MOBILE-2278: Add support for iOS WKWebView

[brianblanchard-wf]

Description

• When an app using this package is run inside a WKWebView and checks the version an error will be throw.
• The error is thrown because there is no Version/X.X.X string in the string returned from navigator.appVersion when running within a WKWebView.

What Was Changed

• Added a stricter check for isSafari
• Added a Browser subclass for the WKWebView

@travissanderson-wf @dustyholmes-wf @kerbyseeley-wf

[aviary2-wf]

Raven

Number of Findings: 0

[rmconsole-wf]

	When this pull is merged I will use the following information:
	Version: platform_detect 1.1.0
	Release Ticket(s): CP-2835

---

Last updated on Tuesday, January 03 09:43 AM CST

[codecov-wf]

Current coverage is 89.87% (diff: 100%)

Merging #8 into master will increase coverage by 1.99%

@@             master         #8   diff @@
==========================================
  Files             2          2          
  Lines            66         79    +13   
  Methods           0          0          
  Messages          0          0          
  Branches          0          0          
==========================================
+ Hits             58         71    +13   
  Misses            8          8          
  Partials          0          0          

Diff Coverage	File Path
•••••••••• 100%	lib/src/browser.dart

Powered by Codecov. Last update 6ab96fe...5220a0f

[travissanderson-wf]

FYI @clairesarsam-wf

[travissanderson-wf]

@clairesarsam-wf @robbecker-wf what do you guys think about detecting these "sub" type browsers like WkWebView is for the most part the same as Safari, similar to Rob's question about separating the detection of IE vs Edge from earlier. When WKWebView is detected, should isSafari return true as well as isWKWebView?

[travissanderson-wf]

@aaronlademann-wf too

[brianblanchard-wf]

@travissanderson-wf We could do that but the issue is if we return true for isSafari in the WKWebView the there would not be a version because the Safari and the WKWebView have different version strings.

[travissanderson-wf]

it could still have a version, currently it would be different depending on if it was parsed out by _Safari or _WKWebView, I just mean we could change the getter for isSafari = this == _safari || this == _wkWebView

[robbecker-wf]

Sure, I could see it being useful to know if you're running in a webview.

[brianblanchard-wf]

@greglittlefield-wf says UIP does not use the dart formatter so it has been removed.

[travissanderson-wf]

what does that have to do with this repo?

[greglittlefield-wf]

Sorry @travissanderson-wf, I was under the impression that UIP was now maintaining this repo.

Brian was running into issues with builds failing due to the formatter enforcing an 80 line length, so I told him to just remove the formatter step from the build as opposed to fixing the issues.

[travissanderson-wf]

No worries, it's presently officially owned by CP but so far it's just been me and I don't expect a whole lot of churn here TBH. In the interest of consistency, I'd prefer to keep dartfmt enforced. It's pretty easy to fix the build with ddev format :)

[travissanderson-wf]

I'm good with this as-is but would be curious to hear others' input on whether WkWebView should report as both Safari and WKWebView or only WKWebView. My suspicion is that code checking for Safari should probably also run when you're in a WKWebView and so it would be desirable as the consumer if isSafari checks would be true in that case. Special cases for WKWebView could still explicitly check for isWKWebView.

[robbecker-wf]

I'm of the opinion that isSafari should return true for wkwebview and a webview property it is a more specific version of the main browser type. For instance, it may be possible that:
isChrome => true && isWebView => true for when running in a chrome webview.

Not to bikeshed, but perhaps we should just renamed wkWebView to just webView ?

UPDATE: To be clear, I'm ok with this PR as-is, just my 2c.

[travissanderson-wf]

Wk is specific to and short for WebKit so it would be mutually exclusive with Chrome as currently defined

[brianblanchard-wf]

@robbecker-wf @travissanderson-wf I think that it would be good to detect webviews for sure but in a different ticket.

[travissanderson-wf]

+1

[travissanderson-wf]

@brianblanchard-wf can you add a QA p1 (assuming this fixes your issue)?

[travissanderson-wf]

@kerbyseeley-wf is also taking a look at it from a datatables perspective

[kerbyseeley-wf]

+10 as far as datatables is concerned. Ran a smoke test in Chrome/Firefox/Safari/IE.

[jordanross-wf]

QA +1

[travissanderson-wf]

rosie

[teresarevious-wf]

RM +1 Manually verified dependencies and configured correctly to scan for them