Re: inputting characters by hexadigit

[Stefan Monnier]

>> So by using chars ? # or \ to mean something else, we get into the
>> situation where the user might not know how to insert #, ?, or \ when
>> those keys are bound to non-self-inserting commands.  I think this is
>> a serious problem.

> Well, that's true enough, though of course already applies to 0-7 or
> 0-9A-F depending on read-quoted-char-radix

That's true, and it bit me already several times.  I'd really prefer not
to make it worse.

> (Though one thing that springs to mind is to allow C-q ## to insert
> a literal #...

Maybe C-q # SPC (and C-q ? SPC, ...) could be used, indeed.  But it's
still yet-another-rule to learn.

> Also, couldn't #?\ pretty easily be made another
> customize option? read-quoted-char-accept-elisp-literals)

I'm only worried about the default behavior, here, but providing this
functionality via an option doesn't seem very satisfactory either.

I think a good solution should start by scrapping the C-q <octal>
corner case, and replace it with something that does not interfere as
much with C-q <char> and which additionally accepts hexadecimal (and
maybe more such as char names, ...).


        Stefan