Closed
Bug 163810
Opened 23 years ago
Closed 23 years ago
Option Object created, own properties added, but Mozilla "forgets" them after some time.
Categories
(Core :: DOM: Core & HTML, defect, P2)
Core
DOM: Core & HTML
Tracking
()
RESOLVED
FIXED
mozilla1.2beta
People
(Reporter: bugzilla, Assigned: peterv)
Details
(Keywords: testcase)
Attachments
(2 files, 1 obsolete file)
|
2.17 KB,
text/html
|
Details | |
|
15.65 KB,
patch
|
sicking
:
review+
jst
:
superreview+
|
Details | Diff | Splinter Review |
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.0) Gecko/20020529
Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.0) Gecko/20020529
The included HTML and Javascript code (see "Steps to Reproduce") is an example
for dynamically creating selectbox options. The code also adds the attribute
"param" to each created option object.
This "param" property is checked
1. after the body has loaded (body onload handler)
2. and again 2 seconds later
On Netscape 4.x and IE both alert boxes show the correct attribut value
(check the text after the "->" arrow), but Mozilla 1.0 fails: The second query
results in an "undefined" value.
Now comes the tricky part: If you remove the comment from the "dummy"
variable this script will work in Mozilla, i.e. keeping a global reference
to options[] will remedy the problem.
Reproducible: Always
Steps to Reproduce:
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<title>Mozilla Option Properties Bug</title>
</head>
<body onload="alert('body onload check:'+checkit());
setTimeout('alert(\'2 seconds later check:\'+checkit())',2000);">
<h2>Mozilla Selecbox Options Scripting Bug</h2>
<form name="dkdk">
<select name="myparam" size=1
onchange="alert('this: '+this.options[this.selectedIndex].param);">
<option>Select . . . . . . .
<option>Select . . . . . . .
<option>Select . . . . . . .
<option>Select . . . . . . .
<option>Select . . . . . . .
<option>Select . . . . . . .
</select>
</form>
<script language="JavaScript">
// Activate line below to get correct results in Mozilla 1.0
// var dummy=document.dkdk.myparam.options;
for (i=0;i<10;i++) {
var nopt=new Option();
nopt.text="Test Option "+i;
nopt.param="Parameter "+ i; // new property
document.dkdk.myparam.options[i]=nopt;
}
// Query attribut "param" of selectbox option #1
function checkit() {
return 'Select #1 -> '+document.dkdk.myparam.options[1].param;
}
</script>
</body>
</html>
Actual Results:
Two alert boxes pop up. The important result is the text after the "->"
First alert box: "Parameter 1"
Second:"undefined"
Expected Results:
First and second alert box: "Parameter 1"
|
Reporter | |
Comment 1•23 years ago
|
||
This bug may relate to bug #138736. Indeed, if you use the script code
provided there and replace "Image" by "Option", you get the same results,
i.e. Option is not a constructor function any more.
|
|
Reporter | |
Comment 2•23 years ago
|
||
Updating References:
Bug #138736 is considered a duplicate of bug #146116 as of today.
(2002-08-21 05:33 that is)
|
||
Comment 3•23 years ago
|
||
Peterv, was this the bug you were thinking about? :-) (on #mozilla)
This should be fixed in nsHTMLOptionElement.cpp, when an option element is
created we should get the node info manager from the document in the context
where it's created, just like we do for new Image();. Easy fix, just copy what
nsHTMLImageElement does. Over to peterv.
Assignee: jst → peterv
Status: UNCONFIRMED → NEW
Ever confirmed: true
|
|
Reporter | |
Comment 4•23 years ago
|
||
|
Assignee | |
Updated•23 years ago
|
Status: NEW → ASSIGNED
OS: Linux → All
Priority: -- → P2
Hardware: PC → All
Target Milestone: --- → mozilla1.2beta
|
|
Assignee | |
Comment 5•23 years ago
|
||
|
|
||
Updated•23 years ago
|
Attachment #99038 -
Flags: superreview+
|
|
||
Comment 6•23 years ago
|
||
Comment on attachment 99038 [details] [diff] [review]
v1
- In nsContentUtils::GetDocumentFromCaller():
+ if (cx) {
...
+ nsCOMPtr<nsIDOMWindowInternal> win(do_QueryInterface(sgo));
+ if (!win) {
+ return;
Sigh! 4-space indentation! :-)
sr=jst
Comment on attachment 99038 [details] [diff] [review]
v1
>+ /**
>+ * Get the docshell through the JS context that's currently on the stack.
>+ * If there's no JS context currently on the stack aDocShell will be null.
>+ *
>+ * @param aDocShell The docshell or null if no JS context
>+ */
> static void GetDocShellFromCaller(nsIDocShell** aDocShell);
>+
>+ /**
>+ * Get the document through the JS context that's currently on the stack.
>+ * If there's no JS context currently on the stack aDocument will be null.
>+ *
>+ * @param aDocument The document or null if no JS context
>+ */
>+ static void GetDocumentFromCaller(nsIDOMDocument** aDocument);
I'd prefer somewhat more distinguishing names for these functions as they do
slightly different things (static vs. dynamic). A comment in the header would
be ok too though
with that, r=sicking
Attachment #99038 -
Flags: review+
|
|
||
Comment 8•23 years ago
|
||
Actually, that's not needed. The functions do exactly the same thing, and this
new function could (and actually should) use GetDynamicScriptGlobal() and not
the static one (even though the script global you get in this case is the same
one) since the dynamic version is potentially a bit faster. It doesn't really
matter though, the point is that both these functions get the doc(shell) from
the caller, that way one could even be based on the other.
Peter, you could change the call to GetStaticScriptGlobal() to
GetDynamicScriptGlobal() when you check in. sr=jst still applies.
|
|
Assignee | |
Comment 9•23 years ago
|
||
Attachment #99038 -
Attachment is obsolete: true
|
|
||
Comment 10•23 years ago
|
||
Comment on attachment 99805 [details] [diff] [review]
v2
sr=jst
Attachment #99805 -
Flags: superreview+
Comment on attachment 99805 [details] [diff] [review]
v2
r=sicking
Attachment #99805 -
Flags: review+
|
|
Assignee | |
Comment 12•23 years ago
|
||
Checked in.
Status: ASSIGNED → RESOLVED
Closed: 23 years ago
Resolution: --- → FIXED
Component: DOM: Core → DOM: Core & HTML
QA Contact: stummala → general
You need to log in
before you can comment on or make changes to this bug.
Description
•