| 00:01.30 | *** join/#brlcad ibot (~ibot@rikers.org) | |
| 00:01.30 | *** topic/#brlcad is BRL-CAD Open Source Solid Modeling || http://brlcad.org || http://sf.net/projects/brlcad || #brlcad logs: http://ibot.rikers.org/%23brlcad/ || BRL-CAD release 7.16.10 is posted! (20100805) | |
| 00:01.53 | brlcad | ibot: wb |
| 00:01.53 | ibot | thx |
| 00:04.05 | starseeker | rle-pix is import, libutahrle is export |
| 00:05.16 | starseeker | wait, now rle-pix builds?? |
| 00:05.20 | brlcad | heh |
| 00:06.09 | starseeker | they build in isolation, but fail with the ALL_BUILD target |
| 00:07.18 | starseeker | I wonder if the presense of both static and dynamic build targets for the libs is messing with things |
| 00:07.20 | brlcad | so maybe with the ALL_BUILD target, a flag isn't getting set |
| 00:10.50 | starseeker | how do I interupt a build in VS? |
| 00:11.02 | starseeker | nevermind, found it |
| 00:12.01 | CIA-42 | BRL-CAD: 03brlcad * r41067 10/brlcad/trunk/src/libbu/ (avs.c fopen_uniq.c log.c malloc.c parse.c): LIKELY/UNLIKELY macros expect an integer argument, so compare against NULL since that's what's implied. fixes strict build warning. |
| 00:13.43 | CIA-42 | BRL-CAD: 03brlcad * r41068 10/brlcad/trunk/src/librt/bundle.c: strict compilation failure, removed unused var |
| 00:15.36 | CIA-42 | BRL-CAD: 03brlcad * r41069 10/brlcad/trunk/src/librt/mkbundle.c: more quellage to fix build. unused vars. |
| 01:29.45 | starseeker | uh oh |
| 01:29.54 | starseeker | dm-tk.obj : error LNK2019: unresolved external symbol _XSync referenced in function _tk_drawEnd |
| 01:33.34 | CIA-42 | BRL-CAD: 03starseeker * r41070 10/brlcad/branches/cmake/ (5 files in 5 dirs): Don't build the static targets with MSVC. |
| 01:36.03 | starseeker | urk. Searching for Xsync in C:\Tcl returns nothing |
| 01:40.36 | ``Erik | xsync is libX11 |
| 01:40.54 | ``Erik | you won't find it on winderz. |
| 01:41.01 | starseeker | hang on - it might be just dm-tk |
| 01:41.22 | starseeker | if that's the case, the original author's assumption that Xutil would be around might be proving false |
| 01:41.36 | starseeker | fortunately, dm-tk isn't done anyway and may end up not being needed at all |
| 01:41.41 | ``Erik | bob may've skipped dm-tk |
| 01:42.57 | starseeker | I doubt it was hooked in - it's never been more than a test dm, at least to date |
| 01:43.06 | starseeker | will turn it off |
| 01:51.16 | brlcad | starseeker: XSync is a tk "symbol" |
| 01:51.22 | brlcad | rather, is *also* a tk symbol |
| 01:51.38 | brlcad | tk guarantees it even for windows in tkWinPort.h |
| 01:52.56 | brlcad | having it come up as an unresolved external symbol probably means a header file is missing |
| 01:53.35 | ``Erik | eh? |
| 01:54.07 | starseeker | tkWinPort.h is there... |
| 01:54.28 | brlcad | ``Erik: they #define it to something else |
| 01:54.39 | starseeker | and does define XSync |
| 01:54.44 | starseeker | what the bleep |
| 01:55.16 | ``Erik | yeah, ok, tkWinPort.h fakes it |
| 01:55.22 | ``Erik | as a macro |
| 01:56.17 | starseeker | brlcad: are you saying I need to include tkWinPort.h somewhere? |
| 01:56.18 | ``Erik | so a missing symbol would be lacking the #include required |
| 01:57.56 | brlcad | starseeker: that the file has to get eventually/somehow included |
| 01:58.01 | brlcad | maybe not directly, maybe indirectly |
| 01:58.07 | brlcad | however tk documents it being provided |
