|
From: Nicholas N. <n.n...@gm...> - 2009-01-15 00:14:54
|
On Wed, Jan 14, 2009 at 7:17 PM, Julian Seward <js...@ac...> wrote: > >> Surely Off64T should be signed, and thus a Long? > > which seems to imply that Off64T should indeed be signed. So I agree. > > I wonder if anything will break if it is changed to being signed. > There are very few places where it's used, so hopefully nothing. In the aspace manager all the functions like VG_(am_notify_client_mmap) use Off64T for an argument, but the 'offset' field of the NSegment type is a ULong. But the regression tests still pass (on AMD64-linux) with the change so I'm inclined to commit it. Or I could try changing those ULongs to Off64T. N |