From: andrew mcelroy <sophrinix@...>
Date: 2012-01-31T22:44:20+09:00
Subject: Re: RUBY_CRITICAL

--f46d04428e04ccb40204b7d328ea
Content-Type: text/plain; charset=ISO-8859-1

On Tue, Jan 31, 2012 at 3:13 AM, Garthy D <
garthy_lmkltybr@entropicsoftware.com> wrote:

>
> Hi Andrew and Tridib,
>
> Okay, some more info:
>
> Ruby 1.8.7 wraps the statements in rb_w32_enter_critical() and
> wb_w32_leave_critical(). With 1.9.2 doing nothing special with the
> statements, my money is on the 1.9.2 series either no longer needing, or no
> longer having, these particular calls.
>
> I haven't dug into which releases in-between do and don't have these extra
> calls, but this might be an angle to explore if investigating the
> RUBY_CRITICAL macro.
>

I bet this is windows specific legacy code then.
EnterCriticalSection function was a windows 2000 call to handle low memory
situations.
It would make sense that you wouldn't want ruby to crash due to a low
memory flux while
virtual memory was being allocated.

http://msdn.microsoft.com/en-us/library/windows/desktop/ms682608(v=vs.85).aspx


It very well might just be a relic. If windows memory management bugs are
still of concern,
then maybe this would be an excellent candidate for a ruby gem.

Andrew McElroy

>
> Garth
>
>
> On 31/01/12 19:38, Garthy D wrote:
>
>>
>> Hi Andrew and Tridib,
>>
>> On 31/01/12 13:21, andrew mcelroy wrote:
>>
>>> On Mon, Jan 30, 2012 at 8:09 PM, Tridib
>>> Bandopadhyay<tridib04@gmail.**com <tridib04@gmail.com>>wrote:
>>>
>>>  Hello all
>>>>
>>>> Can anyone tell me about RUBY_CRITICAL(). I want to know what this
>>>> function is about and what it does and whats the reason of ruby calling
>>>> this function?
>>>>
>>>
>>> it's a ruby c function. There doesn't seem to be much in the way of
>>> documentation. Why not pull the ruby source code and read what the
>>> function
>>> does?
>>>
>>
>> Out of interest, I just did that, and in 1.9.2 at least, it is defined
>> twice, with no comments, with the form:
>>
>> #define RUBY_CRITICAL(expr) do { expr; } while(0)
>>
>> And all references are in win32.c.
>>
>> Eep. :}
>>
>> I'd hazard a guess that it does absolutely nothing at all (apart from
>> running the supplied expression/command) in the 1.9.2 release. Perhaps
>> it is a relic that has since been disabled, or it is to cover code that
>> has not yet been written? I'm now somewhat curious about this as well.
>> Since one reference is in a backward compat header, I'd put my money is
>> on the former.
>>
>> I wonder if somebody knows?
>>
>> Garth
>>
>>
>
>

--f46d04428e04ccb40204b7d328ea--