>>>>> "Alexandre" == Alexandre Oliva <[EMAIL PROTECTED]> writes:
Alexandre> On Jul 17, 2000, Akim Demaille <[EMAIL PROTECTED]> wrote:
>> Maybe AC_DEFINE(`echo FOO`) used to be understood by autoheader,
>> but it was a misfeature. In the general case, this cannot be done,
>> hence it must never work :)
Alexandre> If autoheader doesn't understand it, it's fine: it's
Alexandre> perfectly acceptable, to me, if autoheader deals only with
Alexandre> literals. But people must not be forbidden from computing
Alexandre> their own header files, based on whatever acconfig.h
Alexandre> entries or additional headers.h.in they've generated by
Alexandre> hand.
Right. I'm not saying we will forbid these computations, but still
they are a side effect of the technology which was used. Because they
are a side effect, I'm all for undocumenting this.
Autoconf is sufficiently juggling with the regular issues so that we
don't need additional constraints.
Up to now, I think we have fairly respectful of the needs of Autoconf
users, when something is not possible, we do look for new solutions,
and 2.50 is full of such things.
But there is no featurism either: we solve actual problems.
I'm not saying this is not an actual problem, I'm saying I'd like to
know more to see how Autoconf would best answer the need, or, what is
already in Autoconf which would do the job.
