irclog2html for #brlcad on 20070213

04:13.43*** part/#brlcad deltazap (n=zap@pool-72-64-253-55.tampfl.fios.verizon.net)
04:14.07*** join/#brlcad deltazap (n=zap@pool-72-64-253-55.tampfl.fios.verizon.net)
05:30.12MaloeranThis is lovely.
05:30.16Maloeranlin-lwp.c:1729: internal-error: kill_wait_callback: Assertion `pid == -1 && errno == ECHILD' failed.    A problem internal to GDB has been detected,
11:01.43*** join/#brlcad cad72 (n=549c6492@bz.bzflag.bz)
13:33.56*** join/#brlcad clock_ (n=clock@zux221-122-143.adsl.green.ch)
13:34.11*** join/#brlcad clock_ (n=clock@zux221-122-143.adsl.green.ch)
14:13.51``Erikheh
14:14.39archivist"must try harder" 2 out of 10
14:18.27``Erikmal's progress report, as neutered and de-tech'd by a couple layers of mgmt types... :)
16:27.53*** join/#brlcad ibot_ (i=ibot@pdpc/supporter/active/TimRiker/bot/apt)
16:27.53*** topic/#brlcad is BRL-CAD CSG Modelling... http://www.brlcad.org http://sourceforge.net/projects/brlcad
16:32.26MaloeranCool Erik, can I see that? :)
16:32.38MaloeranOh, you tried to send on Efnet alerady
16:32.41Maloeranalready, too
16:33.01MaloeranAh no, it was the wrapping framework headers
18:10.48``Erika little part, nothing interesting yet...
18:11.14``ErikMAYBE I'll be able to get you the rest this week
18:11.49MaloeranWee, one ray at a time. All right, thanks
18:15.58CIA-5BRL-CAD: 03brlcad * 10brlcad/autogen.sh: report the ident footer (which includes the version of this script now) with --version output
18:29.08CIA-5BRL-CAD: 03brlcad * 10brlcad/autogen.sh: check for autoreconf before autoconv/automake/libtool
18:35.57CIA-5BRL-CAD: 03brlcad * 10brlcad/TODO: improve adrt's integration, ensure compilation/execution works reliably cross-platform -- get latest changes being used by muves teams to libtie. add update of tcl/tk since that is currently going on.
18:54.54``Erikadrt will be equally crippled, dude *shrug* and if we don't get the numbers we want, we'll fix things up better
19:04.21MaloeranADRT was not written to exploit : SSE packed tracing, ray coherency, ray locality
19:04.37MaloeranSo I believe RF will be far more crippled, but we'll see
19:25.19CIA-5libIRC: 03jeffm2501 * 10libirc/ (include/TCPConnection.h src/TCPConnection.cpp): start redoing the TCP server interface to actualy work, and not suck.
20:07.27IriX64make install
20:07.35IriX64sorry :(
20:10.30IriX64wonder if this thing actually runs.
20:11.27IriX64btw if you don't use --prefix, the build puts the html dir in your root dir.
20:12.17IriX64configure reports awf and brlman to have an undefined variable called datarootdir.
20:31.39*** join/#brlcad stunned (n=d1d1dbd1@bz.bzflag.bz)
20:31.44*** part/#brlcad stunned (n=d1d1dbd1@bz.bzflag.bz)
20:33.01*** join/#brlcad stunned (n=d1d1d029@bz.bzflag.bz)
20:38.31*** join/#brlcad clock_ (i=clock@84-72-95-194.dclient.hispeed.ch)
20:51.28brlcadIriX64: what platform/shell has that behavior?
20:51.39IriX64bash
20:51.57brlcadunder which OS?
20:52.03IriX64cygwin
20:52.06brlcadhmm
20:52.30brlcadgrep BRLCAD_DATA doc/html/Makefile
20:52.36brlcadwhat does that output?
20:52.37IriX64ty
20:53.29IriX64just sitting there.
20:53.48brlcader
20:54.03brlcadyou did type the doc/html/Makefile part, yes?
20:54.09IriX64goof (meaning me just a sec.)
20:54.28brlcadthat should be run from the directory that has configure
20:54.44IriX64no such file or directory.
20:54.59brlcad"that should be run from the directory that has configure"
20:55.11IriX64it is
20:55.36brlcadls doc/html
20:56.00brlcad[morrison@hole (Tue Feb 13 15:55:26) ~/brlcad]$ ls  doc/html
20:56.01brlcadCVS             Makefile        Makefile.am     Makefile.in     ReleaseNotes    manuals
20:56.19IriX64ls doc/htmlmake file.am .in releasenotes and manual.
20:56.57brlcadhow come you don't have a Makefile ?  did you run make distclean or something?
