X-Git-Url: https://pd.if.org/git/?a=blobdiff_plain;f=functions%2F_PDCLIB%2Fscan.c;h=ce1dac7b307a0d66b86974b8d54ff6b7f4536543;hb=6089d8ab77f6f7bacbfa6354e8cfecab1b00c8bc;hp=c392a5393727dabd7ebf36a15ae4ccc1c78d3ee5;hpb=5522316811fdce2216c4d1320af68dd54909aed5;p=pdclib diff --git a/functions/_PDCLIB/scan.c b/functions/_PDCLIB/scan.c index c392a53..ce1dac7 100644 --- a/functions/_PDCLIB/scan.c +++ b/functions/_PDCLIB/scan.c @@ -14,6 +14,7 @@ #include #include #include +#include /* Using an integer's bits as flags for both the conversion flags and length modifiers. @@ -37,7 +38,7 @@ type: integer type, used to get the correct type from the parameter stack as well as for cast target. */ -#define ASSIGN( case_cond, type ) \ +#define ASSIGN_VALUE_TO( case_cond, type ) \ case case_cond: \ *( va_arg( status->arg, type * ) ) = (type)( value * sign ); \ break @@ -62,7 +63,7 @@ static int GET( struct _PDCLIB_status_t * status ) if ( rc != EOF ) { ++(status->i); - ++(status->this); + ++(status->current); } return rc; } @@ -82,7 +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; } @@ -114,7 +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; @@ -233,7 +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; @@ -258,25 +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; } @@ -298,6 +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; @@ -316,10 +412,10 @@ const char * _PDCLIB_scan( const char * spec, struct _PDCLIB_status_t * status ) if ( status->base != -1 ) { /* integer conversion */ - uintmax_t value = 0; + uintmax_t value = 0; /* absolute value read */ bool prefix_parsed = false; int sign = 0; - while ( ( status->this < status->width ) && + while ( ( status->current < status->width ) && ( ( rc = GET( status ) ) != EOF ) ) { if ( isspace( rc ) ) @@ -333,7 +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 ) @@ -371,7 +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' ) @@ -428,7 +524,7 @@ const char * _PDCLIB_scan( const char * spec, struct _PDCLIB_status_t * status ) { /* out of input before anything could be parsed - input error */ /* FIXME: if first character does not match, value_parsed is not set - but it is NOT an input error */ - if ( status->n == 0 ) + if ( ( status->n == 0 ) && ( rc == EOF ) ) { status->n = -1; } @@ -439,22 +535,22 @@ const char * _PDCLIB_scan( const char * spec, struct _PDCLIB_status_t * status ) E_intmax | E_size | E_ptrdiff | E_unsigned ) ) { - ASSIGN( E_char, char ); - ASSIGN( E_char | E_unsigned, unsigned char ); - ASSIGN( E_short, short ); - ASSIGN( E_short | E_unsigned, unsigned short ); - ASSIGN( 0, int ); - ASSIGN( E_unsigned, unsigned int ); - ASSIGN( E_long, long ); - ASSIGN( E_long | E_unsigned, unsigned long ); - ASSIGN( E_llong, long long ); - ASSIGN( E_llong | E_unsigned, unsigned long long ); - ASSIGN( E_intmax, intmax_t ); - ASSIGN( E_intmax | E_unsigned, uintmax_t ); - ASSIGN( E_size, size_t ); - /* ASSIGN( E_size | E_unsigned, unsigned size_t ); */ - ASSIGN( E_ptrdiff, ptrdiff_t ); - /* ASSIGN( E_ptrdiff | E_unsigned, unsigned ptrdiff_t ); */ + ASSIGN_VALUE_TO( E_char, char ); + ASSIGN_VALUE_TO( E_char | E_unsigned, unsigned char ); + ASSIGN_VALUE_TO( E_short, short ); + ASSIGN_VALUE_TO( E_short | E_unsigned, unsigned short ); + ASSIGN_VALUE_TO( 0, int ); + ASSIGN_VALUE_TO( E_unsigned, unsigned int ); + ASSIGN_VALUE_TO( E_long, long ); + ASSIGN_VALUE_TO( E_long | E_unsigned, unsigned long ); + ASSIGN_VALUE_TO( E_llong, long long ); + ASSIGN_VALUE_TO( E_llong | E_unsigned, unsigned long long ); + ASSIGN_VALUE_TO( E_intmax, intmax_t ); + ASSIGN_VALUE_TO( E_intmax | E_unsigned, uintmax_t ); + ASSIGN_VALUE_TO( E_size, size_t ); + /* ASSIGN_VALUE_TO( E_size | E_unsigned, unsigned size_t ); */ + ASSIGN_VALUE_TO( E_ptrdiff, ptrdiff_t ); + /* ASSIGN_VALUE_TO( E_ptrdiff | E_unsigned, unsigned ptrdiff_t ); */ default: puts( "UNSUPPORTED SCANF FLAG COMBINATION" ); return NULL; /* behaviour unspecified */