| 00:00.45 | milamber | hello all, are step files something that mged/archer are able to handle? i checked the documentation here: http://brlcad.org/wiki/STEP_Libraries it seems like that is asking for the code to be cleaned up, but the way i interpret that is that the functionality is there. |
| 00:33.13 | brlcad | ``Erik: displacement |
| 00:35.37 | brlcad | milamber: we have a step-g importer tool that will import some step content |
| 00:37.02 | brlcad | namely solid nurbs boundary representation geometry (what most solid modeling systems will export) |
| 00:41.48 | milamber | brlcad: this is the output that i get when i try and run the file through the step-g converter. http://pastebin.com/dcXgPN1A |
| 00:43.07 | milamber | brlcad: and for what it is worth, neither archer nor mged handle it gracefully. archer just crashes and mged says something about malloc and exits |
| 00:49.21 | brlcad | milamber: ah, what format is that .step file? |
| 00:49.36 | brlcad | looks like it might be an AP214 schema file |
| 00:49.43 | brlcad | we support AP203 |
| 00:49.56 | brlcad | (214 is coming later) |
| 00:50.51 | milamber | brlcad: thanks, i can't seem to find anything for linux that will open it. thanks for your time. |
| 00:51.15 | brlcad | milamber: what does the file header say it is? |
| 00:53.39 | milamber | brlcad: ISO-10303-21; |
| 00:53.41 | milamber | HEADER; |
| 00:53.41 | milamber | FILE_DESCRIPTION (( 'STEP AP214' ), |
| 01:13.12 | brlcad | yeah, so that's the problem |
| 01:13.27 | brlcad | it "should" work, or at least we'll eventually get to the point where it does work |
| 01:13.55 | brlcad | but currently, our parser is wired to read STEP AP203 which is why it complains about all those missing/invalid entities |
| 01:14.16 | brlcad | so it ends up with a corrupt .g (which obviously shouldn't happen either) |
| 01:14.39 | brlcad | we have a LOT of STEP work going on this summer, so hopefully we can improve that then |
| 01:15.22 | milamber | sounds good. |
| 01:15.32 | brlcad | what generated the file? |
| 01:15.38 | milamber | i would be happy to beta that stuff if you are looking for volunteers |
| 01:15.46 | brlcad | you should be able to regenerate an ap203 with whatever output the 214 file |
| 01:15.54 | brlcad | sure |
| 01:16.04 | milamber | not sure, it's a file for an ar15 lower |
| 01:16.47 | brlcad | you're not sure how you got the file? |
| 01:16.57 | milamber | http://defcad.org/defdist-ar-15-lower-receiver/ |
| 01:17.05 | milamber | i'm not sure what program created the file |
| 01:17.20 | brlcad | ah, but you didn't, that's the important part |
| 01:17.22 | milamber | i did not |
| 01:17.36 | brlcad | has their whole data set |
| 01:18.26 | brlcad | in theory, we could regenerate our importer with the merged schema and it should read the 214 file better... |
| 01:18.32 | brlcad | maybe a great gsoc task |
| 01:19.39 | milamber | the igs files that are there freecad seems to be able to handle, but freecad won't open this step format either |
| 02:25.58 | brlcad | milamber: does http://www.ida-step.net/components/viewers/basic read the file for you? |
| 02:35.44 | Notify | 03BRL-CAD:brlcad * 54952 brlcad/trunk/src/librt/primitives/bot/tie_kdtree.c: bu_log supports %zu for size_t |
| 02:43.32 | milamber | brlcad: negative - http://pastebin.com/ikX7kKMk |
| 02:47.42 | Notify | 03BRL-CAD:brlcad * 54953 brlcad/trunk/src/util/bwcrop.c: need signed comparison now to match buflines |
| 02:48.37 | *** join/#brlcad milamber (~devlin@d118-75-244-176.try.wideopenwest.com) |
| 03:14.33 | Notify | 03BRL-CAD:brlcad * 54954 brlcad/trunk/src/librt/primitives/arb8/arb8.c: quell gcc false positive due to the parameter getting set within the expression. initialize explicitly for good measure anyways. |
| 03:16.48 | Notify | 03BRL-CAD:brlcad * 54955 brlcad/trunk/src/librt/primitives/arb8/arb8.c: ws indent style cleanup |
| 03:17.33 | brlcad | milamber: what's on line's 1-10 ? |
| 03:17.46 | brlcad | pastebin.ca |
| 03:18.09 | brlcad | (many channel listeners cannot get to pastebin.com (and they suck)) |
| 03:21.21 | Notify | 03BRL-CAD:brlcad * 54956 brlcad/trunk/src/librt/primitives/nmg/nmg.c: rt_nmg_import4_fastf takes a const matp_t |
| 03:21.52 | milamber | http://pastebin.ca/2344562 |
| 03:28.59 | Notify | 03BRL-CAD:brlcad * 54957 brlcad/trunk/src/librt/primitives/nmg/nmg.c: quell warning about preserving constness to a matp_t parameter by casting _before_ passing. |
| 03:30.50 | brlcad | milamber: interesting |
| 03:31.51 | brlcad | that indeed looks like an invalid FILE_NAME line |
| 03:31.57 | brlcad | it's missing a slew of parameters |
| 03:32.03 | brlcad | try changing it to this: |
| 03:35.16 | brlcad | FILE_NAME ('AR-15 - Lower Receiver DD - WW REV 3.3.1-Test 1 (jad).STEP', '2013-01-01T01:01:01', ( 'Whatever' ), ( 'Yeah' ), 'Sean', 'BRL-CAD', '' ); |
| 03:40.30 | milamber | trying . . . |
| 03:42.49 | milamber | ida-step is "triangulating" - which is better than just throwing an error like it was previously doing |
| 03:44.43 | brlcad | progress |
| 03:51.24 | milamber | file loaded in ida-step |
| 03:58.24 | milamber | have to get going for the evening, but i appreciate the help |
| 04:25.51 | brlcad | milamber: glad to help |
| 04:41.59 | *** join/#brlcad merzo (~merzo@239-76-133-95.pool.ukrtel.net) |
| 04:51.14 | *** join/#brlcad Skriptkid (~Skriptkid@59.93.68.250) |
