Do you have patches or solutions to these problems? Obviously, it is nobody's intent to say "only with gcc". It would be quite silly to think that's the position. But you raise a good point: we pretty much depend on 'long long' being available nowadays. Do we have a list of compilers that don't have it? How can we fix them?
Cheers, -g On Thu, May 16, 2002 at 01:06:37PM +0200, Martin Kraemer wrote: > On Wed, May 15, 2002 at 11:00:42AM +0200, Sander Striker wrote: > > > > Pardon? I would have thought that 2.0 is a leap forward in the portability > > department. Care to explain? > > > > Fpr example, did you ever try and build httpd-2.0 on a machine > where the compiler does NOT support "long long" and "unsigned > long long"? > Apr will happily use "long double" for the int64 type, and > "unsigned long double" (care to explain what the hell that is?!?) > as the uint64 type. And hardwired constants (like 0x01B21DD213814000LL > in srclib/apr/misc/unix/getuuid.c) make it impossible to provide a > workaround. > > The requirement "only guaranteed to compile with gcc" _is_ a portability > problem, and 1.3 _did_ try to provide an alternative path for older > systems. After all, the world does not only consist of Win32 and Linux. > > Martin > -- > <[EMAIL PROTECTED]> | Fujitsu Siemens > Fon: +49-89-636-46021, FAX: +49-89-636-47655 | 81730 Munich, Germany -- Greg Stein, http://www.lyra.org/
