Here is the sum total of the code necessary to handle the URL and SOAP
versions of an SRW request. This is so trivial that it makes my heart ache.
It is also intuitively obvious, which is a great virtue.
Please show me how it would be changed to allow it to know about the
database context.
Ralph
import java.io.PrintWriter;
import javax.servlet.http.HttpServletRequest;
import javax.servlet.http.HttpServletResponse;
public class ZNGRPCServer {
// handle the URL form of an SRW searchRequest
public void doGet (HttpServletRequest request, HttpServletResponse
response)
throws javax.servlet.ServletException, java.io.IOException {
PrintWriter out = response.getWriter();
String query=request.getParameter("query");
response.setContentType("text/xml");
out.println(
searchRetrieve(query, request.getParameter("startRecord"),
request.getParameter("maximumRecords"),
request.getParameter("responseSchema"),
request.getParameter("recordSchema")));
out.close();
}
// form the actual XML response record (probably should be using XML
tools here)
private String makeResponse(int postings) {
StringBuffer response=new StringBuffer();
response.append("<?xml version=\"1.0\" encoding=\"UTF-8\"?>");
response.append("<searchRetrieveResponse
xmlns=\"urn:z3950:ZNG_Prototype1\">");
response.append("<totalHits>"+postings+"</totalHits>");
response.append("<records></records>");
response.append("<status><statusCode>0</statusCode></status>");
response.append("</searchRetrieveResponse>");
return response.toString();
}
// the method called by the SOAP RPCRouter
public String searchRetrieve(String query, int startRecord,
int maximumRecords, String responseSchema, String recordSchema) {
// your local database magic happens here
return makeResponse(999);
}
// the method called by doGet() that handles the URL request
private String searchRetrieve(String query, String startRecord,
String maximumRecords, String responseSchema, String recordSchema) {
int max=0, start=0;
if(startRecord!=null)
try {
start=Integer.parseInt(startRecord);
}
catch(NumberFormatException e) {} // take the default
if(maximumRecords!=null)
try {
max=Integer.parseInt(maximumRecords);
}
catch(NumberFormatException e) {} // take the default
return searchRetrieve(query, start, max, responseSchema,
recordSchema);
}
}
> -----Original Message-----
> From: Matthew Dovey [mailto:[log in to unmask]]
> Sent: Tuesday, October 30, 2001 5:32 PM
> To: [log in to unmask]
> Subject: Re: RPC Context
>
>
> Ralph,
>
> The approach, I've outlined would work using the Apache SOAP/Java
> toolkit although it would be the URN defined in the deployment
> descriptor passed via the HTTP SOAP action header which would
> therefore
> identify the database. I think it perfectly legal for the HTTP
> SOAP-Action to determine the database as the URL used. In general, the
> model is that you connect to a different ZNG SOAP Service for a
> different database.
>
> However, if you wanted to determined the database on the with
> the Apache
> SOAP/Java toolkit you would run two instances of the RPC Router on
> different URLS.
>
> Both these approaches are fairly trivial and not particularly
> non-standard uses of the Apache toolkit
>
> Matthew
>
>
>
> > -----Original Message-----
> > From: LeVan,Ralph [mailto:[log in to unmask]]
> > Sent: 30 October 2001 20:59
> > To: [log in to unmask]
> > Subject: Re: RPC Context
> >
> > I can do all sorts of trickery to make this problem go
> away. But, one
> of
> > the points of this exercise is to do something simple. By
> that I mean
> > that
> > it should use standard tools as much as possible.
> >
> > If we are doing standard SOAP RPC, then there are standard ways of
> > generating the RPC client and server. If I use the standard RPC
> server
> > tools, then I don't have access to context.
> >
> > If you know how to do this using standard tools, then please let me
> know.
> > I've been beating my head against this problem for some time now.
> >
> > The problem goes away when we add database to the list of RPC
> parameters.
> > I've been able to put together a very slim client and
> server built on
> the
> > RPC model when I added database.
> >
> > Ralph
> >
> > > -----Original Message-----
> > > From: Matthew Dovey [mailto:[log in to unmask]]
> > > Sent: Tuesday, October 30, 2001 3:50 PM
> > > To: [log in to unmask]
> > > Subject: Re: RPC Context
> > >
> > >
> > > By ID do you mean URL?
> > >
> > > Basically you bind different URLs to different
> instantiations of the
> > > Java class.
> > >
> > > One approach would be to write an abstract class with all the ZNG
> > > handling, something like
> > >
> > > public abstract class BaseZNGObject {
> > > String database = null
> > >
> > > public Test() {
> > > }
> > >
> > > public ZNG(
> > >
> > > ...
> > > // All the ZNG handling, database searching stuff.
> > >
> > >
> > > }
> > >
> > >
> > > Then write classes of the form
> > >
> > > public class Database1ZNGObject extends BaseZNGObject {
> > >
> > > public Database1ZNGObject () {
> > > database="database1";
> > > }
> > > }
> > >
> > > public class Database2ZNGObject extends BaseZNGObject {
> > >
> > > public Database1ZNGObject () {
> > > database="database2";
> > > }
> > > }
> > >
> > > And then bind these classes to different URLs using the SOAP
> toolkit.
> > >
> > >
> > > Matthew
> > >
> > > > -----Original Message-----
> > > > From: LeVan,Ralph [mailto:[log in to unmask]]
> > > > Sent: 30 October 2001 15:20
> > > > To: [log in to unmask]
> > > > Subject: RPC Context
> > > >
> > > > I'm still trying to figure out how to tell what
> database my query
> is
> > > for.
> > > > The problem may be in my toolkit, but I've got to ask this
> question
> > > again.
> > > >
> > > > When a SOAP request is received, a special SOAP servlet
> > > processes the
> > > > message. It looks at the ID for the message, looks
> that ID up in
> a
> > > table
> > > > and determines the Java class that contains the method
> that should
> > > process
> > > > the message. It then pulls the parameters for the method out of
> the
> > > > message. It then creates an instance of the Java class and
> > > calls the
> > > > method
> > > > (searchRetrieve in our case) with the parameters.
> > > >
> > > > Nowhere in there does searchRetrieve have access to the URL that
> > > contained
> > > > the message. That URL somehow encoded the database to be
> searched.
> > > That
> > > > means that my searchRetrieve method does not know what
> database it
> > > should
> > > > be
> > > > searching.
> > > >
> > > > How do we get around this problem without exposing the database
> name
> > > as a
> > > > parameter to searchRetrieve.
> > > >
> > > > Thanks!
> > > >
> > > > Ralph
> > >
>
|
