[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Extension of Built-Using:



Hi,

I've been looking at using the "Built-Using" tag for dh-fortran-mod.

dh-fortran-mod is a debhelper extension for handling Fortran "mod" files (based on an original idea from Sebastian Villemont).

These mod files are effectively pre-compiled header files, in C/C++ terms; normally stored in /usr/include, or something like it. I've been adding  a new Fortran compiler ("flang", based on LLVM) which is in the NEW queue which is part of the driver for this. The trouble is that "mod" files are compiler-specific, and even version of compiler-specific. So when we moved from gfortran-7 to gfortran-8, the 'mod' files were incompatible and needed to be rebuilt. Similarly, flangs mod files are incompatible.

So dh-fortan-mod does two things:
(1)Adds a dependency / track on which compiler was used, to enable tracking for transitions.

(2) Puts the mod files in $fmoddir  where the compiler will get them, allowing co-installation of incompatible files.

(e.g. $fmoddir = /usr/lib/$multiarch/fortran/$fortran-mod-version, by default).

(1) Is the issue today. My initial plan was to add eg. 'gfortran-mod-15' to ${misc:Depends}. The trouble is, this adds a compiler dependency to a package that is unnecessary - some packages have a Fortran interface that has marginal numbers of users; most use the C/C++ interfaces. So, plan B was to use Built-Using, as in:

Built-Using: gfortran-8 (= 8.2.0-6)

The difficulty here is that Policy 7.8 requires that Built-Using: is only used for source package tracking. This is then enforced on the upload package checking which rejects such packages (because gfortran-8 is not a source package; gcc-8 is the source package, but this mostly misses the point).

So, can Built-Using: be safely extended to use this case, and the package checking relaxed ?


best regards

Alastair

--
Alastair McKinstry, <alastair@sceal.ie>, <mckinstry@debian.org>, https://diaspora.sceal.ie/u/amckinstry
Misentropy: doubting that the Universe is becoming more disordered.


Reply to: