I am writing my own custom persistence instance store for WF4, based on the XmlWorkflowInstanceStore found in the .NET 4 WF and WCF samples. This sample is quite simplistic and the xml is produces is quite verbose. I have issues with how some of the objects are serialized.
I have tried using Red Gate Reflector to understand the Sql implementation used, but it is quite complex and difficult to learn from. The MS documentation for this is rather limited - often giving one sentence descriptions for complex methods.
Please could you point me at other examples of WF4 persistence (or proper documentation) around on the web that are not copy and paste versions of XmlWorkflowInstanceStore? Maybe someone else on StackOverflow has written their own?
You are completely correct that the docs are very much lacking here and the sample is of very limited use. I have started work on a custom instance store using the entity framework but, much like you discovered, found it slow going and am nowhere near anything I could use myself, let alone release onto CodePlex.
I am not aware of any blog posts or other information that help solve this.
You've probably seen this already, but I've found the code quite easy to understand: http://xhinker.com/post/WF4Xml-persistence-store.aspx
Ron Jacobs wrote an in memory persistence store for WF unit testing. Check out http://wf.codeplex.com/releases/view/73842
Related
I'm using ScalaJS with Play. Many of the models I'd like to use on both JS and JVM platforms involve dates and times. Given the lack of a cross-platform date/time library, how are people approaching this?
Things I know about:
scalajs-java-time project (https://github.com/scala-js/scala-js-java-time) to port JDK8's java.time api to Scala.js. Unfortunately, it's far from complete and judging by the commit logs, seems to have stalled.
https://github.com/mdedetrich/soda-time is a port of JodaTime to Scala/Scala.js. But it's not ready for production use.
An old post at https://groups.google.com/forum/#!topic/scala-js/6JoJ7x-VxLA suggests storing milliseconds in shared code and then doing implicit conversions on each platform to either js.Date or JodaTime. But we really need a common interface, which this doesn't give.
Li Haoyi's excellent "Hands-on Scala.js" has a simple cross-platform library (http://www.lihaoyi.com/hands-on-scala-js/#ASimpleCross-BuiltLibrary) that could, in theory, be extended to come up with an API in /shared that delegates to JodaTime on the jvm and Momento on js -- but that sounds like a lot of work.
(added later) https://github.com/soc/scala-java-time is based on an implementation of java-time that was contributed to OpenJDK. The README claims that most stuff is working. Right now, this looks like the most promising approach for my needs.
Any advice from those who have gone before me? Right now the fourth options seems like my best bet (with the API limited to stuff I actually use). I'm hoping for something better.
I was in the same boat as you, and the best solution I came up with was cquiroz's scala-java-time library. From reading the comments to your question above, it appears you landed at the same place eventually!
I came here from a google search, and given how much better this solution is than the alternatives you mentions above, let's consider marking this question as resolved for future visitors.
Well, I've been following the NerdDinner tutorial online and I've completed it to a T.
Now I've implemented a project of my own using pretty much the same architecture.
I notice my process is:
Create database.
Create DBML.
Create |TableName|Repository.cs class with data access methods.
Create partial |TableName| class to handle validation.
Rinse and repeat for all tables.
By the time I'm done I'm exhausted and I haven't really done anything except drag and copypaste code and change variable names.
What am I missing here. Where can I improve this boilerplate haze I'm in?
Edit: Sans creating my own T4 template, is there a tool that will help me?
See Visual Studio Templates.
I haven't used them, but I assume if you take the time to customize them, you'll be able to make Visual Studio generate a lot of what you already do instantly specifically using the T4 template language/interpreter thats embedded in Visual Studio.
I haven't used it much but perhaps you could create a codesmith template to handle this. http://www.codesmithtools.com/
I have seen it used to good effect with Nettiers and in some other scenarios.
Have a look at this for ideas. It uses a generic repository and the unit of work pattern.
http://elegantcode.com/2009/12/15/entity-framework-ef4-generic-repository-and-unit-of-work-prototype/
Also here.
