| 01:12.07 | brlcad | updates the ideas list http://brlcad.org/wiki/Google_Summer_of_Code/Project_Ideas |
| 01:12.20 | brlcad | it's closer to a "final initial" form now |
| 02:22.51 | *** join/#brlcad Axman6 (n=Axman6@pdpc/supporter/student/Axman6) |
| 04:01.54 | hippieindamakin8 | nice list man |
| 04:02.47 | brlcad | coming together bit by bit |
| 04:03.01 | yukonbob | looks |
| 04:07.34 | yukonbob | brlcad: interested in my latest build issues, or should I plug away solo? |
| 04:10.07 | yukonbob | http://pastebin.ca/953538 |
| 04:22.06 | yukonbob | that test code needs an "#include <stdlib.h>" for exit() |
| 05:09.24 | yukonbob | (/me assumes introduced at 30538, configure.ac)? But seems completely ac-generated... wft? |
| 05:09.30 | yukonbob | *wtf |
| 05:10.17 | brlcad | yukonbob: yeah weird |
| 05:10.43 | brlcad | the sanity check is missing a header it seems .. that C mode is fine with but C++ mode is adament about |
| 05:11.00 | brlcad | there was a new C++ mode sanity check added just a couple days ago (right after 7.12.0 |
| 05:12.34 | CIA-33 | BRL-CAD: 03brlcad * r30554 10/brlcad/trunk/m4/compiler.m4: bah, need stdlib.h for exit().. C++ sanity check seems insistent on needing it. |
| 05:12.42 | yukonbob | ya -- is BC_SANITY_CHECK a builtin, or user-supplied |
| 05:13.29 | brlcad | all BC_'s are brl-cad macros in m4/ |
| 05:13.33 | yukonbob | whatever you just modded could stand to lose stdio, I believe, too -- probably somebody's finger-memory got the best of them at that point.. |
| 05:13.44 | brlcad | true |
| 05:13.56 | brlcad | stdio was providing exit() on older glibc |
| 05:14.46 | CIA-33 | BRL-CAD: 03brlcad * r30555 10/brlcad/trunk/m4/compiler.m4: don't need stdio.h |
| 05:14.47 | yukonbob | doesn't remember that; if it does/did, so be it, but that won't matter on *BSD, etc. They'll need stdlib |
| 05:15.39 | yukonbob | might end up learning autotools if he's not careful here |
| 05:16.46 | yukonbob | does my commit-bit extend to code as well, or strictly doc? |
| 05:17.00 | brlcad | to anything |
| 05:17.01 | yukonbob | hasn't tried, and we've not talked about this |
| 05:17.32 | brlcad | don't worry about mixing up other stuff, dive right in ;) |
| 05:19.07 | yukonbob | alright; I'm still getting that itcl error that was the initial issue, but working on other stuff too (like the above c++); autogen.sh isn't liking my build env (pkgsrc), but doesn't mind a by-hand run from the shell; I'll explore that too -- (autogen itself hasn't changed, but it must include some changed files, because it used to work np) |
| 05:19.14 | brlcad | all commits are reviewed anyways and announced here so it's easy enough to see what's going on |
| 05:19.43 | yukonbob | nods |
| 05:19.46 | brlcad | does "sh autogen.sh -v" show anything? |
| 05:21.11 | yukonbob | can check; in the "environment" build, it chokes on autoreconf, then tries it's "manual config" (or whatever) and chokes on that, too. |
| 05:21.44 | yukonbob | by-hand, everything works fine... which is strange to me, but I'll track down what the offending difference is... |
| 05:22.13 | brlcad | -v will show the exact commands it's trying to run |
| 05:22.29 | yukonbob | nods, add to wrapping-Makefile |
| 05:22.50 | yukonbob | *adds |
| 05:23.23 | brlcad | which is a version check for autoconf, automake, and libtool .. then something like "autoreconf -v -i -s" .. and then individual steps if that fails |
| 05:24.22 | yukonbob | _has_ autoreconf, and even supplied it directly via the AUTORECONF env var, but no dice |
| 05:25.09 | yukonbob | runs build |
| 05:25.10 | brlcad | autoreconf can fail for lots of reasons |
