OK, thanks. I confirm the report, and will work on a resolution.

The issue226 branch of my fork of MathJax should cause HTML-CSS and SVG output to properly use the style and weight indicated in the mathematics, but the font-family of the surrounding text.

OK, so the Configuration/HTML-CSS/mtextFontInherit/issue226.html is not correct. I should rather test with \textbf and \textit.

Right. I understand the test you had in mind, but in order to make the output using mtextFontInherit work consistently for either value (true or false), it needs to ignore the style and weight of the outer font and use that of the mathematics. Otherwise, if the outer font were bold, say, then \text{A} and \textbf{A} would not be different, and they should be.

So just to be sure that we agree: the issue here is that {mtextFontInherit: true} prevents \textit and \textbf to behave as expected?

OK, I'll write two != reftests

x against \textbf{x}
x against \textit{x}

(with mtextFontInherit = true in both cases)

(I meant \text x)

Configuration/HTML-CSS/mtextFontInherit/issue226a.html
Configuration/HTML-CSS/mtextFontInherit/issue226b.html

Verified with Opera, Chrome and Firefox.

=> In testsuite, Ready for release

Yes, you are correct, \text{x} against \textbf{x} and \text{x} against \textit{x} should be it. Thanks.