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

Re: [PATCH] w3m-lnum add highlight



In [emacs-w3m : No.11253] Andrey Kotlarski wrote:
> Hello, Katsumi-san,

> Thanks for feedback and sorry for slight delay.

> I decided to go a step further in emulating Conkeror's behavior.
> There's a function `w3m-linknum-follow' which follows urls, activates
> form fields or pushes buttons depending on what's selected (not sure
> whether done correctly but works where tested).

Excellent!  It's easy to use, so you who love the w3m-lnum feature
should try the new behaviors soon.  I've already installed your
changes to CVS[1].  `C-c C-l' is what those who have never used
it do first.

> It might be heretic but I changed behavior (along highlighting matches)
> of `w3m-go-to-linknum' allowing it to go to form fields and buttons as
> well, which seems more natural to me as these elements are normally more
> difficult to "isearch".

Agreed.  I also agree to make the `C-c C-l' key be bound to
`w3m-linknum-follow' by default.  I.e., abolish `w3m-go-to-linknum'
and make indexing of forms the default.  Furthermore, I think it's
convenient if `w3m-read-int-interactive' turns off the
`w3m-link-numbering-mode' mode by the `C-c C-l' key input.  IOW,
making the `C-c C-l' key the toggle switch.  WDYT?

> Added a macro `w3m-with-linknum' which can ease this type of functions
> (and bring old behavior of `w3m-go-to-linknum' with a single digit change).

> Made `w3m-link-numbering-mode' more flexible allowing it 2 types of
> enumeration - only url links (old behaviour and default) and urls along
> with form fields and buttons for conveniences pointed above.

> Also added `w3m-linknum-read-url' for comfortable grabbing of urls and
> allowing for such sort of actions:

> (let ((url (w3m-linknum-read-url "Curl on link: ")))
>   (if url (async-shell-command (concat "curl -O " url))))

> A glitch that haven't yet overcome is changing faces of buttons on match.
> Here's the diff against same current cvs w3m-lnum.el and hopefully this
> time will work closer to advertised.

Thanks again.

[1] I've slightly changed your code.  Let me know if there's any
oddity.