Tweaks for NUL chars.

diff --git a/dfa.c b/dfa.c
index f1be016383ef775a44928a4012a5642fa9e29c6c..b8250ba73b93c0663731610dfdae0792a1ec2a0a 100644 (file)
--- a/dfa.c
+++ b/dfa.c
@@ -209,9 +209,9 @@ int state[];
 
     {
     register int i, ec;
-    int out_char_set[CSIZE + 1];
+    int out_char_set[CSIZE];
 
-    for ( i = 1; i <= csize; ++i )
+    for ( i = (uses_NUL ? 0 : 1); i < csize; ++i )
        {
        ec = abs( ecgroup[i] );
        out_char_set[i] = state[ec];
@@ -222,7 +222,7 @@ int state[];
     list_character_set( file, out_char_set );
 
     /* now invert the members of the set to get the jam transitions */
-    for ( i = 1; i <= csize; ++i )
+    for ( i = (uses_NUL ? 0 : 1); i < csize; ++i )
        out_char_set[i] = ! out_char_set[i];
 
     fprintf( file, "\n jam-transitions: EOF " );
@@ -406,14 +406,23 @@ ntod()
 
     {
     int *accset, ds, nacc, newds;
-    int duplist[CSIZE + 1], sym, hashval, numstates, dsize;
-    int targfreq[CSIZE + 1], targstate[CSIZE + 1], state[CSIZE + 1];
+    int sym, hashval, numstates, dsize;
     int *nset, *dset;
     int targptr, totaltrans, i, comstate, comfreq, targ;
     int *epsclosure(), snstods(), symlist[CSIZE + 1];
     int num_start_states;
     int todo_head, todo_next;
 
+    /* note that the following are indexed by *equivalence classes*
+     * and not by characters.  Since equivalence classes are indexed
+     * beginning with 1, even if the scanner accepts NUL's, this
+     * means that (since every character is potentially in its own
+     * equivalence class) these arrays must have room for indices
+     * from 1 to CSIZE, so their size must be CSIZE + 1.
+     */
+    int duplist[CSIZE + 1], state[CSIZE + 1];
+    int targfreq[CSIZE + 1], targstate[CSIZE + 1];
+
     /* this is so find_table_space(...) will know where to start looking in
      * chk/nxt for unused records for space to put in the state
      */
@@ -919,7 +928,8 @@ int symlist[];
        if ( tch != SYM_EPSILON )
            {
            if ( tch < -lastccl || tch > csize )
-               flexfatal( "bad transition character detected in sympartition()" );
+               flexfatal(
+                   "bad transition character detected in sympartition()" );
 
            if ( tch > 0 )
                { /* character transition */

diff --git a/gen.c b/gen.c
index fdc4febf5187d1beda095efabb8c39ca0c5a9042..8bba86c444354c3f427d96af4510dfcfe4e43b7d 100644 (file)
--- a/gen.c
+++ b/gen.c
@@ -214,13 +214,14 @@ genecs()
 
     {
     register int i, j;
-    static char C_char_decl[] = "static const YY_CHAR %s[%d] =\n    {   0,\n";
+    static char C_char_decl[] =
+       "static const YY_CHAR %s[%d] =\n    {   %d,\n";
     int numrows;
     Char clower();
 
-    printf( C_char_decl, "yy_ec", csize + 1 );
+    printf( C_char_decl, "yy_ec", csize, uses_NUL ? abs( ecgroup[0] ) : 0 );
 
-    for ( i = 1; i <= csize; ++i )
+    for ( i = 1; i < csize; ++i )
        {
        if ( caseins && (i >= 'A') && (i <= 'Z') )
            ecgroup[i] = ecgroup[clower( i )];
@@ -233,16 +234,22 @@ genecs()
 
     if ( trace )
        {
+       char *readable_form();
+
        fputs( "\n\nEquivalence Classes:\n\n", stderr );
 
