Only assign known good values of the index to "indexLast" as otherwise
we could end up returning an invalid index value from it if "indexBest"
was never set and the last index wasn't a valid one, as it happened in
wxBitmapBundleImplSet containing an original bitmap and a x2 scaled copy
of it, which shouldn't be used for further scaling.
Add a unit test checking that this bug is fixed now.
Always rounding non-integer scale when upscaling is wrong, as we could
be able to upscale an existing x1.5 image to e.g. 300% DPI scaling with
relatively good results, so only do it if there is no available scale
that exactly divides the requested one.
Add a previously failing unit test which passes now.
It's not necessarily the largest bitmap, but one which may be scaled
using an integer factor.
Fix the tests to expect the correct results, now that they actually
pass.
Check not only that the resulting bitmap has the right size, but also
that it was obtained by scaling the correct bitmap.
Define the required machinery to allow using CATCH CHECK_THAT() macro to
succinctly verify both.
Note that this doesn't change the existing tests, but just rewrites them
in a new form and also adds more tests for a bundle with 3 bitmaps.
Doing anything else results in really poor results, so don't even try
scaling images by 2.25, for example -- scaling by 2 looks much better
and the size difference is relatively insignificant and definitely not
worth the difference in quality.
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.
Don't always return success from this function, NanoSVG just skips
everything until the start of the XML prologue and doesn't return an
error even if it doesn't find it at all, so check that it could parse at
least something to avoid returning a "valid" bundle not containing
anything at all.
Add a unit test checking that we actually can't create an SVG from a
.bmp file (which is something that "worked" before).
The new name, recently introduced in 94716fd801 (Add
wxBitmap::CreateWithLogicalSize(), 2022-01-22), was perhaps more clear,
but also misleading and confusing because the postcondition
CreateWithLogicalSize(size, 2).GetLogicalSize() == size
was not satisfied under MSW, so rename it once again, and hopefully
finally, because the new name is consistent with GetDIPSize() returning
the same size.
Also try to improve the documentation a bit more.
Unlike under Mac (see previous commit), this never worked correctly in
wxMSW at all, only SetScaleFactor() could be used to change the scale
factor of the bitmaps there.
Fix this and make CreateWithLogicalSize(..., scale) result in
GetScaleFactor() returning the same scale for the resulting bitmap, as
expected, under MSW too.
Also add a unit test verifying that this holds.
Optimization of 2ae80673ff (Avoid unnecessarily unsharing bitmaps in
wxMSW, 2022-02-02) broke setting the scale factor because we didn't
update the right bitmap data in case bitmap hadn't had exclusive
ownership of its data before. Fix this by not using a reference which
may refer to another object after AllocExclusive() call.
Also change the other ports to unshared the bitmap when modifying its
scale factor so that the new unit test passes for them too.
This improves the changes of f78db92462 (Avoid bitmap scaling in
wxArtProvider::GetBitmapBundle(), 2021-12-17) and still uses a custom
bundle to avoid scaling the bitmap if possible, but does it in
GetBitmapBundle() itself rather than CreateBitmapBundle().
This allows to also use CreateBitmapBundle() from GetBitmap(), as there
is no possibility of infinite recursion due to calling each of these
functions from the other one any more, and so allows defining art
providers overriding only CreateBitmapBundle() instead of having to
always override both it and CreateBitmap().
Also add a unit test, even if just a trivial one, for these functions,
to at least check that they don't crash.
We often need the logical bitmap size when using it in size computations
involving window size, so add a function returning it directly to
wxBitmapBundle, similarly to wxBitmap::GetLogicalSize(), to avoid using
FromPhys() everywhere.
Also rename the existing wxBitmapBundle::GetPreferredSizeFor() to
GetPreferredBitmapSizeFor() to make it more clear that this is similar
to wxBitmap::GetSize() and so returns the size in physical units.
Closes#22056.
Replace old functions with the new ones in the library code itself.
Note that wxSTC and wxRichText still use GetScaledXXX(), but they're
different functions that might need to be renamed/dealt with separately.
Now wxMSW also stores the scale factor, even if it doesn't use it in its
GetScaledXXX(), so it doesn't seem useful to have this symbol for
distinguishing the platforms with and without bitmap scale factor
support, when we can just use wxHAS_DPI_INDEPENDENT_PIXELS instead in
the only place where this was used.
And as this symbol was added quite recently, in 2c1f4c002d (Add
wxBitmap::SetScaleFactor(), 2021-10-23), we can hopefully just remove it
without breaking any existing code, if we do it right now.
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.
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).
