IRC log for #brlcad on 20090302

00:34.20 *** join/#brlcad elite01 (n=omg@unaffiliated/elite01)
01:30.35 ``Erik *readreadread*
01:30.44 ``Erik hm
01:31.27 ``Erik bah, nascar is still on, I want my cartoons *pout*
01:46.01 *** join/#brlcad Dr_Phreakenstein (n=phreak@216.151.24.198)
01:48.15 *** join/#brlcad BigAToo (n=BigAToo@pool-96-230-124-39.sbndin.btas.verizon.net)
02:13.20 *** join/#brlcad elite01 (n=omg@unaffiliated/elite01) [NETSPLIT VICTIM]
02:13.20 *** join/#brlcad PrezKennedy (n=Matthew@whitecalf.net) [NETSPLIT VICTIM]
02:13.20 *** join/#brlcad kanzure (i=bryan@66.112.232.233) [NETSPLIT VICTIM]
02:36.16 *** join/#brlcad ibot (i=ibot@rikers.org)
02:36.16 *** topic/#brlcad is BRL-CAD Open Source Solid Modeling || http://brlcad.org || http://sf.net/projects/brlcad || GSoC 2008 Highlight: new prototype gui, check it out! || Source Release 7.14.2 is posted (20080207)
04:06.54 *** join/#brlcad Ralith (n=ralith@216.162.199.202)
06:08.29 starseeker brlcad: I made a test manually of a coil coiling inward: http://bzflag.bz/~starseeker/coil_in_ex.png what do you think? Is it a good enough approximation to be worthwhile?
08:57.02 archivist hmm cnc wire bending
09:03.52 brlcad starseeker: I'd still probably say not really -- even that test looks oddly shaped enough that I'd think it was a bug instead of an approximated implementation detail
09:05.17 brlcad maybe if you could control how approximated it was at the expense of speed (like 100 or 1000 bends just to make one turn)
09:05.35 brlcad but yeah, not at the quarter-turn
09:07.07 *** join/#brlcad elite01 (n=omg@unaffiliated/elite01)
10:19.33 CIA-40 BRL-CAD: 03brlcad * r33926 10/brlcad/trunk/src/libged/draw.c: recogized typo and fixd it
10:23.46 *** join/#brlcad Axman6 (n=Axman6@pdpc/supporter/student/Axman6)
10:37.39 *** join/#brlcad _sushi_ (n=_sushi_@84-72-93-63.dclient.hispeed.ch)
11:24.16 *** join/#brlcad mafm (n=mafm@223.Red-83-49-86.dynamicIP.rima-tde.net)
11:50.39 d-lo Mornin all!
11:50.53 d-lo brlcad: Still awake or just woke up?
12:12.33 d-lo ``Erik: Have any essssspertize on decompliation / assembly ?
12:19.18 *** join/#brlcad TC-Rucho (n=tc-rucho@190.191.172.28)
12:19.32 TC-Rucho hello
12:30.51 brlcad d-lo: still awake
12:31.53 d-lo machine....
12:32.14 d-lo I take it you rcurrent work location will be the work location for the day then ? :)\
12:32.18 brlcad trying to fix a stupid resize bug
12:32.49 d-lo as in GUI widget resize?
12:33.06 brlcad it'd be tricky to try and get out right now with the snow depth on top of ice
12:33.24 d-lo how much snow you get so far?
12:33.33 d-lo I got <1" at home :/
12:33.43 d-lo they *said* 5-8"
12:33.49 brlcad it iced through most of the night and has been snowing few a few hours since, maybe 4-6" total
12:34.02 d-lo I got excited and broke out the sleds :/ I think i cursed it.
12:35.47 d-lo brlcad: do you have any exp with disassembly/decompilation?
12:54.09 brlcad sure
12:54.53 brlcad rarely need it these days, though .. usually just to check out what the optimizer did
12:58.09 d-lo I might have to pick your brain when next I see you then.
12:59.24 brlcad some knowledge rot, best to just speak and hit me up with tools at my fingertips, and a wider audience if need be
13:00.39 d-lo Well i have a binary in which I do not have the source for, but I know that it uses a certain open sourced library
13:01.07 d-lo I am trying to find a specific function call so I can hook into it and watch/log the data flow.
13:01.13 d-lo (Personal project btw)
13:01.45 brlcad hook into it..
13:01.58 brlcad java?
13:02.18 d-lo hook's a bad word to use.... um. place a 'watch point' for logging/viewing in a console.
13:02.42 d-lo No, not java. I believe the original language was MSVC6.0
13:04.55 TC-Rucho is there a way to hide/erase the objects that are used for boolean subtraction when drawing a region?
13:05.06 TC-Rucho all this wireframing is driving me nuts
13:05.14 brlcad hm, that's a pretty hefty goal, even to just watch and log without narrowing down and knowing a lot more about that code
13:05.48 d-lo *sigh* i know.
13:06.00 brlcad TC-Rucho: I believe you can change how the negatives are drawn that might help
13:06.28 brlcad "maybe", but there is no way to turn them off outright without switching to a different render mode
13:06.52 TC-Rucho hmm
13:07.00 d-lo I have tracked down what exact lib its using, gotten the source for that and have found (in the dissassembled asm) where the headers I want are imported...but after that I am at a loss :/
13:07.09 d-lo Okay, I stop the OffTopic banter for now ;)
13:07.57 TC-Rucho brlcad: I don't know where to start to change that right now
13:08.03 TC-Rucho any more obvious hint?
