Hacker News Reader

Comment by pricechild | Hacker News Reader

I've been following https://utcc.utoronto.ca/~cks/space/blog/web/Firefox56AddonW...

It's surprised me the amount some people rely on addons for their workflow and the amount of effort they will be going through to resist this change come 57.

For my usage, I've found Firefox Klar/Focus (which I believe already include the speed improvements from 57+?) on Android and have been extremely impressed.

View on HN

I'll be doing the same, staying on 56 for quite awhile. This transition has been very poorly managed by Mozilla. I understand the advantages to Quantum but they have left a very large percentage of their extension developers hanging. Web Extensions is not mature enough for many of the large extensions to be ported but despite that they are proceeding with this transition anyhow.

If I was an affected extension developer I would be walking away now. My user base is about to crumble and the best I can do is maybe offer a pale version of my extension in the new system along with a bunch of excuses out of my control. If perhaps WE does get the features I need, why go to the work to re-write? My users will be long gone.

Quantum is great, don't get me wrong. But the new APIs are too immature to be shutting down a huge number of mature extensions and the users that rely on them. It really makes me question the leadership decisions happening at Mozilla in general.

Not sure if you're aware of the Extended Support Release of FF but it seems better than staying on 56: Firefox 52 ESR will continue getting security patches up through June 2018.

I suspect I'll be on 56 for a very long time as well. I ran on nightlies for many, many years but that ended at 57. At a bare minimum, noScript must be shown to be fully and consistently functional. But I have quite a few other extensions that appear to have been dead ended - self-destructing cookies and session manager are just two.

I think mozilla failed to understand that extensions were the prime differentiator and that a faster Chrome is far down the list of requirements.

The good of the many outweighs the good of the few.

IMO Gutting your developer base is seldom good.

The web developer base is far more interesting and relevant to Mozilla's mission than the extension developer base, and Quantum is the only shot Mozilla has at regaining mindshare there. Looks like it's working, though.

Here's the thing though. Developers understand the technical differences between Chrome and Firefox. To 95% of general users they don't see why they are different. They look about the same and hey, I get Chrome offered to me every time I visit Google so why not use that? The average user doesn't see the difference that gives them any reason to want Firefox. Now the extensions base will be essentially the same, stripping away those that were truly unique. The extensive UI control is disappearing. Firefox (IMO) is slowly stripping away the things that make it unique to 90% of its user base who just see the UI.

Tgis is not true The extension base is not the same. Firefox 57 will have tree style tabs, for example.

What developer base? 99% of new addons are Chrome first/only (like uBlock, uMatrix). And most of the legacy addons were abandoned long ago.

I use several legacy addons and most of those haven't been abandoned? They are about to be though.

Indeed. Just look at PocketPC vs Mobile/Phone/whatever.

Of all companies one would have thought Microsoft understood the value of backwards compatibility...

Speaking of pale, i guess there is always Pale Moon...

Unfortunately that's also where I'll stop during at least a few months after Firefox 57 is released. Quantum is ultra fast, but my extensions are more important. Greasemonkey and Stylish, for example, have equivalent WebExtensions, but they're not as polished. Same with DownThemAll.

I think WebExtensions should have matured a bit longer, with Quantum, electrolysis, and Photon being made available for all with the old extensions.

You can't have your cake and eat it too. I'm fairly sure that if the Firefox developer team could support XUL with the new internals, they would (just look at the insane backlash for a source of motivation). But they didn't, and there must be a reason. Likely it's due to how integrated XUL is with the internals of Firefox and that couldn't gel with the Rust rewrites.

Oh I totally understand why Mozilla did what they did. It was the logical step to make Chrome addons available on Firefox, abandon the structure XUL imposed, and migrate to Servo-based components. But while they promised a better WebExtension standard (which I defended at the time), they didn't deliver. They needed more time to improve the API and make sure the most used extensions were ready[1]. Developers needed more time too to totally change the architecture of their addons.

And today, they're not ready. And they won't be in six weeks either.

[1] https://arewewebextensionsyet.com/

Edit: [1] answers "Are we WebExtensions yet?" as Yes. I don't agree.

Well they gotta make the cut at some point. If Firefox waits forever then Chrome will just eat up all their users.

Don't listen to the stragglers who want all the old Firefox behaviors supported forever. The wast majority of users don't care about Tree style tabs and vim style keyboard navigation or what XUL even is. They just want something snappy and fast like Chrome. That's how you get users to use your browser. That's how Firefox became popular in the first place. That's how any browser ever became popular. Yes even IE was faster than Netscape at one point, maybe because of nefarious practices by MS, but it was faster, and that's what the users wanted. It's all about the speed.

It's because of the killer combo of increased sandboxing, multi-process architecture and performance constraints.

I think they couldn't keep everything together with the degree of access old plugins had to Firefox internals. So at some point they had to rip off the band aid.

Multiprocess already shipped. One of the reasons extension developers are frustrated is that Mozilla told them to make their XUL extensions multiprocess-compatible and then told them they'd have to throw that work away.

Notice how I carefully avoided your strawman. There were several reasons for their action. I listed what I believe are the 3 main reasons.

They did have the choice to not ship the increased sandboxing, multi-process architecture and performance constraints until three months after the new extension APIs were ready. They didn't because they thought the trade-off wasn't worth it.

They really couldn't. People don't do things until pressured. Look what happened with Python 3. It took 10 years for it to reach Python 2 parity in terms of packages...

The problem with WebExtensions has been Mozilla not providing APIs, not extension developers dragging their feet.

I really do miss the slow and methodical Mozilla that was not out to chase the latest Chrome fad...

Absolutely. I won't be upgrading and I'm already looking for alternatives, as long as crucial parts of my workflow (tree tabs, session manager) have no adequate replacement.

