From: Tony Smith <tony@...>
Date: 2002-01-21T23:15:25+09:00
Subject: Re: [RCR] #59 Add fsync method to IO class

Matt Armstrong wrote:

> matz@ruby-lang.org (Yukihiro Matsumoto) writes:
> 
>> We have to make up consensus about fdatasync.  Choose one:
>>
>>   (1) fsync and fdatasync; NOP if not exist
>>   (2) fsync and fdatasync; error if not exist
>>   (3) fsync only; NOP if not exist
>>   (4) fsync only; error if not exist
>>   (5) we don't need both
>>
>> I won't vote for 5; maybe I'm going to vote for 4.
> 
> I vote 4 (2 is my second choice).  Why?
> 
>     ...fsync is widely available, fdatasync is not as common.
> 
>     ...performance difference between fsync and fdatasync is probably
>     not important for Ruby scripts.
> 
>     ...applications that use fsync consider it important and want to
>     know when it does not exist on the platform.

I agree.

> Are we agreed that the implementation of File#fsync will flush all
> data to the OS (like File#flush) before calling fsync()?

Yes. I can't see why anyone would want to do a fsync() without a flush(), 
so it might as well happen in one call. Also it conforms with the principle 
of least suprise.

Tony
-- 
Tony Smith
tony at perforce dot com