[#72745] [Ruby trunk - Misc #11876] [Closed] Scheduled maintenance 2016/01/01 — shibata.hiroshi@...
Issue #11876 has been updated by Hiroshi SHIBATA.
4 messages
2016/01/07
[#72746] Re: [Ruby trunk - Misc #11876] [Closed] Scheduled maintenance 2016/01/01
— Eric Wong <normalperson@...>
2016/01/07
shibata.hiroshi@gmail.com wrote:
[#72824] [Ruby trunk - Bug #11973] IO#advise should raise NotImplementedError on platforms that do not support that call — git@...
Issue #11973 has been updated by Chuck Remes.
3 messages
2016/01/12
[#72954] [Ruby trunk - Feature #12010] [Assigned] Exclude dot and dotdot from Dir#each — naruse@...
Issue #12010 has been reported by Yui NARUSE.
4 messages
2016/01/20
[#72959] Re: [Ruby trunk - Feature #12010] [Assigned] Exclude dot and dotdot from Dir#each
— Eric Wong <normalperson@...>
2016/01/20
naruse@airemix.jp wrote:
[#73313] [Ruby trunk - Bug #12007] [Open] Newly added Unicode data file doesn't get downloaded — shugo@...
Issue #12007 has been updated by Shugo Maeda.
3 messages
2016/01/23
[#73372] [Ruby trunk - Misc #12004] Code of Conduct — benton@...
Issue #12004 has been updated by Benton Barnett.
4 messages
2016/01/24
[#73379] Re: [Ruby trunk - Misc #12004] Code of Conduct
— KOSAKI Motohiro <kosaki.motohiro@...>
2016/01/24
On Sun, Jan 24, 2016 at 5:13 PM, <benton@bentonbarnett.com> wrote:
[#73421] [Ruby trunk - Misc #12004] Code of Conduct — nekocat432@...
Issue #12004 has been updated by Ruby Dino.
6 messages
2016/01/25
[#73425] Re: [Ruby trunk - Misc #12004] Code of Conduct
— Austin Ziegler <halostatue@...>
2016/01/25
I’m sorry, but this, like the code of merit, is merely a derailing tactic.
[#73469] Re: [Ruby trunk - Misc #12004] Code of Conduct
— Martin J. Dürst <duerst@...>
2016/01/26
On 2016/01/26 01:32, Austin Ziegler wrote:
[#73482] Re: [Ruby trunk - Misc #12004] Code of Conduct
— Austin Ziegler <halostatue@...>
2016/01/26
On Tue, Jan 26, 2016 at 12:25 AM, Martin J. Dürst <duerst@it.aoyama.ac.jp>
[#73491] [Ruby trunk - Misc #12004] Code of Conduct — git@...
Issue #12004 has been updated by Chuck Remes.
7 messages
2016/01/26
[#73492] Re: [Ruby trunk - Misc #12004] Code of Conduct
— Andrew Kirilenko <andrew.kirilenko@...>
2016/01/26
They will never provide any numbers because they are not engineers and they
[#73494] Re: [Ruby trunk - Misc #12004] Code of Conduct
— Brandon Weaver <keystonelemur@...>
2016/01/26
Coraline is a panelist on Ruby rogues and a very well respected member of
[#73498] Re: [Ruby trunk - Misc #12004] Code of Conduct
— Andrew Kirilenko <andrew.kirilenko@...>
2016/01/26
OK, sorry for previous comment. Let's try this way.
[#73500] Re: [Ruby trunk - Misc #12004] Code of Conduct
— Austin Ziegler <halostatue@...>
2016/01/26
On Tue, Jan 26, 2016 at 5:15 PM, Andrew Kirilenko <
[#73558] [Ruby trunk - Misc #12004] Code of Conduct — andrew.kirilenko@...
Issue #12004 has been updated by Andrew Kirilenko.
5 messages
2016/01/28
[#73568] Re: [Ruby trunk - Misc #12004] Code of Conduct
— Austin Ziegler <halostatue@...>
2016/01/29
Andrew, please stop digging. Your hole is only getting deeper.
