Decide whether to drop datetime information

[uliska]

As discussed in #12 (comment) retrieving the last-modification timestamp for a given page doesn't actually have a use case within the context of the plugin. Dropping support for it (also from the resulting git_author dictionary in the page context) would make implementation more straightforward.

However, if we decide to keep the Git classes proposed in #12 and to merge the plugin with mkdocs-revision-date-localized it makes sense to also determine that value from the git blame data.

So: if the two decisions are not taken support for that value should be removed (from Commit._populate() or probably Page._process_git_blame() by then).

[timvink]

We have decided to keep the Git classes, and I updated the #16, as it makes sense to combine both git date and author functionality in the future.

There is one more consideration. For date, currently we use git log -n 1 -p <file path> to get the latest commit on a file. Replacing that with parsing git blame --porcelain for the commit with the most recent date seems like we're doing the same thing, but I'm not sure how if there are differences when shallow clones (git clone --depth 1) are made. (background on shallow clone problem in timvink/mkdocs-git-revision-date-localized-plugin#10))

For now, I see no reason yet to drop datetime information from the codebase.