A smart auto-wash? (was: washing and `article-translate-strings' in what hook?)

[Dave Abrahams]

More generally, I find I am constantly fiddling around with articles in
Gnus to get them to be truly readable.  *Many* need `W-Q'. The ones that only
read well as HTML need `K-b K-v'. I found one today that was garbled
without `K-m', which I had to find by experimentation.  Has anybody
written a smart washer that figures it out for you 99% of the time?

on Mon Jun 01 2015, Ben Bacarisse <ben.lists-AT-bsb.me.uk> wrote:

> Emanuel Berg <embe8573@student.uu.se> writes:
>
>> Ben Bacarisse <ben.lists@bsb.me.uk> writes:
>>
>>> I think you need to wrap the body in
>>>
>>>   (gnus-with-article-buffer ...)
>>>
>>> This will have the added effect of making the
>>> interactive function work from the summary buffer
>>> window (provided there is a current article in some
>>> buffer, of course).
>>
>> That's exactly right!
>>
>> But how is anyone to realize this?
>
> Ah, good question.  I don't know.  I learned what I know by reading
> other people's code (some of it the Gnus sources).
>
>> Because there is no article argument to
>> `article-translate-strings', the current article is
>> all it can be applied to (?). So then shouldn't it say
>> there is none, if there isn't?
>>
>> Or did this happen to some *other* article which
>> I have been unaware of?
>
> That's possible.  The code operates on the current buffer, so it was
> probably editing something!
>
>> And why did it work calling it interactively but not
>> doing the same from Lisp?
>> Is `gnus-article-prepare-hook' the wrong place so at
>> that time there isn't a buffer set to work
>> on, interactively?
>
> I don't think the key distinction is interactive/non interactive.  The
> key issue is whether there is a "current buffer" which you can see
> change.  Selecting an article probably makes the article buffer current
> so calling the function interactively works on the article you can see.
>
> <snip>

-- 
-Dave