fix: Bug that on batch fetching string columns cannot be retrieved

[Wakusei]

What does this PR do?

Fix a bug that on batch fetching string columns cannot be retrieved.
With a multi-row cursor SQLSetPos must be called before calling SQLGetData.

This bug happens on the following situation.

auto results= nanodbc::execute(conn, L"SELECT * FROM blahblah", 3);
while(results.next)
{
    auto i = result.get<std::wstring>( 0 );
}

[mloskot]

Thanks for the fix.

https://github.com/nanodbc/nanodbc/blob/master/README.md#tests says:

One of important objectives is to maintain nanodbc covered with tests. New contributions submitted via Pull Requests must include corresponding tests.

So, in order to review and accept this PR, it needs to be updated with test case that reproduces the bug, that's failing due to the bug, and that confirms the fix for the bug.

[lexicalunit]

Thank you for the fix @Wakusei!

Please add a unit test for this as @mloskot suggests.

Then rebase/merge after #230 is merged to master and then run ./utility/style.sh on your PR to ensure that the code has been formatted properly.

[Wakusei]

I see. I'll try to fix them. I try to run style.sh on my local environment by myself, but if I cannot correctly prepare the tools, please help me.

[mloskot]

@Wakusei most important is to add test case for the bug/bugfix. The formatting, we (I) can take care of it, so don't worry if you hit an obstacle.

If you have a problem, just ask, I'll help.

[detule]

@Wakusei this is a neat feature:

I wonder if we shouldn't call SQLSetPos with operation == SQL_REFRESH, rather than SQL_POSITION. This to help with also fetching the data in the columns that have bound buffers.

If I am reading the docs on SQLBindCol correctly (my emphasis added):

TargetValuePtr
[Deferred Input/Output] Pointer to the data buffer to bind to the column. SQLFetch and SQLFetchScroll return data in this buffer. SQLBulkOperations returns data in this buffer when Operation is SQL_FETCH_BY_BOOKMARK; it retrieves data from this buffer when Operation is SQL_ADD or SQL_UPDATE_BY_BOOKMARK. SQLSetPos returns data in this buffer when Operation is SQL_REFRESH; it retrieves data from this buffer when Operation is SQL_UPDATE.

[mloskot]

I've updated the PR's branch with the latest main resolving some conflicts.

[mloskot]

I have just merged main to update to this PR to pick up recent changes to CI and clang-format runs.
I did not want to rebase to not to cause any trouble to your local repo.

This PR still needs test.

[mloskot]

I've taken the liberty to update this PR and catch up with recent CI updates.

It still needs a test case though.

[mloskot]

The CI coverage is now much better and it currently is failing for this PR with this error

-------------------------------------------------------------------------------
test_block_cursor_with_nvarchar
-------------------------------------------------------------------------------
D:\a\nanodbc\nanodbc\test\mssql_test.cpp(480)
...............................................................................

D:\a\nanodbc\nanodbc\test\mssql_test.cpp(502): FAILED:
  REQUIRE( results.next() )
due to unexpected exception with message:
  D:\a\nanodbc\nanodbc\nanodbc\nanodbc.cpp:4019: HY109: [Microsoft][ODBC Driver
  17 for SQL Server]Invalid cursor position 

-------------------------------------------------------------------------------
test_block_cursor_with_nvarchar_and_first_row_null
-------------------------------------------------------------------------------
D:\a\nanodbc\nanodbc\test\mssql_test.cpp(512)
...............................................................................

D:\a\nanodbc\nanodbc\test\mssql_test.cpp(530): FAILED:
  REQUIRE( results.next() )
due to unexpected exception with message:
  D:\a\nanodbc\nanodbc\nanodbc\nanodbc.cpp:4019: HY109: [Microsoft][ODBC Driver
  17 for SQL Server]Invalid cursor position 

-------------------------------------------------------------------------------
test_block_cursor_with_nvarchar_and_second_row_null
-------------------------------------------------------------------------------
D:\a\nanodbc\nanodbc\test\mssql_test.cpp(541)
...............................................................................

D:\a\nanodbc\nanodbc\test\mssql_test.cpp(558): FAILED:
  REQUIRE( results.next() )
due to unexpected exception with message:
  D:\a\nanodbc\nanodbc\nanodbc\nanodbc.cpp:4019: HY109: [Microsoft][ODBC Driver
  17 for SQL Server]Invalid cursor position 

===============================================================================
test cases:    77 |    74 passed | 3 failed
assertions: 28417 | 28414 passed | 3 failed
Errors while running CTest

This needs to be investigated further.
Perhaps it is related to what @detule discussed in #229 (comment)