grom's slam macro sits on SHIFT-1, which his bindings file resolves to
MULTIACTIONBAR1BUTTON1, and pfUI points both `MultiActionButtonDown` and
`MultiActionButtonUp` at the same `pfActionButton` (`modules/actionbar.lua`
:1516-17), which calls `frame:Click()`. So the reference evaluates that macro
twice per press — once on the press, once on the release — and he reported
seeing it evaluate once. That second evaluation is a second roll at a 0.2 s
no-clip window, so it is worth a test of its own.
The test holds a modifier across both edges, as a hand holding shift does, and
asserts the down half on the press and the up half on the release. It passes:
the registry's `Kind::EdgeUpDown` does run both legs through a chord. The
harness says the same one layer up — `MultiActionButtonDown` and
`MultiActionButtonUp` each reach `UseAction` exactly once with pfUI loaded —
so whatever is costing him the second evaluation is not in either of these two
layers, and this test is here to keep it that way while the search goes on.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SgstFWtn1RxEJ5kVrpgYwn
gromandClaude Opus 5
89556c991a