- Added inplace configuration 're2c:yych:conversion'

diff --git a/re2c/CHANGELOG b/re2c/CHANGELOG
index 3ea6cfea576ea5735e509e62df4e1c33c49bb6b7..8f73b3f94d8e5e892c7d05b34365b162a3deb7f6 100644 (file)
--- a/re2c/CHANGELOG
+++ b/re2c/CHANGELOG
@@ -1,5 +1,6 @@
 Version 0.11.2 (????-??-??)
 ---------------------------
+- Add inplace configuration 're2c:yych:conversion'.
 - Fixed -u switch code generation.
 - Added ability to avoid defines and overwrite variable and label names.
 

diff --git a/re2c/code.cc b/re2c/code.cc
index 1d3af0970b67a99e6cdc3ece20227882746525a0..5640ff05f2b4db54f7feb7adc30e3e0b30e1a637 100644 (file)
--- a/re2c/code.cc
+++ b/re2c/code.cc
@@ -287,7 +287,7 @@ void genGoTo(std::ostream &o, uint ind, const State *from, const State *to, bool
 {
        if (readCh && from->label + 1 != to->label)
        {
-               o << indent(ind) << mapCodeName["yych"] << " = *" << mapCodeName["YYCURSOR"] << ";\n";
+               o << indent(ind) << mapCodeName["yych"] << " = " << yychConversion << "*" << mapCodeName["YYCURSOR"] << ";\n";
                readCh = false;
        }
 
@@ -300,7 +300,7 @@ void genIf(std::ostream &o, uint ind, const char *cmp, uint v, bool &readCh)
        o << indent(ind) << "if(";
        if (readCh)
        {
-               o << "(" << mapCodeName["yych"] << " = *" << mapCodeName["YYCURSOR"] << ")";
+               o << "(" << mapCodeName["yych"] << " = " << yychConversion << "*" << mapCodeName["YYCURSOR"] << ")";
                readCh = false;
        }
        else
@@ -342,11 +342,11 @@ static void need(std::ostream &o, uint ind, uint n, bool & readCh, bool bSetMark
 
        if (bSetMarker)
        {
-               o << indent(ind) << mapCodeName["yych"] << " = *(" << mapCodeName["YYMARKER"] << " = " << mapCodeName["YYCURSOR"] << ");\n";
+               o << indent(ind) << mapCodeName["yych"] << " = " << yychConversion << "*(" << mapCodeName["YYMARKER"] << " = " << mapCodeName["YYCURSOR"] << ");\n";
        }
        else
        {
-               o << indent(ind) << mapCodeName["yych"] << " = *" << mapCodeName["YYCURSOR"] << ";\n";
+               o << indent(ind) << mapCodeName["yych"] << " = " << yychConversion << "*" << mapCodeName["YYCURSOR"] << ";\n";
        }
        readCh = false;
 }
@@ -365,7 +365,7 @@ void Match::emit(std::ostream &o, uint ind, bool &readCh) const
        }
        else
        {
-               o << indent(ind) << mapCodeName["yych"] << " = *++" << mapCodeName["YYCURSOR"] << ";\n";
+               o << indent(ind) << mapCodeName["yych"] << " = " << yychConversion << "*++" << mapCodeName["YYCURSOR"] << ";\n";
                readCh = false;
        }
 
@@ -389,7 +389,7 @@ void Enter::emit(std::ostream &o, uint ind, bool &readCh) const
        else
        {
                /* we shouldn't need 'rule-following' protection here */
-               o << indent(ind) << mapCodeName["yych"] << " = *++" << mapCodeName["YYCURSOR"] << ";\n";
+               o << indent(ind) << mapCodeName["yych"] << " = " << yychConversion << "*++" << mapCodeName["YYCURSOR"] << ";\n";
                if (vUsedLabels.count(label))
                {
                        o << labelPrefix << label << ":\n";
@@ -413,7 +413,7 @@ void Initial::emit(std::ostream &o, uint ind, bool &readCh) const
                }
                else
                {
-                       o << indent(ind) << mapCodeName["yych"] << " = *++" << mapCodeName["YYCURSOR"] << ";\n";
+                       o << indent(ind) << mapCodeName["yych"] << " = " << yychConversion << "*++" << mapCodeName["YYCURSOR"] << ";\n";
                }
        }
 
