Lane 11 PR 2b: labels from expressions, first/last/join, and two robustness fixes, in both engines (#737) #967

Merged

claude-bot merged 5 commits from build/lane11-2b-language into main 2026-09-30 10:16:52 +00:00

Conversation 1 Commits 5 Files changed 38 +834 -156

claude-bot commented 2026-09-30 08:58:06 +00:00

Collaborator

Copy link

Lane 11 (#737) PR 2b: a follow-up to the sheet language (#961, #965). Both definition authors (#962, #963) hit the same gaps, which stopped the sheet showing what the drawing shows. Everything below is done in both engines at once, and pinned by shared fixtures.

Additions:

1. Labels on derivation terms: a term can take a label_expr, evaluated in the field's scope. When it isn't null, its text replaces label, so a term can read "Dex 4" rather than "Attribute". note_expr works the same way. A null note is now left off rather than printed as "".
2. Modifier labels: a modifier can take a label_expr, evaluated where its value is, so "Frightened 1" rather than "Frightened". It works for definition modifiers and for those carried by entries and conditions.
3. first(list), last(list) and join(list, sep = ", "): these follow the same list rules as sum/count. Numbers go through the number-to-text rule. This is what lets the sheet show "Chain mail 16" and "d10 → 8, 7".
4. Did-you-mean for unknown function names (jion → "join"), in both the validator and the compiler.

Robustness fixes: both were found porting to JS (#965), and both used to raise where the documented result is null.

• An entry row whose kind isn't a string no longer raises.
• A rank that isn't a finite number reads as "not a rank" (null).

Fixtures:

• expr/first_last_join.json, 33 cases.
• sheet/labels-from-expressions.json: "Dex 4", notes, null fallbacks, "Frightened 1", "Chain mail 16" and "d10 → 8, 7".
• sheet/malformed-rows-and-infinite-ranks.json.
• Totals: both engines run 342 expression cases and 15 sheet cases.
• Fuzz: a differential fuzz extended to the new features (12,000 expressions, 2,040 sheets) found 0 mismatches.

Tests:

• Backend: tests/sheets passes (486), and definition.schema.json is regenerated.
• Frontend: eslint shows 0 errors, src/sheets passes (447), and so does the full vitest suite (3213).
• Lint and changelog: ruff is clean, and changelog.py check passes.

The PF2e and 5e definitions adopt the new labels in a later PR (they validate either way).

Changelog: changelog.d/737-lane11-language-labels.added.md.

🤖 Generated with Claude Code

Lane 11 (#737) PR 2b: a follow-up to the sheet language (#961, #965). Both definition authors (#962, #963) hit the same gaps, which stopped the sheet showing what the drawing shows. Everything below is done in **both** engines at once, and pinned by shared fixtures. **Additions:** 1. **Labels on derivation terms:** a term can take a `label_expr`, evaluated in the field's scope. When it isn't null, its text replaces `label`, so a term can read "Dex 4" rather than "Attribute". `note_expr` works the same way. A null note is now left off rather than printed as "". 2. **Modifier labels:** a modifier can take a `label_expr`, evaluated where its value is, so "Frightened 1" rather than "Frightened". It works for definition modifiers and for those carried by entries and conditions. 3. **`first(list)`, `last(list)` and `join(list, sep = ", ")`:** these follow the same list rules as `sum`/`count`. Numbers go through the number-to-text rule. This is what lets the sheet show "Chain mail 16" and "d10 → 8, 7". 4. **Did-you-mean for unknown function names** (`jion` → "join"), in both the validator and the compiler. **Robustness fixes:** both were found porting to JS (#965), and both used to raise where the documented result is null. - An entry row whose `kind` isn't a string no longer raises. - A rank that isn't a finite number reads as "not a rank" (null). **Fixtures:** - `expr/first_last_join.json`, 33 cases. - `sheet/labels-from-expressions.json`: "Dex 4", notes, null fallbacks, "Frightened 1", "Chain mail 16" and "d10 → 8, 7". - `sheet/malformed-rows-and-infinite-ranks.json`. - **Totals:** both engines run 342 expression cases and 15 sheet cases. - **Fuzz:** a differential fuzz extended to the new features (12,000 expressions, 2,040 sheets) found 0 mismatches. **Tests:** - **Backend:** `tests/sheets` passes (486), and `definition.schema.json` is regenerated. - **Frontend:** eslint shows 0 errors, `src/sheets` passes (447), and so does the full vitest suite (3213). - **Lint and changelog:** ruff is clean, and `changelog.py check` passes. The PF2e and 5e definitions adopt the new labels in a later PR (they validate either way). Changelog: `changelog.d/737-lane11-language-labels.added.md`. 🤖 Generated with [Claude Code](https://claude.com/claude-code)

claude-bot added 2 commits 2026-09-30 08:58:07 +00:00

Lane 11 PR 2b: labels from expressions, first/last/join, two robustness fixes (#737) ... af99f3292d

Both engines (app/sheets and src/sheets):
- derivation term objects take label_expr (non-null text replaces label)
  and note_expr (the same as note; a null note is left off)
- modifiers take label_expr, evaluated where value is
- first(list), last(list), join(list, sep=", ") with the list rules
- rank() of a non-finite number is null; an entry row with a non-string
  kind no longer raises (no kind row defaults, not found by has())
- unknown function names get a did-you-mean in the validator
New conformance: expr/first_last_join.json (33 cases),
sheet/labels-from-expressions.json, sheet/malformed-rows-and-infinite-ranks.json.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

Merge remote-tracking branch 'origin/main' into build/lane11-2b-language 5b44cb221d

claude-bot commented 2026-09-30 09:30:01 +00:00

Author

Collaborator

Copy link

This PR now also carries PR 5c: the PF2e and 5e definitions adopt the new labels, so the sheet reads as drawn. It was stacked directly on this branch, and the previous run failed only on a runner fault (Docker "AlreadyExists" on the phone journeys, run 1701), so one fresh run covers both.

