TEST: Refactor combined id rework 2

[GlancingMind]

Possible improvements:

• Entity::CombineIds should take Entity::Id's as arguments, not strings
• Entity::SeparateIds should return an Entity::Id not a string
• A comment should have a CommentId, which should provide operations to extract the BugId or OperationId. So that using CombeIds/SeparateIds is an implmentation detail.
• RepoCacheBug::ResolveComment should return an Entity::Id not a combined Id, as the user of this function already has a reference to the bug, rendering the CombinedId useless. Unless the CombinedId is the real CommentId, but then this Id should provide functions for easier handling, as mentioned in previous point.

[GlancingMind]

Fixed: Still something todo: https://github.com/GlancingMind/git-bug/pull/57/checks?check_run_id=2712344156#step:5:207
But the WebUI still doesn't display the HistoryMenu and the TermUI terminates with an error when editing a comment.

TermUI error: Error: target hash is invalid: invalid length

[MichaelMure]

Just a quick note (sorry I really can't find the time to work on this at the moment :'-( ):

Entity::SeparateIds should return an Entity::Id not a string

You should not do that. Entity::Id should always be a complete Id. SeparateIds() is only able to output two prefixes and therefore can't be considered as Id. We could consider a special type for the prefixes but for now they are string everywhere.

If you have a prefix but need an Id, you should use one of the ResolveXXX functions. They will deal with the possible collisions as well.

[GlancingMind]

Thanks for the tip, I'll try them out and don't worry about the lack of time. :)

[GlancingMind]

Closing this in Favor of #61
Because recreating the important changes was easier then undoing all irrelevant changes here.