Hacker News Reader

Comment by wvenable | Hacker News Reader

Some automatic updates should really just be different products. The most obvious one at the moment is Android Firefox is getting a big automatic update that, for a lot of people, will be a very big downgrade and an unfriendly UI change.

But Mozilla doesn't want to maintain security patches for X different products. And that's true of most vendors. This model of computing where all users are always on the latest version makes security patching and bug fixing a lot more straight forward.

If you use a web application, this is the model you've always had.

View on HN

I don't know about anybody else, but I know I have a huge motivation cliff when it comes to continuing to pay the consequences for old mistakes that I have already sacrificed to make amends for?

How do I sell my coworkers on making a change that's the Right Thing to Do when we all know damned well that we're still going to have to support the Wrong Way for another four years? I might as well grow thick callouses, learn to Embrace the Suck, or self-select out of this project.

If we could make such large changes while preserving the old behavior, I think we'd get away with it, but there are limits to how much you can refactor, and people like to hitch their agenda to something with momentum, like a rider on the Farm Bill, and the architects of the change might have ulterior motives of their own.

I don't know about anybody else, but I know I have a huge motivation cliff when it comes to continuing to pay the consequences for old mistakes that I have already sacrificed to make amends for?

Sometimes, you just have to be professional and deal with it, if you want to have a quality product and look after your users.

We'd never accept a doctor or an engineer taking this attitude. If they discovered a mistake with potentially serious consequences in their earlier work, we would expect them to make it good as quickly and safely as possible.

Software makes a crazy amount of money and has a huge impact on billions of lives every day. Frankly, we as devs enjoy good and often very well paid jobs. I don't think we should accept that attitude in our industry either. If anyone doesn't like that, maybe invest more in writing better code in the first place and in testing before going into production so there are fewer problems to revisit!

> If you use a web application, this is the model you've always had.

And good lord do I hate it

Now, not only do I have little-to-zero control over the bits on my machine, but I'm fully locked-in to whatever shenanigans Company X has decided is appropriate for me

Yeah, add me to the list of people who are very unhappy with the forced downgrade of Firefox. There was no reason to move the URL bar to the bottom, it makes tab switching more fiddly, previews of tabs seem worse, and I didn't ask for, want, or agree to any of this.

But Mozilla doesn't want to maintain security patches for X different products.

Are they really different products (versions), or are they just branches off a common tree and share so much code in common that it would be close to trivial to apply the fix to them all? My experience with security-related patches is that they're almost all extremely localised --- a missing check, an off-by-one, etc.

It seems like far less effort to do that than a massive rewrite of the UI, in any case...

The issue with this is that lets say in patch 1.0.5 someone used that missing check to get another feature working.

Now if you apply the security fix to 1.0.5 the other feature breaks.

Fennec (the old Firefox for Android) was, as far as I'm aware, a different codebase compared to Fenix (the new one). As in, Fennec was hosted on hg.mozilla.org using the same source tree as desktop Firefox, and Fenix is hosted on GitHub, if I understand things correctly.

Fenix embeds Gecko via GeckoView, which resides in mozilla-central.

This is much less of a problem if you only release new versions at a slower pace, as we used to. There is no good reason that something like a web browser needs to do major updates every month or two. For most users, having a major release annually with just point releases for security fixes in the meantime would be more than sufficient.

I've worked on plenty of professional software that would release a new major version perhaps every 1-2 years and seemed to have no difficulty with backporting security patches and sometimes other important bug fixes for several versions. It's not rocket science. It just doesn't fit nicely into the modern dev culture of pushing everything forwards as fast as possible with little regard for stability or backward compatibility and with the ratchet effects from data format changes and deploying updates to the software seen as positive things.

I was using Firefox Nightly occasionally which already had the controls on the bottom so the UI jump wasn't hugh issue for me, although I still hit my muscle memory habits.

Unfortunately the upgrade also cleared all previously saved form data for me. I know it is cache stuff which could be gone at any time, but still a little glitch in the experience.

I updated and didn't realise a big release was on its way. I hate it and I've swapped to Brave

I did the opposite. I switched from Brave to Firefox. Only reason I was running Brave till now was because Firefox's UI felt slow to me. Now I can get real ublock origin and the UI isn't laggy. Even if the features are less than before, it can replace Brave browser whose only use was "A browser that isn't Google Chrome, having adblocking support".

Is that why Mozilla sent notification spam to announce that the app will be updated?

Correct, since a large number of extensions will stop working, etc... might as well tell people they're in for pain?

The only notification I got was that the app was updated.

Same for me, I would've been happy if they atleast allowed sideloading extensions by flipping a switch in about:config.

Reply on news.ycombinator.com