13:08.13 brlcad TC-Rucho: yeah, hold on
13:08.36 *** join/#brlcad ibot (i=ibot@rikers.org)
13:08.36 *** topic/#brlcad is BRL-CAD Open Source Solid Modeling || http://brlcad.org || http://sf.net/projects/brlcad || GSoC 2008 Highlight: new prototype gui, check it out! || Source Release 7.14.2 is posted (20080207)
13:16.21 brlcad TC-Rucho: hm, i'm not finding the option either, only the one that changes all linewidths
13:16.27 brlcad TC-Rucho: something else you can do though
13:16.30 brlcad two things actually
13:16.53 brlcad one is to utilize ray-tracing more -- it's there to help show the evaluated/shaded result
13:17.18 TC-Rucho I don't like the mged GUI, I find it ugly, so I use the console style
13:17.32 brlcad with the raytrace control panel open, it supposed to be simple enough to render, turn off the framebuffer, turn on and render, repeat, etc
13:17.34 TC-Rucho I do raytrace from time to time to see wtf is going on
13:17.48 TC-Rucho yeah, already doing that
13:18.17 brlcad another thing you can do, which will entirely depend on your objects
13:18.26 brlcad is to display them evaluated instead of unevaluated
13:18.45 TC-Rucho that sounds interesting, keep talking
13:18.46 brlcad what are you using now to display objects?
13:19.02 brlcad draw? e?
13:19.06 TC-Rucho draw
13:19.20 brlcad so instead of draw, you could use either ev or E
13:19.27 brlcad subtlely different results
13:19.50 brlcad it'll evaluate the booleans in polygonal mode and show an evaluated polygon
13:20.05 brlcad some like it more, sometimes too much, just depends on the model
13:20.30 brlcad if the geometry gets too complex, there are limitations to what can be evaluated
13:20.52 brlcad it'll give you more wireframe instead of less, of course, but it might help
13:21.03 TC-Rucho yeah, I'm working with spheres and it's getting really messy
13:21.54 brlcad and if you have an open-gl-enabled mged, there's yet a third thing you can try that displays in shaded mode all the time
13:22.10 TC-Rucho which would be?
13:23.41 brlcad it's an experimental mode, so use at own caution, and only with ogl display manager, but iirc, it is enabled by turning on z-buffering, turning off lighting, run 'set shaded_mode 2' and redraw your objects
13:24.49 TC-Rucho already tried that about an hour ago, and all it does is fill the object with a plain color, no shades, so it's really not useful at all imo
13:25.15 TC-Rucho want to see a screenshot?
13:25.49 brlcad then it's not enabled correctly
13:26.10 TC-Rucho http://tc-rucho.homelinux.net/Scrots/crappy-shade.png
13:26.13 TC-Rucho that's what I get
13:26.16 brlcad those first two enable/disable steps are required
13:26.23 TC-Rucho hmm, let's see
13:26.38 brlcad yeah, that's just not enabled correctly
13:27.16 TC-Rucho I'll try to enable that with mged gui first, brb
13:27.21 brlcad it's not meant to be easy yet. you're using a dev hook at this point, so it's all manual
13:28.27 brlcad i might have the toggling off on one of them too, easy enough to tell when it works though
13:29.35 TC-Rucho z-buffering on
13:29.37 brlcad the real fix for that is in the pipeline, part of the bigger BREP support effort
13:29.41 TC-Rucho now the light
13:30.24 TC-Rucho lighting off and yet the same thing when redrawing
13:30.26 TC-Rucho =/
13:30.39 brlcad there's at least four combinations there
13:30.45 brlcad to try
13:30.48 TC-Rucho yeah
13:30.52 TC-Rucho I'm on my way
13:31.04 brlcad i don't have an ogl-enabled build at the moment
13:31.34 TC-Rucho ok, this can be really haired, but it's Zbuffer off, lighting on
13:31.57 brlcad sounds about right
13:32.24 TC-Rucho ok, got to get it working in the console based mged
13:32.31 TC-Rucho I'll be back in a minute
13:32.52 brlcad what are you working on?
13:33.02 TC-Rucho oh, Z clipping must be on too
13:33.43 TC-Rucho I'm drawing my webcam for the sake of learning how to use BRL CAD. I've had some rough times trying to move objects (actually still haven't figured out how to move regions in space)
13:33.55 brlcad oed
13:34.13 brlcad there's a whole tutorial geared towards teaching that
13:34.14 TC-Rucho I tried sed and ted (which I found awesomely useful)
13:34.35 brlcad sed is for primitives
13:34.43 brlcad oed is for everything else
13:34.51 TC-Rucho I see
13:34.55 brlcad object edit
13:35.15 brlcad oed / my_region/path/to/primitive
13:35.25 brlcad rot 50 0 0
13:35.28 brlcad accept
13:50.24 TC-Rucho wouldn't that commandline be for editing a single primitive instead of a whole region?
14:05.40 brlcad nope
14:05.56 brlcad the primitive is only required to set an explicit keypoint
14:06.53 brlcad so it will rotate about the primary vertex point of the given primitive specified, but apply a matrix between the left-hand and right-hand side for a given /arbitrary/path/to/objects
14:08.12 *** join/#brlcad ibot (i=ibot@rikers.org)
14:08.13 *** topic/#brlcad is BRL-CAD Open Source Solid Modeling || http://brlcad.org || http://sf.net/projects/brlcad || GSoC 2008 Highlight: new prototype gui, check it out! || Source Release 7.14.2 is posted (20080207)
14:23.03 *** join/#brlcad elite01 (n=omg@unaffiliated/elite01)
14:45.06 *** join/#brlcad Axman6 (n=Axman6@pdpc/supporter/student/Axman6)
14:55.50 *** join/#brlcad samrose (n=samrose@c-24-11-214-181.hsd1.mi.comcast.net)
15:03.20 TC-Rucho I was wondering, does brlcad have at least a 1 step-only undo?
