[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

DocEx error trying to read section 'Table of Program Output Facilities']

barmar after originally failing to reproduce the error (as seemed to be the
consensus from most individuals) managed to find the error.  I don't
understand why an error that Symbolics and many others couldn't reproduce
would seem so deterministic: only requiring an Ivory machine and trying to
lookup any of several specific document sections.  I thought I'd share
barmar's findings with the net.  I'm unable to fix my machine because I don't
seem to have the source for this function.

Date: Mon, 25 Jun 90 10:34 CDT
From: Barry Margolin <barmar@Think.COM>
Subject: DocEx error trying to read section 'Table of Program Output Facilities'
To: Donald H. Mitchell <dmitchell@backus.trc.amoco.com>
In-Reply-To: <19900621153814.1.DON@trc.amoco.com>

Found the bug; it only occurs on Ivory machines when looking up a flavor
method.  SAGE::ARGLIST-FOR-FUNCTION-RECORD contains the following:

		    (cond (function-standin nil)
			  ((atom function) nil)
			  ((eq (car function) 'defun-in-flavor) 2)
			  ((member (car function) flavor:*fdefinable-method-types*)
			   #+3600 3))

The bug is that there's no #+imach variant in the last clause.  This is
causing the COND to return what MEMBER returns, which is a tail of
*FDEFINABLE-METHOD-TYPES*, rather than an integer or NIL as expected.