March 2010 - macruby-changes - lists.macosforge.org

[3669] MacRuby/branches/icu/string.c
by source_changes＠macosforge.org 03 Mar '10

03 Mar '10

Revision: 3669 http://trac.macosforge.org/projects/ruby/changeset/3669 Author: lsansonetti(a)apple.com Date: 2010-03-02 17:31:59 -0800 (Tue, 02 Mar 2010) Log Message: ----------- added #succ, #next, #upto + fixed a bug in #<=> Modified Paths: -------------- MacRuby/branches/icu/string.c Modified: MacRuby/branches/icu/string.c =================================================================== --- MacRuby/branches/icu/string.c 2010-03-02 23:57:21 UTC (rev 3668) +++ MacRuby/branches/icu/string.c 2010-03-03 01:31:59 UTC (rev 3669) @@ -875,7 +875,7 @@ return self->length_in_bytes > str->length_in_bytes ? 1 : -1; } - return res > 1 ? 1 : -1; + return res > 0 ? 1 : -1; } static long @@ -3599,6 +3599,166 @@ return str; } +static VALUE +rstr_succ(VALUE str, SEL sel) +{ + if (rb_str_chars_len(str) == 0) { + return str; + } + + if (!str_try_making_data_uchars(RSTR(str))) { + rb_raise(rb_eArgError, + "cannot make receiver data as Unicode characters"); + } + + UChar *chars_buf = (UChar *)malloc(RSTR(str)->length_in_bytes); + UChar *chars_ptr = &chars_buf[1]; + + memcpy(chars_ptr, RSTR(str)->data.uchars, RSTR(str)->length_in_bytes); + + long len = BYTES_TO_UCHARS(RSTR(str)->length_in_bytes); + UChar carry = 0; + bool modified = false; + + for (long i = len - 1; i >= 0; i--) { + UChar c = chars_ptr[i]; + if (isdigit(c)) { + modified = true; + if (c != '9') { + chars_ptr[i]++; + carry = 0; + break; + } + else { + chars_ptr[i] = '0'; + carry = '1'; + } + } + else if (isalpha(c)) { + const bool lower = islower(c); + UChar e = lower ? 'z' : 'Z'; + modified = true; + if (c != e) { + chars_ptr[i]++; + carry = 0; + break; + } + else { + carry = chars_ptr[i] = lower ? 'a' : 'A'; + } + } +#if 0 // TODO: this requires more love + else if (!isascii(c)) { + modified = true; + chars_ptr[i]++; + carry = 0; + break; + } +#endif + } + + if (!modified) { + chars_ptr[len - 1]++; + } + else if (carry != 0) { + chars_ptr = chars_buf; + chars_ptr[0] = carry; + len++; + } + + VALUE newstr = rb_unicode_str_new(chars_ptr, len); + free(chars_buf); + return newstr; +} + +static VALUE +rstr_succ_bang(VALUE str, SEL sel) +{ + rstr_replace(str, 0, rstr_succ(str, 0)); + return str; +} + +/* + * call-seq: + * str.upto(other_str, exclusive=false) {|s| block } => str + * + * Iterates through successive values, starting at str and + * ending at other_str inclusive, passing each value in turn to + * the block. The <code>String#succ</code> method is used to generate + * each value. If optional second argument exclusive is omitted or is <code>false</code>, + * the last value will be included; otherwise it will be excluded. + * + * "a8".upto("b6") {|s| print s, ' ' } + * for s in "a8".."b6" + * print s, ' ' + * end + * + * produces: + * + * a8 a9 b0 b1 b2 b3 b4 b5 b6 + * a8 a9 b0 b1 b2 b3 b4 b5 b6 + */ + +static VALUE +rstr_upto(VALUE str, SEL sel, int argc, VALUE *argv) +{ + VALUE beg = str; + VALUE end, exclusive; + rb_scan_args(argc, argv, "11", &end, &exclusive); + + bool excl = RTEST(exclusive); + StringValue(end); + + if (rb_str_chars_len(beg) == 1 && rb_str_chars_len(end) == 1) { + UChar begc = rb_str_get_uchar(beg, 0); + UChar endc = rb_str_get_uchar(end, 0); + + if (begc > endc || (excl && begc == endc)) { + return beg; + } + while (true) { + rb_yield(rb_unicode_str_new(&begc, 1)); + RETURN_IF_BROKEN(); + if (!excl && begc == endc) { + break; + } + begc++; + if (excl && begc == endc) { + break; + } + } + return beg; + } + + const int cmp = rb_str_cmp(beg, end); + if (cmp > 0 || (excl && cmp == 0)) { + return beg; + } + + SEL succ_sel = sel_registerName("succ"); + + VALUE current = beg; + VALUE after_end = rb_vm_call(end, succ_sel, 0, NULL, false); + StringValue(after_end); + while (!rb_str_equal(current, after_end)) { + rb_yield(current); + RETURN_IF_BROKEN(); + if (!excl && rb_str_equal(current, end)) { + break; + } + current = rb_vm_call(current, succ_sel, 0, NULL, false); + StringValue(current); + if (excl && rb_str_equal(current, end)) { + break; + } + if (rb_str_chars_len(current) > rb_str_chars_len(end) + || rb_str_chars_len(current) == 0) { + break; + } + } + return beg; +} + // NSString primitives. static void @@ -3751,6 +3911,11 @@ rb_objc_define_method(rb_cRubyString, "each_line", rstr_each_line, -1); rb_objc_define_method(rb_cRubyString, "chars", rstr_each_char, 0); rb_objc_define_method(rb_cRubyString, "each_char", rstr_each_char, 0); + rb_objc_define_method(rb_cRubyString, "succ", rstr_succ, 0); + rb_objc_define_method(rb_cRubyString, "succ!", rstr_succ_bang, 0); + rb_objc_define_method(rb_cRubyString, "next", rstr_succ, 0); + rb_objc_define_method(rb_cRubyString, "next!", rstr_succ_bang, 0); + rb_objc_define_method(rb_cRubyString, "upto", rstr_upto, -1); // Added for MacRuby (debugging). rb_objc_define_method(rb_cRubyString, "__chars_count__",

