Re: Multiline arguments to AC_CHECK_HEADERS and AC_CHECK_FUNCS

[Paul Eggert]

Paolo Bonzini <address@hidden> writes:

>> Thanks, that looks better, except it doesn't have any
>> changes to the documentation or to NEWS.  Can you please
>> write those as well?
>
> Why should matters be complicated this way?  Stepan showed it does not
> work with autoheader.  How many packages are there that use this kind
> of quoted argument, and not use autoheader?  Can't we instead *warn*
> on the problem and let the user fix their code?

I'm not opposed to having Autoconf warn about the problem if it
detects it, but that would be more work, no?  I would think that
important thing is to document the revised behavior -- otherwise the
users won't know how to fix their code, because they won't know what
the problem is.