Fix #2924. Use MegaParsec scanner for Markdown files (#2925)

This PR addresses a bug/missing case present since v0.6.2, introduced
specifically by

- PR #2779, 

That PR involves detecting imports in Juvix files before type checking,
and that's the issue.
Detecting/scanning imports is done by running a flat parser (which
ignores the Juvix Markdown structure) and when it fails, it runs a
Megaparser parse. So, for simplicity,
we could just continue using the same Megaparser as before for Juvix
Markdown files.

---------

Co-authored-by: Jan Mas Rovira <janmasrovira@gmail.com>

src/Juvix/Compiler/Concrete/Translation/ImportScanner/FlatParse.hs

@@ -12,7 +12,10 @@ import Juvix.Prelude.FlatParse qualified as FP
 import Juvix.Prelude.FlatParse.Lexer qualified as L
 
 scanBSImports :: Path Abs File -> ByteString -> Maybe ScanResult
-scanBSImports fp = fromResult . scanner fp
+scanBSImports fp
+  -- FlatParse only supports .juvix files
+  | isJuvixFile fp = fromResult . scanner fp
+  | otherwise = const Nothing
   where
     fromResult :: Result () ok -> Maybe ok
     fromResult = \case

tests/positive/Markdown/Test.juvix.md

@@ -1,5 +1,7 @@
 # Example
 
+What is important is seldom urgent. 
+
 A Juvix Markdown file name ends with `.juvix.md`. This kind of file must contain
 a module declaration at the top, as shown below ---in the first code block. 
 

tests/positive/Markdown/markdown/Test.md

@@ -1,5 +1,7 @@
 # Example
 
+What is important is seldom urgent.
+
 A Juvix Markdown file name ends with `.juvix.md`. This kind of file must contain
 a module declaration at the top, as shown below ---in the first code block.