When running 'script/stackprof object', I noticed that it would be helpful to see
GC information. It appears we create a lot of junk -- a source of optimization if
we decide to go down that path.
An average Jekyll build doesn't run a GC, but auto-regeneration likely would eventually
require a GC run and it would be interesting to see if we can reduce how much we throw
away with each call to 'site.process'.
* master: (41 commits)
Fix rubocop offenses on master.
script/fmt: print Rubocop version
Update history to reflect merge of #5030 [ci skip]
rubocop: separate deprecator error messages
Update history to reflect merge of #5031 [ci skip]
Update history to reflect merge of #5032 [ci skip]
rubocop: fix code style
rubocop: fix code style
rubocop: fix code style
Update history to reflect merge of #5024 [ci skip]
Update history to reflect merge of #5025 [ci skip]
utils: check that the object is a hash when merging default_proc
Update history to reflect merge of #5026 [ci skip]
rubocop: refactor modified? method
Add a benchmark for capture vs. assign in Liquid. [ci skip]
Update history to reflect merge of #5027 [ci skip]
Add generator-jekyllized to third party plugins
rubocop: fix code style
rubocop: fix code style
rubocop: fix code style
...
The problem last time was that we removed Pry and Pry brings in CodeRay, we were
testing legacy stuff and didn't have CodeRay in our dependencies, which resulted
in those tests failing.
This also quietly announces the intention to move to RSpec by moving the old
test dependencies to ":test_legacy" and is slightly less agressive in it's
organization than before.
* Move step_definitions/jekyll.rb to just step_definitions.rb
* Rename the formatter to Jekyll::Cucumber::Formatter, it's Jekyll's.
* Add some flair; switch to checks!
* Rename env.rb to helpers.rb
Bogdan
Parker Moore
Frank Taillandier
Ashwin Maroli
XhmikosR
Ben Balter
chrisfinazzo
Jordon Bedwell
jekyllbot
Alfred Xing
AJ Acevedo
Florian Weingarten
Martijn den Hoedt
Nicholas Burlett