X-VM-v5-Data: ([nil nil nil nil nil nil nil t nil]
	[nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil])
Date:         Fri, 6 Jul 90 13:51:00 PDT
Reply-To: LaTeX-L Mailing list <LATEX-L@DHDURZ1.uucp>
From: Don Hosek <DHOSEK@HMCVAX.bitnet>
Subject:      Re: Permanence of .aux files
To: Rainer Schoepf <BK4@DHDURZ1.uucp>
Status: R
X-Status: 
X-Keywords:                  
X-UID: 169

>Barbara Beeton just raised the well-known problem of a
>failing run of LaTeX destroying the good .aux file from a
>previous run.  Many of us resort to shell scripts that
>preserve the old one, and then if the LaTeX run fails,
>replace the bad output one with the old one, so an edit can
>fix the problem, and one can try again.

Well, on VMS, we don't have that problem since the OS keeps old
versions for us. But as Malcolm might put it, "what's an OS?"

I think that any circumstances where a second run of LaTeX bombs
obliterating the old good aux file, either (1) the aux file wasn't
so good after all (cf, the over-long \caption problem--this has
caused many headaches for VMS users who don't realize that the
error they're getting _this_ run is due to something that happened
_last_ run) (2) the aux file information has changed sufficiently
that one more run is necessary anyway or (3) the aux file info
isn't used.

Can anybody come up with an example where one _really_ wants the
aux file from LaTeX run 1 when LaTeX run 2 crashes?

-dh