| 00:01.49 | starseeker | Oooo http://www.mail-archive.com/pdcurses-l@lightlink.com/msg00129.html |
| 00:02.04 | starseeker | http://www.projectpluto.com/win32a.htm |
| 00:04.53 | CIA-43 | BRL-CAD: 03tbrowder2 * r42285 10/brlcad/trunk/AUTHORS: added my name and details to developer list |
| 00:06.31 | CIA-43 | BRL-CAD: 03brlcad * r42286 10/brlcad/trunk/NEWS: |
| 00:06.32 | CIA-43 | BRL-CAD: tom added sorting to the mged attr show command so that attributes are now |
| 00:06.32 | CIA-43 | BRL-CAD: displayed in alphabetical order instead of unsorted creation order. should |
| 00:06.33 | CIA-43 | BRL-CAD: improve readability. rewording NEWS line to fit on one line (two line comments |
| 00:06.33 | CIA-43 | BRL-CAD: are the exception when there are multiple authors) and moving to top of stack so |
| 00:07.03 | CIA-43 | BRL-CAD: it's in proper chronological order. |
| 00:08.33 | brlcad | starseeker: give it a try, pick one of the curses apps and try to compile against it by hand |
| 00:09.00 | starseeker | nods |
| 00:09.34 | starseeker | wonders... if that works, why not combine it with something like the netbsd curses library to have one portable curses |
| 00:10.13 | brlcad | because we don't really use curses.. :) |
| 00:10.35 | starseeker | our build requires curses.h though |
| 00:10.45 | brlcad | we use the predecessor that was absorbed into curses |
| 00:11.02 | brlcad | strictly speaking, we don't need curses.h |
| 00:11.10 | starseeker | blinks |
| 00:11.13 | brlcad | it's sufficient, not required |
| 00:11.22 | brlcad | so we look for it |
| 00:11.36 | starseeker | what do we do if it isn't found? |
| 00:11.58 | brlcad | we look for all the other various alternatives |
| 00:12.06 | brlcad | surely you reviewed that part of configure.ac |
| 00:12.17 | brlcad | cmake needs the same logic |
| 00:12.45 | starseeker | yeah - I think I just call the find_package(Curses) logic |
| 00:13.32 | brlcad | we have four main sections in configure.ac written for the search logic needed to do it right |
| 00:14.00 | brlcad | it's not equivalent to just search for curses by any stretch |
| 00:14.09 | starseeker | mutter... |
| 00:14.22 | starseeker | votes we stuff curses in src/other and call it done |
| 00:15.04 | brlcad | if that's less work than adding in the header/lib checks, then the build system has something wrong with it |
| 00:15.23 | brlcad | curses is way overkill |
| 00:15.32 | brlcad | we already have libtermlib which is sufficient for all but windows |
| 00:15.54 | starseeker | ah, so libtermlib is the fallback if all else fails? |
| 00:16.13 | brlcad | you really should be more familiar with the configure.ac |
| 00:16.15 | brlcad | locig |
| 00:16.19 | brlcad | logic! |
| 00:16.25 | starseeker | that part had me somewhat confused |
| 00:16.42 | brlcad | four big sections, follow "curses" |
| 00:18.56 | starseeker | something's gummed up then, 'cause I did a build on Linux with trunk (autotools) and it died due to not having any of the term related headers (iirc) |
| 00:20.05 | starseeker | ah - both HAVE_NCURSES_H and HAVE_TERM_H were not defined |
| 00:22.21 | brlcad | now if you could get *one* curses impl that was cross-platform, that might be a contender for replacing termlib, but it'd need to be small and basically portable to everything |
| 00:22.28 | brlcad | right now, windows is the only odd-ball out |
