lexer.ll: remove unused patterns, avoid backing up (issue4871041)

[k-ohara5a5a]

Reviewers: ,

Message:
While we are cleaning up warnings, here is a patch for the lexer
warnings.

This is all code cleanup, no behavior change, so I don't think a tracker
issue is appropriate.  Regression tests are unchanged.


http://codereview.appspot.com/4871041/diff/3001/lily/lexer.ll
File lily/lexer.ll (right):

http://codereview.appspot.com/4871041/diff/3001/lily/lexer.ll#newcode23
lily/lexer.ll:23: backup rules
This comment instruction was violated, so this patch brings the code
back in compliance, but it is awkward.

Why are we avoiding backup states?

http://codereview.appspot.com/4871041/diff/3001/lily/lexer.ll#newcode265
lily/lexer.ll:265: <version>{ANY_CHAR}    {
Include \n in the pattern, to avoid warning about fall back to the
default match.  (The \n should never match, because the rules that leave
us in <version> never leave \n on the input stream.)

http://codereview.appspot.com/4871041/diff/3001/lily/lexer.ll#newcode309
lily/lexer.ll:309: <incl>\\{BLACK}*{WHITE}? { /* got the include
identifier */
Accept \include foo<<EOF>> without backing up

http://codereview.appspot.com/4871041/diff/3001/lily/lexer.ll#newcode621
lily/lexer.ll:621: if (YY_START == longcomment)
Error reporting moved from line 287, to avoid warning for two matching
rules for <<EOF>>

http://codereview.appspot.com/4871041/diff/3001/lily/lexer.ll#newcode653
lily/lexer.ll:653: {REAL}|-{UNSIGNED}   {
I'll add comment // backup rule, -123 versus -123.

Description:
lexer.ll: remove unused patterns, avoid backing up

This removes the rest of the warnings from Lilypond `make bin`
... except for a type conversion warning in out/parser.cc generated
by Bison -- I don't see a way to remove that warning.

Please review this at http://codereview.appspot.com/4871041/

Affected files:
  M lily/lexer.ll


Index: lily/lexer.ll
diff --git a/lily/lexer.ll b/lily/lexer.ll
index  
83a7940b6938dc634083fe9936ee389635713412..bd3a2abb6915f3c6e564bc7db10df86ed4842b0e  
100644
--- a/lily/lexer.ll
+++ b/lily/lexer.ll
@@ -135,7 +135,6 @@ PUNCT               [?!:'`]
 ACCENT         \\[`'"^]
 NATIONAL       [\001-\006\021-\027\031\036]
 TEX            {AA}|-|{PUNCT}|{ACCENT}|{NATIONAL}
-WORD           {A}{AN}*
 DASHED_WORD            {A}({AN}|-)*
 DASHED_KEY_WORD                \\{DASHED_WORD}

@@ -148,7 +147,6 @@ E_UNSIGNED  \\{N}+
 FRACTION       {N}+\/{N}+
 INT            -?{UNSIGNED}
 REAL           ({INT}\.{N}*)|(-?\.{N}+)
-KEYWORD                \\{WORD}
 WHITE          [ \n\t\f\r]
 HORIZONTALWHITE                [ \t]
 BLACK          [^ \n\t\f\r]
@@ -165,7 +163,7 @@ BOM_UTF8    \357\273\277


 <*>\r            {
-       // windows-suck-suck-suck
+       // swallow and ignore carriage returns
 }

 <extratoken>{ANY_CHAR}   {
@@ -264,15 +262,15 @@ BOM_UTF8  \357\273\277
 	this->here_input ().get_source_file ()->set_line (here_input ().start (),  
i);
 }

-<version>.       {
+<version>{ANY_CHAR}      {
        LexerError (_ ("quoted string expected after \\version").c_str ());
        yy_pop_state ();
 }
-<sourcefilename>.        {
+<sourcefilename>{ANY_CHAR}       {
        LexerError (_ ("quoted string expected after \\sourcefilename").c_str 
());
        yy_pop_state ();
 }
-<sourcefileline>.        {
+<sourcefileline>{ANY_CHAR}       {
        LexerError (_ ("integer expected after \\sourcefileline").c_str ());
        yy_pop_state ();
 }
@@ -285,12 +283,6 @@ BOM_UTF8   \357\273\277
        "%"+"}"             {
                yy_pop_state ();
        }
-       <<EOF>>     {
-               LexerError (_ ("EOF found inside a comment").c_str ());
-		is_main_input_ = false; // should be safe , can't have \include in  
--safe.
-               if (! close_input ())
-                 yyterminate (); // can't move this, since it actually rets a 
YY_NULL
-       }
 }


