks dude |
18:27.09 |
``Erik |
feck |
18:27.54 |
``Erik |
bparker is downstairs in ed's office, btw...
waiting on computer accounts I think, want anything
relayed? |
18:28.14 |
brlcad |
my dvr actually screwed up the weekend I went
to cali |
18:28.34 |
brlcad |
just pulled them off itunes instead |
18:28.44 |
``Erik |
hm, like $2/ep? |
18:28.52 |
brlcad |
sure, "Hi Bob!" .. see him tomorrow
probably |
18:29.14 |
brlcad |
and i'm gonna undo his chmod
breakage |
18:29.45 |
brlcad |
probably can get away with #define fchmod
_fchmod for winderz |
18:30.22 |
``Erik |
heh, yeah, I think the twist for a lot of
posix on winnt is the _ prefix :/ |
18:30.23 |
brlcad |
i'm about ready to get rid of it |
18:30.37 |
``Erik |
and these days, we can probably assume dos
based windows is irrelevant |
18:30.40 |
brlcad |
it's actually made me watch more tv than
before now that I can watch what I want when I want |
18:30.48 |
brlcad |
thinking of just going itunes-only |
18:31.00 |
brlcad |
buy ala cart for the shows I care to
watch |
18:31.09 |
``Erik |
still drm damaged, though, no? |
18:31.15 |
brlcad |
calculated that it would definitely be cheaper
overall |
18:32.03 |
``Erik |
what, $40/mo opposed to $80? |
18:32.36 |
poolio |
``Erik: nope, sorry. |
18:33.00 |
brlcad |
yeah the video is fuxored, though it can be
burnt/ripped iirc |
18:33.01 |
``Erik |
he probably graduated before you got there,
hoolio :) |
18:33.11 |
brlcad |
not sure if anyone's cracked it outright like
has been done for music |
18:33.36 |
brlcad |
140/mo and you don't have a dvr? |
18:33.42 |
*** join/#brlcad Elperion
(n=Bary@p548743D2.dip.t-dialin.net) |
18:33.50 |
brlcad |
that must be a handful of porn
channels |
18:34.02 |
``Erik |
heh, no porn, ... well, not advertised as
porn |
18:34.11 |
``Erik |
I think I have like showtime or hbo or
smething |
18:34.19 |
``Erik |
that I've watched, like, once in the last 4
years |
18:34.52 |
brlcad |
i think dvr is about 80/mo + whatever base
service, so usually 100-120/mo |
18:35.23 |
brlcad |
i used to have all the movie chans.. sure
didn't cost 140/mo |
18:35.41 |
``Erik |
and I got the digital package for 'discovery
wings' and 'bbc', but 'discovery wings' change to 'discovery
military' (and I was a huge fan of the history and 'learning to
fly' shows), and the bbc here is mostly crap like trading space and
shit, not the onslaught of monty python, black adder, red dwarf,
etc in memphis |
18:36.04 |
``Erik |
and of course the cable modem |
18:36.12 |
brlcad |
ahhh |
18:36.15 |
brlcad |
w/ internet |
18:36.17 |
brlcad |
that's the diff |
18:36.26 |
``Erik |
that's probably $50 of it |
18:36.30 |
brlcad |
yeah |
18:36.33 |
``Erik |
like $45 service and $5 rental |
18:36.34 |
brlcad |
that's more like it |
18:37.52 |
``Erik |
doxy in libbu is good, right? |
18:37.58 |
brlcad |
afaik, yeah |
18:38.06 |
brlcad |
almost complete even |
18:38.40 |
``Erik |
I hate to say it, but I think javadoc stomps
doxygen for html output :/ |
18:40.28 |
poolio |
people here use Doxygen for a lot of
projects |
18:41.52 |
brlcad |
with the defaults it does .. default html
output sucks hairy ones; but it's fairly configurable and can be
made to look as good/better |
18:59.07 |
*** join/#brlcad PrezKennedy
(n=Matt@74.86.45.130) |
19:07.00 |
CIA-4 |
BRL-CAD: 03erikgreenwald *
10brlcad/include/bu.h: convenience wrappers for comparing
vls's |
19:17.04 |
``Erik |
make -s in src/conv puts me in a beavis and
butthead mood O.o |
19:19.44 |
``Erik |
:w |
19:20.15 |
``Erik |
I mean, uh, ^C-w or something |
19:23.46 |
CIA-4 |
BRL-CAD: 03erikgreenwald *
10brlcad/include/bu.h: silly coder, semicolons don't go
there |
19:30.40 |
brlcad |
``Erik: is the bu_vls_strcmp() one really
needed? |
19:30.59 |
brlcad |
I'm in the process of eliminating most
instances of strcmp for flawfinder |
19:31.19 |
``Erik |
not particularly, but the pattern is in
several places (5 that I've noticed) |
