Official autoconf docs contraindicate the use of the $sysconfdir
variable in configure.ac. For reference:
https://www.gnu.org/software/autoconf/manual/autoconf-2.63/html_node/Installation-Directory-Variables.html
The macro SYSCONFDIR (ordinarilly set to "/etc") is required to find
a system configuration file installed with i3. In some build setups,
this may not be defined properly in when the configure script is
compiled.
Instead, define this variable in AM_CPPFLAGS as the documentation
indicates.
fixes#2832
If getopt() returns '?' (option with a missing argument), exit the program with an error code rather than continuing the execution.
Fix the issue #2804 (https://github.com/i3/i3/issues/2804).
This patch fixes the issue #2802 (https://github.com/i3/i3/issues/2802).
1). Revise the state machine for the 'rename workspace' command.
These scenarios are considered:
a). 'rename workspace to to bla'
state transitions: RENAME -> RENAME_WORKSPACE -> RENAME_WORKSPACE_LIKELY_TO -> RENAME_WORKSPACE_LIKELY_TO_NEW_NAME
b). 'rename workspace to tosomething'
state transitions: RENAME -> RENAME_WORKSPACE -> RENAME_WORKSPACE_LIKELY_TO
c). 'rename workspace to to'
state transitions: RENAME -> RENAME_WORKSPACE -> RENAME_WORKSPACE_LIKELY_TO
d). 'rename workspace to bla'
state transitions: RENAME -> RENAME_WORKSPACE -> RENAME_WORKSPACE_LIKELY_TO
e). 'rename workspace bla to foo'
state transitions: RENAME -> RENAME_WORKSPACE -> RENAME_WORKSPACE_TO -> RENAME_WORKSPACE_TO_NEW_NAME
2). Add a test case in 117-workspace.t for the scenario b.
The lookup order stated in the i3 manpage was is contratiction to the
actual lookup order, which was introduced with commit
bfa12a5819.
Since that commit (6.5 years ago) the "traditional" paths took precedence
over XDG_CONFIG paths.
Subset comparison was introduced with the rather large commit
bf3cd41b5d, but I now think we should use
equality.
In other words, the following key binding:
bindsym Mod4+x nop Mod4+x
previously would have been triggered when pressing Mod3+Mod4+x.
Strictly speaking, this is a change of behavior, but it breaks none of our
tests, and using equality instead of subset comparison enables more use-cases.
fixes#2002
Michael Stapelberg