| 00:03.10 | CIA-43 | BRL-CAD: 03bob1961 * r37539 10/brlcad/trunk/src/other/libpng/ (pngpread.c pngrutil.c pngset.c pngwio.c pngwutil.c): Quell a few warnings when compiling for 64-bit Windows. |
| 00:12.06 | CIA-43 | BRL-CAD: 03bob1961 * r37540 10/brlcad/trunk/src/other/libregex/regcomp.c: Quell a few warnings when compiling for 64-bit Windows. |
| 00:17.01 | starseeker | wonders if there is some way to trim down the number of options to rt - we're almost out of single letter arguments |
| 00:17.48 | starseeker | or maybe add support for the --argument syntax? |
| 00:18.48 | mafm | night |
| 00:20.42 | CIA-43 | BRL-CAD: 03bob1961 * r37541 10/brlcad/trunk/src/libtclcad/ (ged_obj.c tclcadAutoPath.c): Quell a few warnings when compiling for 64-bit Windows. |
| 00:20.46 | ``Erik | heh |
| 00:25.32 | starseeker | remembers commenting about that long ago, but can't remember what the conclusion was |
| 00:27.50 | starseeker | I know they're not standard BRL-CAD style, but when you start running out of upper AND lower case letters it seems fair to regulate a few of the less commonly used options to more verbose arguments... |
| 00:28.17 | starseeker | (plus some straight-up more intuitive stuff like --ae or --aet...) |
| 00:28.52 | ``Erik | thinks the --x-{includes,libraries} are part of the aclocal, we just follow along with auto*'s inconsistancies |
| 00:33.10 | brlcad | starseeker: the later, libbu long opt support ftw |
| 00:33.14 | brlcad | er, latter |
| 00:33.35 | brlcad | mafm is probably over his head if he's stuck on basic configure options |
| 00:33.52 | starseeker | feels a certain sympathy... |
| 00:34.12 | brlcad | knew he was trying to extrapolate from the x option, but x is the (only?) exception, not the rule |
| 00:39.15 | starseeker | ahh... |
| 00:39.20 | brlcad | and yeah, it doesn't make sense and generally isn't something we could even mirror out of convenience, but happens to work due to how high-level the x11 checks are automatically hooked in |
| 00:40.02 | starseeker | tries to verify whether GNU's getopt is GPL or LGPL - if the latter, code snarfing may be profitable... |
| 00:42.31 | ``Erik | http://sourceforge.net/projects/freegetopt/ ? |
| 00:42.55 | brlcad | i'm not sure we'd want to snarf getopt_long from gnu regardless |
| 00:43.11 | brlcad | there are subtle differences between gnu and bsd impl |
| 00:43.31 | *** join/#brlcad PrezKennedyII (Matthew@whitecalf.net) |
| 00:43.43 | ``Erik | waits for rt/? |
| 00:43.51 | starseeker | nods - I was thinking more about looking at how they parse - last time I looked at glibc was for search, and ended up hunting up the BSD version |
| 00:44.02 | brlcad | that's nice ``Erik .. interesting |
| 00:44.13 | ``Erik | first hit for 'bsd getopt' |
| 00:44.19 | starseeker | ``Erik: saw that, but I think they only do short and not long? |
| 00:44.40 | ``Erik | at the moment, yeh, but I'd imagine it'd be pretty trivial to help him out and add long |
| 00:44.42 | starseeker | NetBSD has a getopt, but uses 4 clause BSD for some reason |
| 00:44.44 | brlcad | heh, rt /F=/dev/X /o=file.pix /s=1024 |
| 00:44.52 | starseeker | hehe |
| 00:45.02 | ``Erik | rt /F=C:\X |
| 00:45.23 | ``Erik | and don't forget /s=1kibi O.o :D |
| 00:45.37 | starseeker | ``Erik: arguably, our libbu already does short, so it's either add long to freegetopt or to libbu direct (unless he's got goodies we don't have?) |
| 00:46.23 | ``Erik | *shrug* stop adding lame options and it's not an issue :D |
| 00:46.29 | starseeker | hehe |
| 00:46.42 | starseeker | what does FreeBSD use? I haven't due it out yet |
| 00:47.00 | ``Erik | short options, with gnu in ports |
| 00:47.02 | CIA-43 | BRL-CAD: 03bob1961 * r37542 10/brlcad/trunk/src/other/libutahrle/ (6 files): Quell a few warnings when compiling for 64-bit Windows. |
| 00:47.11 | starseeker | ah, phooey |
| 00:47.32 | ``Erik | huh, looks like there IS a getopt_long in fbsd libc |
| 00:47.45 | starseeker | ooo - are they 3 or 4 clause? |
| 00:47.46 | ``Erik | from netbsd 1.5 originally |
| 00:47.46 | brlcad | er, there's are bsd getopt_long |
| 00:48.00 | brlcad | e.g., http://www.koders.com/c/fid5FCCD794DA3E7129AC307C40B5D31C268ED04FF5.aspx |
| 00:48.22 | ``Erik | looks like 3 |
| 00:48.24 | starseeker | ah, good catch |
| 00:48.29 | starseeker | wonders how he missed that |
| 00:48.33 | starseeker | thanks brlcad :-) |
| 00:48.40 | starseeker | *read read read* |
| 00:49.09 | CIA-43 | BRL-CAD: 03bob1961 * r37543 10/brlcad/trunk/src/libwdb/wdb.c: Quell a few warnings when compiling for 64-bit Windows. |
| 00:49.18 | brlcad | that's just netbsd's libc |
| 00:49.23 | ``Erik | ah, huh, that one is 4 clause... http://www.freebsd.org/cgi/cvsweb.cgi/src/lib/libc/stdlib/getopt_long.c?rev=1.15.10.1.2.1;content-type=text%2Fx-cvsweb-markup |
| 00:50.14 | ``Erik | same file, it'd seem :) |
| 00:50.20 | starseeker | has noticed that some of the BSDs updated to 3 clause and others just left it in other cases |
| 00:50.52 | starseeker | dunno why - maybe some of 'em just don't care |
| 00:51.58 | ``Erik | or weren't able to secure all the written permissions for a license change |
| 00:52.12 | starseeker | you'd think if one could they all could though |
| 00:52.21 | brlcad | gnu's version, http://gitorious.org/enca/enca/blobs/master/src/getopt_long.c (but yeah, wouldn't use it without reviewing the differences in detail) |
| 00:52.41 | starseeker | generally speaking, BSD code seems to fit better with libbu |
