| 00:17.43 | louipc | yeah I should probably disable jove |
| 00:47.52 | *** join/#brlcad poolio (n=poolio@c-71-236-67-238.hsd1.pa.comcast.net) | |
| 02:13.14 | *** join/#brlcad louipc (n=louipc@67.68.54.228) | |
| 03:21.05 | *** join/#brlcad SWPadnos_ (n=Me@dsl245.esjtvtli.sover.net) | |
| 03:36.50 | *** join/#brlcad louipc (n=louipc@67.68.54.228) | |
| 04:51.35 | *** join/#brlcad poolio__ (n=poolio@c-71-236-67-238.hsd1.oh.comcast.net) | |
| 05:09.51 | *** join/#brlcad poolio_ (n=poolio@c-71-236-67-238.hsd1.oh.comcast.net) | |
| 05:48.21 | *** join/#brlcad Z80-Boy (i=clock@77-56-68-229.dclient.hispeed.ch) | |
| 07:02.28 | *** join/#brlcad IriX64 (n=mariodot@bas2-sudbury98-1177871835.dsl.bell.ca) | |
| 07:17.30 | *** join/#brlcad Z80-Boy (n=clock@zux221-122-143.adsl.green.ch) | |
| 11:09.19 | *** join/#brlcad d_rossberg (n=rossberg@bz.bzflag.bz) | |
| 11:53.46 | *** join/#brlcad poolio_ (n=poolio@c-71-236-67-238.hsd1.oh.comcast.net) | |
| 12:27.30 | *** join/#brlcad Maloeran (n=maloeran@glvortex.net) [NETSPLIT VICTIM] | |
| 12:27.30 | *** join/#brlcad ``Erik (i=erik@c-69-250-155-85.hsd1.md.comcast.net) [NETSPLIT VICTIM] | |
| 12:27.30 | *** join/#brlcad dtidrow_work (n=dtidrow@host169.objectsciences.com) [NETSPLIT VICTIM] | |
| 12:41.50 | *** join/#brlcad elite01 (n=elite01@dslc-082-082-089-134.pools.arcor-ip.net) | |
| 14:38.11 | *** join/#brlcad IriX64 (n=IriX64@bas2-sudbury98-1177871835.dsl.bell.ca) | |
| 15:08.49 | IriX64 | http://irix32.spaces.live.com/photos/cns!C456D36896FA7A58!141/?startingImageIndex=0&commentsExpand=0&addCommentExpand=0&addCommentFocus=0&pauseSlideshow=0 <---- brlcad and vista64 :) (my new affliction) :) |
| 15:09.52 | IriX64 | mmm maybe this is better ... http://www.irix32.spaces.live.com/photos |
| 15:34.59 | *** join/#brlcad poolio_ (n=poolio@c-71-236-67-238.hsd1.pa.comcast.net) | |
| 15:38.42 | *** join/#brlcad ``Erik (i=erik@c-69-250-155-85.hsd1.md.comcast.net) | |
| 15:38.42 | *** join/#brlcad dtidrow_work (n=dtidrow@host169.objectsciences.com) | |
| 15:41.28 | brlcad | "Erik may be interested in the fact that Jutland is the mainland of Denmark and definitely no part of Germany" |
| 15:43.36 | CIA-27 | BRL-CAD: 03brlcad * 10brlcad/regress/Makefile.am: delete the solids.rt.log file |
| 16:17.22 | *** join/#brlcad dtidrow (n=dtidrow@host131.objectsciences.com) | |
| 16:44.13 | ``Erik | hehehehe http://rfc.net/rfc3514.html |
| 16:46.03 | ``Erik | brlcad: I'm ready too much historical stuff lately :D northern jutland is where archeologists think the cimbri lived, who were a germanic people who caused a lot of grief for the roman empire :D |
| 16:55.57 | *** join/#brlcad Maloeran (n=maloeran@glvortex.net) | |
| 17:04.54 | *** join/#brlcad AcI1 (n=1@net12.mlyniec.gda.pl) | |
| 17:04.56 | *** part/#brlcad AcI1 (n=1@net12.mlyniec.gda.pl) | |
| 19:42.28 | CIA-27 | BRL-CAD: 03brlcad * 10brlcad/src/lgt/lgt.c: remove dead iris code |
| 20:14.30 | *** join/#brlcad Z80-Boy (i=clock@77-56-80-224.dclient.hispeed.ch) | |
| 20:27.29 | *** join/#brlcad kwizart (n=kwizart@fedora/kwizart) | |
| 21:01.06 | *** join/#brlcad louipc (n=louipc@bas8-toronto63-1096782509.dsl.bell.ca) | |
| 21:42.01 | starseeker | For my fellow Gentoo users: BRL-CAD ebuild is bumped to latest version: http://bugs.gentoo.org/show_bug.cgi?id=77197 |
| 21:43.01 | louipc | I think I want a BRL-CAD baseball cap with the big eagle logo :D |
| 21:46.47 | Maloeran | BRL-CAD is in portage now? Neat, I always just compiled |
| 21:48.14 | louipc | heh yeah I compiled too |
| 21:49.27 | louipc | does the new brl-cad website seem kind of crowded to you? |
| 21:49.54 | louipc | I could do without the whole bottom section :/ |
| 21:50.02 | starseeker | Not in portage, on the bugzilla for Gentoo |
| 21:50.16 | starseeker | I'd like it to be in portage, but so far I can't attract the interest of a developer |
| 21:50.27 | starseeker | I could try BECOMING a developer, but that takes time... |
| 21:50.42 | louipc | are there set guidelines to become a developer? |
