BrowserWorks / Waterfox

The official Waterfox 💧 source code repository
https://www.waterfox.net
Other
3.84k stars 342 forks source link

Occasional extension update issues #479

Closed grahamperrin closed 6 years ago

grahamperrin commented 6 years ago

Apologies for the vague title (we can change it to something more specific) … and @jbeich part of this issue might be better placed in FreeBSD Bugzilla … but for now I'll throw outline details into the MrAlex94/Waterfox/ repo because I'm aware of work towards Waterfox-oriented add-on services.

Example: Containers Sync, 2018-02-25

2018-02-25 10 26 42 update not found by about addons

tl;dr version 1.0.1 became available in November 2017 but three months later, the update (from 1.0) was not found by about:addons in Waterfox on FreeBSD-CURRENT. At AMO I see some platform specificity for 1.0.1:

Works with Firefox for Android 55.0 and later, Firefox 55.0 and later

– so I'm reminded of things such as https://github.com/mozilla/addons-frontend/issues/2998 but that was a front end AMO issue.

Example: Ant Video Downloader

tl;dr for https://www.reddit.com/r/waterfox/comments/85w44q/-/dw0xtjs/ I created a new profile named 'elevation' and set preferences to allow installation of 3.2.34,

Works with Firefox for Android 57.0 and later, Firefox 57.0 and later

– but refrained from installation.

Instead I:

  1. began with preferences set to allow installation of legacy 2.4.7.50
  2. installed 2.4.7.50
  3. changed preferences to allow installation of 3.2.34

– about:addons check finds no update (from 2.4.7.50 to 3.2.34). Waterfox 56.0.4.36_3 (56.0.4.1) on FreeBSD-CURRENT.


Maybe tomorrow night I'll make time to compare with Waterfox on a Mac …

grahamperrin commented 6 years ago

Containers Sync: issue not reproducible on a Mac. After installation of outdated 1.0, about:addons finds the update to 1.0.1.

Reproducible with Waterfox on FreeBSD-CURRENT, so maybe it's an AMO back end bug.


Ant Video Downloader: no automated update on the Mac. After manually updating, things made sense. My bad: I forgot that checkCompatibility2 simply disables compatibility checking (doesn't make the browser behave as if it's compatible).