On near-master:

Unhandled coercion
UNREACHABLE executed at /Volumes/Data/swift-public/swift/lib/Sema/CSApply.cpp:6768!
Stack dump:
0.  Program arguments: /Volumes/Data/swift-public/build/ninja/swift-macosx-x86_64/bin/swift -frontend -c -primary-file - -target x86_64-apple-darwin18.5.0 -enable-objc-interop -color-diagnostics -module-name main -o /var/folders/_d/dmrgv26d3bs6lkrks9z825_w0000gn/T/--03f176.o 
1.  While type-checking 'foo(meta1:meta2:)' (at <stdin>:4:1)
2.  While type-checking statement at [<stdin>:4:44 - line:7:1] RangeText="{
  print(meta1.E.self)
  print(meta2.E.self) // crash
"
3.  While type-checking expression at [<stdin>:6:3 - line:6:21] RangeText="print(meta2.E.self"

@xedin, does this look starter-bug-ish to you?

Slava mentions it would be a good starter bug here

Yeah, I agree with Slava on this one, let's see if anybody steps up to fix this.

Comment by Sam Lazarus (JIRA)

I'm happy to pick up the work on this!

slazarus (JIRA User) You most likely want to look somewhere around here

Comment by Sam Lazarus (JIRA)

Just put up this PR (#23787). I'm not sure if this is the best way to go about fixing this though. My patch fixes the crashing coercion, but it could be that the type is being incorrectly deduced which is causing the coercion to happen in the first place. If that's the case, then I treated the symptom, not the problem. @AnthonyLatsis Do you think this is the right way to go about fixing this?

Comment by Sam Lazarus (JIRA)

Looking into this more as I think I was right that my first attempt was treating a symptom, not the problem. I'll take a look in CSSimplify as per your recommendation and see what i can find.

slazarus (JIRA User) You might find this if statement and its comment interesting. Also, you can force push in this case if you want; no need to start a new PR.

Comment by Sam Lazarus (JIRA)

@AnthonyLatsis Reopened and updated the pull request. Thanks for pointing me in the right direction. Hopefully this implementation is a good way of going about this.