| 00:52.49 | starseeker | or any of our libs, for that matter |
| 00:53.06 | ``Erik | they forked a long time ago... ~93, and then there was the ugly legal issue, so'z now all the bsd's are very cautious |
| 00:53.50 | starseeker | ``Erik: you worried the netbsd one isn't OK? |
| 00:54.12 | ``Erik | um, why do you care about 3 vs 4 clause bsd license? O.o |
| 00:54.39 | brlcad | old implementation used by kerberos: http://www.opensource.apple.com/source/Kerberos/Kerberos-47/KerberosFramework/Kerberos5/Sources/util/windows/getopt_long.c |
| 00:54.47 | CIA-43 | BRL-CAD: 03bob1961 * r37544 10/brlcad/trunk/misc/win32-msvc8/ (4 files in 4 dirs): More updates |
| 00:55.09 | starseeker | ``Erik: generally, 4 clause is regarded as not playing nice with LGPL/GPL, iirc |
| 00:55.51 | CIA-43 | BRL-CAD: 03bob1961 * r37545 10/brlcad/trunk/src/conv/asc/asc2g.c: Quell a few warnings when compiling for 64-bit Windows. |
| 00:56.54 | starseeker | brlcad: looks like they define an optional replacement for short getopt, if I'm reading this right - would we want to just keep bu_getopt and have a separate bu_getopt_long ? |
| 00:57.17 | ``Erik | (btw, for the logic migration to libraries for asc2g/g2asc... all the logic in those programs is for v4, just an ugly red herring.. that task can probably be cancelled O.o _get and _adjust do it.) |
| 00:57.56 | starseeker | O.o |
| 00:58.36 | brlcad | latest netbsd version here (1.24): http://ring.nict.go.jp/archives/NetBSD/NetBSD-release-5-0/src/lib/libc/stdlib/ |
| 00:58.37 | starseeker | don't we need to keep that around in case of a v4 asc file? |
| 00:59.19 | ``Erik | yeah, it needs to stay, but it probably doesn't need the refactoring I was talking about, since my intent was to close up a hole for new primitive implementation |
| 00:59.30 | starseeker | ah |
| 00:59.58 | starseeker | bemusedly notes that logic should probably wind up as an --asc option to dbupgrade |
| 01:00.10 | starseeker | brlcad: better and better - 2 clause :-) |
| 01:00.40 | ``Erik | and when I implement librt/primitives/teapot/ for modelling the basic building blocks of the universe, I don't wanna miss anything... :D |
| 01:01.04 | starseeker | checks what nbtool_config.h is... |
| 01:01.05 | brlcad | there's little reason to have both bu_getopt() and bu_getopt_long() other than to mirror the C API (which is not libbu's goal) |
| 01:01.12 | starseeker | nods |
| 01:01.20 | brlcad | it's to wrap concise functionality in the manner that keeps things the most simple for us |
| 01:02.21 | brlcad | bu_getopt() could have it's arguments modified or could be used as an future static to a bu_option() function (that internally just forms up data and calls a getopt_long) |
| 01:03.01 | brlcad | starseeker: I'd still take a look at that sourceforge project -- feature parity on windows is pretty sweet |
| 01:03.18 | brlcad | maybe make it use netbsd's getopt impl instead of whatever it's doing, a fork mod |
| 01:03.43 | ``Erik | heh, the southpark about hybrids is on |
| 01:23.13 | Stattrav | wooh |
| 01:35.56 | starseeker | decides the getopt thing is his next logical project - to get tk framebuffer behaving correctly he needs to understand the details of what rt needs, to do that he needs to get into the rt code, and as long as he's doing that a good starting point is alleviating the problem of option space becoming saturated :-P |
| 01:36.35 | starseeker | reads that again to make sure it makes no sense ;-) |
| 01:48.08 | starseeker | looks like the best starting point is the combination of http://ring.nict.go.jp/archives/NetBSD/NetBSD-release-5-0/src/lib/libc/stdlib/ and http://sourceforge.net/projects/freegetopt/ |
| 01:48.51 | ``Erik | heh, you don't like, uh, the approach used by the photon map lighting model? |
| 01:49.17 | starseeker | is afraid to ask... |
| 01:49.28 | ``Erik | from the manpage: Example: -l7,16384,0,10,60.0,0,0,0,0,1.0,scene.pm. |
| 01:49.44 | starseeker | is that what inspired gqa? |
| 01:50.12 | ``Erik | um, dunno, it was like 7-8 years ago, I think? |
| 01:50.19 | starseeker | heh |
| 01:50.55 | starseeker | supposes he'll get in trouble wanting to change lots of options on commands, but he promises to document it all in docbook... |
| 01:51.13 | ``Erik | heh, what 'lots of options'? O.o |
| 01:52.41 | starseeker | dunno yet really - just have a gut feeling some of the rt/gqa options make much more sense as multi-letter options... |
| 01:55.15 | ``Erik | there're probably already too many options, mere mortal users are skeered by the manpage and don't want/need most options |
| 01:55.19 | ``Erik | *shrug* |
| 01:55.24 | starseeker | yes, that too |
| 01:56.17 | ``Erik | (gqa, for example, may benefit from removing options and mebbe doing some link fu and if(!strncmp(*argv, ... |
| 01:56.21 | starseeker | I'm thinking the most common ones (typed a lot) should be one letter, otherwise a more verbose option (e.g. --use-air) would be both less cryptic and more friendly when it does need to be used... |
| 01:56.22 | ``Erik | )... )) |
| 01:57.12 | starseeker | if they want to set a lot of options to something by default, that's what config files or scripts are all about... |
| 01:57.24 | starseeker | (or preference panels, if you're in a GUI) |
| 01:57.58 | ``Erik | or a swiss army chainsaw shell, like uh, ... btclsh |
| 01:58.22 | starseeker | likes that as a tagline - "btclsh, the swiss army chainsaw" |
| 01:58.37 | ``Erik | *shrug* |
