Re: m4_wrap behavior

[Stepan Kasal]

Hello,

On Tue, Jun 13, 2006 at 09:56:27PM -0600, Eric Blake wrote:
> > It seems to be pity that m4 standrdized for FIFO.
> 
> Well, how about a patch to m4sugar that forces m4_wrap to always be LIFO,
> and documents it as such, regardless of the underlying m4wrap

even though I expressed my disappointment that FIFO was standardized,
I do not think it is worth it to add this code.

m4sugar was always meant as a proposal to ``improve'' or ``enhance''
m4, if I get it right.  I do not think we should try push the world
against the standard in this case, it probably is not that important.

So I'm dismissing one possible motivation: to improve the world by
fixing the broken standard.

Another motivation would be purly utilitary: we might need a sane
wrapper around m4 2.0, so that Autoconf works.

But such a wrapper is not needed for Autoconf proper: the pattern
        m4_diversion_push() m4_wrap(m4_diversion_pop())
is actually wrong; it doesn't check for anything--if we check that
the stack has emptied at the end, it's enough.  (Remember that each
pop checks that the stack is not empty.)

What reminds?  Perhaps some third party macros depend on LIFO
m4_wrap.  (Automake doesn't contain "m4_wrap", BTW.)
Checking for that would require more work.

Unless a reason is found, I would prefer to document that the order
of m4_wrap execution is undefined.
(I apologize if that was not clear from my previous mail.)

Have a nice day,
        Stepan Kasal