@@ -464,11 +464,11 @@ void Save::emit(std::ostream &o, uint ind, bool &readCh) const
        {
                if (bUsedYYMarker)
                {
-                       o << indent(ind) << mapCodeName["yych"] << " = *(" << mapCodeName["YYMARKER"] << " = ++" << mapCodeName["YYCURSOR"] << ");\n";
+                       o << indent(ind) << mapCodeName["yych"] << " = " << yychConversion << "*(" << mapCodeName["YYMARKER"] << " = ++" << mapCodeName["YYCURSOR"] << ");\n";
                }
                else
                {
-                       o << indent(ind) << mapCodeName["yych"] << " = *++" << mapCodeName["YYCURSOR"] << ";\n";
+                       o << indent(ind) << mapCodeName["yych"] << " = " << yychConversion << "*++" << mapCodeName["YYCURSOR"] << ";\n";
                }
                readCh = false;
        }
@@ -528,7 +528,7 @@ void Accept::emit(std::ostream &o, uint ind, bool &readCh) const
 
                if (readCh) // shouldn't be necessary, but might become at some point
                {
-                       o << indent(ind) << mapCodeName["yych"] << " = *" << mapCodeName["YYCURSOR"] << ";\n";
+                       o << indent(ind) << mapCodeName["yych"] << " = " << yychConversion << "*" << mapCodeName["YYCURSOR"] << ";\n";
                        readCh = false;
                }
 
@@ -741,7 +741,7 @@ void Go::genSwitch(std::ostream &o, uint ind, const State *from, const State *ne
 
                if (readCh)
                {
-                       o << indent(ind) << "switch((" << mapCodeName["yych"] << " = *" << mapCodeName["YYCURSOR"] << ")) {\n";
+                       o << indent(ind) << "switch((" << mapCodeName["yych"] << " = " << yychConversion << "*" << mapCodeName["YYCURSOR"] << ")) {\n";
                        readCh = false;
                }
                else
@@ -894,7 +894,7 @@ void Go::genCpGoto(std::ostream &o, uint ind, const State *from, const State *ne
        
        if (readCh)
        {
-               sYych = "(" + mapCodeName["yych"] + " = *" + mapCodeName["YYCURSOR"] + ")";
+               sYych = "(" + mapCodeName["yych"] + " = " + yychConversion + "*" + mapCodeName["YYCURSOR"] + ")";
        }
        else
        {
@@ -1012,7 +1012,7 @@ void Go::genGoto(std::ostream &o, uint ind, const State *from, const State *next
                                        go.unmap(this, to);
                                        if (readCh)
                                        {
-                                               sYych = "(" + mapCodeName["yych"] + " = *" + mapCodeName["YYCURSOR"] + ")";
+                                               sYych = "(" + mapCodeName["yych"] + " = " + yychConversion + "*" + mapCodeName["YYCURSOR"] + ")";
                                        }
                                        else
                                        {
@@ -1691,6 +1691,19 @@ void Scanner::config(const Str& cfg, int num)
        {
                cGotoThreshold = num;
        }
+       else if (cfg.to_string() == "yych:conversion")
+       {
+               if (num)
+               {
+                       yychConversion  = "(";
+                       yychConversion += mapCodeName["YYCTYPE"];
+                       yychConversion += ")";
+               }
+               else
+               {
+                       yychConversion  = "";
+               }
+       }
        else
        {
                fatal("unrecognized configuration name or illegal integer value");
@@ -1748,6 +1761,10 @@ void Scanner::config(const Str& cfg, const Str& val)
        {
                labelPrefix = strVal;
        }
+       else if (cfg.to_string() == "yych:conversion")
+       {
+               yychConversion = mapCodeName["YYCTYPE"];
+       }
        else if (mapVariableKeys.find(cfg.to_string()) != mapVariableKeys.end())
     {
        if (bFirstPass && !mapCodeName.insert(

