This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Update CONFORMANCE
- From: "Joseph S. Myers" <joseph at codesourcery dot com>
- To: libc-alpha at sourceware dot org
- Date: Mon, 24 Aug 2009 20:20:40 +0000 (UTC)
- Subject: Update CONFORMANCE
This patch updates CONFORMANCE to reflect the recent addition of
math_errhandling, together with some other updates extracted from my patch
for bug 10110 (and logically independent of that patch).
2009-08-24 Joseph Myers <joseph@codesourcery.com>
* CONFORMANCE: Update.
diff --git a/CONFORMANCE b/CONFORMANCE
index 6872308..8275aba 100644
--- a/CONFORMANCE
+++ b/CONFORMANCE
@@ -125,7 +125,9 @@ C99 added %a as another scanf format specifier for floating point
values. This conflicts with the glibc extension where %as, %a[ and
%aS mean to allocate the string for the data read. A strictly
conforming C99 program using %as, %a[ or %aS in a scanf format string
-will misbehave under glibc.
+will misbehave under glibc if it does not include <stdio.h> and
+instead declares scanf itself; if it gets the declaration of scanf
+from <stdio.h>, it will use a C99-conforming version.
Compiler limitations
@@ -144,29 +146,21 @@ GCC doesn't support the optional imaginary types. Nor does it
understand the keyword _Complex before GCC 3.0. This has the
corresponding impact on the relevant headers.
-glibc's use of extern inline conflicts with C99: in C99, extern inline
-means that an external definition is generated as well as possibly an
-inline definition, but in GCC it means that no external definition is
-generated. When GCC's C99 mode implements C99 inline semantics, this
-will break the uses of extern inline in glibc's headers. (Actually,
-glibc uses `extern __inline', which is beyond the scope of the
-standard, but it would clearly be very confusing for `__inline' and
-plain `inline' to have different meanings in C99 mode.)
-
glibc's <tgmath.h> implementation is arcane but thought to work
correctly; a clean and comprehensible version requires compiler
builtins.
For most of the headers required of freestanding implementations,
glibc relies on GCC to provide correct versions. (At present, glibc
-provides <stdint.h>, and GCC doesn't.)
-
-Implementing MATH_ERRNO, MATH_ERREXCEPT and math_errhandling in
-<math.h> needs compiler support: see
-
-http://sources.redhat.com/ml/libc-hacker/2000-06/msg00008.html
-http://sources.redhat.com/ml/libc-hacker/2000-06/msg00014.html
-http://sources.redhat.com/ml/libc-hacker/2000-06/msg00015.html
+provides <stdint.h>, and GCC doesn't before version 4.5.)
+
+The definition of math_errhandling conforms so long as no translation
+unit using math_errhandling is compiled with -fno-math-errno,
+-fno-trapping-math or options such as -ffast-math that imply these
+options. math_errhandling is only conditionally defined depending on
+__FAST_MATH__; the compiler does not provide the information needed
+for more exact definitions based on settings of -fno-math-errno and
+-fno-trapping-math, possibly for only some source files in a program.
Issues with headers
--
Joseph S. Myers
joseph@codesourcery.com