bug#50236: 27.2; electric-pair-mode is inconvenient in comint

Ihor Radchenko <[email protected]> writes:

> João Távora <[email protected]> writes:
>
>> Ihor Radchenko <[email protected]> writes:
>>> Paragraphs was just an example. My point is - major mode needs to decide
>>> such things. And fields often serve different purposes, so, IMHO, they
>>> are not the best vehicle to decide about bracket balancing.
>>
>> What purposes do fields often serve?  I've seen them in comint-based
>> shells, widgets, etc.  They don't "decide" about bracket in the same way
>> that don't don't "decide" about movement.  This is just about isolating
>> operation of certain features (delimiter balancing, movement, etc) to a
>> given region.  The uses I've seen of fields all match perfectly.
>
> IMHO, fields are more suitable to mark short spans of the text within
> limited region in buffer. (e.g. tables).
>
>> If you'd rather make a complicated setting of electric-pair-skip-self
>> and electric-pair-inhibit-predicate etc in your major mode (like c-mode
>> infamously did and c-ts-mode fortunately didn't), you can.  I just
>> happen to think that setting a continugous 'field' property in your text
>> source blocks is a nice thing to do for your users.  But I don't use
>> much org.
>
> In Org's case, we would need to fill the whole buffer with
> fields.

I'm not suggesting you delimit your buffer into paragraphs/headings.

But _if_ you wanted to do that, you'd be wrong.  You wouldn't need to
"fill the whole buffer with field [properties]".  A field is just a
region of text with the same ‘field’ property, so if you set a single
char's worth of "field=blargh" in the '*' (where you already set an
overlay anyway aparently) you have the needed demarcations.  Delete that
heading and you magically merge the two fields for free.  Play around
with field-beginning and field-end to see what I mean.  These functions
are in C and they are pretty fast.

> (Here, we do not care about previous section brackets)

Not necessarily.  Currently org matches across different headings and I
haven't seen anyone complaining.  Personally, it's only inside code
blocks I think it makes a little more sense to track balance
independently.

> #+begin_src emacs-lisp
> (message "And here we can Elisp code, that should be isolated")
> #+end_src

Only here would I do the courtesy of marking that a field, because it
comes from another language completely.  You already set a 'src-block'
to 't' property there.  I'd put a 'field' as well, and call it a day.
It's probably a sub-1-loc patch which I would show if I somehow dared to
navigate that petaline code base, but surely you can.

João