On Wed, 10 Mar 2021 12:41:05 GMT, Prasanta Sadhukhan <psadhuk...@openjdk.org> wrote:
>> In java/awt/font/TextJustifier.java justify-method there is a potential code >> path where divison by zero might happen , see also the Sonar finding >> https://sonarcloud.io/project/issues?id=shipilev_jdk&open=AXcqMwpm8sPJZZzONu1k&resolved=false&severities=CRITICAL&types=BUG >> >> >> boolean hitLimit = (weight == 0) || (!lastPass && ((delta < 0) >> == (delta < gslimit))); >> boolean absorbing = hitLimit && absorbweight > 0; >> // predivide delta by weight >> float weightedDelta = delta / weight; // not used if weight == 0 >> >> In case of (weight == 0) the division should not be done because the value >> of weightedDelta is unused in this case anyway. > > src/java.desktop/share/classes/java/awt/font/TextJustifier.java line 159: > >> 157: // predivide delta by weight >> 158: float weightedDelta = 0; >> 159: if (weight != 0) { // not used if weight == 0 > > Can it ever be -ve? Maybe we can do weight > 0 check just as we do for > absorbweight? Hi, I am not sure about the weight > 0 check ; weight is initialized with 0: weight = 0; and later some values are potentially added up to weight: weight += gi.weight; I am not sure about those gi.weight values, maybe they can be negative too ? ------------- PR: https://git.openjdk.java.net/jdk/pull/2912
