RE: Further LC Followup from IE RE: Potential bugs identified in XHR LC Test Suite
It seems we are in agreement, mostly :-)
The point of a fictional browser is that it has no market share, no application
built on top of it and thus would not cause customer pain when it changes
implementation.
In the case there isn't clear technical differences, I don't think we should
pick
the right solution based on implementer's cost. Rather We should base it on
customer impact.
A bank with 6000 applications built on top of IE's current APIs simply would
not be happy
if some applications cannot run due to changes in some underlying object model.
And this is not IE's problem alone since the bank simply cannot upgrade or
change the browser,
if all other browsers result in the same breakage.
-Original Message-
From: Ian Hickson [mailto:[EMAIL PROTECTED]
Sent: Tuesday, June 17, 2008 1:51 AM
To: Zhenbin Xu
Cc: Sunava Dutta; Web API public; IE8 Core AJAX SWAT Team; public-
[EMAIL PROTECTED]
Subject: RE: Further LC Followup from IE RE: Potential bugs
identified in XHR LC Test Suite
On Tue, 17 Jun 2008, Zhenbin Xu wrote:
I am not sure if I understand your question.
responseXML.parseError
has the error information
http://msdn.microsoft.com/en-us/library/aa926483.aspx
Oh, I assumed Sunava meant a conforming Document object was
returned.
A parseError-type object would be what I had in mind, yes. However,
if
we do this, then we should specify it. If we don't specify it, I'd
rather have an exception.
The spec can simply state that a conforming document object is
returned,
which includes out-of-band error information. This is what IE does
today
and is a very reasonable approach that allows rich error information
for
debugging.
I don't believe it is conforming for Document objects to have
parseError
attributes, but I could be mistaken -- is there a spec for parseError?
Even if there isn't, though, I agree that it is a generally good
solution
to the problem, I'm just saying that we should specify it, so that UAs
can be standards-compliant and support it interoperably.
Not really; if the script is expecting an exception, and receives
null
instead, then they'll just get an exception as soon as they
dereference the object, which in almost all cases will be straight
away.
[Zhenbin] I should explain the scenario I talked about. For instance,
if
I am to write a wrapper object myXHR, it makes a difference for me
when
I do the following
myXHR.responseXML
if (!_innerResponseXML)
{
try
{
_innerResponseXML = _innerXHR.responseXML;
}
catch (e)
{
_myexception = e;
return _dummpyResponseXML;
}
}
return _innerResponseXML;
My try catch would not catch null. And the exception would be passed
on
to my callers, which is not what I wanted.
Indeed.
If we are going to spec it to accommodate all existing browsers,
we
would want to make it return null or INVALID_STATE_ERR
exception.
We want interoperable behaviour, so defining it in this way would
be a
bad idea. (I don't really have an opinion either way about
exception
vs null, but it seems that we should just pick whatever is most
commonly implemented, which I'm guessing is what Anne did here.)
Fair enough. So let's pick one.
What is commonly implemented? Is it largest browser market share?
Since the cost to implementations for fixing the problem is independent
of
the size of the user base, it would be based just on the number of
independent implementations.
Is it number of enterprise applications written on top of particular
browser?
All the browsers want to be compatible with the Web, so if there are
mroe
Web sites depending on the exception behaviour than the null behaviour,
then we clearly should do the exception behaviour. And vice versa. Do
we
have any good numbers on this? (That there are widely deployed browsers
that return null instead of throwing an exception tends to suggest that
Web pages don't depend on either behaviour; we'd probably need evidence
to the contrary to decide one way or the other based on compatibility.)
Is it the number of browers, in which case I hope my fictional
home grown personal browser gets a vote :-)
Obviously fictional browsers aren't relevant, since the cost of fixing
a
fictional browser is zero.
From a pure technical point of view, predictably throw exception on
state violations is easier to understand. I hope you would agree
there
is value to change spec for the sake of consistent programming model
(which happens to be the IE model).
Indeed.
Did the spec call out that responseXML returned from XHR should have
equivalent DOM support as UA's object? If it is, that would be a
good
topic for us to debate about.
I believe the spec just says
RE: Further LC Followup from IE RE: Potential bugs identified in XHR LC Test Suite
Inline...
-Original Message-
From: Ian Hickson [mailto:[EMAIL PROTECTED]
Sent: Sunday, June 15, 2008 11:46 PM
To: Zhenbin Xu
Cc: Sunava Dutta; Web API public; IE8 Core AJAX SWAT Team; public-
[EMAIL PROTECTED]
Subject: RE: Further LC Followup from IE RE: Potential bugs
identified in XHR LC Test Suite
On Sun, 15 Jun 2008, Zhenbin Xu wrote:
Ian wrote:
On Wed, 11 Jun 2008, Sunava Dutta wrote:
When Parsing Error happens, IE would still retain responseXML and
put error information on the object. Isnt this better than null
as
there�s more relevant information for the web developer?
How does one distinguish a document returned with parse error
information from one that happens to look like a parse error but
was
well-formed?
