| 03:35.47 | Maloeran | Ahah, I ordered nice indian food so many times from some place that they send me bonus dishes |
| 04:07.11 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/src/other/openNURBS/opennurbs_system.h: gcc 4 at least uses __GNU_C__, encourage ON_COMPILER_GNU to get defined |
| 04:16.30 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/src/other/openNURBS/opennurbs_point.cpp: check for ON_COMPILER_GNU instead of _GNU_SOURCE so we don't need to add configure checks for finite() |
| 04:23.27 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/src/other/openNURBS/Makefile.am: initial automake build file template for openNURBS that correctly generates the libopenNURBS library |
| 06:47.38 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/configure.ac: |
| 06:47.38 | CIA-5 | BRL-CAD: add openNURBS to the configuration and build. this adds |
| 06:47.38 | CIA-5 | BRL-CAD: --enable-opennurbs-build as well as a variety of aliases, while properly |
| 06:47.38 | CIA-5 | BRL-CAD: checking to see if openNURBS is already installed on the system or not so that |
| 06:47.38 | CIA-5 | BRL-CAD: auto as default will work. |
| 06:48.22 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/src/other/Makefile.am: check whether openNURBS has been enabled or not in order to compile/traverse |
| 08:35.28 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/src/rt/main.c: be less harsh about this being a debug build -- just say debug symbols are available and leave it at that |
| 09:36.58 | *** join/#brlcad clock_ (i=clock@84-72-60-48.dclient.hispeed.ch) |
| 09:37.18 | clock_ | Do you need any more information apart from what I posted about the segfault to the bug tracking database? |
| 10:13.51 | brlcad | clock_: no, that was perfect -- thanks |
| 10:13.53 | brlcad | quite detailed |
| 10:21.29 | clock_ | pixelnum=39999, pindex=119997 |
| 10:21.29 | clock_ | nonif |
| 10:21.29 | clock_ | Allocating pixmap to 120000 bytes. |
| 10:21.29 | clock_ | pixelnum=0, pindex=0 |
| 10:21.29 | clock_ | ./moss.rt: line 1: 22600 Segmentation fault (core dumped) /home/clock/brlcad-7.8.4/src/rt/rt -M $* -o moss.pix tetrax.g tetrax 2>>moss.log <moss.rtanim |
| 10:21.36 | clock_ | Now I am puzzling myself why it's crashing |
| 10:22.24 | clock_ | what timezone are you in? |
| 10:36.07 | brlcad | eastern u.s. |
| 10:36.25 | brlcad | yeah, it's late/early :) |
| 11:07.10 | clock_ | eastern us that's GMT-how much? |
| 11:07.42 | clock_ | An array of 120,000 bytes is allocated and not freed. Subsequent touch at the zeroth array element invokes a segfault |
| 11:08.14 | clock_ | The first frame goes through seemingly completely. |
| 11:08.36 | clock_ | Then the array is freed, allocated again to the same size, resulting in the same pointer, and a touch at the very zeroth pixel bombs the program |
| 11:09.07 | clock_ | The same on Gentoo Linux gcc 3.3.6: |
| 11:09.11 | clock_ | clock@server ~/crash $ ./moss.rt -s 200 |
| 11:09.11 | clock_ | *** glibc detected *** double free or corruption (!prev): 0x0809b430 *** |
| 11:09.11 | clock_ | ./moss.rt: line 1: 3383 Aborted /home/clock/brlcad-7.8.4/src/rt/rt -M $* -o moss.pix tetrax.g tetrax 2>>moss.log <moss.rtanim |
| 11:14.49 | clock_ | Do you link BRL-CAD with libefence and try it? |
| 11:51.20 | clock_ | I found the bug |
| 11:51.28 | clock_ | free(pixmap) must be followed by pixmap=0. |
| 11:51.51 | clock_ | It was accessing a freed buffer. |
| 11:55.58 | clock_ | Now it creates all 53 frames :) |
| 15:54.03 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/ (NEWS src/rt/do.c): (log message trimmed) |
| 15:54.03 | CIA-5 | BRL-CAD: awesome.. not only did karel (aka clock3) identify and report on an animation |
| 15:54.03 | CIA-5 | BRL-CAD: bug in rt with extensive detail, he went on to trace down the problem and |
| 15:54.03 | CIA-5 | BRL-CAD: actually identified the bug. a variable that keeps track of frame progress was |
| 15:54.03 | CIA-5 | BRL-CAD: not being reset to null after a frame was completed causing subsequent frames to |
| 15:54.06 | CIA-5 | BRL-CAD: dereference an invalid pointer and segfault. apply the fix from sf bug 1629285 |
| 15:54.08 | CIA-5 | BRL-CAD: (persistent segfaults during animation - with a fix). this affected all |
| 15:55.35 | clock_ | lol :) |
| 16:33.02 | *** join/#brlcad clock_ (i=clock@84-72-60-48.dclient.hispeed.ch) |
| 16:34.22 | clock_ | brlcad: I was filling in a nice detailed bureaucratic patch entry to brl-cad.org but I wanted to verify with a Firefox that the C file I posted is accessible and it told me to select a viewer for a C file (!) so after about 30 seconds of crunching on /usr/local/bin I selected view. Subsequently the X hanged (because Firefox ran the view on the console). |
| 16:34.30 | clock_ | So I lost my nicely detailed bureaucratical entry |
| 16:34.54 | clock_ | So please add pix-y4v into BRL-CAD from http://ronja.twibright.com/utils/pix-y4m.c |
| 16:35.02 | clock_ | I just wrote it and it's handy for making videos. |
| 16:35.11 | clock_ | pix-y4v -> pix-y4m |
| 16:52.39 | *** join/#brlcad _jack- (i=jack@dub.minimal.nl) |
| 17:03.19 | *** join/#brlcad jack- (i=jack@dub.minimal.nl) |
| 18:29.10 | *** join/#brlcad DanielFalck (n=dan@pool-71-111-98-172.ptldor.dsl-w.verizon.net) |
| 18:38.05 | *** join/#brlcad docelic (n=docelic@212.91.112.33) |
| 19:18.57 | *** join/#brlcad ibot (i=ibot@pdpc/supporter/active/TimRiker/bot/apt) |
| 19:46.19 | clock_ | brlcad: here? |