xref: /third_party/glib/patch/backport-regex-Do-not-use-JIT-when-using-unsupported-match-options.patch

From bec68b2d74853de5e23ee40c890433fa336ffbc5 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Marco=20Trevisan=20=28Trevi=C3=B1o=29?= <mail@3v1n0.net>
Date: Fri, 9 Sep 2022 18:30:15 +0200
Subject: [PATCH] glib/regex: Do not use JIT when using unsupported match
 options

Do not store jit status for regex unless during initial compilation.
After that, decide whether to use it depending on matching options.

In fact there are some matching options that are incompatible with JIT,
as the PCRE2 docs states:

  Setting PCRE2_ANCHORED or PCRE2_ENDANCHORED at match time is not
  supported by the just-in-time (JIT) compiler. If it is set, JIT
  matching is disabled and the interpretive code in pcre2_match() is
  run. Apart from PCRE2_NO_JIT (obviously), the remaining options are
  supported for JIT matching.

Fixes: GNOME/gtksourceview#283
---
 glib/gregex.c      | 38 ++++++++++++++++---------
 glib/tests/regex.c | 69 ++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 94 insertions(+), 13 deletions(-)

diff --git a/glib/gregex.c b/glib/gregex.c
index fe7473e628..220a1a11ac 100644
--- a/glib/gregex.c
+++ b/glib/gregex.c
@@ -201,6 +201,13 @@
                               PCRE2_NEWLINE_CRLF |   \
                               PCRE2_NEWLINE_ANYCRLF)
 
+/* Some match options are not supported when using JIT as stated in the
+ * pcre2jit man page under the 芦UNSUPPORTED OPTIONS AND PATTERN ITEMS禄 section:
+ *   https://www.pcre.org/current/doc/html/pcre2jit.html#SEC5
+ */
+#define G_REGEX_PCRE2_JIT_UNSUPPORTED_OPTIONS (PCRE2_ANCHORED | \
+                                               PCRE2_ENDANCHORED)
+
 #define G_REGEX_COMPILE_NEWLINE_MASK (G_REGEX_NEWLINE_CR      | \
                                       G_REGEX_NEWLINE_LF      | \
                                       G_REGEX_NEWLINE_CRLF    | \
@@ -869,7 +876,7 @@ recalc_match_offsets (GMatchInfo *match_info,
   return TRUE;
 }
 
-static void
+static JITStatus
 enable_jit_with_match_options (GRegex   *regex,
                                uint32_t  match_options)
 {
@@ -877,9 +884,13 @@ enable_jit_with_match_options (GRegex   *regex,
   uint32_t old_jit_options, new_jit_options;
 
   if (!(regex->orig_compile_opts & G_REGEX_OPTIMIZE))
-    return;
+    return JIT_STATUS_DISABLED;
+
   if (regex->jit_status == JIT_STATUS_DISABLED)
-    return;
+    return JIT_STATUS_DISABLED;
+
+  if (match_options & G_REGEX_PCRE2_JIT_UNSUPPORTED_OPTIONS)
+    return JIT_STATUS_DISABLED;
 
   old_jit_options = regex->jit_options;
   new_jit_options = old_jit_options | PCRE2_JIT_COMPLETE;
@@ -890,34 +901,34 @@ enable_jit_with_match_options (GRegex   *regex,
 
   /* no new options enabled */
   if (new_jit_options == old_jit_options)
-    return;
+    return regex->jit_status;
 
   retval = pcre2_jit_compile (regex->pcre_re, new_jit_options);
   switch (retval)
     {
     case 0: /* JIT enabled successfully */
-      regex->jit_status = JIT_STATUS_ENABLED;
       regex->jit_options = new_jit_options;
-      break;
+      return JIT_STATUS_ENABLED;
     case PCRE2_ERROR_NOMEMORY:
       g_debug ("JIT compilation was requested with G_REGEX_OPTIMIZE, "
                "but JIT was unable to allocate executable memory for the "
                "compiler. Falling back to interpretive code.");
-      regex->jit_status = JIT_STATUS_DISABLED;
-      break;
+      return JIT_STATUS_DISABLED;
     case PCRE2_ERROR_JIT_BADOPTION:
       g_debug ("JIT compilation was requested with G_REGEX_OPTIMIZE, "
                "but JIT support is not available. Falling back to "
                "interpretive code.");
-      regex->jit_status = JIT_STATUS_DISABLED;
+      return JIT_STATUS_DISABLED;
       break;
     default:
       g_debug ("JIT compilation was requested with G_REGEX_OPTIMIZE, "
                "but request for JIT support had unexpectedly failed (error %d). "
                "Falling back to interpretive code.", retval);
-      regex->jit_status = JIT_STATUS_DISABLED;
+      return JIT_STATUS_DISABLED;
       break;
     }
+
+  return regex->jit_status;
 }
 
