[Patch] Fix libstdc++/11378 (take2) + xsputn optmizations

[Andreas Jaeger]

Nathan Myers <ncm-nospam@cantrip.org> writes:

> Instead, our streamoff and streampos types should always be 64-bit
> types on all platforms.  (That is, "int64_t", not "long long".) 
> Furthermore, we should use the 64-bit versions of system calls 
> automatically wherever they exist.  Where they don't, we should check 
> for range errors where we convert a 64-bit offset to a 32-bit offset.

I would not use int64_t - let's use the libc types directly like
fpos64_t and off64_t - and then using fsetpos64/fseeko64.

> As a matter of staging, we might introduce the 64-bit types and the 
> range checks first, and then work on enabling use of the 64-bit system 
> calls after.  That would probably flush out helpers quicker.

Not sure whether this works.  One problem with the 64-bit types and
system calls is that they are not available everywhere, we have to be
carefull to test for them.  Using int64_t as a fallback might work...

Andreas
-- 
 Andreas Jaeger, aj@suse.de, http://www.suse.de/~aj
  SuSE Linux AG, Deutschherrnstr. 15-19, 90429 Nürnberg, Germany
   GPG fingerprint = 93A3 365E CE47 B889 DF7F  FED1 389A 563C C272 A126