2013/5/8 Peter FELECAN <[email protected]>: > "Maciej (Matchek) Bliziński" <[email protected]> writes: > >> 2013/5/8 Peter FELECAN <[email protected]> >>> >>> I intend to use Qt Core for this. >> >> Do you mean you'd write this in C++? > > Yes. The result can run on all our servers, isn't it?
Sure. >> Qt Core[1] looks interesting. I didn't know QT had non-gui utility classes. > > In my case this is the main usage of Qt: writing servers, back ends, > ETL, &c. > >> What testing framework would you use? QtTest? > > A very good candidate. Not that I need a special framework to write > unit, system and integration tests. However, if you think that using a > test framework should be in the specification, I will gladly conform. Usually you end up with a framework, even if you start without one, so I think picking one right at the start is a good idea. The top-level code could be something like: GetLock(); disk_state = GetDiskState(); db_state = GetDbState(); disk_operation_list = ComputeDiskOperationList(disk_state, db_state); ApplyDiskOperations(disk_operation_list); ReleaseLock(); The Get*State() methods would interact with the environment, so I wouldn't expect much tests for that part. The ComputeDiskOperationList() function on the other hand would do the critical work, so I would expect that there would be a set of sample inputs and expected outputs. This way we'll be able to add support for newly discovered corner cases while ensuring that all the previous cases are still working correctly. Maciej _______________________________________________ maintainers mailing list [email protected] https://lists.opencsw.org/mailman/listinfo/maintainers .:: This mailing list's archive is public. ::.
