Re: Can't call AC_PROG_CXX conditionally

[Ralf Corsepius]

Am Mon, 2002-11-18 um 20.23 schrieb Akim Demaille:
>  >> Something should be done, agreed.
> 
>  Ralf> Should? You must be kidding - face it: autoconf is broken to an extend
>  Ralf> it is not applicable in real world, anymore.
> 
> Well, if that's the way you want to say it, I can't do anything.  My
> opinion is that this is indeed a serious problem, but in itself does
> not suffice to throw away Autoconf.  It's been like this since the
> very first day of Autoconf.
I don't know. As a matter of fact, with autoconf<=2.52 this seemed to
have worked

> I'd like to know what's the behavior people would expect in such a
> case. 
First of all, I would expect a diagnosis at the time autoconf or
automake is run for cases autoconf/automake can't handle.
At present time, autoconf generates configure scripts that work at
random.


>  For instance, what do you expect for
> 
>    AC_INIT
>    if false; then
>      AC_PROG_CXX
>    fi
>    AC_LANG(C++)
>    AC_CHECK_FUNCS(foo)
IMO, this is a pathological case.

A typical Makefile using CXX conditionally looks similiar to this:

AC_INIT
AC_ARG_ENABLE(cxx ...)
if enable_cxx
  AC_PROG_CXX
  AC_CHECK_<something>
  AM_CONDITIONAL( ..., yes)
else
  ....
  AM_CONDITIONAL(...,no)
fi
AM_CONDITIONAL(....)

In your particular case I would expect one of 3 possibilities:
1. Autoconf issuing an error "Can't handle this" (c.f. how automake
treats conditional AM_CONDITIONALs)

2. Autoconf to ignore all AC_CHECK_* if CXX is not defined/unset.

3. Autoconf to abort if CXX is not defined/set.

> Or even what you expect for the Makefiles.
> 
> Autoconf was designed with the fundamental fact that compiler (=
> language support) is not an option: it's a must.  Now people want
> optional language support; it has to be designed.

> It is not at all clear to me that the above example is valid.

Neither to me. But compare it to the example above, or the example I had
attached to my initial report. 

> It
> seems to me that the requirement for a compiler holds for the whole
> configure,
Hmm, I thought it the scope of the compile was defined by using
AC_LANG_PUSH/POP, this would mean if AC_PROG_CC and AC_PROG_CXX are used
consecutively it should apply an implicit PUSH/POP rsp. autoconf should
refuse to accept this.

Ralf