You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
highs duplicate-column-printf.lp with output_flag = false still prints
HighsPostsolveStack::DuplicateColumn::undo Col is nonbasic at zero with upper bound of 1
because it is a bare printf (HighsPostsolveStack.cpp:751). In a debug build the assert(111 == 679) after it fires. The comment above it (line 733) asks "what if it's at zero with a finite upper bound?". This LP is such a case (x5 in (-inf, 1], merged with its duplicate x43). In release, the clean-up solve after postsolve returns the right answer (Optimal, -55, same as presolve off).
Activity
jajhall commentedon Nov 6, 2026
Part 3 of #3558
jajhall commentedon Nov 8, 2026
Closed by #3580