correct handling situation when app reaches connection pool limit

[senid231]

because currently execution is blocked when no available connections in connection pool

[ioquatix]

This looks pretty interesting.

Can we replace the entire pool implementation rather than monkey patching part of it?

In what situations does the existing code fail and can we make a spec?

[coveralls]

Pull Request Test Coverage Report for Build 27

• 0 of 52 (0.0%) changed or added relevant lines in 8 files are covered.
• 12 unchanged lines in 2 files lost coverage.
• Overall coverage decreased (-33.6%) to 18.627%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
lib/async/postgres.rb	0	1	0.0%
lib/async/postgres/connection_pool.rb	0	2	0.0%
lib/async/postgres/patch_reactor.rb	0	2	0.0%
lib/async/postgres/replace_connection_handler.rb	0	2	0.0%
lib/async/postgres/queue.rb	0	9	0.0%
lib/async/postgres/connection_handler.rb	0	11	0.0%
lib/async/postgres/condition.rb	0	12	0.0%
lib/async/postgres/proxy.rb	0	13	0.0%
<!--	Total:	0	52	0.0%	-->

Files with Coverage Reduction	New Missed Lines	%
lib/async/postgres.rb	1	0.0%
lib/async/postgres/connection.rb	11	59.38%
<!--	Total:	12		-->

Totals
Change from base Build 26:	-33.6%
Covered Lines:	19
Relevant Lines:	102

---

💛 - Coveralls