Re: [gcc 4.5 patch] Permit smaller hash tables for hash_map and hash_set

[Ian Lance Taylor <iant at google dot com>]

Paolo Carlini <paolo.carlini@oracle.com> writes:

> Well, it depends on the maintainer ;) For sure *I* do not want to do
> fix regression in such legacy code in ext/. Actually, if you ask me, I
> would have issued a deprecation notice for both hash_map and hash_set
> and then removed both at the first occasion, really, it's something
> obsoleted by the TR1 and C++0x standard facilities.

That is one possible approach.  There is indeed already a deprecation
warning for these classes.  However, codesearch.google.com finds some
4000 uses of __gnu_cxx::hash_map and __gnu_cxx::hash_set, in free
software alone.  I don't think it is wise to quickly break existing
working code.  That does not serve our user base.


> Anyway, If
> Benjamin agrees to keep on maintaining those two containers vs
> regressions caused by new commits, then fine, just go ahead, of
> course.

Thanks.

Ian