-
Notifications
You must be signed in to change notification settings - Fork 47k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Support iframes (nested browsing contexts) in selection event handling (stale version) #7936
Conversation
042f9dd
to
489866b
Compare
3b85ce4
to
f6ee84e
Compare
ee466e5
to
403e4d6
Compare
6b03fb4
to
2a24dac
Compare
You can see this version of the solution in practice with an actual use case in this demo: http://codepen.io/acusti/pen/yajJbZ?editors=0010 The essential point is that I no longer need to manually manage focus in the app code and instead can rely on the ReactInputSelection functionality to take care of preserving visible “focus” (i.e. visible selections). |
If anyone wants to play around with this, I pushed up a copy of the built react and react-dom scripts:
|
|
||
it('gets and restores selection for inputs in an iframe that get remounted', () => { | ||
var iframe = document.createElement('iframe'); | ||
iframe.setAttribute('tabIndex', 0); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Setting tabIndex on the iframe element here and then invoking iframe.focus()
below at line 202 was only necessary because jsdom didn’t handle setting activeElement
in the parent correctly. This is fixed in jsdom v9.7.0, so if we updated jest’s jsdom dependency, I could remove the hacky workaround.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Updating the version of jsdom used by jest has landed (includes a fix of a major perf regression, so it’ll be fun to see how that affects test running times), so when a new version is released, should be able to remove the hacky workarounds in this test.
881a699
to
3dcc0d1
Compare
(EDIT: I updated the commit hash in the react dependency string below because the previous commit I pointed to didn’t have the correct If anyone wants npm-installable versions of this PR to experiment with, you can use these in your package.json (requires the latest "fbjs": "Brandcast/fbjs-built#97b8e54",
"react": "Brandcast/react#9a81d1a", The dependency on |
3dcc0d1
to
7592afd
Compare
@spicyj I’ve put together a Draft.js demo of this PR along with the Draft.js PR to illustrate the utility of making React work across nested browsing contexts. The PR is facebookarchive/draft-js#765, and the demo is here: http://codepen.io/acusti/pen/RGEJZE?editors=0011 In that demo, I used built versions of React and Draft.js which include all the changes for working across nested browsing contexts. Those built versions are usable in node-land or the browser: EDIT See later comment for updated usable depsIn node "draft-js": "Brandcast/draft-js#f9affa3",
"fbjs": "Brandcast/fbjs-built#97b8e54",
"react": "Brandcast/react#9a81d1a", In the browser <script src="https://rawgit.com/Brandcast/react/nested-browsing-contexts-2-built/build/react.js"></script>
<script src="https://rawgit.com/Brandcast/react/nested-browsing-contexts-2-built/build/react-dom.js"></script>
<script src="https://cdn.jsdelivr.net/immutable.js/latest/immutable.min.js"></script>
<script src="https://cdn.jsdelivr.net/es6.shim/0.35.1/es6-shim.min.js"></script>
<script src="https://cdnjs.cloudflare.com/ajax/libs/babel-core/5.8.34/browser.js"></script>
<script src="https://rawgit.com/Brandcast/draft-js/iframe-compatibility-built/dist/Draft.js"></script> I also rebased my branch (for this PR to React) to remove all upstream conflicts. |
Just updated the PR after rebasing with master. There was a conflict in I also verified that I could remove the hacky workaround in the new |
Thought I should mention that I’ve also updated my built versions of everything, including draft-js (v0.9.0), react (v15.4.1), and react-dom (v15.4.1). This is particularly important because the previous built versions were based on the React master branch 26 days ago (i.e. 16.0.0-alpha, or fiber), whereas these built versions are built from 15.4.1. To try out the PRs: npm / yarn "draft-js": "brandcast/draft-js-built#333bf90",
"fbjs": "brandcast/fbjs-built#3961252",
"react": "brandcast/react-built#ab31c46",
"react-dom": "brandcast/react-dom-built#6cd1db3", in the browser <script src="https://rawgit.com/brandcast/react-built/15-stable/dist/react.js"></script>
<script src="https://rawgit.com/brandcast/react-dom-built/15-stable/dist/react-dom.js"></script>
<script src="https://cdn.jsdelivr.net/immutable.js/latest/immutable.min.js"></script>
<script src="https://cdn.jsdelivr.net/es6.shim/0.35.1/es6-shim.min.js"></script>
<script src="https://cdnjs.cloudflare.com/ajax/libs/babel-core/5.8.34/browser.js"></script>
<script src="https://rawgit.com/brandcast/draft-js-built/master/dist/Draft.js"></script> |
There is minor issue in Safari. I have one iframe with my content and one iframe for Google Maps. My website accessed with "http" and Google Maps hosted with "https". In the getElementsWithSelections following code does not throw error: |
Interesting. I suppose just adding an additional early return if doc is falsy would handle that. Thanks for the feedback! |
Thanks for adding this functionality, hope it get merged into master soon(together with draft-js). I have been working with this for about a week now and the only issue I encountered was that one for Safari. |
