Re: utf-translate-cjk-mode

[Dave Love]

Luc Teirlinck <address@hidden> writes:

> I believe that it was decided in a previous thread that if a defcustom
> uses a :set function without providing a way to accomplish exactly the
> same result without using Custom and clearly documenting this in the
> documentation string, then this is a bug which should be fixed.

I don't know anything about that, and I don't see how anyone is
prevented from accomplishing exactly the same result as the :set.

> Kai's change was a bug fix.

I disagree.  I would have made it a minor mode if I thought that DTRT.

> I should clarify that you (or whoever wrote the
> initial defcustom) can definitely can not be blamed for what we, after
> the fact, decided to consider a "bug".

I'm quite happy to take the blame for things I've done wrong and for
them to be fixed, but I'm explaining why this case wasn't wrong.

> Note that I personally do not mind about using Custom.  I believe it
> is extremely useful and I personally do use it.

You are very wise.

> Other people feel extremely strongly that they should be able to use
> Emacs without using Custom and it was decided that they definitely
> should be able to do so.

I'm baffled why they think they can't (or take that attitude in the
first place).

>    That's still not good enough, since customize doesn't act on custom
>    dependencies as I think it should, but it helps.

I think I meant to write `customize-option', not `customize', sorry.

> This sounds rather cryptic.  Do you mean that there is a bug in
> Custom's :set-after machinery?

No, it just doesn't take advantage of the dependency information where
it could, only in `custom-set-variables'.  There are various places
where one Custom option depends on the settings of others and doesn't
get updated when they change, as a user would probably expect.  For
instance, I gave up trying to sanitize some Calendar stuff because of
that.