Convert each plugin to a function from a class to an overriding class

[ghost]

Q	A
Bug fix?	no
Breaking change?	no
New feature?	no
Deprecations?	no
Spec compliancy?	no
Tests added/pass?	tests pass
Fixed tickets
License	MIT

Instead of assigning functions to a parser instance, this creates a class with overridden methods.
The class is a function of the plugins (only "jsx", "flow", and "estree"), and this function is cached.

Based on my earlier comment: #393 (comment)
Like #393, this avoids fixing the indent inside each plugin, in order to have a smaller diff.

This will improve performance because we aren't assigning functions directly to the instance any more.
I ran the perf script again on my home machine just to be sure:

branch	time
master	26.8 26.8 27.1 26.0 26.0
plugin_class	12.8 13.0 13.1 13.3 12.7

[hzoo]

Hey @andy-ms! It looks like one or more of your builds have failed. I've copied the relevant info below to save you some time.

• Node vlts/* - linux
• Node vlts/* - linux

[codecov]

Codecov Report

Merging #459 into master will decrease coverage by 1.28%.
The diff coverage is 97.03%.

@@            Coverage Diff             @@
##           master     #459      +/-   ##
==========================================
- Coverage   98.26%   96.97%   -1.29%     
==========================================
  Files          20       20              
  Lines        3511     3469      -42     
  Branches      934      935       +1     
==========================================
- Hits         3450     3364      -86     
- Misses         22       55      +33     
- Partials       39       50      +11

Flag	Coverage Δ
#babel	?
#babylon	96.97% <97.03%> (-0.04%)	⬇️

Impacted Files	Coverage Δ
src/parser/index.js	100% <100%> (ø)	⬆️
src/plugins/jsx/index.js	97.71% <100%> (-0.03%)	⬇️
src/plugins/estree.js	99.17% <100%> (-0.09%)	⬇️
src/index.js	100% <100%> (ø)	⬆️
src/plugins/flow.js	97.01% <94.44%> (-1.2%)	⬇️
src/parser/expression.js	93.84% <0%> (-3.54%)	⬇️
src/tokenizer/state.js	97.22% <0%> (-2.78%)	⬇️
src/tokenizer/index.js	97.32% <0%> (-1.03%)	⬇️
src/parser/statement.js	98.15% <0%> (-0.93%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 344f070...c5fec5b. Read the comment docs.

[codecov]

Codecov Report

Merging #459 into master will decrease coverage by 0.05%.
The diff coverage is 96.03%.

@@            Coverage Diff            @@
##           master    #459      +/-   ##
=========================================
- Coverage   98.26%   98.2%   -0.06%     
=========================================
  Files          20      20              
  Lines        3506    3462      -44     
  Branches      934     935       +1     
=========================================
- Hits         3445    3400      -45     
- Misses         22      23       +1     
  Partials       39      39

Flag	Coverage Δ
#babel	82.49% <66.66%> (+0.15%)	⬆️
#babylon	96.96% <94.17%> (-0.04%)	⬇️

Impacted Files	Coverage Δ
src/index.js	100% <100%> (ø)	⬆️
src/parser/index.js	100% <100%> (ø)	⬆️
src/plugins/jsx/index.js	97.71% <100%> (-0.03%)	⬇️
src/plugins/flow.js	97.99% <94.07%> (-0.21%)	⬇️
src/plugins/estree.js	99.17% <98.95%> (-0.09%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 2ef4366...20a83d9. Read the comment docs.

[DanielRosenwasser]

Hey @loganfsmyth @hzoo, looks like the tests are fixed. Can you take another look?

[STRML]

Like #393, this avoids fixing the indent inside each plugin, in order to have a smaller diff.

I so wish this were a default, but you can view the diff via https://github.com/babel/babylon/pull/459/files?w=1 (note the ?w=1) to avoid highlighting whitespace diffs.

[ghost]

@STRML Thanks!

[hzoo]

lgtm!

[hzoo]

extra space? return cls; guess we didn't have a rule for this, I guess we should land that prettier pr later

[danez]

I ran my perf script and the results are impressive. (Up to 100% faster 🎉🎉🎉🎉)

Before

name                            run   
./fixtures/angular.js           502ms 
./fixtures/backbone.js          60ms  
./fixtures/ember.debug.js       1468ms
./fixtures/jquery.js            279ms 
./fixtures/react-with-addons.js 576ms 

after

name                            run  
./fixtures/angular.js           298ms
./fixtures/backbone.js          27ms 
./fixtures/ember.debug.js       779ms
./fixtures/jquery.js            141ms
./fixtures/react-with-addons.js 314ms

Thanks a lot