Advice about /usr/ports/math/gmp

[Peter Jeremy]

On 2014-Jan-03 17:53:02 +0100, Torbjorn Granlund <tg at gmplib.org> wrote:
>We have been forced to add three FreeBSD-related items to the releases
>notes:

Your use of "work" in the following notes is ambiguous.  Are you referring
to building gmp, another program linking against gmp or executing a program
that links against gmp?

>  * This release will not work on FreeBSD/amd64 7.x, 8.x or 9 series
>    before 9.3 with a Haswell CPU or any other CPU which supports the
>    BMI2 instructions.  The reason is that the FreeBSD m4 command is not
>    correctly implemented.  (Workaround: Use an older GMP release, or
>    install GNU m4 from /usr/ports and tell GMP to use it.)

Does gmp use m4 other than during the build process?  Using GNU m4 to
build gmp is a relatively painless workaround.

>  * This release will not work on FreeBSD/amd64 before version 10 using
>    the 32-bit ABI.  The reason is broken limits.h and broken dynamic
>    linking.  (Workaround: Use an older GMP release if using the 32-bit
>    ABI on these FreeBSD releases is important.)

Again, what exactly do you mean?  Building code using the 32-bit ABI
on FreeBSD/amd64 was not supported prior to FreeBSD 10 so it was only
a lucky fluke if it previously worked.  If you mean that a 32-bit program
using gmp (ie one compiled on FreeBSD/i386) will not run on FreeBSD/amd64,
please provide more details (ie the PR number).

>  * This release will not work on FreeBSD/amd64 10.0 using the 32-bit
>    ABI.  The reason is bugs in the compiler 'clang'.  (Workaround:
>    Compiling gcc from /usr/ports might work, except that gcc depends on
>    GMP; we have not been able to test that workaround since
>    FreeBSD/i386 10.0 does not work for us under KVM or Xen.)

Please clarify what you mean by "work" here.  Since FreeBSD 10.0 is
the first release where building i386 executables is supported, it is
quite likely there are bugs in this area.  Can you please provide more
details (ie the PR numbers) on the problems you encountered with 32-bit
gmp 5.2 on FreeBSD/amd64 10.0 as well as the problems you encountered
using FreeBSD/i386 10 on KVM and Xen.

On 2014-Jan-03 21:33:12 +0100, Torbjorn Granlund <tg at gmplib.org> wrote:
>This is not the right place for FreeBSD bug reports, and I did not
>indend to do so.

Agreed.  It _is_ expected that you would include references to the
bug reports though.

>  I was only trying to ease the maintenence of
>/usr/ports/math/gmp.

Except that by providing vague and incomplete references to problems,
you aren't doing anything to ease the ports maintenance.  From the
information you provided, it's impossible to determine if the port
should have GNU m4 as a build and/or run dependency.

>I have reported two of these problems.  (I have not reported the
>remaining two problems, and I do not intend to do that.  Sorry.  Again,
>I feel that I would abuse thios mailing list by discussing that here.)

So, what you're saying is that gmp 5.2 has problems on FreeBSD but you
aren't going to provide any details that would assist us to resolve them.
This just wastes everyone's time.

-- 
Peter Jeremy
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 326 bytes
Desc: not available
URL: <http://lists.freebsd.org/pipermail/freebsd-ports/attachments/20140105/5b68db28/attachment.sig>