15:06.09 TC-Rucho guess not
15:06.12 TC-Rucho =/
15:06.27 TC-Rucho note to self: always triple check before doing a killtree
15:15.12 *** join/#brlcad hippieindamakin8 (n=hippiein@202.3.77.38)
15:16.05 *** join/#brlcad _sushi_ (n=_sushi_@84-72-93-63.dclient.hispeed.ch)
15:38.16 *** join/#brlcad madant (n=madant@117.196.140.99)
15:55.56 brlcad TC-Rucho: it's a two-step undo, cp file.g backup.g, later cp backup.g file.g to undo
15:57.39 brlcad best practice is to make backups of your .g file every N timeframes (e.g. daily) so that at worst you've never lost more than N work (e.g. a day's worth)
15:58.52 brlcad that'll be addressed with the new gui, but it's still simply not a high priority since the geometry command layer is designed just like the unix command line and is intentionally unforgiving
15:59.29 brlcad you could just as well create proc overrides that makes backups for you or prompts for confirmation just like some folks do with rm
16:00.20 brlcad again, not something being avoided, just not a priority until shaded displays and the new command layer is in place
16:27.45 ``Erik d'lo: uh, a little, mostly quite antiquated, though
16:30.42 d-lo Hrm, okay.
16:30.53 d-lo Looks like I am on my own :/
16:31.08 d-lo ``Erik: You braving the roads or calling today a wash?
16:31.58 brlcad d-lo: it's usually infinitely easier to snoop the wires
16:32.36 brlcad if there's any net communication, or port protocol i/o, or even file i/o
16:33.02 brlcad a lot easier to meter that doesn't require dissassembly or trying to inject man-in-the-middle code
16:33.19 d-lo Well, thats just the issue. ;) I need to see the data BEFORE it gets encrypted.
16:33.22 ``Erik this is my rdo
16:33.41 ``Erik which is good, as they haven't plowed my neighborhood yet :)
16:34.34 brlcad d-lo: ah, then your best bet is probably to run the app through a debugger, break it right before the data is encrypted, then step through the assembly one instruction at a time watching how all the register data changes
16:34.50 brlcad very tedius, but doable
16:35.21 ``Erik hm, I forget if ndisasm attempts to generate labels or just uses offsets
16:36.15 brlcad yeah, there are some intelligent tools that will rebuild from symbols and labels if they are there
16:36.44 brlcad another toolset to learn though, debugger is simple low-level
16:36.53 d-lo brlcad: Right. Thats the approach I am trying to take. As a step before, since I know its using openSSL, then I am trying to grab the offsets for the functions to make my searching in pure asm easier.
16:37.04 ``Erik so you're picking that old project up again? heh
16:37.10 d-lo yeah. ;)
16:37.53 d-lo I am usin IDA. Its doing a pretty good job thus far, but working on a 20MB binary is kicking my laptop in the nuts :(
16:38.03 ``Erik there should be something like a _call directive going on that should make it a lot more readoable
16:38.44 ``Erik readable, even
16:39.13 ``Erik there should be a "wakethefuckuperik" directive that makes my sentences a lot more readable :D
16:42.19 brlcad yay, finally!
16:42.34 brlcad g'damn that was more difficult than it should have been
16:45.13 CIA-40 BRL-CAD: 03brlcad * r33927 10/brlcad/trunk/src/libged/draw.c: add support to ignore a new '-R' option that implies draw should not resize the view. since the actual resize logic is still in the front-end, we just ignore it (and -A, -o) here.
16:46.53 CIA-40 BRL-CAD: 03brlcad * r33928 10/brlcad/trunk/src/mged/chgview.c: add a -R option to the edit_com() interface affecting e/draw/B so that they don't resize/autoview the view if the -R option is provided
16:52.32 *** join/#brlcad Elrohir (n=kvirc@p5B14E5DC.dip.t-dialin.net)
16:53.29 TC-Rucho brlcad: wouldn't it be more conveniente to have oed to work like this?: oed /path/to/group-or-primitive-to-work-with /path/to/reference-primitive <br/> What's the point in telling oed where is it contained? (the rlh path).
16:54.05 CIA-40 BRL-CAD: 03brlcad * r33929 10/brlcad/trunk/ (NEWS src/mged/chgmodel.c):
16:54.05 CIA-40 BRL-CAD: make the 'make' command use the existing view size when creating new objects by
16:54.05 CIA-40 BRL-CAD: utilizing the new -R option on the draw command. this is a change to make's
16:54.05 CIA-40 BRL-CAD: behavior when there are no objects displayed since previously, it would create
16:54.05 CIA-40 BRL-CAD: an object sized to the view, but then autoview to size the view out. this would
16:54.08 CIA-40 BRL-CAD: result in unexpected view changes that would get compounded if the user called
16:54.10 CIA-40 BRL-CAD: make/draw/kill repeatedly.
16:54.37 brlcad TC-Rucho: it'd be easier to just say "oed object" or "oed /path/to/object" but that would imply automatic keypoints based on "something" (perhaps the center of the bounding box for that object)
16:55.16 TC-Rucho brlcad: what about just oed /path/to/group/and/reference-primitive
16:55.32 brlcad then where does it apply the matrix?
16:55.43 TC-Rucho to the greatest group mentioned in the path
16:55.51 brlcad "greatest"?
16:56.05 TC-Rucho parent?
16:56.13 TC-Rucho for example:
16:56.17 brlcad if "path" is the object you wanted to change, then that'd be no different than oed / /path/to/group/and/reference-primitive
16:56.37 brlcad so you saved two keystrokes and introduced ambiguity :)
16:56.45 brlcad there is just one path
16:57.10 brlcad the arguments are the left-hand side of that one path, and the right-hand side
16:57.40 TC-Rucho if I want to edit group I go with oed group/and/reference-primitive if I want to edit path I go with oed /path/to/group/and/reference-primitive. Is that ambiguous?
