Re: Included Actions
Gregor says: Output of an "included action" in a JSP. I think I understand this. I do something like it for my "Struts Inside of Struts" project. In doing so I am pushing the FormBean concept to its logical (and absurd) conclusion. My FormBean contains only a StringBuffer attribute. The View JSP for this is extirely empty (almost) except for the expression of the StringBuffer. The function of my (outer) Action is to just invoke an HTTP session on an inner Action (a Struts Action in some other application on some other server), then to capture the HTML output of that HTTP session, to modify some of its critical tag attributes and, finally, to stuff that HTML character stream into the StringBuffer of the ActionForm. I really thought that this would have horrible display latency before I tried it. Not really. Its almost bearable. But don't try it 3 deep. Dan Connelly (who is not Dan Connolly) PS. By the way, Struts works nicely for this absurd usage. I will still claim I honored MVC, although my View is vacuous and the Model knows damn well how it is to be rendered (almost). One disappointment in all this is java.net.HTTPSession. It doesn't deal cookies, so this implementation has problems for Session Tracking under Struts. I am not the only person displeased by java.net.HTTPSession. There are alternative classes that produce can transfer cookie values (from Tigris, for example). I just haven't tried any of them yet. N.B., however, with respect to Struts, that if my action form's (viewable) text is really huge, StringBuffer is not goint to work out. Struts would have to be re-architected into producer-consumer threads for the Model logic and the View (rendering) logic, respectively. This is an interesting thought. - Original Message - From: "Gregor Rayman" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Tuesday, July 31, 2001 1:03 PM Subject: Included Actions > Hi all,> > is it somehow possible to include the output of an action in one JSP?> > Now, when I try to include a *.do page, the server complains, it cannot> forward to the action's jsp (of course, since I don't want to forward,> I want to inlcude).> > The only "solution" I've found, is to len the Action's perform method> return null and include the action (*.do) as well as the jsp. (Which> I really do not like, since then the including page plays the role of> the controller)> > Where should I start if I wanted to change the Action servlet, to include> instead of forward?> > (I include actions' output in my templates. Where one part is the content> and another part is the in an action generated menu)> > --> gR> >
Re: OFFTOPIC: Solaris and Tomcat problems!
There is no compiler class (jar) distributed with Tomcat, out of the box. Apparently there are none in open source. Apache will only distribute open source. Of course, Sun (and others) distribute "tools.jar" (or equiv) gratis. Modify your startup.sh (or is it setenv.sh ??) to include the tools.jar on the Tomcat classpath. Then all is well. - Original Message - From: "Mikkel Bruun" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Monday, July 02, 2001 3:09 PM Subject: OFFTOPIC: Solaris and Tomcat problems! > This has nothing to do with struts, but I know this list has a couple of > guru's and demigods lurking... > > I have just installed solaris 8 on an intel machine. > I installed tomcat 3.2.2 and it runs the servlets examples without any > problems...BUT > There seems to be a problem running the jsp files... > Whenever i request a jsp file it just freezes, and i get no respond... > Looking into the matter I saw that the jsp was succesfully compiled to a > .java file...BUT > The jsp compiler seems to be endlessly looping in the translation, because > it keeps writing the same line over and over again (until im out of disk > space)... > I looked into the tomcat mailling archives and saw that somebod had > experienced this problem before, but I wasnt able to find a solution... > > Any suggestions > > thanks in advance... > > Mikkel > > > >
Re: Implementing 2 new scopes using Struts
You would think that the ALL WEB APPs scope (or hierarchicial scopes) would be easy to do. It is not. I had thought that by having one app run another app, the controlling app could "peek". Doesn't work that way either. (Too bad.) One can fetch the sessionid's of other HttpSessions from the the HttpSessionContext. However, this does not grant access to objects in the session scope of the other sessions (as I had once thought). Prohibited by (7.3) of the Servlet Specification. Craig McC. recently posted that the standard mechanisms are EJBs and (persistent) database. I am intrigued with the idea of a hierachically scope managed by an EJB. (The database thing is always possible but seems totally inelegant for this purpose.) Has anyone tried to coordianate objects between app (in Struts) using EJBs? - Original Message - From: Jonathan Asbell To: [EMAIL PROTECTED] Sent: Monday, June 25, 2001 8:23 AM Subject: Implementing 2 new scopes using Struts I will be implementing new scopes using Struts for our unique needs, and they need to be application server independent. One scope will work across (between) web-applications: Ex. user logs on, and an object representing the users values and state is placed into a scope where ALL WEB APPLICATIONS can see it. Another scope is restricted to specified directories inside of an application: Ex. data saved would only be visible when inside a specific sub-directory. I am currently examining the 2.3 servlet spec. Does anyone have any advice as to what to watch out for in implementing this considering I want to be able to be compatible with servlet containers in general and leverage container management and web.xml configuration. Could anyone refer me to some specific sections in the servlet spec that you feel I should pay specific attention to while I am doing this. Thank you Jonathan
External Workflow Engine for Struts ??
