|
View:
New views
12 Messages
—
Rating Filter:
Alert me
|
|
|
how about we honor style sheet in wsdlHi,
For the wsdl like <?xml version="1.0" encoding="UTF-8"?> <?xml-stylesheet type="text/xsl" href="http://api.bioinfo.no/stylesheets/WSDLDocumentation.xslt"?> <definitions xmlns="http://schemas.xmlsoap.org/wsdl/" .... Is it possible that we publish stylesheet as well when we publish the wsdl, I mean if customer get wsdl from http://serviceaddress?wsdl, they can also see the stylesheet their. Or any reason we shouldn't do it? Thanks Freeman |
|
|
Re: how about we honor style sheet in wsdlI don't follow exactly.
If the URL in the PI were relative, then we could imagine serving it up. But as an absolute URL, it's up to the user to configure the container to deliver it, no? On Wed, Jul 23, 2008 at 2:48 AM, Freeman Fang <freeman.fang@...> wrote: > Hi, > > For the wsdl like > <?xml version="1.0" encoding="UTF-8"?> > <?xml-stylesheet type="text/xsl" > href="http://api.bioinfo.no/stylesheets/WSDLDocumentation.xslt"?> > <definitions xmlns="http://schemas.xmlsoap.org/wsdl/" > .... > > Is it possible that we publish stylesheet as well when we publish the wsdl, > I mean if customer get wsdl from http://serviceaddress?wsdl, they can also > see the stylesheet their. > Or any reason we shouldn't do it? > > Thanks > Freeman > > > |
|
|
Re: how about we honor style sheet in wsdlI think the problem is that when served from CXF, the PI is removed
from the original WSDL. On Wed, Jul 23, 2008 at 2:05 PM, Benson Margulies <bimargulies@...> wrote: > I don't follow exactly. > > If the URL in the PI were relative, then we could imagine serving it > up. But as an absolute URL, it's up to the user to configure the > container to deliver it, no? > > > > On Wed, Jul 23, 2008 at 2:48 AM, Freeman Fang <freeman.fang@...> wrote: >> Hi, >> >> For the wsdl like >> <?xml version="1.0" encoding="UTF-8"?> >> <?xml-stylesheet type="text/xsl" >> href="http://api.bioinfo.no/stylesheets/WSDLDocumentation.xslt"?> >> <definitions xmlns="http://schemas.xmlsoap.org/wsdl/" >> .... >> >> Is it possible that we publish stylesheet as well when we publish the wsdl, >> I mean if customer get wsdl from http://serviceaddress?wsdl, they can also >> see the stylesheet their. >> Or any reason we shouldn't do it? >> >> Thanks >> Freeman >> >> >> > -- Cheers, Guillaume Nodet ------------------------ Blog: http://gnodet.blogspot.com/ |
|
|
Re: how about we honor style sheet in wsdlOh! Is there a JIRA for this? It has to be a side-effect of patching
the address, no? On Wed, Jul 23, 2008 at 8:21 AM, Guillaume Nodet <gnodet@...> wrote: > I think the problem is that when served from CXF, the PI is removed > from the original WSDL. > > On Wed, Jul 23, 2008 at 2:05 PM, Benson Margulies <bimargulies@...> wrote: >> I don't follow exactly. >> >> If the URL in the PI were relative, then we could imagine serving it >> up. But as an absolute URL, it's up to the user to configure the >> container to deliver it, no? >> >> >> >> On Wed, Jul 23, 2008 at 2:48 AM, Freeman Fang <freeman.fang@...> wrote: >>> Hi, >>> >>> For the wsdl like >>> <?xml version="1.0" encoding="UTF-8"?> >>> <?xml-stylesheet type="text/xsl" >>> href="http://api.bioinfo.no/stylesheets/WSDLDocumentation.xslt"?> >>> <definitions xmlns="http://schemas.xmlsoap.org/wsdl/" >>> .... >>> >>> Is it possible that we publish stylesheet as well when we publish the wsdl, >>> I mean if customer get wsdl from http://serviceaddress?wsdl, they can also >>> see the stylesheet their. >>> Or any reason we shouldn't do it? >>> >>> Thanks >>> Freeman >>> >>> >>> >> > > > > -- > Cheers, > Guillaume Nodet > ------------------------ > Blog: http://gnodet.blogspot.com/ > |
|
|
