The goal of this is to avoid passing an opaque data structure that needs
to be recursively searched by the host.
I considered having some helper for doing the recursion but I figured it
might be helpful to let the reconciler move this around. For example we
might want to create an instance in beginWork and add to it as we go.
This would let us avoid traversing the tree twice and would solve the IE11
perf issue.
So instead, we create the instance first then call appendChild. I could
just call the normal one but I figured that I would make a special one
just in case. For example if you wanted to perform commits on a separate
thread from creation. This turned out to be useful in ReactNoop where I
can avoid searching the array for an existing one since I know the child
isn't there already. (Although splitting placement into insertion/move
might be better.)
Finally, we need the ability to update an instance after all the children
have been insertion. Such as `<select value={...} />`. I called this
finalizeInitialChildren.
This removes updateContainer and instead uses the regular child mutation
methods to insert into the root container and portals.
Since we're no longer clearing out the container DOM in updateContainer
we have to do that manually during initial mount. This now works on a
document and one of the tests end up unmounting the body when you render
into the document so I had to work around that bit since we don't yet
properly support rendering into the document root.
This misbehavior is particularly egregious because of our current strategy for updating a container in Fiber, but regardless -- if we restore focus to an offscreen node (ex: because it was reordered among its siblings) then we should not scroll the page to reveal it.
We could check the `overflow` computed style values before saving the scroll positions but that's unlikely to be any faster.
I tried to add a temporary check for the correct state in https://gist.github.com/spicyj/338147e989215b6eeaf48a6ce2d68d93 and run our test suites over it, but none of our existing test cases would have triggered that invariant. The new one I added does.
* Fix browser bundle for AMD
* Final fix for standalone browser build.
Much more scientific than the rest so it should stick.
* Throw when we can't find code we need to replace.
We're walking backwards up to the root to find the parent so that we can
propagate events further up to nested React parents. If we don't find a
root, that means that the tree was unmounted and we shouldn't send any
events to it.
This was supposed to fix an issue in refs-test but instead it revealed that
this strategy is broken.
The problem is that Placement effect is not sufficient when the insertion
is completed since the effect flag is reset then and the previous tree
has no effects in it to indicate that this is the wrong tree.
If there is no child there is nothing to tell us that this was the
workInProgress branch. Therefore, we need to look at the other branch
if there are children in it to see if *that* was the workInProgress branch.
We've already been warning for mutating styles so now we'll freeze them
in DEV instead. This isn't as good because we don't have more specific
warnings than the generic error that doesn't even fire in sloppy mode.
This lets Fiber pass the same tests.
This used to be done at the end of the transaction but I made it synchronous.
For this to work it needs to be applied after we have already set the .type
property since it is read by inputValueTracker.
We need this to safely extract the current event listeners from the props.
Unfortunately, we are still not safe to use the return pointer since it
may not point to the current tree when the Fiber is reused. So this is not
fully done yet.
* Use the public ReactDOMServer in tests
We need this because it runs the injection.
* Let Fiber skip react data attributes and comments in SSR tests
This markup is testing implementation details rather than behavior.
* [Fiber] Add ReactDOMFiber.unstable_createPortal()
While #8368 added a version of `ReactDOM.unstable_renderSubtreeIntoContainer()` to Fiber, it is a bit hacky and, more importantly, incompatible with Fiber goals. Since it encourages performing portal work in lifecycles, it stretches the commit phase and prevents slicing that work, potentially negating Fiber benefits.
This PR adds a first version of a declarative API meant to replace `ReactDOM.unstable_renderSubtreeIntoContainer()`. The API is a declarative way to render subtrees into DOM node containers.
* Remove hacks related to output field
My first contribution to React!
While upgrading a React project, I found some suspect SVG that needed updating, so I dug in after checking the docs. I knew that support for some SVG properties had been added (namely `xmlns` and `xmlnsXlink`), but I noticed them missing from the reference's attribute list. This pull request updates `reference-dom-elements.md` by adding said properties.
Only declare the variable once in this scope, instead of declaring them multiple times in the same scope.
This fixes#8318, even though it might technically be a shortcoming in Rollup.
* Update all Jest packages to 17.x, cache babel-jest transforms
* Remove the caching
Looking at the other builds it doesn't seem to actually be that necessary. The bottleneck is executors, not build time.
* Remove unnecessary package, fix fiber test runner
* Regenerate yarn lockfile
I forgot that the normal properties route doesn't do this. We also have
to make sure that the order is 1) insert children (options), 2) set
multiple 3) update the options.