16:57.45 brlcad every where there is a slash, you can inject a transformation matrix
16:58.35 brlcad no, you're still not understanding what it means to be a path I think
16:58.44 TC-Rucho maybe
16:59.02 brlcad the tutorial does a pretty good job with examples to explain it in detail, fwiw
16:59.12 brlcad including the implications and rationale
17:00.23 brlcad but even using a filesystem path metaphor as an example, "oed /path/to/group/and/primitive" could imply that you want to edit the 'path' object or the 'primitive' object in that path'ed context
17:00.58 brlcad the only reason you specify a path at all is to obtain a keypoint -- you still have to say what object in what context no matter what
17:01.32 brlcad e.g. to qualify instance editing versus global editing
17:01.39 TC-Rucho it took me a while to get what was it all about since I was getting a shitty error about "Unable to find solid matching path" (later figured out it was because I did not draw one of the objects)
17:02.15 TC-Rucho hmm, I'll have to experiment about that
17:02.22 brlcad yes, you have to load the objects for editing (via draw/e/B) before you can oed, sed works that way too
17:02.46 brlcad 'who' tells you what objects are editable
17:04.01 *** join/#brlcad ibot (i=ibot@rikers.org)
17:04.01 *** topic/#brlcad is BRL-CAD Open Source Solid Modeling || http://brlcad.org || http://sf.net/projects/brlcad || GSoC 2008 Highlight: new prototype gui, check it out! || Source Release 7.14.2 is posted (20080207)
17:04.03 brlcad another simple example, you have a room with a bed and a chair, constructed with some primitives .. there's a "room", "chair", and "bed" object as well as their primitives in the database
17:04.15 TC-Rucho I don't yet see the difference in doing oed /combination region/primitive and oed / region/primitive
17:04.40 brlcad "oed / room/chair/leg/arb8" means "move the room"
17:04.52 TC-Rucho aye
17:05.07 brlcad "oed /room chair/leg/arb" means "move just that one chair in that room"
17:05.25 brlcad "oed / chair/leg/arb8" would mean "move all chairs"
17:06.11 TC-Rucho really?, I would have bet that "oed / chair/leg/arb" meant "move just that one chair"
17:06.34 brlcad they are named references
17:06.56 brlcad so if you use 'chair' as a name elsewhere and you move /chair, you move everything
17:07.41 brlcad depends how your hierarchy is contstructed
17:08.39 brlcad it's a basic set of directed acyclic graphs where you can apply a matrix to any node in the graph
17:08.51 brlcad same concept as for any other CAD system really
17:09.52 brlcad main difference is different terms, assemblies regions parts combinations primitives groups
17:10.28 TC-Rucho well, I've used AutoCAD, totally different paradigm, and when I reached the oed part, I was confused as hell, because I thought of it as a referencing system like AutoCAD's move/copy/whatever where one selects a group of objects, then a reference point, and an end point
17:10.54 TC-Rucho so I thought of oed as oed /group/to/work/with /path/to/reference-primitive
17:10.58 ``Erik ow
17:11.09 ``Erik got ~6-7" of snow, just finished shoveling
17:11.11 brlcad main difference with a system like autocad is that every object (well, most objects) are automatically wrapped in a container object by default
17:11.34 brlcad so to change all instances, it involves a more explicit action usually
17:12.17 brlcad TC-Rucho: oed is like "oed /group/to/work/with /path/to/reference-primitive" :)
17:12.31 brlcad your words and what those slashes mean are just not connecting
17:12.56 brlcad means that this is some valid path: /group/to/work/with/path/to/reference-primitive
17:13.21 brlcad and you want to edit the 'path' object instance referenced in the 'with' object
17:13.37 brlcad using 'reference-primitive' as the keypoint
17:14.32 brlcad (more specifically, the 'reference-primitive' keypoint in the '/group/to/work/with/path/to' context)
17:15.46 brlcad it really is one of those lightbulb issues that should suddenly go off eventually and you'll wonder how you ever understood it differently
17:16.12 TC-Rucho right, but still, I don't quite see the difference in doing oed / /region/primitive and oed /group region/primitive
17:16.31 TC-Rucho hmm
17:16.49 TC-Rucho I tried both, and they work the same way
17:17.12 brlcad they'll give the same end-result until you have 'region' used in multiple places
17:17.22 brlcad (not in 'group')
17:18.44 TC-Rucho ok, I have just modelled some basic stuff, now I'll complete the webcam model, and maybe I'll get what you mean once the modell get's bigger and more complex (although I think I would need to reuse some regions in order to see what you mean)
17:18.48 brlcad seriously, oed tutorial
17:18.51 brlcad it will explain a lot
17:19.12 TC-Rucho I've been reading it for a long while now
17:19.30 TC-Rucho it's just that it puzzled me out since I had a preconcept about how it should work
17:21.22 brlcad hm, that discussion gave me an idea for a terminology doc
17:22.46 TC-Rucho yeah, that left hand path and right hand path notation is not what I would call "clear" but anyway
17:24.24 brlcad having to specify the path to a primitive for a keypoint is really the point of confusion
17:25.37 brlcad that's a technical limitation by the fact that there is no keypoint and deriving an implicit keypoint based on bounding box sizes is questionable (someone will still probably do it eventually)
17:34.14 d-lo I say: Take the average of the 8 points of the Boundingbox and make that default keypoint for oed.
