[#34030] Hash#select and Hash#reject bug in 1.9.2? — Lin Jen-Shin <godfat@...>
(sorry if this is a duplicated post, i forgot to subscribe)
[#34033] The rights of ruby-core people and Myth of ruby-dev — "NARUSE, Yui" <naruse@...>
Some of you may don't know your rights.
On 04/01/11 at 06:55 +0900, NARUSE, Yui wrote:
Hi,
Hi,
On Tue, Jan 04, 2011 at 06:55:47AM +0900, NARUSE, Yui wrote:
On Tue, Jan 4, 2011 at 1:57 PM, Aaron Patterson
(2011/01/05 3:57), Aaron Patterson wrote:
On Wed, Jan 05, 2011 at 01:01:48PM +0900, Urabe Shyouhei wrote:
Hi,
On Wed, Jan 5, 2011 at 6:52 PM, Shota Fukumori (sora_h)
On Thu, Jan 6, 2011 at 4:37 PM, Charles Oliver Nutter
[#34034] Exception#to_s and non-String Messages — Run Paint Run Run <runrun@...>
I don't know when Redmine will be back up, so:
[#34035] Warn when File.expand_path Ignores its Second Argument? — Run Paint Run Run <runrun@...>
File.expand_path(filename, directory) resolves _filename_ relative to
[#34043] proposal: gem_prelude needs to die — Ryan Davis <ryand-ruby@...>
I think it is time for gem_prelude to die.
2011/1/4 Ryan Davis <ryand-ruby@zenspider.com>:
2011/1/5 Ryan Davis <ryand-ruby@zenspider.com>:
On Thu, Jan 6, 2011 at 1:13 AM, Ryan Davis <ryand-ruby@zenspider.com> wrote:
FYI. JRuby 1.6RC1 is going to be shipping with RubyGems 1.4.1 and
2011/1/10 Ryan Davis <ryand-ruby@zenspider.com>:
> I'd like to drop gem_prelude down to 'require "rubygems"'. If you want, we can profile this against normal and pathological gem setups first.
[#34062] Ruby 1.9* syscall() on 64 bit systems — Niklas Hofer <niklas@...>
Hi,
[#34091] Moving to Git? — Lucas Nussbaum <lucas@...>
Hi,
> Hi,
Hello,
Em 05-01-2011 13:15, U.Nakamura escreveu:
(2011/01/06 0:46), Rodrigo Rosenfeld Rosas wrote:
> > Well, I guess I can help listing some advantages. Using git:
On Wed, Jan 5, 2011 at 11:28, Jon <jon.forums@gmail.com> wrote:
Em 05-01-2011 17:16, mathew escreveu:
On Wed, Jan 5, 2011 at 13:23, Rodrigo Rosenfeld Rosas
On Thu, Jan 06, 2011 at 06:50:24AM +0900, mathew wrote:
On Wed, Jan 5, 2011 at 17:02, Czarek <cezary.baginski@gmail.com> wrote:
On 07/01/11 at 01:05 +0900, mathew wrote:
On Thu, Jan 6, 2011 at 10:36, Lucas Nussbaum <lucas@lucas-nussbaum.net> wrote:
On 07/01/11 at 08:07 +0900, mathew wrote:
On Thu, Jan 6, 2011 at 23:50, Lucas Nussbaum <lucas@lucas-nussbaum.net> wrote:
Dear all,
Hi,
On 08/01/11 at 05:13 +0900, Ryan Davis wrote:
Hi,
What kind of Redmine integration you are talking about? We use Redmine
Hi,
Hi,
On 06-01-2011 04:06, Yukihiro Matsumoto wrote:
On Wed, Jan 5, 2011 at 19:57, Yukihiro Matsumoto <matz@ruby-lang.org> wrote:
Hi,
[#34124] [Ruby 1.9-Bug#4235][Open] svn keywords in code prevent correct building of ruby using git mirror — Stephen Bannasch <redmine@...>
Bug #4235: svn keywords in code prevent correct building of ruby using git mirror
[#34150] [Ruby 1.9-Bug#4237][Open] SSL_shutdown should be called until it returns 0 — Aaron Patterson <redmine@...>
Bug #4237: SSL_shutdown should be called until it returns 0
[#34171] [Ruby 1.8-Feature#4239][Open] Let's begin a talk for "1.8.8" -- How's needed for surviving 1.8? — Shota Fukumori <redmine@...>
Feature #4239: Let's begin a talk for "1.8.8" -- How's needed for surviving 1.8?
