bug#38406: 27.0.50; post-self-insert-hook does not hold its contract in cc-mode derived modes

[yyoncho]

Hi Alan,

> There are other possible "fixes", for example modifying these functions

> so that they don't use self-insert-command at all, but somehow I don't
> think that's what you want.

I don't think that the code that is implemented against the contract listed in
the hook documentation should be rewritten. If electric stuff is so that
important and there is no way to disable it by default then at least a function
to unbind the electric functionality the documentation of post-self-insert-hook
should state: "Don't rely on this hook in cc derived modes because of
{implementation details}. If you still want to use post-self-insert-hook disable
use {implementation details} to turn electric off." 

Thanks,

Ivan

On Sat, Nov 30, 2019 at 4:36 PM Alan Mackenzie <acm@muc.de> wrote:

Hello, Yyoncho.

On Wed, Nov 27, 2019 at 22:00:26 +0200, yyoncho wrote:
> As per post-self-insert-hook documentation.

> > Hook run at the end of `self-insert-command'.
> > This is run after inserting the character.

Yes.  This is a problematic hook, since it is capable of disrupting the
correct functionality of any Lisp program which uses
self-insert-command.  This transpired in CC Mode, so to make
c-electric-brace and friends work, the action of the hook was nullified.

> This does not hold by default in cc-mode due to the following mapped by
> default functions:

> > (define-key c-mode-base-map "#"         'c-electric-pound)
> > (define-key c-mode-base-map "{"         'c-electric-brace)
> > (define-key c-mode-base-map "}"         'c-electric-brace)
> > (define-key c-mode-base-map "/"         'c-electric-slash)
> > (define-key c-mode-base-map "*"         'c-electric-star)
> > (define-key c-mode-base-map ";"         'c-electric-semi&comma)
> > (define-key c-mode-base-map ","         'c-electric-semi&comma)
> > (define-key c-mode-base-map ":"         'c-electric-colon)
> > (define-key c-mode-base-map "("         'c-electric-paren)
> > (define-key c-mode-base-map ")"         'c-electric-paren)

> All of these functions (or at least majority) contain the following lines:

> > (let (post-self-insert-hook) ; Disable random functionality.
> >      (self-insert-command (prefix-numeric-value arg)))

Yes.  This was one way to get self-insert-function to perform its
correct functionality, namely inserting exactly one copy of a typed key
(or alternatively N copies when there's a prefix key).

> Possible fixes:

First question, what's the problem?  What do you want to do that the
above mechanism hinders?

> 1. Do not bind the functions by default.

They are essential to the correct functioning of CC Mode.

> 2. Rewrite them so they do not inhibit post-self-insert-hook functions.

This is difficult.  If there were an easy way to do this, I would have
done it.  Note that, from the point of view of a major mode,
post-self-insert-hook is totally random functionality - it is a bit like
a trojan horse.  The major mode has no way to control what it does, thus
is unable to guarantee the major mode will work.

There are other possible "fixes", for example modifying these functions
so that they don't use self-insert-command at all, but somehow I don't
think that's what you want.

Another fix would be to specify restrictions on what one is allowed to
do in this hook.  I would prefer this, but other people would object
strongly.

I would advise against using post-self-insert-hook, if you possibly can.
after-change-functions may be a good alternative.  Maybe you can't.  So,
again, what is it you're trying to do?

> In GNU Emacs 27.0.50 (build 1, x86_64-pc-linux-gnu, GTK+ Version 3.24.1)
>  of 2019-11-23 built on kyoncho-H87-D3H
> Repository revision: 8934762bb37273e6606097de92dcc2556456acd2
> Repository branch: master
> Windowing system distributor 'The X.Org Foundation', version 11.0.12001000
> System Description: Linux Mint 19.1

[ .... ]

--
Alan Mackenzie (Nuremberg, Germany).