Code review comment for lp:~sil2100/ubuntu-archive-tools/kernel-sru-review-cache

The reason why I was passing only the 'main directory' in the context field was that I knew that I could always get the work directory through getcwd() inside any function, not having to pass it through the context. Since for tarball cache handling I need to know both directories - the work directory (i.e. the temporary dir that we create) and the start directory. The work one for known reasons - we need to know where we're copying the tarballs to. The start directory, on the other hand, is also needed as per Adam's request we are supposed to look for tarballs also in the directory where the script was executed in.

Passing only the start directory allowed me to not skip passing one directory. But yeah, there's no harm in passing it I guess, so I did that now.

After the latest changes, all concerns should have been addressed. Currently tarball caching is not enabled by default but we do try to fetch tarballs from cache directories (or the cwd) if available always. This way we should also satisfy Adam's needs.