| 00:28.36 | CIA-42 | BRL-CAD: 03starseeker * r41028 10/brlcad/branches/cmake/src/rt/CMakeLists.txt: Conditionalize heatgraph.c in libremrt too. |
| 00:33.50 | starseeker | hmm... BU_EXPORT extern int bu_opterr; and friends don't seem to be happy on Windows - at least, terrain.c isn't getting them from linking libbu |
| 00:35.00 | ``Erik | BU_EXPORT is defined by funky msvc fu as msvc requires a funky dll_export flag on funcs |
| 00:35.19 | ``Erik | it might be that the cmake stuff isn't setting something to trip that |
| 00:39.33 | CIA-42 | BRL-CAD: 03starseeker * r41029 10/brlcad/branches/cmake/src/util/CMakeLists.txt: Include the zlib directory in the utils build. |
| 00:45.46 | CIA-42 | BRL-CAD: 03starseeker * r41030 10/brlcad/branches/cmake/src/ (4 files in 4 dirs): Some more zlib includes. |
| 00:46.02 | starseeker | ``Erik: wonder why only on the regex and utahrle stuff |
| 00:46.08 | starseeker | hmm... |
| 00:58.58 | ``Erik | dunno, was just throwing out a common gotcha |
| 00:59.26 | starseeker | you're right, it's gotta be something like that |
| 00:59.34 | starseeker | maybe still missing somethign for libbu's defines... |
| 00:59.39 | ``Erik | I'll be in tomorrie, we can look at it if it's still an issue... indianlarry has some disturbing fu up that alley, too |
| 01:00.08 | starseeker | cool - starting to get down to a few issues like that and the stuff that obviously was never ported in the first place |
| 01:00.19 | starseeker | (anything using libtermlib - that's gonna be some kinda fun) |
| 01:00.22 | ``Erik | in th emean time, I may be feeding this cat to a fish. |
| 01:00.28 | starseeker | hehe |
| 01:00.47 | ``Erik | bob only ported SOME of the BRL-CAD suite, there're probably a lot of executables that got no love |
| 01:01.13 | starseeker | nods - one of the consequences of a cross-platform build system is we're gonna have to do something about the missing ones |
| 01:02.25 | starseeker | ``Erik: hopefully your commercial VS install will be more helpful than the free one |
| 01:02.45 | ``Erik | heh |
| 01:02.54 | starseeker | boots into windows for one last try - see now much the zlib includes fixed |
| 01:03.10 | ``Erik | still a bit annoyed that vs08 is too retarded to think of a static member as a valid function pointer. |
| 01:03.35 | ``Erik | er, vs05, vs8.0 |
| 01:04.11 | ``Erik | has forked both okra and buclet. |
| 01:07.46 | ``Erik | yowza, draft range after pearl harbor went up to 37 O.O |
| 01:09.14 | ``Erik | heh, old stockings were used for casing powder on heavt guns, crazy |
| 01:15.23 | ``Erik | dolittle sure didalot |
| 01:28.55 | starseeker | OK, latest build log: http://pastebin.ca/1965377 |
| 01:29.40 | starseeker | ``Erik: heh - in a WWII situation they'll keep drafting until they have as many as they need |
| 01:30.42 | ``Erik | yes, but this show indicated that immediately, they took everyone from 18 to 37... our curremt law caps at 25 |
| 01:31.13 | starseeker | nods - 'course, these days cannon fodder is less critical than powerful toys |
| 01:46.10 | *** join/#brlcad louipc (~louipc@archlinux/trusteduser/louipc) | |
| 02:21.49 | *** join/#brlcad cosurgi (~cosurgi@atak.bl.pg.gda.pl) | |
| 02:40.53 | brlcad | starseeker: all non-gui apps should probably be _CONSOLE (i.e., everything except bwish, mged, and rtwizard) .. and even those maybe too |
| 02:47.08 | brlcad | starseeker: keep a list of everything that isn't reaily ported/portable on the wiki, grouped by issue if convenient but a simple list should suffice |
