I've checked the ebuild and it's working here. Suggested category: sci-misc, until there are enough ebuilds to justify sci-xtal (prefered) or sci-crystallography. JANA2000 is a system for solving and refinement of regular, modulated and composite structures from monocrystal and powder diffraction data. It is a freeware available for most UNIX workstations like FORTRAN77 and C sources and for PC like a compiled program for Windows. JANA2000 has been developed with Lahey Fortran 95 compiler. Basic features of JANA2000 Integrates all usual tasks of the structure analysis except plotting program and direct methods. For direct methods Jana can directly call SIR97 or EXPO. Works with both single crystal and powder data Works with both X-ray and neutron data Works with ordinary, modulated and composite structures Reads output from many diffractometer types and can merge data sets from various sources Works with various types of twinned crystals and overlapped reflections Can read and export SHELX data and CIF Support for commensurate structures Support for multipole refinement and Anharmonic displacement parameters Enables f' and f'' refinement Enables special modulation functions and orthogonalization method Calculates and plots x4 and t contour maps for modulated structures Calculates and plots (joint) probability density functions Calculates and plots various parameters like a function of t All tools are integrated in a unific user interface Limitations No automatic structure solution Does not contain direct methods and plotting program Neutron TOF powder data not supported Reproducible: Always Steps to Reproduce: Maybe I missed some dependencies.
Created attachment 50210 [details] sci-misc/jana2000-ebuild
Created attachment 50211 [details, diff] patch to select the right CC in the makefile
Perhaps the package should be named jana instead, since they apparently plan to start releasing 2006 versions rather than 2000.
To me it rather sounds like we should at least slot both versions then, if not using two seperate packages, as Jana2006 seems to be a new major release in normal linux terms nad some people might actually want to use the old version to reproduce results from older refinements.
Slots are only for using when the actual package itself is installed in a slottable manner. If that's the case, then sure. Otherwise, people are free to upgrade/downgrade/mask as they please.
Created attachment 197396 [details] ebuild for jana2000 ebuild for the "final" jana2000 verision (20071218)
Created attachment 197397 [details, diff] patch for Makefile This patch sets up the Makefile for g77
Created attachment 197398 [details, diff] patch for Makefile This patch changes the makefile for using gfortran
I would suggest sci-chemistry, since all crystallography related packages are there already. for the package-name I suggest jana2000, since this can stay independent of a future jana2006 package. Alternatively, the package name could be jana with a slot "2000" (?is that possible?) and the future jana2006 will have its own slot.
Created attachment 198646 [details, diff] new patch includes LDFLAGS now
Created attachment 198649 [details, diff] new patch includes LDFLAGS now
Looks good, I will commit it tonight.
The most recent update of the ebuild on 16 Dec 2010 broke it. The problem is that $(tc-getFC) returns a lengthy string such as "x86_64-pc-linux-gnu-gfortran", therefore the string comparison in src_unpack fail and patches are not applied Just by replacing "==" with "=~" the ebuild will be repaired. Can somebody with the permissions to commit plase correct this?! thanks
done.