eugene yokota

> Anyway, I think, as this isn't a SBT bug this ticket can be closed. But I'm not sure weather it would make sense to open a follow up ticket...

For local testing using scripted publishLocal I don't think can error out. So the current status quo I'd say is suboptimal, at least with regards to Metals. otoh I don't...

The relevant code is here - https://github.com/eed3si9n/jarjar-abrams/blob/develop/jarjar/src/main/java/com/eed3si9n/jarjar/ResourceProcessor.java#L61-L63.

I don't understand what's being asked here. Could you follow the https://github.com/sbt/sbt-assembly/blob/develop/CONTRIBUTING.md#how-to-file-a-bug-report format plz?

> This version of C:\Users\HarryXi\AppData\Local\Coursier\data\bin\.sbt.aux.exe ... So when Coursier says it's installing sbt, is it still installing some shim of sbt instead of the genuine sbt?

Yea, sorry about that. That was more of an observation (I'm a maintainer of sbt) that Coursier does that. `bin.sbt.aux.exe` isn't something that we (sbt project) ship.

Use scoop on Windows? - https://www.scala-sbt.org/download

We could load into the memory smaller chunks at a time, and hash the resulting `Vector[Long]`. If there's a 1GB _resource_ file, I don't know if it's worth content hashing...

For text files it's easy to have the same size just by fixing some typo. Also time is almost meaningless when people start sharing Zinc state across CI jobs or...

Thanks for the report. Could you share more details on the environment you're seeing this? Distro and version number etc.