For instance, my bank (Lloyds) has a login screen where you have to select letters from three long drop down lists containing all letters and numbers.

With Safari or Chrome, I can select the dropdown (by clicking it or navigating there using the keyboard) and then enter the letter or number on the keyboard.

With Firefox, I have to click on the dropdown, scroll down looking for the right letter/number and then select it with the mouse. It takes at least 10 times as long as with other browsers. So I'm not using Firefox for that site.

Of all the interactive sites I use regularly about 10% have some sort of bug when visited with Firefox. And this is getting worse at an alarming rate.

I think many mainstream sites not created by actual tech companies have simply stopped testing with Firefox.

(I don't use ad blockers, just browser default settings)

As inconvenient as powering through this bug might be, switching to Chrome to use this site just reinforces their choice. Perhaps emailing them to report the issue might also be beneficial for all?

I changed bank - I had an account with a bank (RBS?) in the late 90s / early 00s, couldn't access it from linux, yet I could access HSBC, so RBS lost my custom.

My current bank (Smile) works in Firefox on linux, so that's fine. If it didn't, they'll be losing my custom.

Sometimes I am astonished by the life other people apparently lead. For example when they'd invest the energy to switch their bank (and all that is attached to that) over using a different browser to access a singular website.

I do not understand. It boggles my mind, in the most curious way.

Sometimes the way it fails gives the impression of a less-than-comfortable level of competence with technology. I also almost switched banks when my first was acquired by another, its webui was so much worse.

Please see the sibling thread (to your comment) for why I'm not going to do that. This "glitch" is probably intentional.

Also, I have actually tried to communicate with that bank about a far more important issue. It was frustrating and pointless, starting with the fact that they don't have a customer support email address.

As it’s a bank I suspect their devs have to use IE6 and have recently upgraded to IE8 and thus change in behaviour.

This used to work on Firefox as well with Lloyds, I don't know what changed but it stopped working about a year ago.

I think it's some sort of anti-keylogger pattern that Lloyds are using. The select options all look like this:

  <option value="&amp;nbsp;c">&nbsp;c</option>

I suspect the nbsp is what blocks the usual keyboard behaviour on Firefox but not on Chrome.

This is quite strange, because it could mean that the anti-keylogger feature only works as intended with Firefox but not with Chrome.

Interesting, that could be something not too complicated to change on Firefox, I'll try to contribute when I'll have a bit more time. At least that's the first time I see something I could realistically improve.

I'm not sure whether Firefox should actually be changed. It looks like Firefox may be conforming to the standard and Lloyds devs may have intended to use that standardised behaviour to thwart keyloggers.

Only they seem to have forgotten to test whether it actually does block keyloggers, and it only does so in Firefox.

So my grand example for a website not working well in Firefox may be just the opposite - Firefox saving my money from getting siphoned off by thiefs. Very bad choice of example on my part! :-)

Here's what the HTML 5 spec says on the <option> element's text content:

The text IDL attribute, on getting, must return the result of stripping and collapsing ASCII whitespace from the concatenation of data of all the Text node descendants of the option element, in tree order, excluding any that are descendants of descendants of the option element that are themselves script or SVG script elements.

https://html.spec.whatwg.org/multipage/form-elements.html#th...

nbsp is not ASCII whitespace and so it should not be stripped. I can't seem to find anything specific on the autocomplete behaviour in this case, but arguably autocomplete should use the text content without further normalisation. Perhaps this is where Chrome and Safari devs would disagree.

In any event, this raises a bigger question. What if a site is actually broken in Firefox, but it's not Firefox's fault? Should Firefox then be changed to non-conforming behaviour, aping any Chrome bugs? That would put a huge question mark on all the work they put into making their own rendering engine instead of just using Blink.

> In any event, this raises a bigger question. What if a site is actually broken in Firefox, but it's not Firefox's fault? Should Firefox then be changed to non-conforming behaviour, aping any Chrome bugs?

I can remember one case of such a thing happening before: css box-sizing.

Before the attribute existed, the spec said content-box was correct, and all browsers except IE implemented it that way. IE implemented border-box, going against spec, but it turned out to make so much more sense it was eventually added to CSS, and nowadays is the recommended way to go (but still not the default value).

Was going to say much the same thing. I've been a full-time Firefox user for 10+ years. I've also worked on web apps and tested against Firefox, Chrome, IE (and later Edge), for 10+ years.

In all that time, I don't think I've ever seen a Firefox-specific bug or layout issue.

I saw them on IE all the time (obviously), sometimes in Edge, rarely in Chrome, and never in Firefox.

Me either. When something doesn't work it's almost always because of Privacy Badger. If I'm really curious I'll whitelist it to see if that does the trick. Otherwise I might report the bug to the site, but typically there's just not enough time in my life to waste on broken web sites.

