Most PRs change code where no tests have been written so far and where
adding tests is difficult. Codecov fails all those PRs.
These values should be adjusted once we have better integrated testing
into our workflow and have written better testing infrastructure for
modules and very tightly coupled components
The previous approach relied on manually updating the version and date
using a Makefile
This approach dynamically gathers the correct version from git and uses
the date of the latest commit in the branch
* Clean up CMake logic
- removed logic to find CppUnit (no longer used)
- removed "dirs" variable used to pass include directories
- removed add_library function (no longer used)
- removed make_executable function
* only used in 2 places (polybar and polybar-msg)
* it was more general than needed, logic is simpler without it
- split polybar into static library and executable
* this allows linking unit tests to the library
* rename library
* add coverage build
- Added a CMake build type "Coverage" that builds C and C++
code with the "--coverage" flag (recognized by both GCC and Clang)
- removed "-Wno-missing-field-initializers" from test flags,
since it didn't seem to be needed any more
- removed logic from tests/CMakeLists to disable "-Werror" and "-pedantic-errors"
since there didn't seem to be any warnings during the build
* fix whitespace
* update travis
* remove O2 from defalt flags
* allow tests to be built by default make target
* disable Werror for debug builds
If FcNameParse returns NULL FcDefaultSubstitute segfaults
We log and throw an error here because the signal emitter swallows all
exceptions so if that exception was thrown, the bar would stop rendering
without any indication as to why.
Fixes#1435
wireless_tools 29 redefines inline in iwlib.h as:
#define inline inline __attribute__((always_inline))
which conflicts with POLYBAR_NS, which is defined as:
#define POLYBAR_NS \
namespace polybar { \
inline namespace APP_VERSION_NAMESPACE {
In version 30.pre9 this #define is moved into a source file and thus
cannot conflict.
The error only occurs when building with clang, so it seems gcc and
clang handle this differently
Fixes#1492
A message is emitted and there the '--force' is still honored. However,
if people want to reuse the build-directory to reduce recompile-times,
this give them the rope they ask for.
If a tailed command is used polybar would generate two action tags, one
with %counter% replaced and one with %pid% replaced, but never both
This is a bug that was introduced in #934
The idea is that pseudo-transparency should behave like real transparency as
much as possible. To achieve this, we now render the bar the same way in both
cases. The only part where pseudo-transparency differs is at the very end, where
the rendered bar is captured and composited against the desktop background
image. This should ensure that both modes behave the same.
This reverts some behaviour differences introduced by the pseudo-transparency
implementation. The new implementation is much closer to the
non-pseudo-transparent case and thus keeps original behaviour.
For the new method we simply fill the bar with the background image fetched from
the root window if in pseudo-transparent mode, otherwise we do nothing. This
means that everything will work as in the fully-transparent mode.
We need to use positions relative to the position of the bar for indexing into
the background image slice, but the code used absolute ones.
This worked fine as long as absolute positions are the same as relative
positions (this is the case for a bar located at (0,0), so if bottom = false).
But for bottom bars (where the bar position is not (0,0)) this was wrong which
caused the tray background to be black (out of bounds for the background slice).
The systray only supports pseudo transparency (real transparency would require
much larger changes) so the real transparency should only be used for the bar itself.
This option is no longer necessary because the tray background color can now
simply be set to any (semi-)transparent color (just like the bar background).
We now take the bar position that the window manager gives us instead of trying
to calculate it ourselves. This is more correct when multiple bars are attached
to the same edge, as the window manager may move some of them in that
case (assuming override redirect is not enabled)
We need to fetch the outer area from the root window, not just the inner area
because we paint the background below the borders as well.
This has the nice effect of supporting semi-transparency for borders as well.
Now all the tokens in the memory module also have ramp and bar counterparts.
These can be used exactly the same as `bar-used` and `ramp-used`, they are named `<bar-swap-used>`, `<bar-swap-free>`, `<ramp-swap-used>`, and `<ramp-swap-free>`
From commit 880896c6f4814f4c7798355a652dc6167be2b75f in googletest cmake
test configuration fails with errors like
CMake Error: install(EXPORT "gtestConfigInternal" ...) includes target
"gtest" which requires target "i3ipc++" that is not in the export set.
The issue was that link_libraries also affected the googletest internal
targets that came after it making the whole thing fail.
This could also have been solved by moving link_libraries after the
configuration of googletest but according to the cmake documentation [1]
target_link_libraries should be prefered over link_libraries anyways.
[1]: https://cmake.org/cmake/help/latest/command/link_libraries.htmlFixes#1393
