Betterbird Blog

We've lamented the squanderous allocation of Thunderbird resources before, so here is the next installment.

This mailing list post just hit the so-called Thunderbird UX mailing list. They author announces a redesign with the aim to simplify the Calendar UI, amongst other things. So what they have done is actually to remove functionality. But they author of the article preempts any concern by stating (quote):

These changes will be completely optional, kept in a separate “New Calendar” space, so those who need legacy features won’t lost access to any of their workflows.

How useful, so the rework will either be hidden, or users will hide it to get the functionality back that they need.

Thunderbird DOES NOT NEED yet another redesign and refactoring, what it needs is that the 2,538 open Calendar bugs be properly triaged and the most important ones fixed, like these:

Full list here, sort by Votes.

The Calendar component has been mostly abandoned since its Module Owner has moved onto greener pastures in so-called Thunderbird services (Thundermail, etc.). So now, instead of providing real functionality improvements, the Thunderbird folks have switched to polishing the turds.

Since the AI bubble is inexorably upon us, it's time to talk about it.

AI advances have boosted productivity in Thunderbird/Betterbird development. In the Thunderbird project, various folks have solved bugs with AI-assisted coding tools, powered by models such as OpenAI's GPT family, Anthropic's Claude models, Google's Gemini models, and DeepSeek models. They have analysed code, proposed and reviewed patches, written tests, and have even run the entire development pipeline via agents right up to running new tests and submitting work to Mozilla's Phabricator platform.

Our project hasn't spend big on AI subscriptions, but with humble subscription to DeepSeek v4 Pro, we've been able to analyse issues (see articles's screenshot), written scripts and automated tedious tasks like localisation, since one new string needs to be added to 15 localisation files and also be made consistent with existing translations in Thunderbird. So far, we haven't produced a Betterbird patch with AI yet, but we're looking forward to the first adequate task.

Update: We had DeepSeek revise our patch for showing native alerts on the application screen. AI pointed out some shortcomings wrt. to screen scaling and added code to cater for monitors aligned vertically. As a reminder, the function can be enabled with preference mail.notification.show_on_app_screen.

But AI is not only helping developers, users also benefit by having AI summarise messages or help when writing message by rephrasing, summarising, explaining, translating, etc. Here's a table of available add-ons.

Disclaimer: We've only surveyed listings on ATN and not tested those add-ons. There is no relationship between us and any of the add-on authors. If you are an add-on author and we misrepresented your add-on, please get in touch.

For e-mail processing, it makes sense to use local models, for example via Ollama. We also gave LM Studio a try, but found the model download unbearably slow on a fibre optic connection. Using a local model, e-mail data is not sent to the AI provider's data centre for processing.

Ollama users should note that in order for their add-on to be able to contact the local server at http://localhost:11434, they need to work around a CORS (Cross-Origin Resource Sharing) issue leading to a 403 error. Ollama needs to send the Access-Control-Allow-Origin: * header which can be enabled by setting the environment variable OLLAMA_ORIGINS to a value of *.

The Thunderbird folks integrated the add-on Close to tray while at the same time dropping the minimise to tray function.

We restored the latter. The article screenshot shows the relevant UI in the settings. And of course it works on Linux, too, where as Thunderbird users have been waiting since 2020, and now, the Windows version doesn't even support minimise to tray any more.

These are the relevant preferences:

• mail.minimizeToTray
• mail.closeToTray
• mail.startupMinimized (in Thunderbird: mail.closeToTray.startInTray)

As widely announced on Thunderbird forums and mailing lists, Thunderbird is diversifying. First they bought the K-9 Android e-mail client and rebadged it as Thunderbird for Android. Next, they are developing an iOS mail application from scratch. The last undertaking is to offer so-called Pro Services, called "TB Pro": www.tb.pro. See the 2025 announcement here.

The mail hosting part of those services is called Thundermail. Currently, it's not yet available, but one can join a waiting list. Expect to pay 12 x USD 6 = USD 72 anually. Not a small price, given that for less money you can resister your own domain and pay a hosting service, or have your mail hosted at our advertising partner ForwardEmail. Of course, there are also long-standing e-mail services, like Fastmail, Posteo or Proton Mail, each backed up be years or decades of experience.

