Re: Strange auto-caching

[kai.grossjohann@uni-duisburg.de (Kai Großjohann)]

Kevin Greiner <kgreiner@xpediantsolutions.com> writes:

> OK, I went back to your original post.  It appears that you always
> fetch articles by selecting them.  That means that you're always
> calling gnus-agent-fetch-selected-article.
> [...]
> I know that you don't want to consider that g-s-u-a-l is the problem.

Yeah, I've been a really stubborn thickhead, haven't I?  (You don't
have to answer.)

I think you're onto something, because I think that the problem with
the bogus `not downloaded' highlighting comes from articles that I
have selected before and then hit M-u on them.  This means that
gnus-summary-update-article-line was in fact called.

So I think you're right after all.

> However, I believe that you can test it fairly easily.
>
> I'd like you to do the following in your summary buffer.
>
> M-: (let ((inhibit-read-only t)) (upcase-region (point-min) (point-max)))
>
> This will upcase your entire summary.  It will look ugly but it will
> sure help ID any changes to the text.  Now, try to read some of the
> unfetched articles.  If g-s-u-a-l is working correctly the selected
> article's line will refresh to mixed-case.  If it isn't working
> correctly, some other line will refresh.
>
> Let me know what happens when the download status is wrong.

I tried it immediately.  But I will try it in a group where the
highlighting is wrong.  For a group where the highlighting is not
wrong, I observe:

(a) the correct line is changed
(b) the change in the current line only affects the color and the `R'
    mark (which appears on selecting and disappears on M-u).  The
    rest of the summary line is not changed -- it stays uppercase.

-- 
A turnip curses Elvis