Anyone tried out of memory
Hi ! I made quick out-of-memory test, where I limited the available heap space for malloc (did not used the fastmalloc). The result was a Segmentation fault, I do not know if this is due to the GTK env or WebKit. But have anyone tried this before? Knows anything about it? Or is there a policy how to handle the case where a allocation is failing? /mikael
On Aug 19, 2008, at 2:54 PM, Mikael Tennhammar wrote:
But have anyone tried this before? Knows anything about it? Or is there a policy how to handle the case where a allocation is failing?
Starting with <http://trac.webkit.org/changeset/35691> (last week), fastMalloc() aborts on failed allocations, so I guess that this is what the policy is. In a few cases where the caller knows how to handle out of memory conditions (e.g. by raising a JavaScript exception), a version of fastMalloc() that doesn't crash is called. Talking about ports that cannot use fastMalloc go, some work towards making it possible to use custom allocators is tracked by <https://bugs.webkit.org/show_bug.cgi?id=20422
.
- WBR, Alexey Proskuryakov
participants (2)
-
Alexey Proskuryakov
-
Mikael Tennhammar