Official Patch #2 for Mush 7.1

[Barton E. Schaefer]

In article <1990Jul18.021543.11830 at eci386.uucp> woods at eci386.UUCP (Greg A. Woods) writes:
} In article <10491 at ogicse.ogc.edu> schaefer at ogicse.cse.ogi.edu (Barton E. Schaefer) writes:
} > This is Official Patch #2 for Mush 7.1.
} >[....]
} >     Handling of "corrupted" spool mailboxes has been improved.  The
} >     size of the spool file is now correctly recorded on new mail checks,
} >     which should reduce the frequency of corruption reports on systems
} >     that rely on DOT_LOCKing.
} 
} Unfortunately the code for DOT_LOCKing is still very closely tied to
} either BSD, or SysVr3.0+ (i.e SVID vol. 3).  It will not work on
} earlier versions of SysV, i.e. those described by SVID vol. 1.

There are two other "unfortunately" comments I want to make about that
patch before we start getting a lot of bug reports.

First, the patch claims:

    The "delete" operation in tool mode will now bring the new current
    message onto the header display if and only if it was changed by
    the deletion.

Due to a reversed logical test, this actually behaves exactly the same
way it did before.  It really will do this, honest, after Patch #3.

Second, the patch also says:

    Addresses containing UUCP !-paths ending in a domain-style @host
    suffix will now be shifted to the right in header summaries to show 
    as at least the last two hosts of the UUCP path tail.  This assumes
    that the correct parsing of host2!host3!user at host1 is the same as
    host1!host2!host3!user.

Another botched boolean check also renders this code a no-op.  This is
what I get for trying to generate a patch and a final draft of a thesis
at the same time.  (I hope my thesis is doing better.)  Fixed in #3.
-- 
Bart Schaefer						schaefer at cse.ogi.edu