Make data a construct

6 files changed, 108 insertions, 90 deletions

diff --git a/readme.md b/readme.md
index 6689c7f..f5ba870 100644
--- a/readme.md
+++ b/readme.md
@@ -84,8 +84,6 @@ cargo doc --document-private-items
 
 #### Refactor
 
-- [ ] (1) Make text data, string data constructs (document in
-      `construct/mod.rs`)
 - [ ] (1) Configurable tokens (destination, label, title)
 - [ ] (1) Configurable limit (destination)
 
@@ -237,6 +235,8 @@ cargo doc --document-private-items
 - [x] (1) Figure out lifetimes of things (see `life time` in source)
 - [x] (1) Use traits for a bunch of enums, e.g., markers
 - [x] (1) Move safe protocols to constants
+- [x] (1) Make text data, string data constructs (document in
+      `construct/mod.rs`)
 
 ### Extensions
 
diff --git a/src/construct/mod.rs b/src/construct/mod.rs
index 407dc6b..9e5da0e 100644
--- a/src/construct/mod.rs
+++ b/src/construct/mod.rs
@@ -7,11 +7,10 @@
 //! For example, [code (fenced)][code_fenced] and
 //! [code (indented)][code_indented] are considered different constructs
 //!
-//! <!-- To do: can these rest things be made into constructs? -->
-//!
-//! Content types also have a *rest* thing: after all character escapes and
-//! character references are parsed, there’s something left.
-//! This remainder is, currently, not called a constructs.
+//! Content types also have a *rest* thing: after all things are parsed,
+//! there’s something left.
+//! In flow, that is a [paragraph][].
+//! In string and text, that is [data][partial_data].
 //!
 //! The following constructs are found in markdown:
 //!
@@ -38,6 +37,14 @@
 //! *   [paragraph][]
 //! *   [thematic break][thematic_break]
 //!
+//! There are also several routines used in different places:
+//!
+//! *   [data][partial_data]
+//! *   [destination][partial_destination]
+//! *   [label][partial_label]
+//! *   [space or tab][partial_space_or_tab]
+//! *   [title][partial_title]
+//!
 //! Each construct maintained here is explained with a BNF diagram.
 //! For example, the docs for [character escape][character_escape] contain:
 //!
@@ -52,6 +59,8 @@
 //! They also contain references to character as defined by [char][], so for
 //! example `ascii_punctuation` refers to
 //! [`char::is_ascii_punctuation`][char::is_ascii_punctuation].
+//!
+//!
 
 pub mod autolink;
 pub mod blank_line;
@@ -68,6 +77,7 @@ pub mod heading_setext;
 pub mod html_flow;
 pub mod html_text;
 pub mod paragraph;
+pub mod partial_data;
 pub mod partial_destination;
 pub mod partial_label;
 pub mod partial_space_or_tab;
diff --git a/src/construct/partial_data.rs b/src/construct/partial_data.rs
new file mode 100644
index 0000000..d83787a
--- /dev/null
+++ b/src/construct/partial_data.rs
@@ -0,0 +1,69 @@
+//! Data occurs in [text][] and [string][].
+//!
+//! It can include anything (including line endings), and stops at certain
+//! characters.
+//!
+//! [string]: crate::content::string
+//! [text]: crate::content::text
+
+// To do: pass token types in?
+
+use crate::tokenizer::{Code, State, StateFnResult, TokenType, Tokenizer};
+
+/// At the beginning of data.
+///
+/// ```markdown
+/// |&qwe
+/// ```
+pub fn start(tokenizer: &mut Tokenizer, code: Code, stop: Vec<Code>) -> StateFnResult {
+    if stop.contains(&code) {
+        tokenizer.enter(TokenType::Data);
+        tokenizer.consume(code);
+        (State::Fn(Box::new(|t, c| data(t, c, stop))), None)
+    } else {
+        at_break(tokenizer, code, stop)
+    }
+}
+
+/// Before something.
+///
+/// ```markdown
+/// |qwe| |&
+/// ```
+fn at_break(tokenizer: &mut Tokenizer, code: Code, stop: Vec<Code>) -> StateFnResult {
+    match code {
+        Code::None => (State::Ok, None),
+        Code::CarriageReturnLineFeed | Code::Char('\r' | '\n') => {
+            tokenizer.enter(TokenType::LineEnding);
+            tokenizer.consume(code);
+            tokenizer.exit(TokenType::LineEnding);
+            (State::Fn(Box::new(|t, c| at_break(t, c, stop))), None)
+        }
+        _ if stop.contains(&code) => (State::Ok, Some(vec![code])),
+        _ => {
+            tokenizer.enter(TokenType::Data);
+            data(tokenizer, code, stop)
+        }
+    }
+}
+
+/// In data.
+///
+/// ```markdown
+/// q|w|e
+/// ```
+fn data(tokenizer: &mut Tokenizer, code: Code, stop: Vec<Code>) -> StateFnResult {
+    let done = match code {
+        Code::None | Code::CarriageReturnLineFeed | Code::Char('\r' | '\n') => true,
+        _ if stop.contains(&code) => true,
+        _ => false,
+    };
+
+    if done {
+        tokenizer.exit(TokenType::Data);
+        at_break(tokenizer, code, stop)
+    } else {
+        tokenizer.consume(code);
+        (State::Fn(Box::new(|t, c| data(t, c, stop))), None)
+    }
+}
diff --git a/src/content/string.rs b/src/content/string.rs
index bae2646..3338c90 100644
--- a/src/content/string.rs
+++ b/src/content/string.rs
@@ -14,8 +14,14 @@
 
 use crate::construct::{
     character_escape::start as character_escape, character_reference::start as character_reference,
+    partial_data::start as data,
 };
-use crate::tokenizer::{Code, State, StateFnResult, TokenType, Tokenizer};
+use crate::tokenizer::{Code, State, StateFnResult, Tokenizer};
+
+const MARKERS: [Code; 2] = [
+    Code::Char('&'),  // `character_reference`
+    Code::Char('\\'), // `character_escape`
+];
 