[#73573] Re: [Ruby trunk - Misc #12004] Code of Conduct
— Andrew Kirilenko <andrew.kirilenko@...>
2016/01/29
>Andrew, please stop digging. Your hole is only getting deeper.
[#73586] [Ruby trunk - Misc #12004] Code of Conduct — andrew@...
Issue #12004 has been updated by Andrew Vit.
4 messages
2016/01/29
[#73593] [Ruby trunk - Bug #12034] RegExp does not respect file encoding directive — nobu@...
Issue #12034 has been updated by Nobuyoshi Nakada.
3 messages
2016/01/30
[ruby-core:73025] [Ruby trunk - Bug #11120] Unexpected behavior when mixing Module#prepend with method aliasing
From:
pschambacher@...
Date:
2016-01-21 13:24:56 UTC
List:
ruby-core #73025
Issue #11120 has been updated by Pierre Schambacher.
Actually I really think that solution 1 is the good one. Here is another code sample:
~~~
module A
def run
puts 'RUN STARTS'
super
puts 'RUN ENDS'
end
def run_without_chain
puts 'RUN_WITHOUT_CHAIN STARTS'
super
puts 'RUN_WITHOUT_CHAIN ENDS'
end
end
class B
def run
puts 'B STARTS'
puts 'B ENDS'
end
prepend A
alias_method :run_without_chain, :run
end
B.new.run_without_chain
~~~
Here I would expect this output:
~~~
RUN_WITHOUT_CHAIN STARTS
B STARTS
B ENDS
RUN_WITHOUT_CHAIN ENDS
~~~
But this is the one I get:
~~~
RUN_WITHOUT_CHAIN STARTS
RUN STARTS
B STARTS
B ENDS
RUN ENDS
RUN_WITHOUT_CHAIN ENDS
~~~
To obtain the result I would like to get, I have to put the alias_method **before** the prepend. That doesn't feel really right for me since aliasing a method and prepending a module in the ancestors are pretty independent actions and they should not be order dependent.
----------------------------------------
Bug #11120: Unexpected behavior when mixing Module#prepend with method aliasing
https://bugs.ruby-lang.org/issues/11120#change-56247
* Author: Pablo Herrero
* Status: Open
* Priority: Normal
* Assignee:
* ruby -v: ruby 2.2.2p95 (2015-04-13 revision 50295) [x86_64-linux]
* Backport: 2.0.0: UNKNOWN, 2.1: UNKNOWN, 2.2: UNKNOWN
----------------------------------------
I'm not completely sure myself if this should be considered a bug, but at least it should be up for discussion.
I stumbled upon this behavior when migrating some code using alias chains to Module#prepend.
Consider the following code:
```ruby
# thingy.rb
class Thingy
def thingy
puts "thingy"
end
end
# thingy_with_foo.rb
module ThingyWithFoo
def thingy
puts "thingy with foo"
super
end
end
Thingy.prepend(ThingyWithFoo)
# thingy_with_bar.rb
class Thingy
alias_method :thingy_without_bar, :thingy # Wont't alias create an alias for Thingy#thingy but ThingyWithFoo#thingy instead
def thingy_with_bar
puts "thingy with bar"
thingy_without_bar # Expected to call original Thingy#thingy method but will call prepended method instead
end
alias_method :thingy, :thingy_with_bar
end
# some_file.rb
Thingy.new.thingy # raises: stack level too deep (SystemStackError))
```
In a nutshell when calling `super` from `ThingyWithFoo#foo` it will call `thingy_with_bar` method, and this method will call back to `ThingyWithFoo#foo` by invoking `thingy_without_bar`, thus producing an endless loop.
This situation arises because `alias_method` is producing an alias not for the Thingy#thingy method the but for the upper method from `ThingyWithFoo` instead. May be this behavior could be considered correct, I'm still not sure, but it will probably became a problem for source code migrating from alias chains to use `Modue#prepend`, specially when other active gems could potentially still be using alias chains themselves without the user knowledge.
--
https://bugs.ruby-lang.org/
Unsubscribe: <mailto:ruby-core-request@ruby-lang.org?subject=unsubscribe>
<http://lists.ruby-lang.org/cgi-bin/mailman/options/ruby-core>