[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: on the glpk patch for octave
From: |
Rafael Laboissiere |
Subject: |
Re: on the glpk patch for octave |
Date: |
Fri, 2 Mar 2007 09:00:22 +0100 |
User-agent: |
Mutt/1.5.13 (2006-08-11) |
* John W. Eaton <address@hidden> [2007-03-02 02:16]:
> On 2-Mar-2007, Rafael Laboissiere wrote:
>
> | I did not test test the patch with 4.14, but I trust it should work with
> | 4.14. I chose the preprocessor macro name as GLPK_PRE_4_15 for
> | pragmatic reasons. If it is well tested that the patch work with 4.14, I
> | have nothing against changing the macro name to GLPK_PRE_4_14.
>
> The patch does a feature test in the configure script, so the macro
> name does not indicate that you are testing for a particular version
> number, correct? Instead of a version, the macro could be named
> HAVE_GLPK_SOME_FEATURE, but what should "SOME_FEATURE" be? Unless
> someone can think of something significantly better, then we might as
> well leave it as it is.
I have nothing against changing the macro name, much to the contrary.
However, I cannot find a goof name for it. No specific feature is being
checked, besides the presence of the glp_lpx_simplex symbol, instead of
_glp_lpx_simplex. This is an (almost) undocumented change in the API done
by the upstream author, in which the public/official API has been reduced
and many symbols deprecated. I think that this change took place at version
4.14 of GLPK (although I only tested the patch with 4.15, which is the last
released version).
If someone has the time to do it, I would suggest to test my patch against
GLPK 4.14 and, if it works, change the name of the macro to GLPK_PRE_4_14.
--
Rafael