| 01:58.54 | ``Erik | (even though I'm polluting getopt namespace heh...) |
| 01:59.12 | ``Erik | facetize -m, for example O:-) |
| 01:59.41 | starseeker | what's -m do? |
| 01:59.51 | ``Erik | marching cubes algo |
| 01:59.56 | starseeker | ah :-) |
| 02:00.11 | ``Erik | ('swiss army chainsaw' was used to refer to perl long ago...) |
| 02:00.28 | ``Erik | aka sysadmin duct tape |
| 02:00.43 | starseeker | would prefer -a marching or --algorithm marching or (maybe) -a m for that... |
| 02:01.21 | ``Erik | yeah, that'd be pretty silly, it's a good thing you're not the one implementing :> *duck* |
| 02:01.30 | starseeker | nice thing about -- options, you can always have them around and use the short ones where you want to, not be forced to use SOME short option even if there's not logical connection... |
| 02:01.47 | starseeker | ``Erik: what's wrong with it? |
| 02:02.13 | ``Erik | it's not what I came up with :D |
| 02:02.24 | starseeker | LOL |
| 02:02.29 | ``Erik | and obviously, my way is right and yours is wrong, by definition |
| 02:02.37 | ``Erik | *duck* :D |
| 02:02.42 | starseeker | very well, Congressman |
| 02:02.50 | ``Erik | nah, it was just a quick and easy way to add it |
| 02:02.56 | starseeker | nods |
| 02:02.59 | ``Erik | "em for .mmmarchingcubes" |
| 02:03.41 | starseeker | no worries now - our docs don't match our commands in a lot of cases, our commands need to be merged in a lot of cases, and our options are "expert friendly" in some cases |
| 02:04.35 | ``Erik | hm, -0x4D43 |
| 02:04.37 | ``Erik | better? |
| 02:04.47 | starseeker | eeep |
| 02:06.46 | starseeker | you know, in some ways I wonder if search shouldn't return an argc, argv setup... |
| 02:07.04 | starseeker | nah, probably not |
| 02:07.30 | ``Erik | wait wait, better, you can do -0x00004D43 on a sane system, but on a little endian, you have to do -0x434D0000 |
| 02:07.38 | starseeker | hehe |
| 02:07.38 | ``Erik | I like it ;D the evil hex -0 option |
| 02:07.53 | starseeker | so do you use the evil option to set the evil bit? |
| 02:08.10 | ``Erik | and it's an easy to use mnemonic, since 0x4D43 is 'M' 'C' |
| 02:08.22 | ``Erik | any newb should just intrinsically know that... :D |
| 02:08.29 | ``Erik | goes mad with insanity |
| 02:08.35 | starseeker | again? |
| 02:09.02 | ``Erik | watch it, boy, or the #define()'s will start again O.o :D |
| 02:09.31 | starseeker | nnoooooo |
| 02:10.33 | ``Erik | could always bust out the partial quotes and intermingling defmacro and define-symbol-macro... |
| 02:10.56 | starseeker | sometimes wonders if someone who demonstrates mad scripting skills with sh, perl, tcsh, zsh, autotools, and a few other such tools deserves some kind of academic degree |
| 02:11.32 | ``Erik | or committed... |
| 02:11.44 | starseeker | I don't remember physics being tremendously more difficult to understand than some of the perl scripts I've seen... |
| 02:11.49 | starseeker | lol |
| 02:11.52 | starseeker | yeah, that too |
| 02:11.57 | ``Erik | ponders digging in his big bag o' coding attrocities to try to melt starseekers brain again |
| 02:12.35 | starseeker | ``Erik: how come you never compete in those "who can make the most wacky C code" contests? |
| 02:12.40 | starseeker | you'd be a natural |
| 02:13.31 | ``Erik | ioccc? nah, those guys pervert the language, I just use it in neat ways |
| 02:15.14 | ``Erik | take, for example, this piece of simple and self-documenting obvious code... |
| 02:15.19 | ``Erik | static int bitcount(unsigned char w) { if (w==0) return 0; return bitcount(w>>1) + w|1; } |
| 02:16.11 | ``Erik | shoulda done { return w==0?0:bitcount(w>>1)+w|1; } |
| 02:16.29 | starseeker | alright, what were you after? |
| 02:16.36 | ``Erik | huh? |
| 02:16.59 | ``Erik | those turds just knocked the power cord off of my cable box O.o |
| 02:17.18 | starseeker | hehe - the cat seek and destroy team? |
| 02:17.37 | ``Erik | yeah, miniature herd of elephants in full throttle play |
| 02:18.07 | ``Erik | takes forever for that thing to sync up, too :/ |
| 02:18.07 | starseeker | gives up and askes - what does the above code do? (i.e. what were you after when writing it?) |
| 02:18.17 | ``Erik | counts the number of bits set |
| 02:18.41 | starseeker | is that superfast? |
| 02:18.48 | ``Erik | nope, slow as hell |
| 02:19.03 | starseeker | erm... O.o |
| 02:19.08 | ``Erik | but a trivial one-liner |
| 02:19.24 | starseeker | ah, so its virtue is brevity |
| 02:19.26 | ``Erik | with both bit ops and recursion :D |
| 02:20.02 | starseeker | ``Erik: by the by, any insights into the marching cubes patterns? |
| 02:20.57 | ``Erik | no, I showed ed and he was trying to think it through... when he asked what my next step was, I answered with "go home, watch tv and drink a beer" |
| 02:21.06 | starseeker | heh |
| 02:21.35 | ``Erik | I'll inject artificial cases into the cube solver with a buttload of bu_log()'s to see what it's trying to do tomorrow |
| 02:21.42 | starseeker | hmm - apparently the IOCCC lost some steam a few years ago |
| 02:22.03 | ``Erik | yeah, they realized that perl had 'em whumped |
| 02:22.15 | starseeker | ``Erik: sounds good |
| 02:22.21 | ``Erik | (larry wall used to win a lot of those competitions iirc) |
| 02:22.32 | starseeker | <snort> no surprise there |
| 02:22.42 | ``Erik | there were some interesting patterns Ed and I noticed, though |
| 02:23.26 | starseeker | it almost looked like it was covering all vertices but not defining all faces or some such... |
