Report kRegexpBadPerlOp for look-behind assertions.

Fixes #468. Change-Id: I9a72db0bbb9a04e5081cc1f0f94399476b63d1c3 Reviewed-on: https://code-review.googlesource.com/c/re2/+/62330 Reviewed-by: Perry Lorier <[email protected]> Reviewed-by: Paul Wankadia <[email protected]>
google · Jan 2, 2024 · c042630 · c042630
1 parent c217103
commit c042630
Show file tree

Hide file tree

Showing 2 changed files with 37 additions and 0 deletions.
diff --git a/re2/parse.cc b/re2/parse.cc
@@ -2070,6 +2070,17 @@ bool Regexp::ParseState::ParsePerlFlags(absl::string_view* s) {
     return false;
   }
 
+  // Check for look-around assertions. This is NOT because we support them! ;)
+  // As per https://github.com/google/re2/issues/468, we really want to report
+  // kRegexpBadPerlOp (not kRegexpBadNamedCapture) for look-behind assertions.
+  // Additionally, it would be nice to report not "(?<", but "(?<=" or "(?<!".
+  if ((t.size() > 3 && (t[2] == '=' || t[2] == '!')) ||
+      (t.size() > 4 && t[2] == '<' && (t[3] == '=' || t[3] == '!'))) {
+    status_->set_code(kRegexpBadPerlOp);
+    status_->set_error_arg(absl::string_view(t.data(), t[2] == '<' ? 4 : 3));
+    return false;
+  }
+
   // Check for named captures, first introduced in Python's regexp library.
   // As usual, there are three slightly different syntaxes:
   //

diff --git a/re2/testing/parse_test.cc b/re2/testing/parse_test.cc
@@ -532,4 +532,30 @@ TEST(NamedCaptures, ErrorArgs) {
   EXPECT_EQ(status.error_arg(), "(?<space bar>");
 }
 
+// Test that look-around error args are correct.
+TEST(LookAround, ErrorArgs) {
+  RegexpStatus status;
+  Regexp* re;
+
+  re = Regexp::Parse("(?=foo).*", Regexp::LikePerl, &status);
+  EXPECT_TRUE(re == NULL);
+  EXPECT_EQ(status.code(), kRegexpBadPerlOp);
+  EXPECT_EQ(status.error_arg(), "(?=");
+
+  re = Regexp::Parse("(?!foo).*", Regexp::LikePerl, &status);
+  EXPECT_TRUE(re == NULL);
+  EXPECT_EQ(status.code(), kRegexpBadPerlOp);
+  EXPECT_EQ(status.error_arg(), "(?!");
+
+  re = Regexp::Parse("(?<=foo).*", Regexp::LikePerl, &status);
+  EXPECT_TRUE(re == NULL);
+  EXPECT_EQ(status.code(), kRegexpBadPerlOp);
+  EXPECT_EQ(status.error_arg(), "(?<=");
+
+  re = Regexp::Parse("(?<!foo).*", Regexp::LikePerl, &status);
+  EXPECT_TRUE(re == NULL);
+  EXPECT_EQ(status.code(), kRegexpBadPerlOp);
+  EXPECT_EQ(status.error_arg(), "(?<!");
+}
+
 }  // namespace re2