On Aug 21 21:16:40, [email protected] wrote:
> On Aug 21 21:04:53, [email protected] wrote:
> > I came across some m4 problems when trying to compile sox
> > (a future version of the audio/sox port), which uses the
> > horrendous autotools to create it's ./configure script.
> > These tools in turn use m4 to define their macros.
> >
> > What should the following print?
> > (Please excuse my m4 ignorance.)
> >
> > divert(-1)dnl
> > changequote([, ])
> > define([dquote], [[$@]])
> > define([argn], [pushdef([_$0],
> > [popdef([_$0])]dquote([$]incr([$1])))_$0($@)])
>
> Could it be this?
>
> The built-ins pushdef and popdef handle macro definitions as a stack.
> However, define interacts with the stack in an undefined way. In this
> implementation, define replaces the top-most definition only. Other
> implementations may erase all definitions on the stack instead.
>
>
> > define([foo], [argn([10], $@)])
> > define([bar], [argn([9], shift($@))])
> > define([baz], [argn([8], shift(shift($@)))])
> > define([numbers], [[1], [2], [3], [4], [5], [6], [7], [8], [9], [10]])
> > divert(0)dnl
> > foo(numbers)
> > bar(numbers)
> > baz(numbers)
> >
> > According to upstream, it should be
> >
> > 10
> > 10
> > 10
> >
> > On current, it's
> >
> > 101
> > 90
> > 10
No really, should these all be 10,
or is that undefined because of the stack interaction?
Jan
> >
> > https://marc.info/?l=sox-devel&m=159803236823541&w=2
> > https://sourceforge.net/p/sox/code/ci/affc279d142f843f3f50d4718798303396ee24b4/
> >
> >
>
>
