Re: Upcoming loss of usability of Emacs source files and Emacs.

[Tassilo Horn]

Alan Mackenzie <address@hidden> writes:

Hi Alan,

>> After this discussion, I'm now pretty much indifferent on these
>> changes and if they should apply to display only or also to actual
>> source code.  But if the main problem with `foo' is its ambiguity,
>> and ‘foo’ is still ambiguous (but less so), why not use something
>> which is even less ambiguous?
>
> Because it might not be typeable, and it might not be displayable.

With typing aids like electric-quote-mode any obscure quote/markup
characters could be typeable.

>> For example, I've seen uses of ⸢foo⸥ for marking up source code
>> snippets.  That seems pretty unambiguous (although you can never be
>> sure), ....
>
> Whatever those quoting characters might be, they display on my
> terminal as inverted question marks.  I haven't a clue what they might
> be, and it would be a lot of work to find out.

They don't need to be displayed as such.  For example, they could be
displayed with special faces with fallbacks for font-lock
disabled/missing terminal capabilities, etc.

The aim of the procedure is to be able to unambiguously refer to
functions/variables and write code snippets in docstrings.  Right now,
this is all done using `...' and in some cases, that doesn't work too
well because at least ` is not too uncommon to be part of snippets, too.
So then we might use ‘...’ which is less ambiguous.  ⸢...⸥ is completely
unambiguous with respect to what we currently have.

But the issue of how special parts of docstrings are enclosed is only
one kind of ambiguity we have.  For example, it would be very useful to
be able to refer to foo-mode the function and not the variable.  And my
bar-mode global bar-method variable could have a possible value
insert-char (the symbol, not the function).

>> BTW: On a related note: I don't get why *Help* buffers display quotes
>> at all when they are already displayed with a face that makes them
>> distinguishable from "normal" text and are clickable.
>
> Help buffers traditionally display verbatim the text from the source
> file (with the exception of explicit directives such as to display
> bindings).  The face might not always be distinguishable from normal
> text (e.g. font-lock might be disabled).  Anyhow, displaying the
> quotes doesn't do any harm.

No, and probably omitting them would break docstrings with ASCII tables
which assume that a cell with `foo' is displayed 5 chars wide.

Bye,
Tassilo