Integer overflow when reading large files

[Chris Coulson]

base::File::Info::size is an int64, but base::FileProxy::Read() takes an int for the size. In oxide::GetFileContentsJob::OnGotFileInfo, we call Read() with the size returned from base::FileProxy::GetInfo(), causing a possible integer overflow for files larger than 2GB (the result of which would be that we wouldn't read the entire file).

This code is currently only used for reading user scripts, and it's unlikely we're going to be loading any this size anyway