• Added note about interpreting error messages.

• Added StringContainsSC and TraversableContainsSC constraints. ...

• Tidied formatting and grammar of error strings.
• Made appropriate use of the *SC assert variants.

• Made running of data tests dependent on correct data types, columns lengths and nullability.

• Removed the modified PHPUnit Autoload.php and switched to explicit loading of custom constraint classes. ...

• Added staff/student mode conditional processing so that the more verbose constraints are used for marking.

• Overrode evaluate() to stop it throwing out a too-informative exception.

• Fixed placeholder typo in printf argument.

• Slightly tweaked output for numeric column length assertion.

• Implemented specialised “non-informative” subclasses of various constraints so that we can test assertions without giving away the answers to the students! (closes #13) ...

• Refactored a couple of variable names.
• Fixed punctuation in assertion output.

• Added customised version of TestCase.php in order to disable output buffering (closes #31 for now).

• Refactored output_stream to outputStream. ...

• Removed spurious extra default: in switch statement.

• Made data tests dependent on the result of testColumnNullability. Closes #11 for all practical purposes (you still get errors when later dependent table load there fixtures, but there isn’t much that can be done about that).

• Fixed incorrectly nested check for whether testPKColumns had succeeded (closes #7).

• Completely re-engineered the way that fixture setup and teardown is managed so that it can be enabled and disabled on demand (closes #3). ...

• Fixed a bug in SimpleTestListener that caused a crash when a table didn’t exist in the schema being tested.
• Added proper error reporting for missing tables.

• Added initial coding for testing UNIQUE constraints. ...

• Removed unused argument to sprintf in assertFKColumns().

• Added support for zero or negative scale values. ...

• Improved reporting of precision/scale vs. length for NUMBER vs. *CHAR.
• Added NVL() to Data_Scale to eliminate spurious errors.
• Added debug output.

• Added constants for verbosity levels. ...

• Added debug verbosity level and status.

• Made verbosity comparisons more explicit. ...

• Made Reporter->report() an abstract method; subclasses must now implement output themselves.
• Switched to explicitly writing to output streams for each subclass rather than just using print or echo. This is because CLI and HTML use different output streams anyway (STDOUT and php://output, respectively).

• Corrected argument name in comment.

• Added NVL() to query on Data_Precision to prevent spurious errors when precision isn’t specified for a NUMBER column (closes #2, I hope!).

Workaround for testing FKs when there is more than one FK in the referenced table.

Added automated true/false value setting to the marking configuration.

Added the ability to specify that a number column should include no length or decimal places.

Added support for alternative true/false values in PHP, placeholders for them in XML, and a quick and dirty shell script to set them.

Corrected reporting of overflow and underflow tests.

Fixed handling of literal text and date default values.

Added default value testing.

Rearranged PK testing code slightly.

Fixed sense of underflow constraint violation test.

Added underflow and overflow testing.

assertColumnLegalValue() now correctly fails if the legal value generates a constraint violation.