1 0

[3668] MacRuby/branches/icu/string.c
by source_changes＠macosforge.org 02 Mar '10

02 Mar '10

Revision: 3668 http://trac.macosforge.org/projects/ruby/changeset/3668 Author: lsansonetti(a)apple.com Date: 2010-03-02 15:57:21 -0800 (Tue, 02 Mar 2010) Log Message: ----------- added #ord Modified Paths: -------------- MacRuby/branches/icu/string.c Modified: MacRuby/branches/icu/string.c =================================================================== --- MacRuby/branches/icu/string.c 2010-03-02 23:44:58 UTC (rev 3667) +++ MacRuby/branches/icu/string.c 2010-03-02 23:57:21 UTC (rev 3668) @@ -2577,6 +2577,24 @@ /* * call-seq: + * str.ord => integer + * + * Return the <code>Integer</code> ordinal of a one-character string. + * + * "a".ord #=> 97 + */ + +static VALUE +rstr_ord(VALUE str, SEL sel) +{ + if (RSTR(str)->length_in_bytes == 0) { + rb_raise(rb_eArgError, "empty string"); + } + return INT2NUM(rb_str_get_uchar(str, 0)); +} + +/* + * call-seq: * str.chomp!(separator=$/) => str or nil * * Modifies str in place as described for <code>String#chomp</code>, @@ -3704,6 +3722,7 @@ rb_objc_define_method(rb_cRubyString, "to_i", rstr_to_i, -1); rb_objc_define_method(rb_cRubyString, "hex", rstr_hex, 0); rb_objc_define_method(rb_cRubyString, "oct", rstr_oct, 0); + rb_objc_define_method(rb_cRubyString, "ord", rstr_ord, 0); rb_objc_define_method(rb_cRubyString, "chomp", rstr_chomp, -1); rb_objc_define_method(rb_cRubyString, "chomp!", rstr_chomp_bang, -1); rb_objc_define_method(rb_cRubyString, "sub", rstr_sub, -1);

