To simplify the config, AppVeyor YML should be managed per-branch

Currently, the appveyor.yml file that’s checked into source control (across all active long-lived branches) has the build configuration for all of the branches. For example in the v2 branch the file has build configs for 1.4.x and 1.5.x as well as the 2.x code. Which part of the yml build config is then chosen by an almost switch/case-like syntax.

To simplify the build configs, the copy of this file in each branch should only contain the build config for the code in that branch. This is OK because when AppVeyor runs a build, it uses the appveyor.yml file from the same commit as it is building. The extra build configs in the version of the file in each branch are actually never used anyway.

Acceptance criteria

In this ticket, we will ‘tidy up’ the YML file in the v2 branch. The master & 1.4 branches can be left as they are for now. When Elsa 2 is finalised and merged into master, it will overwrite and ‘fix itself’ anyway.

_This ticket was created following a brief discussion started by @craigfowler in https://github.com/elsa-workflows/elsa-core/discussions/660_

[sfmskywalker]

As discussed, let’s push a new tag to try it (preview 7). Let’s make sure we use the same tag for the MyGet too since that feed represents the latest bits straight from the Elsa 2 branch.

[craigfowler]

As of build 2.0.0-preview6.1504, I can see that MyGet deployment is working as-intended.