Modenov's computations as a corpus, with the book's errors as rows that must be refused #2153

Reading Modenov's Problems in Geometry for #2131 produced the book's computations in a form AngouriMath can run. That is several hundred identities, determinants, equations and loci, each with what the library does with it today. It also found about 80 errors in the book, each confirmed by computing. Some examples:

• a wrong arc in Chapter III Problem 10: arccos(9/16) where 2·arccos(−1/(2√2)) is meant;
• A = 2π/7 where π/7 is meant in Problem 45;
• Problem 43 is false as stated;
• the general-coordinate angle formulas in Chapter V use the wrong metric.

Wanted: a corpus of these, recorded the way Sources/Tests/UnitTests/Corpus records its problems, as solved, unsolved, wrong, error or timeout.

• Each row reports its own verdict, so A conjugate, with re, im and arg, as functions #2145–Facts about a symbol that rewrite it: abs(z) = 1, a real parameter, a relation such as alpha^7 = 1 #2148 and the defects found alongside them are measured as they land.
• A row from the errata asserts that the printed form is refused.

This is what #1409 did with Sullivan's book.