Re: Flex vs. POSIX 1003.2-1992 repeat operator {} precedence

[Hans-Bernhard Broeker]

On Mon, 29 Apr 2002, Hans Aberg wrote:

> My C/C++ compiler has an "ANSI/ISO strict" button; then one can click other
> options to depart as one wish from that. Flex might work similarly with
> respect to POSIX.

And the environment variable POSIXLY_CORRECT is kind of an established
practice among GNU utilities for doing this on a rather global scale.
People who for whatever reasons are certain they want a 100%
POSIX-conforming system can set that and expect all GNU tools to strive
valiantly to adhere to every letter of the standard --- may it be as
braindead as can be.

> > it's worth pointing out that "flex -l" is an
> >attempt to confirm to AT&T lex, which is a slightly different beast
> >then posix-mandated lex. (The posix-mandated interpretation should be
> >implied by "flex -l", however.)

Maybe the proposed POSIXized flex should fly under a different flag then.
"flex --posix" doesn't sound like a bad idea in that context. It rings
similar enough to gcc -ansi to be remembered easily.  Those SGI guys could
use that as their /usr/bin/lex scriptlet.

I don't think the additional testing burden to keep this flag
well-maintained would be important.  Flex already has enough distinct
modes of operation hiding in all the possible combinations of -C flags
that another factor of two for --posix doesn't really matter, does it?

Or maybe we should just pester the current copyright holder of AT&T lex
(Caldera --- if they haven't been sold out yet another time...) to release
lex to OpenSource, so we can drop flex -l mode for good and never look
back.  They're not entirely opposed to the idea, it seems.

-- 
Hans-Bernhard Broeker (address@hidden)
Even if all the snow were burnt, ashes would remain.