reported by Tomas Hoger <thoger@redhat.com> and patched by nahi.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48402 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* lib/unicode_normalize/tables.rb: commit not to download and
convert Unicode data files every time.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48386 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
remove useless sort. show_files will sort the result and
another branch, files_in_gem, doesn't sort.
it should be removed for consistency.
* test/rubygems/test_gem_commands_contents_command.rb
(test_execute_default_gem): adjust the sort algorithm with
Gem::Commands::ContentsCommand#show_files, which sort items
as array of [prefix, basename] not strings.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48370 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
notify readability on the read side of the pipe.
write_nonblock() is not usable for pipe on Windows.
(cleanup_shutdown_pipe): Rescue IOError for @shutdown_pipe_w.close.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48354 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
to avoid race condition.
(cleanup_shutdown_pipe): New private method.
(cleanup_listener): Extracted from shutdown method.
Call this method from start method to avoid race condition.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48353 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* lib/securerandom.rb (SecureRandom.gen_random): separate
implementation details and select at the load time.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48334 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* lib/securerandom.rb: set the script encoding to make a string
literal in SecureRandom::Kernel32.last_error_message single byte
encoding so msg[] works in bytes, since FormatMessage() returns
the size in TCHARs, not in characters.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48328 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
before cheking `/dev/urandom` because we know windows doesn't have it.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48324 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* template/unicode_norm_gen.tmpl: expand kompatible_table so that
recursive expansion is not needed at runtime.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48311 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* lib/unicode_normalize/normalize.rb (canonical_ordering_one):
use explicit separator, not to depend on $,.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48308 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Reduce the initialize method from 2088 to 1332 bytes on 32-bit x86
Measurement obtained using `memsize_of_all', as `memsize_of'
is non-recursive:
a = ObjectSpace.memsize_of_all(RubyVM::InstructionSequence)
RubyVM::InstructionSequence.compile(src)
b = ObjectSpace.memsize_of_all(RubyVM::InstructionSequence)
p(b - a)
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48292 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* lib/rake/cpu_counter.rb (Rake::CpuCounter#count): separate
Etc.nprocessors implementation, and if is is available other
utility methods are never used.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48279 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* lib/net/imap.rb (Net::IMAP#fetch): [DOC] Document that
Net::IMAP#fetch will return nil instead of an emtpy array.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48207 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
reported by Willis Vandevanter <will@silentrobots.com> and
patched by nahi.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48161 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
as trivially supported encoding (is always normalized,
and may appear mixed in with UTF-8 or other Unicode
encodings).
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48134 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* lib/rake/cpu_counter.rb (count): prefer Etc.nprocessors, which
is hundreds times faster.
* test/rake/test_rake_cpu_counter.rb: add tests for features, and
remove useless tests.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48127 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
bug that can't parse XML correctly when
Encoding.default_internal is different with XML
encoding. REXML::Source converts XML encoding on read. So IO
should not convert XML encoding.
Based on patch by NAKAMURA Usaku.
[ruby-dev:48686] [Bug #10418]
* test/rexml/test_encoding.rb
(REXMLTests::EncodingTester#test_parse_utf16_with_utf8_default_internal):
Add the for the above case.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48109 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
update easier and more predictable, and reducing compilation
time.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48072 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
relevant portion of Unicode standard for Hangul (de)composition
identifiers and algorithm.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48071 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* lib/unicode_normalize/normalize.rb (hangul_decomp_one): use more
descriptive name. leave [SLVT]BASE and [LVTNS]COUNT as they are
vague names.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48055 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
is not hungarian notation. The variable name sIndex is
directly taken from the relevant part of the Unicode
Standard, where it is written SIndex and stands for
'syllable index'. See pp. 144/145 of
http://www.unicode.org/versions/Unicode7.0.0/ch03.pdf.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48052 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* lib/unicode_normalize/normalize.rb (NF_HASH_{D,C,K}): remove
first element by Hash#shift.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48045 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* lib/unicode_normalize/normalize.rb (UnicodeNormalize): use self
instead of module name and remove module name if unnecessary.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48044 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