All that said, the Thunderbird folks are rolling out the integration of their Thundermail service in Thunderbird 153, currently in beta, see screenshot above. Technically, Thundermail is integrated as system add-on, which ships inside the application, packaged into omni.ja at chrome/messenger/builtin-addons/thundermail. It's enabled in the new Account Hub via preference mail.accounthub.thundermail.enabled.

The Thunderbird folks are shipping their version 153.0 beta 2 with Thundermail enabled. We're currently disabling Thundermail via the mentioned preference for various reasons:

• It doesn't yet appear to be available.
• We're not convinced that this is the best possible mail hosting offer.
• Our project does not aim to increase Thunderbird income, but to improve user experience.

Most of the source code of Thundermail is only published in minified/obfuscated form. We did a quick AI analysis of what is publicly visible, and here's the summary of the US services involved in the setup:

Any Betterbird user wishing to try or use Thundermail can easily set the preference to true.

Update 25th June 2026: We've just been made aware that apart from showing "Sign in with Thundermail" during account creation, Thundermail is also forced onto the application menu:

The DOM ID of this element is tbpro-menu-id-root; it is added here.

The add-on code is extracted from this repository and pushed into the Thunderbird tree. Here it is:

https://hg-edge.mozilla.org/comm-central/file/tip/mail/extensions/builtin-addons/thundermail/extension.

This code is not indexed in Searchfox.

Packaging software some or most users don't want into the application via some system add-on sounds like something Mozilla has done before? Well, check this Lunduke video.

Today we're starting our beta program for Betterbird 153 with 153.0b1-bb1. Please refer to the Release Notes for full details.

We're shipping this only in English and for Windows. Language packs for other languages are available.

Note that we're just starting to gain experience with the new 153 series, so please do not use this version in production yet. Also note that Thunderbird migrated password storage from logins.json to logins.db, so when downgrading, you need to re-enter all your passwords (or restore the previous logins.json file).

We've shipped Betterbird 140.12.0esr-bb24 today. Please refer to the Release Notes for full details.

This release contains fixes some important bugs:

Cancel sending had broken in Thunderbird 135, and no one noticed. We've already written about it. This is fixed in Betterbird 140 now, the Thunderbird folks fixed it for release 153.

Secondary sort is not a well-known feature. To make it a bit more obvious, Betterbird shows secondary sort indicators on column headings in table view where appropriate. Those weren't aligned correctly in default and relaxed densities.

Betterbird had a widely misunderstood feature where the user could zoom images in links if they were oversized. We had a few complaints since people didn't recognise the link, and if they did, had to used the context menu to open it. Finally a user sent us a sample e-mail which was rendered quite incorrectly due to the feature. That was the last nail in the coffin and the feature was dropped.

Lastly this release fixes a few issues with reversed sorting and collapsing a thread to the most recent message. This is driven be preferences mailnews.reversed_threading and mailnews.collapse_thread_to_most_recent. Check this article for details about those function.

While the Thunderbird folks are finally making their "account colour" implementation more complete in this ticket, they are discussing amongst themselves whether a colour indicator should only be added to the message list (also called thread pane) for the beloved (irony!) cards view, the one that at default values wastes a lot of space, or whether it should be added to table view as well.

So one developer says to the other (quote):

Let's do this without [table view] and see how many complaints there are.

Surely user complaints are a good design principle.

Needless to say that Betterbird has shipped full account colours for both views since Betterbird 115.6.0 and even full row colour since Betterbird 115.8.0, so for more than two years. And those folks are still discussing whether a partial solution will do.

UPDATE 12th June 2026: So the feature got included in Thunderbird 153, currently in Daily/Alpha stage, going to Beta next week. As mentioned, an account indicator is only displayed in cards view and in the folder list. Apparently, the Thunderbird folks are seeking more complaints, since all people who haven't configured account colours will now get a rather useless additional blue indicator showing the default colour:

(Screenshot taken using some newsgroup content.)