17:37.21 TC-Rucho brlcad: I got it now, the only way to understand the difference is to group some regions, and implicitly reusing the regions by making copies of the group
17:37.29 TC-Rucho it is all clear now
17:40.41 ``Erik hrmmm
17:40.55 ``Erik *snrkt*
17:47.57 brlcad heh
17:48.28 brlcad d-lo: that is the idea, what I meant by "someone will still probably do it eventually"
17:49.10 brlcad the problem then becomes unexpected behavior issues since that keypoint isn't exactly easily worked with until a variety of other commands are updated too
17:49.21 CIA-40 BRL-CAD: 03brlcad * r33930 10/brlcad/trunk/doc/ (2 files in 2 dirs): document the new -R option for e/draw/B in the old htmls docs and for B in the new docbook docs
17:49.45 brlcad rotate and translate something around, add an object, reverse the rotation and you find out it's not reversible
17:50.39 d-lo Well thats just operator error/dumbness. Code can't fix that ;)
17:51.00 ``Erik *nod* plus the issue when you have something like a bigassed box with a tiny little cylinder sticking way out and the keypoint is off in space instead instead of where you think it 'should' be
17:51.49 ``Erik ponders allowing keypoints to be manually selected and saved as attr's to be future default for that comb
17:51.51 brlcad d-lo: right, but usability and interface design can seriously make is a non-issue vs a major/subtle cause for confusion and unexpected behavior
17:52.41 brlcad yeah, some what to manage keypoints more as more than data values, second-class citizens in the db sense would be good
17:53.15 brlcad having things like 'l' and analyze report the keypoints would help, visually showing keypoints, etc
17:53.26 ``Erik ooh, here's a good case to screw the 'center of bb' approach, imagine a cylinder with one side slightly cupped (rcc - sph), that'd come out close to the center of the sphere, not the physical object :)
17:54.45 brlcad negatives are the worst, the expected keypoint is usually the center of mass on the positive evaluated space
17:54.54 brlcad s/mass/volume/
17:55.39 ``Erik and "wait while we run rtweight to find the cm" is probably not tolerable :D
17:55.47 brlcad :)
17:57.14 d-lo ``Erik: that might mess up a center of mass approach, but not for center of BB. especially if the center of BB is being used for a simple default keypoint for oed.
17:57.15 ``Erik out of curiosity, when nurbs is working to satisfaction... will it be 'just another primitive', or are we going to push to be kinda a nurbs only engine (mebbe with construction information stored to allow old style editing)?
17:58.20 ``Erik um, bb is computed in a way that it should be close to the union of all primitives in the tree, even if there're subtractions, iirc
17:59.46 ``Erik if this prims minX < bb minX, bb minX = this prims minX; ...
17:59.57 d-lo so a bounding box is *not* maxX, minX, maxY, minY, max Z, min Z ?
18:00.01 brlcad just another primitive
18:00.10 CIA-40 BRL-CAD: 03brlcad * r33931 10/brlcad/trunk/doc/docbook/system/man1/en/ (Makefile.am g_qa.xml gqa.xml): the mged command is gqa instead of g_qa, rename accordingly
18:00.15 brlcad with a bunch of support to go from csg->nurbs on the fly
18:00.44 brlcad so we can still leverage the optimizations and guarantees if there are implicits
18:01.05 ``Erik try it out, dave, make a sub comb and do make_bb on it :)
18:01.09 brlcad but make it even easier to visualize and analyze anything via the bridge
18:02.18 ``Erik aight, I was under the impression that nurbs could represent any geometry we currently have perfectly (within floating point fuzz), so *shrug* was just wondering. it's my day off, my brain ain't fully in gear, nor will it be
18:03.01 brlcad does a final distcheck, suggests others that might care about release compilation do so as well for their favored environment
18:04.01 brlcad ``Erik: it can for the most part .. it's just wildly more complex to store and deal with than other primitives and about an order of magnitude more data values to represent the same thing
18:04.10 d-lo ``Erik: Hrm, just tried it and it performed as expected it to :/
18:04.26 brlcad sphere goes from having about 6 values to over a hundred for a decent approximation
18:04.36 brlcad similar for arb8's
18:05.12 d-lo ``Erik: ah, I see what you mean now.
18:05.21 brlcad instead of 24 values, suddenly becomes about 300 values
18:22.36 *** join/#brlcad mafm (n=mafm@223.Red-83-49-86.dynamicIP.rima-tde.net)
18:23.10 ``Erik thought nurbs could do a sphere in 2 patches
18:23.34 ``Erik or was it 4 (tetrahedron-like)
18:24.27 ``Erik d-lo: how you expect things to work probably isn't how joe blow off the street does :D you're tainted
18:29.56 d-lo Well, I think the number of tainted users > number of Joe Blow Off the Streets ;)
18:47.26 *** join/#brlcad bitminer (n=bitminer@h96-60-82-113.vrnawi.dsl.dynamic.tds.net)
18:47.33 ``Erik in the tiny self-selecting community you see here, sure, but there are what, 670,000 known downloads? I'm imagining a fairly large number of those are unvoiced first time users :)
18:54.37 d-lo won't mention his script "DownloadMetricPadder.sh"
18:54.43 d-lo :D
18:55.55 d-lo but you are right, there are prolly alot of first timers that end up shying away :/
18:58.48 *** join/#brlcad ibot (i=ibot@rikers.org)
18:58.48 *** topic/#brlcad is BRL-CAD Open Source Solid Modeling || http://brlcad.org || http://sf.net/projects/brlcad || GSoC 2008 Highlight: new prototype gui, check it out! || Source Release 7.14.2 is posted (20080207)
18:58.51 brlcad TC-Rucho: you have a full tcl interp there so that can be used
18:58.56 ``Erik tcl sub-expressions
18:59.00 TC-Rucho I checked some tcl's docs and according to them, [expr {2*2}] should be fine to include in an input line
18:59.02 *** join/#brlcad BigAToo (n=BigAToo@64.74.225.82)
18:59.08 brlcad yep
18:59.11 TC-Rucho but it keeps telling me error
18:59.19 brlcad only caveat is default globbing
18:59.25 TC-Rucho mged> in fuck.s rcc 0 0 0 [expr {2*2}] [expr {2*2*2}] 0 5
18:59.26 TC-Rucho Error: extra characters after close-brace
18:59.31 ``Erik * will mess up globbing
19:00.03 brlcad we have a globbing evaluation mode enabled by default so that things like "draw *.r" will work
19:00.34 brlcad which includes character classes, draw [a-z]*.c
19:00.48 brlcad so those conflict with tcl evaluation
19:00.54 TC-Rucho seems that I should disable globbing to input stuff, but still, is there a more elegant way?
