Make paragraphs really fast

The approach that `micromark-js` takes is as follows: to parse a
paragraph, check whether each line starts with something else.
If it does, exit, otherwise continue.
That is slow, because our actual flow parser does similar things: the work was
being done twice.

To fix this, this commit introduces parsing each line of a paragraph separately.
And finally, when done with flow, combining adjacent paragraphs.

This same mechanism is reused for setext headings.

Additionally, this commit adds support for interrupting things (or not).
E.g., HTML (flow, complete) cannot interrupt paragraphs.
Definitions cannot interrupt paragraphs, and connect be interrupted either,
but they can follow each other.

1 files changed, 14 insertions, 5 deletions

diff --git a/src/construct/html_flow.rs b/src/construct/html_flow.rs
index d0e0558..a1bddad 100644
--- a/src/construct/html_flow.rs
+++ b/src/construct/html_flow.rs
@@ -98,8 +98,10 @@
 //! [html_block_names]: crate::constant::HTML_BLOCK_NAMES
 //! [html-parsing]: https://html.spec.whatwg.org/multipage/parsing.html#parsing
 
-use crate::constant::{HTML_BLOCK_NAMES, HTML_RAW_NAMES, HTML_RAW_SIZE_MAX};
-use crate::construct::{blank_line::start as blank_line, partial_space_or_tab::space_or_tab};
+use crate::constant::{HTML_BLOCK_NAMES, HTML_RAW_NAMES, HTML_RAW_SIZE_MAX, TAB_SIZE};
+use crate::construct::{
+    blank_line::start as blank_line, partial_space_or_tab::space_or_tab_min_max,
+};
 use crate::tokenizer::{Code, State, StateFnResult, TokenType, Tokenizer};
 
 /// Kind of HTML (flow).
@@ -191,7 +193,8 @@ struct Info {
 pub fn start(tokenizer: &mut Tokenizer, code: Code) -> StateFnResult {
     tokenizer.enter(TokenType::HtmlFlow);
     tokenizer.enter(TokenType::HtmlFlowData);
-    tokenizer.attempt_opt(space_or_tab(), before)(tokenizer, code)
+    // To do: allow arbitrary when code (indented) is turned off.
+    tokenizer.go(space_or_tab_min_max(0, TAB_SIZE - 1), before)(tokenizer, code)
 }
 
 /// After optional whitespace, before `<`.
@@ -400,8 +403,10 @@ fn tag_name(tokenizer: &mut Tokenizer, code: Code, mut info: Info) -> StateFnRes
             } else {
                 info.kind = Kind::Complete;
 
-                // To do: do not support complete HTML when interrupting.
-                if info.start_tag {
+                // Do not support complete HTML when interrupting.
+                if tokenizer.interrupt {
+                    (State::Nok, None)
+                } else if info.start_tag {
                     complete_attribute_name_before(tokenizer, code, info)
                 } else {
                     complete_closing_tag_after(tokenizer, code, info)
@@ -784,6 +789,8 @@ fn html_continue_start(tokenizer: &mut Tokenizer, code: Code, info: Info) -> Sta
     match code {
         Code::None => {
             tokenizer.exit(TokenType::HtmlFlow);
+            // Feel free to interrupt.
+            tokenizer.interrupt = false;
             (State::Ok, Some(vec![code]))
         }
         // To do: do not allow lazy lines.
@@ -949,6 +956,8 @@ fn continuation_close(tokenizer: &mut Tokenizer, code: Code, info: Info) -> Stat
         Code::None | Code::CarriageReturnLineFeed | Code::Char('\n' | '\r') => {
             tokenizer.exit(TokenType::HtmlFlowData);
             tokenizer.exit(TokenType::HtmlFlow);
+            // Feel free to interrupt.
+            tokenizer.interrupt = false;
             (State::Ok, Some(vec![code]))
         }
         _ => {