Will Leinweber

Thanks for sending this in. I'm traveling this week and won't be able to properly take a look at this before the new year. Could you please also add a...

The functions in postgres are usually ending in `_send` Here is the range one I think: https://github.com/postgres/postgres/blob/b538c90b1bded5464787e2b8e4431302d24eb601/src/backend/utils/adt/rangetypes.c#L246

You can also call them as functions in postgres if that helps playing around with things faster: ``` will=# select range_send('[1,3]'::int4range); range_send -------------------------------------- \x0200000004000000010000000400000004 (1 row) ```

https://paquier.xyz/postgresql-2/postgres-10-saslprep-scram/

Can you share the code for your `add_db` postgres UDF? Or a simplified version that gives the same crystal error?

Oh interesting. This is defiantly an edge case I didn't come across when implementing the pg protocol, that should be addressed.

that D frame is your data row (1 column, length 3, then ascii 97 for 'a' and so on) but coming at an unexpected time with the block there. This...

Adding those methods and wrapping does sound like a good solution. Is there a particular Errno that you see when the connection is dropped? I'll get to this sometime soon,...

I have no idea what this is. Maybe you mean some other will On Jan 25, 2013, at 6:01 PM, Tom Dignan [email protected] wrote: @will https://github.com/will, thanks, that's a great...

Setting `default_schemas` to default to `{}` instead of `nil` will make the error go away, sure, but it will not make this feature work again. The reason `default_schemas` is nil...