| 00:22.35 | starseeker | nods |
| 00:23.14 | brlcad | pulling in that win32-specific pdcurses and pdcurses would be a bit ridiculous |
| 00:24.00 | starseeker | shakes his head - pdcurses doesn't sound very good for unix/linux terminals |
| 00:24.05 | brlcad | since termlib is already sufficient, already managed, and has been trivial to maintain |
| 00:24.43 | starseeker | perhaps we can bolt the win32 parts of that win32 native pdcurses onto libtermlib |
| 00:25.23 | starseeker | hrm - libcursor is failing because both of those two HAVE_* defines aren't set |
| 00:25.32 | brlcad | possible, but it'd likely amount to a good bit of libtermlib work -- don't see it just dropping in |
| 00:26.03 | brlcad | libcursor is ours |
| 00:26.26 | starseeker | right, but shouldn't it be falling back on libtermlib in the absence of the other two? |
| 00:30.48 | starseeker | O.o |
| 00:30.51 | brlcad | it does |
| 00:30.58 | starseeker | not here - build haulted |
| 00:31.16 | brlcad | which build? |
| 00:31.22 | starseeker | trunk, autotools |
| 00:31.57 | brlcad | you have to look at the configure testing results |
| 00:32.33 | brlcad | whether it thinks there's a usable library, headers, etc |
| 00:33.18 | brlcad | how it ended up enabling/disabling termblib building |
| 00:33.28 | starseeker | oh, it's enabled |
| 00:33.59 | brlcad | it's got the most complex testing logic because it's the oldest dependency |
| 00:34.48 | brlcad | huh, looks like a HAVE_TERMLIB_H define was removed at some point |
| 00:35.31 | brlcad | ah, no I'm wrong -- it's still there |
| 00:35.42 | starseeker | but am I understanding correctly that with libtermlib enabled, ncurses.h and term.h are not required? |
| 00:35.53 | brlcad | right |
| 00:36.02 | starseeker | so something is haywire |
| 00:36.23 | brlcad | it should be hitting line 3380, which turns on libtermlib aka termcap |
| 00:37.26 | brlcad | everything you need should to figure it out should be in the config log |
| 00:43.04 | starseeker | the termlib functionality tests failed, and libtermlib is enabled |
| 00:45.00 | starseeker | it's almost as if it's only looking at HAVE_NCURSES_H and HAVE_TERM_H without going to the else case on the HAVE_NCURSES_H test |
| 00:49.12 | starseeker | brlcad: it IS turning on libtermlib, and building it. |
| 00:50.08 | brlcad | I presume you mean curses.c is failing compilation? |
| 00:50.19 | starseeker | here's the failed configure test that turned it on: http://pastebin.mozilla.org/941860 |
| 00:50.26 | starseeker | cursor.c |
| 00:50.32 | starseeker | in libcursor |
| 00:50.35 | brlcad | right, that's what I meant |
| 00:51.16 | brlcad | well that error is certainly informative |
| 00:51.44 | brlcad | it's exactly why one should never ignore warnings |
| 00:52.16 | *** join/#brlcad juanman (~quassel@unaffiliated/juanman) |
| 00:52.17 | brlcad | it's directly reporting no less than two bugs, possibly more |
| 00:52.26 | starseeker | is the problem using if instead of ifdef? |
| 00:59.10 | starseeker | bingo |
| 01:00.00 | starseeker | shakes his head - I thought perhaps you could get away with if or ifdef... |
| 01:02.43 | *** join/#brlcad crazy_imp (~mj@a89-182-28-13.net-htp.de) |
| 01:03.27 | CIA-43 | BRL-CAD: 03starseeker * r42287 10/brlcad/trunk/src/ (burst/Sc.c libcursor/cursor.c): Fixes to ifdef in cursor.c, expand the logic for Sc.c in burst. |
