Re: [ROOT] Fw: GCC 2.96

From: Dan Pop (Dan.Pop@cern.ch)
Date: Fri Oct 06 2000 - 17:00:47 MEST


On Fri, 6 Oct 2000, Fons Rademakers wrote:

> It has not so much to do with *design* as finding out how the different lib C
> implementers interpreted the standard (or the compiler implementers for that
> sake). You've to read a lot of FM's and header files. ;-)

For this particular issue, finding out how different implementors
implemented the standard was detrimental, rather than helpful, because
it misled you to believe that it is safe to rely on fpos_t being a
scalar type (long, in your case).

The standard itself doesn't say so, but it *guarantees* that fseek and
ftell use long offsets.  Everywhere, no need for peeking in header
files!  So, I can't see any *good* reason for preferring an approach
working by accident to one working by design. 

Dan



This archive was generated by hypermail 2b29 : Tue Jan 02 2001 - 11:50:34 MET