Re: [boost] [thread] semaphore

Subject: Re: [boost] [thread] semaphore
From: Gottlob Frege (gottlobfrege_at_[hidden])
Date: 2013-09-18 12:28:46

On Wed, Sep 18, 2013 at 6:54 AM, Jonathan Wakely
<jwakely.boost_at_[hidden]>wrote:

> On 18 September 2013 11:33, Tim Blechmann wrote:
> >
> > also, checking the condition requires you to acquire a lock, which in
>
> Right, to use a condition_variable you need the CV, a lock and a
> condition. If you don't have those three things you're using it wrong,
>
>
You need these 3 things, and you need to use them in the correct pattern
(while-loop, not an if, etc).

Typically when you have 3 interrelated things that need to be used a
certain way, say maintaining certain invariants, you often wrap them
together into a more easily used class.

The real question is what thwarts wrapping a CV into something simpler.

I think part of the answer is that you need the lock anyhow (ie to
push/pull the queue element in the typical example) and you probably have
the condition anyhow (queue.empty()) so all you are missing is the CV part.

But it still doesn't make it easy to use.  It might have been easier to
somehow wrap the lock and CV together (and a lambda for the condition?) and
tell people to make sure they also lock the same CV+lock thing when
pushing/pulling from the queue.

Of course, nothing stopping anyone from building that on top of CVs.

Tony

Also, dealing with spurious wake-ups is "free" if you have multiple
producers/consumers - ie you wake up - was it spurious, or did another
consumer pop the element before you? Doesn't matter, you need to handle
them the same way. But when you are the only thread on "your side" of the
queue, then spurious wake ups are just more esoterica that *the average
programmer has never heard of*, and worse yet, can typically ignore for
years without seeing the bug come to light (unfortunately).