Status of Safari support for XForms

> Colin,
>
> > just wondering where we are up to with support for Safari in XForms
> > NG?
>
> Ideally, this would be sponsored. If several parties are interested,
> they could even join up together.
>
> > I'm just going live with my first XForm, and as luck would have it,
> > looking at the logs, a significant part of their visitors are using
> > Safari, so looks like we are in for some flack.
> > I could get away with saying latest version of Safari, but no Safari
> is
> > looking problematic.
> > IE users seem to be almost all on IE6 so OK there and Mozilla-derived
> > browsers seem to be no problem.
>
> Yes, that sounds like a good assessment.
>
> > Some research shows Sarissa claims to support Safari, so wondering
> > where the problems are.
>
> > What I am seeing on the latest Safari is that the form displays OK,
> > fills in OK (not doing any fancy constraints or repeats or anything
> yet)
> > and when you click submit, then it just sits there.
> > I found someone saying he had a workround for that, if that is where
> the
> > XForms NG problem is, which may or may not be useful/relevant.
>
> We are wondering too, but this is hard to tell without seriously
> looking into it.
>
> -Erik
>
>
> --
> You receive this message as a subscriber of the
> [hidden email] mailing list.
> To unsubscribe: mailto:[hidden email]
> For general help: mailto:[hidden email]?subject=help
> ObjectWeb mailing lists service home page: http://www.objectweb.org/wws

> Hi Erik,
>
> I don't know that you fixing NG is the answer - OPS is presumably not
> the only system doing Ajax development and hitting problems with Safari
> - the better answer would seem to be to get Safari fixed.
>
> If we could categorize where the failures are, then we could work with
> Apple's WebKit team to get them addressed.
> (Or if they are in one of the Ajax libraries you use, perhaps they might
> have the answer).
>
> It's also presumably possible that there are a number of issues, and so
> maybe we can prioritize them.
> For myself, my concern is that even the most basic of forms, no repeats
> constraints or anything that will need interaction, is not working - the
> external appearance is that the submit does not work. So if that could
> be "fixed", that would actually be a huge leap forward.
> (If that then meant that invalid data could be sent to my pipeline, then
> that is something in my domain that I can take care of).
>
> Hope that makes sense and is more doable
> Thanks again
> Colin
>
> On Feb 1, 2006, at 5:07 AM, Erik Bruchez wrote:
>
>> Colin,
>>
>> > just wondering where we are up to with support for Safari in XForms
>> > NG?
>>
>> Ideally, this would be sponsored. If several parties are interested,
>> they could even join up together.
>>
>> > I'm just going live with my first XForm, and as luck would have it,
>> > looking at the logs, a significant part of their visitors are using
>> > Safari, so looks like we are in for some flack.
>> > I could get away with saying latest version of Safari, but no Safari is
>> > looking problematic.
>> > IE users seem to be almost all on IE6 so OK there and Mozilla-derived
>> > browsers seem to be no problem.
>>
>> Yes, that sounds like a good assessment.
>>
>> > Some research shows Sarissa claims to support Safari, so wondering
>> > where the problems are.
>>
>> > What I am seeing on the latest Safari is that the form displays OK,
>> > fills in OK (not doing any fancy constraints or repeats or anything
>> yet)
>> > and when you click submit, then it just sits there.
>> > I found someone saying he had a workround for that, if that is where
>> the
>> > XForms NG problem is, which may or may not be useful/relevant.
>>
>> We are wondering too, but this is hard to tell without seriously
>> looking into it.
>>
>> -Erik
>>
>>
>> --
>> You receive this message as a subscriber of the
>> [hidden email] mailing list.
>> To unsubscribe: mailto:[hidden email]
>> For general help: mailto:[hidden email]?subject=help
>> ObjectWeb mailing lists service home page: http://www.objectweb.org/wws
>
>
>
> ------------------------------------------------------------------------
>
>
> --
> You receive this message as a subscriber of the [hidden email] mailing list.
> To unsubscribe: mailto:[hidden email]
> For general help: mailto:[hidden email]?subject=help
> ObjectWeb mailing lists service home page: http://www.objectweb.org/wws

