| 00:44.20 | CIA-9 | BRL-CAD: 03brlcad * 10brlcad/sh/shtool: treat screen as vt100, avoid the error message about determining how to do bold |
| 00:59.21 | CIA-9 | BRL-CAD: 03brlcad * 10brlcad/m4/epsilon.m4: so far, only freebsd seems to use doubles for floats .. so revert the leniancy and report when implementations deviate from what should be expected from IEEE compliance. |
| 01:22.41 | CIA-9 | BRL-CAD: 03brlcad * 10brlcad/src/other/blt/src/bltInit.c: Tcl_GetAssocData returns a Clientdata, so have to cast to long to avoid mipspro type check failure |
| 01:30.32 | CIA-9 | BRL-CAD: 03brlcad * 10brlcad/doc/README.IRIX: no longer need to disable blt, build problem fixed (simple cast) |
| 01:34.23 | CIA-9 | BRL-CAD: 03brlcad * 10brlcad/TODO: sgi knobs and example geometry are pushed back as (sigh) clone still isn't done and EF plugin and archer are going to require considerably more attention. prepare for 7.8.2 release, nothing left to do. |
| 01:38.26 | CIA-9 | BRL-CAD: 03brlcad * 10brlcad/src/libpkg/pkg.c: quell unused variable warning |
| 01:47.21 | CIA-9 | BRL-CAD: 03brlcad * 10brlcad/src/conv/bot_shell-vtk.c: quell unused variable warning (norm variable) |
| 02:00.06 | CIA-9 | BRL-CAD: 03brlcad * 10brlcad/src/conv/bot_shell-vtk.c: missed a norm |
| 02:00.21 | CIA-9 | BRL-CAD: 03brlcad * 10brlcad/src/conv/g-adrt.c: quell unused variable warnings |
| 02:01.34 | CIA-9 | BRL-CAD: 03brlcad * 10brlcad/src/rt/viewweight.c: quell unused var warning |
| 02:05.07 | CIA-9 | BRL-CAD: 03brlcad * 10brlcad/src/rt/viewxray.c: quell unsigned type warning |
| 02:08.19 | CIA-9 | BRL-CAD: 03brlcad * 10brlcad/src/libdm/dm_obj.c: USE_FBSERV is now defined by configure, quell set but unused warning on id var. |
| 02:08.20 | brlcad | well then.. |
| 02:08.26 | brlcad | looks like arl just lost power |
| 02:08.47 | brlcad | and comcast.. |
| 02:24.01 | Twingy | wee |
| 02:24.05 | Twingy | no |
| 02:24.10 | Twingy | it's part of the new security protocol |
| 02:24.31 | Twingy | if you turn the power off, you don't get hackers |
| 02:24.38 | Twingy | it's brilliant! |
| 02:28.24 | brlcad | hrm.. seems i only lost connectivity to some servers.. ones that should have survived a browning |
| 02:29.22 | brlcad | ahh, servers are fine, some router is dead |
| 02:29.39 | brlcad | i can hop over with the right combination |
| 02:31.11 | brlcad | and that router seems to be back now |
| 02:32.18 | Twingy | I suspect I'll be painting, sanding, and spackling all day tomorrow |
| 03:16.18 | Twingy | messy |
| 03:25.27 | Twingy | If you have a time machine for sale, I would like to buy it. Send an e-mail. |
| 03:25.27 | Twingy | Please, serious inquiries only |
| 18:37.02 | *** join/#brlcad ibot (i=ibot@rikers.org) |
| 18:37.02 | *** topic/#brlcad is http://brlcad.org/ || BRL-CAD is now Free Software! || Release 7.8.0 is Posted! |
| 20:55.24 | CIA-9 | BRL-CAD: 03brlcad * 10brlcad/src/conv/g-adrt.c: quell more unused variable warnings |
| 20:56.40 | CIA-9 | BRL-CAD: 03brlcad * 10brlcad/src/burst/burst.c: only define sigcld if signal.h doesn't define it |
| 21:00.17 | CIA-9 | BRL-CAD: 03brlcad * 10brlcad/src/gtools/g_transfer.c: remove unused variables |
| 21:01.22 | Twingy | nothing like working on muves on a saturday |
| 21:02.41 | CIA-9 | BRL-CAD: 03brlcad * 10brlcad/src/mged/points/process.c: remove unused variables, quell warnings |
| 21:04.44 | CIA-9 | BRL-CAD: 03brlcad * 10brlcad/src/mged/utility2.c: removed unused variable i |
| 21:06.12 | CIA-9 | BRL-CAD: 03brlcad * 10brlcad/src/nirt/if.c: quell incompatible parammeter type warning |
| 21:07.19 | Twingy | hrmph |
| 21:07.40 | Twingy | this is rather significant speed up |
| 21:08.07 | Twingy | 4.15x speed up, I guaranteed 2x |
| 21:08.22 | Twingy | haven't optimized anything yet either |
| 21:10.14 | CIA-9 | BRL-CAD: 03brlcad * 10brlcad/src/util/bwscale.c: quell type warnings on bu_free |
| 21:10.17 | Twingy | 43% memory savings too |
| 21:14.50 | ``Erik | 4.15 with or w/o ir ? |