| 02:47.26 | brlcad | then the issues can be looked into for those few |
| 02:52.56 | Ralith | ``Erik: forked? Why? |
| 03:19.09 | CIA-42 | BRL-CAD: 03brlcad * r41031 10/brlcad/trunk/TODO: keith verified/fixed vertex fusing changes |
| 03:22.07 | brlcad | sees a bug in brlcad/branches/cmake/src/util/CMakeLists.txt |
| 04:09.25 | brlcad | http://www.sfr-fresh.com/unix/privat/gmsh-2.5.0-source.tgz:a/gmsh-2.5.0-source/utils/converters/brl-cad/README.txt |
| 04:11.10 | brlcad | has support for arb8, tgc, and ell |
| 04:11.21 | brlcad | (and combs of those) |
| 04:56.14 | starseeker | brlcad: I'm not seeing the bug offhand - syntax error or logic error? |
| 04:57.39 | starseeker | must sleep now... |
| 07:33.07 | *** join/#brlcad cosurgi (~cosurgi@atak.bl.pg.gda.pl) | |
| 10:00.51 | d-lo | Mernin all |
| 10:05.49 | brlcad | howdy |
| 10:06.12 | CIA-42 | BRL-CAD: 03brlcad * r41032 10/brlcad/trunk/src/mged/menu.c: remove dead code. |
| 10:14.34 | d-lo | so whats new? |
| 12:27.05 | *** join/#brlcad juanman (~quassel@unaffiliated/juanman) | |
| 12:36.13 | ``Erik | ralith: mods for mac, probably continued dev that aerique is not pursuing... he said to just fork 'em and if the changes are good, he'd fold 'em back to his line |
| 14:45.22 | *** join/#brlcad mafm (~mafm@81.32.97.31) | |
| 14:54.07 | CIA-42 | BRL-CAD: 03starseeker * r41033 10/brlcad/branches/cmake/src/other/CMakeLists.txt: Capitalize Tkhtml target correctly |
| 15:10.29 | *** join/#brlcad cosurgi (~cosurgi@atak.bl.pg.gda.pl) | |
| 15:22.19 | CIA-42 | BRL-CAD: 03starseeker * r41034 10/brlcad/branches/cmake/src/util/CMakeLists.txt: Fix typo in util/CMakeLists.txt |
| 15:29.41 | CIA-42 | BRL-CAD: 03starseeker * r41035 10/brlcad/branches/cmake/src/libged/CMakeLists.txt: Add regex library to libged list. |
| 15:32.09 | brlcad | there you found it |
| 15:32.35 | starseeker | now if only I could figure out what I'm doing wrong with libregex and libutahrle |
| 15:32.54 | brlcad | although the interface flags should not be used outside src/libfb |
| 15:33.19 | starseeker | was copying that from our existing win32 build logic |
| 15:33.43 | brlcad | nods |
| 15:33.49 | starseeker | is stymed for test environment atm - no activestate install |
| 15:33.55 | brlcad | it's probably needed due to the public headers, they use them |
| 15:34.04 | brlcad | they just shouldn't be in the pub headers |
| 15:34.13 | starseeker | ah, gotcha |
| 15:34.34 | brlcad | not your problem but a problem |
| 15:34.49 | starseeker | have we ever had the bu_optind/bu_optarg issue before? |
| 15:34.57 | brlcad | ? |
| 15:35.10 | starseeker | anim_hardtrack.obj : error LNK2001: unresolved external symbol _bu_optind |
| 15:35.15 | starseeker | many other cases |
| 15:35.46 | brlcad | is anim_hardtrack.obj linking libbu? |
| 15:35.53 | starseeker | should be |
| 15:36.03 | brlcad | doesn't sound like it is |
| 15:36.30 | brlcad | either that or the import/export decls are missing the preprocessor toggle |
| 15:36.42 | starseeker | oh, wait - libbn but not libbu |
| 15:36.54 | starseeker | guess libbn isn't pulling in what's needed |
| 15:36.55 | ``Erik | bn should implicitely include bu |
| 15:37.23 | brlcad | not from a dll perspective |
| 15:37.44 | brlcad | those are globals in libbu's dll, so it has to link that dll |
| 15:38.37 | brlcad | there might be a way to specify an __declspec(import) on libbu symbols during libbn compilation, that might make it auto-loadable |
