We Keep Coding

How do I await this callback method to perform logic on the return data?

So I'm working in Silverlight right now unfortunately for the first time. I'm decently familiar with callbacks, but I'm not entirely sure how to convert this method to be synchronous to perform logic on the order data.
I've been frequently told that making this synchronous was ill-advised, but my goal is to check if certain fields have been modified in the XAML UI and are different from what exists in the database. Then prompt for a reason for the change. If there is a better way to go about this, I'd love to know.
I'm in Silverlight 5 with .Net Framework 4.0 in VS 2013
Thank you! Here's the async order provider:
public void GetOrder(string ordNumber, Action<Func<OrderLoadResults>> callback)
{
String exStr = String.Format("{0}.{1}() --> received an empty value for",
this.GetType().Name,
MethodBase.GetCurrentMethod().Name);
if (ordNumber == null)
{
throw new ArgumentNullException("ordNumber", exStr);
}
if (callback == null)
{
throw new ArgumentNullException("callback", exStr);
}
IOrderServiceAsync channel = CreateChannel();
AsyncCallback asyncCallback = ar => GetOrderCallback(callback, ar);
channel.BeginGetOrderByOrdNumber(ordNumber, asyncCallback.ThreadSafe(), channel);
}
And here's what I'm doing with it:
public List<ATMModifiedFieldModel> CheckForATMModifiedFields()
{
if (!_order.Stops.Items.Any(x => x.ModelState == ModelState.Modified))
{
return null;
}
List<StopModel> oldStop = new List<StopModel>();
Provider.OrderProvider orderProvider = new Provider.OrderProvider();
//Looking to convert this method to sync to pull the order info out to compare against
//orderProvider.GetOrder(_order.Item.OrdHdrNumber.ToString(),getResult => OnGetOrderComplete(getResult));
List<ATMModifiedFieldModel> modifiedFields = new List<ATMModifiedFieldModel>();
foreach (StopModel stop in _order.Stops.Items)
{
if (stop.ModelState == ModelState.Modified)
{
foreach (string ATMFieldName in Enum.GetNames(typeof(ATMFields)))
{
string beforeValue = "before value"; //Should check the value in the database
string afterValue = stop.GetType().GetProperty(ATMFieldName).GetValue(stop, null).ToString();
if (beforeValue != afterValue)
{
modifiedFields.Add(new ATMModifiedFieldModel(ATMFieldName, beforeValue, afterValue, stop.StpNumber, "Stop"));
}
}
}
}
return modifiedFields;
}

How to create an EventBookmark when querying the event log

I have an EventLogReader object, and a query in the event log that looks like this:
string query = "*[System[(Level=2) and TimeCreated[#SystemTime>='%LastRun%']]]")
The code basically uses the reader to query for all the events that match the search query since the last time the reader was run.
I would rather use the EventBookmark for this purpose. That's what it is for, after all! But I am having trouble finding any working code.
My existing code run, in part, like this:
// This line replaces the %LastRun% code with the date
var myQuery = myEventLogQuery.Query.Replace("%LastRun%", myEventLogQuery.LastRun.ToString("o"));
var query = new EventLogQuery(myEventLogQuery.Log, myEventLogQuery.PathType, myQuery);
// Now set the LastRun date. I want to avoid this...
myEventLogQuery.LastRun = DateTime.UtcNow;
// ... by making this next line smarter.
var reader = new EventLogReader(query);
// var reader = new EventLogReader(query, ??? new EventBookmark());
EventRecord eventRecord;
while ((eventRecord = reader.ReadEvent()) != null)
{
EventRecords.Add(new EventRecordItem(eventRecord));
}
I should be able to use the EventBookmark property of the first (or last) EventRecord to restrict the next query, but I want to set the EventBookmark initially to be basically the highest record in the log.
When the application runs initially, I don't need it to tell me about all the event log entries from the past, I only care about ones that occur after the application starts.

