The preceding commit includes motivation; specifically,
invoking keyboard and mouse bindings should never fail
unless they perform an action which can fail. For instance,
a `bindcode Ctrl+9 exit` should always exit; of course,
`exec` may fail for external reasons.
The core changes are to bind.c and commands.c. As a single
command may actually, on being parsed, generate multiple
commands to be executed (consider `workspace 1; splith`),
the parsing code in execute_command is moved to a separate
function, parse_command, which returns a list_t of stored
commands; leaving execute_command to parse the command,
execute each resulting list element; and then cleanup.
An additional field for the list of parsed stored commands,
is added to struct sway_binding, usurping the name "command".
This is initialized via parse_command, and then
seat_execute_command is adjusted accordingly.
Side effects of this change include:
* Reference counting `struct criteria`, because a criterion
parsed once may apply to multiple stored commands, and
the alternative of duplicating criteria is even more
complicated.
* Adding reading and active arguments to config_handler,
because when bindsyms parse commands, they require handler
lookups to be performed as though it were runtime.
* Reference counting `struct sway_binding`, because the
alternative of updating sway_binding_dup is even more
complicated when allocation failures are be taken into
account.
The original purpose of this commit is to replace some for loops with
list_find. But while doing this I found the workspace_prev_next_impl
functions to be difficult to read and also contained a bug, so I
refactored them and fixed the bug.
To reproduce the bug:
* Have two outputs, where the left output has workspaces 1, 2, 3 and the
right output has workspaces 4, 5, 6. Make workspace 2 focused_inactive
and workspace 4 focused.
* Run `workspace prev`.
* Previously it would visit the left output, then apply `workspace prev`
to workspace 2, which focuses workspace 1.
* Now it will focus the rightmost workspace on the left output
(workspace 3).
The refactoring I made to the workspace functions are:
* Added the static keyword.
* They now accept an int dir rather than bool, to avoid an unnecessary
conversion.
* Rather than preparing start and end variables for the purpose of
iterating, just iterate everything.
* Replace for loops with list_find.
* Don't call workspace_output_prev_next_impl (this fixes the bug).
* The OP_RESIZE seat operation has been renamed to OP_RESIZE_FLOATING,
and OP_RESIZE_TILING has been introduced.
* Similar to the above, seat_begin_resize and handle_resize_motion have
been renamed and tiling variants introduced.
* resize.c's resize_tiled has to be used, so container_resize_tiled has
been introduced in resize.c to allow external code to call it.
* Add and use lenient_strcat and lenient_strncat functions
* Rename `concatenate_child_titles` function as that's no longer what it
does
* Rename `container_notify_child_title_changed` because we only need to
notify that the tree structure has changed, not titles
* Don't notify parents when a child changes its title
* Update ancestor titles when changing a container's layout
* Eg. create nested tabs and change the inner container to stacking
* No need to store tree presentation in both container->name and
formatted_title
The condition checking if the markup is valid was inverted.
This commit also adds better error handling: if the markup cannot
be parsed, it fallbacks to plain text.
This starts up the event loop and wayland display and shims out the
basic top level rendering concepts. Also includes some changes to
incorporate pango into the 1.x codebase properly.
Increase _POSIX_SOURCE value where needed.
Increase _XOPEN_SOURCE value where needed.
Conditionally link to libcap (only on Linux).
Possibly some trailing whitespace fixes (automatic).
_sway_assert is a variadic function which tries
to delegate to another variadic function. This
requires a vprintf-style variant of the delegate.
https://stackoverflow.com/a/150616
Colors are configured through the command line so that swaylock conforms
to the i3lock fork 'github.com/chrjguill/i3lock-color'. Differences from
it are that one letter options '-r' and '-s' are not implimentend because
'-s' is already used by '--scaling' in swaylock.
This commit also fixed whitespace in 'include/swaylock/swaylock.h' and
changed `parse_color` in 'common/util.h' so that it can accept colors
that do not start with a hash. This was done to keep compatability with
the i3lock fork.