@spicyj What's the status here? We'd need that behaviour but we can't afford to maintain a fork of React. Do you still have plans to merge this? Your last update is unfortunately already a few months old. |
@sebmarkbage To clarify, did you remove the |
@acusti No, we have switched to using the github review feature for all PRs. |
2ca7449
to
b48f472
Compare
I tried to rebase this branch from master and ran into a world of hurt, so I created a fresh branch off latest master cherry-picked over all the commits, and made a new PR: #9184 @aleksey-shmatov, I added a guard for the Safari bug you found in f8b202d, as well as dealing with a Firefox issue I found in my own testing. Thanks for trying it out and reporting back! I’ll hoping to make new built versions of everything off of these commits applied to 15-stable and update the npm / yarn instructions on using those soon. When I do, I will update both this now-stale PR and #9184 with the updated dependency strings. |
React generally handles being rendered into another window context correctly (we have been doing this for a while in native Mac popovers). The main place where there are global window/document accesses are in places where we deal with the DOM selection (window.getSelection() and document.activeElement). There has been some discussion about this in the public React GitHub repo: facebook/fbjs#188 facebook#7866 facebook#7936 facebook#9184 While this was a good starting point, those proposed changes did not go far enough, since they assumed that React was executing in the top-most window, and the focus was in a child frame (in the same origin). Thus for them it was possible to check document.activeElement in the top window, find which iframe had focus and then recurse into it. In our case, the controller and view frames are siblings, and the top window is in another origin, so we can't use that code path. The main reason why we can't get the current window/document is that ReactInputSelection runs as a transaction wrapper, which doesn't have access to components or DOM nodes (and may run across multiple nodes). To work around this I added a ReactLastActiveThing which keeps track of the last DOM node that we mounted a component into (for the initial render) or the last component that we updated (for re-renders). It's kind of gross, but I couldn't think of any better alternatives. All of the modifications are no-ops when not running inside a frame, so this should have no impact for non-elements uses. I did not update any of the IE8 selection API code paths, we don't support it.
React generally handles being rendered into another window context correctly (we have been doing this for a while in native Mac popovers). The main place where there are global window/document accesses are in places where we deal with the DOM selection (window.getSelection() and document.activeElement). There has been some discussion about this in the public React GitHub repo: facebook/fbjs#188 facebook#7866 facebook#7936 facebook#9184 While this was a good starting point, those proposed changes did not go far enough, since they assumed that React was executing in the top-most window, and the focus was in a child frame (in the same origin). Thus for them it was possible to check document.activeElement in the top window, find which iframe had focus and then recurse into it. In our case, the controller and view frames are siblings, and the top window is in another origin, so we can't use that code path. The main reason why we can't get the current window/document is that ReactInputSelection runs as a transaction wrapper, which doesn't have access to components or DOM nodes (and may run across multiple nodes). To work around this I added a ReactLastActiveThing which keeps track of the last DOM node that we mounted a component into (for the initial render) or the last component that we updated (for re-renders). It's kind of gross, but I couldn't think of any better alternatives. All of the modifications are no-ops when not running inside a frame, so this should have no impact for non-elements uses. I did not update any of the IE8 selection API code paths, we don't support it. (cherry picked from commit 94b759b in the 0.14-stable. Appears to work mostly as is, needed to be updated to take 5c5d2ec into account)
For anyone who has been following along with this effort, I’m closing this in favor of a new PR: #11410 If you want to use the changes in the latest PR, which is based off of react 16, in your own project, you can do so by updating the "react": "brandcast/react-built#7c8fc6b",
"react-dom": "brandcast/react-dom-built#017534a", If you’re also using draft.js and want the full latest and greatest cross-iframe experience, use: "draft-js": "brandcast/draft-js-built#003f24b", |
We went with #12037 |
As mentioned in #7866 (comment), this PR takes the approach of checking all nested browsing contexts and saving and restoring their selections regardless of whether or not they are the primary focused element on the page. It is a more complete alternative to the solution in #7866, and like that PR, fixes #427. It is also, like the original, dependent on facebook/fbjs#188, but that PR has been merged into master, so as soon as a new fbjs version is released, this will be ready to go.