19:31.22 |
brlcad |
with vls, you always know the length of
both |
19:32.07 |
brlcad |
you mean pattern where that define is
made? |
19:32.21 |
``Erik |
the strcmp(bu_vls_addr(),
bu_vls_addr()) |
19:32.27 |
brlcad |
ah |
19:32.59 |
``Erik |
and I searched through vls for a bit looking
for the compare functions, seemed there ought to be one |
19:33.34 |
``Erik |
<-- working clone into being vls heavy, has
the 'other 4' ready for commit to use the macro... |
19:33.57 |
``Erik |
if it's excessive, I can up -AC them instead
of commit |
19:34.02 |
brlcad |
should probably make the corresponding
bu_vls_strcmp() func then, so it can do the validity checking like
the other vls str* funcs |
19:34.13 |
``Erik |
yeah, I was thinking of doing that |
19:34.16 |
brlcad |
yeah, just nobody refactored to create one
yet |
19:34.43 |
``Erik |
heh, I was kinda thinking of trying to
optimize them by doing length comparisons first, but then ya don't
know if -1 or 1 is the correct return value |
19:35.08 |
brlcad |
yeah |
19:35.33 |
``Erik |
even though every use is ==0 or !=0 so far...
someone will try pushing it through qsort or something eventually
heh |
19:36.02 |
``Erik |
validation would be good, I was just taking a
baby step that direction for now |
19:36.56 |
``Erik |
do I commit the consumer files, or back out
the macros? |
19:37.33 |
brlcad |
yeah, that'd be a freaking hard bug to
find |
19:37.50 |
brlcad |
whatever suits you |
19:38.13 |
brlcad |
won't matter until time to release
:) |
19:38.28 |
``Erik |
<-- kinda prefers completeness for core
functionality |
19:38.56 |
brlcad |
both probably need to happen, so it's just a
matter of which first :) |
19:39.16 |
CIA-4 |
BRL-CAD: 03erikgreenwald * 10brlcad/src/
(libdm/dm-generic.c mged/cmd.c mged/share.c mged/vrlink.c): move to
using the new bu_vls_strcmp macro |
19:39.28 |
``Erik |
clone.c needs testing, I've done a lot in
there |
19:39.40 |
brlcad |
wonder if flawfinder will bitch on
bu_vls_strcmp .. it seems braindead in its string matching
sometimes |
19:40.06 |
brlcad |
i noticed that it complains about some macros
that call strcmp that it completely gets wrong |
19:40.47 |
``Erik |
hm, [^_a-zA-Z0-9] should alleviate that
:/ |
19:41.23 |
``Erik |
unless it's pushing it through cpp
first |
19:41.28 |
brlcad |
yeah, I think it "tries" to determine context
and just gets confused by cpp stuff |
19:44.27 |
brlcad |
g'dammit .. msvc doesn't seem to have fchmod
at all.. only _chmod |
19:49.19 |
CIA-4 |
BRL-CAD: 03erikgreenwald *
10brlcad/src/mged/clone.c: change get_name() to return a bu_vls
instead of a char* |
19:53.45 |
*** join/#brlcad
MinuteElectron
(n=MinuteEl@silentflame/member/pdpc.base.minuteelectron) |
19:54.12 |
CIA-4 |
BRL-CAD: 03brlcad * 10brlcad/ (5 files in 4
dirs): |
19:54.12 |
CIA-4 |
BRL-CAD: wrong direction bob.. working to
remove _all_ instances of _WIN32 not in a core |
19:54.12 |
CIA-4 |
BRL-CAD: library. besides, using chmod() is
actually has a fundamental race condition |
19:54.12 |
CIA-4 |
BRL-CAD: flaw that is a security violation;
and causes the new flawfinder regression test |
19:54.12 |
CIA-4 |
BRL-CAD: to fail. hopefully msvc has _fchmod()
like the other posix funcs it supports |
19:54.14 |
CIA-4 |
BRL-CAD: prefixed. if not, need to add a libbu
compatibility routine. |
19:56.03 |
``Erik |
(of course, if your'e using win32, race
condition security might not be your biggest priority
*cough*) |
19:56.57 |
MinuteElectron |
Windows is awesome. |
19:59.20 |
CIA-4 |
BRL-CAD: 03erikgreenwald *
10brlcad/src/mged/clone.c: retain 2nd (3rd? 87th?) number if no -c
option is given |
19:59.45 |
brlcad |
aww, now that's freaking cool |
19:59.47 |
brlcad |
"I used to live on the 23rd floor of a
high-rise in Chicago. Groceries were never a problem. The 1st floor
of the building was a grocery store and they delivered with
purchases of $20 or more (excluding alcohol and cigarettes).
