bug-gnu-emacs
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

bug#50005: 28.0.50; silence a couple byte-compiler warnings in ERC


From: Mattias Engdegård
Subject: bug#50005: 28.0.50; silence a couple byte-compiler warnings in ERC
Date: Thu, 12 Aug 2021 13:52:16 +0200

12 aug. 2021 kl. 12.56 skrev J.P. <jp@neverwas.me>:

> Thanks, I spotted the weird double-quote mismatch thing earlier, but I
> guess I'm still missing something else?

The string made it look like backslash were a special character inside [...]. 
It isn't; only ^, ] and - are special, each with their own quirky rules. (Other 
regexp systems have different rules.)

> So, I take this to mean there's no problem with requiring (autoloading)
> rx.el (or regexp-opt.el), right?

Correct. Your original problem was the use of `regexp-opt-charset` which is not 
autoloaded (and somewhat of an internal function to regexp-opt.el, but that's 
debatable). Using plain `regexp-opt` is fine.
Most ordinary uses of rx will not even require rx.el to be loaded at runtime 
since the macros are expanded during compilation, but `rx-to-string` is a plain 
function, autoloaded.

> And I guess the `string-to-list' is necessary because
> 
>  (rx-to-string '(+ (in "abc")) t)
>   ⇒ "[a-c]+"
> 
> is still interpreted as a set of character alternatives, whereas
> 
>  (rx-to-string '(+ (in ?a ?- ?c)) t)
>   ⇒ "[ac-]+"
> 
> combines individual character args into just such a set.

If you meant (in "a-c") in your first example then yes, quite right -- the 
hyphen is special inside strings of the rx `in` construct, but it's not special 
as a character argument. No other characters are special.

> Do you use ERC, Mattias? If so, are you available to review more
> patches?

I'm afraid I rarely use it and know nothing about its innards so I'm probably 
not the right person to ask -- only came by to help out with some regexpy 
things. I have nothing clever to say about your other changes, sorry!

> I'm obviously quite ignorant in the ways of Emacs but am
> pushing hard to improve ERC. Some might say that's a recipe for
> embarrassment/disaster (though I'm pretty sure I already crossed that
> first bridge ages ago).

I'd say it's a good sign! Just dive in.

> So what do you say, can you help?

Wish I could but time probably doesn't permit, sorry. There are much more 
experienced hands about. Of course I'll be happy to answer specifics about 
parts that I happen to know something about.






reply via email to

[Prev in Thread] Current Thread [Next in Thread]