all messages for Emacs-related lists mirrored at yhetil.org
 help / color / mirror / code / Atom feed
From: Arthur Miller <arthur.miller@live.com>
To: Stefan Monnier <monnier@iro.umontreal.ca>
Cc: emacs-devel@gnu.org
Subject: Re: Patch: debug-instrumented predicate
Date: Mon, 04 Oct 2021 23:58:26 +0200	[thread overview]
Message-ID: <AM9PR09MB4977C69522B279DE3039EF1F96AE9@AM9PR09MB4977.eurprd09.prod.outlook.com> (raw)
In-Reply-To: <jwvtuhwimwl.fsf-monnier+emacs@gnu.org> (Stefan Monnier's message of "Mon, 04 Oct 2021 16:00:21 -0400")

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>> Concrete use-case would be to offer a user some kind of gui to instrument or remove
>> instrumentation for debug/edebug/profile/trace.
>
> W.r.t *removal* of debug/profile/trace, this could be done by offering
> a generic removal GUI for advice.

Njah, advice already has add/remove which can be connected to a gui button or
what not. However any generic interface always need a specialisation for a case
at hand anyway, at least a different label on a button. But we are drifting
away.

Edebug, debug, elp and trace already have means to add and remove instrumentation.
I ask/suggest for an API to discover if instrumentation is installed or not.

What they don't offer, except for trace, is an API to check if instrumentation
is on or off. Debug is trivial, see the attached patch; trace has
trase-is-traced, elp has private elp--instrumented-p which I suggested to make public.
The only one that truly is missing a mean to test for on/off state is edebug.


> It wouldn't cover the case where the user wants to add such
> instrumentation, tho.

No, it wouldn't, but it does not have to; see comment above.

> And currently, that wouldn't cover Edebug because Edebug doesn't install its
> instrumented code with `advice-add`.

No edebug does it by other means; it instruments function with wrapped code. I
have posted example of how helpful.el does check, and I have also attached a
patch that uses different method to check for on/off state based on built-in
function as posted with patch.

I don't know which one is more efficient, but I don't like neither of those two,
I don't think any of suggested solutions is efficient. Also they both rely on
internal state that can change whenever.

Since we already talk edebug, I think there is a bug. I meant to report it
when I have done patch to fix it, but I haven't had time with that yet:

When a function is instrumented for edebug, edebug adds some properties into
symbols property list. However when instrumentation is removed, edebug leaves
those properties to scrap behind; that inclusive the position in buffer where it
was active. IMO it is a bug; it shouldn't leave scrap behind. It would also be
trivial to check if a funciton is instrumented or not, if those properties were
removed when instrumentation is removed.

Unless I misunderstand the purpose why properties are left behind after
instrumentation is removed :).



  reply	other threads:[~2021-10-04 21:58 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-10-04 15:20 Patch: debug-instrumented predicate Arthur Miller
2021-10-04 16:24 ` Stefan Monnier
2021-10-04 19:25   ` Arthur Miller
2021-10-04 20:00     ` Stefan Monnier
2021-10-04 21:58       ` Arthur Miller [this message]
2021-10-04 22:14         ` Stefan Monnier
2021-10-04 22:56           ` Arthur Miller

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=AM9PR09MB4977C69522B279DE3039EF1F96AE9@AM9PR09MB4977.eurprd09.prod.outlook.com \
    --to=arthur.miller@live.com \
    --cc=emacs-devel@gnu.org \
    --cc=monnier@iro.umontreal.ca \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
Code repositories for project(s) associated with this external index

	https://git.savannah.gnu.org/cgit/emacs.git
	https://git.savannah.gnu.org/cgit/emacs/org-mode.git

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.