 /**
@@ -1039,6 +1050,7 @@ gboolean
 g_match_info_next (GMatchInfo  *match_info,
                    GError     **error)
 {
+  JITStatus jit_status;
   gint prev_match_start;
   gint prev_match_end;
   uint32_t opts;
@@ -1060,8 +1072,8 @@ g_match_info_next (GMatchInfo  *match_info,
 
   opts = match_info->regex->match_opts | match_info->match_opts;
 
-  enable_jit_with_match_options (match_info->regex, opts);
-  if (match_info->regex->jit_status == JIT_STATUS_ENABLED)
+  jit_status = enable_jit_with_match_options (match_info->regex, opts);
+  if (jit_status == JIT_STATUS_ENABLED)
     {
       match_info->matches = pcre2_jit_match (match_info->regex->pcre_re,
                                              (PCRE2_SPTR8) match_info->string,
@@ -1727,7 +1739,7 @@ g_regex_new (const gchar         *pattern,
   regex->orig_compile_opts = compile_options;
   regex->match_opts = pcre_match_options;
   regex->orig_match_opts = match_options;
-  enable_jit_with_match_options (regex, regex->match_opts);
+  regex->jit_status = enable_jit_with_match_options (regex, regex->match_opts);
 
   return regex;
 }
diff --git a/glib/tests/regex.c b/glib/tests/regex.c
index 26844d63a7..2052ba0204 100644
--- a/glib/tests/regex.c
+++ b/glib/tests/regex.c
@@ -2334,6 +2334,67 @@ test_compile_errors (void)
   g_clear_error (&error);
 }
 
+static void
+test_jit_unsupported_matching_options (void)
+{
+  GRegex *regex;
+  GMatchInfo *info;
+  gchar *substring;
+
+  regex = g_regex_new ("(\\w+)#(\\w+)", G_REGEX_OPTIMIZE, G_REGEX_MATCH_DEFAULT, NULL);
+
+  g_assert_true (g_regex_match (regex, "aa#bb cc#dd", G_REGEX_MATCH_DEFAULT, &info));
+  g_assert_cmpint (g_match_info_get_match_count (info), ==, 3);
+  substring = g_match_info_fetch (info, 1);
+  g_assert_cmpstr (substring, ==, "aa");
+  g_clear_pointer (&substring, g_free);
+  substring = g_match_info_fetch (info, 2);
+  g_assert_cmpstr (substring, ==, "bb");
+  g_clear_pointer (&substring, g_free);
+  g_assert_true (g_match_info_next (info, NULL));
+  g_assert_cmpint (g_match_info_get_match_count (info), ==, 3);
+  substring = g_match_info_fetch (info, 1);
+  g_assert_cmpstr (substring, ==, "cc");
+  g_clear_pointer (&substring, g_free);
+  substring = g_match_info_fetch (info, 2);
+  g_assert_cmpstr (substring, ==, "dd");
+  g_clear_pointer (&substring, g_free);
+  g_assert_false (g_match_info_next (info, NULL));
+  g_match_info_free (info);
+
+  g_assert_true (g_regex_match (regex, "aa#bb cc#dd", G_REGEX_MATCH_ANCHORED, &info));
+  g_assert_cmpint (g_match_info_get_match_count (info), ==, 3);
+  substring = g_match_info_fetch (info, 1);
+  g_assert_cmpstr (substring, ==, "aa");
+  g_clear_pointer (&substring, g_free);
+  substring = g_match_info_fetch (info, 2);
+  g_assert_cmpstr (substring, ==, "bb");
+  g_clear_pointer (&substring, g_free);
+  g_assert_false (g_match_info_next (info, NULL));
+  g_match_info_free (info);
+
+  g_assert_true (g_regex_match (regex, "aa#bb cc#dd", G_REGEX_MATCH_DEFAULT, &info));
+  g_assert_cmpint (g_match_info_get_match_count (info), ==, 3);
+  substring = g_match_info_fetch (info, 1);
+  g_assert_cmpstr (substring, ==, "aa");
+  g_clear_pointer (&substring, g_free);
+  substring = g_match_info_fetch (info, 2);
+  g_assert_cmpstr (substring, ==, "bb");
+  g_clear_pointer (&substring, g_free);
+  g_assert_true (g_match_info_next (info, NULL));
+  g_assert_cmpint (g_match_info_get_match_count (info), ==, 3);
+  substring = g_match_info_fetch (info, 1);
+  g_assert_cmpstr (substring, ==, "cc");
+  g_clear_pointer (&substring, g_free);
+  substring = g_match_info_fetch (info, 2);
+  g_assert_cmpstr (substring, ==, "dd");
+  g_clear_pointer (&substring, g_free);
+  g_assert_false (g_match_info_next (info, NULL));
+  g_match_info_free (info);
+
+  g_regex_unref (regex);
+}
+
 int
 main (int argc, char *argv[])
 {
@@ -2352,6 +2413,7 @@ main (int argc, char *argv[])
   g_test_add_func ("/regex/explicit-crlf", test_explicit_crlf);
   g_test_add_func ("/regex/max-lookbehind", test_max_lookbehind);
   g_test_add_func ("/regex/compile-errors", test_compile_errors);
+  g_test_add_func ("/regex/jit-unsupported-matching", test_jit_unsupported_matching_options);
 
