Theme Migration for Thunderbird 68

Andrei Hajdukewycz sancus at thunderbird.net
Sat Jul 20 04:01:03 UTC 2019


Hi all,

I'm planning to migrate certain lightweight themes to web extension 
themes on Monday(July 22 2018).  After some discussion, I think it's 
best to only migrate the top 1000 by (old) average user count. This 
doesn't mean that the others will disappear into the ether: AMO migrated 
all 450,000+ of them, so you will be able to find old themes there. We 
could even link to AMO via the discovery pane as an additional source of 
themes.

There are two major reasons not to migrate them all on ATN:

 1. It would take a very long time, minimum 3 weeks, possibly more.
    During this time, the website may be slow and errors may happen more
    frequently.
 2. Lightweight themes are very limited in functionality and this lead
    to them being fairly low quality, since they only changed the header
    image and two colors. Web Extension themes can customize much more
    of Thunderbird, and as a result look much better. Having fewer old
    migrated themes will mean that newer themes have a much better
    chance of standing out and there will be much less clutter of old,
    poor quality themes.

AFTER the migration is complete and I've verified that it worked as 
expected, the Themes section of the site will be switched over to 
display only Web Extension themes. That means they'll show up in search 
as well. This will likely happen on Tuesday(July 23, 2018), but I will 
send a followup mail with a specific time on Monday. There should not be 
any downtime.

Note that although AMO themes don't work in Thunderbird 60, our migrated 
themes will -- and it is very simple to "fix" an AMO theme so that it 
can be uploaded to ATN and work in 60. The manifest.json needs to have 
the 'id' (UUID for the add-on) in the applications object, as noted in 
the documentation on developers.thunderbird.net 
<https://developer.thunderbird.net/add-ons/web-extension-themes#use-a-manifest-json-file>. 
Our validator will not accept themes without a UUID set, so no broken 
themes can be uploaded. This is necessary because we don't sign add-ons. 
When AMO signs a theme with no UUID, they generate and set one in the 
signing data. Thunderbird 60.* does not read signing data correctly. 68+ 
does, which is why themes downloaded from AMO work on 68 but not on 60.

Finally, the 1000 number is negotiable. If many people feel we should 
migrate more than that, I will consider it. However I really do think 
that giving new Themes the focus is better. Right now very, very few 
people use lightweight themes at all. I checked by manually querying the 
logs for attempts to hit the theme update service(which we never set up 
because it was deemed not worth the time) and 0.7% of users currently 
use lightweight themes. Yes, less than 1%. My hope is that we'll get 
people uploading much nicer Web Extension Themes and that percentage 
will go up.

P.S. To get an idea of how much you can do with WE themes, check out the 
Firefox documentation 
<https://developer.mozilla.org/en-US/docs/Mozilla/Add-ons/WebExtensions/manifest.json/theme>. 
Our docs are much more basic, but I'd like to expand them over time. One 
cool example I've seen is a dynamic theme that changes background colors 
based on what time of day it is 
<https://github.com/mdn/webextensions-examples/tree/master/dynamic-theme>.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.mozilla.org/pipermail/tb-planning/attachments/20190719/033de549/attachment-0001.html>


More information about the tb-planning mailing list