Issue #4239 has been updated by Zeno Davatz.
Issue #4239 has been updated by Zeno Davatz.
Issue #4239 has been updated by Zeno Davatz.
(2011/01/16 0:11), Zeno Davatz wrote:
On Sat, Jan 15, 2011 at 1:05 PM, NARUSE, Yui <naruse@airemix.jp> wrote:
(2011/01/17 1:49), Rick DeNatale wrote:
On Sun, Jan 16, 2011 at 12:19 PM, NARUSE, Yui <naruse@airemix.jp> wrote:
Issue #4239 has been updated by Zeno Davatz.
Issue #4239 has been updated by Zeno Davatz.
Hi,
[#34172] [Backport92-Backport#4240][Open] segment fault on ruby on rails project — 건하 박 <redmine@...>
Backport #4240: segment fault on ruby on rails project
[#34177] [Ruby 1.9-Bug#4241][Open] IMAPTest#test_imaps_verify_none does not stop — Kazuhiro NISHIYAMA <redmine@...>
Bug #4241: IMAPTest#test_imaps_verify_none does not stop
[#34179] Why is not git mirror mentioned in ruby downloads section? — Michal Suchanek <hramrach@...>
Hello,
On 2011-01-06 , at 07:08, Michal Suchanek wrote:
[#34214] [Ruby 1.9-Feature#4247][Open] New features for Array#sample, Array#choice — Yoji Ojima <redmine@...>
Feature #4247: New features for Array#sample, Array#choice
Hi,
On Fri, Jan 07, 2011 at 08:08:03PM +0900, Yusuke ENDOH wrote:
On 7 January 2011 18:14, Aaron Patterson <aaron@tenderlovemaking.com> wrote:
[#34244] [Ruby 1.9-Bug#4251][Open] Segmentation fault when using ruby-prof and Ruby 1.9.2 — Evan Chaney <redmine@...>
Bug #4251: Segmentation fault when using ruby-prof and Ruby 1.9.2
[#34267] [Ruby 1.9-Feature#4254][Open] Allow method transplanting — Jonas Pfenniger <redmine@...>
Feature #4254: Allow method transplanting
Wouldn't that be unnecessary if matz' trait proposal would be implemented?
2011/1/9 Haase, Konstantin <Konstantin.Haase@student.hpi.uni-potsdam.de>:
Hi,
Hi,
[#34270] Any plans for Rite and Ruby 2.0 ? — "Jonas Pfenniger (zimbatm)" <jonas@...>
Hi Matz,
[#34297] [Ruby 1.9-Bug#4255][Open] When on a case-insensitive filesystem, "loaded features" search in require should ignore case — Charles Nutter <redmine@...>
Bug #4255: When on a case-insensitive filesystem, "loaded features" search in require should ignore case
Issue #4255 has been updated by Nobuyoshi Nakada.
[#34318] ext/bigdecimal/lib/bigdecimal/util.rb — Aaron Patterson <aaron@...>
Hi Murata!
Hi,
Hi Aaron and Matz,
On Tue, Jan 11, 2011 at 01:23:47PM +0900, Kenta Murata wrote:
[#34354] [Ruby 1.9-Feature#4264][Open] General type coercion protocol for Ruby — Charles Nutter <redmine@...>
Feature #4264: General type coercion protocol for Ruby
On Tue, Jan 11, 2011 at 10:27 AM, Jim Weirich <jim.weirich@gmail.com> wrote:
On Jan 11, 2011, at 11:14 AM, Charles Oliver Nutter wrote:
Hi,
On Tue, Jan 11, 2011 at 7:18 PM, Yukihiro Matsumoto <matz@ruby-lang.org> wrote:
Issue #4264 has been updated by Charles Nutter.
On Tue, Jan 11, 2011 at 10:26, Charles Nutter <redmine@ruby-lang.org> wrote:
On 19-01-2011 08:05, mathew wrote:
Issue #4264 has been updated by Rodrigo Rosenfeld Rosas.
