On Nov 9, 7:25 am, casevh <cas...@gmail.com> wrote:
> Hi,
>
> I was testing mpir 1.3.0rc2 and gmpy on FreeBSD 8.0 RC2 AMD64. I
> tracked down a gmpy error to a change in behavior in mpf_set_str: it
> returns an error when passed the string "3.400e+00" but "3.400e+0" is
> accepted.
>
> Both strings are accepted on Windows with mpir-1.3.0rc1. I haven't
> tested Linux yet.
>
> Notes on FreeBSD 8.0: By default, it uses an old version of binutils
> and GCC 4.2.1. Compilation failed with
>
> tmp-addlsh_n.s:75: Error: suffix or operands invalid for 'sahf'
> tmp-addlsh_n.s:85: Error: suffix or operands invalid for 'lahf'
>
> I compiled (via ports) the current version of binutils and GCC 4.4.
> MPIR compiles and tests successfully with this combination.
Hi Case,
The issue is clearly in my rewrite of a part of the old set_str that
we wanted to replace in order to remove any doubt about the LGPL v2.1
status of this file.
The problem is that, as far as I can discover, there is no precise
specification of the syntax and semantics of string input and this
means that its behaviour is only defined by the code itself.
For my part in doing the rewrite I don't want to look at the patch
involved or the resulting code so I have had to guess what behaviour
we want.
If anyone has a precise and complete 'non-code' specification of the
syntax and semantics of string input, I will willingly implement it.
In the meantime I can see what causes your problem is and I will
adjust set_str.c to remove this behaviour (which is more an error on
my part than a specification issue).
Brian Gladman
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"mpir-devel" group.
To post to this group, send email to mpir-devel@googlegroups.com
To unsubscribe from this group, send email to
mpir-devel+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/mpir-devel?hl=en
-~----------~----~----~----~------~----~------~--~---