| 02:23.29 | ``Erik | may've assembled the edge list wrong |
| 02:23.51 | ``Erik | *shrug* it'll get done eventually |
| 02:23.58 | starseeker | nods |
| 02:24.30 | starseeker | always dreads having to deal with assembling BoT structures - the ideas are not yet intuitive to him |
| 02:25.08 | starseeker | OK, time to get outta here - getopt_long exploration to begin tomorrow |
| 02:25.12 | ``Erik | our bots are very similar to the OBJ format |
| 02:25.34 | ``Erik | g2asc a trivial bot and read the resulting file |
| 02:25.35 | ``Erik | :) |
| 03:35.10 | *** join/#brlcad PrezKennedy (Matthew@whitecalf.net) |
| 03:40.43 | brlcad | starseeker: keep in mind that a long option routine refactor is not going to be a quick task and shouldn't be started if it's not going to be finished |
| 03:41.20 | brlcad | I count 329 instances of bu_getopt() that would need to be converted |
| 03:42.08 | brlcad | with each one requiring a struct to be defined with the short and long options at a minimum |
| 03:43.14 | ``Erik | grouses and geom being flakey |
| 03:43.19 | ``Erik | s/nd/t/ |
| 03:43.50 | brlcad | that's actually probably worthy of a little scripting to convert them straight up to the initialization block and new call, so that all that's remaining is potentially moving a block 329 times and filling out the long option names |
| 03:47.40 | brlcad | e.g. to convert bu_getopt(argc, argv, "ab:c"); into something like... |
| 03:50.06 | brlcad | <PROTECTED> |
| 03:50.48 | brlcad | which shows a little while getopt_long is a bit suboptimal, it has you repeat the option string |
| 03:51.35 | brlcad | ours could be simply (argc, argv, opts) and it'd derive the opt string when passing it to getopt_long under the hood |
| 03:57.52 | ``Erik | could also do some sed fu to add ,NULL to the end of arglists, then have something like if(longtops&&argv[i][1]=='0'&&argv[i][2]){do longopts} :/ prolly not worth the effort, though |
| 04:04.10 | CIA-43 | BRL-CAD: 03brlcad * r37546 10/brlcad/trunk/src/libbu/parallel.c: quell unused parameter warning |
| 04:04.56 | brlcad | starseeker: all options can have long AND short names defined simultaneously, it's not necessarily one or the other |
| 04:34.12 | brlcad | ``Erik: actually you should have called count_ones32() instead of rolling your own. 5 shifts, 5 ands, 5 adds, and 5 register writes have to be better than 8 function calls, 7 branches, 7 shifts, 7 adds, and 7 ors |
| 04:59.29 | brlcad | even if you assume it all inlines, should still be faster |
| 05:04.56 | *** join/#brlcad Ralith (~ralith@69.90.48.97) |
| 05:48.55 | CIA-43 | BRL-CAD: 03brlcad * r37547 10/brlcad/trunk/ChangeLog: release prep, update ChangeLog from 2010-01-15 |
| 05:55.49 | CIA-43 | BRL-CAD: 03brlcad * r37548 10/brlcad/trunk/ (5 files in 5 dirs): bump the version numer to 7.16.6 in preparation for release (no non-bugfix changes until tagged (later today)) |
| 06:06.31 | CIA-43 | BRL-CAD: 03brlcad * r37549 10/brlcad/trunk/src/libpkg/pkg.c: quell warning about converting a pid_t to a %d specifier. instead, cast the result from pid_t to an int for quellage. |
| 06:17.48 | *** join/#brlcad Win7_64 (~Warlock@bas2-sudbury98-1128564922.dsl.bell.ca) |
| 07:41.00 | Win7_64 | brlcad: some pictures you haven't seen on http://www3.sympatico.ca/mario.dulisse2 :) |
| 12:37.23 | ``Erik | brlcad: yeh, but it probably woulda taken me longer to find that func, which for a quick sanity check, meh :) I just figured it'd asplode starseekers brain a little |
| 12:40.07 | ``Erik | __popcnt might be better yet (on certain procs) *shrug* |
| 12:59.03 | starseeker | brlcad: would you prefer if I didn't attempt such a refactor for a while? |
| 12:59.16 | starseeker | would understand if it's not something that should be monkeyed with at this time |
| 13:03.20 | ``Erik | (put it on a task card?) |
| 13:26.37 | *** join/#brlcad d_rossberg (~rossberg@BZ.BZFLAG.BZ) |
| 14:15.00 | starseeker | feels like diving into it, but if it's a bad time for it... |
| 14:26.22 | *** join/#brlcad mafm (~mafm@99.Red-83-45-252.dynamicIP.rima-tde.net) |
| 14:43.01 | brlcad | starseeker: it's not a matter of timing, anytime would be a fine time to do that |
| 14:43.14 | brlcad | been a topic of discussion since before I started |
| 14:43.51 | brlcad | it's more an issue of not having yet another work in progress on something like that |
| 14:43.59 | brlcad | it would need to be a full conversion, or it probably shouldn't be started |
| 14:45.18 | brlcad | 300+ refactorings wouldn't take too long, but it'd certainly be very tedious |
| 14:50.38 | brlcad | quick napkin calcs, if that conversion script I mentioned yesterday was written and applied first, that'd result in the conversion taking as little as 11 hours (2min per file) to as much as 55 hours (10min per file), so it's certainly doable |
| 14:51.19 | brlcad | without the script, I think it about doubles the end result |
| 14:51.29 | brlcad | about 20-100 hours |
| 15:13.01 | starseeker | nods |
| 15:13.34 | starseeker | OK, if I tackle it I'll see it through |
| 15:14.22 | starseeker | will try his hand at some scripting foo when he gets in (computer fan making strange noises, so doing backups just in case...) |
| 15:24.26 | brlcad | you could certainly implement the bu routine, get getopt_long into libsysv and hook everything up.. it should just stay HIDDEN until we're ready to convert everyone over |