Both #as (Parslet, Sequel, MetaWhere, hacketyhack, Facets, Ruport, CouchPotato, ...) and #to (RSpec, Ramaze, ActiveSupport) are use in DSLs.
[#34355] [Ruby 1.9-Feature#4265][Open] Provide a core method Kernel#ruby for invoking a new Ruby instance — Charles Nutter <redmine@...>
Feature #4265: Provide a core method Kernel#ruby for invoking a new Ruby instance
> Feature #4265: Provide a core method Kernel#ruby for invoking a new Ruby instance
On Fri, Jan 14, 2011 at 9:07 AM, Roger Pack <rogerdpack2@gmail.com> wrote:
[#34416] ext/openssl development repository — Hiroshi Nakamura <nakahiro@...>
Hi, committers,
[#34483] Windows sockets and CRT fds — ghazel@...
Hi,
Hello,
On Thu, Jan 13, 2011 at 8:03 PM, U.Nakamura <usa@garbagecollect.jp> wrote:
Hello,
[#34504] [Ruby 1.9-Bug#4282][Open] Range#map inconsistency with blocks like {...} and do...end — Andrei Kulakov <redmine@...>
Bug #4282: Range#map inconsistency with blocks like {...} and do...end
[#34534] [Ruby 1.9-Bug#4283][Open] Timeout.timeout may cause application exit unintetionally — Motohiro KOSAKI <redmine@...>
Bug #4283: Timeout.timeout may cause application exit unintetionally
Issue #4283 has been updated by Charles Nutter.
Hi,
On Tue, Jan 18, 2011 at 10:00 PM, Yusuke ENDOH <mame@tsg.ne.jp> wrote:
Hi,
[#34537] [Ruby 1.9-Bug#4285][Open] Ruby don't have asynchrounous exception safe syntax and It should have. — Motohiro KOSAKI <redmine@...>
Bug #4285: Ruby don't have asynchrounous exception safe syntax and It should have.
Issue #4285 has been updated by Motohiro KOSAKI.
Hi,
Issue #4285 has been updated by Charles Nutter.
[#34550] [Ruby 1.9-Feature#4288][Open] Allow invoking arbitrary method names with foo."something" syntax — Charles Nutter <redmine@...>
Feature #4288: Allow invoking arbitrary method names with foo."something" syntax
On Wed, Jan 19, 2011 at 2:58 PM, Gary Wright <gwtmp01@mac.com> wrote:
On Wed, Jan 19, 2011 at 6:03 PM, Gary Wright <gwtmp01@mac.com> wrote:
> It does, but my concern with it is that (in the same way as
On Tue, Jan 18, 2011 at 12:51 AM, Charles Nutter <redmine@ruby-lang.org> wrote:
[#34553] Feature request - Low level bitwise operations on char arrays (or class String) — Martin Asser Hansen <mail@...>
Hello ruby-core,
Hi,
I think that the special cases of Hamming and Levensthein distances are,
What about String#bytes returning a new built-in ByteArray type
[#34558] [Ruby 1.9-Bug#4291][Open] rb_time_new with negative values (pre-epoch dates) on Windows — Kyle Banker <redmine@...>
Bug #4291: rb_time_new with negative values (pre-epoch dates) on Windows
[#34568] [PATCH] Add thread timing to Struct::Tms — Evan Phoenix <evan@...>
Because 1.9 now uses real threads, being able to find out how much time the current thread is using is quite useful, especially for benchmarking.
[#34577] Importing rubygems 1.5.0 (release candidate) into trunk. — Ryan Davis <ryand-ruby@...>
I'm going to be committing rubygems 1.5.0 into trunk in a bit.
Hi,
On Tue, Jan 18, 2011 at 11:22 PM, NARUSE, Yui <naruse@airemix.jp> wrote:
Hello,
[#34593] Autoload treatment of absolute paths in $LOAD_PATH containing . or .. — David Yip <yipdw@...>
Hello all,
[#34610] [Ruby 1.9-Bug#4294][Open] IO.popen ['"ping"', 'localhost -n 3'] fails — red stun <redmine@...>
Bug #4294: IO.popen ['"ping"', 'localhost -n 3'] fails
[#34632] Ruby operator equivalent to Groovy's "?." — Rodrigo Rosenfeld Rosas <rr.rosas@...>
One of the few things I like in Groovy that Ruby doesn't support is
I have definitely had cases where such a feature would have been very
On Thu, Jan 20, 2011 at 1:27 AM, Rodrigo Rosenfeld Rosas
Em 20-01-2011 05:35, Giuseppe Bilotta escreveu:
On Wed, Jan 19, 2011 at 6:27 PM, Rodrigo Rosenfeld Rosas
On Thu, Jan 20, 2011 at 7:01 PM, Jim Weirich <jim.weirich@gmail.com> wrote:
[#34634] Returning from the callee — Rodrigo Rosenfeld Rosas <rr.rosas@...>
Sometimes it is useful to be able to return from the callee method.