OK, I went ahead and experimented a lot, and managed to work out a good system for this. Since this topic isn't really covered, I'm posting my answer here. I hope it's useful!
First challenge is creating the initial EventBookmark. You can't just instantiate one. You need to derive one from an existing EventRecord. To do this, I query for the last item in the log and base my EventBookmark on that.
using System.Diagnostics.Eventing.Reader;
public class MyEventLogQuery
{
public string Log { get; set; }
public PathType PathType { get; set; }
public string Query { get; set; }
public EventBookmark Bookmark { get; set; }
public MyEventLogQuery(string log = "Application", PathType pathType = PathType.LogName, string query = "*[System[(Level=2)]]")
{
Log = log;
PathType = pathType;
Query = query;
Bookmark = GetBookmark(log, pathType); // Query is not important here
}
// This method returns the bookmark of the most recent event
// log EventRecord or null if the log is empty
private static EventBookmark GetBookmark(string log, PathType pathType)
{
var elq = new EventLogQuery(log, pathType) {ReverseDirection = true};
var reader = new EventLogReader(elq);
var record = reader.ReadEvent();
if (record != null)
return record.Bookmark;
return null;
}
}
Next step is to use the bookmark (or lack of bookmark) when subsequently querying the Event Log.
using System.Diagnostics.Eventing.Reader;
public class EventLogTracker()
{
public List<MyEventLogQuery> Queries { get; set; }
// ... snipped some stuff
public int Run()
{
var count = 0;
foreach (var myQuery in Queries)
{
var query = new EventLogQuery(myQuery.Log, myQuery.PathType, myQuery.Query);
// This is the important bit. Must take account that the
// log may have been empty, so bookmark could be null
var reader = myQuery.Bookmark != null ? new EventLogReader(query, myQuery.Bookmark) : new EventLogReader(query);
EventRecord eventRecord;
while ((eventRecord = reader.ReadEvent()) != null)
{
// Do stuff
// ...
// Then update the bookmark
myQuery.Bookmark = eventRecord.Bookmark;
count++;
}
}
return count;
}
And voila, you have code that uses the EventBookmark to only give you events that have occurred since the application was started.

How to determine if a complete batch of Async requests have failed?

A 3rd party has supplied an interface which allows me to search their database for customers and retrieve their details. Eg. Pictures, date of birth etc.
I imported their WSDL into Visual Studio and am using the Async methods to retrieve the customer details.
MyClient Client = new MyClient();
Client.FindCustomersCompleted += FindCustomersCompleted;
Client.GetCustomerDetailsCompleted += GetCustomerDetailsCompleted;
Client.FindCustomersAsync("Jones");
Below are the two events which deal with the responses.
void FindCustomersCompleted(object sender, FindCustomersCompletedEventArgs e)
{
foreach(var Cust in e.Customers)
{
Client.GetCustomerDetailsAsync(Cust.ID);
}
}
void GetCustomerDetailsCompleted(object sender, GetCustomerDetailsCompletedEventArgs e)
{
// Add the customer details to the result box on the Window.
}
So lets assume that my initial search for "Jones" returns no results or causes an error. Its fairly straight forward to tell the user that there was an error or no results found as I will only receive a single response.
However if i say get 50 results for "Jones" then i make 50 GetCustomerDetailsAsync calls and get 50 responses.
Lets say that something goes wrong on the server side and i don't get any valid responses. Each GetCustomerDetailsCompleted event will receive an error/timeout and i can determine that that individual response has failed.
What is the best way to determine that All of my responses have failed and i need to inform the user that there has been a failure?
Alternatively what if 1 out of 50 succeeds?
Should i keep track of my requests and flag them as successful as i receive the response?

Should i keep track of my requests and flag them as successful as i
receive the response?
This is also how I manage multiple requests. Flag if the returned result is without fault and track the flags, evaluating after each return if you already processed all returns. I do not of another way.

I would start by converting Event-based Asynchronous Pattern model to Task based. This would allow to use built in await/async keywords resulting in much easier to use code.
Here is a simple implementation: https://stackoverflow.com/a/15316668/3070052
In your case I would not update UI on each event but gathered all the information in a single variable and only displayed only when I get all the results.
Here is a code to get you going:
public class CustomerDetails
{
public int Id {get; set;}
public string Name {get; set;}
}
public class FindCustomersResult
{
public FindCustomersResult()
{
CustomerDetails = new List<CustomerDetails>();
}
public List<CustomerDetails> CustomerDetails {get; set;}
}
public class ApiWrapper
{
public Task<FindCustomersResult> FindCustomers(string customerName)
{
var tcs = new TaskCompletionSource<FindCustomersResult>();
var client = new MyClient();
client.FindCustomersCompleted += (object sender, FindCustomersCompletedEventArgs e) =>
{
var result = new FindCustomersResult();
foreach(var customer in e.Customers)
{
var customerDetails = await GetCustomerDetails(customer.ID);
result.CustomerDetails.Add(customerDetails);
}
tcs.SetResult(result);
}
client.FindCustomersAsync(customerName);
return tcs.Task;
}
public Task<CustomerDetails> GetCustomerDetails(int customerId)
{
var tcs = new TaskCompletionSource<CustomerDetails>();
var client = new MyClient();
client.GetCustomerDetailsCompleted += (object sender, GetCustomerDetailsCompletedEventArgs e) =>
{
var result = new CustomerDetails();
result.Name = e.Name;
tcs.SetResult(result);
}
client.GetCustomerDetailsAsync(customerId);
return tcs.Task;
}
}
Then you call this by:
var api = new ApiWrapper();
var findCustomersResult = await api.FindCustomers("Jones");
This would fail if any request fails.
PS. I wrote this example in notepad, so bear with me if it does not compiles or contains syntax errors. :)