Re: how about we honor style sheet in wsdlOn Jul 23, 2008, at 8:21 AM, Guillaume Nodet wrote: > I think the problem is that when served from CXF, the PI is removed > from the original WSDL. > Does wsdl4j keep the PI? That's the first question. Actually, just looked in StaxUtils and it wouldn't write the PI: public static void writeDocument(Document d, XMLStreamWriter writer, boolean writeProlog, boolean repairing) throws XMLStreamException { if (writeProlog) { writer.writeStartDocument(); } Element root = d.getDocumentElement(); writeElement(root, writer, repairing); if (writeProlog) { writer.writeEndDocument(); } } That's pretty much all wrong. It needs to iterate over all the children instead of just the DocumentElement. So there's the spot that needs fixing. Dan > On Wed, Jul 23, 2008 at 2:05 PM, Benson Margulies <bimargulies@... > > wrote: >> I don't follow exactly. >> >> If the URL in the PI were relative, then we could imagine serving it >> up. But as an absolute URL, it's up to the user to configure the >> container to deliver it, no? >> >> >> >> On Wed, Jul 23, 2008 at 2:48 AM, Freeman Fang >> <freeman.fang@...> wrote: >>> Hi, >>> >>> For the wsdl like >>> <?xml version="1.0" encoding="UTF-8"?> >>> <?xml-stylesheet type="text/xsl" >>> href="http://api.bioinfo.no/stylesheets/WSDLDocumentation.xslt"?> >>> <definitions xmlns="http://schemas.xmlsoap.org/wsdl/" >>> .... >>> >>> Is it possible that we publish stylesheet as well when we publish >>> the wsdl, >>> I mean if customer get wsdl from http://serviceaddress?wsdl, they >>> can also >>> see the stylesheet their. >>> Or any reason we shouldn't do it? >>> >>> Thanks >>> Freeman >>> >>> >>> >> > > > > -- > Cheers, > Guillaume Nodet > ------------------------ > Blog: http://gnodet.blogspot.com/ --- Daniel Kulp dkulp@... http://www.dankulp.com/blog |
|
|
Re: how about we honor style sheet in wsdlShould WSDLs have stylesheets? After all, as they say, web services are for computers what web browsers are for human beings--but computers don't care about how data looks, so the need for stylesheets is less clear. (Although I certainly understand that the intention of the stylesheet is for the WSDL to be more pleasing to the human.)
Another concern about allowing stylesheets is that they could visually mask or alter the web service's appearance. Hide some operations, for example, possibly even change values (I think)--how much of a concern is that? We have a JAX-WS requirement that the wsdl be viewable at http://endpointurl?wsdl, but arguably that is still met even if the WSDL looks different to the human eye (because of the stylesheet) than it does when processed by the machine. Regards, Glen
|
|
|
Re: how about we honor style sheet in wsdlHi Glen,
Actually I have same concern with you. This request is from a servicemix user but I'm not sure if it's reasonable for wsdl, so I post question here to get more feedback. Regards Freeman Glen Mazza wrote: > Should WSDLs have stylesheets? After all, as they say, web services are for > computers what web browsers are for human beings--but computers don't care > about how data looks, so the need for stylesheets is less clear. (Although > I certainly understand that the intention of the stylesheet is for the WSDL > to be more pleasing to the human.) > > Another concern about allowing stylesheets is that they could visually mask > or alter the web service's appearance. Hide some operations, for example, > possibly even change values (I think)--how much of a concern is that? We > have a JAX-WS requirement that the wsdl be viewable at > http://endpointurl?wsdl, but arguably that is still met even if the WSDL > looks different to the human eye (because of the stylesheet) than it does > when processed by the machine. > > Regards, > Glen > > > Freeman Fang wrote: > >> Hi, >> >> For the wsdl like >> <?xml version="1.0" encoding="UTF-8"?> >> <?xml-stylesheet type="text/xsl" >> href="http://api.bioinfo.no/stylesheets/WSDLDocumentation.xslt"?> >> <definitions xmlns="http://schemas.xmlsoap.org/wsdl/" >> .... >> >> Is it possible that we publish stylesheet as well when we publish the >> wsdl, I mean if customer get wsdl from http://serviceaddress?wsdl, they >> can also see the stylesheet their. >> Or any reason we shouldn't do it? >> >> Thanks >> Freeman >> >> > > |
|
|
Re: how about we honor style sheet in wsdlI don't see it being our job to remove things from a WSDL file. Who
knows what protocol someone might be inventing with PIs? Style sheets are just another PI. I'm going to go ahead and make a fix. |
|
|
