Comment by Jason Knight on "How do you do automation testing for requesting browser permissions?"

I'm not sure what it is you think you'd be automating -- or even quite what it is you are doing that would be "cumbersome" or "time consuming" unless you're wasting time on scripttardery that has ZERO business on a WEBSITE in the first place... and it at least sounds like you're talking about websites since you're talking different browser engines.

But the core of any new features or scripting is features detection, which in most cases shouldn't be anything more than a single line "if" statement unless you're going to polyfill or implement a fallback.

What you don't want is to start "browser sniffing" unless there's really no other way, and by "no other way" I mean legacy IE.

Though your picture has one of the poster-children for things that just piss users off and therein has zero damned business on websites: push notifications that cascade to the OS. SCREW THAT! Number one thing I have people asking me how to disable when some jackass website tricks them into accepting them.

Bottom line though, if you're developing a WEBSITE your job is to deliver content to users in as accessible a manner as possible. That means semantic markup, separation of presentation from content, progressive enhancement so that it gracefully degrades... and to that end as the old saying goes:

If you cannot make a fully functional website without JavaScript FIRST, you likely have zero damned business adding scripting to it!

That's not some cutesy catchphrase, remember you don't work scripting off, you're not meeting WCAG minimums and as such depending on the type of website and industry could land yourself in COURT. Or end up with a money pit the client pays to settle dozens of cases a year rather than fix YOUR stupidity.

Which is why so much of the new "gee ain't it neat" scripted nonsense is just more of the industry flipping the bird at the REAL reason websites exist -- delivering content to users. Same as the asshattery of dicking around drawing PSD's and calling it "design", screwing around with bloated frameworks that undo twenty years of progress, or any of the dozens of other middle-fingers to usability, accessibility, and sustainability that haven't just been hot and trendy the past decade, but have become the 'societal norm'.

But you know what Ron White said; you can't fix stupid.

Well Sidhant Panda then it sounds like you aren't making stuff that belongs on websites... but since all you showed is the default behavior picture for one browserr of something that is inherently different in supported browsers and of which you have ZERO control as a developer, it's pretty vague what it is you're even asking. Some code or actual examples of where you are might help.

Though given that push notifications aren't even out of draft/experimental yet, it really has no business being used in PRODUCTION codebases in the first place. That's why to this day IE is a pain in the arse as they implemented (and people started using) CSS2 before it was even out of draft... which didn't even match the final. Hence "experimental" meaning it is there for learning and testing, NOT production use.

Still, what do you have for code so far? Are you playing with it from a local workers/fetch method? Directly going for the Notification and Push API? What's your fallback plan for older browsers? Are you just going to tell EDGE/IE users to sod off instead of implementing a pull fallback? How are you going to handle users who have push notifications disabled altogether so it NEVER even TRIES to ask? (aka what I have Vivaldi set to do).

Your question just begs more questions, and isn't something some automated tool really should even be handling since to be frank -- like most automated tools in web development -- they just make you work harder without understanding anything (leaving you high and dry when things go bits-up face-down) instead of working smarter.

Though your final sentence spells it out -- it's part of the browser, so NO you don't have control over it directly -- but you should be using Notification.permission (inside a then) to see what the user chose -- or did not choose. It's not exactly rocket science to test for ("Notification" in window) bombing if it doesn't exist, if it does exist call Notification.requestPermission() to ask, using 'then ' to cascade to the "result" which SHOULD be the same as Notification.permission .

I'm not clear in all that what you expect some tool to automate for you since we're not even talking six lines of code. Could you simply clarify what it is you expect to be done for you by such a 'tool'?

Search Hashnode