Feature #5979: deprecate Thread#raise

• Status: Rejected
• Assignee: ko1 (Koichi Sasada)
• Target version: -

[rogerdpack (Roger Pack)]

as discussed in ruby-core, we should probably deprecate Thread#raise since it's dangerous.
http://blade.nagaokaut.ac.jp/cgi-bin/scat.rb/ruby/ruby-core/42318

[matz (Yukihiro Matsumoto)]

Hi,

Although I admit Thread#raise is dangerous, it's not more dangerous
than other asynchronous events, like signals. And I don't think MRI
cannot remove signals, especially keyboard interrupts, as a real-world
language. So there's no reason to deprecate Thread#raise, despite the
fact we need to reimplement Timeout etc.

						matz.

In message "Re: [ruby-core:42416] [ruby-trunk - Feature #5979][Open] deprecate Thread#raise"
on Wed, 8 Feb 2012 05:44:00 +0900, Roger Pack [e-mail] writes:

|as discussed in ruby-core, we should probably deprecate Thread#raise since it's dangerous.
|http://blade.nagaokaut.ac.jp/cgi-bin/scat.rb/ruby/ruby-core/42318

[mame (Yusuke Endoh)]

• Status changed from Open to Rejected
• Assignee set to ko1 (Koichi Sasada)

I agree with matz. Just deprecating Thread#raise does not make sense.
I close this ticket as Rejected.

FYI: Some committers discussed a design of controllable asynchronous
exception at developers' meeting in Akihabara. Ko1 wrote the conclusion
memo in [ruby-dev:45341] (sorry, in Japanese). It is almost based on
Akr's suggestion [ruby-talk:294917]. I think Ko1 will write it soon in
English.

--
Yusuke Endoh [e-mail]