diff --git a/re2c/globals.h b/re2c/globals.h
index e4f527c6b7c5205cfb118616baec61d981d61699..f974a9360e628c94ec91cf099af45c9092dd7928 100644 (file)
--- a/re2c/globals.h
+++ b/re2c/globals.h
@@ -36,6 +36,7 @@ extern bool bUsedYYMarker;
 extern bool bUseStartLabel;
 extern std::string startLabelName;
 extern std::string labelPrefix;
+extern std::string yychConversion;
 extern uint maxFill;
 extern uint next_label;
 extern uint cGotoThreshold;

diff --git a/re2c/htdocs/manual.html b/re2c/htdocs/manual.html
index c6d441cf47f2c9599fada4e3f1c301ce59b8c20e..978779109c365d513bee5fcc3913362fcc706ec7 100755 (executable)
--- a/re2c/htdocs/manual.html
+++ b/re2c/htdocs/manual.html
@@ -385,6 +385,18 @@ placing a "/*!getstate:re2c */" comment.</dd>
 generation of jump tables rather than using nested if's and decision bitfields.
 The threshold is compared against a calculated estimation of if-s needed where 
 every used bitmap divides the threshold by 2.</dd>
+.TP
+<dt><i>re2c:yych:conversion</i> <b>=</b> 0 <b>;</b></dt>
+<dd>When the input uses signed characters and <b>-s</b> or <b>-b</b> switches are
+in effect re2c allows to automatically convert to the unsigned character type
+that is then necessary for its internal single character. When this setting
+is zero or an empty string the conversion is disabled. Using a non zero number
+the conversion is taken from <b>YYCTYPE</b>. If that is given by an inplace
+configuration that value is being used. Otherwise it will be <b>(YYCTYPE)</b>
+and changes to that configuration are  no longer possible. When this setting is
+a string the braces must be specified. Now assuming your input is a <b>char*</b>
+buffer and you are using above mentioned switches you can set <b>YYCTYPE</b> to
+<b>unsigned char</b> and this setting to either <b>1</b> or <b>"(unsigned char)"</b>.</dd>
 <dt><i>re2c:define:YYCTXMARKER</i> <b>=</b> YYCTXMARKER <b>;</b></dt>
 <dd>Allows to overwrite the define YYCTXMARKER and thus avoiding it by setting the
 value to the actual code needed.</dd>

diff --git a/re2c/main.cc b/re2c/main.cc
index 29548ea35332e8461715d0fa0eaff4578b7e5deb..3118b9233967a07fc9c8d7bf83313955f6751a6d 100644 (file)
--- a/re2c/main.cc
+++ b/re2c/main.cc
@@ -45,6 +45,7 @@ bool bUseYYFill    = true;
 
 std::string startLabelName;
 std::string labelPrefix("yy");
+std::string yychConversion("");
 uint maxFill = 1;
 uint next_label = 0;
 uint cGotoThreshold = 9;