| 21:50.54 | kwizart | well how do you solve the tcl tk probem with 8.5 not stable ? |
| 21:51.03 | starseeker | Sort of - you need a mentor and pass some quizzes, etc. |
| 21:51.08 | starseeker | I use the internal copy |
| 21:51.12 | louipc | I don't think my distro has any. It's just "hey we like you, you're a dev" |
| 21:51.17 | starseeker | I install in opt |
| 21:51.24 | kwizart | shared libs ? |
| 21:51.43 | kwizart | x86 or can you test x86_64 also ? |
| 21:51.59 | starseeker | BRL-CAD's configure script is intelligent - if you have a correct tcl/tk installed it will use it, otherwise it builds a local copy for brl-cad |
| 21:52.09 | CIA-27 | BRL-CAD: 03brlcad * 10brlcad/src/ (38 files in 8 dirs): yet another surge of level 4 flawfinder string length checks. somewhere around 30-40% done. |
| 21:52.22 | kwizart | well usually, we still have tcl tk 8.4 |
| 21:52.27 | brlcad | louipc: sure, see the HACKING file |
| 21:53.10 | louipc | brlcad: :D I meant gentoo dev, but thanks |
| 21:53.18 | kwizart | for now it seems to work but i need to genertate sh autgen.sh |
| 21:53.25 | brlcad | louipc: ah :) |
| 21:53.29 | kwizart | and fix: sed -i -e 's|/usr/brlcad/bin/btclsh|%{_bindir}/tclsh|' $RPM_BUILD_ROOT%{_datadir}/tclscripts/mged/points.tcl |
| 21:53.44 | kwizart | and sed -i -e 's|/bin/tclsh|%{_bindir}/tclsh|' $RPM_BUILD_ROOT%{_datadir}/tclscripts/geometree/{geometree.tcl,GeometryBrowser.tcl} |
| 21:54.15 | starseeker | I have the ebuild call autoconf, and there's a patch for tcl/tk needed as well (someone else found the solution to that) |
| 21:54.45 | kwizart | using tcl 8.4 or last 8.5b1 ? |
| 21:54.56 | starseeker | 8.5 |
| 21:55.05 | starseeker | the internal copy in BRL-CAD's cvs |
| 21:55.18 | starseeker | or tarball, actually... |
| 21:56.04 | kwizart | actually i'm using 85b1 system...(shared libs) but i cannot compile itcl itk system so i'm using brlcad for theses |
| 21:56.38 | kwizart | i need also to rename wall |
| 21:56.39 | kwizart | mv $RPM_BUILD_ROOT%{_bindir}/wall $RPM_BUILD_ROOT%{_bindir}/brlcad-wall |
| 21:56.54 | kwizart | since it conflict with wall command |
| 21:57.14 | brlcad | kwizart: curious those fixes .. are they specific to an environment? |
| 21:57.24 | louipc | brlcad: btw I just figured that you wouldn't need the md5sum in the archlinux PKGBUILD included in brlcad src. It would kind of be like making a cvs pkg which doesn't have that check either. I will fix it. |
| 21:57.37 | brlcad | are more importantly, anything that can be done to change our sources so you don't have to make your fixes down the road? |
| 21:57.45 | brlcad | (wall is a known one, just haven't gotten to it) |
| 21:57.56 | brlcad | but the other seds are new |
| 21:57.59 | kwizart | (installed in /usr since it aims to use Fedora guidelines - to be added for F9 maybe ) |
| 21:58.45 | brlcad | kwizart: that is ill-advised still at the moment for fedora if you have conflicting libs |
| 21:58.56 | brlcad | namely librt, libbu, and libbn |
| 21:59.08 | kwizart | brlcad, actually rpm checks for scripts shebang, if the command do not exist, then there is a problem with requires at rpm install |
| 21:59.51 | brlcad | I don't mean commands, I mean specifically libs |
| 22:00.25 | brlcad | librt is the most common conflict |
| 22:00.28 | kwizart | hum... |
| 22:00.45 | brlcad | ours predates, but the kernel compat lib is found on many systems |
| 22:01.02 | kwizart | then that would be fine to change SONAME |
| 22:01.14 | kwizart | [root@kwizatz bin]# rpm -qf /usr/lib64/librt.so |
| 22:01.14 | kwizart | glibc-devel-2.6-4 |
| 22:01.23 | brlcad | ours is a ray-trace library, the other is a real-time kernel library that is deprecated |
| 22:01.53 | kwizart | maybe i will raise this problem, to see if we can remove it.. |
| 22:01.56 | kwizart | for F9 |
| 22:01.59 | brlcad | exactly my point, care should be taken to not clobber the system lib or lots will be broken |