19:00.55 brlcad you can have one or the other, default is globbing
19:00.58 ``Erik [expr {2\*2}] works, or you can turn off globbing
19:01.05 brlcad you can toggle it off with: set glob_compat_mode 0
19:01.12 brlcad or you can escape the globbing chars
19:01.37 brlcad in fuck.s rcc 0 0 0 \[expr {2\*2}\] \[expr {2\*2\*\2}\] 0 5
19:02.11 TC-Rucho I think it would be better to scape globbing rather than tcl expressions
19:02.26 TC-Rucho like draw \*.r
19:02.34 TC-Rucho and keep [expr {2*2}] working
19:02.37 ``Erik odd choice of primitive name O.o
19:02.55 brlcad probably a 'part' primitive ;)
19:03.12 brlcad ah, rcc, guess it's an approximation
19:03.28 ``Erik people are more apt to treat the geometry tree like a filesystem than an expression engine *shrug*
19:03.28 TC-Rucho ``Erik: heh, that's because this tcl expression thing was getting on my nerves
19:03.45 TC-Rucho lol
19:03.58 TC-Rucho @brlcad | ah, rcc, guess it's an approximation <--- lol xD
19:04.06 brlcad TC-Rucho: it's been oft-discussed, how to get the best of both worlds -- the folks that want globbing would cry bloody murder, and they're the expert modelers
19:04.06 ``Erik if you don't use globbing, turn it off, set that to be default in your .mgedrc *shrug* it's all good
19:04.49 brlcad those that know how to write tcl can easily just set the var and be done with it
19:05.50 TC-Rucho tcl == ugly mix of C and... bash? anyway, we'll have lisp bindings eventually and everything will be good
19:07.03 ``Erik I think tcl is older than bash, but it does take cues from the bourne/korn family
19:08.13 ``Erik (dang linux weenies, all not knowin' the history of it all :)
19:09.02 ``Erik huh, bash precedes tcl by a year, neat :)
19:10.20 TC-Rucho (:
19:18.32 brlcad TC-Rucho: awesome -- native bindings?
19:18.49 TC-Rucho brlcad: hm?
19:19.15 *** join/#brlcad hippieindamakin8 (n=hippiein@202.3.77.38)
19:21.06 TC-Rucho brlcad: well, my usability goal is to get a lisp repl for mged, so that one can use lisp directly to input commands, math and scripting
19:22.05 bitminer While were on the topic of TCL... any one have difficulty building lib Bacon Lettus and Tomato ... libBLT in windows. I am getting error : error LNK2019: unresolved external symbol _TclInitSubsystems referenced in function _TkConsoleCreate
19:22.30 bitminer Need me a BLT right now.
19:22.43 TC-Rucho .. bacon, lettus, tomato... that sounds like a salad
19:22.53 louipc sandwich
19:23.00 TC-Rucho right, sandwitch
19:23.03 bitminer It's more like C soup
19:24.01 brlcad TC-Rucho: right, but binding to mged commands how? you can bind straight to the lib, through the mged binary (ugh), through a custom interp, etc
19:24.30 brlcad would be awesome to have a swig interface to libged so there'd be bindings available in everything they support
19:24.36 bitminer it is externed via extern void TclInitSubsystems _ANSI_ARGS_((CONST char *argv0)); in tkConsole.c
19:24.59 bitminer wraped in some suspicous #ifdefs
19:25.14 brlcad bitminer: hm, you're compiling using the msvc build files or autotools?
19:25.36 bitminer Was trying autotools in Cygwin. I have had more luck in msvc
19:25.41 bitminer so msvc for now
19:26.00 TC-Rucho brlcad: I have not even started with that, what I've said is the 1st global usability goal. I'm just giving my first steps modelling with brlcad (you may have noticed already). Once I know it enough so as to make improvements I'll grab the source and/or make the lisp repl using bindings to libmged or whatever I see fits best
19:26.16 brlcad the msvc8 files should be the most up-to-date and are what are used to make releases
19:26.36 bitminer I have done some Swig in the past binding C# to C++ in linux allwoing Windforms apps wirtten in windwos to be run in Linux on Mono
19:26.51 brlcad the 9 files shouldn't be far behind, the other dir relies on cmake but only compiles the libs
19:27.19 bitminer I have noticed that the msvc files only support release
19:27.32 brlcad that's a bob-ism
19:27.39 brlcad he got tired of managing both
19:27.45 bitminer This was one of my first issues and MSVC defaults to debug on start up. I emailed this to the mailing list
19:28.12 bitminer bob-ism?
19:28.24 brlcad bob, one of the devs
19:28.38 brlcad he does a variety of quirky things :)
19:28.40 bitminer Ok
19:29.02 bitminer Are you using CMake to generate the VS project files?
19:29.04 brlcad doesn't talk much, just likes to quietly code
19:29.10 brlcad no
19:29.14 brlcad though that would be cool
19:29.26 bitminer It is a suposed rumored feature of CMake
19:29.33 brlcad yes, it is
19:29.45 brlcad and it works -- that's the other cmake-based build I mentioned that we have
19:29.52 brlcad it just doesn't build everything
19:30.28 brlcad our build system is pretty big.. takes many weeks/months to set everything up that we need with any build tool
19:30.37 bitminer Ok so I could possibly use Cygwin autotolls or CMake, Windows env Cmake, or Windows env Visual Studio
19:30.58 ``Erik hrm, here's a posting claiming that swig doesn't play with lisps (but does with scheme and erlang)
19:31.00 brlcad well yes and no, more no than yes
19:31.11 ``Erik though uffi/cffi around libged would be neat
19:31.17 brlcad bitminer: you can use autotools or the msvc8/msvc9 project files to get a full build
19:31.25 brlcad the cmake build files will only build the libraries
19:31.42 bitminer Ok got it thanks
19:31.43 brlcad we have like two dozen libraries
19:31.48 brlcad and over 400 binaries
19:32.22 bitminer holy cats
19:32.28 brlcad mged is just one of them :)
19:32.46 brlcad granted, it's the biggest by far
19:33.03 d-lo heh, binary envy.
