| Summary: | sci-physics/xfoil-6.97 - make: f77: Command not found | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Bob Johnson <bob> |
| Component: | Current packages | Assignee: | Robin Johnson <robbat2> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | sci-physics |
| Priority: | High | ||
| Version: | unspecified | ||
| Hardware: | x86 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
|
Description
Bob Johnson
2008-08-23 05:20:37 UTC
Well, I tried installing ifc (the binary turns out to be ifort not ifc btw). Then tried "FC="ifort" emerge -uDv xfoil". Nope. It still compiles plotlib with gfortran, then dies when linking because it can't find the gfortran library. If I go to the work directory, I can't compile orrs with 'FC="gfortran" make' as their are unknown command line options. I can get orrs to compile with 'FC="ifort" make'. Trying to compile plotlib with 'FC="ifort" make' doesn't work, as the Makefile includes make.config which has uncommented both Linux and SGI sections, and FC="gfortran" ends up being set no matter what. Commenting out the SGI section allows plotlib to build with ifort. When I then go into xfoil/bin and try 'FC="ifort" make' the build fails with: ld: ../plotlib/libPltDP.a: No such file: No such file or directory That is because the Makefile in xfoil/bin has both the single precision and double precision options commented, and the double precision option is the winner. Unfortunately, plotlib compiles the single-precision version of the library. Commenting out the double-precision section of the xfoil/bin/Makefile allows 'FC="ifort" make' to complete without error. There seem to be two problems here: 1) Systems with only gfortran won't compile due to some command line option issue in orrs. I haven't investigated this any deeper. 2) The ebuild script doesn't set up the makefiles properly no matter what compiler is used. Plotlib will always build with gfortran and single-precision, and xfoil will try to link to the double-precision plotlib. IMO the best option here is to modify the upstream makefiles and have patch files accompany the ebuild. It's late. I'll look into this more tomorrow. It should be fixed in cvs. Please re-emerge after a sync. If it persists, re-open this bug. Thanks. |