Likewise all the local grocery stores would deliver to your
apartment free of charge with a minimum purchase. You could phone
or fax your grocery order in and pay for it on delivery (even pay
with a check) or you could go |
20:00.27 |
MinuteElectron |
It got cut of at "you could g" |
20:00.40 |
dtidrow_work |
yeah |
20:00.46 |
brlcad |
or you could go down and select your items,
pay for them, and one of the box boys would lug the stuff up for
you." |
20:01.11 |
MinuteElectron |
Cool indeed. |
20:04.51 |
``Erik |
shoot, if it was that close, just stop by
every day on the way home |
20:04.56 |
``Erik |
mmmm fresh veggies every day O.o |
20:05.37 |
brlcad |
``Erik: presuming you ever needed to leave
home |
20:05.38 |
``Erik |
everyone still crunching numbers on an xmp,
raise your hands. |
20:05.39 |
``Erik |
:D |
20:06.24 |
``Erik |
what's the xmp shtuff mucking with?
O.o |
20:06.28 |
brlcad |
even if anyone HAS an XMP that will still
power up, I'd imagine that it's probably possible to get an updated
posix lib on there that would remove the need for the
hacks |
20:06.40 |
``Erik |
just readability? |
20:06.46 |
brlcad |
it's all throughout |
20:06.48 |
``Erik |
(heh... "just"...) |
20:07.21 |
brlcad |
some readability, a lot of #if chunks, and
some code that needs updating if I'm to get flawfinder below level
four |
20:07.29 |
brlcad |
so it's finally become a maintenance
burden |
20:07.36 |
dtidrow_work |
gee, my laptop prob. has more performance than
a XMP |
20:08.17 |
brlcad |
dtidrow_work: almost guaranteed -- computed it
once -- iirc, the Apple "Cube" that came out about 8 years ago was
almost the same speed |
20:08.32 |
brlcad |
er, maybe 6 years |
20:08.41 |
dtidrow_work |
heh |
20:08.58 |
brlcad |
"cray in a coffee can" |
20:09.04 |
brlcad |
fanless even |
20:09.32 |
dtidrow_work |
did you see the article about the 'unlocked'
Athlon X2 5000+ that recently came out? |
20:10.02 |
``Erik |
didn't the case crack and become discolored on
the cube? |
20:10.09 |
dtidrow_work |
yuo can overclock it to 3.1GHz without
tweaking the voltage, or dissapating much more power |
20:11.47 |
brlcad |
``Erik: yeah, constant heating and cooling
would cause microfractures in the acrylic/plastic for some of them
.. barely visible, but not good for the image when the visual look
of those boxes was one of their best selling points :) |
20:12.02 |
CIA-4 |
BRL-CAD: 03erikgreenwald *
10brlcad/src/mged/clone.c: don't assume the suffix starts with a
'.' |
20:12.38 |
``Erik |
with jobs stickuptheass about the industrial
design of products, that was a suprising... feature |
20:12.56 |
``Erik |
with a dell or gateway, people woulda just
said "meh, they do that" |
20:13.20 |
brlcad |
hmm.. heh "While your new website seems a
little slicker and is probably more functional it lacks the
patriotic feel of the original." ... "please bring back the
eagle" |
20:13.34 |
MinuteElectron |
bah |
20:13.36 |
brlcad |
the eagle is intentionally
deprecated |
20:14.05 |
``Erik |
echo "the new site sucks, black background is
sooooo 1997" | mail -s 'feedback' brlcad@brlcad.org |
20:14.06 |
``Erik |
O:-) |
20:14.08 |
brlcad |
interesting that it was one of the first
comments, though :) |
20:14.29 |
MinuteElectron |
Did you ever get LDAP running? |
20:15.25 |
brlcad |
MinuteElectron: yeah, ldap is running .. i
just need to focus better -- should now just be a matter of hooking
the web apps into it configwise |
20:15.37 |
brlcad |
i'm assuming that was the problem before where
drupal was being pissy with it |
20:15.55 |
brlcad |
but now it's verified working, and I even
kinda understand it now .. |
20:16.01 |
MinuteElectron |
Oh, sounds good. |
20:16.29 |
brlcad |
i just needed a diversion after hacking on it
