Improve process of updating binary jars in twitter_scrooge tests

[cattibrie]

Problem

Some of the targets in test/src/main/scala/scalarules/test/twitter_scrooge/ depend on scrooge_scala_import rule which takes thrift_jars and scala_jars as input and provides [ThriftInfo, JavaInfo, ScroogeImport].

For exampletest/src/main/scala/scalarules/test/twitter_scrooge:scrooge2_b_imp depends on the thrift3_import target in test/src/main/scala/scalarules/test/twitter_scrooge/thrift/thrift2/thrift3/BUILD.
thrift3_import target as input has thrift3_scrooge.jar and libthrift3.jar. There is no documentation or comment in the code how these jars were created and how to keep them up-to-date.

As a result, changes related to scrooge may break tests and it takes a lot of time to figure out how to fix them.
Example here.

Solution

scrooge_scala_import rule consumes jars as outputs of source targets, or at least have a readme describing how to "fix" the tests for a next Scrooge update.

[ittaiz]

Thanks for this issue!

scrooge_scala_import rule consumes jars as outputs of source targets

I think that would be a problem since this was needed for people who have some of their thrifts outside of bazel (for example migration). WDYT?

have a readme describing how to "fix" the tests for a next Scrooge update.

Agree 100%

[cattibrie]

I think that would be a problem since this was needed for people who have some of their thrifts outside of bazel (for example migration). WDYT?

I thought to update the tests for scrooge_scala_import. For example, create a rule that takes generated jar from thrift sources and passes it to scrooge_scala_import.

[ittaiz]

Got you. Let's give it a try