Main Content

Run Polyspace Analysis by Using MATLAB Scripts

You can automate the analysis of your C/C++ code by using MATLAB® scripts. In your script, you specify your source files and analysis options such as compiler, run an analysis, and read the analysis results to MATLAB tables.

For instance, use this script to run a Polyspace® Bug Finder™ analysis on a sample file: 

proj = polyspace.Project

% Specify sources and includes
sourceFile = fullfile(polyspaceroot, 'polyspace', ... 
    'examples', 'cxx', 'Bug_Finder_Example', 'sources', 'numerical.c');
includeFolder = fullfile(polyspaceroot, 'polyspace', ... 
    'examples', 'cxx', 'Bug_Finder_Example', 'sources');

% Configure analysis
proj.Configuration.Sources = {sourceFile};
proj.Configuration.TargetCompiler.Compiler = 'gnu4.9';
proj.Configuration.EnvironmentSettings.IncludeFolders = {includeFolder};
proj.Configuration.ResultsDir = fullfile(pwd,'results');

% Run analysis
bfStatus = run(proj, 'bugFinder');

% Read results
resObj = proj.Results;
bfSummary = getSummary(resObj, 'defects');

See also polyspace.Project.

Prerequisites

Before you run Polyspace from MATLAB, you must link your Polyspace and MATLAB installations. See Integrate Polyspace with MATLAB and Simulink.

Specify Multiple Source Files

You can specify a folder containing all your source files. For instance, if proj is a polyspace.Project, object, enter:

sourceFolder = fullfile(polyspaceroot, 'polyspace', ... 
    'examples', 'cxx', 'Bug_Finder_Example', 'sources');
proj.Configuration.Sources = {fullfile(sourceFolder,'*')};
You can also specify multiple source folders in the cell array.

You can specify a folder that contains all your source files both directly and in subfolders. For instance:

sourceFolder = fullfile(polyspaceroot, 'polyspace', ... 
    'examples', 'cxx', 'Bug_Finder_Example', 'sources');
proj.Configuration.Sources = {fullfile(sourceFolder,'**')};

If you do not want to analyze all files in a folder, you can explicitly specify which files to analyze. For instance:

sourceFolder = fullfile(polyspaceroot, 'polyspace', ... 
    'examples', 'cxx', 'Bug_Finder_Example', 'sources');
file1 = fullfile(sourceFolder,'numerical.c');
file2 = fullfile(sourceFolder,'staticmemory.c');
proj.Configuration.Sources = {file1, file2};

You can explicitly exclude files from analysis. For instance:

% Specify source folder.
sourceFolder = fullfile(polyspaceroot, 'polyspace', ... 
    'examples', 'cxx', 'Bug_Finder_Example', 'sources');
proj.Configuration.Sources = {fullfile(sourceFolder,'**')};

% Specify files to exclude.
file1 = fullfile(sourceFolder,'security.c');
file2 = fullfile(sourceFolder,'tainteddata.c');
proj.Configuration.InputsStubbing.DoNotGenerateResultsFor = ['custom=' file1 ...
     ',' file2];
However, this method of exclusion does not apply to Code Prover run-time error checking.

Check for MISRA C:2012 Violations

You can customize the Polyspace analysis to check for MISRA C™:2012 rule violations.

Set options for checking MISRA C:2012 rules. Disable the regular Bug Finder analysis, which looks for defects.

If proj is a polyspace.Project object, to run a Bug Finder analysis with all mandatory MISRA C:2012 rules, enter:

% Enable MISRA C checking
proj.Configuration.CodingRulesCodeMetrics.EnableMisraC3 = true;
proj.Configuration.CodingRulesCodeMetrics.MisraC3Subset = 'mandatory';

% Disable defect checking
proj.Configuration.BugFinderAnalysis.EnableCheckers = false;

% Run analysis
bfStatus = run(proj, 'bugFinder');

% Read summary of results
resObj = proj.Results;
misraSummary = getSummary(resObj, 'misraC2012');

Check for Specific Defects or Coding Rule Violations

Instead of the default set of defect or coding rule checkers, you can specify your own set.

If proj is a polyspace.Project object, to disable MISRA C:2012 rules 8.1 to 8.4, enter:

% Disable rules
misraRules = polyspace.CodingRulesOptions('misraC2012');

misraRules.Section_8_Declarations_and_definitions.rule_8_1 = false;
misraRules.Section_8_Declarations_and_definitions.rule_8_2 = false;
misraRules.Section_8_Declarations_and_definitions.rule_8_3 = false;
misraRules.Section_8_Declarations_and_definitions.rule_8_4 = false;

% Configure analysis
proj.Configuration.CodingRulesCodeMetrics.EnableMisraC3 = true;
proj.Configuration.CodingRulesCodeMetrics.MisraC3Subset = misraRules;

See also polyspace.CodingRulesOptions.

To enable Bug Finder defects, use the class polyspace.DefectsOptions. One difference between coding rules and defects class is that coding rule checkers are enabled by default. You disable the ones that you do not want. All defect checkers are disabled by default. You enable the ones that you want.

You can also specify a coding standard XML file that enables coding rules from different standards. When checking for coding rule violations, you can refer to the file. For instance, to use the template XML file StandardsConfiguration.xml provided with the product in the subfolder polyspace\examples\cxx\Bug_Finder_Example\sources, enter:

pathToTemplate = fullfile(polyspaceroot,'polyspace','examples',...
    'cxx','Bug_Finder_Example','sources','StandardsConfiguration.xml');
proj.Configuration.CodingRulesCodeMetrics.EnableMisraC3 = true;
proj.Configuration.CodingRulesCodeMetrics.MisraC3Subset = 'from-file';
proj.Configuration.CodingRulesCodeMetrics.EnableCheckersSelectionByFile = true;
proj.Configuration.CodingRulesCodeMetrics.CheckersSelectionByFile = pathToTemplate;

Find Files That Do Not Compile

If one or more of your files contain a compilation error, the analysis continues with the remaining files. You can choose to stop analysis on compilation errors.

If proj is a polyspace.Project object, to stop analysis on compilation errors, enter:

proj.Configuration.EnvironmentSettings.StopWithCompileError = true;

However, it is more convenient to let the analysis complete and capture all compilation errors from the analysis log file. For more information, see Troubleshoot Polyspace Analysis from MATLAB.

Run Analysis on Server

You can run an analysis on a remote server instead of your local desktop. Once you have set up connection to a server, you can run the analysis in batch mode. For setup information, see Install Products for Submitting Polyspace Analysis from Desktops to Remote Server.

Specify that the analysis must run on a server. Specify a folder on your desktop where results are downloaded after analysis. If proj is a polyspace.Project object, to configure analysis on a server, enter:

proj.Configuration.MergedComputingSettings.BatchBugFinder = true;
proj.Configuration.ResultsDir = fullfile(pwd,'results');

Specify the head node that manages the Polyspace jobs:

proj.Configuration.Advanced.Additional = '-schedular nodeHost'

Run analysis as usual. 

run(proj, 'bugFinder');

Open the results from the results folder location.

pslinkfun('openresults', '-resultsfolder', proj.Configuration.ResultsDir);
If the analysis is complete and the results have been downloaded, they open in the Polyspace user interface.

See Also

| |

Related Topics