Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!umich!samsung!cs.utexas.edu!asuvax!ncar!tank!sophist!goer
From: goer@sophist.uucp (Richard Goerwitz)
Newsgroups: comp.unix.xenix
Subject: fsanalyze-4.1
Message-ID: <7472@tank.uchicago.edu>
Date: 6 Feb 90 08:29:42 GMT
Sender: news@tank.uchicago.edu
Reply-To: goer@sophist.UUCP (Richard Goerwitz)
Organization: University of Chicago
Lines: 29

I was poking through fsanalyze to see if I could get it 
running under Xenix/386.  This was not a problem.  It
took basically three adjustments.  I defined SUPERBOFF
as 1024L, defined the OS as OS_ATT, and then I screwed
with the macro that checks to see if the file system has
been cleaned properly.  Here's the question:

Fsanalyze as it stands wants to read the superblock into
a filsys structure, then check one of its fields (like
I said above) to see if the file system needs to be fsck'd.
The macro (see is_ok in fsconfig.h) looks for a field with-
in the filsys structure called s_state.  This does not
exist in my Xenix superblocks.  Instead, I just poked
around until I found something that looked good:  s_clean.
I then rewrote the macro to be

       ... ((fs)->s_clean = S_CLEAN)

Was this the right decision?  I notice that John Haugh simply
did away with the whole sordid affair by defining XENIX/286
as his os, and indicating that he had a command /etc/fsstat.
I don't have this command.  So I had to take my best guess at
what would be a reasonable equivalent to s_clean under Xenix.
Does anyone have any suggestions on this score?

Should I define SUPERBOFF myself in one of the .h files?

   -Richard L. Goerwitz              goer%sophist@uchicago.bitnet
   goer@sophist.uchicago.edu         rutgers!oddjob!gide!sophist!goer