| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
| |
This fixes an issue where property values that contained markup characters
weren't displayed.
Fixes #8.
|
|
|
|
|
|
|
|
|
| |
commit 3c82a664155695bd4243c2a964b3ec22c544f3f6 added
an idle handler to help process introspection requests
asynchronously. It failed to clear the idle handler id
when the idle finished, however.
This commit fixes that.
|
|
|
|
|
|
|
|
|
|
|
| |
Some projects, like systemd, have truely gigantic introspection trees.
At the moment that makes d-feet lock up.
To address that problem, this commit changes d-feet to process the
incoming introspection data from an idle handler.
https://gitlab.gnome.org/GNOME/d-feet/-/merge_requests/26
|
|
|
|
|
|
|
| |
This will allow to open D-Feet directly to a given address from the
CLI.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
|
|
|
|
| |
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
|
|
|
|
|
|
| |
This will avoid having to switch manually after entering the address.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
|
|
|
|
|
|
|
| |
Reused in the following patch. Also returns whether attempt was
successfull.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Source should be installed by default to the pure location. This is
actually a bug in meson: https://github.com/mesonbuild/meson/pull/6848
This also fixes some ninja install error due to discrepancy between
install_sources() and get_install_dir() locations during post-install
script.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
|
|
|
|
|
| |
The method was added to Meson in 0.50.0 which is the current minimum for
d-feet.
|
|
|
|
|
|
| |
Previously it was run at the point where `meson` was invoked, which
won't catch the case where you introduce a style violation after running
`meson`.
|
|
|
|
|
|
|
|
| |
They are currently timing out in the Flatpak CI, but there's no
indication as to why.
Meson captures the output of tests and hides it if they pass, so it
seems safe to always run them verbosely.
|
|
|
|
|
|
|
|
| |
pycodestyle says:
../../src/dfeet/dbus_utils.py:124:17: E117 over-indented
and it is correct: the indentation is misleading.
|
|
|
|
|
|
|
|
|
| |
meson creates the `d-feet` file from `d-feet.in` input file using
its same file mode bits. However, the input file does not have
the executable mode enabled.
The executable mode bit has been enabled so `d-feet` file is
created with proper modes.
|
|
|
|
|
| |
To avoid the burden of maintaining multiple build systems, this
patch removes autotools support.
|
|
|
|
|
|
| |
meson is a build system focused on speed an ease of use, which
helps speeding up the software development. This patch adds meson
support along autotools.
|
|
|
|
|
|
|
| |
D-Feet's test unit does not specify the required introspection
version.
This adds the required introspection version.
|
|
|
|
|
|
|
|
|
| |
The test unit contains hard coded values for paths. However, these
hard coded variables consider that the source code path and the
build path are the same, something usual in autotools.
These hard coded values has been replaced by build time variables
that take the appropiate values considering different build paths.
|
|
|
|
| |
to match the .desktop file and the updated about menu item label
|
|
|
|
| |
https://gitlab.gnome.org/GNOME/Initiatives/wikis/App-Menu-Retirement
|
|\
| |
| |
| |
| | |
Correctly honour theme foreground colour
See merge request GNOME/d-feet!12
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Rather than try to explicitly look up the colour, don't specify a colour
at all. Argument names are never nested within a <span> applying a
different foreground colour.
This does have one visible effect: with Adwaita (light), selecting the
treeview row causes method argument names to switch from black to white,
just like the method name. I think this is actually more readable (and
the green used for the signature is illegible against the blue
selected-row highlight.)
As discussed in https://gitlab.gnome.org/GNOME/d-feet/merge_requests/4#note_319094 and !8.
|
| |
| |
| |
| |
| |
| |
| | |
This was added to D-Bus almost a decade ago. Without this change it
shows up as 'Error(h)' in method signatures.
(This doesn't mean we support passing them as in-parameters.)
|
| | |
|
|/
|
|
|
| |
It's legal to slice past the end of the string; you get back an empty
string in that case.
|
|
|
|
|
|
|
|
|
| |
Trying to compile this regexp fails with:
re.error: unbalanced parenthesis at position 5
I introduced this bug in 45a1146e1122b376ddac312687e8e27d4e0a8076. Fix
it, and add a small test case that would have caught it.
|
|\
| |
| |
| |
| | |
Fix 'make check', detect 'pycodestyle' (the new name for 'pep8'), use Python 3 in Flatpak version, and set up CI
See merge request GNOME/d-feet!7
|
| |
| |
| |
| |
| |
| |
| | |
../../src/dfeet/settings.py:22:17: W605 invalid escape sequence '\.'
../../src/dfeet/settings.py:23:6: W605 invalid escape sequence '\s'
../../src/dfeet/settings.py:28:10: W605 invalid escape sequence '\s'
../../src/dfeet/settings.py:29:12: W605 invalid escape sequence '\('
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The following code:
self.connection = Gio.DBusConnection.new_for_address_sync(
self.__bus_address,
Gio.DBusConnectionFlags.AUTHENTICATION_CLIENT |
Gio.DBusConnectionFlags.MESSAGE_BUS_CONNECTION,
None, None)
causes newer versions of pycodestyle to complain:
../../src/dfeet/bus_watch.py:139:17: W504 line break after binary operator
However, if I move the the line break to be before the binary operator:
self.connection = Gio.DBusConnection.new_for_address_sync(
self.__bus_address,
Gio.DBusConnectionFlags.AUTHENTICATION_CLIENT
| Gio.DBusConnectionFlags.MESSAGE_BUS_CONNECTION,
None, None)
Older versions complain in the opposite way:
../../src/dfeet/bus_watch.py:140:17: W503 line break before binary operator
The thread around
https://mail.python.org/pipermail/python-ideas/2016-April/039774.html
explains why the tool has changed its mind.