The ugly redesign is also a thorn in my side.

https://addons.mozilla.org/en-US/firefox/addon/tree-style-ta... has a 'Compatible with Firefox 57' badge, and I've been using it with no issues.

I've always been satisfied with the built in session management (restore last session on open) and don't have any feedback on the other.

Not only is the new Tree Style Tabs a step backwards from the old, I'm also not running that addon, but this one: https://addons.mozilla.org/en-US/firefox/addon/tab-tree/ I find it lighter and more responsive.

As for the session manager, I need more features: saving and loading several sessions, etc.

Agreed. The inability to customize button placement and the enforcement of a lowest-common-denominator default placement is painful. Unfortunately, all the other major alternatives are just as bad or worse.

Unsure, but doesn't the 'Customize' option give you exactly the ability to customize button placement?

There are some things you could change that will be going away - such as having separate (and equal sized) forward and back buttons and separating them from the location field; breaking out the stop and reload buttons; changing button appearance; and the like that depended on XUL and are going away with the new "Proton" UI.

Even now adjustments I used to be able to make in seconds by installing an alternative theme now take an hour of wrestling with the Classic Theme Restorer extension to restore and unbreak with each version upgrade.

The session management piece is my big sticking point. I've come to rely on Session Manager, but the developer has already stated it will not be ported because necessary APIs are not available through WebExtensions. Since Chrome has better options (mostly through Session Buddy, which isn't nearly as good as Session Manager, but it'll have to do) I guess I'll be going to Chrome full time.

How's the redesign ugly?

Well, I guess it's subjective, but I find it uglier than the previous design. The obsession with excessive flatness and the deterioration of user experience in favour of "prettier" visuals never sat well with me.

Tree style trends had already been ported

The problem is that electrolysis alone already kept breaking hundreds of addons, so that wouldn't have worked. The only way to prevent breakage, from what I've gathered, would have been to just keep Firefox like it is and not go on with Quantum, which really wouldn't have been great either. And of course, with more users testing WebExtension addons, those new addons hsould mature a lot faster than before.

Oh shit. I had not been following this and I didn't know that DownThemAll wasn't going to work. Now I'm sad :(

At least for the time being, you can use legacy extensions, by enabling the right option [1] in about:config.

[1] extensions.legacy.enabled -> true

You can enable legacy extensions but this is not the same as being able to use them. Many significant and important-to-me legacy extensions (such as FireGestures) are currently non-functional on Firefox Nightly and have been for some time. I can only assume that Mozilla has decided that they don't care about keeping any legacy extensions working other than their own (they have some).

For people who are uber concerned about keeping v56 without losing your extensions as part of the 57 update, you can turn off auto update for Firefox:

Go to about:preferences In Firefox Updates choose "Never Check for Updates"

Breathe easy!

I know this is not recommended, but I love my XUL based extensions a little too much (Tab Groups primary among those) to give them up so easily.

If this is a problem for you, please switch to 52 ESR and use it until June 2018, and file bug reports against those legacy extensions. Disabling updates for a piece of software that is used to run untrusted code is a terrible piece of advice.

Unfortunately, a profile backwards-incompatibility was added in FF55, IIRC, so if you're already on 56 you can't downgrade to 52.

A forum thread[1] suggests it may be possible to use the profile refresh functionality to recover most of the profile. Only problem is that it does not preserve extensions or settings.

[1]: http://forums.mozillazine.org/viewtopic.php?f=38&t=3032457

What does it preserve, then? (Bookmarks? That's the easy thing to restore. Transfer to another browser then transfer back.)

This link should clarify what is preserved by the refresh.

https://support.mozilla.org/en-US/kb/refresh-firefox-reset-a...

I believe this has something to do with moving favicons out of places.sqlite and into favicons.sqlite. This completely destroyed my favicons (thousands) SeaMonkey's Bookmark Manager (and search history going back years).

If Firefox Sync can sync the important bits of your profile, that may be a sufficient solution to this problem?

TBH, Firefox has made great strides in the last 2 releases (55/56) in terms of pure speed and snappiness, which is why the suggestion to keep the best of both worlds. FFx 52 pales performance-wise in comparison with FFx 56. I understand what you are saying and hope that people (on HN of all places) are smart enough to understand the trade-off being made.

I have (sadly) taken this step.

I'm also saddened that a transition with such a large change in functionality for end users didn't warrant syncing up with a Long Term Support version.

There's going to be a bit of a waiting game to see which extensions get ported and with what percentage of their original functionality retained.

Then it will be decision time.

> which I believe already include the speed improvements from 57+?

Interestingly Firefox Focus uses WebKit (/Blink?), since that's what Android's webview provides. Firefox for Android should get the improvements around version 58 or 59.

Also good to note, for people planning to stick with an older Firefox for as long as they can: Firefox 52 is an Extended Support Release, and as such will actually be actively supported until June next year. This will give WebExtensions some extra time to mature for more use cases: https://www.mozilla.org/en-US/firefox/organizations/

Yes, Focus uses WebView. On newer Android versions it's actually provided directly by Chrome (Android 6+ IIRC), on older Android's (Android 5) it's a separate package that's still built from the same sources. (Android 4 is more complicated but also not supported by Focus for that reason.)

I know how he feels. Been sitting on 45 for ages now because of their hamfisted move to GTK3.

As for Focus/Klar, it is just a shell. The engine used is webview, and the oldest supported Android is 5.0 (because of an API and because that was when Webview was made updateable outside of Android proper).

>By 'as long as possible', I mean well beyond when Firefox 56 is officially supported; I hope to keep using it until it actively breaks or is too alarmingly insecure for me.

or you know, use ESR and get 6 more months of support/bugfixes.

Reply on news.ycombinator.com