自动化框架的两种断言设计(pytest 版)

自动化测试断言失败时，根据不同业务场景，可能需要立即终止或继续执行。这里以 Appium + pytest 为例。

一. 断言失败立即终止

用途一：用例的预期结果是其他用例的前提条件时，assert 失败需要立即终止，这是我们设计自动化测试用例时最常见的场景。

用途二：用例中的任何一个步骤执行失败时，立即终止，因为步骤都执行失败了，没有继续执行下一步的必要。方案：思路与用途一类似，这里把每个测试步骤当做都有一个预期结果(True)，封装每一个具体步骤，步骤结果返回布尔值。对该结果进行 assert，False时立即终止。

 def click(self, step):
     method_name = sys._getframe().f_code.co_name
     try:
         element = self.find_element(**step['element_loc'])
         if step.has_key('times'):
             for i in range(step['times']):
                 element.click()
         else:
             element.click()
         return True
     except:
         print u'%s失败' % method_name
         return False

 for step in self.case_steps:
     assert self.operate(step), 'error in step: %s' % step

二. 断言失败继续执行

主要使用了两个函数 expect, assert_expectations 。

Demo: test_delayed_assert.py

 from delayed_assert import expect, assert_expectations

 def test_should_pass():
     expect(1 == 1, 'one is one')
     assert_expectations()

 def test_should_fail():
     expect(1 == 2)
     x = 1
     y = 2
     expect(x == y, 'x:%s y:%s' % (x,y))
     expect(1 == 1)
     assert_expectations()

Module: delayedAssert.py

 '''
 Implements one form of delayed assertions.

 Interface is 2 functions:

   expect(expr, msg=None)
   : Evaluate 'expr' as a boolean, and keeps track of failures

   assert_expectations()
   : raises an assert if an expect() calls failed  

 Usage Example:

     from expectations import expect, assert_expectations

     def test_should_pass():
         expect(1 == 1, 'one is one')
         assert_expectations()

     def test_should_fail():
         expect(1 == 2, 'one is two')
         expect(1 == 3, 'one is three')
         assert_expectations()
 '''

 # ---------------------------------------------------

 def expect(expr, msg=None):
     'keeps track of failed expectations'
     if not expr:
         _log_failure(msg)

 def assert_expectations():
     'raise an assert if there are any failed expectations'
     if _failed_expectations:
         assert False, _report_failures()

 # ---------------------------------------------------

 import inspect
 import os.path

 _failed_expectations = []

 def _log_failure(msg=None):
     (filename, line, funcname, contextlist) =  inspect.stack()[2][1:5]
     filename = os.path.basename(filename)
     context = contextlist[0]
     _failed_expectations.append('file "%s", line %s, in %s()%s\n%s' %
         (filename, line, funcname, (('\n%s' % msg) if msg else ''), context))

 def _report_failures():
     global _failed_expectations
     if _failed_expectations:
         (filename, line, funcname) =  inspect.stack()[2][1:4]
         report = [
             '\n\nassert_expectations() called from',
             '"%s" line %s, in %s()\n' % (os.path.basename(filename), line, funcname),
             'Failed Expectations:%s\n' % len(_failed_expectations)]
         for i,failure in enumerate(_failed_expectations, start=1):
             report.append('%d: %s' % (i, failure))
         _failed_expectations = []
     return ('\n'.join(report))

 # ---------------------------------------------------
 # _log_failure() notes
 #
 # stack() returns a list of frame records
 #   0 is the _log_failure() function
 #   1 is the expect() function
 #   2 is the function that called expect(), that's what we want
 #
 # a frame record is a tuple like this:
 #   (frame, filename, line, funcname, contextlist, index)
 # we're mainly interested in the middle 4,
 # ---------------------------------------------------

参考：Delayed assert / multiple failures per test