Re: how about we honor style sheet in wsdlHave fun. But I don't think PI's are part of the WSDL document--they are external to it, however I'm seeing some open source products that provide stylesheets for WSDLs, so it may not be that big a deal. At any rate, make sure your change does not result in SOAP messages being allowed to have PI's--that remains a no-no (3rd paragraph of http://www.w3.org/TR/2003/REC-soap12-part1-20030624/#soapenv, and also the WS-I Basic Profile). Glen |
|
|
Re: how about we honor style sheet in wsdl> Have fun. But I don't think PI's are part of the WSDL document--they are
> external to it, however I'm seeing some open source products that provide > stylesheets for WSDLs, so it may not be that big a deal. > > At any rate, make sure your change does not result in SOAP messages being > allowed to have PI's--that remains a no-no (3rd paragraph of > http://www.w3.org/TR/2003/REC-soap12-part1-20030624/#soapenv, and also the > WS-I Basic Profile). > Well, I suppose that I've just made it hypothetically possible for a user of CXF to sneak a PI out on the wire, assuming that we have any code path where the user gives us a DOM tree and we pass it along. It's presumably arguable whether it is up to us to step on such things. Does anyone know of such a code path? |
|
|
Re: how about we honor style sheet in wsdlMost likely, in those cases, we would be calling writeDocument with the writePrologue falg set to false as it would be encapsulated in a soap message or something that would have already written that. Most likely, if prologue writing is turned off, we DON'T want to write any PI's either as they would be invalid. Thus, my suggestion would be to use the old behavior if false and your new code if true. Dan On Jul 24, 2008, at 7:06 PM, Benson Margulies wrote: >> Have fun. But I don't think PI's are part of the WSDL document-- >> they are >> external to it, however I'm seeing some open source products that >> provide >> stylesheets for WSDLs, so it may not be that big a deal. >> >> At any rate, make sure your change does not result in SOAP messages >> being >> allowed to have PI's--that remains a no-no (3rd paragraph of >> http://www.w3.org/TR/2003/REC-soap12-part1-20030624/#soapenv, and >> also the >> WS-I Basic Profile). >> > > Well, I suppose that I've just made it hypothetically possible for a > user of CXF to sneak a PI out on the wire, assuming that we have any > code path where the user gives us a DOM tree and we pass it along. > It's presumably arguable whether it is up to us to step on such > things. > > Does anyone know of such a code path? --- Daniel Kulp dkulp@... http://www.dankulp.com/blog |
|
|
Re: how about we honor style sheet in wsdlI'm more likely to remember to add this wrinkle if you reopen the JIRA :-)
On Mon, Jul 28, 2008 at 9:59 AM, Daniel Kulp <dkulp@...> wrote: > > Most likely, in those cases, we would be calling writeDocument with the > writePrologue falg set to false as it would be encapsulated in a soap > message or something that would have already written that. Most likely, > if prologue writing is turned off, we DON'T want to write any PI's either as > they would be invalid. Thus, my suggestion would be to use the old > behavior if false and your new code if true. > > Dan > > > On Jul 24, 2008, at 7:06 PM, Benson Margulies wrote: > >>> Have fun. But I don't think PI's are part of the WSDL document--they are >>> external to it, however I'm seeing some open source products that provide >>> stylesheets for WSDLs, so it may not be that big a deal. >>> >>> At any rate, make sure your change does not result in SOAP messages being >>> allowed to have PI's--that remains a no-no (3rd paragraph of >>> http://www.w3.org/TR/2003/REC-soap12-part1-20030624/#soapenv, and also >>> the >>> WS-I Basic Profile). >>> >> >> Well, I suppose that I've just made it hypothetically possible for a >> user of CXF to sneak a PI out on the wire, assuming that we have any >> code path where the user gives us a DOM tree and we pass it along. >> It's presumably arguable whether it is up to us to step on such >> things. >> >> Does anyone know of such a code path? > > --- > Daniel Kulp > dkulp@... > http://www.dankulp.com/blog > > > > > |
| Free Forum Powered by Nabble | Forum Help |