132217 - delete on global function should not delete the function

Status: VERIFIED FIXED

(Rhino Graveyard :: Core, defect)

Description

[Phil Schwartau]

This is the Rhino version of bug 131964 against SpiderMonkey.
The testcase for this bug is 

          mozilla/js/tests/ecma_3/Function/regress-131964.js

At present, it tests four scenarios:

1. f defined and deleted in global scope
2. f defined and deleted in function scope
3. f defined and deleted in eval scope
4. f defined in eval scope, but deleted in global scope

I'm still trying to sort out what the expected behavior should be
for 3 and 4, but ECMA seems pretty clear about 1 and 2: f should not
be deleted in those cases.

Here are the ECMA references quoted by Thomas Much in the original bug:

> ECMA 262-3, section 13 says that function declarations create a property
> of the current variable object as specified in 10.1.3.

> 10.1.3 says that for function declarations, the property's attributes
> are determined by the type of the code.

> 10.2.1 says that for global code, variable instantiation is performed
> [...] using property attributes { DontDelete }.


And there are similar sections for function scope and eval scope...

Comment 1

[Norris Boyd]

Fixed:

Checking in Interpreter.java;
/cvsroot/mozilla/js/rhino/src/org/mozilla/javascript/Interpreter.java,v  <--  In
terpreter.java
new revision: 1.74; previous revision: 1.73
done
Checking in ScriptRuntime.java;
/cvsroot/mozilla/js/rhino/src/org/mozilla/javascript/ScriptRuntime.java,v  <--
ScriptRuntime.java
new revision: 1.87; previous revision: 1.86
done

Comment 2

[Phil Schwartau]

Reopening bug -


-------------------    SUMMARY OF TEST RESULTS    ---------------------------
The test above checks delete of functions defined in

1. global scope
2. function scope
3. eval scope (delete also done within eval scope)
4. eval scope (delete done in global scope)


-------------------    OUTPUT BEFORE PATCH    ---------------------------
FAILED!: Section 1 of test -
FAILED!: Expected value 'f lives!', Actual value 'f was deleted'

FAILED!: Section 2 of test -
FAILED!: Expected value 'f lives!', Actual value 'f was deleted'


-------------------    OUTPUT AFTER PATCH    ----------------------------
FAILED!: Section 3 of test -
FAILED!: Expected value 'h was deleted', Actual value 'h lives!'

FAILED!: Section 4 of test -
FAILED!: Expected value 'k was deleted', Actual value 'k lives!'


So the patch has fixed the bug in global and function scope, but
has changed the behavior in eval scope. It used to be that functions
defined in eval scope were deletable; now apparently they are not.
According to ECMA, they should be. See discussion in bug 131964.

Comment 3

[Norris Boyd]

Fixed: 

Checking in Interpreter.java;
/cvsroot/mozilla/js/rhino/src/org/mozilla/javascript/Interpreter.java,v  <--  In
terpreter.java
new revision: 1.75; previous revision: 1.74
done

Comment 4

[Phil Schwartau]

Verified FIXED -

The testcase passes in both the rhino and rhinoi shells on WinNT.

Comment 5

[Igor Bukanov]

Targeting as resolved against 1.5R4