| 03:15.36 | pra5ad | damn good 1lb burger for $5 @ inner harbor |
| 03:15.41 | pra5ad | mmm so full =) |
| 03:18.21 | brlcad | mm.. them rats is good eatn |
| 03:18.50 | pra5ad | http://www.koopers.com/ |
| 03:19.00 | pra5ad | jackpot :) |
| 03:19.43 | pra5ad | brlcad, did u fix the mounts yet? |
| 03:21.37 | pra5ad | gutter, mind, bad |
| 03:21.55 | brlcad | bad gutter, no donut for you |
| 04:37.32 | *** join/#brlcad DTRemenak (n=DTRemena@DHCP-170-143.caltech.edu) |
| 14:28.32 | ``Erik | grar, the wench keeps trying to make me go to fells point o.O I'll have to check that place out next time I'm weak |
| 14:28.56 | ``Erik | usually I eat at duclaws or that, uh, 'admirals club' or whatever restraunt on the water by the stagey area |
| 15:41.54 | ``Erik | 2/det |
| 15:42.11 | ``Erik | meh |
| 15:50.05 | brlcad | drinking earlier I see |
| 16:19.24 | *** join/#brlcad amatus_ (n=3fa31caa@bz.bzflag.bz) |
| 16:42.57 | *** join/#brlcad MikeC (n=Mike@pool-138-88-142-43.esr.east.verizon.net) |
| 16:43.06 | MikeC | Hi everyone |
| 16:43.25 | MikeC | I just downloaded the source for brlcad 7.6.4 and compiled it |
| 16:43.39 | MikeC | I went into adrt to compile it as well and it seems a header file is missing: isst.h |
| 16:44.11 | MikeC | isst.h is in a previous version we have compiled successfully 7.6.1 |
| 16:51.11 | MikeC | I'd like to try to get the code from CVS but how do I specify that I would like version 7.6.4? |
| 17:29.54 | MikeC | I'll be back later |
| 19:41.18 | *** join/#brlcad MikeC2 (n=Mike@pool-138-88-142-43.esr.east.verizon.net) |
| 19:42.14 | MikeC2 | hi anyone here, I have some questions about ADRT related things and a CVS question |
| 19:43.45 | *** join/#brlcad raz (n=raz@pool-138-88-142-43.esr.east.verizon.net) |
| 19:46.40 | ``Erik | O.o |
| 19:55.21 | MikeC2 | hi Erik |
| 20:00.33 | MikeC2 | Erik you know if Sean or Justin are around? |
| 20:12.03 | brlcad | howdy |
| 20:12.16 | MikeC2 | hey man! |
| 20:12.41 | MikeC2 | We found in downloading the source code that there is a missing header file named isst.h in the adrt src directory |
| 20:12.55 | MikeC2 | however it seems when we obtain the code through CVS the isst.h file is there |
| 20:13.10 | MikeC2 | also how do we checkout a specific version of brlcad using CVS? |
| 20:15.54 | MikeC2 | I can checkout brlcad with CVS but it always gives the latest version |
| 20:16.02 | CIA-12 | BRL-CAD: 03brlcad * 10brlcad/src/adrt/isst/Makefile.am: isst.h and isst_struct.h were missing from the dist, thanks MikeC |
| 20:16.35 | MikeC2 | ok thanks CIA-12 |
| 20:16.37 | brlcad | thanks |
| 20:17.00 | brlcad | you checkout a specific version using the -r flag to checkout |
| 20:17.28 | MikeC2 | is there a way to get a list of existing revisions available for checkout? |
| 20:17.34 | brlcad | cvs -d blahblah checkout -P -r rel-7-6-2 brlcad |
| 20:17.40 | MikeC2 | oh ok |
| 20:17.52 | brlcad | yeah, do a cvs status some_file |
| 20:17.54 | MikeC2 | and I assume you recommend not using the latest HEAD from CVS right? |
| 20:18.39 | brlcad | cvs head is generally pretty stable, and it's easier for me to fix problems on the fly if you run into any but it all depends on you |
| 20:18.47 | MikeC2 | ok |
| 20:18.57 | brlcad | head should be working fine right now |
| 20:19.14 | brlcad | head 'usually' is working fine |
| 20:19.36 | MikeC2 | we found a problem in the isst slave code that causes a seg fault in normal usage of it |
| 20:20.47 | MikeC2 | you cannot run the isst package successfully as it is |
| 20:20.55 | MikeC2 | we found a solution but would like to bounce it off you guys |
| 20:24.01 | brlcad | excellent, bounce away |
| 20:24.29 | brlcad | does the fix change behavior really or just fix the segfault? |
| 20:24.38 | MikeC2 | fix the segfault |
| 20:24.45 | MikeC2 | ok first lets describe the problem |
| 20:25.09 | MikeC2 | if you look in src/adrt/libcommon/unpack.h |
