Convert into a Nimble package

[markspanbroek]

• added nimble file
• updated readme
• added license
• converted tests to unit tests
• converted under_scores to camelCase
• string to array conversion at compile time instead of runtime
• continuous integration

[bkomuves]

I have bad feelings about this .gitignore file, which, if i understand correctly, is a blanket ignore for anything without a file extension. First, such files are often legitimate (eg. LICENSE files are usually without an extension). Also, doesn't ".gitignore" itself matches this pattern?

But also in general, I would prefer explicit ignores instead of blanket ones.

[dryajov]

.gitignore is quite flexible and allows any pattern combination, it's standard practice to have one in the repos to avoid accidental checkins and cluttering it with unwanted files.

[bkomuves]

my issue is with the actual patterns, not the existence of the .gitignore file

[markspanbroek]

I understand, I also did a double take when I first came across this pattern. It's certainly not ideal, but it's better than not having it. Nim tends to produce a lot of executable files without an extension when running tests, or everytime you run nim c -r ... something else. I'd continuously run into commits that accidentally include these executables on repo's where we didn't have a .gitignore like this.

The ".gitignore" file itself is not ignored by the way, because it has a dot in it.

[bkomuves]

Isn't it possible to convince Nim to produce the artifacts into a subdirectory, and then gitignore that subdirectory?

[markspanbroek]

It's possible to specify command line parameters to that effect, but given the myriad ways in which team members run nim (through nimble, direct invocation of nim, via vscode, etc...) you're still likely to end up with some executable files outside of the directory where you want them

[dryajov]

LGTM

[bkomuves]

Looks good to me too in general.

Maybe we shouldn't export all the helper functions, but this can be changed later when the API stabilizes.