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

[3631] MacRuby/branches/icu/string.c
by source_changes＠macosforge.org 26 Feb '10

26 Feb '10

Revision: 3631 http://trac.macosforge.org/projects/ruby/changeset/3631 Author: lsansonetti(a)apple.com Date: 2010-02-25 18:52:51 -0800 (Thu, 25 Feb 2010) Log Message: ----------- renamed str_substr into rstr_substr, added rstr_splice which sanitizes arguments, added rb_str_update Modified Paths: -------------- MacRuby/branches/icu/string.c Modified: MacRuby/branches/icu/string.c =================================================================== --- MacRuby/branches/icu/string.c 2010-02-26 02:43:04 UTC (rev 3630) +++ MacRuby/branches/icu/string.c 2010-02-26 02:52:51 UTC (rev 3631) @@ -1094,7 +1094,7 @@ } static VALUE -str_substr(VALUE str, long beg, long len) +rstr_substr(VALUE str, long beg, long len) { if (len < 0) { return Qnil; @@ -1118,6 +1118,31 @@ return substr == NULL ? Qnil : (VALUE)substr; } +static void +rstr_splice(VALUE self, long beg, long len, VALUE str) +{ + if (len < 0) { + rb_raise(rb_eIndexError, "negative length %ld", len); + } + + const long slen = str_length(RSTR(self), false); + if (slen < beg) { +out_of_range: + rb_raise(rb_eIndexError, "index %ld out of string", beg); + } + if (beg < 0) { + if (-beg > slen) { + goto out_of_range; + } + beg += slen; + } + if (slen < len || slen < beg + len) { + len = slen - beg; + } + + str_splice(RSTR(self), beg, len, str_need_string(str), false); +} + static VALUE str_trim(VALUE str) { @@ -1507,7 +1532,7 @@ if (TYPE(argv[0]) == T_REGEXP) { return rb_str_subpat(str, argv[0], NUM2INT(argv[1])); } - return str_substr(str, NUM2LONG(argv[0]), NUM2LONG(argv[1])); + return rstr_substr(str, NUM2LONG(argv[0]), NUM2LONG(argv[1])); } if (argc != 1) { @@ -1517,7 +1542,7 @@ VALUE indx = argv[0]; switch (TYPE(indx)) { case T_FIXNUM: - str = str_substr(str, FIX2LONG(indx), 1); + str = rstr_substr(str, FIX2LONG(indx), 1); if (!NIL_P(str) && str_length(RSTR(str), true) == 0) { return Qnil; } @@ -1556,9 +1581,9 @@ case Qnil: return Qnil; default: - return str_substr(str, beg, len); + return rstr_substr(str, beg, len); } - str = str_substr(str, NUM2LONG(indx), 1); + str = rstr_substr(str, NUM2LONG(indx), 1); if (!NIL_P(str) && str_length(RSTR(str), true) == 0) { return Qnil; } @@ -1693,7 +1718,7 @@ rstr_getchar(VALUE self, SEL sel, VALUE index) { const long idx = FIX2LONG(index); - return str_substr(self, idx, 1); + return rstr_substr(self, idx, 1); } /* @@ -2370,7 +2395,7 @@ if (spat != Qnil) { if (spat_len == 0) { do { - VALUE substr = str_substr(str, beg, 1); + VALUE substr = rstr_substr(str, beg, 1); rb_ary_push(result, substr); beg++; if (beg >= len) { @@ -2387,7 +2412,7 @@ if (pos == -1) { break; } - VALUE substr = str_substr(str, beg, pos - beg); + VALUE substr = rstr_substr(str, beg, pos - beg); if (!awk_split || rb_str_chars_len(str_trim(substr)) > 0) { rb_ary_push(result, substr); } @@ -2416,7 +2441,7 @@ if (beg == pos && results[0].beg == results[0].end) { if (last_null) { - rb_ary_push(result, str_substr(str, beg, 1)); + rb_ary_push(result, rstr_substr(str, beg, 1)); beg = start; } else { @@ -2426,7 +2451,7 @@ } } else { - rb_ary_push(result, str_substr(str, beg, pos - beg)); + rb_ary_push(result, rstr_substr(str, beg, pos - beg)); beg = results[0].end; } last_null = false; @@ -2678,7 +2703,7 @@ repl = rb_obj_as_string(rb_yield(rb_reg_nth_match(0, match))); } else { - repl = rb_hash_aref(hash, str_substr(str, results[0].beg, + repl = rb_hash_aref(hash, rstr_substr(str, results[0].beg, results[0].end - results[0].beg)); repl = rb_obj_as_string(repl); } @@ -2796,7 +2821,7 @@ return bang ? Qnil : rstr_dup(str, 0); } str_concat_string(RSTR(dest), - RSTR(str_substr(str, offset, len - offset))); + RSTR(rstr_substr(str, offset, len - offset))); break; } @@ -2811,7 +2836,7 @@ val = rb_obj_as_string(rb_yield(rb_reg_nth_match(0, match))); } else { - val = rb_hash_aref(hash, str_substr(str, results[0].beg, + val = rb_hash_aref(hash, rstr_substr(str, results[0].beg, results[0].end - results[0].beg)); val = rb_obj_as_string(val); } @@ -2826,7 +2851,7 @@ if (pos - offset > 0) { str_concat_string(RSTR(dest), - RSTR(str_substr(str, offset, pos - offset))); + RSTR(rstr_substr(str, offset, pos - offset))); } str_concat_string(RSTR(dest), str_need_string(val)); @@ -2981,7 +3006,7 @@ { do { if (padwidth > width) { - pad = RSTR(str_substr((VALUE)pad, 0, width)); + pad = RSTR(rstr_substr((VALUE)pad, 0, width)); } str_insert(str, index, pad, false); width -= padwidth; @@ -3322,7 +3347,7 @@ substr_len = off - pos + 1; } - VALUE substr = str_substr(str, pos, substr_len); + VALUE substr = rstr_substr(str, pos, substr_len); if (tainted) { OBJ_TAINT(substr); } @@ -4045,7 +4070,7 @@ rb_str_subseq(VALUE str, long beg, long len) { if (IS_RSTR(str)) { - return str_substr(str, beg, len); + return rstr_substr(str, beg, len); } abort(); // TODO } @@ -4059,5 +4084,10 @@ void rb_str_update(VALUE str, long beg, long len, VALUE val) { - abort(); // TODO + if (IS_RSTR(str)) { + rstr_splice(str, beg, len, val); + } + else { + abort(); // TODO + } }

