Intel C++ 8 monster patch

[David Relson]

On Fri, 30 Jan 2004 22:14:04 +0100
Matthias Andree wrote:

> David,
> 
> I am enclosing a diffstat and a cvs diff for pending deltas to quench
> Intel C++ warnings and notes. I have already committed the safe parts,
> but these are a bit more intrusive, so I'd like to negotiate the merge
> date first.
> 
> Please let me know when I should commit these changes.
> 
> They have been test compiled on Solaris 8 SPARC64 with GCC 3.2 and Sun
> Workshop, SuSE Linux 8.2 i386 with GCC 3.3, 2.95.3, 2.7.2 and
> ICC 8.0.055, FreeBSD 4.9 with gcc 2.95.3+FreeBSD ext's, no regressions
> reported.
> 
> Note that the configure.ac changes qualify as a bug fix because the
> SETVBUF_REVERSED check falsely reports "yes" on ICC which leads to
> segfaults (make check exposes these).
> 
> Cheers,

Matthias,

Would tomorrow be a good day for the monster patch?  I've got several
items on my pending list for 0.17.0, specifically:

1 - Remove deprecated code.

2 - long options for bogofilter and bogolexer

3 - bogoutil maintenance - "and" together the different qualifiers. 
"or"ing can be done by multiple bogoutil runs.  As an example, removing
old tokens _and_ tokens with low counts must be done by a script as
bogoutil can't do it.  With the changes, that will become doable. 
Anybody wanting to remove tokens that are old _or_ have low counts can
run bogoutil twice - once for age and once for counts.

I'll do those this evening.  With your changes tomorrow, I'll be able to
build 0.17.0 on Sunday.

OK???

David