I am also of the opinion that the resource type checks are unhappy. We depend on driver internals.

removed unnecessary  setup/teardown-methods

* removed "charset" patch
* implemeted resource_type compare with in_array instead of regex
* implemented functional tests for connection and statement.

Revert "build a correct connection string for mirrored databases, too. In this case "host=ip1:port,ip2:port" servername and dbname are optional but can also be used without host parameter"

This reverts commit 0698cba6.

build a correct connection string for mirrored databases, too. In this case "host=ip1:port,ip2:port"
servername and dbname are optional but can also be used without host parameter

if $persistent=false, resource_type == 'SQLAnywhere connection'

Backport #2650 to 2.5.x

Backport #2651 to 2.5.x

Backport #2637 to 2.5.x
Backport #2645 to 2.5.x

Backport #2636 to 2.5.x

Backport #2646 to 2.5.x

Backport #2638 to 2.5.x

Update Version.php

Backport #2634 into 2.5.*

Backport #2635 to 2.5.*

Backport #2546 to 2.5.x

[DBAL-2546] Fixed the failure of `Functional\StatementTest::testReuseStatementWithParameterBoundByReference()` on ibm_db2

The test is partially skipped due to:

1. Fetching blob as binary string from pdo_sqlsrv requires using `PDOStatement::bindColumn()` and `PDOStatement::fetch(PDO::FETCH_BIND)` which are not supported by the DBAL.
2. Setting encoding on the connection level is not supported: https://msdn.microsoft.com/en-us/library/ff628164(v=sql.105).aspx

[DBAL-2546] Fixed the failure of `Functional\StatementTest::testIncompletelyFetchedStatementDoesNotBlockConnection()` on pdo_mysql + libmysql

The test case being fixed should reset shared connection during teardown as one if its methods modifies connection attributes.