| 15:26.00 | ``Erik | hey, starseeker, it just dawned on me this morning... isst has a normal view... very illuminating, I see what the problem with teh triangles really is now, just gotta solve where the issue is introduced :) |
| 15:26.33 | ``Erik | (why... yes... I did just call a lighting/shading model "illuminating". shuttup :D ) |
| 16:01.42 | CIA-43 | BRL-CAD: 03brlcad * r37550 10/brlcad/trunk/src/mged/mged_dm.h: can't expose X11 types, they might not be available. |
| 16:05.20 | CIA-43 | BRL-CAD: 03brlcad * r37551 10/brlcad/trunk/src/mged/ (mged.c setup.c): move mged_rtCmdNotify() from setup.c to mged.c, renaming to mged_notify() and marking it hidden. |
| 16:09.40 | CIA-43 | BRL-CAD: 03brlcad * r37552 10/brlcad/trunk/src/mged/mged.h: no it's not |
| 16:10.01 | CIA-43 | BRL-CAD: 03brlcad * r37553 10/brlcad/trunk/src/mged/cmd.h: quell shadows, remove argument names from declaration. |
| 16:32.24 | brlcad | looks like a clean build here, just one last thing to test |
| 17:23.09 | starseeker | breaths a sigh of releaf - backup complete |
| 17:23.15 | starseeker | relief even |
| 17:23.29 | starseeker | alrightie, back on the road again |
| 17:45.05 | ``Erik | *burp* |
| 18:08.22 | brlcad | *burp* |
| 18:15.24 | ``Erik | greene turtle, we got to watch the waitress throw ed and jim's food on the floor :D |
| 18:24.46 | starseeker | LOL |
| 18:24.56 | starseeker | missed a show did I? |
| 18:39.01 | CIA-43 | BRL-CAD: 03brlcad * r37554 10/brlcad/trunk/misc/enigma/ (Makefile.am configure.ac): the getpass() function is in -lbsd on some platforms (e.g., haiku), so check for it. |
| 18:49.58 | CIA-43 | BRL-CAD: 03erikgreenwald * r37555 10/brlcad/trunk/src/librt/primitives/metaball/metaball_tri.c: translate between OpenGL and BRL-CAD coordinate systems correctly. |
| 18:51.55 | ``Erik | sunny beaches, undid and redid, but redid it wrong. *sigh* |
| 18:56.43 | CIA-43 | BRL-CAD: 03erikgreenwald * r37556 10/brlcad/trunk/src/librt/primitives/metaball/metaball_tri.c: mehhh, fix it slightly better. |
| 18:58.49 | brlcad | gets a full clean compile on haiku again |
| 18:59.05 | brlcad | that is *everything* except Tk |
| 19:00.37 | starseeker | sweeet |
| 19:00.58 | starseeker | winces at the thought of what a Tk backend for Haiku would take... |
| 19:10.52 | *** join/#brlcad Phurl (~mdupont@ip-81-210-228-126.unitymediagroup.de) |
| 19:16.06 | CIA-43 | BRL-CAD: 03erikgreenwald * r37557 10/brlcad/trunk/src/librt/primitives/nmg/nmg_tri_mc.c: fix winding order |
| 19:21.03 | *** join/#brlcad __monty__ (~toon@78-23-213-229.access.telenet.be) |
| 19:21.17 | brlcad | starseeker: not nearly as much as one would think really, at a glance each platform is about 30k lines of code |
| 19:22.56 | brlcad | about half of that is the basic stub, comment, and template portions |
| 19:25.51 | starseeker | hmm |
| 19:26.11 | brlcad | so about 15k lines of code, with it all pretty much being "make this function do X", all designed out already |
| 19:26.28 | brlcad | probably doable in a month or two for our needs |
| 19:27.04 | brlcad | would also make someone quite the Tk expert |
| 19:37.33 | starseeker | decides first to see if he can comprehend something as simple as getopt_long... |
| 19:37.58 | starseeker | be kinda nice to do something that actually works, even if it is minor... |
| 19:38.56 | brlcad | tire works ;) |
| 19:39.13 | brlcad | coil works (I think, haven't tried it myself yet ;) |
| 19:39.49 | brlcad | EDITOR test did crash hard on me last night in console mode, but I didn't trust the testing host and couldn't get a reliable debug session, so the jury is still out |
| 19:45.27 | CIA-43 | BRL-CAD: 03bob1961 * r37558 10/brlcad/trunk/src/mged/cmd.c: Quell warnings when compiling for 64-bit windows. |
| 20:00.05 | ``Erik | finish up your release so'z I can add successful metaball tesselationi to NEWS O.o BOAH! AH'LL WHUP YA! |
| 20:14.34 | CIA-43 | BRL-CAD: 03erikgreenwald * r37559 10/brlcad/trunk/src/librt/primitives/metaball/metaball_tri.c: remove debugging bu_log |
| 20:14.47 | CIA-43 | BRL-CAD: 03erikgreenwald * r37560 10/brlcad/trunk/src/librt/primitives/nmg/nmg_tri_mc.c: update status |
| 20:15.35 | brlcad | you can update news |
| 20:15.56 | brlcad | working through a bu vls bug |
| 20:32.55 | CIA-43 | BRL-CAD: 03erikgreenwald * r37561 10/brlcad/trunk/src/proc-db/metaball.c: create regions. |
| 20:33.06 | CIA-43 | BRL-CAD: 03brlcad * r37562 10/brlcad/trunk/src/libbu/vls.c: modify the default minimum and step allocation sizes to be word-aligned allocation sizes. make bu_vls_extend() obey the step size and increment ONLY in step-sized increments, not just at least that much. |
| 20:35.59 | mafm | brlcad: so what's the option --x-includes for? |
| 20:39.35 | starseeker | mafm: I believe he said that's a bit of a special case... |
| 20:41.48 | brlcad | mafm: x is the exception not the rule, and none of our doing |
| 20:41.54 | ``Erik | it's the old autoconf way of finding the X11 includes directory... we kinda superceded it with --with-x11= |
| 20:42.10 | brlcad | x options are hooked in very high up in the food chain, that's more a legacy option from more than a decade ago |
| 20:42.23 | mafm | :S |
| 20:42.37 | mafm | so X is literally X? I thought that it was a "variable" |
| 20:43.05 | brlcad | thinks mafm would be much more greatly productive writing code, even fixing bugs, than sorting out build options :D |
| 20:43.07 | mafm | like: if you want to include X package with a special dir, you type --X-includes DIR |