> Colin,
>
> I agree that ideally, all browsers should behave the same, but we all
> know that the real world does not work like that: browsers may have
> bugs, sure, but they also sometime interpret specs differently. In
> this particular case, we don't know which it is yet.
>
> But we do think that modifying things on our side is the answer,
> because if people see that Google can make their Ajax apps work with
> the latest Safari, they will expect OPS to be able to do that too,
> whether Safari is bug-ridden or not.
>
> -Erik

> Hi Erik,
>
> I agree that's a good point - if Google can do it, why can't we -
> though that's unlikely to be the thought going through a visitor's
> mind.
>
> I still think it is possible that the WebKit team can provide a
> solution. They did a lot of work last year to get their CSS
> implementation team fully up to spec. Safari pdates come out
> frequently and I would be surprised if they are not working on
> improving their Ajax support.
>
> Seems like I am the only one interested in Safari support.
> I need to work out some sort of debugging environment for javascript
> in Safari, and I don't know how much I can really achieve, but I am
> willing to do some research on what might be the issue, and maybe even
> look for a fix.
> Are there any other packages apart from sarissa contributing to NG?
>
> Best regards
> Colin
>
> On Feb 1, 2006, at 11:00 AM, Erik Bruchez wrote:
>
>> Colin,
>>
>> I agree that ideally, all browsers should behave the same, but we all
>> know that the real world does not work like that: browsers may have
>> bugs, sure, but they also sometime interpret specs differently. In
>> this particular case, we don't know which it is yet.
>>
>> But we do think that modifying things on our side is the answer,
>> because if people see that Google can make their Ajax apps work with
>> the latest Safari, they will expect OPS to be able to do that too,
>> whether Safari is bug-ridden or not.
>>
>> -Erik
> --
> You receive this message as a subscriber of the
> [hidden email] mailing list.
> To unsubscribe: mailto:[hidden email]
> For general help: mailto:[hidden email]?subject=help
> ObjectWeb mailing lists service home page: http://www.objectweb.org/wws

> Hi Erik,
>
> I agree that's a good point - if Google can do it, why can't we - though
> that's unlikely to be the thought going through a visitor's mind.
>
> I still think it is possible that the WebKit team can provide a
> solution. They did a lot of work last year to get their CSS
> implementation team fully up to spec. Safari pdates come out frequently
> and I would be surprised if they are not working on improving their Ajax
> support.

> Hi Erik,
>
> well, it's always nice to know that a problem was simple to solve.
> So I am sending this email now, in the hope that it will be true in
> this case.
> We still only have one system here running Tiger, with the latest
> Safari, and it's not mine.
> Anyway, I did get on and set up Javascript debugging and this is what
> I see.
> No messages when the form loads.
> If I click in/on a control, just the first time I get
> Value undefined (result of expression request.loadXML) is not object.
> Line 44
>
> It doesn't say, but that is in ops/javascript/xforms.js
>
> The relevant lines are
> function xformsCreateElementNS(namespaceURI, qname) {
>
> var localName = qname.indexOf(":") == -1 ? "" : ":" + qname.substr(0,
> qname.indexOf(":"));
>
> var request = Sarissa.getDomDocument();
>
> request.loadXML("<" + qname + " xmlns" + localName + "='" +
> namespaceURI + "'/>");  <-- line 44
>
> return request.documentElement;
> }
>
> I noticed OPS has version 0.9.5.2 of sarissa.js and the latest is
> 0.9.6.1 so I replaced my sarissa.js with the latest, and the result
> was the same.
>
> Safari doesn't seem to give anything more in terms of javascript
> debugging, and I'm not an expert in javascript and certainly not HTML
> DOM, so I'm hoping this is enough for you to see what the problem is
> and I don't need to go any further.
> (Though if that's not so, just say so, and I will see what else I can
> uncover).
>
> Thanks & regards
> Colin

> Just as a quick follow-up on this.
> I fear that there are good reasons you are still on the older version of
> sarissa and that it will not be trivial to upgrade, and this doesn't
> tell us there is actually an answer.
> Anyway, just checked the changelog for sarissa 0.9.6.1 and it includes
> this earlier entry
>
> 0.9.6
> * Deprecated stuff as warned in 0.9.4
>
> * Big change: loadXML is no longer used, as there's no way to emulate it
> in Opera.
> So i just removed it directly; to load an XML string to a DOM, you have
> to use the
> DOMParser object. No way to do anything like that in Safari though.
>
> * Fixed Opera and Safari issues thanks to Phil Endecott and Conrad Chu
>
> Oh wait, why did I start reading the doc from the middle ;-)
>
> 0.9.6.1
> * Fixed bug in Sarissa.updateContentFromNode thanks to Franky Braem
>
> * Added Safari DOMParser support thanks to a wonderfull idea published
> by Nate Steiner on his blog that utilizes the data: URI scheme and
> Erik Arvidsson that suggested this could easily be wrapped to emulate
> DOMParser. Many thanks to cedric for pointing out the above (between
> other things).
>
> Hope that is meaningful and useful to you
> Colin
>
>
> On Feb 3, 2006, at 12:52 PM, Colin O'Brien wrote:
>
>> Hi Erik,
>>
>> well, it's always nice to know that a problem was simple to solve.
>> So I am sending this email now, in the hope that it will be true in
>> this case.
>> We still only have one system here running Tiger, with the latest
>> Safari, and it's not mine.
>> Anyway, I did get on and set up Javascript debugging and this is what
>> I see.
>> No messages when the form loads.
>> If I click in/on a control, just the first time I get
>> Value undefined (result of expression request.loadXML) is not object.
>> Line 44
>>
>> It doesn't say, but that is in ops/javascript/xforms.js
>>
>> The relevant lines are
>> function xformsCreateElementNS(namespaceURI, qname) {
>>
>> var localName = qname.indexOf(":") == -1 ? "" : ":" + qname.substr(0,
>> qname.indexOf(":"));
>>
>> var request = Sarissa.getDomDocument();
>>
>> request.loadXML("<" + qname + " xmlns" + localName + "='" +
>> namespaceURI + "'/>");  <-- line 44
>>
>> return request.documentElement;
>> }
>>
>> I noticed OPS has version 0.9.5.2 of sarissa.js and the latest is
>> 0.9.6.1 so I replaced my sarissa.js with the latest, and the result
>> was the same.
>>
>> Safari doesn't seem to give anything more in terms of javascript
>> debugging, and I'm not an expert in javascript and certainly not HTML
>> DOM, so I'm hoping this is enough for you to see what the problem is
>> and I don't need to go any further.
>> (Though if that's not so, just say so, and I will see what else I can
>> uncover).
>>
>> Thanks & regards
>> Colin
>
>
>
> ------------------------------------------------------------------------
>
>
> --
> You receive this message as a subscriber of the [hidden email] mailing list.
> To unsubscribe: mailto:[hidden email]
> For general help: mailto:[hidden email]?subject=help
> ObjectWeb mailing lists service home page: http://www.objectweb.org/wws

> Colin,
>
> I entered a bug to track this:
>
> http://forge.objectweb.org/tracker/index.php?
> func=detail&aid=304623&group_id=168&atid=350207
>
> However, you said that upgrading to the latest Sarissa was not enough  
> to make things work, right?
>
> -Erik
>
> Colin O'Brien wrote:
>> Just as a quick follow-up on this.
>> I fear that there are good reasons you are still on the older version  
>> of sarissa and that it will not be trivial to upgrade, and this  
>> doesn't tell us there is actually an answer.
>> Anyway, just checked the changelog for sarissa 0.9.6.1 and it  
>> includes this earlier entry
>> 0.9.6
>> * Deprecated stuff as warned in 0.9.4
>> * Big change: loadXML is no longer used, as there's no way to emulate  
>> it in Opera.
>> So i just removed it directly; to load an XML string to a DOM, you  
>> have to use the
>> DOMParser object. No way to do anything like that in Safari though.
>> * Fixed Opera and Safari issues thanks to Phil Endecott and Conrad Chu
>> Oh wait, why did I start reading the doc from the middle ;-)
>> 0.9.6.1
>> * Fixed bug in Sarissa.updateContentFromNode thanks to Franky Braem
>> * Added Safari DOMParser support thanks to a wonderfull idea published
>> by Nate Steiner on his blog that utilizes the data: URI scheme and
>> Erik Arvidsson that suggested this could easily be wrapped to emulate
>> DOMParser. Many thanks to cedric for pointing out the above (between  
>> other things).
>> Hope that is meaningful and useful to you