| 15:38.40 | starseeker | dunno that that'll fix it, but let's try |
| 15:38.52 | starseeker | (easy to add libbu to the lists) |
| 15:39.04 | brlcad | you might want to check your dll flags |
| 15:39.15 | brlcad | make sure you understand how they're set for the import/export rules |
| 15:39.30 | brlcad | BU_EXPORT |
| 15:39.48 | starseeker | that's set by the -DBU_EXPORT_DLL flag being supplied I though |
| 15:39.52 | starseeker | thought even |
| 15:40.42 | brlcad | do you know when you need that and when you do not? |
| 15:41.13 | brlcad | BU_EXPORT_DLL is the toggle for import/export when BRLCAD_DLL is set |
| 15:41.24 | CIA-42 | BRL-CAD: 03starseeker * r41036 10/brlcad/branches/cmake/src/anim/CMakeLists.txt: Try explicitly adding libbu to the anim tools to see if Windows has better luck fiding opting and friends. |
| 15:41.28 | brlcad | sometimes it HAS to be import .. so it has to be unset |
| 15:41.40 | brlcad | sometimes it HAS to be export .. so it has to be set |
| 15:41.49 | starseeker | manually in the build logic? |
| 15:41.57 | brlcad | right |
| 15:42.11 | brlcad | the build system is telling it how to build .. that's the toggle |
| 15:42.13 | starseeker | well, I've probably not got that right then |
| 15:42.28 | starseeker | just defined it for libbu - I think that's my only use of it |
| 15:42.54 | brlcad | basically, BRLCAD_DLL has to be set everywhere for a windows build iirc |
| 15:42.59 | starseeker | (auuuugh - why does Windows have to suck so bad??? no other platform has these issues) |
| 15:43.12 | starseeker | brlcad: That one I think I do have set toplevel |
| 15:43.13 | starseeker | checks |
| 15:43.15 | brlcad | that sounds right, you export for that lib |
| 15:43.22 | brlcad | but only for that lib |
| 15:44.08 | brlcad | so if you're compiling libbn, it'll have BRLCAD_DLL but BU_EXPORT_DLL is unset and BN_EXPORT_DLL is set |
| 15:44.08 | starseeker | right |
| 15:44.17 | brlcad | then compiling anim_track, neither is set, except BRLCAD_DLL |
| 15:44.29 | starseeker | BRLCAD_DLL is set in src/CMakeLists.txt, and the *_EXPORT_DLL definitions are added by misc/BRLCAD_Util.cmake |
| 15:44.36 | starseeker | right |
| 15:44.44 | brlcad | okay, sounds like you have it right then |
| 15:45.16 | brlcad | so the export during libbn wasn't enough and it's just needing to list all libraries that have symbols actively used |
| 15:46.02 | brlcad | or perhaps runtime loading DOES work, but you still have to list libbu's .lib file during linkage |
| 15:46.09 | starseeker | looks like it - let me swat any others like that I can spot. Still doesn't explain the regex/rle issues though - unless I really messed up those are already explicitly listed where they are needed |
| 15:47.01 | starseeker | 'cept terrain.c already has libbu called out |
| 15:47.06 | starseeker | and still has that failure |
| 15:47.11 | starseeker | growl... |
| 15:47.11 | brlcad | starseeker: for portability, libraries should probably be in decreasing dependency order |
| 15:47.22 | brlcad | i.e., order matters |
| 15:47.28 | brlcad | so rt before bn before bu, etc |
| 15:47.36 | starseeker | oh |
| 15:47.43 | brlcad | all the way down the line |
| 15:47.48 | starseeker | winces |
| 15:48.10 | brlcad | even linux requires that |
| 15:48.36 | starseeker | uh... it must be autosorting then 'cause I sure don't feed 'em in that way |
| 15:48.50 | brlcad | that's why you usually see the libs that have no deps at the end of linkage lists (e.g., -lz) |
