[#48729] [ANN] ruby 2.0.0-preview1 released — Yusuke Endoh <mame@...>
Japanese later; 日本語はあとで
Hi,
Hello Vit,
2012/11/6 Yusuke Endoh <mame@tsg.ne.jp>
[#48745] [ruby-trunk - Bug #7267][Open] Dir.glob on Mac OS X returns unexpected string encodings for unicode file names — "kennygrant (Kenny Grant)" <kennygrant@...>
[#48773] [ruby-trunk - Bug #7269][Open] Refinement doesn't work if using locate after method — "ko1 (Koichi Sasada)" <redmine@...>
(2012/11/03 10:11), headius (Charles Nutter) wrote:
(2012/11/03 10:36), SASADA Koichi wrote:
[#48774] [ruby-trunk - Feature #4085] Refinements and nested methods — "shugo (Shugo Maeda)" <redmine@...>
[#48819] [ruby-trunk - Feature #4085] Refinements and nested methods — "headius (Charles Nutter)" <headius@...>
[#48820] [ruby-trunk - Bug #7271][Assigned] Refinement doesn't seem lexical — "ko1 (Koichi Sasada)" <redmine@...>
[#48847] [ruby-trunk - Bug #7274][Open] UnboundMethods should be bindable to any object that is_a?(owner of the UnboundMethod) — "rits (First Last)" <redmine@...>
[#48882] [ruby-trunk - Feature #4085] Refinements and nested methods — "headius (Charles Nutter)" <headius@...>
[#48964] [Backport93 - Backport #7285][Assigned] some failures on RubyInstaller CI — "usa (Usaku NAKAMURA)" <usa@...>
[#48988] [ruby-trunk - Feature #7292][Open] Enumerable#to_h — "marcandre (Marc-Andre Lafortune)" <ruby-core@...>
[#48997] [ruby-trunk - Feature #7297][Open] map_to alias for each_with_object — "nathan.f77 (Nathan Broadbent)" <nathan.f77@...>
[#49018] [ruby-trunk - Feature #7299][Open] Ruby should not completely ignore blocks. — "marcandre (Marc-Andre Lafortune)" <ruby-core@...>
[#49078] Re: [ruby-cvs:44714] marcandre:r37544 (ruby_1_9_3): merge revisions r33453, r37542: — "U.Nakamura" <usa@...>
Hello,
[#49119] ID_ALLOCATOR ? — Roger Pack <rogerdpack2@...>
Hello.
Can I see ruby-prof code?
On Fri, Nov 9, 2012 at 11:14 AM, SASADA Koichi <ko1@atdot.net> wrote:
[#49196] [ruby-trunk - Feature #7322][Open] Add a new operator name #>< for bit-wise "exclusive or" — "alexeymuranov (Alexey Muranov)" <redmine@...>
[#49211] [ruby-trunk - Feature #7328][Open] Move ** operator precedence under unary + and - — "boris_stitnicky (Boris Stitnicky)" <boris@...>
[#49256] [ruby-trunk - Feature #7336][Open] Flexiable OPerator Precedence — "trans (Thomas Sawyer)" <transfire@...>
[#49267] [ruby-trunk - Feature #7340][Open] 'each_with' or 'into' alias for 'each_with_object' — "nathan.f77 (Nathan Broadbent)" <nathan.f77@...>
[#49268] [ruby-trunk - Feature #7341][Open] Enumerable#associate — "nathan.f77 (Nathan Broadbent)" <nathan.f77@...>
[#49282] Re: [ruby-cvs:44801] tenderlove:r37631 (trunk): * probes.d: add DTrace probe declarations. — "U.Nakamura" <usa@...>
Hello,
Hello,
2012/11/13 U.Nakamura <usa@garbagecollect.jp>:
[#49298] [ruby-trunk - Feature #7346][Open] object(...) as syntax sugar for object.call(...) — "rosenfeld (Rodrigo Rosenfeld Rosas)" <rr.rosas@...>
[#49320] [ruby-trunk - Feature #4085] Refinements and nested methods — "headius (Charles Nutter)" <headius@...>
[#49328] [ruby-trunk - Bug #7349][Open] Struct#inspect needs more meaningful output — "postmodern (Hal Brodigan)" <postmodern.mod3@...>
[#49340] bugs.ruby-lang.org - 500 error — Luis Lavena <luislavena@...>
Hello,
I've been unable to access it since morning EET (about 6 hours now).
