This is the mail archive of the 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: [PATCH] Fix wrong common symbol size in LTO symtabs.

On 10/11/2010 23:02, Richard Guenther wrote:
> On Thu, Nov 11, 2010 at 12:20 AM, Dave Korn <> wrote:

>>  But it is signed, so it fails at 2^31 with 32-bit HWINT, and really 2 gig
>> arrays aren't impossible, we have to cope with them.
> We don't - we have implementation defined limits that rule this out
> (we run into this appearant problem elsewhere).

  Oh!  I didn't know that, and whoever wrote this code to assemble a 64-bit
int out of two (as-it-turns-out-not-necessarily) 32-bit ints presumably didn't
want it to be a limitation either.  The 2^31 limit will be a problem very very
soon if it isn't already, you know what those academic number-crunching
fortran types are like!  So I'd rather not add to it here if I can easily
avoid doing so, it's only leaving work for someone else to tidy up one day in
the future.

  I take it that passing TREE_TYPE(t) into int_size_in_bytes (rather than the
decl "t" itself) would be correct?


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