Make allheaders unit test also enable -Wall, -Wextra and -Wpedantic for
clang to detect any warnings it might give in wx headers.
Don't use -Weverything because there are just too many warnings we'd
need to disable then.
It seems that the default value of NtfsDisable8dot3NameCreation registry
value is 2, and not 0, since a very long time (at least Windows 7), so
the wxPATH_NORM_LONG test was never actually executed.
Change the check guarding it to check if the value of the key is
different from 1 instead of checking that it is 0, as the test can still
succeed on the systems where this is the case.
Of course, it can also fail, if short names are disabled for the current
volume, but let's see if this is a problem in practice before doing
something more complicated to deal with this case.
This file already exists in the test directory, as it's used by another
test, so just use it instead of relying on the existing of mkinstalldirs
in the parent directory, which is not guaranteed as CMake build runs the
tests from a different directory, which is not the immediate child of
the top source directory.
No real changes, just use a more clear and longer (which will be
important for the next commit) name for a test file.
Remove the non-existing samples/console/testdata.gc from make_dist.mk
(there are other non-existing files still referenced there).
Using this ctor is more convenient than using the default ctor and then
calling SetMin() and SetMax().
Document the new ctor and add tests showing that minimum and maximum
values are actually respected.
Closes https://github.com/wxWidgets/wxWidgets/pull/2610
This file was listed in test.bkl but wasn't actually copied because the
value of <files> was overwritten by another <files> just below ever
since the changes of 21fe35aff7 (add wxImage test units to test loading
from both seekable and non-seekable streams, 2009-06-01).
Fix this by splitting this wx-data into 2 different ones, to avoid this
conflict.
This is shorter and more clear than comparing width and height
independently and results in just as informative messages in case of
test failure if asserthelper.h is included.
- Fix handling of delta encoding that change vertical position
- Fix handling of end-of-line encodings that are given mid-line
- Removed unnecessary computation for linepos
Add regression tests for the bitmaps using RLE and loading which
previously didn't work correctly.
Closes https://github.com/wxWidgets/wxWidgets/pull/2590Closes#19318.
This should make the explicit casts such as added in 8db55e9653
(Explicitly cast enum to int inside CHECK(), 2019-09-30) and the
previous commit unnecessary, result in better messages if a test fails
and allow the other tests to still run after the failing tests, as we
use CHECK() instead of CPPUNIT_ASSERT(), which maps to REQUIRE(), now.
Do the same thing for the just added tests as was already done in
8db55e9653 (Explicitly cast enum to int inside CHECK(), 2019-09-30) for
the other ones.
These accelerators are not shown in wxMenuItem label, but still will
work.
Implement support for them in all major ports and XRC.
Co-Authored-By: Vadim Zeitlin <vadim@wxwidgets.org>
It may not be obvious how to build the tests, so add a section
explaining this to the existing file containing the instructions about
writing the tests and add a README to the tests directory itself to
increase chances that this file will be found -- hopefully anybody
interested in the tests will look at the README in the directory
containing them.
Stop throwing std::bad_alloc when BMP has large/negative biClrUsed by
checking that biClrUsed has a reasonable value before attempting to
allocate however much memory it indicates.
Add unit tests showing the loading such invalid bitmaps now correctly
returns an error rather than throwing an exception.
Closes https://github.com/wxWidgets/wxWidgets/pull/2583Closes#19295.
Edge does not support the about: URL using it in the test resulted
in various problems. Just use about:blank for now as the edge backend
does not merge them for the history checks.
These functions used to work in any mode until the changes of 02509cbc39
(Refactor wxGridSelection to store selection as blocks only, 2020-03-03)
and should continue to do so, if only for compatibility.
The changes of b20552116c (Allow wxBitmapBundle to specify its preferred
bitmap size, 2021-10-19) resulted in never rescaling the bitmaps in
standard size in high DPI at all, which isn't the right thing to do: by
default, i.e. if just a single bitmap is specified, we should scale it
up as necessary in order to show the UI elements in the correct sizes.
Use the scaled size, different from the default size, when constructing
wxBitmapBundle from an existing wxBitmap to keep the existing code using
scaled bitmaps working.
Add a unit test checking that this now works as expected under the
platforms where scale factor is used.
This function was defined in wxGTK and wxOSX, but not in wxMSW or the
other ports, but it can be useful there too, so make it public and
define it in common code.
Using bitmaps of preferred size avoids scaling and results in much
better appearance, so add methods allowing querying the bundle about the
bitmaps it supports and implement them in the various implementations.
This is not actually used anywhere yet, but will be soon.
Passing non-const data is inconvenient and error-prone, as data can't be
used again after it was modified by Nano SVG, which resulted in the
button using SVG bitmap not working any longer in the toolbar sample
after recreating the toolbar.
So make it easier to do the right thing, while still keeping the
overload taking non-const data for the situations when avoiding an extra
copy is worth it.
Add nanosvg submodule and use it in the generic implementation of this
function.
This is incomplete yet and, notably, doesn't cache the rasterized
images, but already shows that using SVG images works (at least in the
toolbar sample).
On macOS >= 11 wxListBox has a new layout because underlying
NSTableView has a new style with padding so we need to move
the point to be tested to another position.
Check that Find() of a list declared using WX_DECLARE_LIST_3 (as
wxWindowList is) can be passed a pointer to the base class object and
still compiles and works correctly.
Don't use GetVolumeSeparator() to combine UNC or GUID volumes with
paths, this doesn't work because this kind of paths doesn't contain
colons at all.
Update the documentation to mention this.
This comes at the price of breaking compatibility and returning
"\\share" rather than just "share" from wxFileName::GetVolume() for the
UNC paths. This breakage seems justified because it is required in order
to allow application code to distinguish between paths "x:\foo" and
"\\x\foo", which was previously impossible as GetVolume() returned just
"x" in both cases.
Document this change, adjust the existing checks for the new GetVolume()
semantics and add a new test which passes now, but didn't pass before.
Closes#19255.
This commit is best viewed ignoring whitespace-only changes.
Creating a directory with the leading path separator unexpectedly
created it under the current directory rather than in the root of the
current drive under MSW, due to the path being considered relative, in
spite of starting with the path separator, because it didn't have the
volume.
Fix this by avoiding the use of IsAbsolute() in Mkdir() and checking for
m_relative instead, as it seems the safest possible fix for this bug
because changing IsAbsolute() to return true in this case might change
the behaviour of the existing code.
Closes#4470.
Extend the existing workaround to work not only with explicitly
wxPATH_DOS paths, but with paths implicitly using the DOS format due to
it being the default format for the current platform.
Closes#19261.