| 05:25.46 | brlcad | that's one of many reasons why autogen.sh exists in the first place |
| 05:25.59 | yukonbob | out of comlete curiousity, how much effort do you think it'd take to replace auto* w/ cmake (for example) and what would keep one from doing that, besides the manpower? |
| 05:26.09 | yukonbob | *complete |
| 05:27.07 | brlcad | manpower to replicate the entire build system faithfully, I'd expect it'd take many months |
| 05:27.16 | brlcad | of non-stop effort |
| 05:27.29 | yukonbob | ! |
| 05:28.13 | brlcad | the conversion to autotools took a long time, about that long, and there wasn't really any complicated road blocks |
| 05:28.35 | yukonbob | everybody likes to complain about "autohell"; what's your take on it? |
| 05:28.42 | brlcad | it just takes a lot of work to specify the build rules for 400+ binary targets, the 24 or so libs, and various types of data files and subdependencies |
| 05:30.31 | brlcad | autohell is the way it is for many reasons |
| 05:31.27 | brlcad | it does what it does the way it does it in order to provide strict portability constraints (that really cannot be easily provided most any other way without adding dependencies) |
| 05:32.52 | yukonbob | have you ever played w/ scons (python), or cmake? |
| 05:32.58 | brlcad | yep |
| 05:34.16 | brlcad | i actually rather like cmake, their biggest claim to awesomeness is thier ability to export the various build systems (including msvc files) from one description |
| 05:34.50 | brlcad | scons is nice for small code sets but *seriously* starts to fall apart on large complex build systems -- you basically end up reimplementing autoconf |
| 05:35.16 | yukonbob | ya -- /me is ready to buy the cmake manual (just released (new version)) and check it out... |
| 05:37.29 | brlcad | like a lot of tools though, autotools really do work (and work well once you get over the learning curve) and it really starts to make sense why things are the way they are |
| 05:38.31 | brlcad | their biggest downfall has been practical usability problems (horrible error messages, bad docs, elitist responses to usability) |
| 05:39.17 | yukonbob | speaking of docs, can you recommend any? The GNU-supplied stuff isn't my cup of tea. |
| 05:44.30 | brlcad | I actually recommend just reading existing build files for projects that use the autotools -- it's fairly straight-forward seeing it in other apps as there are really only two or three files types involved |
| 05:44.35 | brlcad | the configure.ac and the Makefile.am files |
| 05:45.38 | brlcad | the Makefile.am files don't really get *any* simpler in any of the autotools/cmake/ant/scons/whatever build systems as it's just a list of targets, sets of files, and how those sets of files relate to those targets |
| 05:47.49 | brlcad | the build/feature/header tests and project identification are somewhat tricky in all the systems to replicate -- that's where configure.ac comes into play for the autotools |
| 05:48.55 | brlcad | which is just shell syntax with macros |
| 05:49.22 | yukonbob | nods -- will probably start poking this stuff in BRLCAD, and probably pick your brain too, where possible ;) |
| 05:49.32 | brlcad | feel free to |
| 05:50.40 | brlcad | I've read the GNU book and the auto* manuals, but I really never learned it until I put it to use |
| 05:51.32 | brlcad | the tools are overly flexible, making it really complicated to document since you can approach things in a variety of ways (even though 90% of projects all do it about the same) |
