Read/write ops for this implementation are done through the java layer via jni, and so for good performance, it's key to avoid numerous repeated small read/write ops due the jni overhead.
The alternative is to allocate a (conversatively-sized) large buffer to reduce the number of read/write ops over the jni boundary.
This is a backport from 4.0 to 3.x that adds 'sort' and 'has' methods
to the following types:
- PoolByteArray
- PoolIntArray
- PoolRealArray
- PoolStringArray
- PoolVector2Array
- PoolVector3Array
- PoolColorArray
For all the types above, the methods 'sort' and 'has' have been
exposed to the GDNative API (v. 1.3)
Since the method 'has' was already implemented in GDScript before,
in this commit it has been only exposed to GDNative API.
The classes documentation is updated.
The method 'sort' uses the exisging class "Sorter".
Pooled arrays in 4.0 are rewritten, that's why this backport is not
completely indentical to the original PR made for 4.0 (see #32144).
Otherwise we would crash if something prints an error before init or
after destruction of the `OS` singleton which handles printing/logging.
(cherry picked from commit c83084fccb)
This was done by refactoring directory and file access handling for the Android platform so that any general filesystem access type go through the Android layer.
This allows us to validate whether the access is unrestricted, or whether it falls under scoped storage and thus act appropriately.
Having a mix of settings with and without subcategory makes the 'Editor'
section stand out with a weird UX, as instead of simply being a foldable
section like the others, it also holds its own top-level settings and is
therefore selectable.
This wasn't the case in 3.4, and is fixed in 4.0 by refactoring, so for
3.5 we should preserve the 3.4 UX, even if it's not the best.
The methods disconnect_points and are_points_connected now have an optional p_bidirectional parameter matching AStar.
(cherry picked from commit 3590cdfd34)
Backports features and bugfixes from current Godot 4.0 to 3.5 and brings functions and codebase of both version largely in sync to make tutorials more compatible and future backports easier.
Some compilers (notably MSVC) were using signed values for bitfield enums. This was causing problems where 2 bits were used to store 4 or less enum values, where they were being treated as negative numbers.
This PR explicitly requests these enums to be treated as unsigned values.
* Replace case-by-case extraction with PNAME & GNAME
* Fix group handling when group hint begins with property name
* Exclude properties that are PROPERTY_USAGE_NOEDITOR
The rationale behind the expected number of bytes is now displayed
in clear (width, height, format, number of mipmaps expected if any).
(cherry picked from commit 18e9e6a81a)
