Startup Tasks

Before you start using the Intel® Inspector, determine if you want to change the default settings for:

About Choosing Baseline Results for Setting States

To avoid investigating the same issues over and over again, set the Intel Inspector to propagate state information from a baseline result the first time you open a new result.

Baseline Result Options

By default, the Intel Inspector propagates state information from the immediately previous result of the same analysis type. However, you may need to configure the Intel Inspector to propagate state information from a specific earlier result. This is most appropriate if your source code is undergoing branched development. You can also instruct the Intel Inspector to not propagate states from a baseline result, which sets the state of all issues in a new result to New.

Typical Usage Model

  1. Use the default Get problem states from previous result of same analysis type baseline result option.

  2. Run an analysis.

  3. Use the filtering function to temporarily limit the displayed issues to only those that are Not investigated.

  4. Set the state of each problem issue you investigate as:

    • Confirmed - Issue requires fixing but has not yet been fixed.

    • Fixed - Issue requires fixing and has been fixed.

    • Not a Problem - Issue does not require fixing.

    • Deferred - Delay further investigation.

  5. The next time you rerun the analysis, verify the issues you expect to be fixed are indeed fixed.

Intel Inspector propagates state information from the baseline result when it determines an issue in a new result corresponds to an issue in the baseline result. For example, if you set the state for a problem in the baseline result to Not a Problem, the Intel Inspector sets the state for the corresponding problem in the new result to Not a Problem.

Caution

Intel Inspector may not recognize an issue as previously investigated when it propagates state information from a baseline result. This is more likely to happen when source code has undergone drastic changes between analysis runs, or when there are several instances of the same problem type in a row in a static analysis result.

Tip

  • Do not delete a result when you are finished interpreting it. You may need it to properly propagate state information to later results.

  • Intel Inspector also propagates annotations in static analysis results.

About Naming Results and Result Directories

You can control - at a cross-project level - how the Intel Inspector names future results and result directories. For example, you may choose a cryptic name for results you consider temporary and plan to delete, or you may choose a more descriptive name for long-term audit purposes.

About Displaying Application Output

When you run a dynamic analysis on an application, the Intel Inspector executes that application. You can send application output to the:

  • Collection Log window

  • Separate console window

About Automatically Launching a Debugger During Interactive Debugging Analysis

When you run an analysis configured for interactive debugging, the Intel Inspector halts application execution and, depending on how the analysis was started, launches a debugging session immediately or when a problem is detected. By default, the Intel Inspector launches the GNU gdb* debugger. You can choose from several debuggers.

For more complete information about compiler optimizations, see our Optimization Notice.