recast 64-bit size_t to 32-bit int to remove warnings

[sfinktah]

There are many things defined as int accepting size_t arguments.

You could change those things to accept size_t arguments, or you can silence the warnings.

I have provided a solution for the latter, since it's not my place to change your structures.

With all my PRs and the two response callbacks that return void removed, this project will now compile without errors or warnings.

[mrozigor]

Sorry for delayed response. Are those error MSVC specific? I tried to build example with all error/warning flags and nothing happened. Also I don't think that workarounds are good solution.

[sfinktah]

@mrozigor There are at least 6 instances of assigning variables defined as size_t to containers defined as int.

The only way that could possibly not cause warnings is if you were compiling on a 32-bit platform where size_t is the same size as int. Try adding -Wconversion, it's a common misconception that -Wall activates all warnings.

[mrozigor]

Mea Culpa, I've forgotten about this. It looks like there are more places with conversion problems. I'll try to run compilation with error/warning flags and fix those. For now - can you change casts from (...) to static_cast? Thanks.

[sfinktah]

Mea Culpa, I've forgotten about this. It looks like there are more places with conversion problems. I'll try to run compilation with error/warning flags and fix those. For now - can you change casts from (...) to static_cast? Thanks.

Sure... fortunately I use ReSharper so I should be able to automate that process.

[sfinktah]

@mrozigor Done. It found a lot more than the ones I added, but it only took 60 seconds for it to change them all to static_cast.

[sfinktah]

@mrozigor wrong place to ask, but never the less -- in actually trying to use crow (after all this work, I felt I should), I noted that it doesn't return control (i.e. doesn't thread itself). that's totally fine, i can launch it in it's own thread -- i was just wondering if there was a recommended method, i.e. boost, std::thread, std::async or whatnot.

And more importantly -- how to terminate the thread (and/or crow) properly afterwards.

[mrozigor]

Again - sorry for late response. With this PR we are still waiting for Travis CI team to respond. Probably will have to setup project on my private Jenkins server and connect it with repo.

Weren't you able to stop() server?

[The-EDev]

@sfinktah could you please merge master into this PR so CI could run?

[sfinktah]

@The-EDev sorry, I have been quite busy.

Weren't you able to stop() server?

I haven't decided the best to launch it in another thread – part of which is the question of how to keep a handle to call the stop() on. Of course a global would work, but it seems rather yuck.