In Haskell it's traditional to support the last two stable releases of GHC, because upgrading can be a time consuming process. Would there be any interest in taking a contribution to allow hie-core to work with GHC 8.4? This would likely mean setting up a separate CI job using something like Travis for hie-core, to ensure it continued to work, and would imply ongoing maintenance. It might also be a good time to split hie-core out, if that is still on the table. To figure out how much work it would take to enable GHC 8.4 support, I started tweaking things, and found:

• There are a handful of lines around the CPP module and invoking it that need to change.
• fakeDynFlags has a few less members and slightly different interface.
• The pretty printing code in findImports would need a slight tweak.
• The CodeActions requires a 1 line tweak.
• Spans.Calculate has a bunch of code that has changed and would need more substantial compatibility work. 5 functions change, requiring approx 14 lines of patching.

The diffstat was:

 .../hie-core/src/Development/IDE/Core/Compile.hs     |  2 +-
 compiler/hie-core/src/Development/IDE/GHC/CPP.hs     |  2 ++
 compiler/hie-core/src/Development/IDE/GHC/Util.hs    |  4 ++--
 .../src/Development/IDE/Import/FindImports.hs        |  6 +++---
 .../hie-core/src/Development/IDE/LSP/CodeAction.hs   |  2 +-
 .../hie-core/src/Development/IDE/Spans/Calculate.hs  | 20 +++++++++++---------

Remember that 20 is 10 lines added and 10 removed, but also that in a few places I block commented or replaced with undefined, so the total diff to maintain would be slightly bigger.

CC @katiejots