1 0

[3667] MacRuby/branches/icu/string.c
by source_changes＠macosforge.org 02 Mar '10

02 Mar '10

Revision: 3667 http://trac.macosforge.org/projects/ruby/changeset/3667 Author: lsansonetti(a)apple.com Date: 2010-03-02 15:44:58 -0800 (Tue, 02 Mar 2010) Log Message: ----------- added upcase, swapcase, downcase methods Modified Paths: -------------- MacRuby/branches/icu/string.c Modified: MacRuby/branches/icu/string.c =================================================================== --- MacRuby/branches/icu/string.c 2010-03-02 21:23:07 UTC (rev 3666) +++ MacRuby/branches/icu/string.c 2010-03-02 23:44:58 UTC (rev 3667) @@ -2946,33 +2946,42 @@ * Note: case replacement is effective only in ASCII region. */ +#define CHAR_ITERATE(str, code) \ + if (str_try_making_data_uchars(RSTR(str))) { \ + for (long i = 0, count = BYTES_TO_UCHARS(RSTR(str)->length_in_bytes); \ + i < count; i++) { \ + UChar __tmp, c; \ + __tmp = c = RSTR(str)->data.uchars[i]; \ + code; \ + if (__tmp != c) { \ + RSTR(str)->data.uchars[i] = c; \ + } \ + } \ + } \ + else { \ + for (long i = 0, count = RSTR(str)->length_in_bytes; \ + i < count; i++) { \ + char __tmp, c; \ + __tmp = c = RSTR(str)->data.bytes[i]; \ + code; \ + if (__tmp != c) { \ + RSTR(str)->data.bytes[i] = c; \ + } \ + } \ + } + static VALUE rstr_downcase_bang(VALUE str, SEL sel) { rstr_modify(str); bool changed = false; + CHAR_ITERATE(str, + if (c >= 'A' && c <= 'Z') { + c = 'a' + (c - 'A'); + changed = true; + }); - if (str_is_stored_in_uchars(RSTR(str))) { - for (long i = 0, count = BYTES_TO_UCHARS(RSTR(str)->length_in_bytes); - i < count; i++) { - UChar c = RSTR(str)->data.uchars[i]; - if (c >= 'A' && c <= 'Z') { - RSTR(str)->data.uchars[i] = 'a' + (c - 'A'); - changed = true; - } - } - } - else { - for (long i = 0, count = RSTR(str)->length_in_bytes; i < count; i++) { - char c = RSTR(str)->data.bytes[i]; - if (c >= 'A' && c <= 'Z') { - RSTR(str)->data.bytes[i] = 'a' + (c - 'A'); - changed = true; - } - } - } - return changed ? str : Qnil; } @@ -2998,6 +3007,154 @@ /* * call-seq: + * str.upcase! => str or nil + * + * Upcases the contents of str, returning <code>nil</code> if no changes + * were made. + * Note: case replacement is effective only in ASCII region. + */ + +static VALUE +rstr_upcase_bang(VALUE str, SEL sel) +{ + rstr_modify(str); + + bool changed = false; + CHAR_ITERATE(str, + if (c >= 'a' && c <= 'z') { + c = 'A' + (c - 'a'); + changed = true; + }); + + return changed ? str : Qnil; +} + +/* + * call-seq: + * str.upcase => new_str + * + * Returns a copy of str with all lowercase letters replaced with their + * uppercase counterparts. The operation is locale insensitive---only + * characters ``a'' to ``z'' are affected. + * Note: case replacement is effective only in ASCII region. + * + * "hEllO".upcase #=> "HELLO" + */ + +static VALUE +rstr_upcase(VALUE str, SEL sel) +{ + str = rb_str_new3(str); + rstr_upcase_bang(str, 0); + return str; +} + +/* + * call-seq: + * str.swapcase! => str or nil + * + * Equivalent to <code>String#swapcase</code>, but modifies the receiver in + * place, returning str, or <code>nil</code> if no changes were made. + * Note: case conversion is effective only in ASCII region. + */ + +static VALUE +rstr_swapcase_bang(VALUE str, SEL sel) +{ + rstr_modify(str); + + bool changed = false; + CHAR_ITERATE(str, + if (c >= 'A' && c <= 'Z') { + c = 'a' + (c - 'A'); + changed = true; + } + else if (c >= 'a' && c <= 'z') { + c = 'A' + (c - 'a'); + changed = true; + }); + + return changed ? str : Qnil; +} + +/* + * call-seq: + * str.swapcase => new_str + * + * Returns a copy of str with uppercase alphabetic characters converted + * to lowercase and lowercase characters converted to uppercase. + * Note: case conversion is effective only in ASCII region. + * + * "Hello".swapcase #=> "hELLO" + * "cYbEr_PuNk11".swapcase #=> "CyBeR_pUnK11" + */ + +static VALUE +rstr_swapcase(VALUE str, SEL sel) +{ + str = rb_str_new3(str); + rstr_swapcase_bang(str, 0); + return str; +} + +/* + * call-seq: + * str.capitalize! => str or nil + * + * Modifies str by converting the first character to uppercase and the + * remainder to lowercase. Returns <code>nil</code> if no changes are made. + * Note: case conversion is effective only in ASCII region. + * + * a = "hello" + * a.capitalize! #=> "Hello" + * a #=> "Hello" + * a.capitalize! #=> nil + */ + +static VALUE +rstr_capitalize_bang(VALUE str, SEL sel) +{ + rstr_modify(str); + + bool changed = false; + CHAR_ITERATE(str, + if (i == 0) { + if (c >= 'a' && c <= 'z') { + c = 'A' + (c - 'a'); + changed = true; + } + } + else if (c >= 'A' && c <= 'Z') { + c = 'a' + (c - 'A'); + changed = true; + }); + + return changed ? str : Qnil; +} + +/* + * call-seq: + * str.capitalize => new_str + * + * Returns a copy of str with the first character converted to uppercase + * and the remainder to lowercase. + * Note: case conversion is effective only in ASCII region. + * + * "hello".capitalize #=> "Hello" + * "HELLO".capitalize #=> "Hello" + * "123ABC".capitalize #=> "123abc" + */ + +static VALUE +rstr_capitalize(VALUE str, SEL sel) +{ + str = rb_str_new3(str); + rstr_capitalize_bang(str, 0); + return str; +} + +/* + * call-seq: * str.ljust(integer, padstr=' ') => new_str * * If integer is greater than the length of str, returns a new @@ -3555,6 +3712,13 @@ rb_objc_define_method(rb_cRubyString, "gsub!", rstr_gsub_bang, -1); rb_objc_define_method(rb_cRubyString, "downcase", rstr_downcase, 0); rb_objc_define_method(rb_cRubyString, "downcase!", rstr_downcase_bang, 0); + rb_objc_define_method(rb_cRubyString, "upcase", rstr_upcase, 0); + rb_objc_define_method(rb_cRubyString, "upcase!", rstr_upcase_bang, 0); + rb_objc_define_method(rb_cRubyString, "swapcase", rstr_swapcase, 0); + rb_objc_define_method(rb_cRubyString, "swapcase!", rstr_swapcase_bang, 0); + rb_objc_define_method(rb_cRubyString, "capitalize", rstr_capitalize, 0); + rb_objc_define_method(rb_cRubyString, "capitalize!", + rstr_capitalize_bang, 0); rb_objc_define_method(rb_cRubyString, "ljust", rstr_ljust, -1); rb_objc_define_method(rb_cRubyString, "rjust", rstr_rjust, -1); rb_objc_define_method(rb_cRubyString, "center", rstr_center, -1); @@ -3773,6 +3937,9 @@ { rb_str_t *str = str_alloc(rb_obj_class(source)); str_replace(str, source); + if (OBJ_TAINTED(source)) { + OBJ_TAINT(str); + } return (VALUE)str; }

