| 00:19.34 | starseeker | ``Erik: yeah, I"m getting the pow thing here too |
| 00:19.54 | ``Erik | it's not nomming -lm right for some reason... and it seems inconsistent O.o |
| 00:20.40 | starseeker | I don't suppose we can just hardcode it in? |
| 00:22.09 | starseeker | here's where we do it: AC_CHECK_LIB(m, cos, m_link_works=yes ; LIBM="-lm") |
| 00:23.16 | ``Erik | or gut the TEA crap and just write a nice clean automake bit to do it O:-) |
| 00:24.00 | starseeker | ``Erik: heh - brlcad wants us to stay with the existing external code whenever possible |
| 00:24.20 | starseeker | would already have tried that, except he wants to learn how to do this sort of thing correctly |
| 00:24.31 | starseeker | it likely won't be the last time we're faced with it |
| 00:25.36 | starseeker | supposes the "ultimate" solution to this would be to make tcl.m4 and automake "play nice", but last time I tried that all I did was expose my own ignorance |
| 00:26.46 | ``Erik | tkpng is a stopgap until 86 though, so *shrug* that one just doesn't matter, I'd imagine |
| 00:28.57 | starseeker | arrrrrgh. |
| 00:28.58 | starseeker | /usr/lib/gcc/i686-pc-linux-gnu/4.3.4/../../../crt1.o: In function `_start': |
| 00:28.58 | starseeker | (.text+0x18): undefined reference to `main' |
| 00:28.58 | starseeker | collect2: ld returned 1 exit status |
| 00:29.07 | starseeker | add in -lm, and THAT happens |
| 00:29.46 | ``Erik | hm, then "ur doin it rong"? :D |
| 00:32.55 | starseeker | why is it looking for main while compiling a library?? |
| 00:34.58 | starseeker | oh, BLAST it |
| 00:35.12 | starseeker | the gcc command from the src/other/Makefile.am isn't including shared |
| 00:55.18 | starseeker | what the bleeeeep.... |
| 01:00.05 | ``Erik | outwitted by the machine yet again? :) |
| 01:00.24 | starseeker | if this sucker isn't AI complete I hate to think what that says about me |
| 01:00.54 | starseeker | mixing libtool and non-libtool config settings just ain't a happy thing |
| 01:02.18 | starseeker | the ironic thing is that --enable-all and the included libz make this significantly harder, instead of simpler :-/ |
| 01:02.23 | starseeker | figures |
| 01:06.23 | CIA-38 | BRL-CAD: 03starseeker * r36935 10/brlcad/trunk/ (configure.ac src/other/Makefile.am): |
| 01:06.23 | CIA-38 | BRL-CAD: Sigh. This probably violates some rule or other, but since we're passing in our |
| 01:06.23 | CIA-38 | BRL-CAD: own custom TKPNGZLIB variable anyway have it point to the .so link instead of |
| 01:06.24 | CIA-38 | BRL-CAD: the .la file and take libtool out of the equation. Also add LIBM to the flags |
| 01:06.24 | CIA-38 | BRL-CAD: we need to feed the tkpng make process. |
| 01:10.47 | brlcad | that wont work starseeker |
| 01:11.09 | brlcad | it's only a .so on shared object library systems |
| 01:11.22 | CIA-38 | BRL-CAD: 03starseeker * r36936 10/brlcad/trunk/src/other/tkpng/Makefile.in: As long as libtool isn't being invoked, restore Makefile.in to pristine state. Now only change is removing explicit -lz in configure.in |
| 01:11.28 | brlcad | could be a .dylib or a .a or a .dll or ... |
| 01:11.39 | starseeker | brlcad: well blast it what do we do then |
| 01:11.47 | starseeker | libtool is screwing up good here |
| 01:12.03 | brlcad | what was the problem? (haven't read backlog) |
| 01:12.07 | brlcad | it was working.. |
| 01:12.18 | starseeker | not here, apparently |
| 01:12.28 | ``Erik | the problem is that TEA and auto* don't live together well :D *duck* |
| 01:12.36 | ``Erik | some machines don't see pow() |
| 01:12.47 | starseeker | that may be fixed with the LIBM flag |
