Sizing the images stored in the list should stick to the convention
adopted in the native wxImageList implemented in wxMSW.
Images stored in the list should have the sizes as it is declared for
the list even if provided bitmaps have different sizes.
In case of discrepancies their dimensions should be adjusted accordingly
(cropped or extended).
If image using bitmap and mask colour is added to nonmasked wxImageList we
need to convert effective mask to alpha channel values prior to adding the
image to the native list to preserve transparency.
Although this was supposed to work, specifying "%l" in wxLog time stamp
format actually didn't because wxLog timestamps were stored as seconds.
Fix this by storing them as milliseconds and add a simple test (not
executed by default) showing that "%l" works correctly now.
Still keep the old wxLogRecordInfo::timestamp field for compatibility,
as it was documented.
See #13059.
Don't use the current locale format which doesn't provide all the
information (e.g. never includes milliseconds) and explicitly set the
format providing it instead when showing wxLogTrace() messages.
Bitmaps stored in the list as a result of replacing existing ones should
conform to the same constraints as bitmaps directly added to the list.
These constraints are applied in the shared GetImageListBitmap() function
called both on adding and replacing the images.
Check that a row's height doesn't change when auto-sizing a column with
an auto-wrapping cell containing newlines.
Also currently will cause an infinite loop which will be fixed in the
next commit.
See #15943.
This convenient method allows to expand all children of the item
recursively.
This is directly supported by both native implementations, so it makes
sense to have this in wxDataViewCtrl itself, even if it could be also
(less efficiently) implemented in the user code.
Failing it is not useful, this results in the CI failure when in 99% of
the cases the transient network problem which resulted in this test
failure has nothing to do with the changes being tested.
Don't loop indefinitely until we run out of memory, possibly after
wrapping around INT_MAX, but impose an arbitrary limit of 128MiB for the
max allocation done by wxString::PrintfV() when the provided format
string or one of the arguments are invalid.
This notably fixes a crash when trying to use "%c" to output an invalid
Unicode character.
Also improve comment explaining DoStringPrintfV() logic and change the
size type to size_t from int.
Co-Authored-By: Arrigo Marchiori <ardovm@yahoo.it>
This makes it more compatible with the standard behaviour of vswprintf()
and allows to use the same logic in the builds using our version of this
function and the normal ones in DoStringPrintfV(), simplifying its
(already rather hairy) logic.
Update the tests to not require any particular return value in case of
buffer overflow, as this differs between Unicode and non-Unicode builds.
When we finally drop the latter, we should just check that it always
returns -1 in this case.
Note that ideal would be to return the actually needed size of the
buffer in case of the error due to buffer being too small, but this
isn't that simple to do and it's probably not worth spending time on
improving this code as long as we still need to use the buffer doubling
strategy in DoStringPrintfV() when using the standard vswprintf().
Reduce duplication and simplify the test code by using the same macro
for any number of printf() arguments.
Note that this relies on "##__VA_ARGS__" gcc extension to work even for
CMP0(), i.e. when there are no parameters at all. If this ever becomes a
problem, we can always reintroduce a separate CMP_NO_ARGS() macro.
This used to work previously but got broken when using C++17 by
c3810da549 (Allow using noexcept methods with event tables macros,
2020-04-09), as cast added to deal with noexcept handlers relied on
argument type deduction which can't be done for overloaded functions.
Fix this by extracting the event argument type from the function pointer
type and specifying it explicitly, while still letting the compiler
deduce the class.
Add a test case checking that using overloaded event handlers compiles.
See #18721.
Closes#18896.
Handling mask when converting wxBitmap to wxImage was accidentally
broken by 7e9afad53a (Add real support for monochrome bitmaps to wxMSW,
2020-10-19). Fix this now and reuse the new wxMonoPixelData to make the
code simpler and more readable.
This commit is best viewed ignoring whitespace-only changes, as it
unindents a large block of code.
Closes https://github.com/wxWidgets/wxWidgets/pull/2125Closes#18974, #18975.
In the context of wxPlatformInfo, 'architecture' means either '32 bit' or
'64 bit'. Anywhere outside the context of wxPlatformInfo, this concept is
known as 'bitness'.
For the rest of the world, 'architecture' generally refers to the CPU
instruction set architecture, i.e. something like x86, x86_64, arm64, whereas
'operating system architecture' refers to the design of separating kernel
space, user space, etc.
This might compile with even earlier MSVS versions, but it definitely
does with the recent ones, so enable the test for this compiler, even if
it still doesn't define __cplusplus to indicate C++11 conformance by
default.
Replace the overloads added in c2e5f3520a (Add a wxRound() overload for
int, 2020-11-05) and 1cf7c47934 (Add more wxRound() compatibility
overloads and improve docs, 2020-11-05) with a template function that
should work for all integer types.
This fixes compilation of existing code using wxRound() with size_t
values: while this doesn't make any sense, it doesn't make much less
sense than using it with int, so let people avoid having to change their
code when upgrading to wx 3.2.
Also add at least some minimal tests for this function.
Closes https://github.com/wxWidgets/wxWidgets/pull/2119
When doing an RGB->HSV->RGB roundtrip, the original value should be
restored (HSV, being double, has sufficient precision).
For e.g. `RGBValue(1,2,3)`, the equivalent resulting code for blue is
`trunc(int * 255.0 / 255.0)` (cast from double to int truncates).
At least with x87 FP math and its immediate 80bit extended precision
the resulting value is ~trunc(2.9999..), i.e. 2, similar problems may
exist on other architectures with other values.
Using proper rounding avoids the error magnification.
Closes https://github.com/wxWidgets/wxWidgets/pull/2078
In order to be able to use monochrome bitmaps as wxMask, improve support
for them in various ways:
1. Implement loading and saving of monochrome BMP files.
2. Add wxMonoPixelData for direct access to monochrome bitmap pixels.
3. Implement conversion from wxImage to monochrome wxBitmap.
Closes https://github.com/wxWidgets/wxWidgets/pull/2032
All still supported versions of MinGW/Cygwin provide w32api.h file, so
there is no need to test for it.
This allows to simplify the code, but also remove the definition of
HAVE_W32API_H from bake- and makefiles and this, in turn, allows to get
rid of extra flags in MinGW format entirely, as we don't support gcc
2.95 for which they were originally needed neither.
Fix issues found by address/leak sanitizers in the unit tests and add
Travis CI build running tests built with ASAN to ensure they don't
reappear in the future.
See https://github.com/wxWidgets/wxWidgets/pull/2086
The version of glibc used under Ubuntu 20.04 doesn't have "%Z" in the
French date-time locale, which broke the test there.
Instead of adding even more tests for glibc version, just ignore "%Z" if
it's present, we don't really care about it, we just want to check that
wxLocale::GetInfo() returns something recognizably different from C and
English locale here.
Also document the need for using WX_CLEAR_LIST() when converting a
list-valued wxVariant to wxAny.
Note that we intentionally don't fix the problem by clearing the list
automatically, even if it could be done, because this would silently
break the existing code which does already clear the list -- and now
would attempt to clear it twice, with fatal consequences. Instead
document the existing behaviour and explain how to avoid memory leaks.
Remove the unnecessary, and actually harmful, overridden GetProduct(),
as we must delete the pointer returned by the base class version to
avoid leaking it.
Also use wxScopedPtr to avoid leaking a test wxHtmlContainerCell.
