Re: I think binary-branch is almost ready for play time

[Bruce Korb]

"Gary V. Vaughan" wrote:
> 
> Hi Bruce,
> 
> On Monday 14 May 2001  1:31 am, Bruce Korb wrote:
> > When this invocation fires,
> > an "anonymous template" is created with that text and then processed.
> > The output is redirected to an unlinked file (no Windows here),
> > and read back in.
> 
> Eeek!  I was already worried about getting guile to compile on Windows to
> support AutoGen-5.x -- which could be worse, but at lease as the resulting
> scripts are okay on Windows.  Unfortunately it makes Cygwin users second
> class citizens, since we won't be able to develop with libtool any more.

"Wont be able to develop ... libtool *ON WINDOWS* any more."
But, on the other hand:

  [= macro-name  arg = "[= another macro =]" =]

is entirely equivalent to:

  [= (out-push-new ".x") =][= another macro =][=
     (out-pop) =][= macro-name arg = `cat .x ; rm -f .x` =]

which is the way I would prefer to see the stuff written
anyways.  It avoids the multiple parsing passes over the
template text.  It also avoids the open/unlink sequence,
but I suppose I could find an alternate implementation for
that anyways....

Alexandre wrote:
> What I need is the
> definition of test-OR-exit to refer to other macros.

Oh.  That is very different.  That is so trivial that I never thought
of it as a possible issue.  Sorry.  Oops.  Template macros are not any
different than the main template body, except that they are surrounded
by "DEFINE" and "ENDDEF" and do not actually emit anything until
they are invoked.  Just for the sake of exercise, go ahead and
insert a macro invocation inside of one of the macros I defined.
You'll see.  I use that technique in the option processing stuff,
but you would likely go blind trying to decipher those hieroglyphics. :-}

Regards,        
Bruce