/* $Id$ */ /* _PDCLIB_scan( const char *, struct _PDCLIB_status_t * ) This file is part of the Public Domain C Library (PDCLib). Permission is granted to use, modify, and / or redistribute at will. */ #include #include #include #include #include #include #include #include #include #ifndef REGTEST /* Using an integer's bits as flags for both the conversion flags and length modifiers. */ #define E_suppressed 1<<0 #define E_char 1<<6 #define E_short 1<<7 #define E_long 1<<8 #define E_llong 1<<9 #define E_intmax 1<<10 #define E_size 1<<11 #define E_ptrdiff 1<<12 #define E_intptr 1<<13 #define E_ldouble 1<<14 #define E_unsigned 1<<16 /* Helper function to get a character from the string or stream, whatever is used for input. When reading from a string, returns EOF on end-of-string so that handling of the return value can be uniform for both streams and strings. */ static int GET( struct _PDCLIB_status_t * status ) { int rc = EOF; if ( status->stream != NULL ) { rc = getc( status->stream ); } else { rc = ( *status->s == '\0' ) ? EOF : (unsigned char)*((status->s)++); } if ( rc != EOF ) { ++(status->i); ++(status->current); } return rc; } /* Helper function to put a read character back into the string or stream, whatever is used for input. */ static void UNGET( int c, struct _PDCLIB_status_t * status ) { if ( status->stream != NULL ) { ungetc( c, status->stream ); /* TODO: Error? */ } else { --(status->s); } --(status->i); --(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; } const char * _PDCLIB_scan( const char * spec, struct _PDCLIB_status_t * status ) { /* generic input character */ int rc; const char * orig_spec = spec; if ( *(++spec) == '%' ) { /* %% -> match single '%' */ rc = GET( status ); switch ( rc ) { case EOF: /* input error */ if ( status->n == 0 ) { status->n = -1; } return NULL; case '%': return ++spec; default: UNGET( rc, status ); break; } } /* Initializing status structure */ status->flags = 0; status->base = -1; status->current = 0; status->width = 0; status->prec = 0; /* '*' suppresses assigning parsed value to variable */ if ( *spec == '*' ) { status->flags |= E_suppressed; ++spec; } /* If a width is given, strtol() will return its value. If not given, strtol() will return zero. In both cases, endptr will point to the rest of the conversion specifier - just what we need. */ char const * prev_spec = spec; status->width = (int)strtol( spec, (char**)&spec, 10 ); if ( spec == prev_spec ) { status->width = UINT_MAX; } /* Optional length modifier We step one character ahead in any case, and step back only if we find there has been no length modifier (or step ahead another character if it has been "hh" or "ll"). */ switch ( *(spec++) ) { case 'h': if ( *spec == 'h' ) { /* hh -> char */ status->flags |= E_char; ++spec; } else { /* h -> short */ status->flags |= E_short; } break; case 'l': if ( *spec == 'l' ) { /* ll -> long long */ status->flags |= E_llong; ++spec; } else { /* l -> long */ status->flags |= E_long; } break; case 'j': /* j -> intmax_t, which might or might not be long long */ status->flags |= E_intmax; break; case 'z': /* z -> size_t, which might or might not be unsigned int */ status->flags |= E_size; break; case 't': /* t -> ptrdiff_t, which might or might not be long */ status->flags |= E_ptrdiff; break; case 'L': /* L -> long double */ status->flags |= E_ldouble; break; default: --spec; break; } /* Conversion specifier */ /* whether valid input had been parsed */ bool value_parsed = false; switch ( *spec ) { case 'd': status->base = 10; break; case 'i': status->base = 0; break; case 'o': status->base = 8; status->flags |= E_unsigned; break; case 'u': status->base = 10; status->flags |= E_unsigned; break; case 'x': status->base = 16; status->flags |= E_unsigned; break; case 'f': case 'F': case 'e': case 'E': case 'g': case 'G': case 'a': case 'A': break; case 'c': { char * c = va_arg( status->arg, char * ); /* for %c, default width is one */ if ( status->width == SIZE_MAX ) { status->width = 1; } /* reading until width reached or input exhausted */ while ( ( status->current < status->width ) && ( ( rc = GET( status ) ) != EOF ) ) { *(c++) = rc; value_parsed = true; } /* width or input exhausted */ if ( value_parsed ) { ++status->n; return ++spec; } else { /* input error, no character read */ if ( status->n == 0 ) { status->n = -1; } return NULL; } } case 's': { char * c = va_arg( status->arg, char * ); 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'; ++status->n; return ++spec; } else { /* matching error */ return NULL; } } else { /* match */ value_parsed = true; *(c++) = rc; } } /* width or input exhausted */ if ( value_parsed ) { *c = '\0'; ++status->n; return ++spec; } else { /* input error, no character read */ if ( status->n == 0 ) { status->n = -1; } 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; // TODO: Like _PDCLIB_print, E_pointer(?) status->flags |= E_unsigned | E_long; break; case 'n': { int * val = va_arg( status->arg, int * ); *val = status->i; return ++spec; } default: /* No conversion specifier. Bad conversion. */ return orig_spec; } if ( status->base != -1 ) { /* integer conversion */ uintmax_t value = 0; /* absolute value read */ bool prefix_parsed = false; int sign = 0; while ( ( status->current < status->width ) && ( ( rc = GET( status ) ) != EOF ) ) { if ( isspace( rc ) ) { if ( sign ) { /* matching sequence terminated by whitespace */ UNGET( rc, status ); break; } else { /* leading whitespace not counted against width */ status->current--; } } else if ( ! sign ) { /* no sign parsed yet */ switch ( rc ) { case '-': sign = -1; break; case '+': sign = 1; break; default: /* not a sign; put back character */ sign = 1; UNGET( rc, status ); break; } } else if ( ! prefix_parsed ) { /* no prefix (0x... for hex, 0... for octal) parsed yet */ prefix_parsed = true; if ( rc != '0' ) { /* not a prefix; if base not yet set, set to decimal */ if ( status->base == 0 ) { status->base = 10; } UNGET( rc, status ); } else { /* starts with zero, so it might be a prefix. */ /* check what follows next (might be 0x...) */ if ( ( status->current < status->width ) && ( ( rc = GET( status ) ) != EOF ) ) { if ( tolower( rc ) == 'x' ) { /* 0x... would be prefix for hex base... */ if ( ( status->base == 0 ) || ( status->base == 16 ) ) { status->base = 16; } else { /* ...unless already set to other value */ UNGET( rc, status ); value_parsed = true; } } else { /* 0... but not 0x.... would be octal prefix */ UNGET( rc, status ); if ( status->base == 0 ) { status->base = 8; } /* in any case we have read a zero */ value_parsed = true; } } else { /* failed to read beyond the initial zero */ value_parsed = true; break; } } } else { char * digitptr = memchr( _PDCLIB_digits, tolower( rc ), status->base ); if ( digitptr == NULL ) { /* end of input item */ UNGET( rc, status ); break; } value *= status->base; value += digitptr - _PDCLIB_digits; value_parsed = true; } } /* width or input exhausted, or non-matching character */ if ( ! value_parsed ) { /* 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 ) && ( rc == EOF ) ) { status->n = -1; } return NULL; } /* convert value to target type and assign to parameter */ if ( ! ( status->flags & E_suppressed ) ) { switch ( status->flags & ( E_char | E_short | E_long | E_llong | E_intmax | E_size | E_ptrdiff | E_unsigned ) ) { case E_char: *( va_arg( status->arg, char * ) ) = (char)( value * sign ); break; case E_char | E_unsigned: *( va_arg( status->arg, unsigned char * ) ) = (unsigned char)( value * sign ); break; case E_short: *( va_arg( status->arg, short * ) ) = (short)( value * sign ); break; case E_short | E_unsigned: *( va_arg( status->arg, unsigned short * ) ) = (unsigned short)( value * sign ); break; case 0: *( va_arg( status->arg, int * ) ) = (int)( value * sign ); break; case E_unsigned: *( va_arg( status->arg, unsigned int * ) ) = (unsigned int)( value * sign ); break; case E_long: *( va_arg( status->arg, long * ) ) = (long)( value * sign ); break; case E_long | E_unsigned: *( va_arg( status->arg, unsigned long * ) ) = (unsigned long)( value * sign ); break; case E_llong: *( va_arg( status->arg, long long * ) ) = (long long)( value * sign ); break; case E_llong | E_unsigned: *( va_arg( status->arg, unsigned long long * ) ) = (unsigned long long)( value * sign ); break; case E_intmax: *( va_arg( status->arg, intmax_t * ) ) = (intmax_t)( value * sign ); break; case E_intmax | E_unsigned: *( va_arg( status->arg, uintmax_t * ) ) = (uintmax_t)( value * sign ); break; case E_size: /* E_size always implies unsigned */ *( va_arg( status->arg, size_t * ) ) = (size_t)( value * sign ); break; case E_ptrdiff: /* E_ptrdiff always implies signed */ *( va_arg( status->arg, ptrdiff_t * ) ) = (ptrdiff_t)( value * sign ); break; default: puts( "UNSUPPORTED SCANF FLAG COMBINATION" ); return NULL; /* behaviour unspecified */ } ++(status->n); } return ++spec; } /* TODO: Floats. */ return NULL; } #endif #ifdef TEST #define _PDCLIB_FILEID "_PDCLIB/scan.c" #define _PDCLIB_STRINGIO #include <_PDCLIB_test.h> #ifndef REGTEST 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; } #endif #define TEST_CONVERSION_ONLY int main( void ) { #ifndef REGTEST char source[100]; #include "scanf_testcases.h" #endif return TEST_RESULTS; } #endif