[PATCH] Fix fold-const breakage (take 2)
Richard Kenner
kenner@vlsi1.ultra.nyu.edu
Fri Feb 23 03:56:00 GMT 2001
Right, ok to commit the updated patch?
Yes.
I have not changed the TYPE_PRECISION test into compare_tree_int,
since if for some reason op1 would be negative, the test would succeed
although it should not.
That's a real reason to change it to using compare_tree_int *and*
tree_int_cst_sgn, so the test is clear. I'd like to phase out
explicit references to TREE_INT_CST_{LOW,HIGH}, since they tend to be
wrong in subtle ways sometimes and we may want to change constant
representation at some point.
But this can be done as a later step.
More information about the Gcc-patches
mailing list