Skip to content

docs(skill): use selected() for the "other" companion of a select_multiple - #80

Merged
jstet merged 1 commit into
mainfrom
docs/other-selected
Sep 25, 2026
Merged

jstet merged 1 commit into
mainfrom
docs/other-selected

Conversation

@jstet

@jstet jstet commented Sep 25, 2026

Copy link
Copy Markdown
Member

Fixes formulaid's report. The skill told authors to use ${q} = 'other' for the "other" companion of a select_multiple, which only matches when "other" is the only answer ticked. It now shows selected(${q}, 'other') for select_multiple and = for select_one.

  • Registry: the unused relevanceTemplate becomes relevanceTemplates, one per type.
  • Example: the select_multiple_other example had the same bug and now uses selected(). Its TSV snapshot changes the companion's relevance.
  • Live: the companion stays hidden either way; the live tests for this entity pass.

That the companion is emitted at all next to the native other=Y is filed as #79.

🤖 Generated with Claude Code

…tiple

The skill's or_other section told authors to set the companion's relevant to
${<question>} = 'other' for both select types. For a select_multiple that
only matches when "other" is the sole answer ticked, since the value is
space-separated; selected(${<question>}, 'other') is right (reported by
formulaid).

- convention:other: the unused relevanceTemplate becomes relevanceTemplates,
  one per type (select_one: =, select_multiple: selected()); the generated
  skill renders both and says why.
- The select_multiple_other example used = as well; it now uses selected().
  Its TSV snapshot's companion relevance changes accordingly. The companion
  stays hidden either way (live round-trip passes); that the companion is
  emitted at all next to the native other=Y is filed as #79.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
@jstet
jstet merged commit 89c1884 into main Sep 25, 2026
3 of 4 checks passed
@jstet
jstet deleted the docs/other-selected branch September 25, 2026 14:56
jstet added a commit that referenced this pull request Sep 25, 2026
…"other" (#81)

#80 changed the select_multiple_other example to selected(${q}, 'other'),
which the forward path writes as (q_other.NAOK == 'Y'): the native "other"
is LimeSurvey's implicit `other` subquestion. The reverse didn't know that
code, read q_other as a separate question and produced
selected(${q_other}, 'Y'), failing the round-trip contract test. #80 merged
with that test failing (my merge chain didn't check the CI exit status).

The selected() lookup now includes the implicit `other` code for every
multiple choice with other=Y, so the relevance reverses to
selected(${q}, 'other') again.

Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant