bug#20385: [PATCH] Support curved quotes in doc strings

[Paul Eggert]

Dmitry Gutov wrote:

> I don't really understand the
> motivation for the original proposal (to switch away from `...'), so it's not
> clear to me if font-locking would satisfy it.

The main motivation is that English text shouldn't use grave accent to quote. 
It looked good decades ago but the underlying encodings changed and now it is 
klunky and offputting.  (It's not as bad as the 1950s syntax 16HTHIS IS A STRING 
but that's a low bar....)  Yes, it was a GNU tradition for many years, but other 
GNU packages (GCC, coreutils, etc.) have largely shifted away from it and it's 
time Emacs made it more convenient to use the more-standard convention of curved 
quotes.

I haven't tried font locking.  As I understand it, though, font locking would 
address the problem only in doc strings.  For example, it wouldn't address 
Emacs's diagnostic messages, which also need to get fixed.  In contrast, the 
sorts of solutions I'm proposing should help support curved quotes nearly 
everywhere.

> "Plain" unicode strings are not that plain, especially if it still takes 4
> keypresses to type the character, and I also need to explain to contributors how
> to do that.

The patch proposed in Bug#20545 largely addresses this problem.  Contributors 
can use the same keypresses as before.  If your contributors type this:

  The value may be `buffered', `retained', or `non-retained'.

the following characters are put into their doc string:

  The value may be ‘buffered’, ‘retained’, or ‘non-retained’.

They won't have to do anything special to get this behavior; just use the 
revised Emacs on its own source code.

> For some reason still unclear to me (I have English locale and language set
> everywhere I can see), it displays a group of cyrillic characters (тАШ) instead
> of the fancy quotes. Which will complicate reading small patches somewhat (ones
> I wouldn't open in an external program otherwise).

I reproduced that problem in Thunderbird by visiting "View > Character Encoding 
> Auto-Detect" and selecting "Russian".  To fix it, I selected "(off)" instead 
of "Russian".