| 22:02.04 | kwizart | but maybe compat will need it |
| 22:02.07 | brlcad | or yeah, remove it |
| 22:02.30 | kwizart | but shouldn't it be easier to change SONAME ? |
| 22:02.53 | brlcad | for us, it's are most fundamental library and will not likely be changed on our end |
| 22:02.58 | kwizart | configure could hadle this properly (saw some sample somewhere ) |
| 22:03.03 | brlcad | our primary user base relies on it and expects it |
| 22:03.40 | kwizart | could do some search to provide a clean way to handle that... |
| 22:03.52 | kwizart | yes i don't mean to remove it |
| 22:03.52 | louipc | kwizart: Fedora doesn't have some standard means of working around conflicts like that? |
| 22:04.11 | brlcad | I mean even to rename it, our librt goes back nearly 25 years of active use |
| 22:04.49 | kwizart | well, the namespace is a problem, if there is a lot, then might be better to have libs installed in /usr/lib(64)/brlcad/ |
| 22:04.53 | brlcad | the approach I'm hoping to eventually add will be something like haveing a /usr/lib/brlcad where our libs would reside |
| 22:05.25 | kwizart | and having a /etc/ld.so.conf.d/brlcad.conf to provide a clean way to say the path to the ldconfig |
| 22:05.27 | louipc | tjat |
| 22:05.31 | brlcad | we have similar conflicts with binaries, but none of those to date are really problematic fortunately (e.g. wall, can be trivially changed) |
| 22:05.32 | louipc | that'd be alright |
| 22:06.01 | kwizart | but since it is mandatory to have shared libs, then the SONAME have to be different |
| 22:06.31 | *** join/#brlcad IriX64 (n=IriX64@bas2-sudbury98-1177871835.dsl.bell.ca) | |
| 22:07.56 | brlcad | mandatory for whom? |
| 22:07.56 | kwizart | librt.so.19 librt.so.1 for glib, so i think they will not conflict, unless for librt.so |
| 22:08.10 | kwizart | mandatory for pacakges within Fedora... |
| 22:08.54 | brlcad | i.e. the most important one will conflict if they're in the same dir |
| 22:09.19 | brlcad | if they're not in the same directly, then it shouldn't be a problem since our binaries will have rpaths to what they need |
| 22:09.28 | kwizart | that why it can add a prefix |
| 22:09.36 | kwizart | rpath are not allowed neither |
| 22:09.54 | brlcad | eh, "rpath are not allowed" .. what does that mean exactly? |
| 22:10.05 | kwizart | rpmbuidl command block them (as this will prevent to use the improvements with the linker...) |
| 22:10.28 | brlcad | it's a compile-time setting |
| 22:10.30 | kwizart | you cannot approve a package to be in Fedora if there is rpaths within |
| 22:10.43 | starseeker | kwizart: It might be useful to check over the gentoo bug history for brl-cad |
| 22:10.48 | brlcad | that's pretty much any tool that links with libtool |
| 22:10.49 | kwizart | yes it is (and i've removed them in my testing package ) |
| 22:11.29 | kwizart | with chrpath since removing them in libtool will prevent them to link at build time (removed after make install ) |
| 22:11.29 | starseeker | kwizart: brlcad generally has specific ways of doing things that are not subject to change, for very long historical reasons |
| 22:11.40 | louipc | heehee |
| 22:11.50 | brlcad | I somehow highly doubt that is an enforceable rule if there really is such a rule.. :) |
| 22:11.51 | starseeker | kwizart: What about puting it somewhere like opt? |
| 22:12.23 | starseeker | Or whatever Fedora uses for opt? |
| 22:12.43 | starseeker | Surely there must be other packages out there that cannot be installed in /usr without problems |
| 22:12.46 | louipc | yeah there's gotta be some kind of provisions within fedora guidelines to deal with such conflicts sanely |
| 22:13.00 | kwizart | Fedora uses /usr :) |
| 22:13.18 | starseeker | I can speak from personal experience - putting brlcad in /usr is a fools game |
| 22:13.19 | kwizart | fedora packages distributed by Fedora |
| 22:13.29 | brlcad | starseeker: I've yet to see a single one of the major distros actually accommodate name collisions outside of forcing one or the other to change their name |
