luax_pushtype keeps the object in a weak table and checks the table before creating a new userdata, so it re-uses the object's existing wrapper userdata if it can, whenever the object is pushed to Lua.
This fixes some subtle issues with love objects, where using them as keys in tables would not always work as expected (https://love2d.org/forums/viewtopic.php?f=4&t=39398&p=112388#p112415). It also decreases the amount of new Lua objects created, saving the garbage collector some work.
- added a new CompressedData type to love.image
- added love.image.isCompressed(file or data)
- love.graphics.newImage automatically tests for compression when loading a file
- added love.graphics.isSupported strings for DXT, BC5, and BC7
Compressed textures meant to be used on the GPU offer huge performance gains when
- loading the texture from a file
- loading the texture from RAM to VRAM
- loading mipmaps
- rendering
--HG--
branch : image-CompressedData
Users (or lovers) do not get to choose which implementation of
a module they would like to use.
The abstraction is for dealing with portability issues, not
for giving people the ability to choose an alternative
(inferior) backend. The best choice for any platform/situation
should at all times be the *default* for that platform/situation.
If this is not the case; it's a bug (that is easily fixed).
Any module should provide the functionality promised in the docs.
The lover (or user) should not care how, and if [s]he does, [s]he
does still not get to choose.
Also, a related note: duplication of wrapper code is something I
especially wanted to avoid when I first (or should I say ...
"eventually") formed the current architecture.
--HG--
branch : minor
Alex Szpakowski
Bart van Strien
rude
Boolsheet
Bill Meltsner
Bart van Strien
bill@Ixion