| 15:49.01 | ``Erik | recent linux should do it's own dep chaining and stuff these days |
| 15:49.10 | brlcad | yeah, you're just getting lucky |
| 15:49.25 | starseeker | and that might be the cause of issues on Windows? |
| 15:49.29 | brlcad | possible |
| 15:49.30 | ``Erik | there was a time that it did lists, then it did queues |
| 15:49.41 | starseeker | braces himself - this is gonna take a while |
| 15:49.41 | brlcad | it'll be a problem for other OS and compilation flags regardless |
| 15:49.43 | ``Erik | um, iirc, you need to be explicit about every dep lib in windows |
| 15:50.02 | ``Erik | I don't think the order was terribly critical, but the name had to be there |
| 15:50.34 | starseeker | does anybody else have a Windows box they can try this out on? |
| 15:50.56 | starseeker | I'm probably the worst guy on the team for debugging on Windows |
| 15:51.36 | ``Erik | not I, I'm using sl today so at home |
| 15:53.37 | CIA-42 | BRL-CAD: 03starseeker * r41037 10/brlcad/branches/cmake/src/anim/CMakeLists.txt: Swap order of anim deps. |
| 15:57.16 | starseeker | that really really sucks - ordering based on deps is a job the computer should be doing |
| 16:01.25 | CIA-42 | BRL-CAD: 03starseeker * r41038 10/brlcad/branches/cmake/src/conv/CMakeLists.txt: Re-order deps in conv |
| 16:04.27 | CIA-42 | BRL-CAD: 03starseeker * r41039 10/brlcad/branches/cmake/src/conv/ (3 files in 3 dirs): Re-order subdirectories of conv |
| 16:13.31 | CIA-42 | BRL-CAD: 03starseeker * r41040 10/brlcad/branches/cmake/src/ (6 files in 6 dirs): reordering up through libgcv - some of these ordering may not be absolutely right and need tuning later. |
| 16:28.54 | CIA-42 | BRL-CAD: 03starseeker * r41041 10/brlcad/branches/cmake/src/ (13 files in 13 dirs): More reordering |
| 16:29.43 | brlcad | wdb is before rt |
| 16:31.59 | brlcad | configure.ac lists all of the deps in proper dependency order around line 3993 |
| 16:32.48 | brlcad | (so you can see who depends on what, not to repeat the logic) |
| 16:33.34 | starseeker | brlcad: do you know what platforms this'll make a difference on? |
| 16:37.21 | starseeker | should test on those if possible, since OSX, BSD and Gentoo all missed it :-( |
| 16:39.42 | brlcad | any slightly older bsd/linux should have it in theory |
| 16:40.06 | CIA-42 | BRL-CAD: 03starseeker * r41042 10/brlcad/branches/cmake/src/ (5 files in 5 dirs): Fix some more ordering issues. |
| 16:40.17 | brlcad | as in probably any kernel/linker a couple years old |
| 16:40.31 | ``Erik | I think one of my cats is retarded. |
| 16:40.34 | brlcad | rhel3's was old enough iirc |
| 16:41.14 | ``Erik | was working on this losi T pro thingy, set the exacto knife with a #11 razor down, she slaps the business end of the knife. she hit the back of the blade so didn't get sliced up, but zomfg wtff |
| 16:41.16 | brlcad | building with a different compiler will likely hit it |
| 16:41.25 | brlcad | could try compiling with intel compiler |
| 16:41.32 | ``Erik | try tendra! |
| 16:42.31 | ``Erik | the worst and oldest fbsd instance you'll see is bz, 5.2.1 was a crappy hotpatch to a bad release |
| 16:42.45 | starseeker | is cmake on bz? |
| 16:42.51 | ``Erik | it can be |
| 16:42.58 | starseeker | let's do it |
| 16:43.16 | ``Erik | um, I might have to reinstall enough of the ports system to do that, if brlcad approves |
| 16:43.21 | starseeker | (if there's disk space, don't think there was last time) |
| 16:43.38 | ``Erik | someone cleaned house, there's a lot of space now |
| 16:43.47 | ``Erik | 7.5 gigs on usr |
| 16:44.28 | starseeker | ``Erik: if you don't need Qt for cmake-gui, hopefully cmake itself won't need much... |
| 16:45.50 | ``Erik | no, the fbsd default is to disable the gui |
| 16:46.24 | starseeker | ``Erik: if you prefer I can try building cmake from source |
| 16:46.49 | ``Erik | qt is only an acceptable default among qt folk... the rest of the universe shuns it |
| 16:46.52 | ``Erik | :> |
| 16:47.18 | starseeker | <snort> - well, on a server any GUI toolkit is kinda outa place |
| 16:47.59 | ``Erik | compiling now, if brlcad has an issue, well, we'll deal with it later |
| 16:48.11 | starseeker | we can un-install once testing is done |
| 16:48.48 | ``Erik | I'll leave it... |
| 16:49.15 | ``Erik | might take a bit, it's an older machine and is quite busy with mysql and www sloppiness |
| 16:49.24 | ``Erik | imma nice 20 it, too |
| 16:49.28 | starseeker | nods |
| 16:50.41 | ``Erik | a bit sad that I'm more productive when I call in sick than when I come in and deal with a shitload of useless bullshit. *sigh* |
| 16:50.46 | ``Erik | anyone wanna be iaso? |
| 16:50.49 | ``Erik | and sa2? |
| 16:50.50 | ``Erik | and ... |
| 16:52.57 | ``Erik | brlcad.org/~erik/rccar |
| 16:53.24 | ``Erik | (it went straight to cmake build no qt crap... but it's going vrry vrry slow) |
| 16:53.53 | starseeker | hah, cool - what do the cats make of it? |
| 16:54.21 | ``Erik | it doesn't move yet |
| 16:54.41 | ``Erik | I need to find my soldering gun, and I need to buy one more part... I'm sure it'll be terrifying |
| 16:55.39 | ``Erik | I have a couple micro-t's, the 1:36 version, those freak them out... I bought a cheap r/c car at target, it sucks and they're more amused at how pathetic it is, I think... a 1:24 car and the turnin radius is like 10 feet??? wtf |
| 16:55.42 | CIA-42 | BRL-CAD: 03brlcad * r41043 10/brlcad/trunk/ (50 files in 2 dirs): |
| 16:55.42 | CIA-42 | BRL-CAD: put the LIKELY/UNLIKELY compiler hints into practice for libbu. these help the |
| 16:55.42 | CIA-42 | BRL-CAD: compiler's branch prediction logic for optimized builds for cases where an |
| 16:55.42 | CIA-42 | BRL-CAD: expression is nearly always true or false. added to most of the checks that |
| 16:55.42 | CIA-42 | BRL-CAD: result in a bomb (should be very unlikely) as well as a lot of function entry |
| 16:55.43 | CIA-42 | BRL-CAD: input sanity tests. |
| 16:55.44 | CIA-42 | BRL-CAD: tests before and after are showing a small consistent performance boost of 2-5% |
| 16:55.52 | ``Erik | "newbright" is pure crap. don't waste your money. |
| 16:57.02 | starseeker | brlcad: nice! |
| 16:57.27 | ``Erik | now if only he'd put some effort into the 3 server migrations he's been saddled with O:-) |
| 16:57.43 | brlcad | is busy preparing another gs brief |
| 16:58.38 | brlcad | i actually want to be working on the servers, just several items came up with much higher priority this past two months |
| 17:03.52 | CIA-42 | BRL-CAD: 03brlcad * r41044 10/brlcad/trunk/NEWS: |
| 17:03.53 | CIA-42 | BRL-CAD: increased the optimized build performance by putting LIKELY/UNLIKELY compiler |
| 17:03.54 | CIA-42 | BRL-CAD: hints into practice for libbu. these help the compiler's branch prediction |
| 17:03.54 | CIA-42 | BRL-CAD: logic for optimized builds for cases where an expression is nearly always true |
| 17:03.55 | CIA-42 | BRL-CAD: or false. added to most of the checks that result in a bomb (should be very |