| 04:54.28 | Notify | 03BRL-CAD:brlcad * 54958 brlcad/trunk/src/libbu/affinity.c: need to explore the portability of this, but go ahead and locally override the _GNU_SOURCE setting so that we can enable processor affinity for pthread systems. after much headache, figured out that _GNU_SOURCE needs to be defined before the common.h header. common includes some system headers and the features.h header is only going to get read once to set |
| 04:54.30 | Notify | the necessary __USE_GNU define and declaring the CPU_* macros and pthread affinity funcs the implementation uses. this is going to jack up linux benchmarks, but it's a nice 5-50% improvement on even small smp systems. |
| 04:57.08 | Notify | 03BRL-CAD:brlcad * 54959 brlcad/trunk/src/libged/tables.c: quellage, go through libbu and a vls to print size_t values to file. |
| 05:23.07 | Notify | 03BRL-CAD:brlcad * 54960 brlcad/trunk/CMakeLists.txt: need to check for a few headers to support thread affinity on bsd. check for pthread_np.h and sys/cpuset.h headers along with the cpu_set_t type (which is linux, bsd is cpuset_t but allegedly may switch to match). |
| 05:23.33 | brlcad | wow, looks like my earlier affinity numbers weren't the whole picture |
| 05:23.48 | brlcad | had reported vgr count going from 120k to 180k |
| 05:24.04 | brlcad | but just tested again now and went from 115k to 290k |
| 05:25.05 | brlcad | so highly dependent on how busy the machine is, but 2.5x faster is just wow |
| 05:26.44 | Notify | 03BRL-CAD:brlcad * 54961 brlcad/trunk/src/libbu/affinity.c: expand support to freebsd, just a few changes needed |
| 05:46.03 | brlcad | ``Erik: you need to create a link_id |
| 05:46.19 | brlcad | need to register your profile |
| 05:50.07 | Notify | 03BRL-CAD:brlcad * 54962 brlcad/trunk/src/libbu/affinity.c: just because we have pthreads doesn't mean we have the cpuset macros, e.g., mac os x. |
| 05:51.51 | Notify | 03BRL-CAD:brlcad * 54963 brlcad/trunk/src/libged/tables.c: pass const |
| 05:53.33 | Notify | 03BRL-CAD:brlcad * 54964 brlcad/trunk/src/libged/tables.c: fastf_t arrays need a little more const persuasion |
| 09:15.16 | *** join/#brlcad jordisayol (~jordisayo@unaffiliated/jordisayol) |
| 10:00.02 | *** join/#brlcad luca79 (~luca@host201-12-dynamic.13-87-r.retail.telecomitalia.it) |
| 10:10.15 | ``Erik | cache miss will kill ya (I'd imagine the 2 core result would also be huge) |
| 10:11.40 | ``Erik | will my old 'erikg' id work? (it all seems to be there) |
| 11:50.20 | *** join/#brlcad ncsaba (~ncsaba@p54983F7E.dip.t-dialin.net) |
| 11:50.50 | *** join/#brlcad merzo (~merzo@231-205-132-95.pool.ukrtel.net) |
| 11:51.07 | ncsaba | Hi all |
| 11:51.23 | ncsaba | is there any human around I can chat with about BRL-CAD ? |
| 11:59.15 | *** join/#brlcad luca79 (~luca@host17-82-dynamic.14-87-r.retail.telecomitalia.it) |
| 12:01.46 | *** join/#brlcad jordisayol (~jordisayo@unaffiliated/jordisayol) |
| 12:06.00 | ``Erik | nope, just us troglodytes |
| 12:31.50 | ncsaba | Hi Erik, sorry, didn't meant it like that, rather though about bots :-) |
| 12:32.30 | ncsaba | look like I'm just not patient enough |
| 12:35.06 | ncsaba | I last used IRC about 13 years ago, and then on some high traffic channels and probably got used to instant answer... which is probably not reasonable on a channel like this |
| 12:36.27 | ``Erik | depends on the time of day, most of us are in the US, so gmt-5 to gmt-8... it's still early :) |
| 12:36.42 | ncsaba | ok, that's a good point too |
| 12:37.30 | ncsaba | well I'm in Europe, gmt+1 |
| 12:38.10 | ncsaba | but I'm glad I found somebody I can talk to :-) |
| 12:38.56 | ``Erik | so are you looking for help doing something with BRL-CAD, trying to decide if it's worth checking out, or ? |
| 12:39.22 | ncsaba | well I just discovered it recently that it exists - and I like it's basic philosophy |
| 12:39.42 | ncsaba | but need some things which it doesn't have - ready to program too if needed |
| 12:39.43 | ``Erik | cool |
| 12:40.21 | ``Erik | what kinda things are you looking for? (there're lots of capabilities that are poorly advertised) |
| 12:40.24 | ncsaba | what I'm actually looking for is discussion - I don't know if the things I need are not already worked on, reasonable at all, or so |
| 12:41.16 | ncsaba | well right now I'm working on some models which use extensively pipes, |
| 12:41.36 | ncsaba | and I want to create some parameterized objects with those pipes |
| 12:41.52 | ncsaba | and used a TCL script to do it |
| 12:41.57 | ``Erik | like shrinkwrapping them or something? |
| 12:43.04 | ncsaba | well I create a fork-like structure out of pipe, and want to parameterize the lengths of the fork sections |
| 12:43.14 | ncsaba | it works - I managed to do it |
| 12:43.19 | ncsaba | that's not the problem :) |
| 12:43.21 | ``Erik | using the pipe meta-primitive? |
| 12:43.25 | ncsaba | yes |
| 12:43.58 | ``Erik | cool, that can be a weird primitive to use... it builds the 'pipe' out of cylinders and torii |
| 12:44.05 | ncsaba | what I found strange is that i couldn't programmatically edit the points in the pipe |
| 12:44.15 | ncsaba | from TCL |
| 12:44.42 | ``Erik | hm, they can be edited with the sed command, I believe |
| 12:44.53 | ncsaba | the "philosophy" part of the idea: whatever can be done from the GUI, should be possible from the scripting part ? |
| 12:44.58 | ncsaba | not really |
| 12:45.09 | ncsaba | or perhaps I'm missing something |
| 12:45.10 | ``Erik | (sed is solid editor, not the unix stream editor) |
| 12:45.28 | ``Erik | all gui components just call a tcl callback, yes |
