The default size of the bitmaps created from bitmap bundles returned by
the Tango art provider for the client strings different from wxART_MENU
and wxART_BUTTON was changed from 24x24 to 16x16 in 766fefeb34 (Use SVG
instead of PNGs in tango art provider, 2022-01-27), see #22067, without
any real need and perhaps even non intentionally at all.
Restore the old default size to restore the old behaviour.
See #22189.
Gnome uses a variant of the "log out" icon for its standard "Exit" icon,
so it should be fine for us to do it too, contrary to what the comment
from back when Tango provider was added in c1d2466a79 (Add wxArtProvider
using Tango icons., 2010-12-31) was saying.
Changing invoking window always needs to update the bitmaps because
the scale factor may need to be changed if this window is on a different
display, using different scaling, so do it automatically whenever
SetInvokingWindow() is called instead of having to remember to call
SetupBitmaps() manually everywhere SetInvokingWindow() is.
This required making the base class function virtual, but allows to get
rid of SetupBitmaps() call in wxWindow::DoPopupMenu(), as it can now
rely on this being already done by wxWindowBase::PopupMenu().
Don't crash if this function is called (which is not the case yet, but
will be after the upcoming commit, when it's going to be implicitly used
when updating the bitmaps for taskbar icon menu), but return the content
scale factor for the main screen, which should be the right thing to do
here.
This currently doesn't work there because the focus rectangle doesn't
fit into the space reserved for it (which is a bug on its own, probably)
and would look very out of place even if it did work because the native
tabs don't use focus rectangle for the selected tab.
Closes#22190.
Since the changes of 7129d2b11c (Accept multiple data formats via
drag-and-drop in wxDataViewCtrl, 2021-08-22) custom data formats were
simply ignored because CreateDataObject() helper didn't do anything for
them.
Fix this by simply continuing to handle them in the same way as before,
i.e. by creating a wxCustomDataObject for the specified format.
Closes#22112.
Set the end iterator to the end of the actually parsed date, instead
of consuming any ultimately unparsed whitespace/delimiters possibly
following the date.
When calling GetDPI() for a window which is not created itself yet and
doesn't have any valid top level parent neither, we have no choice but
to fall back on using the screen DPI, but this doesn't always work
correctly for the systems using multiple monitors with different DPIs,
so warn when this happens -- any situations when it does indicate a
potential bug that should be fixed.
See #22193.
It doesn't make sense to use some hardcoded size for these fields and
we can't use FromDIP() for wxTabFrame which is not a real window (as
it's never created, it doesn't have a valid HWND under MSW, so FromDIP()
can't return the correct result for it), so don't initialize them at
all: they will be set later, before they're actually used, by the code
in wxAuiNotebook using wxTabFrame.
FromDIP() shouldn't be used before the window is created and
m_tabCtrlHeight was already initialized correctly in InitNotebook(),
when it can be used, so simply remove this premature and redundant
initialization.
There is no reason not to allow calling this function with const window,
so do allow it, notably to make it possible to use it with "this" from
const wxWindow methods.
Anything following the dot was mistakenly used as an extension of the
corresponding file and replaced with ".mo", breaking loading of catalogs
with the names like "wxstd-3.1", which is the standard wx catalog name
since ded4da5ce5 (Add version suffix to gettext message catalog files,
2021-02-07).
Closes#22198.
The intent of the implementation clearly is to allow parsing time first,
date second. But this failed, because a time such as "14:30:15" would
successfully parse as a date (as 14th of current month, current year).
Consequently an attempt is made to parse the actual date as time, which
fails, and therefore the whole ParseDateTime() fails.
Adding a failing test case for ensuring times cannot be parsed as dates
does not cause a failure, because partially yet successfully parsed inputs
get silently ignored (in both ParseDate and ParseDateTime tests). Fixing
both of these, too.
Closes#22203.
Fix behaviour of ItemAdded() in generic wxDataViewCtrl implementation
for not yet opened nodes: calling it shouldn't prevent calling the model
GetChildren() when the node is opened, resulting only in explicitly
added children appearing in it.
See #22201.
The RFC accepts 2-digit years, and it makes most sense to interpret
e.g. 95 as 1995. However, this is an incompatible change, as earlier
95 was literally decoded as 95 AD.
Years 00..29 are considered to mean 20xx; 30..99 means 19xx.
Closes#22196.
When using the native fullscreen API by enabling EnableFullScrenView()
allow using hiding (or showing) menu and/or toolbar. An additional style
parameter has been added to EnableFullScrenView() to allow customizing
which style is applied when the user presses the fullscreen button
instead of a call to ShowFullScreen().
Closes#22180.
Allow returning a wxBitmapBundle rather than an individual wxBitmap.
Also make GetLargeIcon() non-pure even in wxOSX, as it now doesn't need
to be overridden if GetIcon() is -- but don't make GetIcon() pure
virtual neither to allow the existing code overriding GetLargeIcon() to
keep working.
This incidentally fixes the icons for the standard pages under macOS
broken by 388d322b68 (carry changes to toolbar over to prefs on osx,
2021-09-28), which replaced return statements with assignments,
resulting in the icon being set to the last value assigned to it instead
of the correct one -- this commit restores the previous control flow in
wxStockPreferencesPage::GetLargeIcon() (now called GetIcon()).
Closes#22187.
Fix wxDataViewMainWindow::ItemAdded() when it used after
wxDataViewCtrl::AssociateModel() (it doesn't create child nodes so we should
not add child items at all to a collapsed node).
wxSplitterWindow can receive multiple size events while the DPI change is being processed.
Keep using the relative sash position while the DPI change is finished.
The implementation implicitly relies, in many places, on the assumption
that the input never ends prematurely. If it does, the iterator
pointing beyond the end of buffer is dereferenced, which is UB.
The solution used here is to append 32 zero bytes to the date string,
which hopefully keeps the code more readable than checking for the end
of string before each deference operation.
Add various syntactically invalid inputs to unit tests.
Closes#22185.
Avoid using, or even mentioning in the comments, __DATE__ and __TIME__
macros as doing this prevents ccache from caching the compilation
results by default (i.e. unless CCACHE_SLOPPINESS=time_macros is used).
As ccache simply scans for the given literal strings, using "##" token
pasting operator is enough to disable this pessimization. Of course,
this does mean that using ccache with the code actually using __TDATE__
or __TTIME__ is not going to work correctly, but there should be no
reason to do it any longer and these macros are not even documented, so
also mention that they shouldn't be used.
Finally do add __DATE__ to the only place in our own code where these
macros are used to ensure that it is not cached incorrectly.
Closes#22156.
