319172 - "function eval must be called directly, and not by way of a function of another name" in loaded sub-scripts.

Status: RESOLVED FIXED

(Core :: XPConnect, defect, P2)

Description

[Lev Serebryakov]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8) Gecko/20051111 Firefox/1.5
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8) Gecko/20051111 Firefox/1.5

When "eval" is called in function, defined in subscript, loaded with JSSubscriptLoader, this error occurs in recent trunk builds.

Reproducible: Always

Steps to Reproduce:
1. Create two JavaScript files:
============== main.js
var loader =
Components.classes['[e-mail]/moz/jssubscript-loader;1'].getService(Components.interfaces.mozIJSSubScriptLoader);
loader.loadSubScript('chrome://application/content/test.js');
TEST();
==============
============== test.js
function TEST()
{
  dump("EVAL: " + eval("1 + 1;") + "\n");
}
==============
2. Load "main.js" into XUL file with <script/> element.
3. Get this error in JavaScript console.

Actual Results:  
"function eval must be called directly, and not by way of a function of another name" error.

Expected Results:  
EVAL: 2
dumped into console.

Comment 1

[Blake Kaplan (:mrbkap) (inactive)]

Attachment: Innerize the global object

Walking up the parent chain outerizes the global object (wrongly). We need to innerize at the top so that the script is loaded with the correct parent chain (pointing at the inner chrome window).

Comment 2

[Lev Serebryakov]

(In reply to comment #1)

> Walking up the parent chain outerizes the global object (wrongly). We need to
> innerize at the top so that the script is loaded with the correct parent chain
> (pointing at the inner chrome window).
 Patch works for me (reporter) on very recent trunk build (updated from CVS about 06 Dec 2005 22:00 GMT+3).

Comment 3

[Brendan Eich [:brendan]]

Comment on attachment 205112 [details] [diff] [review]
Innerize the global object

sr=me, if you want it.

/be

Comment 4

[Blake Kaplan (:mrbkap) (inactive)]

Comment on attachment 205112 [details] [diff] [review]
Innerize the global object

Shaver gave me r=him in person. I also fixed the fprintf to print the correct thing.

Comment 5

[Blake Kaplan (:mrbkap) (inactive)]

Fix checked in. Nominating for branches since this is a relatively large regression from 1.0.x.

Comment 6

[Daniel Veditz [:dveditz] out until Nov 10]

Comment on attachment 205112 [details] [diff] [review]
Innerize the global object

a=dveditz for drivers

Comment 7

[Blake Kaplan (:mrbkap) (inactive)]

Sorry, I checked this into the branches last Friday.

Comment 8

[Jesse Ruderman]

Bug 359159 looks similar.