| 17:03.55 | CIA-42 | BRL-CAD: unlikely) as well as a lot of function entry input sanity tests. |
| 17:03.56 | CIA-42 | BRL-CAD: tests before and after are showing a small consistent performance boost of 2-5% |
| 17:07.21 | starseeker | grabs food while cmake compiles on bz |
| 17:12.21 | CIA-42 | BRL-CAD: 03brlcad * r41045 10/brlcad/trunk/ (NEWS src/util/Makefile.am src/util/query.1 src/util/query.c): |
| 17:12.21 | CIA-42 | BRL-CAD: remove the 'query' command prompt input tool. it's a specialized version of |
| 17:12.21 | CIA-42 | BRL-CAD: 'read' (with defaults) that, while useful for scripting, used SIGALRM and |
| 17:12.21 | CIA-42 | BRL-CAD: alarm() in its implementation making it non-portable to Windows without |
| 17:12.22 | CIA-42 | BRL-CAD: maintenance effort. since it's a burden, not in major use, and has drop-in |
| 17:12.22 | CIA-42 | BRL-CAD: replacements available, remove it from the package. |
| 17:14.34 | Ralith | ``Erik: ah, kk |
| 17:20.32 | CIA-42 | BRL-CAD: 03brlcad * r41046 10/brlcad/trunk/src/util/pixfade.c: remove all globals |
| 17:22.54 | *** join/#brlcad mafm_ (~mafm@81.32.97.31) | |
| 17:25.01 | CIA-42 | BRL-CAD: 03brlcad * r41047 10/brlcad/trunk/src/libcursor/cursor.c: windows compat -- check HAVE_SYS_IOCTL_H before including it. |
| 17:28.32 | *** join/#brlcad cosurgi (~cosurgi@atak.bl.pg.gda.pl) | |
| 17:34.33 | CIA-42 | BRL-CAD: 03brlcad * r41048 10/brlcad/trunk/src/util/pl-tek.c: |
| 17:34.33 | CIA-42 | BRL-CAD: remove the inlined sleep(3) declaration since it breaks the build on windows |
| 17:34.33 | CIA-42 | BRL-CAD: (there's a sleep() macro for windows that converts to Sleep()). instead, |
| 17:34.33 | CIA-42 | BRL-CAD: conditionally include unistd.h like should have been done in the first place. |
| 17:34.33 | CIA-42 | BRL-CAD: also reorder args to avoid all forward decls. |
| 17:40.11 | CIA-42 | BRL-CAD: 03brlcad * r41049 10/brlcad/trunk/src/util/ (11 files): |
| 17:40.12 | CIA-42 | BRL-CAD: remove all of the forward declarations for bu_opt* (i.e., bu_optind, bu_optarg, |
| 17:40.12 | CIA-42 | BRL-CAD: bu_opterr, and bu_getopt()) as including these as extern declarations will break |
| 17:40.12 | CIA-42 | BRL-CAD: the Windows build (missing BU_EXPORT label needed to import the symbol). the |
| 17:40.12 | CIA-42 | BRL-CAD: header already declares them, though, so just remove them. there are probably a |
| 17:40.12 | CIA-42 | BRL-CAD: lot more of these cases. |
| 17:42.30 | CIA-42 | BRL-CAD: 03brlcad * r41050 10/brlcad/trunk/src/util/pixfade.c: inpp name tweak |
| 17:54.33 | *** join/#brlcad cosurgi (~cosurgi@atak.bl.pg.gda.pl) | |
| 18:30.13 | starseeker | ``Erik: is it still compiling? |
| 18:32.15 | CIA-42 | BRL-CAD: 03brlcad * r41051 10/brlcad/trunk/bench/pixcmp.c: |
| 18:32.16 | CIA-42 | BRL-CAD: remove the getopt global var declarations as they bust the windows build. there |
| 18:32.16 | CIA-42 | BRL-CAD: was some platform (maybe solaris?) that required them at some point, but remove |
| 18:32.16 | CIA-42 | BRL-CAD: for now regardless. if readded, they will need __declspec(dllimport). |
| 18:46.24 | CIA-42 | BRL-CAD: 03starseeker * r41052 10/brlcad/branches/cmake/ (73 files in 9 dirs): Sync cmake branch to r41051 |
| 18:46.58 | CIA-42 | BRL-CAD: 03starseeker * r41053 10/brlcad/branches/cmake/src/util/CMakeLists.txt: query is gone. |