1 0

[3666] MacRuby/branches/icu
by source_changes＠macosforge.org 02 Mar '10

02 Mar '10

Revision: 3666 http://trac.macosforge.org/projects/ruby/changeset/3666 Author: lsansonetti(a)apple.com Date: 2010-03-02 13:23:07 -0800 (Tue, 02 Mar 2010) Log Message: ----------- added Symbol#<=>, fixed some bugs in String#<=> Modified Paths: -------------- MacRuby/branches/icu/string.c MacRuby/branches/icu/symbol.c Modified: MacRuby/branches/icu/string.c =================================================================== --- MacRuby/branches/icu/string.c 2010-03-02 20:38:40 UTC (rev 3665) +++ MacRuby/branches/icu/string.c 2010-03-02 21:23:07 UTC (rev 3666) @@ -843,53 +843,39 @@ return 0; } - if (self->length_in_bytes == 0) { - if (str->length_in_bytes == 0) { - // both strings are empty - return 0; - } - else { - // only self is empty - goto bad_length; - } + if (self->length_in_bytes == 0 && str->length_in_bytes == 0) { + // both strings are empty + return 0; } - else if (str->length_in_bytes == 0) { - // only str is empty - goto bad_length; + + if (str_compatible_encoding(self, str) == NULL) { + // incompatible encodings + return -1; } - if (str_compatible_encoding(self, str) != NULL) { - if (str_is_stored_in_uchars(self) == str_is_stored_in_uchars(str)) { - if (self->length_in_bytes != str->length_in_bytes) { - goto bad_length; - } - else { - return memcmp(self->data.bytes, str->data.bytes, - self->length_in_bytes); - } + if (str_is_stored_in_uchars(self) != str_is_stored_in_uchars(str)) { + // one is in uchars and the other is in binary + if (!str_try_making_data_uchars(self) + || !str_try_making_data_uchars(str)) { + // one is in uchars but the other one can't be converted in + // uchars + return -1; } - else { // one is in uchars and the other is in binary - if (!str_try_making_data_uchars(self) - || !str_try_making_data_uchars(str)) { - // one is in uchars but the other one can't be converted in - // uchars - return -1; - } - if (self->length_in_bytes != str->length_in_bytes) { - goto bad_length; - } - else { - return memcmp(self->data.bytes, str->data.bytes, - self->length_in_bytes); - } - } } - else { // incompatible encodings - return -1; - } -bad_length: - return self->length_in_bytes > str->length_in_bytes ? 1 : -1; + const long min_len = self->length_in_bytes < str->length_in_bytes + ? self->length_in_bytes : str->length_in_bytes; + + const int res = memcmp(self->data.bytes, str->data.bytes, min_len); + + if (res == 0) { + if (self->length_in_bytes == str->length_in_bytes) { + return 0; + } + return self->length_in_bytes > str->length_in_bytes + ? 1 : -1; + } + return res > 1 ? 1 : -1; } static long @@ -1931,7 +1917,6 @@ * Two strings are equal if they have the same length and content. */ - static VALUE rstr_eql(VALUE self, SEL sel, VALUE other) { @@ -4216,3 +4201,9 @@ abort(); // TODO } } + +int +rb_str_cmp(VALUE str1, VALUE str2) +{ + return str_compare(str_need_string(str1), str_need_string(str2)); +} Modified: MacRuby/branches/icu/symbol.c =================================================================== --- MacRuby/branches/icu/symbol.c 2010-03-02 20:38:40 UTC (rev 3665) +++ MacRuby/branches/icu/symbol.c 2010-03-02 21:23:07 UTC (rev 3666) @@ -234,6 +234,23 @@ } /* + * call-seq: + * + * str <=> other => -1, 0, +1 or nil + * + * Compares _sym_ with _other_ in string form. + */ + +static VALUE +rsym_cmp(VALUE sym, SEL sel, VALUE other) +{ + if (TYPE(other) != T_SYMBOL) { + return Qnil; + } + return INT2FIX(rb_str_cmp(RSYM(sym)->str, RSYM(other)->str)); +} + +/* * call-seq: * sym == obj => true or false * @@ -370,6 +387,7 @@ rsym_all_symbols, 0); rb_objc_define_method(rb_cSymbol, "==", rsym_equal, 1); + rb_objc_define_method(rb_cSymbol, "<=>", rsym_cmp, 1); rb_objc_define_method(rb_cSymbol, "eql?", rsym_equal, 1); //rb_objc_define_method(rb_cSymbol, "<=>", rsym_cmp, 1); rb_objc_define_method(rb_cSymbol, "inspect", rsym_inspect, 0);