| 22:13.30 | kwizart | meant /usr/bin /usr/lib(64) |
| 22:13.37 | kwizart | not /usr/brlcad |
| 22:13.39 | starseeker | meh |
| 22:13.55 | kwizart | well that's a packager problem actually... |
| 22:14.00 | starseeker | kwizart: I had to re-install my system twice to clean up messes I made with /usr installs |
| 22:14.15 | louipc | brlcad: ouch |
| 22:14.22 | kwizart | i can handle theses problem... if it works fine.. (and i will have time to checks ) |
| 22:15.00 | kwizart | starseeker, well i install thing in /usr with rpm |
| 22:15.21 | starseeker | The name collision is fundamental. |
| 22:15.29 | louipc | yeah manual install should go in /usr/local non? |
| 22:15.32 | starseeker | It's also rather insidious |
| 22:15.44 | kwizart | louipc, yes indeed |
| 22:15.44 | brlcad | kwizart: I do appreciate the efforts, and would strongly prefer finding a solution that doesn't require changing our product names for our libraries -- binaries we can probably adjust depending on which one, but the libraries have extensive heritage behind them (that predates linux itself) |
| 22:15.48 | starseeker | I had my system keep working after brlcad nuked the libraries, which was even stranger than total failure |
| 22:16.42 | brlcad | it'll install usually with no problem, the issue is usually librt getting replaced, which can render a system unusable |
| 22:17.00 | brlcad | or if openssh is installed and their libbn is installed, and our libbn clobbers it |
| 22:17.01 | louipc | I hear of some people caching the whole system in memory hehe |
| 22:17.49 | louipc | brlcad: that shouldn't happen with the pkgs checking for file conflicts |
| 22:17.54 | brlcad | kwizart: if there's build settings that I can change on our part, please let me know (e.g. the sed issues) |
| 22:18.02 | louipc | 'rpm' as kwizart said |
| 22:18.03 | brlcad | louipc: you'd think |
| 22:20.31 | kwizart | brlcad, first step would be to use /usr/lib(64)/brlcad but it would be fine to have also librt-brlcad.so.19 (i knwo you will be disapointed but since it will be provided by brlcad-config or pkgconfig, does it would matter ?...) |
| 22:20.56 | kwizart | just a thought...^^ |
| 22:22.22 | brlcad | /usr/lib(64)/brlcad would be fine, but changing the .so's name would certainly be a last resort |
| 22:22.55 | brlcad | i mean in all fairness, would you consider renaming glibc's deprecated librt.so librt-glibc.so.19 ? it's just as absurd |
| 22:23.15 | kwizart | i will try to see... |
| 22:23.38 | brlcad | especially since theirs is 'dead' last I read, but even if it wasn't |
| 22:23.50 | kwizart | then openssl also : /lib64/libssl.so.6 ( needs to checks) |
| 22:24.30 | brlcad | there are potentially others as well, those are just two pretty popular/common conflicts that I can name |
| 22:24.39 | kwizart | i think glibc in Fedora 8 is enought recent to drop old things... (Fedora 8 will have glibc 2.7 ) |
| 22:26.09 | CIA-27 | BRL-CAD: 03brlcad * 10brlcad/configure.ac: re-enable the checks for tcl8.4, see if we can get both 8.4 and 8.5 working so beta isn't required. |
| 22:26.29 | kwizart | hum nice... |
| 22:29.25 | CIA-27 | BRL-CAD: 03brlcad * 10brlcad/NEWS: merge in the release notes from STABLE |
| 22:30.29 | CIA-27 | BRL-CAD: 03brlcad * 10brlcad/ChangeLog: merge in 7.10.4 ChangeLog from STABLE |
| 22:30.35 | louipc | woo |
| 22:41.05 | brlcad | woo hoo, down to 650 level 4 issues! |
| 22:41.34 | brlcad | (from about 1700) |
| 23:04.18 | ``Erik | heh |
| 23:05.44 | ``Erik | unfortunately, when I do a fresh checkout of my dev tree and delete all the 'imported' stuff, sloccount says I'm a little less impressive :/ |
| 23:09.31 | louipc | hmm |
| 23:14.00 | *** join/#brlcad yukonbob (n=yukonbob@204-225-123-148.xdsl.convoke.net) | |
| 23:14.20 | ``Erik | SILENCE, MAGGOT! |
| 23:22.02 | yukonbob | ~lart maggots |
| 23:22.02 | ibot | does a little 'dpkg -P maggots' action |
| 23:26.13 | ``Erik | heh |