Develop Reference

Is it possible to use method attributes to change how this method is run?

I have this method:
public object LongRunningTask()
{
return SomethingThatTakesTooLong();
}
I wrote the following code so I can transform a normal method in an async one and still get the Exception:
public async Task<object> LongRunningTaskAsync()
{
Exception ex = null;
object ret = await Task.Run(() =>
{
object r = null;
try
{
//The actual body of the method
r = SomethingThatTakesTooLong();
}
catch (Exception e)
{
ex = e;
}
return r;
});
if (ex == null)
return ret;
else
throw ex;
}
When I need to do this in several methods, I have to copy all this code and change only the middle.
Is there a way to do something like this?
[SomeAttributeThatDoesThatMagically]
public async Task<object> LongRunningTaskAsync()
{
return SomethingThatTakesTooLong();
}

Attributes are generally metadata though it is possible to define attributes that can be executed (such as security behaviours in WCF) however, something has to be looking for it first. Your attributes won't just magically run.
I suspect you might have to use a dynamic proxy.
Take a look at how WCF does things for ideas.

How to avoid code duplication inside two methods?

I have two identical methods, but one of them have return statement inside try catch
public void A(Guid agentId)
{
var agent = _agentsProvider.GetAgentById(agentId);
var updateCompletionSource = C(agentId);
try
{
var cacheEntry = UpdateAgentMetadataCacheEntry(agent, true, false);
updateCompletionSource.SetResult(cacheEntry);
}
catch (Exception e)
{
updateCompletionSource.SetException(e);
}
}
private Entry B(IAgent agent)
{
var updateCompletionSource = C(agent.Id);
try
{
var cacheEntry = UpdateAgentMetadataCacheEntry(agent, false, false);
updateCompletionSource.SetResult(cacheEntry);
return cacheEntry;
}
catch (Exception e)
{
updateCompletionSource.SetException(e);
return GetPreviousCacheEntry();
}
}
How to collect identical part and create new method with this part?

Unless GetPreviousCacheEntry could have problematic side-effects, it seems to me that you don't need method A at all.
Just call method B and ignore the return value if you're not interested in it.
As noted in comments, the methods aren't identical other than the return statements though - because they use a different second argument for UpdateAgentMetadataCacheEntry, and they have different parameters too (one has a Guid and one has an Agent). You could refactor this into:
private Entry B(IAgent agent, bool foo)
{
var updateCompletionSource = C(agent.Id);
try
{
var cacheEntry = UpdateAgentMetadataCacheEntry(agent, foo, false);
updateCompletionSource.SetResult(cacheEntry);
return cacheEntry;
}
catch (Exception e)
{
updateCompletionSource.SetException(e);
return GetPreviousCacheEntry();
}
}
... with a meaningful name for foo, obviously. I'll assume the difference in parameter type isn't a problem in reality.

Like Jon said, you don't need method A. Just add another parameter for boolean value.
public void A(Guid agentId)
{
var agent = _agentsProvider.GetAgentById(agentId);
AnotherA(agent, true);
}
private Entry B(IAgent agent)
{
return AnotherA(agent, false);
}
private Entry AnotherA(IAgent agent, bool a)
{
try
{
var cacheEntry = UpdateAgentMetadataCacheEntry(agent, a, false);
updateCompletionSource.SetResult(cacheEntry);
return cacheEntry;
}
catch (Exception e)
{
updateCompletionSource.SetException(e);
return GetPreviousCacheEntry();
}
}

Managing connection with non-buffered queries in Dapper

I have recently started using Dapper, everything seems nice and easy but there is one thing that keeps confusing me: Connection Management.
As per the documentation:
Dapper does not manage your connection's lifecycle, it assumes the
connection it gets is open AND has no existing datareaders enumerating
(unless MARS is enabled)
In light of this I started doing this inside the implementation of my repository methods:
using (var db = new SqliteConnection(connectionString)) {
// call Dapper methods here
}
Then I came across a table with a large number of records, so I though of returning an IEnumerable<T> by passing buffered: false to the Query<> method, and when I started enumerating the enumerable in the front end, boom an exception saying the connection was closed and disposed which is expected since I am wrapping my calls with the preceding using block.
Question: Best way to solve this ?
Side question: Is the way I am managing the connection the preferred way to go about it ?