And in the folder list:

How useful is that?

In Betterbird, we're maintaining the indicator, but we're not showing it, when no colour is set:

Very recently it has come to our attention that cancelling the process of sending a message via the Cancel button on the progress panel wasn't working any more after fixing the very same issue in version 115.

Unfortunately, we picked up a regression from Thunderbird 135, so Betterbird 140 was also affected. Today we shipped a fix in our "latest build 5". The Thunderbird folks are looking into getting a more comprehensive fix shipped for version 153. "More comprehensive" since this area was hit by a total of three regressions, the one affecting Betterbird 140 being the first one of the trilogy. We'll make sure it's all back to working order for Betterbird 153.

It's a bit puzzling that the failure wasn't noticed earlier. We consider being able to cancel an ongoing send operation as important feature, sometimes the last resort when a message was sent to the wrong recipients or with the wrong attachments. The Thunderbird folks pride themselves that most of their functionality is covered by automated tests, but this important feature obviously isn't.

While adjusting our fixes and features to the new Thunderbird 153 code base, we painfully come across bugs that the Thunderbird folks have abandoned and where Betterbird had to come to the rescue to offer users a decent experience. Here are some of them, and yes, we mentioned them before here and here.

Let's ask again (quoted from here): Where is the management who assigns a developer to get the issue fixed? Where is accountability?

Oops, looks like the Director, Desktop & Mobile Apps abandoned some of the bugs himself. Well, there you go!

We are busily adjusting our fixes and features to the new Thunderbird 153 code base as outlined in this article. When working on the "attachments on top" feature for both message display and compose window, we noticed that icons for certain documents are no longer displayed next to the attachment name in using the 153 code base. This also happens in other parts of the system, like here, in the Downloads Manager:

We did a bit of digging in Thunderbird's bug tracker, and lo and behold, found this ticket.

So a regression introduced in Thunderbird 141 which was shipped in July 2025, about 10 months ago. Finally, users noticed and headed to the support forum in November 2025, from where a volunteer filed the ticked we mentioned, also in November 2025. And what did the developers do? They shrugged their shoulders and left the software broken, until yesterday when two volunteers miraculously started showing some activity.

So where is QA to detect such failures before users do? Where is the management that monitors clearly visible regressions? Where is the management who assigns a developer to get the issue fixed? Where is accountability?

Ten years back, the Thunderbird project was run by some volunteers, who contributed as their personal circumstances allowed. The then chairmen of the chairman of the Thunderbird Council used to refer to the team as a "pack of cats" (although the correct term is "clowder of cats"). So with 50+ staff and millions of dollars at their disposal, what has actually changed in the attitude of those running the project?

Old and annoying 20+ year old bugs seem the be exclusively addressed by volunteers, and from them we heard questions and statements like:

• Why do [old bugs] have such little priority that I can find 20 year old low-hanging fruit?
• If I were running the shop, there would be a freeze on new internal features [...] and a systematic revision of all bug reports.

Well, good luck with 16.000+ open tickets. BTW, here is what AI had so say about so-called bug rot:

“Bug rot” is the gradual degradation of a bug tracker’s usefulness as unresolved issues accumulate faster than they are properly triaged, resolved, or retired. Over time, a large backlog—often spanning years or even decades—mixes active defects with obsolete reports, duplicates, unreproducible issues, and context that no longer matches the current codebase. This creates a system where the signal-to-noise ratio steadily worsens: important bugs are harder to identify, prioritization becomes less reliable, and contributors increasingly rely on informal knowledge rather than the tracker itself. The result is not that the project stops working, but that its issue tracker shifts from being a precise engineering tool into a partially archival, partially operational database whose meaning becomes progressively harder to interpret.

Fun fact: Paid Thunderbird staff approach those volunteers asking them to take on further bugs. That's true Mozilla exploitation culture. All those volunteers are part of the business model that counts on highly skilled people to work for free. Of course there are lots of honourable mentions, and some of them even get a free T-shirt. Our project leader used to be one of them, he still has two Mozilla T-shirts. And he usually says: Look at this T-shirt, it's worth about $50,000 of labour.