I was able to get active record running right away. The instructions for getting started were great and in no time I had built a webservice that would let me create and read widgets in my existing db. It was awesome. When it came to updating though, things fell apart. I would edit the object on the client and send it back to the service but when the service saved it, it would just create a new one. I reasoned that this meant that I would need to re-query the db and assign the values sent up to the service from the client but my boss said that would be hacky and that the repository pattern would be better because could use pocos. Unfortunately that's the extent of the guidance I've gotten. So here are my questions.
Are the t4 templates only good for active record or will they build
up your simple repository for you
too. Eg, is there something that
will gen up my pocos too or are they
all 'roll your own'?
Has anyone seen a working example of a subsonic 3 tier
solution? I've read about them but
are there any samples floating
around?
The active record samples/ screencasts were really easy to follow because they started at the same point I was starting with. The simple repository ones seemed to focus more on migrations and other advanced features and being that this stuff is new to me, I don't know enough to connect the dots.
Ugh. There's nothing quite like having a deadline to learn something and have it running by the end of the week. Any advice would be welcome, even if it's rtfm with a link to the manual I should have read.
Thanks in advance
If you want to use a repository pattern you can either use the linq templates or use the simple repository which does not require any t4 templates.
With simple repository you create the pocos yourself. Subsonic can create or update the database scheme for you if you choose:
var repository=new SimpleRepository(SimpleRepositoryOptions.RunMigrations);
but If you ask me I would choose SimpleRepositoryOptions.None and update the database by myself.
Anyway, your initial problem with the ActiveRecord templates could be fixed pretty easy.
I suggest your ActiveRecord object is serialized on the client side and deserialized on the server.
The default constructor of an ActiveRecord object calls the Init function which looks like this:
void Init(){
TestMode=this._db.DataProvider.ConnectionString.Equals("test", StringComparison.InvariantCultureIgnoreCase);
_dirtyColumns=new List<IColumn>();
if(TestMode){
<#=tbl.ClassName#>.SetTestRepo();
_repo=_testRepo;
}else{
_repo = new SubSonicRepository<<#=tbl.ClassName#>>(_db);
}
tbl=_repo.GetTable();
SetIsNew(true);
OnCreated();
}
As you can see, the internal repository is created and SetIsNew(true) is executed.
The only thing you have to do is to call myitem.SetIsNew(false) after the object gets populated with the deserialized values. I suppose that is sufficient to tell subsonic to do an update query during save.
Related
Here is a common scenario: app is installed the first time and needs some initial data. You could bundle it in the app and have it load from a plist or something, or a CSV file. Or you could go get it from a remote store.
I want to get it from CloudKit. Yes, I know that CloudKit is not to be treated as a remote database but rather a hub. I am fine with that. Frankly I think this use case is one of the only holes in that strategy.
Imagine I have an object graph I need to get that has one class at the base and then 3 or 4 related classes. I want the new user to install the app and then get the latest version of this class. If I use CloudKit, I have to load each entity with a separate fetch and assemble the whole. It's ugly and not generic. Once I do that, I will go into change tracking mode. Listening for updates and syncing my local copy.
In some ways this is similar to the challenge that you have using Services on Android: suppose I have a service for the weather forecast. When I subscribe to it, I will not get the weather until tomorrow when it creates its next new forecast. To handle the deficiency of this, the Android Services SDK allows me to make 'sticky' services where I can get the last message that service produced upon subscribing.
I am thinking of doing something similar in a generic way: making it possible to hold a snapshot of some object graph, probably in JSON, with a version token, and then for initial loads, just being able to fetch those and turn them into CoreData object graphs locally.
Question is does this strategy make sense or should I hold my nose and write pyramid of doom code with nested queries? (Don't suggest using CoreData syncing as that has been deprecated.)
Your question is a bit old, so you probably already moved on from this, but I figured I'd suggest an option.
You could create a record type called Data in the Public database in your CloudKit container. Within Data, you could have a field named structure that is a String (or a CKAsset if you wanted to attach a JSON file).
Then on every app load, you query the public database and pull down the structure string that has your classes definitions and use it how you like. Since it's in the public database, all your users would have access to it. Good luck!
I am working on a small hobby project, where I would really like some input and advice.
This is my first "real" node project, and I hope it will teach me a lot about node.js development. I am a .net developer by day, and have been for about 15 years professionally. I have had periods of doing Java as well. I have created small node.js projects to be used as micro services.
But this project can no longer be classified as a micro service ;-)
The purpose of the project is to sample some sensor data, and do some reporting. An idea I got from playing around with a PLC at university. I do that by sampling from a PLC, and emitting the data using ZeroMQ. My node.js server then listens for this sensor data, and stores it in a MongoDB.
I expose that data in a REST api. The REST api also exposes resources like batches and other stuff like authentication etc. On top of that I have an AngularJS app, that creates defines the UI.
The one thing that really annoys me, is that I want to globally assign what batch is running. I have a collection of batches, and one of them is the running one. There are a two ways I see to do this, and both illustrate my novice status in the node.js world. All users should be able to see what batch is running, and I want to be able to easily tell from anywhere in the code as well.
1) Set a flag on the object in Mongo. This has a number of problems. The obvious one being performance. I receive sensor data 10 times a second, and I don't want to ask the database every time what batch to save it under.
2) Save the info on the global object. I really don't like this either. I don't like global state in my code.
What is a good pattern for doing something like this? Does my question make any sense?
Thanks in advance
You can make a simple REST call to set the active batch and call it inside the batch when is started up and ready to accept requests. For example:
app.put('/active-batch', function(req, res, next){
// Make sure req.body is defined
app.set('active-batch', req.body);
res.end();
});
Then everywhere in the code you can use:
app.get('active-batch');
The app.set let you save data globally accessible in your app and app.get let you read previously stored data.
