Daniel, up for a small review in afw?

This adds recursive behavior to the afw::table alias system: instead of just replacing name lookups that begin with registered aliases with the alias target once, we do it repeatedly until we don't find another match. Should be straightforward and well-localized, so I don't think it should be a problem that there's a lot of unfamiliar code nearby.

All changes are on afw branch u/jbosch/DM-1282:

afw:u/jbosch/DM-1282 % git diff --stat master...u/jbosch/DM-1282

 python/lsst/afw/table/Base.i |    8 ++++----

 src/table/AliasMap.cc        |   37 +++++++++++++++++++++++++------------

 src/table/Schema.cc          |   10 +++++-----

 tests/testTableAliases.py    |   17 +++++++++++++++++

 4 files changed, 51 insertions(+), 21 deletions(-)

Jim,
I looked at the code, and it seems sensible, but I was wondering if we care about the efficiency of detecting cycles. Are AliasMaps going to be big? I get the impression that cycles are supposed to be an indication of a (hopefully) rare bug state, so it shouldn't matter that we walk through the cycle up to N hops to report a cycle in an N-sized map. Is this true? If so, then we don't care too much about efficiency, but we might put a comment saying that yes, we could get by with less hops if we're willing to do more bookkeeping, but it's not worth it in this case.

Otherwise, great job, looks pretty clean.

Exactly: it's a rare bug state, and I just didn't want it to result in an endless loop. I'll add the comment.