It's almost 3am in Japan now, don't forget.
On Wed, Nov 14, 2012 at 2:46 PM, Zachary Scott <zachary@zacharyscott.net> wrote:
[#49354] review open pull requests on github — Zachary Scott <zachary@...>
Could we get a review on any open pull requests on github before the
2012/11/15 Zachary Scott <zachary@zacharyscott.net>:
Ok, I was hoping one of the maintainers might want to.
I could add my eyes to monitor the github issues/pull requests, if only to
On Thu, Nov 15, 2012 at 2:11 PM, Marc-Andre Lafortune
On Thu, Nov 15, 2012 at 1:01 PM, Luis Lavena <luislavena@gmail.com> wrote:
On Thu, Nov 15, 2012 at 1:06 PM, Zachary Scott <zachary@zacharyscott.net>
[#49370] [ruby-trunk - Bug #7358][Open] Wrong fd redirection on fork — "felipec (Felipe Contreras)" <felipe.contreras@...>
[#49416] make check: missing psych — Ramkumar Ramachandra <artagnon@...>
Hi,
On Fri, Nov 16, 2012 at 9:58 AM, Ramkumar Ramachandra
Luis Lavena wrote:
[#49463] [ruby-trunk - Feature #7375][Open] embedding libyaml in psych for Ruby 2.0 — "tenderlovemaking (Aaron Patterson)" <aaron@...>
On Sun, Nov 18, 2012 at 03:05:50AM +0900, vo.x (Vit Ondruch) wrote:
Dne 17.11.2012 21:19, Aaron Patterson napsal(a):
On 17 November 2012 21:34, V咜 Ondruch <v.ondruch@gmail.com> wrote:
Hello,
[#49468] [ruby-trunk - Feature #7378][Open] Adding Pathname#write — "aef (Alexander E. Fischer)" <aef@...>
[#49479] [ruby-trunk - Bug #7379][Open] Unexpected result of Kernel#gets on Windows 8 — "phasis68 (Heesob Park)" <phasis@...>
[#49518] [ruby-trunk - Bug #7383][Open] Use stricter cache check in load.c — "funny_falcon (Yura Sokolov)" <funny.falcon@...>
[#49536] [ruby-trunk - Feature #7388][Open] Object#embed — "zzak (Zachary Scott)" <zachary@...>
[#49543] [ruby-trunk - Feature #7390][Open] Funny Falcon Threads — "zzak (Zachary Scott)" <zachary@...>
[#49558] [ruby-trunk - Bug #7395][Open] Negative numbers can't be primes by definition — "zzak (Zachary Scott)" <zachary@...>
[#49868] How to stop spam from ruby-core — Heesob Park <phasis@...>
Hi,
[#49949] [ruby-trunk - Feature #7426][Assigned] Update Rdoc — "mame (Yusuke Endoh)" <mame@...>
(2012/11/27 13:33), drbrain (Eric Hodel) wrote:
On Tue, Nov 27, 2012 at 12:57 AM, SASADA Koichi <ko1@atdot.net> wrote:
On Nov 26, 2012, at 10:09 PM, Luis Lavena <luislavena@gmail.com> wrote:
[#50092] [ruby-trunk - Feature #7434][Open] Allow caller_locations and backtrace_locations to receive negative params — "sam.saffron (Sam Saffron)" <sam.saffron@...>
[#50264] [ruby-trunk - Feature #7457][Open] GC.stat to return "allocated object count" and "freed object count" — "ko1 (Koichi Sasada)" <redmine@...>
[#50306] Towards a better process for changing Ruby — Magnus Holm <judofyr@...>
Hey folks,
What I'd like to see is primarily better communication and release
Hello Magnus,
Endoh-san,
[#50312] How to stop spam message from redmine.ruby-lang.org — Heesob Park <phasis@...>
HI,
Hi,
[#50372] [ruby-trunk - Bug #7476][Open] missing "IP_TRANSPARENT" constant for IP sockets. — "elico (Eliezer Croitoru)" <eliezer@...>
2013/2/24 ko1 (Koichi Sasada) <redmine@ruby-lang.org>:
[ruby-core:49859] [ruby-trunk - Feature #4168] WeakRef is unsafe to use in Ruby 1.9
Issue #4168 has been updated by uggsoutlet (uggsoutlet uggsoutlet).