+       if ( uses_NUL )
+               {
+               fprintf( stderr, "%4s = %-2d\n",
+                        readable_form( 0 ), ecgroup[0] );
+               }
+
        numrows = (csize + 1) / 8;
 
        for ( j = 1; j <= numrows; ++j )
            {
-           for ( i = j; i <= csize; i = i + numrows )
+           for ( i = j; i < csize; i = i + numrows )
                {
-               char *readable_form();
-
                fprintf( stderr, "%4s = %-2d",
                         readable_form( i ), ecgroup[i] );
 

diff --git a/main.c b/main.c
index 81842f8dacca8a5d7e0b775b4251cf17911ca306..95a17516621e69005729ba218739917c4e5122d7 100644 (file)
--- a/main.c
+++ b/main.c
@@ -60,7 +60,7 @@ int current_state_type;
 int variable_trailing_context_rules;
 int numtemps, numprots, protprev[MSP], protnext[MSP], prottbl[MSP];
 int protcomst[MSP], firstprot, lastprot, protsave[PROT_SAVE_SIZE];
-int numecs, nextecm[CSIZE + 1], ecgroup[CSIZE + 1], nummecs, tecfwd[CSIZE + 1];
+int numecs, nextecm[CSIZE], ecgroup[CSIZE], nummecs, tecfwd[CSIZE + 1];
 int tecbck[CSIZE + 1];
 int *xlation = (int *) 0;
 int num_xlations;
@@ -79,7 +79,7 @@ Char *ccltbl;
 char *starttime, *endtime, nmstr[MAXLINE];
 int sectnum, nummt, hshcol, dfaeql, numeps, eps2, num_reallocs;
 int tmpuses, totnst, peakpairs, numuniq, numdup, hshsave;
-int num_backtracking, bol_needed;
+int num_backtracking, bol_needed, uses_NUL;
 FILE *temp_action_file;
 FILE *backtrack_file;
 int end_of_buffer_state;
@@ -96,12 +96,6 @@ static char *outfile = "lexyy.c";
 static int outfile_created = 0;
 
 
-/* flex - main program
- *
- * synopsis (from the shell)
- *    flex [-v] [file ...]
- */
-
 main( argc, argv )
 int argc;
 char **argv;
@@ -206,7 +200,7 @@ int status;
        (void) unlink( outfile );
        }
 
-    if ( backtrack_report )
+    if ( backtrack_report && backtrack_file )
        {
        if ( num_backtracking == 0 )
            fprintf( backtrack_file, "No backtracking.\n" );
@@ -550,7 +544,7 @@ get_next_arg: /* used by -C and -S flags in lieu of a "continue 2" control */
     numecs = numeps = eps2 = num_reallocs = hshcol = dfaeql = totnst = 0;
     numuniq = numdup = hshsave = eofseen = datapos = dataline = 0;
     num_backtracking = onesp = numprots = 0;
-    variable_trailing_context_rules = bol_needed = false;
+    variable_trailing_context_rules = bol_needed = uses_NUL = false;
 
     linenum = sectnum = 1;
     firstprot = NIL;
@@ -565,9 +559,9 @@ get_next_arg: /* used by -C and -S flags in lieu of a "continue 2" control */
     if ( useecs )
        {
        /* set up doubly-linked equivalence classes */
-       ecgroup[1] = NIL;
+       ecgroup[0] = NIL;
 
-       for ( i = 2; i <= csize; ++i )
+       for ( i = 1; i < csize; ++i )
            {
            ecgroup[i] = i - 1;
            nextecm[i - 1] = i;
@@ -578,7 +572,7 @@ get_next_arg: /* used by -C and -S flags in lieu of a "continue 2" control */
 
     else
        { /* put everything in its own equivalence class */
-       for ( i = 1; i <= csize; ++i )
+       for ( i = 0; i < csize; ++i )
            {
            ecgroup[i] = i;
            nextecm[i] = BAD_SUBSCRIPT; /* to catch errors */
@@ -627,7 +621,11 @@ readin()
 
