Advice about /usr/ports/math/gmp

Peter Jeremy peter at rulingia.com
Tue Jan 7 07:18:06 UTC 2014


On 2014-Jan-07 03:11:28 +0100, John Marino <freebsd.contact at marino.st> wrote:
>Yes, the criticism of "well, what exactly is wrong?" is valid, but the
>guy did write a couple of PRs,

Well, he claims to have written a couple of PRs but has so far refused
to divulge any details.  Tijl found one but we have no information on
the other one.

On 2014-Jan-07 04:25:27 +0100, Matthew Rezny <mrezny at hexaneinc.com> wrote:
>in my mind. Specifically this:
>"I was about to fix it, but found that the license of the code in
>question was unacceptable to me. I only contribute to Free Software,
>not to Open Source code bearing a license analogous to a law system
>that permits people to sell themselves as slaves."

I liked that wording and made a note that next time I wanted to send a
bug report regarding some GPL'd code, I should state:
"I was about to fix it, but could not understand the license of the
code in question.  I only contribute to Free Software, not to possibly
Open Source code bearing a license that requires a law degree to
understand."

>From his email, he is simply unhelpful. Essentially just "you're
>different from GNU/Linux so you're broken but I won't bother to say
>how". Great, so helpful.

I agree.  And, since he's posting from using a GMP Project address, I
presume this is official GMP policy.  I guess the approved mechanism
for reporting bugs in GMP is to include a vague "GMP is broken"
statement, with an ambiguous explanation, in the release notes of
software that uses or is used by GMP.  If pressed, you should state
that you can't be bothered to do anything further.

On 2014-Jan-07 04:37:02 +0100, John Marino <freebsd.contact at marino.st> wrote:
>Yes, if that's an accurate quote, it doesn't reflect well on Torbjorn.

It was.

>Okay, so let's say he's correct and FreeBSD's base m4 can't build the
>latest gmp.  Even if he won't tell us what the problem is, at least he
>told us there is a problem which could be easily reproduced.

That's one problem.  The other problems he mentions are:
:>  * 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.

There's no indication of what "will not work" means.  It's not clear
whether he's talking about compiling (which wasn't supported before
10.x) or running GMP.

:>  * This release will not work on FreeBSD/amd64 10.0 using the 32-bit
:>    ABI.  The reason is bugs in the compiler 'clang'.

Again, there's no indication of what "will not work" means.

:>    FreeBSD/i386 10.0 does not work for us under KVM or Xen.)

That's hardly a recipe for reproducing a problem.

-- 
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/20140107/b6b84907/attachment.sig>


More information about the freebsd-ports mailing list