Re: [PATCH] libstdc++/8399

[Loren James Rittle <rittle at latour dot rsch dot comm dot mot dot com>]

In article <3DD2E9D6.1080101@unitus.it>,
Paolo Carlini<pcarlini@unitus.it> writes:

> However, there is a risk that systems without isatty will
> experience a performance regression for having 8399 fixed.

> Ok?

Fine by me.  For the record, I applied your patch and rebuilt
libstdc++-v3 in addition to a quick visual review.

> 2002-11-13  Paolo Carlini  <pcarlini@unitus.it>
>             Loren J. Rittle  <ljrittle@acm.org>

Thanks for adding my name after you did the heavy lifting.

This would be my plan: Install on mainline; queue for 3.2.2 since this
is a fix for a QoI regression (if we declare no interactive support)
or worse (if we declare ttys to be interactive, as expected).

Speaking of declaring.  Phil, should we document what we consider to
be an interactive stream per section 1.9 clause 11 point 3?
docs/html/17_intro/howto.html reminds us that section 1.3.5 requires
that `each implementation shall document' implementation-defined
behavior and we do document many such issues.  Perhaps saying that
"For the purposes of section 1.9 clause 11 point 3, isatty() implies
interactive stream support by the library" would be a nice start. ;-)

Should it go in docs/html/1_general/[...] or docs/html/27_io/howto.html?

Regards,
Loren