This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: PATCH: Check LONG_MAX > INT_MAX for __WORDSIZE


On 4/13/2012 12:04 AM, Roland McGrath wrote:
>> I can change x32 psABI to define  __ILP32__ if it helps.
> I certainly agree with Paul that this is far cleaner, even if the
> limits.h trick were technically acceptable (which it's not, for the
> reasons Paul gave).  I was going to ask if there were really x86_64
> systems with GCC < 3.3 where libc actually works.  But if you make x32
> compilers predefine __ILP32__ and test that then we don't need to figure
> out the answer.
>
> I think it would be best to be consistent in the sense of the tests we
> use.  So once you've changed the compiler, please make all the tests in
> libc be on __ILP32__ rather than !__LP64__.

I don't think HJ would likely touch this, but note that there are a few
uses of __LP64__ in ports for the tile architecture, which we need.  I like
the idea of __ILP32__ in any case (I've forwarded it on to our gcc folks)
but currently we only have __LP64__.

-- 
Chris Metcalf, Tilera Corp.
http://www.tilera.com


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