I wouldn't mind including more information but it seems like it
should
be out-of-band.
I am not sure if I understand your question. responseXML.parseError
has
the error information
http://msdn.microsoft.com/en-us/library/aa926483.aspx
Oh, I assumed Sunava meant a conforming Document object was returned. A
parseError-type object would be what I had in mind, yes. However, if we
do
this, then we should specify it. If we don't specify it, I'd rather
have
an exception.
[Zhenbin] The spec can simply state that a conforming document object
is returned, which includes out-of-band error information. This is what
IE does today and is a very reasonable approach that allows rich error
information for debugging.
The test is expecting us to return NULL in case open() has not
been
called. We throw an exception in IE. I�d pre fer if the spec
says �MUST return null OR an exception� otherwise I fear sites
today will be broken.
If a site is expecting an exception and gets null, then they'll get
an
exception when they try to dereferene the null, so in most cases it
seems like this would work anyway.
Properly written sites would have no problem one way or the other.
However if someone is writing a wrapper on top of XMLHTTP, clearly it
would make a difference on how to expose wrapped properties.
Not really; if the script is expecting an exception, and receives null
instead, then they'll just get an exception as soon as they dereference
the object, which in almost all cases will be straight away.
[Zhenbin] I should explain the scenario I talked about. For instance, if
I am to write a wrapper object myXHR, it makes a difference for me when I
do the following
myXHR.responseXML
if (!_innerResponseXML)
{
try
{
_innerResponseXML = _innerXHR.responseXML;
}
catch (e)
{
_myexception = e;
return _dummpyResponseXML;
}
}
return _innerResponseXML;
My try catch would not catch null. And the exception would be passed on
to my callers, which is not what I wanted.
If we are going to spec it to accommodate all existing browsers, we
would want to make it return null or INVALID_STATE_ERR exception.
We want interoperable behaviour, so defining it in this way would be a
bad
idea. (I don't really have an opinion either way about exception vs
null,
but it seems that we should just pick whatever is most commonly
implemented, which I'm guessing is what Anne did here.)
[Zhenbin] Fair enough. So let's pick one.
What is commonly implemented? Is it largest browser market share?
Is it number of enterprise applications written on top of particular browser?
Is it the number of browers, in which case I hope my fictional home grown
personal browser gets a vote :-)
From a pure technical point of view, predictably throw exception on state
violations is easier to understand. I hope you would agree there
is value to change spec for the sake of consistent programming model (which
happens to be the IE model).
I think it's important that we test that the DOM returned from XHR
is
DOM Core conformant just like any other, so this seems like an
important and relevant testing area for XHR.
That is not necessarily a good idea because you would then have to
mandate which level of DOM Core support is required. And if the spec
requires DOM level 3, that is big barrier for new user agent that
wants
to be compliant with XHR spec.
getElementById requires DOM Level 2. At the least the testing case
can
be changed to use getElementByTagName, which is DOM level 1.
I think expecting DOM Level 3 is the least of our worries -- after all,
that's a 3+ year old spec. So testing just DOM Level 2 is really not a
problem as far as I can tell. However, I agree that it would make sense
to
make the test pass if the UA didn't support that level of DOM on
regular
DOM objects too. The key is just to make sure that the objects returned
by
XHR are of equivalent DOM support as the rest of the UA's objects.
[Zhenbin
RE: Further LC Followup from IE RE: Potential bugs identified in XHR LC Test Suite
Exceptional situation in general doesn't hinder interoperability. It only becomes an issue if it is clearly stated in W3C spec and used as a way to measure if UA is compliant. We have provided our feedback to write spec in a away all current browsers are complaint. If the general sentiment is to pick one model, then lets pick the exception model, which is the model original XHR inventors picked. Technically because all other XHR methods/properties throw exceptions in case of state violation, exception for responseXML/responseText is better. If original XHR picked the null model, then we should go null model. The important thing really is consistency, thus predictability. Thanks! Zhenbin -Original Message- From: Jonas Sicking [mailto:[EMAIL PROTECTED] Sent: Monday, June 16, 2008 2:13 PM To: Zhenbin Xu Cc: Sunava Dutta; Web API public; IE8 Core AJAX SWAT Team; public- [EMAIL PROTECTED] Subject: Re: Further LC Followup from IE RE: Potential bugs identified in XHR LC Test Suite Zhenbin Xu wrote: The issue of return null or an exception is simply a compromise here. IE would throw an exception for state violations. Accessing responseXML before open() is a state violation so it would trigger exception. Other browsers may return null in such situation. In order to accommodate all browsers, the spec would have to be rewritten in some way. Please note that it is not a goal for the spec to be written in such a way that all existing browsers are conforming to the spec. It turned out that it was impossible to write a spec with that goal while still keeping the spec useful. So we no longer try to accomodate all browsers, but instead write a spec that leads to interoperability between browsers. We would certainly love to have the spec change to MUST throw INVALID_STATE_ERR exception, which is consistent with other INVALID_STATE_ERR cases. For instance, the spec says if send() is called before OPENED, it should trigger INVALID_STATE_ERR exception. Another example is that user agent must raise INVALID_STATE_ERR if status is not available. responseText and responseXML are the outlier in the spec. Personally I think it makes more sense to return 'null' from .responseXML. We at mozilla have not had any interoperability problems with this behavior. Exceptions are better left for exceptional circumstances. However I can't say that I think the behavior is very important to me one way or another, as long as it's usefully defined. Best Regards, Jonas Sicking