| 20:43.19 | ``Erik | no, it's for X, aka the X windowing system... |
| 20:43.24 | brlcad | heh, .. yes literally that is the option to set X11 build flags |
| 20:44.37 | ``Erik | looking at the output of proc-db/metaball in isst right now... too effin' neat :D |
| 20:45.17 | mafm | $ ls /usr/include/tcl8.5/{tcl,tk,itcl}.h |
| 20:45.23 | mafm | <PROTECTED> |
| 20:45.39 | ``Erik | ./configure CPPFLAGS=-I/usr/include/tcl8.5 |
| 20:45.42 | mafm | CPPFLAGS="-I/usr/include/tcl8.5/" ./configure --disable-documentation --disable-all --enable-urt-build --enable-opennurbs |
| 20:45.49 | mafm | and fails |
| 20:46.31 | brlcad | ``Erik: pics or it didn't happen |
| 20:46.32 | mafm | checking for Tcl configuration... configure: WARNING: Can't find Tcl configuration definitions |
| 20:46.43 | ``Erik | gotta tell us more than that... and http://www.freebsd.org/cgi/cvsweb.cgi/ports/cad/brlcad/Makefile?rev=1.40 might help |
| 20:48.54 | brlcad | mafm: you can read the configure.ac file to see all the checks that get run in the order they are run, it's broken out into sections |
| 20:49.11 | brlcad | config.log has the juicy details on everything that happens |
| 20:52.12 | starseeker | Hey, cool - 400+ square mile ranch - that's gotta be fun for something :-) http://www.landsofamerica.com/america/index.cfm?Detail=&INV_ID=680063 |
| 20:53.03 | ``Erik | http://brlcad.org/~erik/mb-isst.png uNF |
| 20:54.01 | brlcad | cool! |
| 20:55.37 | starseeker | that looks like a brlcad.org headline news image if I ever saw one :-) |
| 20:55.48 | ``Erik | need to cook up a few images with some tech details for a status report briefing, then I can start doing it right rays on arbitrary regions (and watch it asplode on edges) |
| 20:56.00 | ``Erik | nmg's are a bitch, btw. |
| 20:56.38 | ``Erik | should create a simple 'add triangle to nmg' function that takes point_t[3] |
| 20:57.56 | brlcad | and now.. they're YOUR bitch |
| 20:58.35 | brlcad | ``Erik: can you send me those details too? i'll put them up on the site |
| 20:59.07 | brlcad | or you can send an announcement to brlcad-news and I'll grab from there |
| 20:59.24 | brlcad | or to devel, then I can turn that into a news, whatever works |
| 20:59.27 | ``Erik | sure, I'm thinking about a 'jot 100' script, was kinda thinking about making a movie as the -r changes |
| 21:00.57 | brlcad | starseeker: the mged manual pages definitely cannot be .1 files |
| 21:01.07 | brlcad | (from doxygen) |
| 21:04.26 | starseeker | brlcad: ok, what should they be? |
| 21:04.36 | starseeker | you mean the docbook stuff? |
| 21:04.44 | brlcad | notes the considerable operlap between our "loop" tool, gnu's "seq" tool, and bsd's "jot" tool |
| 21:05.10 | ``Erik | yes, I came across the linux seq one a bit ago |
| 21:05.21 | brlcad | I mean manual pages for mged commands should *definitely* not be in the same space as system commands |
| 21:05.40 | ``Erik | very... linuxy... it's almost a copy, just different enough to ... not work. and missing features. And verbose options. |
| 21:05.49 | brlcad | otherwise things like installing cp.1 and mv.1 and apropos.1 .. bad |
| 21:09.26 | mafm | mm, you have two libraries together, jama and tnt |
| 21:09.38 | brlcad | yes |
| 21:09.56 | brlcad | they were too tiny to separate in a meaningful way |
| 21:10.06 | brlcad | they work together, one is an extension of the other |
| 21:10.26 | mafm | dunno, but in Debian they're two separated packages |
| 21:10.36 | brlcad | jama is like 5 .h files and nothing else |
| 21:10.42 | brlcad | that's fine |
| 21:11.03 | brlcad | they could separate them into a package per .h, doesn't change anything for us :) |
| 21:11.26 | brlcad | just makes life suck for the package maker |
| 21:12.04 | starseeker | brlcad: ok, I can see that if someone tries to install into a system man page area - should the MGED commands not generate the man page version of their output? |
| 21:14.02 | *** join/#brlcad R0b0t1 (~Enigma@unaffiliated/r0b0t1) |
| 21:16.42 | mafm | I certainly feel miserable :P |
| 21:17.07 | CIA-43 | BRL-CAD: 03brlcad * r37563 10/brlcad/trunk/src/util/loop.c: basic cleanup, ws. |
| 21:19.45 | mafm | http://paste.debian.net/58746/ -- I didn't enable tkhtml3 and NIST STEP, are they enable by default? are tkimg and jove important for a regular installation? are opengl or librtserver important for regular installation? |
| 21:20.12 | CIA-43 | BRL-CAD: 03brlcad * r37564 10/brlcad/trunk/bench/run.sh: quell the sanity check output from ELP |
| 21:20.52 | starseeker | mafm: Did STEP and tkhtml3 build even when disabled? |
| 21:20.59 | starseeker | can check if there's a problem there |
| 21:21.48 | mafm | well, my line was: Options & variables: --disable-documentation --disable-all --enable-urt-build --enable-opennurbs --with-tcl=/usr/lib/tcl8.5/ --with-tk=/usr/lib/tk8.5/ --with-tnt=/usr/lib/tnt --enable-termlib-build --with-tkinclude=/usr/include/tcl8.5/ |
| 21:21.50 | louipc | jove is not really important |
| 21:22.19 | mafm | urt, termlib and opennurbs are the only ones not present in Debian |
| 21:24.53 | mafm | if I don't need those, it'll be much easier to get it into Debian officially |
| 21:26.13 | mafm | both for legal and technical matters :) |
| 21:26.40 | brlcad | starseeker: manual pages are fine and good, they just need to be separate |
| 21:26.42 | starseeker | you only lose the MGED html help browser when tkhtml3 is disabled, and it's not really "prime time" yet |
