from website: The CFD General Notation System (CGNS) provides a general, portable, and extensible standard for the storage and retrieval of computational fluid dynamics (CFD) analysis data. It consists of a collection of conventions, and free and open software implementing those conventions. It is self-descriptive, machine-independent, well-documented, and administered by an international steering committee. It is also an American Institute of Aeronautics and Astronautics (AIAA) Recommended Practice. Reproducible: Always I tried to create ebuilds for sci-libs/cgnslib and sci-misc/cgnstools but they do not work, please help me to fix them.
Created attachment 117635 [details] sci-libs/cgnslib-2.4.5.ebuild ebuild for stable version. 1. failed on install, have no idea why, it package uses autotools, but "emake DESTDIR=${D} install" and einstall do not work. 2. configure cannot find fortran it would be good if sombody can help me...
Created attachment 117637 [details] sci-libs/cgnslib-2.5.1.ebuild this is for newer version, which considered to be beta the problems are the same...
Created attachment 117638 [details] sci-misc/cgnstools-2.5.1.ebuild tools for viewing/editing ADF file (Tcl/Tk based), cannot not check/install because of dependance on cgnslib. some questions anyway: how to make proper dependance on X? how to obtain tcl/tk version installed on the system, since configure need it?
Created attachment 125985 [details] working cgnslib-2.5.1.ebuild ugly hack for ${D} variable, ebuild installed with notice message: * QA Notice: Pre-stripped files found: * /var/tmp/portage/sci-libs/cgnslib-2.5.1/image/usr/lib/libcgns.so * QA Notice: The following files contain insecure RUNPATH's * Please file a bug about this at http://bugs.gentoo.org/ * with the maintaining herd of the package. * /var/tmp/portage/sci-libs/cgnslib-2.5.1/work/cgnslib_2.5/LINUX usr/lib/libcgns.so Auto fixing rpaths for /var/tmp/portage/sci-libs/cgnslib-2.5.1/work/cgnslib_2.5/LINUX usr/lib/libcgns.so * QA Notice: The following shared libraries lack a SONAME * /var/tmp/portage/sci-libs/cgnslib-2.5.1/image/usr/lib/libcgns.so what does it mean and how to fix it? thanks...
Created attachment 125987 [details, diff] files/DESTDIR.patch patch for Makefile.in in order to use DESTDIR variable
Created attachment 136548 [details] Version bump. This is a working version bump of the cgnslib-2.5.2
Created attachment 136550 [details] working version bump This is a working version bump for cgnstools-2.5.2
Created attachment 136652 [details] change wired $SYSTEM directory LINUX64 directory name was wired in ebuild and it did not worked on x86 where LINUX directory expected
Created attachment 137080 [details] Improved ebuild. The system.patch won't work for amd64 systems!
This two ebuild's are now also in the sunrise overlay: http://overlays.gentoo.org/proj/sunrise/browser/reviewed/sci-libs/cgnslib and http://overlays.gentoo.org/proj/sunrise/browser/reviewed/sci-libs/cgnstools
it is still old version in sunrise overlay (with LINUX64 bug). one more note: could you please add || die "install failed" string for lines in src_install(), it would be easy to debug on other architectures (now there is no error if installation failed, you have to check log manually)
The ebuild's should now be in the reviewed sunrise tree.
Created attachment 160393 [details] Version bump to 2.5-3 Added some additional functionality. But I think there are some remaining issues with the right soname.
These ebuilds are from now on in the science overlay: http://overlays.gentoo.org/proj/science/browser/overlay/sci-libs/cgnslib http://overlays.gentoo.org/proj/science/browser/overlay/sci-libs/cgnstools
The science overlay ebuilds need the tk/tcl update mentioned in the recent post to gentoo-dev about converting Tcl_Interp->result to use the new accessor function.
(In reply to comment #15) > The science overlay ebuilds need the tk/tcl update mentioned in the recent > post to gentoo-dev about converting Tcl_Interp->result to use the new > accessor function. Could you please use githubs fork and pul-request and add the fix to this package?
Just to confirm, you are asking for a pull request for the overlay and not for cgns upstream, yes?
(In reply to comment #17) > Just to confirm, you are asking for a pull request for the overlay and not > for cgns upstream, yes? right, the overlay.
OK. It will take a while. The way cgnstools uses the result pointer makes the conversion non-trivial.
Or not. It turns out upstream already fixed this, so I just copied the required changes (only). Pull request forthcoming.