| 12:45.35 | ncsaba | ok, so I put the object in edit mode with sed |
| 12:45.48 | ncsaba | then how do I tell that I want to add a point ? |
| 12:46.20 | ncsaba | the GUI's menu updates with pipe specific commands |
| 12:46.32 | ``Erik | I'd have to dig into the code, I tend to avoid gui stuff myself, let me take a quick look |
| 12:46.37 | ncsaba | but what's the corresponding comamnd in the scripting then ? |
| 12:46.50 | ncsaba | well if you tell me where to look, I can do that myself too ? |
| 12:46.52 | ncsaba | :) |
| 12:47.24 | ``Erik | well, in the tclscripts for mged, there should be the tk menu entry to add the point, just look at the function it calls |
| 12:47.30 | ncsaba | BTW, I'm using mged, not archer - that was the first thing I found documented when installing BRL-CAd, so I got used to that |
| 12:47.41 | ncsaba | ok |
| 12:47.58 | ``Erik | archer is an alpha stage 'next gen' interface, mged is still the 'officially supported' interface |
| 12:48.09 | ncsaba | aha, then it's fine I guess... |
| 12:48.29 | ``Erik | nice, I can't even start mged due to a recent boost thread change, heh |
| 12:48.41 | ncsaba | :-) |
| 12:49.16 | *** join/#brlcad indianlarry (~indianlar@66-118-151-70.static.sagonet.net) |
| 12:50.48 | ncsaba | ok, I'm in the tclscripts/mged dir - any hint what I should look for ? I admit I have no experience with TK |
| 12:50.52 | ``Erik | indianlarry probably knows the tcl command to add a pipe element O.o |
| 12:51.50 | ncsaba | well frankly I would prefer to figure out the TK links, that's like difference between getting a fish and learning how to catch it ;-) |
| 12:53.13 | ``Erik | 'k, looks like the adjust command is what you want |
| 12:53.44 | ``Erik | just reference an index one past the last and it automatically adds it to the list |
| 13:00.33 | Notify | 03BRL-CAD:bob1961 * 54965 brlcad/trunk/src/libged/mater.c: Putting back the double-quotes around "light invisible=1" so that the example is once again correct. Will fix the double-quote issue on the Tcl side. |
| 13:12.21 | ncsaba | thanks Erik - I'm looking in the docs, plus the src/tcslscripts/mged/edit_solid.tcl, and I suppose the "db adjust ..." command is what I need to use there ? |
| 13:27.34 | brlcad | ``Erik: you have to set up your profile (basically confirm your address, the terms, etc) |
| 13:27.58 | ncsaba | Managed to do it with "db adjust..." :-) |
| 13:28.25 | brlcad | ncsaba: we're here all the time and quite active, but not always responsive ;) |
| 13:28.42 | brlcad | depends on the time of day, how many of us have are heads down in code at the moment, etc |
| 13:28.54 | ncsaba | ok, no problem - I just need to know that :-) |
| 13:29.24 | ncsaba | the first impression was a ghost town, but as said, I was used to busy chat channels... |
| 13:29.29 | brlcad | we have tons of projects and welcome the activity |
| 13:29.58 | brlcad | it fluctuates |
| 13:30.03 | ncsaba | OK, regarding activity - what I'm missing in BRL-CAD are helper constructs to set up constraints between objects |
| 13:30.08 | brlcad | sometimes it's non-stop streaming |
| 13:30.12 | brlcad | sometimes it'll be quiet for hours |
| 13:30.25 | brlcad | really depends what a dozen different people are doing, what time of day, what time of year |
| 13:30.36 | ncsaba | ok, understood :-) |
| 13:30.38 | brlcad | but yeah, not 100's of people, so a lil diff |
| 13:30.51 | brlcad | we certainly welcome the interest ;) |
| 13:31.03 | ncsaba | I'm busy too normally - today I have public holiday here in Germany |
| 13:31.15 | ncsaba | so I can work on my hobby stuff :-) |
| 13:31.25 | brlcad | so here's some basic docs to help organize thoughts |
| 13:32.35 | ncsaba | do any of you tried VARKON ? |
| 13:32.36 | brlcad | not entirely useful to an individual developer since devs can work on whatever you like, but the overall "big picture" priorities are http://brlcad.org/BRL-CAD_Priorities.png |
| 13:32.42 | *** join/#brlcad caen23 (~cezar@92.85.92.51) |
| 13:33.08 | ncsaba | ok, I've seen already that :-) |
| 13:33.17 | brlcad | cool |
| 13:33.28 | brlcad | and the Deuces and Quickies pages? |
| 13:33.39 | ncsaba | as said, I like the basic philosophy behind BRL-CAD |
| 13:33.47 | ncsaba | that one not sure |
| 13:33.51 | brlcad | they're sort of 1-day and 1-week scoped projects |
| 13:33.55 | brlcad | just some ideas |
| 13:34.36 | brlcad | I keep a list of bigger projects (multiweeks to multiyear) here: http://brlcad.org/~sean/ideas.html |
| 13:35.11 | brlcad | http://brlcad.org/wiki/Deuces and http://brlcad.org/wiki/Quickies are the aforementioned |
| 13:35.37 | brlcad | ``Erik: the boost stuff is undone |
| 13:37.23 | brlcad | ncsaba: and you are right, philosophy is everything from the GUI should be doable from the command line (and is as far as I know) |
| 13:38.24 | brlcad | if you try to go after one of the short tasks, might ask about it first as a lot of them were recently completed by students in a competition |
| 13:38.30 | brlcad | have to update the pages |
| 13:39.18 | ncsaba | ok, I'm looking at the tasks page - what about suggestions about things which are not there ? |
| 13:39.35 | ncsaba | have you heard about VARKON ? |
| 13:39.58 | ncsaba | that's a 3D CAD software too, which I also tried |
| 13:40.43 | ncsaba | I didn't like it that much (mostly due to their decision to create a new language for the scripting), |
| 13:41.03 | ncsaba | but it has one nice feature: |