| 18:55.34 | CIA-42 | BRL-CAD: 03brlcad * r41054 10/brlcad/trunk/src/anim/ (8 files): clean things up for Windows. remove all extern declarations of the bu_getopt globals. remove all forward decls, for that matter, including a slew of unnecessary ones provided by anim.h; reordering accordingly. |
| 18:58.28 | CIA-42 | BRL-CAD: 03brlcad * r41055 10/brlcad/trunk/ (7 files in 2 dirs): replace RTOD/DTOR with vmath's RAD2DEG/DEG2RAD accordingly |
| 19:32.13 | *** join/#brlcad Ralith (~ralith@d142-058-092-056.wireless.sfu.ca) | |
| 19:51.08 | CIA-42 | BRL-CAD: 03starseeker * r41056 10/brlcad/branches/cmake/src/ (3 files in 3 dirs): Try linking in some windows libraries for the symbol __imp__UuidCreate@4 error... |
| 20:19.12 | CIA-42 | BRL-CAD: 03starseeker * r41057 10/brlcad/branches/cmake/src/CMakeLists.txt: Let's see if explicitly setting this define helps any with regex/rle. |
| 20:22.44 | brlcad | http://www.unlogic.se/projects/openicons |
| 20:25.19 | starseeker | brlcad: did you want to swap out some of the Archer icons with those? |
| 20:25.35 | brlcad | nah, just an interesting icon project to note |
| 20:25.56 | starseeker | nods |
| 20:26.22 | starseeker | brlcad: thanks for tackling those windows issues |
| 20:26.32 | brlcad | still working on others |
| 20:26.36 | brlcad | you shouldn't be defining _WIN32 ... that should be coming from the compiler |
| 20:26.53 | starseeker | WIN32 seems to be, but not _WIN32 |
| 20:27.36 | brlcad | WIN32 shouldn't be used (even though our current msvc build files defines it) |
| 20:28.06 | brlcad | http://msdn.microsoft.com/en-us/library/b0084kay(VS.80).aspx |
| 20:29.09 | starseeker | well, rle.h and regex.h both use _WIN32 in the conditional for the DLL_EXPORT define, and unless I'm missing something I'm not sure what other reason there could be for all of the unresolved external symbol errors pertaining to those two libs |
| 20:29.24 | brlcad | really shouldn't be defining *any* preprocessor symbols that begin with an underscore, as the spec says they are reserved by the compiler |
| 20:30.18 | brlcad | you're undoubtedly missing something because defining _WIN32 shouldn't fix it -- and even if it did, then there's badness somewhere else being compounded |
| 20:30.54 | starseeker | brlcad: I'll take another look once I get home... it's frustrating the heck out of me |
| 20:30.55 | brlcad | add a #ifdef _WIN32 .. #error guess it's defined .. #endif and I bet it was already defined |
| 20:31.23 | brlcad | (regardless, defining it is still outright wrong) :) |
| 20:31.55 | starseeker | OK... I'm kinda out of ideas then |
| 20:34.18 | starseeker | O.o server error from sf |
| 20:35.09 | CIA-42 | BRL-CAD: 03starseeker * r41058 10/brlcad/branches/cmake/src/CMakeLists.txt: Sigh - not allowed to define _WIN32, that's up to the compiler. |
| 20:39.30 | brlcad | same two questions as the other libs |
| 20:39.35 | brlcad | make sure the lib is exporting |
| 20:39.39 | brlcad | make sure the binary is importing |
| 20:40.16 | starseeker | -DREGEX_EXPORT_DLL and -DBRLCAD_DLL are explicitly defined in the regex CMakeLists.txt |
| 20:40.29 | brlcad | if you're sure of both (and both can be verified) .. then it's just a matter of listing the .lib during compilation and having the .dll available during runtime |
| 20:41.01 | brlcad | that's not making sure they're defined |
| 20:41.19 | starseeker | you mean in the msvc project file itself? |
| 20:41.32 | brlcad | nope, not even that is certain |
| 20:41.51 | starseeker | uh... |
| 20:41.55 | brlcad | put a #error on the line that defines export or check the actual value with #ifequality testing |
