Minutes from telcon 20020920

All, below please find the minutes from our minutes (first half minuted 
by me, second half minuted by Philippe).

If you have sent no comments within one week, I'll post the link to this 
mail on the main DOM TS page.

/Dimitris

---
dom ts telcon 20020920

participants:
johhny stenback, netscape (jst)
rick rivello, nist (rri)
plh, w3c (plh)
dd, ontologicon (dd)
glenn adams (ga)
curt arnold (ca)
mary brady, nist (mb)
brad pettit, ms (bp)
rob releya, ms (rre)

1. status report
rri: traversal/range
jst: level 2 html
plh: dom wg (licenses)
dd: general
mb: still available, rick carries the larger workload
ca: running the suite on mozilla TODO: patches to jsunit
bp: two people at Microsoft have the suite (forwarded email from dd)

2. process
start using the issue tracker

3. rre: microsoft are not able to commit to the dom test suite

4. dom level 2 html
plh: quickly go over mozilla issues

htmldocument08 (DOMWG: test is wrong). the test expects 4. the impl 
returns 2. 2 are objects not containng applets
js: spec looks correct
document.applet collection should be a collection of active applets in 
the doc. active means instantiated
plh: neither ie nor mozilla will be consistent here
mozilla displays the gif, does not load the applet, ie does the other 
way around
bp: would an image that apperas in an object appear in the applet array?
jst: yes
jst: change the definition in the spec to say all objects and applets. 
then we match what ie is doing. we can change mozilla to do the same
brad: in the case of a plugin, is the plugin going to appear in the 
applets aray? something could then appear both in the applets and images 
array
jst: there's a lot to be wished for object handling in the dom.
jst: what happens in ie if you have two nested object tags that both 
laod applets?
ca: should the test be taken out?
plh: no the test is right

(rest from dom wg f2f minutes)

IE

1. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLAppletElement07.html:HTMLAppletElement07 
failed

Comment: The hspace attribute specifies the horizontal space to the left
and right of this image, applet, or object. Retrieve the hspace
attribute and examine it's value.

Johnny: IE uses an integer instead of a string.

Glenn: What about % length?

Johnny: IE4.x always reported as pixels throught the DOM.

Glenn: inconsistency in the specification. vspace/hspace are strings on
applets and objects and long for images.

ACTION: Johnny to look at hspave, vspace, border, height, width for 
applets/images/objects.

2. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLAppletElement09.html:HTMLAppletElement09 
failed

Comment: The vspace attribute specifies the vertical space above and 
below this
image, applet or object. Retrieve the vspace attribute and examine it's
value.

[ditto]

3. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLAppletElement11.html:HTMLAppletElement11 
failed

Johnny: document.applets[0].object, in Mozilla it's a string that's the
URI of the Java class, in IE it's an object of unknown type. The spec
says it should be "Serialized applet file. See the object attribute
definition in HTML 4.01. This attribute is deprecated in HTML 4.01.", I
have no idea what this means.

"objectLink"
Expected org/w3c/domts/DOMTSApplet.class (string) but was  (object)

Johnny: from the HTML spec: "This attribute names a resource containing
a serialized representation of an applet's state. It is interpreted
relative to the applet's codebase. The serialized data contains the
applet's class name but not the implementation. The class name is used
to retrieve the implementation from a class file or archive."

Rick: Curt sent some emails on the mailing list on this...

Johnny: should it be the name of the property specified?

Glenn: I should expect it should return a string. symetrical to the code
attribute. IE3 did not support the applet object.

Brad: MSDN says it retrieves the contained object? pretty vague
here. doesn't offer a type.

Brad: I think it should the value of the object attribute and contains
the name of the resource.

Johnny: if no value? empty?

Brad: yes.

Johnny: not sure if we use it in Mozilla... don't know what we do with
it.

Glenn: HTML says it's CDATA and case sensitive.

-> match the Mozilla behavior.

