This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [RFC] Canonical Standards Names
On Wed, 28 Jun 2017, Rical Jasan wrote:
> On 06/21/2017 04:03 AM, Joseph Myers wrote:
> > We need standard needs for everything feature test macros distinguish.
>
> Would a good first step be syncing creature.texi to features.h? The
> manual doesn't document 201112L or 200809L for _POSIX_C_SOURCE, 600 or
> 700 for _XOPEN_SOURCE, or _ATFILE_SOURCE, _FORTIFY_SOURCE, or
> _ISOC11_SOURCE. Should features.h contain a comment for developers to
> update creature.texi if any feature test macros change?
Yes, I think all the supported feature test macros should be documented.
(I don't think use of _ISOC99_SOURCE or _ISOC11_SOURCE should be
*encouraged*; normally people should get those features via compiling with
appropriate -std options to define the right __STDC_VERSION__ value; GCC
>= 5 defaults to -std=gnu11 and so makes the C11 features visible by
default, while _DEFAULT_SOURCE implies _POSIX_C_SOURCE=200809L which
implies C99 features. But in a year or two I expect we'll have
_ISOC2X_SOURCE to enable features from the following C major revision
while it's in development.)
> Then, a set of standards could be derived from all the possible feature
> test macros. Would we then document with each macro (and any specific
> values) in creature.texi the corresponding name used in @standards?
Yes, I think creature.texi should, when @standards is rendered visibly in
the manual, say what names are used in the rendering of @standards.
--
Joseph S. Myers
joseph@codesourcery.com