Index: libcfa/src/fstream.cfa =================================================================== --- libcfa/src/fstream.cfa (revision ca995e3090a31786ce34129347ad66e6dd17c3a9) +++ libcfa/src/fstream.cfa (revision e0dc038fbe4972034fd6a262cc70c8bb4258cd32) @@ -10,6 +10,6 @@ // Created On : Wed May 27 17:56:53 2015 // Last Modified By : Peter A. Buhr -// Last Modified On : Fri Aug 18 10:41:17 2023 -// Update Count : 541 +// Last Modified On : Tue Oct 17 08:38:49 2023 +// Update Count : 544 // @@ -311,4 +311,5 @@ } // if va_end( args ); +// if ( len == 0 ) throw ExceptionInst( missing_data ); return len; } // fmt Index: libcfa/src/fstream.hfa =================================================================== --- libcfa/src/fstream.hfa (revision ca995e3090a31786ce34129347ad66e6dd17c3a9) +++ libcfa/src/fstream.hfa (revision e0dc038fbe4972034fd6a262cc70c8bb4258cd32) @@ -10,6 +10,6 @@ // Created On : Wed May 27 17:56:53 2015 // Last Modified By : Peter A. Buhr -// Last Modified On : Fri Aug 18 10:41:15 2023 -// Update Count : 258 +// Last Modified On : Fri Oct 13 13:55:21 2023 +// Update Count : 260 // @@ -119,14 +119,13 @@ void ends( ifstream & ); int fmt( ifstream &, const char format[], ... ) __attribute__(( format(scanf, 2, 3) )); +ifstream & ungetc( ifstream & is, char c ); +bool eof( ifstream & is ); bool fail( ifstream & is ); void clear( ifstream & ); -bool eof( ifstream & is ); void open( ifstream & is, const char name[], const char mode[] ); // FIX ME: use default = "r" void open( ifstream & is, const char name[] ); void close( ifstream & is ); - ifstream & read( ifstream & is, char data[], size_t size ); -ifstream & ungetc( ifstream & is, char c ); void ?{}( ifstream & is ); Index: libcfa/src/iostream.cfa =================================================================== --- libcfa/src/iostream.cfa (revision ca995e3090a31786ce34129347ad66e6dd17c3a9) +++ libcfa/src/iostream.cfa (revision e0dc038fbe4972034fd6a262cc70c8bb4258cd32) @@ -1,3 +1,2 @@ - // // Cforall Version 1.0.0 Copyright (C) 2015 University of Waterloo @@ -11,6 +10,6 @@ // Created On : Wed May 27 17:56:53 2015 // Last Modified By : Peter A. Buhr -// Last Modified On : Sun Oct 8 12:10:21 2023 -// Update Count : 1564 +// Last Modified On : Tue Oct 17 20:57:05 2023 +// Update Count : 1795 // @@ -966,13 +965,17 @@ char fmtstr[ sizeof("%*[]") + nscanset ]; int pos = 0; - fmtstr[pos] = '%'; pos += 1; - fmtstr[pos] = '*'; pos += 1; - fmtstr[pos] = '['; pos += 1; + strcpy( &fmtstr[pos], "%*[" ); pos += 3; strcpy( &fmtstr[pos], f.scanset ); pos += nscanset; - fmtstr[pos] = ']'; pos += 1; - fmtstr[pos] = '\0'; - fmt( is, fmtstr, (void*)0 ); // last arg is dummy: suppress gcc warning - } - else for ( f.wd ) fmt( is, "%*c" ); + strcpy( &fmtstr[pos], "]" ); + fmt( is, fmtstr, "" ); // skip scanset + } else { + char ch; +// fprintf( stderr, "skip " ); + for ( f.wd ) { // skip N characters + if ( eof( is ) ) break; + fmt( is, "%c", &ch ); +// fprintf( stderr, "`%c' ", ch ); + } // for + } // if return is; } @@ -980,77 +983,89 @@ istype & ?|?