1 0

[3630] MacRuby/branches/icu
by source_changes＠macosforge.org 26 Feb '10

26 Feb '10

Revision: 3630 http://trac.macosforge.org/projects/ruby/changeset/3630 Author: lsansonetti(a)apple.com Date: 2010-02-25 18:43:04 -0800 (Thu, 25 Feb 2010) Log Message: ----------- added #%, added remaining NSString primitives Modified Paths: -------------- MacRuby/branches/icu/objc.h MacRuby/branches/icu/objc.m MacRuby/branches/icu/sprintf.c MacRuby/branches/icu/string.c Modified: MacRuby/branches/icu/objc.h =================================================================== --- MacRuby/branches/icu/objc.h 2010-02-26 01:49:34 UTC (rev 3629) +++ MacRuby/branches/icu/objc.h 2010-02-26 02:43:04 UTC (rev 3630) @@ -234,6 +234,8 @@ #define RB2OC(obj) (rb_rval_to_ocid((VALUE)obj)) #define OC2RB(obj) (rb_ocid_to_rval((id)obj)) +void rb_objc_exception_raise(const char *name, const char *message); + #if defined(__cplusplus) } #endif Modified: MacRuby/branches/icu/objc.m =================================================================== --- MacRuby/branches/icu/objc.m 2010-02-26 01:49:34 UTC (rev 3629) +++ MacRuby/branches/icu/objc.m 2010-02-26 02:43:04 UTC (rev 3630) @@ -608,6 +608,14 @@ return rb_exc_new2(rb_eRuntimeError, buf); } +void +rb_objc_exception_raise(const char *name, const char *message) +{ + assert(name != NULL && message != NULL); + [[NSException exceptionWithName:[NSString stringWithUTF8String:name] + reason:[NSString stringWithUTF8String:message] userInfo:nil] raise]; +} + size_t rb_objc_type_size(const char *type) { Modified: MacRuby/branches/icu/sprintf.c =================================================================== --- MacRuby/branches/icu/sprintf.c 2010-02-26 01:49:34 UTC (rev 3629) +++ MacRuby/branches/icu/sprintf.c 2010-02-26 02:43:04 UTC (rev 3630) @@ -12,10 +12,7 @@ #include <stdarg.h> #include "ruby/ruby.h" -#include "ruby/node.h" #include "ruby/encoding.h" -#include "vm.h" -#include "compiler.h" /* * call-seq: Modified: MacRuby/branches/icu/string.c =================================================================== --- MacRuby/branches/icu/string.c 2010-02-26 01:49:34 UTC (rev 3629) +++ MacRuby/branches/icu/string.c 2010-02-26 02:43:04 UTC (rev 3630) @@ -1747,6 +1747,31 @@ /* * call-seq: + * str % arg => new_str + * + * Format---Uses str as a format specification, and returns the result + * of applying it to arg. If the format specification contains more than + * one substitution, then arg must be an <code>Array</code> containing + * the values to be substituted. See <code>Kernel::sprintf</code> for details + * of the format string. + * + * "%05d" % 123 #=> "00123" + * "%-5s: %08x" % [ "ID", self.object_id ] #=> "ID : 200e14d6" + */ + +static VALUE +rstr_format(VALUE str, SEL sel, VALUE arg) +{ + VALUE tmp = rb_check_array_type(arg); + + if (!NIL_P(tmp)) { + return rb_str_format(RARRAY_LEN(tmp), RARRAY_PTR(tmp), str); + } + return rb_str_format(1, &arg, str); +} + +/* + * call-seq: * str << fixnum => str * str.concat(fixnum) => str * str << obj => str @@ -3316,6 +3341,27 @@ // NSString primitives. +static void +check_bounds(void *rcv, long pos, bool can_be_end) +{ + const long len = str_length(RSTR(rcv), true); + if (pos >= 0) { + if (can_be_end) { + if (pos <= len) { + return; + } + } + else if (pos < len) { + return; + } + } + + char buf[100]; + snprintf(buf, sizeof buf, "Position (%ld) out of bounds (%ld)", + pos, len); + rb_objc_exception_raise("NSRangeException", buf); +} + static CFIndex rstr_imp_length(void *rcv, SEL sel) { @@ -3325,9 +3371,38 @@ static UniChar rstr_imp_characterAtIndex(void *rcv, SEL sel, CFIndex idx) { + check_bounds(rcv, idx, false); return str_get_uchar(RSTR(rcv), idx, true); } +static void +rstr_imp_getCharactersRange(void *rcv, SEL sel, UniChar *buffer, CFRange range) +{ + check_bounds(rcv, range.location + range.length, true); + if (range.length > 0) { + if (str_try_making_data_uchars(RSTR(rcv))) { + memcpy(buffer, &RSTR(rcv)->data.uchars[range.location], + sizeof(UniChar) * range.length); + } + else { + for (long i = range.location, j = 0; + i < range.location + range.length; + i++, j++) { + buffer[j] = RSTR(rcv)->data.bytes[i]; + } + } + } +} + +static void +rstr_imp_replaceCharactersInRangeWithString(void *rcv, SEL sel, CFRange range, + void *str) +{ + check_bounds(rcv, range.location + range.length, true); + rb_str_t *spat = str_need_string((VALUE)str); + str_splice(RSTR(rcv), range.location, range.length, spat, true); +} + void Init_String(void) { @@ -3367,6 +3442,7 @@ rb_objc_define_method(rb_cRubyString, "rindex", rstr_rindex, -1); rb_objc_define_method(rb_cRubyString, "+", rstr_plus, 1); rb_objc_define_method(rb_cRubyString, "*", rstr_times, 1); + rb_objc_define_method(rb_cRubyString, "%", rstr_format, 1); rb_objc_define_method(rb_cRubyString, "<<", rstr_concat, 1); rb_objc_define_method(rb_cRubyString, "concat", rstr_concat, 1); rb_objc_define_method(rb_cRubyString, "==", rstr_equal, 1); @@ -3418,13 +3494,11 @@ (IMP)rstr_imp_length); rb_objc_install_method2((Class)rb_cRubyString, "characterAtIndex:", (IMP)rstr_imp_characterAtIndex); -#if 0 - rb_objc_install_method2(rb_cRubyString, "getCharacters:range:", + rb_objc_install_method2((Class)rb_cRubyString, "getCharacters:range:", (IMP)rstr_imp_getCharactersRange); - rb_objc_install_method2(rb_cRubyString, + rb_objc_install_method2((Class)rb_cRubyString, "replaceCharactersInRange:withString:", (IMP)rstr_imp_replaceCharactersInRangeWithString); -#endif rb_fs = Qnil; rb_define_variable("$;", &rb_fs);

