Currently the Mozilla CI pipeline for Linux(32/64) runs on a 'hacked' Ubunt=
u 16.04 LTS image, last updated sometime around 6-8 months ago.
Meanwhile, the world has moved on to Ubuntu 18.04 LTS (released April 2018)=
with Ubuntu 20.04 LTS approaching fast (projected April 2020).
Within the Linux ecosystem, it can be argued that Ubuntu represents a large=
chunk (if not the majority) of users that use Firefox on the desktop. Whil=
e continuing to test on Ubuntu 16.04 will suffice for the time being, it is=
time to being exploring an upgrade path such that our CI pipeline better r=
eflects the user base.
I had began work in https://bugzilla.mozilla.org/show_bug.cgi?id=3D1565332 =
to upgrade the Ubuntu 16.04 images to Ubuntu 18.04 when a comment was recei=
ved suggesting to use Debian as an alternative.=20
The main advantage of switching to Debian is supposedly the more determinis=
tic builds that are easier to maintain. It is claimed that using Debian wou=
ld eliminate the sort of issues that led to https://bugzilla.mozilla.org/sh=
ow_bug.cgi?id=3D1503756 in the first place, for example.
Conversely, the main disadvantage of switching to Debian would be the lack =
of drivers and other required packages that Ubuntu provides. The full scope=
of work needed to make a Debian-based image workable with current CI tests=
is unknown at this time.
For comparison, please take a look at the Treeherder pushes for the followi=