Add godocs and free memory in more places

[gmlewis]

This PR is my attempt to more fully document the Go PDK with idiomatic function and variable names along with godoc-style comments.

This is also basically a replacement for PR #34 but I've kept it in two PRs (and left #34 open) in case you want to proceed with #34 separately.

[zshipko]

Thanks, I appreciate all the docs!

I don't have strong feelings about the naming of internal functions, does godoc complain? Or is this just easier to read for Go developers? What do you think @nilslice? If everyone is on-board with the renaming changes we can just close #34

As I mentioned in #34 - I have a PR to fix the issue with vars getting freed to early here: extism/js-sdk#71 so we can remove the fix for that from here.

[zshipko]

I just updated CI, so if you rebase it should run the tests

[nilslice]

Thanks!

I'm fine renaming the internal functions to be more idiomatic Go names. Since they don't impact the API, it's a pretty harmless change. The main reason I had them in the PDK code named like this was to make them more recognizable when I used them, but it's really no big deal to me.

I think most Go developers would expect them as @gmlewis has them.

[gmlewis]

Now that I have a better understanding of the memory model thanks to your comments, there are more Memory blocks that should be freed that were created with extismAlloc, like all the extismOutput functions...

I'll work on that later tonight if you agree.

EDIT: I believe I'm done addressing this change.

[zshipko]

Yeah, that sounds good!

And thanks @nilslice, I will close #34

[gmlewis]

One other thing I was very tempted to change here (GetVarInt) was the return signature from int to (int, bool) to be more idiomatic Go (where the bool typically is assigned to a variable called ok and that is used to determine if the key actually exists)... but that would have been a breaking API change, so I left it alone.

If this were changed, then the signature for GetVar should probably also change, but I'm leaving them alone in this PR since I believe it is an orthogonal issue to the memory freeing issue.

Just FYI.

[zshipko]

Yeah that's a good point!