 /// Before string.
 ///
@@ -33,49 +39,11 @@ pub fn start(tokenizer: &mut Tokenizer, code: Code) -> StateFnResult {
     }
 }
 
-/// Before string, not at a character reference or character escape.
-///
-/// We’re at data.
+/// At data.
 ///
 /// ```markdown
 /// |qwe
 /// ```
 fn before_data(tokenizer: &mut Tokenizer, code: Code) -> StateFnResult {
-    match code {
-        Code::None => (State::Ok, None),
-        Code::CarriageReturnLineFeed | Code::Char('\r' | '\n') => {
-            tokenizer.enter(TokenType::LineEnding);
-            tokenizer.consume(code);
-            tokenizer.exit(TokenType::LineEnding);
-            (State::Fn(Box::new(start)), None)
-        }
-        _ => {
-            tokenizer.enter(TokenType::Data);
-            tokenizer.consume(code);
-            (State::Fn(Box::new(in_data)), None)
-        }
-    }
-}
-
-/// In data.
-///
-/// ```markdown
-/// q|w|e
-/// ```
-fn in_data(tokenizer: &mut Tokenizer, code: Code) -> StateFnResult {
-    match code {
-        Code::None | Code::CarriageReturnLineFeed | Code::Char('\r' | '\n') => {
-            tokenizer.exit(TokenType::Data);
-            before_data(tokenizer, code)
-        }
-        // To do: somehow get these markers from constructs.
-        Code::Char('&' | '\\') => {
-            tokenizer.exit(TokenType::Data);
-            start(tokenizer, code)
-        }
-        _ => {
-            tokenizer.consume(code);
-            (State::Fn(Box::new(in_data)), None)
-        }
-    }
+    tokenizer.go(|t, c| data(t, c, MARKERS.to_vec()), start)(tokenizer, code)
 }
diff --git a/src/content/text.rs b/src/content/text.rs
index 6a30d4c..857e9a0 100644
--- a/src/content/text.rs
+++ b/src/content/text.rs
@@ -21,8 +21,17 @@ use crate::construct::{
     character_reference::start as character_reference, code_text::start as code_text,
     hard_break_escape::start as hard_break_escape,
     hard_break_trailing::start as hard_break_trailing, html_text::start as html_text,
+    partial_data::start as data,
 };
-use crate::tokenizer::{Code, State, StateFnResult, TokenType, Tokenizer};
+use crate::tokenizer::{Code, State, StateFnResult, Tokenizer};
+
+const MARKERS: [Code; 5] = [
+    Code::Char(' '),  // `hard_break_trailing`
+    Code::Char('&'),  // `character_reference`
+    Code::Char('<'),  // `autolink`, `html_text`
+    Code::Char('\\'), // `character_escape`, `hard_break_escape`
+    Code::Char('`'),  // `code_text`
+];
 
 /// Before text.
 ///
@@ -49,49 +58,11 @@ pub fn start(tokenizer: &mut Tokenizer, code: Code) -> StateFnResult {
     }
 }
 
-/// Before text.
-///
-/// We’re at data.
+/// At data.
 ///
 /// ```markdown
 /// |qwe
 /// ```
 fn before_data(tokenizer: &mut Tokenizer, code: Code) -> StateFnResult {
-    match code {
-        Code::None => (State::Ok, None),
-        Code::CarriageReturnLineFeed | Code::Char('\r' | '\n') => {
-            tokenizer.enter(TokenType::LineEnding);
-            tokenizer.consume(code);
-            tokenizer.exit(TokenType::LineEnding);
-            (State::Fn(Box::new(start)), None)
-        }
-        _ => {
-            tokenizer.enter(TokenType::Data);
-            tokenizer.consume(code);
-            (State::Fn(Box::new(in_data)), None)
-        }
-    }
-}
-
-/// In data.
-///
-/// ```markdown
-/// q|w|e
-/// ```
-fn in_data(tokenizer: &mut Tokenizer, code: Code) -> StateFnResult {
-    match code {
-        Code::None | Code::CarriageReturnLineFeed | Code::Char('\r' | '\n') => {
-            tokenizer.exit(TokenType::Data);
-            before_data(tokenizer, code)
-        }
-        // To do: somehow get these markers from constructs.
-        Code::Char(' ' | '&' | '<' | '\\' | '`') => {
-            tokenizer.exit(TokenType::Data);
-            start(tokenizer, code)
-        }
-        _ => {
-            tokenizer.consume(code);
-            (State::Fn(Box::new(in_data)), None)
-        }
-    }
+    tokenizer.go(|t, c| data(t, c, MARKERS.to_vec()), start)(tokenizer, code)
 }
diff --git a/tests/autolink.rs b/tests/autolink.rs
index 3882264..f0486ef 100644
--- a/tests/autolink.rs
+++ b/tests/autolink.rs
@@ -8,7 +8,7 @@ const DANGER: &Options = &Options {
 };
 
 #[test]
-fn code_fenced() {
+fn autolink() {
     assert_eq!(
         micromark("```\n<\n >\n```"),
         "<pre><code>&lt;\n &gt;\n</code></pre>",