<\/p>\n
<\/p>\n
There are more bugs in real-world programs than human programmers can
\nrealistically address. Automatic program repair has been a longstanding
\ngoal in software engineering, yet debugging remains a largely manual
\nprocess.<\/p>\n
Debugging brings an implementation closer to its specification: we
\nthus desire to know that specification. We briefly discuss recent advances
\nin dynamic invariant detection and specification mining. Using techniques
\nas diverse as polyhedra, equation solving, and software quality metrics, we
\ncan discover critical aspects of what a program “does do” or “should do”
\nwith fewer than 5% false positives.<\/p>\n
We then introduce a fully automated approach for crafting patches
\nthat repair bugs in off-the-shelf legacy systems. An extended form of
\ngenetic programming is used to evolve repairs that leverage existing
\nprogrammer expertise and preserve required functionality. In a
\nlarge-scale, systematic study of high-priority defects from programs
\ninvolving over 5 million lines of code and over 10,000 test cases, we
\nfind that our technique can repair 60 of 105 bugs, freeing developer
\neffort. Publicly-available cloud computing prices allow our results to be
\nreproduced for $356; a successful repair completes in 88 minutes and
\ncosts $5.92 on average, a result that is strongly human competitive.
\n<\/p>","protected":false},"excerpt":{"rendered":"
On Thursday 10 October 2013, 9:30-10:30 INRIA Lille Salle Pl\u00e9ni\u00e8re (aka du Conseil), Westley Weimer (University of Virginia) will give a talk on “Discovering Program Invariants and Automatic Program Repair “. Abstract: There are more bugs in real-world programs than human programmers can realistically address. Automatic program repair has been a longstanding goal in … <\/p>\n