Firefox 77 won't connect to non-domain address bar entries with periods anymore (will search instead)
Mozilla plans to change the processing of address bar entries with periods (dots) that are not domain names when it releases Firefox 77. Current versions of Firefox attempt to connect to entries with periods when they are entered in the address bar of the browser. If you enter a phrase like "my.bat" or "console.log", you will get a "Hmm. Weâ€™re having trouble finding that site." in Firefox currently as Firefox adds https:// in front of the query because it interprets the input as being a domain name that it should connect to. Note that Firefox interprets the phrase as a search term if it contains spaces.
The error is quite common for filename searches. I prepend ? to the query whenever I need to run a search phrase with a period to make Firefox search for the term instead of it attempting to connect to it as it considers the term a domain.
Starting in Firefox 77, Firefox uses a different logic when it comes to address bar entries that contain periods. Basically, if the term is not a domain, e.g. ghacks.net, it handles the term as a search. Means: Firefox maintains a list of top level domains (using the public suffix lists); if the typed string contains a period it checks it against that list to determine whether it should try to connect to it or run a search instead.
Firefox users who run the cutting edge Nightly version of the web browser will notice the change already. A search for console.log runsÂ a search in the latest version instead of connecting to it.
Administrators and users may add custom extensions, those not found in the public suffix listing, in the following way to force Firefox to connect to these sites:
- Load about:config in the Firefox address bar.
- Confirm that you will be careful.
- Type browser.fixup.domainwhitelist.DOMAIN.EXTENSION
- Make sure you replace DOMAIN and EXTENSION with the values that you require, e.g.Â browser.fixup.domainwhitelist.example.local
- Make it a Boolean
- Set its value to True.
You may also type the protocol when you need to access locally available sites
Firefox runs the search using the default search engine in that case. It is still possible to prepend ? to the query to make sure that Firefox runs a search for it. If you type ?ghacks.net, Firefox will run a search for the domain name instead of connecting to it.
Google Chrome, and other Chromium-based browsers, use the same technique to determine whether a user's input should be resolved or redirected to the search provider configured in the web browser.
Mozilla plans to release Firefox 77 on June 2, 2020.
This is just another example of why it’s best to have a separate search box. The address bar is for addresses, and the search box is for searches. When you try to merge the two, things go wrong.
@Matthew: they won’t anymore in FF77.
I’m a full believer in keeping the two separate. The search box is also a nice way to keep track of what you have been searching, or what window was being used for what (for history tracking).
So, no more web site on your intranet? That’s … disturbing.
Btw, I had a look at the list, I don’t understand why they list top level names and also a few names below: they list .ac, why bother listing com.ac, edu.ac, etc. it’s already covered?
You may add custom domains/extensions to a whitelist. I added the info to the article.
If it’s anything like the handling of single words both Chrome and Firefox have been doing for ages (i.e. searching for that word instead of going to an intranet site) I’m pretty sure all you have to do is append a ‘/’ to the URL to load it anyway instead of searching.
It’s not already covered by .ac. Purpose of public suffix list is not what you think. https://en.wikipedia.org/wiki/Public_Suffix_List
I wonder if this is going to be a problem for common internal FQDNs, such as ones ending in .local.
They probably thought of that, along with .lan
For more specific situations, there should be a user definable whitelist
I wonder if they’re going to consider internal domain names, such as .local.
@bsod: You can still append a slash to get to the internal DNS name.
This is amazing. I hate this so much when doing day to day IT work. GG Firefox
This seems like a good idea, although I’m not sure why this was a priority for their development team.
It wasn’t, the related bug ticket was open for 6 years.
I installed Nightly 77.0a1, and this issue doesnt happen for me.
Can this be turned off?
Why are we constantly redesigning things to send potentially sensitive data to people who have no business collecting it? Examples are searching the web in the OS’s search function by default (without users opting into this), ISPs hijacking your 404 error pages to send you to shady looking advertising pages, and now, the address bar not just behaving like an address bar, the way that it always has.