Encapsulation with multiple classes C#

So my question is over basic encapsulation. I know I am setting up my getters and setters right (I actually have a question later about this) but I have multiple classes. I have another class and I understand that I am making pieces of my code view-able to my outside class by making certain things public. So I think I set up my first code file right. (Some background, I have a class that is connecting to a database and then another that is encapsulating all the data. The first code section posted is the encapsulating part, I then post my three methods I was messing up on.)
I feel okay with the getting and setting, I feel a little unsure of my constructor. I feel like I put my variables in the parameter list so that I put values in them from my outside class? Right? or should I be putting public forms of my private variables in my other code file and then passing those into my constructor in that same file?
/ this my first code file
using System;
public class clsEncapsulate
{
private int mID;
private string mName;
private string mClassification;
private DateTime mConvocationDate;
private string mLocation;
public int ID
{
get
{
return mID;
}
set
{
mID = value;
}
}
public string Name
{
get
{
return mName;
}
set
{
mName = value;
}
}
public string Classification
{
get
{
return mName;
}
set
{
mName = value;
}
}
private DateTime ConvocationDate
{
get
{
return mConvocationDate;
}
set
{
mConvocationDate = value;
}
}
private string Location
{
get
{
return mLocation;
}
set
{
mLocation = value;
}
}
public clsEncapsulate(int id, string name, string classification, DateTime convocationDate, string location)
{
bool running = false;
while(running == false)
{
ID = mID;
Name = mName;
Classification = mClassification;
ConvocationDate = mConvocationDate;
Location = mLocation;
running = true;
}
}
}
In my second code file I am just going to put the methods that I am having trouble with.
private void refreshbox()
{
string formattedConvocation;
string formattedDateTime;
string formattedConvocationName;
lstConvocations.Items.Clear();
foreach (clsEncapsulate currentConvocation in mConvocationAL)
{
formattedConvocationName = currentConvocation.Name;
formattedConvocationName = truncateString(formattedConvocationName, 30);
formattedConvocation = formattedConvocationName.PadRight(33);
formattedConvocation += currentConvocation.Classification.PadRight(17);
formattedDateTime = currentConvocation.ConvocationDate.ToShortDateString().PadRight(10)
+ currentConvocation.ConvocationDate.ToShortTimeString().PadLeft(8);
formattedConvocation += formattedDateTime;
lstConvocations.Items.Add(formattedConvocation);
}
}
Alright, so in order for my second code file to manipulate the variables in the first code file, I need to expose them to this method. I didn't know if I should be putting my public variables in the constructor, or if I should be declaring them somewhere in my first code file. I was very unsure of how to expose these variables to this method. I've fiddle around with it but my book doesn't address this situation exactly and I was having trouble figuring it out.
If someone does answer this question please break down why you're going to put what you're going to put! I want to understand why, say, I put my public variables in one place, and not another. Or why I declare an object of my encapsulate class in one place and not another. I was trying to declare an encapsulate object in my method so it would give this method access to the variables, but it wasn't working! Please tell me what I was doing wrong or if you want me to post more of my code.
Below are the two other methods I was messing up on.
/ second method from my second code file I was messing up on:
private void displayProperties(int index)
{
if (index == -1)
{
return;
}
clsEncapsulate selectedValue = (clsEncapsulate)mConvocationAL[index];
txtConvocationName.Text = selectedValue.Name;
txtConvocationClassification.Text = selectedValue.Classification;
txtConvocationDate.Text = selectedValue.ConvocationDate.ToShortDateString();
txtConvocationTime.Text = selectedValue.ConvocationDate.ToShortTimeString();
txtConvocationLocation.Text = selectedValue.Location;
txtID.Text = selectedValue.ID.ToString();
}
/ last method I was messing up on:
private void readConvocations(string filterConstraint, string sortField, string sortOrder)
{
OleDbConnection connection = null;
OleDbDataReader reader = null;
try
{
connection = new OleDbConnection();
connection.ConnectionString = mConnectionString;
connection.Open();
string statement = "SELECT ID, Name, Classification, Location, Date FROM Convocations ";
if(filterConstraint != "")
{
statement += "WHERE Name LIKE " + toSQL(filterConstraint, true) + " ";
}
string statement2 = statement;
statement = string.Concat(new string[]
{
statement2, "ORDER BY ", sortField, " ", sortOrder
});
OleDbCommand oleDbCommand = new OleDbCommand(statement, connection);
reader = oleDbCommand.ExecuteReader();
mConvocationAL.Clear();
while(reader.Read())
{
clsEncapsulteconvocation = new clsEncapsulate();
convocation.ID = (int)reader["ID"];
convocation.Name = (string)reader["Name"];
convocation.Classification = (string)reader["Classification"];
convocation.Location = (string)reader["Location"];
convocation.ConvocationDate = (DateTime)reader["Date"];
mConvocationAL.Add(convocation);
}
}
finally
{
if (reader != null)
{
reader.Close();
}
if (connection != null)
{
connection.Close();
}
}
}
Tell me if you need me to elaborate more to help you understand my situation. I am new at learning vocabulary and want to understand this! Thank you for helping. :)

