latest in portage is 0.2.5.4!!
there's a big difference between k3d and k3b.
*** Bug 51150 has been marked as a duplicate of this bug. ***
Yeah. K3b is for CD writing. K-3D is for 3D graphics and it is not tightly integrated into KDE....
Actually, K-3D is desktop agnostic ... not KDE-specific at all. A better category is "graphics". Sorry for the confusion, but the name predates my involvement with Free Software ;) Cheers, Tim Shead
*** Bug 57273 has been marked as a duplicate of this bug. ***
Created attachment 35583 [details] New ebuild for k3d-0.4.1.0
Actually, 4.2 is now out. Can we get this into Portage soon?
0.4.3.0 is out. Any news on portage integration?
Created attachment 39408 [details] media-gfx/k3d-0.4.3.0.ebuild Fixed ebuild for 0.4.3.0 (now depends on >=media-gfx/netpbm-10)
K-3D 4.4.0 is out :) http://sourceforge.net/project/showfiles.php?group_id=11113&release_id=280211
Created attachment 45303 [details] media-gfx/k3d-0.4.4.0.ebuild Updated ebuild for K-3D 0.4.4.0
K-3D is a very advanced modeling and animation package actively developed so why give others the impression that it is dead by only including a very old, now obsolete ebuild? I guess that's the price to pay for not being a blender kid! Cheers, -Giuseppe
Just before hacking the configure and the ebuild some questions: - Does it require ode? - Does it work with graphviz? - Why it is using gtk-1.2?
Luca: I am the original author and maintainer of K-3D. I'm also a big fan of Gentoo (I use it at home and at work), so I'm keen to get up-to-date versions of K-3D and its sister project Aqsis into portage. Please feel free to contact me directly with any concerns or suggestions you may have. I know that the K-3D 0.2.5.4 ebuild currently in portage required some patching, but my understanding of autotools has matured a lot since then, and you shouldn't need to make any changes at all to get a clean ebuild, let me know if you find otherwise. Our ODE code is prototypical at best, so I recommend passing --without-ode to configure. We optionally use graphviz in our doxygen-generated developer docs, which aren't installed, so I recommend passing --without-graphviz to configure. The 0.5 development branch of K-3D uses gtk2. That K-3D 0.4 uses gtk 1.2 is partly conservatism on my part, partly a reflection of how far behind the ebuild is - we will be releasing a stable 0.6 build in the coming year. A word on Aqsis - K-3D is a 3D modeling and animation system - it cannot render images on its own. Although there are no compile- or link- dependencies, I recommend making Aqsis a mandatory requirement for the K-3D ebuild. See Bug #34525 for an Aqsis ebuild that I've submitted. Many Thanks! Tim Shead
Current issue I'm patching now jsprf.c: In function `BuildArgArray': jsprf.c:620: error: incompatible types in assignment make[5]: *** [jsprf.lo] Error 1 make[5]: Leaving directory `/var/tmp/portage/k3d-0.4.4.0/work/k3d-0.4.4.0/modules/javascript/js/src' that code is exactly the same as mozilla one, I guess there is a configure issue somewhere.
the solution is to add the defines in order to use va_copy. -DHAVE_VA_COPY -DVA_COPY=va_copy
Interesting, I just did a complete install of the submitted 0.4.4.0 ebuild without any problems at all ... what architecture/compiler are you using? I am using gcc (GCC) 3.3.4 20040623 (Gentoo Linux 3.3.4-r1, ssp-3.3.2-2, pie-8.7.6) on x86 here. In any case your patch sounds harmless enough, if you want to send it along I could test it and get a release out ASAP so you don't have to do any patching in the ebuild. Many Thanks! Tim
Nice to see the original author of K-3D is also using Gentoo :) I hope K-3D will soon be officially in portage! (And aqsis too)
Suggestion to the K-3D author: Tomorrow (8th Jan. 2005) is a gentoo bug day: http://bugday.gentoo.org/ The developers will be reachable on IRC: #gentoo-dev Maybe it is a good change to ask them once again to include K-3D and Aqsis again! Sincerly Art
tested on ppc, the problem due a missing config.h-like include in the javascript engine (see how mozilla defines va_copy), sorry for the delay
Created attachment 47927 [details, diff] PPC HAVE_VA_COPY patch Patch to fix Javascript build issue on PPC. If it works I will apply it upstream to simplify future ebuilds.
Luca: Let me know if the patch I've attached fixes this issue for you. If yes, I'll release 0.4.4.1 ASAP in hopes of getting it into Portage. Cheers, Tim
It fixes the issue but could lead to some problems with arches/system in which VA_COPY is defined as __va_copy or in systems in which you have simple pointer operations. A better solution is add the relative am macros and generate the proper config.h file/Makefile CFLAGS. Have someone test it on x86 and amd64 (just to name a couple)
I would test it on a pentium 4 notebook, if there would be instructions on what to test. I don't know what part of the application I've to use, to trigger the possible "bug" of the application. If there would be an instruction or an demo file, I would do it.
I will figure out how to do this "the right way" and release 0.4.4.1 in the next day or so. Am getting a crash course in va_copy() and co.
In the mean time I tried k3d and aqsis. just a problem after closing the first tutorial: Gdk-ERROR **: BadValue (integer parameter out of range for operation) serial 28092 error_code 2 request_code 142 minor_code 9 /usr/bin/k3d: line 182: 2815 Segmentation fault /usr/bin/k3d-bin --dialogpath /usr/share/k3d/dialogtemplates --libpaths /usr/lib/k3d --shadersourcepaths /usr/share/k3d/shaders --shadercachepath /home/lu_zero/.k3d/shadercache --sharepath /usr/share/k3d --tutorialspath /usr/share/k3d/tutorials --optionsfile /home/lu_zero/.k3d/options.k3d
We have a complete set of regression tests if you run "make check" that include running the tutorials (among other things), so I'm confident that this is a PPC-specific issue. Since you're the first person I know to try to run K-3D on PPC, could you send me a backtrace? Cheers, Tim
Created attachment 48271 [details, diff] Better HAVE_VA_COPY patch Works fine on x86, requires testing on ppc and other platforms.
make[2]: Entering directory `/var/tmp/portage/k3d-0.4.4.0/work/k3d-0.4.4.0/application' ../application/k3d-bin --gui none --shadercachepath . --sharepath . --optionsfile . lt-k3d-bin: sdpxmlsax.cpp:34: char <unnamed>::getchar(std::istream&, long unsigned int&, long unsigned int&): Assertion `Stream.good()' failed. make[2]: *** [check-local] Aborted
Luca: Seems as if we're at an impasse; obviously there are some serious issues on ppc. I don't have access to ppc hardware so I welcome any assistance you can provide on that platform. In the meantime the x86 build is extremely stable, can we set the appropriate masks and get it into portage? Many thanks, Tim
Created attachment 48337 [details] K-3D 0.5.0.0.ebuild (live cvs) This is a shot in the dark, but so much has changed in the 0.5 branch that I'd be interested to hear if it's any improvement over 0.4 on platforms other than x86. Cheers, Tim
Marked as fixed: k3d-0.4.4.0 is in portage as ~x86 and we are working on getting it fixed for ppc