| 13:41.12 | brlcad | have followed pretty much every open source CAD-related softare that's been put on a page in the last 10 years :) |
| 13:41.26 | ncsaba | ok :-) |
| 13:41.55 | brlcad | varkon's an oldie |
| 13:42.12 | ncsaba | so, in VARKON I liked the possibility to get points on primitives using parametric selection |
| 13:42.52 | ncsaba | like get the point say 0.4xlength on a segment |
| 13:43.01 | ncsaba | or on an arc |
| 13:43.07 | brlcad | nods |
| 13:44.03 | brlcad | one of the "problems" with our current primary representation format (implicit geometry) is surface boundary selection |
| 13:44.05 | ncsaba | or get a tangent to a curve on a point |
| 13:44.27 | brlcad | because the surfaces are implicit, a lot of surface-based operations become super hard or impractical |
| 13:44.48 | brlcad | which is why we've spent the last few years implementing support towards becoming fully hybrid representation |
| 13:45.07 | ncsaba | ok, what about providing with the solids also some predefined key 2D shapes / points which can be selected ? |
| 13:45.11 | brlcad | implicit and brep/nurbs under the hood, so we can get access to surface features without loosing geometry detail |
| 13:45.32 | ncsaba | like a cilinder always has a center axis |
| 13:45.37 | brlcad | yep |
| 13:45.54 | brlcad | that's on our list of things to do |
| 13:46.11 | ncsaba | ok, that's the direction I would like to help out then if possible |
| 13:46.15 | brlcad | make every primitive define a set of parameters that can be selected/edited |
| 13:46.41 | brlcad | to some extent that happens already, but it happens up in tcl land in the gui itself (year, really bad) |
| 13:47.05 | brlcad | the idea is to have each primitive define their set of parameters (and constraints) intrinsic to them |
| 13:47.18 | brlcad | then the gui just probes them for what to allow for selection/editing |
| 13:47.49 | ncsaba | yes, cool |
| 13:48.13 | brlcad | working on that quickly gets into the center of our geometry kernel |
| 13:48.35 | ncsaba | aha |
| 13:48.44 | brlcad | so that's a great area to start, but will have to maybe work towards it from some feature perspective |
| 13:48.56 | brlcad | like creating a new command-line command that just lists parameters |
| 13:49.29 | ncsaba | I'm coming from a java background, just to mention it |
| 13:49.35 | brlcad | if you understand that layer (simple C api), getting into the callbacks and primitive functions will be easy |
| 13:49.51 | brlcad | okay |
| 13:50.27 | ncsaba | but I have no problems with other languages, even if I prefer the object oriented ones :-) |
| 13:50.54 | brlcad | no worries |
| 13:51.06 | brlcad | our nurbs work is all OO C++ |
| 13:51.19 | ncsaba | ok |
| 13:51.24 | brlcad | our core geometry is heavily OO C |
| 13:51.39 | brlcad | (there is such a beast) ;) |
| 13:52.30 | ncsaba | yes, I imagine - I tried to apply OO design to MS foxpro back on my programming beginnings - it kind of works but everything has to be done manually :-) |
| 13:52.51 | brlcad | yeah, so there's a lot of that |
| 13:53.15 | ncsaba | I guess you have a defined interface for what primitives need to provide ? |
| 13:53.24 | brlcad | if you look in src/librt/primitives, you'll see each primitive is in a subdir where each defines a set of callbacks (which are itemized in src/librt/primitives/table.c) |
| 13:53.32 | ncsaba | ok |
| 13:53.52 | brlcad | yeah, that's all private API, but it's the rt_functab |
| 13:54.24 | brlcad | we have several maintenance tasks to encapsulate and push more down to that layer, but it's pretty easy to follow |
| 13:55.50 | brlcad | before you dive in too far there, though, I'd suggest starting with something you can wrap your head around |
| 13:56.04 | ncsaba | will have a look, but for the moment I think there's something else |
| 13:56.11 | brlcad | there's a LOT of code, so it's really easy to get lost |
| 13:56.14 | ncsaba | yes, exactly my thought :)= |
| 13:56.26 | ncsaba | but good to know |
| 13:56.49 | ncsaba | ok, then another question: is there a way already to get the length of a pipe ? |
| 13:57.05 | ncsaba | if not that is something I would add - I need it :-) |
| 13:57.10 | brlcad | I've been pushing to get all of our comments and docs and api cleaned up so it's easier to navigate (and remove some of our legacy complexity), but as you can imagine that takes a lot of time and effort across 1M lines of code |
| 13:57.26 | brlcad | length of a pipe, hmm |
| 13:57.38 | brlcad | i'm not sure, lemme check |
| 13:57.48 | ncsaba | just a function pipe_length(pipe_object) |
| 13:58.17 | brlcad | I highly doubt it exists |
| 13:58.30 | ncsaba | OK, then that's a good task for starters I guess |
| 13:58.36 | brlcad | I mean what exactly would you measure? the centerline? exterior line? inner line? |
| 13:59.58 | ``Erik | brlcad: updated, link id erikg |
| 14:00.03 | brlcad | yeah, I don't see anything |
| 14:01.09 | brlcad | linear pipe segment lengths are recorded, but I don't see anything for overall length (the places to look would be in src/libged (e.g., edpipe.c), src/librt/primitives/pipe, and src/tclscripts/mged (unlikely)) |
| 14:01.20 | ncsaba | I would measure the center-line, that's the best approximation of how much material I need for a pipe |
| 14:01.34 | brlcad | ncsaba: that does sound like a great function |
| 14:01.46 | brlcad | and could be exposed to users via the 'analyze' command |
| 14:01.54 | ncsaba | ok |
| 14:02.11 | brlcad | analyze reports calculations on objects in a tabulated text output |
