Eric Blake <[EMAIL PROTECTED]> writes:
> Is isblank portable?
Not universally, no.
> And if not, should we have a public gnulib header that provides it,
> rather than regex_internal.h?
For isblank, it's so simple that one wonders whether it's worth the
effort of gnulibizing it. But if we're go
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
I noticed that POSIX requires m4 to skip only leading and
, which is a subset of in the LC_CTYPE category (at
least, in the POSIX locale). I also noticed that regex_internal.h
provides a fallback in case isblank is not provided. Is isblank portable
I installed this:
2006-10-21 Paul Eggert <[EMAIL PROTECTED]>
* lib/stdint_.h [defined _AIX]: Don't include .
Problem reported by Perry Smith and Ville Laurikari.
--- lib/stdint_.h 11 Oct 2006 05:58:47 - 1.34
+++ lib/stdint_.h 21 Oct 2006 22:09:46 -
@@ -
Bruno Haible <[EMAIL PROTECTED]> writes:
> + /* Mingw doesn't have SSIZE_MAX. */
> + #ifndef SSIZE_MAX
> + # define SSIZE_MAX ssize_t)1 << (sizeof (ssize_t) * CHAR_BIT - 2)) - 1)
> * 2 + 1)
> + #endif
A small point: getdelim.c and other modules already define SSIZE_MAX
in a different way, a
I expect to make a stable coreutils-6.4 release very soon (maybe even
tomorrow), so if you know of remaining portability (or other) problems
and want to see them fixed in that release, report them now.
Simon Josefsson <[EMAIL PROTECTED]> wrote:
> On Sat, 2006-10-21 at 10:39 +0200, Jim Meyering wrote:
>> Paul, Simon,
>> Any objection to the changes below?
>
> Seems generally fine to me, thanks!
>
>> + short int port = ntohs (((const struct sockaddr_in *) sa)->sin_port);
>
> Shouldn't 'short i
On Sat, 2006-10-21 at 10:39 +0200, Jim Meyering wrote:
> Paul, Simon,
> Any objection to the changes below?
Seems generally fine to me, thanks!
> + short int port = ntohs (((const struct sockaddr_in *) sa)->sin_port);
Shouldn't 'short int' be either 'unsigned short int' or 'uint16_t'?
My
Paul, Simon,
Any objection to the changes below?
I noticed that getaddrinfo uses snprintf to perform a simple
short-to-string conversion. For such a simple job, snprintf
and its dependent, vasnprintf, are overkill.
Instead, how about using a new shorttostr function instead?
Why a new function?
