From 9968aedd7412e9987d100cecf1c61d058432fc51 Mon Sep 17 00:00:00 2001
From: Tom Alexander <tom@fizz.buzz>
Date: Sat, 22 Apr 2023 22:34:37 -0400
Subject: [PATCH] Make sure text markup doesn't have interior spaces.

---
 src/parser/text_markup.rs |  6 ++++++
 src/parser/util.rs        | 22 ++++++++++++++++++++++
 2 files changed, 28 insertions(+)

diff --git a/src/parser/text_markup.rs b/src/parser/text_markup.rs
index db3ce057..1c36c51f 100644
--- a/src/parser/text_markup.rs
+++ b/src/parser/text_markup.rs
@@ -2,9 +2,11 @@ use nom::branch::alt;
 use nom::bytes::complete::tag;
 use nom::character::complete::anychar;
 use nom::character::complete::line_ending;
+use nom::character::complete::multispace1;
 use nom::character::complete::one_of;
 use nom::character::complete::space0;
 use nom::combinator::map;
+use nom::combinator::not;
 use nom::combinator::peek;
 use nom::combinator::recognize;
 use nom::combinator::verify;
@@ -30,6 +32,7 @@ use crate::parser::Object;
 use crate::parser::StrikeThrough;
 use crate::parser::Underline;
 use crate::parser::Verbatim;
+use crate::parser::util::preceded_by_whitespace;
 
 #[tracing::instrument(ret, level = "debug")]
 pub fn text_markup<'r, 's>(context: Context<'r, 's>, input: &'s str) -> Res<&'s str, Object<'s>> {
@@ -112,6 +115,7 @@ fn _text_markup_object<'r, 's, 'x>(
 ) -> Res<&'s str, Vec<Object<'s>>> {
     let (remaining, _) = pre(context, input)?;
     let (remaining, open) = tag(marker_symbol)(remaining)?;
+    let (remaining, _peek_not_whitespace) = peek(not(multispace1))(remaining)?;
     let text_markup_end_specialized = text_markup_end(open);
     let parser_context =
         context.with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
@@ -148,6 +152,7 @@ fn _text_markup_string<'r, 's, 'x>(
 ) -> Res<&'s str, &'s str> {
     let (remaining, _) = pre(context, input)?;
     let (remaining, open) = tag(marker_symbol)(remaining)?;
+    let (remaining, _peek_not_whitespace) = peek(not(multispace1))(remaining)?;
     let text_markup_end_specialized = text_markup_end(open);
     let parser_context =
         context.with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
@@ -208,6 +213,7 @@ fn _text_markup_end<'r, 's, 'x>(
     input: &'s str,
     marker_symbol: &'x str,
 ) -> Res<&'s str, &'s str> {
+    not(parser_with_context!(preceded_by_whitespace)(context))(input)?;
     let (remaining, _marker) = terminated(
         tag(marker_symbol),
         peek(parser_with_context!(post)(context)),
diff --git a/src/parser/util.rs b/src/parser/util.rs
index b5847f2e..fe1169ef 100644
--- a/src/parser/util.rs
+++ b/src/parser/util.rs
@@ -151,6 +151,28 @@ pub fn start_of_line<'r, 's>(context: Context<'r, 's>, input: &'s str) -> Res<&'
     Ok((input, ()))
 }
 
+/// Check that we are at the start of a line
+#[tracing::instrument(ret, level = "debug")]
+pub fn preceded_by_whitespace<'r, 's>(
+    context: Context<'r, 's>,
+    input: &'s str,
+) -> Res<&'s str, ()> {
+    let document_root = context.get_document_root().unwrap();
+    let preceding_character = get_one_before(document_root, input)
+        .map(|slice| slice.chars().next())
+        .flatten();
+    match preceding_character {
+        Some('\n') | Some('\r') | Some(' ') | Some('\t') => {}
+        // If None, we are at the start of the file which is not allowed
+        None | Some(_) => {
+            return Err(nom::Err::Error(CustomError::MyError(MyError(
+                "Not preceded by whitespace.",
+            ))));
+        }
+    };
+    Ok((input, ()))
+}
+
 /// Pull one non-whitespace character.
 ///
 /// This function only operates on spaces, tabs, carriage returns, and line feeds. It does not handle fancy unicode whitespace.