diff --git a/re2c/re2c.1.in b/re2c/re2c.1.in
index 172a54fe5ab85c4eaa36f1c717efb180adf3a27c..0b5bf4f7acd753672c3fa118d2272f21f3493df8 100644 (file)
--- a/re2c/re2c.1.in
+++ b/re2c/re2c.1.in
@@ -457,6 +457,18 @@ generation of jump tables rather than using nested if's and decision bitfields.
 The threshold is compared against a calculated estimation of if-s needed where 
 every used bitmap divides the threshold by 2.
 .TP
+\fIre2c:yych:conversion\fP \fB=\fP 0 \fB;\fP
+When the input uses signed characters and \fB-s\fP or \fB-b\fP switches are 
+in effect re2c allows to automatically convert to the unsigned character type 
+that is then necessary for its internal single character. When this setting 
+is zero or an empty string the conversion is disabled. Using a non zero number
+the conversion is taken from \fBYYCTYPE\fP. If that is given by an inplace 
+configuration that value is being used. Otherwise it will be \fB(YYCTYPE)\fP 
+and changes to that configuration are  no longer possible. When this setting is
+a string the braces must be specified. Now assuming your input is a \fBchar*\fP
+buffer and you are using above mentioned switches you can set \fBYYCTYPE\fP to
+\fBunsigned char\fP and this setting to either \fB1\fP or \fB"(unsigned char)"\fP.
+.TP
 \fIre2c:define:YYCTXMARKER\fP \fB=\fP YYCTXMARKER \fB;\fP
 Allows to overwrite the define YYCTXMARKER and thus avoiding it by setting the
 value to the actual code needed.

diff --git a/re2c/test/config9.b.c b/re2c/test/config9.b.c
new file mode 100755 (executable)
index 0000000..65d902a
--- /dev/null
+++ b/re2c/test/config9.b.c
@@ -0,0 +1,184 @@
+/* Generated by re2c */
+#line 1 "config9.b.re"
+#include <cstring>
+#include <iostream>
+
+struct Scanner
+{
+       Scanner(char *txt)
+               : cur(txt), lim(txt + strlen(txt))
+       {
+       }
+       
+       char    *cur;
+       char    *lim;
+       char    *ptr;
+       char    *ctx;
+       char    *tok;
+};
+
+enum What
+{
+       UNEXPECTED,
+       KEYWORD,
+       NUMBER,
+       EOI
+};
+
+char * tokens[] = { "UNEXPECTED", "KEYWORD", "NUMBER", "EOI" };
+
+void fill(int)
+{
+}
+
+int scan(Scanner &s)
+{
+       char *cursor = s.cur;
+
+       if(cursor == s.lim)
+               return EOI;
+       
+std:
+       s.tok = cursor;
+
+{
+       static const unsigned char yybm[] = {
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+               128, 128, 128, 128, 128, 128, 128, 128, 
+               128, 128,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+                 0,   0,   0,   0,   0,   0,   0,   0, 
+       };
+
+#line 81 "<stdout>"
+       {
+               unsigned char curr;
+
+               if((s.lim - s.cur) < 3) fill(3);
+               curr = (unsigned char)*s.cur;
+               if(curr <= ' ') {
+                       if(curr <= 0x09) {
+                               if(curr <= 0x08) goto xx9;
+                               goto xx6;
+                       } else {
+                               if(curr <= 0x0A) goto xx8;
+                               if(curr <= 0x1F) goto xx9;
+                               goto xx6;
+                       }
+               } else {
+                       if(curr <= '9') {
+                               if(curr <= '/') goto xx9;
+                               goto xx4;
+                       } else {
+                               if(curr <= '`') goto xx9;
+                               if(curr >= 'c') goto xx9;
+                       }
+               }
+               s.ctx = s.cur + 1;
+               ++s.cur;
+               if((curr = (unsigned char)*s.cur) <= '/') goto xx3;
+               if(curr == '1') goto xx15;
+               if(curr <= '9') goto xx12;
+xx3:
+#line 68 "config9.b.re"
+               {
+               return UNEXPECTED;
+       }
+#line 115 "<stdout>"
+xx4:
+               ++s.cur;
+               curr = (unsigned char)*s.cur;
+               goto xx11;
+xx5:
+#line 58 "config9.b.re"
+               { return NUMBER;  }
+#line 123 "<stdout>"
+xx6:
+               ++s.cur;
+xx7:
+#line 61 "config9.b.re"
+               {
+               if(s.cur == s.lim)
+                       return EOI;
+               cursor = s.cur;
+               goto std;
+       }
+#line 134 "<stdout>"
+xx8:
+               curr = (unsigned char)*++s.cur;
+               goto xx7;
+xx9:
+               curr = (unsigned char)*++s.cur;
+               goto xx3;
+xx10:
+               ++s.cur;
+               if(s.lim == s.cur) fill(1);
+               curr = (unsigned char)*s.cur;
+xx11:
+               if(yybm[0+curr] & 128) {
+                       goto xx10;
+               }
+               goto xx5;
+xx12:
+               ++s.cur;
+               if(s.lim == s.cur) fill(1);
+               curr = (unsigned char)*s.cur;
+               if(curr <= '/') goto xx14;
+               if(curr <= '9') goto xx12;
+xx14:
+               s.cur = s.ctx;
+#line 57 "config9.b.re"
+               { return KEYWORD; }
+#line 160 "<stdout>"
+xx15:
+               ++s.cur;
+               if((curr = (unsigned char)*s.cur) <= '/') goto xx16;
+               if(curr <= '9') goto xx12;
+xx16:
+               s.cur = s.ctx;
+#line 56 "config9.b.re"
+               { return KEYWORD; }
+#line 169 "<stdout>"
+       }
+}
+#line 71 "config9.b.re"
+
+}
+
+int main(int,char**)
+{
+       Scanner s("a77 a1 b8 b1");
+       
+       int t, n = 0;
+       while ((t = scan(s)) != EOI)
+       {
+               std::cout << (++n) << ": " << tokens[t] << " = \""; std::cout.write(s.tok, s.cur-s.tok); std::cout << "\"" << std::endl;
+       }
+}

