Krzysztof Romanowski

It took me over 2 years to finally create full version. My Bad.

I run few tests and is seems that that is only the case where we are dealing with classes that comes from packages (package objects or plain objects are handled...

I think scope of dependency (class based name hashing is all about) does not matter since it is also present if we have only one class per source file (and...

I fear root of problem may be the same since every class has a 'wildcard import' from its package(s) (but I need to confirm that).

Problem manifested in another form (please correct me if I am not right): https://github.com/romanowski/zinc/commit/82d660cf2f245da400ab68a145b1f3ed21f83046 No wildcard used here :) Do want me to PR this test?

> always auto-imported in sources defined in the same package name Not in this case. The problem here is that we've got unused import on type `foo,Foo` and later on...

I think we are will be forced to hash packages entries. I think we can do it without much overhead (for sure there will be some). Anyway we need to...

PR created: https://github.com/sbt/zinc/pull/289

I don't know if we need another issues for this but we should also consider at least a hook (or any other way) to capture changes done by compiler plugins...

I've set up something similar and we've get a tons of differences (I've compared crc in jar entries). After further examination it turns out that we've got a lot of...