| 14:02.27 | brlcad | can give it a try in mged to see what it outputs, src is src/libged/analyze.c |
| 14:02.44 | ncsaba | ok, I'll try right now |
| 14:02.53 | ``Erik | be neat to some day have a 'generate materials list' utility |
| 14:02.54 | brlcad | that hooks into librt (and is one of those prime places needing cleanup, to push logic into librt) |
| 14:03.19 | ncsaba | analyze: unable to process ID_PIPE solid |
| 14:03.40 | brlcad | ncsaba: so maybe start with adding a func (rt_pipe_length()) to src/librt/primitives/pipe and calling it from analyze |
| 14:03.44 | brlcad | well there ya go ;) |
| 14:03.51 | brlcad | not even hooked up to that primitive yet |
| 14:03.56 | ncsaba | good :-) |
| 14:04.01 | brlcad | make tgc tgc ; analyze tgc |
| 14:04.16 | brlcad | or make arb8 arb8 ; analyze arb8 for some examples |
| 14:04.35 | ncsaba | yep, got it |
| 14:05.05 | ``Erik | woohoo analyze: unable to process ID_METABALL solid |
| 14:05.27 | brlcad | ``Erik: yeah, that's one of those that really needs to get pushed down into the functab |
| 14:05.37 | ncsaba | ok, so this is a good area to work on |
| 14:05.40 | brlcad | otherwise people writing new prims like yours don't even know it exists |
| 14:06.00 | brlcad | basically anywhere there's a switch statement over prim types should be pushed |
| 14:06.06 | brlcad | ncsaba: yeah, great area |
| 14:06.28 | brlcad | as soon as you get a tiny thing working, I'd suggest submitting it as a patch |
| 14:06.38 | ncsaba | sure |
| 14:06.59 | brlcad | the smaller the better, not worrying about functionality so much as "correctness" and format |
| 14:07.11 | ncsaba | I need to get brl-cad compile first, I have old Ubuntu and some library version problems |
| 14:07.25 | brlcad | getting that sorted out helps get commit access set up, read HACKING if you haven't when you're ready |
| 14:07.32 | ncsaba | now I use the precompiled stable version (and looking in trunc sources) |
| 14:07.42 | brlcad | nods |
| 14:08.02 | ``Erik | interesting, my ubuntu box seems to have cpu_set_t instead of cpuset_t |
| 14:08.36 | brlcad | ``Erik: did it not detect correct? |
| 14:08.42 | brlcad | linux uses cpu_set_t |
| 14:08.52 | brlcad | bsd uses cpuset_t |
| 14:08.59 | ``Erik | guess it didn't detect, then |
| 14:09.13 | brlcad | oh, did you rerun cmake? |
| 14:09.36 | brlcad | someone pointed out the difference to the bsd guys and they said they can change to match linux |
| 14:09.43 | brlcad | apparently posix didn't define the type |
| 14:09.48 | ``Erik | yeah, but I'll force it again |
| 14:10.06 | brlcad | the test may have failed for other reasons |
| 14:10.24 | brlcad | it's a simple: BRLCAD_TYPE_SIZE("cpu_set_t" "sched.h") |
| 14:10.30 | brlcad | but other foo may be needed |
| 14:10.35 | ``Erik | I did the automagic rerun earlier, I'm doing a forced rerun right now |
| 14:11.21 | brlcad | ncsaba: just a word of caution, gsoc2013 is about to begin and if we're accepted (it'll be our sixth year), it's going to get a bit crazy in here |
| 14:11.55 | brlcad | ~seen skriptkid |
| 14:12.01 | ibot | skriptkid <~Skriptkid@117.202.103.23> was last seen on IRC in channel #brlcad, 30d 23h 2m 9s ago, saying: 'Ahh..was looking forward to meeting you too :)'. |
| 14:12.16 | ncsaba | aha |
| 14:12.26 | ``Erik | not quite http://www.youtube.com/watch?v=fpivIYJFjd4 crazy, but close :D |
| 14:13.21 | ncsaba | well I can't normally hang out here as long as today, but will get back regularly |
| 14:13.50 | ncsaba | what is the best time to join here on the IRC ? |
| 14:13.55 | brlcad | no worries |
| 14:14.41 | ncsaba | here's 15:15 now - workdays I can get in ~20:00 my time (gmt+1) |
| 14:14.56 | ``Erik | that's a pretty good time |
| 14:14.59 | ncsaba | ok |
| 14:15.35 | ncsaba | I'll stay around now for a while, I need to get BRL-CAD compiled |
| 14:15.42 | ncsaba | maybe I need some help there |
| 14:15.56 | ``Erik | use cmake, not automake |
| 14:16.17 | ncsaba | CMake Error at CMakeLists.txt:97 (cmake_minimum_required): |
| 14:16.17 | ncsaba | <PROTECTED> |
| 14:16.23 | brlcad | really any time usually, but 13:00 UTC to 7:00 UTC is when I'm usually around |
| 14:16.25 | ncsaba | old ubuntu :) |
| 14:17.03 | brlcad | one of the devs is in germany, so he's obviously a closer sync to you |
| 14:17.12 | brlcad | a few others are in india, so they'll be on the other tail end |
| 14:17.24 | ncsaba | ok |
| 14:17.33 | brlcad | fortunately cmake is really easy to compile ;) |
| 14:18.03 | ``Erik | sudo apt-get install cmake |
| 14:18.05 | ``Erik | boom, done |
| 14:18.07 | *** join/#brlcad merzo_ (~merzo@231-205-132-95.pool.ukrtel.net) |
| 14:19.10 | ncsaba | well the system's cmake is old - apt-get is not really helping, I will need to either compile or get an ubuntu package from a newer verison |
| 14:19.50 | ``Erik | hm, might have to do 'sudo apt-get update', been a while... I use aptitude on my ubuntu box for managing versions/upgrades/etc |
| 14:20.10 | ncsaba | eventually I will have to upgrade ubuntu, but it will be a pain - I'm running on a VM and unitiy is not playing well there |
| 14:20.59 | ``Erik | if you're on windows and have visual studio, you can grab the windows cmake and do that |
| 14:21.55 | ncsaba | my windows is only there because my employer forces me to have one - everything really useful on it is still linux :-) |
| 14:29.34 | brlcad | ncsaba: if you haven't seen it, some more big picture info at http://www.ohloh.net/p/brlcad |
