Recent releases of Visual Studio 2015 have left all our cshtml markup in a very fragile state. Many times we have attribute values ββthat exceed the usable width, and we need to split the attribute into several lines to make the code more readable.
Example: wide data binding
<div data-bind="someBinding: { hasErrors: bindingObject.HasErrors(duration), hasWarnings: bindingObject.HasWarnings(duration), parameterKey: bindingObject.SelectedObjectKey }, anotherBinding: value, anotherBinding2: value2, AndSoOn: yaddayadda.the.point.is.this.can.get.really.long">
Let's go back to VS 2010/2013, if we start just entering a carriage return inside the attribute value to break it, the designer will usually no longer parse the html accordingly, and the DOM elements after that will no longer be valid ... for example, broken table.
So, we found that by encapsulating the attribute value inside @ ("..."), it actually allows the designer to work smoothly.
Example: with attribute @ () binding
<div data-bind="@("someBinding: { " + "hasErrors: bindingObject.HasErrors(duration), " + "hasWarnings: bindingObject.HasWarnings(duration), " + "parameterKey: bindingObject.SelectedObjectKey }, " + "anotherBinding: value, " + "anotherBinding2: value2, " + "AndSoOn: yaddayadda.the.point.is.this.can.get.really.long")">
Well, the good news here is that the IDE will even help with the extension of this line, and when the page reformatted itself, it would be happy and the DOM was in great shape. The server will use this C # line frame and display it on one line during production, and everything was fine ...
Until we updated. VS2015 seems to hate this, but for completely different reasons. Now it splits everything, and the prompt says "Missing attribute name." When the page is reformatted, it messes up ...

So, I am walking on the glass here. Is this some kind of error or is there an answer, because we have a large code base that does not want to process this attribute formatting.
Thoughts?