RE: Further LC Followup from IE RE: Potential bugs identified in XHR LC Test Suite
Inline... -Original Message- From: Ian Hickson [mailto:[EMAIL PROTECTED] Sent: Wednesday, June 11, 2008 9:34 PM To: Sunava Dutta Cc: Web API public; IE8 Core AJAX SWAT Team; [EMAIL PROTECTED] Subject: Re: Further LC Followup from IE RE: Potential bugs identified in XHR LC Test Suite Some quick comments on some of the comments regarding the tests: On Wed, 11 Jun 2008, Sunava Dutta wrote: http://tc.labs.opera.com/apis/XMLHttpRequest/responseXML/009.htm When Parsing Error happens, IE would still retain responseXML and put error information on the object. Isnt this better than null as there�s more relevant information for the web developer? How does one distinguish a document returned with parse error information from one that happens to look like a parse error but was well-formed? I wouldn't mind including more information but it seems like it should be out-of-band. [Zhenbin] I am not sure if I understand your question. responseXML.parseError has the error information http://msdn.microsoft.com/en-us/library/aa926483.aspx http://tc.labs.opera.com/apis/XMLHttpRequest/responseXML/001.htm The test is expecting us to return NULL in case open() has not been called. We throw an exception in IE. I�d pre fer if the spec says �MUST return null OR an exception� otherwise I fear sites today will be broken. If a site is expecting an exception and gets null, then they'll get an exception when they try to dereferene the null, so in most cases it seems like this would work anyway. [Zhenbin] Properly written sites would have no problem one way or the other. However if someone is writing a wrapper on top of XMLHTTP, clearly it would make a difference on how to expose wrapped properties. OTOH they likely already have handled both cases (currently IE throws exceptions and others return null). If we are going to spec it to accommodate all existing browsers, we would want to make it return null or INVALID_STATE_ERR exception. If we are going to spec it the right way, then I think throw exception is a more consistent design (see my reply to Jonas). http://tc.labs.opera.com/apis/XMLHttpRequest/responseXML/012.htm http://tc.labs.opera.com/apis/XMLHttpRequest/responseXML/013.htm This test really doesn�t test XHR here. It seems to be focused on manipulating the XML DOM. (I also don�t think Microsoft.XMLDOM supports getElementById for an XML document FYI). Also, if I'm barking up the wrong tree here please let me know! I think it's important that we test that the DOM returned from XHR is DOM Core conformant just like any other, so this seems like an important and relevant testing area for XHR. [Zhenbin] That is not necessarily a good idea because you would then have to mandate which level of DOM Core support is required. And if the spec requires DOM level 3, that is big barrier for new user agent that wants to be compliant with XHR spec. getElementById requires DOM Level 2. At the least the testing case can be changed to use getElementByTagName, which is DOM level 1. -- Ian Hickson U+1047E)\._.,--,'``.fL http://ln.hixie.ch/ U+263A/, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
RE: Further LC Followup from IE RE: Potential bugs identified in XHR LC Test Suite
The issue of return null or an exception is simply a compromise here. IE would throw an exception for state violations. Accessing responseXML before open() is a state violation so it would trigger exception. Other browsers may return null in such situation. In order to accommodate all browsers, the spec would have to be rewritten in some way. We would certainly love to have the spec change to MUST throw INVALID_STATE_ERR exception, which is consistent with other INVALID_STATE_ERR cases. For instance, the spec says if send() is called before OPENED, it should trigger INVALID_STATE_ERR exception. Another example is that user agent must raise INVALID_STATE_ERR if status is not available. responseText and responseXML are the outlier in the spec. Thanks! Zhenbin -Original Message- From: Jonas Sicking [mailto:[EMAIL PROTECTED] Sent: Thursday, June 12, 2008 11:20 AM To: Sunava Dutta Cc: Web API public; IE8 Core AJAX SWAT Team; [EMAIL PROTECTED] Subject: Re: Further LC Followup from IE RE: Potential bugs identified in XHR LC Test Suite _http://tc.labs.opera.com/apis/XMLHttpRequest/responseXML/001.htm_ The test is expecting us to return NULL in case open() has not been called. We throw an exception in IE. I'd pre fer if the spec says **MUST return null OR an exception** otherwise I fear sites today will be broken. How would that help sites that expect an exception, since it would still be conforming for the UA to return null? If anything, your proposal seems to make it harder for sites to code against the spec. / Jonas
