Hi Gary, On Wed, 2005-11-16 at 15:10 +0000, Gary Benson wrote: > The checks at the end of java.io.FilePermission.implies are backward. > They're supposed to be checking that fp's actions are a subset of this > object's actions, but they're actually checking that this object's > actions are a subset of fp's. The attached patch fixes.
Oops. > 2005-11-16 Gary Benson <[EMAIL PROTECTED]> > > * java/io/FilePermission.java (implies): Correct the sense > in which action checks are applied. And you created mauve tests for this. Very nice! Patch checked in. > Also, as an aside, the (simple) action checks are performed after the > (complex) pathname checks. We'd refuse permission slightly faster if > we did the action checks first. Yeah, but I assume that in general programs only try to open files that they have permission for, so we have to do all checks anyway except if they have ALL_FILES permission. But maybe that isn't true in practise. Please feel free to supply another patch to change the order. Thanks again, Mark
signature.asc
Description: This is a digitally signed message part
_______________________________________________ Classpath-patches mailing list Classpath-patches@gnu.org http://lists.gnu.org/mailman/listinfo/classpath-patches
