XML Test Case Import - sebiboga/testlink-upgraded GitHub Wiki

XML Test Case Import — REST BFF API

Endpoint: POST /api/testcasesimport/?action=import_xml&tproject_id=<id> Auth: TestLink session (cookie), permission mgt_modify_tc on target project Import engine: legacy lib/testcases/tcImport.php helpers (function defs extracted + eval'd in the BFF) Format reference: TestLink XML export format (<testcases> / <testsuite> root)

Request

Param Where Notes
tproject_id query target test project (falls back to session project)
containerID query parent suite id; defaults to project root
uploadedFile multipart file field the XML file to import
hit_criteria query name (default) | internalID | externalID
action_on_hit query skip (default) | update_last_version | generate_new | create_new_version
useRecursion query 1 = import nested <testsuite> structure (root must be <testsuite>)
intoProject query 1 = import suites under project root

Limits: file larger than import_file_max_size_bytes → HTTP 413; malformed XML → HTTP 422 JSON.

Semantics

  1. The uploaded file is saved to a temp file and validated with a safe parser (libxml_use_internal_errors + simplexml_load_string, LIBXML_NONET).
  2. Flat import (testcases/testcase root) → importTestCasesFromSimpleXML; the recursive import (testsuite root, useRecursion=1) → importTestSuitesFromSimpleXML.
  3. Duplicate handling follows hit_criteria + action_on_hit (legacy parity).
  4. Response normalizes the legacy flat [title, message] resultMap into a result array of {name, message} rows for a report table.

Example

curl -b cookies.txt -X POST \
  -F "[email protected]" \
  "http://host/api/testcasesimport/?action=import_xml&tproject_id=1&containerID=1&hit_criteria=name&action_on_hit=skip"

{"status":"ok","tproject":{"id":1,"name":"Demo Project"},"useRecursion":false,
 "intoProject":false,"options":{"hit_criteria":"name","action_on_hit":"skip"},
 "result":[{"name":"Login with valid credentials","message":"ok"},
           {"name":"Logout","message":"ok"}]}

Legacy parity notes (vs lib/testcases/tcImport.php importTestCaseDataFromXML)

Same permission model, size cap, hit criteria and action set. The legacy page mixed page-rendering code (config require, Smarty display) with the import helpers, so the BFF extracts only the pure function definitions (from function importTestCaseDataFromXML( onward) and eval's them — the helper libs (common.php, xml.inc.php, csv.inc.php) are already loaded by the BFF.

Unlike the legacy page (which die()s with raw HTML on malformed XML via simplexml_load_file_wrapper()), the BFF pre-validates and returns a clean HTTP 422 JSON error.

Modernized UI (gui/templates/testcases/tcImport.html)

The XML import path (submitLegacyXml → AJAX importXml) posts to the BFF above.

  • File type selector: XML (BFF action=import_xml) vs Markdown (.md) (BFF action=import_md).
  • Duplicate hit criteria (name / internal ID / external ID) + action for duplicates.
  • Report table shows per-row title + result message; counts chip for created rows.
  • Client-side size guard before upload; server 413 backup; malformed XML → error toast.

Import screen

XML import report

Malformed XML error toast

UI regression suite

Suite 819 in tmp/TLU_Test_Cases.md (render, XML flat import + DB verify, duplicate skip on re-import, recursive suite import, malformed XML → 422 toast, MD toggle regression, Event Viewer cleanliness).

⚠️ **GitHub.com Fallback** ⚠️