JavaScript:SpiderMonkey:JS Tests: Difference between revisions
Jump to navigation
Jump to search
No edit summary |
m (Reverted edit of Mcripper, changed back to last version by Crowder) |
||
Line 1: | Line 1: | ||
== Rules for Writing JavaScript Test Cases == | |||
* Use reportCompare() to report success/failure! (except for e4x cases) | |||
* If your test will always throw an uncaught exception then you should name it with a suffix -n; it is always preferable to catch and handle exceptions directly. | |||
* Don't use print(), use printStatus() or writeLineToLog(), etc. | |||
* The tests will also be run in the browser in addition to the shell. If you use shell only features, detect that and let the browser report success. | |||
* Do the converse, if you have a browser-only test. | |||
* Each sub-suite has its own style -- use the template.js in that sub-suite's directory to as a starting-point for new tests. | |||
* Each new sub-suite directory must have an empty shell.js and browser.js | |||
When in doubt about this, ask for help. |
Latest revision as of 19:27, 27 November 2006
Rules for Writing JavaScript Test Cases
- Use reportCompare() to report success/failure! (except for e4x cases)
- If your test will always throw an uncaught exception then you should name it with a suffix -n; it is always preferable to catch and handle exceptions directly.
- Don't use print(), use printStatus() or writeLineToLog(), etc.
- The tests will also be run in the browser in addition to the shell. If you use shell only features, detect that and let the browser report success.
- Do the converse, if you have a browser-only test.
- Each sub-suite has its own style -- use the template.js in that sub-suite's directory to as a starting-point for new tests.
- Each new sub-suite directory must have an empty shell.js and browser.js
When in doubt about this, ask for help.