1 0

[3629] MacRuby/branches/icu/string.c
by source_changes＠macosforge.org 26 Feb '10

26 Feb '10

Revision: 3629 http://trac.macosforge.org/projects/ruby/changeset/3629 Author: lsansonetti(a)apple.com Date: 2010-02-25 17:49:34 -0800 (Thu, 25 Feb 2010) Log Message: ----------- added #each_line Modified Paths: -------------- MacRuby/branches/icu/string.c Modified: MacRuby/branches/icu/string.c =================================================================== --- MacRuby/branches/icu/string.c 2010-02-26 01:25:31 UTC (rev 3628) +++ MacRuby/branches/icu/string.c 2010-02-26 01:49:34 UTC (rev 3629) @@ -3167,7 +3167,7 @@ */ static VALUE -rstr_lstrip(VALUE str) +rstr_lstrip(VALUE str, SEL sel) { str = rstr_dup(str, 0); rstr_lstrip_bang(str, 0); @@ -3204,13 +3204,116 @@ */ static VALUE -rstr_rstrip(VALUE str) +rstr_rstrip(VALUE str, SEL sel) { str = rstr_dup(str, 0); rstr_rstrip_bang(str, 0); return str; } +/* + * Document-method: lines + * call-seq: + * str.lines(separator=$/) => anEnumerator + * str.lines(separator=$/) {|substr| block } => str + * + * Returns an enumerator that gives each line in the string. If a block is + * given, it iterates over each line in the string. + * + * "foo\nbar\n".lines.to_a #=> ["foo\n", "bar\n"] + * "foo\nb ar".lines.sort #=> ["b ar", "foo\n"] + */ + +/* + * Document-method: each_line + * call-seq: + * str.each_line(separator=$/) {|substr| block } => str + * + * Splits str using the supplied parameter as the record separator + * (<code>$/</code> by default), passing each substring in turn to the supplied + * block. If a zero-length record separator is supplied, the string is split + * into paragraphs delimited by multiple successive newlines. + * + * print "Example one\n" + * "hello\nworld".each {|s| p s} + * print "Example two\n" + * "hello\nworld".each('l') {|s| p s} + * print "Example three\n" + * "hello\n\n\nworld".each('') {|s| p s} + * + * produces: + * + * Example one + * "hello\n" + * "world" + * Example two + * "hel" + * "l" + * "o\nworl" + * "d" + * Example three + * "hello\n\n\n" + * "world" + */ + +static VALUE +rstr_each_line(VALUE str, SEL sel, int argc, VALUE *argv) +{ + VALUE rs; + if (rb_scan_args(argc, argv, "01", &rs) == 0) { + rs = rb_rs; + } + + RETURN_ENUMERATOR(str, argc, argv); + + if (NIL_P(rs)) { + rb_yield(str); + return str; + } + + StringValue(rs); + + rb_str_t *rs_str = str_need_string(rs); + if (rs_str->length_in_bytes == 0) { + rs_str = str_need_string(rb_default_rs); + } + + const long len = str_length(RSTR(str), false); + const bool tainted = OBJ_TAINTED(str); + + long pos = 0; + do { + const long off = str_index_for_string(RSTR(str), rs_str, pos, + false, false); + + long substr_len = 0; + if (off < 0) { + if (pos == len) { + break; + } + substr_len = len - pos; + } + else { + substr_len = off - pos + 1; + } + + VALUE substr = str_substr(str, pos, substr_len); + if (tainted) { + OBJ_TAINT(substr); + } + rb_yield(substr); + RETURN_IF_BROKEN(); + + if (off < 0) { + break; + } + pos = off + 1; + } + while (true); + + return str; +} + // NSString primitives. static CFIndex @@ -3300,6 +3403,8 @@ rb_objc_define_method(rb_cRubyString, "strip!", rstr_strip_bang, 0); rb_objc_define_method(rb_cRubyString, "lstrip!", rstr_lstrip_bang, 0); rb_objc_define_method(rb_cRubyString, "rstrip!", rstr_rstrip_bang, 0); + rb_objc_define_method(rb_cRubyString, "lines", rstr_each_line, -1); + rb_objc_define_method(rb_cRubyString, "each_line", rstr_each_line, -1); // Added for MacRuby (debugging). rb_objc_define_method(rb_cRubyString, "__chars_count__",

