Jarmo, Please go thru the old commits. It's a hairy situation trying to support all kinds of stuff. Yes, if we get support from you all in terms of patches, we can do things better. Please remember this is a open source, volunteer project :(
http://marc.theaimsgroup.com/?l=axis-dev&w=2&r=1&s=unwrapArrays&q=b http://marc.theaimsgroup.com/?l=axis-dev&w=2&r=1&s=wrapArrays&q=b thanks, dims On 11/10/05, Pete Hendry <[EMAIL PROTECTED]> wrote: > > > >>-----Original Message----- > >>From: Jarmo Doc [mailto:[EMAIL PROTECTED] > >>Sent: Thursday, 10 November 2005 5:52 p.m. > >>To: [email protected] > >>Subject: RE: When to use undocumented wsdl2java --wrapArrays? > >> > >>Thanks for the comprehensive answer Pete. > >> > >>Having switched to using wrapped arrays (and changing server > >>code, client > >>code, and WSDL!) my code now works. I was surprised to find > >>that I had to > >>change the server-side code so that each class contained an > >>ArrayOfXXX > >>rather than XXX []. Without this (seemingly unnecessary) layer of > >>indirection I couldn't get the code to work at all. For > >>example: using > >>simple document/literal rather than wrapped caused me to lose > >>all 2nd and > >>subsequent parameters to all my operations. > >> > > This is an issue with Axis. I believe .NET and others (including ours) > recognize a wrapped array and map to [] accordingly without the wrapping > Java class. This is obviously a drawback which may want you to try to get it > working without wrapping if that makes the code easier. > > >><rant>Who would specify an interface where this kind of mess > >>was even a > >>possibility never mind a likelihood? If you want to surround > >>arrays with > >><array> ... </array> or something else then that's fine but > >>surely you don't > >>have to involve me (and my server-side code, client-side > >>code, and WSDL!) in > >>the decision. We've had arrays since about the Stone Age so > >>it's not as if > >>the people specifying these interfaces didn't know about them.</rant> > >> > > There is, unfortunately, a number of areas in schema that map poorly to some > (all in some cases) programming languages, or present alternatives (such as > arrays) which each have different benefits/drawbacks. Don't hold your breath > for this to be resolved as schema 1.1 is not touching on these problems I > don't believe. > > Pete > > -- Davanum Srinivas : http://wso2.com/blogs/
