autotools transition report: State of the disUnion

[Benjamin Kosnik]

>As a side effect, I've found a couple of bugs and am fixing them en passant.
>For example, if --enable-clocale was not given, we were reporting "yes" as
>the choice, and going with generic without ever trying the OS-based probe.
>In the real world, this doesn't seem to have been a problem, since distro
>vendors were specifying a model.

Ouch! Thanks for fixing this.

>Also, option handling is a lot easier.  Picking the concept checks as an
>example, before:
>
>    AC_DEFUN(GLIBCXX_ENABLE_CONCEPT_CHECKS, [dnl
>    define([GLIBCXX_ENABLE_CONCEPT_CHECKS_DEFAULT], ifelse($1, yes, yes, no))dnl
>    AC_ARG_ENABLE(concept-checks,
>    changequote(<<, >>)dnl
>    <<  --enable-concept-checks use Boost-derived template checks [default=>>GLIBCXX_ENABLE_CONCEPT_CHECKS_DEFAULT],
>    changequote([, ])dnl
>    [case "$enableval" in
>     yes) enable_concept_checks=yes ;;
>     no)  enable_concept_checks=no ;;
>     *)   AC_MSG_ERROR([Unknown argument to enable/disable concept checks]) ;;
>     esac],
>    enable_concept_checks=GLIBCXX_ENABLE_CONCEPT_CHECKS_DEFAULT)dnl
>    dnl Option parsed, now set things appropriately
>    if test x"$enable_concept_checks" = xyes; then
>      AC_DEFINE(_GLIBCXX_CONCEPT_CHECKS)
>    fi
>    ])
>
>After:
>
>    AC_DEFUN(GLIBCXX_ENABLE_CONCEPT_CHECKS, [
>      GLIBCXX_ENABLE(concept-checks,$1,[use Boost-derived template checks])
>      if test $enable_concept_checks = yes; then
>        AC_DEFINE(_GLIBCXX_CONCEPT_CHECKS)
>      fi
>    ])

Nice. I can actually understand this now...

>Even the --help output is sane now.

Whee. Does this mean that it's aligned, without having to hand-space the
acinclude.m4 text? That was really funky.

-benjamin