1 0

[3628] MacRuby/branches/icu
by source_changes＠macosforge.org 26 Feb '10

26 Feb '10

Revision: 3628 http://trac.macosforge.org/projects/ruby/changeset/3628 Author: lsansonetti(a)apple.com Date: 2010-02-25 17:25:31 -0800 (Thu, 25 Feb 2010) Log Message: ----------- better string/symbol interaction Modified Paths: -------------- MacRuby/branches/icu/encoding.h MacRuby/branches/icu/string.c Modified: MacRuby/branches/icu/encoding.h =================================================================== --- MacRuby/branches/icu/encoding.h 2010-02-26 01:07:51 UTC (rev 3627) +++ MacRuby/branches/icu/encoding.h 2010-02-26 01:25:31 UTC (rev 3628) @@ -290,7 +290,9 @@ STRING_VALID_ENCODING); } +VALUE mr_enc_s_is_compatible(VALUE klass, SEL sel, VALUE str1, VALUE str2); VALUE str_inspect(rb_str_t *str, bool dump); +VALUE rb_str_intern_fast(VALUE str); // The following functions should always been prefered over anything else, // especially if this "else" is RSTRING_PTR and RSTRING_LEN. @@ -303,8 +305,6 @@ void rb_str_append_uchar(VALUE str, UChar c); unsigned long rb_str_hash_uchars(const UChar *chars, long chars_len); -VALUE mr_enc_s_is_compatible(VALUE klass, SEL sel, VALUE str1, VALUE str2); - // Return a string object appropriate for bstr_ calls. This does nothing for // data/binary RubyStrings. VALUE rb_str_bstr(VALUE str); Modified: MacRuby/branches/icu/string.c =================================================================== --- MacRuby/branches/icu/string.c 2010-02-26 01:07:51 UTC (rev 3627) +++ MacRuby/branches/icu/string.c 2010-02-26 01:25:31 UTC (rev 3628) @@ -1035,8 +1035,17 @@ static rb_str_t * str_need_string(VALUE str) { - if (TYPE(str) != T_STRING) { - str = rb_str_to_str(str); + switch (TYPE(str)) { + case T_SYMBOL: + str = rb_sym2str(str); + break; + + case T_STRING: + break; + + default: + str = rb_str_to_str(str); + break; } return IS_RSTR(str) ? (rb_str_t *)str : str_new_from_cfstring((CFStringRef)str); @@ -1948,8 +1957,7 @@ if (OBJ_TAINTED(self) && rb_safe_level() >= 1) { rb_raise(rb_eSecurityError, "Insecure: can't intern tainted string"); } - str_make_data_binary(RSTR(self)); - return ID2SYM(rb_intern(RSTR(self)->data.bytes)); + return rb_str_intern_fast(self); } /* @@ -3441,6 +3449,13 @@ return str; } +VALUE +rb_str_intern_fast(VALUE str) +{ + // TODO: this currently does 2 hash lookups, could be optimized. + return ID2SYM(rb_intern_str(str)); +} + // MRI C-API compatibility. VALUE

