Digit separators mess with comments

[Eisenwave]

To reproduce, try something like

\begin{codeblock}
unsigned x = 0xffff'ffff; // comment
\end{codeblock}

This will treat // comment as having code font, presumably because the C++ parser is not aware of digit separators, so it thinks that 'ffff; // comment is one long character literal with no closing '.

[jensmaurer]

I'm not sure we can do something here.

Since we don't use digit separators often in our examples in the standard, are there specific instances where we have a problem and might want to move the comment a line above or below or similar?

[Eisenwave]

One would think that https://eel.is/c++draft/time.format#example-1 is affected, but it seems like digit separators in decimal numbers are rendered properly.

It might be an issue limited to digit separators in hexadecimal literals.

I cannot find any improperly rendered examples. It's just something annoying I've discovered when creating some new wording that could have contained 0xff'ff'ff'ff or 0xffff'ffff.

What extension are we using for this purpose anyway? Is that something we could fix after all?