Add support for trimming whitespace around string, text

This commit introduces trimming initial and final whitespace around the
whole string or text, or around line endings inside that string or text.

*   Add `register_resolver_before`, to run resolvers earlier than others,
    used for labels
*   Add resolver to merge `data` events, which are the most frequent token
    that occurs, and can happen adjacently.
    In `micromark-js` this sped up parsing a lot
*   Fix a bug where a virtual space was not seen as an okay event
*   Refactor to enable all turned off whitespace tests

1 files changed, 8 insertions, 1 deletions

diff --git a/src/tokenizer.rs b/src/tokenizer.rs
index fe69366..817c1de 100644
--- a/src/tokenizer.rs
+++ b/src/tokenizer.rs
@@ -1796,6 +1796,13 @@ impl<'a> Tokenizer<'a> {
         }
     }
 
+    pub fn register_resolver_before(&mut self, id: String, resolver: Box<Resolver>) {
+        if !self.resolver_ids.contains(&id) {
+            self.resolver_ids.push(id);
+            self.resolvers.insert(0, resolver);
+        }
+    }
+
     /// Prepare for a next code to get consumed.
     fn expect(&mut self, code: Code) {
         assert!(self.consumed, "expected previous character to be consumed");
@@ -1901,7 +1908,7 @@ impl<'a> Tokenizer<'a> {
         let point = self.point.clone();
 
         assert!(
-            current_token != previous.token_type || previous.point != point,
+            current_token != previous.token_type || previous.index != self.index,
             "expected non-empty token"
         );