@ornicar i idk how to fix this for that case
==> X lila.study.PgnDumpTest.force variation after last mainline move should render in brackets 0.069s munit.ComparisonFailException: modules/study/src/test/PgnDumpTest.scala:151
150: )
151: assertEquals(rootToPgn(tree).value, "1. e4 c5 (1. d4)")
values are not the same
=> Obtained

1. e4 (1. d4) 1... c5
   => Diff (- expected, + obtained)
   -1. e4 c5 (1. d4)
   +1. e4 (1. d4) 1... c5

idk how to fix this for that case

You mean for the one case you tested?

That new test looks wrong to me, or I don't understand what you're trying to do.

idk how to fix this for that case

You mean for the one case you tested?

That new test looks wrong to me, or I don't understand what you're trying to do.

that one test is wrong in ci for my update in force variations

The expectation in the test appears logically incorrect from the point of view of chess notation.

Indeed. And yet you did write that test.

So how can you say that it "appears logically incorrect"?

I don't get it.

Indeed. And yet you did write that test.

So how can you say that it "appears logically incorrect"?

I don't get it.

In chess notation (and in the step-by-step data of a chess tree), the rule is that a variation immediately follows the move from which it deviates.

Move 1. e4 is White's first move along the main line.
Variation (1. d4) is an alternative to White's first move.
Move 1... c5 is Black's response, which is only possible after 1. e4.

Parser Logic:
The parser reads 1. e4. The board position is White played e4.
The parser sees an opening bracket (). It saves the current position (after 1. e4) to memory and begins to play the variation.
Inside the brackets is 1. d4. The parser realizes, "Aha, instead of e4 it could have been d4."
The parser sees a closing bracket (). It restores the position to what it was before the bracket (it becomes 1. e4 again).
The parser reads c5. This is Black's move after 1. e4.

The test expects: "1. e4 c5 (1. d4)"
If you feed this string to the PGN parser:

It reads 1. e4.
It reads c5. The position is: 1. e4 c5.
It sees the bracket (. It preserves the position after 1... c5.
Inside the bracket, it reads 1. d4.
Conflict: The parser expects Black to move (since White just played e4 and Black played c5, the next move should be White #2), but it reads 1. d4 (White's move #1). Or, if you ignore the numbers, a logical error occurs: the d4 variation is applied to the position after c5, which is impossible, since d4 is an alternative to the very first
Thats why i think about bad test , so maybe im wrong