Fabrizio Ferrai

Is this really a problem? I don't think we gain anything in allowing multiple jobs to run concurrently, and we should just limit the amount of running jobs to a...

There are several barrels of worms that would be cracked open by allowing the whole of unicode in these names - no one can type them directly, a series of...

> leaves non-English speakers having to make compromises on package names I'm not a native English speaker - Italian living in Finland, so I get to use all the various...

@hdgarrood to get back to your points: 1. if the package index is a "cache" then we don't have to ensure integrity and availability, which will be best-effort instead. I.e....

@hdgarrood do you have any more input on this, or should I move forward with the plan detailed in the last comment above?

> which is that I think handling metadata fixes with patch versions basically amounts to using yanking rather than revisions That's a good point, in trying to figure out a...

Having implemented a part of this, I am now of the idea that having a separate `revision` field is a great deal of unnecessary complexity, and we should avoid it...

- we'd detail in the spec that this is not strictly SemVer but a variant of it, so if you'd try to use `purescript-versions` and expect correct ordering that would...

@hdgarrood addressing some of your other concerns: - package authors do not need to publish revisions at all since they can just cut a new patch release - we're not...

> requiring people to do a patch version for these changes basically amounts to yanking instead In my understanding yanking is when there's a way to tell the solver "don't...