Added support for "format_first_only" new format flag

[SVN r10119]

changes.txt

@@ -3,6 +3,9 @@ FIXED:    Support for STLPort 4.1b6.
 FIXED:    Library calling convention for VC6 debug builds.
 FIXED:    Borland compiler support when using non-default calling convention.
 CHANGED:  Reorganised and simplified config setup.
+ADDED:    Support for replacing only the first occurance during a regex_merge 
+          using the format_first_only flag.
+CHANGED:  Removed use of deprecated api set_expression in one of the examples.
 
 Version 310:
 ADDED:    Support for static linking with VC6 + dynamic runtime.
@@ -259,5 +262,6 @@ BUG:      character sets don't function correctly when regbase::char_classes
 
 
 
+
 
 

example/snippets/regex_merge_example.cpp

@@ -57,8 +57,9 @@ void load_file(std::string& s, std::istream& is)
 
 int main(int argc, const char** argv)
 {
-   e1.set_expression(expression_text);
+   try{
-   e2.set_expression(pre_expression);
+   e1.assign(expression_text);
+   e2.assign(pre_expression);
    for(int i = 1; i < argc; ++i)
    {
       std::cout << "Processing file " << argv[i] << std::endl;
@@ -80,6 +81,9 @@ int main(int argc, const char** argv)
       boost::regex_merge(out, s.begin(), s.end(), e1, format_string);
       os << footer_text;
    }
+   }
+   catch(...)
+   { return -1; }
    return 0;
 }
 
@@ -129,3 +133,4 @@ const char* footer_text = "</PRE>\n</BODY>\n\n";
 
 
 
+

hl_ref.htm

@@ -15,8 +15,9 @@ content="C:\PROGRAM FILES\MICROSOFT OFFICE\OFFICE\html.dot">
 
 <table border="0" cellpadding="7" cellspacing="0" width="100%">
     <tr>
-        <td valign="top" width="50%"><h3><img src="../../c++boost.gif"
+        <td valign="top" width="50%"><h3><img
-        alt="C++ Boost" width="276" height="86"></h3>
+        src="../../c++boost.gif" alt="C++ Boost" width="276"
+        height="86"></h3>
         </td>
         <td valign="top" width="50%"><h3 align="center">Regex++,
         RegEx Class Reference. </h3>
@@ -112,8 +113,8 @@ are allowed. </p>
     </tr>
     <tr>
         <td valign="top" width="7%">&nbsp;</td>
-        <td valign="top" width="43%">RegEx(<b>const</b> RegEx&amp;
+        <td valign="top" width="43%">RegEx(<b>const</b>
-        o);</td>
+        RegEx&amp; o);</td>
         <td valign="top" width="42%">Copy constructor, all the
         properties of parameter <i>o</i> are copied.</td>
         <td valign="top" width="7%">&nbsp;</td>
@@ -196,7 +197,8 @@ are allowed. </p>
     <tr>
         <td valign="top" width="7%">&nbsp;</td>
         <td valign="top" width="43%"><b>bool</b> Match(<b>const</b>
-        <b>char</b>* p, <b>unsigned</b> <b>int</b> flags = match_default);</td>
+        <b>char</b>* p, <b>unsigned</b> <b>int</b> flags =
+        match_default);</td>
         <td valign="top" width="42%">Attempts to match the
         current expression against the text <i>p</i> using the
         match flags <i>flags</i> - see <a
@@ -221,7 +223,8 @@ are allowed. </p>
     <tr>
         <td valign="top" width="7%">&nbsp;</td>
         <td valign="top" width="43%"><b>bool</b> Search(<b>const</b>
-        <b>char</b>* p, <b>unsigned</b> <b>int</b> flags = match_default);</td>
+        <b>char</b>* p, <b>unsigned</b> <b>int</b> flags =
+        match_default);</td>
         <td valign="top" width="42%">Attempts to find a match for
         the current expression somewhere in the text <i>p</i>
         using the match flags <i>flags </i>- see <a
@@ -306,7 +309,8 @@ are allowed. </p>
         <td valign="top" width="7%">&nbsp;</td>
         <td valign="top" width="43%"><b>unsigned</b> <b>int</b>
         Grep(std::vector&lt;<b>unsigned int</b>&gt;&amp; v, <b>const</b>
-        <b>char</b>* p, <b>unsigned</b> <b>int</b> flags = match_default);</td>
+        <b>char</b>* p, <b>unsigned</b> <b>int</b> flags =
+        match_default);</td>
         <td valign="top" width="42%">Finds all matches of the
         current expression in the text <i>p</i> using the match
         flags <i>flags </i>- see <a
@@ -435,11 +439,16 @@ are allowed. </p>
         replace operation: searches through the string <i>in</i>
         for all occurrences of the current expression, for each
         occurrence replaces the match with the format string <i>fmt</i>.
-        Uses <i>flags</i> to determine what gets matched. If <i>copy</i>
+        Uses <i>flags</i> to determine what gets matched, and how
-        is true then all unmatched sections of input are copied
+        the format string should be treated. If <i>copy</i> is
-        unchanged to output. Returns the new string. See <a
+        true then all unmatched sections of input are copied
+        unchanged to output, if the flag <em>format_first_only</em>
+        is set then only the first occurance of the pattern found
+        is replaced. Returns the new string. See <a
         href="format_string.htm#format_string">also format string
-        syntax</a>.</td>
+        syntax</a>, <a href="template_class_ref.htm#match_type">match
+        flags</a> and <a
+        href="template_class_ref.htm#format_flags">format flags</a>.</td>
         <td valign="top" width="7%">&nbsp;</td>
     </tr>
     <tr>
@@ -451,18 +460,23 @@ are allowed. </p>
         replace operation: searches through the string <i>in</i>
         for all occurrences of the current expression, for each
         occurrence replaces the match with the format string <i>fmt</i>.
-        Uses <i>flags</i> to determine what gets matched. If <i>copy</i>
+        Uses <i>flags</i> to determine what gets matched, and how
-        is true then all unmatched sections of input are copied
+        the format string should be treated. If <i>copy</i> is
-        unchanged to output. Returns the new string. See <a
+        true then all unmatched sections of input are copied
+        unchanged to output, if the flag <em>format_first_only</em>
+        is set then only the first occurance of the pattern found
+        is replaced. Returns the new string. See <a
         href="format_string.htm#format_string">also format string
-        syntax</a>.</td>
+        syntax</a>, <a href="template_class_ref.htm#match_type">match
+        flags</a> and <a
+        href="template_class_ref.htm#format_flags">format flags</a>.</td>
         <td valign="top" width="7%">&nbsp;</td>
     </tr>
     <tr>
         <td>&nbsp;</td>
         <td valign="top"><b>unsigned</b> Split(std::vector&lt;std::string&gt;&amp;
-        v, std::string&amp; s, <b>unsigned</b> flags = match_default,
+        v, std::string&amp; s, <b>unsigned</b> flags =
-        <b>unsigned</b> max_count = ~0);</td>
+        match_default, <b>unsigned</b> max_count = ~0);</td>
         <td valign="top">Splits the input string and pushes each
         one onto the vector. If the expression contains no marked
         sub-expressions, then one string is outputted for each

