From: "Suraj N. Kurapati" <sunaku@...>
Date: 2008-04-08T15:07:17+09:00
Subject: ruby_init_stack() & ruby_set_stack_size() after ruby_init()?

Hello,

What are the consequences of setting the (GC?) stack size or setting
a completely new base stack address after ruby_init() has already
been called?

  ruby_init();
  ruby_init_stack( the_new_addr );
  ruby_set_stack_size( the_new_size );

I ask because ruby_init_stack() is calibrating itself using
information from getrlimit().  This is perfectly fine if Ruby is
running at the top-level of the main process (in main() or a
subroutine called from main()).

However, in my use case, I run ruby_init() inside a pthread (whose
stack size is much less than that of the main process).  Since this
is an uncommon use case:

* Should I be worried about Ruby's default getrlimit() based stack
calibration?

* Should I correct the default calibration by setting Ruby's stack
limit equal to the stack limit of the pthread?

* Would it be safer/easier if I malloc() some memory and make Ruby
use that for its stack + heap?

Thanks for your consideration.