Problem 1770. singularity 2.0 (hard)

Created by Alfonso Nieto-Castanon

Solve Later

{"relationships":[{"relationshipType":"http://schemas.mathworks.com/matlab/code/2013/relationships/document","targetMode":"","relationshipId":"rId1","target":"/matlab/document.xml"},{"relationshipType":"http://schemas.mathworks.com/matlab/code/2013/relationships/output","targetMode":"","relationshipId":"rId2","target":"/matlab/output.xml"}],"parts":[{"partUri":"/matlab/document.xml","relationship":[],"contentType":"application/vnd.mathworks.matlab.code.document+xml","content":"<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n<w:document xmlns:w=\"http://schemas.openxmlformats.org/wordprocessingml/2006/main\"><w:body><w:p><w:pPr><w:pStyle w:val=\"text\"/></w:pPr><w:r><w:t>This problem is the continuation of problem</w:t></w:r><w:r><w:t> </w:t></w:r><w:hyperlink w:docLocation=\"http://www.mathworks.com/matlabcentral/cody/problems/1769-singularity-2-0-easier\"><w:r><w:t>Singularity 2.0 (easier)</w:t></w:r></w:hyperlink></w:p><w:p><w:pPr><w:pStyle w:val=\"text\"/></w:pPr><w:r><w:t>If you have been in Cody long enough you have probably run across some badly constructed problems and test suites.</w:t></w:r></w:p><w:p><w:pPr><w:pStyle w:val=\"text\"/></w:pPr><w:r><w:t>This problem, I believe, represents an apparently impossible scenario. Yet, it is possible to solve it. Do you know how?</w:t></w:r></w:p><w:p><w:pPr><w:pStyle w:val=\"text\"/></w:pPr><w:r><w:rPr><w:b/></w:rPr><w:t>Description:</w:t></w:r></w:p><w:p><w:pPr><w:pStyle w:val=\"text\"/></w:pPr><w:r><w:t>This is this problem's testsuite:</w:t></w:r></w:p><w:p><w:pPr><w:pStyle w:val=\"code\"/></w:pPr><w:r><w:t><![CDATA[ myfunction();\n [a,b]=1; % oops...\n assert(isequal(a,b));]]></w:t></w:r></w:p><w:p><w:pPr><w:pStyle w:val=\"text\"/></w:pPr><w:r><w:t>As you may notice, the second line is not a proper assignment, so the testsuite will break at that point, and will return an error message:</w:t></w:r></w:p><w:p><w:pPr><w:pStyle w:val=\"text\"/></w:pPr><w:r><w:rPr><w:rFonts w:cs=\"monospace\"/></w:rPr><w:t>??? Too many output arguments.</w:t></w:r></w:p><w:p><w:pPr><w:pStyle w:val=\"text\"/></w:pPr><w:r><w:t>Typical hacks work by overloading some function in the evaluation code (e.g. assert hack). Unfortunately, as far as I know, assignment operators cannot be overloaded in Matlab. Is there something</w:t></w:r><w:r><w:t> </w:t></w:r><w:r><w:rPr><w:rFonts w:cs=\"monospace\"/></w:rPr><w:t>myfunction</w:t></w:r><w:r><w:t> could do to still solve this problem?</w:t></w:r></w:p><w:p><w:pPr><w:pStyle w:val=\"text\"/></w:pPr><w:r><w:rPr><w:b/></w:rPr><w:t>Hint</w:t></w:r><w:r><w:t>:</w:t></w:r></w:p><w:p><w:pPr><w:pStyle w:val=\"text\"/></w:pPr><w:r><w:t>You may use a similar trick as before but now targeting a vulnerability in the function</w:t></w:r><w:r><w:t> </w:t></w:r><w:r><w:rPr><w:rFonts w:cs=\"monospace\"/></w:rPr><w:t>verifyCode</w:t></w:r><w:r><w:t> instead of in the testsuite. For simplicity, here is the relevant portion of the verifyCode.m file (the function that evaluates and scores Cody problems):</w:t></w:r></w:p><w:p><w:pPr><w:pStyle w:val=\"code\"/></w:pPr><w:r><w:t><![CDATA[ function testSuiteResults = verifyCode( sourceFile, testFile )\n %VERIFYCODE Evaluates and verifies MATLAB Code against provided test suite.