Figure 1. Test Execution Environment
The software under test must be verified
at different levels. The first level of software
verification is the unit level. A unit is defined
to be the smallest compilable piece of software.
At this level an individual module is subjected
to test cases developed from the requirement
model. The unit under test is placed in an
environment that will extract the module's
specification and generate a driver capable of
calling the unit under test. All modules that are
called by the Module Under Test (MUT) will
be automatically stubbed if they are not
available. The Graybox methodology supports
stub data value return through the MTIF file.
The Graybox methodology supports automated
code generation of all test unique software
(drivers, stubs, results verification). During
software integration, the Graybox toolset is
once again invoked to generate the driver to
test the module in relationship to its
subordinate modules.
As the module under test completes one
invocation and returns, the data values on the
module's interface are verified against the
requirement based test case verification data.
"Comparison of the algebraically couched
results of symbolic testing with external
specifications given in similar terms constitutes
an exact verification process" [Dunn84]. This
process is repeated for each test case in the test
case input file. In the Graybox methodology
the tester never steps inside the module under
test. All data is viewed external to the module
under test.
The module is probed after being
introduced into the test environment. This allows the test environment to gain visibility to
the path coverage information. A module is
considered tested when it has passed all
requirement-based test cases and has traversed
all paths. Test verification is given a score of
one and every path from top to bottom of the
module has a score of one. Combining the
verification and path coverage scores gives a
test completion score. A module that has
passed all requirement-based test cases and did
not traverse all paths has code that has not been
executed by any test case. "In its most
primitive form, the monitor will report, at the
conclusion of each dynamic test, the decisionto-
decision paths that have been exercised. A
static analysis system automatically
instruments the source code with probes for the
gathering of branch execution statistics"
[Dunn84]. Once the score reaches loo%, the
module is deemed functionally and structurally
tested. The Graybox methodology includes
steps to capture and verify code coverage.
"Coverage monitors are aids that are applicable
to white-box testing activities" [Myers79]. The
Graybox methodology can be used for both
Top Down (stub generation) and Bottom Up
Testing (driver generation). A module can be
integration tested by replacing all stubbed
modules with the actual modules when they are
coded and repeating the test case execution
step.
Figure 1. Test Execution Environment
The software under test must be verified
at different levels. The first level of software
verification is the unit level. A unit is defined
to be the smallest compilable piece of software.
At this level an individual module is subjected
to test cases developed from the requirement
model. The unit under test is placed in an
environment that will extract the module's
specification and generate a driver capable of
calling the unit under test. All modules that are
called by the Module Under Test (MUT) will
be automatically stubbed if they are not
available. The Graybox methodology supports
stub data value return through the MTIF file.
The Graybox methodology supports automated
code generation of all test unique software
(drivers, stubs, results verification). During
software integration, the Graybox toolset is
once again invoked to generate the driver to
test the module in relationship to its
subordinate modules.
As the module under test completes one
invocation and returns, the data values on the
module's interface are verified against the
requirement based test case verification data.
"Comparison of the algebraically couched
results of symbolic testing with external
specifications given in similar terms constitutes
an exact verification process" [Dunn84]. This
process is repeated for each test case in the test
case input file. In the Graybox methodology
the tester never steps inside the module under
test. All data is viewed external to the module
under test.
The module is probed after being
introduced into the test environment. This allows the test environment to gain visibility to
the path coverage information. A module is
considered tested when it has passed all
requirement-based test cases and has traversed
all paths. Test verification is given a score of
one and every path from top to bottom of the
module has a score of one. Combining the
verification and path coverage scores gives a
test completion score. A module that has
passed all requirement-based test cases and did
not traverse all paths has code that has not been
executed by any test case. "In its most
primitive form, the monitor will report, at the
conclusion of each dynamic test, the decisionto-
decision paths that have been exercised. A
static analysis system automatically
instruments the source code with probes for the
gathering of branch execution statistics"
[Dunn84]. Once the score reaches loo%, the
module is deemed functionally and structurally
tested. The Graybox methodology includes
steps to capture and verify code coverage.
"Coverage monitors are aids that are applicable
to white-box testing activities" [Myers79]. The
Graybox methodology can be used for both
Top Down (stub generation) and Bottom Up
Testing (driver generation). A module can be
integration tested by replacing all stubbed
modules with the actual modules when they are
coded and repeating the test case execution
step.
การแปล กรุณารอสักครู่..
