Closed simoncozens closed 3 years ago
It’s easier to get your head around the code if there’s less of it.
Well, it could stay in your own fork - that's why I have my opencompiler branch (and contains things that are not yet accepted by upstream, and some materials probably won't be acceptable to upstream, and I don't have to be too careful about not breaking visual studio builds), and have separate cherry-pick to new branches for pulls which contains "no need to debate" changes...
It would be nice if the rasterizer is also released, even if in a binary-only/plug-in external-add-on modular form.
We would like to fully open source VTT at some point. The rasterizer as binary-only/plug-in external could be a possibility.
FWIW, there is an early idea of a project, https://github.com/HinTak/fscaler/ , for re-creating a windows-binary-interface-compatible of the Microsoft rasterizer backend of FontVal using some wine-bits + FreeType. The current FreeType re-implementation is only source-compatible. The idea is to transition from source-compatible FreeType to windows-binary-compatible wine-bits + FreeType, to allow eventually a Microsoft binary-only rasterizer backend to be used...
I'll personally prefer not to push for this kind of "premature tidying" change though - it could be useful functionality that Microsoft might re-add later, for example, and it interfere with others' work-in-progress: somebody else might actually want to fix warnings etc inside those removed code, even if that section is not used.
If it bothers somebody, a pair of "#ifdef 0" "#endif" is preferable.