i18n can still go wrong

[turt2live]

Resurgence of #10859

Possible cause being #10965

[bwindels]

If it's caused by missing cache, one way to prevent this would be to make all of the assets available offline through our service worker. This would also preload icons, etc. The hardest part probably being creating a list of all of our assets (during build) without causing a maintenance headache, although we might be able to get rid of the content hashes with this approach.

[t3chguy]

There's https://www.npmjs.com/package/serviceworker-webpack-plugin to auto-populate it with a blacklist

[bwindels]

Probably best to keep the content hashes though, in case the browser removes your cache but not your (outdated) service worker, you'd end up rerequesting but getting the asset for a newer version.

[ara4n]

I'm getting reports on this on app.element.io (i.e. not to do with cachebusting going wrong, unless it happened to coincide with a release getting pushed)?

[mcepl]

@ara4n And BTW, this is me from the email report on this issue.

[t3chguy]

We had to pull it back out due to other issues it caused: #15352

[Yoric]

Got it twice this week, once on Firefox Stable and once on Firefox Nightly, both times on app.element.io, both times when starting Firefox. In both cases, a reload fixed the issue.

[lutoma]

This happens to me constantly as well on Firefox.

[robintown]

Since the only actual reports on this issue are of this occurring in Firefox, closing this as a duplicate of #9422