2011/1/20 Rodrigo Rosenfeld Rosas <rr.rosas@gmail.com>:
Em 20-01-2011 07:59, Jonas Pfenniger (zimbatm) escreveu:
On Thu, Jan 20, 2011 at 19:27, Rodrigo Rosenfeld Rosas
Em 20-01-2011 18:51, Nikolai Weibull escreveu:
[#34648] [Ruby 1.9-Bug#4298][Open] Duration of calling String#[] with the same index is strangely related to string length. — Radosław Bułat <redmine@...>
Bug #4298: Duration of calling String#[] with the same index is strangely related to string length.
[#34659] MRI file architecture doc — "Jonas Pfenniger (zimbatm)" <jonas@...>
Hi ruby folks,
[#34679] [Ruby 1.9-Bug#4300][Open] Merge keys not working as expected in psych yaml parser — Matthew Conway <redmine@...>
Bug #4300: Merge keys not working as expected in psych yaml parser
[#34690] [Ruby 1.9-Bug#4301][Open] Off-by-one line number in Psych parse error — Charles Nutter <redmine@...>
Bug #4301: Off-by-one line number in Psych parse error
[#34739] Redmine spam — "Jonas Pfenniger (zimbatm)" <jonas@...>
I really hate spammers, is it something we can do to prevent them
2011/1/21 Jonas Pfenniger (zimbatm) <jonas@pfenniger.name>:
[#34752] Is evaluation order of arguments specified? — Nikolai Weibull <now@...>
Is the evaluation order of arguments specified? I couldn’t find
[#34832] [Ruby 1.9-Bug#4314][Open] spwn cause Segmantation Fault if take a long string. — Motohiro KOSAKI <redmine@...>
Bug #4314: spwn cause Segmantation Fault if take a long string.
[#34861] [Ruby 1.9-Feature#4326][Open] Fiber should respond to call() and [] — Aaron Patterson <redmine@...>
Feature #4326: Fiber should respond to call() and []
Issue #4326 has been updated by Charles Nutter.
On Sat, Jan 29, 2011 at 02:58:46AM +0900, Charles Nutter wrote:
On Fri, Jan 28, 2011 at 5:29 PM, Aaron Patterson
On Sat, Jan 29, 2011 at 09:45:30AM +0900, Charles Oliver Nutter wrote:
Hi,
[#34869] make ruby support line continuations ? — Marc Chantreux <khatar@...>
hello,
You can escape line endings with backslash:
> I love it so much i tried it in ruby. trying to rewrite:
hello,
Dne 27.1.2011 7:15, Marc Chantreux napsal(a):
On Thu, Jan 27, 2011 at 04:16:50PM +0900, V咜 Ondruch wrote:
[#34888] [Ruby 1.9-Feature#4328][Open] export rb_thread_call_with_gvl() — Eric Wong <redmine@...>
Feature #4328: export rb_thread_call_with_gvl()
Hi,
SASADA Koichi <ko1@atdot.net> wrote:
[#34911] The ruby-lang.org downloads page should include RVM for OS X — Andrew Vos <andrew.vos@...>
(I sent this before I subscribed and I'm not sure if it bounced. Sorry if
RVM is not official, and makes problem more difficult. (magically
What do you mean by "official"? Also, what does it make more difficult? Do
return mail is gmail thing. I have same problem.
