[Conkeror] flailing about on upgrade

[Mark Eichin]

On Sun, Mar 23, 2008 at 2:34 AM, Jeremy Maitin-Shepard
<jeremy at jeremyms.com> wrote:
> "Mark Eichin" <eichin at gmail.com> writes:
>
>  [snip]
>
>
>  > I think I was getting bit by the usual "X paste doesn't actually work
>  > half the time" bugs; c-y does get a value to paste from *somewhere*
>  > but it isn't what I select in emacs or xterm, at least most of the
>  > time; whatever m-space does was more reliable, I think.
>
>  I don't remember what M-space used to do, but I'll look into it and see
>  if that is a command to add back again.  There isn't so much an "x paste
>  doesn't actually work half the time" bug but there are several different
>  notions of a clipboard in X, which can create confusion and
>  inconsistency.  Most commonly used are the PRIMARY selection, which is
>  generally used by terminal emulators, and by emacs by default, and by
>  other programs (i.e. gtk programs) when you simply select text with the
>  mouse, and the CLIPBOARD selection, which is used when you explicitly
>  choose e.g. a cut or copy command or paste command in most programs
>  (whereas middle click uses the PRIMARY selection typically.

In fact, I just found a case where c-y did nothing (other than expand
the minibuffer to show selection possibilities) but middle-click
worked.  Hmm, looks like shift-insert is bound to
paste-x-primary-selection, which may be what I'm looking for instead
of c-y...

http://www.jwz.org/doc/x-cut-and-paste.html is a pretty good rant about it :-)

>  You can set the keymap to pass through unbound keys, but as more
>  functionality is added to Conkeror, more keys will likely be bound in
>  the default keymaps, and so your Google Mail will gradually become less
>  functional.  You could also bind keys like M-n and M-p and C-x C-f even
>  in the key binding that is used for C-M-q.

Ahh, the string concatenation trick in
define_content_buffer_input_mode is why I couldn't find where that was
in code, at first.  Yeah, I'll try that, especially since it would let
me map out what things I actually think are needed there (I'd assumed
it was a lot more magic than that :-)

>  You could probably get that to work easily enough.  JavaScript may also
>  be able to do what you need directly, though.

Well, what I've got in JS now is a machine-generated block of translations:

non_ascii_replacements = {
"\u200c": "&zwnj;",
"\xe5": "&aring;",
etc. that get applied to .innerHTML (though it looks like the current
code also passes it to btoa, I think that's a separate issue.) I
couldn't find js primitives to do something sane at time time (I think
without the quoting, btoa isn't enough either, for a string passed to
XMLHttpRequest...) though since I'm recasting it for modern conkeror,
I'm rewriting it anyway and open to suggestions.  [This is part of my
web-clipping commands.]

-- 
_Mark_ <eichin at thok.org> <eichin at gmail.com>