Subchapter 73.2
references/multi-configuration-runs.mdMarkdown2 KBView on GitHub
Read this before creating or reviewing any spec that carries a config_id. These runs need a manual
Testmo UI pass after the script finishes, and the pass is mandatory.
Some platforms ship several same-named runs distinguished only by Testmo Configuration (for example
one Mobile run per OS). Each variant is its own spec file with the same run_name and a different
config_id. Look config ids up via GET /projects/{id}/configs.
Testmo’s /cases API cannot filter by configuration and exposes no per-case config assignment, so
these are always two-step runs:
config_id._comment.Until step 2 is done the run holds the wrong case set. The script lists every affected run under ; pass that list on to the user.
MANUAL STEP 2 REQUIREDA common split is broad + narrow: one variant covers a whole folder, and the others are driven by a
tag that marks the essential cases. A narrow, tag-only spec has no folder or test-type filter behind it,
so verify the tag against GET /projects/{id}/tags before trusting it — a wrong or duplicated tag name
yields the wrong case set rather than an error.
"config_id": null is a deliberate placeholder meaning “this run needs a Configuration that has not
been looked up yet.” Dry-runs still work so case counts can be validated, but --create refuses, since
the variants would be indistinguishable.