Handle escaped character for substring rule

ppomes · Aug 27, 2024 · d0aaa41 · d0aaa41
1 parent 55b587a
commit d0aaa41
Show file tree

Hide file tree

Showing 4 changed files with 40 additions and 11 deletions.
diff --git a/main/myanon.c b/main/myanon.c
@@ -77,6 +77,34 @@ char *mystrcpy(char *dest, const char *src, size_t size)
     return dest;
 }
 
+char *myescapedstrcpy(char *dest, const char *src, size_t size)
+{
+    size_t srccount = 0;
+    size_t dstcount = 0;
+    short backslash = 0;
+
+    memset(dest, 0, size);
+    while (src[srccount] != '\0' && dstcount < size - 1)
+    {
+        if (src[srccount] == '\\')
+        {
+            backslash++;
+
+            if (backslash % 2 == 0)
+            {
+                backslash = 0;
+                dest[dstcount++] = src[srccount++];
+            }
+        }
+        else
+        {
+            dest[dstcount++] = src[srccount++];
+        }
+    }
+
+    return dest;
+}
+
 unsigned long get_ts_in_ms()
 {
     struct timeval tv;
@@ -203,7 +231,7 @@ anonymized_res_st anonymize_token(bool quoted, anon_base_st *config, char *token
     case AM_SUBSTRING:
         res_st.len = MIN(worktokenlen, config->len);
         DEBUG_MSG("%d, %d, %d", worktokenlen, config->len, res_st.len)
-        mystrcpy((char *)&(res_st.data[0]), worktoken, res_st.len + 1);
+        myescapedstrcpy((char *)&(res_st.data[0]), worktoken, res_st.len + 1);
         break;
 
 #ifdef HAVE_PYTHON

diff --git a/main/myanon.h b/main/myanon.h
@@ -201,6 +201,7 @@ EXTERN unsigned long anon_time;
 
 /* some safe malloc/strpcy wrappers */
 void *mymalloc(size_t size);
+char *myescapedstrcpy(char *dest, const char *src, size_t size);
 char *mystrcpy(char *dest, const char *src, size_t size);
 
 /* function to anonymize a single field 'token' which length is 'tokenlen'

diff --git a/tests/test_substring.sql b/tests/test_substring.sql
@@ -1,8 +1,8 @@
--- MySQL dump 10.13  Distrib 8.3.0, for macos14.2 (arm64)
+-- MySQL dump 10.13  Distrib 8.0.37, for Linux (x86_64)
 --
--- Host: 127.0.0.1    Database: minimal
+-- Host: localhost    Database: test_substring
 -- ------------------------------------------------------
--- Server version	8.3.0
+-- Server version	8.0.37-0ubuntu0.24.04.1
 
 /*!40101 SET @OLD_CHARACTER_SET_CLIENT=@@CHARACTER_SET_CLIENT */;
 /*!40101 SET @OLD_CHARACTER_SET_RESULTS=@@CHARACTER_SET_RESULTS */;
@@ -33,7 +33,7 @@ CREATE TABLE `minimal_table` (
 
 LOCK TABLES `minimal_table` WRITE;
 /*!40000 ALTER TABLE `minimal_table` DISABLE KEYS */;
-INSERT INTO `minimal_table` VALUES ('minimaal zo veel'),('🥳');
+INSERT INTO `minimal_table` VALUES ('minimaal zo veel'),('🥳'),('	test escape');
 /*!40000 ALTER TABLE `minimal_table` ENABLE KEYS */;
 UNLOCK TABLES;
 /*!40103 SET TIME_ZONE=@OLD_TIME_ZONE */;
@@ -46,4 +46,4 @@ UNLOCK TABLES;
 /*!40101 SET COLLATION_CONNECTION=@OLD_COLLATION_CONNECTION */;
 /*!40111 SET SQL_NOTES=@OLD_SQL_NOTES */;
 
--- Dump completed on 2024-08-13 15:47:46
+-- Dump completed on 2024-08-27 14:58:38
diff --git a/tests/test_substring_anon.sql b/tests/test_substring_anon.sql
@@ -1,8 +1,8 @@
--- MySQL dump 10.13  Distrib 8.3.0, for macos14.2 (arm64)
+-- MySQL dump 10.13  Distrib 8.0.37, for Linux (x86_64)
 --
--- Host: 127.0.0.1    Database: minimal
+-- Host: localhost    Database: test_substring
 -- ------------------------------------------------------
--- Server version	8.3.0
+-- Server version	8.0.37-0ubuntu0.24.04.1
 
 /*!40101 SET @OLD_CHARACTER_SET_CLIENT=@@CHARACTER_SET_CLIENT */;
 /*!40101 SET @OLD_CHARACTER_SET_RESULTS=@@CHARACTER_SET_RESULTS */;
@@ -33,7 +33,7 @@ CREATE TABLE `minimal_table` (
 
 LOCK TABLES `minimal_table` WRITE;
 /*!40000 ALTER TABLE `minimal_table` DISABLE KEYS */;
-INSERT INTO `minimal_table` VALUES ('minim'),('🥳');
+INSERT INTO `minimal_table` VALUES ('minim'),('🥳'),('	test');
 /*!40000 ALTER TABLE `minimal_table` ENABLE KEYS */;
 UNLOCK TABLES;
 /*!40103 SET TIME_ZONE=@OLD_TIME_ZONE */;
@@ -46,4 +46,4 @@ UNLOCK TABLES;
 /*!40101 SET COLLATION_CONNECTION=@OLD_COLLATION_CONNECTION */;
 /*!40111 SET SQL_NOTES=@OLD_SQL_NOTES */;
 
--- Dump completed on 2024-08-13 15:47:46
+-- Dump completed on 2024-08-27 14:58:38