Recently I stumbled upon a very common interview question (I think it can be found both in AlgoExpert database as explained in this mock interview) and CoderPad examples database. The gist of the problem is that you’re asked to write a calendar app, which given two calendars will return a list of free time slots between the two. Each calendar is a list of meetings.
This is a small feature added with c++11 revision which I think is super fancy. In short, you can define your own unit system when declaring variables and each literal’s value, prior to its usage will be put through the operator appropriate to the types and suffixes used. cppreference page describes everything in great details. This sounds a bit unclear but the examples will summarise everything, I promise.
Today’s post is gonna be a short one. I’d like to recommend a framework which I used personally on a number of occasions.
Golang comes with a great environment and a set of tools. Tests are first class citizens. Why would you need additional framework for your tests then? Purely for convenience.
is is a mini framework, at the moment of writing this post, the API is comprised of four functions:
C++17 has brought some interesting additions to STL algorithms. One of which are execution policies. This is a set of tags that some STL algorithms accept as first argument (there’s an additional overload provided), instructing how container elements will be accessed and processed. C++17 added three distinct execution policies and one more came along with C++20. As of now, the list contains the following:
class sequenced_policy,class parallel_policy,class parallel_unsequenced_policy,class unsequenced_policy.The description of these policies can be a bit intimidating and unfamiliar at first glance. Let’s take the description for parallel_policy:
CMake has become a de facto industry standard as a natural ancestor superseding autotools. But is it actually an improvement? Personally, after spending signifficant amount of time with CMake projects, I’m inclined to conclude that no, not really. Below, I present why CMake sucks and why you shouldn’t use it for any of your projects.
Just as a disclaimer, I’m presenting my personal views here, which are very subjective. You’re entitled to have your own opinions. You have been warned :).
Yep! This headline says it all. For some reason, for quite a while, this was (hopefully it’s no longer) a very popular problem given during technical interviews. Why that might be? Probably because it’s a simple problem to solve, interviewers don’t have to prepare themselves to use it and it has a couple of solutions so, it’s a good conversation starter about algorithms and performance in regards to time & space complexity. On top of that there’s possibility to follow up on the question, ask about detecting cycles in the list, cache friendliness, so on and so forth. But obviously, prior to solving the problem itself, the candidate has to come up with the implementation of the linked list.
Managing complexity in large code base is all about decoupling. Observer is a perfect tool to do that. The Observed object has no idea about its observers, no dependency injection is required at all, as a result a loose coupling is achieved between two classes of objects where the only common part is the interface they agree upon.
It’s a classic design pattern and I’m sure majority (if not all) of engineers know the details behind it, most of us implement it in a slightly different way, depending on the requirements and the code base we are working with though. Here’s my flavour of this design pattern.