Reworked scanf() testing. General cleanups.
[fw/pdclib] / functions / _PDCLIB / scan.c
index 0d46f9ecfd53efe0f8a13fdd901487f482198c55..b346ffc2bb81dc7943d1db696e168241e2a77370 100644 (file)
@@ -63,7 +63,7 @@ static int GET( struct _PDCLIB_status_t * status )
     if ( rc != EOF )
     {
         ++(status->i);
-        ++(status->this);
+        ++(status->current);
     }
     return rc;
 }
@@ -83,7 +83,45 @@ static void UNGET( int c, struct _PDCLIB_status_t * status )
         --(status->s);
     }
     --(status->i);
-    --(status->this);
+    --(status->current);
+}
+
+
+/* Helper function to check if a character is part of a given scanset */
+static bool IN_SCANSET( const char * scanlist, const char * end_scanlist, int rc )
+{
+    // SOLAR
+    int previous = -1;
+    while ( scanlist != end_scanlist )
+    {
+        if ( ( *scanlist == '-' ) && ( previous != -1 ) )
+        {
+            /* possible scangroup ("a-z") */
+            if ( ++scanlist == end_scanlist )
+            {
+                /* '-' at end of scanlist does not describe a scangroup */
+                return rc == '-';
+            }
+            while ( ++previous <= (unsigned char)*scanlist )
+            {
+                if ( previous == rc )
+                {
+                    return true;
+                }
+            }
+            previous = -1;
+        }
+        else
+        {
+            /* not a scangroup, check verbatim */
+            if ( rc == (unsigned char)*scanlist )
+            {
+                return true;
+            }
+            previous = (unsigned char)(*scanlist++);
+        }
+    }
+    return false;
 }
 
 
@@ -115,7 +153,7 @@ const char * _PDCLIB_scan( const char * spec, struct _PDCLIB_status_t * status )
     /* Initializing status structure */
     status->flags = 0;
     status->base = -1;
-    status->this = 0;
+    status->current = 0;
     status->width = 0;
     status->prec = 0;
 