1 0

[3665] MacRuby/branches/icu/re.cpp
by source_changes＠macosforge.org 02 Mar '10

02 Mar '10

Revision: 3665 http://trac.macosforge.org/projects/ruby/changeset/3665 Author: lsansonetti(a)apple.com Date: 2010-03-02 12:38:40 -0800 (Tue, 02 Mar 2010) Log Message: ----------- change an assert by an exception, this way we can progress in the specs Modified Paths: -------------- MacRuby/branches/icu/re.cpp Modified: MacRuby/branches/icu/re.cpp =================================================================== --- MacRuby/branches/icu/re.cpp 2010-03-02 20:26:41 UTC (rev 3664) +++ MacRuby/branches/icu/re.cpp 2010-03-02 20:38:40 UTC (rev 3665) @@ -408,7 +408,9 @@ int rb_reg_search(VALUE re, VALUE str, int pos, bool reverse) { - assert(!reverse); // TODO + if (reverse) { + rb_raise(rb_eRuntimeError, "reverse searching is not implemented yet"); + } const long len = rb_str_chars_len(str); if (pos > len || pos < 0) {

1 0

[3664] MacRuby/branches/icu/spec/frozen/core/matchdata
by source_changes＠macosforge.org 02 Mar '10

02 Mar '10

Revision: 3664 http://trac.macosforge.org/projects/ruby/changeset/3664 Author: lsansonetti(a)apple.com Date: 2010-03-02 12:26:41 -0800 (Tue, 02 Mar 2010) Log Message: ----------- skipping specs using named captures Modified Paths: -------------- MacRuby/branches/icu/spec/frozen/core/matchdata/element_reference_spec.rb MacRuby/branches/icu/spec/frozen/core/matchdata/names_spec.rb Modified: MacRuby/branches/icu/spec/frozen/core/matchdata/element_reference_spec.rb =================================================================== --- MacRuby/branches/icu/spec/frozen/core/matchdata/element_reference_spec.rb 2010-03-02 07:00:25 UTC (rev 3663) +++ MacRuby/branches/icu/spec/frozen/core/matchdata/element_reference_spec.rb 2010-03-02 20:26:41 UTC (rev 3664) @@ -17,4 +17,5 @@ end end -language_version __FILE__, "element_reference" +# MacRuby does not support named captures yet. +#language_version __FILE__, "element_reference" Modified: MacRuby/branches/icu/spec/frozen/core/matchdata/names_spec.rb =================================================================== --- MacRuby/branches/icu/spec/frozen/core/matchdata/names_spec.rb 2010-03-02 07:00:25 UTC (rev 3663) +++ MacRuby/branches/icu/spec/frozen/core/matchdata/names_spec.rb 2010-03-02 20:26:41 UTC (rev 3664) @@ -1,3 +1,4 @@ require File.dirname(__FILE__) + '/../../spec_helper' -language_version __FILE__, "names" +# MacRuby does not support named captures yet. +#language_version __FILE__, "names"

