When there is a new build we perform smoke testing,in few projects they call it as sanity testing.Can someone please elaborate on the difference between the both.
Tell me more
×
Software Quality Assurance & Testing Stack Exchange is a question and answer site for
software quality control experts, automation engineers, and software testers. It's 100% free, no registration required.
|
|
|
sanity testing means we have to go through every main part in detail we can avoid parts who have low severity and in smoke testing we have to check each and every part in cursory way...each and every parts should be tested |
|||
|
|
|
Sanity and Smoke testing are conflated. They are not the same and they have a different purpose! Smoke tests determines whether it is possible to "continue" testing and sanity tests determines whether it is "reasonable" to test.
where as
As Joe suggests, they are often used as a gate to make sure that time isn't wasted on testing a build that is fundamentally flawed. |
||||
|
|
To remember easily, Smoke test - you just test all pages are accessed without errors and whether input is submitted and you get results. You do this across the application. Sanity test - You take 1 ore 2 modules. probably the more no. of fixes it has or the important one and test it deep by going through all functions. You understand that you do this only to accept the build for further testing, ie. regression, system, etc. |
|||
|
|
|
Just to muddy the waters even more I'm going to throw in "it depends". If you're looking for a practical answer within your own company then I'd ask them, as they are terms that change for different domains. Sometimes they are the same, sometimes they differ, and often they are defined at the whim of whomever writes the tests. Stuartf's answer is as close to an "offical" definition as you will get. A little history and context will help display the difference: Smoke tests refer orignally to hardware testing, wherein a device would be plugged in and if it started to smoke then it was unplugged as no further testing was required. It doesn't work, there is no point in testing it. Sanity tests are, traditionally (in other realms of testing) tests that establish whether or not a result could be true, rather than if it is true, that is to say if the output is "sane". If you put "9" into the function "squareroot()" and get "badger" you have failed a sanity test. If you get, say, a positive number less than 9 then you have passed it, even if the number is not "3". This is why they refer to tests that ensure that the system is working enough to be tested - it is giving a sane response to basic tests, therefore it is not a waste of time to run full functional tests. The confusion comes when lost in the complexity of software - at what point is the system testable? How can we see if smoke is coming out of it when it only breaks when a certian section of the code is run? It differs from project to project, evidently. Also in order to do a smoke test you need to use your testing framework (manual or otherwise) to actually check if the system is running and responsive; so in many cases they ARE a subset of functional tests. So the definitions are different, but are (often erroneously) used interchangably on a practical level. Hopefully knowing the context will give some meaning to the words, even if they both reduce (practically and contextually) to "give it a kick and see what happens". Their purpose is similiar in software testing, to save time on running a full regression suite and finding out that the system falls over because of a base error. The reason being that a basic error in the system has to be fixed, and that fix affects the system, and the testing should then be restarted, so the quicker that larger, high priority errors can be found the better. For example a web application may be spidered for 404 errors and broken links before long, slow, functional tests are run. Hope that's helpful! |
|||||||||||||||||
|
|
In most shops, there is no difference. Often, they are two names for the same process. In both cases, they are a subset of all your test cases, covering the main functionality of your system, to ascertain that the most crucial features are working, but not bothering with finer details. Many times the smoke test/sanity test is used as a gateway, through which the build must pass before it's deemed worthy of more in-depth testing. If the build fails this test, it goes back to the dev team for fixing. If it passes, the test team then digs in. |
|||||||||||||||
|
