[SR-4294] Warn about overly-nested generic type specifiers

[jckarter]

Previous ID	SR-4294
Radar	None
Original Reporter	@jckarter
Type	Bug

Additional Detail from JIRA

Votes	0
Component/s	Compiler
Labels	Bug, DiagnosticsQoI, StarterBug
Assignee	None
Priority	Medium

md5: c1aab4167ab2d03975e533b9a662698f

Issue Description:

https://twitter.com/steipete/status/843896945718546434

The tweet is in C++, but the same problem could affect Swift. A warning about overly-nested generic type specifiers that have enough piled-up closing >>>>>> brackets to be mistaken for a conflict marker in version control might be helpful, not only for tooling interaction, but for code cleanliness, since you should probably consider using a type alias at that point.

[belkadan]

Resetting assignee for all Starter Bugs not modified since 2018.

[swift-ci]

Comment by Minhyuk Kim (JIRA)

I'd like to take on this issue, but I'm not sure where to start with. @jckarter could you please give me few pointers on which files/folders to look?

[jckarter]

A good place to start looking might be in the TypeResolver, in TypeCheckType.cpp. The compiler represents the source-level spelling of a type in a TypeRepr object, and the TypeResolver contains the code that resolves this into a reference to a specific type, raising diagnostics if there are issues such as references to nonexistent types. You could look at where we resolve GenericIdentTypeRepr (which is the subclass for the representation for a generic type reference like Foo<....>) and check whether the TypeRepr for its final argument is also a nested GenericIdentTypeRepr, repeating up to 6 or so levels of nesting, and then check the `AngleBrackets` source range of all of them to see whether the closing brackets line up in a row like >>>>>>.