[mvdbeek]

Galaxy rejects a workflow at import when two of its steps share a label, and that includes input steps:

400: Duplicated step label 'Strip version from reference sequence ids' in request.  (err_code 400011)

Neither lint_ga nor lint_format2 checked for this. As a result, planemo workflow_lint --iwc --fail_level error passed galaxyproject/iwc#310, where a parameter_input and a tp_sed_tool step had the same label, and CI only failed once Galaxy was running.

Change

• New structural lint rule StepLabelNotUnique, an error added to the structural profile. It mirrors Galaxy's check in lib/galaxy/managers/workflows.py:

  • labels must be unique within one workflow level, inputs included;
  • each subworkflow has its own namespace;
  • empty or missing labels are ignored;
  • comparison is exact-string.

  The message lists every occurrence, e.g. Duplicated step label 'X' (step 15, step 16). Native errors carry json_pointer /steps/<key>/label.

• Format2 gets the same check. It catches three cases:

  • an input and a step with the same key;
  • repeated label: in list-form steps;
  • repeated input ids.

  To avoid false positives or misses, the lint reuses the labelling that conversion applies. _step_label and the input-label logic from _build_input_step are now shared as native_step_label and native_input_label. This also covers the Input N fallback and the _unlabeled_* sentinels.

Tests

Declarative cases in lint_native.yml and lint_format2.yml, with new fixtures:

• Should error:
  • an input and a tool step share a label (the IWC #310 shape);
  • a duplicate inside a subworkflow;
  • Format2: an input/step key clash;
  • Format2: repeated labels in list-form steps.
• Should not error:
  • the same label in a parent workflow and its subworkflow;
  • several steps with null, empty or missing labels.

The four positive cases fail without the lint change.

Checks

• Test suite: passes with GXFORMAT2_TEST_IWC_DIRECTORY set, 1895 tests.
• IWC corpus: none of the 126 currently merged IWC .ga workflows trip the new rule, natively or after conversion to Format2.
• Repro: the bad revision from iwc#310 now reports Duplicated step label 'Strip version from reference sequence ids' (step 15, step 16). The fixed revision lints clean.

Not included (open for discussion)

• Duplicate workflow-output labels. Galaxy only emits an upgrade message for these ("Ignoring duplicate labels: ...") and still imports the workflow. They are not an import error, so I left them out rather than make them a lint error. A warning-level rule could be a follow-up.

Follow-up

For IWC CI to benefit, this needs a gxformat2 release and a planemo release that picks it up.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 42.45%. Comparing base (28273b0) to head (9548c45).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #268   +/-   ##
=======================================
  Coverage   42.45%   42.45%           
=======================================
  Files          51       51           
  Lines        1974     1974           
  Branches      441      441           
=======================================
  Hits          838      838           
  Misses       1102     1102           
  Partials       34       34           

Flag	Coverage Δ
unittests	42.45% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.