| 21:16.29 | Twingy | w/o |
| 21:16.45 | Twingy | and half the code |
| 21:17.00 | Twingy | 4.15x speed up, 43% memory savings, half the code |
| 21:17.12 | Twingy | and I'm not even trying :) |
| 21:17.36 | Twingy | I've met my requirements though, so I should be done this week |
| 21:18.13 | Twingy | erik, you should base your M3 benchmarks off the ahem newest version of s2 |
| 21:22.26 | ``Erik | heh, indeed |
| 21:22.37 | ``Erik | although that task has been transfered to the new tl |
| 21:22.46 | Twingy | ah |
| 21:23.10 | ``Erik | apparently, if someone gives them bad news, they move to find someone else for a 'second opinion'... until they hear what they want to hear. |
| 21:23.18 | ``Erik | which is a good chunk of what drove mike to ssvt |
| 21:23.20 | ``Erik | :) |
| 21:23.25 | Twingy | haha, so dilbertesque |
| 21:23.47 | Twingy | atleast we don't get locked in a hole and shot |
| 21:24.11 | ``Erik | yet |
| 21:24.53 | Twingy | hrm |
| 21:25.12 | Twingy | if there were a way to know how many triangles were in the .g file that would save a good chunk of prep time |
| 21:25.41 | Twingy | save on alot of realloc's |
| 21:26.09 | Twingy | instead I'll set tie_init to something like 256k triangles by default, cater to small models and super fast prep |
| 21:29.00 | ``Erik | if you realloc 2*prev each step, would that cut the # of reallocs down enough to be usable? |
| 21:29.26 | ``Erik | or if you loaded into dynamic store, like a linked list of 'pages', then counted pages to generate the 'final' area? |
| 21:29.55 | Twingy | 2* is good for small values |
| 21:30.07 | Twingy | I have had several versions in there |
| 21:30.13 | Twingy | you're not teaching me anything new |
| 21:30.28 | ``Erik | heh, 2* is good for big values, too... physical pages aren't wired until they're written to |
| 21:30.31 | Twingy | I think I will go back to 8MB hunks |
| 21:30.36 | Twingy | no... |
| 21:30.48 | Twingy | if you're at 1GB and your algorithm goes, whomp 2GB |
| 21:30.53 | Twingy | and the model is 1.1GB |
| 21:31.12 | Twingy | you just ran out of memory |
| 21:31.12 | ``Erik | um, ... |
| 21:31.37 | Twingy | so now you gotta add code to set an upper bound based on your memory limit |
| 21:31.54 | Twingy | 8MB chunks, and that's that |
| 21:33.20 | ``Erik | no oom |
| 21:33.21 | ``Erik | sorry |
| 21:33.30 | ``Erik | memory isn't wired until it's written to |
| 21:34.44 | Twingy | for windows, for linux, for bsd, for irix, for... ? |
| 21:34.51 | ``Erik | fbsd |
| 21:34.58 | ``Erik | it SHOULD be the same on linux, and probably irix |
| 21:35.01 | ``Erik | I d'no about windows, heh |
| 21:35.44 | ``Erik | it's a byproduct of the mmu and COW, any system with virtualized memory SHOULD behave in a similar fashion... :/ |
| 21:35.53 | ``Erik | (which means old dos would break violently) |
| 21:36.08 | CIA-9 | BRL-CAD: 03brlcad * 10brlcad/src/util/ (pl-X.c pl-sgi.c): quell lots of warnings. |
| 21:38.09 | CIA-9 | BRL-CAD: 03brlcad * 10brlcad/src/mged/points/process.c: ws |
| 21:42.58 | CIA-9 | BRL-CAD: 03brlcad * 10brlcad/src/mged/points/process.c: remove the usage if you're going to remove the decl |
| 21:58.23 | Twingy | getting close to dinner time, and another round of sanding in the garage |
| 22:07.50 | brlcad | time to go drinking |
| 22:08.19 | ``Erik | <-- sips his beer |
| 22:10.00 | Twingy | I have soda and iced-tea |
| 22:10.03 | ``Erik | my kitchen is a lot nicer when clean |
| 22:10.09 | Twingy | the possibilities are endless! |
| 22:11.09 | Twingy | let's see, 2 hours of sanding and painting, 1 hour of basement bathroom work, followed by an hour of kitchen floor work |
| 22:11.22 | Twingy | what a fun night |
| 22:11.49 | Twingy | back at 10pm |
| 22:54.49 | *** join/#brlcad brick_ (n=brick@dsl-146-236-32.telkomadsl.co.za) |
| 22:55.23 | brick_ | hi |
| 22:56.05 | brick_ | Does BRL include tetrahedral meshing capabilities, or is some external app needed? |