Empty lines in compiler generated methods

[daveMueller]

Partially closes #799

[daveMueller]

Moved the code to skip nested ranges from Coverage (#1006) to the Instrumenter. Thus I could
omit the empty line from the Document. -> see latest comments

[daveMueller]

/azp run

[azure-pipelines]

Commenter does not have sufficient privileges for PR 1023 in repo coverlet-coverage/coverlet

[daveMueller]

Somebody has an idea what went wrong on the macOS Debug build with the integration test Msbuild_SourceLink?

[MarcoRossignoli]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[daveMueller]

Mhh OK this is strange. I just disabled 2 integration tests and enabled them back again and now all tests passed. I also don't have any problems when executing the integration tests locally. I guess we should discuss this when someone finds time to review this. Maybe it has something to do with serialization/deserialization speed because I extended the InstrumenterResult by a new DataMember.

[MarcoRossignoli]

@daveMueller I need to deeply check...that new prop serialized could be very big, have you compared before and after size to find out some sort of ratio?
Can we avoid to persist it?
I mean if you use it during instrumentation to skip some sequences and after we reload it for other reason(nested accounting) should work, right?
Pay attention in tests, when you debug inprocess you need to clean that list https://github.com/coverlet-coverage/coverlet/blob/master/test/coverlet.core.tests/Coverage/InstrumenterHelper.cs#L120 the prepared result is in memory so we need to simulate the serialization(on file) and deserialization(from disk).

[daveMueller]

@MarcoRossignoli I now thought of a better solution. I always assumed the documents list is immutable in Coverage and that's why I made the changes in the Instrumenter. Just to answer your questions real quick.

1. No changes regarding serialization anymore
2. Avoided to persist it because everything is done in Coverage
3. Yes this was my intention to reuse the data after reloading but this also is obsolete now.

Also thanks for tip with the inprocess debugging.

[MarcoRossignoli]

@daveMueller is this PR still needed?

[daveMueller]

@MarcoRossignoli I will check this again and get back to you.

[daveMueller]

I converted this to draft. I look into it again when I have some more free time.