| 14:29.49 | brlcad | the in a nutshell and activity sections would have probably answered your mailing list question ;) |
| 14:31.03 | ncsaba | perhaps... I'm kind of the impatient type unfortunately |
| 14:31.32 | ncsaba | unpatien |
| 14:31.33 | ncsaba | t |
| 14:31.36 | ncsaba | well |
| 14:32.26 | brlcad | :) |
| 14:32.56 | ``Erik | can't even wait to finish the word before hitting return, sheesh :D |
| 14:34.02 | ncsaba | :) |
| 14:43.57 | ncsaba | one more question: how well it works to have 2 versions of brlcad in parallel ? I have currently the stable via package manager, and the one I'm just compiling... will they coexist peacefully, or better use only 1 ? |
| 14:49.39 | brlcad | the way we work, they're not really in parallel |
| 14:50.08 | ncsaba | ok |
| 14:50.09 | brlcad | our STABLE branch is usually validated snapshots of trunk |
| 14:50.31 | brlcad | trunk is stabilized, cleaned up, verified, and sync'd to stable during a release |
| 14:51.00 | brlcad | we only evern release off of stable (after passing all our tests and other release steps) |
| 14:51.45 | ncsaba | what I meant is: if I install both a stable _and_ the compiled trunk to the same machine, will it work ? |
| 14:52.08 | brlcad | depends but they should install to different directories |
| 14:52.11 | ncsaba | for some things that works, for others not... I have 3 different versions of python here |
| 14:52.21 | brlcad | i do it all the time |
| 14:52.33 | ncsaba | ok, thanks ! |
| 14:52.52 | brlcad | the only problem would be if you installed the trunk version on TOP of another existing version |
| 14:53.01 | brlcad | even that will work, but it'll basically be the trunk version |
| 14:53.23 | brlcad | where's your current install? |
| 14:53.30 | brlcad | /usr/brlcad or /usr/brlcad/rel-something? |
| 14:54.06 | ncsaba | I just uninstalled it :-) |
| 14:54.48 | ncsaba | will only use trunc, better so |
| 14:54.52 | brlcad | if it's from our deb installer, it was probably /usr/brlcad/rel-7.22.0 or similar |
| 14:54.57 | ncsaba | I don't have anything mission critical depending on it... |
| 14:55.07 | brlcad | and trunk would then install as /usr/brlcad/dev-7.23.1 at the moment |
| 14:55.55 | ``Erik | <-- likes to keep /usr/brlcad/trunk/ /usr/brlcad/trunkopt/ /usr/brlcad/rel-7.22.0/ ... |
| 14:56.46 | brlcad | ``Erik: which reminds me, I removed all of those from .bz recently |
| 14:56.58 | brlcad | apparently libc got upgraded, none resolved |
| 14:57.01 | ncsaba | well if trunc will cause problems, I will install back the deb package - for now I will try using trunc (to have my own modifications immediately) |
| 14:57.40 | brlcad | ncsaba: you can read about trunk and STABLE in the top-level HACKING file |
| 14:57.49 | ``Erik | there were local builds on bz? I was just making sure the ports version was installed :) |
| 14:57.55 | jordisayol | deb/rpm packages installs at /usr/brlcad/ |
| 14:58.19 | brlcad | basically, trunk is a work in progress area that only tends to "settle down" as we near a release (monthly) |
| 14:58.44 | brlcad | trunk won't even necessarily compile sometimes |
| 14:58.46 | ncsaba | ok, fine for me - as long as it doesn't break completely |
| 14:59.00 | brlcad | but it is *expected* to, and any breakage is usually fixed quickly |
| 14:59.38 | ncsaba | as said, no mission critical things - if something breaks I can wait till fixed |
| 14:59.48 | brlcad | any breakage is usually going to be really trivial portability things, it compiled cleanly on mac but gcc on linux warns about something different and the build stops |
| 15:00.18 | brlcad | we not only treat all warnings as errors (so even benign non-issues will halt the build), but we also turn on a lot of extra warnings |
| 15:01.15 | brlcad | that will definitely be different from what you dealt with in java |
| 15:01.56 | ncsaba | well it's definitely easier to get C code wrong than java - at least for me :-) |
| 15:02.21 | ``Erik | 'wrong' can be a loaded word, I've seen plenty of java that works, but is sooooo wrong... :D |
| 15:02.26 | ncsaba | ok, scratch that last sentence, I don't want to hear any C - against java here :-) |
| 15:02.32 | brlcad | this is way beyond right vs wrong |
| 15:02.36 | ``Erik | it's too late! pandoras box is open! |
| 15:02.43 | brlcad | get him! |
| 15:02.58 | ncsaba | :) |
| 15:03.12 | ncsaba | ok, cmake problem fixed - I have now libpng, libregex and zlib off, despite having the dev packages installed ? Is there some specific version checked ? |
| 15:03.33 | ``Erik | breakage in trunk is usually a portability issue and usually fixed within a couple hours if not minutes |
| 15:04.30 | ``Erik | you mean cmake says that building libpng is off? |
| 15:04.35 | brlcad | ncsaba: OFF means their *compilation* is off, so it'll use the system version |
| 15:04.50 | ncsaba | aha, then I should perhaps check the rest which is ON |
| 15:04.50 | brlcad | that's a list of what to build and not build |
| 15:06.37 | brlcad | for getting started, I'd suggest just enabling everything |
| 15:07.02 | brlcad | it won't affect anything else (nothing else is going to find our libs) and it'll be a guaranteed configuration |
| 15:07.56 | brlcad | of the three configurations we try to sustain (everything on, everything off, and auto-detected), everything on is the one that is continually tested the most since it's a controlled environment |
| 15:08.59 | ncsaba | ok |
| 15:15.52 | *** join/#brlcad merzo_ (~merzo@40-90-132-95.pool.ukrtel.net) |
