Code refactoring for the repo pruning pass.

[MaggieYingYi]

This is a spin-off pull request from #109 .

[paulhuggett]

We could really use an explanation of what's being proposed (what it being changed and why). I know that this information exists elsewhere, but it would be kind to future code archaeologists to include it here.

[MaggieYingYi]

We could really use an explanation of what's being proposed (what it being changed and why). I know that this information exists elsewhere, but it would be kind to future code archaeologists to include it here.

This commit is a code refactoring for the repo pruning pass. In the master branch, two kinds of pruning exist in the pass: the repository level and the module level.

1. Repository level pruning. If the global object is compiled before and exists in the repository, it will be pruned. The repository-level pruning removes code and data that are unchanged since a previous compile.
2. Module level pruning. If the global object is not already present in the repository but is in the same compilation unit (module) as another global object has the same digest, it will be pruned as well. The module-level pruning is just an early optimization to reduce the amount of stuff going through the remaining parts of the compiler.

However, they are handled in a single function. This commit is just a code factoring to make two kinds of pruning more obvious and noticeable.

[paulhuggett]

Makes much more sense now. Relatively minor comments.

[MaggieYingYi]

@paulhuggett , Are you happy with the change?

[FlameTop]

LGTM