| 01:13.17 | ``Erik | ah, new problems arose? :) |
| 01:13.18 | starseeker | my libtool call is looking for main because (for whatever reason) it's not including -shared in the gcc call |
| 01:13.40 | starseeker | that's what the "undefined reference to main" is |
| 01:13.53 | starseeker | flippin thing is trying to build a program, not a shared library |
| 01:14.07 | starseeker | and I can't even track down where that -shared is supposed to come from yet |
| 01:14.51 | ``Erik | usually it comes from it seeing the automake target having an LT in the name |
| 01:15.09 | ``Erik | lib_LIBRARIES vs lib_LTLIBRARIES iirc |
| 01:15.24 | brlcad | sounds like you're making the wrong libtool call then |
| 01:15.40 | brlcad | what's the actual error? |
| 01:15.55 | brlcad | and what's the compile/link line look like |
| 01:16.45 | ``Erik | got his sushi set today, can finally roll his maki with a real sushi mat instead of an old chinese calendar |
| 01:17.06 | starseeker | one sec... |
| 01:17.12 | CIA-38 | BRL-CAD: 03starseeker * r36937 10/brlcad/trunk/ (configure.ac src/other/tkpng/Makefile.in): Good point Sean - .so extension is not universal. Keep looking... |
| 01:17.14 | ``Erik | might have to sharpen up my knife this weekend :) |
| 01:23.58 | CIA-38 | BRL-CAD: 03starseeker * r36938 10/brlcad/trunk/src/other/Makefile.am: Put back the libtool stuff in src/other/Makefile.am |
| 01:24.16 | starseeker | my machine is so slow.... |
| 01:25.32 | starseeker | http://pastebin.bzflag.bz/m1b73e737 |
| 01:25.35 | starseeker | there we go |
| 01:26.08 | starseeker | the pow thing is handled by adding in LIBM, but the main is due to the missing -shared after gcc (manually inserting it and running that command succeeds) |
| 01:26.57 | brlcad | where is that -shared coming from? |
| 01:27.06 | brlcad | that's not a libtool option |
| 01:27.26 | brlcad | possibly preventing it from adding the flag(s) it needs |
| 01:27.41 | CIA-38 | BRL-CAD: 03starseeker * r36939 10/brlcad/trunk/src/other/Makefile.am: OK, looks like the LIBM inclusion is legit. |
| 01:29.05 | brlcad | you basically just need to invoke libtool the same way it would be invoked if you wrote the Makefile.am with an LT_LIBRARIES declaration |
| 01:29.42 | brlcad | look at one of our other existing Makefiles that gets generated from Makefile.am to see what it adds or, if you have to, whip up a quick testing Makefile.am |
| 01:30.35 | starseeker | actually I think the -shared IS supposed to be after the gcc command, but gets lost in translation somehow? http://pastebin.bzflag.bz/d4d5ba3c |
| 01:45.27 | brlcad | you're not supposed to figure out the gcc line |
| 01:45.33 | brlcad | you're supposed to figure out the libtool line |
| 01:45.40 | brlcad | remember, libtool is what you're actually running |
| 01:45.50 | brlcad | it takes a variety of options to tell it how to run |
| 01:46.05 | brlcad | it then figures out whether to use -shared or a variety of other possibilities |
| 01:46.19 | brlcad | depending on the flags you told it, the compiler being used, the OS, etc |
| 01:48.41 | starseeker | is experimenting with libtool now |
| 01:51.32 | brlcad | look at one of the other lib dirs, that has the logic in it lalready |
| 01:53.14 | starseeker | nods |
| 01:53.37 | starseeker | misses asdf right about now... |
| 01:58.18 | CIA-38 | BRL-CAD: 03starseeker * r36940 10/brlcad/trunk/src/other/Makefile.am: Start playing with libtool options |
| 01:58.58 | starseeker | that's all the options from one of the other makefiles before --mode-link |
| 02:04.49 | starseeker | good grief. this is more work than doing our own system |