The code you provided is one public object and a bunch of private methods so its difficult to get the overall picture of how your code it working together but there are a few principles you can apply to make your code better structured, now and in the future.
Have a read about SOLID (http://en.wikipedia.org/wiki/SOLID_(object-oriented_design)). The S and the D would apply well to your example.
Also you mentioned construtors and private properties. Try looking into Imutable types. That means once the object is created you cannot change it. For your clsEncapsulate class that would mean making your fields read only and remove the public setters.
Good luck.

Adding new record to db through silverlight Model

I have a silverlight mvvm with ria project. I have a UI in which admin users can enter info to create new work orders. However, I am having trouble calling the db and adding a new record to the table. I have no code-behind for the UI, the controls are tied to the model through Commands and Command Parameters. So when a user clicks, 'Add new job' it comes here,
public class EditJobViewModel : ViewModelBase
{
private Job _job;
public Job CurrentJob
{
get { return _job; }
set
{
_job = value;
OnPropertyChanged("CurrentJob");
}
}
public ICommand NewJob
{
get
{
return new DelegateCommand(BeginNewJob, (o) => true);
}
}
public void BeginNewJob(object o)
{
_job = new Job();
//_job.JobNumber = _job.JobID.ToString();
_job.AssignedTo = App.userID;
_job.AddedBy = App.userID;
_job.FileTypeJob = "PDF";
_job.AddedTS = DateTime.Now;
_job.OpenDate = DateTime.Now;
BeginSave(o);
}
}
Where Im having trouble is creating a new record in the 'Job' table. On my breakpoint it returns all the columns it needs to, just not a new 'JobID' which is my primary key. This is how I was supposedly trying to create a new record.
public void BeginSave(object o)
{
if (!IsDesignTime)
{
try
{
if (CurrentJob.EntityState == EntityState.New)
{
CurrentJob.AddedBy = App.userID;
CurrentJob.AddedTS = DateTime.Now;
}
CurrentJob.UpdatedBy = App.userID;
CurrentJob.UpdatedTS = DateTime.Now;
// This is here because of a bug in infragistics grid/Entity Framework
foreach (JobFileType ft in CurrentJob.JobFileTypes)
{
if (ft.EntityState != EntityState.Unmodified)
(ft as IEditableObject).EndEdit();
}
foreach (JobTag tag in CurrentJob.JobTags)
{
if (tag.EntityState != EntityState.Unmodified)
(tag as IEditableObject).EndEdit();
}
//(CurrentJob as IEditableObject).EndEdit();
SubmitOperation s = _context.SubmitChanges();
if (s.HasError)
{ }
}
catch (Exception ex)
{ }
}
}
Except that it never hits the EntityState.new. That's just the way I thought to try it. Im thinking there a way to do it from the 'BeginNewJob' command but unable to find a way to create a new JobID or record in general. The Database already has 10000 records and has multiple users creating jobs, so I need a way to get the last job created (getMaxID()??) and increment appropriately, creating a new job on the spot.

İf you use guid type for id column, you will not need to find next id and this approach will decouple new objects from previous objects.