| 01:58.14 | brlcad | see what all includes it |
| 01:58.33 | brlcad | it being tkWinPort.h because that's possibly a private header too.. |
| 02:00.40 | brlcad | yeah, it's looking on the surface to be a private header, meaning there is some other call that should be made instead of a direct XSync() |
| 02:01.32 | starseeker | tkPort.h |
| 02:02.47 | starseeker | I'm more inclined to just turn off the tk dm for now - it's not really production code anyhow |
| 02:04.44 | starseeker | to really work it'd have to be paired with a tk framebuffer, which in turn needs C-side threading in Tcl... |
| 02:06.40 | brlcad | right, and tkPort.h is included via tkInt.h, so it's not installed |
| 02:06.55 | starseeker | Well, including WinPort directly gets XSync, but not XDrawSegments or XDrawPoint |
| 02:07.14 | starseeker | which I don't see in that file |
| 02:08.18 | brlcad | probably because it was branched off of dm_X and wasn't fully decoupled from X calls |
| 02:08.31 | brlcad | even the XSync was probably inadvertent |
| 02:08.45 | brlcad | I wouldn't think to use that to sync if I were writing a tk interface... |
| 02:08.53 | starseeker | nods |
| 02:09.05 | starseeker | I believe it was approached that way, from what I recall of the code |
| 02:18.43 | CIA-42 | BRL-CAD: 03starseeker * r41071 10/brlcad/branches/cmake/src/libdm/CMakeLists.txt: Don't build DM-TK on Windows - not quite portable yet. |
| 02:21.06 | brlcad | starseeker: what was the problem building static libraries on windows? |
| 02:33.06 | starseeker | not quite sure - things are going a lot smoother without them though. |
| 02:33.37 | starseeker | I'd like to merge the windows specific config.h and brlcad_config.h, if that's technically possible |
| 02:54.45 | starseeker | I have a feeling there's some name conflict at play on Windows given how I defined the macros - i'll have to check png to see how they deal with it |
| 02:55.07 | starseeker | lower priority than getting the basic build working though - need to at least achieve parity with our current windows build |
| 02:59.41 | brlcad | hopes it's not shotgun debugging, should understand why it's not working so having it off isn't added complexity |
| 03:00.40 | brlcad | things would go even more smooth if you disable shared ones too, it's not a race to the finish in any form possible.. |
| 03:04.49 | starseeker | I don't intend to leave them off unless there's some sound reason to on Windows |
| 03:06.43 | starseeker | but I was spinning my wheels - the issues I can see right now are clear cut, and I have at least a rough idea how to approach some of them. It all has to get dealt with, but hopefully it will be easier to see what's going on once I have fewer errors of other sorts joining the party |
| 03:07.51 | starseeker | png has a static target that works and we aren't getting png related errors, so it's clearly possible to do right |
| 03:10.32 | starseeker | here's where matters currently stand: http://pastebin.ca/raw/1966492 |
| 03:12.04 | starseeker | come to think of it, I need to check and see if our current msvc logic can tell me how to do static libs - forgot to check |
| 03:44.05 | CIA-42 | BRL-CAD: 03brlcad * r41072 10/brlcad/trunk/doc/deprecation.txt: move the minimally impacting docs down with that section in order to keep them in localized context. add bu_ptbl() changes. |
| 03:51.35 | CIA-42 | BRL-CAD: 03brlcad * r41073 10/brlcad/trunk/ (10 files in 8 dirs): |