| 15:19.10 | ncsaba | BTW, this we page is giving error: http://more.brlcad.org/ |
| 15:19.55 | ncsaba | it is linked as "BRL-CAD Model Repository" from the wiki page |
| 15:20.21 | Notify | 03BRL-CAD:brlcad * 54966 brlcad/trunk/CMakeLists.txt: check for the mach kernel thread_policy header |
| 15:20.30 | brlcad | yeah, we just recently migrated our entire web infrastructure to fancy new faster hardware |
| 15:21.00 | brlcad | some of our websites still need to be configured |
| 15:21.05 | ncsaba | a model repository is a nice idea |
| 15:21.23 | ncsaba | do you know "Art of Illusion" ? |
| 15:21.29 | brlcad | of course :) |
| 15:21.35 | ncsaba | :) |
| 15:22.04 | ``Erik | brlcad: that libbu affinity cpu type issue was a style cmake cache |
| 15:22.24 | brlcad | ``Erik: good to know |
| 15:22.35 | brlcad | and a shame it didn't figure it out to invalidate the cache |
| 15:23.50 | ncsaba | AOI has the disadvantage it's interface is too GUI oriented for my taste, and it is not CAD-ready (units are not well supported) |
| 15:24.09 | brlcad | ncsaba: another great view on activity, ohloh has nice per-dev graphs: http://www.ohloh.net/p/brlcad/contributors/summary |
| 15:24.40 | ncsaba | what I like on AOI are are the procedural editors, and the plugin infrastructure |
| 15:24.44 | brlcad | ncsaba: most of the open source options out there aren't CAD-ready or are problematic licensing (from our perspective) |
| 15:24.52 | brlcad | but yeah, AOI has lots of nice stuff |
| 15:25.08 | brlcad | I'd love to collaborate more with a group like that, but we'd have to work out license differences |
| 15:25.15 | brlcad | they're a GPL project iirc |
| 15:25.53 | ncsaba | yes |
| 15:26.00 | ncsaba | and convinced one :) |
| 15:26.16 | brlcad | unfortunately or fortunately, we cannot look at anything GPL |
| 15:26.46 | brlcad | I'd like to move our own codebase to entirely MIT, but right now we're LGPLv2 for the forseeable future |
| 15:26.53 | ncsaba | but ideas are not under copy-right protection :) |
| 15:27.25 | ncsaba | so unless patented, ideas can be re-implemented |
| 15:28.02 | ncsaba | the animation features of AOI are also quite nice |
| 15:28.32 | ncsaba | the very idea of animating almost any of the parameters of an object is nice... |
| 15:29.14 | brlcad | yeah, our animation features are just terrible ;) |
| 15:31.25 | ncsaba | ok, how can I disable cmake to take my installed zlib ? I get: |
| 15:31.36 | ncsaba | make[2]: *** No rule to make target `/usr/local/lib/libz.so', needed by `lib/libopenNURBS.so.2010.04.95'. Stop. |
| 15:31.37 | ncsaba | make[1]: *** [src/other/openNURBS/CMakeFiles/openNURBS.dir/all] Error 2 |
| 15:34.23 | ncsaba | ah, ok - there was a broken symlink to the libz installed by brlcad last time: |
| 15:34.24 | ncsaba | lrwxrwxrwx 1 root root 23 2013-02-27 17:34 /usr/local/lib/libz.so -> /usr/brlcad/lib/libz.so |
| 15:34.47 | ncsaba | still left there after I uninstalled brlcad package - is that a bug in the uninstallation script ? |
| 15:35.01 | brlcad | interesting, sounds like it |
| 15:35.10 | brlcad | jordisayol: is it? :) |
| 15:36.59 | jordisayol | brlcad: yes, they do not allow to install multiple version at a time |
| 15:42.44 | Notify | 03BRL-CAD:bob1961 * 54967 brlcad/trunk/src/libtclcad/tclcad_obj.c: Update libtclcad's to_more_args_func() to protect any special characters that may be in the prompt strings. |
| 15:45.09 | *** join/#brlcad libero (~luca@host145-110-dynamic.0-79-r.retail.telecomitalia.it) |
| 15:47.47 | Notify | 03BRL-CAD:bob1961 * 54968 (brlcad/trunk/src/tclscripts/archer/Archer.tcl brlcad/trunk/src/tclscripts/archer/ArcherCore.tcl brlcad/trunk/src/tclscripts/lib/Command.tcl): Update the Command widgets invoke method to store the actual command the was executed (i.e., this command may be one that gets built up by the more-args mechanism that prompts the user for input). Also added a cmd_history_callback so that the |
| 15:47.49 | Notify | application can be notified of the actual command that was executed. |
| 15:51.50 | Notify | 03BRL-CAD:brlcad * 54969 brlcad/trunk/src/libbu/affinity.c: stub in an implementation of thread affinity for mac os x, to the extent that mac supports the notion. they don't really do affinity, but will let you group/ungroup threads to help/discourage cache sharing. here we put every thread into their own group so the default affinity scheduling should help minimize cpu movement. |
| 15:55.47 | jordisayol | brlcad: deb install/uninstall scripts do not create any symlink. |
| 15:58.33 | jordisayol | just set default application for the new "brlcad-v4" and "brlcad-v5" mime types, and update the environment PATH and MANPATH variables |
| 15:59.53 | jordisayol | BTW, is archer still in pre-alpha state? |
| 16:18.10 | Notify | 03BRL-CAD:brlcad * 54970 brlcad/trunk/src/libbu/affinity.c: document the mac interface since it's substantially different than the pthread api |
| 16:18.24 | brlcad | jordisayol: yes pre-alpha |
| 16:18.50 | ncsaba | jordisayol: hmm, is it possible I used update-alternatives ? that would then be a ubuntu/debian generic problem to clean up alternatives on uninstalling a package - if it is possible at all |
| 16:18.51 | brlcad | it's arguably alpha, but I haven't poked its stability recently |
| 16:19.16 | brlcad | yeah, something else must have made the symlink in /usr/local |
| 16:20.04 | brlcad | also "they do not allow" is apt and dpkg do not allow |
| 16:20.16 | brlcad | our tools don't care |