include/boost/regex/detail/regex_format.hpp

@@ -33,7 +33,8 @@ enum format_flags_t{
    format_sed = match_max << 1,         // sed style replacement.
    format_perl = format_sed << 1,       // perl style replacement.
    format_no_copy = format_perl << 1,   // don't copy non-matching segments.
-   format_is_if = format_no_copy << 1   // internal use only.
+   format_first_only = format_no_copy << 1,   // Only replace first occurance.
+   format_is_if = format_first_only << 1   // internal use only.
 };
 
 namespace re_detail{
@@ -468,7 +469,7 @@ public:
          oi_assign(out, re_copy_out(*out, iterator(m[-1].first), iterator(m[-1].second)));
       oi_assign(out, _reg_format_aux(*out, m, f, flags, *pt));
       *last = m[-2].first;
-      return true;
+      return flags & format_first_only ? false : true;
    }
 };
 

test/regress/parse.cpp

@@ -82,6 +82,7 @@ flag_info flag_data[] = {
                           { BOOST_RE_STR("format_sed"), 10, format_sed, 3 },
                           { BOOST_RE_STR("format_perl"), 11, format_perl, 3 },
                           { BOOST_RE_STR("format_no_copy"), 14, format_no_copy, 3 },
+                          { BOOST_RE_STR("format_first_only"), 17, format_first_only, 3 },
 
                           { BOOST_RE_STR("REG_NO_POSIX_TEST"), 17, REG_NO_POSIX_TEST, 4 },
                           { BOOST_RE_STR("REG_UNICODE_ONLY"), 16, REG_UNICODE_ONLY, 4 },

test/regress/tests.txt

@@ -813,6 +813,13 @@ a+(b+) "...aaabb,,,ab*abbb?" $1 "...bb,,,b*bbb?"
 (a+)|(b+) "...aaabb,,,ab*abbb?" (?1A:B)C "...ACBC,,,ACBC*ACBC?" 
 (a+)|(b+) "...aaabb,,,ab*abbb?" ?1:B "...B,,,B*B?" 
 
+- match_default normal REG_EXTENDED REG_STARTEND REG_MERGE format_first_only
+; move to copying unmatched data, but replace first occurance only:
+a+ "...aaa,,," bbb "...bbb,,," 
+a+(b+) "...aaabb,,," $1 "...bb,,," 
+a+(b+) "...aaabb,,,ab*abbb?" $1 "...bb,,,ab*abbb?" 
+(a+)|(b+) "...aaabb,,,ab*abbb?" (?1A)(?2B) "...Abb,,,ab*abbb?" 
+
 ;
 ; changes to newline handling with 2.11:
 ;
@@ -871,3 +878,4 @@ a+(?#b+)b+ xaaabbba 1 7
 
 
 
+