What steps will reproduce the problem?

read ("eq97.fas") set (root: "Seq1") transform ((all, tcm:(1,1)), gap_opening:1) build (branch_and_bound, trees:100) transform (static_approx) calculate_support (bremer, build(100), swap(trees:100)) report ("eq97.bremer", supports) report (supports)

What is the expected output?

Bremer Supports: Support tree: (Seq1 (Seq2 (Seq3 Seq4)[3.])[1.#INF])[1.#INF]

What do you see instead?

Bremer Supports: Support tree: (Seq1 (Seq2 (Seq3 Seq4)[1.#INF])[1.#INF])[1.#INF]

What version of the product are you using? On what operating system?

release 2318 ncurses_poy on Windows

Please provide any additional information below.

This 2nd batch file works fine. The only difference is the "transform" command has been removed.

read ("eq97.fas") set (root: "Seq1") transform ((all, tcm:(1,1)), gap_opening:1) build (branch_and_bound, trees:100) calculate_support (bremer, build(100), swap(trees:100)) report ("eq97.bremer", supports) report (supports)

Thanks for checking this out.

Mark

Comment deleted

Comment deleted

The problem is caused by the comparison of infinite floating point numbers. It has been corrected in build _4_0_2339.

The problem is unlikely to occur with a decent number of builds per clade (even a small number of builds like build (10) will not show the problem), so the bug is unlikely to have major relevance in real analyses).