| 01:49.11 | *** join/#brlcad Pimpinella (~frank@p5481A1A9.dip0.t-ipconnect.de) | |
| 03:44.13 | narnia | brlcad, have you every used sourcenav/snavigator for development work? |
| 03:47.00 | brlcad | I have not |
| 03:48.24 | brlcad | closest would probably be doxygen's call tree displays |
| 04:01.30 | narnia | brlcad, i did not realize that redhat had released it under the gpl. it is now on sourceforge. i had used it in the past when it was still cygnus. pretty nice. |
| 04:02.04 | narnia | brlcad, i have just installed it. makes working through the old nist scl3-2 much easier. |
| 04:03.11 | narnia | http://sourcenav.sourceforge.net |
| 04:11.19 | brlcad | yeah, I know -- just never used it |
| 04:39.01 | narnia | brlcad, how are you doing? anther late night? |
| 04:54.52 | brlcad | not too late |
| 04:55.04 | brlcad | trying to build up some steam for a long weekend of work |
| 04:55.12 | brlcad | otherwise doing alright |
| 04:58.59 | narnia | i am plodding along with part11 and part 21. part21 has alot of exceptions. entity tracking and relationships are something else. |
| 05:02.15 | narnia | looking at how to build the symbol table to be more efficient. |
| 05:02.54 | brlcad | is this still part of libexpress or something else you're working with? |
| 05:03.39 | narnia | this is for part21. |
| 05:04.29 | brlcad | no, I mean code-wise |
| 05:04.37 | narnia | basically a completely new part21. |
| 05:04.37 | brlcad | or do you mean that you're reading the spec |
| 05:05.35 | narnia | been reading the spec and coding. entity inheritance in express is nothing like c++. |
| 05:05.54 | brlcad | k |
| 05:06.59 | narnia | express editon 2 is taking longer. i do not want to break express edition 1. |
| 05:08.39 | narnia | getting part21 correct requires reading the part21 pdf many many times to full understand what they are attempting to convey. |
| 05:11.43 | narnia | the part21 parser is not static. it is really dynamic. a new entity defined in a step file becomes a new keyword in the data section. |
| 05:18.03 | narnia | so the entity declaration and the entity keyword in the data section need to be validated. |
| 05:19.01 | narnia | i can see why some people programmed in lisp, prolog, etc. |
| 05:20.20 | narnia | i am tempted to say that part21 may have to be multi-pass parsing. |
| 05:21.49 | narnia | i do know that as the size of the step file grows the part21 source code as it stands now gets slower and slower. |
| 05:31.16 | narnia | i tried to feed it a 40mbytes step file of a complete yanmar diesel engine. it was painful. |
| 05:33.45 | narnia | not as bad as opencascade but close. |
| 05:35.00 | narnia | opencascade is still the worse case. |
| 05:38.39 | narnia | i read a good write up about part21 parsing. i should have saved the link. the author talked about parsing and entity validation in large step files being a problem. the author was suggesting a parallel approach. |
| 05:47.19 | narnia | brlcad, found the links. http://www.c-lab.de/~wolfgang/ecip017.pdf and http://www.c-lab.de/~wolfgang/ecip013.pdf |
| 14:58.27 | brlcad | cool, I'll take a look |
| 14:58.44 | brlcad | sorry, had to jump out early last night |
| 15:17.44 | narnia | brlcad, no probelm. |
| 15:21.08 | narnia | been using soucenav to look at the new part21 code and the nist scl3-2 source. the problem with the new part21 code appears to be the hash table routines. i am reading a pdf writing by Per-?ke Larson, Dynamic hashing. a similar approach was used in the old nist scl3-2 code. think i will borrow the nist scl3-2 hash code and test it out. |
| 15:28.20 | narnia | it is also clear that part21 is going to have to be at least 2 passes. |
| 15:57.50 | brlcad | narnia: in regards to your unicode comments yesterday, there is a new database I/O layer that is in development that will support unicode |
| 15:58.28 | brlcad | it'll be quite a while before that's ready for use, but it's part of the new modeler system |
| 16:18.57 | narnia | brlcad; that is good to know, thank you |
| 16:21.26 | brlcad | yeah, it'll be pretty interesting.. working out the details of integrating geometry into an actual database and revision managemnet system |
| 16:24.33 | narnia | brlcad, which database? mysql? |
| 16:26.53 | brlcad | the current plan, which may change mind you, is to have a thin layer that will sit on top of pluggable backends |
| 16:27.17 | brlcad | so one could use the existing .g file format, an sql database, a db2 database, etc |
| 16:27.29 | narnia | ah, |
| 16:27.59 | narnia | maybe reusable for part 22. |
| 16:28.45 | narnia | always nice to be able to recycle/reuse code. |
| 16:29.12 | brlcad | could write in the hooks for any file format really, but the current plan is to focus on .g, sql, and flat files |
| 16:29.43 | narnia | sql and flat files would work for part22. |
| 16:29.48 | brlcad | that should ensure the api is easy to extend |
| 16:30.20 | brlcad | even .g could work for part22, I'd just need to do some testing |
| 16:30.51 | narnia | i do not know enough about .g to have a clue. ;-) |
| 16:31.11 | brlcad | at the absolute worst, the .g can stash the entire contents of other formats |
| 16:31.26 | brlcad | sort of akin to sql blobs |
| 16:31.34 | narnia | ah |
| 16:31.35 | brlcad | at best, it'll all just work :) |
| 16:31.43 | brlcad | with compact representations |
| 16:32.21 | brlcad | I suspect that unicode will actually work, barring maybe some minor bugs |
| 16:32.41 | brlcad | it was designed to support it, just never needed/tested |
| 16:33.49 | *** join/#brlcad asdasxca (~oper@83.148.76.210) | |
| 16:34.36 | brlcad | hello asdasxca |
| 16:36.17 | asdasxca | guu |
| 16:36.42 | brlcad | hrm |
| 16:37.19 | asdasxca | brlcad |
| 16:37.57 | brlcad | yes? |
| 16:38.08 | asdasxca | bg |
| 16:38.08 | asdasxca | ? |
| 16:38.58 | brlcad | can you form a complete sentance? :) |
| 16:40.53 | brlcad | well, I'm off to the gym |
| 16:40.56 | narnia | i wonder if it would be worth exploring a syntax tree for part21? treat express language like perl. almost everything else in 10303 is based in part11. would not solve the problem with the entities and the subtype and supertype. |
| 16:41.11 | narnia | brlcad, say hello to jim. ;-) |
| 16:41.18 | brlcad | narnia: if asdasxca causes any trouble, here's the hammer |
| 16:41.21 | *** mode/#brlcad [+o narnia] by brlcad | |
| 16:41.31 | narnia | ah |
| 16:41.36 | narnia | understand |
| 16:42.54 | narnia | asdasxca, are you using brl-cad? do you have any questions about brl-cad? |
| 18:27.20 | brlcad | ahh, good no trouble ;) |
| 18:29.20 | brlcad | fyi, lots of folks (including a wide variety of kids) know me through bzflag |
| 18:30.14 | brlcad | that can bring in a random assortment of personalities from time to time |
| 18:30.24 | brlcad | not a big deal, just might come up as it has in the past |
| 18:58.19 | narnia | okay |