[webkit-reviews] review denied: [Bug 113168] LayoutTests/fast/js/large-expressions.html crashes on Linux : [Attachment 194976] patch

[bugzilla-daemon at webkit.org]

Allan Sandfeld Jensen <allan.jensen at digia.com> has denied linzj
<manjian2006 at gmail.com>'s request for review:
Bug 113168: LayoutTests/fast/js/large-expressions.html crashes on Linux
https://bugs.webkit.org/show_bug.cgi?id=113168

Attachment 194976: patch
https://bugs.webkit.org/attachment.cgi?id=194976&action=review

------- Additional Comments from Allan Sandfeld Jensen <allan.jensen at digia.com>
View in context: https://bugs.webkit.org/attachment.cgi?id=194976&action=review


> Source/WTF/wtf/StackBounds.cpp:152
> +    // FIXME: Not sure if the libc of other unix is using the same
implementation as linux does.If anyone does,please remove the OS(LINUX) macro.
> +#if OS(LINUX)

pthread_attr_getguardsize is a generic POSIX call, please enable it for all
UNIX. Whether this code is correct or not only depends on whether the stack
grows up or down, but we seem to only handle stacks growing dow here.

> Source/WTF/wtf/StackBounds.cpp:161
> +    stackBase =
reinterpret_cast<void*>(reinterpret_cast<uintptr_t>(stackBase) + guardSize);
> +    stackSize -= guardSize;

GuardSize does not affect stackSize. This hack is only necessary because of how
m_origin is set lower down. Remove it and just set
m_bound as stackBase + guardSize.