Removing this feature causes boost in performance when using Ruby 1.9.
Ruby 1.9 started to do implicit conversions using `to_ary` and `to_str`
in some STDLIB methods (like Array#join). To do such implicit conversions,
Ruby 1.9 always dispatches the method and rescues the NoMethodError exception
in case one is raised.
Therefore, since the whiners feature defined NilClass#method_missing, such
implicit conversions for nil became much, much slower. In fact, just defining
NilClass#method_missing (even without the whiners feature) already causes a
massive slow down. Here is a snippet that shows such slow down:
require "benchmark"
Benchmark.realtime { 1_000.times { [nil,nil,nil].join } }
class NilClass
def method_missing(*args)
raise NoMethodError
end
end
Benchmark.realtime { 1_000.times { [nil,nil,nil].join } }
This commit vastly reduces the impact of auto
explain logging when enabled, while keeping
a negligible cost when disabled.
The first implementation was based on the idea
of subscribing to "sql.active_record" when
needed, and unsubscribing once done. This is
the idea behind AR::Relation#explain. Subscribe,
collect, unsubscribe.
But with the current implementation of notifications
unsubscribing is costly, because it wipes an internal
cache and that puts a penalty on the next event.
So we are switching to an approach where a long-running
subscriber is listening. Instead of collecting the
queries with a closure in a dedicated subscriber, now
we setup a thread local.
If the feature is disabled by setting the threshold
to nil, the subscriber will call a method that does
nothing. That's totally cheap.
Rationale: As discussed with José and Jon, this convenience
shortcut is not clearly justified and it could let the user
thing the disabled EXPLAINs are related to the model instance
rather than being globally disabled.
Implement a mini state machine for serialized attributes. This means we
do not have to deserialize the values upon initialization, which means
that if we never actually access the attribute, we never have to
deserialize it.
'Type' is a reserved column for STI. Changed conditions example to
avoid using that column name as an example. The example isn't
STI-related (and mentioning STI here is needless clutter), so changing
to avoid accidentally encouraging users to use 'type' as a column name
for other purposes.
This fixes a situation I encountered where a subclass would cache the
name of a generated attribute method in @_defined_class_methods. Then,
when the superclass has it's attribute methods undefined, the subclass
would always have to dispatch through method_missing, because the
presence of the attribute in @_defined_class_methods would mean that it
is never generated again, even if undefine_attribute_methods is called
on the subclass.
There various other confusing edge cases like this. STI classes share
columns, so let's just keep all the attribute method generation state
isolated to the base class.
