This is the mail archive of the gcc-patches@gcc.gnu.org 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]
Other format: [Raw text]

Re: Fix powerpc64 g77


>>>>> Alan Modra writes:

>> low = INTVAL (x) << (POINTER_SIZE - GET_MODE_BITSIZE (mode));
>> high = (HOST_WIDE_INT) low >> 32;
>> low &= 0xffffffff;

Alan> Well, only with other changes to this function.  I was writing a
Alan> minimal patch.  Using the above is rather confusing to someone who
Alan> doesn't know that the CONST_DOUBLE case never occurs for integer
Alan> constants when HOST_WIDE_INT is 64 bits.

	I didn't mean *only* that.  I meant using INTVAL(x) on one branch
of the conditional.  What other changes would be necessary?

Alan> I'm happy to write patches that clean up code, if fact I'd done so
Alan> in the first patch I wrote to fix this problem, but I thought your
Alan> life as a maintainer would be easier if I separated bug fixes from
Alan> code tidying.  (And that the patch would go in quicker, with less
Alan> discussion.  Heh.)

	I'm so glad that we avoided a long discussion ;-).

	I don't care.  Check in either version -- with INTVAL(x) or
low |= high << 32 .

David


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