| 01:04.57 | CIA-43 | BRL-CAD: 03starseeker * r42288 10/brlcad/trunk/configure.ac: fix if->ifdef in configure.ac too... |
| 01:11.24 | starseeker | O.o - that line in libcursor dates back to 2005... apparently a long time since someone's had a system config that hit that case |
| 01:11.42 | starseeker | gets outta here before he gets in any worse trouble... |
| 01:16.53 | brlcad | yeah, that's very VERY curious that the preprocessor skips both the if AND the else clause because it's not defined |
| 01:17.06 | brlcad | but then it warned about it too ;) |
| 01:18.05 | brlcad | that's the first I've seen a preprocessor do that actually.. usually "if empty" will hit the else, that's probably why it never got caught |
| 01:40.26 | ``Erik | interesting, those metaball type errors I was seeing only show up in --enable-optimized |
| 01:49.31 | *** join/#brlcad juanman (~quassel@unaffiliated/juanman) |
| 02:26.57 | brlcad | yeah, all the uninitialized use warnings only occur during optimized too |
| 02:54.52 | Ralith | hm |
| 02:54.56 | Ralith | can't seem to compile latest svn |
| 03:09.45 | CIA-43 | BRL-CAD: 03tbrowder2 * r42289 10/brlcad/trunk/doc/docbook/system/mann/en/oed.xml: add note about using in a script |
| 03:31.26 | CIA-43 | BRL-CAD: 03brlcad * r42290 10/brlcad/trunk/AUTHORS: rewrite the introduction with more details on the purpose and content of the authorship file. |
| 03:39.14 | CIA-43 | BRL-CAD: 03brlcad * r42291 10/brlcad/trunk/AUTHORS: |
| 03:39.14 | CIA-43 | BRL-CAD: not quite so fast... :) getting listed as a developer is reflective not |
| 03:39.15 | CIA-43 | BRL-CAD: prospective. generally takes several hundred commits over sustained effort. |
| 03:39.15 | CIA-43 | BRL-CAD: you can track progress at http://www.ohloh.net/p/brlcad/contributors in the |
| 03:39.16 | CIA-43 | BRL-CAD: meantime (and you're already credited under code contributors). |
| 03:39.31 | brlcad | Ralith: you have the power to fix that ;) |
| 03:39.44 | Ralith | brlcad: yep, 'cept it didn't seem to be giving me an error O.o |
| 03:39.48 | Ralith | does make redirect stdout/stderr? |
| 03:40.07 | brlcad | make doesn't, but gcc sends different messages to each |
| 03:45.56 | CIA-43 | BRL-CAD: 03brlcad * r42292 10/brlcad/trunk/doc/docbook/system/mann/en/oed.xml: the 'e' command is discouraged in documentation, instead referring to its synonym, 'draw' |
| 03:52.29 | Ralith | well, make was failing with no obvious output whatsoever :/ |
| 03:52.40 | Ralith | is GBS still supported? |
| 03:53.21 | CIA-43 | BRL-CAD: 03starseeker * r42293 10/brlcad/branches/cmake/ (CMakeLists.txt db/CMakeLists.txt doc/CMakeLists.txt): Few left over updates that didn't get committed from the home machine. |
| 03:59.10 | starseeker | woot - broke into top 5 on ohloh |
| 03:59.24 | starseeker | did they start tracking the CMake branch or something? :-P |
| 04:09.02 | Ralith | debug mode didn't compile either O.o |
| 04:20.34 | brlcad | Ralith: --disable-warnings |
| 04:20.45 | brlcad | or fix whatever warning is halting the build |
| 04:21.36 | brlcad | strict build is now enabled on several more directories so it'll be a little while as platforms we don't regularly test on get weeded out |
| 04:27.47 | starseeker | loooves merging in date changes... |