Advantage of creating a generic repository vs. specific repository for each object?
I've implemented something based on these ideas. You may also want to look at using a service layer more interfaces etc than the nerd dinner example shows.
As good as the tutorial is, it isn't IMO really fit for use in an enterprise application as it is still quite tightly coupled.
Hope this helps.
I use my own tool for that: http://github.com/Necroskillz/NecroNetToolkit
It basically bypasses all the steps that u mentioned (except validation, because I have validation on my view models).
While its not directly MVC focused, I've enjoyed using NetTiers http://nettiers.com for this kind of foundation. We start with the Db tables and indexing and then generate the stored procedures and their related data and service layer code. After that we add our own custom stored procedures, creating a rich (and easily understood) foundation. Future changes are easily accommodated through this mechanism.
We also bolt the code generation into our Nant build procedures, providing a very quick way of updating all of the mundane 'plumbing' code, allowing us to concentrate on the interesting stuff.
We've just started working with MVC and have found the netTiers service layer works nicely with it - time will tell as our MVC experience grows. I hope that helps!
Create a framework!
I am a newbie of Zend Framework.
I downloaded the Zend Framework and then followed the official quickstart tutorial to build a very simple registration form. But after that, I find very hard to learn different elements of Zend Framework.
Many tutorials mention frontController and registerAutoload() in the bootstrap file. However, it seems that I can't see it again in the code in version 1.8 (both in the official quickstart tutorial and Getting Started with Zend Framework 1.8). Many tutorials contain the old version of code and it seems very different in v1.8.
I start with writing registration form with password confirmation. I read the official document and find the custom validator for password confirmation. It just puts the class there and doesn't mention how I can use it. I don't know where I should put this file and I can't find any hints in Google. The "class not found" error always makes me sick (I've tried addPrefixElement, set_include_path but they don't work).
User registration, email activation, login, access control are very common tasks. But I don't even find a piece of sample code in v1.8 that I can run in my machine. I am very frustrated about Zend Framework.
Does anyone give me some advices?
I have also just started learning the zend framework. I also find alot of old tutorials that make learning confusing and hard,
but there are some resources that I find help in learning the zend framework:
http://www.zendcasts.com/ - they have tutorials on ZF1.8 already, like Bootstrapping using Zend_Application, etc.
Some useful blogs that might help:
http://weierophinney.net/matthew/
http://www.thomasweidner.com/flatpress/index.php
http://smartycode.com/zf/
http://codeutopia.net/blog/
Forums to ask your questions:
http://www.nabble.com/Zend-Framework-Community-f16154.html - alot of ZF developers seem to be here to answer your questions :) much more than the official Zend Forums below ...
http://forums.zend.com/viewforum.php?f=69
As to class not found, I think you didn't autoload?
http://framework.zend.com/manual/en/zend.loader.autoloader-resource.html
Parts of Zend Framework have changed slightly in the 1.8 release but everything should be backwards compatible so all code example written using the 1.* release series should still work with little or no modification. The main areas that are different are
Zend Loader (related to register_autoload) which now has several improvements related to performance and some slight differences in its API (with a PHP warning if you use the old way). I wouldn't worry about this too much as a beginner - just follow the tutorial or continue to use include/require statements until you become annoyed with adding them all the time!
Zend Application which basically allows you to create your set-up and initialisation code with less fuss. Again, you don't have to use it and you can happily write a manual bootstrap class to get you used to how the framework is put together.
With regards to getting started I would highly recommend getting your head around the Zend Controller component, particularly the front controller part and this diagram.
For authentication look at Zend Auth, for e-mail see Zend Mail, for access control see Zend Acl and for forms and validation see Zend Form, Zend Validate and Zend Filter. The form and validation components will explain about paths and adding your own custom validators and form elements.
The other area of importance is Zend Db which allows you to persist your data to a database. There is also Zend View which represents the view layer in the MVC stack. In my opinion these are the key components of a basic ZF application. From there you can explore the other components as and when you need them.