( istype & is, _Istream_Cstr f ) { - const char * scanset = f.scanset; + const char * scanset; + size_t nscanset = 0; if ( f.flags.delimiter ) scanset = f.delimiter; // getline ? - - size_t len = 0; - if ( scanset ) len = strlen( scanset ); - char fmtstr[len + 16]; - int start = 1; + else scanset = f.scanset; + if ( scanset ) nscanset = strlen( scanset ); + + char fmtstr[nscanset + 32]; // storage for scanset and format codes fmtstr[0] = '%'; - if ( f.flags.ignore ) { fmtstr[1] = '*'; start += 1; } - // no maximum width necessary because text ignored => width is read width - if ( f.wd != -1 ) { + + int pos = 1; + int args; + bool check = true; + + if ( f.flags.ignore ) { check = false; fmtstr[1] = '*'; pos += 1; } + int rwd = f.wd; + if ( f.wd != -1 ) { // => just ignore versus ignore with width // wd is buffer bytes available (for input chars + null terminator) // rwd is count of input chars - int rwd; - if (f.flags.rwd) { - verify (f.wd >= 0); - rwd = f.wd; + // no maximum width necessary because text ignored => width is read width + if ( f.flags.rwd ) check = false; + else rwd = f.wd - 1; + pos += sprintf( &fmtstr[pos], "%d", rwd ); + } // if + + if ( ! scanset ) { // %s, %*s, %ws, %*ws +// fprintf( stderr, "cstr %s\n", f.s ); + strcpy( &fmtstr[pos], "s%n" ); + int len = 0; // may not be set in fmt + if ( f.flags.ignore ) args = fmt( is, fmtstr, &len ); // no string argument for '*' + else args = fmt( is, fmtstr, f.s, &len ); +// fprintf( stderr, "cstr %s %d %d %d %s\n", fmtstr, args, len, f.wd, f.s ); + if ( check && len >= rwd && ! eof( is ) ) { // might not fit + char peek; + fmt( is, "%c", &peek ); // check for whitespace terminator +// fprintf( stderr, "peek %d '%c'\n", args, peek ); + if ( ! eof( is ) ) { + ungetc( is, peek ); + if ( ! isspace( peek ) ) throw ExceptionInst( cstring_length ); + } // if + } // if + } else { + if ( f.flags.delimiter ) { // getline +// fprintf( stderr, "getline\n" ); + sprintf( &fmtstr[pos], "[%s%s]%%n", f.flags.inex ? "^" : "", scanset ); +// fprintf( stderr, "getline %s %d\n", fmtstr, f.wd ); + int len = 0; // may not be set in fmt + if ( f.flags.ignore ) args = fmt( is, fmtstr, &len ); // no string argument for '*' + else args = fmt( is, fmtstr, f.s, &len ); +// fprintf( stderr, "getline %s %d %d %d\n", fmtstr, args, f.wd, eof( is ) ); + if ( check && len == rwd && ! eof( is ) ) { // might not fit + char peek; + fmt( is, "%c", &peek ); // check for delimiter +// fprintf( stderr, "peek %d '%c'\n", args, peek ); + if ( ! eof( is ) ) { + if ( peek != f.delimiter[0] ) { + ungetc( is, peek ); + throw ExceptionInst( cstring_length ); + } // if + } // if + } else fmt( is, "%*c" ); // remove delimiter } else { - verify (f.wd >= 1); - rwd = f.wd - 1; + // incl %[xxx], %*[xxx], %w[xxx], %*w[xxx] + // excl %[^xxx], %*[^xxx], %w[^xxx], %*w[^xxx] + sprintf( &fmtstr[pos], "[%s%s]%%n", f.flags.inex ? "^" : "", scanset ); +// fprintf( stderr, "incl/excl %s %d\n", fmtstr, f.wd ); + int len = 0; // may not be set in fmt + if ( f.flags.ignore ) args = fmt( is, fmtstr, &len ); // no string argument for '*' + else args = fmt( is, fmtstr, f.s, &len ); +// fprintf( stderr, "incl/excl %s \"%s\" %d %d %d %d %d %c\n", fmtstr, f.s, args, f.wd, len, eof( is ), check, f.s[f.wd] ); + if ( check && len == rwd && ! eof( is ) ) { // might not fit +// fprintf( stderr, "overflow\n" ); + char peek; + fmt( is, "%c", &peek ); // check for whitespace terminator +// fprintf( stderr, "peek %d '%c'\n", args, peek ); + if ( ! eof( is ) ) { + ungetc( is, peek ); + if ( f.flags.inex ^ strchr( f.scanset, peek ) != 0p ) throw ExceptionInst( cstring_length ); + } // if + } // if } // if - start += sprintf( &fmtstr[start], "%d", rwd ); - } - - if ( ! scanset ) { - // %s, %*s, %ws, %*ws - fmtstr[start] = 's'; fmtstr[start + 1] = '\0'; - // printf( "cstr %s\n", fmtstr ); - } else { - // incl %[xxx], %*[xxx], %w[xxx], %*w[xxx] - // excl %[^xxx], %*[^xxx], %w[^xxx], %*w[^xxx] - fmtstr[start] = '['; start += 1; - if ( f.flags.inex ) { fmtstr[start] = '^'; start += 1; } - strcpy( &fmtstr[start], scanset ); // copy includes '\0' - len += start; - fmtstr[len] = ']'; fmtstr[len + 1] = '\0'; - // printf( "incl/excl %s\n", fmtstr ); - } // if - - int check = f.wd - 2; - if (! f.flags.ignore ) { - if ( ! f.flags.rwd ) f.s[check] = '\0'; // insert sentinel - } - len = fmt( is, fmtstr, f.s ); - //fprintf( stderr, "KK %s %zd %d %c %s\n", fmtstr, len, check, f.s[check], f.s ); - - if ( ! f.flags.ignore && ! f.flags.rwd && f.s[check] != '\0' ) { // sentinel overwritten ? - // buffer filled, but would we have kept going? - if ( ! eof( is ) ) { - char peek; - fmt( is, "%c", &peek ); - ungetc( is, peek ); - bool hasMore; - if (f.flags.delimiter) { // getline - hasMore = (peek != f.delimiter[0]); - } else if (f.scanset) { // incl/excl - bool peekMatch = strchr(f.scanset, peek) != 0p; - hasMore = f.flags.inex ? (!peekMatch) : (peekMatch); - } else { // %s - hasMore = !isspace(peek); - } - if (hasMore) throw (cstring_length){ &cstring_length_vt }; - } // if - } // if - - if ( f.flags.delimiter ) { // getline ? - if ( len == 0 ) f.s[0] = '\0'; // empty read => argument unchanged => set empty - if ( ! eof( is ) ) { // ignore delimiter, may not be present because of width - char delimiter; - fmt( is, "%c", &delimiter ); - if ( delimiter != f.delimiter[0] ) ungetc( is, delimiter ); - } // if - } //if + } // if + if ( args == 1 && eof( is ) ) { // data but scan ended at EOF +// fprintf( stderr, "clear\n" ); + clear( is ); // => reset EOF => detect again on next read + } // if return is; } // ?|? Index: libcfa/src/iostream.hfa =================================================================== --- libcfa/src/iostream.hfa (revision ca995e3090a31786ce34129347ad66e6dd17c3a9) +++ libcfa/src/iostream.hfa (revision e0dc038fbe4972034fd6a262cc70c8bb4258cd32) @@ -10,6 +10,6 @@ // Created On : Wed May 27 17:56:53 2015 // Last Modified By : Peter A. Buhr -// Last Modified On : Sun Oct 8 12:02:55 2023 -// Update Count : 568 +// Last Modified On : Mon Oct 16 21:12:01 2023 +// Update Count : 573 // @@ -324,4 +324,5 @@ istype & ungetc( istype &, char ); bool eof( istype & ); + void clear( istype & ); }; // basic_istream @@ -329,5 +330,5 @@ trait istream { bool fail( istype & ); - void clear( istype & ); + void open( istype & is, const char name[], const char mode[] ); void open( istype & is, const char name[] ); void close( istype & is ); @@ -402,7 +403,9 @@ ExceptionDecl( cstring_length ); +ExceptionDecl( missing_data ); // *********************************** manipulators *********************************** +// skip does not compose with other C string manipulators. struct _Istream_Cskip { const char * scanset;