Traditional Approach . . . Static Analysis Tools Some “Defects” . . . Correcting Non- . . . How to Gauge Repair Risk? We Need to Gauge . . . First Risk Factor: How . . . Francisco Zapata 1 and Vladik Kreinovich 2 Second Risk Factor: . . . Resulting Criteria for . . . 1 Department of Industrial, Need to Go Beyond . . . Manufacturing, and Systems Engineering 2 Department of Computer Science Home Page University of Texas at El Paso Title Page 500 W. University El Paso, Texas 79968, USA ◭◭ ◮◮ fazg74@gmail.com, vladik@utep.edu ◭ ◮ Page 1 of 43 Go Back Full Screen Close Quit
Traditional Approach . . . Static Analysis Tools 1. Traditional Approach to Software Testing Some “Defects” . . . • The main objective of software is to compute the de- Correcting Non- . . . sired results for all possible inputs. We Need to Gauge . . . First Risk Factor: How . . . • From this viewpoint, a reasonable way to test the soft- Second Risk Factor: . . . ware is: Resulting Criteria for . . . – to run it on several inputs for which we know the Need to Go Beyond . . . desired answer, and Home Page – to compare the results produced by this software Title Page with the desired values. ◭◭ ◮◮ • This was indeed the original approach to software test- ◭ ◮ ing. Page 2 of 43 Go Back Full Screen Close Quit
Traditional Approach . . . Static Analysis Tools 2. Experts Can Detect Some Software Defects With- Some “Defects” . . . out Running the Program Correcting Non- . . . • Once it turns out that on some inputs, the program is We Need to Gauge . . . not producing the desired result, the next step is: First Risk Factor: How . . . Second Risk Factor: . . . – to find – and correct – the defect Resulting Criteria for . . . – that leads to the wrong answer. Need to Go Beyond . . . • After going through this procedure many times, pro- Home Page grammers started seeing common defect patterns. Title Page • For example, a reasonably typical mistake is forgetting ◭◭ ◮◮ to initiate the value of the variable. ◭ ◮ • In this case: Page 3 of 43 – we may get different results depending on Go Back – what was stored in the part of the computer mem- Full Screen ory which is allocated for this variable. Close Quit
Traditional Approach . . . Static Analysis Tools 3. Software Defects (cont-d) Some “Defects” . . . • This defect is even more dangerous if the variable is a Correcting Non- . . . pointer , i.e., crudely speaking, if We Need to Gauge . . . First Risk Factor: How . . . – it stores not the actual value of the corresponding Second Risk Factor: . . . object, Resulting Criteria for . . . – but rather the memory address at which the actual Need to Go Beyond . . . value is stored. Home Page • In this case, if we do not initialize the pointer, not only Title Page can we access the wrong value, but: ◭◭ ◮◮ – we may also end up with a non-existing address ◭ ◮ – or an address outside the memory segment in which Page 4 of 43 your program is allowed to operate, Go Back – at which point the program stops; – this is known as segmentation fault . Full Screen Close Quit
Traditional Approach . . . Static Analysis Tools 4. Software Defects (cont-d) Some “Defects” . . . • Many programming language do not automatically check Correcting Non- . . . the array indices. We Need to Gauge . . . First Risk Factor: How . . . • Then, a typical defect is asking for a value a [ i ] of an Second Risk Factor: . . . array a for an index i which is outside the array’s range. Resulting Criteria for . . . • In this case, the compiler obediently finds the corre- Need to Go Beyond . . . sponding space in the memory. Home Page • However, this space is beyond the place of the original Title Page array. ◭◭ ◮◮ • This can overwrite important information; this is known ◭ ◮ as buffer overrun . Page 5 of 43 Go Back Full Screen Close Quit
Traditional Approach . . . Static Analysis Tools 5. Static Analysis Tools Some “Defects” . . . • Programmers realized long time ago that there are cer- Correcting Non- . . . tain patterns of code typical for software defects. We Need to Gauge . . . First Risk Factor: How . . . • So, they started to come up with automatic tools for Second Risk Factor: . . . detecting such patterns. Resulting Criteria for . . . • These tools warn the user of possible defects of different Need to Go Beyond . . . potential severity. Home Page • At present, there are many such tools – Coverity, For- Title Page tify, Lint, etc. ◭◭ ◮◮ • Most of these tools are efficiently used in practice. ◭ ◮ Page 6 of 43 Go Back Full Screen Close Quit
Traditional Approach . . . Static Analysis Tools 6. Some “Defects” Found by Static Analysis Tools Some “Defects” . . . Do Not Harm the Program’s Functionality Correcting Non- . . . • Not all “defects” uncovered by a static analysis tool We Need to Gauge . . . are actually hurting the program. First Risk Factor: How . . . Second Risk Factor: . . . • For example, some programs have extra variables , i.e., Resulting Criteria for . . . variables which are never used. Need to Go Beyond . . . • This happens if: Home Page – a programmer originally planned to use the vari- Title Page able, ◭◭ ◮◮ – started coding with it, ◭ ◮ – then changed her mind but forgot to delete all the occurrences of this variable. Page 7 of 43 • Static analysis tools mark it as a possible detect. Go Back Full Screen • Indeed, in some situations, it is indeed an indication that some important value is never used. Close Quit
Traditional Approach . . . Static Analysis Tools 7. Some “Defects” Do Not Harm the Program’s Some “Defects” . . . Functionality (cont-d) Correcting Non- . . . • However, in many other cases: We Need to Gauge . . . First Risk Factor: How . . . – it may be syntactically clumsy, Second Risk Factor: . . . – but it does not cause any problem for the program. Resulting Criteria for . . . • Another defect that may not necessarily be harmful is Need to Go Beyond . . . the logically dead code, when: Home Page – a branch in a branching code Title Page – is never visited. ◭◭ ◮◮ • For example: ◭ ◮ Page 8 of 43 – if as part of the computations, we compute a square root of some quantity, Go Back – it makes sense to make sure that this quantity is Full Screen non-negative. Close Quit
Traditional Approach . . . Static Analysis Tools 8. Some “Defects” Do Not Harm the Program’s Some “Defects” . . . Functionality (cont-d) Correcting Non- . . . • When this quantity appears as result of long computa- We Need to Gauge . . . tions, it may happen that, First Risk Factor: How . . . Second Risk Factor: . . . – due to rounding errors, Resulting Criteria for . . . – a small non-negative value becomes small negative. Need to Go Beyond . . . • In this case, it makes sense, if the value is negative, to Home Page replace this with 0. Title Page • However, if we write a code this way, ◭◭ ◮◮ – but we only use it to compute √ x of a non-negative ◭ ◮ x (e.g., of the weight), Page 9 of 43 – then the branch corresponding to a negative value Go Back is never used. Full Screen Close Quit
Traditional Approach . . . Static Analysis Tools 9. Some “Defects” Do Not Harm the Program’s Some “Defects” . . . Functionality (cont-d) Correcting Non- . . . • In some cases, this may be a real defect, indicating We Need to Gauge . . . that: First Risk Factor: How . . . Second Risk Factor: . . . – we may have missed something Resulting Criteria for . . . – that would lead to the possibility of this condition. Need to Go Beyond . . . • However, in cases described above, this “defect” is mostly Home Page harmless. Title Page • Yes another example of a possible defect is indentation. ◭◭ ◮◮ • In some programming languages like Python, indenta- ◭ ◮ tion is used to indicate: Page 10 of 43 – the end of the condition or Go Back – the end of the loop. Full Screen • However, in most other programming languages, in- dentation is ignored by the compiler. Close Quit
Traditional Approach . . . Static Analysis Tools 10. Some “Defects” Do Not Harm the Program’s Some “Defects” . . . Functionality (cont-d) Correcting Non- . . . • Indentation simply helps people better understand each We Need to Gauge . . . other’s code. First Risk Factor: How . . . Second Risk Factor: . . . • A static analysis tool: Resulting Criteria for . . . – will detect the discrepancy between the indentation Need to Go Beyond . . . and the actual end of the condition or of a loop, and Home Page – indicate it as a defect. Title Page • It indeed may be a serious defect. ◭◭ ◮◮ • However, in many cases, it is just a sloppiness of a pro- ◭ ◮ grammer that does not affect the program’s execution. Page 11 of 43 Go Back Full Screen Close Quit
Recommend
More recommend