20:57.07IriX64yep
20:57.20brlcadthat's kinda important to tell me ;)
20:57.30IriX64whup apologies :)
20:57.51brlcadgrep BRLCAD_DATA doc/html/Makefile.in
20:58.15brlcadand .. grep BRLCAD_DATA doc/html/manuals/Makefile.in
20:58.28brlcador just the latter even
20:58.38brlcadshould be two lines
20:58.38brlcadBRLCAD_DATA = @BRLCAD_DATA@
20:58.39brlcadhtmldocsdir = $(BRLCAD_DATA)/html/manuals
20:58.54brlcaddo you see something different?
20:59.01IriX64BRLCAD_DATA  = @BRLCAD_DATA@
20:59.17IriX64only one line
20:59.30brlcadin which file
20:59.37IriX64.in
20:59.40brlcaddoc/html will be just one line
20:59.47brlcaddoc/html/manuals should have two
21:00.23IriX64has none.
21:00.51brlcadhuh?  you ran: grep BRLCAD_DATA doc/html/manuals/Makefile.in
21:00.55brlcadand it reported no lines?
21:01.02IriX64reported one line.
21:01.14brlcadread carefully, you're telling me two things
21:01.36brlcadgrep BRLCAD_DATA doc/html/manuals/Makefile.in    <-- note the manuals part
21:01.58IriX64my mistake both only have one line.
21:02.15IriX64not used to grep man ;)
21:02.30brlcadgrep BRLCAD_DATA doc/html/manuals/Makefile.am    outputs what?
21:03.42IriX64htmldocsdir = $(BRLCAD_DATA)/html/manuals
21:04.23brlcadthen "grep BRLCAD_DATA doc/html/manuals/Makefile.in" cannot have only one line ...
21:04.45brlcads/cannot/really should not/
21:05.05IriX64specifying --prefix is a workaround.
21:05.42brlcadi got that much, trying to figure why something that basic would fail
21:06.40brlcadwhen it "puts it in the root" .. what dir does it create and where?
21:06.58IriX64html in the root dir of the system.
21:07.22IriX64<PROTECTED>
21:07.29brlcadhmmm.. maybe if...
21:08.21*** join/#brlcad SWPadnos (n=Me@dsl245.esjtvtli.sover.net)
21:08.27IriX64just went by again, i'm configuring.
21:09.35brlcadgrep usr/brlcad configure
21:09.59brlcadis that this?  ac_default_prefix=/usr/brlcad
21:10.24IriX64ac_default_prefix=/usr/brlcad
21:11.13IriX64shall i paste the warning?
21:11.17brlcadorsure
21:11.33IriX64config.status: WARNING: src/brlman/awf contains a reference to the variable `dat
21:11.33IriX64arootdir'
21:11.33IriX64which seems to be undefined.  Please make sure it is defined.
21:11.33IriX64config.status: creating src/brlman/brlman
21:11.33IriX64config.status: WARNING: src/brlman/brlman contains a reference to the variable `
21:11.36IriX64datarootdir'
21:11.38IriX64which seems to be undefined.  Please make sure it is defined.
21:13.04brlcadcan you pastebin or send me your awf or brlman ?
21:13.52IriX64brlman? brlman.1? or brlman.in?
21:14.13brlcadbrlman
21:14.23IriX64how do i deliver it to you?
21:14.32brlcadwhat works for you?
21:14.46brlcad~pastebin
21:14.55ibotpastebin is, like, a place to paste your stuff without flooding the channel - try  http://pastebin.ca, or http://channels.debian.net/paste, or http://rafb.net/paste/
21:14.55brlcadwould be the most convenient
21:14.55IriX64dunno haven't transfreed files in s *long time :)
21:15.55brlcadif you go to one of those three urls, you can paste the contents of brlman
21:16.00brlcador awf
21:17.09IriX64pastebin.ca has it.
21:17.31brlcadheh, you're supposed to provide me that url it tells you
21:17.40IriX64arrgh.
21:18.22brlcadafter you hit submit post
21:18.39brlcadyou're not one of the recent ones.. doesn't look like you posted it yet
21:19.19IriX64just posted it again.
21:19.49IriX64http://www.pastebin.ca/354242
21:20.14brlcadthere we go.. that's better :)
21:20.21IriX64:)
21:21.05IriX64bbiab nature yells :)
21:21.35``Erikheh
21:21.47brlcadpeculiar.. your BRLCAD_DATA is getting set to ${datarootdir}
21:21.50``Erikpaste.lisp.org is cooler, and has a 'report' bot on this network... O:-)
21:24.02brlcadIriX64: grep datadir=\' configure
21:32.18brlcadIriX64: grep BRLCAD include/brlcad_config.h
21:32.19IriX64${datarootdir}
21:32.34brlcadyeah, that's wonky..