=begin
would be an excellent rule. Such an attitude would emphasize ((<uggs on sale|URL:http://www.gooduggboots.org/>)) sharply the value of life. Every day we should with gentleness, vigor, hold ((<cheap ugg boots|URL:http://www.gooduggboots.org/>)) the heart of thanksgiving to life. But when the time for endless days, months and years passed in ((<uggs outlet|URL:http://www.gooduggboots.org/>)) front of us, we are often not the seed feeling. Of course, there is also " eat, drink, enjoy .458gyu854
=end
----------------------------------------
Feature #4168: WeakRef is unsafe to use in Ruby 1.9
https://bugs.ruby-lang.org/issues/4168#change-33472
Author: bdurand (Brian Durand)
Status: Closed
Priority: Normal
Assignee: nobu (Nobuyoshi Nakada)
Category: lib
Target version: 2.0.0
=begin
I've found a couple issues with the implementation of WeakRef in Ruby 1.9.
1. WeakRef is unsafe to use because the ObjectSpace will recycle object ids for use with newly allocated objects after the old objects have been garbage collected. This problem was not present in 1.8, but with the 1.9 threading improvements there is no guarantee that the garbage collector thread has finished running the finalizers before new objects are allocated. This can result in a WeakRef returning a different object than the one it originally referenced.
2. In Ruby 1.9 a Mutex is used to synchronize access to the shared data structures. However, Mutexes are not reentrant and the finalizers are silently throwing deadlock errors if they run while new WeakRefs are being created.
I've included in my patch a reimplementation of WeakRef called WeakReference that does not extend Delegator. This code is based on the original WeakRef code but has been rewritten so the variable names are a little clearer. It replaces the Mutex with a Monitor and adds an additional check to make sure that the object returned by the reference is the same one it originally referred to. WeakRef is rewritten as a Delegator wrapper around a WeakReference for backward compatibility.
I went this route because in some Ruby implementations (like Ruby 1.8), Delegator is very heavy weight to instantiate and creating a collection of thousands of WeakRefs will be slow and use up far too much memory (in fact, it would be nice to either backport this patch to Ruby 1.8 or the delegate changes from 1.9 to 1.8 if possible). I also don't really see the value of having weak references be delegators since is very unsafe to do anything without wrapping the call in a "rescue RefError" block. The object can be reclaimed at any time so the only safe method to be sure you still have it is to create a strong reference to the object at which point you might as well use that reference instead of the delegated methods on the WeakRef. It also should be simpler for other implementations of Ruby like Jruby or Rubinius to map native weak references to a simpler interface.
Sample code with WeakReference
orig = Object.new
ref = WeakReference.new(orig)
# ...
obj = ref.object
if obj
# Do something
end
I also have a version of the patch which just fixes WeakRef to work without introducing a new class, but I feel this version is the right way to go.
Also included are unit tests for weak references but the test that checks for the broken functionality is not 100% reliable since garbage collection is not deterministic. I've also included a script that shows the problem in more detail with the existing weak reference implementation.
ruby show_bug.rb 100000 # Run 100,000 iterations on the current implementation of WeakRef and report any problems
ruby -I. show_bug.rb 100000 # Run 100,000 iterations on the new implementation of WeakRef and report any problems
=end
--
http://bugs.ruby-lang.org/