Don't forget that the Zend Framework is specifically designed to be loosely coupled and it is very easy to use a different component from another project for a particular task. Sometimes it is necessary to write your own components too. If there is a component you would prefer to use over the Zend component then go ahead and do so.
Honestly, the documentation is very good although it is hard to navigate for beginners. Once you get your head around the core components everything will become a lot clearer.
Also try following this tutorial. It will help explain a lot of the basics.
I recommend these video tutorials pu up by a guy called Alex, he keeps them up to date and releases new ones all the time,they are all relevant to ZF 1.8 and 1.9.
He has even a detailed video tutorial covering ACL, login forms etc. (my first steps with Zend were easier thanks to him!)
http://alex-tech-adventures.com/development/zend-framework.html?start=15
- his site is a bit comfusing but worth the time in figuring it out :)
Cheers
Roman
Some tips:
Here is a nice diagram of the Dispatch process
http://nethands.de/download/zenddispatch_en.pdf
It is much easier to understand ZF if you are experienced in OOP / OOP Design Patterns
When I first stumbled across ZF I had little knowledge of OOP. ZF prompted me to learn OOP which was a very good thing.
Distinguish between the 'magic' of ZF and the actual ZF Architecture
Although I have a good understanding of ZF Classes I still get confused when ZF adds its 'magic' to the mix.
When I say 'magic' I am referring to things like
the default objects loaded and used if none specified, and
url-controllername-filepath-filename inflections / naming conventions.
config settings-to-class-instantiation mappings
ZF is a bit like a crack-dealer in the sense that the 'Getting Started Tutorials' get you hooked straight away. They perform so much magic that it makes you think
.oO(Wow - how simple was that! This framework is cool!)
Then as soon as you start to try to do your own thing - the magic starts to get really confusing.
So - my tactic is to remove as much ZF magic as possible. If you have the option to specify a class to use, or a router to use, then specify it. Don't let Zend do it for you.
Then when something weird is happening you can find the bugs more easily.
Also, if you specify the locations of classes/ paths as (opposed to letting ZF magically inflect paths and find scripts), then you do not have to worry about questions of 'where to put this? , where to put that?' - you just put things where you want them to be and explicitly point ZF to them.
As you get more familiar with ZF you might then want to let it take over and do some magic for you. Only then can you really understand why that magic is convenient.
As Tim Wardle said in another answer - favour require statements over zend loader until you really need it.
If you are a PHP design house that churns out 10 website projects a week, then ZF Magic can be really useful. If you are designing your first ZF app - then 'reducing complexity' is a more important design goal than 're-usability'.
Browse the code
It might seem obvious but one of the best ways to learn about the ZF is to look at the code.
Again the ZF magic can get in the way - I often want to look at a method's arguments to discover what i need to pass in only to discover that it expects an ambiguous 'options array' (not exactly helpful). After a while, though, you start to notice the conventions used in how an 'option key' marries up to a getter/setter method. So, keep looking at the code and familiarising yourself with the 'ZF way'.
I 'ope that 'elps.
The Front Range PHP Users Group website has some presentations on Zend Framework which may be of use.
I have so much sympathy with you Billy. I am a newbie and the confusion between the old way and the new way is almost just too much to handle. Also see here for someone who points out some of the main differences:
http://crossfunctional.wordpress.com/2009/05/
All documentation, particularly Zend's documentation is absolutely horrible. It all assumes that you "know" where to put the snippets they are referring to and how they work together. If I already knew those things, I wouldn't need the documentation. No full example applications with authentication etc. that you can download and modify. I'm seriously regretting putting any time into trying to learning this framework. All examples that I can find are apparently for older versions. I followed one for authentication on youtube that was designed for 1.8 to the letter. It fails to work in 1.9.x. Can't get it to work for the life of me. Completely frustrated with Zend Framework!
Zend Framework has shifted from being a relatively accessible and simple system to a more complex entity. It has gone through a number of rapid developments over the past couple of years which have left a lot of older documentation around.
