Paul Eggert wrote:
> * doc/posix-headers/limits.texi (limits.h): Document this.
> * lib/limits.in.h (SSIZE_MAX): Define if not already defined.
When I add a unit test for this, I see that the assertion
static_assert (TYPE_MAXIMUM (ssize_t) == SSIZE_MAX);
fails on 64-bit MSVC 14.
on SSIZE_MAX
* lib/areadlink-with-size.c, lib/areadlinkat-with-size.c:
- * lib/careadlinkat.c:
+ * lib/careadlinkat.c, lib/getdelim.c:
(SSIZE_MAX): Do not define.
* modules/areadlink-with-size, modules/areadlinkat-with-size:
- * modules/careadlinkat
| 1 +
4 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/ChangeLog b/ChangeLog
index c3a0b34130..b9c3e49552 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -2,8 +2,10 @@
limits-h: other modules can rely on SSIZE_MAX
* lib/areadlink-with-size.c, lib/areadlinkat-with-size.c
@@
2023-05-20 Paul Eggert
limits-h: other modules can rely on SSIZE_MAX
- * lib/areadlink-with-size.c (SSIZE_MAX):
- Do not define.
- * modules/areadlink-with-size (Depends-on):
- Add limits-h.
+ * lib/areadlink-with-size.c, lib/areadlinkat-with-size.c
* lib/areadlink-with-size.c (SSIZE_MAX):
Do not define.
* modules/areadlink-with-size (Depends-on):
Add limits-h.
---
ChangeLog | 6 ++
lib/areadlink-with-size.c | 4
modules/areadlink-with-size | 1 +
3 files changed, 7 insertions(+), 4 deletions(-)
diff --git a
rely on SSIZE_MAX
* lib/areadlink-with-size.c, lib/areadlinkat-with-size.c:
* lib/careadlinkat.c, lib/getdelim.c, lib/getndelim2.c:
+ * lib/regex_internal.h:
(SSIZE_MAX): Do not define.
* modules/areadlink-with-size, modules/areadlinkat-with-size
* doc/posix-headers/limits.texi (limits.h): Document this.
* lib/limits.in.h (SSIZE_MAX): Define if not already defined.
* m4/limits-h.m4 (gl_LIMITS_H): Also generate limits.h
if does not define SSIZE_MAX.
---
ChangeLog | 6 ++
doc/posix-headers/limits.texi | 6
rely on SSIZE_MAX
* lib/areadlink-with-size.c, lib/areadlinkat-with-size.c:
- * lib/careadlinkat.c, lib/getdelim.c:
+ * lib/careadlinkat.c, lib/getdelim.c, lib/getndelim2.c:
(SSIZE_MAX): Do not define.
* modules/areadlink-with-size, modules/areadlinkat-with-size
On 07/13/2016 06:12 AM, Eric Blake wrote:
>> +The macro @code{SSIZE_MAX} has the wrong type,
>> +albeit with the correct value:
>> +glibc 2.24.
>> +@item
>
> Only 32-bit glibc is buggy; 64-bit glibc is okay.
Worse, it's platform-dependent. On 32-bit s390,
doc/posix-headers/limits.texi
> @@ -23,6 +23,10 @@ AIX 5.1, HP-UX 11, IRIX 6.5, OSF/1 5.1 with gcc.
> The macros @code{WORD_BIT}, @code{LONG_BIT} are not defined on some
> platforms:
> glibc 2.11 without @code{-D_GNU_SOURCE}, Cygwin, mingw, MSVC 9.
> @item
> +The macro @code{SS
l Eggert
Date: Wed, 13 Jul 2016 11:42:40 +0200
Subject: [PATCH] doc: mention glibc SSIZE_MAX buglet
* doc/posix-headers/limits.texi (limits.h): Document the bug.
---
ChangeLog | 5 +
doc/posix-headers/limits.texi | 4
2 files changed, 9 insertions(+)
diff --git a/Cha
On 07/12/2016 11:06 AM, Eric Blake wrote:
> I just tripped up on this glibc bug:
>
> https://sourceware.org/bugzilla/show_bug.cgi?id=13575
>
> where glibc incorrectly defines SSIZE_MAX as a long even though ssize_t
> is an int on a 32-bit platform.
>
> I'm
I just tripped up on this glibc bug:
https://sourceware.org/bugzilla/show_bug.cgi?id=13575
where glibc incorrectly defines SSIZE_MAX as a long even though ssize_t
is an int on a 32-bit platform.
I'm planning on tweaking gnulib's replacement to work around
it in the meantime.
--
to Derek's full-header-path.m4 macro
and because I spent a day on that last week-end.
> Is that also true for SSIZE_MAX?
SSIZE_MAX is expected to be defined in , and
#include
is portable nowadays. So all we need is a
#define SSIZE_MAX ...
in config.h on those systems that l
Ben Pfaff <[EMAIL PROTECTED]> writes:
> If ssize_t doesn't exist then I wouldn't expect SSIZE_MAX to
> exist either. The current ssize_t module doesn't define
> SSIZE_MAX, but, in my opinion, it should.
That sounds reasonable, yes. Sorry, I was confused by your earlier
messages.
> headers on different systems. Is that also true for SSIZE_MAX?
>
> I don't offhand recall SSIZE_MAX being defined anywhere other than its
> standard location . If so, then we wouldn't need a
> ssize_max.h header, and we wouldn't need to overload size_max.h; we
>
Ben Pfaff <[EMAIL PROTECTED]> writes:
> I really meant ssize_max.h, analogous to size_max.h. Or they
> could be combined into one header file.
I think size_max.h is there only because SIZE_MAX is in different
headers on different systems. Is that also true for SSIZE_MAX?
I d
Ben Pfaff <[EMAIL PROTECTED]> writes:
> Gnulib includes a module for ssize_t, but it doesn't ensure that
> SSIZE_MAX is available. It seems that an ssize_t.h header file
> with the following contents would be sufficient:
I really meant ssize_max.h, analogous to size_max.
Gnulib includes a module for ssize_t, but it doesn't ensure that
SSIZE_MAX is available. It seems that an ssize_t.h header file
with the following contents would be sufficient:
#include
#ifndef SSIZE_MAX
# define SSIZE_MAX ((ssize_t) (SIZE_MAX / 2))
#endif
A few of the s
19 matches
Mail list logo
