gmane.comp.java.dwr.user

RE: class mapping E5 & E6: Offer a "lighter" bi-directional class-mapping scheme & Allow alternative class-mapping schem

Mike Wilson — 2008-09-06T19:50:25

Yes, several of the recent class mapping additions have shown to be quite simple and straight-forward in the end. You get that warm and fuzzy feeling :-) Btw, by all means rename added functions and members according to your liking. Best regards Mike _____ From: joseph.walker-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org [mailto:joseph.walker-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org] On Behalf Of Joe Walker Sent: den 6 september 2008 12:08 To: users-EyPigyGktj4FDOXUYO6UHQ< at >public.gmane.org Subject: Re: [dwr-user] class mapping E5 & E6: Offer a "lighter" bi-directional class-mapping scheme & Allow alternative class-mapping schem Thanks Mike. The change looks nice and simple, and even with added tests. Thanks. Joe. On Fri, Sep 5, 2008 at 9:53 PM, Mike Wilson public.gmane.org> wrote: E5 is now implemented (we will not do E6) and I have resolved http://bugs.directwebremoting.org/bugs/browse/DWR-269: I have now checked in the following enhancement: It is now possible t

Re: Re: Build 110

Thomas Manson — 2008-09-06T19:24:56

Sorry Joe, my mistake... I didn't read carrefuly your optimisation ;) Thomas. On Sat, Sep 6, 2008 at 1:59 PM, Joe Walker public.gmane.org> wrote:

Re: Spring change/patch

Jose Noheda — 2008-09-06T15:46:18

Not sure that I agree here. The destroy method is common to all containers (and hence it's placed in DefaultContainer). It's failing in Spring but it can fail in any other conainer and for the same reasons. DWR just needs to invoke the close method on the beans it owns and leave the rest untouched. So DefaultContainer seems like the best place to do it to me. On Sat, Sep 6, 2008 at 2:16 PM, Joe Walker public.gmane.org> wrote:

Re: DWR and Maven

Patrick Lightbody — 2008-09-06T14:06:22

Brendan, This looks great! I'd love to see poms for DWR available with each release. Please let me know if you want any help or need to bounce ideas of of me. The multi-module approach makes a lot of sense. Patrick On Sat, Sep 6, 2008 at 5:26 AM, Joe Walker public.gmane.org> wrote:

Re: Script Session ID, setActiveReverseAjax, Allow New Script Session Callback method

Joe Walker — 2008-09-06T13:24:54

I've created this: http://bugs.directwebremoting.org/bugs/browse/DWR-289 Currently we kill script sessions far too often, and we need to work out why. I think you're seeing just one symptom of this. It is my hope that we can be so successful at this that you don't ever need the callback, but we'll see how we get on. Joe. On Fri, Sep 5, 2008 at 4:32 AM, saturnism public.gmane.org> wrote:

Re: E14 / DWR-274: Allow overloaded method access

Joe Walker — 2008-09-06T13:12:57

So, with the change that I'm about to check-in we decide that the last parameter is a callback object if it contains one of the following as a function: - callback - callbackHandler - exceptionHandler - errorHandler - warningHandler Joe. On Wed, Sep 3, 2008 at 8:18 PM, Mike Wilson public.gmane.org> wrote:

Re: Safari vs. Firefox

Joe Walker — 2008-09-06T12:33:03

Thanks. http://bugs.directwebremoting.org/bugs/browse/DWR-288 Joe. On Wed, Sep 3, 2008 at 6:30 AM, saturnism public.gmane.org> wrote:

Re: response.encodeURL does not work as expected after WebContext.for wardToString()

Joe Walker — 2008-09-06T12:30:47

Yes - done. Thanks for the report. Joe.

Re: DWR and Maven

Joe Walker — 2008-09-06T12:26:37

One thought - before we can really properly have DWR builds being Spring only (for example) we would need to fix that core/impl/main/java/org/directwebremoting/defaults.properties Declares the implementations of all the required interfaces. I suspect things will work, but that you'll get quite a few errors/warnings on startup. Joe. On Tue, Sep 2, 2008 at 1:16 AM, Brendan Grainger public.gmane.org

Re: Spring change/patch

Joe Walker — 2008-09-06T12:16:01

I think that what I'm not keen on is that we're fixing the problem away from where the problem pops up. When you look at the DefaultContainer it doesn't make any sense what you are doing. Putting the fix into SpringContainer makes more sense because that's where the problem occurs. I'm not saying this is Spring fault or SpringContainer's fault, just that that's at the boundary, and so it seems like it makes sense to fix it there. Joe. On Mon, Sep 1, 2008 at 7:45 PM, Jose Noheda public.gmane.org> wrote:

Re: JettyContainerAbstraction and Jetty 6.1.12 rc1

Joe Walker — 2008-09-06T12:12:13

Greg/etc, From the DWR mailing list On Thu, Sep 4, 2008 at 11:20 PM, saturnism public.gmane.org> wrote: Did we do the wrong thing in checking for "jetty-"? Joe.

Re: Re: Build 110

Joe Walker — 2008-09-06T11:59:46

Very confused. The source to appendCall on both ScriptBuffer and ScriptSessions is very similar: appendScript(funcName); appendScript("("); for (int i = 0; i < params.length; i++) { if (i != 0) { appendScript(","); } appendData(params[i]); } appendScript(");"); I don't see how it can possibly miss out the (), and from my testing it seems like it's working just fine. () are part of ASCII, so I don't see that it can be an encoding problem. Any ideas? Joe. On Fri, Sep 5, 2008 at 7:27 PM, Thomas Manson public.gmane.org>wrote:

Re: Tomcat doesn't shut down after upgrading to 3.0 m1

Joe Walker — 2008-09-06T11:10:07

In the year 2050 when servlet 3 is release we should be able to... Joe.

