Checkboxes on Linux
I wonder if it will be able to render checkboxes as something other than 1x1 pixel in this version?
Mozilla has released Firefox 78 with a new Protections Dashboard and a bunch of updates for web developers. This is also the last supported version of Firefox for macOS El Capitan (10.11) and earlier. Firefox is on a "rapid release plan", which means a new version every four to five weeks. This means that major new features …
"Users often have an instinct to uninstall and reinstall when an application is not working correctly. Anticipating this, Mozilla has put a Refresh button into the uninstaller, which resets the settings. Refresh may well work better than reinstall..."
If only. No website owrked today since the update. Tried refresh to no avail. Uninstalled and reinstalled and hey presto everything is working again.
Well, regexes really count as DSLs (domain specific languages), which is why the two browsers have been using the same approach for years. The change presumably includes some kind of binding so that the V8 engine can be called directly and doesn't have to be ported.
The same does not necessarily apply to the JS runtime. But I do think we will start to move towards consolidation other areas such as HTML, CSS and JS parsing, areas where Mozilla's use of Rust might have advantages.
It's very annoying that the only browsers that handle WebRTC in an acceptable way are Chromium/Chrome & derivatives and Vivaldi.
It's one of the simplest way to set up decent videoconferencing, but it sucks in Firefox.
Update: it seems to get decent test results at https://test.webrtc.org/ (apart from the whinging about lack of IPv6 connectivity, but that's because I disabled it as I don't yet know if that is flatout public, or is still at least behind the router NAT).
Well, there is hope.
If you have IPv6 working on your router - presumably by default on an ISP supplied router - then it's pretty well a given that it will be using a statefull firewall. The default for the firewall will be to block all inbound connections - so you are as safe behind IPv6 as you are behind IPv4 and NAT. In some ways you will be safer as the default these days is for end user devices to change addresses every so often - so an attacker would be targeting something that's using just a handful of addresses out of a block of 2^64 addresses (itself only part of the wider internet of 2^128 addresses.) In principle, your device could use a different address for every outbound connection if it wanted to and thus present a "never in the same place twice" moving target - but nothing has got to that stage yet as it would break too many things (such as websites that expect relatively static client IPs).
In any case, probably the largest attack vector these days isn't from the outside working in, it's from a beachhead (via a compromised device) on the inside where it has full access and a full view of what's on the network. [sarcasm]All these devices that maintain outbound connections to all sorts of stuff they don't tell you about - yeah, that's a really good idea and not open to abuse [/sarcasm]
Disappointed at the slow pace of adding safe to implement API's to WebExtensions to add features the legacy add-ons could do. Most developers got bored of waiting, and gave up waiting for missing features needed to port.
In the meantime, I switched to Chrome: It runs faster on my hardware, and with much less customisation to streamline the workflow, there wasn't really an advantage to stick with Firefox.