[Top][All Lists]

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

Re: Multiple next-error sources

From: Daniel Colascione
Subject: Re: Multiple next-error sources
Date: Fri, 07 Nov 2014 17:10:29 +0000
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.2.0

On 11/07/2014 04:55 PM, Alan Mackenzie wrote:
> Hello, Stefan.
> On Fri, Nov 07, 2014 at 09:59:12AM -0500, Stefan Monnier wrote:
>>> The advantage of hooks is that it makes it easier for users to
>>> customize the behavior by adding and removing various entries.
>> While it's true that
>>     (add-hook 'next-error-functions #'my-function)
>> is shorter than
>>     (add-function :before-until next-error-function #'my-function)
>> I don't think it warrants the addition of a next-error-functions.
> Why not?
>> If the :before-until is the problematic part, then I guess we should
>> look for ways to improve that (e.g. a better name, or some way for
>> a variable to say that :before-until is the default when adding
>> functions to it?).
>>> Is add-function intended to replace hooks like this in general?
>> Somewhat, yes.  I have no intention to go around and replace existing
>> hooks in the forseeable future (except for those rare hooks that used
>> with-wrapper-hook), but I'll favor new foo-function over new
>> foo-functions.
> Why?  I'm adding my voice to the clamour of dissent.
> It would seem that the use of single functions, with `add-function' is
> inferior to the conventional hook mechanism in every way.  What am I
> missing?  In addition to the things cited by Daniel, there's:
> (i) the danger (near certainty) that somebody is going to use `setq'
> rather than `add-function' to configure it;

The same critique applies to regular hooks, doesn't it?

> (ii) the additional incompatibility with other Emacsen;

I'm not sure that compatibility with other Emacsen is as important as it
once was. AIUI, GNU Emacs is receiving the vast majority of development

> (iii) the difficulty (or perhaps clumsiness) in looking at the contents
> of an advised function.  There would seem to be nothing equivalent to
> M-: after-change-functions.
> So why are you changing from the conventional hook mechanism, which works
> so well?  What is the advantage of the new scheme.
> Incidentally, I had a look at the documentation for add-advice, and
> there's a problem with it.  "Advice" in English has no plural - there's
> no such word as "advices".  If it's necessary to emphasize the plurality,
> then "pieces of advice" can be used.  There's no such thing as "an
> advice", rather you'd say "some advice".  It's a bit like you wouldn't
> refer to a lake as "a big water"; you'd say it contains "a lot of water".
> I think there's a term in linguistics for such a word, but I don't know
> it off hand.

I think "advise" works like "code" in the software sense and
"furniture". The term is "mass noun".

> Incidentally 2, the verb corresponding to the noun "advice" is "to
> advise".

Isn't English fun?

Attachment: signature.asc
Description: OpenPGP digital signature

reply via email to

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