Title: new window browser for each request Has anyone tried connecting a Struts application to a back-end workflow engine? For instance, WfTK is a workflow engine with a Source Forge distribution. It demos with a Perl front-end but many web-ends are possible. A Struts front end would be nice, probably nicer than most. (But current work is on PHP. Go figure.) WfTK is written in C/Tcl./Tk. There are some pure Java engines listed on Source Forge, but no distros out yet for them. Doesn't matter. I need only a loose coupling at this point. WfTK has the essential features of Workflow. Seems to work OK. An exercise like this would help to clarify Workflow requirements within Struts (Release 1.1), I think.
Re: Cookies and Clusters (Cookies)
I will reply to Malcolm Davis in two parts: Cookies, then (later) Clusters: Cookies: Yes, my presentation of the DHTML and JS stuff was a bit confusing. And unnecessary. Let me boil it down. What I was musing about was a hack (??) I had just seen in a non-Struts MVC app whereby the current page gets updated, in situ, through the cookie(s) that head-up an SC_NO_CONTENT page just sent from the server. The JavaScript in the current page reads the cookie(s) and updates its own document model. No screen flash. Very smooth. A nice MVC screen effect. Think of the Color Chooser, the canonical MVC application. Smooth updates. Is it a hack? Would we consider such a technique for Struts? Of course, having the client code read cookies might be considered weird. But why not. This is the web. We survived the frameset mish-mash. Many B2B apps have captive clients on dedicated PCs. If the app wants cookies, he should get cookies, right? These are session cookies, not persistent cookies. Privacy is not a big concern. If the browser is loading up on cookies from other apps and maxes out, he gets a nice warning to quit screwing around. I did not write this MVC-ish app that I am talking about. I am merely testing it, aggresively, looking for hacks. I don't have access to the servlet code but I can see the cookies arriving at the browser. No JSESSIONID there. Just 4k of encoded state. You ask, does this app use HttpSession? Apparently not. It bakes its own cookies, complex cookies. Bad. Bad. Bad. But I like it. It contributes to MVC. Why not do it, do it, do it?? - Original Message - From: malcolm davis To: [EMAIL PROTECTED] Sent: Monday, April 23, 2001 1:47 AM Subject: RE: Cookies and Clusters Dan, > Cookies get a lot of bad press. But, cookies (or complex URL encoding) are the fall-back > when session beans are not going to be used. Many JSP developers will shun session > beans because the app might get deployed on clustered servers (reverse proxied). > Each POST might then be dispatched to a different JVM. Session beans were not shared > correctly among clustered JVMs until Servlet 2.3. (Correct me if I am wrong.) > Nobody is there yet in the commercial offerings. iPlanet does, and I'm sure other vendors do also. I'm a little lost here about the DHTML & JS stuff, but you might have a misunderstanding about cookies. You can turn cookies off in the browser, but the Appserver should implement them via the jsessionid. Cookies are fine for a cluster environment. The servlets must be marked as sticky, meaning each subsequent request will return to the correct JVM. Of course, this might depend on the Appserver being used, and how your cluster is configured.At my present job, we have 3 machines running 10 JVM’s, a total of 30 JVM’sin the cluster. If your not using session beans, are you using HttpSession? - malcolm
Cookies and Clusters