19:33.23 brlcad most of that 400 are unix-style commands that do one thing and are streamable so you can tie them together for much more powerful functionality
19:34.01 brlcad ala cat | awk | sed | grep, ours focus on geometry, images, and data file processing
19:34.22 bitminer How do they communicate, MPI, sockets, carrier pigions?
19:34.27 starseeker unix pipes
19:34.35 starseeker bites us when working on Windows though
19:34.40 brlcad pipes and sockets
19:34.54 bitminer got it
19:35.06 bitminer Look a boost for IPC?
19:35.27 brlcad the SGI_Cube example on the wiki shows a couple commands being used for image processing/conversion
19:35.43 brlcad yeah, I've looked at it before .. what about it? :)
19:35.51 bitminer Inter Process Communication (IPC)
19:36.24 ``Erik yeah, we know what ipc is... :)
19:36.33 bitminer Most of the code looks to be C. What is your take on C++ in this project and Boost C++ and 0x support
19:37.11 ``Erik the work in rt^3 is using boost at the moment, and c++ is starting to creep into the codebase :/
19:37.15 brlcad most of the bigger project infrastructure for the new modeling interface is being done in C++ (using boost and stl heavily)
19:37.26 brlcad all the new brep/nurbs work is c++
19:37.43 bitminer So it is an option then?
19:38.11 brlcad all of our existing core libs are going to stay pure C (libbu, libbn, librt, libwdb, etc)
19:38.21 brlcad but new code, it's an option
19:38.37 bitminer OK. Will Boost compile on all your platforms?
19:39.10 brlcad it's just not cool to be half-assed about it, using little tidbits of C++ throughout a code that is 95% plain C isn't cool
19:39.16 brlcad just makes for bad C++ and bad C
19:39.34 brlcad yeah, most of boost compile's fine -- at least no issues so far
19:39.55 bitminer Yes I can understand wanting to maintain consistency.
19:39.56 ``Erik yeah, if someone commits c++ stuff in a .c file, I'll break their kneecaps O.o I'll try to be gentle about kneecap breaking if it's c99
19:39.58 ``Erik :D
19:40.07 brlcad the new parametric constraint solving system that madant has been working on is the closest to causing boost to snap, but it's been fine
19:43.52 brlcad ponders gsoc projects
19:44.12 bitminer Yes I think I read about this using Spirit? GSOC proj? for parametric modeling
19:44.18 brlcad yep
19:45.06 bitminer Well I'll keep banging on the Win32 build.
19:45.07 brlcad ``Erik: irix64 said he has a patch for you to review on his site if you're interested, dunno when he'll be back on
19:45.30 louipc hehe he pms me about the patches too
19:45.36 ``Erik ok
19:45.38 brlcad bitminer: send the error to the devel mailing list, To Bob or whomever ;)
19:45.39 louipc but he disappears before I can respond
19:45.46 ``Erik I'll keep an eye out for him
19:45.54 brlcad sounds like it might be something new -- tcl was just upgraded
19:46.11 brlcad might be easier to install activestate's tcl and then just change the linkage to use that instead of building
19:46.18 bitminer Ok will do.. There were multiple. I fixed most just stuck on libBLT at the moment.
19:46.22 brlcad louipc: really? jeez
19:46.32 brlcad louipc: let me know if it gets to be a problem
19:46.41 louipc it's all cool
19:46.49 brlcad k
19:47.20 brlcad starseeker: the tcl folks liked your archer screenie
19:47.33 ``Erik he needs a lot of steering and handholding, I'd hate to see his enthusiasm crushed, though :)
19:47.38 starseeker oh, the tire wizard? :-)
19:47.43 brlcad they (jokingly) said the gui looked dated, referring to ttk updates
19:48.14 brlcad but that the tire looked great ;)
19:48.17 ``Erik (does ttk mean that aquatk is now a dead-end)
19:48.20 starseeker would love to take a stab at using ttk
19:48.27 brlcad ``Erik: hardly
19:48.43 ``Erik <-- has managed to avoid that entire chunk for the most part
19:49.16 ``Erik I made a button in wish once, otherwise I just imitate code already there for my patches and hope I guessed right :)
19:50.22 starseeker had to fight the busting of Archer with the lastest tcl/tk as an excuse to take Archer apart and put it back together using ttk widgets
19:50.39 starseeker er, fight using it as an excuse
19:54.14 starseeker figures Archer is having enough trouble without me messing with it like that ;-)
19:56.51 starseeker brlcad: Are you close to tagging for the release?
20:04.22 *** join/#brlcad Twingy (n=justin@74.92.144.217)
20:11.19 ``Erik reboots brlcad's new machine again
20:11.32 starseeker ``Erik: is that the new bz server?
20:11.51 ``Erik yeah
20:12.18 ``Erik I'm being aggressive about keeping it bleeding edge until everything is migrated to it, so we have a good launch point
20:12.26 ``Erik bleeding stable edge, that is
20:12.35 starseeker ah
20:12.51 starseeker should migrate his stuff to that one and free up some space
20:13.37 ``Erik and it's back up
20:13.57 starseeker ``Erik: At some point can you show me how to get to that box?
