|
|
||||||
|
#1
02-15-08, 11:50 PM
|
|
|
|
|
Sorry if this is a newbie question.
My application sends data (via TCP/IP sockets) to a 3rd party application which dispenses the data to various tables in a SQL Server 2005 database based on its own business logic. I am in the process of writing Unit tests for my application that mimics the sending of this data. However on test Teardown I would like to delete this test data so that next time I run the tests, the database is in a known state. I can think of a few options, none of which sound ideal to me: 1. Run the profiler on the database and note all the tables it touches and delete the those rows in the reverse order in which it was inserted. This is less than ideal because insertion is happening via stored complex procedures and it would be an arduous task going through the code of all the stored procs and coming up with a list such the referential constraints are not violated on deletion. 2. Backup and Restore the database.This is quite a time consuming operation and would be a huge bottleneck to do that everytime a test suite is run. I liked the article by Roy Osherove (http://msdn.microsoft.com/msdnmag/ issues/05/06/UnitTesting/#S5) but unfortunately I do not have control over the transaction because the transaction was commited by that different third party application. Is there some other mechanism by which I can tell SQL Server: "Note the state of the database now, and essentially rollback to it when I tell you so (in test teardown)"? Thanks. |
|
|
|
#2
02-16-08, 09:24 AM
|
|
|
|
|
On 16 ÆÅ×, 01:50, sprash <spras> wrote:
[..] > I liked the article by Roy Osherove ([..] > issues/05/06/UnitTesting/#S5) but unfortunately I do not have control > over the transaction because the > transaction was commited by that different third party application. > > Is there some other mechanism by which I can tell SQL Server: "Note > the state of the database now, and essentially rollback to it when I > tell you so (in test teardown)"? > > Thanks. Hi, I think you describe an integration test. In some cases it is possible to design it in such a way that test doesn't depend on data already in database. It just creates new data. If you just need to test sending functionality I think you need to mock your 3rd party component in some way. HTH, Sergey |
|
|
| Similar Threads | |
| Running Unit Test Remotely Hi, Some of my future unit / component test may run op to several hours. I would like to run them simultaneously and on a different machine. Arguably, I misuse the unit... |
|
| Restoring state of database after running a Unit test Sorry if this is a newbie question. My application sends data (via TCP/IP sockets) to a 3rd party application which dispenses the data to various tables in a SQL Server 2005... |
|
| Why does running a Test::unit test in irb work so weirdly? Why can't I run a Test::Unit::TestCase from IRB and have it run before I quit? I have found this problem mentioned by others. For an example, see this tutorial about... |
|
| detecting running under UNIT TEST Hi there, I use Visual Studio 2005, c# projects... I like the new unit test framework very much, it helps a lot mantaining and growing up a project with the test driven... |
|
| Running one test with test/unit Hi, As per this page[1], it is possible to run one test from a testcase by using.. # ruby [programfile] -n [testmethod] I'm probably missing something obvious but I really... |
|
|
All times are GMT. The time now is 08:59 AM. | Privacy Policy
|
Restoring state of database after running a Unit test