21:32.48IriX64not your problem ?
21:33.00brlcadI think I can compensate for it
21:33.04brlcadwhat's: grep BRLCAD include/brlcad_config.h
21:34.15IriX64$ grep BRLCAD include/brlcad_config.h
21:34.15IriX64#define BRLCAD_DATA "${datarootdir}"
21:34.16IriX64#define BRLCAD_ROOT "/usr/solariscad"
21:34.19IriX64#define BRLCAD_VERSION "7.8.4"
21:35.24brlcadeek
21:35.32brlcadgood to know, that's a separate problem
21:35.38IriX64$... ?
21:36.35IriX64autoconf?
21:36.47brlcadyeah
21:36.57IriX64version 2.6
21:37.50brlcadthe autoconf you're using with cygwin is setting datadir to a variable that must be evaluated else you end up with variable turds like you're seeing
21:38.18IriX64ill get me a new autoconf or fix this one, i have source.
21:38.23brlcadnot so much a bug on their part but inconsistent behavior (they don't do that elsewhere afaihs) that we'll need to compensate for
21:38.29brlcad2.6 doesn't sound right
21:38.42IriX642.60
21:38.46brlcadahh
21:38.56brlcadokay, that could/would be why
21:39.04brlcadthat's a newer (protocol-breaking) version
21:39.17IriX64thanks for this one i set prefix=solariscad.
21:39.35brlcadthis should fix it
21:39.49IriX64does already tested it.
21:40.29brlcadheh, no.. i mean for noticing the problem.. I have a fix here going in
21:40.38IriX64oh sorry.
21:40.41IriX64:)
21:44.52CIA-5BRL-CAD: 03brlcad * 10brlcad/configure.ac:
21:44.52CIA-5BRL-CAD: compensate for an autoconf inconsistency issue where apparently newer 2.60+
21:44.52CIA-5BRL-CAD: autoconf is setting datadir to datarootdir as a default, causing a need to
21:44.53CIA-5BRL-CAD: evaluate datadir before using it with things like AC_DEFINE and AC_SUBST values.
21:44.53CIA-5BRL-CAD: do an explicit eval echo like is done for the summary printing.
21:44.54CIA-5BRL-CAD: thanks to IriX64 for noticing the problem and helping pinpoint the cause. the problem affected most BRLCAD_DATA installation paths as well as awf and brlman (which use BRLCAD_DATA).
21:48.48``Erikw00t
21:51.55CIA-5BRL-CAD: 03brlcad * 10brlcad/configure.ac: reorganize so the eval expansion happens only once.
22:32.23IriX64question.. is there some issue I don't know about that affects shared library builds? every time I try to build shared libraries, I get undefined references.
22:36.17CIA-5libIRC: 03jeffm2501 * 10libirc/ (include/TCPConnection.h src/TCPConnection.cpp): get the TCP server class to read data and call it's listeners.
23:02.39*** join/#brlcad b0ef (n=b0ef@084202025057.customer.alfanett.no)
23:43.56CIA-5libIRC: 03jeffm2501 * 10libirc/ (include/TCPConnection.h src/TCPConnection.cpp):
23:43.56CIA-5libIRC: data send methods for server connected peers.
23:43.56CIA-5libIRC: disonection functions for server connected peers.
23:43.56CIA-5libIRC: disconection notification for servers.
23:51.51CIA-5BRL-CAD: 03brlcad * 10brlcad/BUGS:
23:51.51CIA-5BRL-CAD: there does seem to be a ray-trace consistency failure in the solids.sh
23:51.51CIA-5BRL-CAD: regression test case where it fails on Mac OS X with one pixel exhibiting an
23:51.51CIA-5BRL-CAD: off-by-many error on the edge of the middle ARB8. this only seems to occur when
23:51.51CIA-5BRL-CAD: using an optimized -O3 compile (and using -fno-unsafe-math-optimizations seems
23:51.53CIA-5BRL-CAD: to make no difference) possibly indicating a floating point/tolerance issue that
23:51.55CIA-5BRL-CAD: needs to be accounted for so that results are exactly repeatable regardless.
23:55.36MaloeranExactly repeatable? Picky :)
23:56.09MaloeranThe compiler could just be shuffling instructions around, meaning the ray that was hitting the primitive no longuer does
23:58.06``Erikwhich means the optimizer broke it, and that's important to know...
23:59.41MaloeranOptimizers shuffle instructions around when it leads to the "same" result, mathematically speaking

Generated by irclog2html.pl by Jeff Waugh - find it at freshmeat.net! Modified by Tim Riker to work with blootbot logs, split per channel, etc.