Add support for using aliases

[matt-gretton-dann]

My first DB Schema change - can that be checked please.

[nickrolfe]

It might be nice to beef up the new test with an alias template, e.g. something like

template<typename T>
struct S {
  using X = T*;
};

using Y = S<int>::X;

And perhaps include t.getUnderlyingType() in the query's select statement.

[matt-gretton-dann]

It might be nice to beef up the new test with an alias template, e.g. something like:

Done - commit b524297

[jbj]

On upgraded dbs it will look like every typedef is C-style. What are the applications you have in mind for this feature, and how will they be affected?

[ian-semmle]

I don't think there are any specific plans to use it right now, it just seemed like something the AST ought to distinguish. It seems plausible that someone might want to distinguish them in a query, for example when codifying a style guide. I don't think we need to be concerned about upgraded databases, though.

[jbj]

LGTM, except for the conflicts.

[matt-gretton-dann]

The conflicts are due to @ian-semmle's DB schema upgrades - I'm rebasing to clear these out.

[jbj]

The QLDoc could be formatted better to align with our new style, but I won't let that hold up this PR. It has to be synced with the internal repo, so it's hard to get merged without conflict.