On 01/02/2013 10:04 AM, Joseph S. Myers wrote:
On Wed, 2 Jan 2013, Andreas Jaeger wrote:
The baseline is needed for adding new code so that we all agree what might
need special handling and what not.
Well - there are plenty of cases that don't have fallbacks for older or
non-GCC compilers, even when they would be easy. E.g. complex.h only
defines the C11 CMPLX macros for GCC >= 4.7, though for older GCC you
could do an approximation (not usable in static initializers) with a
temporary variable and __real__ and __imag__ assignments. Or uchar.h not
defining char16_t and char32_t for GCC versions before 4.4 (when the
built-in __CHAR16_TYPE__ and __CHAR32_TYPE__ macros were added), or for
non-GCC, although the types are required by C11 to be the same as
uint_least16_t and uint_least32_t, so in fact glibc always has the
information to define them correctly.
Agreement does not automatically fix all of our problems, but it
does guide us in determining what is or is not a valid bug report.
I say we agree that 2.95.3 is the oldest compiler we will support
for compiling userspace applications using glibc headers.
Once we agree to this it would be valid to submit conformance
bugs for some of the above issues.