1 0

[3627] MacRuby/branches/icu/string.c
by source_changes＠macosforge.org 26 Feb '10

26 Feb '10

Revision: 3627 http://trac.macosforge.org/projects/ruby/changeset/3627 Author: lsansonetti(a)apple.com Date: 2010-02-25 17:07:51 -0800 (Thu, 25 Feb 2010) Log Message: ----------- added #<=> Modified Paths: -------------- MacRuby/branches/icu/string.c Modified: MacRuby/branches/icu/string.c =================================================================== --- MacRuby/branches/icu/string.c 2010-02-26 00:29:31 UTC (rev 3626) +++ MacRuby/branches/icu/string.c 2010-02-26 01:07:51 UTC (rev 3627) @@ -825,36 +825,36 @@ self->length_in_bytes = new_length_in_bytes; } -static bool -str_is_equal_to_string(rb_str_t *self, rb_str_t *str) +static int +str_compare(rb_str_t *self, rb_str_t *str) { if (self == str) { - return true; + return 0; } if (self->length_in_bytes == 0) { if (str->length_in_bytes == 0) { // both strings are empty - return true; + goto bad_length; } else { // only self is empty - return false; + goto bad_length; } } else if (str->length_in_bytes == 0) { // only str is empty - return false; + goto bad_length; } 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) { - return false; + goto bad_length; } else { - return (memcmp(self->data.bytes, str->data.bytes, - self->length_in_bytes) == 0); + return memcmp(self->data.bytes, str->data.bytes, + self->length_in_bytes); } } else { // one is in uchars and the other is in binary @@ -862,20 +862,23 @@ || !str_try_making_data_uchars(str)) { // one is in uchars but the other one can't be converted in // uchars - return false; + return -1; } if (self->length_in_bytes != str->length_in_bytes) { - return false; + goto bad_length; } else { - return (memcmp(self->data.bytes, str->data.bytes, - self->length_in_bytes) == 0); + return memcmp(self->data.bytes, str->data.bytes, + self->length_in_bytes); } } } else { // incompatible encodings - return false; + return -1; } + +bad_length: + return self->length_in_bytes > str->length_in_bytes ? 1 : -1; } static long @@ -1795,11 +1798,63 @@ } return rb_equal(other, self); } - return str_is_equal_to_string(RSTR(self), str_need_string(other)) + return str_compare(RSTR(self), str_need_string(other)) == 0 ? Qtrue : Qfalse; } /* + * call-seq: + * str <=> other_str => -1, 0, +1 + * + * Comparison---Returns -1 if other_str is less than, 0 if + * other_str is equal to, and +1 if other_str is greater than + * str. If the strings are of different lengths, and the strings are + * equal when compared up to the shortest length, then the longer string is + * considered greater than the shorter one. In older versions of Ruby, setting + * <code>$=</code> allowed case-insensitive comparisons; this is now deprecated + * in favor of using <code>String#casecmp</code>. + * + * <code><=></code> is the basis for the methods <code><</code>, + * <code><=</code>, <code>></code>, <code>>=</code>, and <code>between?</code>, + * included from module <code>Comparable</code>. The method + * <code>String#==</code> does not use <code>Comparable#==</code>. + * + * "abcdef" <=> "abcde" #=> 1 + * "abcdef" <=> "abcdef" #=> 0 + * "abcdef" <=> "abcdefg" #=> -1 + * "abcdef" <=> "ABCDEF" #=> 1 + */ + +static VALUE +rstr_cmp(VALUE self, SEL sel, VALUE other) +{ + long result; + + if (TYPE(other) != T_STRING) { + if (!rb_respond_to(other, rb_intern("to_str"))) { + return Qnil; + } + else if (!rb_vm_respond_to(other, selCmp, false)) { + return Qnil; + } + else { + VALUE tmp = rb_vm_call(other, selCmp, 1, &self, false); + if (NIL_P(tmp)) { + return Qnil; + } + if (!FIXNUM_P(tmp)) { + return rb_vm_call(LONG2FIX(0), selMINUS, 1, &tmp, false); + } + result = -FIX2LONG(tmp); + } + } + else { + result = str_compare(RSTR(self), str_need_string(other)); + } + return LONG2NUM(result); +} + +/* * call-seq: * str.eql?(other) => true or false * @@ -1816,7 +1871,7 @@ if (TYPE(other) != T_STRING) { return Qfalse; } - return str_is_equal_to_string(RSTR(self), str_need_string(other)) + return str_compare(RSTR(self), str_need_string(other)) == 0 ? Qtrue : Qfalse; } @@ -3204,6 +3259,7 @@ rb_objc_define_method(rb_cRubyString, "<<", rstr_concat, 1); rb_objc_define_method(rb_cRubyString, "concat", rstr_concat, 1); rb_objc_define_method(rb_cRubyString, "==", rstr_equal, 1); + rb_objc_define_method(rb_cRubyString, "<=>", rstr_cmp, 1); rb_objc_define_method(rb_cRubyString, "eql?", rstr_eql, 1); rb_objc_define_method(rb_cRubyString, "include?", rstr_includes, 1); rb_objc_define_method(rb_cRubyString, "to_s", rstr_to_s, 0);

