Re: [Boost-users] [Asio] Socket Read/Write Thread-Safety (TimothyLiang)

Subject: Re: [Boost-users] [Asio] Socket Read/Write Thread-Safety (TimothyLiang)
From: Timothy Liang (timothy003_at_[hidden])
Date: 2011-01-31 07:04:55

> RTFM. Really, RTFM.
> Use boost::asio::io_service::strand and don't worry about synchronization.
>
> http://www.boost.org/doc/libs/1_45_0/doc/html/boost_asio/overview/core/strands.html
>
> You can use async_read and async_write simultaneously (but note "Thread 
> Safety" for used asio class) with strands easily. Yes, with multiple 
> threads running boost::asio::io_service::run on the same 
> boost::asio::io_service object.
>
> What amazing laziness in reading asio docs :) - all of the latest 
> questions about Boost.Asio have respective answers in documentation.
>
> Marat Abrarov.

Thanks for your reply.  I've read the docs already, and didn't find a good 
explanation.  After a couple of days of searching, I got frustrated and 
asked in the IRC channel.  Still without an answer, I looked through the 
docs again and then took the time to subscribe to this mailing list and type 
up a detailed and concise question.  And then waited patiently for the 
moderators to review my mail for approval.

I think that page could present a better example.  I did come across it, but 
it used the term 'intermediate handler', which I thought was referring to 
the completion-condition handler.  It also didn't say that I had to directly 
pass my wrapped handler to the composed operations as the completion 
handler, without encapsulating it in another function object, which I was 
doing since I thought that the wrapped handler type only served as a 
functor.  I wasn't aware that there was a hooking mechanism for all handlers 
called by the io_service that used ADL to find custom invocation code.  The 
example on the page makes perfect sense now that I know about it, but I 
couldn't for the life of me understand it before.