| 00:06.42 | *** join/#brlcad Ralith (n=ralith@216.162.199.202) | |
| 01:16.34 | *** join/#brlcad iday (n=jlowens@c-76-111-141-77.hsd1.de.comcast.net) | |
| 01:18.29 | PrezKennedy | i use vim/jedit |
| 01:25.40 | ``Erik | .de, jlo? |
| 01:26.08 | ``Erik | is have her dur grace linked up to a deleware cx? |
| 01:36.38 | *** part/#brlcad Nikty (n=nicholas@89.163.117.94) | |
| 03:40.53 | yukonbob | hello, cadheads |
| 03:56.49 | *** join/#brlcad cad98 (n=4463b899@bz.bzflag.bz) | |
| 04:04.42 | *** join/#brlcad Ralith (n=ralith@216.162.199.202) | |
| 04:59.25 | *** join/#brlcad CIA-4 (n=CIA@208.69.182.149) | |
| 05:48.23 | *** join/#brlcad PrezKennnedy (i=Matthew@whitecalf.net) | |
| 07:53.26 | *** join/#brlcad bolly (n=5190e8a3@bz.bzflag.bz) | |
| 08:47.54 | brlcad | ``Erik: he's at the beach |
| 10:41.43 | *** join/#brlcad prasad_1 (n=psilva@static-70-108-244-218.res.east.verizon.net) | |
| 11:09.56 | *** join/#brlcad yukonbob_ (i=1000@s142-179-54-198.bc.hsia.telus.net) | |
| 11:49.25 | *** join/#brlcad thing0 (n=ric@123.208.95.171) | |
| 11:56.40 | *** join/#brlcad elite01 (n=elite01@unaffiliated/elite01) | |
| 11:59.47 | starseeker | is starting to get the hang of glibc's fnmatch code a bit better |
| 12:13.41 | *** join/#brlcad andrecastelo (n=chatzill@189.71.31.249) | |
| 12:29.25 | *** join/#brlcad archivist_emc (n=archivis@host81-149-119-172.in-addr.btopenworld.com) | |
| 14:07.35 | *** join/#brlcad mafm (n=mafm@elnet-111.lip.pt) | |
| 14:07.52 | mafm | hallo |
| 14:09.08 | *** join/#brlcad thing0 (n=ric@124-169-23-145.dyn.iinet.net.au) | |
| 14:09.42 | ``Erik | oi |
| 14:10.53 | brlcad | ugh, 130 hour build time.. |
| 14:11.00 | mafm | brlcad: re:argc for commands: maybe I missed some of them |
| 14:11.14 | mafm | I think that one that starseeker commited recently had the same problem |
| 14:11.34 | brlcad | mafm: you missed a bunch of them iirc :) |
| 14:11.54 | brlcad | starseeker's not using the ged interface just yet contrary to the function name |
| 14:11.59 | mafm | uhm, then maybe was a regex too restrictive |
| 14:12.02 | brlcad | using the wdb interface |
| 14:16.29 | mafm | hmm |
| 14:16.34 | mafm | so do you want me to fix the rest? |
| 14:24.04 | brlcad | i already did |
| 14:24.15 | brlcad | that same day, iirc even |
| 14:24.39 | brlcad | ahh, cia is down... dagnabit |
| 14:25.33 | mafm | oh |
| 14:30.04 | mafm | I saw a bit of activity in the mail lately |
| 14:35.28 | yukonbob_ | brlcad: I'd change that hard-coded value to something like sizeof(int) |
| 14:35.34 | yukonbob_ | it's not portable. |
| 14:35.43 | yukonbob_ | ^--in Tcl patch |
| 14:41.01 | mafm | so, any news lately? I took a little break with a friend for the last week, I still have a lot of mail to read |
| 14:41.37 | brlcad | yukonbob_: i know .. it's not really a proper fix (wasn't intended to be really) .. just hopefully enough to provoke them to fix it properly |
| 14:42.29 | brlcad | the bigger issue is really prevalent throughout the API .. there are unqualified ints used everywhere that can overflow in many/most of the functions that aren't being checked |
| 14:42.30 | yukonbob_ | is also a bit surprised Tcl is using malloc, and not some tcl_malloc(), where a test for this type of thing could be inserted once, and fix all cases... |
| 14:42.51 | brlcad | tcl does have their own wrapper, but it doesn't check |
| 14:43.08 | yukonbob_ | that's where the fix should be, by my one-minute evaluation... |
