toJSONFilter no longer supports non-monad list result

[elliottslaughter]

I have been using the following filter on some of my documents. As of pandoc-types 1.23.0.1, this appears to no longer compile:

{-# LANGUAGE OverloadedStrings #-}
import Text.Pandoc.Definition
import Text.Pandoc.JSON

-- This pass strips <span class="remove">...</span>.
removeSpan :: Inline -> [Inline]
removeSpan (Span (_, ["remove"], _) _) = []
removeSpan x = [x]

main :: IO ()
main = toJSONFilter removeSpan

Error:

my_filter.hs:11:8: error:
    • No instance for (ToJSONFilter IO (Inline -> [Inline]))
        arising from a use of ‘toJSONFilter’
        (maybe you haven't applied a function to enough arguments?)
    • In the expression: toJSONFilter removeSpan
      In an equation for ‘main’: main = toJSONFilter removeSpan
   |
11 | main = toJSONFilter removeSpan
   |        ^^^^^^^^^^^^

After staring a bit at the documentation and source code for toJSONFilter, I came to the conclusion that this form is no longer supported and I need to add the IO monad to unbreak it. And indeed, this works:

{-# LANGUAGE OverloadedStrings #-}
import Text.Pandoc.Definition
import Text.Pandoc.JSON

-- This pass strips <span class="remove">...</span>.
removeSpan :: Inline -> IO [Inline]
removeSpan (Span (_, ["remove"], _) _) = return []
removeSpan x = return [x]

main :: IO ()
main = toJSONFilter removeSpan

It's not the most horrible change to need to make to my code, but it is (I think) a breaking change, and can cause users some needless effort, especially because the list form is not clearly demonstrated in the documentation of either the filters page or Hackage for toJSONFilter.

Versions, etc.:

$ pandoc --version
pandoc 3.1.4
Features: +server +lua
Scripting engine: Lua 5.4
User data directory: .../pandoc
Copyright (C) 2006-2023 John MacFarlane. Web:  https://pandoc.org
This is free software; see the source for copying conditions. There is no
warranty, not even for merchantability or fitness for a particular purpose.

macOS 13.5

[jgm]

See this change.

It sure looks like this commit removed a pure instance and added IO. I hadn't noticed that. If so it should be changed!

[jgm]

@KommuSoft may want to comment.

[jgm]

Ah, I think the problem was not @KommuSoft's commit but my revision. Will fix.