I prefer the old style, so let's just ignore the new warning.
|
| |
| |
| |
| | |
These were all reported by flake8.
|
| |
| |
| |
| | |
These are considered Bad Style, and pycodestyle reports them as such.
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The 'pep8' tool has been renamed upstream to 'pycodestyle'. On Debian,
the Python 3 version is packaged and installed as pycodestyle. On
Fedora, it's packaged as python3-pycodestyle and installed as
pycodestyle-3; the Python 2 version is installed as pycodestyle-2 with a
symlink from pycodestyle. Make configure search a variety of such names,
and use the path it finds.
|
| |
| |
| |
| |
| |
| |
| | |
Since c28acbf326bb79543b285f371bdf068d8831d9a7 the application has used
the Python version detected at configure time, but tests were still
unconditionally run with the system version of Python 2. This made it
impossible to run the tests on systems with only Python 3.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
These were introduced in e6488e6c0fd0416c54ecff1dd682f0ae8aa4c6fb and
e6488e6c0fd0416c54ecff1dd682f0ae8aa4c6fb:
src/dfeet/introspection_helper.py:9:1: E302 expected 2 blank lines, found 1
src/dfeet/introspection_helper.py:14:100: E501 line too long (107 > 99 characters)
src/dfeet/introspection_helper.py:17:1: E302 expected 2 blank lines, found 1
src/dfeet/introspection_helper.py:20:1: E302 expected 2 blank lines, found 1
src/dfeet/introspection_helper.py:23:1: E302 expected 2 blank lines, found 1
|
|/
|
|
|
|
| |
This dialog is not particularly short of space, and it's nice to have
the full (bus name, object path, interface, method name) quad visible in
one place.
|
|
|
|
|
|
| |
See: !4
Spotted-by: Johannes Sasongko
|
|\
| |
| |
| |
| |
| |
| |
| |
| | |
'6-display-statistics-and-match-rules-from-the-new-stats-interface' into 'master'
Display statistics and match rules from the Stats interface
Closes #6
See merge request GNOME/d-feet!3
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This requires dbus 1.9.0 for the new GetAllMatchRules call on the Stats
interface and for compiling the Stats interface by default.
The Stats interface is enabled by default on the session bus. On the system
bus, it is not enabled for non-root users by default but it can be enabled as
shown in /usr/share/doc/dbus/examples/example-system-enable-stats.conf
If the feature is not available in D-Bus, D-Feet ignores the error and
continues as before.
https://bugzilla.gnome.org/show_bug.cgi?id=735167
|
|/
|
|
| |
This makes property names readable on dark themes like adwaita-dark.
|
| |
|
|
|
|
|
| |
This was being set to "d-feet", but the icon is now called
org.gnome.dfeet.
|
|
|
|
|
|
|
|
|
| |
The convention for DBus and Flatpak is to not use hyphens in the last
segment of a name.
In fact, this is even illegal in object paths.
This commit moves everything from org.gnome.d-feet to org.gnome.dfeet
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=758889
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Without this, we see:
Gtk-Message: GtkDialog mapped without a transient parent. This is discouraged.
You currently cannot actually interact with the toplevel application
window while an ExecuteMethodDialog is visible; this change also makes
this fact visible to the user.
https://bugzilla.gnome.org/show_bug.cgi?id=734514
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
If a D-Bus method returns multiple results, d-feet previously only
pretty-printed the first. Print them one per line, so a method with
return signature 'isa{sv}' might end up showing:
123,
"hello, world!",
{}
Signed-off-by: Simon McVittie <smcv@collabora.com>
https://bugzilla.gnome.org/show_bug.cgi?id=784387
|
|
|
|
|
|
|
|
|
| |
With pep8, or rather its successor pycodestyle 2.2.0, the test-suite
fails with:
../../src/tests/tests.py:105:1: E305 expected 2 blank lines after class or function definition, found 1
https://bugzilla.gnome.org/show_bug.cgi?id=774735
|
|
|
|
|
|
|
| |
Correctly test whether a value has been retrieved (not None) instead of
testing whether the value is logically True to decide on showing it.
https://bugzilla.gnome.org/show_bug.cgi?id=783006
|
| |
|
|
|
|
|
|
|
|
|
| |
Since we are completely ignoring Wnck on non-X11, there is no need to
guard against the default Wnck.Screen being None.
This reverts commit c9d202ede2474f4a748be78fac78ba077e958399.
https://bugzilla.gnome.org/show_bug.cgi?id=763615
|
|
|
|
|
|
|
|
|
|
|
| |
Wnck is an X11-only API. While Wnck.Screen.get_default() might have
elegantly returned None in the absence of X11 earlier, we cannot rely
on that behaviour. In fact, the same version of libwnck has started to
crash inside Wnck.Screen.get_default().
Let's not touch Wnck when X11 is absent.
https://bugzilla.gnome.org/show_bug.cgi?id=763615
|