Fixed
Status Update
Comments
al...@google.com <al...@google.com>
as...@google.com <as...@google.com> #2
Sorry forgot to change the issue title to something that makes more sense.
Main question is, can the same instance of a worker be used in multiple places? And if so, it looks like things are broken.
Main question is, can the same instance of a worker be used in multiple places? And if so, it looks like things are broken.
as...@google.com <as...@google.com> #3
I think the issue here is that we should explicitly disallow more than one instance of the same work request in a chain. What's happening here is correct - each work request runs exactly once.
as...@google.com <as...@google.com> #4
Can you give us a sample app where you see the problems occur? It's a little hard to tell from your logs because both sets of output are using the same exact work requests and it's hard to tell when you enqueued them and what order you're running them in. Did you use them for both the unique and non-unique work?
Description
[With nonSkippingGroupOptimization feature enabled]
The following code,
does not produce a group around the
if
statement.This code produces,
but should produce,
A non-unit returning function, like
remember
cannot be conditionally called in a group. That is, if the group executes, it must execute in a group that always exists if it is called and is removed when it isn't.Without this group, the runtime will only detect that new content is being inserted into composition when the
Wrap
function is called. It currently will record the newremember
calls but gets confused about what order to send theonRemember
andonForgotten
calls.