Index: libcfa/src/collections/string_res.cfa =================================================================== --- libcfa/src/collections/string_res.cfa (revision 64c4b4ddd59133fe62b3f4efdef17ebc2c51119b) +++ libcfa/src/collections/string_res.cfa (revision 68cf1d49c3f5d0ac8afe843ba072540cf6db0db7) @@ -10,6 +10,6 @@ // Created On : Fri Sep 03 11:00:00 2021 // Last Modified By : Peter A. Buhr -// Last Modified On : Tue Jan 16 22:19:27 2024 -// Update Count : 35 +// Last Modified On : Mon Jan 22 23:12:42 2024 +// Update Count : 43 // @@ -263,5 +263,5 @@ bool cont = false; - _Istream_Cstr cf = { cstr, (_Istream_str_base)f }; + _Istream_Cwidth cf = { cstr, (_Istream_str_base)f }; if ( ! cf.flags.rwd ) cf.wd = wd; Index: libcfa/src/iostream.cfa =================================================================== --- libcfa/src/iostream.cfa (revision 64c4b4ddd59133fe62b3f4efdef17ebc2c51119b) +++ libcfa/src/iostream.cfa (revision 68cf1d49c3f5d0ac8afe843ba072540cf6db0db7) @@ -10,6 +10,6 @@ // Created On : Wed May 27 17:56:53 2015 // Last Modified By : Peter A. Buhr -// Last Modified On : Wed Jan 3 10:53:13 2024 -// Update Count : 1898 +// Last Modified On : Thu Jan 25 08:39:31 2024 +// Update Count : 1901 // @@ -984,86 +984,61 @@ } - istype & ?|?( istype & is, _Istream_Cquoted f ) with( f ) { - char fmtstr[32]; // storage scanset and format codes + istype & ?|?( istype & is, _Istream_Cquoted f ) with( f.cstr ) { + int args; + fini: { + args = fmt( is, "%*[ \f\n\r\t\v]" ); // remove leading whitespace + if ( eof( is ) ) break fini; + char rfmt[4] = { delimiters[0], '%', 'n', '\0' }; + int len = 0; // may not be set in fmt + args = fmt( is, rfmt, &len ); // remove leading quote + if ( len == 0 || eof( is ) ) break fini; + + // Change the remainder of the read into a getline by reseting the closing delimiter. + if ( delimiters[1] != '\0' ) { + delimiters[0] = delimiters[1]; + delimiters[1] = '\0'; + } // if + flags.delimiter = true; + return is | *(_Istream_Cstr *)&f; + } // fini + if ( ! flags.ignore && args == 0 ) s[0] = '\0'; // read failed => no pattern match => set string to null + if ( args == 1 && eof( is ) ) { // data but scan ended at EOF + clear( is ); // => reset EOF => detect again on next read + } // if + return is; + } + + istype & ?|?( istype & is, _Istream_Cstr f ) with( f.cstr ) { + const char * scanset; + size_t nscanset = 0; + if ( flags.delimiter ) scanset = delimiters; // getline ? + else scanset = f.cstr.scanset; + if ( scanset ) nscanset = strlen( scanset ); + + char fmtstr[nscanset + 32]; // storage for scanset and format codes fmtstr[0] = '%'; - int pos = 1; int args; bool check = true; - if ( cstr.flags.ignore ) { check = false; fmtstr[1] = '*'; pos += 1; } - int rwd = cstr.wd; - if ( cstr.wd != -1 ) { // => just ignore versus ignore with width + if ( flags.ignore ) { check = false; fmtstr[1] = '*'; pos += 1; } + int rwd = wd; + if ( wd != -1 ) { // => just ignore versus ignore with width // wd is buffer bytes available (for input chars + null terminator) // rwd is count of input chars // no maximum width necessary because text ignored => width is read width - if ( cstr.flags.rwd ) check = false; - else rwd = cstr.wd - 1; + if ( flags.rwd ) check = false; + else rwd = wd - 1; + assert( rwd > 0 ); pos += sprintf( &fmtstr[pos], "%d", rwd ); } // if - int len = 0; // may not be set in fmt - char enddelim; - if ( ! cstr.flags.inex ) { // => quoted getline - args = fmt( is, "%*[ \f\n\r\t\v]" ); // remove leading whitespace - if ( eof( is ) ) goto Eof; - char rfmt[4] = { cstr.delimiters[0], '%', 'n', '\0' }; - args = fmt( is, rfmt, &len ); // remove leading quote - if ( len == 0 || eof( is ) ) goto Eof; - } // if - enddelim = cstr.delimiters[1] == '\0' ? cstr.delimiters[0] : cstr.delimiters[1]; - sprintf( &fmtstr[pos], "[^%c]%%n", enddelim ); - if ( cstr.flags.ignore ) args = fmt( is, fmtstr, &len ); // no string argument for '*' - else args = fmt( is, fmtstr, cstr.s, &len ); - if ( check && len == rwd && ! eof( is ) ) { // might not fit - char peek; - fmt( is, "%c", &peek ); // check for delimiter - if ( ! eof( is ) ) { - if ( peek != enddelim ) { - ungetc( is, peek ); - throwResume ExceptionInst( cstring_length ); - } // if - } // if - } else fmt( is, "%*c" ); // remove delimiter - Eof: ; - if ( rwd > 0 && args == 0 ) cstr.s[0] = '\0'; // read failed => no pattern match => set string to null - if ( args == 1 && eof( is ) ) { // data but scan ended at EOF - clear( is ); // => reset EOF => detect again on next read - } // if - return is; - } - - istype & ?|?( istype & is, _Istream_Cstr f ) with( f ) { - const char * scanset; - size_t nscanset = 0; - if ( flags.delimiter ) scanset = delimiters; // getline ? - else scanset = f.scanset; - if ( scanset ) nscanset = strlen( scanset ); - - char fmtstr[nscanset + 32]; // storage for scanset and format codes - fmtstr[0] = '%'; - - 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 - // 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 ); + // fprintf( stderr, "cstr %s\n", 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 ( flags.ignore ) args = fmt( is, fmtstr, &len ); // no string argument for '*' + else args = fmt( is, fmtstr, s, &len ); + // fprintf( stderr, "cstr %s %d %d %d %s\n", fmtstr, args, len, wd, s ); if ( check && len >= rwd && ! eof( is ) ) { // might not fit char peek; @@ -1076,18 +1051,17 @@ } // if // FIX ME: CFA strings need to be modified to NOT change the argument for this case, then this can be removed. - if ( ! f.flags.ignore && rwd > 0 && args == 0 ) f.s[0]= '\0'; // read failed => no pattern match => set string to null + if ( ! flags.ignore && args == 0 ) s[0]= '\0'; // read failed => no pattern match => set string to null } else { - if ( f.flags.delimiter ) { // getline + if ( flags.delimiter ) { // getline int len = 0; // may not be set in fmt - sprintf( &fmtstr[pos], "[^%c]%%n", f.delimiters[0] ); - if ( f.flags.ignore ) args = fmt( is, fmtstr, &len ); // no string argument for '*' - else args = fmt( is, fmtstr, f.s, &len ); + sprintf( &fmtstr[pos], "[^%c]%%n", delimiters[0] ); + if ( flags.ignore ) args = fmt( is, fmtstr, &len ); // no string argument for '*' + else args = fmt( is, fmtstr, s, &len ); if ( check && len == rwd && ! eof( is ) ) { // might not fit - fmtstr[0] = f.delimiters[0]; fmtstr[1] = '%'; fmtstr[2] = 'n'; fmtstr[3] = '\0'; - fmt( is, fmtstr, &len ); // remove delimiter + char peek; + fmt( is, "%c", &peek ); // check for delimiter if ( ! eof( is ) ) { -// if ( peek != f.delimiter[0] ) { - if ( len != 1 ) { -// ungetc( is, peek ); + if ( peek != delimiters[0] ) { + ungetc( is, peek ); throwResume ExceptionInst( cstring_length ); } // if @@ -1097,10 +1071,10 @@ // 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 ); + sprintf( &fmtstr[pos], "[%s%s]%%n", flags.inex ? "^" : "", scanset ); + // fprintf( stderr, "incl/excl %s %d\n", fmtstr, 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 ( flags.ignore ) args = fmt( is, fmtstr, &len ); // no string argument for '*' + else args = fmt( is, fmtstr, s, &len ); + // fprintf( stderr, "incl/excl %s \"%s\" %d %d %d %d %d %c\n", fmtstr, s, args, wd, len, eof( is ), check, s[wd] ); if ( check && len == rwd && ! eof( is ) ) { // might not fit // fprintf( stderr, "overflow\n" ); @@ -1110,9 +1084,9 @@ if ( ! eof( is ) ) { ungetc( is, peek ); - if ( f.flags.inex ^ strchr( f.scanset, peek ) != 0p ) throwResume ExceptionInst( cstring_length ); + if ( flags.inex ^ strchr( scanset, peek ) != 0p ) throwResume ExceptionInst( cstring_length ); } // if } // if } // if - if ( ! f.flags.ignore && rwd > 0 && args == 0 ) f.s[0]= '\0'; // read failed => no pattern match => set string to null + if ( ! flags.ignore && args == 0 ) s[0]= '\0'; // read failed => no pattern match => set string to null } // if if ( args == 1 && eof( is ) ) { // data but scan ended at EOF @@ -1123,6 +1097,11 @@ } // ?|? - istype & ?|?( istype & is, _Istream_Char f ) { - fmt( is, "%*c" ); // argument variable unused + istype & ?|?( istype & is, _Istream_Char f ) with(f) { + if ( ignore ) { + fmt( is, "%*c" ); // argument variable unused + } else { + int len = -1, args = fmt( is, fmt, &c, &len ); + if ( args != -1 && len == -1 ) throwResume ExceptionInst( missing_data ); + } // if return is; } // ?|? Index: libcfa/src/iostream.hfa =================================================================== --- libcfa/src/iostream.hfa (revision 64c4b4ddd59133fe62b3f4efdef17ebc2c51119b) +++ libcfa/src/iostream.hfa (revision 68cf1d49c3f5d0ac8afe843ba072540cf6db0db7) @@ -10,6 +10,6 @@ // Created On : Wed May 27 17:56:53 2015 // Last Modified By : Peter A. Buhr -// Last Modified On : Sun Jan 21 22:42:40 2024 -// Update Count : 655 +// Last Modified On : Thu Jan 25 08:39:04 2024 +// Update Count : 696 // @@ -404,37 +404,51 @@ }; // _Istream_str_base -struct _Istream_Cstr { +struct _Istream_Cwidth { char * s; inline _Istream_str_base; }; // _Istream_Cstr +// Restrict nesting of input manipulators to those combinations that make sense. + +struct _Istream_Cstr { + _Istream_Cwidth cstr; +}; // _Istream_Cstr + struct _Istream_Cquoted { - _Istream_Cstr cstr; + _Istream_Cwidth cstr; }; // _Istream_Cquoted static inline { - // width must include room for null terminator - _Istream_Cstr wdi( unsigned int wd, char s[] ) { return (_Istream_Cstr)@{ .s : s, { {.scanset : 0p}, .wd : wd, {.all : 0} } }; } - _Istream_Cstr wdi( unsigned int wd, unsigned int rwd, char s[] ) { - if ( wd <= rwd ) throw (cstring_length){ &cstring_length_vt }; - return (_Istream_Cstr)@{ .s : s, { {.scanset : 0p}, .wd : rwd, {.flags.rwd : true} } }; + // width must include room for null terminator, (gcc) scanf does not allow a 0 width => wd > 1 (1 char and null) and rd > 0 (1 char); + _Istream_Cwidth wdi( unsigned int wd, char s[] ) { + if ( wd <= 1 ) throw (cstring_length){ &cstring_length_vt }; // minimum 1 character and null terminator + return (_Istream_Cwidth)@{ .s : s, { {.scanset : 0p}, .wd : wd, {.all : 0} } }; } - _Istream_Cquoted & quoted( _Istream_Cstr & fmt, const char Ldelimiter = '"', const char Rdelimiter = '\0' ) { - fmt.delimiters[0] = Ldelimiter; fmt.delimiters[1] = Rdelimiter; fmt.delimiters[2] = '\0'; - return (_Istream_Cquoted &)fmt; + _Istream_Cwidth wdi( unsigned int wd, unsigned int rwd, char s[] ) { + if ( wd <= 1 || wd <= rwd ) throw (cstring_length){ &cstring_length_vt }; // minimum 1 character, null terminator, plus subset + return (_Istream_Cwidth)@{ .s : s, { {.scanset : 0p}, .wd : rwd, {.flags.rwd : true} } }; } - _Istream_Cstr & getline( _Istream_Cstr & fmt, const char delimiter = '\n' ) { - fmt.delimiters[0] = delimiter; fmt.delimiters[1] = '\0'; fmt.flags.delimiter = true; fmt.flags.inex = true; return fmt; + _Istream_Cquoted & quoted( _Istream_Cwidth & f, const char Ldelimiter = '"', const char Rdelimiter = '\0' ) { + f.delimiters[0] = Ldelimiter; f.delimiters[1] = Rdelimiter; f.delimiters[2] = '\0'; + return (_Istream_Cquoted &)f; } - _Istream_Cstr & incl( const char scanset[], _Istream_Cstr & fmt ) { fmt.scanset = scanset; fmt.flags.inex = false; return fmt; } - _Istream_Cstr & excl( const char scanset[], _Istream_Cstr & fmt ) { fmt.scanset = scanset; fmt.flags.inex = true; return fmt; } - _Istream_Cstr ignore( char s[] ) { return (_Istream_Cstr)@{ .s : s, { {.scanset : 0p}, .wd : -1, {.flags.ignore : true} } }; } - _Istream_Cstr & ignore( _Istream_Cstr & fmt ) { fmt.flags.ignore = true; return fmt; } + _Istream_Cstr & getline( _Istream_Cwidth & f, const char delimiter = '\n' ) { + f.delimiters[0] = delimiter; f.delimiters[1] = '\0'; f.flags.delimiter = true; return (_Istream_Cstr &)f; + } + _Istream_Cstr & incl( const char scanset[], _Istream_Cwidth & f ) { f.scanset = scanset; f.flags.inex = false; return (_Istream_Cstr &)f; } + _Istream_Cstr & excl( const char scanset[], _Istream_Cwidth & f ) { f.scanset = scanset; f.flags.inex = true; return (_Istream_Cstr &)f; } + _Istream_Cstr ignore( char s[] ) { return (_Istream_Cwidth)@{ .s : s, { {.scanset : 0p}, .wd : -1, {.flags.ignore : true} } }; } + _Istream_Cstr & ignore( _Istream_Cwidth & f ) { f.flags.ignore = true; return (_Istream_Cstr &)f; } + _Istream_Cquoted & ignore( _Istream_Cquoted & f ) { f.cstr.flags.ignore = true; return (_Istream_Cquoted &)f; } + _Istream_Cstr & ignore( _Istream_Cstr & f ) { f.cstr.flags.ignore = true; return (_Istream_Cstr &)f; } } // distribution forall( istype & | basic_istream( istype ) ) { - istype & ?|?( istype & is, _Istream_Cstr f ); istype & ?|?( istype & is, _Istream_Cskip f ); istype & ?|?( istype & is, _Istream_Cquoted f ); + istype & ?|?( istype & is, _Istream_Cstr f ); + static inline { + istype & ?|?( istype & is, _Istream_Cwidth f ) { return is | *(_Istream_Cstr *)&f; } + } // distribution } // distribution