Paul Moore

> it actually makes more sense to fix this as part of a broader transition +1 IMO changes in this area will be sufficiently disruptive (they directly affect user's experience...

> but they forgot to remove the fsync call You've asserted this a few times now, but what's your proof? You said yourself that this is your "best guess". But...

Not at all. It is recorded as a potential fundable piece of work because it's complex, and it seems likely that the only way of getting sufficient effort made on...

The question of how this affects memory usage has still not been addressed. Also, how does this affect performance/memory use of *small* installs? For instance, does the large cache slow...

I agree with @ichard26 here - @notatallshaw you added this to the 25.1 milestone, do you plan on merging it? If not, can you remove the milestone and it can...

Thanks, I hadn't spotted that when I did a search.

In the worst case, why can you not simply set *every* field to be dynamic, except for `Name`, `Version` and `Metadata-Version`? The only complexity for those fields[^1] is with versions...

I'm assuming your concern here is with the statement in PEP 643 that says > Backends MUST NOT mark a field as Dynamic if they can determine that it was...

Sounds good to me. Are plugins permitted to mess with data read from `pyproject.toml`? If so, you might have to weaken (3) to say "only if no plugins are present"....

Cool. I think your response to the plugin author in the case you linked to was exactly right (and to be fair, the author seemed happy with that response). Ideally,...