Bug #2246: segfault instead of SystemStackError in 1.8.x

• Status: Rejected
• Assignee: -
• Target version: -
• ruby -v: ruby 1.8.7 (2008-08-11 patchlevel 72) [i486-linux]

[rogerdpack (Roger Pack)]

=begin
class Go
def method_missing meth, *args
[].send meth, *args { |*incoming_args| yield(*incoming_args) }
end
end

Go.new.pop

results in
rdp@rdp-vm:/free_space/rdp/ruby/all_gems_rdocs$ ruby -v attempt_recreate.rb
ruby 1.8.7 (2008-08-11 patchlevel 72) [i486-linux]
Segmentation fault

on 1.8.7 linux. (ok in 1.9, windows 1.8.6).
Thanks.
=end

[matz (Yukihiro Matsumoto)]

=begin
Hi,

In message "Re: [ruby-core:26216] [Bug #2246] segfault instead of SystemStackError in 1.8.x"
on Wed, 21 Oct 2009 23:56:00 +0900, Roger Pack [e-mail] writes:

|Bug #2246: segfault instead of SystemStackError in 1.8.x

It's know issue. Ruby interpreter checks stack depth periodically, so
when stack grows too fast, it causes SEGV before next depth check.
Frequent checks would prevent the problem, but slows down the
interpreter a lot.

						matz.

=end

[rogerdpack (Roger Pack)]

• Status changed from Open to Rejected

=begin
Thanks for the answer.
Interesting that it reports the overflow appropriately in 1.9 but not 1.8. Perhaps 1.9 should be made to check less often, to make it faster?
Another thought would be to possibly create a guard page at the end of each stack, mentioned here
http://timetobleed.com/fixing-threads-in-ruby-18-a-2-10x-performance-boost/
then not have to check at all.
Thoughts?
-r
=end