20:14.00 ``Erik looks like it has an 80 gig drive in it
20:14.09 starseeker er, well nevermind :-)
20:14.25 starseeker will just suck it up and get a terabyte drive
20:14.33 ``Erik hm, I don't see you in the passwd file
20:14.41 d-lo I thought it was supposed to have >80
20:15.13 ``Erik oh, wait, 120g
20:15.41 ``Erik that's odd, there may be an unallocated partition
20:18.35 ``Erik aah, reading it wrong, 80g for home
20:18.49 ``Erik refers back to where he stated his brain would not be functioning today
20:25.43 starseeker was dbbinary the one that got renamed to bo?
20:26.26 starseeker checks NEWS
20:29.15 starseeker ah, yes
20:39.35 *** join/#brlcad ibot (i=ibot@rikers.org)
20:39.35 *** topic/#brlcad is BRL-CAD Open Source Solid Modeling || http://brlcad.org || http://sf.net/projects/brlcad || GSoC 2008 Highlight: new prototype gui, check it out! || Source Release 7.14.2 is posted (20080207)
20:41.17 CIA-40 BRL-CAD: 03starseeker * r33932 10/brlcad/trunk/doc/docbook/system/man1/en/ (9 files): Add more MGED docbook man pages by Janine and Cliff
20:59.39 *** join/#brlcad ibot (i=ibot@rikers.org)
20:59.39 *** topic/#brlcad is BRL-CAD Open Source Solid Modeling || http://brlcad.org || http://sf.net/projects/brlcad || GSoC 2008 Highlight: new prototype gui, check it out! || Source Release 7.14.2 is posted (20080207)
21:12.56 *** join/#brlcad cad33 (n=424ca4ee@bz.bzflag.bz)
22:24.34 TC-Rucho hey guys, if you like to do inline math just like me and hate that \[expr {mathstuff}\] gay sh*t, you can type in => \[proc unknown args {set ::that \[expr $args\]}\] so you can just do \[mathstuff\] (:
22:25.37 TC-Rucho this works by setting the unknown handler to perform an [expr ...] to the screwed up input (math without preceding expr)
22:25.44 TC-Rucho hope you like it
22:26.35 TC-Rucho note: and desabling globbing it get's even better
22:28.02 TC-Rucho brlcad: should I add this to the brlcad-wiki as a tip?
22:28.30 *** join/#brlcad samrose (n=samrose@c-24-11-214-181.hsd1.mi.comcast.net)
22:34.41 *** join/#brlcad hippieindamakin8 (n=hippiein@202.3.77.38)
22:40.51 *** join/#brlcad BigAToo (n=BigAToo@pool-96-230-124-39.sbndin.btas.verizon.net)
23:03.58 *** join/#brlcad tc-rucho1 (n=tc-rucho@190.191.172.28)
23:05.02 *** join/#brlcad tofu (n=sean@bz.bzflag.bz)
23:07.56 *** join/#brlcad tofu (n=sean@bz.bzflag.bz) [NETSPLIT VICTIM]
23:07.56 *** join/#brlcad hippieindamakin8 (n=hippiein@202.3.77.38) [NETSPLIT VICTIM]
23:07.56 *** join/#brlcad bitminer (n=bitminer@h96-60-82-113.vrnawi.dsl.dynamic.tds.net) [NETSPLIT VICTIM]
23:07.56 *** join/#brlcad mafm (n=mafm@223.Red-83-49-86.dynamicIP.rima-tde.net) [NETSPLIT VICTIM]
23:07.56 *** join/#brlcad Axman6 (n=Axman6@pdpc/supporter/student/Axman6) [NETSPLIT VICTIM]
23:07.56 *** join/#brlcad Ralith (n=ralith@216.162.199.202) [NETSPLIT VICTIM]
23:07.56 *** join/#brlcad smurfette (n=user@c-69-242-189-29.hsd1.mo.comcast.net) [NETSPLIT VICTIM]
23:07.56 *** join/#brlcad archivist (n=archivis@host81-149-119-172.in-addr.btopenworld.com) [NETSPLIT VICTIM]
23:07.56 *** join/#brlcad d-lo (n=claymore@bz.bzflag.bz) [NETSPLIT VICTIM]
23:07.56 *** join/#brlcad MinuteElectron (n=MinuteEl@unaffiliated/minuteelectron) [NETSPLIT VICTIM]
23:07.56 *** join/#brlcad alex_joni (n=juve@emc/board-of-directors/alexjoni) [NETSPLIT VICTIM]
23:07.56 *** join/#brlcad yukonbob (i=1000@s142-179-54-198.bc.hsia.telus.net) [NETSPLIT VICTIM]
23:09.27 *** join/#brlcad elite01 (n=omg@cl-213.dus-01.de.sixxs.net)
23:11.06 *** join/#brlcad IriX64 (n=IriX64@bas2-sudbury98-1177871980.dsl.bell.ca)
23:27.23 *** mode/#brlcad [+o tofu] by ChanServ
23:27.50 ``Erik *yawn*
23:28.21 ``Erik bar commercials on tv heh
23:28.32 elite01 some movie on vlc heh
23:29.04 ``Erik I was laughing at the existance of, not just commenting on what I was seeing :D vlc is good stuff, though
23:32.57 *** join/#brlcad bitminer69er (n=bitminer@h96-60-82-113.vrnawi.dsl.dynamic.tds.net)
23:33.05 brlcad tc-rucho1: hah, that's a really fantastic way to abuse tcl
23:36.21 brlcad great hack/tip
23:57.18 ``Erik \[mathstuff\] makes me think LaTeX
23:57.55 ``Erik that'd be awesome if tcl were kicked to understand tex math mode, one step closer to literate programming :D
23:58.35 ``Erik (or generate tex from the tcl)

Generated by irclog2html.pl Modified by Tim Riker to work with infobot.