Consecutive big delimiters are misaligned if \smash appears in their subscripts

[EdwinChan]

I'm using version 2017/01/27 v0.8d. The source

\documentclass{article}
\usepackage{amsmath,unicode-math}
\begin{document}
\[\biggl(\frac XY\biggr)_{\smash Z}\biggl(\frac XY\biggr)_{\smash Z}\]
\[\biggl[\frac XY\biggr]_{\smash Z}\biggl[\frac XY\biggr]_{\smash Z}\]
\[\biggl\{\frac XY\biggr\}_{\smash Z}\biggl\{\frac XY\biggr\}_{\smash Z}\]
\end{document}

typesets to the following; clearly the second pair of delimiters on each line is incorrectly shifted downward:

[godbyk]

The fraction bar is heavier, too.

[davidcarlisle]

You are using xetex? (I see this with xetex but not luatex)

[EdwinChan]

@davidcarlisle That's correct; I'm using XeLaTeX 3.14159265-2.6-0.99996 that comes with TeX Live 2016.

[davidcarlisle]

This seems to be a xetex engine issue, not fontspec as I can reproduce with plain Tex

\font\m= "[latinmodern-math.otf]/OT:script=math;language=DFLT;" at 10pt
\font\sm= "[latinmodern-math.otf]/OT:script=math;language=DFLT;" at 7pt
\font\ssm= "[latinmodern-math.otf]/OT:script=math;language=DFLT;" at 5pt

\textfont0=\m
\scriptfont0=\sm
\scriptscriptfont0=\ssm
\textfont1=\m
\scriptfont1=\sm
\scriptscriptfont1=\ssm
\textfont2=\m
\scriptfont2=\sm
\scriptscriptfont2=\ssm
\textfont3=\m
\scriptfont3=\sm
\scriptscriptfont3=\ssm

$$
\biggl({X\over Y}\biggr)_{\smash Z}
\biggl({X\over Y}\biggr)_{\smash Z}
$$

\bye

[u-fischer]

With miktex there is no problem. The xetex version number is the same as in texlive but the miktex has been compiled with newer libraries. E.g. harfbuzz version 1.4.1 versus version 1.2.6.

Imho the problem it is part of a number of questions and bug reports regarding leaking fontdimens (e.g. https://sourceforge.net/p/xetex/bugs/75/#b012 and https://tex.stackexchange.com/questions/281549/why-is-the-fraction-off-the-math-axis-in-xetex) which have been be fixed in the development version and that miktex already has this fixes but the newer libraries could be also the cause that miktex behaves better.

[wspr]

I'm going to close this as an issue here, although discussion can by all means continue; unfortunately I don't have the capacity to follow up engine problems at this stage...