382123 - Add a unit test for nsIException catching

Status: RESOLVED INACTIVE

(Core :: XPConnect, defect)

Description

[Alex Vincent [:WeirdAl]]

We actually don't have any unit tests for nsIException.  It took me a little while to realize that's what XPConnect was throwing when returning to my code.

This patch adds a simple unit test for nsIException being returned, and the expected result code, to the test_import.js file.  Because it's test code, it shouldn't need sr.

Comment 1

[Alex Vincent [:WeirdAl]]

Attachment: patch

whoops, forgot the actual patch :)

Comment 2

[Robert Sayre]

looks ok, but is this the right file to include it?

Comment 3

[Alex Vincent [:WeirdAl]]

I don't know.  I'll take suggestions for better spots (including a new file).

Comment 4

[Robert Sayre]

yeah, new file seems right.

Comment 5

[Alex Vincent [:WeirdAl]]

I'll try a new file later.  I would test the other properties of nsIException, but that would require a JS component file that was pretty much guaranteed to never change.  Any suggestions?

Comment 6

[Robert Sayre]

the same component you use for the import test?

Comment 7

[Alex Vincent [:WeirdAl]]

Wow, this is odd.  I dumped the properties of the caught nsIException and its stack trace, expecting to find component_import.js in there.  I didn't get it.  This leads me to wonder if I have misinterpreted how nsIException objects are created, and what they should have.  I expected to get a full stack trace, essentially, from the JS component throwing the exception, on down.

exc.result: 2147500034
exc.QueryInterface: function QueryInterface() {
    [native code]
}
exc.message: Component returned failure code: 0x80004002 (NS_NOINTERFACE) [nsISupports.QueryInterface]
exc.name: NS_NOINTERFACE
exc.filename: ../../../../_tests/xpcshell-simple/TestXPC/unit/test_nsIException.js
exc.lineNumber: 52
exc.columnNumber: 0
exc.location.QueryInterface: function QueryInterface() {
    [native code]
}
exc.location.language: 2
exc.location.languageName: JavaScript
exc.location.filename: ../../../../_tests/xpcshell-simple/TestXPC/unit/test_nsIException.js
exc.location.name: run_test
exc.location.lineNumber: 52
exc.location.sourceLine: null
exc.location.caller.QueryInterface: function QueryInterface() {
    [native code]
}
exc.location.caller.language: 2
exc.location.caller.languageName: JavaScript
exc.location.caller.filename: c:/trunk/mozilla/tools/test-harness/xpcshell-simple/tail.js
exc.location.caller.name: null
exc.location.caller.lineNumber: 41
exc.location.caller.sourceLine: null
exc.inner: null
exc.data: null
exc.initialize: function initialize() {
    [native code]
}

Comment 8

[Robert Sayre]

Comment on attachment 266199 [details] [diff] [review]
patch

needs a new patch, right?

Comment 9

[Alex Vincent [:WeirdAl]]

Yes, it does, but I probably won't submit one unless someone can clarify what the expected behavior of nsIException is - why, in particular, I didn't get what I expected in comment 7.

Comment 10

[Alex Vincent [:WeirdAl]]

Re comment 9:  Components.Exception leaves the throwing line in there.  Throwing a nsresult code leaves one line missing.  That's what I missed.

Comment 11

[BMO Automation]

Per policy at https://wiki.mozilla.org/Bug_Triage/Projects/Bug_Handling/Bug_Husbandry#Inactive_Bugs. If this bug is not an enhancement request or a bug not present in a supported release of Firefox, then it may be reopened.