A findings list is triaged, not merely read. The goal of a first pass is to sort every entry into fix, waive with evidence, or needs-design-decision, and to clear the unambiguous structure bugs before spending review time on the architectural ones.
The excerpt below is fabricated and labelled illustrative: it shows the shape of triage thinking, not any real report. Work the generic shapes from the previous lesson, never message text.
illustrative findings excerpt (fabricated triage example, not tool output)
R1 different-reset, ctrl_a -> ctrl_b ......... no protection ............ FIX: synchronise
R2 different-reset, cfg_static -> ctrl_c ...... static after boot ........ WAIVE: qualify + re-check intent
R3 combinational gate on reset net rst_b ..... glitch risk .............. FIX: remove gate
R4 release straight from async source, dom D . ....................... FIX: add synchroniser
R5 producer P released before consumer C ............................... DECIDE: sequence or prove harmless
R6 reset used as enable in domain E .................................... FIX: synchronise firstFix the structural rows first (R3, R4, R6, R1): each has one standard fix, and each fix can change what the remaining findings mean.
Re-analyse after the fixes: some ordering and different-reset findings disappear or change shape once the reset nets are clean.
Decide the architectural rows (R5) with the designer: sequence the pair, add a ready flag, or run the adversarial release-order test as evidence.
Waive last (R2): a waiver records the claim (static after boot), the evidence (qualification logic plus intent), the reviewer, and the condition that reopens it (any change to the reset structure).
Because each structural fix changes the reset nets the remaining findings were reported against: ordering analysis and waivers written before the fixes may describe circuitry that no longer exists.
The evidence behind the claim, an independent reviewer, and the condition that reopens it, typically any change to the reset structure. A reason alone is an opinion, not a waiver.