I'd offer this repository pattern:
public class Repository
{
private readonly string _connectionString;
public Repository(string connectionString)
{
_connectionString = connectionString;
}
protected T GetConnection<T>(Func<IDbConnection, T> getData)
{
using (var connection = new SqlConnection(_connectionString))
{
connection.Open();
return getData(connection);
}
}
protected TResult GetConnection<TRead, TResult>(Func<IDbConnection, TRead> getData, Func<TRead, TResult> process)
{
using (var connection = new SqlConnection(_connectionString))
{
connection.Open();
var data = getData(connection);
return process(data);
}
}
}
For buffered queries you want to use first overload of GetConnection method, for non-buffered you use second, specifing callback for processing data:
public class MyRepository : Repository
{
public MyRepository(string connectionString) : base(connectionString)
{
}
public IEnumerable<MyMapObject> GetData()
{
return GetConnection(c => c.Query<MyMapObject>(query));
}
public IEnumerable<ResultObject> GetLotsOfData(Func<IEnumerable<MyMapObject>, IEnumerable<ResultObject>> process)
{
return GetConnection(c => c.Query<MyMapObject>(query, buffered: false), process);
}
}
Very basic usage:
static void Main(string[] args)
{
var repository = new MyRepository(connectionString);
var data = repository.GetLotsOfData(ProcessData);
}
public static IEnumerable<ResultObject> ProcessData(IEnumerable<MyMapObject> data)
{
foreach (var record in data)
{
var result = new ResultObject();
//do some work...
yield return result;
}
}
But keep in mind - connection may be opened for too long time in this case...

#Sergio, AWESOME! Thanks for such a great pattern. I modified it slightly to be async so that I can use it with Dapper's async methods. Makes my entire request chain async, from the controllers all the way back to the DB! Gorgeous!
public abstract class BaseRepository
{
private readonly string _ConnectionString;
protected BaseRepository(string connectionString)
{
_ConnectionString = connectionString;
}
// use for buffered queries
protected async Task<T> WithConnection<T>(Func<IDbConnection, Task<T>> getData)
{
try
{
using (var connection = new SqlConnection(_ConnectionString))
{
await connection.OpenAsync();
return await getData(connection);
}
}
catch (TimeoutException ex)
{
throw new Exception(String.Format("{0}.WithConnection() experienced a SQL timeout", GetType().FullName), ex);
}
catch (SqlException ex)
{
throw new Exception(String.Format("{0}.WithConnection() experienced a SQL exception (not a timeout)", GetType().FullName), ex);
}
}
// use for non-buffeed queries
protected async Task<TResult> WithConnection<TRead, TResult>(Func<IDbConnection, Task<TRead>> getData, Func<TRead, Task<TResult>> process)
{
try
{
using (var connection = new SqlConnection(_ConnectionString))
{
await connection.OpenAsync();
var data = await getData(connection);
return await process(data);
}
}
catch (TimeoutException ex)
{
throw new Exception(String.Format("{0}.WithConnection() experienced a SQL timeout", GetType().FullName), ex);
}
catch (SqlException ex)
{
throw new Exception(String.Format("{0}.WithConnection() experienced a SQL exception (not a timeout)", GetType().FullName), ex);
}
}
}
Use with Dapper like this:
public class PersonRepository : BaseRepository
{
public PersonRepository(string connectionString): base (connectionString) { }
// Assumes you have a Person table in your DB that
// aligns with a Person POCO model.
//
// Assumes you have an existing SQL sproc in your DB
// with #Id UNIQUEIDENTIFIER as a parameter. The sproc
// returns rows from the Person table.
public async Task<Person> GetPersonById(Guid Id)
{
return await WithConnection(async c =>
{
var p = new DynamicParameters();
p.Add("Id", Id, DbType.Guid);
var people = await c.QueryAsync<Person>(sql: "sp_Person_GetById", param: p, commandType: CommandType.StoredProcedure);
return people.FirstOrDefault();
});
}
}

Repeating code pattern

I have certain code pattern (which log performance and other variable for each function) which need to be present in every function, and I do not want to repeat the code over and over again. Here is what the code looks like:
public OutClass FUNC-X
{
if (IsDebugEnabled)
{
Logger("Start DataLibrary: FUNC-X");
}
try
{
CheckInitSucceeded();
GetAuthenticationTokens();
var dm = new Manager();
/**
* THIS SINGLE LINE IS THE VARIABLE PART
**/
var output = dm.FUNC-X(...);
if (IsDebugEnabled)
{
var data = Serialize(output);
Logger(output);
}
return output;
}
catch (WebFaultException)
{
throw;
}
catch (OtherException ex)
{
if (Logger.IsErrorEnabled)
{
Logger.LogError("Exception in FUNC-X", ex);
}
throw new OtherException("Some Message");
}
catch (Exception ex)
{
if (IsErrorEnabled)
{
Logger("Exception in FUNC-X", ex);
}
throw new Exception("Generic Exception");
}
finally
{
if (IsDebugEnabled)
{
Logger("End FUNC-X");
}
}
}
Essentially, I just need to replace FUNC-X with FUNC-Y or FUNC-Z everywhere this name occurs, is there some kind of design pattern which I can follow?
I'm sorry if the question is vague, I shall be happy to provide any details you ask.