1 0

[3626] MacRuby/branches/icu
by source_changes＠macosforge.org 26 Feb '10

26 Feb '10

Revision: 3626 http://trac.macosforge.org/projects/ruby/changeset/3626 Author: lsansonetti(a)apple.com Date: 2010-02-25 16:29:31 -0800 (Thu, 25 Feb 2010) Log Message: ----------- make sure File.join and File.expand_path return RubyStrings (for now) Modified Paths: -------------- MacRuby/branches/icu/file.c MacRuby/branches/icu/objc.m Modified: MacRuby/branches/icu/file.c =================================================================== --- MacRuby/branches/icu/file.c 2010-02-25 23:58:38 UTC (rev 3625) +++ MacRuby/branches/icu/file.c 2010-02-26 00:29:31 UTC (rev 3626) @@ -2575,13 +2575,11 @@ static VALUE rb_file_join(VALUE ary, VALUE sep) { - CFMutableStringRef res = CFStringCreateMutable(NULL, 0); - CFStringRef sep_cf = (CFStringRef)sep; + VALUE res = rb_str_new(NULL, 0); const long count = RARRAY_LEN(ary); if (count > 0) { - long i; - for (i = 0; i < count; i++) { + for (long i = 0; i < count; i++) { VALUE tmp = RARRAY_AT(ary, i); switch (TYPE(tmp)) { case T_STRING: @@ -2595,9 +2593,9 @@ FilePathStringValue(tmp); } - CFStringRef tmp_cf = (CFStringRef)tmp; - - if (i > 0) { + if (i > 0 && !NIL_P(sep)) { +#if 0 +// TODO: we should probably mimic what 1.9 does here instead of this if (CFStringHasSuffix(res, sep_cf)) { if (CFStringHasPrefix(tmp_cf, sep_cf)) { // Remove trailing slash from res if tmp starts with a @@ -2609,14 +2607,15 @@ else if (!CFStringHasPrefix(tmp_cf, sep_cf)) { CFStringAppend(res, sep_cf); } +#endif + rb_str_concat(res, sep); } - CFStringAppend(res, tmp_cf); + rb_str_concat(res, tmp); } } - CFMakeCollectable(res); - return (VALUE)res; + return res; } /* Modified: MacRuby/branches/icu/objc.m =================================================================== --- MacRuby/branches/icu/objc.m 2010-02-25 23:58:38 UTC (rev 3625) +++ MacRuby/branches/icu/objc.m 2010-02-26 00:29:31 UTC (rev 3626) @@ -197,32 +197,35 @@ NSString *res = (NSString *)FilePathValue(fname); if ([res isAbsolutePath]) { - NSString *tmp = [res stringByResolvingSymlinksInPath]; - // Make sure we don't have an invalid user path. - if ([res hasPrefix:@"~"] && [tmp isEqualTo:res]) { - NSString *user = [[[res pathComponents] objectAtIndex:0] substringFromIndex:1]; - rb_raise(rb_eArgError, "user %s doesn't exist", [user UTF8String]); - } - res = tmp; + NSString *tmp = [res stringByResolvingSymlinksInPath]; + // Make sure we don't have an invalid user path. + if ([res hasPrefix:@"~"] && [tmp isEqualTo:res]) { + NSString *user = [[[res pathComponents] objectAtIndex:0] + substringFromIndex:1]; + rb_raise(rb_eArgError, "user %s doesn't exist", [user UTF8String]); + } + res = tmp; } else { - NSString *dir = dname != Qnil ? - (NSString *)FilePathValue(dname) : [[NSFileManager defaultManager] currentDirectoryPath]; + NSString *dir = dname != Qnil + ? (NSString *)FilePathValue(dname) + : [[NSFileManager defaultManager] currentDirectoryPath]; - if (![dir isAbsolutePath]) { - dir = (NSString *)rb_file_expand_path((VALUE)dir, Qnil); - } + if (![dir isAbsolutePath]) { + dir = (NSString *)rb_file_expand_path((VALUE)dir, Qnil); + } - // stringByStandardizingPath does not expand "/." to "/". - if ([res isEqualTo:@"."] && [dir isEqualTo:@"/"]) { - res = @"/"; - } - else { - res = [[dir stringByAppendingPathComponent:res] stringByStandardizingPath]; - } + // stringByStandardizingPath does not expand "/." to "/". + if ([res isEqualTo:@"."] && [dir isEqualTo:@"/"]) { + res = @"/"; + } + else { + res = [[dir stringByAppendingPathComponent:res] + stringByStandardizingPath]; + } } - return (VALUE)[res mutableCopy]; + return rb_str_new2([res fileSystemRepresentation]); } static VALUE

1 0

[3625] MacRuby/branches/icu/string.c
by source_changes＠macosforge.org 25 Feb '10

25 Feb '10

Revision: 3625 http://trac.macosforge.org/projects/ruby/changeset/3625 Author: lsansonetti(a)apple.com Date: 2010-02-25 15:58:38 -0800 (Thu, 25 Feb 2010) Log Message: ----------- fixed #[] to use rb_range_beg_len() Modified Paths: -------------- MacRuby/branches/icu/string.c Modified: MacRuby/branches/icu/string.c =================================================================== --- MacRuby/branches/icu/string.c 2010-02-25 22:26:18 UTC (rev 3624) +++ MacRuby/branches/icu/string.c 2010-02-25 23:58:38 UTC (rev 3625) @@ -1536,24 +1536,21 @@ default: { - VALUE rb_start = 0, rb_end = 0; - int exclude_end = false; - if (rb_range_values(indx, &rb_start, &rb_end, - &exclude_end)) { - long start = NUM2LONG(rb_start); - long end = NUM2LONG(rb_end); - if (exclude_end) { - --end; - } - return str_substr(str, start, end - start + 1); - } - else { - str = str_substr(str, NUM2LONG(indx), 1); - if (!NIL_P(str) && str_length(RSTR(str), true) == 0) { + long beg = 0, len = 0; + switch (rb_range_beg_len(indx, &beg, &len, str_length(RSTR(str), + false), 0)) { + case Qfalse: + break; + case Qnil: return Qnil; - } - return str; + default: + return str_substr(str, beg, len); } + str = str_substr(str, NUM2LONG(indx), 1); + if (!NIL_P(str) && str_length(RSTR(str), true) == 0) { + return Qnil; + } + return str; } } }

1 0

[3624] MacRuby/branches/icu/string.c
by source_changes＠macosforge.org 25 Feb '10

25 Feb '10

Revision: 3624 http://trac.macosforge.org/projects/ruby/changeset/3624 Author: lsansonetti(a)apple.com Date: 2010-02-25 14:26:18 -0800 (Thu, 25 Feb 2010) Log Message: ----------- make sure the evil RSTRING_PTR macro always return a null-terminated string Modified Paths: -------------- MacRuby/branches/icu/string.c Modified: MacRuby/branches/icu/string.c =================================================================== --- MacRuby/branches/icu/string.c 2010-02-25 21:56:37 UTC (rev 3623) +++ MacRuby/branches/icu/string.c 2010-02-25 22:26:18 UTC (rev 3624) @@ -689,15 +689,33 @@ str_resize_bytes(rb_str_t *self, long new_capacity) { if (self->capacity_in_bytes < new_capacity) { - char *bytes = xrealloc(self->data.bytes, new_capacity); - if (bytes != self->data.bytes) { - GC_WB(&self->data.bytes, bytes); + if (self->data.bytes == NULL) { + GC_WB(&self->data.bytes, xmalloc(new_capacity)); } + else { + char *bytes = xrealloc(self->data.bytes, new_capacity); + if (bytes != self->data.bytes) { + GC_WB(&self->data.bytes, bytes); + } + } self->capacity_in_bytes = new_capacity; } } static void +str_ensure_null_terminator(rb_str_t *self) +{ + assert(!str_is_stored_in_uchars(self)); + + if (self->length_in_bytes > 0 + && (self->capacity_in_bytes == self->length_in_bytes + || self->data.bytes[self->length_in_bytes] != '\0')) { + str_resize_bytes(self, self->length_in_bytes + 1); + self->data.bytes[self->length_in_bytes] = '\0'; + } +} + +static void str_splice(rb_str_t *self, long pos, long len, rb_str_t *str, bool ucs2_mode) { // self[pos..pos+len] = str @@ -3459,11 +3477,15 @@ rb_str_cstr(VALUE str) { if (IS_RSTR(str)) { + if (RSTR(str)->length_in_bytes == 0) { + return ""; + } str_make_data_binary(RSTR(str)); + str_ensure_null_terminator(RSTR(str)); return RSTR(str)->data.bytes; } - // CFString code path, hopefully this should not happen a lot. + // CFString code path, hopefully this should not happen very often. const char *cptr = (const char *)CFStringGetCStringPtr((CFStringRef)str, 0); if (cptr != NULL) { return cptr;

