Fix bugs in SQLScriptScanner with big String literals and PostgreSQL identifiers (as introduced by #7646) (#7818)

Co-authored-by: Eddú Meléndez Gonzales <eddu.melendez@gmail.com>

Author: inponomarev

modules/database-commons/src/main/java/org/testcontainers/ext/ScriptScanner.java

@@ -28,11 +28,7 @@ class ScriptScanner {
 
     private final Pattern whitespace = Pattern.compile("\\s+");
 
-    private final Pattern identifier = Pattern.compile("[a-z][a-z0-9_]*", Pattern.CASE_INSENSITIVE);
-
-    private final Pattern singleQuotedString = Pattern.compile("'(\\\\'|[^'])*'");
-
-    private final Pattern ansiQuotedString = Pattern.compile("\"(\\\\\"|[^\"])*\"");
+    private final Pattern identifier = Pattern.compile("[a-z][a-z0-9_$]*", Pattern.CASE_INSENSITIVE);
 
     private final Pattern dollarQuotedStringDelimiter = Pattern.compile("\\$\\w*\\$");
 
@@ -54,7 +50,8 @@ private boolean matches(String substring) {
 
     private boolean matches(Pattern regexp) {
         Matcher m = regexp.matcher(script);
-        if (m.find(offset) && m.start() == offset) {
+        m.region(offset, script.length());
+        if (m.lookingAt()) {
             currentMatch = m.group();
             offset = m.end();
             return true;
@@ -99,6 +96,26 @@ private boolean matchesMultilineComment() {
         return false;
     }
 
+    private boolean matchesQuotedString(final char quote) {
+        if (script.charAt(offset) == quote) {
+            boolean escaped = false;
+            for (int i = offset + 1; i < script.length(); i++) {
+                char c = script.charAt(i);
+                if (escaped) {
+                    //just skip the escaped character and drop the flag
+                    escaped = false;
+                } else if (c == '\\') {
+                    escaped = true;
+                } else if (c == quote) {
+                    currentMatch = script.substring(offset, i + 1);
+                    offset = i + 1;
+                    return true;
+                }
+            }
+        }
+        return false;
+    }
+
     private boolean matchesDollarQuotedString() {
         //Matches $<tag>$ .... $<tag>$
         if (matches(dollarQuotedStringDelimiter)) {
@@ -124,7 +141,7 @@ Lexem next() {
                 return Lexem.SEPARATOR;
             } else if (matchesSingleLineComment() || matchesMultilineComment()) {
                 return Lexem.COMMENT;
-            } else if (matches(singleQuotedString) || matches(ansiQuotedString) || matchesDollarQuotedString()) {
+            } else if (matchesQuotedString('\'') || matchesQuotedString('"') || matchesDollarQuotedString()) {
                 return Lexem.QUOTED_STRING;
             } else if (matches(identifier)) {
                 return Lexem.IDENTIFIER;

modules/database-commons/src/test/java/org/testcontainers/ext/ScriptScannerTest.java

@@ -0,0 +1,54 @@
+package org.testcontainers.ext;
+
+import org.apache.commons.lang3.StringUtils;
+import org.junit.Test;
+
+import java.util.regex.Pattern;
+
+import static org.assertj.core.api.Assertions.assertThat;
+
+public class ScriptScannerTest {
+
+    @Test
+    public void testHugeStringLiteral() {
+        String script = "/* a comment */    \"" + StringUtils.repeat('~', 10000) + "\";";
+        ScriptScanner scanner = scanner(script);
+        assertThat(scanner.next()).isEqualTo(ScriptScanner.Lexem.COMMENT);
+        assertThat(scanner.next()).isEqualTo(ScriptScanner.Lexem.WHITESPACE);
+        assertThat(scanner.next()).isEqualTo(ScriptScanner.Lexem.QUOTED_STRING);
+        assertThat(scanner.getCurrentMatch()).matches(Pattern.compile("\"~+\""));
+    }
+
+    @Test
+    public void testPgIdentifierWithDollarSigns() {
+        ScriptScanner scanner = scanner(
+            "this$is$a$valid$postgreSQL$identifier  " +
+            "$a$While this is a quoted string$a$$ --just followed by a dollar sign"
+        );
+        assertThat(scanner.next()).isEqualTo(ScriptScanner.Lexem.IDENTIFIER);
+        assertThat(scanner.next()).isEqualTo(ScriptScanner.Lexem.WHITESPACE);
+        assertThat(scanner.next()).isEqualTo(ScriptScanner.Lexem.QUOTED_STRING);
+        assertThat(scanner.next()).isEqualTo(ScriptScanner.Lexem.OTHER);
+    }
+
+    @Test
+    public void testQuotedLiterals() {
+        ScriptScanner scanner = scanner("'this \\'is a literal' \"this \\\" is a literal\"");
+        assertThat(scanner.next()).isEqualTo(ScriptScanner.Lexem.QUOTED_STRING);
+        assertThat(scanner.getCurrentMatch()).isEqualTo("'this \\'is a literal'");
+        assertThat(scanner.next()).isEqualTo(ScriptScanner.Lexem.WHITESPACE);
+        assertThat(scanner.next()).isEqualTo(ScriptScanner.Lexem.QUOTED_STRING);
+        assertThat(scanner.getCurrentMatch()).isEqualTo("\"this \\\" is a literal\"");
+    }
+
+    private static ScriptScanner scanner(String script) {
+        return new ScriptScanner(
+            "dummy",
+            script,
+            ScriptUtils.DEFAULT_STATEMENT_SEPARATOR,
+            ScriptUtils.DEFAULT_COMMENT_PREFIX,
+            ScriptUtils.DEFAULT_BLOCK_COMMENT_START_DELIMITER,
+            ScriptUtils.DEFAULT_BLOCK_COMMENT_END_DELIMITER
+        );
+    }
+}