@@ -314,7 +306,7 @@ BOM_UTF8    \357\273\277
        new_input (s, sources_);
        yy_pop_state ();
 }
-<incl>\\{BLACK}*{WHITE} { /* got the include identifier */
+<incl>\\{BLACK}*{WHITE}? { /* got the include identifier */
        string s = YYText () + 1;
        strip_trailing_white (s);
        if (s.length () && (s[s.length () - 1] == ';'))
@@ -333,7 +325,7 @@ BOM_UTF8    \357\273\277
            scm_display (sid, err);
          }
 }
-<incl>\"[^"]*   { // backup rule
+<incl,version,sourcefilename>\"[^"]*   { // backup rule
        error (_ ("end quote missing"));
        exit (1);
 }
@@ -420,7 +412,13 @@ BOM_UTF8   \357\273\277
                yylval.scm =  scan_fraction (YYText ());
                return FRACTION;
        }
-
+       {UNSIGNED}\/    {
+               // handle two tokens in "5/ " without using backup states
+               yylval.i = String_convert::dec2int (string (YYText ()));
+               unput ('/');
+               char_count_stack_.back ()--;
+               return UNSIGNED;
+       }
        {DIGIT}         {
                yylval.i = String_convert::dec2int (string (YYText ()));
                return DIGIT;
@@ -465,6 +463,13 @@ BOM_UTF8   \357\273\277
                yylval.scm =  scan_fraction (YYText ());
                return FRACTION;
        }
+       {UNSIGNED}\/    {
+               // report erroneous input "2/ 3" without backing up
+               yylval.i = String_convert::dec2int (string (YYText ()));
+               unput ('/');
+               char_count_stack_.back ()--;
+               return UNSIGNED;
+       }
        {UNSIGNED}              {
                yylval.i = String_convert::dec2int (string (YYText ()));
                return UNSIGNED;
@@ -505,6 +510,13 @@ BOM_UTF8   \357\273\277
                yylval.scm =  scan_fraction (YYText ());
                return FRACTION;
        }
+       {UNSIGNED}\/    {
+               // report erroneous input "7/c" without backing up
+               yylval.i = String_convert::dec2int (string (YYText ()));
+               unput ('/');
+               char_count_stack_.back ()--;
+               return UNSIGNED;
+       }
        {UNSIGNED}              {
                yylval.i = String_convert::dec2int (string (YYText ()));
                return UNSIGNED;
@@ -606,7 +618,14 @@ BOM_UTF8   \357\273\277
 }

 <*><<EOF>> {
-       if (is_main_input_)
+       if (YY_START == longcomment)
+       {
+               LexerError (_ ("EOF found inside a comment").c_str ());
+		is_main_input_ = false; // should be safe , can't have \include in  
--safe.
+               if (!close_input ())
+                 yyterminate (); // can't move this, since it actually rets a 
YY_NULL
+       }
+       else if (is_main_input_)
        {
                /* 2 = init.ly + current file.
                   > because we're before closing, but is_main_input_ should
@@ -631,13 +650,7 @@ BOM_UTF8   \357\273\277
        }
 }

-{WORD} {
-       return scan_bare_word (YYText ());
-}
-{KEYWORD}      {
-       return scan_escaped_word (YYText () + 1);
-}
-{REAL}         {
+{REAL}|-{UNSIGNED}     {
        Real r;
        int cnv = sscanf (YYText (), "%lf", &r);
        assert (cnv == 1);
@@ -646,6 +659,12 @@ BOM_UTF8   \357\273\277
        yylval.scm = scm_from_double (r);
        return REAL;
 }
+-\.    {
+       /* "-." is an error in context where REAL is allowed;
+          match it to report the error without backing up.
+       */
+       return '.';
+}

 {UNSIGNED}     {
        yylval.i = String_convert::dec2int (string (YYText ()));