Added an ASSERT to catch an implausible but theoretically possible leak.
Reviewed by Dan Bernstein. In theory, if malloc allocated a UChar buffer directly after a StringImpl, the StringImpl might incorrecly assume that the UChar buffer was inline, and fail to delete it. This ASSERT is somewhat academic, since we don't use the same allocator in debug builds, but oh well. * platform/text/StringImpl.cpp: (WebCore::StringImpl::StringImpl): (WebCore::StringImpl::createUninitialized): * platform/text/StringImpl.h: Separated the inline buffer StringImpl constructor from the out-of-line buffer StringImpl constructor. Made the former ASSERT that its buffer was indeed inline, and the latter ASSERT that its buffer was indeed not inline. git-svn-id: svn://svn.chromium.org/blink/trunk@54460 bbb929c8-8fbe-4397-9dbb-9b2b20218538
Showing
Please register or sign in to comment