| 20:25.43 | MikeC2 | notice the prototype for the function: extern void common_unpack_free(); |
| 20:25.43 | brlcad | eww, fixed array sizes |
| 20:25.44 | brlcad | okay |
| 20:25.57 | MikeC2 | now look in src/adrt/libcommon/unpack.c |
| 20:26.10 | MikeC2 | look for the common_unpack_free function |
| 20:26.30 | MikeC2 | notice that the prototype in the c file does not match the prototype in the header file |
| 20:26.50 | MikeC2 | now look in src/adrt/isst/slave/slave.c |
| 20:27.24 | MikeC2 | look at the function: isst_slave_Free |
| 20:27.29 | MikeC2 | oops isst_slave_free |
| 20:27.35 | MikeC2 | notice it calls common_unpack_free |
| 20:27.38 | MikeC2 | with no parameters |
| 20:27.51 | CIA-12 | BRL-CAD: 03brlcad * 10brlcad/src/adrt/libcommon/unpack.h: prototype should match, common_unpack_free -- thanks to MikeC |
| 20:28.00 | brlcad | heh |
| 20:28.02 | MikeC2 | somehow this compiles |
| 20:28.13 | brlcad | the prototype let that compile |
| 20:28.17 | brlcad | which is bogus ;) |
| 20:28.27 | MikeC2 | we are using gcc (GCC) 4.0.0 20050519 |
| 20:28.38 | brlcad | results in an anonymous argument (which C allows) |
| 20:28.44 | MikeC2 | and also gcc 3.4.4 on another computer |
| 20:29.04 | MikeC2 | ok that causes a major problem |
| 20:29.13 | MikeC2 | now the question is, how has it been running in the past? |
| 20:29.17 | brlcad | it's like how I can call malloc(); without a size and it'll compile |
| 20:29.38 | brlcad | presuming I declare a matching prototype |
| 20:29.46 | MikeC2 | ok well I guess I been in the C++ world too much |
| 20:29.52 | brlcad | it's just undefined as to what it'd do (and in malloc's case it'd crash probably) |
| 20:30.22 | MikeC2 | since the slave crashes how has it been running before? |
| 20:30.25 | brlcad | C++ carrys that same baggage, but you don't see it much since you also have polymorphic function args |
| 20:30.52 | MikeC2 | I understand that part now |
| 20:30.53 | brlcad | it's a free routine.. so it's like passing a bogus pointer to free() |
| 20:30.56 | brlcad | it may or may not crash |
| 20:31.08 | brlcad | in your case it does |
| 20:31.17 | brlcad | he's apparently gotten lucky ;) |
| 20:31.37 | MikeC2 | ok |
| 20:31.48 | MikeC2 | we'll fix it for now in our stuff |
| 20:31.54 | MikeC2 | we got it running after fixing it |
| 20:32.04 | MikeC2 | but we still have to actually do aa shotline |
| 20:32.09 | MikeC2 | aa = a |
| 20:34.24 | CIA-12 | BRL-CAD: 03brlcad * 10brlcad/src/adrt/libcommon/unpack.c: make sure db is a valid pointer so common_unpack_free doesn't crash |
| 20:34.25 | brlcad | did you do anything to isst_slave_free? |
| 20:34.45 | MikeC2 | yes |
| 20:35.06 | MikeC2 | there is a global variable called db |
| 20:35.20 | MikeC2 | we pass in &db to common_unpack_free |
| 20:35.44 | brlcad | ahh, there it is |
| 20:37.03 | CIA-12 | BRL-CAD: 03brlcad * 10brlcad/src/adrt/isst/slave/slave.c: free the global db node |
| 20:37.18 | brlcad | who may I credit for the fix in the release notes? |
| 20:37.33 | MikeC2 | Mike Caruso and Razak Arunaogun - Qryon LLC |
| 20:39.17 | brlcad | people get the credit when the announcements go out, not organizations -- though I'll note them in the cvs commit log and authors file |
| 20:39.33 | brlcad | thanks for finding that |
| 20:39.47 | MikeC2 | no problem, we'll let you know if we find more as we go |
| 21:51.08 | CIA-12 | BRL-CAD: 03brlcad * 10brlcad/NEWS: fixed adrt/isst client crash bug |
| 21:54.11 | CIA-12 | BRL-CAD: 03brlcad * 10brlcad/AUTHORS: special thanks to Mike Caruso and Razak Arunaogun - Qryon LLC for fixes and feedback |
| 22:05.04 | CIA-12 | BRL-CAD: 03brlcad * 10brlcad/AUTHORS: Qryon not Qryan |
| 22:15.45 | CIA-12 | BRL-CAD: 03brlcad * 10brlcad/AUTHORS: dtremmy to special thanks for his build fixes |
| 23:57.31 | *** join/#brlcad DTRemenak (n=DTRemena@DHCP-170-143.caltech.edu) |