regression-tests
From the beginning of a project - already in the phase of specification - we need good test-sets. With test-sets we specify and describe the functionality, the input and the output of a software-process. Good test-sets are the treasure in the software-lifecycle.
So it is indispensable, to adapt test-cases to meet the requirements of a modified-software and to protect and preserve our treasure.
May be that sound's simple but every expert knows:
The real trick is, to solve this in an efficient and not at all financial affordable manner.
Therefore in our customer's projects we use our own proofed batch-test-software.
With this batch-test-software, we are able to
- specify input and output, describe the processing-rules.
- define the location of input and output (files, database-tables, ...).
- print test-cases with input and output as pdf-documents.
- modify input- and output-structures, without loosing test-data.
- compare the actual output with the specified output.
- execute regression-tests.
- group test-cases to test-groups.
- define simple and complex comparing rules.
- exclusion of fields from comparing; for example:
- timestamps, batch-numbers
- fields, content of them has changed because of software-change
- define joins between record-structures.
- problem: join-fields are often automatic numbers, which cannot be simply compared
- define comparison-rules for xml-data.
- exclusion of fields from comparing; for example:
The xml-comparison-component of our batch-test-software is available for customers as a standalone-tool called <xml>cmp.