| 21:27.00 | brlcad | for starters, can make them just 'n' pages |
| 21:27.07 | brlcad | but then we even have conflicts with tcl |
| 21:27.18 | starseeker | brlcad: ok, will do... |
| 21:27.18 | brlcad | tcl groups there n pages so tcl and tk don't collide |
| 21:27.28 | brlcad | with the .n and .ntcl pages |
| 21:27.47 | starseeker | we could make up our own man convention - manm for mged... |
| 21:27.48 | brlcad | so we could do something similar with .ncad or .nged pages |
| 21:27.55 | brlcad | heh |
| 21:27.58 | brlcad | no |
| 21:28.02 | brlcad | they didn't make up 'n' |
| 21:28.37 | starseeker | mafm: you won't miss the STEP library at the moment - the functionality isn't quite ready - so don't worry too much about it |
| 21:28.44 | starseeker | brlcad: oh, I know :-) |
| 21:28.59 | mafm | ok, so I could disable them... the thing is that I did and didn't work :) |
| 21:29.31 | starseeker | it built them anyway? (that was my earlier question) |
| 21:30.07 | brlcad | mafm: debian has termlib |
| 21:30.15 | brlcad | it's just got a variety of names over the years |
| 21:30.28 | mafm | starseeker: yes, I pasted you the lines above just after you asked |
| 21:30.52 | starseeker | right, but with those lines it still built tkhtml3 and step, correct? |
| 21:30.59 | mafm | brlcad: there's no termlib, no termcap, and some libterm-*-perl |
| 21:31.29 | mafm | starseeker: yes, with that option it built what I put in the line just before you asked: http://paste.debian.net/58746/ |
| 21:31.43 | brlcad | look for libtermlib or libtermcap or termcap or termlib or terminfo or tinfo and failing all of those, curses or ncurses will provide them indirectly |
| 21:31.58 | starseeker | brlcad: what about things like rt where I hope to have one man page both for command line and for mged? |
| 21:32.05 | brlcad | mafm: probably because it's installed by default on linux now |
| 21:32.05 | louipc | oh --disable-tkhtml3 wasn't covered in --disable-all eh? |
| 21:32.27 | starseeker | is surprised it isnt... |
| 21:32.43 | louipc | should --disable-doc be covered as well? |
| 21:33.02 | brlcad | mafm: I'd be very surprised if you didn't already have it installed, base-essentials or something similar that provides it |
| 21:33.46 | brlcad | "locate termcap" and/or "locate terminfo", and I bet you'll see you already ahve it |
| 21:33.58 | brlcad | can then lookup those files to see what package they belong to |
| 21:34.38 | starseeker | mafm: hang on, I'm testing to see what happens here with those options... (jeez I wish the distros would lighten up sometimes) |
| 21:34.50 | louipc | yeah use ncurses |
| 21:34.53 | brlcad | starseeker: that's a sticky bit wrt 'rt', I'd just start keeping them separate (as their usage really is distinct in both) |
| 21:35.08 | brlcad | the mged one should be fairly short and sweet, referencing the other |
| 21:35.24 | starseeker | ick. Well, I guess if we have to... |
| 21:37.49 | mafm | if it only needs termcap.h, it's in libncurses |
| 21:38.00 | brlcad | the mged one should just explain what is unique to mged, namely that you don't specify the .g or objects, it uses whatever is displayed, that some options don't work as expected (e.g. -M) because you're not in a terminal, and other options that are auto-provided unless overridden like -n/-w |
| 21:38.41 | brlcad | mafm: speaking of combining two packages into one... |
| 21:39.08 | brlcad | termcap is technically separate :) |
| 21:39.15 | louipc | heheh |
| 21:39.15 | brlcad | terminfo is newer/better |
| 21:39.27 | brlcad | who owns your terminfo files? |
| 21:39.35 | mafm | well, but that's fine, it saves work! |
| 21:39.51 | brlcad | exactly why tnt+jama .. saved work :) |
| 21:40.20 | mafm | but tnt+jama is *bad* because it puts more work on *me* :P |
| 21:40.36 | mafm | now, since they're packaged in debian, that's not much of a problem |
| 21:40.43 | louipc | stop using debian :D |
| 21:41.11 | louipc | ah that's convenient |
| 21:41.26 | starseeker | mafm: if you just treat BRL-CAD as one big atomic package, no extra work! :-) |
| 21:41.39 | mafm | starseeker: disabling tkhtml3 and step, makes things work, so I guess that they're missing in disable-all |
| 21:42.01 | mafm | http://paste.debian.net/58749/ -- the result of configure |
| 21:42.50 | mafm | starseeker: in that case there would be a .deb for brl-cad, but it would get rejected, for sure |
| 21:42.51 | louipc | hmmm |
| 21:43.00 | mafm | rejected in Debian I mean |
| 21:43.31 | CIA-43 | BRL-CAD: 03brlcad * r37565 10/brlcad/trunk/ (configure.ac misc/Makefile.am misc/debian/): remove misc/debian .. those files are so out-of-date and wrong that it's just misleading and bloat. many of the files still referenced glpong and had wrong license info. |
| 21:44.19 | brlcad | mafm: er, how does that put more work on you wrt tnt/jama? you just specify them, no? our disable-tnt option turns both off |
| 21:44.27 | louipc | the build files used glpong as a template? |
| 21:44.48 | mafm | so... are opengl or librtserver important for regular installation? |
| 21:44.53 | brlcad | mafm: you haven't read INSTALL still, I gather |
| 21:45.50 | brlcad | because it explains how --disable-all .. is really --disable-almost-everything |
| 21:46.37 | brlcad | the prior is just a convenience alias for those in the know-how that call it frequently so they can type less |
| 21:48.37 | mafm | in fact I have, just not very carefully, still, I can't even find the pattern "disable-all" |
