Hacker News Reader

Comment by f055 | Hacker News Reader

I find it interesting that the fact so many comments mention config flags mean not many here are actually using the same Firefox, and not even close to the default one used by the 99% of its users.

View on HN

I don't even know how default Firefox looks like. For all the changes going on, my configuration seems to persist fine.

Being highly user-configurable used to be the best reason to use Firefox, but Mozilla seem to be have become openly hostile towards that flexibility in recent times, even when it made the experience far worse for lots of users. These days, being probably more trustworthy when it comes to privacy and phoning home is more-or-less the only benefit Firefox has compared to just using Chrome/Safari/Edge that comes out of the box with your device.

Supporting all that flexibility is challenging for development, and could cause issues for the end user.

I can imagine that people over-customize Firefox, forget about their changes, then have bugs, finally abandon Firefox thinking it's unstable.

No doubt some people do. On the other hand, that flexibility was Firefox's main selling point.

Even today, there are several things I used to use on a frequent basis that are missing or limited because the Quantum changes broke so many add-ons. The equivalent functionality hasn't been or in some cases can't be implemented by add-ons using the new APIs instead.

I'm a programmer. I understand the technical arguments for why they wanted to make the internal changes. But when you are dumping your #1 competitive advantage because of technical details, it's probably time to start asking some tough questions.

Being fair, they dumped their #1 competitive advantage in order to relieve their #1 competitive disadvantage - keeping XUL made desperately needed performance improvements impossible to implement. Calling that "technical details" rather understates the case.

FWIW, I didn't find Firefox noticeably slow on any device I was using before, nor did I notice any big speed up after those changes. I understand that others had varying experiences, but unfortunately I can't relate to them at all myself.

The only time pre-e10s Firefox performed well was anecdotally, in threads relating to pre-e10s Firefox not performing well.

That might be your experience; I have no reason to doubt you. Mine is that no-one I know personally ever expressed any dissatisfaction with the speed of Firefox before anyway, and the only time people complain about the performance of pre-e10s Firefox is when they're trying to justify all the broken useful functionality that was caused by the change.

You're right that people weren't all that likely to complain loudly about Firefox's pre-e10s performance. Much more often, they just quietly switched to Chrome.

You're making a significant assumption there. Let's not pretend that people only switch to Chrome because of some perception about performance. If that were the main or only reason, shouldn't lots of people have switched to Firefox later, if Quantum made it so much faster?

Who's pretending anything of the sort? I hope it doesn't really need to be explained that saying this is a reason why some people switched to Chrome doesn't preclude other people doing it for other reasons, or that once there, they're likely to stick instead of going through the headache of switching back again.

None of this needs to be taken on faith. There were Mozillians here and elsewhere talking at length around the time of the change about why it was necessary, and maintaining market share against Webkit browsers and especially Chrome was a major reason. I don't care enough to spend any of my Friday night digging up those references from a couple of years ago, but they shouldn't take all that much finding if you do.

Firefox comes pretty much broken out of the box if you don't use config flags.

Hmm? The only pref I change is "image.animation_mode=none" to disable animated GIFs and videos, and I can understand why they don't set that to the default. What do you find is broken about Firefox's default state? Do you really think most users feel the same way you do?

>What do you find is broken about Firefox's default state?

Firefox uses a self-made fullscreen implementation by default on macOS, which is buggy and causes menubar issues. Firefox has code to use the macOS native fullscreen feature, but it's actually hidden behind "full-screen-api.macos-native-full-screen" about:config flag. So if you're on macOS and want to fullscreen something on another monitor (and actually have it work proper), you need to figure out that the fix is hidden there.

Surely you can understand that an issue in a platform-specific implementation of a fairly obscure feature doesn't really support your assertion that it is "broken out of the box". It's likely that config option is not set by default because it's known to be buggy in some other way. You can read some discussion of this setting here, which has seen some activity in the past year, so it is being worked on: https://bugzilla.mozilla.org/show_bug.cgi?id=1403085

A fairly obscure feature, what? Fullscreen is a fairly obscure feature?

Uh, this was the first I'd heard about this setting, so I went ahead and tried it.

I don't know what's happening for you, but when I toggle that setting, full-screen becomes quite broken: https://i.ibb.co/7tJYHPf/Screen-Shot-2020-04-10-at-9-30-25-A... (Note all the empty space at the top and bottom.) Breaks the fullscreen animation too.

I don't have problems with the standard full screen in Firefox—I never knew it wasn't using the native mac method—although I only have one monitor.

The one big problem I do have with Firefox on Mac is that it intercepts the system keyboard shortcuts. So if you set up custom keyboard shortcuts in System Preferences for menu items, they won't work in Firefox.

That's not happening to me with full-screen-api.macos-native-full-screen and single screen.

Still though, you can probably understand now why it's not the default.

I can't actually for a myriad of reasons. Not only has the native fullscreen functionality been around for years (eight to be exact), it's pretty mature. Other applications, like mpv, have migrated over to the native fullscreen functionality just fine. And as I already mentioned, the current Firefox implementation is just downright broken.

> And as I already mentioned, the current Firefox implementation is just downright broken.

Yes, and as others mentioned, the "native" implementation in Firefox is also downright broken. Without knowing which one is more broken, there doesn't seem to be a strong reason to change the default. I understand that the native one works better for you, but you're not the only user.

The strong reason to change the default is that only one of them is supported on macOS and it's the one that's not default.

Not to mention the fact that Firefox already uses the native fullscreen functionality. It just doesn't use it fully and consistently.

And what's with the quotes around native?

I'm trying to indicate that I'm using your words. I have roughly zero experience with macos, and I don't know what native fullscreen means. In windows, I don't think browsers use any separate apis for rendering fullscreen, so I just don't really understand it.

> one of them is supported on macOS

Assuming "support" means it actually works, there seems to be conflicting information about which one is actually supported.

Support means that Apple has APIs for it.

https://developer.apple.com/documentation/appkit/nswindowdel...

It seems that in some cases, the way Firefox is calling those APIs is incorrect, or causes undesirable behavior.

Who the fuck knows what's happening in Wowfunhappy's screenshot. I doubt it's because of the native fullscreen because it's something that's been a part of macOS for the past 8 years. And again, Firefox is already using the native fullscreen functionality in some cases. The issue is that some of the time it's not actually a fullscreen application and is merely just pretending to be one.

My own pet feature, and prior to 75, the only config I changed, is "middlemouse.contentLoadURL=true" -- the default middle button action is some goofy scrolling feature copied from IE; much more useful to use it to load links.

Thank you for mentioning this. Even I've never really understood the Firefox dev team's rationale behind making the DEFAULT action of middle mouse click of that of auto-scrolling.

In my years of using the web, never have I ever come across a scenario where I'd need middle-mouse-click triggered auto-scrolling.

Oh man, the middle click autoscrolling is one of my favorite features. To scroll all over a huge page, you just middle-click once and nudge your hand a couple pixels up and down. Boom, free scrolling with just tiny nudges of your hand. It's probably how I do 90% of my scrolling in Firefox. (No opinion on the default setting; it's easily findable in the Settings dialog.)

It's working fine for me.

That is quite subjective...

Using old, broken system APIs without going into about:config is not really that subjective.

I've been using firefox for over ages and I've pretty much never changed a config flag. No complaints.

Reply on news.ycombinator.com