| 03:51.35 | CIA-42 | BRL-CAD: remove bu_ptbl() since it conflicts with 'struct bu_ptbl' causing a compilation |
| 03:51.35 | CIA-42 | BRL-CAD: shadow warning on its constructor for c++ codes. the interface is actually |
| 03:51.35 | CIA-42 | BRL-CAD: duplicitous so code can be trivially updated to the various equivalent |
| 03:51.36 | CIA-42 | BRL-CAD: bu_ptbl_*() calls that it was wrapping. |
| 03:53.49 | brlcad | starseeker: that's good to hear then |
| 03:54.12 | brlcad | as for static libs on windows, the current build doesn't really address it (though daniel's build may) |
| 03:54.34 | brlcad | the .lib files are the static libs .. basically it's the same build but the symbols neither need to be imported or exported |
| 04:18.40 | CIA-42 | BRL-CAD: 03brlcad * r41074 10/brlcad/trunk/ (3 files in 3 dirs): rename nmg_struct_counts() to nmg_pr_m_struct_counts() since it conflicts with the 'struct nmg_struct_counts' constructor during c++ compilation. |
| 04:19.24 | CIA-42 | BRL-CAD: 03brlcad * r41075 10/brlcad/trunk/TODO: renamed bu_ptbl() and nmg_struct_counts() so they no longer hide the struct constructors during c++ compilation |
| 04:25.25 | CIA-42 | BRL-CAD: 03brlcad * r41076 10/brlcad/trunk/src/bwish/input.c: check for sys/time.h since it's posix.1 |
| 04:29.56 | CIA-42 | BRL-CAD: 03brlcad * r41077 10/brlcad/trunk/include/cmd.h: include bio.h so we can get windows.h included so we can get struct timeval defined |
| 04:36.51 | CIA-42 | BRL-CAD: 03brlcad * r41078 10/brlcad/trunk/src/proc-db/surfaceintersect.h: do not directly include stdint.h .. it's included with protections via common.h |
| 04:37.09 | CIA-42 | BRL-CAD: 03brlcad * r41079 10/brlcad/trunk/src/proc-db/surfaceintersect.cpp: assert.h is a system header, remove stale comments |
| 04:46.57 | CIA-42 | BRL-CAD: 03brlcad * r41080 10/brlcad/trunk/src/rt/rtshot.c: quick cleanup, ws, and move variable decls to the top of their scopes for Windows |
| 04:51.07 | CIA-42 | BRL-CAD: 03brlcad * r41081 10/brlcad/trunk/src/ (12 files in 3 dirs): |
| 04:51.07 | CIA-42 | BRL-CAD: replace the 'i' bu_structparse type identifier with '%p' instead so that it's |
| 04:51.07 | CIA-42 | BRL-CAD: more consistent with the other types for representing the pointer indirection to |
| 04:51.07 | CIA-42 | BRL-CAD: another bu_structparse structure and run-on sentences can be awesome on tuesdays |
| 04:51.07 | CIA-42 | BRL-CAD: with developer deprecation warnings firmly in place to blather so any missed |
| 04:51.08 | CIA-42 | BRL-CAD: entries can be readily identified and updated. |
| 04:51.21 | CIA-42 | BRL-CAD: 03brlcad * r41082 10/brlcad/trunk/src/remrt/remrt.c: netdb.h is ancient, remove and hope |
| 04:57.35 | CIA-42 | BRL-CAD: 03brlcad * r41083 10/brlcad/trunk/configure.ac: check for arpa/inet.h and netdb.h for remrt portability protections |
| 05:00.40 | starseeker | brlcad: hmm - maybe I need to do per-target definitions for the *_EXPORT_DLL settings then - they're currently being enabled per-directory, which is probably wrong for the static libs |
| 05:01.56 | CIA-42 | BRL-CAD: 03brlcad * r41084 10/brlcad/trunk/src/remrt/ihost.c: protect all the funky networking headers so windows has a chance. blindly stab in an include for winsock2.h (even though it will have to have init calls added too for things like gethostbyname() to work. |
