defer activating the package until the window is loaded.

[akonwi]

Using the latest version of atom (1.34.0 at this time), this package is consistently the slowest to activate. I've recorded the average activation time for this package in two scenarios;

1. A window with one project open and reloaded 4 times: average activation time is 321ms
2. A window with no project open and reloaded 4 times: average activation time is 222ms

Using this activation hook, atom will defer activating this package until the window is ready to use and then there's no impact to startup time. Since this package doesn't need to do heavy setup, I think it should wait as long as possible to activate anyway.

[codecov-io]

Codecov Report

Merging #485 into master will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master     #485   +/-   ##
=======================================
  Coverage   75.91%   75.91%           
=======================================
  Files          27       27           
  Lines         436      436           
  Branches       42       42           
=======================================
  Hits          331      331           
  Misses         89       89           
  Partials       16       16

---

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 f324f1d...fd658ef. Read the comment docs.

[robwise]

Nice, this solves a recurring issue we've had where we keep getting startup performance regressions from new features being added.

Can you please rename the commit to use the conventional changelog standard, something like perf(startup): defer activating the package until the window is loaded?

[akonwi]

Woops. Sorry, I skimmed the contributing guide and missed that part. I've updated the commit

[robwise]

No worries, thanks so much for the PR :)

[akonwi]

I just noticed #330. I guess this resolves that as well. After reading the suggestions there and looking at the code, I'm pretty confident that all the lazy* functions can be removed because they don't gain much when the package is activated outside of the initial window startup, which is so messy and suboptimal. I'll test and may open another PR