I'm researching BreezeJS for a big upcoming project.
Our goal is a offline first web app.
But here is what I can't fully understand (and would take to much time to test) - Does BreezeJS allow for the backend to be a REST API (built with NodeJS and Express)?
We need this because we don't want to simply sync to a remote DB (in our case Mongo), but use a remote REST API so that we can embed some business logic. Things like workflow triggering on a POST to a particular entity.
Is this possible with BreezeJS? If not what would be a good option?
Thanks in advance
It is certainly possible, simply take the breeze-mongo server implementation and strip out the mongo specific code. This should be fairly straight forward, express and mongo are pretty well separated in the code.
That said, you would lose or have to rewrite much of the server side code that converts an OData query string into a mongo query, but if you are going pure 'REST' you probably don't want that anyway.
You would have to do something similar on the save/POST side, but this is presumably something you are already familiar with.
I just started using node, backbone and mongoose not so long ago to create my first web app.
At the very beginning, I followed tutorials, and used backbone client side to define models. Those models mirror my mongoose schemas server side.
When I run schema.save() on one of my models, my data is automatically sent back to the client, with an _id.
But now that my app is almost finished, I realize that I don´t really need to save anything, as the only thing I do is query an api, and the data doesn´t have to be reused.
So my question is, what is the best way to keep the same mechanisms, but without saving anything?
The end reason is that I plan to run the app on an ec2 instance, and knowing that I don´t need to save anything, I think it is more beneficial to reduce the IO usage by not having any database.
Thanks, and sorry if the question seems dumb.
We have a working website using ServiceStack as the back end that amounts to a complex data-entry form.
My users have requested an "offline editor" for the forms. To use the offline program, the user will have to connect to the ServiceStack service, create empty instances of the forms, and then I will save the POCOs from the service to disk using ServiceStack's JSON serializer. From there the user can log off the service and edit the POCOs. When they're done, they reconnect to the service, and post/put the edited POCO object.
This all works great. My question involves validation. The validation logic is built into my Service.Interface library, which isn't available offline. The winforms program references only the POCO library and the ServiceStack "common" libraries, which do not look like they include the ServiceStack.Validation namespace.
Is there a way I can rearrange my project so that both the service and the Winforms client can run Validation against the POCOs, so that they can have data validation while offline?
UPDATE:
getting closer, I think - I moved all of the Validation classes into their own project. From my Winforms project, I can now manually set up a validator for a POCO class like this:
ServiceStack.FluentValidation.IValidator<SomePOCO> IValidator;
IValidator = new Tonto.Svc.Validation.SomePOCOValidator();
ServiceStack.FluentValidation.Results.ValidationResult vr =
IValidator.Validate(_rpt);
I can see the validator constructor being set up and the rules being initialized, but the .Validate method doesn't seem to do anything. (object comes back as valid, and breakpoints into custom validator code never get there).
UPDATE #2
I discovered my validator code wasn't running from Winforms because my validators all specify a servicestack ApplyTo Put/Post only (see sample code below). When I remove the entire Ruleset clause, though, then validation happens in my service on GETs - something I never want.
Can anyone think of a way to configure the validator rules to run for POST/PUT only when called from ServiceStack, but to also always run when NOT in servicestack? So close!
public class SomePOCOValidator : AbstractValidator<SomePOCO>
{
public SomePOCO()
{
RuleSet(ApplyTo.Put | ApplyTo.Post, () =>
{
(rules)
});
}
}
If your validation is doing anything interesting, then it probably HAS to be done "online".
Maybe just allow your client to save the POCOs locally until they go back online, at which point you send them up to your server. Any transactions that are okay, get processed normally, and any that fail, get returned for the user to edit (so your client will need some smarts to have a working set of POCOs for editing)...
If you don't want ANY extra stuff on the client, just have the transactions that fail to validate get stuffed into a "needs_corrections" table on the server, and then code up a supervisor-sort of screen to manage that table.
The validation framework that ServiceStack uses is named FluentValidation. There is no WinForms support in it. Jeremy Skinner the creator of FluentValidation answerd a question about this back in 2010 on his forum here.
Personally I don't use FV with WinForms - the vast majority of my projects are web-based with the occasional WPF project.
However, if I was going to do this then I probably wouldn't validate the controls directly, but instead use a ViewModel which is bound to the controls. I'd use a fairly strict convention where the names of the controls would match the names of the properties that they're bound to. Then, after validation completes I'd walk the control hierarchy to find the control with the name that matches the property that failed validation (I'm not sure how you'd do this in WinForms, but in WPF I'd use LogicalTreeHelper.FindLogicalNode) and then use the ErrorProvider to set the appropriate error.
Jeremy
I was able to work out a solution that allowed me to use ServiceStack validation libraries on both a ServiceStack client and an offline client. Here are the details.
Move all AbstractValidators to their own project: Proj.Svc.Validation.
get rid of all RuleSets in your AbstractValidators.
Reference Proj.Svc.Validation from Proj.Svc.Interface and Proj.OfflineWinformsClient projects.
Turn OFF the ValidationFeature() plugin in your service. All validation will have to be done manually. This means no iOC injected validators in your service classes.
When it's time to validate, either from your service or the offline client, manually declare the validator and use it like this.
IValidator validator = new
Tonto.Svc.Validation.SomePOCOValidator();
ServiceStack.FluentValidation.Results.ValidationResult vr =
validator.Validate(poco);
if (!vr.IsValid)
(throw exception or notify user somehow);