Define a preprocessor flag for enabling warnings on unused ConnectionHandle

[phyBrackets]

Resolving #43

[LeonMatthesKDAB]

Overall a good change.
The way it's currently implemented, it would make CMake required, which I would like to avoid.

That should easy to fix though.

[phyBrackets]

Thanks @LeonMatthesKDAB , Looks good to me now!

[LeonMatthesKDAB]

@phyBrackets cool, can you please incorporate my last review, so we can merge the PR?

[LeonMatthesKDAB]

@phyBrackets Thanks for the update, the code is good now 👍

It just seems we have loads of unused ConnectionHandles in our tests 🤔
I guess we could use std::ignore to fix this.

It does bring up the question of how annoying this would be for end-users though.
@seanharmer If we enable this by default, do you think our existing user code would explode with warnings about unused ConnectionHandles?
Because then I may need to revert my opinion about making this ON by default.

If it's just our testing that discards connection handles that often, I'm fine with leaving it ON though.

[seanharmer]

I'd suggest enabling it by default. It usually indicates something we're doing wrong so good to know and fix them.

[LeonMatthesKDAB]

@seanharmer thanks for the feedback 👍 .

@phyBrackets can you please add the std::ignore calls where required, or use the ConnectionHandle if it makes sense, so CI passes.
Then we can merge the PR 🥳

Side note: To test locally, you should be able to just run cmake --preset=clazy && cmake --build --preset=clazy && ctest --preset=clazy, which enables all warnings.

[phyBrackets]

@phyBrackets Thanks for the update, the code is good now 👍

It just seems we have loads of unused ConnectionHandles in our tests 🤔 I guess we could use std::ignore to fix this.

It does bring up the question of how annoying this would be for end-users though. @seanharmer If we enable this by default, do you think our existing user code would explode with warnings about unused ConnectionHandles? Because then I may need to revert my opinion about making this ON by default.

If it's just our testing that discards connection handles that often, I'm fine with leaving it ON though.

Hi @LeonMatthesKDAB , Thanks for the idea. I guess it would be fine with std::ignore as well but it seems more intended to be used with tuples and it is also not a recommended way by standard to be used for suppressing warning in this case as it could potentially break in future version of the spec. Casting to simply void seems fine but I am not sure about if it looks good for our examples? (fine for tests!)