A JSArray may have a filler map in the elements pointer.
authormvstanton@chromium.org <mvstanton@chromium.org@ce2b1a6d-e550-0410-aec6-3dcde31c8c00>
Fri, 28 Feb 2014 12:29:19 +0000 (12:29 +0000)
committermvstanton@chromium.org <mvstanton@chromium.org@ce2b1a6d-e550-0410-aec6-3dcde31c8c00>
Fri, 28 Feb 2014 12:29:19 +0000 (12:29 +0000)
commitb1ffc7901f6df183dfff995bfc1855df6c62f8ad
tree6df67721c298ad3db59798cca08726ee96ecbb4b
parent5c186bb19792a0078381cdd9649066682c0858a6
A JSArray may have a filler map in the elements pointer.

We already have code that expects this, but incorrectly asserted that the
filler map case would never happen when allocation folding is turned on.
However, even folding has it's limits, bailing out of continued folding
when the object size grows too large. Therefore, it's a general problem
when verifying JSArray objects, that we might encounter a filler map
in elements().

Discovered by ClusterFuzz crbug 347903.

R=hpayer@chromium.org
LOG=N
BUG=347903

Review URL: https://codereview.chromium.org/184493002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19604 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
src/objects-debug.cc
test/mjsunit/regress/regress-crbug-347903.js [new file with mode: 0644]