<t>QuoteOriginally posted by: katastrofaQuoteOriginally posted by: CluelessCppDo you also regularly do pair programming on the job or is this just something you would do for the interview?On the job? Rarely. But it's a good way of interviewing a programmer.Really? That programmer gets a completely d...

QuoteOriginally posted by: katastrofaYou can also do pair programming during interviews.Do you also regularly do pair programming on the job or is this just something you would do for the interview?

<t>QuoteOriginally posted by: katastrofaQuoteOriginally posted by: CluelessCppQuoteIf you're any good as a programmer, you can write FizzBuzz or a function to reverse an array in-place on a piece of paper. If you can't, go back to school.And what does that tell you?It tells me that they're not one o...

<t>QuoteOriginally posted by: katastrofaQuoteOriginally posted by: CluelessCppLook at some of the candidate's Open Source code, maybe?1. What if they didn't write any? I don't have time to contribute to Open Source projects, and I can't show the stuff I at work because it belongs to my employer, not...

<t>QuoteOriginally posted by: katastrofaPS. How do you check that a candidate is capable of writing correct, clean code which solves a problem without asking them to write code during the interview? Talking the talk is one thing, walking the walk quite another.Look at some of the candidate's Open So...

<t>QuoteOriginally posted by: CuchulainnQuote LOL, just looked at the first one, and that's already sneaky! f(1)+f(2)*f(3)So, what's your answer?Depends on the question - in terms of the C++ Standard I'd say it's unspecified (as the evaluation of operands is unsequenced, the evaluation of the bodies...

<t>QuoteOriginally posted by: outrunDo you get a RT error when you add a little print statement inside doit? that would force it to be called.Now it might get removed entirely (the doit() code is written inside the class body -> hence it's inlined), which is incorrect: it should give a RT instead IM...

QuoteOriginally posted by: CuchulainnOf course, it won't work in const member functions.Why do you think so - it "works" just the same as in non-const member functions.

<t>QuoteOriginally posted by: CuchulainnQuoteOriginally posted by: CluelessCppBTW, the _array variants are not needed [...]Unless you using raw legacy C arrays?Sorry, should have put that into the context of the std smart pointers - these are specialised for array types to call the correct delete op...

std also has a unique_ptr.BTW, the _array variants are not needed as unique_ptr and shared_ptr can handle arrays as well.

IntelliSense uses a different C++ parser (EDG) than the Visual C++ compiler. And the CTP only updated the command line compiler, but not IntelliSense - so a different set of C++11 features is supported in IntelliSense compared to the command line compiler.

<t>QuoteOriginally posted by: MiloRambaldiI'm not too familiar with C++11, but it does not seems to allow:vector<T> v {t1, t2, t3};This is allowed in g++.They have implemented the core language support in the VC++ November CTP, but haven't updated the standard library yet - so you could use it in yo...

I have now noticed a few times when applying for a job via LinkedIn that the application apparently doesn't even get viewed (the job shows up as "Applied" instead of "App viewed"). I wonder on what basis this could have been filtered?Does anyone have any ideas or views on this?

Do you insist on using explicit specializations for print_type?if not, http://ideone.com/xMubC could be another approach.

QuoteOriginally posted by: CuchulainnThe problem with C++ templates is that you cannot prescribe constraints on the template parameters. So, in some cases you are left at the mercy of the compiler.But you can use SFINAE - see std::enable_if