| 16:23.35 | Notify | 03BRL-CAD:brlcad * 54971 brlcad/trunk/src/libbu/affinity.c: fix style |
| 16:23.52 | jordisayol | ups, there is a "shlib" file on deb packages... |
| 16:24.24 | jordisayol | s/shlib/shlibs/ |
| 16:24.57 | jordisayol | http://man.he.net/man5/deb-shlibs |
| 16:25.34 | jordisayol | I this that this is not a good idea. I'll remove it for future releases |
| 16:26.17 | jordisayol | s/this/think/ pebcak :-/ |
| 16:35.36 | jordisayol | this was fixed few months ago, so ncsaba, if you build your own deb package from trunk, no symlinks against /usr/brlcad/lib libraries will be created. |
| 16:39.54 | Notify | 03BRL-CAD:bob1961 * 54972 brlcad/trunk/src/tclscripts/lib/Command.tcl: A few minor tweaks to the Command widget. Updated Command::print_more_args_prompt to make the insertion point visible. Updated Command::invoke to print the prompt string whether we have a more-args interruption or not. |
| 16:45.38 | Notify | 03BRL-CAD:brlcad * 54973 brlcad/trunk/src/libbu/heap.c: no longer indexing from 1 |
| 16:46.45 | Notify | 03BRL-CAD:brlcad * 54974 brlcad/trunk/src/libbu/affinity.c: take a stab at implementing thread affinity for windows as well (hah, in advance of windows threading). using bu_parallel_id() to set the cpu number. earned fewest line count badge (albeit untested) |
| 16:55.16 | Notify | 03BRL-CAD:brlcad * 54975 brlcad/trunk/src/libbu/affinity.c: need to test this theory, but we shouldn't need to look up the current affinity mask to set a cpu preference. bu_parallel_id() modded with the number of avail cpus should set all threads with the appropriate affinity mask. |
| 16:57.24 | Notify | 03BRL-CAD:brlcad * 54976 brlcad/trunk/src/libbu/affinity.c: remove the unused/unnecessary variables |
| 17:18.47 | ``Erik | is parallel_set_affinity going to get a bu_ entry in bu.h soon? be interesting to plug it into adrt/librender/camera.c and see what happens |
| 17:24.04 | Notify | 03BRL-CAD:n_reed * 54977 brlcad/trunk/src/libged/draw.c: save pointers to the display lists that we want to draw so we don't have to search for them later |
| 17:25.17 | brlcad | ``Erik: wasn't planning on exposing it, hence no bu_ prefix, but it'd be quick and easy to test -- if it gives a boost, sure |
| 17:25.36 | brlcad | just #include "../libbu/parallel.h" and try calling it |
| 17:26.12 | brlcad | though at the moment, I just changed it to call bu_parallel_id(), which is stubbed empty after removing boost_thread |
| 17:26.33 | brlcad | have to implement that (later today) |
| 17:26.54 | ``Erik | I'll wait a bit, would need to scrap up a benchmark case or something :) |
| 17:27.51 | ``Erik | I'm mucking with a trainwreck of maven dependancies and about to head out for the weekend *shrug* it'll still be there next week, I assume |
| 17:35.46 | Notify | 03BRL-CAD:r_weiss * 54978 brlcad/trunk/src/other/step/cmake/Generated_Source_Utils.cmake: Fix a break in the Windows build. |
| 17:41.09 | *** join/#brlcad yiyus (1242712427@je.je.je) |
| 17:47.33 | Notify | 03BRL-CAD:r_weiss * 54979 brlcad/trunk/src/other/step/cmake/Generated_Source_Utils.cmake: Fix to windows built, again. |
| 17:55.51 | *** join/#brlcad luca79 (~luca@net-188-216-230-48.cust.dsl.vodafone.it) |
| 18:03.42 | *** join/#brlcad ncsaba (~ncsaba@p54983F7E.dip.t-dialin.net) |
| 18:59.11 | brlcad | stops tweaking the application |
| 19:00.53 | Notify | 03BRL-CAD Wiki:Sean * 4956 /wiki/Google_Summer_of_Code: we've applied for 2013 |
| 19:01.16 | brlcad | So.... GSoC mentors! Time to update http://brlcad.org/wiki/Google_Summer_of_Code/Project_Ideas |
| 19:06.49 | Notify | 03BRL-CAD Wiki:Sean * 4957 /wiki/Google_Summer_of_Code/2013: stub in the 2013 page |
| 20:24.39 | *** join/#brlcad vaibhav_ (75cdb036@gateway/web/freenode/ip.117.205.176.54) |
| 20:46.02 | ``Erik | the pages impact statements almost read like starbucks sizes O.o |
| 20:50.44 | Notify | 03BRL-CAD:r_weiss * 54980 brlcad/trunk/src/other/step/cmake/Generated_Source_Utils.cmake: Change to cmake build logic for md5 express. This change appears to work for both the linux and windows build. More testing is needed. |
| 21:01.25 | Notify | 03BRL-CAD Wiki:Sean * 4958 /wiki/Google_Summer_of_Code/2013: /* Accepted Projects */ |
| 21:01.33 | vaibhav_ | I may be needing loads of help. |
| 21:02.00 | vaibhav_ | Firstly, I'm good with computer graphics. |
| 21:02.14 | vaibhav_ | Now what should I do next, to accelerate my learning curve? |
| 21:08.31 | brlcad | vaibhav_: hello and who are you? :) |
| 21:25.14 | brlcad | vaibhav_: that said, welcome and there are lots of ways to get started with brl-cad ... see our quickies and dueces pages on our wiki for some ideas |
| 21:25.28 | brlcad | discussion warranted, though, as some are done and we need to update the page |
| 21:46.29 | *** join/#brlcad caen23 (~cezar@92.81.220.160) |
| 22:05.01 | *** join/#brlcad ncsaba (~ncsaba@p54983F7E.dip.t-dialin.net) |
| 22:18.19 | ncsaba | OK guys, I will thank you for your help today - I managed finally to compile the sources |
| 22:19.22 | ncsaba | had a short look to pipe.c and analyze.c; next days will do the rt_pipe_length function - might need some help here and there though... |
| 22:29.47 | starseeker | brlcad: does libpc make use one one of the "pre-defined" constraint input languages like minizinc or flatzinc? |
| 22:35.16 | kanzure | win 102 |
| 22:35.21 | kanzure | oops :( please excuse me |
| 23:14.20 | *** join/#brlcad velociostrich (~anonymous@c-24-0-153-224.hsd1.pa.comcast.net) |
