[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Sat, 13 Oct 2001 22:36:07 +0200
On Sat, Oct 13, 2001 at 04:10:53PM +0100, Gary V. Vaughan wrote:
> so presumably that code has already stripped unquoted leading spaces by
> the time the builtin receives its argv?
No, this is different. Think of eval([ 1 + 1]).
> I think we are in agreement
> here, that normal m4 whitespace rules should apply to numeric
> arguments. I am slightly worried about old code that relies on this
> "feature" though... perhaps Rene wrote it this way for compatibility
> with vendor M4?
This is dead broken, extremely bad, puke! Never ever put features for
novices which can make it harder for real hard core programmers. M4
programmers are aware of the rules, and having an exception is a means
to have unwanted phenomenons remain hidden.
I really don't want soft corners. But that's my HO.
> Do you have access to other implementations for compatison?
> If this is a peculiarity of GNU M4, then you have my okay. Otherwise I
> guess we need to keep it in :-( The skip_space function should be using
> the syntax table to decide what is skipped anyway!
I thought it did via IS_SPACE, but I didn't look at its meaning since I
was to remove it.