| 20:42.27 | brlcad | then make sure it's defined on one of the lines before it's used |
| 20:42.54 | brlcad | if those two hold, then you can be reasonably certain that it exported |
| 20:43.05 | starseeker | <PROTECTED> |
| 20:43.30 | brlcad | think of it like a printf() saying "I got here with the preprocessor" |
| 20:43.57 | brlcad | #if 0 |
| 20:44.02 | brlcad | #error this will never print |
| 20:44.03 | brlcad | #else |
| 20:44.06 | brlcad | #error this will print |
| 20:44.08 | brlcad | #endif |
| 20:44.22 | starseeker | k, cool |
| 20:44.43 | brlcad | so you make sure it's defined right .. then you can make sure it's still defined where it's used |
| 20:45.19 | starseeker | nods |
| 20:48.29 | brlcad | it might be easier to use #pragma warning ( some general non-halting debug message here ) as an alternative if you don't want it to halt |
| 20:49.43 | brlcad | er, maybe not .. that's right, msvc doesn't support arbitrary messages |
| 20:51.35 | brlcad | ahh, here we go: #pragma comment( user, "some general non-halting message" ) |
| 21:17.06 | CIA-42 | BRL-CAD: 03brlcad * r41059 10/brlcad/trunk/src/anim/ (Makefile.am anim_track.c cattrack.c cattrack.h): |
| 21:17.06 | CIA-42 | BRL-CAD: break out the functions from cattrack.c that are used by anim_track.c into a |
| 21:17.06 | CIA-42 | BRL-CAD: private header file so that the function signatures can be declared in one |
| 21:17.06 | CIA-42 | BRL-CAD: place. move and doxygenify the function comments, reordering the |
| 21:17.07 | CIA-42 | BRL-CAD: implementations to avoid the need for forward decls and calling out the HIDDEN |
| 21:17.07 | CIA-42 | BRL-CAD: ones that are just internal to the implementation. |
| 21:24.23 | CIA-42 | BRL-CAD: 03brlcad * r41060 10/brlcad/trunk/src/anim/ (8 files): quell verbose compilation warnings, reorder to eliminate forward decls |
| 21:26.30 | *** join/#brlcad R0b0t1 (~Enigma@64-136-219-55.dyn.everestkc.net) | |
| 21:26.30 | *** join/#brlcad R0b0t1 (~Enigma@unaffiliated/r0b0t1) | |
| 21:36.15 | *** join/#brlcad Ralith (~ralith@d142-058-092-056.wireless.sfu.ca) | |
| 21:45.29 | CIA-42 | BRL-CAD: 03brlcad * r41061 10/brlcad/trunk/src/anim/ (17 files): more restructuring and cleanup including ws/style consistency cleanup along with forward decl elimination and comment restructuring. |
| 21:49.29 | starseeker | ah, putty CAN do this |
| 21:49.31 | starseeker | excellent |
| 21:58.30 | CIA-42 | BRL-CAD: 03starseeker * r41062 10/brlcad/branches/cmake/src/librt/CMakeLists.txt: Whoops, forgot the quotes. |
| 22:18.50 | CIA-42 | BRL-CAD: 03brlcad * r41063 10/brlcad/trunk/src/tab/ (script-tab.c tabsub.c): quellage and de-k&r cleanup |
| 22:19.12 | CIA-42 | BRL-CAD: 03brlcad * r41064 10/brlcad/trunk/TODO: bu_cv_optimize() should be using bu_byteorder() |
| 22:22.25 | CIA-42 | BRL-CAD: 03brlcad * r41065 10/brlcad/trunk/src/ (16 files in 11 dirs): remove the remainder of places where the bu_opt* family was being declared as extern and were a problem with Windows portability. removed since bu.h declares them portably and properly. |
| 22:33.50 | CIA-42 | BRL-CAD: 03starseeker * r41066 10/brlcad/branches/cmake/ (38 files in 14 dirs): sync cmake branch to trunk r41065 |
| 23:03.59 | *** join/#brlcad Ralith (~ralith@S010600221561996a.vc.shawcable.net) | |
| 23:32.45 | ``Erik | *yawn* cmake failed on bz |
| 23:34.31 | ``Erik | I think it's my fault, not cmakes |
| 23:36.43 | ``Erik | ok, it's installed now |