We don't have the check the range when the value is coming from the DB, so
override type_cast_from_database to short-circuit the extra work.

type_cast_from_database (small)  3437507.5 (±29.2%) i/s -   14223135 in   4.996973s
type_cast_from_database (large)  3158588.7 (±28.3%) i/s -   13265628 in   4.992121s
type_cast (small)                 481984.8 (±14.2%) i/s -    2352012 in   5.005694s
type_cast (large)                 477331.8 (±14.2%) i/s -    2332824 in   5.012365s

Comparison:
type_cast_from_database (small):  3437507.5 i/s
type_cast_from_database (large):  3158588.7 i/s - 1.09x slower
type_cast (small):                 481984.8 i/s - 7.13x slower
type_cast (large):                 477331.8 i/s - 7.20x slower

The difference is huge but the absolute gain is quite small. That being said
this is a hotspot and it showed up on the radar when benchmarking discourse.

Oh hey, we got to remove some code because of that!

Arel handles this for us automatically. Updated tests, as BindParam is
no longer a subclass of SqlLiteral. We should remove the second argument
to substitute_at entirely, as it's no longer used

If something causes the teardown block to not get run (errors,
interrupting test runs), we have to manually delete them, which is a
pain.

Fix typo. [ci skip]

This caused a pretty major performance regression for 4.2, as this is a
hotspot for query construction. We're still slightly slower than 4.1,
but it's much less significant.

Makes life easier for people profiling on master.

This class no longer has any private methods

PERF: optimise type lookup to avoid invoking procs

Skip spring install in Cygwin due to fork() bad support.

Fix for assigns(:..) resetting template assertions

Update to Unicode 7.0.0

Minor English fixes in docs [ci skip]

This is a more conservative approach to 2602a49a. Also changed the comment to be
more inline with everything else in the file (describing what the config value
is doing and why). People should just read the docs for alternatives.

This reverts commit 2602a49a, reversing
changes made to 5d7c1057.

The explicit default was introduced in 21f6d723, so apps created with Rails < 4
have the commented out version, which means that this change would break those
apps.

When calling assigns(:...) or cookies(:...), template assertions would be reset, which they obviously shouldn't be.

Rename #source_extract to #source_extracts in ExceptionWrapper

Don't double check trace origin in ExceptionWrapper#traces

Don't show full trace on routing errors

See also: https://www.cygwin.com/faq.html#faq.using.fixing-fork-failures

If a trace isn't an application one, then it comes from a framework.
That's the definition of framework trace. We can speed up the traces
generation if we don't double check that.

Since dbcbbcf2 the full trace is shown
by default on routing errors. While this is a nice feature to have, it
does take the attention off the routes table in this view and I think
this is what most of the people look for in this page.

Added an exception to the default trace switching rule to remove that
noise.

add Nokogiri to docs generation guides [ci skip]

It returns multiple source extracts since 1ed264bc. Also cleaned its
result structure, as we no longer need the file in a code extract.

[ci skip] Missing space in Active Job guide for referring to Action Mailer

Don't let #{application,framework,full}_trace be nil

Those three can be nil when exception backtrace is nil. This happens and
that forced a couple of nil guards in the code. I'm proposing to make
those always return an array, even on nil backtrace.

7.0.0 was released on June 16, 2014

http://unicode-inc.blogspot.com.ar/2014/10/unicode-version-70-complete-text-of.html

ruby bin/generate_tables

[ci skip] add "Qu" to Backends Features list