Correctly model overflowing of integer literals
commit5c744572055a2ca1dd5e14aa5f242e99686f5b46
authorAndrew Kennedy <akenn@fb.com>
Thu, 28 Sep 2017 14:33:29 +0000 (28 07:33 -0700)
committerHhvm Bot <hhvm-bot@users.noreply.github.com>
Thu, 28 Sep 2017 14:47:01 +0000 (28 07:47 -0700)
treeac3c8ef9342bbdd269251d390def430f6bcabef7
parente964870ece63986b2606130f1f5bc76ad84a41e9
Correctly model overflowing of integer literals

Summary: If an integer literal (whether decimal, hex or octal) is larger than 64-bit positive maxint, we should either cap it at positive maxint, or overflow to float, depending on the IntsOverflowToInts setting.

Reviewed By: hubyrod

Differential Revision: D5931443

fbshipit-source-id: a8c75f390f874a7313740f2be2013efa18f41782
hphp/hack/src/hhbc/ast_constant_folder.ml
hphp/test/hhcodegen_failing_tests_quick
hphp/test/hhcodegen_failing_tests_slow