Re: Autoconf 2.54 is released

[Paul Eggert]

> Sascha>     The two build systems I authored (for the Apache HTTP
> Sascha> Server and PHP 4.x) fully support and make use of autoconf
> Sascha> 2.13, because our volunteers prefer the rapid compile & test
> Sascha> cycles of autoconf 2.13.  Unscientific tests show a 5 to 6
> Sascha> times speed decrease in 2.5x

Are these unscientific tests available somewhere, e.g., in an email
archive?  Or can you point to a .tar.gz file somewhere where autoconf
2.13 is much faster than autoconf 2.5x?  Either bit of info would help
us track down the problem.

Without knowing the symptoms, I suspect that Autoconf 2.54 is
overenthusiastically using m4, and that it's an m4 combinatorial
explosion somewhere; but without test cases it's hard to be sure.

Thanks.