| 05:11.46 | CIA-42 | BRL-CAD: 03brlcad * r41085 10/brlcad/trunk/bench/ (Makefile.am pixcmp.c): finally cave in to portability. pixcmp now utilizes libbu in order to leverage bu_getopt() for Windows portability. |
| 05:14.28 | brlcad | starseeker: the decls "should" get ignored for static libs .. but dunno, would have to test |
| 05:18.28 | CIA-42 | BRL-CAD: 03brlcad * r41086 10/brlcad/trunk/src/proc-db/ (makebuilding/makebuilding.c mkbuilding.c): remove gratuitous blank lineage and move rgb decl to top of scope for Windows portability |
| 05:19.13 | CIA-42 | BRL-CAD: 03brlcad * r41087 10/brlcad/trunk/src/proc-db/mkbuilding.c: move vars to top of scope for Windows |
| 05:20.13 | CIA-42 | BRL-CAD: 03brlcad * r41088 10/brlcad/trunk/src/proc-db/metaball.c: protect unistd.h for portability |
| 05:30.29 | CIA-42 | BRL-CAD: 03brlcad * r41089 10/brlcad/trunk/src/util/ (pixblend.c ttcp.c): undoubtedly others, but respond to a volley of Windows compilation header inlusion failures |
| 05:32.32 | CIA-42 | BRL-CAD: 03brlcad * r41090 10/brlcad/trunk/src/lgt/extern.h: don't declare errno. we get the linkage wrong on Windows. |
| 05:33.26 | CIA-42 | BRL-CAD: 03brlcad * r41091 10/brlcad/trunk/src/lgt/reflect.c: quell with UNUSED() instead of the hack. it was before one of the variable decls anyways. |
| 05:41.23 | CIA-42 | BRL-CAD: 03brlcad * r41092 10/brlcad/trunk/include/config_win.h: windows has winsock2.h too. |
| 05:42.09 | CIA-42 | BRL-CAD: 03brlcad * r41093 10/brlcad/trunk/src/ (12 files in 10 dirs): more header cleanup for Windows checking for sys/time.h and others. |
| 05:48.50 | CIA-42 | BRL-CAD: 03brlcad * r41094 10/brlcad/trunk/src/conv/jack/g-jack.c: reorder to avoid forward decls, fix nmg_eue_dist() linkage on Windows, pull from header |
| 05:58.07 | CIA-42 | BRL-CAD: 03brlcad * r41095 10/brlcad/trunk/src/fbed/ (extern.h fbed.c popup.h prnt.c): rename Rectangle to Rect2D to avoid naming conflicts on Windows with Rectangle() |
| 05:59.53 | CIA-42 | BRL-CAD: 03brlcad * r41096 10/brlcad/trunk/misc/enigma/enigma.c: try to stay portable, key on _WIN32 for unistd.h |
| 06:02.40 | CIA-42 | BRL-CAD: 03brlcad * r41097 10/brlcad/trunk/src/conv/dem-g.c: remove inclusion of stdbool.h for Windows compatibility. update references to true/false/bool accordingly. |
| 06:05.45 | CIA-42 | BRL-CAD: 03brlcad * r41098 10/brlcad/trunk/src/proc-db/csgbrep.cpp: add fixme since these are not supposed to be public functions and are not exported. |
| 06:10.50 | CIA-42 | BRL-CAD: 03brlcad * r41099 10/brlcad/trunk/src/burst/ (13 files): remove boolean, pointer, and DEGRAD. replace with the usual. |
| 06:13.32 | *** join/#brlcad cosurgi (~cosurgi@atak.bl.pg.gda.pl) | |
| 06:14.19 | CIA-42 | BRL-CAD: 03brlcad * r41100 10/brlcad/trunk/src/proc-db/brepintersect.h: tsk tsk jdoliner... get rid of the insane #include lines. someone apparently didn't know how cppflags work, undo the relative paths. |
| 06:16.08 | CIA-42 | BRL-CAD: 03brlcad * r41101 10/brlcad/trunk/src/adrt/master/master.c: check for pthread.h before including |
| 06:18.57 | CIA-42 | BRL-CAD: 03brlcad * r41102 10/brlcad/trunk/src/adrt/master/ (compnet.c master.c): more windows header inclusion protections. |
| 06:22.47 | CIA-42 | BRL-CAD: 03brlcad * r41103 10/brlcad/trunk/src/adrt/master/ (dispatcher.c tienet_master.c): and yet even more Windows header inclusion protections. |
| 06:26.35 | brlcad | and that should be all of the easily fixable errors from the Windows build log |