I have recently ran across this interesting, web app. It is a quite MVC-ish with a Controller servlet and Model beans, but too much logic in the JSPs to be strictly Views (no taglibs). This app features a page with a hierarchical, walking menu, somewhat like a "DHTML Labs" HierMenu. However, each step through this menu hiearchy requires (??) an interaction with the web server. [I leave open the possibility that the "walking" requests to the server might be overkill in this case.] Fetching each successive level of the walking menu from the server is handled in a complex JavaScript in this app. This JS sends a request, a POST, to the controlller, giving the level in the hierarchy and the OPTION picked in the last traversed SELECT list. The browser gets back a Status Code=204, NO CONTENT. The screen is not updated, except as noted below. The document model (including JavaScript) is not replaced. The JS now examines the Cookies where it finds the encoded options for the next select list. It then fills in that select list with the newly received options. Very smooth screen changes. The MVC ideal. (So OK, the graphics aren't so great. It not really a HierMenu. It doesn't really "walk" The "menu" is fully extended to begin with. The "unwalked" lists on the right are empty. But its the same idea. And it could scale up nicely beyond what DHTML could do with wimpey static hierarchies.) Is this legal? I think so. It does seem to be what 204 (NO CONTENT) is intended to be used for. So that's kosher. Cookies get a lot of bad press. But, cookies (or complex URL encoding) are the fall-back when session beans are not going to be used. Many JSP developers will shun session beans because the app might get deployed on clustered servers (reverse proxied). Each POST might then be dispatched to a different JVM. Session beans were not shared correctly among clustered JVMs until Servlet 2.3. (Correct me if I am wrong.) Nobody is there yet in the commercial offerings. Is there any one of us working on a taglib for Sruts with cookie hacks like this? Or, am I going to get a well-deserved flaming for even suggesting the possiblity. Dan Connelly (Not the famous guy at W3C).
Re: A Server-side Color Chooser, the Struts v. Barracuda Bake-off
The interesting thing about building a Struts Color Chooser app is that I am being forced to program the multiple update dependencies. Each of my Color Views (five of them at last count) will be an Observer of all of the others, on the server-side. This is not a problem in such a trivial case, but I am hoping to find a more general way of dealing with this. (In another application I may want to modify an collections on the fly this way. Seems like this should work.) For the Color Chooser I will render each Color View through a JSP into a separate mini-frame on the browser window. I cannot resist referring to these dependent mini-frames as "framelets" (forgive me). Each framelet will hold a minimal . When any of these formlets (ugh!) gets submitted, I will need to update all the other framelets. Since the HTTP allows only one response I will probably select a Master Color View for the single response to any form submission. (But in other apps, the update pattern may be much more complex.) The standard references tell me that, in order to provide an update to each of the other framelets, I will have to supply an "onLoad" JavaScript as a knowledgeable loader in the rendering of the Master Color View. That JavaScript program will explicitly load the subordinate Color Views. (How to generalize??) I am beginning to understand the relevance of CRM's remark, in another context, that the JSP is a way to get a program to write a program. (BTW. I want to use Microsoft's "HyperText Component" slider widget, available at http://msdn.microsoft.com/workshop/c-frame.htm?/workshop/author/default.asp , in my Struts Color Chooser. All the classic choosers use a slider. It does not bother me that only IE 5.5 supports HTCs, also known as DHTML Behaviours. Of course, there are potentially a huge number of these things. Creating a new Struts tag for each one is not an option. More problems of generalization.)