What PR 5c changes:

• Only labels. Every value is unchanged, and all 19 PF2e and 5e fixtures pass in both engines and through the server.
• PF2e:
  • Class DC, spell DC/attack and strikes read "Dex 4"/"Wis 4", not "Attribute 4 (Dex)".
  • HP reads "Ancestry 8 + 6 levels 60 (Class 8 + Con 2 each)".
  • Condition modifiers read "Frightened 1", "Stupefied 2" and "Drained 1".
• 5e:
  • AC reads "Chain mail 16 + Shield 2".
  • HP reads "10 + 2 × (d10 → 8, 7) + Con 2 × 3".
  • Saves, skills and attacks read "Str 3 + Prof 2".
  • Spells read "Prof 3 · Wis 3".
  • Also "Exhaustion 2" and "Chain mail, needs Str 13".
• Short ability names come from a tables.ability_short; armour names use first(...) and HP rolls use join(...).
• Fragments: 737-lane11-pf2e-labels.changed.md and 737-lane11-5e-labels.changed.md.

**This PR now also carries PR 5c:** the PF2e and 5e definitions adopt the new labels, so the sheet reads as drawn. It was stacked directly on this branch, and the previous run failed only on a runner fault (Docker "AlreadyExists" on the phone journeys, run 1701), so one fresh run covers both. **What PR 5c changes:** - **Only labels.** Every value is unchanged, and all 19 PF2e and 5e fixtures pass in both engines and through the server. - **PF2e:** - Class DC, spell DC/attack and strikes read "Dex 4"/"Wis 4", not "Attribute 4 (Dex)". - HP reads "Ancestry 8 + 6 levels 60 (Class 8 + Con 2 each)". - Condition modifiers read "Frightened 1", "Stupefied 2" and "Drained 1". - **5e:** - AC reads "Chain mail 16 + Shield 2". - HP reads "10 + 2 × (d10 → 8, 7) + Con 2 × 3". - Saves, skills and attacks read "Str 3 + Prof 2". - Spells read "Prof 3 · Wis 3". - Also "Exhaustion 2" and "Chain mail, needs Str 13". - **Short ability names** come from a `tables.ability_short`; armour names use `first(...)` and HP rolls use `join(...)`. - **Fragments:** `737-lane11-pf2e-labels.changed.md` and `737-lane11-5e-labels.changed.md`.

rbrooks added 3 commits 2026-09-30 09:30:02 +00:00

Lane 11 PR 5c: PF2e derivations and condition labels read as drawn (#737) ... 9e3ed2ed08

Uses PR 2b's label_expr/note_expr: the varying attribute term is labelled
by its short name ("Dex" + value 4 renders "Dex 4") for class DC, spell
DC/attack and Strikes; HP reads "Ancestry 8 + 6 levels 60 (Class 8 + Con 2
each)"; valued conditions label their modifiers "Frightened 1" etc.
Values unchanged; fixtures' derivation and modifier labels updated.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

Lane 11 PR 5c: 5e derivations and modifiers read as drawn (label_expr, first/join) (#737) ... 1bc1fafdd9

AC 'Chain mail 16 + Shield 2', HP '10 + 2 × (d10 → 8, 7) + Con 2 × 3',
'Str 3 + Prof 2', proficiency '2 + ⌊(3 − 1) / 4⌋', 'Exhaustion 2',
'Chain mail, needs Str 13'. Values unchanged; fixture labels updated.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

Merge branch 'build/lane11-5c-5e-labels' into build/lane11-5c-pf2e-labels 0a19f3d12b

claude-bot merged commit 78bd49c02b into main 2026-09-30 10:16:52 +00:00

claude-bot referenced this pull request from a commit 2026-09-30 10:16:54 +00:00

Merge PR #967: lane 11 PR 2b + 5c — labels from expressions, first/last/join; the definitions read as drawn (#737)

Sign in to join this conversation.

Reviewers

No reviewers

Labels

Clear labels   

backend

Issue related to the backend

  

bot

Issue related to the bot

  

bug

Something is not working

  

duplicate

This issue or pull request already exists

  

enhancement

New feature

  

frontend

Issue related to the frontend

  

help wanted

Need some help

  

invalid

Something is wrong

  

question

More information is needed

  

recording

Issue related to the recording system

  

security

Security hardening / vulnerability work

  

wontfix

This won't be fixed

No labels

backend

bot

bug

duplicate

enhancement

frontend

help wanted

invalid

question

recording

security

wontfix

Milestone

Clear milestone

No items

No milestone

Projects

Clear projects

No items

No project

Assignees

Clear assignees

No assignees

2 participants

Notifications

Subscribe

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference

rbrooks/Quest-Board!967

No description provided.