   /* TEST_NEW(pattern, compile_opts, match_opts) */
   TEST_NEW("[A-Z]+", G_REGEX_CASELESS | G_REGEX_EXTENDED | G_REGEX_OPTIMIZE, G_REGEX_MATCH_NOTBOL | G_REGEX_MATCH_PARTIAL);
@@ -2488,6 +2550,7 @@ main (int argc, char *argv[])
   TEST_MATCH_SIMPLE("a", "ab", 0, G_REGEX_MATCH_ANCHORED, TRUE);
   TEST_MATCH_SIMPLE("a", "a", G_REGEX_CASELESS, 0, TRUE);
   TEST_MATCH_SIMPLE("a", "A", G_REGEX_CASELESS, 0, TRUE);
+  TEST_MATCH_SIMPLE("\\C\\C", "ab", G_REGEX_OPTIMIZE | G_REGEX_RAW, 0, TRUE);
   /* These are needed to test extended properties. */
   TEST_MATCH_SIMPLE(AGRAVE, AGRAVE, G_REGEX_CASELESS, 0, TRUE);
   TEST_MATCH_SIMPLE(AGRAVE, AGRAVE_UPPER, G_REGEX_CASELESS, 0, TRUE);
@@ -2947,6 +3010,12 @@ main (int argc, char *argv[])
   TEST_REPLACE("\\S+", "hello world", 0, "\\U-\\0-", "-HELLO- -WORLD-");
   TEST_REPLACE(".", "a", 0, "\\A", NULL);
   TEST_REPLACE(".", "a", 0, "\\g", NULL);
+  TEST_REPLACE_OPTIONS("(\\w+)#(\\w+)", "aa#bb cc#dd", 0, "\\2#\\1", "bb#aa dd#cc",
+                       G_REGEX_OPTIMIZE|G_REGEX_MULTILINE|G_REGEX_CASELESS,
+                       0);
+  TEST_REPLACE_OPTIONS("(\\w+)#(\\w+)", "aa#bb cc#dd", 0, "\\2#\\1", "bb#aa cc#dd",
+                       G_REGEX_OPTIMIZE|G_REGEX_MULTILINE|G_REGEX_CASELESS,
+                       G_REGEX_MATCH_ANCHORED);
 
   /* TEST_REPLACE_LIT(pattern, string, start_position, replacement, expected) */
   TEST_REPLACE_LIT("a", "ababa", 0, "A", "AbAbA");
-- 
GitLab