Hi,
I wasn't aware of this thread until now, but we ended up fixing the same
issues recently (see below).
On Sun, Mar 02, 2014 at 08:35:05PM +0100, Martin Pohlack wrote:
> Hi,
>
> the attached patch seems to fix the problem for me. I basically invoke
> the relevant hooks early in key_press() and key_release() to allow the
> embedded terminals to do their own event processing.
We made this change for key_press in 9.21 because otherwise the key
press hook does not behave as documented, namely
"If the hook returns true, then the event will be ignored by rxvt-unicode."
Incidentally, this change broke the mark-yank-urls ext and possibly
other exts which rely on the undocumented and buggy behaviour that urxvt
will process a key press event even if the key_press hook returns a true
value.
We did not do the same change to key_release but the only thing that is
performed before invoking the key_release hook is ISO 14755 mode, which
should have higher priority anyway. Do you need the same change also in
key_release and, if so, why?
> The perl wrapper functions needs to do an additional refresh_check(),
> similar to the x_cb() that usually calls the key_press() and
> key_release() functions.
Concerning this point, we fixed it two days ago by adding a
refresh_check perl method and calling it in tabbed after forwarding a
key_{press,release} event to the active tab. Whoever is the current
tabbedex maintainer can apply the same change to tabbedex after the next
urxvt release.
Emanuele
_______________________________________________
rxvt-unicode mailing list
[email protected]
http://lists.schmorp.de/cgi-bin/mailman/listinfo/rxvt-unicode