A Server-side Color Chooser, the Struts v. Barracuda Bake-off
The canonical MVC application in the GUI world is the Color Chooser. You will see the Color Chooser in standard texts on both Motif and Swing, but with very different implementations of the "Observer" pattern. The Patterns people claim that the Observer is at the heart of MVC loose coupling. I agree, although that pattern is an arch wide enough to drive several implementation buses through. Your Java 2 jdk distribution probably contains a Color Chooser application under demo/jfc/SwingSet2. However, that app is quite complex, more demo than tutorial. The O'Reilly book by Eckstein, Lay and Wood has a more approachable example. Source code is available in either case. I propose that we have a "contest" to implement a Struts Color Chooser, with a goal being to provide as much as possible of the liveliness that one sees in the Java/Swing app and to support a server-side Color Database. This is an exercise left to you, the reader. It is not to be part of the Struts codebase. You can find numerous examples of simple Javascript Color Choosers which operate totally on the client-side using very static color tables. These "apps" do not have a sever-side except for the HTML pages. To show off Struts, I want to add a server-side. The Struts Color Chooser must support a database of Color Names. However, that term database is loose. It need not be persistent, but it must be accessible to all concurrent Color Chooser web apps. So There will be a Color Model, with various Color Views (e.g., RGB Sliders, RGB Numerics, HSV Numerics and Color Names). The various Views must update each other in timely fashion (somehow). There should be a "ComboBox" allowing the user to scroll a (dynamic) database of Color Names. There must be Color Actions for adding, editing and deleting Color Names on the shared database (which need not be persistent across server-side reloads of the Struts Color Chooser). We can try this in Struts, Baracuda, Turbine and Hammock (about which I know nothing). You may also want to try it in your favorite scriptable template engine, but this would be unlikely to beat out any of the above four. BTW, on closer reading of the Barracuda "Overview" I see that Barracuda does not (now) claim any mechanism to map client-side events to server-side handlers. This is more limiting than I was expecting, given the rhetoric from Barracuda that would support a Swing-style Event/Listener model. (But not for the Barracuda View ??) I incorrectly imputed to Barracuda a desire for propagating client-side events when i posted comparing Struts to Barracude. This is not true. Like Struts, the only client-side event that Barracude builds in is an HTTP Request generated by a Form/Submit.) On further consideration of the issue of event notification, I am less pessimistic about the "mud". I do think that both Struts and Barracuda, in their current distributions, could host applications which implement client-side events that are notifiers to server-side actions. This can be done with slightly less fuss than I previously imagined because the set of client-side events to be accommodated is already identified and limited. There is a clear path to the application code to send notifications for these events. But this code may be repetitious and inelegant. That is the crux of the matter. In Struts, the onXXX JavaScript handlers in the html tags be coded as event notifiers that interact with server actions. I would guess (just guessing) that Baracuda wants to use Java applets for this. The SOAP/Biztalk books now filling your local bookshop have examples of client-side events being sent server-side using JScript/SOAP/XML. The examples there are even more repetitious and inelegant, being totally in static HTML. But this is how Microsoft captures a market and then sells its Visual tools. As far as my contest is concerned, almost anything is fair. I requre that at least one of the two "standard" browsers be supported in one of its standard realeases. You can require any browser setting you need. But, no plug-in or ActiveX downloads are permitted. Also, you must build the application on top of some standard distribution of your chosen framework. You may use any extensions that this framework permits, but you must not alter the framework in the cases where it is open source (Struts, Turbine, Barracuda). The contest will be "judged" on the basis of both functionality and elegance. The time limit should be very generous so that we can share ideas. Say one year. Dan Connelly
Re: How does Struts compare to Barracuda?
Okay, that's a good question.
Here is my instant analysis of Barracuda v. Struts.
("Instant" and therefore probably worthless. BTW, I haven't used
Barracuda. And my experience with Struts is confined to toy
applications.)
Barracuda has apparently bitten of the Event/Listener model
as its top-priority task. Struts has deferred this design decision
until a later release. And there certainly are a ton of more mundane
concepts to work through.
