Misc #10553: Ruby 2.2.0 release engeneering

• Status: Closed
• Assignee: naruse (Yui NARUSE)

[naruse (Yui NARUSE)]

For the management of release blockers.

Current Phase: feature freeze; changing feature is not allowed unless naruse explicitly allowed.
Next release: RC1 (scheduled on 2014-12-1x)

Q: When ruby_2_2 branch is created?
A: On RC1

Q: Can I add/change a feature?
A: Get explicit permission from naruse

[usa (Usaku NAKAMURA)]

Nobu and I want to change a feature.

Ruby 2.1 and earlier, on Windows, parsing commandline was with ANSI encoding (it's locale encoding of ruby).
This means that a script never get the correct byte sequence of a character if locale encoding does not have the representation of the character and even if suitable -E option is specified.
And, since such a character is replaced by '?', in some cases it would point an incorrect file by commandling globbing.

In trunk, since commandline globbing uses Unicode, the part is almost OK.
But a script still cannot get the correct byte sequence of such a character with any -E options.
We hope to fix this problem with keeping compatibility.

Can we go for Ruby 2.2?

[naruse (Yui NARUSE)]

Usaku NAKAMURA wrote:

Nobu and I want to change a feature.

Ruby 2.1 and earlier, on Windows, parsing commandline was with ANSI encoding (it's locale encoding of ruby).
This means that a script never get the correct byte sequence of a character if locale encoding does not have the representation of the character and even if suitable -E option is specified.
And, since such a character is replaced by '?', in some cases it would point an incorrect file by commandling globbing.

In trunk, since commandline globbing uses Unicode, the part is almost OK.
But a script still cannot get the correct byte sequence of such a character with any -E options.
We hope to fix this problem with keeping compatibility.

Can we go for Ruby 2.2?

ok

[seki (Masatoshi Seki)]

コマンド版のerbの起動引数からスクリプトへ引数を渡す機能を追加したいです。
エラー時の行番号の間違い（バグ）もなおる感じです。

Nobu's patch
https://github.com/nobu/ruby/compare/Feature%2310395-erb-variables-from-commandline

発端
https://bugs.ruby-lang.org/issues/10395

[naruse (Yui NARUSE)]

Masatoshi Seki wrote:

コマンド版のerbの起動引数からスクリプトへ引数を渡す機能を追加したいです。
エラー時の行番号の間違い（バグ）もなおる感じです。

Nobu's patch
https://github.com/nobu/ruby/compare/Feature%2310395-erb-variables-from-commandline

発端
https://bugs.ruby-lang.org/issues/10395

わかりました、どうぞ

[ko1 (Koichi Sasada)]

Give me approval for [Feature #10590]

[naruse (Yui NARUSE)]

Koichi Sasada wrote:

Give me approval for [Feature #10590]

ok

[josb (Jos Backus)]

Hi,

On Dec 11, 2014 6:16 PM, [e-mail] wrote:

Issue #10553 has been updated by Yui NARUSE.

Koichi Sasada wrote:

Give me approval for [Feature #10590]
ok

Thank you for adding this erb enhancement!

Jos

[luislavena (Luis Lavena)]

Got some failures on Windows (MinGW-w64, GCC 4.7.2, 32bits):

https://gist.github.com/luislavena/086d39660851ba49d70f

test_float failure seems a regression from the fixes made for Ruby 2.1

There are other failures that weren't present before.

[ko1 (Koichi Sasada)]

How about [Bug #8984] ?

• nobody against it
• matz approved it

Maybe correct document can be added after rc1.

[naruse (Yui NARUSE)]

Koichi Sasada wrote:

How about [Bug #8984] ?

• nobody against it
• matz approved it

Maybe correct document can be added after rc1.

ok

[hsbt (Hiroshi SHIBATA)]

Please judge about #10601

r48748 introduce to break faraday . I hope to revert or fix before 2.2.0 release.

[funny_falcon (Yura Sokolov)]

#10585 - speedup struct.c - first 2 patches ?

[vo.x (Vit Ondruch)]

#9591 should be resolved. And this issue applies to other bundled gems as well.

[naruse (Yui NARUSE)]

Luis Lavena wrote:

Got some failures on Windows (MinGW-w64, GCC 4.7.2, 32bits):

https://gist.github.com/luislavena/086d39660851ba49d70f

test_float failure seems a regression from the fixes made for Ruby 2.1

There are other failures that weren't present before.

Make a issue for each failure.
Anyway I suppress TestSprintf#test_float by r48844.

[vo.x (Vit Ondruch)]

Vit Ondruch wrote:

#9591 should be resolved. And this issue applies to other bundled gems as well.

Thanks for bumping the io-console version. The bigdecimal should be bumped as well IMO.

[naruse (Yui NARUSE)]

• Related to Feature #5458: DL should be removed added

[vo.x (Vit Ondruch)]

Please backport r48888. But this is incomplete IMO. There should be fixed also this location:

$ git diff
diff --git a/test/ruby/test_rubyoptions.rb b/test/ruby/test_rubyoptions.rb
index 2bd1a30..8d3da54 100644
--- a/test/ruby/test_rubyoptions.rb
+++ b/test/ruby/test_rubyoptions.rb
@@ -85,7 +85,7 @@ class TestRubyOptions < Test::Unit::TestCase
 
   def test_verbose
     assert_in_out_err(["-vve", ""]) do |r, e|
-      assert_match(/^ruby #{RUBY_VERSION}(?:[p ]|dev).*? \[#{RUBY_PLATFORM}\]$/, r.join)
+      assert_match(/^ruby #{RUBY_VERSION}(?:[p ]|dev|rc).*? \[#{RUBY_PLATFORM}\]$/, r.join)
       assert_equal RUBY_DESCRIPTION, r.join.chomp
       assert_equal([], e)
     end

[vo.x (Vit Ondruch)]

Vit Ondruch wrote:

Please backport r48888.

Actually, if there was not other RC, there is no need to backport ;)

[usa (Usaku NAKAMURA)]

Vit Ondruch wrote:

Please backport r48888. But this is incomplete IMO. There should be fixed also this location:

Oops, I had overlooked it.

[akr (Akira Tanaka)]

Please backport r48905 and r48906 to fix Dir.mktmpdir(nil). ruby-core:66943 [Bug #10616]

[ko1 (Koichi Sasada)]

Could you backport r48958 for [Bug #10623]?

[hsbt (Hiroshi SHIBATA)]

• Status changed from Assigned to Closed

Ruby 2.2.0 is already released :)