Re: 2.8.0 and newly strict checking

From: Murray S. Kucherawy <msk_at_blackops.org>
Date: Tue, 5 Mar 2013 12:13:02 -0800 (PST)

On Mon, 4 Mar 2013, Doug Barton wrote:
> I agree that seems to be what should happen. What actually _did_ happen is
> that the tests for the directories passed, but when it got to the test for
> the file itself it failed on "does the group that this file has permissions
> for have members other than myself?" test; in spite of the fact that the file
> is mode 0600.
>
> Thus, I rather strongly suspect that there is a bug somewhere, but I was not
> able to unravel where.

If there were no group permissions, then the "if" testing them should
never have passed, so I'm stumped. That's basic bitwise comparison with
no special magic of any kind.

Regardless, if I attempt to reproduce this, I'd be guessing at your
environment and might miss a key detail. If you're able to restore it to
the failing state and then execute the filter in test mode with the
attached patch, it should tell you (via syslog) what part of the test is
failing and why. You may need to replace LOG_DEBUG with something like
LOG_NOTICE if you're not logging at debug level already.

-MSK


Received on Tue Mar 05 2013 - 20:13:23 PST

This archive was generated by hypermail 2.3.0 : Tue Mar 05 2013 - 20:18:01 PST