On Jan 27, 2011, at 5:54 PM, Shota Fukumori (sora_h) wrote:
Hi,
On Jan 27, 2011, at 7:49 PM, Shota Fukumori (sora_h) wrote:
On Fri Jan 28 2011 @ 10:57, James Edward Gray II wrote:
On Jan 27, 2011, at 8:04 PM, Peter Aronoff wrote:
[#34938] Setting libpath for AIX — Perry Smith <pedzsan@...>
Hi,
[#34959] [Ruby 1.9-Bug#4340][Open] Encoding of result string for String#gsub is not consistent — Eric Hodel <redmine@...>
Bug #4340: Encoding of result string for String#gsub is not consistent
Issue #4340 has been updated by Yui NARUSE.
Can I ask why regexps are not affected by
[#34970] [Ruby 1.9-Bug#4343][Open] Dir.glob does match files without extension — Vit Ondruch <redmine@...>
Bug #4343: Dir.glob does match files without extension
Issue #4343 has been updated by Nobuyoshi Nakada.
On 01/29/2011 10:19 AM, Nobuyoshi Nakada wrote:
Dne 29.1.2011 17:27, Jeremy Bopp napsal(a):
On 01/29/2011 10:33 AM, Vテュt Ondruch wrote:
Dne 29.1.2011 17:53, Jeremy Bopp napsal(a):
On 01/29/2011 11:09 AM, V鱈t Ondruch wrote:
On Sat, Jan 29, 2011 at 11:49, Jeremy Bopp <jeremy@bopp.net> wrote:
On 01/30/2011 06:30 PM, mathew wrote:
On Sun, Jan 30, 2011 at 23:07, Jeremy Bopp <jeremy@bopp.net> wrote:
On 02/03/2011 09:38 PM, mathew wrote:
On Thu, Feb 3, 2011 at 22:29, Jeremy Bopp <jeremy@bopp.net> wrote:
[#34991] Understanding realpath_rec behavior — Luis Lavena <luislavena@...>
Hello,
On Sun, Jan 30, 2011 at 03:01, Luis Lavena <luislavena@gmail.com> wrote:
[#35001] file_expand_path performing unneeded IO operations on native Windows — Luis Lavena <luislavena@...>
Hello,
Hello,
On Sun, Jan 30, 2011 at 9:39 PM, U.Nakamura <usa@garbagecollect.jp> wrote:
[#35010] Question: Windows versions - supported and deprecated platforms. — Luis Lavena <luislavena@...>
Hello,
Hello,
[ruby-core:34114] [Ruby 1.9-Bug#4168] WeakRef is unsafe to use in Ruby 1.9
Issue #4168 has been updated by Brian Durand. File weakref.rb added File weakref_patch.diff added I found what may be a bug in my patch. When testing similar code on Rubinius, I found the process deadlocking. I believe what was happening is that a thread was stopped in order to run garbage collection when it was inside a synchronize block on a monitor. On garbage collection the finalizers were run which locks on the same monitor to cleanup weak references which resulted in the deadlock and hung process. Can anyone shed some light on if this might be a problem with the garbage collector in YARV? I've attached updated patch files that removes the synchronization in the finalizers and is should still be thread safe. Also, I believe there are really two issues and that #2 could use some more discussion (and maybe a ticket of its own): 1. BUG - WeakRef is broken on YARV 1.9 2. ENHANCEMENT - The Ruby specification should include a WeakReference and SoftReference implementation Because soft and weak referencers are fairly tightly tied to the VM's memory manager, having them as part of the specification would make it easier for gem authors to count on them being present and having a consistent interface. I'm not as sure about the utility of phantom references or porting reference queues straight from Java. I'm not sure they fit into Ruby the same way they fit into Java. Specifically, is there a difference in Ruby between a weak reference and a phantom reference with the way Ruby finalizers work and is there a better Rubyesque interface for reference queues given the way finalizers can be attached to objects? I don't think WeakRef should be the standard going forward because it encourages buggy code by mixing in the delegator pattern. If you want to safely use a weakly referenced object, you should first establish a strong reference to it. Once you have this strong reference, there is no reason to use a delegate anymore. ---------------------------------------- http://redmine.ruby-lang.org/issues/show/4168 ---------------------------------------- http://redmine.ruby-lang.org
Attachments (2)
require 'monitor'
