+ bases = self._common.bases
+ if callable(bases):
+ # This happens for Rust exported object, and we don't know
+ # at the moment how to make an equivalent of property for them
+ bases = bases()
+

• 1. The presence of nullrev will confuse heads(). So filter it out. return set(self._repo.revs('heads(%ld)',
• self._common.bases - {nullrev}))

+ (r for r in bases if r != nullrev)))

@lothiraldan Any idea on how to make this less ad-hoc?

If we'll never need bases of set type, maybe we can add baseslist() to both
Python and Rust implementations.

@yuja I talked with @lothiraldan about this today. It turns out that it'd be better to provide a basesheads returning a set directly from the MissingAncestors object (I do have a Rust implementation for heads).

This would make the call from setdiscovery uniform, and our measurements show that this would be a significant performance gain in the Rust version.
Preliminary measurements indeed show a performance increase with this approach over x30 within a perfdiscovery on the PyPy repo for this commonheads() method, which was abnormaly slow in the version you commented upon and ate most of the Rust boost.
It brings the Rust overall improvement for that discovery at 28% faster.