#10034 System doens't assing added bonds to defined Preset component

[oguzhanduran]

How the feature works? / How did you fix the issue?

(Screenshots, videos, or GIFs, if applicable)

Check list

• unit-tests written
• e2e-tests written
• documentation updated
• PR name follows the pattern #1234 – issue name
• branch name doesn't contain '#'
• PR is linked with the issue
• base branch (master or release/xx) is correct
• task status changed to "Code review"
• reviewers are notified about the pull request

[claude]

The fix is correct — the previous code seeded the traversal from the atom already assigned to the component, so it walked through component atoms first and could incorrectly trip hasConnectionToOriginalStructure for connections that have nothing to do with the new bond. Two suggestions while this line is being touched:

1. Rename atomIdToStartFrom → unassignedAtomId (or similar). The intent — "start from the unassigned endpoint" — would then be self-documenting, and the ternary's direction obvious by inspection. This is exactly the kind of bug that name-driven clarity prevents from regressing.
2. Consider adding a focused unit test for this function. The three branches (no component / one component / two components) plus the "unassigned side reaches the original structure" case are highly testable, and a regression test would lock in this silent semantic inversion.

[claude]

Reviewed the one-line fix in canAssignRnaComponent. The change is correct: traversal must start from the unassigned endpoint so it explores only the new fragment's external connections. The previous code started from the atom already in the component, which walked through component atoms first and could falsely flag hasConnectionToOriginalStructure for pre-existing connections unrelated to the new bond. As a side benefit, the new behavior is also strictly faster (O(fragment size) instead of O(component size + fragment size)).

Main concern: no test is being added for a silent semantic-inversion bug. Inline suggestions cover a rename to make the intent self-documenting, an outer-loop early-exit, and a small dead-code cleanup.

[claude]

A couple of pre-existing nits in the same function, optional but cheap to fix while touching it:

• Outer loop doesn't early-exit on disqualifying flags (Editor.ts:2112): once hasConnectionToAnotherComponent or hasConnectionToOriginalStructure is set, the forEach short-circuits per-neighbor at lines 2125–2130, but the outer while keeps draining atomsToCheck. Tightening the loop condition to while (atomsToCheck.length > 0 && !hasConnectionToAnotherComponent && !hasConnectionToOriginalStructure) would early-exit the whole traversal in the disqualifying case.
• Redundant visitedAtomIds.add (Editor.ts:2176): neighborAtomId is already added to the set at line 2145, so the second add at 2176 is dead.