This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: Generate string.h and time.h if nonexistent; kill POSIX

On Sat, May 12, 2001 at 11:07:31AM +0100, Joseph S. Myers wrote:
> On Fri, 11 May 2001, Zack Weinberg wrote:
> > Supporting pre-C89 targets does take more effort, and I'm definitely
> What is a "pre-C89 target"?  The vendor compiler might be pre-C89, but if 
> we're building a cross-compiler then that's irrelevant, so the only 
> support for such targets is surely that for fixincludes to fix their 
> headers and such things as TARGET_MEM_FUNCTIONS?

Target without all the C89-specified functions in its libc.  Causes
us to have to maintain caution through all the code compiled for the
target, including the compiler itself since host==target most of the
time.  Look at the portability contortions in libf2c, or libgcc.

On the other hand, libstdc++3 and libjava may just punt.  I can make a
much better case for not supporting non-C on pre-C89 targets, than I
can for not supporting C...

zw        This is, no doubt, the rational strategy; quite possibly the
          only one that will work.  But it ignores the exigiencies of
          the tenure system and is therefore impractical.
          	-- Jerry Fodor, _The Mind Doesn't Work That Way_

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]