Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 23 Jul 2013 20:12:28 -0500
From:      Mark Linimon <linimon@lonesome.com>
To:        David Chisnall <theraven@FreeBSD.org>
Cc:        svn-src-head@freebsd.org, svn-src-all@freebsd.org, src-committers@freebsd.org
Subject:   Re: svn commit: r253563 - head/contrib/libstdc++/include/c_std
Message-ID:  <20130724011228.GB20455@lonesome.com>
In-Reply-To: <201307231023.r6NANhGf065713@svn.freebsd.org>
References:  <201307231023.r6NANhGf065713@svn.freebsd.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, Jul 23, 2013 at 10:23:43AM +0000, David Chisnall wrote:
> A surprising number of configure checks rely on this.  It was broken by recent
> cleanups to math.h.

Once you have the experiences with the ports tree that I have had, you
will no longer assume anything about how ports configure checks work
(or many other similar items.)  Whatever the number, I would hardly find
it surprising.

The quality of code in the ports collection varies wildly.  Some of it
is truly professional-quality.  Some of it is written by people who
cannot even tie their own shoelaces.  The bulk of it is somewhere in
the middle -- and many of those people simply do not have the patience
or aptitude to understand the multitude of build and configure systems
that are out there.

We simply don't have the several thousand people that it would probably
take to audit the tens of millions of lines of code involved.

I would like to very politely suggest that regression testing such
changes beforehand is a far more effective strategy -- both technical
and inter-personal -- than simply assuming that either port authors
or maintainers will get such issues correct.  They can be subtle, and
there are an indefinite number of them.

mcl



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20130724011228.GB20455>