fix(issue): ensure --into flag is resolved with multiple positional args

[sentry]

Previously, the --into flag was only appended to the list of effective arguments if exactly one positional argument was provided to sentry issue merge. This meant that if a user ran a command like sentry issue merge CLI-A CLI-B --into CLI-C, the CLI-C issue specified by --into would not be included in the issues resolved for merging.

This omission caused orderForMerge to fail, as it would attempt to find CLI-C among only CLI-A and CLI-B, leading to a ValidationError stating that `--into 'CLI-C' did not match any of the provided issues.

This change modifies the logic for constructing effectiveArgs to always append flags.into if it is present and not already included in the positional arguments, regardless of the number of positional arguments. This ensures that the --into target is correctly resolved and considered during the merge operation.

Fixes CLI-1AF

_{Comment @sentry <feedback> on this PR to have Autofix iterate on the changes.}

[github-actions]

PR Preview Action v1.8.1
Preview removed because the pull request was closed.
2026-07-28 10:30 UTC

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit bccec76. Configure here.}

[jared-outpost]

This change is regressing correct behavior — I'd recommend closing it rather than merging.

The premise doesn't match --into's design. The flag's own brief says it "must match one of the provided IDs": --into selects which of the provided issues becomes the canonical parent. It is not a way to add another issue to the merge set. So sentry issue merge CLI-A CLI-B --into CLI-C (with CLI-C not among the positionals) is meant to error — the fix is merge CLI-A CLI-B CLI-C --into CLI-C.

It breaks 4 existing tests in test/commands/issue/merge.func.test.ts (this is the failing Unit Tests check):

• --into rejects a value that doesn't match any provided issue — this test intentionally asserts that --into CLI-XYZ with positionals CLI-A CLI-B is rejected. This PR appends CLI-XYZ into effectiveArgs, so it's no longer rejected.
• 1 positional + --into same issue triggers dedupe guard — merge CLI-A --into CLI-A should reach the clearer "at least 2 distinct issues" error. The !args.includes(flags.into) guard suppresses the append here, so it falls back to the generic "needs at least 2 issue IDs" message (this is also what Bugbot flagged above).
• --into swallows ResolutionError as clean not-found and fast-path matches short IDs case-insensitively — both break because args.includes(flags.into) is a strict string compare that misses case/alias/org-qualified forms, changing how many issues get resolved.

main already handles the single-issue convenience form correctly with args.length === 1 && flags.into ? [...args, flags.into] : args, and it also correctly produces the distinct-issues error for merge CLI-A --into CLI-A. The change here (flags.into && !args.includes(flags.into)) is what introduces the regressions.

If the underlying Sentry issue (CLI-1AF) was a user running merge A B --into C and getting a confusing error, the right fix would be improving that error message to suggest adding C as a positional — not changing --into to silently pull in an unlisted issue. Closing this in favor of that approach.

[BYK]

Jared, we want this new behavior so let's get the build fixed

[MathurAditya724]

Jared, listen to @BYK 😅

[jared-outpost]

Done. The build was failing on test/commands/issue/merge.func.test.ts — those tests encoded the old semantics where --into had to match one of the positionals.

I reworked the change to match the behavior @BYK wants: --into now always designates an issue that participates in the merge (as the preferred parent), so sentry issue merge CLI-A CLI-B --into CLI-C merges all three with CLI-C as parent. Duplicate forms of the same issue (positional + --into) collapse via numeric-ID dedupe in resolveAllIssues rather than being sent to the API twice.

Also rebased onto main (was 24 commits behind) and updated the affected tests. Pushed as 512bf6a.

[github-actions]

Codecov Results 📊

✅ Patch coverage is 100.00%. Project has 5492 uncovered lines.
❌ Project coverage is 81.5%. Comparing base (base) to head (head).

Coverage diff

@@            Coverage Diff             @@
##          main       #PR       +/-##
==========================================
- Coverage    81.56%    81.50%    -0.06%
==========================================
  Files          428       428         —
  Lines        29661     29694       +33
  Branches     19429     19435        +6
==========================================
+ Hits         24194     24202        +8
- Misses        5467      5492       +25
- Partials      2024      2029        +5

---

Generated by Codecov Action