SmarTest: Accelerating the detection of faults in Drupal
Ana Belén Sánchez
http://www.isa.us.es/members/anabelen.sanchez
Ana Belén Sánchez, Sergio Segura, Antonio Ruiz-Cortés, Gabriel Hidalgo
Research Group of Applied Software Engineering
(www.isa.us.es)
Escuela Técnica Superior de Ingeniería Informática University of Seville
I am a researcher and a teacher
Why am I here, in a DrupalCon?
Ph.D. in automated testing on variability-intensive systems
Variability-intensive systems
Variability-intensive systems
Mandatory Optional
Alternative (choose 1) Or
Excludes
Requires
A Feature Model represents all the possible configurations of the Variability-intensive system in terms of features and relationships among them.
= {MobilePhone, Calls, Screen, High resolution, Media, Camera}
Feature
Testing problems
Too many configurations!
Time-consuming
Testing variability-intensive systems
Test case selection techniques reduce the test space by selecting a portion of the configurations to be tested.
Testing variability-intensive systems
Test case prioritization techniques schedule test cases for execution in an order that attempts to increase their effectiveness at meeting some performance goal.
Testing variability-intensive systems
How to evaluate these testing techniques?
Artificial models to test
Towards real highly-configurable tools
Artificial models to test
Studied Tools
Testing real highly-configurable tools
1. Available code
2. Detailed fault reports
3. Automated test cases
4. Extensive documentation
5. > 630,000 users and developers
6. > 30,000 modules
Testing real highly-configurable tools
Artificial models to test
Ana B. Sánchez, Sergio Segura, Jose A. Parejo and Antonio Ruiz Cortés. Variability Testing in the Wild: The Drupal Case Study Published in Software and Systems Modeling Journal, 2015
(http://www.isa.us.es/publications/type/article-journal/9998/variability-testing-wild-drupal-case-study)
Our approach: The Drupal case study
Artificial models to test
Module .info files
Drupal documentation
Module JIT
The Drupal variability model Drupal 7.23
Node Filter CtoolsFile
Options
Path
Blog Forum Ctools access ruleset
Ctools custom content
Views content
Field Field UI
Field SQL storage
ImageSystem UserViews
Views UITaxonomy
Comment
Text
Date
Token
Forum Requires Taxonomy Views content Requires Views Panels IPE Requires Ctools Rules Requires Entity tokens Forum Requires Options Taxonomy Requires Options Panel Nodes Requires Ctools Rules scheduler Requires Entity APIForum Requires Comment Date views Requires Ctools Pathauto Requires Token Rules scheduler Requires Entity tokensImage Requires File Date views Requires Views Pathauto Requires Path Rules UI Requires Entity API Views Requires Ctools Panels Requires Ctools Rules Requires Entity API Rules UI Requires Entity tokens Views UI Requires Ctools
PathautoIMCE
Date views
Date popup
Date API
WebForm
Link EntityAPI
Entitytokens
CKEditor Captcha
ImageCaptcha
Features Panels
Panel nodes
Panels IPE
LibrariesAPI
JqueryUpdate
GoogleAnalytics
Rules
Rules UIRulesscheduler
Backupmigrate
MandatoryOptional
48 modules 21 dependencies 2,000 millions of configurations
Drupal information
LoC
Module Size
Changes Commits
Tests 352 test cases 24152 assertions
Drupal information
CC
Cyclomatic Complexity
Drupal Installations
Drupal Contributors
Drupal faults
Bug tracking System
3.232 Faults in v7.23
160 integration faults:
• 132 caused by 2 modules
• 25 caused by 3 modules
• 3 caused by 4 modules
Statistical study
Module Size Faults Module
Changes Faults
Faults v7.22
Faults v7.23
Module CC Faults
C1 C2
C3 C4
Statistical study
Module Developers
Faults Tests ? Faults?
Optional Modules Faults
91 %
Core Mandatory Modules Faults
41 %
C5 C6
Test case prioritization criteria
Size–driven Criterion
Fault–driven Criterion
CC-driven Criterion
Change–driven Criterion
Priori%za%on criteria based on previous sta%s%cal study
Test case prioritization criteria
94.2 % Test case selection
Fault-driven Prioritization
2.
Could we accelerate the detection of faults in Drupal using previous information?
86.2 % Test case selection
1.
Applying our proposal to Drupal
SimpleTest
Anayzing the SimpleTest module
Main view SimpleTest
Our proposal: SmarTest
Main view SmarTest
Our proposal: SmarTest
Our proposal: SmarTest
Our proposal: SmarTest
Test Prioritization SmarTest
Our proposal: SmarTest
Continuos feedback in real time SmartTest
Our proposal: SmarTest
Our proposal: SmarTest
Conclusions
¿What improvements do we add to SimpleTest?
SimpleTest ü Designed to make testing easier
ü A powerful automated testing
ü Selector to select tests to execute
ü Detailed inform. about test execution
ü Possibility to follow the trace of errors
SmarTest ü Customizable Dashboard with run-time extracted data to guide the testing
ü Test prioritization to detect faults faster
ü Automated testing with feedback in real time
ü Time-out option to automatically stop the test execution
Conclusions
Things that we are working on …
• Add new widgets to the dashboard.
• Include multi-prioritization criteria (based on genetic algorithms to detect faults faster).
• Study the integration of SmarTest with other test systems: Behat, phpUnit… • Get SmarTest is accepted in Drupal. Now is in Sandbox: https://www.drupal.org/sandbox/annasan/2503695
• ¿Any suggestions?
Acknowledgments
This work was partially supported by the European Commission (FEDER), the Spanish and the Andalusian R&D&I programmes (Grants IPT-2012-0890-390000 (SAAS FIRE- 46 WALL), TIN2012-32273 (TAPAS), TIC-5906 (THEOS), TIC-1867 (COPAS)).
“Research is to see what everybody else has seen, and to think what nobody else has thought".
– Albert Szent-Gyorgyi, 1893–1986 Hungarian Biochemist, 1937 Nobel Prize for Medicine
Sprint: Friday
https://www.flickr.com/photos/amazeelabs/9965814443/in/faves-38914559@N03/
Sprint with the Community on Friday. We have tasks for every skillset. Mentors are available for new contributors. An optional Friday morning workshop for first-time sprinters will help you get set up. Follow @drupalmentoring.