nom-sql: Parse escaped quotes in table column comments

We now handle doubled quotation marks, both single and double (MySQL
only).  We add several tests of the various combinations of these
things, including of table comments, which were already working.

Release-Note-Core: Correctly handle escaped quotes in table column
  comments.
Fixes: REA-4446
Change-Id: I40d56e5b01880a182db1cc73b2e7e6fd6ff0ebfd
Reviewed-on: https://gerrit.readyset.name/c/readyset/+/7626
Tested-by: Buildkite CI
Reviewed-by: Marcelo Altmann <[email protected]>

Author: rs-sac

nom-sql/src/column.rs

@@ -292,24 +292,29 @@ pub fn column_specification(
     dialect: Dialect,
 ) -> impl Fn(LocatedSpan<&[u8]>) -> NomSqlResult<&[u8], ColumnSpecification> {
     move |i| {
-        let (remaining_input, (column, field_type, constraints, comment)) = tuple((
+        let (i, (column, field_type, constraints)) = tuple((
             column_identifier_no_alias(dialect),
             opt(delimited(
                 whitespace1,
                 type_identifier(dialect),
                 whitespace0,
             )),
             many0(column_constraint(dialect)),
-            opt(parse_comment),
         ))(i)?;
 
+        let (i, comment) = if matches!(dialect, Dialect::MySQL) {
+            opt(parse_comment(dialect))(i)?
+        } else {
+            (i, None)
+        };
+
         let sql_type = match field_type {
             None => SqlType::Text,
             Some(ref t) => t.clone(),
         };
 
         Ok((
-            remaining_input,
+            i,
             ColumnSpecification {
                 column,
                 sql_type,

nom-sql/src/common.rs

@@ -6,7 +6,7 @@ use std::str::FromStr;
 
 use itertools::Itertools;
 use nom::branch::alt;
-use nom::bytes::complete::{tag, tag_no_case, take_until};
+use nom::bytes::complete::{tag, tag_no_case};
 use nom::character::complete::{digit1, line_ending};
 use nom::combinator::{map, map_res, not, opt, peek};
 use nom::error::{ErrorKind, ParseError};
@@ -810,17 +810,18 @@ pub(crate) fn if_not_exists(i: LocatedSpan<&[u8]>) -> NomSqlResult<&[u8], bool>
 }
 
 // Parse rule for a comment part.
-pub fn parse_comment(i: LocatedSpan<&[u8]>) -> NomSqlResult<&[u8], String> {
-    map(
-        preceded(
-            delimited(whitespace0, tag_no_case("comment"), whitespace1),
-            map_res(
-                delimited(tag("'"), take_until("'"), tag("'")),
-                |i: LocatedSpan<&[u8]>| str::from_utf8(&i),
+pub fn parse_comment(
+    dialect: Dialect,
+) -> impl Fn(LocatedSpan<&[u8]>) -> NomSqlResult<&[u8], String> {
+    move |i| {
+        map(
+            preceded(
+                delimited(whitespace0, tag_no_case("comment"), whitespace1),
+                map_res(move |i| dialect.string_literal()(i), String::from_utf8),
             ),
-        ),
-        String::from,
-    )(i)
+            String::from,
+        )(i)
+    }
 }
 
 pub fn field_reference(
@@ -980,12 +981,6 @@ mod tests {
         );
     }
 
-    #[test]
-    fn comment_data() {
-        let res = parse_comment(LocatedSpan::new(b" COMMENT 'test'"));
-        assert_eq!(res.unwrap().1, "test");
-    }
-
     #[test]
     fn terminated_by_semicolon() {
         let res = to_nom_result(statement_terminator(LocatedSpan::new(b"   ;  ")));
@@ -1204,6 +1199,22 @@ mod tests {
                 }
             );
         }
+
+        #[test]
+        fn table_column_comment() {
+            let res = test_parse!(parse_comment(Dialect::MySQL), b"comment 'foo'");
+            assert_eq!(res, "foo");
+            let res = test_parse!(parse_comment(Dialect::MySQL), b"comment \"foo\"");
+            assert_eq!(res, "foo");
+            let res = test_parse!(parse_comment(Dialect::MySQL), b"comment 'f''oo'");
+            assert_eq!(res, "f'oo");
+            let res = test_parse!(parse_comment(Dialect::MySQL), b"comment 'f\"\"oo'");
+            assert_eq!(res, "f\"\"oo");
+            let res = test_parse!(parse_comment(Dialect::MySQL), b"comment \"f\"\"oo\"");
+            assert_eq!(res, "f\"oo");
+            let res = test_parse!(parse_comment(Dialect::MySQL), b"comment \"f''oo\"");
+            assert_eq!(res, "f''oo");
+        }
     }
 
     mod postgres {

nom-sql/src/create_table_options.rs

@@ -421,4 +421,24 @@ mod tests {
             vec![CreateTableOption::Comment("foobar".to_string())],
         );
     }
+
+    #[test]
+    fn create_table_option_comment_escape() {
+        should_parse_all(
+            "COMMENT='foo''bar'",
+            vec![CreateTableOption::Comment("foo'bar".to_string())],
+        );
+        should_parse_all(
+            "COMMENT=\"foo\"\"bar\"",
+            vec![CreateTableOption::Comment("foo\"bar".to_string())],
+        );
+        should_parse_all(
+            "COMMENT='foo\"\"bar'",
+            vec![CreateTableOption::Comment("foo\"\"bar".to_string())],
+        );
+        should_parse_all(
+            "COMMENT=\"foo''bar\"",
+            vec![CreateTableOption::Comment("foo''bar".to_string())],
+        );
+    }
 }