René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"wrong precision for floats","articleBody":"Some of my tests showed that K-Java has a wrong precision for float values. I assume that float values were just encoded as double precision values. \r\nAs a result, I could observe that in some calculations, K-Java produces results that significantly diverge from the results of Java. Moreover, K-Java sometimes produces a concrete value, whereas Java returns Infinity, which is shown in the following example:\r\n```\r\nfloat a = (float) 1 - 1473150328275393513L / 0.16f * 6854191149854506640L * -128033296 * 23412341234123414L;\r\nSystem.out.println(\"\" + a); //K-Java shows 1.8916947877681303e+62, Java shows Infinity\r\n\r\ndouble b = (double) 1 - 1473150328275393513L / 0.16 * 6854191149854506640L * -128033296 * 23412341234123414L;\r\nSystem.out.println(\"\" + b); //K-Java and Java shows 1.8916947877681303e+62\r\n```\r\nAdditionally, I noticed that K-Java wrongly represents the infinity value (e.g. -Infinity.0) of double or float calculations and also the exponent representation is a bit different to Java.","author":{"url":"https://github.com/rschumi0","@type":"Person","name":"rschumi0"},"datePublished":"2020-02-29T23:53:20.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":0},"url":"https://github.com/59/java-semantics/issues/59"}