1 0

[3663] MacRuby/branches/icu
by source_changes＠macosforge.org 02 Mar '10

02 Mar '10

Revision: 3663 http://trac.macosforge.org/projects/ruby/changeset/3663 Author: lsansonetti(a)apple.com Date: 2010-03-01 23:00:25 -0800 (Mon, 01 Mar 2010) Log Message: ----------- unicode string formats (a work in progress) Modified Paths: -------------- MacRuby/branches/icu/encoding.h MacRuby/branches/icu/sprintf.c MacRuby/branches/icu/string.c Modified: MacRuby/branches/icu/encoding.h =================================================================== --- MacRuby/branches/icu/encoding.h 2010-03-02 06:50:51 UTC (rev 3662) +++ MacRuby/branches/icu/encoding.h 2010-03-02 07:00:25 UTC (rev 3663) @@ -302,6 +302,8 @@ UChar rb_str_get_uchar(VALUE str, long pos); void rb_str_append_uchar(VALUE str, UChar c); unsigned long rb_str_hash_uchars(const UChar *chars, long chars_len); +long rb_uchar_strtol(UniChar *chars, long chars_len, long pos, + long *end_offset); // Return a string object appropriate for bstr_ calls. This does nothing for // data/binary RubyStrings. Modified: MacRuby/branches/icu/sprintf.c =================================================================== --- MacRuby/branches/icu/sprintf.c 2010-03-02 06:50:51 UTC (rev 3662) +++ MacRuby/branches/icu/sprintf.c 2010-03-02 07:00:25 UTC (rev 3663) @@ -13,6 +13,7 @@ #include "ruby/ruby.h" #include "ruby/encoding.h" +#include "encoding.h" /* * call-seq: @@ -289,13 +290,10 @@ VALUE rb_enc_sprintf(rb_encoding *enc, const char *format, ...) { - VALUE result; va_list ap; - va_start(ap, format); - result = rb_enc_vsprintf(enc, format, ap); + VALUE result = rb_enc_vsprintf(enc, format, ap); va_end(ap); - return result; } @@ -308,13 +306,10 @@ VALUE rb_sprintf(const char *format, ...) { - VALUE result; va_list ap; - va_start(ap, format); - result = rb_vsprintf(format, ap); + VALUE result = rb_vsprintf(format, ap); va_end(ap); - return result; } @@ -356,42 +351,50 @@ width -= slen; do { CFStringInsert((CFMutableStringRef)arg, start, pad); - } while (--width > 0); + } + while (--width > 0); } static long -cstr_update(char **str, unsigned long start, unsigned long num, char *replace) +cstr_update(UChar **str, long *str_len, long start, long num, VALUE replace) { - unsigned long len = strlen(*str) + 1; - unsigned long replace_len = strlen(replace); + const long len = *str_len; + long replace_len = replace == 0 ? 0 : rb_str_chars_len(replace); if (start + num > len) { num = len - start; } if (replace_len >= num) { - char *new_str = (char *)xmalloc(len + replace_len - num); - memcpy(new_str, *str, len); - *str = new_str; + *str_len = len + replace_len - num; + *str = (UChar *)xrealloc(*str, + sizeof(UChar) * (len + replace_len - num)); } if (replace_len != num) { - bcopy(*str + start + num, *str + start + replace_len, len - start - - num); + bcopy(*str + start + num, *str + start + replace_len, + sizeof(UChar) * (len - start - num)); } if (replace_len > 0) { - bcopy(replace, *str + start, replace_len); + UChar *replace_chars = NULL; + bool need_free = false; + rb_str_get_uchars(replace, &replace_chars, &replace_len, &need_free); + assert(replace_len > 0); + bcopy(replace_chars, *str + start, sizeof(UChar) * replace_len); + if (need_free) { + free(replace_chars); + } } return replace_len - num; } -VALUE -get_named_arg(char *format_str, unsigned long format_len, unsigned long *i, +static VALUE +get_named_arg(UChar *format_str, long format_len, unsigned long *i, VALUE hash) { if (TYPE(hash) != T_HASH) { rb_raise(rb_eArgError, "hash required for named references"); } - char closing = format_str[(*i)++] + 2; - char *str_ptr = format_str + *i; + UChar closing = format_str[(*i)++] + 2; + UChar *str_ptr = &format_str[*i]; while (*i < format_len && format_str[*i] != closing) { (*i)++; } @@ -399,26 +402,32 @@ rb_raise(rb_eArgError, "malformed name - unmatched parenthesis"); } - format_str[*i] = '\0'; - hash = rb_hash_aref(hash, rb_name2sym(str_ptr)); - format_str[*i] = closing; - return (hash); + VALUE substr = rb_unicode_str_new(str_ptr, str_ptr - format_str); + hash = rb_hash_aref(hash, ID2SYM(rb_intern_str(substr))); + return hash; } -// XXX -// - this method uses strtol to read numbers from the format string, so -// extremely large numbers get silently truncated. this should be fixed -// - switch to a cfstring format string to allow for proper encoding support - // XXX look for arguments that are altered but not duped VALUE rb_str_format(int argc, const VALUE *argv, VALUE fmt) { bool tainted = OBJ_TAINTED(fmt); - fmt = rb_str_new3(fmt); - char *format_str = (char *)RSTRING_PTR(fmt); - unsigned long format_len = strlen(format_str); - long num; + + UChar *format_str = NULL; + long format_len = 0; + bool need_free = false; + rb_str_get_uchars(fmt, &format_str, &format_len, &need_free); + if (format_len == 0) { + goto bail; + } + UChar *tmp = (UChar *)xmalloc(format_len * sizeof(UChar)); + memcpy(tmp, format_str, format_len * sizeof(UChar)); + if (need_free) { + free(format_str); + } + format_str = tmp; + + long num, pos; int j = 0; int ref_type = 0; @@ -427,7 +436,7 @@ continue; } if (format_str[i + 1] == '%') { - cstr_update(&format_str, i, 1, (char *)""); + cstr_update(&format_str, &format_len, i, 1, 0); continue; } @@ -444,7 +453,6 @@ int base = 0; CFStringRef negative_pad = NULL; CFStringRef sharp_pad = CFSTR(""); - char *str_ptr; unsigned long start = i; while (i++ < format_len) { @@ -464,16 +472,17 @@ i--; break; } - num = strtol(format_str + i, &str_ptr, 10); - if (str_ptr == format_str + i--) { + + num = rb_uchar_strtol(format_str, format_len, i, &pos); + if (pos == i--) { SET_REF_TYPE(REL_REF); width = NUM2LONG(rb_Integer(GETNTHARG(j))); j++; } - else if (*str_ptr == '$') { + else if (format_str[pos] == '$') { SET_REF_TYPE(ABS_REF); width = NUM2LONG(rb_Integer(GETNTHARG(num - 1))); - i = str_ptr - format_str; + i = pos; } } if (width < 0) { @@ -513,9 +522,9 @@ case '7': case '8': case '9': - num = strtol(format_str + i, &str_ptr, 10); - i = str_ptr - format_str; - if (*str_ptr == '$') { + num = rb_uchar_strtol(format_str, format_len, i, &pos); + i = pos; + if (format_str[pos] == '$') { if (num == 0) { rb_raise(rb_eArgError, "invalid absolute argument"); } @@ -543,23 +552,26 @@ i--; break; } - num = strtol(format_str + i, &str_ptr, 10); - if (str_ptr == format_str + i--) { + + num = rb_uchar_strtol(format_str, format_len, + i, &pos); + if (num == i--) { SET_REF_TYPE(REL_REF); precision = NUM2LONG(rb_Integer(GETNTHARG(j))); j++; } - else if (*str_ptr == '$') { + else if (format_str[pos] == '$') { SET_REF_TYPE(ABS_REF); precision = NUM2LONG(rb_Integer(GETNTHARG( num - 1))); - i = str_ptr - format_str; + i = pos; } } } else if (isdigit(format_str[i])) { - precision = strtol(format_str + i, &str_ptr, 10); - i = str_ptr - format_str - 1; + precision = rb_uchar_strtol(format_str, format_len, + i, &pos); + i = pos - 1; } else { rb_raise(rb_eArgError, "invalid precision"); @@ -653,7 +665,7 @@ break; } - arg = rb_str_new(format_str + i, 1); + arg = rb_unicode_str_new(&format_str[i], 1); if (precision_flag) { rb_str_update(arg, 0, 0, rb_big2str(LONG2NUM(precision), 10)); @@ -677,9 +689,10 @@ } rb_str_update(arg, 0, 0, (VALUE)CFSTR("%")); - asprintf(&str_ptr, RSTRING_PTR(arg), value); - arg = rb_str_new2(str_ptr); - free(str_ptr); + char *ptr; + asprintf(&ptr, RSTRING_PTR(arg), value); + arg = rb_str_new2(ptr); + free(ptr); break; } @@ -732,8 +745,10 @@ arg = rb_big2str(num, base); if (!sign_pad && IS_NEG(num) && negative_pad != NULL) { - char neg = *RSTRING_PTR(negative_pad); - str_ptr = (char *)RSTRING_PTR(arg) + 1; + break; // TODO +#if 0 + UChar neg = CFStringGetCharacterAtIndex(negative_pad, 0); + char *str_ptr = (char *)RSTRING_PTR(arg) + 1; if (base == 8) { *str_ptr |= ((~0 << 3) >> ((3 * strlen(str_ptr)) % (sizeof(BDIGIT) * 8))) & ~(~0 << 3); @@ -744,11 +759,13 @@ rb_str_update(arg, 0, num_index, (VALUE)negative_pad); rb_str_update(arg, 0, 0, (VALUE)CFSTR("..")); num_index = 2; +#endif } if (precision_flag) { - pad_format_value(arg, num_index, precision + (IS_NEG(num) && - (sign_pad || negative_pad == NULL) ? 1 : 0), + pad_format_value(arg, num_index, + precision + (IS_NEG(num) + && (sign_pad || negative_pad == NULL) ? 1 : 0), zero_pad); } if (sharp_flag && rb_cmpint(num, Qfalse, Qfalse) != 0) { @@ -773,14 +790,17 @@ } pad_format_value(arg, minus_flag ? -1 : 0, width, CFSTR(" ")); - num = cstr_update(&format_str, start, i - start + 1, - (char *)RSTRING_PTR(arg)); - format_len += num; + num = cstr_update(&format_str, &format_len, start, i - start + 1, + arg); i += num; break; } } - fmt = rb_str_new2(format_str); - return tainted ? OBJ_TAINT(fmt) : fmt; +bail: + fmt = rb_unicode_str_new(format_str, format_len); + if (tainted) { + OBJ_TAINT(fmt); + } + return fmt; } Modified: MacRuby/branches/icu/string.c =================================================================== --- MacRuby/branches/icu/string.c 2010-03-02 06:50:51 UTC (rev 3662) +++ MacRuby/branches/icu/string.c 2010-03-02 07:00:25 UTC (rev 3663) @@ -22,6 +22,8 @@ #include "ruby/node.h" #include "vm.h" +#include <unicode/unum.h> + VALUE rb_cString; VALUE rb_cNSString; VALUE rb_cNSMutableString; @@ -4152,6 +4154,25 @@ } long +rb_uchar_strtol(UniChar *chars, long chars_len, long pos, long *end_offset) +{ + assert(chars != NULL && chars_len > 0 && pos >= 0); + + UErrorCode status = U_ZERO_ERROR; + UNumberFormat *nf = unum_open(UNUM_DEFAULT, NULL, -1, NULL, NULL, &status); + assert(nf != NULL); + + int32_t parse_pos = (int32_t)pos; + int64_t val = unum_parseInt64(nf, chars, chars_len, &parse_pos, &status); + unum_close(nf); + + if (end_offset != NULL) { + *end_offset = (long)parse_pos; + } + return val; +} + +long rb_memhash(const void *ptr, long len) { CFDataRef data = CFDataCreate(NULL, (const UInt8 *)ptr, len);