The Event/Listener Model appears on the Struts 1.1 to-do
list. My guess is that it will actually not make 1.1, that it will be
delayed until 2.0 because of the massive refactoring required. Just
guessing.
The Barracuda docs say that the Event/Listener Model is the
crux of the matter for leveraging MVC on the Web. Yes. Yes.
Yes. Very True, IMHO. You screw it up, you lose.
But also Very Difficult given the client-server division of
labor in HTTP and the on-going Browser War. (Just one war,
obviously.)
It does seem wise to me that Struts should not make a hasty
decision on its Event/Listener Model, but not past the point where it would
break everything to put it in.
When I worked on MVC apps in C++/Motif, we addressed this
issue less generally, as Update Dependencies. We
compiled the UD declarations into Mapping objects using a
preprocessor. This seemed adequate for building lively apps and gave
quite good performance, but Motif widgets had a much finer granularity of
interaction in the application logic than do FormBeans.
Of course, the Swing-style Event/Listener Model is much
more run-time oriented in its approach than is a UD
preprocessor. Java favors late binding. C++ favors early
binding. It is a continuum. Each Framework needs to pick its
spot on this continuum regardless of its implementation language.
My guess, and it is just a guess, is that Struts will end up
with some flavor of static UDs in the struts-config.xml, and that will be
about the best you can do for HTTP Web apps.
Barracuda will have to stick its hands very deeply into the
DHTML mud to map the client-side events out to server-side objects. This
is a mess. Once Barracuda is stuck there, then it falls victim to the
"real" barracudas (the proprietary vendor interests) .
If you need something very dynamic, very programmatic in
your application's Event/Listener Model, then you need a different protocol
than HTTP over the wire. Period. IMHO. Surely X-style
"GUI Servers" will become part of the Web in due time. Same goes for
Critrix-style "Terminal Servers". (Were it not for firewalls, this
would already be true.) And if that doesn't happen, then look who's
betting on SOAP (nee XML-RPC) to put events (as RPCs) on the
wire. Clever marketing can easily morph SOAP into an
Event/Listener Model on its own, eventually escaping HTTP
altogether.
Struts is wise to avoid this fray, for now.
But, I'll bet that others in the Struts community feel strongly just the other
way, that early engagement would be best. Maybe Barracuda
has already done succeeded in mapping client-side events. ??
Dan Connelly
- Original Message -
From: "Johan Compagner" <[EMAIL PROTECTED]>
To: "Struts" <[EMAIL PROTECTED]>
Sent: Tuesday, February 20, 2001 7:28 AM
Subject: How does Struts compare to
Barracuda?
> Here is the link:>
http://barracuda.enhydra.org/Barracuda/>
> Can we learn something from it?> Is it better or worse?>
> Johan> >
Re: Strange Error with
This one came up earlier this month. Nothing to do with per se. A large number of tags on a page can cause this. Check the archive at http://marc.theaimsgroup.com/ Search struts-user for the topic "Illegal target of branch or jump" The last post on that topic says that the cure is to give yourself more heap in the JVM with -mx and -ms (I wonder if its really more stack space per thread that on needs for a large number of tags on a JSP (??). I haven't looked into it, but stack(s) are usually more fragile than heap in multi-threaded system.) - Original Message - From: "Neal Kaiser" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Monday, February 19, 2001 9:45 AM Subject: Strange Error with > I am using Orion 1.4.7. I don't know if this is an Orion bug, or possibly> Struts? I have a jsp which uses the include directive to read in a list of> countries:> > > The jsp looks like this:> > > <%@ include file="../util/countries.jsp" %>> > > > And the countries.jsp is something like this:> ..> Tuvalu> U.S. Virgin Islands > Uganda> > If I have 223 lines in my countries.jsp file, it works, no problem. However,> if I go over 223 lines (no matter what line it is) I get this error:> > Error parsing JSP page /galacy/register/layout.jsp> > Error creating jsp-page instance: java.lang.VerifyError: (class:> __jspPage28_register_regstep2_jsp, method: _jspService signature:> (Ljavax/servlet/http/HttpServletRequest;Ljavax/servlet/http/HttpServletRespo> nse;)V) Illegal target of jump or branch> > I just tried putting the select list in the .jsp rather than using an> include -- same error.> > I am leaning towards this being a container bug, not a Struts one. But I> just wanted to verify and make sure this wasn't know with Struts first...> > Thanks, Neal>