@Tom – To avoid Firefox searching from the address bar without an explicit cue, set keyword.enabled to false in about:config.
Will it also affect IP addresses? Like http://192.168.0.1
That’s an “un-routeable” local address. Lots of dots, no domain name – like ethernet was designed to work with.
I suspect this is being mis-reported; it would make the Firefox web browser incompatible with a basic ethernet network. You would not be able to administer a local media server or local network storage except from the command line.
If you dropped to a command prompt, your computer would be able to ping and tracert to it. The PC would have its mac address – the real address of the device, Normal network traffic would build up a local address table of everything on your LAN segment.
That’s possible but so 20th century. For more than two decades, every maker of consumer network equipment has built their stuff to be administered through a web interface, using an IPv4 address. Only a few home routers have pseudo-domain names prior to their being initially configured.
The idea of a browser forcing the use of domain names is ridiculous. A web browser block the attempt by forcing a DNS lookup is absurd.
Eh, sounds like something that needs to be fixed with an addon when mozilla have “fixed” it.
There is a searchbox for a reason. Like someone said. Way easier to split the adres bar and the search bar both for usability and codewise.
How will they handle an address like https://xxx.xxx.xxx.xxx?
DNS was supposed to make things easier, but with ever-growing length domain names, that is not always the case. Administering a local gateway or router for instance is done without a domain name.
It just wouldn’t make sense to handle it any differently, unless there were some ulterior motive to do with forcing or monitoring dns (more granularly than an ip addy can).
Sensibly I think no change.. but there is that nagging doubt these days.
Doesn’t anyone get it? mozilla knows what you should be doing, you on the other hand do not!
Complete search bar removal incoming.
Will Firefox still be able to connect to a modem or router config interface page?
Yeah I’d imagine that it’d add the local domain to whatever whitelist exists
It would make most sense and be the easiest implementation of this feature for Firefox to compare the whitelist, check for a responses from the potential domain and if no server is found, instead of showing the doom-and-gloom screen, fallback on the search screen instead.
this is great. I’ve resorted to having to quote strings like this in order to avoid having them processed, but that’s not ideal. In the meantime thanks for the tip of using a question mark character
If you set about:config preference browser.fixup.dns_first_for_single_words to true, Firefox will try to resolve every query first as URL using DNS servers and only fall back to search if the resolution wasn’t successful, resulting in similar behaviour as before.
Martin Brinkmann, could you add that to the article? It is quite useful for users who want old behaviour.
1. Use an old, reliable ESR
2. Install privoxy, add the instruction to change User Agent String to latest bloatest Firefox.
4. Enjoy unbloated browsing.
I like this. Have frequently run into the problem when searching for a file name so nice to see a smart, automatic solution. That said, I really wish I had read about Martin’s ? workaround years ago.
I use a search engine as my home and new tab pages. Searches can be entered into the urlbar (it’s still called that) but it spits out the default search engine result. Absolutely cannot stand that stupid bouncing bar in any browser and that behavior is always disabled along with search suggestions.
I know what I’m looking for; search engine results are almost the same as, usually better, than bouncing suggestions.
Mozilla’s strategy, if there is one, for FF is befuddling. In one move, they bloat it with “Who cares?” features; in the next, functionality is removed. It’s evolving into a whale that can’t swim.
So, this is irrelevant to me unless IP v4 addresses don’t work. Maybe Mozilla is being threatened by the IPver6 Illuminati or something. The Group of Six. Bizarre.
If you activate the search bar separately from the address bar, it’s obvious that you don’t want to mix up searches with addresses. Why do so many settings have to be made for everything to work the way you want it to?
I do not agree with people who prefer two bars. It is aesthetically very ugly. Doing two things in one bar is much better when it comes to agility in doing things.
It is so ugly that I have even tried to remove the search bar from windows, having two divided bars is alkaline. You have to innovate and if you do not like it, it is better that you create your own browser and use it to your liking.