Merge release commits to 33.x

PiperOrigin-RevId: 823185820

…olves

#23306

Since there's a measurable performance cost (~12% for a long varint heavy message) on little and mid size cores, these instructions aren't unconditionally emitted even though they're compatible (CPUs lacking FEAT_BTI interpret them as HINT/NOP). Even with these extra instructions, the assembly path is still much faster than the generic one.

PiperOrigin-RevId: 834093313

PiperOrigin-RevId: 835386232

This was always intended to be possible, but was accidentally broken by a recent improvement in C++.  This expands the generated defaults to all features, adding an implicit requirement that the message type of each feature extension is unique.  The only violation of this today was in our test features.

PiperOrigin-RevId: 836350099

85507b9 introduced a new call to Marshal.Sizeof(Type). This call is not AOT compatible, which is a regression as prior versions were AOT compatible, even if not directly supported.

This conditionally uses the generic version of Marshal.SizeOf, which is AOT compatible to fix the issue.

Closes #21824

Closes #21964

COPYBARA_INTEGRATE_REVIEW=#21964 from ThadHouse:genericsizeof a12294e
PiperOrigin-RevId: 818694739

Co-authored-by: Thad House <thadhouse1@gmail.com>

Backport fixes for 33.2

PiperOrigin-RevId: 830474777

Add EDITION_UNSTABLE for new edition development