| 02:05.57 | starseeker | brlcad: I'm going to revert to tkimg for now so the build will work again while I beat my head on this |
| 02:15.47 | ``Erik | .asd files have their own quirks, though... |
| 02:19.54 | starseeker | ``Erik: they do |
| 02:21.13 | starseeker | but they don't have all the bizarre syntax fun |
| 02:22.27 | starseeker | hrm... since talcite has a tkimg package on Fedora, perhaps the smart move is just to wait for Tk8.6 on this one |
| 02:23.38 | starseeker | is of the opinion that a general solution for passing info to subconfigures is in order, but so far looks to be too dim a bulb to get a grasp of it... |
| 02:28.39 | CIA-38 | BRL-CAD: 03starseeker * r36941 10/brlcad/trunk/ (124 files in 12 dirs): |
| 02:28.39 | CIA-38 | BRL-CAD: Merge back in tkimg - tkpng integration is taking more time to get right, and |
| 02:28.39 | CIA-38 | BRL-CAD: doesn't offer new functionality. Problem of subconfigures using local BRL-CAD |
| 02:28.39 | CIA-38 | BRL-CAD: src/other copies of libraries needs to be considered, but this isn't the case |
| 02:28.39 | CIA-38 | BRL-CAD: that justifies the time. |
| 05:06.28 | *** join/#brlcad talcite (n=matthew@69-196-191-166.dsl.teksavvy.com) |
| 05:34.17 | *** join/#brlcad LCACpn (n=user@12.187.130.2) |
| 05:51.12 | *** join/#brlcad PrezKennedy (i=Matthew@whitecalf.net) |
| 11:13.36 | *** join/#brlcad talcite (n=matthew@69-196-191-166.dsl.teksavvy.com) |
| 11:43.18 | CIA-38 | BRL-CAD: 0388.232.184.140 07http://brlcad.org * r2053 10/wiki/MediaWiki_talk:Ipb-blocklist-addr: New page: 192.168.2.1 |
| 12:29.24 | *** join/#brlcad parigaudi (n=quassel@pd95b7f5e.dip0.t-ipconnect.de) |
| 14:21.21 | ``Erik | heh "I don't suffer from insanity... I enjoy every minute of it!" |
| 14:24.33 | *** join/#brlcad cosurgi (n=cosurgi@atak.bl.pg.gda.pl) |
| 14:54.48 | starseeker | ``Erik: then you've got the right job ;-) |
| 14:55.17 | ``Erik | tar... flavored... icecream... I don't think I'm ever going to finland O.o |
| 14:57.05 | starseeker | wonders who verified it tasted like tar... |
| 14:57.41 | ``Erik | http://thedailywtf.com/Articles/Souvenir-Potpourri-Surprise!.aspx about 3/4 down |
| 14:57.55 | ``Erik | "Do not anger the Finns, for they have strange ways and a twisted sense of humor," he wrote, "you described salmiak as tar-like, but we're already way ahead of you: the leijona salmiak is actually tar-flavored salmiak candy. I'm not joking. We also have tar-flavored ice cream here, too." |
| 15:15.37 | CIA-38 | BRL-CAD: 03erikgreenwald * r36942 10/isst/trunk/src/gui.c: pack more info (code, hostname, .g filename) into the saved shotline |
| 15:29.01 | *** join/#brlcad LCACpn` (n=user@12.187.130.2) |
| 16:11.47 | ``Erik | "Do not anger the Finns, for they have strange ways and a twisted sense of humor," he wrote, "you described salmiak as tar-like, but we're already way ahead of you: the leijona salmiak is actually tar-flavored salmiak candy. I'm not joking. We also have tar-flavored ice cream here, too." |
| 16:11.52 | ``Erik | damnit |
| 16:21.58 | starseeker | Hmm, this might be interesting to the Qt/Ogre integration folks: http://www.mail-archive.com/qt4-preview-feedback@trolltech.com/msg01194.html |
| 16:25.52 | starseeker | or I guess this is the actual qt archive of the message: |
| 16:25.54 | starseeker | http://lists.trolltech.com/pipermail/qt4-preview-feedback/2009-October/000964.html |
| 16:26.09 | CIA-38 | BRL-CAD: 03d_rossberg * r36943 10/rt^3/trunk/ (4 files in 2 dirs): C++ interface to the eto primitive (elliptical torus) |
| 16:27.51 | CIA-38 | BRL-CAD: 03d_rossberg * r36944 10/brlcad/trunk/misc/win32-msvc/Dll/CMakeLists.txt: included the elliptical torus (eto) in the brlcad.dll |