\n % TESTSUITERESULTS = VERIFYCODE(SOURCEFILE, TESTFILE) evaluates SOURCEFILE against\n % TESTFILE and returns the results.\n \n % Score it\n score = calculateSize(sourceFile);\n % Load the test file\n code = fileread(testFile);\n % Split it into cells\n code2 = regexprep(code,'\\n%%','\\nxxx-cellbreak-xxx%%');\n cellList = regexp(code2,'xxx-cellbreak-xxx','split');\n % Test each cell\n % Each testpoint structure\n % testPoint = struct(...\n % 'pass', false, ...\n % 'code', '', ...\n % 'output', '');\n testSuite = struct([]);\n [containsIllegalFcnFlag, illegalFcnMessage] = containsIllegalFcn(sourceFile);\n for countVariable = 1:length(cellList)\n cellCode = cellList{countVariable};\n cleanCellCode = cleanCode(cellCode);\n % Test is guilty until proven successful\n pass = false;\n % Run it\n try\n output = evaluateCode(cleanCellCode);\n pass = true;\n catch me\n output = ['Error: ' me.message];\n end\n if containsIllegalFcnFlag\n pass = false;\n output = illegalFcnMessage;\n end\n cellOutput = cleanCode(output);\n testSuite(countVariable).code = cleanCellCode;\n testSuite(countVariable).output = cellOutput;\n testSuite(countVariable).pass = pass;\n end\n % Set up data structure to return\n testSuiteResults = struct(...\n 'pass', false, ...\n 'score', 0, ...\n 'functions', [], ...\n 'testPoints', struct([]));\n % Populate the result struct\n % If any test point is failed, sets the overall pass status to false.\n testSuiteResults.pass = all([testSuite.pass]);\n % sets the score\n testSuiteResults.score = sprintf('%d',score);\n % Set test points\n testSuiteResults.testPoints = testSuite;\n % Set functions points\n testSuiteResults.functions = findFcns(sourceFile,'file');\n end]]></w:t></w:r></w:p><w:p><w:pPr><w:pStyle w:val=\"text\"/></w:pPr><w:r><w:t>Previous problem in this series:</w:t></w:r><w:r><w:t> </w:t></w:r><w:hyperlink w:docLocation=\"http://www.mathworks.com/matlabcentral/cody/problems/1769-singularity-2-0-easier\"><w:r><w:t>Singularity 2.0 (easier)</w:t></w:r></w:hyperlink></w:p><w:p><w:pPr><w:pStyle w:val=\"text\"/></w:pPr><w:r><w:t>Next problem in this series:</w:t></w:r><w:r><w:t> </w:t></w:r><w:hyperlink w:docLocation=\"http://www.mathworks.com/matlabcentral/cody/problems/1764-singularity-2-0\"><w:r><w:t>Singularity 2.0 (really hard)</w:t></w:r></w:hyperlink></w:p><w:p><w:pPr><w:pStyle w:val=\"text\"/></w:pPr><w:r><w:rPr><w:i/></w:rPr><w:t>Copyright 1984-2012 The MathWorks, Inc.</w:t></w:r></w:p></w:body></w:document>"},{"partUri":"/matlab/output.xml","contentType":"text/xml","content":"<?xml version=\"1.0\" encoding=\"UTF-8\" standalone=\"no\" ?><embeddedOutputs><metaData><evaluationState>manual</evaluationState><layoutState>code</layoutState><outputStatus>ready</outputStatus></metaData><outputArray type=\"array\"/><regionArray type=\"array\"/></embeddedOutputs>"}]}

