Using gin::ObjectTemplateBuilder::SetMethod() causes leaks. This seems to happen from the CallbackHolder that we allocate not getting correctly destroyed.
Some of these are suppressed - namely, the versions of SetMethod that take a function address. However, using SetMethod() and passing a base::Bind'd callback causes lsan to complain.
Ideally, we should be able to accomplish this without leaking, tying the lifetime of the callback holders to the isolate (since they conceptually are one-per-isolate).
I'll take this for now, but depending on how hairy things get, might need to rope in some more experts.
Comment 1 by rdevlin....@chromium.org
, Feb 23 2017