We have projects that are stuck on Version 1.7.x because the differences between that and the latest version are too great which makes the dev and testing time too expensive.
I like Zend Framework because it has so much depth to it but it is definitely a two edged sword for beginners.
I'd certainly recommend making use of the expertise of other users and search out some simple framework examples. ZF is highly adaptable but you need to try and work from the simplest case possible for your needs.
I think Matthew Weier o'Phinney, (search for 'Phly, boy, phly') is one of the more approachable members of the dev team and has loads of examples and ideas on his website.
Good luck!
I've been trying to find a good form of RPC to standardize on, but so far I've ran into a ton of walls and was wondering what the stackoverflow communities view was.
My ideal RPC would provide the following:
Somewhat wide support in other languages, in that people shouldn't have to write a custom stack to use our server
Input validation
Ideally, some way to turn the above input validation into some sort of automated documentation to distribute
Clean and maintainable code
I am a fan of the catalyst framework and would prefer to stick to it, but if there is a clearly better alternative for RPC servers I'd be open to that as well.
So far I have looked at the following:
Catalyst::Controller::SOAP
Doesn't appear to support returning of complex data structures, only string('literals'). I could probably serialize data on top of that, but that seems very hacky. It also lets you return a pre-formed XML object, but I couldn't get that to work and it looks like you'd need to re-create a lot of SOAP data structure for it to work.
I do like the idea of WSDLs, but the complexity of the overall spec also makes me wonder how well support for communicating with other languages will be.
Custom POSTing XML based controller
We tried to roll our own by hand in a similar way to how we've seen two other projects do it recently where there is a dispatch url that you post XML to. This lets you have XSD validation/documentation, but required creating a lot more code than we want to maintain at this point.
Catalyst::Plugin::Server::XMLRPC
Gave a warning about using a deprecated method that will be removed in a future version of Catalyst.
No input validation or doc creation, but otherwise the best I've found
JSONRPC
Looks pretty similar to XMLRPC only the module is actually updated. I'll probably go with this next unless someone suggests something better
There also appears to be two different modules for catalyst that do JSONRPC
I realize that REST isn't pure RPC (only a subset), but...
I would recommend the Catalyst::Controller::REST and Catalyst::Action::REST modules. They're frequently updated and the documentation is fairly good. There is also a good (but rather dated) example of using Catalyst::Controller::Rest in the 2006 Catalyst Advent calendar titled Day 9 - Web Services with Catalyst::Action::REST.
FWIW, Catalyst::Controller::SOAP does support returning complex data. Take a look at the documentation http://search.cpan.org/~druoso/Catalyst-Controller-SOAP-1.23/lib/Catalyst/Controller/SOAP.pm, which will show you that you can use a WSDL to describe your service. Also, see http://daniel.ruoso.com/categoria/perl/soap-today.html.en for a more detailed step-by-step process.
I'm faced with writing a framework to simplify working with a large and complex object library (ArcObjects). What guidelines would you suggest for creating a framework of this kind? Are static methods preferred? How do you handle things like logging? How do you future proof your framework code from changes that a vendor might introduce?
I think of all of the various wrappers and helpers I've seen for NHibernate, log4net, and code I've read from projects like NLog and NetTopologySuite and I see so many good approaches, but honestly I'm at a loss where to start.
BTW - I'm working in C# 3.5 but it's more about recommended approach rather than language.
Brad Abrams' Framework Design Guidelines book is all about this. Might be worth a look.
Try to write code to be more flexible. For example, if you have a method that accepts an array as a parameter, would you be able to accept an IEnumerable or IList instead?
I think that you're consistent is more important than what conventions you go with. As far as future-proofing yourself, that's a matter of the code that you're making a framework for. It's a lot easier to build on a brick house than a sand one.
Writing code for framework is absolutely very different from writing application code.
I have always consulted (and have others consult) the Design Guidelines for Class Library Developers when writing framework level code.