     else if ( useecs )
        {
-       numecs = cre8ecs( nextecm, ecgroup, csize );
+       if ( uses_NUL )
+           numecs = cre8ecs( nextecm, ecgroup, csize, 0 );
+       else
+           numecs = cre8ecs( nextecm, ecgroup, csize - 1, 1 );
+
        ccl2ecl();
        }
 

diff --git a/misc.c b/misc.c
index 119658a8d3ad67e4a990511ab0bfe7d07116efc4..e7d8398075b071e394043fbd60145cc0c2c89ed1 100644 (file)
--- a/misc.c
+++ b/misc.c
@@ -610,12 +610,6 @@ Char array[];
 
            array[sptr] = c;
 
-           if ( esc_char == '\0' )
-               {
-               synerr( "escape sequence for null not allowed" );
-               return ( 1 );
-               }
-
            return ( esc_char );
            }
 

diff --git a/parse.y b/parse.y
index 07ce77fdf8187ba8caa828450a0c992bc164834a..12479a8b36198b707ac7fc2e2e0cb62af72b0e57 100644 (file)
--- a/parse.y
+++ b/parse.y
@@ -503,7 +503,7 @@ singleton       :  singleton '*'
                        ++rulelen;
 
                        if ( $1 == '\0' )
-                           synerr( "null in rule" );
+                           uses_NUL = true;
 
                        if ( caseins && $1 >= 'A' && $1 <= 'Z' )
                            $1 = clower( $1 );
@@ -531,6 +531,9 @@ fullccl             :  '[' ccl ']'
 
 ccl             :  ccl CHAR '-' CHAR
                         {
+                       if ( $2 == '\0' || $4 == '\0' )
+                           uses_NUL = true;
+
                        if ( $2 > $4 )
                            synerr( "negative range in character class" );
 
@@ -559,6 +562,9 @@ ccl             :  ccl CHAR '-' CHAR
 
                |  ccl CHAR
                        {
+                       if ( $2 == '\0' )
+                           uses_NUL = true;
+
                        if ( caseins )
                            if ( $2 >= 'A' && $2 <= 'Z' )
                                $2 = clower( $2 );
@@ -579,6 +585,9 @@ ccl             :  ccl CHAR '-' CHAR
 
 string         :  string CHAR
                         {
+                       if ( $2 == '\0' )
+                           uses_NUL = true;
+
                        if ( caseins )
                            if ( $2 >= 'A' && $2 <= 'Z' )
                                $2 = clower( $2 );

diff --git a/scan.l b/scan.l
index dfc1d4734a9b521f73af3e582f288744d9376a7c..ee9a2be9ab1eb2db619a1601421e1440d5920cd7 100644 (file)
--- a/scan.l
+++ b/scan.l
@@ -135,15 +135,15 @@ ESCSEQ            \\([^\n]|[0-9]{1,3}|x[0-9a-f]{1,2})
                        char *malloc();
 
                        ++linenum;
-                       xlation = (int *) malloc( sizeof( int ) * (csize + 1) );
-
-                       for ( i = 1; i <= csize; ++i )
-                           xlation[i] = 0;
+                       xlation = (int *) malloc( sizeof( int ) * csize );
 
                        if ( ! xlation )
                            flexfatal(
                                "dynamic memory failure building %t table" );
 
+                       for ( i = 0; i < csize; ++i )
+                           xlation[i] = 0;
+
                        num_xlations = 0;
 
                        BEGIN(XLATION);

diff --git a/tblcmp.c b/tblcmp.c
index 840736d573637419eb4398f91ab5a56c7bb29a13..ac822eb690f221369c44da7005bbabd700ebffdd 100644 (file)
--- a/tblcmp.c
+++ b/tblcmp.c
@@ -233,7 +233,7 @@ cmptmps()
         * transitions
         */
 
-       nummecs = cre8ecs( tecfwd, tecbck, numecs );
+       nummecs = cre8ecs( tecfwd, tecbck, numecs, 1 );
        }
     
     else