| 07:39.12 | CIA-42 | BRL-CAD: 03brlcad * r41104 10/brlcad/trunk/bench/Makefile.am: blasted tcl includes are needed for bu.h |
| 08:45.34 | *** join/#brlcad merzo (~merzo@smartbussiness.mobicom.net.ua) | |
| 09:22.40 | CIA-42 | BRL-CAD: 03d_rossberg * r41105 10/brlcad/trunk/misc/win32-msvc/Dll/BrlcadCore.def: bu_ptbl() was removed |
| 09:50.16 | *** join/#brlcad Ralith (~ralith@S010600221561996a.vc.shawcable.net) | |
| 10:03.18 | *** join/#brlcad mafm (~mafm@81.32.97.106) | |
| 10:33.03 | brlcad | hmm |
| 11:07.47 | *** join/#brlcad d_rossberg (~rossberg@BZ.BZFLAG.BZ) | |
| 11:09.26 | d-lo | Mernin |
| 12:14.34 | brlcad | howdy |
| 12:15.17 | brlcad | d_rossberg: thanks for the (more carefully worded) help responding to the floating point forum person |
| 12:18.12 | d_rossberg | but i don't think he understood |
| 12:26.47 | brlcad | yeah, I got that impression as well |
| 13:37.24 | d_rossberg | i tried the cmake build on windows but got nothing (i.e. only a tree of mainly empty directories) |
| 13:39.30 | *** join/#brlcad Zaebos (~irc@pd95b7f5e.dip0.t-ipconnect.de) | |
| 15:42.40 | *** join/#brlcad mafm (~mafm@81.32.97.106) | |
| 15:49.29 | *** join/#brlcad Elrohir (~kvirc@p4FC5AB47.dip.t-dialin.net) | |
| 16:32.03 | CIA-42 | BRL-CAD: 03starseeker * r41106 10/brlcad/branches/cmake/ (85 files in 29 dirs): Update cmake branch to r41105 |
| 16:43.50 | CIA-42 | BRL-CAD: 03starseeker * r41107 10/brlcad/branches/cmake/bench/CMakeLists.txt: Makefile.am added Tcl includes, so probably should do the same for CMake... |
| 17:11.23 | CIA-42 | BRL-CAD: 03starseeker * r41108 10/brlcad/branches/cmake/CMakeLists.txt: Need to check for netdb for adrt. |
| 17:17.19 | CIA-42 | BRL-CAD: 03starseeker * r41109 10/brlcad/branches/cmake/ (4 files in 4 dirs): |
| 17:17.20 | CIA-42 | BRL-CAD: If the static libraries really are .lib files that aren't doing the |
| 17:17.20 | CIA-42 | BRL-CAD: dll_import/dll_export thing, then it's quite plausible that the static builds |
| 17:17.20 | CIA-42 | BRL-CAD: were actually overwriting the dynamic .lib linking files, whereas on other |
| 17:17.20 | CIA-42 | BRL-CAD: platforms the different extension avoids any issue. Let's try that and enable |
| 17:17.20 | CIA-42 | BRL-CAD: static libs on WIN32 again. |
| 17:32.46 | CIA-42 | BRL-CAD: 03bob1961 * r41110 10/brlcad/trunk/src/tclscripts/archer/ArcherCore.tcl: Modified ArcherCore::updateTreeTopWithName to account for mPNode2CList() possibly not existing. |
| 18:41.24 | *** join/#brlcad cosurgi (~cosurgi@atak.bl.pg.gda.pl) | |
| 19:13.11 | CIA-42 | BRL-CAD: 03bob1961 * r41111 10/brlcad/trunk/src/tclscripts/archer/ (DataUtils.tcl tclIndex): Added DataUtils::measureLastDataPoints for measuring the distance between the last two data axes. |
| 19:58.55 | starseeker | ``Erik: Cmake doesn't seem to be working on bz |
| 20:33.08 | *** join/#brlcad R0b0t1 (~Enigma@64-136-219-55.dyn.everestkc.net) | |
| 20:33.08 | *** join/#brlcad R0b0t1 (~Enigma@unaffiliated/r0b0t1) | |
| 20:33.20 | *** join/#brlcad Ralith (~ralith@d142-058-093-168.wireless.sfu.ca) | |
| 20:42.45 | *** join/#brlcad Ralith (~ralith@d142-058-093-168.wireless.sfu.ca) | |
| 21:36.28 | *** join/#brlcad Ralith (~ralith@d142-058-093-168.wireless.sfu.ca) | |
| 22:31.40 | *** join/#brlcad Ralith (~ralith@d142-058-093-168.wireless.sfu.ca) | |
| 22:49.13 | starseeker | brlcad: looks like dumpbin might be something like the nm command... |