| 04:28.04 | starseeker | one at work apparently didn't go through, trying again |
| 05:03.55 | CIA-43 | BRL-CAD: 03starseeker * r42294 10/brlcad/branches/cmake/ (2749 files in 159 dirs): Update cmake branch to trunk r42293. Also got some preliminary efforts at improving the libtermlib handling in the mix, but they aren't ready yet (probably don't work.) |
| 05:08.11 | Ralith | ah. |
| 05:29.26 | *** join/#brlcad juanman (~quassel@unaffiliated/juanman) |
| 06:09.14 | *** join/#brlcad R0b0t1 (~Enigma@64-136-219-55.dyn.everestkc.net) |
| 06:09.14 | *** join/#brlcad R0b0t1 (~Enigma@unaffiliated/r0b0t1) |
| 07:10.37 | *** join/#brlcad PrezKennedy (MK@whitecalf.net) |
| 08:34.30 | *** join/#brlcad WhiteCalf (~MK@whitecalf.net) |
| 08:41.45 | *** join/#brlcad PrezKennedy (~MK@whitecalf.net) |
| 08:53.40 | *** join/#brlcad WhiteCalf (MK@whitecalf.net) |
| 08:58.52 | *** join/#brlcad WhiteCalf (MK@whitecalf.net) |
| 09:18.51 | *** join/#brlcad epileg (~epileg@unaffiliated/epileg) |
| 12:42.28 | CIA-43 | BRL-CAD: 03tbrowder2 * r42295 10/brlcad/trunk/doc/docbook/articles/en/oed.xml: add info on used of oed when scripting |
| 12:44.20 | CIA-43 | BRL-CAD: 03tbrowder2 * r42296 10/brlcad/trunk/doc/docbook/system/mann/en/qorot.xml: correct typo |
| 12:50.22 | CIA-43 | BRL-CAD: 03tbrowder2 * r42297 10/brlcad/trunk/src/burst/Hm.c: quell warning about unused function HmPrntLList |
| 12:58.09 | CIA-43 | BRL-CAD: 03tbrowder2 * r42298 10/brlcad/trunk/src/fb/cat-fb.c: remove duplicate assignment (which also quells compiler warning) |
| 12:59.32 | CIA-43 | BRL-CAD: 03tbrowder2 * r42299 10/brlcad/trunk/src/librt/tree.c: correct typo |
| 13:00.51 | CIA-43 | BRL-CAD: 03tbrowder2 * r42300 10/brlcad/trunk/src/util/bwcrop.c: add cast to unsigned to quell compiler warnings |
| 13:02.11 | CIA-43 | BRL-CAD: 03tbrowder2 * r42301 10/brlcad/trunk/HACKING: correct typo; add missing word |
| 13:04.18 | CIA-43 | BRL-CAD: 03tbrowder2 * r42302 10/brlcad/trunk/INSTALL: correct typo; grammar |
| 13:50.39 | CIA-43 | BRL-CAD: 03tbrowder2 * r42303 10/brlcad/trunk/src/conv/g-xxx_facets.c: change main to standard form; add informative output about tesselation parameters |
| 15:50.55 | CIA-43 | BRL-CAD: 03brlcad * r42304 10/brlcad/trunk/src/burst/Hm.c: dead code can be eliminated. revision control has our back. |
| 15:51.13 | brlcad | awesome browder :) |
| 19:06.45 | CIA-43 | BRL-CAD: 03starseeker * r42305 10/brlcad/branches/cmake/src/other/CMakeLists.txt: That's not the way to handle libtermlib. |
| 19:09.04 | starseeker | is going to have to face the fact that he has at least one more major piece of work ahead of him for CMake... |
| 19:14.56 | starseeker | brlcad: I've been studying the termlib logic, and one thing puzzles me a bit - why does the term.h header test need the other headers included? |
| 19:50.18 | CIA-43 | BRL-CAD: 03erikgreenwald * r42306 10/brlcad/branches/bottie/ (3115 files in 225 dirs): MFC 42305 |
| 20:36.05 | *** join/#brlcad waprat (~packrator@99-67-225-40.lightspeed.livnmi.sbcglobal.net) |
| 20:42.05 | *** join/#brlcad epileg1 (~epileg@188.119.210.222) |
| 21:08.08 | *** join/#brlcad CIA-29 (~CIA@208.69.182.149) |
| 22:41.20 | *** join/#brlcad R0b0t1 (~Enigma@unaffiliated/r0b0t1) |
| 23:44.08 | *** join/#brlcad mafm_ (~mafm@12.Red-80-26-128.dynamicIP.rima-tde.net) |
| 23:47.35 | *** join/#brlcad juanman (~quassel@unaffiliated/juanman) |