Not sure what's going on, but an index-state bump allowed HLS to pull in newer cabal-gild, which broke our tests. I manually tracked it down to cabal-gild-1.3.2, so I guess #84.

Here's what we have in the failing test, I don't know why this would have started failing now? https://github.com/haskell/haskell-language-server/blob/master/plugins/hls-cabal-gild-plugin/test/testdata/commented_testdata.cabal#L8