This problem is the continuation of problem <a href = "http://www.mathworks.com/matlabcentral/cody/problems/1769-singularity-2-0-easier">Singularity 2.0 (easier)</a>If you have been in Cody long enough you have probably run across some badly constructed problems and test suites.This problem, I believe, represents an apparently impossible scenario. Yet, it is possible to solve it. Do you know how?Description:This is this problem's testsuite:<pre> %%
 myfunction();
 [a,b]=1; % oops...
 assert(isequal(a,b));</pre>As you may notice, the second line is not a proper assignment, so the testsuite will break at that point, and will return an error message:<tt>??? Too many output arguments.</tt>Typical hacks work by overloading some function in the evaluation code (e.g. assert hack). Unfortunately, as far as I know, assignment operators cannot be overloaded in Matlab. Is there something <tt>myfunction</tt> could do to still solve this problem?Hint:You may use a similar trick as before but now targeting a vulnerability in the function <tt>verifyCode</tt> instead of in the testsuite. For simplicity, here is the relevant portion of the verifyCode.m file (the function that evaluates and scores Cody problems):<pre> function testSuiteResults = verifyCode( sourceFile, testFile )
 %VERIFYCODE Evaluates and verifies MATLAB Code against provided test suite.
 % TESTSUITERESULTS = VERIFYCODE(SOURCEFILE, TESTFILE) evaluates SOURCEFILE against
 % TESTFILE and returns the results.
 % Copyright 1984-2012 The MathWorks, Inc.
 % Score it
 score = calculateSize(sourceFile);
 % Load the test file
 code = fileread(testFile);
 % Split it into cells
 code2 = regexprep(code,'\n%%','\nxxx-cellbreak-xxx%%');
 cellList = regexp(code2,'xxx-cellbreak-xxx','split');
 % Test each cell
 % Each testpoint structure
 % testPoint = struct(...
 % 'pass', false, ...
 % 'code', '', ...
 % 'output', '');
 testSuite = struct([]);
 [containsIllegalFcnFlag, illegalFcnMessage] = containsIllegalFcn(sourceFile);
 for countVariable = 1:length(cellList)
 cellCode = cellList{countVariable};
 cleanCellCode = cleanCode(cellCode);
 % Test is guilty until proven successful
 pass = false;
 % Run it
 try
 output = evaluateCode(cleanCellCode);
 pass = true;
 catch me
 output = ['Error: ' me.message];
 end
 if containsIllegalFcnFlag
 pass = false;
 output = illegalFcnMessage;
 end
 cellOutput = cleanCode(output);
 testSuite(countVariable).code = cleanCellCode;
 testSuite(countVariable).output = cellOutput;
 testSuite(countVariable).pass = pass;
 end
 % Set up data structure to return
 testSuiteResults = struct(...
 'pass', false, ...
 'score', 0, ...
 'functions', [], ...
 'testPoints', struct([]));
 % Populate the result struct
 % If any test point is failed, sets the overall pass status to false.
 testSuiteResults.pass = all([testSuite.pass]);
 % sets the score
 testSuiteResults.score = sprintf('%d',score);
 % Set test points
 testSuiteResults.testPoints = testSuite;
 % Set functions points
 testSuiteResults.functions = findFcns(sourceFile,'file');
 end</pre>Previous problem in this series: <a href = "http://www.mathworks.com/matlabcentral/cody/problems/1769-singularity-2-0-easier">Singularity 2.0 (easier)</a>Next problem in this series: <a href = "http://www.mathworks.com/matlabcentral/cody/problems/1764-singularity-2-0">Singularity 2.0 (really hard)</a>