Yeah, there are many ways to provide nice log code.
Use aspect-oriented programming. There is PostSharp and Spring.NET. You can use log4net library. All of them after config supports writing of method name.
You can use T4 and generate code for all funcs before compile.
You can write one global logging method that accepts Func<object> method and string methodName. Suppose you named your logging method LogAndExecute. Then to call you must write smth like:
LogAndExecute("func-x", () => dm.Func-X(/*your args*/))
If you got problems with different different return types of your funcs, use generics

You can use simple delegate to accept the variable part as parameter (if no. of parameters to function call in single line are same)
public void Function_X(Func<object,..> func)
{
if (IsDebugEnabled)
{
Logger("Start DataLibrary: FUNC-X");
}
try
{
CheckInitSucceeded();
GetAuthenticationTokens();
var dm = new Manager();
/**
* THIS SINGLE LINE IS THE VARIABLE PART
**/
// var output = dm.FUNC-X(...);
var output = func(...);
if (IsDebugEnabled)
{
var data = Serialize(output);
Logger(output);
}
return output;
}
catch (WebFaultException)
{
throw;
}
catch (OtherException ex)
{
if (Logger.IsErrorEnabled)
{
Logger.LogError("Exception in FUNC-X", ex);
}
throw new OtherException("Some Message");
}
catch (Exception ex)
{
if (IsErrorEnabled)
{
Logger("Exception in FUNC-X", ex);
}
throw new Exception("Generic Exception");
}
finally
{
if (IsDebugEnabled)
{
Logger("End FUNC-X");
}
}
}

You can create a common function that accepts a Func delegate:
static public TOutClass CommonFunc<TOutClass>(Func<Manager, TOutClass> func)
{
if (IsDebugEnabled)
{
Logger("Start DataLibrary: FUNC-X");
}
try
{
CheckInitSucceeded();
GetAuthenticationTokens();
var dm = new Manager();
TOutClass output = func(dm);
if (IsDebugEnabled)
{
var data = Serialize(output);
Logger(output);
}
return output;
}
catch
[...]
}
Your would write your functions as:
public OutClass FUNC-X(...)
{
return CommonFunc(dm=>dm.FUNC-X(...));
}

how do I update entity object passed from another class?

This is my query where I'm returning an IEnumerable<CreditCardTransaction> to iterate through.
public partial class CreditCardTransaction
{
public static IEnumerable<CreditCardTransaction> GetUnprocessedTransactions()
{
try
{
using (var context = new SuburbanEntities())
{
return from trans in context.CreditCardTransactions
where trans.IsPublished == false
select trans;
}
}
catch (Exception ex)
{
Logging.Log("An error occurred.", "GetUnprocessedTransactions",Apps.ServicesConfig, ex);
return null;
}
}
}
This is where I am modifying those transactions once I have processed them:
public void ProcessFile()
{
try
{
_client = new TruckServiceClient();
_globalSetting = new GlobalSetting();
var unprocesstransactions = CreditCardTransaction.GetUnprocessedTransactions();
foreach (var creditCardTransaction in unprocesstransactions)
{
creditCardTransaction.IsPublished = ProcessTransaction(creditCardTransaction);
}
}
catch (Exception ex)
{
Logging.Log("An error occurred.", "ProcessCreditCardTransactions.ProcessFile", Apps.RemoteServices, ex);
}
}
I am modifying the transactions here:
creditCardTransaction.IsPublished = ProcessTransaction(creditCardTransaction);
But once I have saved them, can I update the entity directly or do I need to create another method where I pass this information back in?

The problem you don't have access to the context. Here you have some examples how to do:
https://github.com/geersch/EntityFrameworkObjectContext
If you're developing Asp.Net app, you'll have some drawbacks illustreted in this article:
http://dotnetslackers.com/articles/ado_net/Managing-Entity-Framework-ObjectContext-lifespan-and-scope-in-n-layered-ASP-NET-applications.aspx#managing-objectcontext-instantiation