Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
20 changes: 20 additions & 0 deletions rust/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -25,12 +25,26 @@ every command group reaches parity.
| `runtime:check`: native build identity and compilation target | Ported, intentional output/help difference |
| `doctor`: redacted connection sources, server metadata, warnings and next steps | Ported |
| `workflow:list`: one namespace/page, filters, Server envelopes, JSON and JSONL | Ported |
| `workflow:list-runs`: one workflow response page, envelope/cursor, JSON and JSONL | Ported |
| Workflow, namespace and schedule commands | Next slices |
| `--tls-verify=false` against HTTPS | CLI support pending |

Server calls go through the published `durable-workflow` crate, pinned to
the latest qualified release.

`workflow:list-runs` sends its workflow ID as one percent-encoded SDK path
segment. IDs containing slashes, query delimiters or fragments stay within
that segment; they cannot select another route. Empty IDs and `.` or `..`
are refused with exit 2 before the runs request. This deliberately replaces
the PHP command's interpolation of the ID into the request path.

The command returns one Server response and leaves any `next_page_token`
unconsumed. JSON preserves that envelope and cursor; JSONL emits only the
returned records. Successful responses must identify the requested workflow
and `list_runs` operation. Explicit row workflow IDs must match; omitted IDs
in the published PHP Server shape inherit the verified envelope without
adding fields. Contradictory identities or operations are refused.

## Build and test

Requires Rust 1.86 or newer.
Expand Down Expand Up @@ -85,6 +99,12 @@ option definitions and help text), generated from the pinned reference by
read it, so they match 2.x even for commands not ported yet. The harness
also checks `--help` for every catalog command.

Catalog regeneration applies the reviewed description/help changes in
`parity/catalog-updates.json`. Comparisons reconcile only those exact fields
on descriptor surfaces; command-result JSON and unrelated definitions remain
unchanged. The run-list help describes one response page instead of promising
that every run has been retrieved.

A case can record an intentional difference with a reason and the exact
fields it excuses; any other differing field still fails. One exists today:
2.x prompts "Do you want to run …?" on stdout even when stdin is not a
Expand Down
Loading
Loading