--- Begin Message ---
|
Subject: |
Emacs 29: Edebug fails to instrument a parameter whose name begins with _ |
|
Date: |
Sat, 12 Nov 2022 09:35:58 +0000 |
Hello, Emacs.
In Emacs 29 (not started with -Q, but...),
I instrumented for edebug a function which looked like:
(defun c-trim-found-types (beg end _old-len) ....)
, the compilation being with lexical-binding: t.
During the edebug session, I attempted
e _old-len RET.
Instead of giving me the value of _old-len (which was 3) it gave the
error message
Error: Symbol's value as variable is void: _old-len
.. This is a bug.
Just because a function doesn't use a particular argument (here
_old-len) doesn't mean the person debugging it isn't interested in its
value. In this particular case, it was extremely interesting, because
beg and end were unequal, and _old-len was 3.
In the end, I found out the info with the d command (backtrace), but I
shouldn't have to.
--
Alan Mackenzie (Nuremberg, Germany).
--- End Message ---
--- Begin Message ---
|
Subject: |
Re: bug#59213: Emacs 29: Edebug fails to instrument a parameter whose name begins with _ |
|
Date: |
Sat, 11 Feb 2023 11:17:24 +0000 |
Hello, Stefan.
On Fri, Feb 10, 2023 at 17:05:32 -0500, Stefan Monnier wrote:
[ .... ]
> > What do you think?
> I think that's good enough for `emacs-29`, yes.
I've committed the fix (with modifications) to master, as requested by
Eli, and I'm closing the bug with this post.
> I don't like the use of a dynbound variable to control this, but it's
> not clear how to do better. One thing that occurred to me right now is
> that we could mark the Edebug closures themselves, e.g. by replacing
> (function (lambda () ,@forms))
> with
> (function (lambda () :closure-dont-trim-context ,@forms))
> and then have `cconv-make-interpreted-closure` look for this
> tell-tale sign.
Interesting. Isn't there a good chance this would foul up programs which
analyse the structure of a closure? (A bit like
testcover-analyze-coverage analyses the calling structure of
edebug-enter).
> A few minor comments about your patch below.
[ .... ]
> Hmm... any reason why we can't just replace
> (if (null lexvars)
> with
> (if (or cconv-dont-trim-unused-variables (null lexvars))
> and be done with it?
No, no reason at all. I think I was concerned to preserve the macro
expansion, but seeing as how that wouldn't even be in the uninstrumented
form, this is clearly not a valid concern. So I modified the patch to do
that before committing it.
Thanks!
> Stefan
--
Alan Mackenzie (Nuremberg, Germany).
--- End Message ---