Meego Wiki
Views

Quality/TestSuite/MCTS/MCTS Development Guideline

From MeeGo wiki
< Quality | TestSuite | MCTS
Revision as of 13:57, 9 June 2010 by Jwang (Talk | contribs)
Jump to: navigation, search

Contents

MeeGo Core Test Suite(CTS) Development Guideline

Test Case Naming

  • Show the purpose clearly, avoid simply use number.
  • Name length should less than 30
  • Use '_' to connect words

Test Definition

File header

  • Copyright
  • License
  • Author
  • Changelog

Check list

  • README
  • src/ : contain source files
  • utils/: contain utility and tools if any
  • Makefile
  • tests.xml
  • data/: contain small data files (For big data such as media content, need separate package)
  • pack.sh: the script to make package
  • test.spec: spec files to create package

Makefile (??? If need automake)

  • make : build test package.
  • make install: install test binary or script into /usr/bin, tests.xml into /usr/share/<testpackagename>/tests.xml
  • make clean: remove all object file and tempory files
  • make uninstall: remove test binary or scripts, tests.xml under target installation directory.
  • make package: create rpm for current test package

Coding Style

Test Case Design

  • Common Rule:
    • Automate test under condition of stability
    • Main function should keep simple and structure keep neat.
    • Define common function file shared between test cases.
    • Have no dependency on other test case.
    • Have no dependency on UX or vertical specific apps.
    • Have no dependency on test framework.
    • Platform irrelevant
    • Clean environment after test complete.
    • Avoid hard coding - String, Digital
    • If parameters number greater than 3, use getopts
    • Suitable comments
  • Test Level Specific Rules:
    • Thorough
      • All APIs should be covered.
      • A test case targets an usage (API combination) instead of an API itself
      • Test need cover param pairs with equivalence partitioning and boundary condition (All-pairs optional)
    • Average
      • Legacy or low risk APIs are covered only with positive tests
    • Lightweight
      • Test most critical functionalities
      • With most typical parm-pairs
Personal tools