VCS Interface classes

[Matthew Woehlke]

Jakob Petsovits wrote:
> Btw, I had a short look on how MonoDevelop do their VCS stuff.
> As one could expect, they are working with abstract classes instead of 
> interfaces, so they add a lot of default code into their template classes.
> [snip]
> Specific stuff that could be interesting for us:
> 
> - Instead of edit() and unedit(), they've got RequestFileWritePermission() and 
> NotifyFileChanged() which is called on saving the file. 

Naming. Also, we already talked about how to be notified on write. So 
we're the same, here.

> - Overall, they've got a good amount of Can*(localPath) methods:
> CanAdd(), CanUpdate(), CanCommit(), CanRemove(), CanRevert(),
> CanLock() and CanUnlock().

That, I think, is exactly what we are trying to /avoid/ by defining 
multiple sets of optional interfaces. This way, you get this information 
by asking 'do you implement IFoo?'.

> - Move and Delete actions get an additional "bool force" parameter.
> I think that one would come in handy for us as well, probably as an enum 
> instead of a bool.

What would "force" do? Is it needed for some reason?

> - Their DiffInfo class is a structure consisting of string fileName and
> string content, where content is an actual unified diff, if I read that 
> correctly.

Check out the latest version of diff(). :-)

> - Their Revision class contains a DateTime as only "revision name", and also 
> stores author and commit message as well as the Repository it belongs to, and 
> (optionally) a changedFiles array, the latter one consisting of string path, 
> commit action (enum with Add, Delete, Replace, Modify, Other) and
> string actionDescription (??).

Yes, we need a method to get the list of files, if supported. Andreas 
and I were talking about this, it should be in there soon.

-- 
Matthew
Disadvantage: Bad Puns [-5]
You constantly utter puns so egregious as to cause mental distress to 
anyone hearing them. This can, however, be used to distract enemies.