@@ -234,7 +272,7 @@ const char * _PDCLIB_scan( const char * spec, struct _PDCLIB_status_t * status )
                 status->width = 1;
             }
             /* reading until width reached or input exhausted */
-            while ( ( status->this < status->width ) &&
+            while ( ( status->current < status->width ) &&
                     ( ( rc = GET( status ) ) != EOF ) )
             {
                 *(c++) = rc;
@@ -259,25 +297,28 @@ const char * _PDCLIB_scan( const char * spec, struct _PDCLIB_status_t * status )
         case 's':
         {
             char * c = va_arg( status->arg, char * );
-            while ( ( status->this < status->width ) && 
+            while ( ( status->current < status->width ) && 
                     ( ( rc = GET( status ) ) != EOF ) )
             {
                 if ( isspace( rc ) )
                 {
+                    UNGET( rc, status );
                     if ( value_parsed )
                     {
                         /* matching sequence terminated by whitespace */
                         *c = '\0';
-                        return spec;
+                        ++status->n;
+                        return ++spec;
                     }
                     else
                     {
-                        /* leading whitespace not counted against width */
-                        --(status->this);
+                        /* matching error */
+                        return NULL;
                     }
                 }
                 else
                 {
+                    /* match */
                     value_parsed = true;
                     *(c++) = rc;
                 }
@@ -299,6 +340,60 @@ const char * _PDCLIB_scan( const char * spec, struct _PDCLIB_status_t * status )
                 return NULL;
             }
         }
+        case '[':
+        {
+            const char * endspec = spec;
+            bool negative_scanlist = false;
+            if ( *(++endspec) == '^' )
+            {
+                negative_scanlist = true;
+                ++endspec;
+            }
+            spec = endspec;
+            do
+            {
+                // TODO: This can run beyond a malformed format string
+                ++endspec;
+            } while ( *endspec != ']' );
+            // read according to scanlist, equiv. to %s above
+            char * c = va_arg( status->arg, char * );
+            while ( ( status->current < status->width ) && 
+                    ( ( rc = GET( status ) ) != EOF ) )
+            {
+                if ( negative_scanlist )
+                {
+                    if ( IN_SCANSET( spec, endspec, rc ) )
+                    {
+                        UNGET( rc, status );
+                        break;
+                    }
+                }
+                else
+                {
+                    if ( ! IN_SCANSET( spec, endspec, rc ) )
+                    {
+                        UNGET( rc, status );
+                        break;
+                    }
+                }
+                value_parsed = true;
+                *(c++) = rc;
+            }
+            if ( value_parsed )
+            {
+                *c = '\0';
+                ++status->n;
+                return ++endspec;
+            }
+            else
+            {
+                if ( rc == EOF )
+                {
+                    status->n = -1;
+                }
+                return NULL;
+            }
+        }
         case 'p':
             status->base = 16;
             status->flags |= E_unsigned;
@@ -318,11 +413,9 @@ const char * _PDCLIB_scan( const char * spec, struct _PDCLIB_status_t * status )
     {
         /* integer conversion */
         uintmax_t value = 0;         /* absolute value read */
-        uintmax_t limit;             /* max. value allowed */
-        uintmax_t threshold;         /* overflow threshold */
         bool prefix_parsed = false;
         int sign = 0;
-        while ( ( status->this < status->width ) &&
+        while ( ( status->current < status->width ) &&
                 ( ( rc = GET( status ) ) != EOF ) )
         {
             if ( isspace( rc ) )
@@ -336,7 +429,7 @@ const char * _PDCLIB_scan( const char * spec, struct _PDCLIB_status_t * status )
                 else
                 {
                     /* leading whitespace not counted against width */
-                    status->this--;
+                    status->current--;
                 }
             }
             else if ( ! sign )
@@ -356,53 +449,6 @@ const char * _PDCLIB_scan( const char * spec, struct _PDCLIB_status_t * status )
                         UNGET( rc, status );
                         break;
                 }
-                switch ( status->flags & ( E_char | E_short | E_long | E_llong | E_intmax | E_size | E_ptrdiff | E_unsigned ) )
-                {
-                    case E_char:
-                        limit = ( sign == 1 ) ? CHAR_MAX : ( CHAR_MIN * sign );
-                        break;
-                    case E_char | E_unsigned:
-                        limit = UCHAR_MAX;
-                        break;
-                    case E_short:
-                        limit = ( sign == 1 ) ? SHRT_MAX : ( SHRT_MIN * sign );
-                        break;
-                    case E_short | E_unsigned:
-                        limit = USHRT_MAX;
-                        break;
-                    case E_long:
-                        limit = ( sign == 1 ) ? LONG_MAX : ( LONG_MIN * sign );
-                        break;
-                    case E_long | E_unsigned:
-                        limit = ULONG_MAX;
-                        break;
-                    case E_llong:
-                        limit = ( sign == 1 ) ? LLONG_MAX : ( LLONG_MIN * sign );
-                        break;
-                    case E_llong | E_unsigned:
-                        limit = ULLONG_MAX;
-                        break;
-                    case E_intmax:
-                        limit = ( sign == 1 ) ? INTMAX_MAX : ( INTMAX_MIN * sign );
-                        break;
-                    case E_intmax | E_unsigned:
-                        limit = UINTMAX_MAX;
-                        break;
-                    case E_size:
-                    case E_size | E_unsigned:
-                        limit = SIZE_MAX;
-                        break;
-                    case E_ptrdiff:
-                    case E_ptrdiff | E_unsigned:
-                        limit = ( sign == 1 ) ? PTRDIFF_MAX : ( PTRDIFF_MIN * sign );
-                        break;
-                    case E_unsigned:
-                        limit = UINT_MAX;
-                        break;
-                    default:
-                        limit = ( sign == 1 ) ? INT_MAX : ( INT_MIN * sign );
-                        break;
-                }
             }
             else if ( ! prefix_parsed )
             {
@@ -421,7 +467,7 @@ const char * _PDCLIB_scan( const char * spec, struct _PDCLIB_status_t * status )
                 {
                     /* starts with zero, so it might be a prefix. */
                     /* check what follows next (might be 0x...) */
-                    if ( ( status->this < status->width ) &&
+                    if ( ( status->current < status->width ) &&
                          ( ( rc = GET( status ) ) != EOF ) )
                     {
                         if ( tolower( rc ) == 'x' )
@@ -468,20 +514,9 @@ const char * _PDCLIB_scan( const char * spec, struct _PDCLIB_status_t * status )
                     UNGET( rc, status );
                     break;
                 }
-                // SOLAR
-                // if ( ( ( limit - ( digitptr - _PDCLIB_digits ) ) / status->base ) >= value )
-                //if ( ( ( limit / status->base ) >= value ) && ( ( limit - ( digitptr - _PDCLIB_digits ) ) >= ( value * status->base ) ) )
-                {
-                    /* no overflow */
-                    value *= status->base;
-                    value += digitptr - _PDCLIB_digits;
-                    value_parsed = true;
-                }
-                //else
-                //{
-                //    value = limit;
-                //    threshold = 0;
-                //}
+                value *= status->base;
+                value += digitptr - _PDCLIB_digits;
+                value_parsed = true;
             }
         }
         /* width or input exhausted, or non-matching character */
@@ -529,14 +564,34 @@ const char * _PDCLIB_scan( const char * spec, struct _PDCLIB_status_t * status )
 
 
 #ifdef TEST
+#define _PDCLIB_FILEID "_PDCLIB/scan.c"
+#define _PDCLIB_STRINGIO
+
 #include <_PDCLIB_test.h>
-#include <limits.h>
 
+static int testscanf( char const * s, char const * format, ... )
+{
+    struct _PDCLIB_status_t status;
+    status.n = 0;
+    status.i = 0;
+    status.s = (char *)s;
+    status.stream = NULL;
+    va_start( status.arg, format );
+    if ( *(_PDCLIB_scan( format, &status )) != '\0' )
+    {
+        printf( "_PDCLIB_scan() did not return end-of-specifier on '%s'.\n", format );
+        ++TEST_RESULTS;
+    }
+    va_end( status.arg );
+    return status.n;
+}
+
+#define TEST_CONVERSION_ONLY
 
 int main( void )
 {
-    /* Testing covered by fscanf.c */
+    char source[100];
+#include "scanf_testcases.h"
     return TEST_RESULTS;
 }