Debuggers

Debuggers

Advantages/Disadvantages

• Advantages
  • Single step through the code
  • Stop execution at a given point to investigate where it goes and what the values are
  • Attach to an already running program
• Disadvantages
  • Not running real-time, so may not expose all problems

DDD

How to Compile for Debug

Compile with the option: -ggdb -O0

For the statgen library, you can compile with: make OPTFLAG="-ggdb -O0"

How to Start DDD

On a linux command line, type: ddd pathToYourExecutable/yourExecutable &

Do not specify command line options at this point. That will be done later when you type "run"

How to Use DDD

Debugging Tools

• Run: start the program with the previously specified options
• Interrupt: stop the program from running
• Step: Go into the function call (or go to next line of code)
• Next: Go over a function call (execute it, but do not step into it)
• Finish: Continue execution until the end of the current method
• Cont: Continue execution until the next breakpoint or the end of the program is reached.
• Kill: stop the program from running.

Basic Usage

Bring up a file in the viewer

• At the gdb prompt:

  l <filename>:<line#>

  -or -

  l <class>::<method>

Examples:

l SamRecord.cpp:1

l SamRecord::getCigar

Set a breakpoint

• Use mouse right-click on the line number

  Set Breakpoint (can set properties – break after hit X number of times, etc)

-or-

• At the gdd prompt:

  b <class>::<method>

Attach to already running process

• On the Menu Bar:

  File->Attach to Process

Run with options

• At the gdb prompt run your program with options, replacing your executable name with "run":

  run <options>

Backtrace (see where you are in execution, look up/down the call stack)

• On the Menu Bar:

  Status->Backtrace

See a variable's value

• Right click the variable in the source code window and click “Print” (or to keep it tracked, click “Display”)

-or-

• At the gdb prompt:

  p <variable>

• To Print in hex, At the gdb prompt:

  p/x <variable>

Other Testing Advice

• Reduce test size from one that takes hours to one that is much quicker.
  • Reduce file sizes
  • Turn off unnecessary sections
• Write a set of automated tests that test the different cases so they can be re-run each time the library changes
• When you find a bug, write a test that exposes the bug (fails), fix the bug, rerun the test (succeeds)