Re: Re: cvs: pear /MDB2/tests MDB2_usage_testcase.php
| From: | Lukas Kahwe Smith | Date: | Sun, 20 Aug 2006 18:06:14 +0000 |
| Subject: | Re: Re: cvs: pear /MDB2/tests MDB2_usage_testcase.php | ||
| References: | 1 2 3 | Groups: | php.pear.dev |
| Request: | Send a blank email to pear-dev+get-43744@lists.php.net to get a copy of this message | ||
Daniel Convissor wrote:
Hey Lukas: On Sun, Aug 20, 2006 at 12:22:41PM +0200, Lukas Kahwe Smith wrote:Not sure I understand what you are saying. MDB2 expects things to be portable. The test suite aims to cover this. It runs through most of the API, including some edge cases, and compares the actual with the expected data. The question was about if I should also have tests where the necessary portability features are disabled which lead to inconsistent behavior between RDBMS. All these tests would illustrate would be for example: RDBMS xyz does not automatically RTRIM data coming from a CHAR field. RDBMY abc does not convert empty strings to NULL. However like I said if people are interested in this they can just run the test suite with the portability option set to MDB2_PORTABILITY_NONE and check which test cases fail. regards, LukasThere is the question of how to deal with test cases that essentially just illustrate the default behavior of the given RDBMS without any portability handling.It's not just illustrative. It's important to test that the expected default behavior is actually happening. For example, say I inadvertently put in some portability code where I shouldn't have. There should be a test showing that goof. Isn't it possible to include expected values for each DMBS?