| 14:43.47 | yukonbob_ | tct will probably come to same conclusion themselves, though, unless I'm just wrong. |
| 14:43.58 | brlcad | we'd still be stuck at it failing on commands that exceed 1073741823 |
| 14:44.21 | brlcad | the patch pushes it all the way up to 2147483647 before it'll fail |
| 14:44.44 | brlcad | there's a length = offset * 2 that'll overflow it half-way |
| 14:44.53 | yukonbob_ | would have to look harder to completely understand; trusts brlcad |
| 14:45.17 | yukonbob_ | would the "test for completeness of command" help wrt parsing the db? |
| 14:45.52 | yukonbob_ | is also working to getting itcl up/running with his -current OS, and Tcl 8.5.4, then back on the brl-cad wagon ;) |
| 14:46.08 | brlcad | also, by the time it's to tcl's wrapper, it's an unsigned int (like ours in libbu) .. so you lose the fact that it overflowed |
| 14:54.48 | *** join/#brlcad CIA-4 (n=CIA@208.69.182.149) | |
| 15:01.50 | yukonbob_ | [info complete $x] is what I was thinking of, perhaps allowing one to parse the db in smaller chunks. |
| 15:01.55 | yukonbob_ | brlcad: --^ |
| 15:04.58 | yukonbob_ | heads to work. |
| 15:24.00 | brlcad | yukonbob_: brlcad/src/conv/asc2g.c if you want to look at improving the converter further :-) |
| 15:24.20 | brlcad | not sure what info complete would translate to on Tcl's C API |
| 15:24.34 | brlcad | right now, we just call Tcl_EvalFile iirc |
| 15:42.31 | *** join/#brlcad mafm (n=mafm@elnet-111.lip.pt) | |
| 15:43.18 | mafm | great, my machine at work is borked |
| 15:43.19 | mafm | :) |
| 15:44.55 | brlcad | excellent |
| 15:45.18 | mafm | I love when things run smoothly :P |
| 15:59.56 | *** join/#brlcad thing0 (n=ric@203-206-101-244.dyn.iinet.net.au) | |
| 17:08.41 | ``Erik | http://www.cyriak.co.uk/lhc/lhc-webcams.html |
| 17:16.43 | *** join/#brlcad thing0 (n=ric@203-206-101-244.dyn.iinet.net.au) | |
| 17:18.57 | mafm | heh |
| 17:19.23 | mafm | LHC news are having a big impact |
| 17:46.54 | *** join/#brlcad mafm (n=mafm@elnet-111.lip.pt) | |
| 18:09.30 | *** join/#brlcad quentusrex (n=quentusr@c-71-197-244-228.hsd1.or.comcast.net) | |
| 18:45.37 | starseeker | brlcad: It looks like including the collate features of GNU's fnmatch will mean including some locale code too - is it worth it? |
| 19:13.42 | starseeker | growls in annoyance and starts re-thinking plan B, adding a few features to the BSD version |
| 20:32.21 | *** join/#brlcad iday (n=jlowens@c-76-111-141-77.hsd1.de.comcast.net) | |
| 21:54.26 | *** join/#brlcad prasad_ (n=psilva@h-72-245-122-226.mclnva23.covad.net) | |
| 22:39.47 | brlcad | starseeker: nope |
| 22:40.03 | brlcad | just rip that part out? |
| 22:40.09 | starseeker | It' |
| 22:40.12 | ``Erik | o.O |
| 22:40.16 | starseeker | s a bit tricky |
| 22:40.21 | starseeker | hang on, take a look: |
| 22:41.11 | CIA-4 | BRL-CAD: 03starseeker * r32641 10/brlcad/trunk/src/libbu/fnmatch.c: Start of an implementation of character classes. |
| 22:42.06 | starseeker | I've got a little more debugging to do, but I think I'm getting close. |
| 22:47.01 | starseeker | "and an ominous silence descended over IRC" ;-) |
| 22:53.27 | starseeker | decides to head home before he gets in any more trouble :-{ |
| 22:53.36 | brlcad | heh |
| 22:53.39 | starseeker | how'd it get to be 7 already?? |
| 22:53.47 | brlcad | yeah.. |
