Check Generator: Fill & Download for Free

Yes, but only if you add the generated files to git.It is generally good practice to avoid checking generated files into a version control system, precisely for this reason - the creation of inevitable merge conflicts.So, the issue isn’t the use of TypeScript (or git) per se - the issue is deciding to check generated artefacts into a version control system.

On the Google Webmaster Team we initially had many difficulties around code quality (here: HTML/CSS) but managed to turn things around through a variety of means, including* enforcing coding guidelines (we were also involved in Google’s HTML and CSS guidelines [1]),* requiring peer reviews and code approvals,* establishing quality control through accessibility, performance, validation, and link checks, and even self-written maintainability, semantics, and typography checks [2] (Google cleared some of this for release, making it possible for e.g. UITest.com [3] to share some of what we did),* performing regular code scoring (grading web pages on a 0–10 scale, based on said checks),* requiring linting,* automating code compression,* running regular QA events to clean up and improve code,* offering discussion and expert lists,* running office hours,* &c.Some of these means we developed on our own, others we borrowed from other Google teams.A key to the improvements was measuring code quality, which is representative of the Google spirit. By looking at all the different factors that contribute to good code quality, and writing tools for those factors that were not readily measurable, we knew where we stood and could improve in the areas we weren’t doing well in.[1] Page on googlecode.com[2] Validation: measuring and tracking code quality[3] Web Development Tools (Validators, Checks, Generators, &c.)

Test against the sample IO provided in the problem statement. This is often enough for trivial problems. It’s very cheap time wise. For some very trivial problems I used to skip even this step.A good way to test is write a brute force solution too and check the return value for both function is same. Over the months, I had my own collection of brute force implementation of popular algorithms / patterns. This is however not very feasible in short format contest. It’s more feasible when you are practicing or doing a research kind of solution mixing different kind of data structure.Let another teammate decide a corner case and expected output. This works well for implementation-critical problems where you are likely to “miss” the corner case the requires a special conditional check.Generating random input might be a good way to test for efficiency of your implementation. Correctness of a solution cannot be tested this way. This also helps encounter dangling pointer and memory related exceptions.Sometimes corner cases are evident from the problem constraint itself. For example, whenever an input value can be 64 bit number, try to test it with at least one case of that input value higher than what is possible to store in 32 bit.