as it diffs them.

This makes diff-tree usable again in the new world order.

It now requires the "--add" flag before you add any new files, and
a "--remove" file if you want to mark files for removal. And giving
it the "--refresh" flag makes it just update all the files that it
already knows about.

It's got some debugging printouts etc still in it, but testing on the
kernel seems to show that it does indeed fix the issue with huge tree
files for each commit.

The "diff-tree" program needs it.

This is needed for the change to make "read-tree" just read into the
cache (and then you do a "checkout-cache" to update your current dir
contents).

Like the cache filename finder, it's a generically useful function,
rather than something specific to the current "show-diff" thing.

It finds the cache entry position for a given name, and is
generally useful. Sure, everybody can just scan the active
cache array, but since it's sorted, you actually want to
search it with a binary search, so let's not duplicate that
logic all over the place.

Patches from Dave Jones and Ingo Molnar, but since I don't have any
infrastructure in place to use the old patch applicator scripts I
am trying to build up, I ended up fixing the thing by hand instead.

Credit where credit is due, though. Nice to see that people are
taking a look at the project even in this early stage.

And fix up header declarations.

This allows us to also actually check the sha1 hash using these
routines. Needed for the "fsck" thing.

New syntax: "cat-file -t <sha1>" shows the tag, while "cat-file <tag> <sha1>"
outputs the file contents after checking that the supplied tag matches.