| 22:54.47 | starseeker | brlcad: if we opt out of the collate feature, I don't think it's too much difference between adding to the working one and stripping out the unneeded stuff from the other one |
| 22:54.55 | brlcad | what is the collate feature? |
| 22:55.23 | starseeker | As near as I can tell, it's some kind of ordering awareness based on locale information |
| 22:55.51 | brlcad | there is no geoemtry locale, so that doesn't seem to be anything of use regardless |
| 22:55.56 | brlcad | you'd just rip that out if you did use it |
| 22:56.11 | starseeker | nods |
| 22:56.20 | brlcad | sort of like -atime |
| 22:56.30 | starseeker | It might be things like "a comes before z" though - that's what I'm not sure of yet |
| 22:56.57 | brlcad | either way, sounds like a gnu extension |
| 22:57.06 | starseeker | I'm pretty close to getting the character classes working, which I remember you missing |
| 22:57.33 | brlcad | the point isn't to get all features, it's to get whatever makes most sense for geometry processing |
| 22:57.53 | brlcad | which is going to be all ascii-based |
| 22:58.00 | brlcad | standard posix globbing |
| 22:58.26 | starseeker | Here's where it pops up on opengroup: http://www.opengroup.org/onlinepubs/000095399/basedefs/xbd_chap09.html#tag_09_03_05 |
| 23:00.28 | starseeker | I got a simple test case of [[:digit:]] to work, so now it's just a question of making sure I understand how its used and making sure what I've got in there has the correct behaviors |
| 23:02.16 | starseeker | I guess it's a sign of coding immaturity but I found it much more straightforward to add to the BSD code than ripping stuff out of the GNU code :-/ |
| 23:02.29 | brlcad | <PROTECTED> |
| 23:02.40 | brlcad | never used collating |
| 23:03.08 | brlcad | not sure equivalence classes are useful to us, unless it's easy to implement |
| 23:03.33 | starseeker | Unless I'm very much mistaken, it's quite hard to implement - I think it depends on locale support in some fashion. |
| 23:04.00 | brlcad | collating actually sounds rather useful, but I don't see anyone getting that advanced |
| 23:04.10 | starseeker | nods |
| 23:04.12 | brlcad | it's a way to match substrings |
| 23:04.32 | starseeker | I was thinking if anyone got that advanced regex could step in |
| 23:05.49 | brlcad | looks like bsd doesn't support/implement it |
| 23:05.55 | starseeker | nope |
| 23:06.12 | brlcad | yeah, forget about it |
| 23:06.38 | starseeker | ok. If someone specifically requests we can do an estimate of the cost/benefit ratio, but I'm thinking it's pretty high |
| 23:07.03 | starseeker | would probably mean including locale to do it portably |
| 23:07.05 | brlcad | yeah, i don't see the need |
| 23:07.11 | brlcad | you could just use a regex, even easier |
| 23:07.18 | starseeker | nods |
| 23:08.14 | starseeker | Now I just need to hunt up some advanced glob test cases using :digit: and friends, but after I get home :-) |
| 23:08.21 | starseeker | really leaves, bbl |
| 23:09.04 | brlcad | and ranges |
| 23:09.13 | brlcad | find . -name [0-9]* |
| 23:09.29 | starseeker | actually, that should work |
| 23:09.55 | brlcad | find . -name [x-z]* |
| 23:10.27 | starseeker | will test, but I think it does have at least basic ranges already |
| 23:10.31 | brlcad | find . -name [a-zA-Z_][a-zA-Z0-9_-]* |
| 23:10.48 | starseeker | dunno about that one ;-) |
| 23:10.53 | starseeker | ok, later all! |
| 23:11.12 | brlcad | cya |