Update inline docs

Author: bgrgicak

tests/WP_SQLite_Translator_Tests.php

@@ -1253,7 +1253,7 @@ public function testColumnWithOnUpdate() {
 	}
 
 	public function testTypeKeywordsAsKeyNames() {
-		// CREATE TABLE with ON UPDATE
+		// CREATE TABLE with a data type as a key name
 		$this->assertQuery(
 			'CREATE TABLE `_tmp_table` (
 				`id` bigint(20) unsigned NOT NULL AUTO_INCREMENT,

wp-includes/sqlite/class-wp-sqlite-translator.php

@@ -1004,12 +1004,21 @@ private function parse_create_table() {
 			$this->rewriter->replace_all( array() );
 
 			/*
-			 * Decide how to parse the current line. We expect either:
+			 * Decide how to parse the current line. We ehttps://github.com/WordPress/sqlite-database-integration/blob/eba8ab5a92d95d332cdbcdf008c7328906027cc2/wp-includes/sqlite/class-wp-sqlite-translator.php#L1007xpect either:
 			 *
 			 * Field definition, e.g.:
 			 *     `my_field` varchar(255) NOT NULL DEFAULT 'foo'
 			 * Constraint definition, e.g.:
 			 *      PRIMARY KEY (`my_field`)
+			 *
+			 * Lexer does not seem to reliably understand whether the
+			 * first token is a field name or a reserved keyword, so
+			 * alongside checking for the reserved keyword, we'll also
+			 * check whether the second non-whitespace token is a data type.
+			 *
+			 * By checking for the reserved keyword, we can be sure that
+			 * we're not parsing a constraint as a field when the
+			 * constraint symbol matches a data type.
 			 */
 			$current_token = $this->rewriter->peek();
 			$second_token  = $this->rewriter->peek_nth( 2 );