This fixes things for `netcoreapp3.1` and fixes the outstanding #1205.

Behavior is expected, but doc isn't clear - clarifying!

* Set initial capacity for collections
* Revert Massive changes
Co-authored-by: Nick Craver <nrcraver@gmail.com>

Simplifying and improve build speed.

- Decreases builds from 8-13 minutes down to ~2 minutes (while running more tests)
- Moves to SQL Server 2019 on AppVeyor
- Moves the 2 longest running tests with params to `[FactLongRunning]`
- Bumps tests up to `netcoreapp3.1`
- Also fixes .Contrib tests
- Moves to `Build.csproj`
- Simplifies `build.ps1`
- Removed defunct `build.sh`
- Nukes `IsAppveyor` (moves to environmental variable overloads)
- Builds now work in Linux/macOS

I add update `Set` method and unit test.

e.g: 
```C#
var sb = new SqlBuilder()
       .Set("vip = @vip", new { vip })
       .Set("updatetime = @updatetime", new { updatetime })
       .Where("id = @id", new { id })
;
var template = sb.AddTemplate("update #Users /**set**/ /**where**/");
connection.Execute(template.RawSql, template.Parameters);
```

RawSql will generate :　
```
update #Users Set vip = @vip , updatetime = @updatetime\n WHERE id = @id\n
```

using Dapper Execute in sqlserver it'll reqeust by below sql command
```
exec sp_executesql N'update #Users Set vip = @vip , updatetime = @updatetime
 WHERE id = @id
',N'@vip bit,@updatetime datetime,@id int',@vip=1,@updatetime='2020-01-01 00:00:00',@id=1
```
Co-authored-by: Nick Craver <nrcraver@gmail.com>
Co-authored-by: Nick Craver <craver@stackoverflow.com>

`SingleOrDefault<(int, int)?>(query)` always returned null because the IsValueTuple did not handle nullable types and thus the `GenerateValueTupleDeserializer` method was never called.

I've added test for the case when the query does not return something (the tuple should be null) and when it does return something. Additionally added tests for the edge-case where you select 8 or 15 element tuples to make sure nothing broke there. 

* Added EF6 tests in netcoreapp3.0

* Fix up EF Core ref to always be 6.3.0
Co-authored-by: Alex Bagnolini <alex.bagnolini@ericsoft.com>
Co-authored-by: Nick Craver <craver@stackoverflow.com>

Co-authored-by: Youssef Victor <31348972+Youssef1313@users.noreply.github.com>

Fix small typo and grammar

We are now providing it programmatically so we don't disrupt SQL Server.

The comment suggested that more than the number specified in InListStringSplitCount and it would start to use string_split.  However, it is greater-than-or-equal-to (GTE) the number specified by InListStringSplitCount.

* - enable deterministic build
- lib updates for core packages

* set source root if needed

* Restore VS test runner

xUnit won't run in VS without it :)
Co-authored-by: Nick Craver <craver@stackoverflow.com>

In Sqlite, DateTime is stored as String, and custom serializers ultimately use Convert.ChangeType to do the conversion. The generated code however doesn't set the Invariant culture, and the conversion can fail if the current culture doesn't match the standard DateTime serialization format from Sqlite.

* Use loop variable instead of 0

* Changed unit test, added additional nesting level

framework and tooling lib updates; does *NOT* include the full corpus of 3rd party ADO.NET providers

This is a branch cleaning up old `#if` clauses, `.csproj` bits, Linux build compatibility, and a few other tid-bits (broken out by commit).

In general: we're on 3.0 tooling and all previous cruft has been removed.

…cording to their names.

Hand coded version did not operate under the same circumstances as others.

- Add CommandBehavior.SingleResult and CommandBehavior.SingleRow. All other framworks does this under the hood.
- Run SqlCommand.Prepare() (perhaps it should be benchmarked as "Compiled"?)
- Use GetNullableValue because it's easier to read and slightly faster
- _table.Rows was incrementally added, which made the benchmark run slower the longer it executed. Added _table.Rows.Clear() to prevent the growth.
- Changed select to use * like other benchmarks

Hi!
This PR tries to fix Dapper.Performance.Tests, actually you can't get [MemoryDiagnoser](https://benchmarkdotnet.org/api/BenchmarkDotNet.Diagnosers.MemoryDiagnoser.html)  stats because 
```IDiagnoser.GetColumnProvider()``` method was gone on BenchmarkDotnet 0.11.2. I choose ```DefaultColumnProviders.Metrics``` because of [this](https://github.com/dotnet/BenchmarkDotNet/issues/1184#issuecomment-507377695), i think it's the most simple way to add stats now.

**Before**
![image](https://user-images.githubusercontent.com/34173061/65264060-1a5f7780-db0e-11e9-9c8c-bb4c69a808df.png)

**After**
![image](https://user-images.githubusercontent.com/34173061/65264158-4f6bca00-db0e-11e9-9313-597b09070ae6.png)

On the other hand i choose removing ```[Ordered]``` as temporal solution of https://github.com/dotnet/BenchmarkDotNet/issues/1238, and yeah, the results are still ordered because of https://github.com/StackExchange/Dapper/blob/9282ef2872e1c82c68c86f31e93e8f1ceaacd409/Dapper.Tests.Performance/Config.cs#L42

**Before**
![image](https://user-images.githubusercontent.com/34173061/65264660-6232ce80-db0f-11e9-9046-9c2d8fc5e2d8.png)

**After**
 (No errors)

Thanks for all!!

Fix issue #418 for DeleteAsync (with tests)

* Handle char correctly

* Add PostgreSQL char test

* Test for string before testing for char