# WeakReference is a class to represent a reference to an object that is not seen by
# the tracing phase of the garbage collector. This allows the referenced
# object to be garbage collected as if nothing is referring to it.
#
# The difference between this class and WeakRef is that this class does not
# use the delegator pattern and so has an interface more suited for detecting
# if the referenced object has been reclaimed.
#
# Usage:
#
# foo = Object.new
# ref = WeakReference.new(foo)
# ref.object # should be foo
# ObjectSpace.garbage_collect
# ref.object # should be nil
class WeakReference
attr_reader :referenced_object_id
# Map of references to the object_id's they refer to.
@@referenced_object_ids = {}
# Map of object_ids to references to them.
@@object_id_references = {}
@@monitor = Monitor.new
# Finalizer that cleans up weak references when an object is destroyed.
@@object_finalizer = lambda do |object_id|
reference_ids = @@object_id_references.delete(object_id)
if reference_ids
reference_ids.each do |reference_object_id|
@@referenced_object_ids.delete(reference_object_id)
end
end
end
# Finalizer that cleans up weak references when references are destroyed.
@@reference_finalizer = lambda do |object_id|
referenced_id = @@referenced_object_ids.delete(object_id)
if referenced_id
obj = ObjectSpace._id2ref(referenced_object_id) rescue nil
if obj
backreferences = obj.instance_variable_get(:@__weak_backreferences__) if obj.instance_variable_defined?(:@__weak_backreferences__)
if backreferences
backreferences.delete(object_id)
obj.send(:remove_instance_variable, :@__weak_backreferences__) if backreferences.empty?
end
end
references = @@object_id_references[referenced_id]
if references
references.delete(object_id)
@@object_id_references.delete(referenced_id) if references.empty?
end
end
end
# Create a new weak reference to an object. The existence of the weak reference
# will not prevent the garbage collector from reclaiming the referenced object.
def initialize(obj)
@referenced_object_id = obj.__id__
ObjectSpace.define_finalizer(obj, @@object_finalizer)
ObjectSpace.define_finalizer(self, @@reference_finalizer)
@@monitor.synchronize do
@@referenced_object_ids[self.__id__] = obj.__id__
add_backreference(obj)
references = @@object_id_references[obj.__id__]
unless references
references = []
@@object_id_references[obj.__id__] = references
end
references.push(self.__id__)
end
end
# Get the reference object. If the object has already been garbage collected,
# then this method will return nil.
def object
obj = nil
begin
if referenced_object_id == @@referenced_object_ids[self.object_id]
obj = ObjectSpace._id2ref(referenced_object_id)
obj = nil unless verify_backreferences(obj)
end
rescue RangeError
# Object has been garbage collected.
end
obj
end
private
def add_backreference(obj)
backreferences = obj.instance_variable_get(:@__weak_backreferences__) if obj.instance_variable_defined?(:@__weak_backreferences__)
unless backreferences
backreferences = []
obj.instance_variable_set(:@__weak_backreferences__, backreferences)
end
backreferences << object_id
end
def verify_backreferences(obj)
backreferences = obj.instance_variable_get(:@__weak_backreferences__) if obj.instance_variable_defined?(:@__weak_backreferences__)
backreferences && backreferences.include?(object_id)
end
end
require "delegate"
# WeakRef is a class to represent a reference to an object that is not seen by the tracing
# phase of the garbage collector. This allows the referenced object to be garbage collected
# as if nothing is referring to it. Because WeakRef delegates method calls to the referenced
# object, it may be used in place of that object, i.e. it is of the same duck type.
#
# If you don't need to use the delegator pattern, you can use WeakReference instead.
#
# Usage:
# foo = Object.new
# foo = Object.new
# p foo.to_s # original's class
# foo = WeakRef.new(foo)
# p foo.to_s # should be same class
# ObjectSpace.garbage_collect
# p foo.to_s # should raise exception (recycled)
class WeakRef < Delegator
class RefError < StandardError
end
def initialize(obj)
super
end
def __getobj__
obj = @reference.object
Kernel::raise(RefError, "Invalid Reference - probably recycled", Kernel::caller(1)) unless obj
obj
end
def __setobj__(obj)
@reference = WeakReference.new(obj)
end
def weakref_alive?