ACTION: Johnny: update the specification for object to match Mozilla's
behavior.

4. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLBaseFontElement03.html:HTMLBaseFontElement03 
failed

Comment: The size attribute specifies the base font's size.

Johnny: change the spec to long HTMLBaseFontElement.size.

Glenn: the calculated value if the attribute value is relative?

Glenn: unsigned vs signed is appropriate thing here.

ACTION: johnny: check HTMLBaseFontElement.size

5. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLButtonElement08.html:HTMLButtonElement08 
failed

Comment: The value attribute specifies the current control value. 
Retrieve the
value attribute and examine it's value.

Johnny: The tests has a <button value="foo">bar</button> and it expects 
.value
to be "foo" even if the button is displayed with the text "bar" in
it. Mozilla returns "foo" and IE returns "bar".

Rob: the initial value of the button should be the content of the button
element. you first set the value and then you use the content.

Glenn: so if you change the value, do you change the text node?

Rob: don't know.

Glenn: I view this as setting the display of the button, no the value.

Glenn: the value of the name returned should be original

Johnny: if you submit, you submit what's visible.

Philippe: sounds like to be more an HTML problem than a DOM problem.

Johnny: we can say that we won't define which value will be sent.

Mary: should we remove the test?

Johnny: no, it should accept both value as ok.

ACTION Johnny: to clarify HTMLButtonElement.value in the specification.

Brad: You can probably reference the value. it doesn't say what happen.

Johnny: we can try to make it consistency with how we deal with
options value.

6. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLDocument04.html:HTMLDocument04 
failed

Johnny: Test problem? document.URL is the absolute URI in IE and
Mozilla, per spec. on a http server, it works. on file:, it is
inconsistency. you get c: instead of file:.

Mary: dependent on how you run the test suite. if we make the test
available from a server, it will work.

Johnny: we can accept the fact that the test is kind of odd.

Rick: I can add something "isn't appropriate from a local file system".

7. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLFrameElement09.html:HTMLFrameElement09 
had an error

Comment: The contentDocument attribute specifies the document this frame
contains, if there is any and it is available, or null
otherwise. Retrieve the contentDocument attribute of the first FRAME
element and examine it's TITLE value.

Error message is:
"'title' is null or not an object"

Johnny: IE doesn't support frame.contentDocument, bug in IE

Philippe: this is the CR phase. we're missing one implementation.

Glen: remove it from the spec.

-> remove from the specification.

8. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLIFrameElement11.html:HTMLIFrameElement11 
had an error

IE doesn't support iframe.contentDocument, bug in IE

Error message is:
"'title' is null or not an object"

[ditto]

9. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLImageElement04.html:HTMLImageElement04 
failed

Comment: The border attribute specifies the width of the border around 
the image.

"borderLink"
Expected 0 (number) but was 0 (string)

Johnny: we should change it back. Mozilla will have to change it back.

Glen: applet/image/object will need to be consistent.

Johnny: only a problem when you do mathematical operations.

10. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLInputElement06.html:HTMLInputElement06 
failed

<input TYPE="CHECKBOX" ALIGN="bottom"...>

"alignLink"
Expected bottom (string) but was  (string)

Johnny: maybe ie does not support bottom. maybe IE does not does
vertical or horizontal alignement as well...

Glen: was introduced in DOM1.

Johnny: IE does something with it since it returns an empty string.

Glen: ATSC published SMPTE366M-2002 (DOM Level 0) recently btw. it's
dom1, I think it should be keep.

Brad: on MSDN, the align property is not listed.

Resolution: no change.

11. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLIsIndexElement01.html:HTMLIsIndexElement01 
failed

Comment: The form attribute returns the FORM element containing this
control. Retrieve the form attribute and examine its value.

Johnny: There are two <isindex> tags in the testcase, one's inside a
form, one's not. IE doesn't display the one that's not in a form, nor
does it include it in its DOM. the missing isIndex is a bug in IE.