| 16:34.36 | starseeker | eyes option b) in that email... hmm... |
| 16:38.39 | CIA-38 | BRL-CAD: 03erikgreenwald * r36945 10/isst/trunk/src/ (gui.c isst.h local_worker.c net_worker.c): window is now resizable. GUI uses adjustable panes instead of fixed positions. |
| 16:41.00 | ``Erik | yes, the implications of b) are truely terrifying O.o if I'm guessing at the QT bits of it correctly, the performance might be on par with remote X over a dialup :D |
| 16:41.14 | *** join/#brlcad docelic (n=docelic@78-2-120-22.adsl.net.t-com.hr) |
| 16:41.59 | ``Erik | every pixel would come covered in stickers from every bit of hardware it'd been to, like an old steamer trunk :D |
| 16:45.31 | ``Erik | well, poop, I thought there was a card for that |
| 18:49.01 | *** join/#brlcad mafm (n=mafm@249.Red-88-11-185.dynamicIP.rima-tde.net) |
| 18:59.51 | *** join/#brlcad R0b0t1 (n=Enigma@unaffiliated/r0b0t1) |
| 19:49.18 | CIA-38 | BRL-CAD: 0392.112.171.107 07http://brlcad.org * r2054 10/wiki/Google_Summer_of_Code/Project_Ideas: /* Project Ideas */ |
| 20:10.20 | ``Erik | reads up on vEB trees |
| 20:27.54 | starseeker | ``Erik: erm... you really think it would be that slow? |
| 20:30.32 | ``Erik | it's basically schlepping crap to the gpu for fast rendering in ogre, then copying the buffer back to ram, then sending it back to the video card for display, right? |
| 20:31.00 | ``Erik | using qt's funky crap instead of actual ogl 'render to texture' goodness |
| 20:31.07 | starseeker | dunno |
| 20:31.21 | starseeker | don't know if they've got that optimized or not |
| 20:31.21 | ``Erik | and render to texture can drag down a gpu without that read/write pair in the middle |
| 20:31.55 | starseeker | you think slower than merging the two opengl contexts like g3d is doing now? |
| 20:33.44 | ``Erik | I've no idea what g3d is doing :D |
| 20:34.22 | starseeker | Ogre and Qt each have their own context |
| 20:34.26 | starseeker | then they get merged |
| 20:34.52 | CIA-38 | BRL-CAD: 03erikgreenwald * r36946 10/brlcad/trunk/src/librt/ (4 files in 2 dirs): break metaball tesselation out into it's own file |
| 20:35.09 | ``Erik | kinda gathered, the details of the second are what I don't know... :D |
| 20:35.35 | starseeker | checks... |
| 20:36.37 | starseeker | http://www.ogre3d.org/forums/viewtopic.php?f=2&t=42733&sid=76d05fd1184616e02c51ab4734ffb9b7&start=25 |
| 20:36.52 | starseeker | Marseyeah's post starts it off as far as we're concerned |
| 20:38.29 | starseeker | ah yes, MakeCurrent is the problem child |
| 20:39.06 | starseeker | Ralith made it tolerable via some sort of lazy approach, but MakeCurrent remains expensive |
| 20:39.51 | starseeker | functional though - we can probably proceed using that as an approach until we have a real need to do radical surgery for performance reasons... |
| 20:42.21 | CIA-38 | BRL-CAD: 03erikgreenwald * r36947 10/brlcad/trunk/src/librt/primitives/metaball/ (metaball.c metaball.h metaball_tri.c): quell warnings |
| 21:56.59 | ``Erik | jams his finger up his nose |
| 21:57.27 | CIA-38 | BRL-CAD: 03erikgreenwald * r36948 10/brlcad/trunk/src/librt/ (3 files in 2 dirs): import marching cube vertex and edge tables from Cory Bourke's public domain demo. |
| 22:33.28 | *** join/#brlcad Don_ (n=Don@c-71-238-51-148.hsd1.mi.comcast.net) |
| 22:34.53 | *** join/#brlcad yukonbob_ (i=1000@s142-179-54-198.bc.hsia.telus.net) |
| 22:56.29 | *** join/#brlcad Nohla (n=jesica@168.226.177.29) |
| 23:17.27 | *** join/#brlcad jesica__ (n=jesica@168.226.176.86) |
| 23:20.45 | ``Erik | all stocked up and ready for the blizzard, huzzah |
| 23:52.17 | brlcad | heh |