[Bug 211152] benchmarks/iozone: Build fails on typedef redefinition with different types ('long long' vs '__off64_t' (aka 'long'))

[bugzilla-noreply at freebsd.org]

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=211152

--- Comment #42 from Mark Millard <markmi at dsl-only.net> ---
Notes from Don Capps (the upstream owner)



Most recent note:

        Revision 3.446 will appear on the Iozone web site within 24 hours.  



Slightly older material:

        The changes for :
        * FreeBSD handling of definition of __off_t 
        * Definition CACHE_LINE_SIZE changed to MY_CACHE_LINE_SIZE
        Have been merged into the official Iozone source tree and will
appear in 3.445 and later releases.

        It is unclear how you wish to handle the printf and scanf issues
that are related to long longs.  It would be not be acceptable to change the
format so that only more modern compilers would work.  Many of the systems
that compile and run Iozone span 2 decades of compilers.  Breaking backwards
compatibility would not be permitted.

        If you have a way to handle this without breaking previous builds,
please send the diffs, and we'll put this in the queue of pending merges. I
suspect that we may be doing a bunch of casting, so that we can maintain
full backwards compatibility.  



I'll note that 2016-20==1996 so anything post-dating C89/C90 is unacceptable
for upstream. And anything not commonly implemented a few years after C89/C90
would also likely be unacceptable.

-- 
You are receiving this mail because:
You are the assignee for the bug.