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]

Re: alpha block clear enhancements


    I'm clearing MEM_ALIAS_SET when it is known that the access
    is going to overlap something else.  I don't think that is
    handled elsewhere.

OK, I wasn't sure that was the case.  The AND case I believe *is* handled
elsewhere, though.

    Because unsigned types are more trouble than they are worth.
    IMO, you should actually _need_ that last bit of representation
    before you use them.

Sometimes.  Another advantage of using them is that it makes things
more "type safe" in that you are more likely to get an error message if
you put the wrong kind of thing in the wrong place.  Moreover, it's more
efficient for division.  Finally, from a clarify point of view, it tells
the reader that the value is always non-negative.

Indeed, there is a methodology that says you should always use datatypes
that match as closely as possible the actual range of the data.  This is seen
more commonly in strongly-typed languages, but can be good practice for C
as well.


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