This problem is the continuation of problem <http://www.mathworks.com/matlabcentral/cody/problems/1769-singularity-2-0-easier Singularity 2.0 (easier)>

If you have been in Cody long enough you have probably run across some badly constructed problems and test suites.

This problem, I believe, represents an apparently impossible scenario. Yet, it is possible to solve it. Do you know how?

*Description:*

This is this problem's testsuite:

%%
 myfunction();
 [a,b]=1; % oops...
 assert(isequal(a,b));

As you may notice, the second line is not a proper assignment, so the testsuite will break at that point, and will return an error message:

|??? Too many output arguments.|

Typical hacks work by overloading some function in the evaluation code (e.g. assert hack). Unfortunately, as far as I know, assignment operators cannot be overloaded in Matlab. Is there something |myfunction| could do to still solve this problem?

*Hint*:

You may use a similar trick as before but now targeting a vulnerability in the function |verifyCode| instead of in the testsuite. For simplicity, here is the relevant portion of the verifyCode.m file (the function that evaluates and scores Cody problems):

function testSuiteResults = verifyCode( sourceFile, testFile )
    %VERIFYCODE Evaluates and verifies MATLAB Code against provided test suite.
    %   TESTSUITERESULTS = VERIFYCODE(SOURCEFILE, TESTFILE) evaluates SOURCEFILE against
    %   TESTFILE and returns the results.
    %   Copyright 1984-2012 The MathWorks, Inc.
    % Score it
    score = calculateSize(sourceFile);
    % Load the test file
    code = fileread(testFile);
    % Split it into cells
    code2 = regexprep(code,'\n%%','\nxxx-cellbreak-xxx%%');
    cellList = regexp(code2,'xxx-cellbreak-xxx','split');
    % Test each cell
    % Each testpoint structure
    % testPoint = struct(...
    %    'pass', false, ...
    %    'code', '', ...
    %    'output', '');
    testSuite = struct([]);
    [containsIllegalFcnFlag, illegalFcnMessage] = containsIllegalFcn(sourceFile);
    for countVariable = 1:length(cellList)
        cellCode = cellList{countVariable};
        cleanCellCode = cleanCode(cellCode);
        % Test is guilty until proven successful
        pass = false;
        % Run it
        try
            output = evaluateCode(cleanCellCode);
            pass = true;
        catch me
            output = ['Error: ' me.message];
        end
        if containsIllegalFcnFlag
            pass = false;
            output = illegalFcnMessage;
        end
        cellOutput = cleanCode(output);
        testSuite(countVariable).code = cleanCellCode;
        testSuite(countVariable).output = cellOutput;
        testSuite(countVariable).pass = pass;
    end
    % Set up data structure to return
    testSuiteResults = struct(...
        'pass', false, ...
        'score',   0, ...
        'functions', [], ...
        'testPoints', struct([]));
    % Populate the result struct
    % If any test point is failed, sets the overall pass status to false.
    testSuiteResults.pass = all([testSuite.pass]);
    % sets the score
    testSuiteResults.score = sprintf('%d',score);
    % Set test points
    testSuiteResults.testPoints = testSuite;
    % Set functions points
    testSuiteResults.functions = findFcns(sourceFile,'file');
 end

Previous problem in this series: <http://www.mathworks.com/matlabcentral/cody/problems/1769-singularity-2-0-easier Singularity 2.0 (easier)>

Next problem in this series: <http://www.mathworks.com/matlabcentral/cody/problems/1764-singularity-2-0 Singularity 2.0 (really hard)>

Solve

Solution Stats

82 Solutions
17 Solvers

Last Solution submitted on Mar 09, 2023

Last 200 Solutions

Problem Comments

6 Comments

6 Comments

Show 3 older comments

Jean-Marie Sainthillier on 3 Aug 2013

Where do you find all these informations ?

Alfonso Nieto-Castanon on 3 Aug 2013

dbstack('-completenames') will tell you the function call stack, then you can just read the appropriate files to see what is in them...

Jean-Marie Sainthillier on 3 Aug 2013

You should write a book with all your Matlab experience (seriously).

Jean-Marie Sainthillier on 3 Aug 2013

Do you suggest it is possible to modify directly the testsuite in the function ?

Alfonso Nieto-Castanon on 3 Aug 2013

well, not exactly. When myfunction() is being evaluated it is being called by the testsuite code (and that happens inside a evalc command inside the evaluateCode() subfunction) so it is already too late to change the testsuite itself. You can do two things: a) either assume the the call to 'evaluateCode()' is going to fail, and see if you can do something about it (this is the simpler way to solve this problem); or b) try to modify the testsuite *before* myfunction() is being evaluated (that is the solution needed for the next problem -'singularity 2.0 (really hard)'- in this series

Alfonso Nieto-Castanon on 3 Aug 2013

ps. and thanks for the compliment :)

Solution Comments

Show comments

Problem Recent Solvers17

Problem Tags

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Problem 1770. singularity 2.0 (hard)

Solution Stats

Last 200 Solutions

Problem Comments

Solution Comments

Problem Recent Solvers17

Suggested Problems

More from this Author38

Problem Tags

Community Treasure Hunt

Problem 1770. singularity 2.0 (hard)

Solution Stats

Last 200 Solutions

Problem Comments

Solution Comments

Problem Recent Solvers17

Suggested Problems

More from this Author38

Problem Tags

Community Treasure Hunt

Players