Fix expressions indent

Closes GH-150.
Closes GH-152.

Reviewed-by: Titus Wormer <[email protected]>

src/construct/partial_mdx_expression.rs

@@ -56,15 +56,16 @@
 //! [mdx_expression_text]: crate::construct::mdx_expression_text
 //! [interleaving]: https://mdxjs.com/docs/what-is-mdx/#interleaving
 
-use crate::construct::partial_space_or_tab::space_or_tab_min_max;
 use crate::event::Name;
 use crate::message;
 use crate::state::{Name as StateName, State};
 use crate::tokenizer::Tokenizer;
-use crate::util::{constant::TAB_SIZE, mdx_collect::collect};
+use crate::util::mdx_collect::collect;
 use crate::{MdxExpressionKind, MdxExpressionParse, MdxSignal};
 use alloc::boxed::Box;
 
+pub const INDENT_SIZE: usize = 4;
+
 /// Start of an MDX expression.
 ///
 /// ```markdown
@@ -180,7 +181,6 @@ pub fn eol_after(tokenizer: &mut Tokenizer) -> State {
             }
         )
     } else if matches!(tokenizer.current, Some(b'\t' | b' ')) {
-        tokenizer.attempt(State::Next(StateName::MdxExpressionBefore), State::Nok);
         // Idea: investigate if we’d need to use more complex stripping.
         // Take this example:
         //
@@ -200,12 +200,27 @@ pub fn eol_after(tokenizer: &mut Tokenizer) -> State {
         // the start of the expression and move past whitespace.
         // For future lines, we’d move at most to
         // `line_start_shifted.column + 4`.
-        State::Retry(space_or_tab_min_max(tokenizer, 0, TAB_SIZE))
+        tokenizer.enter(Name::LinePrefix);
+        State::Retry(StateName::MdxExpressionPrefix)
     } else {
         State::Retry(StateName::MdxExpressionBefore)
     }
 }
 
+pub fn prefix(tokenizer: &mut Tokenizer) -> State {
+    tokenizer.tokenize_state.size_c += 1;
+    if matches!(tokenizer.current, Some(b'\t' | b' '))
+        && tokenizer.tokenize_state.size_c < INDENT_SIZE - 1
+    {
+        tokenizer.consume();
+        return State::Next(StateName::MdxExpressionPrefix);
+    }
+
+    tokenizer.exit(Name::LinePrefix);
+    tokenizer.tokenize_state.size_c = 0;
+    State::Retry(StateName::MdxExpressionBefore)
+}
+
 /// Parse an expression with a given function.
 fn parse_expression(tokenizer: &mut Tokenizer, parse: &MdxExpressionParse) -> State {
     // Collect the body of the expression and positional info for each run of it.

src/event.rs

@@ -3376,6 +3376,8 @@ pub enum Name {
     ///     ^ ^ ^
     /// ```
     ThematicBreakSequence,
+
+    LinePrefix,
 }
 
 /// List of void events, used to make sure everything is working well.

src/state.rs

@@ -360,6 +360,7 @@ pub enum Name {
 
     MdxExpressionStart,
     MdxExpressionBefore,
+    MdxExpressionPrefix,
     MdxExpressionInside,
     MdxExpressionEolAfter,
 
@@ -835,6 +836,7 @@ pub fn call(tokenizer: &mut Tokenizer, name: Name) -> State {
         Name::MdxEsmAtEnd => construct::mdx_esm::at_end,
 
         Name::MdxExpressionStart => construct::partial_mdx_expression::start,
+        Name::MdxExpressionPrefix => construct::partial_mdx_expression::prefix,
         Name::MdxExpressionBefore => construct::partial_mdx_expression::before,
         Name::MdxExpressionInside => construct::partial_mdx_expression::inside,
         Name::MdxExpressionEolAfter => construct::partial_mdx_expression::eol_after,

tests/mdx_expression_flow.rs

@@ -144,6 +144,19 @@ fn mdx_expression_flow_agnostic() -> Result<(), message::Message> {
         "should support mdx expressions (flow) as `MdxFlowExpression`s in mdast"
     );
 
+    assert_eq!(
+        to_mdast("  {`\n    a\n  `}", &mdx.parse)?,
+        Node::Root(Root {
+            children: vec![Node::MdxFlowExpression(MdxFlowExpression {
+                value: "`\n  a\n`".into(),
+                position: Some(Position::new(1, 3, 2, 3, 5, 15)),
+                stops: vec![(0, 3), (1, 4), (2, 7), (5, 10), (6, 13)]
+            })],
+            position: Some(Position::new(1, 1, 0, 3, 5, 15))
+        }),
+        "should support indent in `MdxFlowExpression` in mdast"
+    );
+
     Ok(())
 }