On Tue, 07 Sep 2010 13:08:37 -0700, Ian Romanick <i...@freedesktop.org> wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Luca Barbieri wrote: > > Well, the alternative is to copy&paste this into ir_lower_jumps and > > other similar passes. > > > > I added a new class instead of changing ir_visitor so that subclasses > > that need to examine all leaves will get a compile error if not > > updated when new leaves are introduced. This is debatable though, > > since you need to add the function to ir_visitor anyway. > > The problem with the default navigation is that code breaks in > mysterious ways when new node types are added. We've encountered this a > couple times with the hierarchical visitor. If the visitor base class > is pure virtual, the compiler will tell you all the places that need to > be updated when a new node type is added. The cases where we've had > this problem with the hierarchical visitor have been a bitch to debug. > The extra bit of cut-and-paste seems like a small price to pay.
The only value I've seen from the pure virtual methods is that you don't mis-type them (which is a problem I've had with the HV, since the signatures don't all match). People would still regularly cut-and-paste the wrong junk in to make the compile succeed.
pgpTKFCPeKi3O.pgp
Description: PGP signature
_______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
