Martin Ek

This would make it impossible to have nested aliases in one expression, but I can't think of any usecases where that would make a difference (e.g. `((foo AS a) /...

@ms705: saw that you added aliases to `Literal` in https://github.com/ms705/nom-sql/commit/2b3f18809dad08687452e1b584a3e8a8b7d45bd7. If you think that is a better approach in general I'll close this :) I guess it'd be useful to...

No worries, but let me know and I'll rebase. It does also change existing behavior slightly (see https://github.com/ms705/nom-sql/pull/15#discussion_r171262574) so probably requires some thought nonetheless.

Yep - that definitely sounds reasonable!

`DataType::None` sounds reasonable too - changed to that. MySQL allows both, but the 0 behavior can be disabled by enabling [NO_AUTO_VALUE_ON_ZERO](https://dev.mysql.com/doc/refman/5.7/en/sql-mode.html#sqlmode_no_auto_value_on_zero). I also pushed a commit to shard `DataType::None` keys...

Got a fair share of thesis writing to do, but I'll take a look at in a few days! Still need to implement support for sending back the newly assigned...

Thanks @jonhoo! I'll let you know if there's anything in particular 👍

Sorry about the delay! I've been traveling for the last few weeks. I've merged in master now. This changes the current `None` sharding behavior from always going to 0 to...

Better late than never! Update: * The first auto increment ID generated is now returned as a part of the `send_back` acknowledgments. They're exposed to the clients through a new...

@ms705: The `previous_shard` is there to be able to do round-robin of the shard given to an auto incremented ID when an `AutoIncrementRequest` is given, [here](https://github.com/mit-pdos/distributary/pull/66/files#diff-44ad6ceaef9c8e35e5780492693e2047R41). The alternative (at least...