for half a day and haven't gotten back on it |
20:20.21 |
brlcad |
ah, interesting, cray was also pseudo-big
endian |
20:20.39 |
brlcad |
ints were big, floats were proprietary
format |
20:21.07 |
``Erik |
ieee754 came far later than 2's
compliment |
20:22.12 |
brlcad |
with a hardware pin on the cray el to toggle
between big and little |
20:30.25 |
MinuteElectron |
Whoever designed the gallery2 skinning system
needs shooting. |
20:44.41 |
brlcad |
can only go to user www |
20:44.55 |
brlcad |
(chown) |
20:45.28 |
brlcad |
if don't want to type commands over and over
via sudo, can sudo -s -u www |
20:46.04 |
MinuteElectron |
Ahh, right. Thanks. |
20:46.13 |
MinuteElectron |
I thought you didn't watch the logs
;) |
20:46.27 |
brlcad |
I don't watch user logs |
20:46.34 |
brlcad |
but I do get sudo notifications |
20:46.46 |
MinuteElectron |
Oh, I must be annoying then :P |
20:46.55 |
brlcad |
ctrl-l is my friend :) |
20:47.05 |
MinuteElectron |
:D |
20:47.43 |
brlcad |
I get a lot more than sudo notifications, not
a big deal |
20:47.58 |
MinuteElectron |
ahh, kk |
20:48.24 |
brlcad |
just about any security/access/system-level
related event |
21:00.36 |
CIA-4 |
BRL-CAD: 03brlcad * 10brlcad/ (28 files in 12
dirs): |
21:00.36 |
CIA-4 |
BRL-CAD: the old cray cos and unicos code for
the cray xmp and friends finally became a |
21:00.36 |
CIA-4 |
BRL-CAD: burden to maintain while resolving
various flawfinder issues. remove the old |
21:00.36 |
CIA-4 |
BRL-CAD: code, particularly that which wasn't
__STDC__. however, intentionally leave the |
21:00.36 |
CIA-4 |
BRL-CAD: code for calculating byteoffsets and
converting float formats to/from network |
21:00.39 |
CIA-4 |
BRL-CAD: order .. could still come in handy at
some point down the road. |
21:02.33 |
MinuteElectron |
Back to Plan A. |
21:07.21 |
brlcad |
MinuteElectron: how hard is it to swap right
and left menus on the site so that the config matches on the
right/left side correctly? |
21:07.29 |
CIA-4 |
BRL-CAD: 03brlcad * 10brlcad/ (21 files in 8
dirs): removed the non-ansi/stc compliant noalias.h and
noalias-prag.h headers that were used with the cray (and alliant)
to make pragma decls before loops. made obsolete long ago with the
move to ansi compliance. |
21:08.14 |
MinuteElectron |
I could do it with little effort. Give me five
minutes or so. |
21:08.17 |
brlcad |
cool |
21:09.27 |
brlcad |
I'm also thinking to maybe change the header
to not require alpha transparency .. i (annoyingly) got a few IE6
complaints about it rendering badly already |
21:09.42 |
brlcad |
more prevalent that I thought still
apparently |
21:11.58 |
MinuteElectron |
I will look into pngfix again, I've tried it
before, but it didn't work. |
21:12.54 |
MinuteElectron |
Also there are methods of single color
transparency, but I'm sure you know about them. |
21:13.05 |
MinuteElectron |
I can also try converting images to gif, that
might work. |
21:13.10 |
CIA-4 |
BRL-CAD: 03brlcad *
10brlcad/src/libdm/dm-pex.c: whoops, here's to reviewing all commit
e-mails even when I make them.. apparently edited out the wrong
line for noalias.h |
21:13.34 |
brlcad |
pngfix requires javascript iirc, no? |
21:13.42 |
brlcad |
I think these are systems that have scripting
turned off |
21:14.15 |
MinuteElectron |
:( |
21:14.24 |
brlcad |
I was thinking of redoing the header
regardless so that the menu can expand verically too |
21:14.27 |
MinuteElectron |
Ok, GIF it is, I will sort it out after
left\right columns |
21:14.41 |
brlcad |
(or you can, have at it) :) |
21:15.36 |
brlcad |
I'm thinking that it might actually look even
better to do what's done on the right side panel with the ovals for
the menu, just not beveling the bottom two corners |
