A custom term list should act like removable review tape, not an automatic correction engine. Rune Studio's current Mac documentation says that terms in the selected list are highlighted automatically. The retained STUDIO-312 run proves a narrower set of facts: the temporary list name and active state, removal of term A, and readback of term B. It did not operate the graphical highlighting surface or observe a rendered color. This text editor highlight custom word list procedure therefore treats configuration evidence and a reader's on-screen confirmation as separate acceptance checks.
Give the Temporary List a Name That Cannot Be Confused with Production
The test list is named Evidence-STUDIO-312-1788369720. It contains Term-STUDIO-312-A and Term-STUDIO-312-B. Before creating it, record the number and names of existing lists. That baseline is needed at cleanup; merely deleting something at the end does not prove that the production dictionaries were left intact.
Place A and B in a short synthetic passage. A is the removal target and B is the control that must survive. The purpose is not to judge whether either word is good prose. It is to determine whether one named list can be activated, changed by one membership, read back, and removed without affecting neighboring list data.
Verify activeIndex 1 Before Removing Term A
After activating the temporary list, read the state and match both activeIndex=1 and the full list name. Stop if either value differs. An index alone is fragile because adding or removing another list may shift positions. The index and name together identify the intended target for this fixed workspace.
Remove only Term-STUDIO-312-A, save the state, and read the list again. The measured result contains one member: Term-STUDIO-312-B. No manuscript sentence was rewritten during this operation. The controlled change occurred in list membership, and B's survival acts as the positive control.
Keep Membership Evidence Separate from Color Evidence
The command interface exposed the list name, active position, and member strings. It did not expose the exact color painted in the graphical editor, transparency, theme interaction, or precedence when selected text overlaps a registered term. Therefore, this run cannot truthfully say that B appeared blue, yellow, or any other specific color.
For an on-screen check, open the retained workspace and temporary list in the current Mac build. Place an A occurrence, a B occurrence, an unrelated word, and a selected B occurrence in the same view. Compare the two-member state with the saved state after A is removed: B alone should remain highlighted as a registered term. Record the application version, theme, active list name, highlight setting, and screenshot under one run identifier. If B is not highlighted, stop before judging color and reread the active list name and exact stored string. Until that visual record exists, the exact color, transparency, and selection precedence remain explicitly unmeasured.
Treat Every Highlight as a Question for the Editor
Even after graphical highlighting is confirmed, a marked word may be an approved name, a quotation, or a deliberate exception. Read the complete sentence and mark the occurrence as change, keep, or unresolved. The list points to locations efficiently; it does not own editorial intent.
This boundary also prevents an unrelated connection specimen from becoming evidence. The relevant input is the named temporary list and its two members, not an EPUB path rewrite or a search-and-replace result. Every claim in this article maps to list creation, activation, member removal, state readback, or cleanup.
Recover a List When the Control Term Disappears
If the reread returns zero members, preserve that failed state. Return to the saved two-member receipt, confirm the exact code points of B, and add B back to the same temporary list. Read the state again. Recovery passes when A is absent and B is present exactly once.
Do not edit the specimen sentence to make the apparent count look right. A missing highlight candidate can be caused by list membership, spelling, normalization, or the chosen active list. Changing prose would hide which condition failed. Diagnose the state first, and only then perform editorial work on a separate manuscript copy.
Remove the Evidence List and Confirm the Baseline Count
Before cleanup, save the accepted one-member state. Deactivate Evidence-STUDIO-312-1788369720, delete it, and list the remaining dictionaries. The measured cleanup condition is that the list count returns to its starting value. If it does not, stop and compare list names rather than deleting another entry by position.
The final receipt contains the starting list inventory, two-member state, active name and index, one-member readback, cleanup inventory, and any unmeasured visual fields. Acceptance means B alone remained before cleanup and the original list count was restored afterward. It does not claim that the interface color was seen or that a marked term was automatically an error.
Compare Stored Strings, Not Their Apparent Shape
Two entries can look alike while differing by trailing space, character width, or Unicode normalization. Preserve the exact two-member state and the exact one-member reread in the evidence. Acceptance is not “one colored shape remained”; it is that A is absent, B occurs once, and total membership is one in the named list.
If a B value with a trailing space survives by mistake, do not add another visually identical B. Compare the stored value with the original input, remove the malformed member, add the intended string once, save, and reread. Case sensitivity and substring matching are still separate display behaviors. They require dedicated variants in a graphical run and must not be inferred from membership state.
Promote a Term Only Through a Separate Review
Success in a temporary list does not authorize copying B into a production dictionary. An editor should decide whether the term is reusable, title-specific, or only a one-time query. If promotion is approved, record the production list name and member inventory before and after the addition.
When a neighboring production term disappears during promotion, recover by exact list name and member string, not by position. Reread the complete list and require every original member plus one approved B. The current test stops earlier: it accepts temporary membership, preserves the receipt, removes the evidence list, and restores the baseline count. Production promotion and exact rendered color remain explicitly outside this measured run.