Re: Suggestion: AC_DEFINE -> m4_pattern_allow

[Ralf Wildenhues]

Hello Akim,

* Akim Demaille wrote on Tue, Jun 13, 2006 at 12:08:23PM CEST:
> >>> "Ralf" == Ralf Wildenhues <address@hidden> writes:
>  > * Akim Demaille wrote on Wed, May 31, 2006 at 04:26:17PM CEST:
>  >> >>> "Akim" == Akim Demaille <address@hidden> writes:
>  >> 
>  >> Some time ago I sent this message for which I had no answer.
> 
>  >> > I suggest that the AC_DEFINE family explicitly allow its $1.
> 
>  > Not true,
> 
> Hm, yes, I have not received answered: apparently I was not in CC.

Yes.  Seems we have to get used to the fact that you are not subscribed
to all lists any more...

>  > I agreed with it back then, although I do have a vague feeling that all
>  > the m4_pattern_allow make for a significant runtime overhead.
> 
> There are not that many AC_DEFINE, AC_SUBST.

This is not true in general.  I help with packages that have several
hundred AM_CONDITIONALs, and around 700 AC_SUBSTs, and several hundred
AC_DEFINEs.  Probably you're right though in that the m4_pattern_allow
is not a big overhead.

>  >> I suggest that Automake provide the same feature for AM_CONDITIONAL:
> 
>  > But you are not likely to use the first argument of AM_CONDITIONAL
>  > elsewhere in your configure.ac, are you?  (I have probably just
>  > misunderstood this part.)
> 
> I meanr that this first arguments appears in the expansion of
> AM_CONDITIONAL, so it should be un-forbidden.

Stepan and I had two different ideas:
http://lists.gnu.org/archive/html/automake/2006-06/msg00009.html
http://lists.gnu.org/archive/html/automake/2006-06/msg00015.html

Cheers,
Ralf