!!@reference.object
end
end
if __FILE__ == $0
foo = Object.new
p foo.to_s # original's class
foo = WeakRef.new(foo)
p foo.to_s # should be same class
ObjectSpace.garbage_collect
ObjectSpace.garbage_collect
p foo.to_s # should raise exception (recycled)
end
13c13
< require 'thread'
---
> require 'monitor'
22,30c22,42
< @@mutex = Mutex.new
< @@final = lambda {|id|
< @@mutex.synchronize {
< rids = @@id_map[id]
< if rids
< for rid in rids
< @@id_rev_map.delete(rid)
< end
< @@id_map.delete(id)
---
> @@monitor = Monitor.new
>
> @@object_finalizer = lambda do |id|
> rids = @@id_map.delete(id)
> if rids
> rids.each do |rid|
> @@id_rev_map.delete(rid)
> end
> end
> end
>
> @@reference_finalizer = lambda do |id|
> rid = @@id_rev_map.delete(id)
> if rid
> obj = ObjectSpace._id2ref(referenced_object_id) rescue nil
> if obj
> backreferences = obj.instance_variable_get(:@__weak_backreferences__) if obj.instance_variable_defined?(:@__weak_backreferences__)
> if backreferences
> backreferences.delete(object_id)
> obj.send(:remove_instance_variable, :@__weak_backreferences__) if backreferences.empty?
> end
32,36c44,47
< rid = @@id_rev_map[id]
< if rid
< @@id_rev_map.delete(id)
< @@id_map[rid].delete(id)
< @@id_map.delete(rid) if @@id_map[rid].empty?
---
> refs = @@id_map[rid]
> if refs
> refs.delete(id)
> @@id_map.delete(rid) if refs.empty?
38,39c49,50
< }
< }
---
> end
> end
42,45c53,56
< @__id = orig.object_id
< ObjectSpace.define_finalizer orig, @@final
< ObjectSpace.define_finalizer self, @@final
< @@mutex.synchronize {
---
> @__id = orig.__id__
> ObjectSpace.define_finalizer orig, @@object_finalizer
> ObjectSpace.define_finalizer self, @@reference_finalizer
> @@monitor.synchronize do
47,49c58,61
< }
< @@id_map[@__id].push self.object_id
< @@id_rev_map[self.object_id] = @__id
---
> @@id_map[@__id].push self.__id__
> @@id_rev_map[__id__] = @__id
> __add_weakref_backreference__(orig)
> end
54,56c66
< unless @@id_rev_map[self.object_id] == @__id
< Kernel::raise RefError, "Invalid Reference - probably recycled", Kernel::caller(2)
< end
---
> obj = nil
58c68,71
< ObjectSpace._id2ref(@__id)
---
> if @@id_rev_map[self.__id__] == @__id
> obj = ObjectSpace._id2ref(@__id)
> obj = nil unless __verify_weakref_backreference__(obj)
> end
60c73
< Kernel::raise RefError, "Invalid Reference - probably recycled", Kernel::caller(2)
---
> # Object has been garbage collected.
61a75,76
> Kernel::raise(RefError, "Invalid Reference - probably recycled", Kernel::caller(1)) unless obj
> obj
62a78
>
67c83,90
< @@id_rev_map[self.object_id] == @__id
---
> if @@id_rev_map[self.__id__] == @__id
> obj = ObjectSpace._id2ref(@__id)
> __verify_weakref_backreference__(obj)
> else
> false
> end
> rescue RangeError
> false
68a92,113
>
> private
>
> # Add a backreference to the weakref object id in the referenced object to protect
> # against the object id being recycled before the finalizers have a chance to run.
> def __add_weakref_backreference__(obj)
> backreferences = obj.instance_variable_get(:@__weak_backreferences__) if obj.instance_variable_defined?(:@__weak_backreferences__)
> unless backreferences
> backreferences = []
> obj.instance_variable_set(:@__weak_backreferences__, backreferences)
> end
> backreferences << object_id
> end
>
> # Verify that the object is the one the weakref thinks it is. On runtimes where
> # object ids are recycled after garbage collection and reallocated to new objects,
> # there is a chance that the reference could find an object at the specified
> # location in the ObjectSpace that is not the one it originally referenced.
> def __verify_weakref_backreference__(obj)
> backreferences = obj.instance_variable_get(:@__weak_backreferences__) if obj.instance_variable_defined?(:@__weak_backreferences__)
> backreferences && backreferences.include?(self.__id__)
> end
72c117
< # require 'thread'
---
> # require 'thread'