Re: w3m-clear-display-while-reading vs w3m-view-header on link

[Boruch Baum <boruch_baum@xxxxxxx>]

On 2018-05-25 18:03, Kevin Ryde wrote:
> Boruch Baum <boruch_baum@xxxxxxx> writes:
> >
> > I still have no luck reproducing your report, but upon re-reading maybe
> > we have different expectations of what the command should do.  Per the
> > docstring, function `w3m-view-header` ...  Display the header of the
> > *current page*, not the header of the url at point.  Is that the
> > confusion?
>
> Maybe, but if you nose around the code in `w3m-about-header' you can see
> how it's supposed to report info from the current page concerning link
> and/or image at point.  Fixed in any case now in w3m-process-stop hopefully.

I just did now take a look at the code with the debugger, and see what's
going on with me, which is puzzling. My copy of the function includes
the snippet:

1   (let (anchor anchor-title
2                image-url image-alt image-size)
3     (with-current-buffer w3m-current-buffer
4       (when (equal url w3m-current-url)
5         (setq anchor (w3m-anchor)
6               anchor-title (w3m-anchor-title)
7               image-url (w3m-image)
8               image-alt (w3m-image-alt)
9               image-size (w3m-get-text-property-around 'w3m-image-size))))
10     (if anchor
11         (insert "\nCurrent Anchor: " anchor))
12     (if anchor-title
13         (insert "\nAnchor Title:   " anchor-title))
14     (if image-url
15         (insert "\nImage:      " image-url))
16     (if image-alt
17         (insert "\nImage Alt:  " image-alt))
18     (if image-size
19         (insert (format "\nImage Size: %sx%s"
20                         (car image-size) (cdr image-size)))))

For me, the 'equal' test on line 4 always fails because the value of
`w3m-current-url' is nil(!) at that point, in that context. I was
surprised, but when I double-checked in the debugger by switching to the
`w3m-current-buffer', the location bar at the top was blank! Also, point
had moved to (point-min), so all the assignments on lines 5-9 returned
nil. All was restored at some point before the command completed,
though. So, at some point, the code is for me temporarily messing with
`w3m-current-url'.

That explains why my output doesn't include the header data that you
get, but I haven't found why the behavior is happening.

>
> > In order to get a link's page header information, we would
> > need to post a GET request for that page, which would give us the entire
> > page along with the header, so it would be the same as loading the link.
>
> Yes, no, no fetching, just what the current page has about the link.

The fetching may possibly happen because the call to w3m-goto-url is
being made without the RELOAD arg set to 'redisplay, which I understand
means that if the page has expired, it will fetch. That seems the
correct behavior, but just so you know.

> I sometimes like a HEAD to look at a link to check how massive before
> downloading but let's not confuse ourselves with a separate topic :-)

Nice idea, especially for us North Americans living in a country with
third-world minus internet speeds.

> Personally I'm happy in the same buffer, and it's uncommon anyway.

Fine by me.

-- 
hkp://keys.gnupg.net
CA45 09B5 5351 7C11 A9D1  7286 0036 9E45 1595 8BC0