This is implemented by allocating one of the (few!) remaining bits in
the cells’ attribute struct to indicate the cell should be “URL
highlighted”.
render_cell() looks at this bit and draws an underline using the color
from colors.urls (defaults to regular3 - i.e. yellow).
A new function, url_tag_cells(), iterates the currently detected URLs
and sets the new ‘url’ attribute flag on the affected cells.
Note: this is done in a separate function to keep urls_collect() free
from as many dependencies as possible.
urls_reset() is updated to *clear* the ‘url’ flag (and thus implicitly
also results in a grid refresh, _if_ there were any URLs).
We now exit URL mode on *any* client application input. This needs to
be so since we can’t know if the URLs we previously detected are still
valid.
This allows us to update the jump label positions when the viewport
changes.
This in turn allows us to stay in URL mode while the user is using the
mouse to scroll in the scrollback history.
Scrolling with the keyboard is currently not possible, since input
handling in URL mode does not recognize “regular” key bindings. We
_could_ add scrollback up/down bindings to URL mode too, but lets not,
for the time being.
(Note: an alternative to this patch is to disallow mouse scrolling
too. Then we could have kept the URL start/end as viewport local
coordinates).
* colors.jump_labels configures the foreground and background colors
used when rendering URL jump labels. Defaults to “regular0
regular3” (i.e. black on yellow).
* colors.urls configures the color to use when highlighting URLs in
URL mode. Note that we aren’t currently doing any
highlighting... Defaults to regular3 (i.e. yellow).
This works just like show-urls-launch, except that instead of opening
the URL (typically using xdg-open), it is placed in the clipboard when
activated.
While parenthesis and brackets _are_ valid URL characters, there are
many times when we do *not* want them to be part of the URL.
For example, in markdown we write “[text](url)”, or even
“[](url-2)”.
Here, the URLs are clearly *not* “url)” or “url-1)](url2)”.
action_print() is in the hot path, and having if-statement here *does*
have an impact on performance.
Much more so when that if-statement involves a functional call to
wcwidth().
Closes#330
The main surface is the *only* IME capable surface we have. If the
compositor is sending an ime::enter event for another surface, it is a
compositor bug, not a foot bug.
When the application is grabbing the mouse, we require ‘shift’ to be
pressed to be able to use foot’s mouse bindings. Otherwise the mouse
event is sent to the application.
This wasn’t being enforced consistently; each and every mouse **only**
action was checking the mouse grab state, but e.g. *primary
paste* (which can be used as both a key binding and a mouse binding)
was not.
The fix is actually quite simple: don’t even try to execute *any*
mouse bindings when the application is grabbing the mouse (unless
shift is pressed).
This fixes an issue where e.g. middle mouse button events (using the
default mouse bindings) were never seen by any mouse grabbing client
applications.