1 0

[3662] MacRuby/branches/icu/compiler.cpp
by source_changes＠macosforge.org 02 Mar '10

02 Mar '10

Revision: 3662 http://trac.macosforge.org/projects/ruby/changeset/3662 Author: lsansonetti(a)apple.com Date: 2010-03-01 22:50:51 -0800 (Mon, 01 Mar 2010) Log Message: ----------- fixed unicode strings compilation Modified Paths: -------------- MacRuby/branches/icu/compiler.cpp Modified: MacRuby/branches/icu/compiler.cpp =================================================================== --- MacRuby/branches/icu/compiler.cpp 2010-02-28 14:53:32 UTC (rev 3661) +++ MacRuby/branches/icu/compiler.cpp 2010-03-02 06:50:51 UTC (rev 3662) @@ -2790,8 +2790,7 @@ // // 10.times { s = 'foo'; s << 'bar' } // - const size_t str_len = RSTRING_LEN(val); - if (str_len == 0) { + if (rb_str_chars_len(val) == 0) { if (newString3Func == NULL) { newString3Func = cast<Function>( module->getOrInsertFunction( @@ -2805,6 +2804,7 @@ bool need_free = false; rb_str_get_uchars(val, &chars, &chars_len, &need_free); + assert(chars_len > 0); GlobalVariable *str_gvar = compile_const_global_ustring(chars, chars_len); @@ -2830,7 +2830,7 @@ std::vector<Value *> params; params.push_back(load); - params.push_back(ConstantInt::get(Int32Ty, str_len)); + params.push_back(ConstantInt::get(Int32Ty, chars_len)); return CallInst::Create(newString2Func, params.begin(), params.end(), "", bb);

1 0