HPCC-32795 Additional optimizations to replaceString

rtl/eclrtl/eclrtl.cpp

@@ -6066,7 +6066,7 @@ void rtlAddExceptionTag(StringBuffer & errorText, const char * tag, const char *
 void rtlSubstituteEmbeddedScript(size32_t &__lenResult, char * &__result, size32_t scriptChars, const char *script, size32_t outFieldsChars, const char *outFields, size32_t searchChars, const char *search)
 {
     StringBuffer result;
-    ::replaceString(result, rtlUtf8Size(scriptChars, script), script, rtlUtf8Size(searchChars, search), search, rtlUtf8Size(outFieldsChars, outFields), outFields);
+    ::replaceString(result, rtlUtf8Size(scriptChars, script), script, rtlUtf8Size(searchChars, search), search, rtlUtf8Size(outFieldsChars, outFields), outFields, true);
     __lenResult = result.lengthUtf8();
     __result = result.detach();
 }

system/jlib/jstring.cpp

@@ -935,13 +935,10 @@ StringBuffer & StringBuffer::replace(char oldChar, char newChar)
 }
 
 // Copy source to result, replacing all occurrences of "oldStr" with "newStr"
-StringBuffer &replaceString(StringBuffer & result, size_t lenSource, const char *source, size_t lenOldStr, const char* oldStr, size_t lenNewStr, const char* newStr)
+bool replaceString(StringBuffer & result, size_t lenSource, const char *source, size_t lenOldStr, const char* oldStr, size_t lenNewStr, const char* newStr, bool forceCopy)
 {
     if (lenOldStr && lenSource >= lenOldStr)
     {
-        // Avoid allocating an unnecessarly large buffer and match the source string
-        result.ensureCapacity(lenSource);
-
         size_t offset = 0;
         size_t lastCopied = 0;
         size_t maxOffset = lenSource - lenOldStr + 1;
@@ -951,6 +948,10 @@ StringBuffer &replaceString(StringBuffer & result, size_t lenSource, const char
             if (unlikely(source[offset] == firstChar)
                 && unlikely((lenOldStr == 1) || memcmp(source + offset, oldStr, lenOldStr)==0))
             {
+                // Wait to allocate memory until a match is found
+                if (!lastCopied)
+                    result.ensureCapacity(lenSource); // Avoid allocating an unnecessarly large buffer and match the source string
+
                 // If lastCopied matches the offset nothing is appended, but we can avoid a test for offset == lastCopied
                 result.append(offset - lastCopied, source + lastCopied);
                 result.append(lenNewStr, newStr);
@@ -960,13 +961,16 @@ StringBuffer &replaceString(StringBuffer & result, size_t lenSource, const char
             else
                 offset++;
         }
-        // Append the remaining characters
-        result.append(lenSource - lastCopied, source + lastCopied);
+
+        if (lastCopied || forceCopy)
+            result.append(lenSource - lastCopied, source + lastCopied); // Append the remaining characters
+
+        return lastCopied;
     }
-    else
+    else if (forceCopy)
         result.append(lenSource, source); // Search string does not fit in source or is empty
 
-    return result;
+    return false;
 }
 
 StringBuffer &replaceVariables(StringBuffer & result, const char *source, bool exceptions, IVariableSubstitutionHelper *helper, const char* delim, const char* term)
@@ -1056,13 +1060,29 @@ StringBuffer &replaceStringNoCase(StringBuffer & result, size_t lenSource, const
 // this method will replace all occurrences of "oldStr" with "newStr"
 StringBuffer & StringBuffer::replaceString(const char* oldStr, const char* newStr)
 {
-    if (curLen)
+    if (curLen && oldStr)
     {
-        StringBuffer temp;
-        size_t oldlen = oldStr ? strlen(oldStr) : 0;
+        size_t oldlen = strlen(oldStr);
+        if (oldlen > curLen)
+            return *this;
+
         size_t newlen = newStr ? strlen(newStr) : 0;
-        ::replaceString(temp, curLen, buffer, oldlen, oldStr, newlen, newStr);
-        swapWith(temp);
+        if (oldlen == curLen)
+        {
+            if (memcmp(buffer, oldStr, oldlen) == 0)
+            {
+                if (newlen > curLen)
+                    ensureCapacity(newlen);
+                memcpy(buffer, newStr, newlen);
+                curLen = newlen;
+            }
+        }
+        else
+        {
+            StringBuffer temp;
+            if(::replaceString(temp, curLen, buffer, oldlen, oldStr, newlen, newStr))
+                swapWith(temp);
+        }
     }
     return *this;
 }

system/jlib/jstring.hpp

@@ -407,7 +407,7 @@ extern jlib_decl void decodeXML(ISimpleReadStream &in, StringBuffer &out, unsign
 extern jlib_decl int utf8CharLen(unsigned char ch);
 extern jlib_decl int utf8CharLen(const unsigned char *ch, unsigned maxsize = (unsigned)-1);
 
-extern jlib_decl StringBuffer &replaceString(StringBuffer & result, size_t lenSource, const char *source, size_t lenOldStr, const char* oldStr, size_t lenNewStr, const char* newStr);
+extern jlib_decl bool replaceString(StringBuffer & result, size_t lenSource, const char *source, size_t lenOldStr, const char* oldStr, size_t lenNewStr, const char* newStr, bool forceCopy = false);
 
 interface IVariableSubstitutionHelper
 {

testing/unittests/jlibtests.cpp

@@ -913,6 +913,26 @@ void testEncodeCSVColumn()
         source.set("abbabab");
         source.replaceString("ab", "xxx");
         CPPUNIT_ASSERT_EQUAL_STR("xxxbxxxxxx", source.str());
+
+        // Search string has same length as source string and matches
+        source.set("ababab");
+        source.replaceString("ababab", "xxxxxx");
+        CPPUNIT_ASSERT_EQUAL_STR("xxxxxx", source.str());
+
+        // Search string has same length as source string and replace is smaller than source
+        source.set("ababab");
+        source.replaceString("ababab", "xxx");
+        CPPUNIT_ASSERT_EQUAL_STR("xxx", source.str());
+
+        // Search string has same length as source string and replace is larger than source
+        source.set("ababab");
+        source.replaceString("ababab", "xxxxxxxxx");
+        CPPUNIT_ASSERT_EQUAL_STR("xxxxxxxxx", source.str());
+
+        // Search string has same length as source string and does not match
+        source.set("ababab");
+        source.replaceString("ababac", "xxxxxx");
+        CPPUNIT_ASSERT_EQUAL_STR("ababab", source.str());
     }
 };