GTK needs to be told the data on the clipboard can be stored
before the store is attempted.
Also, document that the clipboard behavior on GTK is very
dependent on the configuration of the user's machine.
Closes#10515
No real changes, just move m_textLines and m_heightLine from wxTipWindow
to wxTipWindowView as they were only used by the latter and so can just
as well be its members.
Also use this as an opportunity to replace wxArrayString with wxVector.
It doesn't seem to make sense to provide wxFrame-based implementation of
wxTipWindow, if anything we should use similar code to implement
wxPopupTransientWindow itself on the platforms not providing it.
So simplify wxTipWindow code by leaving only a single implementation,
using wxPopupTransientWindow. This should have arguably been done back
in 8962e1d938 (wxTipWindow is now a wxPopupTransientWindow instead of a
wxFrame., 2001-11-30) which introduced the use of wxPopupTransientWindow
in this code.
Allow wxPopupTransientWindow to work without wxPU_CONTAINS_CONTROLS too
in wxMSW and don't use wxPU_CONTAINS_CONTROLS by default, notably
allowing wxTipWindow to be shown without stealing focus.
See https://github.com/wxWidgets/wxWidgets/pull/1942Closes#18636.
The iterator passed to erase() was off by 1, but this worked correctly
in the default build using wxVector, and not std::vector, because its
base() implementation was off by 1 too.
Now that wxVector bug is fixed (see the parent) commit, fix the code
using it too, which makes it work both in the default and STL builds.
Closes#18765.
This is embarrassing, but the iterator returned by this method seems to
have always been wrong, ever since it was added back in 946954d3bf
(Added reverse iterator to wxVector<T>, 2008-09-16). Moreover, it was
also broken in its const_reverse_iterator counterpart where it was
copy-and-pasted in f7ef20685f (Add wxVector<>::const_reverse_iterator,
2013-05-08).
Finally fix this to return the correct value and add at least a simple
unit test check that this method behaves as expected.
This is consistent with wxToString(wxColour) and seems more useful than
either returning black string representation (as wxMSW used to do) or
asserting (as wxGTK did).
Document this behaviour and add a test checking for it.
Closes#18623.
Also define wxColour-specific matchers to allow comparing RGB(A)
channels to the expected values, replacing the old ASSERT_EQUAL_RGB(A)
macros.
No real changes.
Fix setting focus in GUI tests using wxUIActionSimulator with GTK 2.
Fixing the actual problem allows to re-enable activating the TLW in
SetFocus() (done in d06e97e8d9 (Make sure toplevel is active in
SetFocus(), 2020-07-08), see #18783) and remove the now unnecessary
workarounds in the tests too.
See https://github.com/wxWidgets/wxWidgets/pull/1945
We probably could drop support for some gcc 4.x versions too, but we
definitely don't need to support gcc 3.x any longer and not doing it
simplifies the code a bit, so just assume gcc >= 4.
Closes https://github.com/wxWidgets/wxWidgets/pull/1943
It wasn't used since f8252483ec (Applied patch [ 803473 ] wxListCtrl
header height bugfix, 2003-09-11) but remained, commented out, in the
code for some reason.
This test was apparently supposed to check if the mouse was inside the
header, vertically, but this should always be the case unless the mouse
is captured and when it is captured in this code, m_isDragging is set to
true meaning that the code in the "else" branch of the test for it can't
be executed at all, so checking the vertical position seems completely
unnecessary.
Worse, it was actively harmful when using GTK 2 with DPI scaling, as in
this case the height of the window could be different from 22 (e.g. 44
for 2x scaling).
Closes#18713.
Don't duplicate practically the same code in all ports, just add m_paths
itself to the base class. The only drawback of doing this is that it's
unused in the ports not (yet) using it, but this saves enough code in
the aggregate to be worth it.
This is confusingly similar to std::string::empty() which doesn't do the
same thing, so prefer using clear() instead.
And simply remove Empty() calls before the assignment, as they're
useless.
No real changes.
OSX 10.11+ doesn't actually display the title, so update documentation
to reference SetMessage instead. For pre-10.11 override the SetTitle
method to set the title of the NSOpenPanel instead of the window.
Also change so the directory is not updated unless it is provided.
Closes#15143.
Add wxDirDialog::GetPaths() similar to the existing member of
wxFileDialog with the same name and also taking, for consistency,
wxArrayString as the output parameter.
Popups not using this style work too now (since the parent commit), so
this style should be used only if the popup actually requires focus and
not always, unconditionally.
Turn on this style explicitly in the sample which shows a popup window
with wxTextCtrl inside it (which requires focus to work).
Popups not using wxPU_CONTAINS_CONTROLS were not automatically dismissed
at all any longer, as the only auto-dismissal mechanism related on
getting WM_ACTIVATE, which they never did, so implement a different
logic for dismissing them: do it on any change of focus and also any
mouse press (but not move and not key press neither).
This will allow not using wxPU_CONTAINS_CONTROLS for popups that don't
need focus, but still must disappear on their own.
Under MSW only popup windows with wxPU_CONTAINS_CONTROLS can have focus,
attempting to set it to a [child of a] popup without it will just result
in a debug error message from wxWindow::SetFocus() and nothing else, so
just avoid doing it entirely.
At least under MSW this results in an activation loss for the previously
active TLW, as it activates wxTipWindow itself, which looks bad. And, of
course, setting focus is completely unnecessary in the first place, as
this window doesn't accept any input and wxEVT_KILL_FOCUS is handled
only when not using wxPopupWindow.
The window should be notified about its dismissal, as it happens in the
result of the user action and not due to a call to Dismiss() from the
program itself.
Using "pushd tests && test-command && popd" loses the exit code of the
test command, as it's not taken into account by "set -e" when the
command is part of a "&&" list.
Fix this by simply splitting such lists in their individual commands.
We need to set the new owner for the TLW, instead of using the new
parent as the actual parent, in the MSW sense, as this results in a
weird situation in which the TLW becomes a child (i.e. non-TLW) window.
Closes#18785.
Limit the changes of d06e97e8d9 (Make sure toplevel is active in
SetFocus(), 2020-07-08) to GTK 3 only as they break GTK 2 GUI test
suite, resulting in many CI failures.
See #18783.
