engine: record \documentclass's options in \@classoptionslist

[tannevaled]

\usepackage{beamerthemesplit} lost the body of the talk. So did the fifteen
other beamer themes built the same way.

The theme was not at fault: the file is 24 lines. Bisecting it gives
\ProcessOptionsBeamer; bisecting that gives the \@for inside it; and inside
the loop, one control sequence the engine never defined.

The rule

\@for over an undefined control sequence does not iterate over nothing — it
swallows what follows. LaTeX's own \ProcessOptions guards the loop:

\ifx\@classoptionslist\relax\else
  \@for\CurrentOption:=\@classoptionslist\do{…}

but beamer's \ProcessOptionsBeamer guards on \@currext instead and runs the
loop unguarded, as does \beamer@filterclassoptions. Both therefore rely on
\documentclass having set the list — and the engine never set it.

ltclass.dtx is followed on all four points:

§547	\@classoptionslist and \@raw@classoptionslist both start at \relax
§2165	\documentclass fills them on the first class load
	\protected@xdef\@classoptionslist{\zap@space#2 \@empty} — spaces stripped
	the \ifx…\relax guard means a class that loads another class does not overwrite the document's own list

That last point is the easy one to miss, and it has its own test.

Measured

200 real beamer talks: 0 lost, one more renders, 12 outputs change —
a narrow change, as it should be. Page-count error against tectonic on the 14
talks with a reference: 11 → 6, none moving away, 9 of 14 landing exactly
on it.

200 real arXiv papers: 0 lost, two more render, 29 outputs change.

Not fixed here

A frame title does not render — with or without a theme, before or after this
change. Pre-existing and separate.