21:16.01 |
brlcad |
that would let it expand horizontally and
vertically, and could be more readily tied into drupal's menu
management |
21:16.19 |
brlcad |
looks like atm it's ignoring/not using
drupal's built-in menu management, just having the menu
hard-coded |
21:16.47 |
brlcad |
maybe something similar for the bevel for the
two search tabs too, so they expand correctly |
21:17.03 |
MinuteElectron |
The problem with tieing the menu system
management in with Drupal is the wiki anf gallery2. |
21:17.14 |
brlcad |
ah, true dat |
21:17.29 |
brlcad |
fair nuff |
21:17.42 |
brlcad |
but still would be nice if it expanded
correctly :) |
21:18.09 |
brlcad |
I got horizontal working, just need veritcle
(which means making the bevel automatic like on the other
boxes) |
21:19.50 |
MinuteElectron |
The problem with vertical is the background
images. |
21:20.39 |
brlcad |
? |
21:20.41 |
brlcad |
howso? |
21:20.49 |
MinuteElectron |
oh, wait |
21:20.50 |
brlcad |
it's got a solid background color |
21:20.52 |
MinuteElectron |
you got rid of them lol |
21:22.28 |
brlcad |
the way the bezels are now, it's a one pixel
white hilight on left, top-left, and top edge; and then with a dark
gray 1 pixel edge on the bottom, bottom-right, and right
edges |
21:22.43 |
brlcad |
that pretty much translates directly to the
trick done for the right panel bezel |
21:23.14 |
MinuteElectron |
oh cool |
21:23.15 |
brlcad |
just that one is 3 pixels thick and includes
bottom-left and top-right corners |
21:23.32 |
brlcad |
completely image-less |
21:23.41 |
brlcad |
just css tricks |
21:24.16 |
MinuteElectron |
Why is it you want vertical
expansion? |
21:24.36 |
brlcad |
so it's not reliant upon expected font
sizes |
21:24.49 |
MinuteElectron |
ahh, ok |
21:24.51 |
brlcad |
font size isn't guaranteed with any browser,
user can override |
21:25.23 |
brlcad |
cmd-shift-+ on a mac in safari or firefox will
dynamically adjust the size |
21:26.55 |
MinuteElectron |
hmm |
21:27.08 |
brlcad |
example: http://my.brlcad.org/~sean/tmp/fontsize.png |
21:27.23 |
MinuteElectron |
yeah |
21:27.24 |
brlcad |
everything resizes correctly example the menu
because of that fixed height |
21:34.26 |
MinuteElectron |
ok, this is weird |
21:35.34 |
brlcad |
don't know if you noticed, but I tweaked a lot
of the style settings on the section headers |
21:36.15 |
MinuteElectron |
yes |
21:36.58 |
MinuteElectron |
blocks fixed |
21:42.51 |
*** join/#brlcad poolio_
(n=poolio@c-71-206-215-46.hsd1.pa.comcast.net) |
21:43.14 |
MinuteElectron |
gifs don't work well with some of the sliding
transparency you have |
21:45.16 |
*** join/#brlcad poolio__
(n=poolio@c-71-206-215-46.hsd1.pa.comcast.net) |
21:49.36 |
brlcad |
yep, I know .. I was thinking entirely
changing that so it basically doesn't need any
transparency |
21:50.03 |
MinuteElectron |
aww, ok |
21:50.33 |
poolio__ |
geez. my internet is out of
control...sorry |
21:50.58 |
brlcad |
i think it's pretty cool, but not worth it if
it's going to look like total ass on a major subset of
visitors |
21:51.35 |
brlcad |
alternatively, could leave it like it is, and
just have a separate fixed header with no truck for the lame
browsers |
21:52.04 |
MinuteElectron |
It would be possible to use IE if statements
to do so. |
22:19.09 |
*** join/#brlcad Z80-Boy
(i=clock@77-56-82-71.dclient.hispeed.ch) |
23:03.03 |
*** join/#brlcad poolio
(n=poolio@c-71-206-215-46.hsd1.pa.comcast.net) |
23:03.05 |
Maloeran |
This is unexpected. This chunk of code
performs 18 times faster with -O2 -mtune=nocona in comparison to
binaries produced without optimisation |
23:03.18 |
Maloeran |
I had never seen such a gap before, I didn't
even think it possible |