When has submitting addons to AMO become such a nuisance?

I am currently battling with the manual review process for almost 2 weeks now and start to get really irritated by whatever policy it is Mozilla is trying to enforce here…

You can check my handle on AMO, i have 9 listed addons including minified code so i am somewhat accustomed to how this usually works, or at least so i thought.

The addon in question is https://addons.mozilla.org/en-US/firefox/addon/tracker-beeper/

Currently unlisted because Mozilla

It’s a lose implementation of Bert Hubert’s Tracker Beeper https://github.com/berthubert/googerteller in pure JS.

Everything was cool and dandy until i made the grave mistake to add Chart.js to plot the number of beeps per day to the options page in version 0.0.8

At first i included the minified version from https://github.com/chartjs/Chart.js/releases which is when the horror started, because i include minified code in the addon, well da.

Coincidentally Mozilla in their infinite wisdom decided to disable every version, including 0.0.7 and older not even including Chart.js, i quit looking for any logic and reason about a week ago so whatever.

Anywho, then i submitted a new version pulling the script from deliverjs with the SRI checksum set, which of course got also rejected because all scripts have to be bundled, i am starting to notice a pattern here.

Then i made the build process 1000% more complicated than it has to be by adding webpack to bundle Chart.js into an UML and ship that, so basically version 0.0.8 all over again except much much much more complicated.

So complicated that Mozilla failed to build this version because despite my best efforts including a README with all 3 steps how to build it Mozilla failed because they obviously don’t know building addons with web-ext implies installing web-ext…i assume whoever does these builds doesn’t have any background in STEM considering the amount of hand holding i am doing and all my replies to the reviewer falling on completely deaf ears, but that’s just my observation after 2 weeks.

Again, i then submitted another version including web-ext as a devDependency which is now producing a different UML for Mozilla, why that is will be again up to me to figure out by wasting another hour of my life because all i ever get from Mozilla is some cryptic stdout.

I am this close to admit defeat and revert the code to 0.0.7 without Chart.js and call it quits.

Make sure to submit also package-lock.json file.
Also, don’t use any global libraries.
That way both of you should get the same result.

Ideally your build steps should be:

npm i
npm run your_build_script

But yeah, these review issues can be pretty frustrating, but it still feels better than what I experienced in Chrome, Opera and Safari stores.

I think they are dealing with a lot of malicious actors so they are super strict about reviews.

Make sure to submit also package-lock.json file.

I am going to try this next and see how it goes, despite my better judgment committing the file.

I also noticed web-ext is unable to reproduce artifacts between Windows and Linux, but that seems to be by-design: https://github.com/mozilla/web-ext/issues/2381