Re: More HTML element examples []
Robert Taylor wrote: > I've been through the examples included in the struts0.5 release.> Specifically I'm looking for examples of using the element with> dynamically generated . I have seen the element and read> its definition but still am unclear on how to use it.> I, too, found the documentation inadequate in this area. However, the question has apparently been brought up on this listserv before. There are examples in the listserv archive. The fundamental idea is the same as is used for the "User Database" in the Example example. Have a look at http://marc.theaimsgroup.com/ Its a searchable Mail ARChive (MARC) covering this listserv (and lots of other interesting ones too). Just search on "options" for struts-user. The thread "Need some help please" explores the use of the tag. (The "Subject" text here could use some help! But, Hey, the same goes for "More HTML element examples".) Continuing... The "Need some ... " thread is slightly misleading. It has examples that use the tag. I believe that that tag has now been replaced by , a switch which undoubtedly contributes to the confusion on this topic. Also, have a look at the "Options Tag" thread (Yes, good Subject text here), but some indications in that thread that there are existing bugs in indexed array accesses. (??) Ted H's reply on the "Need some.." thread indicates that a Struts "collection" should be synthesized from a "database." Maybe, maybe not. I can see that a "static" collection, which is often we really want, would be better derived from a Properties file. However, in either case, the preferred method in Struts for instantiating such a collection (for use with ) is to have an made-to-purpose application Servlet which you code up to do the "heavy lifting". The collection get instantiated in that Servlet's init method, when the application is loaded into the server container. Bottom line, you roll your own. See "DatabaseServlet.java" in the Example app, for an example. If you are reading Hans Bergten's book on JSPs, which discusses a Struts-like architecture, you might naively assume that the Struts built-in ActionServlet would contain some packaged support for cons-ing up your collections for you. Not in Rev 1.0 by what I see. I don't see it on the Rev 1.1 ToDo list either, although "Better Database Support" could be stretched in that direction. Maybe, in that case, we should be looking at better J2EE Support instead. (IMHO Application Scope Bean collections are not at all the same thing as Persistent EJB collections, which is where "databases" really fit in. Just my opinion.) OK. OK. I'm rambling. isn't the only area where coding examples that use (ordered) collections under Struts would be helpful. Has anyone got an example of Previous and Next submit buttons to scan over a collection for FormBeans? That should be simple, and nice. Better yet, how about an example of an IE5.5-style Floating Popup containing a scrolling list of FormBean ids that would dynamically update of the parent page when clicked ...? OK. I'm still rambling. And futuring. (Except that it should be possible to do a very sweet, very MVC Floating Popup now, with Struts 1.0, if you limit yourself to IE5.5) Dan Connelly South Burlington, Vermont
Re: which development tool to use? (XML Editor)
Tag editor for struts: Try Xeena from IBM Alphaworks. Its free; its Java; its fragile. Create/edit like a Bean?? Dunno. But Xeena works great for struts-config.xml, web.xml and deployment XMLs in general. - Original Message - From: "Steven D. Wilkinson" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Thursday, February 08, 2001 6:43 PM Subject: Re: which development tool to use? > Malcolm Davis wrote:> > > > 3. Do you have a preference on Tag editors/creators?> > I'm looking for something that creates/edit tags like I would a bean.> > Does JBuilder Enterprise do this and I'm just missing something?> > > I only have JBuilder4 Professional, so I don't know what else the Enterprise> brings to the table other than EJB wizards. > > I too am interested in something that creates/edits tags like a bean. I haven't> found one yet. If you hear of one please let us know on the mail list.> > Thanks,> Steve
