621276 - Script tag onerror event does not fire if src attribute points to non-existing file://

Status: RESOLVED WORKSFORME

(Core :: DOM: Events, defect)

Description

[st777]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.2.13) Gecko/20101203 Firefox/3.6.13
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.2.13) Gecko/20101203 Firefox/3.6.13

I'm attempting to inject a new script tag dynamically into a head element using DOM appendChild. When I inject a script tag that points to a non-existent file, onerror event of the injected script tag does not fire.

Reproducible: Always

Steps to Reproduce:
1. Create a new script tag: var newScript = document.createElement("script");
2. Bind an event to new script element: newScript.onerror = function() { alert('error!');}

3. Set src attribute of new script element to a bogus(non-existent) file: newScript.src = "foo.js";

4. Append new script element to head element: document.getElementsByTagName('head')[0].appendChild(newScript);
Actual Results:  
onError handler in step 2 does not fire

Expected Results:  
onError event must fire to let me know that something is wrong with loading a file

Test script is running locally on a desktop machine and not from a hosted location. Not sure if that makes a difference.

This behavior works correctly in all major browsers such as chrome, safari, IE 7-9

Comment 1

[Boris Zbarsky [:bzbarsky]]

> Test script is running locally on a desktop machine and not from a hosted
> location. Not sure if that makes a difference.

It does.  For file:// URIs this is a known issue.

This is similar to bug 269125, and I was pretty sure we had a duplicate of this for <script> already...

Ms2ger, want to fix?

Comment 2

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

Bug 608809?

Last time I looked at our script execution code (or the spec, for that matter), it looked sufficiently scary not to attempt a fix. I could try if you pointed me at the place I'd need to patch, though.

Comment 3

[Boris Zbarsky [:bzbarsky]]

> Bug 608809?

That bug wfm.

> I could try if you pointed me at the place I'd need to patch, though.

We need to fire an error event if AsyncOpen on the script's channel returns failure.

Comment 5

[Nickolay_Ponomarev]

Attachment: testcase from bug 646214

Comment 6

[douglyuckling]

As more and more people are unit testing their JavaScript (and often using flat file:// URLs), this bug is pretty annoying.  There is some discussion of workarounds at Stack Overflow: http://stackoverflow.com/questions/4150430

Comment 7

[O. Atsushi (Torisugari)]

JFYI, according to Asqueella's testcase, WFM.

Comment 8

[Boris Zbarsky [:bzbarsky]]

This got fixed in bug 789856, I would bet...