Updated code to ES6 #85
Conversation
|
Well, there are lies, darn lies and benchmarks :-) Lets look at CI results first: comparing the previous CI results listed as ES5 to the CI results for this PR listed as ES6
From this comparison we would be stupid to migrate from ES5 to ES6, but remember this is cloud and VM performance can vary quite some and Github does not promise any performance on its Actions :-( It also does not make a lot of sense that the V8 team would be improving their performance for ES5 but not for ES6 nor that v10 would be way faster than v14 (the module is not that exotic imho). So I redid the tests on my local machine (linux container on x86 chromebook). There are probably processes running in the background as well but the results are more stable and logical than from Github Actions:
Here it shows that the difference in performance is marginal as to be expected. Running the same npm ci for ES6 on v14 5 times results in values between 4.82 and 4.97. So my conclusion would be that the performance difference is minimal. Kind regards, |
|
FYI: This PR is a more recent version of #46 which can now also be closed imho. |
I have updated all code to ES6 in two commits:
All code, including examples in the README, has been converted using the ES5 to ES6 extension of Visual code and then checked by
standard --fix.(the code examples in the README were extracted to tempfiles, converted and pasted back)
I only did manual updates to code to make
standard --fixsucceed.This process leaves room for further optimizations (e.g. creating javascript classes, removing some dependencies). I'm happy to try those as well but I thought I'd first do a most basic conversion and work from that.
In the README I also updated the Node versions tested in CI ;-)
Kind regards,
Hans