I would like to print various .fo.ftl PDF forms automatically using ECA services. I have written (using modified Sun code) a PDF print writer. All I need as a way to get a ByteArrayOutputStream from the .fo.ftl. Using org.ofbiz.widget.screen.ScreenFopViewHandler.render as an example, I can easily get a ByteArrayOutputStream and print it. Unfortunately, this routine requires a request and a response, although the response is never really used and can probably be null.
The question is how to get the request in the eca service. I read the entries at: http://ofbiz.markmail.org/message/54mvqvnmmhnwoahy?q=Session+in+Services where Rishi Solanki sez to just add: <attribute name="request" mode="IN" type="javax.servlet.http.HttpServletRequest"/> However, that did not work. You get an error "The following required parameter is missing: [IN] [printInvoicePDF.request]." So, I modified the code in in framework\webapp\src\org\ofbiz\webapp\event\ServiceEventHandler.java around line 220: // don't include locale, that is also taken care of below if ("locale".equals(name)) continue; //Added by skipd to set up a request and response if ("request".equals(name)) { serviceContext.put("request", request); continue; } if ("response".equals(name)) { serviceContext.put("responset", response); continue; } This all works now, but there are two problems. I am told that it breaks the job sandbox. Second, I have to modifiy the service definition of the service on which the eca runs to also include the request which kind of messes with the purpose of an ECA. Here is the code in ScreenFopViewHandler.render ScreenRenderer screens = new ScreenRenderer(writer, null, htmlScreenRenderer); screens.populateContextForRequest(request, response, servletContext); ... screens.getContext().put("formStringRenderer", new FoFormRenderer(request, response)); screens.render(page); .... ByteArrayOutputStream out = new ByteArrayOutputStream(); TransformerFactory transFactory = TransformerFactory.newInstance(); Fop fop = fopFactory.newFop(contentType, out); Transformer transformer = transFactory.newTransformer(); // set the input source (XSL-FO) and generate the output stream of contentType Reader reader = new StringReader(writer.toString()); Source src = new StreamSource(reader); ... Then, instead of: // write to the browser try { out.writeTo(response.getOutputStream()); response.getOutputStream().flush(); } catch (IOException e) { throw new ViewHandlerException("Unable write to browser OutputStream", e); } This PDFPrint(out, printerdefinition); This all works. No browser is involved. The only issue is that the calls to populateContextForRequest, and FoFormRenderer both require request and response arguments. The response is only used (as far as I can tell) for populating a beanshell, and it is never used in any of the .bsh scripts for PDF generation. It is never used in FoFormRenderer. In any case, I would not be writing to the browser, only capturing the output stream and sending it on to the PDFPrint routine that I have written. In other words, what I want to do is use the various ....fo.ftl screens for the various forms, render them to a ByteArrayOutputStream, and send that ByteArrayOutputStream to a routine that knows how to print the resulting PDF stream directly to a printer. I expect that this would be useful to lots of people. All of my customers want it as it saves them several minutes per order, i.e., accounts payable does not have to look up all the completed orders, and print invoices. Pick and packing slips get printed automatically, when an ecommerce order is placed, a pick slip is automatically printed, etc, etc. I also would like to contribute the code back to the project and so would like whatever method used to be "approved". Skip