| 05:52.01 | yukonbob | that might be my problem -- the gnu manuals are drier than toast (and technical manuals don't have to be that way); maybe those in conjunction w/ some interesting "real life" work will click |
| 05:52.29 | brlcad | in hindsight after you've put it to use, it does all make sense |
| 05:52.36 | brlcad | even becomes somewhat overly obvious |
| 05:52.46 | brlcad | but is totally opaque until you put it to use |
| 05:56.38 | yukonbob | gets ready for sleep |
| 05:56.38 | brlcad | usually much more effective to just look at a project that is already set up (like brl-cad or something even more simple) and follow their example |
| 05:56.54 | yukonbob | nice chatting again, brlcad; probably talk tmo, or soon at least. |
| 05:56.56 | yukonbob | :) |
| 05:57.00 | brlcad | aiight |
| 05:57.01 | brlcad | cheers |
| 06:47.50 | CIA-33 | BRL-CAD: 03brlcad * r30556 10/brlcad/trunk/HACKING: the dutch cad4linux site seems to be now defunct |
| 07:10.07 | CIA-33 | BRL-CAD: 03brlcad * r30557 10/brlcad/trunk/HACKING: add devmaster to the notification list |
| 08:57.35 | *** join/#brlcad homovulgaris (n=ca3fe93d@bz.bzflag.bz) |
| 09:00.44 | brlcad | uploads some new images to the gallery |
| 09:43.46 | brlcad | ensleeps |
| 09:51.41 | *** join/#brlcad cad22 (n=52f7631d@bz.bzflag.bz) |
| 09:56.28 | *** part/#brlcad cad22 (n=52f7631d@bz.bzflag.bz) |
| 14:15.47 | *** join/#brlcad Elperion (n=Bary@p54875E2A.dip.t-dialin.net) |
| 15:58.51 | *** join/#brlcad vedge (n=vedge@vedge.org) |
| 16:14.45 | brlcad | time for a sunday feast! |
| 16:36.04 | yukonbob | morning brlcad... happy Easter |
| 16:37.06 | poolio | happy purim :) |
| 16:37.30 | yukonbob | looks up purim |
| 16:46.01 | brlcad | likewise |
| 16:46.33 | brlcad | bah, can't find the original bradley rendering... |
| 17:10.30 | yukonbob | would love to see that |
| 17:17.23 | yukonbob | http://my.brlcad.org/gallery/gallery/s/historic/particles.jpg.html |
| 17:18.03 | yukonbob | ^-- what is this, and it reminds me of a question I've had for a while: what's the 'atom' for? |
| 17:27.23 | *** join/#brlcad Elperion (n=Bary@p54875E2A.dip.t-dialin.net) |
| 17:28.11 | brlcad | frankly I don't know really :) the image is an old one from bill laut .. don't know what he used it for, I'd have to ask him |
| 17:28.27 | brlcad | as for the atom? which/what atom? |
| 17:33.11 | *** join/#brlcad Elperion (n=Bary@p54875E2A.dip.t-dialin.net) |
| 17:35.53 | brlcad | few more images uploaded |
| 17:39.02 | yukonbob | is sure he's seen an "atom" primitive |
| 17:39.45 | yukonbob | also doesn't have a copy of BRL-CAD installed atm, and is heading out for easter brunch... chat later cadheads |
| 17:41.50 | *** join/#brlcad Elperion (n=Bary@p54875E2A.dip.t-dialin.net) |
| 17:53.27 | *** join/#brlcad Elperion (n=Bary@p54875E2A.dip.t-dialin.net) |
| 20:15.48 | *** join/#brlcad Elperion (n=Bary@p54875E2A.dip.t-dialin.net) |
| 21:56.20 | yukonbob | http://pastebin.ca/954422 |
| 21:56.34 | yukonbob | ^--- build failure at g-3dm |
| 22:03.40 | *** join/#brlcad louipc (n=louipc@bas8-toronto63-1128543675.dsl.bell.ca) |
| 22:12.44 | yukonbob | traces to src/conf/.deps/g-3dm.Po... |
| 22:13.03 | yukonbob | ...or does he? |
| 22:17.28 | yukonbob | wonders if 3dm is only in development... |
| 22:19.02 | yukonbob | looks like g-3dm isn't ready, but it included (erroneosly) in build... |
| 22:19.31 | yukonbob | *erroneously |
| 22:26.51 | yukonbob | puts this on hold -- will return, test idea, and commit fix (if found) later. |
| 23:14.19 | *** join/#brlcad Elperion (n=Bary@p54875E2A.dip.t-dialin.net) |
| 23:18.05 | *** join/#brlcad Elperion (n=Bary@p54875E2A.dip.t-dialin.net) |
