Open abehring opened 4 months ago
While discussing this with @abehring I bisected this to c5a24b7c2b476a3c3a65f524d4b14ed4bf3a7059 . It is somewhat surprising that this behaviour has been in the code for so long, without being noticed.
There are several issues on the tracker related to negative vectors, but I don't think any of them are quite the same as this one.
I've noticed that the current version 5 of FORM sometimes drops a minus sign when matching a single negative vector. A minimal working example is
With an ancient FORM 3.3 binary I get the expected
while the current master branch yields:
@jodavies also observed that the exact syntax on the right-hand side of the
id
statement matters for whether the bug is triggered:id Delta(v1?) = v1;
is OK (removing the second arg)id Delta(v1?) = D.v1;
tooid Delta(v1?,du1?) = (D.v1)^du1
too ??