engine: \pmb is amsbsy's poor man's bold, and there is a real one here

[tannevaled]

\pmb does not switch font. It typesets its argument three times, offset by a fraction of an ex, so the ink looks heavier:

\def\pmb@#1#2{\setbox8\hbox{$\m@th#1{#2}$}%
  …\mkern-.8mu\copy8 \kern\dimen@\mkern.4mu\raise\pmbraise@\copy8 …}
                                              % amsbsy.sty:40-57

It is the fallback the package offers when no bold math version exists. One exists here: the maths layer renders \boldsymbol, which is amsbsy's own real bold (\mathversion{bold}, amsbsy.sty:28) and which go-tex/math knows (parse.go:443).

So \pmb{x} now expands, on the maths retry path, to the thing \pmb was imitating rather than costing the formula. 348 of the 4172 formulas the arXiv corpus drops are this one command.

Measured — 200 arXiv papers, 200 beamer talks

4 documents change and 2 repaginate — one closer to tectonic (27 → 29 against 33), one further (17 → 18 against 17) — for +3373 glyphs of formulas that were simply absent. The page error against the 157 tectonic renders is 670 → 669. beamer does not move by one glyph.

Test

Added to mathaliases_test.go beside \mathds and \Tilde: $\pmb{\alpha}+\pmb{x}$ must render with neither pmb nor \pmb in the skip tally. Fails on main with map[\pmb:1].