Blank Newbie (on Tomcat)
Here is a solution, not a problem ! I recently posted to the struts-dev list that the new application, struts-blank, should include a blank form to be submitted, that I had trouble when I tried to add even the simplest form (although the un-edited application worked OK). I suspected that I was making some Newbie mistake putting the form together. I was. Should you also have trouble with this, the Newbie problem (my problem, now solved) is that you must re-deploy the application if you make a change to the struts-config.xml. (Is this obvious? It wasn't to me. In fact, the splash screen on struts-blank implies otherwise.) The splash screen on struts-blank tells you that you can just edit, seemingly everything, seemingly after deployment, while Tomcat is running.. Yes. You can do this for index.jsp. It works. No problem. So you press on. You can also write the code and add your BlankForm class and your BlankAction class in situ to /WEB-INF/classes while Tomcat is running. No problem. However, when you edit struts-config.xml (as you must do to incorporate the BlankForm and the BlankAction into the application), then you must re-deploy. (Well, its slightly easier on Tomcat.) The struts-config.xml is ingested on deployment, not whenever index.jsp runs (as a Newbie might suppose). However, on Tomcat you can still edit struts-config.xml in situ, instead of remaking the WAR, which you could also do. But, in either case, you must shutdown and (re) startup Tomcat to get have your changes to struts-config.xml made effective. (J2EE platforms allow dynamic deployment.)
Struts/Motif
Struts for Motif. Okay, wash my mouth out with SOAP! But, look, lots of neat, old X/Motif applications were written using the MVC architecture. What good did it do X, Motif, ... or MVC? And, don't try blaming it all on the X Server being such a heavy "client". On the other hand, don't try to claim that the Web-way is the only way. [We could also consider Struts for Swing, but that clouds the argument I am about to make.] A Motif MVC app I recently worked on had MVC features similar to Struts, like an action map "digester". Well, the Motif app digester was written in awk and it was a preprocessor. The mappings got compiled into the code. But it was similar, a nice high level view of the internal data architecture. Unfortunately, I also see some disturbing similarities between the weaknesses of MVC Motif (as commonly programmed) and Struts, such as 1. A difficult and impoverished widget set. For instance, do we have a high-level Tree widget to plug into Struts? Having third-party Tree widgets and scripting hacks for Trees didn't make Motif competitive with the Win-mill. ( ?? "COMpetitive" ??) 2. Ad hoc Database-to-Model mappings. For instance, if a number of browser sessions, perhaps hierarchically related browser sessions, are looking at the same Persistent EJBs, how do these sessions coordinate their separate (hierarchical) Models? Their database updates? Their screen Updates? Is this to be left to the realm of "Business Logic"? It's system logic, for sure.It invites trouble to let the bit jockeys be responsible for getting this right, for buggering the MVC. The Motif MVC app I worked on used the windowing hieirachy in X/Motif to build a Model hierarchy of database views. This was very ad hoc, very quirky and quite buggy, but nice. Or so I thought. The nuances escaped many in the user community. And, eventually the programmers also lost track of the ad hoc programming needed to sustain the Model hierarchy. So it goes. But, jeez, it did make for a very lively GUI when it all went well, with model changes rippling from window to window. Made Win-doze apps look absolutely static in comparison. Unfortunately, I don't see much (yet) in Struts to make a more dynamic display out of (hierarchical) web browser sessions. Or, is this just a feature in search of an audience?? Dan Connelly
Ted's FAQ "What Web sites are already Powered by Struts?"