Brad: the isIndex is treated as an input object inside a form element.

Glen: HTML 4.01 as an example of a rewrite with an input object.

Brad: do we want to support deprecated elements?

Johnny: it is a deprecated element but so are many other ones which we
are dealing with as well. in the HTML DTD, it can appear anywhere where
a block is allowed. I would claim this is an IE bug.

Glen: this is in the transitional DTD and we are supporting here I
think.

Brad: anything that requires a transitional should be separated.

Glen: it is necessary to introduce transitional to support legacy
content such as the name attribute.

Johnny: and this is a level 1 property.

Resolution: bug in IE.

12. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLIsIndexElement02.html:HTMLIsIndexElement02 
failed

Comment: The form attribute returns null if control is not within the
context of form. Retrieve the form attribute and examine its value.

Johnny: This test fails in IE due to the bug exposed by the previous 
test.

"notnull"
Expected not to be null

13. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLIsIndexElement03.html:HTMLIsIndexElement03 
failed

Comment: The prompt attribute specifies the prompt message. Retrieve the
prompt attribute of the 1st isindex element and examine its value.

Johnny: IE fails this due to the bug in IE explained in test
HTMLIsIndexElement01. IE would pass this test if there was only one
isindex element in this test.

"Asize"
Expected 2 (number) but was 1 (number)

ACTION: Rick will rewrite the test to avoid problem from test
HTMLIsIndexElement01.

14. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLObjectElement11.html:HTMLObjectElement11 
failed

Comment: The hspace attribute specifies the horizontal space to the left 
and
right of this image, applet or object.

Johnny: change the spec

"hspaceLink"
Expected 0 (string) but was 0 (number)

-> long attribute hspace

[see above]


15. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLObjectElement16.html:HTMLObjectElement16 
failed

Comment: The vspace attribute specifies the vertical space above or
below this image, applet or object.

Johnny: Change the spec

"vspaceLink"
Expected 0 (string) but was 0 (number)

[ditto]

16. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLObjectElement19.html:HTMLObjectElement19 
had an error

Comment: The contentDocument attribute specifies the document this
object contains, if there is any and it is available, or null
otherwise. Retrieve the contentDocument attribute of the first OBJECT
element and examine it's TITLE value.

Johnny: IE doesn't support object.contentDocument, bug in IE.

Johnny: property added in DOM 2 and should be removed from the spec.

17. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLPreElement01.html:HTMLPreElement01 
failed

Comment: The width attribute specifies the fixed width for
content. Retrieve the width attribute and examine it's value.

Johnny: The spec says pre.width is an unsigned long, but IE returns it
as a string???

"widthLink"
Expected 277 (number) but was 277 (string)

Johnny: I'm happy be to say that this should get back to a string even
if it doesn't make sense... Netscape didn't expose the pre in NS3.

-> change

18. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLTableCellElement11.html:HTMLTableCellElement11 
failed

Comment: The char attribute specifies the alignment character for cells
in a column of table header cells(TH). Retrieve the char attribute from
the second TH element and examine its value.

Johnny: IE does not support property at all on any of the table
elements. Given that it was in DOM1, I suggest that it is a bug in IE.

Johnny: ch (ECMAScript) maps to char (HTML) and "char" can't be because
of the language conflict.

Johnny: 2 options: bug in IE or remove from the spec. maybe we should
flag these as optional?

Brad: ch and chOff are documented in MSDN but says that the property is
not available... recorded as being new...

let's keep them then.

26. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLTableElement34.html:HTMLTableElement34 
failed

Comment: The insertRow() method throws a INDEX_SIZE_ERR DOMException if
the specified index is greater than the number of rows.

"HTMLTableElement34"
Call to assertTrue(boolean) with false

Johnny: let's match the spec to match IE and we will change Mozilla.

28. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLTableElement36.html:HTMLTableElement36 
failed

Comment: The deleteRow() method throws a INDEX_SIZE_ERR DOMException if
the specified index is greater than the number of rows.

"HTMLTableElement36"
Call to assertTrue(boolean) with false

-> no op

29. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLTableElement37.html:HTMLTableElement37 
failed

Comment: The deleteRow() method throws a INDEX_SIZE_ERR DOMException if
the specified index is equal the number of rows.

"HTMLTableElement37"
Call to assertTrue(boolean) with false

-> no op


30. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLTableElement38.html:HTMLTableElement38 
failed

Comment: The deleteRow() method throws a INDEX_SIZE_ERR DOMException if
the specified index is negative.

"HTMLTableElement38"
Call to assertTrue(boolean) with false

-> no op

31. 32. (ch)

33 -> no op

34 -> no op

35 -> no op

36 -> no op

37 -> no op

38. ch

39. ch

40. ch

41. chOff

42. chOff

43. chOff

44. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLTableSectionElement21.html:HTMLTableSectionElement21 
had an error

Comment: The insertRow() method inserts a new empty table row. The new 
row is
inserted immediately before the current indexth row in this section. If
index is equal to the number of rows in the section, the new row is
appended.

Error message is:
"Invalid argument."

Johnny: insertRow does not take an index where to put. it takes an
existing and insert it (before or after).

Brad: it says it takes an index ... default value is -1 which happen at
the end of the collection.

Brad: to add at the end, specify -1 or the length of the row.

Johnny: let's do some testing and update the spec.

Brad: will check with my team. for deleteRow, we return Invalid argument
if there is no row.  for insertRow, we always append it.

ACTION: Johnny to look at insertRow/deleteRow.

45. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLTableSectionElement25.html:HTMLTableSectionElement25 
failed

Comment: The insertRow() method throws a INDEX_SIZE_ERR DOMException if 
the
specified index is greater than the number of rows.

"HTMLTableSectionElement25"
Call to assertTrue(boolean) with false

Johnny: verify that the spec defines what IE does.

46. file:///C:/cygwin/home/Administrator/2001/DOM-Test-
Suite/build/level2/html/ecmascript/tests/HTMLTableSectionElement26.html:HTMLTableSectionElement26 
failed

"HTMLTableSectionElement26"
Call to assertTrue(boolean) with false

Comment: The insertRow() method throws a INDEX_SIZE_ERR DOMException if 
the specified index is negative.

ditto

47. ditto

48. ditto.

Official issues list:

pettit1 : HTMLTableRowElement.rowIndex

Brad: within the table definition, order is different from how they
appear.

-> OPEN ISSUE for WG.

hasFeature (new issue):
Brad: returns false for 1.0 and true for 2.0

Johnny: don't know how Mozilla is going to do that but you can't be
compatible with both.

Glen: in Java you can since the package is different.

Johnny: in ECMAScript, they are close enough...

-> OPEN ISSUE for WG

pettit2 : usemap referencing external map definitions

Philippe: issue with the HTML WG?

Glen: how would it be represented if loaded from external resources?

Brad: not clear.

Glen: sounds like a backdoor for a fragment.

Johnny: yes, would be anonymous content in Mozilla.

Brad: they changed the definitions of what's allowed in maps.

Brad: would it more clear if you couldn't specify those externally but
there is a need for it... it has no parent in the tree and that's the
problem.

Johnny: this is definitively something weird in terms of the DOM.

-> OPEN ISSUE

Rick: some don't have HTMLDOMImplementation, do you create documents?

Johnny: Mozilla does not support that. IE does not have a notion of
HTMLDOMImplementation. we should consider removing it...

-> OPEN ISSUE

blur and focus within select, ...

Rick: onfocus and onblur

Brad: no way an onfocus object would give the focus to it.

Glen: it is not something we can address at this level imho.

Rick: so without looking at thr ui, it is not possible to test focus and
blur.

Johnny: for submit, you use iframes.