compute.att_gt trimming due to pscore_problems_likely not applied when panel=FALSE

[ransiw]

With balanced panels when overlap is violated (lines 218 to 221 in compute.att_gt) the relevant treatment unit is excluded.

did/R/compute.att_gt.R

Line 218 in 8b2201d

if (reg_problems_likely | pscore_problems_likely) {

But this is not applied with unbalanced panels when panel=FALSE. After line 318 in compute.att_gt, the estimation starts and no checks for overlap happen.

Is there a reason why this was not implemented for when panel=FALSE?