E.g. trying to press mouse button 1.5 via root.fake_input() doesn't make sense.
Previously the code silently truncated the number to an integer. Now it
complains about this instead.
Signed-off-by: Uli Schlachter <psychon@znc.in>
The only remaining calls are for a window's opacity and in the DBus type
handling. Everything else wants integers, not something with a comma.
Signed-off-by: Uli Schlachter <psychon@znc.in>
This opens xterm, closes it and makes sure that the client object representing
xterm is GC'able at the end. The test will fail currently.
Signed-off-by: Uli Schlachter <psychon@znc.in>
Any awful.titlebar.widget.button widget (e.g. floatingbutton or closebutton)
decides on the currently visible symbol based on several factors. One of them is
"is the client currently focused?" and thus the button has to be updated when
the client is focused/unfocused.
The way the code did this was to use client.connect_signal("focus", f) and
client.connect_signal("unfocus", f). However, these signals are never
disconnected and kept alive forever. The callback function had a strong
reference to the client (as an upvalue) and thus this also prevented the client
from being garbage collected.
Fix this by using c:connect_signal("focus/unfocis", f) instead. These kind of
signals are only kept alive by the client object and don't prevent it from being
garbage collected.
This fixes the new test that the previous commit added.
Signed-off-by: Uli Schlachter <psychon@znc.in>
ICCCM specifies when the WM has to send a ConfigureNotify. Java does not care
and wants one all the time. Meh.
Fixes: #248
Signed-off-by: Uli Schlachter <psychon@znc.in>
Until now, this layout was "append only". There was no official
APIs to remove, replace, insert and swap widgets. This is fine
for the usual wibox + sensors widget used by the majority of
users, but lack flexibility necessary to use the layout system
to place dynamic elements such as clients.
The methods introduced by this commit are also recursive. This
allow widgets to be decorated, wrapped and splitted without
having to add boilerplate code everywhere.
This remove duplicated code and will allow more "collection"
style layouts to be implemented without logic duplication.
This commit also do some small cleanup to remove duplicated
code now present in `awful.util`.
Fixes https://github.com/awesomeWM/awesome/issues/617
Go around a limitation of the lua language spec. The return value
of this method guaranteed `ipairs()` correctness.
Please note that both the official Lua and Luajit implementation
provide a sparse table compatible `ipairs()` and __len implementation
by default.
It is an internal API and is used by `gears.shape`, `gears.pattern`
and `gears.composition` only.
This commit also add `:rotate_at` and `:copy` methods.
Via this commit, any warnings from LDoc are fatal and make "make" fail. The
intention is to cause failures on Travis for PRs that introduce broken
documentation.
Closes: https://github.com/awesomeWM/awesome/issues/643
Signed-off-by: Uli Schlachter <psychon@znc.in>
This also removes some obsolete dependencies and it also and creates
`/usr/bin/lua` even when using luajit. Without this, our lgi-check.sh
script complains about a missing lgi installation because it cannot find
a lua binary.
Closes https://github.com/awesomeWM/awesome/pull/652.
Signed-off-by: Uli Schlachter <psychon@znc.in>
A quote from the XKeyboard specification:
The server notifies interested clients of keyboard map changes in one of two
ways. It sends XkbMapNotify to clients that have explicitly selected them and
core protocol MappingNotify events to clients that have not. Once a client
requests XkbMapNotify events, the server stops sending it MappingNotify events
to inform it of keyboard changes.
This commit moves the code that we had for handling MappingNotify events to the
place where we handle XkbMapNotify events. This might even fix some bugs where
parts of awesome continued to use old key binding "stuff"!
Signed-off-by: Uli Schlachter <psychon@znc.in>
The benchmarks in tests/test-benchmark.lua have two modes. When CI=1 is set in
the environment, only a "quick" and less exact test is done. Otherwise, a slower
and more exact measurements is taken. This was added so that we do not waste CPU
time on travis.
However, most of the time the user running "make check" doesn't want exact
measurements either. So instead of only being quick when CI=1 is set, this
commit changes the logic to always being quick unless BENCHMARK_EXACT=1 is set.
Additionally, a message is printed next to the benchmark results so that the
user is reminded to set this var if the measurements should actually mean
something.
Signed-off-by: Uli Schlachter <psychon@znc.in>
It is necessary to have it beforehand when creating layout objects
for unselected layouts.
In the current layout system, there is no layout object, but to allow
tabs and dynamic tagging features like ion3, layouts cannot be stateless.
Before, it was the caller job to make sure the client wasn't floating.
This limitation is unecessary. awful.client.idx now return nil instead
of an error. awful.rules setting the master width factor are now
foolproof.
actionless
Uli Schlachter
Emmanuel Lepage Vallée
Daniel Hahler
Cory Burgett