Re: DWR & tomcat 6

Joe Walker — 2008-09-06T10:32:01

Eh? $ wget https://dwr.dev.java.net/files/documents/2427/109056/dwr-3.0.0.110.dev-src.zip--no-check-certificate [blah blah] $ unzip dwr-3.0.0.110.dev-src.zip [blah blah] $ find ui/gi/main/java/ -type f ui/gi/main/java//jsx3/GI.java ui/gi/main/java//jsx3/lang/Object.java ui/gi/main/java//jsx3/net/URIResolver.java ui/gi/main/java//jsx3/net/URIResolverConverter.java ui/gi/main/java//jsx3/xml/CdfDocument.java ui/gi/main/java//jsx3/xml/CdfDocumentConverter.java ui/gi/main/java//jsx3/xml/Node.java ui/gi/main/java//jsx3/xml/Record.java ui/gi/main/java//org/directwebremoting/gi/gi.js ui/gi/main/java//org/directwebremoting/gi/GiHandler.java Joe. On Fri, Sep 5, 2008 at 9:22 PM, Robert Dietrick public.gmane.org> wrote:

Re: class mapping E5 & E6: Offer a "lighter" bi-directional class-mapping scheme & Allow alternative class-mapping schem

Joe Walker — 2008-09-06T10:08:17

Thanks Mike. The change looks nice and simple, and even with added tests. Thanks. Joe. On Fri, Sep 5, 2008 at 9:53 PM, Mike Wilson public.gmane.org> wrote:

Death of TypeHintContext

Joe Walker — 2008-09-06T10:06:33

DWR used to have a rather confusing class called TypeHintContext. The idea was to provide a way to get generic type information even when type erasure was making it hard to get at. We also had 2 other classes TypeInfo and Property which did related jobs. And they were all distinct from InboundVariable, InboundContext and the Converters. I'm just deleting TypeHintContext and TypeInfo. I'm just checking in a simplification which came about from supporting ConstructorInjection. Here's what the classes are for now. Property: Describes some way to get or set a bit of data on an object. There are obvious Properties for a getter/setter pair (PropertyDescriptorProperty) and for a field (FieldProperty) and also for a constructor parameter (ConstructorProperty). There is a PlainProperty for when we have a collection of Properties of a bean, but need to augment it in some way and NestedProperty which is used when we're digging into nested generic type info. It would be nice if {< at >link PropertyDescriptor}, and the vari

Re: Re: DWR/Guice Integration Question

Joe Walker — 2008-09-06T09:17:17

We've not branched for 3.0 yet, and I expect we will only just before the release. I'm not a big fan of going off to work on something else before the release is done. Joe. On Fri, Sep 5, 2008 at 11:29 PM, Tim Peierls public.gmane.org> wrote:

Re: RE: Re[dwr-user] verse Ajax poll failed

Joe Walker — 2008-09-06T09:12:54

Excellent - we've got something that we can fix now. Please could you raise a big report, so we don't lose it. http://bugs.directwebremoting.org/bugs/secure/CreateIssue!default.jspa Thanks. Joe. On Sat, Sep 6, 2008 at 12:11 AM, James Mortensen public.gmane.org>wrote:

RE: Re[dwr-user] verse Ajax poll failed

James Mortensen — 2008-09-05T23:11:13

Hi All, I found what I believe is the problem. Apparently, Reverse Ajax fails in cases where gzip compression is used, but only with certain versions of DWR. I use the following CompressionFilter in my web application:

CompressingFilter

com.planetj.servlet.filter.compression.CompressingFilter

CompressingFilter

pjl-comp-filter-1.6.6.jar With DWR 2.0.1, there are no issues with gzip compression; however, 2.0.5 fails, as does 3.0+. The next step I will try is to use another means of http compression. Meanwhile, it would be helpful if someone could provide some insight into why compression is failing for the newer versions of DWR. Please let me know if anyone has any more questions. James Mortensen James Mortensen wrote:

Re: Re: DWR/Guice Integration Question

Tim Peierls — 2008-09-05T22:29:32

Checked in on trunk. Does that mean it won't get into 3.0 final? --tim On Fri, Sep 5, 2008 at 3:14 PM, Patrick Lightbody public.gmane.org>wrote:

RE: class mapping E5 & E6: Offer a "lighter" bi-directional class-mapping scheme & Allow alternative class-mapping scheme

Mike Wilson — 2008-09-05T20:53:24

E5 is now implemented (we will not do E6) and I have resolved http://bugs.directwebremoting.org/bugs/browse/DWR-269: I have now checked in the following enhancement: It is now possible to use class mapping without the generated JavaScript classes. When a generated class is not included in the current page, the class-mapping system will use the "light" mapping mode for this class. This means that mapped objects sent from server to client will be created as ordinary JavaScript Objects annotated with a $dwrClassName property to identify the object's class. When sending objects from client to server the same property is used to attach type information on objects. This property on instances takes precedence over class information from the JavaScript class, so the instance property may be used to override class identity for objects even when they belong to a loaded JavaScript class. Naturally, this lighter class-mapping scheme does not support the use of instanceof to det