On Sat, Jul 05, 2014 at 12:45:51PM +0200, Jerome BENOIT wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Hello Bill, > > On 05/07/14 12:29, Bill Allombert wrote: > > On Sat, Jul 05, 2014 at 11:24:37AM +0200, Jerome BENOIT wrote: > >> -----BEGIN PGP SIGNED MESSAGE----- > >> Hash: SHA1 > >> > >> Hello Bill, > >> > >> thanks for your prompt reply. > >> > >> gives, as concerned GAP: > >> > >> > >> checking for GAP root directory... /usr/lib/gap > >> checking for GAP architecture... x86_64-pc-linux-gnu-gcc-default64 > >> checking for GAP include files... /usr/lib/gap/src/compiled.h > >> checking for GAP config.h... > >> /usr/lib/gap/bin/x86_64-pc-linux-gnu-gcc-default64/config.h > >> checking for GAP's gmp.h location... not found, GAP was compiled without > >> GMP > > > >> I am agree that this issue may be not relevant for the GAP IO package > >> itself, but it may be for > >> other GAP packages (e.g., float). The non location of the GMP header > >> introduces some inconsistencies > >> in the building process, as such it may be fixed. > > > > GAP does not provide a file gmp.h. Instead, this file is part of libgmp-dev. > > GAP is configured in with --with-gmp=system which is a supported > > configuration. > > I am agree. > > > AC_FIND_GAP is not shipped with GAP so it is a bug in IO. > > AC_FIND_GAP is used by some GAP package, so the issue does not concern only > GAP-IO. > Even if I am agree to say that the implemented location of gmp.h is weak, it > seems > nevertheless that an external include hierarchy is expected in > `/usr/lib/gap/bin/<GAP triplet>/extern'. > If this assertion is true, then the --with-gmp=system in GAP does not > complete its jobs properly, > so it might be a bug in the GAP source itself --- after all, it is GAP itself > that may point the GMP material with which it was built.
I do not see how gap can generate the extern hierarchy when using --with-gmp=system and doing so would be very awkward, and I would rather not support it in Debian anyway. > What do you think ? I reported the bug upstream. Cheers, Bill. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
