Apidays Singapore 2024 - Building Digital Trust in a Digital Economy by Veron...
Â
Partitioned Based Regression Verification
1. Partition-Based Regression
[1]
Verification
1
PRESENTED BY:
AUNG THU RHA HEIN(5536871)
BOONYA SUWANMANE(5436284)
NATTACHART TAMKITTIKHUN (5637378)
[1] MARCEL BOâHME, BRUNO C. D. S. OLIVEIRA, ABHIK ROYCHOUDHURY
SCHOOL OF COMPUTING,NATIONAL UNIVERSITY OF SINGAPORE
PUBLISHED ON ICSEâ13, SAN FRANCISCO, USA
2. Outline
2
ï Introduction
ï Partition-Based Regression Verification
ï Empirical Study
ï Results and Analysis
ï Threats to Validity
ï Related Works
ï Discussion & Conclusion
3. Introduction-Software Regression
3
ï Software Regression- software bugs occur after
changes to software functionalities
ï Regression Testing- selective retesting of a
software system
ï Regression Verification-verify the correctness of
the program relative to earlier versions
6. Introduction-Research Contributions
6
ï Introduces (PRV) partition-based regression verification
ï Proposes a differential partitioning technique
ï Provides another way to regression test generation
techniques
7. Introduction-PRV
7
ï A gradual approach to RV based on the exploration of
differential partitions
ï Verify inputs by partitions
ï Shares the advantages of RV and RT
10. PRV: A) Computing Differential
Partitions
10
Base on the value detect of test suite(input) which lead to regression error.
ï Output a set of test suite and as set condition
ï
11. PRV: B) Computing Reachability
Conditions
11
Base on the detect condition.
ï Output is a condition depend on your proposal input value be the criteria
ïĄ Reachable condition or
ïĄ Unreachable condition
ï
12. PRV: C) Computing Propagation
Conditions
12
Base on the detect where the differential
program states converge.
ï Output: Statement instance at line Ni(i substitute line number)
ï
13. PRV: D) Computing Difference
Conditions
13
Base on the detect where the differential
program states converge.
ï Algorithm as same as C) plus the change value process , check does the
converge output are different?
ï Output: A set of change statement instance at line Ni(i substitute line
number) which lead the output at converge line are different.
ï
14. PRV: E) Generating Adjacent Test Cases
14
Base on the detect where the differential
program states converge.
ï Algorithm: If adjacent condition have been deleted after compute beyond the
existing condition does program compute the output.
: If place or add or reorder condition does program compute the
output.
If can compute how the output different or same at the converge statement.
ï
15. PRV: F) Theorems
15
ï
In practice, the absence of regression errors can be guaranteed for all inputs
to the same extent as symbolic execution can guarantee the absence of
program errors. SpeciïŹcally, they assume deterministic program execution.
16. Empirical Study
16
ï Evaluate relative efficiency of PRV and discuss practicability based on
authorsâ experience.
ï Do not prove the scalability of PRV.
ï It suffers from the same limitation as symbolic execution.
ï However, it can benefit from optimizations like domain reduction,
parallelization, and better search strategies.
17. Empirical Study â Setup and Infrastructure
17
ï Built into authorsâ dynamic backward slicing tool JSlice.
ï The differential partitions are explored in a breadth-first manner
starting from the same initial input within 5 minutes, unless stated
otherwise.
ï Every version of the same subject uses the same test driver to construct
necessary inputs.
ï Subject programs are analyzed on a desktop computer with an Intel 3
GHz quad-core processor and 4 GB of memory.
18. Empirical Study â Subject Programs
18
ï Subject programs in the experiments are chosen according to the
following 2 criteria:
ïĄ They represent a variety of evolving programs.
ïĄ They are discussed in related work (which allows the comparison
with our own experimental results).
ï There are 83 versions of programs ranging from 20 to almost 5000
lines of code.
ïĄ Some are derived by seeding faults, called mutants, of the original
versions.
ïĄ Some are real versions that were committed to a version control
system.
20. Empirical Study â Subject Programs
20
ï The authors compare the empirical results of the references discussing
regression verification and regression test generation.
ï No empirical results available for the regression test generation
techniques and differenctial symbolic execution
ï All programs are tested as whole programs, except for Apache CLI.
ïĄ
For Apache CLI, command line component was tested for regression.
21. Empirical Study â Research Questions
21
ï RQ1:
How efficiently does PRV find the first input that exposes semantic
difference?
ï RQ2:
ïĄ How efficiently does PRV find the first input that exposes software
regression?
ï RQ3:
ïĄ How practical is PRV in an example usage scenario?
ïĄ
22. Results and Analysis â RQ1: Efficientcy â Semantic
Difference
22
ï Measure 2 aspects when searching for the first difference-revealing
input:
ïĄ average time
ï· If > 5 min, not included.
ïĄ mutation score.
ï· The fraction of versions for which a difference-revealing input can
be found within 5 minutes.
24. Results and Analysis â RQ1: Efficientcy â Semantic
Difference
24
ï Answer to RQ1.
ïĄ
PRV generates a difference-revealing test case on average for 21%
mor version pairs in 41% less time, than the eXpress-like approach
that analyzes only the changed version Pâ.
25. Results and Analysis â RQ2: Efficientcy â Software
Regression
25
ï In practice, not every difference-revealing test case reveals software
regression.
ï A difference-revealing test case can be checked formally on informally
against the programmerâs expectation.
27. Results and Analysis â RQ2: Efficientcy â Software
Regression
27
ï Answer to RQ2.
ïĄ
PRV generates a regression-revealing test case on average for 48%
more version pairs in 63% less time than the eXpress-like approach
that analyzes on the changed version Pâ.
28. Results and Analysis â RQ3 Practicability â Usage
Scenario: Apache CLI
28
ï Apache CLI is used to evaluate PRV in a practical usage scenario.
ï PRV generates difference-revealing test cases within the bound of 20
minutes for every version pair.
ï A developer checks these test cases for regression and relates the
regression revealing test cases to changes that semantically interfere.
29. Results and Analysis â RQ3 Practicability â Usage
Scenario: Apache CLI
29
ï Answer to RQ3.
ïĄ
For the evolution of Apache CLI over 6 years, tests generated as
witnesses of differential behavior of 2 successive versions suggest:
ï· An average progression of 49%, regression of 18% and
intermediate semantic changes of 33% towards the latest revision.
30. Threats to Validity
30
ï Main threat to internal validity:
Implementation of PRV into JSlice
ïĄ Tried to mitigate by using the same implementation to gather results
for the DART-like and eXpress-like approches.
ï Main threat to external validity
ïĄ The generalization of the results
ï· Limited choice and # of subjects does not suggest generallizability.
ï· The subjects are served mainly as comparison to relavant works,
and give an idea about practibility of PRV.
ïĄ
31. Related Works
31
ï Regression Verification(RV)
based on semantic equivalence
ïĄ time consuming
ïĄ no intermediate guarantees
ï Differential Symbolic Execution (DSE)
ïĄ based on symbolic summaries
ïĄ less scalable
ï Regression Test Generation (RTG)
ïĄ construct sample input that can expose software regression
ïĄ