diff --git a/re2c/test/config9.b.re b/re2c/test/config9.b.re
new file mode 100755 (executable)
index 0000000..c8a338a
--- /dev/null
+++ b/re2c/test/config9.b.re
@@ -0,0 +1,83 @@
+#include <cstring>
+#include <iostream>
+
+struct Scanner
+{
+       Scanner(char *txt)
+               : cur(txt), lim(txt + strlen(txt))
+       {
+       }
+       
+       char    *cur;
+       char    *lim;
+       char    *ptr;
+       char    *ctx;
+       char    *tok;
+};
+
+enum What
+{
+       UNEXPECTED,
+       KEYWORD,
+       NUMBER,
+       EOI
+};
+
+char * tokens[] = { "UNEXPECTED", "KEYWORD", "NUMBER", "EOI" };
+
+void fill(int)
+{
+}
+
+int scan(Scanner &s)
+{
+       char *cursor = s.cur;
+
+       if(cursor == s.lim)
+               return EOI;
+       
+std:
+       s.tok = cursor;
+
+/*!re2c
+
+re2c:define:YYCTYPE            = "unsigned char";
+re2c:define:YYCURSOR   = s.cur;
+re2c:define:YYLIMIT            = s.lim;
+re2c:define:YYMARKER   = s.ptr;
+re2c:define:YYCTXMARKER        = s.ctx;
+re2c:define:YYFILL      = fill;
+
+re2c:variable:yych      = curr;
+re2c:labelprefix        = xx;
+
+re2c:yych:conversion    = 1;
+
+("a"|"b")/[1]    { return KEYWORD; }
+("a"|"b")/[0-9]+ { return KEYWORD; }
+[0-9]+           { return NUMBER;  }
+
+[ \t\n]
+       {
+               if(s.cur == s.lim)
+                       return EOI;
+               cursor = s.cur;
+               goto std;
+       }
+.
+       {
+               return UNEXPECTED;
+       }
+*/
+}
+
+int main(int,char**)
+{
+       Scanner s("a77 a1 b8 b1");
+       
+       int t, n = 0;
+       while ((t = scan(s)) != EOI)
+       {
+               std::cout << (++n) << ": " << tokens[t] << " = \""; std::cout.write(s.tok, s.cur-s.tok); std::cout << "\"" << std::endl;
+       }
+}