1 0

[3623] MacRuby/branches/icu/include/ruby/re.h
by source_changes＠macosforge.org 25 Feb '10

25 Feb '10

Revision: 3623 http://trac.macosforge.org/projects/ruby/changeset/3623 Author: lsansonetti(a)apple.com Date: 2010-02-25 13:56:37 -0800 (Thu, 25 Feb 2010) Log Message: ----------- add ruby/re.h file for MRI compat Added Paths: ----------- MacRuby/branches/icu/include/ruby/re.h Added: MacRuby/branches/icu/include/ruby/re.h =================================================================== --- MacRuby/branches/icu/include/ruby/re.h (rev 0) +++ MacRuby/branches/icu/include/ruby/re.h 2010-02-25 21:56:37 UTC (rev 3623) @@ -0,0 +1 @@ +// This file is there for backward compatibility with MRI.

1 0

[3622] MacRuby/branches/icu
by source_changes＠macosforge.org 25 Feb '10

25 Feb '10

Revision: 3622 http://trac.macosforge.org/projects/ruby/changeset/3622 Author: lsansonetti(a)apple.com Date: 2010-02-25 13:53:42 -0800 (Thu, 25 Feb 2010) Log Message: ----------- regexps are now AOT compilable Modified Paths: -------------- MacRuby/branches/icu/compiler.cpp MacRuby/branches/icu/compiler.h MacRuby/branches/icu/encoding.h MacRuby/branches/icu/include/ruby/ruby.h MacRuby/branches/icu/re.cpp MacRuby/branches/icu/re.h MacRuby/branches/icu/string.c Modified: MacRuby/branches/icu/compiler.cpp =================================================================== --- MacRuby/branches/icu/compiler.cpp 2010-02-25 21:13:29 UTC (rev 3621) +++ MacRuby/branches/icu/compiler.cpp 2010-02-25 21:53:42 UTC (rev 3622) @@ -18,12 +18,13 @@ #include "ruby/ruby.h" #include "ruby/encoding.h" #include "ruby/node.h" -#include "ruby/re.h" #include "id.h" #include "vm.h" #include "compiler.h" #include "objc.h" #include "version.h" +#include "encoding.h" +#include "re.h" extern "C" const char *ruby_node_name(int node); @@ -498,10 +499,12 @@ GlobalVariable * RoxorCompiler::compile_const_global_ustring(const UniChar *str, - const size_t len, CFHashCode hash) + const size_t len) { assert(len > 0); + const unsigned long hash = rb_str_hash_uchars(str, len); + std::map<CFHashCode, GlobalVariable *>::iterator iter = static_ustrings.find(hash); @@ -2797,22 +2800,18 @@ return CallInst::Create(newString3Func, "", bb); } else { - UniChar *buf = (UniChar *)CFStringGetCharactersPtr( - (CFStringRef)val); - bool free_buf = false; - if (buf == NULL) { - buf = (UniChar *)malloc(sizeof(UniChar) * str_len); - CFStringGetCharacters((CFStringRef)val, - CFRangeMake(0, str_len), buf); - free_buf = true; - } + UChar *chars = NULL; + long chars_len = 0; + bool need_free = false; - GlobalVariable *str_gvar = compile_const_global_ustring(buf, - str_len, CFHash((CFTypeRef)val)); + rb_str_get_uchars(val, &chars, &chars_len, &need_free); - if (free_buf) { - free(buf); - buf = NULL; + GlobalVariable *str_gvar = compile_const_global_ustring(chars, + chars_len); + + if (need_free) { + free(chars); + chars = NULL; } std::vector<Value *> idxs; @@ -5621,8 +5620,10 @@ RubyObjTy, PtrTy, NULL)); Function *newRegexp2Func = - cast<Function>(module->getOrInsertFunction("rb_reg_new_retained", - RubyObjTy, PtrTy, Int32Ty, Int32Ty, NULL)); + cast<Function>(module->getOrInsertFunction( + "rb_unicode_regex_new_retained", + RubyObjTy, PointerType::getUnqual(Int16Ty), Int32Ty, + Int32Ty, NULL)); Function *newBignumFunc = cast<Function>(module->getOrInsertFunction("rb_bignum_new_retained", @@ -5669,28 +5670,33 @@ case T_REGEXP: { - struct RRegexp *re = (struct RRegexp *)val; + const UChar *chars = NULL; + long chars_len = 0; + regexp_get_uchars(val, &chars, &chars_len); + Value *re_str; - if (re->len == 0) { + if (chars_len == 0) { + re_str = ConstantPointerNull::get( + PointerType::getUnqual(Int16Ty)); re_str = compile_const_pointer(NULL, NULL); } else { - GlobalVariable *rename_gvar = - compile_const_global_string(re->str, re->len); + GlobalVariable *re_name_gvar = + compile_const_global_ustring(chars, chars_len); std::vector<Value *> idxs; idxs.push_back(ConstantInt::get(Int32Ty, 0)); idxs.push_back(ConstantInt::get(Int32Ty, 0)); - re_str = GetElementPtrInst::Create(rename_gvar, + re_str = GetElementPtrInst::Create(re_name_gvar, idxs.begin(), idxs.end(), ""); } std::vector<Value *> params; params.push_back(re_str); - params.push_back(ConstantInt::get(Int32Ty, re->len)); + params.push_back(ConstantInt::get(Int32Ty, chars_len)); params.push_back(ConstantInt::get(Int32Ty, - re->ptr->options)); + rb_reg_options(val))); Instruction *call = CallInst::Create(newRegexp2Func, params.begin(), params.end(), ""); Modified: MacRuby/branches/icu/compiler.h =================================================================== --- MacRuby/branches/icu/compiler.h 2010-02-25 21:13:29 UTC (rev 3621) +++ MacRuby/branches/icu/compiler.h 2010-02-25 21:53:42 UTC (rev 3622) @@ -314,7 +314,7 @@ return compile_const_global_string(str, strlen(str)); } GlobalVariable *compile_const_global_ustring(const UniChar *str, - const size_t str_len, CFHashCode hash); + const size_t str_len); Value *compile_arity(rb_vm_arity_t &arity); Instruction *compile_range(Value *beg, Value *end, bool exclude_end, Modified: MacRuby/branches/icu/encoding.h =================================================================== --- MacRuby/branches/icu/encoding.h 2010-02-25 21:13:29 UTC (rev 3621) +++ MacRuby/branches/icu/encoding.h 2010-02-25 21:53:42 UTC (rev 3622) @@ -301,7 +301,7 @@ long rb_str_chars_len(VALUE str); UChar rb_str_get_uchar(VALUE str, long pos); void rb_str_append_uchar(VALUE str, UChar c); -unsigned long rb_str_hash_uchars(UChar *chars, long chars_len); +unsigned long rb_str_hash_uchars(const UChar *chars, long chars_len); VALUE mr_enc_s_is_compatible(VALUE klass, SEL sel, VALUE str1, VALUE str2); Modified: MacRuby/branches/icu/include/ruby/ruby.h =================================================================== --- MacRuby/branches/icu/include/ruby/ruby.h 2010-02-25 21:13:29 UTC (rev 3621) +++ MacRuby/branches/icu/include/ruby/ruby.h 2010-02-25 21:53:42 UTC (rev 3622) @@ -641,13 +641,6 @@ # define RARRAY_PTR(a) (rb_ary_ptr((VALUE)a)) #endif -struct RRegexp { - struct RBasic basic; - struct re_pattern_buffer *ptr; - long len; - char *str; -}; - #if !WITH_OBJC struct RHash { struct RBasic basic; Modified: MacRuby/branches/icu/re.cpp =================================================================== --- MacRuby/branches/icu/re.cpp 2010-02-25 21:13:29 UTC (rev 3621) +++ MacRuby/branches/icu/re.cpp 2010-02-25 21:53:42 UTC (rev 3622) @@ -1336,6 +1336,29 @@ "finalize", (IMP)match_finalize_imp); } +// Compiler primitives. + +void +regexp_get_uchars(VALUE re, const UChar **chars_p, long *chars_len_p) +{ + assert(chars_p != NULL && chars_len_p != NULL); + + UnicodeString *unistr = RREGEXP(re)->unistr; + assert(unistr != NULL); + + *chars_p = unistr->getBuffer(); + *chars_len_p = unistr->length(); +} + +VALUE +rb_unicode_regex_new_retained(UChar *chars, int chars_len, int options) +{ + VALUE str = rb_unicode_str_new(chars, chars_len); + VALUE re = rb_reg_new_str(str, options); + GC_RETAIN(re); + return re; +} + // MRI compatibility. VALUE Modified: MacRuby/branches/icu/re.h =================================================================== --- MacRuby/branches/icu/re.h 2010-02-25 21:13:29 UTC (rev 3621) +++ MacRuby/branches/icu/re.h 2010-02-25 21:53:42 UTC (rev 3622) @@ -23,6 +23,8 @@ VALUE rb_reg_regcomp(VALUE str); int rb_reg_search(VALUE re, VALUE str, int pos, bool reverse); +void regexp_get_uchars(VALUE re, const UChar **chars_p, long *chars_len_p); + typedef struct rb_match_result { unsigned int beg; unsigned int end; Modified: MacRuby/branches/icu/string.c =================================================================== --- MacRuby/branches/icu/string.c 2010-02-25 21:13:29 UTC (rev 3621) +++ MacRuby/branches/icu/string.c 2010-02-25 21:53:42 UTC (rev 3622) @@ -3731,7 +3731,7 @@ // Unicode characters hashing function, copied from CoreFoundation. // This function might have some performance issues on large strings. unsigned long -rb_str_hash_uchars(UChar *chars, long len) +rb_str_hash_uchars(const UChar *chars, long len) { #define HashNextFourUniChars(accessStart, accessEnd, pointer) \ {result = result * 67503105 + (accessStart 0 accessEnd) * 16974593 + (accessStart 1 accessEnd) * 66049 + (accessStart 2 accessEnd) * 257 + (accessStart 3 accessEnd); pointer += 4;}

1 0