Qui rogat, non errat
Mark Grechanik Ph.D., University of Texas at Austin
© Copyright Mark Grechanik 2012

Generating Integration Tests Automatically Using Frequent Patterns

of Method Execution Sequences.

My   paper   won   the   first   place   and   the   best   paper   award   at   SEKE   2019 .   We created a      novel      framework      for      automatically      synthesizing      effective system/integration   tests   that   find   bugs   efficiently.   This   paper   is   a   first   small step   in   my   three-pronged   research   program   for   Automatic   Synthesis   of   System and   Integration   Software   Tests   (ASSIST).   New   algorithms   and   techniques   will be   created   and   evaluated   as   part   of   the   work   on   this   project   for   automatically obtaining   models   that   describe   interacting   components,   thus   reducing   the   number of   synthesized   system   and   integration   tests   and   increasing   their   fault-finding power.   Also,   a   novel   way   will   be   created   in   which   static   and   dynamic   analyses   and machine   learning   are   used   to   obtain   test   input   data   as   well   as   oracles   for   the synthesized system and integration tests..

College of Engineering Research Award From UIC

Testing Applications with Data Anonymization (TaDa!)

T esting   Applications   with   Data   Anonymization   (TaDa!)   targets   a problem    of    improving    the    quality    of    testing    when    pr ivacy    laws prohibit   data   owners   to   share   data   with   testers.   Please   read   our award-winning  ISSRE paper to get more information.

Inferring Types of References to GUI Objects in Test Scripts

We offer    a    novel    approach    for    Type    Inference    of    GUI    Object    References (TIGOR)   in   test   scripts.   TIGOR   makes   types   of   GUI   objects   explicit   in   the source   code   of   scripts,   enabling   test   engineers   to   reason   more   effectively   about the   interactions   between   operations   in   complex   test   scripts   and   GUI   objects that   these   operations   reference.   We   describe   our   implementation   and   give   an algorithm   for   automatically   inferring   types   of   GUI   objects.   We   built   a   tool   and evaluated   it   on   different   GAPs.   Our   experience   suggests   that   TIGOR   is   practical and efficient, and it yields appropriate types of GUI objects.