Mike Beaton

I've been continuing to work on this. I have basic function support in Postgres: ```csharp public class dateArgs { public DateTime d { get; set; } } static void Main(string[]...

In order to have non-breaking SP support I think it is necessary to have a method name like ExecuteSqlBlock (or ExecuteWithParams, or something), in addition to the existing Execute, to...

I have coded in my SP project a non-conflicting variant of Query, following the existing Massive pattern (`yield return` `IEnumerable`), but supporting parameter names, types & directions. As expected, the...

Update (new research): Npgsql used to support that pattern. If they still supported that pattern, then reading back cursor results in PostgreSQL would just work, using the `yield return` `IEnumerable`over...

This is definitely (or almost definitely, see last para below...) a breaking change. The obvious new method signature is `public virtual dynamic Single(string where, string columns = "*", params object[]...

`TryInvokeMember` is obscure. It removes IntelliSense from `DynamicModel` for some, but not all, of it's possible calling patterns - which means that I think those patterns are bound to be...

For the reason of adding IntelliSense, and not having to get everything right before it gets tested at runtime, I agree (for what it's worth - maybe not much) that...

I wonder if `TryInvokeMember()` could be refactored to look for an explicit method to call, after doing only pre-processing to support named arguments. That might be a neat solution? (It...

That does sound clever, and useful as far as it goes, but still too limited, I agree. Re my previous comment, I am thinking that if you split the methods...

Just looking at the code again, I reminded myself that `TryInvokeMember` is written so that `FindPartridge`, `GetBalance`, `LastZebraOnEarth`, `SingleWhiteFemale` will work as aliases for `Find`, `Get`, `Last`, `Single`, etc. My...