Introduction to Static Analysis

From Foss2Serve
Revision as of 08:46, 1 August 2016 by Kevin.buffardi (Talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search
Title

Introduction to Test-Driven Development (TDD)

Overview

Participants write a unit test (in GoogleTest) and corresponding C++ code to pass the test to verify the test passes

Prerequisite Knowledge

GoogleTest needed to have been installed (as instructed in Unit Test With GoogleTest Activity) and intermediate C++ expertise.

Learning Objectives

Participant should be able to learn how to follow "Red light, green light" Test-Driven Development process and reflect over the approach.

Directions:

Students:

  • In this exercise, we will set up and start exploring a static analysis tool packaged with the Clang compiler.
Downloading Clang and dependencies =

For OSX

  • You already have Clang and all the other dependencies installed. However, you need to install the Static Analyzer by downloading the packaged build of checker here.
  • After fully downloaded, I recommend then extracting the files from the tarball (checker-278.tar.bz2) as a new directory in your /usr/local folder, such as /usr/local/checker
  • If you are using sh, ksh, or bash shell, type the command: export PATH=$PATH:/usr/local/checker-278/bin but replace /usr/local/checker-278/bin with the location of where the static analyzer was installed, followed by /bin since that is where the binary (executable) files are

For *nix

  • You will have to first download and build Clang and LLVM by following these directions.

Once you are done, follow the instructions to build the analyzer from source.

Running scan-build

  • Navigate to where you stored (or freshly clone) your ConnectX code
  • To start with a fresh build, first make clean

Run: scan-build make

If you get a command not found error, your PATH is not correctly set (see instructions above).

Once run, it should take a moment to analyze the code. Does it identify any bugs? What does the scan-build output results means?

Exploring bug types

  • Navigate to your Minefield project or make a new clone of the github repo so you have it locally
  • Run scan-build on the Minefield project and confirm that it has no bugs found
  • Read over the different checks that the scan-build static analyzer available checks.
  • Within the Field class, try to create a bug as described by one of the available checks. However, don't just use the example code the documentation provides. Instead, try to incorporate one of those types of errors into the Field.cpp file so that if someone else looked at the code they might not immediately recognize the bug. Try to make a bug but "hide" it in a clever way.
  • Once you've created a "hidden bug," run scan-build on the project and make sure it identifies the bug that you think it should.

Deliverables:

A revised C++ class with a bug injected into it

Assessment:

The bug the student created should force a static analysis message if run again

Additional Information:

Knowledge Area/Knowledge Unit

Softward Engineering

Topic

Static Analysis

Level of Difficulty

Understanding types of static analysis messages and creating a new bug requires intermediate programming expertise

Estimated Time to Completion

Installation may take up to 30 minutes for computers running *nix operating systems and is suggested to assign for homework before class. The rest of the activity should take 15-30 minutes, not counting discussion.

Materials/Environment

Need to have a working C++ development environment

Author

Kevin Buffardi

Source

Kevin Buffardi

License

Licensed CC BY-SA 4.0



This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License

CC license.png

Personal tools
Namespaces
Variants
Actions
Events
Learning Resources
HFOSS Projects
Evaluation
Navigation
Toolbox