Correct usage of jira-editor-core from Atlaskit (Richt-Text-Editor) in p2 environment

Hey there,

Currently I’m about to refactor the UI of my plugin, I read a few blogs where some developer use react to implement the UI, which inspiried me to do it as well. Actually it works pretty fine, but I have an issue when I try to implement a RTE (Rich-Text-Editor).

First of all, I implemented a component for the RTE based on the atlaskit component

In my view, the RTE is rendered just like I expected and it actually works good.

Now, when I’m typping into the RTE and hit the ‘c’ key, the “Create-an-Issue”-View will pop up…So every key which is an actual shortcut will open a new dialog. Does anyone knows how to use the editor-core component so that it will be rendered as a valid input-field?

Here is my react-component:

export default class EditorTest extends Component{

    render(){
        return (
            <EditorContext>
                <div>
                    <div style={{ padding: '20px' }}>
                        <Editor
                            className="editor"
                            appearance="comment"
                            placeholder="What do you want to say?"
                            analyticsHandler={analyticsHandler}
                            allowCodeBlocks={true}
                            allowLists={true}
                            allowRule={true}
                            allowTables={{
                              allowColumnResizing: true,
                              allowMergeCells: true,
                              allowNumberColumn: true,
                              allowBackgroundColor: true,
                              allowHeaderRow: true,
                              allowHeaderColumn: true,
                              permittedLayouts: 'all',
                              stickToolbarToBottom: true,
                            }}
                            allowDate={true}
                            contentTransformerProvider={schema =>
                            new JIRATransformer(schema)
                            }
                            />
                    </div>
                </div>
            </EditorContext>
        );
    }
}

regards

Hi, @M.Abdel-Mola, were you able so solve this issue?

Hi @M.Abdel-Mola

Yes we’ve had the same problem. I believe the issue is that the editor uses an editable div rather than a text area so Jira doesn’t recognise it as a field.

A solution we use is to wrap usages of the editor in another component that disables the shortcuts for its children. Something along the lines of this:

import React, { PureComponent } from "react";

class ShortcutBoundary extends PureComponent {
    constructor(props) {
        super(props);
        this.wrapperNode = React.createRef();
    }

    componentDidMount() {
        const { current } = this.wrapperNode;
        if (current && typeof AJS !== "undefined") {
            AJS.$(current).on("aui:keypress", this.preventKeyboardShortcut);
        }
    }

    componentWillUnmount() {
        const { current } = this.wrapperNode;
        if (current && typeof AJS !== "undefined") {
            AJS.$(current).off("aui:keypress", this.preventKeyboardShortcut);
        }
    }

    render() {
        return (
            <div ref={this.wrapperNode}>
                {this.props.children}
            </div>
        );
    }

    private preventKeyboardShortcut = e => e.stopPropagation();
}

export default ShortcutBoundary;

Cheers,
Jared

3 Likes

We recently ran into this in Confluence as well, FWIW. I’m not sure if the root cause is the same or similarly solvable with Jared’s (very elegant) solution, but there it’s a problem with shortcut keys typed into an RTE being swallowed by jquery’s hotkeys plugin, which does not suppress such events in contenteditable components. Our workaround is an ugly hack, but I will post it here just in case it helps anyone coming at this from a different angle:

      var add = window.jQuery.event.special.keypress.add;
      window.jQuery.event.special.keypress.add = function (handleObj) {
        var result = add.call(this, handleObj);
        var handler = handleObj.handler;
        handleObj.handler = function (e) {
          if (e.target && this !== e.target && e.target.contentEditable === 'true') return;
          return handler.call(this, e);
        };
        return result;
      };

Running that as early as possible on page load (not waiting for DOMReady) will fix it for all contenteditables on the page (at least in Confluence 6).

3 Likes

Does this still work? I can’t seem to get it to work now