When test can stop _ test coverage rate

Every time a new version is shipped, it is often asked if the test is over. Do you have confidence in quality? What are the criteria you are basing on?
I think a lot of people will find it difficult to answer this question. Basically, you can decide whether the test can be completed according to the following five conditions.
1. Boss decides
Basically, the boss is invincible. He said when it was time. I think we will not, and dare not disagree. Xd
2. The team has a consensus to stop
When the team is finished talking, decide when to stop the test, which can also end
3. When the price is too high
If the cost of finding the next bug is outweighed by the loss of the bug, there is no need to do it again, and it can be done immediately.
4. If the percentage of bugs found falls to the desired target
Sometimes you will see how many bugs you find each time, and if you find that it has been falling and below your target, then you can ship. Like a bugs/per day below 5, and these bugs are not serious bugs
5. If the expected test coverage target has been achieved
How you measure your test case, how many things you have covered, and then you can tell if your test range is adequate. Like the% line coverage, 75% branch coverage and so on. When you reach a set goal, naturally you can also say that the test can be done.
At present, it looks like only the back two, compared with the data to refer to the first three comparison is free heart certificate. In fact, these are the words of peace of the word. Because as long as give you time and resources, in fact, can still find bugs.
Because, < 1 bugs/per day or 100% line/branch coverage, there is no guarantee. At most just on the book gives you confidence. After the shipment was caught problem, QA or wait to be scolded did not do well. It's not fair, but it's true. Xd
This article is from: 51Testing Software test Network http://www.51testing.com/html/46/n-863746.html