dijit/Editor nested in dijit/form/Form causes error in IE

[adros]

If Editor is nested in Form, getValue function is called before the Editor is loaded (called by connectChildren of _FormMixin). getValue function applies filters, and one of the filter (_stripBreakerNodes, registred for IE) uses this.window, which is undefined before onLoad is called.

I think that another condition needs to be added to getValue function. For example:

if (this.textarea) {
    if (this.isClosed || !this.isLoaded) {
        return this.textarea.value;
    }
} else if (!this.isLoaded) {
    return ...
}

[bill]

Duplicate of #15030.

[Colin Snover]

This is not really a duplicate of #15030, or the fix for #15030 was wrong, one or the other. People are adding new methods (_stripTrailingEmptyNodes now) and keep causing this error to happen because it needs to be fixed once here, not many times elsewhere.

[Colin Snover]

For more information, this issue was reintroduced in a4b34a345d3daa981ed0199d1795ca477817e082

[Colin Snover]

Ed did you have a patch for this?

[Ed Hager]

Here is a pull request: ​https://github.com/dojo/dijit/pull/21

[bill]

The patch looks good, assuming that the correct editor value is returned. It just needs an automated test that creates an Editor and then immediately queries the value. (The Editor doesn't need to be inside a form per se.)

Obviously, #15030 shouldn't have been checked in w/out a test case, or we wouldn't have had this regression in the first place.

[dylan]

Bill, can we make this change for 1.9.3, or do you want an additional test before it gets made. We would like to get 1.9.3 released this week and this is the last blocker.

[bill]

The behavior looks wrong, get("value") is returning "" instead of the actual value of the editor.

[Ed Hager]

Updates:

• I made the editor return this.value before the iframe loads
• removed a line from open that was deleting this.value
• added test cases retrieving the editor's value before the iframe loads

Commit: ​https://github.com/edhager/dijit/commit/1e5f39cc4811b3b4b53e1c9ac2f2ec1f6ffb217a Pull Request: ​https://github.com/dojo/dijit/pull/21

[dylan]

Bill, I think this addresses everything, and I also had Ed remove the spurious delete this.value as well. Ready for your review/testing.

[Bill Keese <[email protected]…>]

In 4dc879a433b3c3db4310011cf28e13edd0801139/dijit:

Error: Processor CommitTicketReference failed

Unsupported version control system "git": Can't find an appropriate component, maybe the corresponding plugin was not enabled? 

[Bill Keese <[email protected]…>]

In 73c826455d90aefa088a165b6fc240edbcf80199/dijit:

Error: Processor CommitTicketReference failed

Unsupported version control system "git": Can't find an appropriate component, maybe the corresponding plugin was not enabled?