sheaf

I don't think the current Happy grammar handles the lines in the OP, unfortunately.

Thanks for the bug report. I think the rule for SPECIALI(S|Z)E should check that it is preceded by a space and not just a word boundary.

Hmm, sorry about that. What happens if you run: ```sh npx vscode-tmgrammar-test "-s" "source.haskell" "-g" "/home/zliu41/language-haskell/syntaxes/haskell.json" "-t" "/home/zliu41/language-haskell/test/tests/ADTs.hs" ``` Thanks for taking the time to help, I appreciate it.

By the way, to change the syntax grammar you should be editing the `haskell.YAML-tmLanguage` file, and then generating `haskell.json` using `npx js-yaml haskell.YAML-tmLanguage > haskell.json`. I will add instructions to...

Some improvements to Haddock support in a0b8ea7, detecting more Haddock comment blocks than before (such as those starting with `-- ^`).

Yes a separate haddock grammar file is what I had in mind too. Do you think you could start one?

We should also include the Haddock grammar to highlight the `description` field in `.cabal` files.

I agree that it wouldn't be very useful for the large structs where one usually wants to access a single field. Does the matching force reading all the fields? Adding...

That's, nice, I didn't know about that. To spell out the obvious, if we write ```haskell {-# COMPLETE VkSurfaceFormatKHR #-} pattern VkSurfaceFormatKHR :: VkFormat -> VkColorSpaceKHR -> VkSurfaceFormatKHR pattern VkSurfaceFormatKHR...

Seems it was part of GHC 8.0, see [GHC issue #8582](https://gitlab.haskell.org/ghc/ghc/issues/8582#note_107857). Concerning duplicate record fields I came across [ticket #11228](https://gitlab.haskell.org/ghc/ghc/issues/11228), and [ticket #14630](https://gitlab.haskell.org/ghc/ghc/issues/14630) also has some interesting discussion on the...