Re: Why yacc_EOF is used as a list_terminator for for_command?

[Chet Ramey]

On 5/12/21 10:00 AM, Peng Yu wrote:
> On Wed, May 12, 2021 at 7:18 AM Koichi Murase <myoga.murase@gmail.com> wrote:
> > 2021年5月12日(水) 21:11 Alex fxmbsw7 Ratchev <fxmbsw7@gmail.com>:
> > > oh i see
> > > so whats the conclusion
> >
> > I'm not sure if it is already concluded. Peng pointed out that it's
> > unnecessary in the syntax rule, which I agree with. Peng claims it's
> > ``incorrect'', which I don't agree with. I think it's just a kind of
> > ``overloading'' of the concept ``list_terminator'' for convenience,
> > which is reused in the definition of `for' statement. I'm not sure
> > what Peng would say about this way of thinking in the next reply if
> > any.
>
> I am not referring to anything in whether the current bash should be
> changed or not. After all, that is Chet's call. If he had
> intentionally (there is also a chance that he didn't notice it, but
> only he can answer which case it is) decided to include yacc_EOF in
> the for-statement and bash has worked with it for so long without a
> problem, I'd expect that it should continue to work properly in
> practice. There is no point to change the bash source code at this
> point of time as changing the code can only introduce chances to break
> other source code of bash.

If anyone's curious, the list_terminator production including yacc_EOF
dates from at least 1989.

(Again, if anyone has bash versions before bash-1.02, I would be very
interested in getting copies. We didn't pay as much attention to software
preservation in those days as we should have.)

> What I refer to is, if I were to rewrite bash from scratch, whether
> yacc_EOF should be included or not. I would not include it as it has
> been shown to be unnecessary. 

There are a lot of hypotheticals here.

--
``The lyf so short, the craft so long to lerne.'' - Chaucer
                 ``Ars longa, vita brevis'' - Hippocrates
Chet Ramey, UTech, CWRU    chet@case.edu    http://tiswww.cwru.edu/~chet/