[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Problems with advice...
From: |
David Kastrup |
Subject: |
Re: Problems with advice... |
Date: |
Thu, 17 Mar 2005 11:41:42 +0100 |
User-agent: |
Gnus/5.11 (Gnus v5.11) Emacs/22.0.50 (gnu/linux) |
David Kastrup <address@hidden> writes:
> I have the problem that preactivated advice no longer gets
> preactivated. However, the same problem occurs now with a current
> compilation of Emacs-21.3! So I suspect that it might be
> compiler-related (gcc-4.0). It worked at one point of time in the
> past.
Update on that: seemingly related to loading a completely unrelated
package (that does a "require" on the file containing the function to
be advised), so not a compiler error after all. I have to debug this
further.
In the mean time, the following finding still looks bad:
> Whatever. In the search for the problem I have also come across the
> following code snippet in advice.el:
>
> ((and orig-subr-p
> orig-interactive-p
> (not interactive-form)
> (not advised-interactive-form))
> ;; Check whether we were called interactively
> ;; in order to do proper prompting:
> `(if (called-interactively-p)
> (call-interactively ',origname)
> ,(ad-make-mapped-call orig-arglist
> advised-arglist
> origname)))
> ;; And now for normal functions and non-interactive subrs
> ;; (or subrs whose interactive behavior was advised):
> (t (ad-make-mapped-call
> advised-arglist orig-arglist origname)))))
>
> This looks so wrong that I want somebody with more of a clue to take a
> look at it: ad-make-mapped-call is called in two branches of a cond,
> and the order of its first two arguments is interchanged in those two
> calls!
--
David Kastrup, Kriemhildstr. 15, 44793 Bochum