Re: [asio-users] Cancelling composed operations

[Marsh R. <ma...@ex...>]

On 01/07/2011 11:40 AM, Marat Abrarov wrote:
>
>> As far as I am concerned, the documentation for asio_handler_invoke lacks
>> clarity.
>> [...]
>> It took reading the source code to find out that [...wrapped_handler...]
>> is special in that asio has
>> a private asio_handler_invoke overload for this type class.
>
> Isn't this enough?
> http://www.boost.org/doc/libs/1_45_0/doc/html/boost_asio/overview/core/stran
> ds.html

Speaking as someone who doesn't understand all the details here, maybe I 
have a useful other perspective on the docs.

I've used ASIO without strands and it never occurred me that various 
composed operations would require an extra strand. Probably I didn't 
think about it hard enough. :-) I've also worked on code where another 
developer used ASIO and put in a strand where I don't think one was 
necessary.

Perhaps it would be better if we didn't need to read the "strands" page 
to learn about the properties of builtin operations in general?

Perhaps the info on strands does such a good job describing what they do 
that it seems to overstate the case if you look at the big picture (or 
approach it as a new user)?

- Marsh