| 21:48.56 | louipc | autogen creates those |
| 21:49.36 | starseeker | line 631 |
| 21:49.56 | starseeker | all leads to set_everything, apparently |
| 21:50.09 | starseeker | which seems to have step and tkhtml3 in there... hmm... |
| 21:50.17 | starseeker | (configure.ac) |
| 21:50.28 | louipc | yeah |
| 21:51.05 | brlcad | mafm: that's because very early on it explains that there is a disable for every enable and then proceeds to itemize all of the enable options |
| 21:52.22 | brlcad | understandable for the many instances where we don't even bother to document, but can't help you if you don't read it carefully when the information IS there .. :) |
| 21:52.50 | mafm | well, you told me yesterday that I could disable all with... disable-all |
| 21:52.55 | brlcad | if you have a suggestion for a way it could have been stated more clearly, i'd love to hear it ;) |
| 21:53.00 | mafm | and that's what I did |
| 21:53.48 | brlcad | i also said to read INSTALL, which I suppose I should have caveated with "and don't just skim"? |
| 21:53.59 | mafm | it works for everything except those libraries, which seems to have been overlooked or something... so thatnks to my incompetence/lazyness I found a bug... aren't you glad? :P |
| 21:54.27 | brlcad | actually, it's not a bug |
| 21:54.40 | brlcad | that's why the real name is "almost" everything |
| 21:57.15 | mafm | anyway, compiling now |
| 21:57.28 | mafm | I hope that ccache works fine :P |
| 21:58.07 | brlcad | certainly unexpected/unintentional, so we'll see |
| 21:58.27 | brlcad | never tried ccache on brl-cad yet, interesting to hear how well it works ;) |
| 22:02.40 | mafm | http://paste.debian.net/58751/ -- failed creating .libs, might be a kind of race condition of -j2? |
| 22:03.08 | CIA-43 | BRL-CAD: 03louipc * r37566 10/brlcad/trunk/configure.ac: ws |
| 22:05.11 | brlcad | finds six options not documented, modifies distcheck |
| 22:05.48 | brlcad | mafm: huh, yeah, that does look like a race |
| 22:06.31 | brlcad | seen those cannot create dir messages before but they never stop the build |
| 22:07.39 | *** join/#brlcad mafm (~mafm@99.Red-83-45-252.dynamicIP.rima-tde.net) |
| 22:09.27 | mafm | ops, X crashed :P |
| 22:09.35 | starseeker | ow |
| 22:10.06 | mafm | mine stopped the build, had to use -k after... |
| 22:10.40 | mafm | well, now I have to stop for a while but will torture my CPUs a bit later :P |
| 22:37.27 | brlcad | :) |
| 22:41.56 | CIA-43 | BRL-CAD: 03brlcad * r37567 10/brlcad/trunk/regress/repository.sh: |
| 22:41.57 | CIA-43 | BRL-CAD: add an additional release check to make sure all of our configure options are |
| 22:41.57 | CIA-43 | BRL-CAD: documented in the INSTALL file. don't halt the build just yet, though, as there |
| 22:41.58 | CIA-43 | BRL-CAD: are at least 5 that have crept in without documentation. (to be dealt with after |
| 22:41.58 | CIA-43 | BRL-CAD: 7.16.6 is tagged) |
| 22:45.25 | CIA-43 | BRL-CAD: 03brlcad * r37568 10/brlcad/trunk/include/bu.h: add a little note to make it clear that the vls may be extended more than extra length requested. |
| 22:47.10 | ``Erik | awesome, I forgot to remove all of teh glpong references heh |
| 22:51.51 | ``Erik | got a problem with glpong, louipc? O.o it was one of my debian thingies that worked, so'z I was in the process of importing and modifying for BRL-CAD when I lost my debian box |
| 22:52.41 | louipc | ``Erik: glpong is cool |
| 22:53.03 | ``Erik | O.o are you thinking of the right one? heh |
| 22:54.22 | louipc | errr not sure |
| 22:56.21 | CIA-43 | BRL-CAD: 03brlcad * r37569 10/brlcad/trunk/src/libged/move_all.c: |
| 22:56.22 | CIA-43 | BRL-CAD: clean up the usage statement to use consistent notation for required/optional |
| 22:56.22 | CIA-43 | BRL-CAD: parameters. '[]' are optional, '{}' is a logical grouping (to indicate mutually |
| 22:56.23 | CIA-43 | BRL-CAD: exclusive options), '<>' is optional markup indicating substituted user input |
| 22:56.23 | CIA-43 | BRL-CAD: (normally with italics, e.g. man page, but this is plain text). |
| 22:56.27 | ``Erik | http://brlcad.org/~erik/files/glpong-1.2.tar.gz is where the debian crud came from |
| 22:57.11 | ``Erik | or mebbe I lost the debian machine when I was trying to get 1.3 ready for release, ah ferget |
| 22:57.50 | ``Erik | yeh, I think that was in 1.3, which has been sitting all distchecked for several years now |
| 22:57.51 | louipc | hey it does look cool |
| 22:58.04 | louipc | http://www.downbroad.com/images/linux/linux2506.jpg |
| 22:58.43 | ``Erik | yup, that's it |
| 23:00.11 | ``Erik | was a 1 day hack, shirked off all possible job leads but one really amazing sounding one as I was finishing college, was too nervous to sleep or eat for the entire trip and night before the interview, totally blew it, so after getting home and sleeping, knocked out the first version in a day as an "I'm not a loser" act :D |
| 23:01.37 | ``Erik | heads home O.o |
| 23:16.37 | *** join/#brlcad Ralith (~ralith@69.90.48.97) |
| 23:40.54 | mafm | brlcad: there are loads of failures, I don't know if caused by ccache or not |
| 23:44.28 | brlcad | mafm: paste? |
| 23:45.07 | brlcad | you probably want to disable strict building while you're at it so you're not chasing compiler warnings |
| 23:45.14 | brlcad | --disable-strict |
| 23:47.06 | mafm | I'm reacompiling now with --disable-strict and without ccache |
| 23:50.14 | mafm | bbiab, and paste if there're errors again |
