feat(list): flatten custom fields in --json output #39

Merged

David merged 1 commit from feat/list-flatten-custom-fields into main

2026-05-15 13:09:58 +02:00

David commented

2026-05-15 13:07:54 +02:00

Owner

Summary

yt list --json previously serialized each issue with the raw customFields array, which renders in nu | from json as an unreadable list of {name, value} records (per attached screenshot).
This PR flattens that array: each custom field name is promoted to a top-level key alongside idReadable, summary, and sprints, with the value rendered via the same presentation logic the text table uses (State -> "In Review", Assignee -> "NiceGuyIT", etc.). The result is one nushell column per field.
New --custom-fields flag appends the original customFields array to each row in addition to the flattened keys (not as a replacement), for callers that need the raw structure (e.g. enum bundle ids, full assignee object).
Custom fields named idReadable / summary / sprints / customFields are skipped so they cannot shadow the built-in row keys.
Enabled serde_json's preserve_order feature so the JSON key order is stable (idReadable, summary, custom fields in server order, sprints, optional customFields) instead of the alphabetical order BTreeMap would otherwise produce.
Behavior of the non-JSON table output is unchanged.

Test plan

cargo fmt --check
cargo clippy --all-targets -- -D warnings
cargo test --all-targets (212 passed; three new tests cover flatten ordering, the --custom-fields flag, and shadow-key skipping)
Smoke test against a real YouTrack instance: yt list --query "#Unresolved for:me" --json | from json shows one column per custom field; rerun with --custom-fields to confirm the raw array is appended.

## Summary - `yt list --json` previously serialized each issue with the raw `customFields` array, which renders in `nu | from json` as an unreadable list of `{name, value}` records (per attached screenshot). - This PR flattens that array: each custom field name is promoted to a top-level key alongside `idReadable`, `summary`, and `sprints`, with the value rendered via the same presentation logic the text table uses (`State` -> "In Review", `Assignee` -> "NiceGuyIT", etc.). The result is one nushell column per field. - New `--custom-fields` flag appends the original `customFields` array to each row in addition to the flattened keys (not as a replacement), for callers that need the raw structure (e.g. enum bundle ids, full assignee object). - Custom fields named `idReadable` / `summary` / `sprints` / `customFields` are skipped so they cannot shadow the built-in row keys. - Enabled `serde_json`'s `preserve_order` feature so the JSON key order is stable (`idReadable`, `summary`, custom fields in server order, `sprints`, optional `customFields`) instead of the alphabetical order BTreeMap would otherwise produce. - Behavior of the non-JSON table output is unchanged. ## Test plan - [x] `cargo fmt --check` - [x] `cargo clippy --all-targets -- -D warnings` - [x] `cargo test --all-targets` (212 passed; three new tests cover flatten ordering, the `--custom-fields` flag, and shadow-key skipping) - [ ] Smoke test against a real YouTrack instance: `yt list --query "#Unresolved for:me" --json | from json` shows one column per custom field; rerun with `--custom-fields` to confirm the raw array is appended.

David added 1 commit

2026-05-15 13:07:55 +02:00

feat(list): flatten custom fields in --json output

Check / fmt + clippy + build + tests (pull_request) Successful in 18s

Details

Create release / Create release from merged PR (pull_request) Has been skipped

Details

39ff3d14fe

Each issue in `yt list --json` now serializes as a flat object: `idReadable`, `summary`, every custom field promoted to a top-level key (presentation string), then `sprints`. This lets `from json` in nushell render one column per field instead of a nested `customFields` list of records that was unreadable in the table view.

The raw `customFields` array is no longer in the default output. Add `--custom-fields` to append it (in addition to, not in place of, the flattened keys) for callers that need the full structure.

Enable serde_json's `preserve_order` feature so the JSON key order matches the insertion order above instead of being sorted alphabetically.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>