People hold on to this belief that Google is infallible because they're rich and they put together a marketing campaign some years back to convince people that only smart people work there. I believe the truth is exactly the opposite.

Also a full time Firefox user.

Another bug that I believe is actually Firefox doing the right thing: if you make an event handler function but give it no parameters, Firefox will have no local named event. Chrome will create an event variable for you. Caused several "Firefox bugs" in our app.

The behaviour you're referring to is actually a global property `window.event` [0]. Although it's part of the DOM spec, it's only there for historical reasons and should definitely be avoided.

[0]: https://developer.mozilla.org/en-US/docs/Web/API/Window/even...

Also full time Firefox user. Most of my issues are either blocking related so that is on me or apparently sites bugs. At work I use chrome for beyondtrust because I cannot get ff to automatically download and launch rdp sessions. Apparently the content type is wrong in downloaded file so ff will not associate an application with the file so adds several steps to launching rdp sessions. Easier to keep chrome off to the side around for that task.

The other issue is that it does not use Windows ca store for certificates. When company uses internal certificate authority for internal websites this requires users to manually update the browser cert store. Chrome and Edge will use windows ca store on windows. My it dept will not automate this but will manage the windows ca store.

>The other issue is that it does not use Windows ca store for certificates

By default. But can be configured to use the cert store [0].

> Apparently the content type is wrong in downloaded file so ff will not associate an application with the file

If it is consistent that shouldn't be a real problem.

[0] https://wiki.mozilla.org/CA/AddRootToFirefox

Skype Web is an example from the top of my head, but I don't run into them a lot either (or don't notice that they are only broken in Firefox, because I rarely care enough to try another browser).

I do see quite a few sites that apparently never have been tested with ublock or privacy badger, but they break on all browsers where those are enabled.

Skype web works if you switch user agent

For example's sake, try opening payment/billing setting pages of Google products. For example, pay.google.com doesn't load on FF latest on MacOsX latest.

Thanks, though it works for me on Linux and Firefox 68.0.2.

Loads for me - latest Safari, latest macOS. Maybe it's a specific page?

I know for example I can use FF everywhere but specific Google apps (new sites) don't work well in FF, though that's improved lately as well.

Just now I came across an info table [0] for some game that should allow sorting by columns, but doesn't work on my Firefox, only Chrome. Yeah, its just some fan site for a game and thankfully doesn't happen that often to me, yet. But I feel like those many smaller sites, that I'd prefer to exist over all content centralized on a few giant platforms, are especially susceptible to not testing on many clients. And yes, I'll try to contact the author, but probably hadn't if not for this discussion today.

[0] https://oni-assistant.com/database/elements

Interesting, this is the only example given that I was able to confirm does not work. I'm glad this discussion is making you consider contacting the author. I feel like we've dug ourselves into an IE-shaped hole all over again and this kind of grassroots action is very important for getting out.

As an aside, immediately upon seeing the website, I had a split-second thought of the content reminding me of Don't Starve. Amusingly, the website turns out to be about Oxygen Not Included, another Klei game. :-)

Maybe it’s the combination with ad blockers, but there are a lot of times I get stucked in the payment part.

And then I have to redo all the ordering process on Chromium.

It for long failed on Visa 3D secure (so not website dependant) until I found why with Badger blocking it.

But often when the payment system is included in the website it fails (and that’s really annoying)

Blocking fingerprinting can cause anti-fraud to think you’re a bot.

As it turns out in recent news, credit card companies sell weakly anonymized information about cardholder purchases so ¯\_(ツ)_/¯

I solve this problem by having a clean profile with only uBlock that I can open up when I want to make sure a purchase is going to work. uMatrix has a glitch where even when you enable something it sometimes won't work / load even when you refresh the page multiple times, but I don't think I've encountered problems on many pages with just uBlock.

I have never had payment fail in Firefox.

One of the two bike registry websites only worked in Chrome, with some JS error in Firefox. Some function was always defined in Chrome but never in Firefox by the time somescript tried to use it. Wasn't addon-related. Seems fixed now.

I had trouble with imgur on firefox mobile. it doesn't allow uploads until you request the desktop version (which fakes another useragent string).

As a (mostly) full-time firefox user, I commonly find that sites have bugs and some are unusable in FF. What I find fascinating is that every time I mention this on HN or elsewhere, somebody also chimes in with this "doesn't happen to me" message and demands to know which sites.

Every damn time.

You mentioned your anecdote multiple times, but expect others to only mention theirs once? Not to mention it was likely a different person responding to you each time. It seems like a very natural pattern, not surprising.

It seems like this is a good thing since at least one other user said this thread prompted him to contact the site developer.

Sounds like tracking protection blocks some elements.