Brewtrade - http://java-genius.com/ TeaTimeJ - http://www.webappcabaret.com/teatimej/ What! Only two public sites powered by Struts? And, java-genius.com always gives me a DNS error anyway. So there's really only one site in this list that I would dare suggest to people who ask me about Struts in the real world. There must be more public sites than this. There's http://jakarta.apache.org/struts at least (although it doesn't do anything showy). Please help me explain to MVC-clueless colleagues and bosses why I am spending so much time getting into Jakarta Struts (as opposed to the superficially similar, released, commercial, freely downloadable JSP taglibs, like Allaire's JRun Tag Library or WDIG's TeaServlets). Please post up some additions to Ted's list. Also, how about adding more "Struts competitors" to my list of two? (And, while you're at it, do the competitive analysis.) Dan Connelly (who is not Dan Connolly, W3C)
WEB-INF/lib problems
I am new here, and new J2EE deployment issues. Forgive me if these observations are naive. I gather from recent complaints on this list that deployment of struts WARs into J2EE environments (with the WARs packaged into EAR files) is not working as expected. The struts.jar isn't getting cpath'ed from the demo apps WEB-INF/lib at deployment time, or gets put in some sub-directory of WEB-INF/lib, where it is not found. But, the struts demos do work OK on stand-alone Tomcat (if you use Xerces and avoid JAXP). One workaround people try is to put struts.jar into the j2ee server's CLASSPATH, but this is a bad idea because j2ee is an "embedded environment", ie, a total platform into which apps are installed dynamically. Anyway, juking the CLASSPATH just causes other problems (in "example" for example). (Rough summary of various threads.) I am seeing this too. I my case the problems are on the freebee j2eesdk1.2.1 server that I downloaded from Sun. This may be related to the WL and iPlanet problems that others have reported. (Except that you paid good money for that software and you are getting excellent support from your vendor, right? No annoyance issues, right?) The Sun Bug Report I have posted below seems relevant to these problems. I kept thinking that Sun must know about (and would quickly fix ??) such an obvious problem in their own SDK implementation. So I kept searching their site for such a fix. Apparently not. Perhaps deadlines are tight at Sun. Anyway, . This particular Bug Report is quite old and is listed as closed. But, skeptic that I am, I am not convinced that it was fixed anywhere but on Tomcat. And it could be that various commercial implementations of j2ee (inlcuding the Sun "reference") employed workarounds for this bug, which ended up as regressions. Also, what does "javaxservlet:api 3.0beta" mean in this context, given that the latest Servlet API I know of is Release 2.3. Dan Connelly (who is not Dan Connolly, W3C, but likes his ideas anyway) Bug Id 4282385 Votes 0 Synopsis ServletLoader cannot load jars from WEB-INF/lib directory. Category javaxservlet:api Reported Against 1.1beta Release Fixed 3.0beta State Closed, fixed Related Bugs Submit Date Oct 18, 1999 Description jcheck tests are failing with both tomcat and j2ee. You did not see it fail when running under tomcat because moo.jar is in the classpath when you started tomcat web-server. So, the tomcat server found all the classes. But on J2EE, (where there is no classpath related issues) you saw all of them fail because the ServletLoader is unable to load up classes. Here is the problem with the ServletLoader as it exists today: It can load classes from the WEB-INF/classes and WEB-INF/lib directories but it *cannot* load *jars* from the WEB-INF/lib directory. I am going to file a bug against the tomcat implementation. Meanwhile, I have a suggestion to whoever is responsible for jcheck. Please *do not* start the tomcat server with any more classes than whats required. Can you or anyone else in your team remove moo.jar from the classpath? Workaround None. Evaluation in an embedded environment (ie j2ee) and for war docbases tomcat was not able to access WEB-INF/lib/*jar classes due to the fact that the war protocol handler registration moved from server package to the optional shell package and thusly was not made available to the hosting environment. the fix involved modifying the protocol handler to manage both war and jar uri's alike and update the ServletLoader to generate "jar" uri references (vs war uri references) so that for environments which have a specified war protocol handler (eg tomcat/jdk1.1.x) will get work as well as environments which which run on jdk1.2.x where JarUrlConnection is natively available. cool fix. x@x 1999-10-28
