Hi, > One reading is that since step 1 fails, the entire rmdir command fails at > that point without attempting step 2.
Yeah, well... I admit that for the first read one can easily interpret the standard this way. However, I can't see it saying that a failure causes the whole procedure to fail. So according to my interpretation this is "step 1; step 2" rather than "step 1 && step 2". It is actually similar to "rmdir nonexistent existent". The standard says "For each dir operand, the rmdir utility shall perform actions equivalent to the rmdir() function..." and really, it does not abort on an error, keeps on processing subsequent arguments. > The only question now is what do other implementations (such as BSD or > Solaris) do in this case? Sorry, I only have access to GNU/Linux systems so I can't help you in answering this question. :) Thanks very much, Egmont _______________________________________________ Bug-coreutils mailing list [email protected] http://lists.gnu.org/mailman/listinfo/bug-coreutils
