Refactor dependency resolution out of deployment handler

[sameerzuberi]

Issue #, if available:

Description of changes:

• Remove all dependency resolution logic from deployment_handler. This logic is now in a new dependency_resolver which will be called from within deployment_handler.
• Refactor all deployment configuration logic into deployment_configuration. This makes it easier to get any necessary config values and streamlines deployment handler code.

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[AniruddhaKanhere]

We need to verify that the buffer is at least as long as resp.data before copying over the data. Similarly in other places

[AniruddhaKanhere]

I think that this assert is wrong. Should it not be thing_name->len > resp.len?

we should also fail gracefully in case the lengths do no match.
I think it can be done as such:

if(thing_name->len <= resp.len) {
    assert(false);
    return GGL_ERR_FAILURE;
}

This would allow us to fail gracefully even when the assert is not defined in production environments.

[AniruddhaKanhere]

This assert also seems wrong no?

[rawalexe]

PR #762 covers it